Speaking of Colors and Odors (Converging Evidence in Language and Communication Research)

Speaking of Colors and Odors

Converging Evidence in Language and Communication Research (CELCR) Over the past decades, linguists have taken a broader view of language and are borrowing methods and findings from other disciplines such as cognition and computer sciences, neurology, biology, sociology, psychology, and anthropology. This development has enriched our knowledge of language and communication, but at the same time it has made it difficult for researchers in a particular field of language studies to be aware of how their findings might relate to those in other (sub-)disciplines. CELCR seeks to address this problem by taking a cross-disciplinary approach to the study of language and communication. The books in the series focus on a specific linguistic topic and offer studies pertaining to this topic from different disciplinary angles, thus taking converging evidence in language and communication research as its basic methodology.

Editor Marjolijn H. Verspoor University of Groningen

Wilbert Spooren

Vrije Universiteit Amsterdam

Advisory Board Walter Daelemans

Leo Noordman

Cliff Goddard

Martin Pütz

University of Antwerp University of New England

Tilburg University University of Koblenz-Landau

Roeland van Hout

Radboud University Nijmegen

Volume 8 Speaking of Colors and Odors Edited by Martina Plümacher and Peter Holz

Speaking of Colors and Odors

Edited by

Martina Plümacher Technical University Berlin

Peter Holz University of Bremen

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Speaking of colors and odors / edited by Martina Plümacher and Peter Holz. p. cm. (Converging Evidence in Language and Communication Research, issn 1566-7774 ; v. 8) Includes bibliographical references and index. 1. Language and color. 2. Language and smell. I. Plümacher, Martina, 1958II. Holz, Peter. P120.C65S68 2007 418--dc22

2006048028

isbn 978 90 272 3895 5 (Hb; alk. paper)

© 2007 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

JB[v.20020404] Prn:3/11/2006; 14:57

F: CELCR8CO.tex / p.1 (49-117)

Table of contents

Speaking of colors and odors Martina Plümacher and Peter Holz

1

Color, smell, and language: The semiotic nature of perception and language Wolfgang Wildgen

19

How can language cope with color? Functional aspects of the nervous system Manfred Fahle

35

Color perception, color description and metaphor Martina Plümacher

61

Attractiveness and adornment: Reference to colors and smells in Papuan speech communities Volker Heeschen

85

Color terms between elegance and beauty: The verbalization of color with textiles and cosmetics Siegfried Wyler

113

Color names and dynamic imagery Andrea Graumann

129

From blue stockings to blue movies: Color metonymies in English Susanne Niemeier

141

Odor memory: The unique nature of a memory system Gesualdo M. Zucco

155

From psychophysics to semiophysics: Categories as acts of meaning. A case study from olfaction and audition, back to colors Danièle Dubois

167

Cognition, olfaction and linguistic creativity: Linguistic synesthesia as poetic device in cologne advertising Peter Holz

185

JB[v.20020404] Prn:3/11/2006; 14:57



F: CELCR8CO.tex / p.2 (117-134)

Table of contents

Understanding synesthetic expressions: Vision and olfaction with the physiological = psychological model Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

203

Olfactory and visual processing and verbalization: Cross-cultural and neurosemiotic dimensions Tatiana V. Chernigovskaya and Victor V. Arshavsky

227

Contributors

239

Index

241

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.1 (44-151)

Speaking of colors and odors Martina Plümacher and Peter Holz

.

Introduction to the topic

We are confronted with a huge manifold of diverse sensory impressions which we have to classify and interpret in regard to our former experiences and intentions in acting. We also have to coordinate our sensory impressions with other individuals in order to perform purposeful joint actions and meaningful interactions. Categorization is a cognitive activity that serves both to develop personal experience to an extent that makes quick orientation in new situations possible and to establish inter-individual understanding in processes of joint action. Language is involved in these processes of categorization. What exactly is its specific part? This is the central question of this book. Most of its contributions start from the assumptions that language is complementory to other forms of communication, such as joint attention, regular forms of joint action, gestures and other non-verbal semiotic forms, and that linguistic forms of categorization are based on perceptual discernment and non-linguistic forms of classifying things, events and properties. The articles of this book originate from an international and interdisciplinary conference dealing with the question “How can language cope with color and smell?”. ‘Color’ and ‘smell’ were chosen as subjects to compare the extensively studied field of color perception and color categorization with the less investigated fields of perception and categorization of odors. The neurobiological and neurophysiological conditions of color perception are known to a large extent, and cross-linguistic research on color categorization and lexical coding of color has developed since the 1950s. The sensory system of olfaction and categorization of odors have been of comparatively minor interest. – This can be asserted at least with regard to science. Within the large fields of industrial production of perfume, scents of cosmetics and flavor of food, the interest is immense, and a huge amount of money is spent on research, evaluation and advertising. – We supposed that comparing these differently investigated fields of sensory systems and linguistic representation of sensory impressions could broaden the view on the issue. It might help to clear controversies, especially as to the prevailing theses that colors

JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.2 (151-178)

Martina Plümacher and Peter Holz

can be considered a quasi-continuous spectrum of hues that is completely represented in languages, although in different ways, whereas the manifold of odors cannot be conceived in a systematic arrangement and is not linguistically represented. As the contributions of this book show, this opposition does not grasp the semiotic reality. In many contexts, the linguistic representation of colors refers to colored objects, to practices of color production or to specific experiences of particular colors, just as much as linguistic descriptions of odor refer to objects as sources of odor and to odor experiences. The idea of a systematically arranged spectrum of colors is a relatively late product in the history of ideas. It could also be established that, despite the lack of ‘basic terms’ of odors, there is linguistic practice of odor description. The fact that many odors are not consciously perceived might be caused both by specific characteristics of the neurophysiological olfactory system and by a less developed cultural need to reflect odors. This fact, however, does not prevent us from analyzing the different strategies to describe olfactory experiences and the forms of categorization that this linguistic representation reveals. Especially studies on linguistic representation of odors indicate the functional role of language within the spectrum of diverse cognitive activities. Language does not express each act of discerning nor does it represent each form of categorization. In some contexts, there are joint activities of discriminating odor experiences between subjects. Unless there is no need to communicate linguistically about a particular type of odor, specific lexemes do not have to be established. Language seems to be involved when non-linguistic activities become subject to reflection. This reflection, however, applies to particular aspects with regard to tasks or problems of activities or to special purposes of communication. It is an intrinsically perspectival activity. Those contributions to this book which deal with descriptions of odors and colors in advertising focus on a specific power of language, namely its power to compose complex images of multiple sensory experience in connection with ideals of what count as beautiful or desirable. Lexicalization or stability of word meaning and reference is another aspect in the comparison of descriptions of colors and odors. At first glance, the difference seems to be enormous. With respect to color, speakers have at their disposal lexicalized color words as well as common strategies to describe more particular hues of color. In contrast to this, many types of odors are without a specific name, and there is no common strategy of description. How should we, for instance, speak about well-known differences between odors of perspiration? If subjects try to articulate their experiences, they are aware of the lack of common linguistic expressions as well as of the individuality of perception and its evaluation. They often emphasize personal experience, express the vagueness of descriptions, and resort to metaphors. The problem is more difficult: as empirical studies on odor identification and description show, subjects identify the same odor differently in time, and their descriptions differ considerably. Differences between color and

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.3 (178-244)

Speaking of colors and odors

odor experiences and their descriptions diminish a little when one looks at descriptions of specific hues of color and color effects on subjects. With respect to the latter, metaphors are frequently used, descriptions diverge, and subjects also refer to their personal experience. Comparative analyses of speaking about colors and odors bring into question factors of language stabilization. What enables people to share linguistic expressions and underlying categories? The contributions of this book are approaches to the topic from different disciplinary angles. They stress diverse aspects of the relation between perception, cultural activity and linguistic representation. Neurobiologists point to biological individuality which indicates differences between individuals in the processing of sensory information, neuropsychologists confirm the individuality of odor experiences, whereas linguists emphasize the function of language to establish interpersonal forms of categorization. This leads to the question of how to relate individual perception, individual abilities to differentiate phenomena, individual forms of categorization and interpersonal forms of categorization which are mediated by language and joint activities. The contributions of this book are pioneering in the sense that they show important factors that should be addressed in further research in order to answer this question sufficiently.

. The contributions in the context of previous research In the past, studies on speaking about sensory impressions concentrated on color words. Forms of speaking about other sensory impressions have been less investigated (with respect to odors, exceptions are Harper et al. 1968; David et al. 1997; Dubois 1997a, 1997b, 2000; Gschwind 1998; Rouby et al. 2002; Holz 2005). One reason for this concentration is the order which has been already established in the area of color: the phenomenon ‘color’ is explained to a large extent, and colors are systematically ordered in schematic representations, such as the color sphere. Schematic representations do not show the real totality of hues, nor those properties of color which are characteristic of colored substances, such as shine, luster, texture or structures of color application. They serve the purpose of presenting an overview of the set of possible colors that, strictly speaking, is a concept of color order. In psycholinguistic research on linguistic representations of colors and language and thought relations, these schematic representations were used as an instrument of objectivization, especially with respect to inter-cultural methods. Colored chips seemed to be perfect reference objects to compare the scope of applicability of color words in cross-linguistic studies. A first step was made by Roger W. Brown, Eric H. Lenneberg and colleagues who methodologically based their studies on codability of colors on the Munsell array of colors (cf. Brown & Lenneberg 1954; Lenneberg & Roberts 1956). Brent Berlin and Paul Kay, whose



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.4 (244-311)

Martina Plümacher and Peter Holz

cross-linguistic study on ‘basic color terms’ (1969) became the most famous work within the field of research on color words, also made use of the colored chips of the Munsell Book of Color. These chips consist of 40 hues in 8 degrees of brightness and 9 shades of gray between black and white and represent the color schema by the American painter Albert Henry Munsell (1859–1918). Odors and other sensory impressions are not systematized into a similar order that provides a system of measuring units. Psycholinguistic research on identification and descriptions of odors operates with selected odors which are presented in the form of smelling cards. Beside the problem of producing the same odor under same contextual conditions with respect to temperature, air quality and air movement, there is the issue of how to define the boundaries of an olfactory stimulus (cf. Dubois this volume). From an ethnolinguistic perspective, however, the use of the Munsell color chips is called into question because these chips represent a model of colors which cannot be considered universal. Rolf Kuschel and Torben Monberg (1974: 241) called it “a Westernised system of classification” (cf. Lucy 1992: 155, 1997; Lehmann 1998: 189, 198). It emerged from European culture and is customary in many places but not in every culture, hence its culture-dependent aspects should be taken into account. The characterization of color represented by this model is such an aspect: colors are classified according to three properties – hue, brightness and saturation. Especially people in industrialized civilizations are used to this classification of color and corresponding representations, which are frequently consulted when a particular color among the wide variety of industrially produced colors has to be chosen. In industrialized civilizations, color is known as an artificially produced substance with which an object can be coated. Since color as a film or as a coat on objects has gained considerable practical importance, Lehmann (1998: 198f.) supposes that “we tend to consider ‘color’ as an abstract substance being painted on [something]”. When linguists used the Munsell color chips in their cross-linguistic studies, they took it for granted that color chips are known as samples of colors that can be found in nature. According to Kuschel and Monberg (1974: 241), the Polynesian Bellona speaking people do not have such an abstract concept of color but innumerable ‘color words’, most of them closely connected with specific objects. The particular cultural character of the color classification according to hue, brightness and saturation is plainly recognizable with comparison to alternative forms of color systematization. Erika Friedl (1979) pointed out a systematization in accordance with knowledge about dyeing. Within the Iranian language Luri, most modifiers of color terms refer to processes of dyeing. Informants who were asked to denominate a light red color chip often described it as ‘red, not well dyed’ (Friedl 1979: 55; cf. Lehmann 1998: 190f.). Robert E. MacLaury discovered in his fieldwork on color categorization in Mesoamerica and the Pacific Northwest that diverse languages prefer categories of color bright-

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.5 (311-389)

Speaking of colors and odors

ness to categories of hue (MacLaury 1992). In accordance with him, Ronald W. Casson (1997) states a shift from brightness to hue in the evolution of the English color vocabulary. This made clear that hue is not the only aspect of color categorization. John A. Lucy (1992: 146f.) opposes the prevailing belief that the diverse languages only encode reality differently. It is suggested “that reality is the primary phenomenon in the interaction and that language classifications are secondary, merely encoding ‘it’ in diverse ways. But since the reality is in fact conceived of here as the raw sense data of perception, and perception is ultimately taken as an unproblematically prior part of the cognition, this leads irresistibly to the conclusion that it is cognition that is being encoded by language and cognition that is, therefore, ultimately shaping language” (ibidem: 146). These concepts of perception, cognition and language are problematic since they do not take into consideration cultural activities that have an influence on what is attentively perceived, reflected and considered to be worth communicating. The diversity of culture-dependent forms of speaking about colors have not yet been analyzed sufficiently. The methodology of determining color terms by using colored chips does not record the diversity of specific color terms and color classifications with regard to particular objects or activities. For example, German, English or French people being asked to denominate a light yellowish-brown color chip would not term it blond(e), since this expression is restricted to hair. Neither does the use of color chips record color terms which refer to processes of a change in color, such as in “the hair turns gray” (German has a special verb: ergrauen, see also grünen, whereas English uses the adjective “turn green”) or “the sky reddens”, “his face reddens with shame” (German: erröten). Color terms are not only adjectives (see Heeschen this volume, for languages of New Guinea). There is also a functional use of color terms that is not noticed in color chip based studies. In some cases, color appears as a distinguishing marker in expressions, such as “white versus black coffee”, “white versus black people”, “red versus green apples”, “white or red wine” (cf. Wyler 1992: 38–42, 92f.). Those expressions belong to contexts of talk in which the real color does not matter – wine or coffee, for instance, are not white, and ‘black people’ are not really black. Descriptions of color effects have scarcely been studied, just as connotations or further meanings of object-related color designations, such as “green tomato” meaning ‘unripe tomato’. In that case, a color term is used to indicate a stage within a process (see further examples in Heeschen this volume). Often, it depends on the purpose of communication how a color term is used and which aspects of color experience are highlighted thereby. The reference function of color terms is not of main interest in every case. In advertisement and commercial contexts, for example, color names are carefully chosen or created in order to achieve the right connotations with respect to a special group of customers (cf. Wyler 1992, this volume; Graumann this volume).



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.6 (389-447)

Martina Plümacher and Peter Holz

The prevailing form of research on color categorization and lexical coding of color is criticized in particular for three reasons: focusing only on color lexemes, it does not turn attention to the whole diversity of forms and functions of language with regard to color. A result of this is that culture-dependent meaning extensions and complex networks of meaning are not analyzed. Finally, ‘basic color terms’ in Berlin and Kay’s sense are hardly distinguishable from other monolexemic and object-independent color words (for criticism of this stream of research see Lucy 1992, 1997; Wyler 1992; Lehmann 1998; Dedrick 1998; Dubois & Grinevald 2000, and in this volume Heeschen; Niemeier; Plümacher). In the last decade, some studies have tried to remedy the deficiency. Siegfried Wyler (1992, 2006) analyzed forms of color classification in different branches as well as morphology and usage of color terms in English and German. Beat Lehmann (1998) gave an overview of forms and function of color terms in German and analyzed the usage of color terms in text and speech. Many contributions to this volume also bring into focus the different forms and functions of linguistic representation of color and odor experience. Semantic networks, metaphoric expressions, connotations, and the different language uses of experts are analyzed. In comparison to linguistic studies on color terms, odors have been scarcely subject to linguistic investigations. Studies, especially from psychology and psycholinguistics, emphasize the problematic relationship between odor perception and its linguistic representation. As mentioned above, there is a lack of specific odor names: the perceptual discrimination between odors is hardly represented linguistically, and ‘basic terms’ of odors, i.e., terms abstracted from their reference to particular objects, are missing – at least in the most commonly studied languages. Usually, the source of an odor is named, provided that it is well-known. Asked about their olfactory perception, people give the name of the object that is known as the cause of the odor, for instance vanilla, lemon, leather, etc. If odors cannot exactly be connected to a source, expressions accentuate similarities: “it smells like an over-ripe banana”. Also categorizations referring to groups of objects are common, such as lemon-like, perfume, candy (cf. this volume Dubois; Holz). People often evaluate odors, unknown mixtures in particular, according to personal effects, such as pleasant or unpleasant. Descriptions frequently turn to metaphors: something smells floral, sweet, hot, spice, pungent, heavy, etc. Perfumers developed several classificatory systems that make partial use of similar metaphors (cf. Gschwind 1998). To some extent, there is a language on odors. In many cases, however, we cannot take it for granted that people agree on their descriptions of odors. This condition of a shared language is not met in many language communities, despite some exceptions when people refer to substances of salient, well-known odors in contexts of everyday life, for example, when they notice the smell of just mowed grass, coffee, roast, etc. As psychological tests on odor identification show, people

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.7 (447-522)

Speaking of colors and odors

have difficulties identifying odors correctly (cf. Zucco this volume). The difference, however, between the unreliability of odor identification in experimental situations and its reliability in everyday life has yet to be explained. In psychological literature, the fact that linguistic representations of odors are less developed is put down to three different causal relations (mainly from neurophysiology). One explanation, based on an evolutionary perspective of cognition, emphasizes that humans developed their visual and acoustic perception as primary sources of their orientation in space. The olfactory sense, on the other hand, then receded in importance among the information giving senses (cf. Zimmer 2001: 58; Lorig 1998: 391). It seems to have been restricted to particular fields, such as sexuality, the evaluation of food (Heeschen this volume, gives examples supporting this thesis) and religious practices too (see Detienne 1972; Douglas 1984; Albert 1996; Musset & Fabre-Vassas 1999; Candau 2000).1 As a result of this reduced human odor response, only a limited language on odors has developed. Tyler S. Lorig (1998: 392) suggests we should also consider repercussions of language on odor perception: “Since so much human behavior, and arguably all cognition is language-mediated, our limited language for odors may be a cause for our disregard of this sense rather than an effect of getting our noses off the ground. Possibly, it may have worked the other way: bipedality reduced the olfactory information load and allowed language to flourish.” A second explanation, from a neurophysiological angle, draws our attention to cortical structures of processing odor information. Odor information is first analyzed in the evolutionary oldest part of the brain, the so-called limbic system, before it is further processed in parts of the neocortex, especially in the right hemisphere. Direct connections to the left neocortex, where language is normally processed, are less developed (Bear et al. 2001: 576ff.; Vroon et al. 1996: 25; Zucco & Tressoldi 1989). For this reason, the cerebral organization of odor perception might not be adequately synchronized with language processing. Neurophysiologists, such as Burdach (1988: 22), consider that the lack of a stable lexicon of olfaction may be a result of this. Since the limbic system in particular is involved in the organization of visceral body functions, the autonomic nervous system, and the development of emotions, the fact that olfactory perception affects human behavior unconsciously seemed to be explained. For not all odor stimuli may reach a conscious level if they end at the level of the limbic system (cf. Zucco this volume; Chernigovskaya & Arshavsky this volume). The third explanation, neurophysiological as well, points out differences between olfactory perception on the one hand and acoustic and visual perception on . We are grateful to Danièle Dubois for making us aware of this.



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.8 (522-577)

Martina Plümacher and Peter Holz

the other hand. Visual and acoustic stimuli can be scaled according to wavelength, hue and brightness or volume of noise. In contrast to this, odors cannot be classified according common physical or chemical properties (Vroon et al. 1996: 62). There is no linear proportional relation between strength of olfactory stimuli and intensity of perception. In the case of some substances, it is known that a change of molecular concentration changes the perceived quality of odor (cf. Vroon et al. 1996: 83; Burdach 1988: 23). It is not clear how to classify odors according to their molecular sources. This, again, is considered as the very reason for the lack of systematic odor categorization and lexicalization. Despite all mentioned problems concerning the linguistic representation of odor perception, descriptions of odors found in everyday language, advertisement and technical language of perfumers should be analyzed with respect to systematic principles that might underlie the linguistic structures. As contributions to this volume show, there are rewarding avenues of research.

. On the contributions to this volume Wolfgang Wildgen (linguistics) inquires into the “architecture of sensibility and sense” in the interaction of perception and thought. He opposes the historical idea of progress which begins with psychophysical inputs to the senses and their cognitive processing, including their multisensory integration into a unifying experience of objects or events in space and time as well as formation of memory and imagination, and finally reaches the level of linguistic articulation. This line of progress is an illusion because a fundamental conflict exists between individual perception and action on the one hand and the social character of language on the other hand. Use of language requires the adaptation of personal experiences to a culturally transmitted system of linguistic rules, social beliefs and knowledge, social action and perception. Two dimensions have to be distinguished – the individual organization of the brain and the organization of forms of social life. The difference is shown with respect to odor and color perception and their linguistic representation. Wildgen points out that humans react to odors and are able to distinguish a large set of odors but, nevertheless, are unable to communicate those perceptions. He answers the question about conditions of the development of linguistic odor representation by referring to the need for communication, i.e., the need to ask or warn somebody or to order something with reference to an odor. According to him, the problem of a proper categorization and characterization of odors occurs if there is the need to communicate about them when they are not present as objects of reference. Wildgen states that shared reference of linguistic expressions has two conditions: a stable grounding of labels in personal perceptual experience and the efficient coordination of perception and language

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.9 (577-651)

Speaking of colors and odors

use between partners of communication. With respect to color, he takes up the well known fact that experts, such as painters, and non-experts perceive colors differently. He agrees with Fahle (this volume) that differences in the individuals’ color perception, even physiological differences, are hidden and overridden by language. Finally, stages in the evolution of cognitive and symbolic capacities are considered. Manfred Fahle (neurobiology) describes mechanisms in the neurophysiological processing of visual stimuli, i.e., light of different wavelengths and intensities. His neurophysiological view does not allow him to state that language refers directly to external objects. Strictly speaking, language labels particular cortical activation patterns. Fahle gives several reasons why we have to consider that even in the case of color perception, these patterns vary from individual to individual just as patterns of emotions do. Language use disguises the resulting vagueness of word meaning the same way as in communication about emotions. The most important reason to assume the individuality of color perception is evidence for the fact that the relative amount of those retinal cones that adopt either ‘red’ or ‘green’ light differs between subjects. Hence, we can conclude that people see the same red or green differently. Fahle explicates various physiological mechanisms from which we have to reason discrepancies between ‘objective’ properties of external objects and ‘subjective’ impressions. For instance, the absorption characteristics especially of the so-called red and green retina cones overlap. The subjective impression of ‘red’ color arises from the ratio of activation between different types of ‘color’ receptors in the retina. Further examples are the well-known phenomenon of perception constancy – a constancy of color in changing light for instance –, our failure to subjectively notice the lack of ‘blue’ wavelength information in the center of the visual field, after-images and their influence on subjective color impressions, evidence for unconsciously processed sensory information, and not least hallucinations and synethesias. To a certain extent, we can say that due to cerebral processing of visual information images of the colored world are created or constructed. Martina Plümacher (philosophy) discusses conditions of referential functions of color names and color descriptions in case reference objects are not present. In the first part of her paper, she analyzes the linguistic possibilities of denominating single hues, and points out two strategies that are particularly successful. On the one hand, there is a traditional and cross-cultural strategy of designating particular shades by referring to the prototypical color of a well-known object. On the other hand, a second strategy developed in the 18th century in Europe when a systematic relational order of colors, represented by circles and spheres of colors, was established. It was primarily based on knowledge about mixing colors, although additionally inspired by the development of physical optics. The relational ordering of colors according to hue, brightness and saturation enables people to



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.10 (651-701)

Martina Plümacher and Peter Holz

differentiate between various shades linguistically, without referring to colored objects. The late development of a systematic order of colors is evidence that even in the visual sphere, the right criteria of systematization had to be found in order to gain an abstract model of colors that extends the possibilities of linguistic color representation. The second part of Plümacher’s paper deals with descriptions of effects of color interaction in artistic color compositions. It refers to the endeavors of painters at the turn of the 20th century to develop a common form of communication about their new interest in abstract color compositions. Plümacher shows that metaphorical expressions were used to develop, to some extent, a technical language that does not work on a metaphorical basis but on knowledge about psychological insights into regularities of visual perception as well as about systematic effects of color compositions. This story is an example of the development of a technical language by experts and its conditions. Volker Heeschen (linguistics) shows the challenge to linguistics by studies of non-European languages. He reports on his studies of Papuan languages, Eipo and Yale in particular. In view of these linguistic cultures, it becomes clear that language has to be considered as a particular part within a broader cultural semiotic system. Heeschen brings into question the specific social role of language. For in the cultures, he is concerned with, people are well aware of numerous different odors and colors that are essential in their everyday life and their rituals. In contrast with this, there are only scarce linguistic means referring to these practiced distinctions. The awareness of characteristic phenomena does not necessarily lead to their linguistic representation. Heeschen points to the fact that activities are accompanied by non-linguistic signs, especially gestures, and emphasizes ritualization as a factor in reducing the need for linguistic forms of understanding. In a culture where many concrete things are negotiated non-verbally, language need not to represent all the practiced distinctions. It can be used in a distanced, reflexive mode within which experiences are generalized. According to Heeschen, there is especially one point where linguistic terms have to be specific, namely when dangerous peculiarities must be denominated. He points out that Papuan languages have many words and collocations of decay and rottenness including information about different degrees of edibility. Since also many color terms are used to indicate degrees of ripeness and decay, Heeschen assumes that the contrast between unripe and ripe, on the one hand, and young and old, on the other hand, is a reference system that partly underlies the terms of color and smell. He opposes Berlin and Kay’s characterization of linguistic cultures, such as Eipo and Yale, which had to be assigned to Berlin and Kay’s stage two where only three ‘basic color terms’ could be found. This characterization does not take into consideration numerous diverse forms of linguistic representation of colors, such as for instance specific object-related or event-related color terms, and the extension of meaning of those terms or collocations including a metaphorical use. According to Heeschen, languages should

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.11 (701-787)

Speaking of colors and odors

be analyzed in relation to cultural practices and their specific semiotic circles, and regarded a purposeful addition to them rather than a complete representation. Siegfried Wyler (linguistics) analyzes the verbalization of color in connection with textiles and cosmetics, as it can be found in particular in Great Britain today. In the first place, he expounds forms of linguistic color representation: the grammatical representation of colors as properties of objects as well as deviations from pure color designation, such as radicalized color names used to distinguish objects or degrees of ripeness and unripeness. Like Wildgen, Wyler emphasizes the cultur-dependent application of color terms. Butter, for instance, is white for the English child, but yellow for the Continental child. In the case of products, such as textiles or cosmetics, the distinguishing capacity of color names is essential. For their colors are carefully chosen in expectations of the prospective salability of the product and customer behavior. Wyler shows that color naming, too, is involved in strategies of sale and advertisement. He analyzes semantic and pragmatic principles of color naming as well as the morphological structure of color designations for textiles and cosmetics. Beside their role as markers of different kinds of products, color names function as trademarks of companies; they indicate quality or exclusiveness of products, present trends in fashions, and, most important to the salability of products, they appeal to the emotion and the attitude of the prospective customer by generating positive associations of desirable phenomena of life. In contrast to color names for textiles, color names of cosmetics are less accurate regarding the actual color of the product in question. More fancy names are used, some of them without any association to color. They only intend to be affective, emotional or merely easy to remember. It is a peculiarity of cosmetics that many color names appeal to synesthetic or multi-sensory associations by referring to objects of a pleasant aroma. Such names are involved in subjective olfactory interpretations. Wyler further indicates differences between color names used in sales talks and in magazines, and thereby confirms Heeschen’s thesis that language use is part of complex semiotic systems. In sales talk, when people can directly refer to texture and color of textiles, an accuracy in color designation is not required, and people content themselves with some general color names. In contrast, catalogues of mail-order companies are maximally explicit concerning color designation. Present-day fashion magazines, however, show a new trend in comparison to former issues. Photos take on the task of informing about color, as a result the number of color descriptions has shrunk. Andrea Graumann (linguistics) inquires into cognitive and psychological functions of color terms. She explicates how knowledge about these functions is used in advertisements, in particular in connection with the automobile industry. With reference to the concept of ‘semantic frames’, which has been established in cognitive semantics, Graumann shows that color terms do not only designate colors, but also arouse further connotations dependent on lexical and



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.12 (787-866)

Martina Plümacher and Peter Holz

empirical background knowledge of addressees. In addition, specific connotations emerge in the context of personal, often emotional experiences. Color names also have connotations that stem from the psycho-physiological effects of colors, that have been proven by color psychology. Graumann assumes that especially psychophysiological color impressions play a relevant role in composing complex color terms of metaphorical, often synesthetic meaning, such as cold blue, dynamic red or cheerful yellow. She asks about constraints of modifying a basic color term, and suggests that the semantic features of adjective and noun must be compatible regarding psycho-physiological color impressions or other forms of background knowledge. Imola red, a color term of an automobile, is one of her examples that show that metonymies mediate the activation of knowledge that is required to understand the collocation. Some collocations used in advertisements can hardly be decoded by the average language user but go down well with a specific social group. Graumann explicates how complex color terms, cryptic at first glance, appeal to specific life style images. Like Wyler, she emphasizes the function of color terms to hint at quality and luxury of the product. Susanne Niemeier (linguistics) concentrates on metonymies and thereby shows that playing with language is an important factor in expansions of meaning. Her analysis of metonymies is an example of how to draw up the semantic network of meaning that is connected with a single term. In her case, this term is the word blue in British English. In the first place, she explains metonymy as it is considered in cognitive linguistics. In this view, metonymy is one of the strategies of meaning extension that makes lexemes polysemous. Metonymies are highly culture-dependent, for specific background knowledge is required to understand metonymic expressions. Niemeier emphasizes that colors have motivated a lot of metonymies due to their immediate availability as sensory input. Even expressions that seem to be only descriptive, such as blue eyes, can be considered as metonymic in so far as they are based on a generalization of a basic or core meaning of the lexical item in question. According to Niemeier, the core meaning of blue refers to the natural blue of the sky or the blue sea. Meaning extensions can develop in different directions. Thereby contradictory meanings can be created. Using the example of bluestocking, Niemeier shows different forms of metonymic meaning extensions, such as part-for-whole metonymy, generalization, specialization. Knowledge of the history of different metonymic steps is often important in order to understand a specific expression. The meaning extensions of ‘blue’ presented in Niemeier’s paper make clear that translation as a one-to-one meaning is not possible in most cases. Therefore, discussions of radial networks of meaning appear to be a motivating instrument of foreign language teaching. Gesualdo Zucco (psychology) explains the unique nature of the olfactory memory system. With respect to this, he considers following phenomena as indicative: odors are poorly remembered initially, but well retained over time, i.e., in

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.13 (866-917)

Speaking of colors and odors

recognition tasks where memorized odors are presented again, the rate of recognition remains high, relatively unaffected by the interval of time. In contrast, memory of pictures or words wanes with time. Zucco explains the low initial acquisition level of odor memory by the specific perceptual character of odors. They are perceived as events with few attributes that could be used as traces of coding. Other characteristics of odor memory are that it is resistant to retroactive interference and seems to be unaffected by the conditions of learning as well as by subjective factors, such as the familiarity or the pleasantness of odor substances. Zucco explicates two of his experiments which show that learning strategies, such as verbal labeling, creating an image, describing a life episode, etc., are without significant effects on odor recognition, and that inter- and intra-inference affects the recognition of acoustic and visual material but not the recognition of odors. Another characteristic seems to be the weak relationship between language and odor memory. Zucco sees evidence to support this thesis from experiments that prove the people’s difficulties to identify, i.e., to label, presented odors adequately. The problem is not only that different subjects label the same odor differently, but in addition, that the same subject frequently encodes the same odor differently on successive presentations. Further, there seems to be evidence that odor memories cannot be intentionally recalled, that is to say that odors cannot be consciously represented. According to Zucco, this might explain the lack of effects of interference and learning strategies on odor memory because they work at the level of conscious representations. He concludes that odor memory seems to be mainly perceptual in nature and resistant to manipulations, for instance to language. Danièle Dubois (psycholinguistics) reports on studies of categorization of odors, sounds and noises, research that was conducted within an interdisciplinary program of cognitive science. She explicates the reasons why the procedures of explorations on color categorization, used by Berlin and Kay or Rosch, could not be used for this purpose. Research on odor perception is especially critical. There is, for instance, the lack of ‘measurement’ of odors and problems of how to clearly define the boundaries of an olfactory stimuli and the concept of ‘odor’ – should it be regarded as a physical or a psychological concept? Dubois opposes the prevailing assumption in identification tasks that there are ‘correct odor names’, i.e., names the experimenter expects with regard to the given sources of odor, i.e., odorants. Instead, odors should be considered as psychological entities. This means taking the subjects’ linguistic representations as indications of olfactory perception and its categorization. Any linguistic response to an odorant can then be regarded as the speaker’s attempt to communicate something about his/her odor memory and identification of the presented odorant. Dubois presents some results of her explorations of linguistic devices used for referring to odors and sounds. It is shown that descriptions could be classified according to different generic levels of conceptualization and with respect to forms of objectification or to ‘subjective’ versus



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.14 (917-995)

Martina Plümacher and Peter Holz

‘objective’ utterances. Especially odors are more often conceptualized as effects on the subject than as ‘objective’ properties of particular objects. The differentiation between forms of objectification is also important with respect to color conceptualization because colors too can be perceived with regard to their psychological effects on the subject as well as to properties of objects in the world or signals for objects. The study on odor descriptions shows the important influence of aromatized cultural artifacts over the subjects’ odor experiences. Dubois emphasizes that analyzing the diversity of linguistic representations of sensory impressions provides a foundation to clarify the relation between two oppositions – on the one hand, subjective and individual representations versus collective or shared lexical meanings, and on the other hand, natural sciences versus cultural sciences. Peter Holz (linguistics) analyzes descriptions of perfumes in advertising. In contrast to the meager and deficient forms of odor communication in everyday life, perfume advertising displays a highly elaborated way of speaking about odors. Like Dubois, he considers the linguistic form of odor representation to be indicative of odor conceptualizing and the organization of odor memory. Several forms of descriptions are distinguished – hedonistic judgements, descriptions with iconic reference to a similar memorized odor, descriptions with indexical reference to a source of odor, and metaphoric descriptions, especially linguistic synesthesias. The latter are explicated in greater detail because they are frequently used in perfume advertisement. Linguistic synesthesia can activate multi-sensory associations, for instance a holistic bodily feeling. An expression, such as a sickly sweet scent, refers not to an odor alone but also to a personal body feeling. Holz shows that perfume advertising carefully uses metaphoric descriptions to develop poetic dimensions of language. Referring to Roman Jakobson’s criteria of the poetic function of language, he is able to prove that in texts of advertising this form of language is prevalent. Holz emphasizes that the language usage in perfume advertisements demonstrates the important role of linguistic cross-modality-references with respect to the complexity of odor perception. If scents evade precise referential access and are perceived as effects on the subject’s body, these metaphorical, synesthetic and poetic descriptions are examples of an adequate form of speaking about odors, although advertisement is a very specific example, because it appeals to desires of customers. Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru (linguistics) discuss synesthetic expressions of language, which are not considered as metaphors but rather as linguistic representations of cross-modal sensory associations that stem from a multi-sensory experience of world. The authors assume that linguistic expressions, the word apple for example, arouse associations of different sensory modalities connected with apples. ‘Meaning’ is identified with the activation of memorized multi-sensory experiences, i.e., associations, in the strict sense. This idea is represented by the “physiological-psychological model”. It presupposes

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.15 (995-1041)

Speaking of colors and odors

that all the cortical areas involved in processing sensory information play a part in processes of language production and language reception. Allport’s theory of distributed semantic memory is the theoretical background of this model. The authors consider the question of why only particular synesthetic collocations occur but not every theoretically possible word combination. If the form of collocations is only due to a co-activation of different senses, the lexemes should occur in any sequence. Word combinations, such as smelly taste or reddish warmth, should be as common as sweet smell or warm red. The strength of sensory relations in daily experiences is considered as a reason: whereas the tactile sense is almost always activated in combination with the visual sense, the opposite case does not hold. This thesis leads to the question of how to explain synesthetic expressions, such as fragrant music, which refer to sensory modalities rarely activated together. Such word combinations are explained by reference to emotions which link up different events. Tatiana Chernigovskaya and Viktor Arshavsky (psycholinguistics) report on their experiments on the importance of cerebral right hemispheric activity to the type of human behavior towards odors. The authors argue for more differentiated psychological research methods that take into account cultural differences and different personal cognitive styles, since there is evidence that the development of right or left hemispheric functions is different, depending on cultural styles of life, especially on the role that is given to language or non-verbal communication and orientation in space. There is further evidence that differences in odor perception are due to different hemispheric activities. For instance, it is known from studies on cerebral hemispheric asymmetry that the right hemisphere deals with global Gestalt recognition which is especially required in the case of odor perception, since odors can hardly be analyzed according to their components. The authors emphasize the synesthetic nature of odor perception that is never separated from the accompanying sensory signals and more general contexts. In their experiments, they differentiate between right and left hemispheric personalities in order to prove significant differences in their reactions to odors and colors. A further point of interest is the correlation between these reactions and changes in alarm level, as proven in the Spilberger test. Chernigovskaya and Arshavsky assume that, even at an unconscious level, odors can have signal meaning and impaired odor perception might cause alarm or depressive states. Their experimental results support their hypotheses concerning the important role of the right hemisphere with respect to odor perception. According to their data, most right hemispheric personalities demonstrate a specific odor memory and the ability to speak about odors.



JB[v.20020404] Prn:26/01/2007; 14:21



F: CELCR801.tex / p.16 (1041-1147)

Martina Plümacher and Peter Holz

. Acknowledgements We are grateful for the financial support we received from the Fritz-Thyssen Foundation and the University of Bremen, which gave us the opportunity to organize a conference on the question “How can language cope with color and smell?”, autumn 2002 at the University of Bremen, – the conference which led to this book. We thank each of the contributors for their willingness to point out converging evidence and diverse or even contradictory theses and for their valuable corrections to this introduction. Not least, we wish to thank Kathryn Nussdorf for correcting our English.

References Albert, J.-P. (1996). Odeurs de sainteté: La mythologie chrétienne des aromates (1 1990). Paris: Editions de L’EHESS. Bear, M. F., Connors, B. W., & Paradiso, M. A. (2001). Neuroscience – Exploring the Brain. Baltimore et al.: Williams and Wilkins. Berlin, B., & Kay, P. (1969). Basic Color Terms – Their Universality and Evolution. Berkeley: University of California Press. Brown, R. W., & Lenneberg, E. H. (1954). A Study in Language and Cognition. International Journal of Abnormal and Social Psychology, 49, 454–462. Burdach, K. (1988). Geschmack und Geruch: Gustatorische, olfaktorische und trigeminale Wahrnehmung. Bern et al.: Huber. Camporese, P. (1995). Les effluves tu temps jadis. Paris: Plon. Candau, J. (2000). Mémoires et expériences olfactives: Anthropologie d’un savoir faire sensoriel. Paris: PUF. Casson, R. W. (1997). Color shift: Evolution of English color terms from brightness to hue. In C. L. Hardin & L. Maffi (Eds.), Color Categories in Thought and Language (pp. 224–239). Cambridge: Cambridge University Press. David, S., Dubois, D., Rouby, C., & Schaal, B. (1997). L’expression des odeurs en français: Analyse lexicale et représentation cognitive. Intellectica, 24, 51–83. Dedrick, D. (1998). Naming the Rainbow, Colour Language, Colour Science, and Culture. Dordrecht and London: Kluwer. Detienne, M. (1972). Les jardins d’Adonis: La mythologie des aromates en Grèce. Paris: Gallimard. Douglas, M. (1984). Purity and Danger. An Analysis of the Concept of Pollution and Taboo (1 1966). London: Routledge. Dubois, D. (1997a). Catégorisation et cognition: de la perception au discourse. Paris: Kimé. Dubois, D. (1997b). Cultural beliefs as nontrival constraints on categorization: Evidence from colors and odors. Behavioral and Brain Sciences, 20, 188. Dubois, D. (2000). Categories as acts of meaning: The case of categories in olfaction and audition. Cognitive Science Quarterly, 1, 35–68. Dubois, D., & Grinevald, C. (2000). Denominations of colors in practices. In A. K. Melby & A. R. Lommel (Eds.), Lacus Forum XXVI The Lexicon, Lacus (2000) (pp. 143–151). Friedl, E. (1979). Colors and culture change in Southwest Iran. Language in Society, 8, 51–68.

JB[v.20020404] Prn:26/01/2007; 14:21

F: CELCR801.tex / p.17 (1147-1239)

Speaking of colors and odors

Gschwind, J. (1998). Repräsentation von Düften, P. Neumann & M. Schleidt (Eds.). Augsburg: Wißner. Harper, R., Bate-Smith, E. C., & Land, D. G. (1968). Odour Description and Odour Classification. London: Churchill. Holz, P. (2005). Die Sprache des Parfums – Eine empirische Untersuchung zur Grammatik, Metaphorik und Poetizität des Parfumwerbetextes. Hamburg: Verlag Dr. Kovac. Kuschel, R., & Monberg, T. (1974). ‘We don’t talk much about colour here’: A study of colour semantics on Bellona Island. Man, 9 (n.s.), 213–242. Lehmann, B. (1998). Rot ist nicht ‘rot’ ist nicht [rot]. Eine Bilanz und Neuinterpretation der linguistischen Relativitätstheorie. Tübingen: Narr. Lenneberg, E. H., & Roberts, J. M. (1956). The Language of Experience: A Study in Methodology [Indiana University Publications in Anthropology and Linguistics, Memoir 13]. Bloomington: Indiana University Press. Lorig, T. S. (1998). On the similarity of odor and language perception. Neuroscience and Biobehavioral Reviews, 23, 391–398. Lucy, J. A. (1992). Language Diversity and Thought. A Reformulation of the Linguistic Relativity Hypothesis. Cambridge: University Press. Lucy, J. A. (1997). The linguistics of ‘color’. In C. L. Hardin & L. Maffi (Eds.), Color Categories in Thought and Language (pp. 320–346). Cambridge: Cambridge University Press. MacLaury, R. E. (1992). From brightness to hue. An explanatory model of color-category evolution. Current Anthropology, 33 (2), 137–163 (comments, reply and references: 163– 186). Musset, D., & Fabre-Vassas, C. (Eds.). (1999). Odeurs et parfums. Paris: Éditions du CTHS. Rouby, C., Schaal, B., Dubois, D., Gervais, R., & Holley, A. (Eds.). (2002). Olfaction, Taste and Cognition. Cambridge University Press. Vroon, P., Amerongen, A. van, & Vries, H. de. (1996). Psychologie der Düfte. Zürich: Kreuz. Wyler, S. (1992). Colour and Language. Colour Terms in English. Tübingen: Gunter Narr. Wyler, S. (2006). Colour Terms in the Crowd – Colour Terms in Use. Tübingen: Gunter Narr. Zimmer, R. (2001). Handbuch der Sinneswahrnehmung. Freiburg: Herder. Zucco, G. M., & Tressoldi, P. E. (1989). Hemispheric differences in odour recognition. Cortex, 25, 607–615.



JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.1 (47-137)

Color, smell, and language The semiotic nature of perception and language Wolfgang Wildgen The historical roots of the propositional versus the imagistic view on meaning in perception and language can be found in Aristotle’s Categories and in the medieval and Renaissance models for mnemonic networks of sign systems (Lullus, Bruno). The architecture of sensibility and meaning has been sketched in Condillac’s Traité des sensations (1754) and it demonstrates a fundamental conflict between sensation on one side and communication/ language on the other. Their relation has been further specified after Darwin in the context of an evolutionary theory of human cognition and language. In the final section, the contributions of different disciplines (neurophysiology, psychology, ethnology, linguistics, art history) to the question are compared in order to conceive a strategy of synthesis.

.

Is the propositional nature of language an Aristotelian myth?

Aristotle’s Categories and Porphyry’s commentary on it were the standard logical texts through the middle ages and into the Renaissance. Aristotelian logic therefore had a deep influence on European philosophy. The first pages of Aristotle’s Treatise on Categories introduce the proposition as the foundation of logic. According to Aristotle’s Analytics, single words or word pairs, e.g., compounds and phrases are not sufficient and are not the proper entity for syllogistic argumentation. The logical machinery starts only when there are entities which are asserted or negated, or, to express the same thing in modern terminology, when some entity has the value true or false. This specification has been modified by multi-valued logic (cf. Ballmer & Pinkal 1983) and the assertion character of logic has been challenged by speech act theory (cf. Austin 1962). The holistic nature of truth conditional logic was further challenged by situation-semantics which focuses on propositional attitudes (believe, know, a.o.) and partial interpretations which take situations instead of worlds as reference domains. In spite of all these challenges to Aristotelian logic, the basic assumption that sentences, with their predicative structure and criteria of completeness, are the proper basis of human discourse, has not been seriously questioned, and alternative proposals by Charles Sanders

JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.2 (137-181)

Wolfgang Wildgen

Peirce1 and the early Wittgenstein (Notebooks 1914–1916) for an image-model of sentence meaning were ignored by the mainstream. Despite its long and honorable heritage, the Aristotelian position is a barrier to the integration of percept-based (e.g., visual and olfactory) and linguistic semiotics. The treatise on categories prepares the central Aristotelian endeavor: formulating an epistemology of scientific thinking and arguing and establishing the necessary preconditions for deductive reasoning. The propositional structure of such an (artificial) language is necessary, because syllogistic schemata must guarantee the truth of a conclusion given the truth of the premises irrespective of the meaning of single terms in the premises (and conclusions). The following sentences with a subject (bold) and a predicate (italics) have identical (formal) structures: (1) This apple is red (2) This apple is sweet (3) This apple smells nice

and they have a truth value (true or false). The image (prototype) of apple and the sensorial correlates of: red, sweet, nice are irrelevant for the application of syllogistic schemata, such as (1) ∧ (2) → (1) → (2) (1)

→ it is false that not (1)

The mechanistic deductive schemata of Aristotle neglect the basic question of how these complexes like “this” + “apple” + “is” + “red” build up a semantic/ cognitive representation or ‘image’ in a more abstract sense, given the semantic content of “apple” and “red” and the operational features of “this” and “is”. The first two are considered as logically opaque, the latter are eliminated by the translation into a proper logical expression. Thus, the demonstrative “this” helps to choose the mode of quantification and “is” stands for the predicative function itself. Our questions in this context are different: 1. What is the semantic space of “apple” (some prototype of apples in the community) and of “red” and how can these ‘mental spaces’ be mapped or blended to produce: ‘red apple’, ‘apple (is) red’, ‘red(ness) (of) apples’?2 . Peirce had devised his diagrammatic logic at the turn of the century, but his ideas only began to be known when he was rediscovered by Morris in the thirties and then understood from a different perspective. . The terms ‘map’ and ‘blend’ are used intuitively. The elaborated models of ‘mental mapping’ and ‘blending’ of Fauconnier and Turner (2002) do not cover the imagistic aspects of meaning and are formally variants of logical semantics in the classical sense.

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.3 (181-282)

Color, smell, and language

2. What makes the completion of a semantic gestalt containing a number of elements (morphemes, words)? What makes these complete configurations apt for being framed by ‘propositional attitudes’ and speech acts verbs and thus contribute to communication/ social action via language? Examples: (1) I see that this apple is red. (2) I believe this apple is red. (3) I hope that this apple is red. (4) I promise this apple will be red. (5) Give me an apple which is red!

The question of semantic completion or gestalt is not linked to the proposition by definition, as it applies also to compounds (even to any morphologically complex entity), to phrases like NPs, PPs, VPs, and to paragraphs, narratives, arguments, conversational sequences. It is the basic question of linguistic semantics. One can even say that the view of the proposition as the basic unit in language is a myth. It is only in the context of Aristotelian syllogistics and modern deductive logical proofs that truth-functional units gain a prominent role. The fundamental problem is that of the mapping/ blending of semantic and functional spaces and the stability and regularity of a speech act which refers to these spaces.

. Is there a language of memory? The “ars memoriae” is as old as Aristotelian logic and I shall introduce the former here as a historical alternative to the latter although in their historical development both traditions have been mixed. Explicit models of artificial memory were introduced by Raymundus Lullus in the 13th century and by Giordano Bruno in the 16th (cf. Wildgen 1998). Both used ideal geometrical entities: regular polygons and polyhedrons, circles and spheres segmented by in- and circumscribed regular polygons and polyhedrons3 along with regular operations on these figures. Lullus rotated one, two or three circles with nine subdivisions (linked to three inscribed regular triangles) and Bruno used regular paths in the quadratic houses defined by a geometrical grid of 3 × 3 (× 3) cells (cf. Wildgen 1998, 2000: 204–212). . Polyhedrons, which are called “Platonic”, are only referred to by Bruno but not systematically used, because his mnemonic schemata are mostly two-dimensional. Kepler took up the Platonic view of nature in his first writing: “Mysterium cosmographicum” (1596), which refers to the complete list of Platonic solids.



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.4 (282-282)

Wolfgang Wildgen

Figures 1 and 2. The circles of Lullus and the house of Bruno. The circle of Lullus is subdivided into nine categories (B-K), which are enumerated in the order of 3×3 by triadic categories like: prime (E), medial (F), final (G). The atrium of Bruno is subdivided into 8×3=24 rooms which are inhabited by pictures represented by nouns. The first atrium (Altaris) is followed by 23 atria. Thus the whole system of pictures contains 24×24=576 pictures (nouns). If one follows pathes in these houses one can generate texts.

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.5 (282-337)

Color, smell, and language

What does this entail for semantics? The semantics of artificial memory are based on words (or images, i.e., words with an imagistic content); the operative, functional parts of speech are presented by motion in the system, e.g., a path in the square grid (Bruno) or a rotation of the circle (Lullus) and the reading out of a combination thus produced. This technique basically maps the ‘images’ of words, using a geometrical procedure, to ordered pairs or triads of images, which build the imagistic skeleton of a story.4 The basic entities are therefore: – –

‘images’ which have a position in a geometrical framework, a house, a town, a landscape, etc. ‘motion’ in the spatial frame which reproduces a scene, where something happens, i.e., a primitive story.

In contraposition to Aristotelian deductive logic, artificial memory produces semantic gestalts in an imagined space by means of ‘imagistic motion’ in this space. The regularity of the space and the completion of a path respond to the selection of well-formed maps/ blends. As the motion may have two, three, four, . . . steps, there is no preference for dyadic subject-predicate constellations. In consequence, the different valence patterns of verbs (with valence 1, 2, 3, 4) fit this framework much better than the Aristotelian subject-predicate schema (cf. Wildgen 2002). The disadvantage of mnemonic semantics in the style of Lullus and Bruno is their license for very strange and fantastic combinations, a kind of ‘anything goes’ in semantics. Bruno develops a labyrinth of mnemonic combinations in which finally any text can be transformed into any other applying metaphorical and metonymical replacement.5 One could oppose the mnemonic semantics of ‘creativity’ to the logical semantics of ‘control’. In a psycho- or neuro-semantic context the individual’s imagination in thinking and dreaming could fit the dynamics shown by Bruno, the control of a (scientific) community over accepted argumentation (proofs) would fit the Aristotelian model. In the context of our topic “How Can Language Cope with Color and Smell?” Bruno’s model seems to be more promising as it does not eliminate individual creativity and the (free) operation of imagination. Both models have to cope with another basic problem: the grounding of meaning (in the users’ bodies, in their environment insofar as they are interacting with it). Computer semioticians have . Cf. Wildgen (1994: Chapter 1) for a critique of proposals for ‘imagistic’ semantics and the proposals for a realistic model of semantics (including imagistic aspects). . Cf. Wildgen (1998: 181–184) and the last text Bruno published: De imaginum, idearum et signorum compositione (1591) or the English translation: On the Composition of Images, Signs and Ideas (Bruno 1991).



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.6 (337-397)

Wolfgang Wildgen

become aware of the problem of grounding as a basic question in the simulation of language behavior. Thus Cangelosi, Greco, Harnad (2002) write: The symbol system’s property of systematic semantic interpretability implies that any part of the system, and the whole system itself, can be assigned a meaning. Therefore a fundamental question must be asked: How is a symbol given a meaning? This is the problem of symbol grounding. The type of link that exists between symbols and objects is of central importance when using symbol systems as models of language and cognition. Cognitive semantics avoid this problem by ignoring it or trivializing it. (Cangelosi, Greco, & Harnad 2002: 193)6

The problem is a traditional one and I shall discuss the solution proposed by Condillac in 1746, which brings us to the title of this conference “Sense and Sensibility”. How is “sense” understood as ‘meaning’ organized in symbolic forms (language, myth, art) grounded in human sensibility.7

. Is there an architecture of sensibility and sense? The composition of larger structures of words/ imagistic units asks for largescale organizational principles, which may be called the architecture of sense. The simplest architectures are: –

–

–

Bottom-up: Sensations produced by our sense organs, triggered by external events are stabilized (transformed from a continuous flow to discrete entities), combined, abstracted; they are memorized and come to conscience and thus make up the operations of thinking. This view has two critical zones: the psychophysical transition in the sensory organs (and correlated neural areas) and the combination and abstraction of perceptual ‘atoms’ (produced in the first transition). Top down: Any perception, memory, imagination, combination of ideas presupposes the mode of thinking and thus incorporates principles and axioms of mental processing. The psychophysical contact-zone immediately selects those data which fit the mental processing (eliminating or forgetting everything else). Self-referential: Thinking itself is shaped in the processing of sense-data. Piaget’s equilibrium of assimilation (top down) and accommodation (bottom-

. Further aspects of an evolutionary theory of language are discussed in Wildgen (2004: Chapter 8). . Cf. the Philosophy of Symbolic Forms proposed by Ernst Cassirer between 1921 and 1930. Its relation to language philosophy and linguistics is discussed in Wildgen (2003; cf. 2004: Chapter 10).

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.7 (397-470)

Color, smell, and language

up) clearly expresses this view, which is already prefigured in Locke’s and Condillac’s essays (1690 and 1746). Locke makes a first and Condillac a decisive second move by including the problem of signs (words) in the sequence: word → (sign of an) idea → (sign of an) external event

Ideas are constructs at an intermediate level and not independent of the stabilizing and organizing function of words (at least after the acquisition of language). Perception is itself a semiotic activity. The 18th century tradition was taken up by Ernst Cassirer (via his preferred authors Herder, Goethe, W. von Humboldt). The “symbolic form” cuts the continuous flow of sensations, categorizes it for memory and thus links individual perception to collective structures like: language, myth, art, technical skill, ethics, etc.8 In spite of these advances, the following major problems were not solved: –

–

How are meanings grounded in the activity of our (subconscious) sensory activities, which are themselves adapted by evolution to selected features of our environment? The quality of this grounding (the ‘correct’ selection, the stable transfer of relevant structural relations up to the highest levels of cognition) is crucial for the functional fitness of the “animal symbolicum”, as man is called by Cassirer (1946). How can individual thinking (as representation) be both economical and still representative in reference to external reality? Locke and Condillac have made clear that the empirical sciences like psychology (Condillac is considered the founder of French psychology), sociology (John Locke is one of the founders of political science); anthropology (which was developed in the late 18th century, e.g., Herder 1784) will have to contribute to an answer, which cannot be found by philosophical speculation.

In his Traité des sensations (1754) Condillac presents the following thought experiment. A marble statue is successively endowed with sensual experience. Condillac begins with olfaction. The statue first is what it smells, second it distinguishes pleasant and unpleasant odors. Finally, memory continues the move of attention after the stimulus is seized. The difference is one of activity, if motivated by memory, and passivity, if motivated by a stimulus (cf. Condillac 1754: 49).9 With the two basic principles pleasure/ aversion and memory/ comparison, Condillac develops a system of ideas just based on olfaction. In a similar vein he treats hearing, taste, and finally sight, consisting of light and color. All of them allow for positive/ . In Wildgen (2004: Chapter 9), they are called “genres of collective symbolic forms”. . Activity is linked to a force inside the individual and passivity to a force outside (ibidem).



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.8 (470-578)

Wolfgang Wildgen

negative evaluation and memory/ comparison and thus for order. Their common limit is established by the idea of extension in space, solidity, and identity. This major ingredient is only ‘taught’ to the four senses by touch (he follows Buffon 1749 in this perspective). The architecture Condillac proposed separates olfaction, hearing, taste, and sight, which allow for evaluation and comparison. The integration into a world of conceived objects in space is, however, due to touch. The transition from sensibility to meaning/ sense is defined by the evaluative and abstractive capacity of our mind and by the integration of different types of experience: the sensibility of touch is for Buffon and Condillac the origin of the cognitive integration of the different types of sensibility.

. Sensation vs. communication: A field of conflict Condillac considered perception as the first semiotic level and signs as the necessary stage to stabilize and organize memory and thinking. Thus a straight line of progress in complexity is proposed: – – – –

psychophysical input to the senses, perception as formation of signs for memory and imagination (dream), language as intermediate stage of stabilization and necessary for the organization at a high cognitive level, (scientific) knowledge based on a linguistically organized universe of discourse.

This line of progress is, however, an illusion because a fundamental conflict exists between: – –

individual perception and action, and social communication.

The first field has as its domain the individual brain (including the sensory organs) and the individual development (maturation and basic learning in an adequate environment, interaction with the environment). The second field concerns the adaptation to a culturally transmitted system of linguistic rules, social beliefs and knowledge, social action and perception. Although processes of learning and socialization link both domains, the basic mechanisms are different. The organization of the brain cannot be identified with the organization of a community, and communication underlies different conditions of coding, of transmission, of functional adequacy, and economy than those operative in sensory organs and cortical centers of perception and motorplanning. I shall exemplify the differences in relation to olfaction and color.

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.9 (578-602)

Color, smell, and language

. Perception and communication in olfaction Humans are constantly exposed to odors. They may react to them positively or negatively, and their attention and memory may be controlled by olfactory impressions. Nevertheless, the level of communication is only reached if someone has the need to speak to somebody, to ask somebody, to warn somebody, to order something with reference to an odor. If both the speaker and the audience are exposed to the same odor, it is enough to call the attention of the audience to the odor and then to perform a speech-act referring to the (supposed) common perception. It is only if the odor is not present and must be represented, that the problem of a proper categorization and characterization occurs. It can only work if a stable system of meanings grounded in (almost) identical olfactory experiences is given. In a community, where this demand does not occur frequently, communication with reference to odors will be very insecure, unstable, poor or even meaningless. Thus, in the extreme case every participant in a communication may be perfectly able to perceive and if necessary distinguish a large set of odors, but totally unable to communicate those perceptions. The necessary ability concerns first the labels (lexical or periphrastic), second the stable grounding of these labels in personal perceptual experience and third an efficient coordination with the perceptual patterns in those, with whom one wants to communicate. Thus, it does not help just to learn words for odors, and even if these labels are individually grounded in olfactory experience, olfactory communication remains vague because no common grounding has been achieved. In expert communication such a grounding may be achieved in training situations and by explicit conventions on language use. This is demonstrated in the contributions by Martina Plümacher und Siegfried Wyler (in this volume). The report of Danièle Dubois on experimental research in odor, sound and color categorization (in this volume) makes clear that the grounding of odors is difficult or even impossible in experiments, because the chemical structure of odors is very complex and behavioral reactions are very context-dependent. There is nothing which corresponds to the Munsell classification of colors in the domain of odors. Another problem already mentioned by Condillac (1754) concerns the organization of the field of odors and its dependence on the conditions and contexts of memorization.10 Thus, if one experiences a series of odors, e.g., of flowers, coffee, and a nearby factory, a network which links and partially mixes these impressions and the evaluative reactions to them, is created in memory. As the neighborhood . Cf. Condillac (1754/1821: 81): “Prenons par exemple cette suite, jonquille, rose, violette. Dès que ces odeurs sont constamment liées dans cet ordre, une d’elles ne peut affecter son organe, qu’aussitôt la mémoire ne lui rappelle les autres dans le rapport où elles sont à l’odeur sentie.”



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.10 (602-662)

Wolfgang Wildgen

of odors in time and space, the amount of attention paid to them, the pleasure or displeasure associated with them change from one individual to the next, the network-structure of olfactory memories and associated values is basically different from one individual to the other. Possibly, it is more different for individuals living in different places and social and cultural contexts. The cultural complexity of discourse on odors is shown in the contribution of Volker Heeschen (in this volume) based on Eipo, Yale and other Papuan languages. All these complexities and interdependencies cannot be communicated without prior cognitive adaptations between speakers and hearers. Insofar as olfaction is to a large degree subconscious, the adaptation or social coordination is difficult. Therefore communication can only use the universe of individual olfactory perceptions and associated reactions to a minimal extent or these may even be inaccessible to communication.11 The results in the psychology of odor memory are discussed by Gesualdo Zucco (in this volume) and aspects of linguistic reference to odors in texts about perfumes are treated by Peter Holz (in this volume). The relation between different modes of categorization for odor and color (sound) is discussed by Danièle Dubois (in this volume). . Perception and communication of colors Condillac (1754) argues that the perception of colors alone does not constitute colored places, situations, motions, or colored objects, it just constitutes a continuum of colored surfaces limiting one against the other. It is our memory which allows us to see two or three different colors and to make a journey from one color to the other. Nevertheless, the perception of colors contributes an idea of extension, a space of colors. He also points to the difference between a naive observer of a painting and the painter himself. A painter and I see all the parts of the painting; but, whereas he extricates them quickly, I have such difficulties discovering them, that I have the impression that I discover in any moment parts which I have not yet seen. (Condillac 1754: 111)12

Again, the astonishing differentiation of human color vision does not automatically entail a rich system of ‘ideas’ of colored surfaces in our mind. Attention, . This case is clear for animals that have an even poorer communicative repertoire, like apes, dogs, fishes. Social coordination can in these cases only be enabled by instincts, or by very selective, learned reactions. . Translation by the author. Cf. Condillac (1754: 111). “Un peintre et moi nous voyons également toutes les parties d’un tableau : mais tandis qu’il les démêle rapidement, je les découvre avec tant de peine, qu’il me semble que je voie à chaque instant ce que je n’avais point encore vu.”

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.11 (662-713)

Color, smell, and language

memory, the organization of the color space (surface) and paths of visual attention are necessary to produce a selected set of ‘color ideas’. Higher levels are only achieved if the other types of sensibility are integrated and form an organization linked to objects, events, and actions.13 This insight shows up in many contributions of this book. In his contribution to this volume Manfed Fahle shows how individually variable color vision is. The differences between the physical world and the subjective experience are dramatic. However, language “can disguise physiological differences between observers by normalizing the responses to given stimuli, irrespective of the differences in internal representations” (Fahle, in this volume). Again, the critical line is the transition to communication. As the divergence of color terminologies in the languages of the world (cf. Berlin & Kay 1969) shows, the need for a basic (stable) set of color terms is very variable between languages even on a similar level of cultural development; i.e., communication with reference to colors is not a universal and central concern of linguistic communities. This does not preclude that the cognitive relevance of color perception is very high in individual perception, i.e., perception and communication do not share the same functional pattern, do not respond to identical (or similar) needs. They follow their own principles and laws. This topic is discussed under many different aspects in the contributions to this volume by Danièle Dubois, Andrea Graumann, Susanne Niemeier, Martina Plümacher and Siegfried Wyler. In a given linguistic community some sub-communities, i.e., those of painters or other professionals develop specific visual competences and techniques of teaching or even arguing about colors which modify the demands on communication about colors. Models of colors like those developed for and by artists help to enrich and stabilize professional codes in communication about colors and point to the underlying categorical architecture of communication. The contribution of Martina Plümacher (in this volume) presents different relational orderings of colors since the 17th century and color descriptions by artists like Paul Klee and other modern artists. . Some consequences Our reflections point to three major problems or problematic transitions between “sensibility” (perception) and “sense” (linguistic meaning): . Condillac argues that the sensibility of touch is able to control this higher reorganization, we can generalize this claim towards a synesthetic reorganization of particular sensibilities (the old idea of a ‘sensus communis’); cf. the contribution of Shibuya, Nozawa & Kanamaru in this volume.



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.12 (713-790)

Wolfgang Wildgen

a. The transition from single sensibilities, e.g., olfaction or color sensibility, to an integrated perception. b. The transition between presentation and representation. The latter has to be specified in relation to attention, memory and spontaneous imagination. c. The transition between perception, which is basically governed by principles of human neural architecture and dynamics, and communication, which is based on principles of social interaction and organization. The most dramatic transition and the evolutionarily most recent one is that of (b) to (c).

. Is the evolution of symbolic communication based on human sensibility? The levels and transitions distinguished in the last section recall stages in the evolution of cognitive and symbolic capacities (cf. Wildgen 2004). From an evolutionary perspective different aspects can be distinguished (a* is the correlate of a in the list above from an evolutionary perspective). a*. Evolutionary changes in the profile of sensory capacities, e.g., the evolution of color or 3-D-vision, the (relative) regression of olfaction as compared to audition in humans. The problem of integration in view of a unitary representation of biologically crucial aspects of the environment is given from the beginning, e.g., in uni-cellular beings. With further specialization, concentration and complexity of perception, these divergent functions have to be integrated again and again in order to guarantee a unitary (holistic) representation, on which action can be based. b*. Memory is surely a cognitive capacity of many animals and there are many ways in which memory may be stored (in the brain or in lower neural structures). Memory and its interaction with perception and motor control are very basic biological faculties. What is new in higher primates is “displaced reference”, i.e., “the feature of language that allows humans to establish reference to objects, events, and ideas that are not present at the time of interaction, obviating a dependence on environmental conditions for communication” (Morford & Goldin-Meadow 2001: 173 based on Hockett 1960). Communication about displaced referents and events must be preceded (not only in evolution, but also in child development) by the awareness (consciousness) of past (and future) objects and events. It is here that displaced reference which organizes the present in function of past and future becomes crucial. The imagination of future actions, the design of utopia, the conscious forming of human society and man himself is the promising and at the same time frightening new phenomenon. An implicit knowledge (without consciousness) of past events

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.13 (790-843)

Color, smell, and language

is certainly present in many animals, e.g., horses, dogs, cats, etc.; the imagination of future objects and events which allows for planning, for (spontaneous) novel solutions based on past experience is more restricted, but chimpanzees are certainly able to achieve it. Therefore the systematic organization of past experience (products of sensibility) and their projection into future actions is a cognitive level which is a precondition for human language (cf. Wildgen 2004: Chapter 2). c*. Life in animal societies is possible without language. Social communication and social competence may be organized on the basis of instinctive patterns, adaptations (behavioral learning) and natural signs (of strength, desire, need, etc.). The rather quick social dynamics apparent since late Paleolithic, Mesolithic and Bronze Age societies only became possible because they unfolded a whole range of symbolic capacities, which were part of the biological heritage of Homo sapiens (probably since 150 to 200,000 y BP). They were distributed over different media of communication and language is just one of them. Observable motion/ behavior in space/ time (places or paths frequented/ avoided, proxemic behavior), gestures and mimics, laughing and humor; dance, music, religious and magic beliefs are symbolic forms different from, although related to language. If we assume a standard repertoire of action patterns typical for a community, then speech-acts are only on sector of it. The profile of symbolic forms (media) may show universal functional preferences: orientation in space/ time, symbolic categorization of resources (plants, animals, materials like stone, bone or wood, water, shelters), narrative recapitulation of past (collective) experience and preparation of future actions, etc. However, communities may show rather divergent preferences as in the case of people living in tropical forests, arid plains or in arctic zones. In a larger island, populations living on the shores or in the low lands experience a different ecology from those living in the highlands (e.g., the populations in New Guinea; cf. the contribution of Volker Heeschen in this volume). The major functions of linguistic communication can vary with these contexts and as a consequence the symbolic forms may have different profiles. One can assume that the ecological contexts of Homo sapiens diverged only dramatically after their migration out of Africa (after 70,000 BP) and that the functional profile of human language was already well established at this time. This could explain the universal core of human languages (for more details on the evolution of human language cf. Wildgen 2004). The communicational profile of human language defines the field left for other types of communication (visual, olfactory, tactile), although intersections and multimodal communication are frequent. All contributions in this book contain examples of multimodality and intersection between visual, olfactory and linguistic (acoustic) modes of perception and communication.



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.14 (843-910)

Wolfgang Wildgen

From this evolutionary perspective, the different contributions to the topic: olfaction and color in this book refer to different levels in the evolution of man, which are still co-present in the architecture and dynamics of sensibility and sense: –

–

–

–

Neurophysiology of perception is nearer to the biological basis which man shares with animals. For this reason olfaction and color vision may (also for methodological and ethical reasons) be analyzed in animals. The results can (with a number of restrictions) be transferred to humans. Level a* is perhaps the most natural home for these endeavors (cf. the contributions by Fahle, and those by Holz; Shibuya, Nozawa & Kanamaru; and Plümacher, who relate the neural level to the levels b* and c*). Psychology of perception and cognition. Psychology, rather than mapping the neural processes measured in experiments (with animals), prefers to deal with more complex cognitive capacities, including memory and learning. A typical example is given by Gestalt psychology or by computer simulation of cognitive capacities. Behavioral experiments with human individuals are probably the prototype, although brain imaging during perceptual and other tasks on the one hand and socio-psychological experiments on the other are relevant for the field of psychology. Level b* is at the center of this research (most of the contributions in this volume refer more or less to the cognitive level) although only some authors are professional psychologists. Ethnology and anthropological linguistics consider the geographical, cultural and environmental diversity of cognition and language. Although differences in perceptual integration are found, the major results concern level c*, i.e., the symbolic organization of an ethnic entity, with its rituals, beliefs, social techniques and language. Attention is mainly given to the ethnical/ social contexts of symbol use and not so much to the organization of symbolic systems per se. This level is specifically addressed by Volker Heeschen (in this volume) and passim in other contributions, e.g., Tatiana Chernigovskaya & Viktor Arshavsky. Linguistics was traditionally concerned with the grammar, lexicon, and literature of written languages standing in a long historical tradition, which can be studied on the basis of texts. Since W. von Humboldt and the rise of anthropological linguistics (beginning of the 20th century), comparative and typological studies try to cover all language families. General linguistics and semiotics propose models for the universal linguistic capacity underlying all languages and the principles which govern all grammars (and lexicons). Linguistics operate mainly on level c* and only in some fields, such as cognitive linguistics does it consider the transition from level b* to level c*. Aspects of cognitive linguistics concerning odor and color are dealt with in the contributions of Andrea Graumann and Susanne Niemeier. Another cognitive

JB[v.20020404] Prn:26/06/2007; 10:15

F: CELCR802.tex / p.15 (910-982)

Color, smell, and language

–

framework, the PP (physiological = psychological) model for linguistics, is focused by Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru (in this volume). Academic disciplines concerned with art study high level achievements in the use of symbols and in the device of systems of literary, artistic, musical, architectural representations. In the arts, a level of consciousness and exploration of symbolic means is arrived at which allows for deep insights into the potential of symbolic systems. This field with its foregrounding of innovation and experimental designs allows a view of future possibilities. In this sense it goes beyond c* and explores future symbolic systems. The chapter by Martina Plümacher refers to the visual arts, whereas the text of Siegfried Wyler discusses the verbalization of colors in magazines (in texts which describe textiles and cosmetics).

. Conclusion The basic idea behind the conference “Sense and Sensibility” documented in this book is to bring together specialists from all of the above fields in order to be able to reconstruct the puzzle of human cognition and symbol use. We chose the aspects: olfaction and color, because they seem to be nearer to the basically continuous (chaotic or probabilistic) pole and nearer to level a* in the evolutionary hierarchy which I proposed in the last section. Implicitly, however, language and other human symbolic capacities are the perspective from which more fundamental aspects are looked at. The communication about colors and odors has to refer to such fundamental fields of research like physics, physiology, neurobiology and is of primeval relevance for cultural studies (ethnology, linguistics, art disciplines). It involves the evolutionary dynamics which produced the human species and the cultural dynamics which produced art and science.

References Austin, J. L. (1962/1975). How to Do Things with Words, J. O. Urmson & G. J. Warnock (Eds.) (2nd ed.). London: Oxford University Press. Ballmer, T. T., & Pinkal, M. (1983). Approaching Vagueness. Amsterdam: North-Holland. Berlin, B., & Kay, P. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley: Berkeley U.P. Bruno, G. (1991). On the Composition of Images, Signs and Ideas, D. Higgins (Ed.). New York: Willis, Locker & Owens. Buffon, G. L. L. (1749/1971). De l’homme, M. Buffon Duchet (Ed.). Paris: Maspero.



JB[v.20020404] Prn:26/06/2007; 10:15



F: CELCR802.tex / p.16 (982-1076)

Wolfgang Wildgen

Cangelosi, A., Greco, A., & Harnad, S. (2002). Symbol grounding and the symbolic theft hypothesis. In A. Cangelosi & D. Parisi (Eds.), Simulating the Evolution of Language (pp. 191–210). Leipzig: Springer. Cassirer, E. (1923–1929). Philosophie der symbolischen Formen, Vol. 1 to 3. Reedited in Ernst Cassirer. Gesammelte Werke [Hamburger Ausgabe], Vol. 11 to 13. Hamburg: Meiner, 2001/2002. Cassirer, E. (1946/1974). The Myth of the State. New Haven: Yale U.P. Condillac, E. de. (1746/1973). Essai sur l’origine des connaissances humaines. Auvers-sur-Oise: Editions Galilée. Condillac, E. de. (1754/1821). Traité des sensations. In Œuvres complètes, 1821–1822, Vol. III. Geneva: Slatkine Reprints, 1970. Fauconnier, G., & Turner, M. (2002). The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Herder, J. G. (1784/1896). Ideen zur Philosophie der Geschichte der Menschheit. In H. Kurz (Ed.), Herders Werke, Vol. 3. Leipzig: Bibliographisches Institut. Hockett, C. (1960). The origin of speech. Scientific American, 20 (03), 88–96. Kepler, J. (1596/1938). Prodomus Dissertationum cosmographicarum, continens mysterium cosmographicum, de admirabili proportione orbium calestium, de que causis caelorum numeri, magnitudinis, motuumque periodicorum genuinis & propriis, demonstratum, per quinque regularia corpora Geometrica. In M. Caspar (Ed., 1938), Johannes Kepler, Gesammelte Werke (pp. 1–80). München: Beck. Locke, J. (1690/1949). An Essay Concerning Human Understanding. New York: Dover Publications. Morford, J. P., & Goldin-Meadow, S. (2001). Time and again: Displaced reference in the communication of linguistic isolates. In G. Györi (Ed.), Language Evolution. Biological, Linguistic, and Philosophical Perspectives (pp. 173–197). Frankfurt: Lang. Wildgen, W. (1994). Process, Image, and Meaning. A Realistic Model of the Meanings of Sentences and Narrative Texts. Amsterdam: Benjamins. Wildgen, W. (1998). Das kosmische Gedächtnis. Kosmologie, Semiotik und Gedächtnistheorie im Werke von Giordano Bruno (1548–1600). Frankfurt: Lang. Wildgen, W. (2000). The history and future of field semantics. From Giordano Bruno to dynamic semantics. In L. Albertazzi (Ed.), Meaning and Cognition [Series: CERLC] (pp. 203–226). Amsterdam: Benjamins. Wildgen, W. (2002). Dynamical models of predication. STUF, 4, 403–420. Wildgen, W. (2003). Die Sprache – Cassirers Auseinandersetzung mit der zeitgenössischen Sprachwissenschaft und Sprachtheorie. In H. J. Sandkühler & D. Pätzold (Eds.), Kultur und Symbol. Eine Einführung in die Philosophie Ernst Cassirers (pp. 148–174). Stuttgart: Metzler. Wildgen, W. (2004). The Evolution of Human Language. Scenarios, Principles, and Cultural Dynamics [Series: Advances in Consciousness Research]. Amsterdam: Benjamins. Wittgenstein, L. (1984). Tractatus logico-philosophicus, Tagebücher 1914/1916, Philosophische Untersuchungen. Frankfurt: Suhrkamp.

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.1 (45-115)

How can language cope with color? Functional aspects of the nervous system Manfred Fahle A naïve observer does not see any problem with naming colors – don’t we generally all agree about the naming of colors? Certainly, a given distribution of wavelengths usually stipulates the same denomination in different individuals, such as red or green. (Actually, due to our limitation to three classes of receptors, a huge number of luminance distributions will all lead to exactly the same impression of red). I here argue that the strict relation between wavelength distribution and color name hides the fact that the actual brain states of different individuals when seeing the same red probably differ significantly, a fact masked by the convention learned during childhood to label certain wavelength distributions with a specific word.

.

Language and color: Is there a problem?

Language, for a naïve biologist such as myself, seems to be produced by specialized muscles under the control of certain (motor) areas of the cortex. The motor areas, in turn, are influenced by other cortical areas such as Broca’s and Wernicke’s, by the frontal cortex (believed to be involved in planning), and maybe indirectly, by sensory areas such as the visual cortex. Different wavelengths of the electromagnetic spectrum differentially influence the retina and retinal activations are transmitted to the visual cortex and analyzed there. So language labels particular cortical activation patterns occurring in the brain rather than objects. Social feedback ensures that a given wavelength distribution is labeled with a certain adjective or noun by all members of a given verbal community, if they are able to discriminate this distribution from other wavelengths that are required to evoke the expression of a different adjective or noun. (Not all members of the population will be necessarily able to make this distinction, as will be detailed below, and different cultures may use different classification schemes.) So for most members of the populace, language seems to perfectly cope with color: wavelength distributions will generally be associated with the commonly used verbal expression for that wavelength. Unlike emotional states such as anger

JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.2 (115-176)

Manfred Fahle

or joy – where we can never really be sure that what one person experiences when describing their state as “joyful” or “sad” corresponds to that which another person experiences when labeling their state as “joyful” or “sad”, the perception of color seems to offer a much closer match between a word and the resulting subjective experience since both are anchored at the same physical stimulus.

. Language and color: There is a problem! In the following, I will try to convince you that this impression of a close match between language and color is a too optimistic one, as one might suspect from the existence of patients showing anomalies of color perception. The outer world (if it exists) contains objects. Electro-magnetic waves coming, until very recently, mostly from stars such as the sun, ‘shine’ on these objects and are partly reflected, partly absorbed. The wavelength spectrum reaching the eye of an observer is determined by the product of the wavelength spectrum of the light source (stemming usually from sources such as the sun, stars, a fire, light bulbs or a flash) and the surface reflectance of the object. Humans perceive wavelengths between about 400 nm (blue) to 700 nm (red) (Figure 1a). This spectrum reaching the retina can be described very precisely by physical methods and physiologists and psychologists investigating color vision are able to predict very accurately what color-word will be associated with this wavelength-spectrum (Figures 1a and b). But at the retina, we leave the realm of physics and enter that of physiology. The electromagnetic waves hitting the outer segments of the four types of photoreceptors evoke graded electric potentials in these receptors. If the potential changes in these receptors reach a specific threshold, they evoke a so-called action potential via bipolar cells in the ganglion cells. This action potential is an all-or-nothing type of signal, not a graded one as in the photoreceptors. From the level of the ganglion cells onwards, the strength of illumination is coded in terms of numbers of action potentials per time unit. Hence, language does not refer directly to physical

450

500

550 Wavelength (nm)

600

650

Figure 1a. Spectrum of visible wavelengths, from blue (ca. 440 nm) over green (550 nm) to red (600 nm).

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.3 (176-201)

How can language cope with color? 2 white

spectral sensitivity (log. units)

1

red

0

–1 green –2 blue –3

–4 400

500 600 wavelength (nm)

700

Figure 1b. Absorption spectra of the three human cone types, and their superposition (‘white’) (after Fahle 2003b).

properties of objects, but to activation patterns in the sensory parts of the brain which may differ between observers as much as for states such as joy and anger. Even though color names are usually used to denote properties of the outer world – as opposed to “joy” and “anger” – they are still triggered by specific states of the brain, and hence in specific brain states such as hallucinations, we may experience a color without it being present in the outer world.

. A short description of some principles of color vision Humans generally ‘own’ four types of photoreceptors. The rods, specialized for low levels of luminance, serve our black-and-white night vision. The three types of cones, each with a different absorption preference, subserve color vision. So-called



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.4 (201-243)

Manfred Fahle

red cones are most strongly activated by wavelengths around 600 nm, ‘green’ cones prefer wavelengths around 550 nm, and ‘blue’ cones 450 nm (see Figure 1b). But already on the retinal level, the correspondence between the physical world and its neuronal representation is not equivocal. The absorption characteristics especially of the ‘red’ and ‘green’ cones overlap considerably. This is to say that a monochromatic light with a wavelength of 600 nm not only activates the ‘red’ cones, but also the ‘green’ cones. The probability that a given number of 600 nm photons will activate a single ‘red’ cone is just larger by about a half log unit; i.e., a factor of 3, than is the case for a ‘green’ cone. Hence, only by statistics, i.e., averaging over many photons, can the visual system discriminate between different colors. Hence the subjective impression of a ‘red’ color arises from the ratio of activation between the different types of ‘color’ receptors in our retina, and the color names, in the end, are defined by relative activation strength, such as 50% red cone activation, 30% green cone activation and 0.1% blue cone activation. But things are even more complex due to the fact that humans possess simultaneously with the ability to discriminate between colors, the ability to compensate for moderate changes in the composition of the illuminant. We achieve color constancy, i.e., perceive a banana as ‘yellow’, for example, over the course of a day, from the somewhat reddish light at sundown to the more bluish light at midday or the clearly yellow tinge of light bulbs. Color constancy is a remarkable achievement since it ensures that we perceive an object as relatively constant in color and correctly identify an object’s color, irrespective of the time of day and the corresponding changes in illuminant even if constancy is not perfect as everybody knows who has purchased colored garments in a shop under artificial lighting. And object recognition clearly is the main purpose of visual analysis, after all. We have to identify the objects in the world around us in order to survive.

. A short description of some principles of brain function Our brains achieve the analysis of the visual world, including object colors, with amazing speed and precision, in spite of their rather slow elementary electrical processes. Neurons in the brain have maximum discharge frequencies of around 1,000 Hz, while a modern computer performs 2,000,000,000 operations per second, being more than a million times faster. Humans, nevertheless, are faster in analyzing complex natural scenes than the best computers. Why? I believe that four principles of brain organization make possible this amazing performance. 1. Specialization: Different parts of the cortex are specialized for assessing certain aspects of sensory information, for example, color. Certain cortical areas are known to become more active when a subject has to deal with color discrimina-

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.5 (243-288)

How can language cope with color? Central visual pathways

Visual field and its projections Visual field

Left

Center Periphery: near the center

Right Optic nerve

Lateral geniculate body

A B

C D

Periphery

Optic chiasm Optic tract

dorsal A B D ventral C

Optic radiation Superior colliculi posterior Primary visual cortex (Area 17, V1) Calcarine sulcus

B D A C

Figure 2. Projection of the visual world onto the visual cortex; its distribution over different parts of the optic nerve and visual cortex, and the central visual pathways (after Fahle 2003b).

tions, and defects in roughly the same cortical areas may lead to acquired inability to perceive color (achromatopsia). At present, more than 40 different and separate cortical areas have been identified which all deal with different aspects of visual analysis. 2. Order: This principle is intrinsically related to specialization. For example, the primary visual cortex receives an ordered representation of the visual surround. Adjacent points in space are projected onto neighboring points of the retina and (mostly) represented by neighboring neurons of the (primary) visual cortex (Figure 2). This order, together with specialization, allows for a most beneficial feature, distributed parallel processing. 3. Parallel processing: The cortex simultaneously processes the information from as many as one million different positions in the visual field. This is possible by the more than one million fibers in the optic nerve, diverging to about 500 million neurons in the primary visual cortex. Hence, the primary visual cortex alone performs massive parallel processing of the visual input, with about 500 of its neurons dealing with every incoming fiber of the optic nerve and optic tract. In addition, due to the specialization of cortical areas, different types of analysis are performed



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.6 (288-338)

Manfred Fahle Detail analysis from focused attention

Master map

Feature maps Color

Orientation

Size

Distance

Visual field

Figure 3. Different aspects, such as color, orientation, size and distance of different objects are simultaneously analyzed by different cortical areas (and finally combined on the level of the “master-map” (after Treisman 1986).

simultaneously in different cortical areas, increasing the extent of parallel processing by another order of magnitude and allowing different features of stimuli, such as color, line orientation, size and distance to be computed simultaneously and partly independently. (The different features are subsequently combined e.g., in a “master-map” to form object representations; Figure 3). 4. Lateral (horizontal) integration between neighboring neurons as well as feedbackbased vertical integration between different levels of processing ensure an optimal processing of visual inputs in a complex, semi-hierarchical network with different levels being active simultaneously.

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.7 (338-403)

How can language cope with color?

. Methods to study the function of the brain As a rule, results presented by scientists should be judged not the least on the basis of the methods employed to gain these results. This is necessary to judge how plausible and robust the results presented may be. Five main methods allow researchers to gather insights into the mechanisms and principles underlying the function of the brain. The first one, though not officially allowed in the natural sciences, is introspection: trying to analyze how oneself is experiencing different situations, stimuli, and tasks. Of course, this method is prone to errors and self-delusions, but often powerful in giving hints as to which effects are worthwhile investigating quantitatively and objectively, e.g., by psychophysical experimentation. The second is to observe and quantify behavioral features, especially by psychophysical methods and to perform a type of black-box analysis, i.e., to find out how subjects detect and discriminate between stimuli. In this way, it is possible to infer functions inside the brain, a bit like ascertaining drop-down commands of a newly purchased computer program and noting their consequences on the monitor. This third method’s aim is to investigate how these functions are distributed over the cortex and how they deteriorate after defects in the brains of patients as caused, for example, by cortical infarctions. This neuropsychological method, allows one to correlate specific functions with defined parts of the cortex. The fourth method is to record the electrical phenomena ‘accompanying’ cortical activity either by means of single neuron recordings in animals or by sum-potential recordings through the intact scalp in humans or animals. This method allows for the clarification of, for example, the temporal sequence of activation of different cortical layers or of different cortical areas (cf. Figure 4). This fourth method reveals especially telling insights into the type of stimuli able to activate different parts of the cortex – indicating that the respective parts are particularly interested in this type of stimulus. The main advantage of this method is its high temporal resolution which is completed by the much higher spatial resolution (albeit at much lower spatial resolution) of the fifth method, recording local neuronal activity of the cortex by means of functional magnetic resonance imaging (fMRI) (see Figure 5). Those cortical areas analyzing a given stimulus increase their neuronal activity causing increased need of oxygen leading to increased levels of blood perfusion in this restricted part of the cortex. The increased blood perfusion can be detected with high spatial resolution in the brains of human observers by means of fMRI. Together, these methods have greatly enlarged our knowledge about the functioning of the (human) brain and have lead to a far better understanding of the basic principles of its function as outlined above.



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.8 (403-403)

Manfred Fahle

a)

Orbitofrontal (14) TS (2) TAa (2) TE1 (2) TE2 (2) IPa (2) PGa (2) TPo (2) IT (15) STS (11) Prefrontal (5) TEa (2) TEm (2) LIP (1) 7a (3) FEF (3) FEF (13) MST (7) MT (12) V2 (10) V1 (10) V1 (4) V1 (14) V1 (8) V1 (9)

0

20

40

80

60

100

120

140

160

180

200

220

Latency (msec) b) V1

V2

1

1

2/3

2/3

4B

4

4Cá 4Câ

5

5 6

6

P

M

Figure 4. Increase of activity, i.e., number of discharges of individual neurons is measured by means of single–cell recordings in animals. a) Time periods during which neurons in different cortical areas are active after presentation of a visual stimulus. b) Two different types of neurons, ‘P’ and ‘M’ enter the primary visual cortex, V1, in different layers, 2/3 and 4B, respectively, and are relayed to different layers of area V2 (after Bullier 2003).

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.9 (403-403)

How can language cope with color? a)

Unstimulated tissue Blood flow

Deoxyhemoglobin Oxyhemoglobin Stimulated tissue

b)

Figure 5. a) Increase of activity in cortical areas leads to an increased blood flow and hence more oxygen which can be detected by functional magnetic resonance imaging (fMRI) (after Saper, Iversen & Frackowiak 2000). b) Different levels of increase in cortical activity can be symbolized by different colors or shades of grey superimposed on the pattern of cortical sulci and gyri.



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.10 (403-456)

Manfred Fahle

. Single cells, neuronal assemblies, and behavior Natural science, in many cases, has the advantage of being able to test hypotheses and predictions. For example, neuroscience is based on the assumption that information is coded in the nervous system as trains of electrical impulses (‘spikes’) traveling within individual neurons and being eventually transmitted to other neurons. I referred to this basic assumption above when stating that words refer to patterns of electrical activity in the brain and only indirectly to the objects of the outer world that (may) have caused these patterns. So when I say “red” I refer to those particular electrical activities that happen in my brain and not necessarily to the apple on the table which may or may not be there. Let us look at the evidence available for this basic assumption. Several arguments support the theory of an electrico-chemical basis of brain function. First, activation of sensory organs by their appropriate stimuli leads to (sometimes dramatic) increases in the frequency of nerve impulses, or ‘spikes’, as revealed by single-neuron recordings in animals. Second, appropriate stimulation leads to changes in the sum potential that can be recorded as the so-called ElectroEncephalo-Gram (EEG) from the scalp of human subjects. Third, these sum potentials are a sure sign of life such that if no electrical activity can be recorded from the scalp, a comatose patient is considered to be ‘brain-dead’! But could this electrical activity be a by-product instead of directly causing the proper function of the brain? (A hard-working person may produce sweat, but this sweat is neither the basis nor the cause for the work to be done.) Again, several reasons argue against this suspicion. First, stimulating a peripheral nerve by strong magnetic fields through the intact skin, or by small needle-electrodes inserted in the skin produces a constriction of the muscles to which these nerves lead. Similarly, electrical stimulation of sensory nerves can induce rough sensory impressions. Second, the same is true for direct cortical stimulation. Both the transcranial stimulation, through the intact skull and direct electrical stimulation of cortex, for example during brain operations, lead to specific excitations, depending on the part of cortex stimulated. Transcranial magnetic stimulation (TMS) by strong magnetic fields induces short electrical currents in the cortical areas underneath the coil of the TMS-stimulator. Stimulation of the motor cortex, for example, leads to short activations of the muscle (groups) represented underneath the coil position, while stimulation of the visual cortex can produce small flashes of light (phosphenes) being perceived by the subject that are sometimes even colored. Penfield (1965) directly stimulated the parietal and temporal cortex of patients during brain surgery by injecting small currents. A proportion of his patients experienced some type of visual, auditory or bodily hallucination, and some even recalled of events past during this stimulation. Salzman, Britten and Newsome (1990) went one step further in monkeys by injecting small currents into the cortex by miniature electrodes, influencing ar-

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.11 (456-510)

How can language cope with color?

eas of the cortex less than 1 mm in diameter, in an area known to analyze visual motion. By selectively stimulating the cortex at a site where neurons preferentially detected one direction of motion, say to the left, and by combining visual stimulation with direct electrical stimulation of the cortex, these researchers were able to make a stimulus moving slightly to the right appear as if moving to the left. Hence, if electrical current was simultaneously injected, the monkey indicated ‘movement to the left’ while the stimulus was actually moving just noticeably to the right. Without any current, the monkey correctly indicated ‘movement to the right’. These experiments show, in my opinion, beyond reasonable doubt that indeed, the mechanisms underlying brain function rely on electrochemical reactions. These can be influenced by appropriate substances such as drugs, alcohol, and hallucinogens. Moreover, both the reports of patients and specific control experiments in monkeys demonstrate that the neuronal activation produced by external currents is not experienced as something forced upon the brain, but as a genuine sensory activation. Hence, by electrically stimulating the visual areas dealing with color perception, one may produce an impression of color leading the subject to experience a color without any corresponding object in the visual world. This is a somewhat extreme example for the argument put forward above, i.e., that language does not primarily deal with object properties of the outer world, but with states of the brain. (Luckily, evolution saw to it that these external brain states usually correspond well to some important properties of the outer world.) But stimulation experiments as well as disorders of perception, visual illusions and hallucinations all demonstrate that the correspondence between brain states and the physical world is not always as close as one usually assumes. We will come back to the implication of this fact shortly. Here, it suffices to assume that language relates directly to brain states and sensory impressions, as coded by neuronal activations and only indirectly to properties of objects in the outer world.

. Single cells versus cell assemblies A last point to be considered is whether objects of the outer world are coded by a few cortical neurons or by large neuronal assemblies with dozens or even hundreds, thousands, or millions of cells. This question, at present, cannot be answered equivocally and there are proponents for both points of view. However, it is clear that a colored object is not coded by just one cortical neuron – otherwise, what would happen if this neuron would die during the process of aging? On the other extreme, it is clear that the number of neurons dealing with each object is a small subset of all cortical neurons, one reason being the specialization of cortex. The cortex is not just one big, uniform general-purpose computing machine, quite opposite to the influential views of Lashley and his colleagues.



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.12 (510-574)

Manfred Fahle

A rough estimate of the number of neurons involved in the representation of objects comes from experiments injecting small currents into the cortex of experimental animals as described above. Obviously direct influence on a number of neurons in the order of hundreds rather than thousands out of the billions of cortical neurons suffices to change an important aspect of a perceived object, namely its direction of motion, at least close to discrimination threshold. Additional recordings of the spiking behavior of such neurons revealed an amazing property: given sufficient integration times, these neurons were as good, and sometimes better, in discriminating between different directions of motion than the animal was. In this experiment, an algorithm, i.e., a program performed by a computer, received the train of action potentials recorded from a single neuron or a few neurons during a presentation of a stimulus which moved either to the right or left. On the basis of this information about the activity of just one (specifically selected!) neuron, the program made a guess about which direction the stimulus had moved. Simultaneously, the monkey made a decision regarding this same question indicating the direction of stimulus motion by performing an eye movement in the corresponding direction. As mentioned above the program achieved higher percentages of correct responses based on the spike activity of some of the neurons than the behaving monkey did. Several reasons can explain this finding: the program integrated over a longer time, and the experimenter had chosen, on purpose, a neuron optimally suited for exactly this type of discrimination between exactly these two directions. Hence, the program had the advantage of this a-priori knowledge of the near-optimal detector for the task at hand, taking into account only the output of this neuron well-suited to decide the actual task, while the monkey might not know which neuron is suited best to solve the specific perceptual task. In any case, the result of this experiment shows clearly that information about objects of the visual world is coded, in a very precise way, even by single cortical neurons. Higher numbers of neurons might be required to discriminate between more closely related features, such as similar directions of motion which will stimulate any given neuron to a very similar amount; but obviously, individual cortical neurons can contain a relatively precise description of certain features of objects. Similar experiments probing the coding of color in the cortex would probably show that the discrimination between colors by single cortical neurons is similarly precise. But how exactly is color coded in the nervous system in the first place?

. Processing of color information in the nervous system We have been reminded above that the human retina usually contains three types of photoreceptors with peak sensitivities for different wavelengths (see Figure 1b).

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.13 (574-675)

How can language cope with color?

From the relative excitation of each of these receptor types, the brain very precisely calculates and reconstructs the (mean) wavelength underlying this excitation in the case of monochromatic stimulation, i.e., if a single wavelength is present. As a first step, the retinal ganglion cells show an antagonism between their receptive field center and its surround (Figure 6a). The receptive field of a ganglion cells is the small region of the outer world from which stimuli can influence the activation level of this cell. A ganglion cell may increase its spike activity following the presentation of say, a long-wavelength (‘red’) stimulus in its center. Typically, this neuron would decrease its activity following presentation of a medium-wavelength (‘green’) stimulus in the surround of its receptive field (Figure 6b). This is called a center-surround antagonism. Such a cell will be maximally activated if a red stimulus is presented at its center and no green wavelengths are present in its surround, hence for example by a red dot on a black surround. Other neurons have complementary center-surround antagonism, or react to the disappearance (off-center) rather than the appearances of a stimulus (on-center). A homogeneous white area that contains a broad range of wavelengths and hence stimulates all receptors, will not activate such a ganglion cell with antagonistic center-surround characteristics since activation from the center and inhibition from the surround will roughly cancel each other out. This antagonism between different stimulus features is a general principle of function in our nervous system, from the antagonism between bright versus dark parts of images (Mach-Bands and simultaneous contrast; cf. for example Figure 7a), to figure-ground reversals (Figure 7b), and to the dialectics of thesis versus antithesis. A similar antagonism exists between long wavelengths, exciting the blue cones, and wavelengths around 600 nm which excite both red and green cones, leading to the perception of yellow. These antagonisms accentuate color borders and are characteristic also for cortical neurons. A certain proportion of these cortical neurons, however, code colors not according to the physical wavelength stimulating the photoreceptors, but try to accommodate the illuminant thus achieving color constancy (see above). The center-surround antagonism, together with the fact that nerve cells usually adapt to a continuously presented stimulus, explains in a straight-forward way a well-known illusion consisting of a mismatch between physical stimulus on one side and perception and language on the other side. If we fixate for half a minute a large homogeneous stimulus and then change our view to a homogeneous white area, this area will appear as greenish for seconds or even tens of seconds. A somewhat simplified explanation is that while looking at the red stimulus, the red cones adapt as all photoreceptors do, that is, they respond less vigorously to the stimulus. (This is not usually a problem since normally we move our eyes to a new stimulus at least once per second.) The white area reflects equal amounts of long and medium wavelengths. But the receptors have been exposed to ‘red’ light for too long and hence have adapted to it. In consequence, the neuronal response to



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.14 (675-688)

Manfred Fahle A M cells

On-center

Off-center +

– –

+

– ++ +

–

– –

+

–

–

+ +

+

–

+

–

+

On-center

Off-center

B P cells –

– – – –

–

– R+ G– – –

G+

– –

– –

–

+

Y– –

– –

–

+ G+

+ + +

+

B– –

+ +

– B– – +

+ R+ +

+

+ +

+ G–

+

– + Y+ +

+ R–

+

– –

– + B+ +

R– –

–

– –

+

–

+ + Y+ +

+ + +

– Y– –

+ B+ +

+

Figure 6. Center-surround antagonism of ganglion cell receptive fields. (A) M-cells have larger receptive fields than P-cells. (B) On-center cells are stimulated by shining a light into the center of their receptive field, Off-center cells respond when the light is switched off. Light falling onto the periphery of an on-center cells inhibits this cell. Off-center cells have a corresponding antagonism, and P-cells usually show a color-antagonism, between red (R) and green (G) or else between blue (B) and yellow (Y).

the medium wavelengths is stronger than to the long wavelengths and though the wavelength distribution is actually balanced and would normally produce the impression of ‘white’, we perceive a greenish surface in the same shape as the red object previously observed. We will come back to other examples of non-correspondence between ‘objective’ and ‘subjective color’ below when dealing with visual illusions in a broader sense. Suffice it to say here that due to the characteristics of receptor physiology (adaptation) and neuronal circuitry already on the retinal level (center-surround color antagonism), our color perception might not correspond to object properties, and may vary at times. But before dealing in more detail with such discrepan-

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.15 (688-688)

How can language cope with color?

Figure 7a. Squares of identical intensity appear far darker on a bright surround than on a dark surround, due to simultaneous contrast (upper row). At the borders between rectangles of differing grey levels, darker or brighter halos may be perceived, the so-called Mach bands (middle row). A third phenomenon related to lateral inhibition is the scintillating Grid by Schrauf et al. (1995) (lower part).



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.16 (688-714)

Manfred Fahle

Figure 7b. Rivalry or antagonism between figure and ground, as exemplified by Rubin’s faces versus vase (from A. Seckel: The Art of Optical Illusions).

cies between the properties of objects in the world and our subjective impressions of these objects, let’s consider some surprising properties of human color vision.

. Peculiarities of human color perception and naming By now, the reader has grown accustomed to the idea that our perception of color does not necessarily correspond to the physical surface properties of objects. So it should not come as too much of a surprise that the periphery of the human visual field does not contain color-sensitive elements. This is to say that objects further away from the center of gaze than about halfway to the brink of the visual field are only detected by black-and-white detectors. This certainly does not correspond to our subjective impression, since peripheral objects seem to be colored as well as more central objects. Even more surprising, the central part of the visual field lacks receptors for the short wavelengths (blue) cones or may contain only very few of them. This is to say that we should be unable to see ‘blue’ all that clearly

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.17 (714-770)

How can language cope with color?

for objects projected upon this area. One reason is probably that due to chromatic aberrations the ‘blue’ part of the spectrum is focused at a different plane than the red/ green part, and hence it is impossible to focus both parts at the same time sharply on the retina. Since the red/ green part is more important, obviously evolution decided not to sample the blurry blue image at the center of gaze, at least not at any detail. Sounds like a good idea. But why do we fail to notice subjectively this lack of ‘blue’ wavelength information in the center of the visual field? (Part of the answer certainly relies on the fact that we seem to ‘fill’ in colors at least partly from the nearest contour). Subjective perception does not exactly correspond to the physical properties of physical objects. This fact has far reaching consequences for philosophy, and has implications also for semiotics. If nouns and adjectives are not directly linked to objects but instead to internal brain states and are only indirectly related to the objects proper – namely to the extent that these brain states correspond to the objects proper, then a deep divide separates the objects of the world on one side and our subjective representations of those objects on the other side. This loose correspondence between words and objects allows for the possibility that different persons might experience quite different internal states when describing the same object with exactly the same words! In this sense, color perception is not an objective phenomenon, as will be shown in more detail in the next section that deals with visual illusions and hallucinations. And indeed, physiological research during the last years provided evidence for the surprising fact that the relative frequencies or proportions of the long (‘red’) versus medium (‘green’) cones differ strongly between different subjects. A new optical method borrowed from astrophysics allows us to compensate for many imperfections of the human eye’s optics and thus to produce images of the living human retina with unprecedented resolution. Bleaching either the long or middle-wavelength cones and subsequently taking photographs of the retina allows us to identify the individual photoreceptors belonging to each of these classes. As is illustrated in Figure 8, the relative amount of these cones differs widely between subjects even though only very few individuals have been investigated so far due to the difficult and time-consuming procedure required. Even in these individuals, the ratio of ‘red’ versus ‘green’ cones varies between 1:1 and 4:1. This is to say that the same object, say of a brilliant red, will activate the cones of one individual who possesses a high number of ‘red’ cones significantly stronger than the retina of another individual with a lower number of ‘red’ cones. This larger retinal activation should lead to a different, and probably stronger activation of visual cortex in the subject possessing more ‘red’ cones. Though both individuals are perfectly normal and healthy, and though they both refer to the same object using the same word, their brain states, and possibly their subjective experience differ. Hence, in spite of the fact that reference to external objects is an advantage when



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.18 (770-859)

Manfred Fahle

assessing the inter-individual reliability of words and language in general, this is only a relative advantage. In a recent study using both electrophysiological (sum potential) and fMRI methods, we were able to demonstrate that the same stimuli activate the brains of different observers in significantly different ways (Fahle & Spang 2003). The differences between the subjective experiences of different observers can be even more pronounced when patients are included, or, in the case of color vision, in subjects with color deficiencies. Moreover, illusions teach us vividly about discrepancies between the physical world and our perceptions.

. Discrepancies between the physical world and its subjective experience, or perception A first example of a discrepancy between physical stimuli and our perception of these stimuli is the so-called unconscious priming. A stimulus presented first may influence the perception of a second stimulus presented shortly afterwards – even if the first stimulus is not consciously perceived. For example, presenting for 14 ms an arrow pointing either left or right and a mask after a variable interval between 14 and 84 ms will mask the direction of the arrow: observers are unable, even in a forced-choice task, to correctly identify the direction in which the arrow points. The mask, which is itself displayed for 140 ms, also contains an arrowlike leading edge pointing either to the left or right. Hence, there are two possible types of stimulus configuration: the first and second arrows point in the same direction, or else they point in opposite directions. As is obvious from Figure 9, observers achieve faster reaction times in responding to the second arrow (the ‘mask’) if directions of both stimuli correspond while reaction times are slower for contradicting directions. The difference in reaction times increases with the time difference between presentations of the first and second stimulus. It seems that the visual system analyses the direction of the first stimulus and prepares for a response without the owner of the brain being aware of this analysis and preparation taking place. If the second stimulus points in the same direction, the prepared reaction can be executed quickly. But, if the second stimulus points in the opposite direction, the prepared reaction has to be cancelled first before the correct reaction can be planned – reaction times are longer than without any priming at all. This subconscious priming demonstrates that we do not perceive (consciously) everything presented to our eyes (and analyzed partly at least on ‘lower’ cortical levels), but that these unconscious stimuli may nevertheless influence our behavior. Alas, this is true only for simple choice-reaction tasks, apparently not for more complex tasks such as buying ice cream, contrary to some earlier reports from the newspapers.

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.19 (859-859)

How can language cope with color?

A

B

C

Figure 8. Distribution of different types of cones in three different observers (light gray = long wavelength = ‘red’ cones; dark gray = middle wavelength = ‘green’ cones). The ratio between the cone types differs markedly between individuals (A-C) (after Roorda et al. 2001).



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.20 (859-890)

Manfred Fahle

700 ms

14 ms

RT (ms) 450 incongruent 400

0–70 ms

350 300

congruent

140 ms 250

14 28 42 56 70 84 SOA (ms)

Figure 9. A priming stimulus, in the form of an arrow, points either to the right or left. After a variable interstimulus interval SOA, a larger arrow is displayed for a longer time, masking the first arrow. Subjects are unable to indicate the direction of the first arrow. When asked to indicate the direction of the second arrow, they require longer when the directions of the arrows are incongruent than when they are congruent (after Vorberg et al. 2003).

Hallucinations are in some respects the opposite of unconscious priming. A hallucination is the emergence of a (subjective) conscious perception without a corresponding stimulus in the outer world. (In the priming experiments cited above, in reverse, the object exists, but no conscious perception results). Wellknown examples of hallucinations are the ones experienced during the aura of some epileptic seizures, often with the perception of vivid colors, or the ones accompanying ophthalmic migraine that are sometimes also associated with (faint) color impressions. In these cases, the nouns used to describe the subjective perceptions are not related at all to objects in the outer world, but exclusively to brain states and therefore remain most evidently private. The same is not true for visual illusions. They share with hallucinations the fact that the subjective perception does not correspond to the stimulus. However, this is true for (more or less) everybody at any given time making it possible for language to reach an inter-individual agreement about perception of the stimulus, unlike for hallucinations. A well known example was mentioned above: in afterimages, we perceive a color not present in the outer world. Other examples rely on the filling-in of colors from colored border-lines into neighboring areas, for example, the so-called neon-spreading.

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.21 (890-938)

How can language cope with color?

Figure 10. Benham’s Top. When the top rotates, neighboring receptors in the retina receive transitions from bright to dark at slightly differing times. These time differences are perceived as tinges of different colors, probably originating on the level of the retinal ganglion cells.

A most peculiar illusion relating to color are the flicker-induced colors produced by rotating Benham’s top. This top is available today as a (not only) children’s toy. It consists of circular dark lines on bright surround that are shifted, or rotated relative to each other (Figure 10). When the top rotates, the dark/ bright transitions at the start and end of the lines reach neighboring photoreceptors at slightly different times. These temporal phase differences in the order of one hundredth of a second are present at each rotation of the disc. The differences in the onset and offset times at neighboring retinal positions obviously create a color impression in the retinal network responsible for creating the antagonistically organized receptive field of ganglion cells (Figure 6). These colors are relatively unsaturated but nevertheless, clearly identifiable and observers have no problems associating words such as red, greenish, or yellow to the perceptions they experience for different types of onset- and offset asynchronies.



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.22 (938-981)

Manfred Fahle

Visual illusions are another indication for a general and basic principle of brain function. The information reaching the brain from the sense organs is underdetermined. For example, we receive two flat (two dimensional) views of the three dimensional world from our two eyes. From this information, the brain synthesizes a plausible and usually adequate representation of the three-dimensional outer world. But sometimes, the neuronal networks achieving this synthesis fail, or the synthesis is based on principles not applying to a special sensory situation. And these accidents of perception, the illusions, often allow important insights into the principles of brain function (cf. Fahle 2003a). Synesthesias, which are mentioned in several other articles of this book, represent a specific type of illusion – the association between representations from different sense modalities, such as vision and hearing. Our language indicates that such associations exist even in so-called normal subjects. Compare, for example, such expressions as a loud color or a fullbodied wine. In a small percentage of otherwise completely normal subjects, there seem to exist specific, orderly connections between cortical areas subserving different sense modalities, such as the auditory and visual domains. Hence, a given color will, in these subjects, activate not only the appropriate part of the visual cortex but moreover a similarly restricted part of the auditory cortex, thus producing a strong association between the representation of a color (as produced by a visual stimulus) and an association of a tone. This is to say that this cortical association is just that: an association. But the presentation of a specific color will, in this subject, produce a relatively vivid impression of a well defined tone which is, of course, not present in the external world. Another way to gain insights into the function of the normal brain is the study of patients with clearly defined disturbances of function such as disturbances of color vision. Perception of the visual world may fail in patients in a number of different ways. The most severe form is blindness, i.e., the complete failure of objects to impose an impression at all onto the visual system. But there are other, less severe forms of visual disorders, such as the inability to bind together the plethora of borders between different intensities and/or colors to form outlines and surfaces, a disorder called apperceptive agnosia. Or else the inability to associate the form presently perceived with a stored representation, leading to a failure to recognize (familiar) objects, a disorder called associative agnosia. Finally, some patients recognize visual objects but are unable to find their names. This disorder, quite related the to topic of this book, is called visual aphasia, or anomia. It clearly demonstrates that visual perception as such – and even analysis of objects as well as their purposes – is separated from language, and that only after analysis of visual scenes is more or less completed, words are connected to the results of this analysis. A completely different type of disorder, but even more closely related to the topic of this book, are deficiencies of color perception. The most severe, and extremely rare form, is cerebral achromatopsia, with patients experiencing the world

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.23 (981-1016)

How can language cope with color?

as shades of black and white. The quite common disorders (especially in males) of deuteranomalia and protanomalia refer to individuals with decreased numbers and/or function of the long- or middle- wavelength cones. These individuals have problems to discriminate between red and green colors, and those lacking the long wavelength (red) receptor are unable to perceive faint red lights clearly visible for normal subjects. In certain tasks, such as finding red berries among green leaves, these individuals are clearly handicapped but are otherwise perfectly normal. Clearly, the internal states corresponding to the words “red” and “green” in these individuals differ from those in normal observers. In their case, this difference becomes obvious since their naming sometimes disagrees with that of other people. The more than five percent of the male population suffering from more or less severe disorders of color perception are a good demonstration that language labels internal brain states and only indirectly labels features of the outer world such as object ‘colors’ (more precisely: object reflectances).

. Perception as synthesis of an internal representation, and its relation to language Let’s recapitulate how our knowledge about the brain helps us to understand the relationship between the perception of color on one side and language about color on the other side. Wavelength-distributions from objects of the outer world reach the retina and, at most central positions, activate three classes of cones to differing degrees. Activities of different cone-types are combined with one another locally, partly in antagonistic ways to form Gegenfarben (complementary colors). This mechanism of extracting information about ‘color’ from wavelength distributions can be fooled already on the retinal level, by appropriate timing of repetitive dark-bright transitions as present in Benham’s top. Moreover, different individuals (let alone patients) may code the same stimuli quite differently already on the retinal level due to differences in photoreceptor types and distributions, leading, with high probability, to differences in cortical activity patterns (see, e.g., Fahle & Spang 2003). In the brain, sensory information is coded by spike trains in (probably small) groups mainly of cortical neurons in the sensory cortices. Electrical stimulation from outside, as well as internally generated activity such as in epileptic seizures or migraine attacks, can lead to the subjective impression of sensory features such as color without any corresponding object in the outer world. This is partly due to the specialization of brain areas, combined with parallel processing. These properties of the brain ensure that stimulation of different parts of cortex leads to different subjective experiences, such as in different sense modalities, and of different types within each modality (such as a color, or a motion percept), and



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.24 (1016-1087)

Manfred Fahle

at different positions in the outer world, due to the orderly cortical representation of the outer world. Our subjective, personal experiences are based on cortical activation patterns. They represent the physical basis for our subjective experience and these patterns can be changed by relatively crude influences such as electrical or magnetical stimulation, but also by subtle ones such as a priming stimulus, without our awareness. Language labels these internal states rather than the objects of the outer world. This is true both in general and also specifically regarding color. Hence, language is perfectly able to cope with color, since “color” denotes a subjective percept, which itself is influenced by language, such as in the at least partly learned classification schemes of color. There may or may not, depending on circumstances, a considerable gap between our percept and the object in the outer world – but usually evolution has seen to it that this gap is not too wide. The gap is mirrored in language, in so far as we discriminate between colors (subjective) and wavelengths (objective) of the outer world. Even as a biologist, I have to admit that it is not just evolution that ensures that color naming is rather consistent between observers (apart from those subjects evolution did not prevent from developing deviating types of color perception such as missing one type of receptor). Language, to a large extent, can disguise physiological differences between observers by ‘normalizing’ the responses to given stimuli, irrespective of the differences in internal representations. Language, of course, is learned in a social context, and so the child learns to say the word “red” whenever it is asked about the “color” of an object sending out predominately long wavelengths – irrespective of the exact nature of his or her cortical activation, as long as this representation of a long-wavelength object is sufficiently different from that evoked by an object sending out medium wavelengths (see the article by Wildgen, this volume, defining language as an “ensemble of social rules”). And it is exactly this failure to discriminate between wavelengths evoking clearly discernible brain states in normal observers that sets individuals with deficits in color perception apart.

. Conclusions In conclusion, it has become clear that the information we receive about objects in the visual world is underdetermined and moreover sense organs seem to differ between individuals in more or less subtle ways. Therefore, all of us have to synthesize, in our brains, a representation of the outer world based on sensory information and using ‘reasonable’ assumptions about the structure of the world. A pure analysis, without assumptions filling in the gaps, is impossible. Together, the differences in the structure and composition of individual sense organs (even in

JB[v.20020404] Prn:11/04/2007; 13:55

F: CELCR803.tex / p.25 (1087-1140)

How can language cope with color?

so-called normals), the differences in the anatomical and physiological structure of the central nervous system, and the differences in the algorithms used by different individuals in constructing the cortical representation of the outer world will lead to clearly differing cortical activation patterns in different observers looking at the same scene even if we abstract from differences in mood, attentional state, arousal, fatigue, interest and similar effects. Hence, the same word, pronounced by different observers, in response to the same object may nevertheless correspond to widely differing cortical states, and my ‘red’ may be quite different from your ‘red’. Language, to a naïve biologist, seems to be a social convention to share information about the internal states of our brains and that through evolution and education (i.e., both nature and nurture) corresponds relatively well to the object denominated by the ‘fitting’ word. But in the end, language seems not to be more precise in conveying my subjective feelings about a color than it is in conveying my feelings about seemingly more internal states such as appetite or feelings even if the latter lack an ‘anchor’ in the outer world. So, like it or not, language is able to cope with color, but no better than with seemingly more private sensations such as emotions.

Acknowledgment Supported by German Research Council Grant (SFB 517). Dedicated to Werner Maihofer.

References Bullier, J. (2003). Cortical connections and functional interaction between visual cortical areas. In M. Fahle & M. Greenlee (Eds.), The Neuropsychology of Vision. Oxford: Oxford University Press. Fahle, M. (2003a). Visuelle Täuschungen. In O. Karnath & P. Their (Eds.), Neuropsychologie. Berlin: Springer. Fahle, M. (2003b). Gesichtssinn und Okulomotorik. In R. F. Schmidt & K. Unsicker (Eds.), Lehrbuch Vorklinik. Köln: Deutscher Ärzte Verlag. Fahle, M., & Spang, K. (2003). Heterogeneity of brain responses to identical stimuli. International Review of Sociology/ Revue Internationale de Sociologie, 13, 507–532. Penfield, W. (1965). Conditioning the uncommitted cortex for language learning. Brain, 88, 787–798. Roorda, A., Metha, A. B., Lennie, P., & Williams, D. R. (2001). Packing arrangements of the three cone classes in primate retina. Vision Research, 41, 1291–1306. Salzman, C. D., Britten, K. H., & Newsome, W. T. (1990). Cortical microstimulation influences perceptual judgments of motion direction. Nature, 346, 174–177.



JB[v.20020404] Prn:11/04/2007; 13:55



F: CELCR803.tex / p.26 (1140-1161)

Manfred Fahle

Saper, G. B., Iversen, S., & Frackowiak, R. (4 2000). Integration of sensory and motor function: The association areas of the cerebral cortex and the cognitive capabilities of the brain. In E. R. Kandel, J. H. Schwartz & T. M. Jessell (Eds.), Principles of Neural Science. New York: McGraw-Hill. Schrauf, M., Lingelbach, B., Lingelbach, E., & Wist, E. R. (1995). The Hermann grid and the scintillation effect. Perception (Supplement), 24, 88–89. Treisman, A. (1986). Features and objects in visual processing. Scientific American, 255, 106–115. Vorberg, D., Mattler, U., Heinecke, A., Schmidt, T., & Schwartzbeck, J. (2003). Different time courses for visual perception and action priming. Proceedings of the National Academy of Sciences of the United States of America, 100, 6275–6280.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.1 (45-146)

Color perception, color description and metaphor Martina Plümacher The contribution discusses prerequisites of linguistic representation of colors and color perception. In its first part, it points out strategies of color naming that aim at designating particular shades of colors as precisely as possible. The second part deals with descriptions of regular effects of interacting colors, for instance effects of artistic color compositions. The endeavor of painters to develop a technical language to speak about artistic and psychological effects of color compositions is analyzed. It is shown that speaking of particular shades of color and designating regular, inter-individual color effects requires established forms of categorizing them. Cultural techniques of illustrating colors and specific color effects are of crucial importance.

.

Introduction

. Colors without a name – color names without color information The world is colored for human beings. Although we can distinguish millions of colors visually, only very few of them have special names. It is to be remembered that linguistic designation of color is accompanied by a categorization of color, i.e., establishing classes of shades of colors. Each form of categorization is based on a particular principle of organization and thereby emphasizes specific aspects, socalled distinguishing marks, which are cognitively relevant.1 We do not designate every shade which is visually identifiable, but confine ourselves to names with a wide range of application – to names that refer to more or less restricted areas of colors. The individuality of a particular shade of color is not nameable. Of course, standardization in the production of colors makes it possible to provide a particular pigment or paint with a unique name, as is customary in var. Berlin and Kay’s research on the universality of focal colors (Berlin & Kay 1969) leads one to suppose that in addition to linguistic classification in connection with color naming there may be a pure perceptual categorization which might be prototypically organized around salient shades.

JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.2 (146-194)

Martina Plümacher

ious commercial contexts. In the paint industry, for instance, every paint has a name. Most of these color names refer – at least with one of their components – to a well-known object with a particular color, such as olive green, green turquoise or lemon yellow, or to the materials of pigments or dyes, such as vermilion, zinc white, cadmium yellow, cobalt blue, madder lake light rose, alizarin blue lake, etc. Several proper names are also used to define a specific shade, for example Van Dyck brown, Venetian red, Chinese orange, Mars yellow, etc. (cf. the catalogue of oils and oil pastel colors by Sennelier, in: boesner. Großhandel für Künstlerbedarf 1995: 148, 286f.). As Wyler shows (this volume and Wyler 1992), especially companies producing cosmetics prefer fancy names, even some that have no color component, for example night watch for a lipstick or spicy punch for a nail polish. These names are not informative regarding color, i.e., they do not provide addressees with a color image until they see the product. Color names with unusual modifiers as in cyber green2 or Mars yellow are irritating because addressees know that the name refers to a particular color – within the range of green or of yellow respectively – but ‘cyber’ or ‘Mars’ give no clue as to the shade. In commercial contexts and advertising, color names frequently have a poetic and suggestive function in addition to their referential function (cf. in this volume: Graumann and Wyler; Holz demonstrates the poetic and suggestive use of words in olfactory descriptions of perfumes). Wyler (this volume) emphasizes that the possibilities for inventing names are in no way limited. A standardly produced color, however, does not appear as the same shade in every situation. Color is a changeable property, and depends on light and illumination as well as on the color of the surrounding area. Normally, we do not notice that a color is changing due to a change in light, but perceive the color of an object as a constant property. In the psychology of perception, this phenomenon is called color constancy (Fahle, this volume; Pokorny, Shevell & Smith 1991; Stadler & Haynes 2001: 548) and is known as a special case of the cognitive ability of humans to perceive constant properties and relations of objects in spite of changes in the conditions of perception, as for instance the change of color in light or the change of position of a moving perceiver, according to which the form of objects perceived in perspective varies. Nevertheless, within a particular attitude color can also be perceived as an effect of light. We adopt this attitude, for instance, when watching the colorfulness of shadows. This has inspired artists to paint the multiple refraction of light, the variations in the color of an object and its shadows. The fact that the colors of the surrounding area affect the impression of a single color has also been recognized by artists for a long time. They term this phe. Cyber green denominates a specific green of a particular car model produced by the Volkswagen AG; see the contribution by Graumann in this volume.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.3 (194-232)

Color perception, color description and metaphor

nomenon the ‘interaction of colors’.3 For example, one color may appear darker or brighter and more brilliant depending on the colors surrounding it. One of the effects of interacting colors is the so-called simultaneous contrast of color. In this case, the perception of a color is influenced by the psychological projection of the complementary color.4 For instance, a gray figure tends to look greenish in a red environment. The perception of two neighboring regions, one red and the other yellow, involves the tendency for the red to become violet and for the yellow to become green (cf. Saint-Martin 1990: 44; Gage 2001: 191).5 With respect to the variation of color in the light and the psychological effects of interacting colors, artists emphasize the vividness of colors, their motion and their dynamics (Albers 1970: 12; Rompza 1998: note on October 3, 1988). Communication about a very specific shade of color is impossible as long as this shade is not present as the object of reference for each participant. Without there being a colored object of reference present, communication about color can only appeal to general knowledge of the linguistic designation of colors. This includes knowledge of patterns of descriptions of colors as well as of psychological color effects. . Strategies of naming colors and describing color impressions This article is about these patterns or strategies to circumscribe particular shades as well as impressions of colors that are due to an interaction of colors. It restricts itself to referential functions of color descriptions and excludes the broad dimension of various associations that colors and color names evoke in people – for instance associations with feelings, emotional appreciation, culture-dependent symbolic meanings of colors, etc. In the first part, it is argued that, in addition to the traditional strategy of designating particular shades by referring to the prototypical color of a well-known object, the development of the relational scheme of colors in the well-known color circle in particular has provided new principles to circumscribe shades. The relational ordering of colors enables one to differentiate between various shades linguistically, without referring to colored objects, and to designate particular features of colors. In other words, it frees people from having to refer to objects and to create specific names for particular shades. Even so, one . “Interaction of color” is also the title of a famous book by Josef Albers (1970), whose main subject is the particular effects of a colored environment on a single color. . See the contribution by Fahle on the neurophysiological basis of this phenomenon (this volume). . The French chemist Michel Eugène Chevreul was the first theorist of color to describe the observations of the law of simultaneous contrasts in greater detail.



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.4 (232-322)

Martina Plümacher

is not prevented from using very specific names as a starting point to speak about relations of colors and shades. In the second part, the article turns to the patterns of metaphorization used in descriptions of interacting colors, for instance in descriptions of color compositions in the fine arts. Talking about the interaction of colors is less standardized than color denotation. It will be shown that some of the customary patterns of speaking metaphorically about subjective impressions of colors were picked up by artists at the beginning of the 20th century in order to develop a technical language for describing abstract color compositions. For this purpose, metaphors had to be transformed into ‘technical terms’ by defining their range of application with respect to particular categorial segmentations of the color continuum and certain laws of interindividually perceivable psychological effects of interacting colors. The formerly effective word meaning – metaphorical analogies to a source domain or metaphorical synesthesia – was considered nothing else but a starting point to generate a new meaning, a meaning which was free of metaphorical imagery or synesthetic feelings and referred only to a sense of color. For our purpose, this development of ‘technical language’ is of some interest, because it shows a strategy to lexicalize wordings that had been irregularly used by individuals in order to describe their impressions of color compositions. Synesthetic metaphors are very common. For instance, colors are subdivided into warm and cold colors or provided with acoustic and tactile qualities, such as in the following expressions: loud red, soft blue, heavy dark green, etc. These expressions may have arisen based on real synesthetic processes. But as soon as expressions become frequent, their usage is going to be regulated by interindividualistic negotiations about reference and criteria of reference. Criteria of reference are more or less strong, the definition of reference being dependent on interests and practical necessities. The more an interindividual system of description can be established, the more idiosyncratic meanings diminish. Within the community of painters there has been great interest in defining classifications. Metaphorical expressions have not been rejected, but rather sorted and filled with non-metaphorical meaning.6 As Fahle (this volume) argues, individuals perceive colors differently. In addition, we have to take into consideration that many individuals who do not have the knowledge of specialists use wordings creatively in order to express their subjective . Shibuya, Nozawa and Kanamaru (this volume) suggest a model that explains the first occurrence of synesthetic metaphors and gives some reasons for the linguistic stabilization of these terms. However, meaning processes are inappropriately characterized as processes of associations, because they are primarily categorial processes. It is not frequency of coinciding entities that determines categorization and use of language but the importance of categorization for the organization of experience (which is not only bodily experience).

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.5 (322-355)

Color perception, color description and metaphor

experiences. Processes of the free application of words seem only to be restricted by the individual’s understanding of how linguistic categories mutually limit their range of application and her or his feeling of the need to be understood. Speaking about colors, especially about the psychological effects of interacting colors, cannot be completely free of subjectivity for several reasons. It is not just that individuals define the limits of the range of application of a color name differently. Sometimes disagreement about the proper color name is caused by the choice of different systems of distinctions between colors. The same reddish orange might be called red as well as orange or orange tinged with red – depending on the degree of gradation the person prefers at the very moment of speaking.7 According to a system that roughly differentiates colors around certain focal colors that work as prototypes, a color in the border zone of the linguistic categories ‘orange’ and ‘red’ can be named either way. Only with respect to a finer gradation does the shade have to be denoted more precisely. Sometimes, individuals are divided on the power of the psychological effects of colors, and this might not only be caused by the individuality of seeing colors. We also have to consider that individuals are trained differently in observing and describing the effects of interacting colors.

. Patterns of denoting colors Color denotation is so familiar to us that underlying patterns are hardly recognized. Only if we ask about these patterns do we notice them. In the European culture, the following four strategies of denoting shades of colors are especially common, although they are only used in particular contexts. In most contexts of ordinary, everyday life the colors of things are denoted by a short color name, if there is no special need to be more specific.8 Normally, a pullover is just referred . Dubois (this volume) shows that different levels of categorization also appear in descriptions of odors. . The monolexemic color names whose application is not restricted to a narrow class of objects are called ‘basic color terms’. Berlin and Kay (1969) define this term in order to compare the color concepts of various speech communities and to prove that there are certain universals despite all linguistic relativity. Critics have pointed out that there is no notice taken of how color terms are used by speakers in actual speech situations and various linguistic contexts. As Wyler (1992) and Dubois and Grinevald (2000) show, practices of naming colors differ in specific linguistic contexts and branches of industry, depending on the particular purposes and traditions of naming. The so-called basic color terms are not the most frequently used terms in each field of practice. Language users do not differentiate between basic color terms and other short and frequently used names like turquoise, beige, violet, rose, olive, auburn, etc. That is why Wyler considers it difficult from a linguistic point of view to follow Berlin and Kay in singling out the eleven basic color terms (cf. Wyler 1992: 53).



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.6 (355-433)

Martina Plümacher

to as red, a pair of jeans as blue, etc. Sometimes the color name is modified with the attributes light or dark. . Denotation of a hue in analogy to the typical color of a well-known object To be more precise a particular shade has to be circumscribed. For this purpose, it is not common to use the technical terms of the paint industry, which are well known only to painters – for instance vermilion, ultramarine, cerulean blue, chromium oxide green, etc. The frequently used strategy is to name a shade by pointing to the prototypical9 color of a well-known object. For example, shades are specified by calling the colored object lemon-colored or more briefly lemon, straw-colored, olive, violet, salmon, peach, pitch-black, jet-black, cornflower blue, grass-green, etc. More extensive descriptions of colors are also possible, for example a red like coagulated blood, the reddish brown of roof tiles, the blue of thunderclouds, etc. The advantage of this method of naming is a relatively high precision of describing a shade. An enormous number of very particular shades are nameable in this way. These object-based names are highly informative in regard to the particular shade, because they provide addressees with a relatively concrete color image provided that they know the typical color of the object in question. Graumann (this volume) suggests differentiating between complex color names that give relatively concrete information about the shade in question, one of whose components refers to the color of a well-known object, and complex color names, which do not mobilize such common general knowledge. Her example panther black vs. tarantella black shows the significance of the culture-dependent background knowledge for the efficiency of color names in communication. . Color naming with regard to dyes and pigments Before the revolution in the chemical production of colors in the 1850s and 60s, color names referred more often than today to the pigments or dyes used to produce the color (Gage 2001: 35; many examples are given in Schweppe 1993). Some of our present color names come from this tradition. Examples are the Latin purpura (purple), which was the name of the dye obtained from the secretion of a special kind of snail, or cochineal, a name that refers to the female scale insect that provides carmine. There are also names of pigments which are produced from specific kinds of earth, such as raw and burnt sienna, raw and burnt umber, ochre, and . Several examples show that color names refer to the prototypical color of an object. A lemon, for instance, is not always bright yellow – unripe lemons are green, old lemons that have dried up are a brownish yellow, moldy lemons are turquoise.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.7 (433-487)

Color perception, color description and metaphor

green earth. Vermilion is an example of a mineral pigment (German: Zinnober), while madder (German: Krapp) is the name of a vegetable dye. Several color names refer to the particular location of the production of the pigments or dyes. This strategy of naming is a kind of metonymy (on this issue see Niemeier, this volume). Indigo (Latin: indicum) for instance, one of the oldest organic dyes, is named with reference to its East Indian home. Indian yellow is also a traditional name that goes back to an unusual production technique in India. There, it was produced from the urine of cows that had been fed mango leaves (Schweppe 1993: 37). Within commercial contexts poetic names have been invented which refer to a characteristic color production of a particular geographic location. For instance, in the late Middle Ages the Flemish dyers termed their beautiful dark blue, produced from woad, manteaux de Frise. The deep-black produced in Flanders from the bark of a particular alder (named sticky alder tree or Schwarzerle in German, i.e., ‘black alder’) was called teinture aux racines (Schweppe 1993: 71). In all these cases, participants in communication have to know the typical color that is denoted by these names. . Relational ordering of colors in the color circle and color sphere It was not until the 17th century that the relational order of colors, which is so familiar to us today, was developed. At that time, scholars began to think about how to bring the variety of colors into a proper order. The oldest attempt at a systematization that led to the present color sphere was discovered in the Royal Library of Stockholm. In 1611, the priest and astronomer Aron Sigfrid Forsius presented the idea of a diagram of colors in a manuscript on perception (Silvestrini & Fischer 2002: 21ff.; Gage 2001: 166f.). Forsius had picked up the idea from classical antiquity that colors arise from the mixture of light and darkness, and should therefore be arranged spatially between white and black. But he was not satisfied with an arrangement in sequence. Like many astronomers, he considered the sphere to be the ideal form of orders found in nature. In his color sphere white and black are the poles; the central area of the sphere is coated with red, yellow, gray, green and blue, and between them and the poles are the lighter or darker hues. In contrast to the present color sphere, gray was set between yellow and green, and orange was placed a degree away from black. A precursor of the well-known color circle was a circle which was published by the English physician Robert Fludd in 1626. It represented only seven colors in a nowadays remarkable arrangement, in which black and white are neighbors, as are red and green opposite black and white. Besides these four colors, three other colors are presented: blue, between green and black, light yellow (flavus), next to white, and then an orange-yellow like saffron (croceus, the Latin name of saffron), between the light yellow and red. Inside the circle this order is explained: black and white are the poles of darkness and light – black is without light, white



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.8 (487-509)

Martina Plümacher

Figure 1. Color circle by Robert Fludd (1626) (from Gage 2001: 9)

without darkness; in green and red darkness and light are kept in balance; blue contains more darkness than light; in the yellow the whitish and the reddish are equally distributed; croceus is less whitish, and more reddish. Fludd’s idea of seven primary colors might go back to Aristotle’s De sensu et sensibilibus, where Aristotle mentions five colors besides black and white: yellow, crimson, purple, green and dark blue (Aristotle 1955: 442a; Ross 1955: 22, 203). Fludd picked up the idea of a circular arrangement of colors from a circle of colors of urine which had been published in an anonymous treatise in the 15th century (Gage 2001: 162, 171). As far as we know, this was the first pictorial diagram of a relational ordering of colors. During the 17th century, the idea of three primary colors – red, blue and yellow – became accepted (Gage 2001: 35f.). The observation made by artists that these colors are the only ones that cannot be produced by mixture was decisive. Green, orange, and violet – the so-called secondary colors – occur if two of these

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.9 (509-594)

Color perception, color description and metaphor

primary colors are mixed. Further shades, every intermediate stage of two given colors, can be produced by mixture. Different shades of gray, brown and even black occur if two complementary colors – namely green and red, yellow and violet, or blue and orange – are mixed. The result of these insights was that black and white were excluded from the circular arrangement of primary and secondary colors. But again around 1776, a new attempt to include black and white as paints by which colors can be darkened and lightened was presented by the entomologist Moses Harris. In his treatise Prismatic Colours he published a circle which represents the darkening of the colors towards the black center of the circle. This was a first step towards the well-known sphere of colors, which was eventually presented by the German artist Philipp Otto Runge. It reintegrates white and black as poles of the sphere. Today we distinguish between chromatic and achromatic colors. Physical research on the division of light into its colored spectrum strengthened the idea of a color continuum and a systematic order based on primary colors. Isaac Newton himself presented a color circle in his Opticks (1704) which represented the colors of the spectrum in sequence and in proportion. Quite a series of similar systematic orders of colors were published in the 18th and 19th centuries: besides the circle and the sphere there were star- and pyramidshaped diagrams. They differed especially in the shades that should represent the primary and secondary colors as well as in the segmentation of the color continuum. In 1788, the German painter Christoph Frisch, for instance, suggested a circle of eight colors, which included additional segments in comparison to the common circle of three primary and three secondary colors: a segment between violet and red and another one between blue and green, on the grounds that the human eye notices greater distances between blue and red and between blue and yellow (Gage 2001: 172). In the 19th century, Thomas Young’s thesis that the receptors of the eye are sensitive to red, blue and green light called common assumptions about the three primary colors, red, blue and yellow, into question. The debate about this issue finally led to the well-known distinction between additive and subtractive colors. The establishment of a relational order of colors was a great advance, even though some issues remained concerning the special shades of primary and complementary colors in the color circle as well as color harmony. The relational order of colors has made it easier to denote shades which have no well-known name, for it represents the knowledge of mixing colors, i.e., knowing how to produce the diversity of hues and their tonality, which works as background knowledge for communication – at least in societies in which this knowledge is regularly taught in school. On this basis, the following color denotations mobilize proper mental images or at least the ability of an addressee to define the proper range of application of these expressions by choosing suitable examples: a light, but yellowish



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.10 (594-614)

Martina Plümacher

Figure 2. Color circle by Moses Harris (around 1776) (from Gage 2001: 190)

white; an intense yellow, slightly tinged with green; a dark yellow toned down towards an olive; a light whitish yellow with a slightly brown tone, etc. In addition, the circle and the sphere of colors have also been taught as models of certain further characterizations of color: saturation, brightness, luminosity and complementarity (cf. Albers 1970; Itten 1970). Based on the knowledge of these features the possibilities for describing colors have increased. In painting, saturation designates the point at which further layers of paint do not change the color any more. Saying a color is not saturated means it is transparently applied. Saying it is not yet saturated means it would appear darker with a further layer of paint. So, a color might be described as dark for two reasons – because the paint is saturated or because it is toned with black. With a little knowledge of paint mixture one can differentiate between these cases and

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.11 (614-651)

Color perception, color description and metaphor

Figure 3. Color circle by Isaac Newton (1704) (from Gage 2001: 171)

describe a shade in question more precisely. Even a hue toned with black may be transparently applied. The degree of brightness means the degree of color intensity, i.e., the light which a single-colored surface seems to emit. It can be demonstrated with a black and white photograph, which shows colors with the same degree of brightness in the same shade of gray. However, avoid jumping to the conclusion that this could be an absolutely objective method! Even in the case of a black and white photograph the perception of a color is influenced by the interaction of other colors. For example, a physically identical gray appears lighter if it is seen lying on black stripes on a white background than if it is seen as an area behind these black stripes. There is no strong ‘objectivity’ in color perception if ‘objectivity’ means that a perceived color exactly corresponds to reflected light (cf. Fahle, this volume; Stadler & Haynes 2001). The term luminosity refers to the relative brightness that colors have in a colored environment and in different light (cf. Pokorny, Shevell & Smith 1991: 44; Saint-Martin 1990: 36ff.): a color seems to be more or less bright, depending on the neighboring colors. White, for instance, weakens the luminosity of colors, whereas black or dark brown tones in many cases increase it. The bright red in Rembrandt’s painting glows in the environment of dark shades, while if surrounded by yellow



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.12 (651-710)

Martina Plümacher

Figure 4. A physically identical gray appears different (Gekeler 2003: 48–49)

or white it would appear darker. Likewise, blue glows on a dark, black background and appears darker on a white ground (Itten 1970: 17, 42). The impression that a color gives also depends on illumination. In the daylight, people with normal vision consider a greenish yellow to be the lightest and a violet to be the darkest color in the spectrum; between these are in sequence orange, red, green and blue. But in the semi-darkness of the twilight the perception of red, green and blue changes: red seems to be darker and to have more depth than in the daylight, whereas blue and green shades become whitish and seem to come to the front. Painters have discovered that some colors – shades of green and blue, and even a bluish red – are more intense in candlelight than in daylight, whereas carmine and vermilion seem to be duller (Arnheim 2000: 330; Gage 2001: 192). This phenomenon was described in greater detail by the physiologist J. E. Purkinje in Beobachtungen und Versuche zur Physiologie der Sinne, II, 1825. Adjectives like bright, brilliant, powerful, intense, glowing or dull, muted, subdued, and deadened as modifiers of color names refer in many cases to the luminosity of a color relative to other colors. The complementarity of colors is defined today with regard to a physiological property of perception, namely the so-called after-image of a color. This phenomenon occurs if a person concentrates on a given color for about 30 seconds and then looks at a white surface: the fluorescent color which he or she then sees is the complementary one. For most people, the complementary color of a prototypical red is not a neutral green, but a bluish green; violet has a greenish yellow as its complement, a reddish violet has a yellowish green, dark blue has a neutral yellow, a neutral green corresponds to a slightly bluish red, etc. This definition of complementarity differs from the earlier idea that had become widespread among the artists of the 17th and 18th century, according to which the complement of each of

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.13 (710-755)

Color perception, color description and metaphor

the three primary colors is obtained by the mixture of the two other primary colors. Then, red would simply have the secondary color green as its complement.10 Several psychological effects of color perception are due to the complementarity of colors. For example, while one is looking at juxtaposed complementary colors their borders ‘blur’ and the colors ‘shimmer’ like colors in the heat (that is why this effect is called the Flimmereffekt in German). Knowing about this effect entitles the person to assert “blurring borders of colors”; this is no longer a merely individual impression. As a component part of knowledge, the systematic order of colors, shades, tints and tones represented in the color circle and similar diagrams is a cognitive model that enables us to speak about colors and shades without using color names that derive from objects. A large number of shades, or to be more precise, of different ranges of shades are linguistically identifiable with regard to the fixed relations of colors, their intermediate stages and features of colors like saturation, brightness, luminosity and complementarity. We can talk about degrees of tonality, for example about a more or less reddish or bluish violet, that is or should be more or less tinted with white or toned with black. A speaker may take a present object of reference as a starting point to explain the deviation of the color she has in mind: “Look here at this color, my color is slightly more tinged with red”. Speaking about deviations and transformations of colors is quite common in many contexts of choice of colors. Color diagrams represent relatively pure, clear, and brilliant shades (German: reine, klare, brilliante Farben), i.e., the brightest saturated examples of primary colors and their intermediates. Deviations, formed by the admixture of black or tertiary colors, for instance a yellow with a tinge of a grayish olive, are called dark and dirty. The attribute “dirty” in a dirty yellow, for example, refers to quite broad areas of different shades. The name is less informative about the particular shade, but not without meaning since in its usage it is related to a prototypical hue of yellow. To call a color dirty is to use an expression that sets an accent on deviation. Often it is simply used instead of a more extensive description of the shade in question. In case the admixture is slight, but as a result the luminosity is reduced, the color can be described as toned down or dingy11 (German: schmuddelig). In a German description it can also be called getrübt, gedämpft or gebrochen. These three words refer to the reduced luminosity through analogy to other domains: . In fact, Newton had already defined complementary colors. He had noticed that pairs of prismatic colors occur if two thin glass plates are pressed together. Every pair, if mixed, results in an uncolored light. The notion ‘complementary color’ probably occurred for the first time in 1794 in a treatise by the American physicist Benjamin Thomson, who had been inspired by Robert W. Darwin’s explanations of ‘after-images’ (cf. Gage 2001: 172). . I would like to thank the American art historian Ian Verstegen for pointing this out.



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.14 (755-811)

Martina Plümacher

water that is getrübt is not clear, light which is gedämpft is reduced and a beam of light can be gebrochen, i.e., it is refracted. To sum up, we can say that the relational ordering of colors, which represents the knowledge of how to produce different shades, has provided numerous new possibilities for speaking about colors without referring to the typical color of an object. In particular, it has provided the possibility of circumscribing the specific features of a given shade more exactly. . Hybrid forms: Color description with regard to the relational order of colors and to typical colors of objects Innumerable so-called tertiary colors have no specific name. Even the abovedescribed method provides only several rough approaches to a definition of an individual shade. An additional strategy of characterizing a particular tertiary color somewhat more precisely is a hybrid form of color naming: a color description with regard to the relational order of colors as well as to typical colors of objects. Examples are glowing coppery red brown, olive green with a tinge of gray, olive brown tinged with a yellowish orange, etc. Extended descriptions are possible, such as a greenish pale yellow, partly tinged with pink like a white peach. Here “white” in white peach does not denote a shade, but identifies an object – a more whitish peach in comparison to the normal peach. As Wyler says, the color becomes a distinguishing marker of objects in this case (Wyler 1992: 38–42). Some modifying adjectives of color names describe the shine of the color in question, such as in metallic red brown, which refers to a metallic shimmering; silky shimmering pale yellowish orange (German: seidig schimmerndes blasses Gelborange) points to the typically dull shine of silk. The denomination of colors also sometimes takes into account the material.12

. Descriptions of interacting colors Speaking about the effects of interacting colors, within a color composition for example, is less standardized than descriptions of single shades. Because of this, metaphors are frequently used. Artists who have studied color effects systematically and tried to verbalize their experiences have systematized metaphors and partly transformed them into technical forms. This has especially been the case since the second half of the 19th century, when many painters no longer felt ob. Dubois and Grinevald (2000) have also proved this in their study on the names of artists’ and decorative paints.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.15 (811-902)

Color perception, color description and metaphor

ligated to represent objects and their colors naturalistically, choosing instead to apply colors freely, with respect only to the internal color relations of a painting. The effects of color relations became the main point of interest and thus the need arose to speak about them. Subjective sensory impressions of colors and the feelings that colors elicit were studied. Parallel to this practical research by artists ran physiological studies on the influence of colors on our emotional life. These led – first in France, then in Germany as well – to the so-called Chromotherapie, a curative treatment of hysteria which was based on the use of colored light. Artists were more interested in the intersubjectivity of color effects than in psychologically observed reactions of individuals, because to understand the regularities of color effects meant to establish the basis of a non-representational art. They compared their insights into the regularities of color perception with the results of the psychology of perception, especially with those of the Gestalt theory, which developed at the beginning of the 20th century. Metaphors could be transformed into ‘technical terms’, because they were considered merely a mode of speaking, of verbalizing a particular impression of the interacting colors of a picture, and referred to ‘laws’ of color perception, to specific groupings of colors, i.e., general color concepts, or even to general rules of harmony of a color composition. The simultaneous contrast or ‘blurring borders’ of juxtaposed complementary colors, for example, are based on psychological laws of color perception, whereas the division of colors into ‘warm’ and ‘cold’ colors is a specific grouping, a sort of general categorization. The description of juxtaposed complementary colors as “shimmering” colors like colors “in the heat” is a metaphor supported by the analogy to the specific perception of colors in a very sunny atmosphere. But as soon as one knows that shimmering is an effect of directly juxtaposed complementary colors, this analogy is no longer essential for the meaning of the term “shimmering” in this context. Many metaphors inherent in our general language only have an indirect relation to an alleged original domain, as Lakoff and Johnson show in their Metaphors We Live By (1980). In many cases, metaphors in art descriptions are used to designate particular phenomena of color and color perception. The more they are used not only to define the phenomenon of the target domain in concepts of a source domain, but also to set general categorial distinctions between phenomena of the so-called target domain, the more they gain a meaning of their own, i.e., they receive an original meaning, although an analogy to another domain persists. For example, painters have borrowed vocabulary from music composition, like composition, rhythm, tone, accord of colors, and color harmony, but they read a meaning into it which has been defined by their own, purely pictorial experiences. Many art theorists among the painters, such as Kandinsky, Klee, Itten and Albers, who were teachers at the ‘Bauhaus School’, used metaphors in order to



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.16 (902-939)

Martina Plümacher

develop a ‘technical language’ which enabled them to speak about quite regular, interindividual color effects. For this reason they had to determine the range of application of metaphorical expressions. It is a striking phenomenon that many of the metaphorically used adverbs which are added to color names accentuate opposing aspects of color effects. If these linguistic expressions reflect structures of perception, then there is reason to assume that the perception of color compositions is oriented towards diametrically opposed values in order to grasp the whole of different effects that a complex abstract color composition produces. Like the psychologists of the Gestalt theory, artists assumed that perception is not a passive reception of perceptible data, but an active structuring. So they picked up several metaphorical expressions and linked them with concepts which they needed in order to reflect structures of their color compositions: concepts of different types of contrasts of color, of spatial effects of interacting colors, of balance and symmetry of a color composition, etc. These partly refer to Gestalt laws of perception, for instance the concept of spatial relations of colors refers to the perceptual division of foreground and background. The concepts of contrasts are linked with psychological insights into the relevance of contrasts to the process of perceptual structuring, or – as a complement to this – with the Gestalt law of similarity, i.e., the general perceptual tendency to combine the similar into a group of connected elements. As the following will show, artists pointed out some oppositions that are made by common metaphorical descriptions of interacting colors and filled them with non-metaphorical meaning. . Opposition: warm – cold Descriptions of colors of a picture or a room that characterize the colors as warm or cold have been quite frequent for a long time. The designations warm red and cool or cold blue (German: warmes Rot; kühles or kaltes Blau) are prototypical examples, because the differentiation between ‘cold’ and ‘warm’ colors indicates – so to speak in a non-metaphorical manner – a division of colors into a yellow-red and a green-blue tonality. A phrase from a description of a painting by the art historian Lorenz Dittmann (1987: 223) shows that the opposition between warm and cold colors is not only used to differentiate the yellow-red from the green-blue range of colors, but also to mark relative tensions: “from a brown ground emerges the warm leaf-green of her jacket” (German: “aus braunem Grunde entsteht das warme Laubgrün ihrer Jacke”). A ‘warm’ green is a green with a yellow tint that is accentuated by other ‘colder’ colors surrounding it, or a neutral green glowing in the environment of really dark brown or black shades. Painters, such as Ernst Ludwig Kirchner and Johannes Itten, pointed out that the distinction between warm and cold colors is not a determination of absolute properties of colors but a description of the relative effects of interacting colors (cf. Dittmann 1987: 356; Itten 1970: 45).

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.17 (939-993)

Color perception, color description and metaphor

They did not refer to really synesthetic feelings, which some observers may have, but to relations between colors along the division of a yellow-red and a green-blue range of colors. A color that context-independently belongs to one of these two groups may in a particular context show a tendency towards the opposite group if this context stresses a subtle nuance of shade that combines it with this group. In case of a neutral green on a black ground, its luminosity contributes to its relation to ‘warm’ colors. . Opposition: active – passive Artists, such as Edvard Munch and Wassily Kandinsky, distinguished between active, i.e., intense colors of high luminosity, and passive colors (cf. Dittmann 1987: 322, 394f.; Itten 1970: 17). Several metaphorical expressions fit this opposition, for example13 exciting red, loud red, garish red, aggressive red; soothing blue, pale green, dull blue (German: erregendes Rot, aufdringliches Rot, aggressives Rot; beruhigendes Blau, fahles Grün, mattes Blau). Outside the context of specialists in the fine arts, these metaphors often refer to subjective feelings evoked by colors. Linked with the distinction between ‘active’ and ‘passive’ colors, they are divided into two groups of expressions that can be attributed to only one or the other group of colors. Then, they lose their connection to feelings. Only pure, brilliant, saturated, and therefore salient colors which seem to come to the fore can be called active and in this sense intense, powerful or strong. In a non-metaphorical way of speaking one can say that the metaphor of ‘active’ colors accentuates the prominence of a pure color with regard to its luminosity within a composition. One never finds expressions that ascribe activity to a toned down color. ‘Activity’ also corresponds to a certain degree to spatial effects. The general rules of spatial effects are described by Itten as follows: If the six colors yellow, orange, red, violet, blue and green are lined up with no spaces between on a black background, you clearly see that the bright yellow seems to come to the front, and that the violet floats in the depths of the black ground. All the other colors form intermediate stages in depth between yellow and violet. If a white background is used, the depth effects change. Violet is repelled by the white ground and seems to come to the front, whereas the white holds back the yellow as its ‘light relative’. (Itten 1970: 77; my translation, MP)

Hence, colors never appear as merely juxtaposed areas, but rather seem to have different degrees of depth and seem to lie one on top of the other. Brilliant red . Many examples of metaphorical descriptions of colors can be found in Dittmann (1987) and Wyler (1992). I would like to thank the linguist Kirsten Brock for assistance with the English translations.



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.18 (993-1057)

Martina Plümacher

and yellow are special colors because it is almost impossible to bring them into a spatial background position within a picture. . Opposition: soft – strong / pale – intense The distinction between soft and strong or pale and intense colors may seem very similar to the above-mentioned distinction between active and passive colors, but it actually refers to different features of colors. A color is soft or pale or delicate if it is transparently applied or tinted with white. The white paint increases the brightness of the hue, but it also tones the hue. Only a color that is not tinted with white, but is saturated and of some prominence within a composition can be called strong, hard, intense or powerful (cf. Albers 1970: 97; Itten 1970: 17, 34, 36ff., 55, 58). It is ‘strong’ or ‘hard’ with respect to its power to create chromatic contrasts. The theory of colors in the context of art distinguishes between two types of contrasts of color: the chromatic and the tonal contrasts of color. A chromatic contrast develops by juxtaposing pure colors, such as brilliant red, blue, green, etc. A tonal contrast is a contrast of colors of different degrees of brightness, for instance light and dark blue. Modifiers, such as intense or powerful, refer not only to the contrast producing feature of a color, but in addition to its luminosity and ‘activity’ in the above-described manner. The following expressions are examples of describing colors within the framework of the opposition between ‘soft’ and ‘strong’ or ‘pale’ and ‘intense’: a soft pale yellow, a subtle blue, a delicate pink, a gentle, delicately tinted white; a strong green, an intense orange, a fiery orange, a glowing orange, a blazing red, a powerful violet, a shrill red, a hard black, a harsh black (German: ein zartes blasses Gelb, ein weiches Hellgelb, ein sanftes Rosa, ein mildes zart gebrochenes Weiß; ein kräftiges Grün, ein scharfes Grün, ein intensives Orange, ein feuriges Orange, ein glühendes Orange, Feuerrot, Glutrot, ein mächtiges Rotblau, ein schrilles Rot, ein hartes Schwarz). . Opposition: heavy – light According to artists, colors with a low degree of brightness and some spatial extent within a composition have weight. It is possible to describe the pictorial relation between a pink and a green as a contrast of ‘light’ and ‘heavy’ colors: a delicate pale pink against the heavy dark green of the ground (German: ein zartes Rosa gegen das schwere, dunkle Grün des Grundes) (cf. Dittmann 1987: 294). Paul Klee tried to explain this metaphor by showing how balance, i.e., mainly symmetry, of forms and colors within a composition can be produced or destroyed. In this sense, a large, dark colored form can be called heavy in relation to smaller, brightly colored forms, because several smaller and brighter forms are needed as its counterweight in order to bring the composition into balance (cf. Klee 1991: 120–133). Also, a

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.19 (1057-1138)

Color perception, color description and metaphor

broad area of bright color might appear ‘heavy’ relative to small colored forms. Like the oppositions described above, ‘heaviness’ and ‘lightness’ are also relative qualities of interacting colors (Klee 1991: 133). . Opposition: deep – flat Sometimes it makes sense to qualify colors as deep or flat if one wants to point out the spatial effect of a particular color or a group of colors. We normally perceive a spatial structure when looking at an abstract color composition. However, colors of almost the same degree of brightness all appear flat relative to one another. All the above-mentioned attributes do not refer to properties of single shades, but to the features that a shade or a group of shades gains from its colored environment. The words indicate several aspects of salience by contrasts: the relationship to the warm or green spectrum of colors, the feature of coming to the fore of the pictorial space combined with a remarkable degree of luminosity, the feature of increased brightness by tinting, the feature of being the cause of chromatic contrasts, the feature of spatial dominance, and the relationship of colors in a decreased contrast of tonality. . The metaphor of the ‘rhythm of colors’ Analogies between the fine arts and music go back to classical antiquity. Especially since the end of the 19th century, it has become a frequent practice among painters to speak about color compositions in analogy to compositions of music (cf. Gage 2001: 227–246; many examples in: Hess 1981; Dittmann 1987). In many of his works and in his lectures at the Bauhaus school, Paul Klee analyzed the phenomenon which he and other painters called the rhythm of colors and forms. He described it by pointing to the repetition of similar figures, which the perceiver combines into a whole group – either immediately at the first glance at the picture or step by step while observing the pictorial structure. According to Klee, the spatial distances can be set in analogy to time in music. Since according to Gestalt psychology, perception is always a process of grouping forms, Klee could claim that the Gestalt of the reiterated similar forms is comparable to a melody in music (Klee 1991: 132). He exemplified the ‘rhythmic’ effect of reiteration in many pictures. One of the best and most complex examples is his Camel in rhythmic landscape. Indeed, similar forms or colors are perceived as connected; they form a Gestalt. One of the Gestalt laws of perception says that similarity is an important factor in



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.20 (1138-1146)

Martina Plümacher

Figure 5. Camel in rhythmic landscape (1920) by Paul Klee (from Herzogenrath et al. 2004: 41)

binding figures and colors visually (cf. Stadler & Haynes 2001: 545).14 Not only similar forms or colors provoke a linkage, but also an equal degree of brightness makes colors similar, and perception connects them by structuring the visible. Itten says: “The same brightness or the same darkness creates a relationship between colors. Colors are linked to each other and combined by the same values of tone” (Itten 1970: 41; my translation, MP; cf. Albers 1970: 104; Metzger 1975: 93). A clear arrangement of intense contrasts of light and dark colors forms a pictorial structure which is easy to grasp, because the lighter areas on the one side and the darker ones on the other side are connected to larger units. Contrasts of form, however, can loosen these bindings of color – a deviant figure, for instance, is salient. Different Gestalt laws compete within complex color compositions. Since perception always focuses on particular aspects of similarity, a deliberate regrouping, i.e., a change of accents, is possible. Especially the complex configura. Stadler and Haynes (2001) describe perceptual representation as a continuous generation of patterns, i.e., as a process which includes ‘interpretation’ of perceptual data.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.21 (1146-1205)

Color perception, color description and metaphor

tions in paintings offer different possibilities for visual groupings of colors and forms. Itten calls the perceptual concentration of a particular group of colors, forms or textures a “simultaneous form” (Simultanform): The function of seeing is such that one is inclined to connect similar things and to see them together. These similarities may consist of very similar colors, very similarly sized patches [of color], very similar degrees of darkness, very similar textures or accents. During the process of looking [at the picture] a visual ‘reference figure’ emerges. This figure is called ‘simultaneous form’, because it manifests itself through the relationship of the given similarities, without being materially existent. Simultaneous forms may arise in the case of two similarly sized patches of different colors. In contrast, the eye is inclined to see very similar colors together, so that in the case of multiple colors several simultaneous forms may emerge. The effect of the composition depends on the character of the simultaneous forms, their directions and the distance between them. (Itten 1970: 92; my translation, MP)

In the case of Klee’s picture Camel in rhythmic landscape, ‘rhythm’ occurs in several respects: for one thing, there is the repetition of similar forms, colors and directions which invites the eye to jump from one similar element to the next, for instance from one dark green or blue spot to the other. This visual grasping of forms one after the other is one type of ‘rhythm’. Then, there are different possible simultaneous forms – perceptual groupings of the yellowish or the reddish-brown spots, the dark green and blue spots, the vertical lines, the horizontal rows or the reddish or yellowish or greenish or bluish toned areas of the background. These changes of simultaneous forms are the second type of ‘rhythm’.

. Conclusions Language sets landmarks for an intersubjective orientation according to structures of perception as well as to mutual knowledge. This is possible because linguistic expressions accentuate particular relations, and focus attention on these relations. Object-based color names, for instance, refer to relations between objects and their prototypical color. Color names with regard to dyes and pigments accentuate the relations between shades and essentials of their production. Names that refer to particular locations normally stress the relation between a particular type of color and the location of its production or the origin of its color essentials. Sometimes they refer to specific colored clothes or other things that are connected with a specific location. The relational order of colors that developed in the 17th century in Europe is a cognitive model by which single shades can be described primarily through their relations to pure, brilliant colors, the so-called primary and secondary colors, and to black and white. This order of colors increases the



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.22 (1205-1277)

Martina Plümacher

possibilities for circumscribing a shade, but even in that case, linguistic expressions are only rough approaches to a definition of an individual shade. Language cannot express the finer perceptible distinctions of color. Therefore, it needs to be complemented by other semiotic systems like diagrams, pictures, catalogues, etc. Actually, language is applied to these other systems as well as to practices of differentiation between phenomena (cf. Heeschen, this volume). The practice of producing colors and color compositions has contributed to a refinement of perceptible distinctions. As shown above, the relational order of colors, together with distinctions between particular features of colors, namely brightness, luminosity, saturation, complementarity, etc., developed in connection with the practice of painting and interest in systematizing both the wide variety of colors and interindividual experiences of color perception. Scientists and artists contributed to color theories that to a certain degree complement each other. On the whole, they increased the knowledge of colors in several contexts. This has had an effect on conceptual and semantic distinctions. Even diverse metaphors that had been used to describe subtle effects of interacting colors were grouped and used to refer to a few relations of interacting colors by artists and art theorists. In the context of theoretical background knowledge, metaphors lose their metaphorical power. Within contexts of experts of art, they change in part into ‘technical terms’ that refer to a common knowledge of relations of colors and several laws of color perception. It is no longer necessary to think about analogies and a transfer of structures from a source domain to a target domain, because there is – at least for artists and art theorists – a developed conceptual structure to which metaphorical expressions have to apply. Looking at the results of investigations into the verbalization of odor perception, I am quite sure that strategies of verbalization of odor perception and those of color perception have a lot in common: like colors, odors are denoted with reference to a prototypical property of a well-known object. Sometimes odors are linked with an event, like burnt milk, freshly cut wood, etc. (cf. Dubois; Holz; Zucco, in this volume). In these cases the expressions describe a specific feature of an object under particular conditions. As Zucco (this volume) says, people sometimes describe a specific odor by mentioning a specific location and episode in their life. This strategy is quite comparable with the method of naming colors by referring to the location of their production or the origin of the color essentials. There is a sense in which even the hedonistic distinction between good and bad or pleasant and unpleasant odors can be considered as comparable to the distinction between ‘warm’ and ‘cold’ colors. In both cases it is a specific type of classification and therefore categorization, and both distinctions, one at the olfactory level, the other at the visual one, divide the whole range of phenomena into two classes. Both are conceptual strategies that reduce the variety of phenomena to two values of personal and interindividual relevance.

JB[v.20020404] Prn:26/01/2007; 15:17

F: CELCR804.tex / p.23 (1277-1324)

Color perception, color description and metaphor

Wildgen (this volume) argues that communication on specific odor perceptions is unstable and unsure as long as there is no demand for a society to establish a common system of labeling that goes beyond the method of naming an odor by referring to a proper object. The history of systematization of colors and features of colors shows that even in the sphere of visual phenomena, which is asserted to give closer clues to linguistic categorization, the variety of phenomena had to be ordered. Ordering is a process of standardizing in which several elements are marked as prototypes of a broad conceptual group; through this process characteristic features are highlighted, which then work as values of scaling. The system of classification of wines is another example that shows the possibility of standardizing several features of wines, e.g., the odor and taste of wines, with the result that experts can communicate about their odor and taste perceptions. The judgment of an expert on wine – for instance that a particular wine has an aroma of raspberry – cannot be checked by a non-expert. She or he might agree with the expert or not. As long as non-experts do not share the expert’s system of evaluation they have to rely on their momentary, unstable impressions and associations as the basis of their judgment. There really is a gap between experts and non-experts. The question of whether naming and denotation of sensations could be grounded in perception cannot be examined by referring to abilities of nonexperts. With the historical development of color descriptions in mind it seems to me that the development of a common language on sensual qualities depends on whether there are starting points for an interindividual and systematic form of grouping of qualities, i.e., categorization, on the one hand, and of specification of the transformation of qualities under specific conditions on the other hand. The ability to imagine a particular sensual quality does not seem to be decisive in this respect.15 Even the mental images of colors – at least of non-experts – are often unreliable, for they are vague and inexact. Nevertheless, everybody who is educated in the system of describing colors that has been developed in the fine arts can systematize and refine his or her experiences of color perception by recognizing the learned distinctions and using the methods of color description while looking at colors.

Acknowledgment I wish to thank Kirsten Brock for correcting my English. . Zucco (this volume) states that the impossibility of forming conscious mental representations of particular odors not only hinders communication – which I think is true – but also works against a systematic language on odors.



JB[v.20020404] Prn:26/01/2007; 15:17



F: CELCR804.tex / p.24 (1324-1440)

Martina Plümacher

References Albers, J. (1970). Interaction of Color. Grundlegung einer Didaktik des Sehens (1963). Nürnberg and Köln: DuMont. Aristotle. (1955). De Sensu et Sensibilibus. In D. Ross (Ed.), Aristotle. Parva Naturalia. Oxford: Clarendon Press. Arnheim, R. (2000). Kunst und Sehen. Eine Psychologie des schöpferischen Auges (1974), translated by H. Herrmann. Berlin and New York: Walter de Gruyter. (Original: Art and Visual Perception. A Psychology of the Creative Eye. Expanded and revised edition 1974). Berlin, B., & Kay, P. (1969). Basic Color Terms – Their Universality and Evolution. Berkeley: University of California Press. Dittmann, L. (1987). Farbgestaltung und Farbtheorie in der abendländischen Malerei. Eine Einführung. Darmstadt: Wissenschaftliche Buchgesellschaft. Dubois, D., & Grinevald, C. (2000). Denomination of colors in practices. In A. Lacus, K. Mely, & A. R. Lommel (Eds.), Lacus Forum XXVI The Lexicon (pp. 143–151). Gage, J. (2001). Kulturgeschichte der Farbe. Von der Antike bis zur Gegenwart, translated by M. Moses & B. Opstelten. Leipzig: E. A. Seemann. (Original: Colour and Culture, 1993). Gekeler, H. (2003). Handbuch der Farbe. Systematik, Ästhetik, Praxis (1 2000). Köln: Dumont. Herzogenrath, W., Buschhoff, A., & Vowinckel, A. (Eds.). (2004). Paul Klee – Lehrer am Bauhaus [Exhibition catalogue, Kunsthalle Bremen]. Bremen: H.M. Hauschild. Hess, W. (1981). Das Problem der Farbe in den Selbstzeugnissen der Maler von Cézanne bis Mondrian (1953). Mittenwald: Mäander Kunstverlag. Itten, J. (1970). Kunst der Farbe. Subjektives Erleben und objektives Erkennen als Wege zur Kunst (1961) [Studienausgabe], A. Itten (Ed.). Ravensburg: Otto Meier. Klee, P. (1991). Beiträge zur bildnerischen Formlehre (Vorträge im Wintersemester 1921/22). In G. Regel (Ed.), Paul Klee, Kunst-Lehre. Aufsätze, Vorträge, Rezensionen und Beiträge zur bildnerischen Formlehre (pp. 91–313). Leipzig: Reclam. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press. Metzger, W. (3 1975). Gesetze des Sehens. Frankfurt am Main: Kramer. Pokorny, J., Shevell, S. K., & Smith, V. C. (1991). Colour appearance and colour constancy. In P. Gouras (Ed.), The Perception of Colour. (Vision and Visual Dysfunction, Vol. 6) (pp. 43–61). London: The Macmillan Press. Rompza, S. (1998). arbeitsnotizen 1980–1998. Mannheim and Ladenburg: märz galerien. Ross, D. (1955). Introduction / Commentary. In D. Ross (Ed.), Aristotle. Parva Naturalia (pp. 1– 68/183–340). Oxford: Clarendon Press. Saint-Martin, F. (1990). Semiotics of Visual Language (1987). Bloomington, Indianapolis: Indiana University Press. Schweppe, H. (1993). Handbuch der Naturfarbstoffe. Hamburg: Nikol. Silvestrini, N., & Fischer, E. P. (2002). Farbsysteme in Kunst und Wissenschaft, K. Stromer (Ed.). Köln: DuMont. Stadler, M. A., & Haynes, J.-D. (2001). Dynamics of perceptual representation. In M. Matthies, H. Malchow & J. Kriz (Eds.), Integrative Systems. Approaches to Natural and Social Dynamics: Systems Science 2000 (pp. 540–549). Berlin: Springer. Wyler, S. (1992). Colour and Language. Colour Terms in English. Tübingen: Gunter Narr Verlag.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.1 (45-129)

Attractiveness and adornment Reference to colors and smells in Papuan speech communities Volker Heeschen This contribution is based on the author’s lexicographical work in two Papuan speech communities as well as on published and unpublished narratives, songs, and instances of everyday speech. About a dozen dictionaries of Papuan languages and the ethnographic literature was checked. There is a discrepancy between the important messages transferred via the olfactory and visual channels of communication and the poor linguistic means referring to them. The signals used in olfactory and visual communication are not necessarily transformed into meaningful acoustic signals, they form part of the signaling behavior in the functional cycles of eating, pair formation, fight and threat. Words and idioms come into being in relaxed meta-practical activities like the artist’s work or during speaking about being attracted or repelled, examining food, growth, freshness and decay and by exploring the environment for signs of dangerous or useful and beautiful things and humans.

.

Introduction: Functional cycles and naming

The Kwoma speaking Papuans, who live in the north-east part of the island of New Guinea, have a word mukushe gwonya referring to the “characteristic smell and aroma of a pubescent girl, especially the smell emanating from the armpit and genitals which men find highly attractive sexually” (Bowden 1997: 138). The Eipo living in the mountains of West Papua say that they do not like the smell of an old vagina, a fresh or new one is good (wise kwate tong malye, winiryuk telebe).1 Like . All Eipo and Yale data without references are based on my own research. Eipo and Yale, members of the (Papuan) Mek language family, are spoken in the eastern, central mountains of Papua Barat (formerly West New Guinea and Irian Jaya), Indonesia. See Heeschen 1992b for information on the composition of that family and its relationship to the Trans New Guinea Phylum. For short introductions into the culture and for further references see Heeschen 1990a, 1992a, 1998; as well as Eibl-Eibesfeldt et al. 1989. Between 1974 and 2002, I spent more than five

JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.2 (129-180)

Volker Heeschen

the courting Yellow-breasted Bowerbird (Chlamydera lauterbachi)2 presenting red fruits and colorful objects to the female visiting the male’s dancing-ground men present themselves during dance festivals with beautifully painted faces, golden and white feather adornments, red leaves, white chains, dyed netbags, and bundles of fresh, green leaves. The men present their arms, display the cohesion of their group and are ready to defend their territory. One could be tempted to think of the natives’ world as one of strange smells apparently beyond the western people’s sensibility and a multiplicity of colors suited for disguise and camouflage during carnival. However, listening to narratives, songs, and everyday talk only yields a very limited number of terms, the rare examples being scattered over vast stretches of ‘texts’. There is a discrepancy between the seemingly important messages transferred via the olfactory and visual channels of communication and the poor and scarce linguistic means referring to them. The signals used in olfactory, visual, and tactile communication are not necessarily transformed into meaningful acoustic signals or speech, they form part of the signaling behavior in the functional cycles of pair formation, territoriality, and fight and threat. Language does not focus on them, and speakers may even be unaware of them and of the ways they tune and direct behavior. First of all, then, this contribution has to delimit the role of language in the whole of human communicative behavior. Second, the various and different color naming systems in the languages of the world have been partly tied to, or, at least, used as examples in, discussing the problem of linguistic relativity;3 I believe that contributing to this problem may depend on answering the first question. Third, narrowed down to the linguist’s framework exploring how smells and colors are years in the Mek language area. Research was founded by the Deutsche Forschungsgemeinschaft and the Max-Planck-Gesellschaft. In 1974 and 1976 I worked among the Eipo in the center of the Mek area as a member of the Interdisciplinary German Research Project “Man, Culture, and Environment in the Central Mountains of West New Guinea, Indonesia”, and from 1978 to 1981 for the Vereinigte Evangelische Mission, Wuppertal, among the Yale people, the westernmost Mek people. Between 1983 and 2002 I revisited the Yale as well as the Eipo eleven times for periods of up to four months. Eipo and Yalenang were first contacted in the seventies of the last century. Their main work consists in making gardens, the sweet potato being their staple diet. They live in hamlets of up to 200 inhabitants. Social and political life is characterized by virilocal, clan exogamic marriage rules, trading partnerships, dance and exchange feasts, conflict managing by warfare or informal talks, and by non-institutionalized big-men leadership. Religious and spiritual life centers around the men’s houses, the cult of the ancestors, and the ritual re-enactments of the founding-deeds of ‘those who came first and took possession of the area’. . See Eibl-Eibesfeldt 1999: 234 and Table III. . A careful and balanced account of the older literature in Gipper 1972; recent discussions in Lehmann 1998; Lucy 1996, 1997; Levinson 2000; the problem seen from the ethnologist’s point of view with rediscovery of nineteen century sources in Fischer 2000.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.3 (180-242)

Attractiveness and adornment

named could probably be part of a lexical typology “but we may wonder if a (lexical-)semantic typology exists at all, because the lexicon seems to be too full of interlingual diversity and of idiosyncrasies to lend itself to systematic typological studies” (Koch 2001: 1142; see also Brown 2001). Indeed, as one may guess from the unique and peculiar introductory example, lexicographic work contrasts to morphology and syntax, which try to capture the essence of language by abstracting from utterances, a dictionary providing “at its best a joyful sense of business with language. One is immersed in the details of language as in no other field” (Landau 1984: 4). It may be worthwhile starting work from details, idiosyncrasies and uses and not from known patterns: it shows how meaning is built into language from a variety of sources, how “cross-modale Assoziationen” are synthesized in “Sprachzeichensystemen” (Jäger 2001: 26), or how the reflexiveness of language turns the signals of the other codes into representational symbols distanced from ongoing interaction and freed from the primary functional cycles (see Heeschen 1990b, 2003a). Unfortunately, I have done no systematic work and testing on references to colors and smells during all the times in the field. My “joyful business with language” did not extend to the themes cut to shape for those who study structures of la langue not being “immersed in the details of language”. The data come from dictionary-making, everyday uses of speech and collections of texts.4 Fortunately, then, the instances are genuine showing that naming is embedded in activities and originates from functional cycles.

. From composite signals to distanced communication Recent research suggests that smells trigger emotions and that emotions are indispensable for memory, attention, and successful communication (see Schleidt 1995; Wassmann 2002). Nevertheless Schleidt (1995: 36), summarizing older research, says Für höhere kognitive Prozesse sind Informationen, die auf der Basis von Gerüchen erworben wurden, wenig bedeutsam. Korrespondierend dazu ist der Geruchssinn nur vergleichsweise lose mit der Sprache verbunden [...].

. The present work is mainly based on my own lexicographical work (Heeschen 1992a; Heeschen & Schiefenhövel 1983), published texts, see Heeschen 1990a; and unpublished texts, (a) about 300 pages of new Eipo texts and (b) about 1000 pages of Yale texts. Furthermore, I have read and checked a dozen dictionaries of Papuan languages as well as the ethnographic literature.



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.4 (242-301)

Volker Heeschen

Human beings have an individual smell. “Die chemische Signatur jedes Menschen ist auf der Geruchsebene genauso spezifisch wie auf der Immunebene, ja beide scheinen zusammenzuhängen.” (Schleidt 1995: 38). Smell plays its role in mating, bonding, cohesion of a group, attractiveness of one’s own territory, and evaluation of food (Schleidt 1995; see also Grammar 1995: 252). Eibl-Eibesfeldt (1999: 740) and Eibl-Eibesfeldt et al. (1989: 20) describe some rites and ceremonies in which the smell of the sweat, especially that of the armpit, signals identity, cohesion, or feeling of fellowship. Eibl-Eibesfeldt watched among the Gidjingali in Australia how a man, who was going to leave, touched his armpits and rubbed the palms of his hand on his partner’s body (Eibl-Eibesfeldt et al. 1989: 21). Fragrances attract healing and beneficent spirits, the Yale people use the wood of the kulun heing, a liana species, the spirits are believed to follow its fragrance when it is burnt and carried away. Repulsive smells ward off bad spirits. In order to convict a man of theft, the Eipo invite him to sit down in the men’s house, light a big fire, watch him intensively, and when he begins to move and to draw back from the fire, he is said to be guilty: probably the thief perspires the bad smell of those who are afraid. In religious ceremonies and rites fragrances and incense are well-known. Most languages referring to the vast realm of olfactory signals are satisfied with simple interjections and expressions like Eipo malye tong ‘a bad smell’ and teleb tong ‘a good smell’. A good number of languages have grammaticalized the category of gender. However, the ‘fireworks’ of olfactory, visual, and tactile signals in the functional cycle of courtship and mating cannot be linguistically systematized, and, from the grammarian’s point of view, there is no need to focus one’s attention on the names of smells. The same kind of discrepancy shows up with respect to the color names: a good number of languages have classificatory systems referring to the form and shape of objects, yet color is not central to the semantic organization of any known language, and [...] is never grammaticalized [....] In this light, it is indeed strange that colors are a canonical topic for lexical semantics, for discussions of the general relationship between language and cognition [...]. (Goddard 1999: 261–262)

The paradigm set by Berlin and Kay (1969) is still underlying current research. They say (a) that there is a set of basic color terms, (b) that there are six primary foci (black, white, red, green, blue, yellow), and (c) that the evolution from primitive systems naming only black and white to more elaborated semantic fields follows strict rules, for instance a stage following black/ white/ light/ dark systems differentiates the white/ light area into white versus red (including yellow).5 Criticism, besides being directed against the sample of languages, the number of . See references in Note 3.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.5 (301-359)

Attractiveness and adornment

informants, and the bad documentation, was resumed and carried on by Lehmann (1998) and Lucy (1997): (a) the distributional and grammatical properties of the terms are not investigated. Those who research color terms do as if the language the terms belong to has no grammar, which conveys special meanings to special structures within the structural whole of that language. No linguist would expect one range of meanings to be always expressed by one and the same grammatical property, for example color terms by adjectives. (b) The standards of semantic analysis are neglected. The objects referred to, that is the denoted object or the extra-linguistic reality, and the meaning are constantly confused. (c) No attention is paid to usages, styles, and ways and fashions of speaking. (d) The everyday activities like touching and examining a fruit, preparing, drying, and burning soils, dyeing and decorating, from which color terms originate, are neglected.6 Having researched usages among the Bellona speaking peoples, and “contextualised colour terms”, which designate objects which have changed color and surface, and having emphasized that the terms under investigation are “so closely connected with specific objects, emotions, and other cultural aspects that they can hardly be claimed to constitute a separate colour category in our sense of the word”, Kuschel and Monberg (1974: 241, quoted by Lehmann 1998: 189) conclude: Bellonese colour terms could thus be analysed as belonging to the Berlin and Kay Stage 2, having only three colour terms. They could also be considered an extremely sophisticated system of colour notation, with innumerable ‘colour words’, way beyond the Western system, and thus more sophisticated.

(e) The distinctive values and symbolic meanings of colors in other functional cycles and symbolic systems as well as their relationship with smells (for example blushing in nonverbal communication, signs on the men’s house, or while examining fresh, green, black or rotten food) cannot be captured by so-called basic color terms. They are an outcome of insufficient testing, which relied on “two narrow methodological strategies: restricting the stimuli to color chips and the task to labeling” (Goddard 1999: 263). One could well do without basic color terms, if one discovered the semiosis of signs, indices, icons, and symbols, developing in the four codes (olfactory, visual, tactile, acoustic) and being synthesized in distanced “Sprachzeichen”. Maybe a dancing young man, emanating a smell from his armpits, which excites young women (Eibl-Eibesfeldt 1997: 604), who has painted his face and his body, who looks and stares at his beloved girl (though he is not supposed to do so in Papuan societies, a cursory glance is sufficient), who exercises strange movements and displays unusual postures, and whose emotions are expressed by . The only counter-examples are Conklin 1955; Kuschel & Monberg 1974 (quoted by Lehmann 1998: 186–189); Fischer 2000: 127–133.



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.6 (359-404)

Volker Heeschen

paralinguistic features, probably conveys just one meaning: ‘I love you’. During the night, separated from the group of his fellow dancers (in Papuan societies men and women do not dance together and never touch each other), secluded from the on-lookers and bystanders, he may sing a song of love. The composite signal is transformed into a verbal message, detached from actual commerce with other human beings, exempted from the specific moment and place and also from the coercions of life in a group, no longer bound to the individual case, but generalized and subjected to the listeners’ norm and control. What was desire and arousal, is longing, love, or sorrow. ‘Love’ is not iconically and indexically expressed, but thematized by sequences of representational symbols. In the same way an original composite signal expressing disgust is turned into a theme one is able to talk about. If you present to somebody a repulsively smelling object disgust is expressed by turning off one’s head, shutting eyes and mouth, and turning up one’s nose: all channels of perception are blocked (see Eibl-Eibesfeldt et al. 1989: 56–60). The expressive movements can be accompanied by an interjection. Trabant (1998: 121) calls interjections “Spezialzeichen für Emotionen”. It is remarkable that linguistics does not investigate the expression of emotions by means of language. This is left to ethologists and psychologists. Jetzt wo die Sprache völlig ‘kognitiv’ geworden ist, hat die Linguistik natürlich noch weniger mit Gefühlen zu tun als in den guten alten Zeiten, wo die Sprache auch noch ‘kommunikativ’ war und die Linguistik daher etwa mit den Bühlerschen Dimensionen von “Kundgabe” und “Appell” einen Raum für die Emotionen bereitstellte. (Trabant 1998: 122)

The young Eipo man, who was exposed to a bad smell, could later smilingly and ironically talk about that smell. Instead of an indescribable interjection [s x::], he used na lik ‘I did not like it’. Words for emotions are derived from gestures and movements of the body. There is only a limited set of words referring to emotions. Their meaning is rather general, bleached, degraded into mere signs of assertion and refusal. Thus Eipo lik is not “disgust”, but anything from “I don’t want or like it” to “I say ‘no”’. Again, a multi-channel expression, a composite signal, turned into a verbal symbol. Kundgabe and Appell occur in face-to-face communication, they favor, or even rely on, nonverbal communication. In discourse and face-toface communication composite signals abound, during relaxed talk and narration representational symbols prevail. The movement from composite signals to distanced symbols is also one from Kundgabe and Appell to the representational function of language (Bühler), from multi-channel communication to the generalized code of language. In face-to-face communication the different codes supplement each other. During relaxed story-telling and ironic accounts of events the relation of language to the other codes and semiotic systems develops into one of reflexiveness. The

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.7 (404-464)

Attractiveness and adornment

language can describe someone who is courting, it can name rites and ceremonies carried out by expressive signaling in all codes, it puts into words the code of nonverbal communication: the sweating of a thief, the gaze of a lover, the disgust of someone exposed to a bad smell. This may explain the paucity of the verbal material in a considerable number of semantic fields in small societies in which a multitude of codes and semiotic systems is used in well delimited spaces and temporary tasks and events. Consequently, while looking for basic color terms or for names of smells one has to be well aware of the fact that the use speakers make of their vocabulary, and the relationship of the lexicon to other symbolic systems should not be neglected. How behavior and culture, on the one hand, and language, on the other, lie athwart so that the two domains do not even correlate with each other can be illustrated by two instances, by reference to number and sexual antagonism. Sometimes I had the impression that even big men were not able to count up to three (when interviewed and asked inappropriately). Though the counting system seems to be restricted, the Eipo and Yalenang love to count. Asking for the number of pigs or children seemed to be hopeless. However, in fairy tales up to one hundred cuscus are mentioned. In real life wounds, arrows, casualties, the number of hits, and earth-ovens are counted. It is tabooed to speak about pigs, yet the number of pigs killed for an exchange feast, their size, the quality of their fat and meat is well known. An old man looking back at his life may be proud of having produced more male offspring than the interviewer. Counting the number of mornings after sleep and the number of moons or seasons is essential for measuring distances and sequences of events. After dance the number of taros and other gifts are carefully counted. The speakers are fully aware of the quality and countability of objects and living beings as well as the flow of time. Style and ways of speaking make the best out of a set of words, which seems to be poor if one looks for numbers detached from objects and used for calculations. The same relationship between the paucity of the linguistic material and the richness of rites, the degree of awareness, and the constant preoccupation shows up in references to sexual antagonism, the strict separation of the sexes, which is said to be common in other Highland societies as well as in the whole of Melanesia. Eipo and Yale have only two words to refer to this distinction, ‘man, male’ and ‘woman, female’. Beyond language the antagonism is shown in taboos, rites, and ceremonies. Women come from, and dance at, the periphery in primordial times, men establish themselves in the center. The men’s house, which the women should not approach and which they are not allowed to enter, is thought of as the very center of one’s world. Coming and going in the hamlet, the use of certain paths in the garden areas and in the hunting grounds is strictly regulated. The initiation is, above all, a separation of the boys from the women and their genuine fertility. Dirt, dust, and food characteristic of that sphere are ritually driven out of the body



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.8 (464-511)

Volker Heeschen

of the boys during initiation. Men are reborn by cultural means, and the food they eat, the objects they handle, and the actions they perform while constructing a hut or making a new garden, are purified by rites, that is to say by cultural acts. The antagonism is transposed into nonverbal behavior and ritual action. The corresponding worldview could not be guessed from lexicon or grammar. Conversely, we know societies whose language constantly classifies nouns according to genus or sexus, but where the antagonism is absent or far less developed.7

. Reference to smell in Eipo, Yale, and some other Papuan languages A good number of Papuan languages have a verb meaning ‘to perceive’; in collocations it goes together with nouns like “eye”, “smell”, “ear” in order to specify the organ of sense and the perceptive faculty. The pattern runs like this: ‘eye perceive’ would be “to look”. For example Middle Wahgi dakil pi- ‘to smell a bad odor’ from dakil ‘a bad odor or aroma’ and pi- ‘to understand, know, hear, perceive, feel’ (Ramsey 1975: 39, 240). Though Eipo yubren- means ‘to perceive’ as well as ‘to smell’, it is usually constructed with tong ‘smell, odor, aroma, fragrance’, such as tong yubrenman ‘I am smelling sth.’. I have also found tong gekebren- ‘to smell’, though the verb usually refers to acoustic perception: ‘to hear’. Yale song covers the same meanings as Eipo tong. There is no simple verb. We find song kaeb- and song lib- ‘to smell’, with kaeb- ‘to hear, understand, perceive’ and lib- ‘to fill sth., to put into the netbag’. All other expressions, idioms, or collocations are associated with special organs, objects, states of arousal, and, last not least, ways of speaking, for instance: (Yale) silimu buluktongue tear, start anew ‘to play with one’s tongue, to smell (of snakes)’ Pam song (or sao) ngeknel pig smell (or taste) it.has.sniffed.at.me ‘The smell, or the appetite for, pork has entered my nose.’ nonomnonom amarking or indicating make ‘to smell, perceive’ (idiom used in fairy tales which delight in finding new expressions and metaphors)

. This passage is slightly altered and taken from Heeschen 2003b.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.9 (511-590)

Attractiveness and adornment

(Eipo) Sikin ubmanil. feeling (in the mouth) it.is.to.me ‘I smell or taste sth. good.’

Smell and tastes are not differentiated. They emanate from leaves, trees, body parts, and food, for example Eipo asulye (Yale kwada’au) ‘plant similar to parsley (Oenanthe javanica)’, yek or yeik ‘a small tree with fragrant flowers’, kúma ‘ginger (Zingiber officinale)’, kutmut ‘plant species with aromatic leaves (Blumea sylvatica)’, kinkin (Yale kenkenak) ‘pleasant smell or taste (e.g., that of salt or sugar cane)’, ua ‘saliferous plant’, saboka ‘tobacco’ (see also Yale kulun heing in Section 2). These terms only imply reference to smell, other terms refer to the characteristic evaporations of the referents: Kwoma kwowa ‘aroma; strong smell (e.g., the characteristic smell of a person’s body)’ (Bowden 1997: 106), Yareba ofe urara ‘body odor’, from ófe ‘skin, bark, body’ and urára ‘stinky’, sisámu ‘good smell of meat’ (H. and N. Weimer 1974: 214–215, 245, 275), Hua beptu “smell of vagina [...]; a mushroomlike smell resulting according to a fable from the swallow’s revenge on a group of girls who refused to sleep with him. While they slept he crept into their house and stuck zokoni mushrooms to their vaginas.” (Haiman 1991: 13, see also Section 1) Whether there are terms for the smells of plants, animals, the human body or body parts, is more or less accidental. The awareness of characteristic smells does not necessarily lead to a complete list of words and wordings designating what the olfactory faculty is able to detect. The references to plants depends on the habitat and the uses the people make of these plants in rites and ceremonies, the references to the functional cycle of pair formation may depend on whether young men and young women are allowed to come together during dancing festivals; usually they are strictly separated. However, there is one sphere for which almost all languages have words, collocations, or idioms, namely that of decay and rottenness, which includes the different degrees of edibility. We will see that these references are partly connected with the terms for ‘ripe, green, fresh, black, old’ (see Section 4) as well as with the references to the state of the skin. The survival value of being able to discuss the edibility of meat and fruits is quite evident. The skin may be covered by wounds, indicate the primordial unfinishedness of human beings or symbolize beauty and healthiness which originate from the community’s ritual activities. There are terms for the state of wood, food, leaves, meat, skin, and hair. Some of the terms are associated with smells, others with ‘having become cold or old’. Only a limited number of Eipo and Yale examples shall illustrate the range of meanings. Examples from other Papuan languages would yield the same ‘picture’ (one should say the same atmosphere of odors).



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.10 (590-644)

Volker Heeschen

(Eipo) albabna bulong bitbit tong fubun tong fun suk wise

‘cold, rotten, decayed’ ‘cooled off, stale’ ‘a bad smell’ (like the one of cold smoke or burnt old netbags) ‘a bad smell, smell of decay, of cold smoke or old netbags’ ‘rotten (of wood)’ ‘foul smelling spot on a sweet potato’ ‘ripe’ (of food which is ready to be eaten, of fruits which have become colored or soft), ‘old’ (age is associated with a special smell and with being soft)

In everyday speech speakers would say malye tong ‘a bad smell’, malye dayuk ‘it has become bad’, lyubuk ‘it has become soft or rotten’. (Yale) bolon obos sok mohon suk eda song

‘odor of putrefaction’ ‘withered, rotten’ (e. g., wood exposed to rain) ‘rotten (of food)’ ‘long, slow, bad or penetrating of smell’ ‘rotten (of trees), dried up leaf ’ ‘the smell of plenty of rain’ (probably because the smell of rotten things becomes stronger during rain)

Again, in every day speech some very general terms and idioms are used. Terms like obos or sok, which do not seem to refer to smell, are used together with an interjection of disgust or with mali song ‘a bad smell’, when a rotten or withered object is presented to a Yale speaker. In accounts of everyday life and in fairy tales ancestors, spirits and people are attracted by the smell of earth ovens, especially by pig’s fat. Yet ‘smelling food’ only forms part of the background information concerning the coming and going of the protagonists. A human-like being covered all over with tiny, evil-smelling wounds and without skin quite often makes his appearance in fairy tales. The name of one of the bad protagonists is Yale Kabumseleang ‘the one who is covered with wounds’. The smell of a decaying body constitutes a recurrent, though minor motive within the series of episodes in a fairy tale: somebody has killed a friend or relative by chance and buries him. The people look in vain for the absent person. Birds and other animals are attracted by the body, the people watch them and discover the dead person. However, information on smell is never important, it must be extracted from the other pieces of information and is always part of something otherwise forbidding or unfinished. Considering these restrictions one may look at the following episodes as typical instances of how the speakers pay attention to,

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.11 (644-695)

Attractiveness and adornment

or, so to say, pass over smell.8 This passing over is stressed by the fact that the human visual faculty makes out the spot where animals move and gather attracted by rottenness and decay. (Eipo) Yakob Bolmerin: Kwemdina nukna Diduemna asik kil ton, el makalnye ton, el makal kil ton, ora, el makalnye kil ab, sik winilye ulamek. Sik winilye ulamik ane, el makalnye ninye asik beyuk. Ninye asik beyuk-buk, el makal kil ab, el basi kil, el basi kil, sik betinye wa wilberik. Keting weik deibmuk obora, el makalnye kil keting-uk ateb maning kankan mabuk. Mabmuk-buk, el basi kil wa weibnin balamuk akarye, el makalnye kil ya sekuk ora, dilelamuk-ak gum, fi deibuk. Debmuk obora, wiribnamal-do tenen kanye teninmuk-une [...] kwaning wa falilamak atonun ubuka, falinmuke falibukabukye telibdongobuk. Tunim amuk obora, “an basi kil ton ara wiribuk-do?” winyalamuk unyak, “na walwal, na gum bikse,” winyabuk. Winyabuk obora, engkiriklamuk, el akanye birye fi elamolamuk, gum eibuk obora, a-wa windam balamuk ob unyak, lyulamukce, kamselenge, kwasire, ayame, make, yukyuk nirya lyulamuk ate nin tong9 yulyangalamik akarye, yan ara wik ulamik. [...] Dildelamuk unyak, el kil akonum obukabukye tokwe falibuke yala kouamuk obora, eibuk. Yakob Bolmerin: A myth-like narrative In the hamlet of Diduemna a woman, her brother, his sister, the brother and his wife, the three of them were living. The three of them were living there, and then the brother visited a men’s hamlet (a hamlet where partners and friends were living). While he was gone, his sister and her sister-in-law, the two of them went in order to make gardens. The sun was shining very hot, and because of that the brother’s wife slept in the grass. While she was sleeping, her sister-in-law went making gardens, and by that she cut her brother’s wife with a stone adze, and where she looked, there was nothing, she died. After that she was oppressed thinking about what she should do, and it was as if she had made gardens for sweet potatoes, and she made these garden, and having . The first episode is taken from an unpublished manuscript “Neue Eipo-Texte”. The second from Heeschen 1990a: 53–56. In Heeschen 1990a the complete narrative, comments and other versions are found. Another version will be published in “Neue Eipo-Texte”. The third episode was published in Heeschen 1998: 361–363; other versions of the creation of the Tengket people in “Neue Eipo-Texte”. The third example is also one of the rare instances of reference to colorful objects in on-going discourse or pieces of narration, see Section 4. . nin tong is an enigma, it consists of nin ‘1. mother, mother’s sister, 2. smell, feeling’ and tong ‘smell’. Probably ‘mother’s smell’ is a metonym for ‘feeling home, home, the smell of one’s home or one’s relatives’, hence generalized ‘feeling, smell’. See also Eipo nongyape, ninyape ‘that which belongs to the mothers, the essential part’, hence probably ‘tradition, oral tradition’.



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.12 (695-743)

Volker Heeschen

made these gardens, she covered her. After that he was asking: “Your sister-inlaw, how and what has she done?” But she answered: “I am ignorant, I don’t know”. After that he looked for her, her husband was very much looking for her, and when he did not find her, he was going around in the garden area, and as to the one who was decaying, the kamselenge-birds, the kwasire-, the ayame-birds, and all the other birds, they came smelling the mother’s smell (the smell of a close relative), because she was decaying, and that is why there were a good number of them [...] While he was looking, he saw that someone had killed his wife, worked the soil and planted the seedlings.

The same sequence of motives is found in the next example: death, decaying, smelling, attraction. Yet here it forms part of a well known mythical context: the killing of an ancestor in order to assure food and a habitable world. – The men tried to kill the Yulye who annoyed human beings by making strange noises and by shining like red fruits in the forest. They did not succeed in killing him, finally the eryon skink came from the west and killed him. On his flight the Yulye created the human world and his scrumbling parts nourished the animals which humans hunt and use. (Eipo) Bingde ab, Kwengkweng ab: Yulye yupe Dake singibuka binamuk-ba, yin bide liblib yanamolomuk-ak motokwe doke kibkib yanamuk-ane, winebrak tokwe-ak yane foubuk. El Keme dam, Borumna dam yane foubuk-ba, yin dakamuk, burburumna yin dake binamolomuk, dake binamuka ak-uk yane malekuk. Malekuk dara, yane foubukabukye, yane foubuk. Lyubnamak dara, winebuk. Lyubnaman tenen el lyubnamukye winibmuk-ba, ninye wine lyubuk-ba, winebra kulib arye, make arye, sanib arye, malye, mana, koude, ayam, kusinge, make lolublamle nirye amubik. Amobra tokwe weriamik. Bingde und Kwengkweng: Story of Yulye And it (the skink who succeeded in killing the Yulye, an ancestor of a clan in the west) carved and draw (the bow) and it went there, and he tried to avoid the bow, thereby he came creating the valleys and passes, but then he fell on the earth and fell apart. He scrumbled at the Keme mountain and at the Borborumna mountain, and where it had carved the bow and was gone, he fell down. After that he fell apart and decayed. He decayed, and there, where he was destined to decay, in the way human beings decay, birds of paradise, all birds, the casuary, all of them ate, and where he had gone bad and was dissolving, the cuscus, the koude-and ayam-rats, the crickets, and the birds gathered and they buried him in the soil.

In the last example the flies signal a bad smell. The ancestor of the Nabyal clan discovers the ancestor of the Tengket clan in the earth, he digs him up, shoulders

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.13 (743-832)

Attractiveness and adornment

him and carries him to the valley the Nabyal man was going to take possession of. They will be allies, trading-partners and exchange women. The Nabyal man will assign garden areas to the Tengket man. The unready and unfinished state of the Tengket man’s body will be ritually purified. (Tani, a dialect spoken east of Eipomek) Kabrob Tengket: Tengketnang kwemdina [....] gekenmau ate Tengketnye bida Atapekwendam kwiwidim atonun, ore boba a-pferamde atonun, ore Nabyalne daksaksak banmau, “dantam, dantam pfenmar-o?” tenen binmau-buk, Kolesolul bisibukmaua, Ongkongoble bisibukmaua, a-binmau-e, Bonamde koubwe, Soanye a-binmau-e, Atapekwendam baramwe ocak, “ninye diba bukmar-o?” tenen baramwe ocak, kabum ongobmau, yilanye kabum ongobmaua, ora, dara bum arye ore manmalya co arye kisok dub nalebra antananmai. Kisok dub-ai doukwe bibiya banmau, nalebra antananmai, widim bo ana, ki bo ana atonun talebmau-e, a-tokwe fotokra buronmau. Biri tokwe fotonmau. Fotonmau ora, oupe ono dena kum. Kabrob Tengket: Narrative of the creation of the Tengket people [....] He was listening; the Tengket man, he was as if it were buzzing, like the turning (recorder) here does, and then the Nabyal man rushed and was going, because he thought: “Where does this noise come from?” He went to the Kolesolul and Ongkongoble mountain, there he crossed the Bonamde and Soanye rivers, and at the place to which he went at the Atapekwendam mountain, at the place to which he went thinking: “Is there a real man?” there he (the Tengket man) was covered all over with small wounds, there the bad man was without skin, and flies and bad scrub were buzzing around his head. At the top of his head bibiya-rhododendron scrub was moving, the flies were buzzing and humming, and the earth was cracking. The yellow (bad) earth was cracking. After that, there was no noise.

According to what dictionaries and collections of texts show, reference to smell seems to be a secondary system. Though it is ubiquitous and puts into the moods of attraction and disgust, approach and flight, it is expressed and mediated in language by visual and acoustic signals, the colors of fruits, the state of the skin, the gathering of animals, the buzzing of flies, a strange noise in the earth. What humans passively perceive, as undergoers and patients subject to the evaporations and emanations in the world, is signaled by what they are able to do, watch, touch, and examine as agents. Ironically, one of the few shining colors in the unchanging and permanently green grasslands, garden areas and virgin forests, namely the beautiful colors of the endemic rhododendron species, are not man-made and only symbolize the unfinished primordial state of an emerging ancestor.



JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.14 (832-861)

Volker Heeschen

. Reference to color in Eipo, Yale, and some other Papuan languages Eipo bibiya mentioned at the end of the last section is a generic name for different Rhododendron species (corresponding to Yale lelbelakna – Rhododendron sp. div.). It could be derived from Eipo bi ‘red, reddish, orange, reddish-blue’ by way of reduplication and the suffix -a (see Heeschen 1998: 203). Its meaning probably could be paraphrased by ‘what is remarkably colored’. The same formation is found in Yale uwonguwonga ‘color’ from uwong ‘red, reddish (all medium-red color-shades)’. Indeed, in a world in which shining flowers are rare, in which green grasses, forests, and black and green mosses dominate and where all colors turn into dark hues when clouds climb up the mountains and rain descends from the high forests, red objects stand forth in the continuum of shades ranging from white teeth, foaming rivers and distant waterfalls to the muddy soils and dark forests. When the rain comes, the Yale people say moksoo midibal ‘the mountains have become dark’. The name for color comes from the prototype of what is visually conspicuous. However, in other language communities a general concept is hardly different from terms referring to activities, such as painting, decorating, dying, tattooing, drawing lines, and carving ornaments, for instance Kwoma now “1. paint [...], 2. earth pigment (of kind used to make paint [...], 3. colour” (Bowden 1997: 155; see also Fischer 2000: 126–133), Halia toptop “tattoo, skin marking, coloured” (Allen et al. 1982: 162), Fore awanayé “decoration, adornment, ornament, colour” and pirásiwé “flower, decoration, colourfulness” (Scott 1980: 29, 84). Sometimes the collective meaning is taken from a whole of colors as found in nature, for example Yareba furúru “lines of varying colours (as on a spotted cuscus)” (H. and N. Weimer 1974: 126), very special Managalasi anuakatao “colors of the rainbow seen in a reflection, i.e., on snakes when the sun shines on them” (J. and J. Parlier 1981: 24), and very complex Telefol kaal “skin [of person], peelings [of vegetable], bark [of tree], colour” (P. and H. Healey 1977: 93). Though I do not believe in the existence of basic color systems, because all so-called underlying and/or abstract systems result from concrete speech and are a way of economically summarizing the references to previous activities, the simplified Yale and Eipo systems can be said to be similar to Berlin and Kay’s stage two10 . See also Heider 1972, where the system of the Dani language is treated. The Dani languages are the Western neighbors of the Mek languages. Though they are grammatically quite different, terms like Dani mol, Eipo mele, and Yale mol ‘bright, light, white’ may represent the same etymon. However, as ornaments and pigs are traded the colors of which are carefully differentiated, the words may have been borrowed from one community to the other. E.g., white or light skinned pigs are said to come from the west, see below Yale solom, Eipo sulum.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.15 (861-933)

Attractiveness and adornment

‘white, light, bright’ ‘red, reddish, orange’ ‘dark, black’

Eipo korun bi merin

Yale kul, solom uwong midin

Such a system can be constructed for a good number of Papuan languages. Its reality is, so to say, bleached in discourse, narration, judgements by different divisions, connotations, the grammar of word classes and collocations, the metonomy of characteristically colored objects (and people), and by homeric comparisons. Eipo korun extends into ‘red’, while Yale kul, solom are somewhat restricted to ‘what appears light, gray, or white skinned’. Consequently, the western people are called in Eipo korun-nang ‘the white or red people’ and in Yale uwong-nang ‘the red people’. Eipo korun has the connotation of ‘friendly, open(hearted)’, so Eipo could say someone has a korun kanye ‘an open mind’, that is ‘he has a friendly mood’, while Yale solom, kul as well as uwong do not extend into metaphorical uses. Quite often black and dark are associated with dirt. Obviously the basic color terms appear to be adjectives. Yet there is no linguistically justifiable rule which says that colors should be expressed by adjectives. Even the status of Eipo bi and Yale midin can be doubted. The correct translation of bi tokwe would be ‘a soil which is red, a soil of the class of red soils’, while im bi is ‘sky red’, that is the ‘redness of the sky, dawn, evening glow’. Furthermore there are collocations like bi foronmal ‘(the sky) is red’ with forob- ‘to dye, to show color, to become old, to bleach’. In the first case bi is really an adjective, in the second it ‘behaves’ like a noun, in the third it is positionally well-defined, it is preverbal, a position usually filled with unclassifiable words in Eipo as well as in Yale. Yale midin formally is a verbal noun of midib- ‘to get dark or black’, but the verb is rarely used (see above in this section), while the so-called adjective turns up in numerous statements and compounds, for instance midin mak ‘black water’ or Midin-nang ‘black people/ the name of a lineage or clan’. The equivalents or synonyms, too, of the basic color terms are not abstracted from the objects, but the attributes follow from the properties of the objects. There are different words for ‘black’ in Manalagassi: black of the skin, black-colored teeth, black (said of an ornament), very black (like the trunk of a fern tree) (see J. and J. Parlier 1981). Different ‘reds’ are implied or signaled by the skin, being too close to the fire, being red in the face with rage, red dye, blood, reddish saplings, red micro-organisms (algae?) used for paint, rust (J. and J. Parlier 1981; R. and A. Loving 1975; P. and A. Healey 1977). Different ‘whites’ quite often come from the feathers of the white cockatoo, the whiteness of the eyeball, the skin or fur of animals, white limestone (J. and J. Parlier 1981; R. and A. Loving 1975; P. and A. Healey 1977; Heeschen & Schiefenhövel 1983; Heeschen 1992b). The Eipo use the bukbuk (Melastoma affine) bush with black berries in order to blacken the netbags: aleng bukbukman ‘I blacken the netbag’. The red leaves of



JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.16 (933-999)

 Volker Heeschen

a tree, Eipo degit, Yale dikis (Elaeocarpus sp.), are used as decoration. For painting the face red ochre (Eipo mure, Yale keinangmud), is used. The most shining red comes from the sisilya bush, the slimy sap around the kernel in the fruit turns red upon touching and rubbing. The verb bukbuk- is generalised to ‘to dye’: sisilya bukbukman ‘I am dyeing with the sisilya-kernels (not: I am blackening)’. White are the feathers of the ngarye (white cockatoo) and the flowers of the ome tree. I have not found words for the waves of golden and yellow feathers of the bird of paradise used as head decoration during dancing. In narratives colorful phenomena are expressed by elaborated comparisons, for instance: (Eipo) Bingde: Yalenye yupe11 Yalenye el nong dare, degit atonun, mal atonun winebuka, kulib mak olokna on dam atonun, batareng im atonun liklobukmana winyabra, a-motokwe yin sekmal arye balamuk. El nong ara, ukwe bok lebmal atonun a-ubmal-to, sing a-ubmal-to ubuke, el winebuke el nong noitam anirye luluklamuk. Bingde: Story of the Yalenye As to the body of the Yalenye (the ancestor of a clan in the east), it is like the red leaf of the degit tree, like an arrow, with the feathers of the bird of paradise threaded into its bundle of ornaments, like the (white) nassa (shell) headband being like the sky after clearing up, like that he went as if his bow is lighting this world. As to his body, it is like that here, as if the flame of the fire is shining up, like the Job’s tears, and thus by the middle of his body (by what appeared from the middle of or around his body) it (the earth) was shivering.

Furthermore the importance of pure color terms is on principle overridden by the semantics of competing systems; all terms are likely to be, or to function as, a synthesis of cross-modal associations and symbolic activity. (a) The terms usually translated as “black” quite often signify ‘burnt, old, dark, rotten, stinky’. What appears to be “red” also is ‘ripe, old’. There are numerous collocations like Yale kolae dalamla ‘it is turning red, it is ripening’, kolae originally meaning ‘red leaf ’; red and ripe things develop into ‘withered, yellow, old’. There are special terms for ‘withered leaf ’ in Eipo as well as in Yale, fuluma and lesa. Almost all languages have terms for ‘green, unripe, fresh, young’.12 Eipo sukun, a verbal noun, signifies ‘brown soil, light green colors’, the corresponding verb sukub- means ‘to cramp, to get wrinkled or fissured, to get black’. ‘Light green’ is associated with ‘hard, tight’, hence kanye sukulamle means ‘his mind is cramping, he thinks of sth. with envy’. The notion ‘young, fresh’ is complex, it implies the . Taken from Heeschen 1990a: 63–64. . E.g., Kwoma neekiirii ‘blue, green, yellow’ (Bowden 1997: 145).

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.17 (999-1045)

Attractiveness and adornment 

dangers of ‘unfinished states’. When the men bring parts for the reconstruction of the men’s house, these parts are wrapped in green leaves, something which was nature turns into a sacred object, hence ‘green, new’ also implies ‘sacred’ (see Flierl & Strauss 1977: 98, 383). On the other hand, ‘green’ develops into ‘cramped, fissured, black’. In the dabo mot, the ‘dabo dance’, which is sung and danced during dawn, the Eipo sing Yale kil sukun teleb-o, aik dam burublo bilam-do? ‘The black woman of the east is good, have you gone off to the hut, disappearing?’ After the dance the men should be prepared to resume everyday activities like going ‘to an old and black woman’ instead of being in high spirits. – I believe that the reference systems, which partly underlie color terms, are ‘unripe – ripe’, ‘young – old’,13 which seem to be much more important than basic color terms. As we have seen, these systems overlap with terms related to smell. The semantics and the associations of the terms are built up regardless of the systematics of pure color terms. (b) The spectrum of colors is ‘outshined’ or ‘drowned’ by the references to dark and light. Homo sapiens is an animal active during the day. During the night the spirits roam about, come back to the hamlets, cause sickness, eat and consume the flesh and the healthiness of the living, and destroy human order. One is afraid of darkness. There are numerous stories which tell of children or lonely grown-ups who get lost in the forest after dusk and in the night. The Eipo say moke cene ton, doa wik ton kune wik ulamle ‘clouds and fog14 and rain cause thick shadows’, after dusk they say bisik dantam-do? winyabra, taruk aryuk wininmakye, walwal ublye ora, yo deye gum tekmak ‘where is the way? and they grope their way with their hand, but they do not know, they no longer stand at the stem of the trees (they no longer find support or a foothold)’. There are numerous15 words for ‘pale light, different hues of brightness, dawn’, for instance: (Eipo) inibmorubmele im mele likmal mile feibmal til

‘to shed a pale light, to dawn slowly’ ‘light’ ‘the sky is painted with light color’ ‘(the sky) is painted over with lightness’ (paraphrase for the Milky Way) ‘bright’

. And probably ‘soft – hard’ as well as ‘cold – hot’. Eipo and Yale have special terms for ‘young man, young woman’, e.g., Yale kabuni ‘young man’, saden kel ‘young woman’. These terms are associated with fragrance, brightness, beauty and freshness. . Eipo doa ‘cloud, fog’. . Quantity must be evaluated considering the size of the lexicon of small languages: in dictionaries of Papuan languages one finds from 2,000 to 8,000 entries (against more than 100,000 and up to 500,000 in English or German).

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.18 (1045-1104)

 Volker Heeschen

wale til donokmal fala foron foron im keite

‘the moon is shedding bright light’ ‘whitish, light’ (e.g., the pig’s belly, lower side of a leaf) ‘reddish, yellow, pale’ ‘yellow-whitish color of the sky’ ‘liquid, white, light’ (e.g., the ventral side of animals, the palm of the hand, the sky in the morning, the white water of a distant river16 )

Besides that there are words for the inner part of plants, for example Eipo kwalye tin ‘the white inner fibrous bark of a banana stem’. The dancing songs ardently and longingly salute the day: (Eipo) ururub bona min daknenibmai-o! ‘It should be quickly dawning where the sky is supported (at the eastern horizon)!’ See also the following complex example for the paraphrases and metaphors signaling the morning. (Eipo, Tani dialect) Dabo mot Dabo im keakea fobre, yalye im keakea fobre. Dabo nerme tarin teleb. Nener malye, nener wise ne bise bir, ou-bise kubreidika bir. Bi bon im-do, ong bona im-do? Bi bon im dibre, batum bon im dibre. Metekmetek oupe febkur-o, ore berekde, im kaner, ong bo im likde, bi sukul im likde. Dabo (lowland’s) dancing song The sky of the lowlands has opened, the sky of the east has opened. The young girl of the lowlands is good. (The day or morning is presented as the girl of the eastern lowlands; tarin etymologically same as Yale saden ‘young’, see Note 13.) My wife is bad, my wife is old, she has gone my way, she is gone the way down. (The old wife and the way down: images for the passing of the night, see also kel sukun above in this section.) Is it the sky which carries red, is it the sky which carries the ong rope? (The rope is thought to be tied from the top of the men’s house to the top of the mountains. Along that rope the spirits shall go home.) . See also the beginning of this section.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.19 (1104-1162)

Attractiveness and adornment 

The sky which carries the redness has opened, the sky which carries the batum-river has opened. (A mythical river along the line of the horizon where sky and earth unite and where the waters gather.) Small voices will rise toward you, it has dawned, the sister-in-law’s sky, the sky of the ong rope has been lighted, the sky which has put on redness has been lighted. (The small voices come from tiny insects, which announce dawn.)

In the literal sense of the word brightness dominates color: for dancing festivals and initiations the men rub pig’s fat on their dark and black skin. The shining brightness (and shining black) is a symbol of health and growth. In Kwoma kepi has a wide range of meaning ‘good, right, attractive (of woman), healthy, bright’; Bowden (1997: 87) comments Brightness is a greatly-admired quality in colours; because European paints are generally much brighter, and more durable, than their indigenous counterparts Kwoma artists use store paints whenever they can obtain them.

(c) This comment leads to another sphere of meanings which competes with the color terms. The artists and artisans focus on preparing the colors, extracting them from fruits, burning and roasting soils, dyeing, painting lines and forms, and, finally, assigning symbolic values to the colors. The colors are the very outcome of all these activities, but they must not be named, you can see the results and enjoy them. Awareness, however, is enkindled by the preparatory steps and by thinking about the colors. Lines and forms seem to be as important as colors. The symbolic values are not inherently given, but actively assigned. Thus they may differ from one community to the other. The Eipo say that black means ninye ibarye wik ublul tenen ‘for the sake of men being enraged and strong’ or ninye bol mirimmirim ablul tenen ‘for the sake of men’s skin being light and bright’. Red signals isa bindoblul tenen ‘that the spirits may go off ’, and yellow ninye bol kurublul tenen ‘that men’s skin turns into bright hues’. What Strathern (1981: 26) says about blackness in the Medlpa society (Mount Hagen, Papua New Guinea), is conform to what the Eipo think, while the values assigned to yellow and red are quite different. “[...] black face paint represents the internal group solidarity of males and their aggressiveness toward outsiders, while the bright colours of red and yellow stood for female values, sexual appeal, and intergroup friendship, affinity and exchange.”17 In Eipo ‘rage’ is associated with activity and taking the initiative. It may well be part of aggressiveness toward offenders and traditional enemies. As to red and yellow, I . See also C. Kaufmann in Kaeppler et al. 1994: 327, 330; Crawford 1981: 183–203; Forge 1970; Gell 1975.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.20 (1162-1208)

 Volker Heeschen

believe that the gender antagonism referred to by Strathern has no bearing on how the Eipo assign meaning to the colors. The same differences may be found if one compares the paintings on the walls and planks of the men’s houses. In the Sepik area the cosmos of the ancestral world is depicted, while the Yale people paint tiny pictures of the sun, ceremonial shields, and cockroaches on the wall hoping that women will be simply astonished at seeing them and not think about the real sense of the men’s houses. While joy can be expressed by a color, grief and mourning are symbolized by putting ashes on one’s head, looking down, being dirty, and moving slowly. The symbolic meanings of colors are supplemented by the semiosis of movement and nonverbal behavior. – Colors, then, are embedded in coloring activities and in ‘thinking for coloring’. (d) As one may have already guessed the real place for naming colors, speaking about colors and assigning symbolic meanings to colors or colored objects is narrative discourse and poetry: references to concrete objects with hidden and veiled meanings or play with the colors and hues of the world distanced from the functional cycles and everyday activities. Even younger narrators ironically make a note of how a bald head is shining in the dark: Yale (dela) kaiskais18 alamok ‘(a bald head) flashed up’. The author of a song artfully mentions colored flowers. (Eipo, Tani dialect) Kutaya lumudamde ei-bindobdum fenman, balting lumunmar moram ei-bindobdum fenman. I say, go up there, where you bend down to the kutaya-flowers, I say, go up there, where you bow to the fern tree. (kutaya is a small flower growing in the grass: bending and bowing is done by those who meet unexpectedly and make love.)

Mentioning flowers can also express mourning: the dead persons turn into birds and, while flying from bush to bush and eating the flowers, they slowly move from the hamlet to the realm of the dead at the shining tops of the mountains. A dancing-song depicts the play of the bower birds who delimit a place for mating and decorate it with flowers: ma ererek bate dibre ‘the ererek-bird has placed the colored bundle of ornaments (or: the prohibitive taboo-sign)’. Somewhere in the enumeration of bower birds and what they do the names of humans who love each other are hidden. A song sung and composed by women pictures moving and hopping bundles like the ornaments of bows (consisting of colored stripes and feathers of the bird of paradise) and bands made of yellow orchid fibers: they are worn by the fighting men, yet the women pretend that the men come dancing to them. They are longing for peace and the returnees. Another song plays . Yale kais etymologically same as Eipo keite ‘liquid, white, light’, see above in this section.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.21 (1208-1260)

Attractiveness and adornment 

with the references to white objects. The author, a woman of a neighboring valley, takes pleasure in the ornaments worn by two young men. Both of them are said to have white teeth so that the two meanings of lenmal are systematically mixed up, namely, ‘he says’ and ‘he smiles’. (Eipo, Fa valley) O Kelwe meley ong, lenmal ong, Leklo lulay ong, lenmal ong. Ya ukwe danamuk, aik ukwe danamuk, bam ukwe danamuk. Ouda lulay leble, meley leble, Kelwe meleye lenmale, Leklo lulay lenmale [....]. O Kelwe (white and shining) cymbrium chest-decoration ong,19 he is speaking (smiling) ong. Leklo (white and shining) bone-dagger (used as decoration) ong, he is speaking (smiling) ong. It turned into the fire of the stone-adze, it turned into the fire of the hut, it turned into the fire of the pig. (Symbols of wealth, stone-adze, home, and pig, are mentioned and presented as if they are shining and glowing in the fire.) Down there, the bone-dagger has spoken (smiled), the cymbrium has spoken (smiled), Kelwe speaks (smiles) cymbrium, Leklo speaks (smiles) bonedagger.

Maybe red is the color of warning (and not of affinity and friendship). In myth and in tales of origin the early settlers saw something like the red fruit of the pumpkinplant (Trichosanthes sp.). It made a thundering noise and alarmed the men. The Yulye, an ancestral figure (see above Section 3), had turned into something threatening, kisang atonun unmuk-ak ninye mentebre balamik ‘where he was like the red kisang-fruit, they went missing him’. We may return to the beginning of this section: in the world of greenness one pays attention to a colored object, it should be touched and examined, however, it is new, strange, probably sacred, besides that it makes a thundering noise, which, in myth, is the sign of the arrival of the sacred, the going over from nature to culture. After killing the Yulye the men will place some of his remains in the sacred netbags in the men’s houses in order to assure fertility and growth. Such totems, bones or skulls, are painted red, a sign of warning, isa bindoblul tenen ‘that the spirits shall go’. The symbolic values of colors and coloring are woven into the context of myth and cult. The people know the stories, colors are rarely mentioned. When I heard of the Yulye for the first time, the narrator simply said “and he was dark in

. A meaningless syllable.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.22 (1260-1315)

 Volker Heeschen

the bush”. Later explanations and elaborated versions added colors, comparisons, and ideas about the meaning of the story. In the way colors are added to known and visible ornaments, the references to colored objects or beings in stories are embellishments, ways of explanatory discourse, play with meaning and wording.

. Conclusion: Signs of danger, signs of beauty There is some evidence for references to smell and colors in Papuan languages regardless of what could be thought to be basic terms: Kwoma mukushe gwonya ‘characteristic smell and aroma of a pubescent girl’, the shining red in the bush announcing the Yulye ancestor, the rhododendron redness on top of the Tengket clan’s ancestor symbolizing an unfinished state. The ways from the sphere of smells and the realm of colors into language however, are not systematic. The evidence must be anecdotal. You smell and act, stimulated by attraction or disgust, you see and pay attention to something which has to be touched and evaluated. The simulative mode of language is separated from the functional cycles of pair formation as well as from looking for, and examining, food. The speaker is freed from drives, feelings, and sensibility; he ‘reviews’ the signals which come from the functional cycles, taking the observer’s point of views, and he combines, and transfers, into a cross-modal and permanent synthesis what was momentary perception. The relationship of language to smells and colors is reflexivity. The distancing from the functional cycles leads to synthesis and reflexivity. At the same time the ideas, notions, or concepts, developing from sensibility, are semantized, generalized, and made suitable for thematization in speech and discourse. The relation from the functional cycles to the themes in discourse is not straight. Usually the olfactory and visual code, on the one hand, and language, on the other, supplement each other. There is no need to have a word for the aroma of a young girl or the redness of a fruit or flower. Each speech act is a means of focussing; whenever speakers pay attention to what they have done or perceived the signals of the olfactory and visual code may by chance and casually creep into language. The reflexive role and the thematizing function of language do not at all warrant a complete system of references to the other codes. Consequently the references are expressed by a good number of means, by nouns, verbs, adjectives, collocations, idioms, metonyms, and comparisons. A world view is pieced together by what people do and results from how the different codes supplement each other and how language reflexively mirrors the nonverbal codes. Surprisingly, the sphere of smells and the world of colors are more or less related. If one comes from the functional cycles, the relation is quite plausible. The prototypes are being attracted to, or repelled from, a person of the opposite gender, being seduced or repelled by food, examining food, wood, growth,

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.23 (1315-1382)

Attractiveness and adornment 

healthiness, and the degree of freshness or decay, and finally the exploration of the environment for signs of dangerous or useful things, animals, and humans. The olfactory faculty is not limited to nearby objects and living beings in face-to-face communication, you can smell the smoke of a faraway fire, you can be attracted by it, because it is (Eipo) asik tulum ‘the smoke of a hamlet or home’, you can be appalled, because it indicates a spirit’s home. The Yale people have a word for the smell of the approaching rain, and the hunters also smell the animal they hunt (see Smith 1998). If one accepts the logic of so-called basic color terms and does not stress the richness of lexical and grammatical means, one may say that Papuan languages usually have two or three color adjectives. The descriptive paucity is supplied by numerous references to materials showing color and being used for ornaments and body paintings. The beauty of young men or woman or the appearance of the ancestor is expressed by images referring to such materials, the colors of the morning sky, the feathers of the bird of paradise, the brightness of rivers, teeth, the different hues and colors of the sky at dawn and dusk, kinds of soils, fruits, and so on. Speakers are fully aware of the ‘colors of the world’. They are made aware of the colors by actively producing colors from soils, berries and fruits, by coloring netbags, carved arrow tips, planks of the men’s houses, ancestral sculptures, and ceremonial shields, by stressing lines and surfaces, and by decorating their faces and bodies. The world of colors is a human world, because the environment is dominated by the permanent greenness of grass and forests and the hues of dark and light. The concept of beauty combines the sphere of smells and the world of colors. The smell attracts, the shining skin pleases, the painted body excites, and the mask or painted face hide the person’s identity: the Eipo say that, having momentarily succeeded in concealing one’s identity and by being finally rediscovered, is a highlight of life, a momentum of happiness, and an act of renewal. Thus coloring and self-decoration are the symbols for the enactment of personhood. In rites the growth of plants and human beings and the wealth of the society are associated with the brightness and ‘renewal’ of the skin. The Eipo dancers show ibarye ‘steam, brightness’ and beauty. Kwoma hirika associates ‘smoke’ and ‘steam’ with the aesthetic quality or aura that makes people and other entities (e.g., ceremonial sculptures) visually attractive when they are painted, decorated with shell and feather ornaments and have magical substances [...] applied to them in ceremonial contexts [. . .]. (It is the application of magical substances in particular that is thought to make people and other objects aesthetically attractive.) (Bowden 1997: 69; see also Strathern 1981; A. and M. Strathern 1971; Gell 1975; Forge 1970)

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.24 (1382-1446)

 Volker Heeschen

However, the smells, colors, feathers and other ornaments quickly fade away. The symbolic meanings of the colors are not bound to a permanent state of beauty. The people ‘think for coloring’, what they do, is meaningful, not what the colors, symbols, and objects of art mean per se, independent of human activities and practices (see Gell 1998). The symbols and their meaning unfold during these activities, their significance is made permanent by the reflexive, simulative, and distanced mode of language, whose signifiers have lost the immediateness of signals which warn and attract. They are made suitable for discussing, assigning meanings, and being reminiscent of past smells and faded colors. As I have suggested in the two previous sections, the set of grammatically marked words for smells and colors is a by-product of these secondary or meta-practical activities, namely of speaking, narration, composing poems, and play with language. The terms of smell and color are certainly not ‘central to language’, yet central to singling out dangerous peculiarities and glancing at attractive phenomena in a relaxed narrative mode of action.

References Allen, J., Latu, M., Koesana, M., & Tsirumits, M. (1982). Dictionaries of Papua New Guinea. Vol. 6. Halia Language. Halia to English. English to Halia. Ukarumpa: Summer Institute of Linguistics. Berlin, B., & Kay, P. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Bowden, R. (1997). A Dictionary of Kwoma: A Papuan Language of North-East New Guinea. Canberra: Australian National University. Brown, C. H. (2001). Lexical typology from an anthropological point of view. In M. Haspelmath, E. König, W. Oesterreicher & W. Raible (Eds.), Sprachtypologie und sprachliche Universalien. Ein internationales Handbuch (pp. 1178–1190). Berlin: Walter de Gruyter. Brown, H. A. (1968). A Dictionary of Toaripi with English-Toaripi Index. Part 1–2. Sydney: University of Sydney. Conklin, H. C. (1955). Hanunóo color categories. Southwestern Journal of Anthropology, 11, 339– 344. Crawford, A. L. (1981). AIDA. Life and Ceremony of the Gogodala. Bathurst: The National Cultural Council of Papua New Guinea/ Robert Brown. Dutton, T. (1992). A first Dictionary of Koiari. Canberra: Australian National University. Ebberfeld, I. (1998). Botenstoffe der Liebe. Über das innige Verhältnis von Geruch und Sexualität. Frankfurt am Main: Campus. Eibl-Eibesfeldt, I. (1997). Die Biologie des menschlichen Verhaltens. Grundriß der Humanethologie. Third, revised and expanded edition. Weyarn: Seehamer. Eibl-Eibesfeldt, I. (1999). Grundriß der vergleichenden Verhaltensforschung. Ethologie. Eighth, revised edition. München: Piper.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.25 (1446-1557)

Attractiveness and adornment 

Eibl-Eibesfeldt, I., Schiefenhövel, W., & Heeschen, V. (1989). Kommunikation bei den Eipo. Eine humanethologische Bestandsaufnahme im zentralen Bergland von Irian Jaya (WestNeuguinea), Indonesien. Berlin: Reimer. Fischer, H. (2000). Wörter und Wandel. Ethnographische Zugänge über die Sprache. Berlin. Reimer. Flierl, W., & Strauss, H. (Eds.). (1977). Kâte Dictionary. Canberra: Australian National University. Forge, A. (1970). Learning to see in New Guinea. In P. Mayer (Ed.), Socialization. The Approach from Social Anthropology (pp. 269–291). London: ASA Monographs. Franklin, K. J., & Franklin, J., assisted by Kirapeasi, Y. (1978). A Kewa Dictionary with Supplementary Grammatical and Anthropological Materials. Canberra: Australian National University. Gell, A. (1975). Metamorphosis of the Cassowaries. Umeda Society, Language, and Ritual. London: Athlone Press and New Jersey: Humanities Press. Gell, A. (1998). Art and Agency. An Anthropological Theory. Oxford: Clarendon. Gipper, H. (1972). Gibt es ein sprachliches Relativitätsprinzip? Untersuchungen zur Sapir-WhorfHypothese. Frankfurt am Main: S. Fischer. Goddard, C. (1999). Review article. Color categories in thought and language. Linguistic Typology, 3, 259–269. Grammer, K. (1995). Signale der Liebe. Die biologischen Gesetze der Partnerschaft. München: Deutscher Taschenbuch Verlag. Haiman, J. (1991). Hua-English Dictionary with an English-Hua Index. Wiesbaden: Otto Harrassowitz. Hardin, C. L., & Maffi, L. (Eds.). (1997). Color Categories in Thought and Language. Cambridge: Cambridge University Press. Healey, P., & Healey, A. (1977). Telefol Dictionary. Canberra: Australian National University. Heeschen, V. (1990a). Ninye bún. Mythen, Erzählungen, Lieder und Märchen der Eipo. Berlin: Reimer. Heeschen, V. (1990b). Human ethology and semiotics. In W. A. Koch (Ed.), Semiotics in the Individual Sciences (pp. 91–121). Bochum: Brockmeyer. Heeschen, V. (1992a). A Dictionary of the Yale (Kosarek) Language (With Sketch of Grammar and English Index). Berlin: Reimer Heeschen, V. (1992b). The position of the Mek languages of Irian Jaya among the Papuan languages: History, typology, and speech. Bijdragen tot de Taal-, Land- en Volkenkunde, 148, 465–488. Heeschen, V. (1998). An Ethnographic Grammar of the Eipo Language Spoken in the Central Mountains of Irian Jaya (West New Guinea), Indonesia. Berlin. Reimer. Heeschen, V. (2003a). Semiotische Aspekte der Ethnologie: Ethnosemiotik. In R. Posner, K. Robering & T. A. Sebeok (Eds.), Semiotik. Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur, Vol. 3 (pp. 3278–3296). Berlin: Walter de Gruyter. Heeschen, V. (2003b). Linguist and anthropologist as translators. In T. Maranhão & B. Streck (Eds.), Translation and Ethnography. The Anthropological Challenge of Intercultural Understanding (pp. 115–134). Tucson: The University of Arizona Press. Heeschen, V., & Schiefenhövel, W. (1983). Wörterbuch der Eipo-Sprache. Eipo – Deutsch – Englisch. Berlin: Reimer. Heider, E. Rosch. (1972). Probabilities, sampling and ethnographic method: The case of Dani colour names. Man, 7, 448–466.

JB[v.20020404] Prn:26/01/2007; 15:28



F: CELCR805.tex / p.26 (1557-1672)

Volker Heeschen

Jäger, L. (2001). Sprache als Medium. Über die Sprache als audio-visuelles Dispositiv des Medialen. In H. Wenzel, W. Seipel & G. Wunberg (Eds.), Audiovisualität vor und nach Gutenberg. Zur Kulturgeschichte der medialen Umbrüche (pp. 19–42). Wien: Kunsthistorisches Museum. Jäger, L., & Plum, S. (1988). Historisches Wörterbuch des deutschen Gefühlswortschatzes. Theoretische und methodische Probleme. In L. Jäger (Ed.), Zur historischen Semantik des deutschen Gefühlswortschatzes. Aspekte, Probleme und Beispiele seiner lexikographischen Erfassung (pp. 5–55). Aachen: Alano. Kaeppler, A. L., Kaufmann, C., & Newton, D. (1994). Ozeanien. Kunst und Kultur. Freiburg, Basel and Wien: Herder. Koch, P. (2001). Lexical typology from a cognitive and linguistic point of view. In M. Haspelmath, E. König, W. Oesterreicher & W. Raible (Eds.), Sprachtypologie und sprachliche Universalien. Ein internationales Handbuch (pp. 1142–1178). Berlin: Walter de Gruyter. Landau, S. I. (1984). The Art and Craft of Lexicography. New York: Scribner. Lang, A. (1973). Enga Dictionary with English Index. Canberra: Australian National University. Lehmann, B. (1998). ROT ist nicht “rot” ist nicht [rot]. Eine Bilanz und Neuinterpretation der linguistischen Relativitätshypothese. Tübingen: Narr. Levinson, S. C. (2000). Language as nature and language as art. In R. Hide, J. Mittelstrass & W. Singer (Eds.), Proceedings of the Symposium on changing concepts of nature at the turn of the Millennium (pp. 257–287). Vatican City: Pontificia Academia Scientiarum. Lieberman, P. (2000). Human Language and Our Reptilian Brain: The Subcortical Bases of Speech, Syntax, and Thought. Cambridge, MA: Harvard University Press. Lindauer, M. (1990). Botschaft ohne Worte. Wie sich Tiere verständigen. München. Piper. Loving, R., & Loving, A. (1975). Awa Dictionary. Canberra: Australian National University. Lucy, J. A. (1996). The scope of linguistic relativity. An analysis and review of empirical research. In J. J. Gumperz & S. C. Levinson (Eds.), Rethinking linguistic relativity (pp. 70–97). Cambridge: Cambridge University Press. Lucy, J. A. (1997). The linguistics of ‘colour’. In C. L. Hardin & L. Maffi (Eds.), Color Categories in Thought and Language (pp. 320–346). Cambridge: Cambridge University Press. Parlier, J., & Parlier, J. (1981). Dictionaries of Papua New Guinea. Vol. 4: Managalasi Language. Ukarumpa: Summer Institute of Linguistics. Ramsey, E. M. (1975). Middle Wahgi Dictionary. Mount Hagen: Church of the Nazarene. Schleidt, M. (1995). Gerüche als semiotische Zeichen. In U. L. Figge (Ed.), Mosaik: Die Kultur und ihre Evolution in humanethologischer und semiotischer Perspektive (Acta colloquii) (pp. 35–51). Bochum: Brockmeyer. Scott, G. (1980). Fore Dictionary. Canberra: Australian National University. Smith, D. M. (1998). An Athapaskan way of knowing: Chipewyan ontology. American Ethnologist, 25, 412–432. Strathern, A. (1981). Dress, decoration, and art in New Guinea. In Man as Art. New Guinea. Photographs by Malcolm Kirk. Introduction by Andrew Strathern (pp. 15–36). New York: The Viking Press. Strathern, A., & Strathern, M. (1971). Self-Decoration in Mount Hagen. Toronto: University of Toronto Press. Trabant, J. (1998). Artikulationen. Historische Anthropologie der Sprache. Frankfurt am Main: Suhrkamp. Wassmann, C. (2002). Die Macht der Emotionen. Wie Gefühle unser Denken und Handeln beeinflussen. Darmstadt: Wissenschaftliche Buchgesellschaft.

JB[v.20020404] Prn:26/01/2007; 15:28

F: CELCR805.tex / p.27 (1672-1685)

Attractiveness and adornment

Weimer, H., & Weimer, N. (1974). Dictionaries of Papua New Guinea. Vol. 2: Yareba Language. Ukarumpa: Summer Institute of Linguistics. Wierzbicka, A. (1995). Universal semantic primitives as a basis for lexical semantics. Folia Linguistica, 29, 149–169. Wierzbicka, A. (1998). Anchoring linguistic typology in universal semantic primes. Linguistic Typology, 2, 141–194.



JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.1 (45-133)

The rich colours that we see are inventions of the nervous system rather than properties of light itself. Colour, like beauty, is in the eye and brain of the beholder. (Denis Baylor 1995: 103)

Color terms between elegance and beauty The verbalization of color with textiles and cosmetics Siegfried Wyler

The paper investigates formation and use of color designations in two specialized areas: textiles and cosmetics, lipsticks in particular. Whereas language users in everyday communication apply easily available basic terms such as red or blue, the chemical industry and paint manufacturers have developed highly differentiated systems of color designations, either lexically or numerically. Equally differentiated systems are found with the color names for textiles and cosmetics: ample use is made of the property of color names being ‘nameables’, practically every lexeme or collocation of lexemes of the language system of a speech community can be used to designate a color or a shade of color. An attempt is made to classify this extensive use of color designations. Further, the use of color names for textiles by language users is discussed: the oral use by customers or the written use in catalogues, as well as recent trends in fashion catalogues to replace language by high quality photography. Finally questions of the saleability of products and psychological aspects of customer behavior have been considered.

.

Axiomatic preliminaries

Considering the color names used with textiles and cosmetics, two fundamental statements must be made, which can actually be considered as axiomatic. a. Humans have a concept of ‘color’ or in other words: we live in a colorful world. Some theorist, was it Gerritsen or Itten, said “Our world is colorful”, and b. as Harrison (1973) in Form and Content stated, “colors are nameables”, i.e., that the color array can be segmented and be attributed so-called ‘constructed color names’.

JB[v.20020404] Prn:29/01/2007; 16:32



F: CELCR806.tex / p.2 (133-194)

Siegfried Wyler

Such reflections are based on the fact that we live in an environment in which we learnt in our childhood and later on that, by the experience provided by our optical-neural system, the objects around us are characterized by material, color, size and shape and/or taste (Minsky 1985, rev. 1988).1 That is, color is an inherent property of objects as, for example, with an orange, a tree, a flower, a rock, a stretch of open ground. This can also be supported linguistically by the fact that in multiple attributes the color adjective is always immediately in front of the reference noun or head, as in an unripe big green apple or an oval white table as opposed to a *green unripe big apple or a *white oval table, which are not part of the language. In other words: the color, no matter whether or not it is only the surface or the color of the material itself, in multiple attribution is inseparably connected with the object (cf. Wyler 1989: 218). Moreover, we learnt in our childhood and later on to attribute names to all these objects, red, blue, pink, just as it is said in the little poem for children from Bob Gill’s book (1962): What colour is the world? Suppose you asked a gardener? “Simple” he would say “my world is green” referring to cabbages, plants. Suppose you asked a milkman? “Here” he would say “my world is white” referring to milk, butter, and so on.2

All this is said with reference to the grown environment and natural objects. Colors, as the human eye perceives them, are inherent properties of nature: snow is white, grass is green, a donkey is gray. But, as the little poem above shows, human beings organize their world by attributing colors to objects. Also we must note that, in this organizing process, based on or making use of, the ‘inherent natural colors’ of environmental entities, humans bundle the actually occurring colors and shades of colors to ‘focal colors’ (‘basic colors’ to use Berlin and Kay’s nomenclature or ‘cardinal colors’ as it is used below in this article) and in certain . Cf. Minsky (1985, rev. 1988: 200): “a polyneme sets [...] agencies for color, shape, and size into unrelated states that represent the independent properties of being red, round and size” where an apple is concerned. We differ, however, from Minsky’s view insofar as we do not consider color and size as an independent property but one inherent to matter. . Cf. the passage from the philosopher Franz Brentano quoted by LeRider (2000: 200): “Namen der einfachen Ideen [...] [sind] in der Regel weniger zweifelhaft und unsicher [...]. Da sie nämlich nur eine einzige einfache Wahrnehmung bezeichnen, stimmen die Menschen in ihrer Ansicht über ihre Bedeutung meistens leicht und vollständig überein. Für Irrtümer und Streitigkeiten über ihren Sinn bleibt deshalb nur wenig Spielraum. Wer einmal gelernt hat, dass ‘weiss’ der Name der Farbe ist, die er an der Milch beobachtet hat, der wird, solange er jene Idee festhält, dazu neigen, dieses Wort falsch anzuwenden.”

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.3 (194-237)

Color terms between elegance and beauty

cases even radicalize color names, that is they ignore the actual color or shade of color and classify the entity under a focal color name (cf. Bennett 1988: 30; Wyler 1992: 89, 2006: 112f.). That is why we speak of red wine and white wine, white people and black people, or coffee can be black or white. Color names help humans distinguish, together with other properties, between objects. It is one of the characteristic properties of color designations to act as distinguishers. A child learns to distinguish between objects by being told that bananas are yellow, butter is white and marmalade is orange and that one does not eat berries when they are green (which means ‘not yet ripe’) but only when they are red (which stands for ‘ripe’). Wildgen (in this volume) refers to this process, seen as a general phenomenon in the development of individual and social communication, as “basic learning in an adequate environment”. Butter is white reveals that this organizing process may, in certain cases, be culture-sensitive: butter is white for the English child, for the Continental child it may be yellow. The distinguishing capacity of color names is, of course, also essential, if not decisive, with artefacts. These are objects which have been given a color in the process of their being produced by man, not by nature. The coloring is done with the help of paints or dyes. It is not a natural process as it is with a flower or a fruit, but an artificial process carried out based on reflections and with very specific intentions. Speaking of textiles and cosmetics, their production is based on a number of converging parameters which may be an underlying philosophy, a strategy and the prospective salability of the product as well as on consumer behavior. This would, quite obviously, also apply to other fields of commodities, furniture, wall decoration, porcelain, etc., etc. Before we discuss these parameters in greater detail, we will take a closer look at the color names used for textiles and cosmetics.

. Color naming with textiles Colors are nameables. As experiments with a fairly large number of informants have shown, for most people, in everyday life, between ten and fifteen different color names are readily available: black, white, red, green, blue, etc., and, on further reflection, the easily available names are completed by further names, such as turquoise, mauve, scarlet, etc. The availability differs from person to person although a certain consensus exists within a more or less unified community. Generally the color names that are easily available are focal colors, red, blue, orange, and where variations or shades of these colors are concerned, the focal color name serves as a macro-color name. These colors and their names, in most cases, are not only easily available but also most easily remembered. An exception to this principle is where a less frequently used color or shade of color is remembered in connection with a definite object, e.g., turqoise with certain jewelry or, within the



JB[v.20020404] Prn:29/01/2007; 16:32



F: CELCR806.tex / p.4 (237-328)

Siegfried Wyler

context of textiles, camel with a lady’s (or a gentleman’s) overcoat. Plümacher (in this volume) speaks of “a prototypical color of a well-known object”. Relating to the memory and communication of shades of colors named or of word-creations by textile or cosmetics manufacturers without the actual presence of the respective color chip or object, there applies what Wildgen states for communication of odors, namely that if “these labels are individually grounded” also communication on, as well as memory of, color “remains vague”. (Also see Wildgen’s remarks on a “selected use of color ideas” and “higher levels” of color perception in this volume.) It is generally only in special, definite contexts, e.g., paints, wall papers, textiles, that names of shades of colors are referred to. In these cases, as Plümacher points out in this volume, “the particular shade has to be circumscribed”. The textile industry, and the producers of cosmetics, lipsticks in particular, need more color designations than, say, twenty or thirty easily available color names in order to verbally specify the different shades and mixtures of colors which are characteristic of their products. Niemeier, this volume, refers to Brown and Lenneberg (1954), who maintained that the human eye could discriminate 7.5 million color differences, an estimate which is still maintained by present-day physiology (Eysel & Grüsser-Cornehls 2005). Language, as a social system, does not comply with this physiological precision and, as a social product, has, within each speech community, developed a flexible system of color designations which answers the practical needs of language users.3 Of course shades of color or colorimetric positions within the color continuum or color array can be specified by indicating nanometer frequencies, or just numerically. Imperial Chemical Industries (ICI) classified the Vibrant Jade Cluster 1 (a saturized dark green) as Dulux 42GG 08/250 or Groupe Carlin, Paris, apply what they call the Pantone Textile System to classify the colors of their textiles numerically by the mark 13-0850 TP (soleil) or 17-1525 TP (terre rouge), or Article 58185, model ombe, size 42, a ladies’ pullover, has as color description a bare 014 (designating a kind of beige). However, numerical notations mean very little to the customer. He or she is used to using color names to denote colors. Beside red, blue, yellow, which we may call cardinal color names (Grundfarben, in German), a great number of color names is required to denote the colors and shades of colors actually occurring in the textile trade. The same applies to cosmetics. Lipsticks are, very rarely, just say, focal red or focal brown, they are based on red or brown, but are very sophisticated varia. Cf. also Fahle’s conclusion in this volume: “[...] language seems not to be more precise in conveying subjective feelings about a color than it is in conveying about seemingly more internal states [. . .] language is able to cope with color, but no better than with seemingly private sensations such as emotions.”

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.5 (328-362)

Color terms between elegance and beauty

tions of their basic hue. In order to specify the almost unlimited possible positions in the color continuum between 380 and 720 nanometers verbally, a great many new names and types of names are required. The paint manufacturer Sanderson, for instance, uses such names as Rosy Morn (light pink) or Blue Daisy (light blue) to specify the 1352 different paints they produce for their Spectrum Programme.4 Names, such as cherry, sand, Sahara, green apple, light pistachio or khaki taupe are designations used with textiles, with cosmetics we find names, such as Exotic Terracota, Café Latte, Racy Purple or Rich Cocoa. Since color terms are constructed nameables, there is no reason not to use any lexeme as a color designation, especially with cosmetics. Examples are Dollar, Lullaby or Chocolate Opera, which seem to have very little to do with color. Actually on the whole, the lexemes used to denote color stem from nature or the world people live in or know. This seems to be important because only then can the names chosen elicit a color association in a person’s mind. Thus color names are coined from the areas of plants (palm, begonia, strawberry), products made from plants or fruit (cocoa, lemonade), minerals (jade), types of landscape (desert), geographic references (Aegean blue, tropical, burgundy), the atmosphere (frost blue, sunset orange), animals (chamois, coral), with cosmetics drinks, such as coffee, cocoa, sangria or cider occur beside food, such as sugar or honey. This list is not exhaustive; occasionally, certain terms occur individually, as for instance, admiral in admiral blue or newport in pale newport. The English language is extremely flexible as to creating color denominations, e.g., burgundy, cinnamon, mustard or plum can freely be used as color names whereas German would have to use zimtfarben, mustard could be senfgelb, burgundy has no equivalent in German, plum could only be expressed by a comparison. It is interesting to note the objects and their names chosen. The color designations used for textiles and cosmetics are very close to the “spectrum of hues and values entirely related to natural objects” as they are shown, for instance, in the paintings of the English non-representational painter Winifred Nicholson (quoted from Gage 1995: 27–28). In linguistic terms: the names chosen are such that they would have a feature color in componential analysis which would evoke a color sensation within the range of a corresponding cardinal or focal color. Admiral, newport or Dollar would be exceptions to this general principle.5

. Macintosh Computers advertise new iMac notebooks which can show films and digital photos in “millions of colors” (Internet information http://www.apple.com/chde/ibook/). . For a classification of color names of textiles based on Methuen Handbook of Colours and the classification suggested by Gipper (1956) see Wyler (1992: 65), MacLaury (1997: 392, A theory of categorization).



JB[v.20020404] Prn:29/01/2007; 16:32



F: CELCR806.tex / p.6 (362-414)

Siegfried Wyler

. Classification of color names for textiles The following color names used with textiles are based on such publications as the magazines Vogue, Marie Louise, the publication of the Austrian Fashion Institute, the Land’s End Catalogues, the Quelle Catalogue; the terms for Cosmetics on Color Charts published by Estée Lauder, Clinique, Virgin, Marks & Spencer. According to Harrison’s classification, the colors black, gray, white, red, yellow, blue would be called ‘natural nameables’. They belong to the deep structure from which all surface structures, the ‘constructed nameables’ are generated. The ‘constructed nameables’ are then the surface applications. Since colors are nameables their names are willfully chosen. In principle they need not depend on any given precondition. As is well-known, color names were in most cases derived from objects with a specific surface color, e.g., orange from the corresponding fruit, turquoise from the corresponding mineral. This even applies for example to yellow although it is derived from an Indo-European stem *ghel- or ghôl- found in the words gall and gold, which we do not necessarily connect with the yellow of the surface of a lemon. As has become evident from attempts at classifying color names by the German scholars Weisgerber, Gipper and Schmitter, linguistic oriented systems of classification are little suited for this purpose. More successful are either formal classifications, such as Munsell’s or that of the ICI Colour Atlas, or then practical color handbooks of the type of the Methuen Handbook of Colour where shades of hues are represented by plates of color samples with a numerical notation to which, in a separate list, verbal names are attributed. For colors of textiles or cosmetics a classification must be adapted for specific purposes. These, in the case of textiles, do not only serve accuracy but also psychological, emotional and economic purposes. In other words, the color designations have to contribute to effectuating the producer’s intentions. As we read in Dedrick (1998: 179): “Colour ordering systems [...] have no special ontological or epistemological status. Different systems of colour ordering have entirely different functions and purposes, and they record different types of data”.6 In the following we attempt a classification of color designations for textiles and cosmetics which may be called pragmatic, its parameters being on the one hand semantic, on the other hand morphological. We distinguish three main categories: (1) cardinal color names, (2) converted color names, and (3) modified color names.

. Cf. color or color name ordering with flowers, plants in the Horticultural Atlas or the adaptations of the Munsell color system for paints.

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.7 (414-504)

Color terms between elegance and beauty

(1)

Cardinal color names

(1.1)

Cardinal color names black, gray, white, red, blue, yellow7

(1.2)

Extended list of cardinal color names green, orange, pink, purple, khaki, navy, scarlet, vermilion, etc.

(2)

Converted color names: object becomes color term

(2.1)

Organic objects (flowers, fruit, etc.) cherry, herb, lemon, lime, lupine, rhubarb, sage, wisteria

(2.2)

Anorganic or atmospheric objects (minerals, sky, water) amethyst, coral, horizon, mist, ocean, sea, sky, stone, turquoise

(2.3)

Collective color names evoking color sensation within the scope of the collective expression: (hues and shades of...) autumn, forest, sahara

(3)

Modified color names: in particular modification to denote shades, tonality (brightness) and saturation of hue

(3.1)

Modification by adjectives expressing shades of colors, brightness or saturation

(3.1.1) of color term (1.1, 1.2) light gray, dark blue, soft blue, deep brown, bright red (3.1.2) of object used as color term dark amethyst, deep olive, soft begonia, pale jonquil (3.2)

Modification by object assuming color component with color term (type 1.1, 1.2) desert orange, desert khaki, admiral blue, midnight navy

(3.3)

Double modification (3.1 and 3.2 combined) dark lagoon blue, dark mineral sage, light antique indigo, light sea blue, pale Aegean blue

No names with lexemes that cannot develop a color component or fancy names outside the morphemic system of the language can be found.8

. The List is subject to alteration depending on textile manufacturers’ usage or the personal color term vocabulary of individuals. It can therefore be reduced or extended as shown in (1.2.). . This classificatory system would also apply to German color names, but – due to the German morphemic system, expressions, such as light blue would be a compound hellblau and, more importantly, an expression, such as cherry would be kirschrot or kirschfarben, violet would be veilchenblau or veilchenfarben, also orangefarben. The formation orange-coloured could not be found in the above mentioned English material.



JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.8 (504-570)

 Siegfried Wyler

In industry and manufacturing, various chemical processes are used to reach the desired hue. They are processes which result in a great many more colors, i.e., shades, blends, degrees of tonality and saturation, than is foreseen by Harrison’s system of one-lexeme color names. Yet, with a certain degree of adaptation, Harrison’s system retains its value also for textiles and cosmetics, as it generates the colors and their names which serve as starting point for the colors used in textile and cosmetics production.

. Classification of color names for cosmetics The classification of color names for cosmetics, lipsticks in particular, does not differ greatly from that of textiles, but needs some modification. This is not surprising, since certain terms used for textiles are also used for cosmetics, although, from the point of view of hue there are differences. The same name is not necessarily the same hue or shade of color. Examples are sand, heather, pistachio. Also it has to be noticed that the ‘natural nameables’ underlying the ‘constructed nameables’ with cosmetics are not black, gray, white – red, yellow, blue. In cosmetics they are red and brown, also violet. Moreover each color name for cosmetics is accompanied by a numeric notation, e.g., 09 Berry Kiss (Clinique), 397-44 Tiger Lily (Virgin), P-116 Candy (Estée Lauder). The morphological patterns used with lipsticks also occur with eye shadows or powder blush. The listings according to the classification of Chapter 3, 1.1 (red), 1.2 (scarlet), 2.1 (lemon), 2.2 (stone), are also valid for textiles and cosmetics. 2.3 Collectives (forest, all colors connected with a forest) are not found with cosmetics, since in cosmetics each item is a specific representation of a hue. 3.1.1 and 3.1.2, modification with adjectives denoting tonality or saturation, do generally not occur in cosmetics. On the other hand, the category ‘modification of color name by an adjective derived from object acquiring a color component’ is extremely fertile with cosmetics: 3.2 roselit beige, Indian red, hot purple, passionate red, starlit pink, matte scarlet, earthy red, velvet lilac, China red, icy brown. Characteristic of cosmetics, but not occurring with textiles, is the modification of a noun denoting an object by an adjective which is neither a color name nor an adjective referring to tonality or saturation: 3.2.1 rich cocoa, spiced cider, gilded mist, exotic terracotta, rouge gipsy, whereas the modification of a noun by another noun is a type found with textiles, cosmetics as well as with industrial color names (paints): 3.2.2 chocolate cream, coffee bean, honey bunch, raspberry jam, apple butter, air kiss, plum jam. Double modification (3.3), frequent with textiles, is rare or does not occur at all with cosmetics, an exception may be Red Red Red or Park Lane Purple. On the other hand, contrary to color names of textiles, cosmetics use any number

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.9 (570-633)

Color terms between elegance and beauty

of fancy color names which do not necessarily evoke any color sensation or even are, as such, not semantically intelligible in the context of color. Thus they have, in fact, just a labeling function, but one which intends to be affective or even emotional, if not merely being easy to remember: 4. tenderheart, lullaby, sinful, Clémentine, nude, kir royale, almost kissed, honeybunch, applesauce, chilli roll, all heart, forever fudge. There may exist a certain relationship between visual perception and olfactory sensation with color designations which are derived from flowers or plants, such as lilac, begonia, olive, rhubarb or products, such as chocolate or coffee. They are, in the context of textiles or cosmetics, names derived from pleasant objects, pleasant to the eye and objects with a pleasant aroma. Also note Shibuya et al.’s statement (in this volume): “Sensory co-occurrence between two sense modalities forms a sensory association of the two senses. A sensory association, however, does not only specify what types of senses are involved, but it also specifies the strength of the association.” With color designations for textiles and cosmetics, the visual sense would definitely prevail. Also note the contribution to this volume by Holz, who subsumes visual and olfactory perception under synesthesia. Moreover, we may add, as mental awareness, the suggestive effect of color names, in particular with lipsticks, but also other cosmetics such as creams, powders, etc., which produce what we might call an olfactory interpretation in the user’s or customer’s mind. Such interpretation may be “a discreet fragrance”, “a pleasant odor”, “a fragrance promising well-being”, “an odor making attractive, seducing, erotic”, and so forth. Olfactory interpretations of this nature will always be subjective. However, they lend themselves for advertising on the side of the producing firms.

. Language use and color names for textiles We may now ask: How are color names for textiles and cosmetics actually used, how do they occur. Speaking first of textiles, we notice a most interesting development in the use of color names. Here we must distinguish between oral use and written use. . Oral use or customer behavior We notice that speakers, if they refer to the color of a garment, generally stick to the color names they use in everyday life, or in other words, to the ten to fifteen easily available color names. Thus, when buying a dress or a jacket or a blouse, the customer generally asks for something in blue, black, beige or green, perhaps also camel or navy, depending on the presently prevailing trends in fashion. Accuracy



JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.10 (633-689)

 Siegfried Wyler

in color designation is not required in this type of sales talk; general indication of hue, if made at all, is sufficient. Also modification of these terms may occur, i.e., modification with adjectives indicating tonality or saturation a bright green, a soft pink. More specific naming is practically never used, be it by the customer or the sales person, with the exception of one color, namely blue: dark blue, light blue, marine, navy, royal, etc.9 If one considers the actual situation in larger fashion shops or boutiques, we notice that, in fact, very little color language is used since the items on sale are presented in rows on hangers, and the customer scrutinizes the garments offered and on show and then chooses from the displays, no reference to color or any kind of language is used or needed at all. The ensuing conversation with the sales person generally concentrates on size, material and, possibly, in general terms on the color or on the effect of the particular color (“You can wear red”; “This color makes you look pale”, etc.). . Written use Similar to non-literary texts, there occur comparatively few color names in texts of present-day magazines, such as Vogue, Annabelle, Marie Louise, etc.10 Our investigation in three copies of Vogue March, May, June 2002 produced 23 different color names in all. All of them were cardinal color names (1.1) or extended cardinal color names (1.2): black, white, blue, brown, gray, green, orange, pink, purple, red, yellow,11 and a few of category (2.1) object as color term: coral, cream, gold, indigo, navy, orange, purple, rose, turquoise, jade, lemon, lime,12 plus a small number of modified terms: jade green, pale blue, hot pink, sea green. In recent years, a great change in the style of these fashion magazines seems to have taken place. Color photography, and fashion photography in particular, . It may be interesting to note an observation made by Antal Pablé in a seminar paper (University of Zurich) that women (22 informants) referred to “freedom, independence” as predominant metaphoric implications with blue, as opposed to answers by men, who referred to “hope” as predominant metaphoric implication of ‘blue’. This may also suggest that blue garments may be a lady’s preferential choice, independent of the occasion these garments are worn for. . On the scarcity of color terms in texts see Gage (1995: 188), in particular his remark: “[. . .] it [a report] makes clear that colour as hue is not everybody’s interest, and in many contexts, we can, of course, do perfectly well without it”. . Examples are: Christian Dior: orange-brown-green chiffon garment/ Louis Vullion: gray jacket (Vogue: March 2002). . The lack or extreme scarcity of object as color term or modified color terms seems to go hand in hand with the predominance of photos as a communicative medium.

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.11 (689-742)

Color terms between elegance and beauty

has made great progress. Thus fashion photos have become, more and more, and in high-grade quality, a medium of communicating colors in textiles in the printmedia. The reproduction of color is one of the outstanding achievements of photography, besides accuracy. Instead of describing color, the focus is now on design and materials. This may help explain why the number of color terms occurring in publications on fashion and cosmetics has shrunk enormously.13 Compared to a study carried out in 1979 when six copies of Vogue yielded 271 different color names, it seems that magazines, apart from partly replacing color names by color photographs, appear to have introduced a practice which is described in Wyler (1992: 64): Regarding colour memory, Lenneberg (1954), Brown and Lenneberg (1954) found that focal or salient colour areas are most easily and most lastingly remembered which, in turn, reduces the number of memorized colours to a much smaller number than those actually ‘counted’ in the magazines. We may argue that in the light of frame theory the many colours ‘counted’ or named in magazines largely reduce themselves in the reader’s mind to assumption values or simple or basic colour.

These colors then produce in the reader’s awareness the color or group of colors which designers favor for the respective season and they are all color names which are easily available and can actually be kept in mind. Now the situation is completely different when catalogues of mail-order companies are investigated. Their catalogues are maximally explicit giving the color in print and the adhering color designation. Customers at home wish to get maximally accurate information of what they intend to order. Thus, for instance, the catalogue of Lands’ End, Direct Merchants, lists 117 different color names in the English edition and 130 in the German Edition. These names belong to all categories of the above classification of color names. Also, there are especially many modified terms, we choose the modification of blue as example: admiral blue, Aegean blue, clear blue, colonial blue, dark cobalt, blue light indigo, frost blue, blue light, sea blue, pale soft blue, true blue14 (11 tokens of ‘blue’). The difference between oral and written use clearly shows that the use of color terms depends on customer behavior and customer expectation. It also depends on cultural factors, i.e., customers’ habits, namely whether or not it is customary . The Aigner fashion catalogue for 2002 is an extreme example of this trend: it contains only illustrations of the various garments and no text at all. . ‘Blue’ in the German edition of the Lands’ End catalogue is even more differentiated: Aegeis blau, Atlantik blau, Azorenblau, blass blau, blaue Iris, China blau, Delftblau, denim, eisblau, graublau, washed tief indigo, island blau, kobalt, marine, mittelblau, Navajo blau, navy, ozeanblau, schwarzblau, tiefseeblau, wolkenblau (21 tokens).



JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.12 (742-811)

 Siegfried Wyler

in a certain country or area or group or class of people to buy clothes in large stores, in boutiques or order them by mail, phone or the internet. This again has its influence on the knowledge and use of color names by the respective individual person: some rely on macro-terms, others use more and more accurate and thus more extended cardinal, object related or modified, color names.

. Consumer behavior and function of color names This raises the questions “What are the incentives for a person to buy textiles and cosmetics?” and “Do color names have a place or a function in consumer behavior?” Of course some of the answers are trivial: people need and buy clothes, people go with the fashions, people are influenced by advertising, etc., as to cosmetics: ladies take care of their skin, they want to look good, etc. Kroeber-Riel and Weinberg (2003) distinguish between three factors which determine active consumer behavior: emotion, motivation, attitude. These raise the following questions in the consumer: a. Emotion: Is it pleasant/ do I feel well? b. Motivation: Would I like to do it? c. Attitude: Do I consider X good or would I prefer Y? Based on such – conscious or unconscious – questions, the manufacturer assumes that he or she produces a product which they can sell and that the customer may reflect what he or she will buy. In this process, color names may play a role together with other factors such as material, design or/and prestigious manufacturers or designers, such as Dior, Armani, Calvin Klein, Givenchy and others. We may now ask how color names can contribute to this process: 1. colour names are markers: they help to distinguish products or kinds of products by their color: red, green, pink, or 2. color names hint at the quality or exclusiveness of the product. Example: The color designations of the fabric of the cheaper type of office chairs of the internationally active firm Girsberger (Switzerland, etc.) are turquoise, red, ultramarine, dark blue, gray, green whereas the color names of the chairs with leather seats are more select: Atlantic, Muscat, Inka, Cherry, Royal Blue, Hop, Ivoire, Cognac, etc. The fabric of the seats of a BMW is not just green, blue or black but English Green, Marine Blue, Anthracite (an exception is Gray). This would also apply to the exterior colors of a BMW (models 2001): Titan Silver, Sapphire Black, Slate Green, Glacier Green, Toledo Blue, Alpine White, etc. An exception is Black uni. For a more

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.13 (811-906)

Color terms between elegance and beauty

extensive discussion of the use of complex color names in the automotive industry see Graumann’s article in this volume, in particular Chapter 4. With clothes black implies ‘elegance’ (cf. “black tie”, Vogue, March 2002: 261f.; German ed.: “Farbe der Diva: suggestives Schwarz. Schwarz verleiht Ihren Bewegungen Nachdruck und einen Hauch Pathos”) but may also imply mourning, intellectuality, non-establishment. 3. Names of cardinal colors or color combinations may be the trademark of a manufacturer or designer; an example would be the Burberry pattern for Burberry. 4. Color names stand for the present trend in fashions which is created by outstanding dress designers. Thus certain colors or shades of colors become the color of the fashion of the respective season, promoted by the respective designer. These colors do not have a simple cardinal color name but a name which appeals to the emotion and the attitude of the prospective customer. Therefore more resounding, more exclusive and, as the case may be, also more romantic color names are chosen: Vert amazone (Issey Miake), Rosso Chianti (Ungaro), Brouillard de Mer and Orange brûlée (Claude Montana). Of special interest are the lists of colors and colour names published for the respective season ahead by the Color Marketing Group, an international, not-forprofit association of 1,700 Color Designers. Their aim is to enhance the function, saleability and/or quality of a product and in this endeavor color names also play a role. Thus, for instance, they predict two blues, one Turquoise and the other Deep Arctic for 2003 Fashion. The names imply the trend towards nature colors, (with an underlying appeal to environment awareness), saturated chromatics and an influence of technology (with an underlying awareness of modernity). Also Ornella Bignami (Swiss Textile and Clothing Association) in a published paper read in Milan (20.09.01) on Swiss Textiles equally predicts “a mix of High-Tech and ‘pristine’ nature.” The color names planned by bodies concerned with the creation and information on trends in fashions must be meaningful. They must be bearers of hidden, implied thoughts, ideas, messages, intellectual or emotional contents.15 This becomes evident if one notes the descriptions CMG gives of the above-mentioned colors and their names chosen from Trends in 2003: Cinder Blue Silver sideswipes this mechanical Blue driving it into the cool Greys; Blue Air technology melts retro Blue in this fresh breeze from the 60’s classic cars; Deep Arctic seriously conservative, this dusty navy anchors our spirit in a safe harbor.

. Cf. also Table 1: Psycho-physiological effects evoked by colors, in Graumann’s contribution to this volume.



JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.14 (906-932)

 Siegfried Wyler

. Underlying philosophy and strategy The underlying philosophy of color names of textiles is to help present textiles, in compliance with the three principles of active consumer behavior, as a pleasant object: something that pleases and, where clothes in any form are concerned, a color in connection with material and design which contributes to the appearance of a person. As to fashion, it is also essential that the color should be in the trend or even new (e.g., New Blue). In this philosophy and in this process, the color name has its function. It may add to the emotion and the attitude of the consumer as well as to the motivation to buy. The names beside the cardinal color names are based on pleasant objects which generate positive associations: flowers and other plants, fruit, ocean, sunset, amethyst and so forth. The names chosen generally belong to the more desirable phenomena of life. Thus they assume a semiotic quality, being signs, as is said in a publication of the Benetton family: “Colour is magic, simple and bold colours not only attract people but they symbolize energy and vigour. They are the key to feelings and happiness and confidence” or as a reviewer commented on the fashion color for winter 2002/2003: “Red as expression of Joy and Optimism”. It is obvious where color denominations also have to stimulate a prospective buyer: the colors Kandinsky and Munch called “active colors” (as opposed to “passive colors”) with regard to painting are also of paramount importance in motivating a purchase with regard to fashions. Plümacher (in this volume) quotes such metaphorical expressions as exciting red, aggressive red, which, one may say, belong to the everyday vocabulary of sales persons in the fashion or cosmetics business. Obviously such statements and color names coined in the service of such interpretations are without scientific foundation, they belong to sale strategies, advertising and psychology. The names for these colors, for their part, must ring in the prospective buyer’s ear.16 This is even more obvious with the color names of cosmetics. They are even more suggestive, pointing to enhancements of a person’s appearance, bringing a person’s intentions regarding how he or she wants to look more to the foreground: to appear reserved, modest or to appear undertaking, extravagant or provocative. The almost unlimited number of shades and mixtures of colors serve this purpose, and the names ranging from descriptive to provocative play their part in this endeavor: pure orange, dark brown on the one hand, blushing nude or hot kiss at the other end of the scale. Mella (1990: 9) in the chapter “The Fascination of Colour”, comments on the physiological impact of colors in the context of textiles and cosmetics: . For criticism see Davidoff (1991: 113). Note especially his remark: “In particular, one ought to be cautious about the widely held belief that different parts of the color space are necessarily associated with particular emotions”.

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.15 (932-983)

Color terms between elegance and beauty 

The effects of colour on our minds and bodies is a subject of increasing interest. Scientific studies show that red raises the blood pressure, quickens the pulse, and increases the rate of breathing. Blue, by contrast, slows down body activity and stimulates the mind. These facts, along with other scientific and empirical evidence, are already widely used by the fashion and advertising industries for profit.

Regarding psycho-physiological effects evoked by colors, see Table 1 in Graumann’s contribution to this volume. It may be added that, with textiles, fashion in particular, or with cosmetics, the psychological impression is not only evoked by the color itself, but also by its name. The physiological impression, or effect, is important with fashion, but also with cosmetics, since the buyer of a dress or a lipstick must feel physically good when using it.

. Conclusion In conclusion one may reflect how color names of textiles and cosmetics verbalize color sensation and how they relate to color names in general. That is to say the way they are used in everyday life, orally and in writing. It is obvious that in general use the color designations of the colors we called cardinal and extended cardinal color names (categories 1.1 and 1.2) are the most prevalent together with modified terms where the modification indicates degrees of tonality and saturation (3.11). Modification by objects are, generally, very rare in general speech. It is especially in this category that color names for textiles and cosmetics differ from general use, together with color names for paints and dyes for industrial purposes. With cosmetics there is even room for free choice of words and combinations of the type adjective or noun plus noun at will. Baby Kiss is an example of this type. It can safely be said that, within the classification of color names for textiles and cosmetics we have outlined above, there is no limitation to new name formation. New seasons and new products will also, within the respective language system, create new names, although in many cases, and with color language in particular, language is presently being ousted by photographic illustrations. Verbalization of color in the context of textiles and cosmetics is not part of color science but a process of color designation in concrete situations of usage, that is beside description, in the service of bringing over intentions and effects on the buyers and users of the offered products, or as the title of this paper suggests: color names are in the service of elegance and beauty.

JB[v.20020404] Prn:29/01/2007; 16:32

F: CELCR806.tex / p.16 (983-1079)

 Siegfried Wyler

References Baylor, D. (1995). Colour mechanism in the eye. In T. Lamb & J. Bourriau (Eds.), Colour: Art & Science (pp. 103–126). Cambridge: Cambridge University Press. Bennett, T. J. A. (1988). Aspects of English Colour Collocations and Idioms. Heidelberg: Carl Winter Universitätsverlag. Benetton. Publication The United Colours of Benetton (sine die). Brown, R. W., & Lenneberg, E. H. (1954). A study of language and cognition. Journal of Abnormal and Social Psychology, 49, 454–462. Davidoff, J. B. (1991). Cognition through Color. Boston: MIT Press. Dedrick, D. (1998). Naming the Rainbow, Colour Language, Colour Science and Culture. Dordrecht and London: Kluwer. Eysel, U., & Grüsser-Cornehls, U. (2005). Sehen und Augenbewegungen. In R. F. Schmidt, F. Lang & G. Thews (Eds.), Psychologie des Menschen. Mit Pathophysiologie, Chapter 8. Heidelberg: Springer Verlag. Gage, J. (1995). Colour and Culture. In T. Lamb & J. Bourriau (Eds.), Colour: Art & Science, (pp. 175–193). Cambridge: Cambridge University Press. Gipper, H. (1956). Die Bedeutung der Sprache im Umgang mit Farben. Physikalische Blätter, 12, 540–548. Harrison, B. (1973). Form and Content. Oxford: Basil Blackwell. Imperial Chemical Industries Colour Atlas (1971). Kornerup, A., & Wanscher, J. H. (1978). Methuen Handbook of Colour. London: Eyre Methuen. Kroeber-Riel, W., & Weinberg, P. (2003). Konsumverhalten. München: Verlag Franz Vahlen. LeRider, J. (2000). Farben und Wörter: Geschichte der Farben von Lessing bis Wittgenstein. Wien: Böhlau. MacLaury, R. E. (1997). Color and Cognition in Mesoamerica: Constructing Categories as Vantages. Austin: University of Texas. Minsky, M. C. (1985, rev. 1988). The Society of Mind. New York: Simon & Schuster. Mella, D. L. (1990). The Language of Colour. London: Michael Joseph. Weisgerber, L. (1962). Die sprachliche Gestaltung der Welt. Düsseldorf: Pädagogischer Verlag Schwann. Wyler, S. (1989). The position of colour adjectives in multiple attribution. In A. Fischer (Ed.), The History of Dialects (pp. 207–220). Heidelberg: Carl Winter Universitätsverlag. Wyler, S. (1992). Colour and Language. Tübingen: Gunter Narr. Wyler, S. (2001). Colour names and text. ZAA Zeitschrift für Anglistik und Amerikanistik XLIX (Heft 1, 1.Vierteljahr), 3–16. Wyler, S. (2006). Colour Terms in the Crowd – Colour Terms in Use. Tübingen: Gunter Narr.

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.1 (45-128)

Color names and dynamic imagery Andrea Graumann The de- and encoding of simple, modified or complex color names draw back on different cognitive domains and involve different mental strategies. While basic color names directly evoke an idea about a shade and color terms modified by adjectives reinforce a specific semantic feature of a given color term, complex color terms composed of a noun and a basic color term excite complex images, in which different parts of the mental lexicon are involved. Starting out form an analysis of 250 color terms taken form a range of color pallets from different car manufactures it is shown how basic, modified and complex color terms are used in product design to stress distinct aspects of a product and reach different consumer groups.

.

Introductory remarks

The gap between the 10 million colors that the human eye can distinguish according to hue, saturation and lightness (cf. Stiles & Wyszecki 2000) and the 11 basic colors or color terms by which we categorize the entire color continuum (cf. Berlin & Kay 1969) is striking. Although language provides a wide range of means to create new terms and collocations it actually seems as if we not only categorize the color spectrum according to the 11 focal colors but also primarily use the 11 color terms in everyday conversations when naming the color of an object. Even if different psychological tests have shown (cf. Heller 2002) that people associate certain psycho-physiological sensations with a color, these sensations are not evoked or do not play a salient role in ordinary communication. That is, saying that an apple is red or a car is green does not generate a dynamic or calm sensation, but rather enables a listener to imagine a red apple or a green car. Thus, while we are not creating specific cognitive images when decoding basic color names (red, blue, green, etc.), modified color terms (warm red, calm blue, fresh green, etc.) and complex color names (tornado red, aragonite blue, cyber green, etc.) create mental images, in which different cognitive domains (imagination, lexical and empirical knowledge, social and cultural context, etc.) are involved. In advertisement and product design, mental images evoked by complex color names are used to support the overall concept of a product that by itself is aiming

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.2 (128-202)

 Andrea Graumann

at a certain group of consumers. Especially in this context the color compounds cannot be arbitrary but should be semantically motivated. My contribution is aiming at: 1. demonstrating that different mental images – drawing on distinct cognitive domains – are evoked when de- or encoding basic, modified or complex color terms and 2. illustrating how these mental images are deliberately used in the automotiveindustry to stress the overall concepts for different car models and to reach specific consumer groups. (This investigation is based on 250 color terms taken from a range of color pallets from different car manufactures such as Volkswagen, BMW, Mercedes, Renault, Citroen, Ford, etc. The different basic and complex color terms are all taken from this corpus.)

. Color classification schemes Throughout the last two thousand years researchers from different scientific areas have investigated the phenomenon of colors and have – according to their specific scientific interest – come up with different classification schemes. One of the first systematic color schemes based on physical research was presented by Isaac Newton who demonstrated that sunlight consists of rays that can be refracted by a prism. According to Newton, these refracted rays evoke certain color sensations in the human eye. In order to demonstrate the qualities and proportions of the seven simple colors (red, orange, yellow, green, blue, indigo and violet) he designed a color circle, which was first presented in his Opticks in 1704 (cf. Kuehni 1997: 133–134). Newton’s findings were not left uncriticized and especially Johann Wolfgang Goethe came up with his own Farbenlehre (1840). Yet, Goethe’s own color circle, which started out from different scientific premises and was meant to refute Netwon’s circle, shows mainly the same schematization (cf. Kuehni 1997: 135). The color schemes that have been introduced within the last three hundred years have become more and more complex and range from color cubes or hexahedrons (Alfred Hickethier) and double cones (Wilhelm Ostwald) to octahedrons (Harald Küppers). These – as well as a range of further – representational schemes of the color continuum depict the qualities and quantities of the different colors as such, show the boundaries of the distinct color spheres and demonstrate transitions between colors which color categories separate. (See also Martina Plümacher’s contribution which gives a broader overview about Relational ordering of colors in the color-circle and color-sphere.)

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.3 (202-264)

Color names and dynamic imagery

A different approach towards the categorization of the color continuum is undertaken by psychology. Within the last hundred years various psychological investigations have focused on the relation between colors and their psychophysiological effects on people. The results of these investigations have been widely used in very different areas of human lives: heliotherapy, art, interior design, marketing, etc. Already Wolfgang Goethe tried to ascribe sensual-ethical affections to colors, and the German physicist Harald Küppers claims, that Goethe’s “sensualethical classification” can be regarded as the “origin of color psychology” (Küppers 1987: 164, my translation).1 According to Goethe, colors “produce a corresponding influence on the mind” and “excite particular states of feelings” (Sloane 1991: 22). Goethe splits the color circle into two sides: a plus or positive side (encompassing yellow, orange and red) and a minus or negative side (including green, blue and purple). The colors on the positive side are said to evoke quick, lively and aspiring sensations while those on the negative side are claimed to generate restless, susceptible and yearning feelings. In a further step, Goethe describes how and what correlations between color preferences and different personalities can be drawn and differentiates the color circle into quadrants according to the following four temperaments: sanguine (yellow, green), melancholic (red, purple), choleric (yellow, orange, red) and phlegmatic (green, blue, purple) (cf. Goethe 1953 (1840); for an illustration see Gage 2001: 160). Research on sensual-ethical or – in modern terminology – on psychophysiological effects has thus a long tradition and has not only been of interest to psychologists but also to artists (cf. Plümacher in this volume) and in more recent times to product designers and advertising experts (cf. Siegfried Wyler’s contribution about color denotations for textiles and cosmetics in this volume). Within psychology first results were presented by Kurt Goldstein and Otto Rosenthal (1930), who investigated the stimulating effects of colors and showed that physical tension was lower when perceiving dark, cold colors (black, blue, green) instead of warm, bright colors (yellow, red, orange). Heinrich Frieling’s (1968) research concluded similar results. Frieling’s extensive investigations, during which he questioned about 10,000 subjects, showed that the basic colors can be segmented according to Wundt’s wind-rose of emotions and be classified as stimulating or dynamic colors (red and yellow shades), as soothing or unexciting colors (blue shades) and as stabilizing colors (green shades) (cf. Frieling 1990: 187). As a further example of investigations within psychology those of the German psychologist Max Lüscher have become widely known. During the 1970s the Lüscher-Color-Test . “Nur sein [Goethes] Beitrag zur sinnlich-sittlichen Wirkung ist bis auf den heutigen Tag von Bedeutung geblieben. Er darf als der Ursprung der Farbenpsychologie betrachtet werden” (Küppers 1987: 164).



JB[v.20020404] Prn:29/01/2007; 9:20



F: CELCR807.tex / p.4 (264-346)

Andrea Graumann

Table 1. Psycho-physiological effects evoked by colors

Red Blue Green Yellow Black White Gray

Psychological Impressions

Physiological Impressions

Dynamic, active, powerful, dangerous, desirous Distant, vast, eternal, longing Natural, lively Cheerful, happy, exuberant Powerful, sad, elegant, conservative, withdrawn Perfect, innocent Sad, distant

Warm, hot, dry, stimulating Calm, cold, wet, soothing Cool, wet, fresh, quiet Warm, light, stimulating Heavy, strength Cold, clean Fast, dynamic

was used to evaluate the personality of individuals: according to color preferences people were designated to have different personalities, different characters as well as different attitudes towards life (cf. Lüscher 1971). A more recent investigation about the connection between color preferences and psycho-physiological effects can be found in Eva Heller (2002). Heller questioned about 2,000 people about their associations between colors and 200 terms related to different topics (culture, politics, emotions, etc.). Heller’s results can be considered as further proof for the earlier findings within psychology. As some of the clearest correlations drawn between colors and psycho-physiological impressions consider the following examples: aggression – red (58% = percentage of total number of subjects questioned who drew the according correlations), heat – red (46%), energy – red (38%), desire – red (34%), dynamic – red (25%), coldness – blue (58%), calmness – blue (29%), longing – blue (27%), liveliness – green (38%), freshness – green (34%), quietness – green (40%), conservative – black (40%), strength – black (29%), power – black (48%), happiness – yellow (16%), objective – white (27%), lightness – white (37%) (cf. Heller 2002). Although Jules Davidoff stresses that “one ought to be cautious about the widely held belief that different parts of the color space are necessarily associated with particular emotions” (Davidoff 1991: 113), the comparison of a range of psychological investigations about the relation between color and psychophysiological impressions and associations reveals mainly the same classification (see Table 1). As will be shown in the next section (Section 3), these psycho-physiological effects play a salient role when modifying basic color names and composing complex color terms, which are in turn deliberately used in advertisement and marketing to reach specific consumer groups (cf. Section 4).

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.5 (346-416)

Color names and dynamic imagery

. Color names and dynamic imagery Siegfried Wyler (this volume) states in his contribution, “since color names are constructed nameables there is no reason not to use any lexeme as a color designation.” Different linguistic investigations about color names and color term modifications underline this statement and show the range of linguistic means to create new color collocations (for an extensive representation of linguistic research on color terms see Wyler 1992). So, at first sight there seem to be no constraints concerning linguistic modifications of basic color terms: the linguistic possibilities of specifying basic color terms by any adjective or noun do not seem to be restricted. And what is more, since modified color names and complex color terms are composed of lexical units that are generally known to the speakers of a language community, the de- and encoding of specific color collocations should be as easy as the de- and encoding of the basic terms. What has to be kept in mind though is that basic color words (just as other words) do not label specific and objective colors in the empirical world but rather notions or concepts that are characterized by blurred boundaries. This conception about the fuzziness of meaning has been advocated by various linguists, psychologists and neurobiologists and Siegfried Wyler’s description about the relation between mental representations evoked by (and respectively verbalized through) color terms picks these assumptions up: Colour imagination and colour memory are so to speak fuzzy notions in a speaker’s brain, default values of basic colours, which when excited and verbalized are either attached to a vocabulary of acquired colour names or then brought together within an excited frame with objects or imaginations or associations that in some way or other help to create some sort of comprehension in the listener that corresponds to the speaker’s intended colour description. (Wyler 1992: 79)

Communication about color concepts does thus not presuppose that the enunciation of the word red describes a specific and clearly determined color but rather a fuzzy idea about a shade within the entire continuum of red. The interpersonal exchange of color impressions between interlocutors does not mean that an objective truth is transferred but rather a subjective image about a specific color sphere (cf. Manfred Fahle about the subjectivity of perception in this volume). Since the same accounts for lexical units as such and neither adjectives nor nouns label an objective truth but rather vague concepts about things and events in the world, complex color names – composed of basic color words and an adjective or a noun – should not be considered as describing a color more precisely than basic color terms. That communication is possible although mental concepts and representations are vague and differ from individual to individual is founded in the fact that the different notions are equally accessible to everybody. That is, although



JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.6 (416-481)

 Andrea Graumann

people might conceptualize things and events differently, they basically have the same biological apparatus with which they perceive and experience the world and are accordingly equally constrained by this apparatus (cf. Manfred Fahle about deficiencies of color perception in this volume). Furthermore they have similar experiences in the world and in their individual cultural or social environments. What follows from this is that while our perceptive system is an innate apparatus and basic bodily experiences are universally the same, the cultural backgrounds between different language communities differ. Especially these cultural differences can be cited when it comes to explaining language varieties and distinctions. The theory that a correlation between cultural experiences and language exists is specifically linked to the Sapir-Whorf-hypothesis of linguistic relativity. Following Edward Sapir’s assumptions about the influence of language on thought, his student Edward Lee Whorf investigated Amerindian languages in order to show that language – or rather language structure – determines the mental structures and the world view of a language user (cf. Whorf 1954). While the strong form of the Sapir-Whorf-hypothesis has only few defenders, the weaker form of linguistic relativity has been widely accepted in linguistics (cf. Lucy 1996). In its weaker form the Sapir-Whorf-hypothesis states that the culture and the specific language that a person is raised and educated in is the language that that person will think and perceive the world in. Thus, culture and language are insofar two sides of a coin, as the cultural environment influences language, which in turn influences the ideas or thoughts about the world and the experiences in it. The theory of linguistic relativity has been taken up by various linguists and been mainly embodied in the investigations of cognitive semantics. The basic theory of cognitive semantics does not only encompass the idea that language is biased by culture, but also that speakers of a language society have the possibility to conceptualize empirical things and events differently, and accordingly verbalize them in distinct ways (cf. Lakoff 1986; Langacker 2002). As far as the meaning of single words is concerned, cognitive linguists have shown that the understanding of a word requires more than the knowledge of its lexical meaning. The meaning of a word is described as a semantic frame that encompasses semantic features shared by the language community as well as distinct connotations that vary from individual to individual. The semantic knowledge that a speaker of a given language community has stored depends to a great part on his empirical knowledge and on his personal experiences. Or to put it differently: “Semantics is constrained by our models of ourselves and our worlds” (Turner 1987: 7). The aforementioned psycho-physiological sensations (cf. Section 2) related to basic color terms and the complex matrixes – as Langacker (2002: 4) says – of lexemes can be put forward for the common intelligibility of complex color names as well as for their different associative features and dynamic representations. Keeping this double premises in mind, it can be demonstrated that the creation of

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.7 (481-547)

Color names and dynamic imagery

complex color collocations is not only restricted by conceptual constraints but as well that the de- and encoding of basic, modified and complex color terms rely on different cognitive domains (lexical and empirical knowledge, personal ideas, psycho-physiological impressions, etc.). First of all, it can be demonstrated, that due to the meaning extensions of the basic color terms, basic color terms will not occur with just any adjective in everyday conversation. If a basic color term is modified by an adjective, the semantic features of the chosen adjective must be compatible or overlap with the semantic features of the modified basic color term (cf. Table 1 above), since otherwise it would come to a semantic clash. In this sense modified color terms such as warm red, cold blue, fresh green, soft yellow, heavy black or clean white in which the modifying adjectives refer to physiological sensations, as well as color terms such as dynamic red, sad blue, lively green, cheerful yellow, gloomy black or innocent white in which the modifying adjectives refer to psychological states, can be considered as compounds in which modifier (adjective) and modified element (basic color name) semantically cross and can thus be grouped together. Comparing these color compounds with those in which the modifier conceptually does not harmonize with the modified color term reveals the boundaries of color term modifications. Hence, color name collocations such as cold red, warm blue, hot green, cold yellow, bright black or heavy white and calm red, dynamic blue, sad green, melancholic yellow, happy black or sad white should – from a semantic point of view – be considered as exceptions. In contrast to the color modifications that reinforce a specific semantic feature or psycho-physiological sensation ascribed to the basic colors, modifications through nouns should be judged differently. That is, complex color names composed of a noun and a basic color term (tornado red, cyber green, panther black, etc.) do not primarily aim at our psychological and/or physiological sensations but create or generate multidimensional dynamic images that encompass a different and above all a wider conceptual frame than those color terms modified by adjectives (dynamic red, lively green, heavy black, etc.). Following this line of argumentation, another point must be considered: the semantic classification of a noun. That is, some of the nouns used to compose complex color terms can be considered as common vocabulary, that each individual of a given language society has similar access to. Contrary to this, a wide range of modifying nouns makes only sense to a specific group within a society. Nouns used to modify basic color terms can thus be distinguished according to their different complex matrixes and their semantic intelligibility for different individuals. Those complex color terms constructed with nouns that label concrete objects in the empirical world and are common in our environment or experiences should be differentiated from those color compounds with nouns labeling objects, places or persons that are generally not considered as belonging to the general knowl-



JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.8 (547-612)

 Andrea Graumann

edge. As examples compare: lemon yellow vs. sari yellow, wheat beige vs. Kalahari beige or panther black vs. tarantella black. A further distinction has to be made between color modifying nouns that label objects with a proper surface color and those that no color can be ascribed to. Color collocations like tomato red, lemon yellow, water blue, pine green or bamboo gray are semantically more intelligible than complex color names like Imola red, urban gray, tarantella black, cyber green or hologram gray (these and the following complex color names are all taken from the above mentioned corpus). While the former compounds are dissolvable in such a way that the modifying noun can be used without the specific color term (tomato colored, lemon colored, water colored, etc.) or be paraphrased (as red as a tomato, as yellow as a lemon) those collocations composed of less specific or less known nouns fail to make sense as color denotations when used without the abstract color word (*Imola colored, *urban colored, *tarantella colored, etc.) or paraphrased (*as red as Imola, *as gray as urban, *as black as tarantella). These linguistic features can be explained by the different conceptual structures and mental processes involved in decoding the nouns. When decoding a complex color term the meaning of which does not refer to concrete and wellknown objects but to unspecific, abstract or rare objects, the combinatorial properties and correspondences between color word and modifying noun are not disclosed at first sight. Consider for example the relation between the compounds tornado and red or Imola and red. The semantic relation between the compounds tornado or Imola and red will not be disclosed as easily as the relation between tomato and red. The lexical meaning of the noun tornado does not reveal any specific correlation to the color red or to any color at all. Yet, the multidimensional meanings of the noun tornado encompass more than the lexical meaning, and our encyclopedic knowledge as well as our cognitive images include features about this natural phenomenon that can for example be described as powerful, dynamic or fast. And it is this knowledge that will directly evoke the psycho-physiological sensations of speed, dynamic and power in a hearer – granted of course that these features are associated with the noun tornado. Thus, upon hearing and decoding the complex color term tornado red, a hearer will automatically and spontaneously fuse his idea about the color red, his encyclopedic knowledge and personal ideas about a tornado and the psycho-physiological impressions into one image. The color collocation tornado red establishes a dynamic image, which embraces different features from different cognitive domains and integrates them into one representation (also see Susanne Niemeier about mental networks in this volume). The same accounts for the complex color name Imola red. A further point comes into play here though: the proper name Imola does in this context not primarily refer to the Italian town as such but rather to the event of the Formula One Race taking place there. The mental picture thus established encompasses our em-

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.9 (612-675)

Color names and dynamic imagery 

pirical knowledge about Imola being a town where Formula One Races take place, the idea about the race as such, the specific imagination of racing cars (in its ambiguous sense) and following from this the sensation of acceleration, speed and dynamic. The use of town names and geographical regions as modifiers for color names is in general quite common in color modifications. While some of the modifying nouns are chosen because a product with the actual shade is produced in that town (Murano red) or refers to a specific color used in a town’s heraldic (Ravenna blue) or in a nation’s flag (Japan red), others display an atmospheric scene and aim directly at the suggestive dimension exciting romantic, yearning or happy impressions: Monaco blue, Venetian green, Biscayan blue, Kalahari beige, Cyclade blue, Spanish red or even urban gray. The complexity of the semantic frames grants on one side that the basic meaning of the lexemes is understandable and contributes furthermore to a multidimensional impression that varies from individual to individual. What has to be granted though is that the meaning of the modifiers is at least in some way or the other intelligible. While this premise is mostly fulfilled, some of the complex color terms or rather some of the modifying nouns are too abstract or too specific to reveal the correspondence between modifier (noun) and modified element (color term) to an average language user. As examples consider the following color collocations: cyber green, hologram gray, tarantella black, aragonite blue or andradite green. Most of these complex color names will only be understood by a specific social group. While for example younger people will establish a complex mental image upon hearing the word cyber green older people might not associate anything at all with this color term. The same accounts for the terms tarantella black or andradite green, which will only make sense to a person who is well versed in music or in geology. In cases like these, the modifier does not contribute to establish any associations or to generate dynamic images in the above described sense. If a modifying noun is not intelligible to a speaker or not discernable by a hearer it will either be left out when naming the color of an object or it will be ignored in the process of decoding – the speaker or respectively the hearer will directly focus on the basic color term.

. Complex color terms in advertising – examples from the automobile industry One question emerges though “Why do product designers and marketing experts come up with cryptic color terms, which will not be decodable by an average consumer?”

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.10 (675-736)

 Andrea Graumann

Already Gipper remarked some fifty years ago that certain “connotations, emotional- and sound-moments” (Gipper 1955: 143) connected with complex or modified color terms are a common instrument in advertisement to excite impressions of glamour and wealth and more recent investigations show that imaginative color terms have a bigger attraction than abstract color terms (cf. Küthe & Venn 1995). That the imaginative or suggestive impetus of complex color terms is deliberately used in advertisement and product design becomes clear when investigating color descriptions from the automotive industry. That is, designing and marketing a new car successfully not only require that the technical specifications of the car meet newest standards but as well that the overall concept of the car meets the consumers’ needs and life-styles (also see Siegfried Wyler’s contribution for a discussion of color terms in adverts for cosmetics and textiles, this volume). Thus, a car advertised as fast and powerful targets at a different group than a car advertised as safe and comfortable. The technical features of a car, its name as well as the colors it is available in will influence the consumer’s perception of the car and his evaluation about this car as fitting to his own life-style (cf. Hölscher 1998 classification of consumer groups according to different life-style-types). Already the names chosen for the different car models are in many cases not arbitrary. While car producers like Daimler-Chrysler or BMW differentiate their models by plain figures (related to the engine size), other car producers borrow the car names from objects or compose new denotations. As examples consider some names used by the German car producer Volkswagen: Touran is derived from the French word tour and is supposed to excite images of long, yet comfortable voyages. Sharan is meant to associate nomadism and traveling with the family. In the eighteenth and nineteenth century luxurious carriages were named Phaetons and at the beginning of the twentieth century the word denoted a popular car-body design. Although the New Beetle refers back to the traditional German Käfer, it is deliberately called New Beetle in order to stress its modern design. (Special thanks to Mr. Kleine-Besten from Volkswagen AG (Germany) for these descriptions.) Thus, the associations that a consumer has upon hearing the name of a car model are no more left at random than the color terms used to label the color of a car. Analyzing the different color denotations for cars especially two points can be observed: while a wide range of colors and thereby of color terms are used interchangeable, a variety of colors and color terms is limited to specific car models. The reasons for this are obvious: in order to appeal to different consumer groups, car models are usually available in different colors and shades. Especially for middle-class cars the potential market is manifold and the marketing concepts for these cars should not be too specific. Hence, by offering a middle-class car in different designs (different engines, interiors and colors) a car manufacturer

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.11 (736-800)

Color names and dynamic imagery 

grants that different consumer groups are reached. Yet, while most car models aim at heterogeneous consumer groups, some car types target at a very specific group of consumers. The color denotations used for the distinct colors and shades partly mirror the underlying concepts attributed to a particular car model. As examples consider the color pallets for two very distinct car models from the Volkswagen fleet: while the New Beetle is designed as being a car for young and modern individualists, the Phaeton mainly targets at people who primarily set value on high quality and luxury. The color names used for these different car types contribute to establish these life-style images. The New Beetle is not only available in standard black, red or royal navy but as well in cyber green, speed blue and sunflower. The Phaeton on the other side is primarily available in achromatic colors and the color denotations mainly draw back on traditional musical elements: bolero beige, mazeppa gray, tarantella black, apassionata blue or nocturne aubergine. Thus for both car models the colors and color denotations are selected: while the color denotations for the New Beetle reflect the young and modern concept of the car, the color names for the Phaeton stress images of quality and tradition. Together with a multitude of other factors, the colors and color denotations used for different cars and models contribute to excite specific images and associations in a consumer’s mind. These images – and thus the products connected with these images – will be evaluated as either fitting into ones own life-style or be rejected.

. Conclusion Cognitive domains and processes differ when complex color names, modified color terms or simple color names are de- or encoded. Basic color names directly evoke an idea about a color shade and color terms modified by adjectives reinforce a specific semantic (primarily psycho-physiological) feature of the given color term. Contrary to this, the multidimensional meanings of nouns embrace a wider range of cognitive domains and thereby excite more complex mental images. When decoding a basic color name a speaker does accordingly not establish a mental chain of associations or a dynamic image. By contrast, complex color names composed of a noun and an abstract color term create or generate mental images, in which different parts of the mental lexicon are involved and must be involved in order to mentally visualize and comprehend the specific color collocation. The specific images, associations or impressions evoked by complex color terms are deliberately used in advertisement to underline the overall concept of a designed product – in doing so, advertising and marketing experts count on, and play with, the complex semantic networks establishing meaning and the denotational level of language.

JB[v.20020404] Prn:29/01/2007; 9:20

F: CELCR807.tex / p.12 (800-894)

 Andrea Graumann

References Berlin, B., & Kay, P. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Brown, R. W., & Lenneberg, E. H. (1954). A study of language and cognition. Journal of Abnormal and Social Psychology, 49, 454–462. Davidoff, J. (1991). Cognition through Color. Cambridge and London: MIT Press. Frieling, H. (1990). Das Gesetz der Farbe (1968). Göttingen and Zürich: Muster-Schmidt. Gage, J. (1994). Kulturgeschichte der Farbe: Von der Antike bis zur Gegenwart. Ravensburg: Maier. Gipper, H. (1955). Die Farbe als Sprachproblem. Sprachforum, I/2, 135–145. Goethe, J. W. (1953). Farbenlehre (1840). Tübingen: Wissenschaftliche Buchgemeinschaft. Heller, E. (2002). Wie Farben wirken. Farbpsychologie, Farbsymbolik, Kreative Farbgestaltung (1 1998). Reinbek bei Hamburg: Rowohlt. Hölscher, B. (1998). Lebensstile durch Werbung? Zur Soziologie der Life-Style-Werbung. Opladen and Wiesbaden: Westdeutscher Verlag. Kuehni, R. G. (1997). Color. An Introduction to Practice and Principles. New York: John Wiley & Sons, Inc. Küppers, H. (1987). Farbe: Ursprung – Systematik – Anwendung: Einführung in die Farbenlehre (1 1972). München: Callwey. Küthe, E., & Venn, A. (1995). Marketing mit Farben. Kampen: Dumont. Lakoff, G. (1986). Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago and London: University of Chicago Press. Langacker, R. W. (2002). Concept, Image, and Symbol. The Cognitive Basis of Grammar. Berlin and New York: Mouton de Gruyter. Lucy, J. A. (1996). The scope of linguistic relativity: An analysis and review of empirical research. In J. J. Gumperz & S. Levinson (Eds.), Rethinking Linguistic Reality (pp. 37–70). Cambridge: University Press. Lüscher, M. (1971). Der Lüscher-Test: Persönlichkeitsbeurteilung durch Farbwahl. Reinbek bei Hamburg: Rowohlt. Sloane, P. (Ed.). (1991). Primary Sources: Selected Writings on Color from Aristoteles to Albers. New York: Design Press. Stiles, W. S., & Wyszecki, G. (2000). Color Science (1 1982). New York: John Wiley & Sons Inc. Turner, M. (1987). Death is the Mother of Beauty. Mind, Metaphor, Criticism. Chicago and London: University of Chicago Press. Whorf, B. L. (1956). Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf, J. B. Carroll (Ed.). Cambridge, MA: MIT Press. Wyler, S. (1992). Colour and Language: Colour Terms in English. Tübingen: Narr.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.1 (45-121)

From blue stockings to blue movies Color metonymies in English Susanne Niemeier The contribution discusses metonymic color expressions in English, with a focus on the color ‘blue’, as meaning extensions of basic color terms. Color metonymies can on the one hand be seen as culture dependent, but on a deeper level of analysis seem to rely on universal natural features, which may predestine them for being used in emotional expressions. After a summary of research results on color terms and a discussion of the concept of metonymy, examples concerning the color ‘blue’ are analyzed within a radial network of meaning. In a last step, the potential of such networks as an inroad to intercultural learning within foreign language teaching is pointed out.

.

Introductory remarks

As indicated by the title, my contribution focuses on color expressions in language; however, it will concentrate on only one color, namely blue, on only one language, namely English, and on one example of how colors are used in language, namely in metonymies. Similar analyses could be carried out for other colors (see Niemeier 1998) and for other languages, and could then be used as a starting point for discovering intercultural differences as well as similarities. Although culture dependence is a major issue when comparing the actual metonymies in question, there is still some indication that similarities between cultures are discernible as well. This is especially so when the problem is seen from a more abstract angle, since color metonymies ultimately seem to rely on universal natural features and furthermore seem to have a tendency to be used for emotional expressions. Colors can be considered – and have been considered – as concepts that are relatively easy to describe, at least within a specific culture. According to Brown and Lenneberg (1954), 7,5 million minimal color differences can be discriminated by the human eye according to hue, luminosity, and saturation. However, in this three-dimensional continuum, no discrete color categories can be distinguished. Are the discrete categories that are used by people then the product of their cultures and languages (cf. Lucy 1992: 185ff.)? In this case, we should be able

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.2 (121-199)

 Susanne Niemeier

to discover culture-specific color expressions, for example in metonymies. Color can be regarded as a concept widely internalized and widely shared within a language community and therefore as firmly rooted in our minds, prone to give rise to numerous meaning extensions in diverse types of metonymy. In my contribution, I will address four issues. First, I will summarize some results from research on color terms that are relevant for my topic. This is followed by a brief characterization of the concept of metonymy as it is discussed in Cognitive Linguistics (see also Niemeier 2000). Next, I will discuss one specific kind of metonymy (i.e., color metonymy – using examples for blue) as well as explain the radial network of meaning among the diverse metonymic expressions presented. Finally, I will relate this network to various trends within foreign language teaching methodology as it seems to have considerable potential for enhancing intercultural learning as well as vocabulary learning.

. Research on color terms Research on the lexical expressions for colors has attracted many scholars (cf. Brown & Lenneberg 1954, or the famous Berlin & Kay study from 1969). Berlin and Kay claimed to have found semantic universals in the area of color coding by showing that color terms are not located on a semantic continuum but that there are certain focal points in this continuum which are conceptually more salient than others. One of Berlin and Kay’s (1969: 2) results is that all languages incorporate a relatively small set (between 2 and 11 – or even 12, as Berlin and Kay claimed later on) of what they call a “universal inventory of basic color categories” and which they claim to be perceptual referents, i.e., “focal colors”. Their claim reversed the arguments of earlier research on color terms insofar as for them, these focals [...] served as the referential basis for linguistic color categories rather than its being the case, as previously proposed, that the linguistic color categories served as the basis for segmenting the color continuum. (Lucy 1992: 179)

Berlin and Kay also pointed out that these basic color terms appear in a fixed sequence, thus in a language with only two color terms, these terms would designate something like the English terms black + cool colors, i.e., blue, green and white + warm colors, i.e., red, yellow, orange, purple, pink. A language with a third term would add red, the fourth and fifth terms would be yellow or green (no fixed sequence was detected here), then blue, then brown, and then gray, orange, purple, and pink in no fixed sequence either. However, this does not mean that the color terms in those languages are readily transferable to other languages; Russian, for example, has two focal terms for blue: goluboy for a light, pale blue and siniy for a dark, bright blue (cf. Paramei 2005 or

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.3 (199-269)

From blue stockings to blue movies 

http://en.wikipedia.org/wiki/Blue#Blue_in_Russian_language). These two colors are considered to be two different colors and not just two distinctly different shades of one color, as is the case in English or German, for instance – a view that is, however, not shared by all researchers. Furthermore, Welsh has the focal color term glas, referring to a color in the range of the English blue, green, and gray. We find a similar distribution of color terms in Irish Gaelic, where glas refers to green and gray (cf. Lazar-Meyn 2004). For the present purpose, we will leave these non-coincidences aside as English is the target language. However, these results were not left uncriticized. Lucy (1992: 186) argues that these earlier studies (i.e., Brown & Lenneberg 1954; Berlin & Kay 1969) on color terms are based on Westernized (i.e., English) concepts and thus implicitly presuppose that the English worldview is a ‘natural’ given. He further claims that the meanings of those color terms which were elicited in experiments might not necessarily be the central meanings of the words used. Lucy points out that the studies focus only on the denotational meanings of the color terms in question and neglect their possible meaning extensions, their relations to other lexemes, and the general status of one color term in the context of the set of color terms established in a given language. Indeed, Lucy heavily criticizes the very concept of ‘denotational meaning’. Meaning extensions that exist in a language, e.g., in the form of metaphors and metonymies, necessarily contribute to the central meaning of the term as well insofar as when this term is processed in the brain, the network of meaning extensions becomes activated simultaneously. In line with this argument, Lucy also laments that the pragmatic use of these color terms (i.e., the cultural differences in their usage) has never been analyzed, which might have led to a considerable amount of blurring of the differences between languages. An analysis focussing on color metonymies should aim at integrating all the subtle meaning nuances of the color terms that may appear in the language analyzed. These findings would of course make even more sense in cross-linguistic studies, especially of languages which are not closely related. Such a cross-cultural perspective cannot as yet be offered here, but we will try to work out the possible semantic network of meaning of the basic English color term blue and at least at some points contrast the findings to German.

. Metonymy as seen within Cognitive Linguistics I will proceed by briefly defining metonymy as it is seen within the framework of Cognitive Linguistics. Most lexemes are highly polysemous (i.e., have more meanings besides their basic meaning). This basic meaning can be extended, and one of these meaning extension strategies is metonymy (cf. also Dirven & Verspoor 2004: 32ff.). In metonymy, we conceive of a whole scene but mention just one

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.4 (269-321)

 Susanne Niemeier

aspect of it, usually one that is particularly salient or highlighted. If we say, for example, “the White House has passed a new law”, our world knowledge tells us that a house does not pass laws, but that the expression “the White House” refers to the people working there (i.e., the President and the President’s advisors) and also to the work processes taking place there. Or if we imagine a restaurant scene in which one waiter says to another: “The ham sandwich at Table 4 has asked for the bill”, we understand this automatically as referring to the person at Table 4 who consumed a ham sandwich – because we would hardly ever think of ham sandwiches sitting in restaurants and asking for bills. Thus, such a sentence is easily understandable and makes perfect sense in a restaurant scenario. Sometimes, the item in question does not even have to be present in the scene, but is still part of our mental picture of this scene. Thus, in a hospital scenario a nurse might say: “The appendix in room 33 is calling for the doctor.” Now, presumably, this appendix has already been removed, but patients in a hospital are frequently differentiated in a metonymic way by the names of their illnesses or affected organs. Thus, it is obvious that metonymy is a pervasive strategy which we use all the time without generally even being aware of it. However, although we have a large choice of possible metonymies, this choice is not an unlimited one, as many metonymies rely on the background culture or subculture. Thus, the “White House” makes no sense in a culture where we cannot connect it with the President’s domicile, and a real appendix lying in a hospital bed and calling for the doctor would indeed be highly unusual, presumably in every culture. . Color metonymy: Blue – different aspects of meaning In the following paragraphs, I will focus on color metonymy. I have chosen color terms as an example because one might be tempted to think that colors are quite easy to describe by their wavelengths, saturation, luminosity and hue, but in everyday language we only rarely use them in such a precise way. What we generally do when we use a color term is to use metonymy. Colors have motivated a lot of metonymies due to their immediate availability as sensory input. These metonymic meanings are not ad hoc meanings, as they are all motivated by different aspects of the culture in question, even though the colors themselves are not culture dependent, as generally all of us humans are equipped with the same physiological apparatus to perceive them. We normally already use metonymy when we think we are using a color term in a very precise way, for example, when we say of somebody that s/he has blue eyes. Normally, this kind of blue is not exactly the prototypical, basic kind of blue, it may be somewhat darker or somewhat lighter or interspersed with other colors – therefore, we generalize when we just call it blue. Generalization is considered as one kind of metonymy (cf. Dirven & Verspoor 2 2004, especially Chapter 2), as

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.5 (321-378)

From blue stockings to blue movies 

we refer to more than the exact referent or the basic (core) meaning when using the lexical item in question, but stay within the same domain, we just widen the focus to a certain extent. However, the range of possible shades of blue is not an unlimited one, as certain marginal kinds of blue – such as cornflower blue, for example – are excluded, a phenomenon which is based on the speakers’ world knowledge and which may be due to the fact that the object relationship already hints at the range of possible nuances of blue to the exclusion of others (cf. the contributions by Plümacher; Wyler). Within Western culture, the basic (core) meaning of blue refers to the natural blue surrounding us every day, e.g., the blue sky or the blue sea (even if both of them are not always that blue – in this case we are again dealing with generalization). But there are a lot of metonymies (sometimes also referred to as idioms) that are meaning extensions from this basic meaning. Thus, we may come up with color expressions that even contradict each other, but this only highlights the fact that the meaning extensions have gone in different directions and that they rely on different aspects concerning the background culture. . Metonymy vs. metaphor I want to illustrate the claim that meaning extensions can go into quite different directions by using the example of bluestocking, a term which today is somewhat outdated but still in use, not only in English, but also in German (Blaustrumpf ), in French (bas bleu) as well as in other European languages, due to cultural transfers. If we trace back the term to its original meaning (cf. http://concise.britannica.com/ebc/article-9357534/Bluestocking), we find as its basic meaning the reference to silk stockings of blue color, which a certain Mr. Benjamin Stillingfleet wore to the meetings of a literary circle, where the usual dress code prescribed black silk stockings. In a first metonymization, Mr. Stillingfleet himself was then referred to as bluestocking (part-forwhole metonymy). After a while, all people attending literary circles at a given place and time (i.e., London in the eighteenth century) were called bluestockings (generalization). Again after a while, this meaning narrowed down and from then on only referred to the women attending literary circles (specialization, another type of meaning extension via metonymy). Finally, in a last metonymic step, the term started to be used to refer to a certain type of women with pedantic literary and/ or academic interests. According to Crook (cf. http://www.rc.umd.edu/features/features/chambermusic/crook.html), these bluestocking feminists characteristically downplayed sexuality and could therefore be seen as prudish, as this was a hypothesized characteristic of women who were interested in books and literary discussions.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.6 (378-447)

 Susanne Niemeier

Comparing this notion of prudishness to the expression blue movies (i.e., pornographic films), the latter one seems to entail quite an opposite meaning, but it is a metonymy as well, just having gone into a different direction. This expression dates back to the days of Puritanism in Northern America, where certain laws against sinful behaviors, which happened to be written down on blue paper, were passed. These laws referred to drinking, gambling, sex, and business hours. After a while, this original meaning was narrowed down and referred just to laws dealing with sex and pornography, which then gave rise to expressions such as blue laws, blue movies, blue jokes, to turn the air blue and the like. What is important here is that we can only understand these metonymies as metonymies if we know how they developed (i.e., if we are aware of the different metonymic steps). If we do not know these steps, we rather perceive these expressions as metaphors (i.e., as not objectively related to the basic meaning). At least in a Cognitive Linguistic perspective, this seems to be the basic difference between metaphor and metonymy – it is, so to speak, in the eye of the beholder if s/he understands something as a metaphor or as a metonymy, as – possibly – there is no basic difference between these two meaning extension strategies because they may be seen as the two poles of a continuum. Some cognitive linguists even go so far as to claim that every metaphor ultimately relies on metonymy (see Barcelona 2000). . Radial network of blue metonymies in English So far, we have seen just two examples of meaning differences concerning expressions with blue. I would now like to present the whole radial network, which is certainly by no means complete because language constantly changes and new meaning extensions are created all the time, some of them getting entrenched and others not. In the following paragraphs, I will explain some of the examples in detail and show why they can be considered to be metonymies. We will also see that many of the metonymic links are culture dependent and that some even rely on mere coincidence (as the blue paper for the blue laws) and that therefore intercultural translatability is not automatically a given. Collins Cobuild lists 25,087 collocates for blue, and there is a further node for blueprint with another 1,058 entries, which may be added to the total number. In the BNC (British National Corpus), there are 10,476 instances of blue to be found. Thus, it is definitely a term that is widely used in present-day English and certainly not only in a denotative way. Although blue is quite often used to denote nothing else but the color blue, this does not necessarily refer to a focal blue, as mentioned above, and furthermore one can never rule out the tacit co-presence of other – metonymic – meanings, depending on the cotext and the context.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.7 (447-494)

From blue stockings to blue movies 

The basic meaning of blue refers to natural entities like the sea or the sky which are highly variable and therefore the meaning spectrum of blue is not as firmly fixed as that of other colors which do not show such a lot of variation in nature – this is probably why some languages, such as Welsh, have grouped the color ranges of blue, green and gray together under the label glas (sometimes rendered in British English by the term grue). Another facet of the basic meaning is the reference to eyes, which are usually among the first items to be discovered when looking at a person’s face. This is what I have called ‘universal blue’ as it is present in every culture. Beyond these universal meanings, we find notions that are relatively wide-spread throughout the Western world, and finally we find metonymies that specifically rely on an Anglo-British cultural background. I will now explain some of the examples in more detail. Based on the universal meanings, it may be seen as a logical consequence that the most frequent collocation to be found is blue eyes. This expression can be taken in its quasi-basic meaning as just designating the color of somebody’s eyes. On the other hand, there are also metonymic variations (i.e., blue-eyed boy), referring to somebody special and outstanding, probably because blue eyes are less frequent than for example brown ones. Consequently, to receive the blue-eyed boy treatment means to be treated very favorably. In German, however, we find the expression blauäugig etwas tun (literally: ‘to do sth. in a blue-eyed way’), denoting naiveté and thus highlighting another aspect of the concept, namely that of presumed innocence and lack of experience, which may have to do with the fact that all babies are born with blue eyes. Thus, in both examples, one salient aspect of a person is named in order to refer to the whole person or process, which is a clear instance of metonymy. There is another expression in German, blaues Auge (literally: ‘blue eye’), which is rendered as black eye in English and which refers to the fact that somebody was hit on the eye and that there are bruises all around it as a consequence. Germans tend to highlight the bluish undertones of these bruises, whereas English focuses on the blackish undertones. None of these expressions refers to the actual color around the eye itself, but they are motivated by the diverse colors of the skin around the eye. The second most frequent collocation encompasses another aspect referring to natural blue, namely blue sky. Again, this in itself is not necessarily a metonymy, but in the English language has given rise to metonymies, such as the US usage of blue sky laws. These are laws aiming at protecting investors from buying valueless securities which are metaphorically seen as pieces of blue sky, of which there is a great deal. Other examples in this respect are to blue one’s money (i.e., to spend lots of money), or blue sky research (i.e., research for ideas only). The sky is also the domain of reference in phrases such as out of the blue and a bolt from the blue, both expressing a high degree of unexpectedness of events, which has to do with the belief that God’s actions in heaven are unforeseeable. Here, only the sky’s color

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.8 (494-567)

 Susanne Niemeier

is enough to refer to the whole scenario and we are thus dealing with a part-forwhole metonymy. The prototypical color of the sky is certainly one of the first ideas that come to mind when we hear the word blue, the sight of it being also one of our first visual human experiences. The abundance of available blue sky motivating the above-mentioned metonymies stands in contrast to another group of metonymies denoting value. The best-known examples of this group are blue diamond or blue rose, both referring to something impossible in nature, as well as blue chip, originally denoting a poker chip of great value but now generally meaning ‘value’, or blue ribbon meaning referring to quality or first place. Here, the unattainability or scarceness of the objects makes them especially desirable and thus valuable. In these cases, we are dealing with a generalization of meaning and thus, again, with metonymy. The color ‘blue’ also plays a role in several illnesses, such as blue tongue (a sheep’s disease), blue baby, or blue devil (delirium tremens induced by alcohol), where the symptoms metonymically give these illnesses their common names. We might also want to mention the expression blue balls as a metonymy for a quasiillness, referring to the pain a male person feels after having had an erection for a very long time without an opportunity for release. There are also other bodily features which are metonymically used to refer to specific entities or events: apart from the blue-eyed boy, we find the expression blue blood which is widely used in the Western world and which refers to the generally fair skin of Spanish aristocrats that lets the veins shine through, as compared to the Arabian conquerors (a meaning that is close to the meaning of ‘value’). These bodily features may include mental features as well: death is visible from the skin (i.e., in this case livid, bloodless skin), therefore mental depression is seen as a mental death, when one no longer feels anything. Thus, mental depression is referred to by the term blue as well, such as in to feel blue or to have the blues. This usage of blue(s) has been carried over to the domain of music where we find expressions such as blue note or to sing the blues, music which is generally associated with the above-mentioned emotions. Emotions are also at stake when people scream blue murder, or when they experience an attack of blue despair or blue fear. Here it seems that the concept of an emotional excess may have motivated the choice of the term blue in these meaning extensions, which would at the same time also connect them to the blue sky group of expressions mentioned before, where the notion of abundance is present as well. Another frequently used metonymic source is clothing, where the salient aspect of color stands for the whole concept. The best-known example is perhaps blue-collar worker, referring to only a part of the worker’s outfit in order to designate the person and their social status. Other examples in this respect are bluejacket for a sailor, bluecoat or blue flu for the police as well as the already mentioned bluestocking.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.9 (567-639)

From blue stockings to blue movies 

A further notion shared in the Western cultural tradition seems to be the connection between the color ‘blue’ and the idea of romanticism. The concept of the Blue Flower or the blue moon is widely used in poetic language – these examples are again connected to the ‘scarceness’ group of expressions, as the objects in question hardly ever occur. Thus, the expression once in a blue moon refers to a very rare incident. Another example would be the use of the color ‘blue’ to speak of liberalconservative politics and politicians. Possibly, as the color blue is used to refer to wide expanses (sky, sea), it also encompasses an idea of ‘freedom’. Liberal parties tend to call themselves Party of Freedom or Free Democrats and thus, again, the color term metonymically stands for a larger concept than just the mere denotation of color, and, just as was demonstrated for bluestocking, there are different steps of metonymization involved. In most of the examples, it is evident that one needs cultural background knowledge in order to be able to explain the metonymies. Then it becomes obvious that the word blue may cover even such contrasting notions as sex vs. prudishness, or abundance vs. value. Blue itself does not change its meaning, but the expressions as a whole, which incorporate this lexeme, may orientate their meanings in different directions. This again goes to show that even totally ‘innocent’ and supposedly easy words are not culture-free concepts and also that translation as a one-to-one meaning transfer is just not possible. Still, the notion of ‘blue’ is accessible to everybody because humans are daily surrounded with blue entities. And this is why this concept is very much on their minds and thus offers itself to metonymic usage. The core meaning of the blue metonymies seems to lie in their reference to ‘vast expanses’ and, connected to this meaning aspect, to ‘liberty’. Whereas the first-mentioned aspect is definitely used in many other cultures as well, the second aspect is much more culture-specific and as such more closely related to the culture’s language and social history. The polysemy of the term blue is definitely a given, whereas its representation in a radial network goes back to ideas from Cognitive Linguistics which analyzes lexico-grammatical entities as being organized around prototypes. Thus, what I have called the ‘universal meaning’, can also be seen as the prototypical meaning. The other meanings (i.e., the meaning extensions), which I have summarized in the above paragraphs, are more marginal exemplars of the category ‘blue’, but they are nevertheless still connected to the core meaning. Although these marginal exemplars need not be connected amongst themselves (as we have seen when discussing bluestocking against blue movies), they are still all explainable via the core meaning or basic meaning.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.10 (639-652)

 Susanne Niemeier WESTERN WORLD

color of nobility blue blood

color of liberal politics blue politics blue party

color of romanticism “Blue Flower” blue moon

UNIVERSAL blue

clothes color/people blue collar worker bluejacket bluecoat blue flue bluestocking

ccc

color of value blue rose blue diamond blue chip blue-eyed boy color of abundance blue sky laws blue sky research blue one’s memory color of excess blue murder blue despair blue fear

color of naturally blue entities: -sea -sky

color of unexpectedness a bolt from the blue out of the blue

ANGLO-BRITISH WORLD color of illness physical blue devil blue tongue blue baby blue balls mental/depression to leave the blues to feel blue

color of immorality blue laws blue movie blue joke turn the air blue

color of editing blueprint blue-pencil

color of music blues blue note

Figure 1. Radial network of metonymic extensions of blue

. Connection to newer approaches in foreign language teaching methodology Representing meaning connections in radial networks seems to be a valuable instrument when it comes to foreign language teaching, not at least because visualization is an important means for enhancing learners’ understanding. But more important is presumably the relation of the single expressions to a basic meaning. Generally, metonymic or idiomatic expressions seem difficult to teach or to learn, but once they are connected to a core meaning, learners have the possibility of creating their own mental networks of meaning and thus also to create ‘meaning’ for opaque expressions. And if meaningful items are connected to other meaningful items, learning becomes a lot easier. It has been shown in a couple of experiments that retrieval is a lot quicker that way (cf. Lowie & Verspoor 2004; Boers 2004). Being able to see the meaning behind the different expressions entails understanding and maybe even a different perspective on language altogether, not only of the foreign language but also of the mother tongue. What could for example be done with such a radial network in the foreign language classroom is to use metonymy as an inroad to showing up intercultural

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.11 (652-709)

From blue stockings to blue movies

differences between the native language and the target language, as intercultural learning is one of the foremost aims of – at least – German foreign language education. Language is closely connected to culture; thus, knowing the language or knowing about the language also entails knowledge about the culture. The similarities and differences that show up may raise the learners’ curiosity concerning the other culture. A radial network can also be the basis for classroom exercises concerning vocabulary, either leading to vocabulary extension or to a restructuring of already existing knowledge. Learners are enabled to see how meaning is created and changed, and they will then presumably pay more attention to the words and expressions they use on a daily basis. Furthermore, the concept of radial network can be of use concerning lexicographical work, i.e., learners (especially advanced learners) could be asked to find out the meanings of the expressions in question in dictionaries (traditional ones or etymological ones) or thesauruses and to paraphrase them. Sometimes of course, these explanations will remain mere speculations, but that is not the crucial point, as ‘playing with language’ and building up one’s own constructs of meaning is one of the most important aspects of foreign language learning and teaching, as it sensitizes the learners to meaning connections within the language and to the various ways meaning extensions can take within a given culture. At the same time this way of working with words caters to the affective side of learning, as it is fun for the learners to come up with self-constructed explanations and analyzes. Based on this, working with metonymy in the foreign language classroom could also be extended to creative tasks, such as drawing similar diagrams for other languages and discussing the differences, or to creative writing, for example creating ‘intercultural jokes’, such as an English boy asking a German one “Do you like blue movies?” and the German boy answering “No, I actually prefer them in black-and-white”. As affective and creative ways of learning enhance the learners’ motivation, they should not be neglected in any classroom. Finally, I would like to widen the focus and connect the above reflections to current trends in foreign language teaching methodology, because when I claim that the concepts of color metonymy and radial networks are useful in the foreign language classroom, this does not entail that they should stand apart from other approaches, but that they are compatible with some of them and can therefore supplement them. First, I would like to focus on what is called ‘holistic learning’, following Bach and Timm (3 2004) or Weskamp (2001). This approach maintains that – in contrast to traditional, mainly cognitively oriented methods, aiming just at the left hemisphere – the whole person should be taken into account, with all senses involved and both hemispheres active. This is highly compatible with the ideas advanced in this paper, which can also be seen as holistic – although not geared towards the learner, but towards language – as they do not separate between language and



JB[v.20020404] Prn:29/01/2007; 9:30



F: CELCR808.tex / p.12 (709-786)

Susanne Niemeier

culture because in analyzing the metonymies in question, both these aspects are present: on the one hand, humans’ everyday knowledge is highlighted and on the other hand, language is seen as an integral part of culture. Second, I want to dwell on the role of affection in language learning. Already Krashen (1981) mentioned the importance of the so-called affective filter for learning. Learners are more motivated when learning is fun and when they feel at ease. Concerning foreign language learning, fun is more easily achieved when it is connected to stories, games, creativity and authentic language usage than merely with grammar drills and word lists, where the learners may have trouble seeing the sense of what they are supposed to learn. The approach advanced in this paper suggests asking learners to be creative and to explore the motivation for peculiarities in the foreign language and thus to look for ‘sense’ and meaning. As Langacker (2000: 9) said, “It is not the linguistic system per se that constructs and understands novel expressions, but rather the language user, who marshals for this purpose the full panoply of available resources”. Thus, the language learner him-/herself is in the center of interest and should, in an ideal case at least, learn to consider the foreign language as a flexible instrument. We may also want to consider the notion of action-oriented or processoriented learning (in German: Handlungsorientierung with a slightly different meaning than the translations just given, cf. Bach & Timm 3 2004). In this approach, language is seen as an instrument for linguistic action or activity, and learners are to be exposed to language not in an abstract, rule-governed way but they are expected to use it, according to communicative strategies and always oriented towards a result. Especially when we think of the metonymic and/ or idiomatic use of language, it becomes obvious that this kind of language belongs more to the oral competence than to the written one: it is language that is rather to be used in contact with native speakers than in (quasi-) academic writing. Therefore, enabling learners to deal with metonymies/ metaphors/ idioms is enhancing their communicative competence, which again is connected to their dealing actively and flexibly with the foreign language. Another aspect worth mentioning in this respect is language awareness raising. Language awareness has been defined as “a person’s sensitivity to and conscious awareness of the nature of language and its role in human life” (Donmall 1985: 7). When dealing with metonymy analysis, learners are made aware of the cognitive strategies underlying language usage and meaning extension and they are thus guided towards noticing these strategies in the language input that they receive. As a side-effect, they will be enabled to think about language and are provided with explanations for linguistic phenomena. This is not only true for lexis and for cultural aspects, but for grammar as well. Language is culture and transmits culture and therefore an understanding of the language entails an understanding of the culture.

JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.13 (786-829)

From blue stockings to blue movies

This brings us immediately to the last point, namely intercultural learning, understanding of ‘the other’ (e.g., Kramsch 1993). The goal of intercultural learning or intercultural competence has so far been mainly pursued by texts and sometimes audio and video materials, but within the approach under discussion (or, in a wider perspective, within an Applied Cognitive Linguistics approach) it can also be part of language analysis and vocabulary work. If, for example, metonymies are dealt with in a contrastive way, learners see the different ways in which the cultural background may influence language change and thus, learn about cultural similarities and differences. At the same time, they may perceive the ‘situatedness’ of language usage and may acquire schematic social knowledge about the target culture, as the choice for a particular linguistic element or phenomenon is at the same time always a reflection of its social usage and thus a symptom of its cultural contextualization. Summing up, a close analysis of color metonymies in English does not only aim at a better understanding of vocabulary and idioms, but also has a functional side insofar as it may be used for foreign language teaching in such a way that learners of the language may get an insight into how the foreign language – as well as language in general – works. This may make it easier for them to use the foreign language adequately in diverse communication situations – which is one of the most important goals in the foreign language classroom. Foreign language teaching and learning is of course not the only applied field in which color metonymies are of interest. As other papers in this volume point out (cf. the contributions by Graumann; Holz; or Wyler), decoding metonymic relationships is also relevant in many other contexts, as for example in advertising. Playing with language is not only fun, but on a deeper level refers to the speakers’ world knowledge and cultural socialization, which are often metonymically encoded.

Acknowledgement I would like to thank David A. Victor (Eastern Michigan University, USA) for checking my English. All remaining flaws and inconsistencies are of course entirely my own responsibility.

References Bach, G., & Timm, J.-P. (Eds.). (3 2004). Englischunterricht. Grundlagen und Methoden einer handlungsorientierten Unterrichtspraxis. Tübingen and Basel: Francke.



JB[v.20020404] Prn:29/01/2007; 9:30

F: CELCR808.tex / p.14 (829-936)

 Susanne Niemeier

Barcelona, A. (Ed.). (2000). Metaphor and Metonymy at the Crossroads. Berlin: Mouton de Gruyter. Berlin, B., & Kay, P. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Blank, A. (1993). Polysemie und semantische Relationen im Lexikon. In W. Börner & K. Vogel (Eds.), Wortschatz und Fremdsprachenerwerb (pp. 22–56). Bochum: AKS. Boers, F. (2004). Expanding learners’ vocabulary through metaphor awareness: What expansion, what learners, what vocabulary? In M. Achard & S. Niemeier (Eds.), Cognitive Linguistics, Second Language Acquisition, and Foreign Language Teaching (pp. 211–232). Berlin: Mouton de Gruyter. Brown, R. W., & Lenneberg, E. H. (1954). A study of language and cognition. Journal of Abnormal and Social Psychology, 49, 454–462. Dirven, R. (1993). Metonymy and metaphor: Different mental strategies of conceptualisation. Leuvense Bijdragen, 82, 1–28. Dirven, R., & Verspoor, M. (2 2004). Cognitive Exploration of Language and Linguistics. Amsterdam: Benjamins. Donmall, B. G. (1985). Language Awareness. NCLE Papers and Reports, 6. London: Centre for Information on Language Teaching and Research. Kramsch, C. (1993). Context and Culture in Language Teaching. Oxford: OUP. Krashen, S. (1981). Second Language Acquisition and Second Language Learning. Oxford: Pergamon. Langacker, R. W. (2000). Grammar and Conceptualization. Berlin: Mouton de Gruyter. Lazar-Meyn, H. A. (2004). Color naming. “Grue” in the Celtic languages of the British Isles. Psychological Science, 15, 288. Lowie, W., & Verspoor, M. (2004). Input versus transfer: The role of frequency and similarity in the acquisition of L2 prepositions. In M. Achard & S. Niemeier (Eds.), Cognitive Linguistics, Second Language Acquisition, and Foreign Language Teaching (pp. 77–94). Berlin: Mouton de Gruyter. Lucy, J. A. (1992). Language Diversity and Thought. A Reformulation of the Linguistic Relativity Hypothesis. Cambridge: Cambridge University Press. Niemeier, S. (1998). Colourless green ideas metonymise furiously. In F. Ungerer (Ed.), Kognitive Lexikologie und Syntax – Rostocker Beiträge zur Sprachwissenschaft, 5 (pp. 119–146). Rostock: Universität Rostock. Niemeier, S. (2000). Straight from the heart: Metonymic and metaphorical explorations. In A. Barcelona (Ed.), Metaphor and Metonymy at the Crossroads (pp. 195–211). Berlin: Mouton de Gruyter. Paramei, G. V. (2005). Singing the Russian blues: An argument for culturally based color terms. Cross-Cultural Research, 39, 10–34. Weskamp, R. (2001). Fachdidaktik: Grundlagen & Konzepte. Anglistik – Amerikanistik. Berlin: Cornelsen. Webpages: http://concise.britannica.com/ebc/article-9357534/Bluestocking (last accessed: August 11, 2006) http://en.wikipedia.org/wiki/Blue#Blue_in_Russian_language (last accessed: August 11, 2006) http://www.rc.umd.edu/features/features/chambermusic/crook.html (last accessed: August 11, 2006)

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.1 (45-169)

Odor memory* The unique nature of a memory system Gesualdo M. Zucco Olfactory memory has some important distinguishing characteristics which may suggest its uniqueness in cognition. Evidence to support this hypothesis is the following: (a) odor memory is only slightly affected by the length of the retention intervals; (b) it is very resistant to retroactive interference (i.e., to forgetting produced by subsequent learning experiences); (c) odor memory presents a lower initial acquisition level compared to visual and verbal material (this led some authors to assume that odors are represented in memory as distinctive events and learned in an all or none fashion); (d) the relationship between odorants and words seems to be very weak; (e) no differences appear in recognition tasks for odorants learned intentionally or incidentally; (f) neither strategies nor interferences seem to affect recognition memory for odorants. Such peculiarities of odor memory will be discussed and tentatively integrated in a single model. The main assumption is that people lack a conscious representation for odorants, which are stored in memory at an implicit – unconscious – level of knowledge.

Odors represent powerful cues, able to remind us of sometimes very distant memories. Such memories are often characterized by unusually strong emotive connotations. This is due to the direct anatomical links between the primary olfactory cortex and some structures in the limbic system. Visual and verbal sensory systems do not share similar peculiarities. Olfactory memory, however, has some other important distinguishing characteristics which may suggest its uniqueness in cognition, and although this matter is yet an open question, there are several experimental evidences to support this hypothesis. One of the more important peculiarity of odor recognition memory is that it is only slightly influenced by the length of retention intervals relative to that for pictures and words. This effect was observed for short intervals, for example a few seconds and minutes (Bromley & Doty 1996; Engen & Ross 1973; Jehl, Royet, & Holley 1994; Jones, Moskowitz, & Butters 1975), as well as for longer retention pe* The present contribution is partially based on the paper by Zucco (2003).

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.2 (169-234)

 Gesualdo M. Zucco

riods such as a week (Lawless & Cain 1975; Rabin & Cain 1984; Wood & Harkins 1987; Zucco 1983), a month (Lawless & Cain 1975), four months (Lawless 1978) and over a year (Engen & Ross 1973; Goldman & Seamon 1992). Thus, if one is exposed to a series of relatively common and uncommon odorants, and he is asked to try to recognize them immediately, probably he will be able to recognize about seventy percent of them; however, this performance will be more or less the same even after one year. Engen and Ross (1973), for instance, in a seminal paper observed that the hit recognition rate for odorants, after one year, was only five percent lower than few seconds following acquisition. Therefore, the forgetting curve resulting from these experiments is relatively flat, unlike for verbal and visual material (see Ebbinghaus’ pioneering experiments 1885; Nickerson & Adams 1979; Shepard 1967). We may consider this effect as a psychological law. – Olfactory memory presents a relatively low initial acquisition level compared to visual and verbal material (Engen 1991; Herz & Engen 1996; Lawless 1978; Richardson & Zucco 1989; Schab 1991; Zucco 2003, 2005). Thus, at immediate recognition one may be able to recognize about seven odorants out of 10 but nine or ten pictures or words. Why are odorants so poorly remembered initially and well retained over time? Engen and Ross (1973) suggested that pictures and words contain many attributes which serve as a basis for coding, producing a rich trace. Such a rich trace can be recognized easily on an immediate test, being richer in accessible features, but is subject to interference from later stimuli which share some of its attributes. Odors, at odds, are represented in memory as unitary perceptual events with little attributes redundancy (Engen 1991; Engen & Ross 1973; Herz & Engen 1996). They have fewer perceptual features to encode. This limits acquisition, but it results in minimal loss over time. In fact, the less rich trace leads to more errors at immediate recognition, but the absence of attributes held in common with other stimuli makes the trace resistant to interference (Schab 1991). In Underwood’s (1969) memory model, odorants would be stored by means of relatively inefficient forms of mnemonic coding. In fact, memory for one event may be seen as the retrieval of its attributes; then the more the attributes of an event the better the recall. Odors seem to be acquired holistically (in an all-or-none fashion). – Another characteristic of odor memory is that it is very resistant to retroactive interference, namely to forgetting produced by subsequent learning experiences. To give an example, in a seminal experiment, Lawless and Engen (1977) asked two groups of participants to associate one set or two sets of 12 pictures to a single set of 12 odorants. On the first session, both control and experimental group associated the odorants to one set of 12 pictures. On the second session the control group associated the 12 odorants again to the set of 12 pictures, while the exper-

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.3 (234-313)

Odor memory

imental group associated the odorants to a new set of 12 pictures. On the third session, that took place three weeks later, both groups were asked to recognize the odor-picture association presented on the first session. The authors did not find any difference between the experimental and the control groups in remembering the first set of pictures, despite the task demand was higher for the first group than the second. Thus, the first association to an odor seems to be relatively impervious to forgetting and difficult to modify through subsequent associations (you may think, for instance, how it is difficult to modify a learned aversion to food). Generally speaking, such absence of a retroactive interference effect could explain the endurance of olfactory traces over time (see Herz & Engen 1996). – Odor recognition memory seems to be unaffected by the conditions of learning (i.e., intentional or incidental). Thus, if one group of subjects is explicitly requested to smell a set of odorants in order to recognize them later, while a second group is simply asked to perform a task without mention of the subsequent memory test, then no differences will emerge between the two groups at recognition (Engen & Ross 1973). – Furthermore, Engen and Ross (1973) and Lawless and Cain (1975) found that odor memory seems to be unaffected by the familiarity of the substances used, as well as by the pleasantness of the substances. Thus, one may be able to recognize the odor of the garlic as well as the odor of an unknown chemical compound (Rabin & Cain 1984 for evidence suggestive of opposite results). As far as the relationship between odorants and language is concerned, it seems to be very weak (see also the contributions by Dubois and Holz in this volume). Evidence to support this hypothesis is the following: – To give a name to an odor seems to be a very difficult cognitive task (see end note 1 as regards the difference between identification and recognition of odorants1 ). Experimental subjects, in fact, show difficulties in the identification of the odorants (as observed by Herz & Engen 1996; Larsson 1997; Richardson & Zucco . Identification versus Recognition of odorants: These tasks involve different mental operations which are more demanding at the identification level than at the recognition level. Identification, in fact, presupposes retrieval from semantic memory of the meaning of the labels provided by the examiner in order to identify an odor (or to retrieve without help the name of the odor); while recognition presupposes only the correct matching of an odor with its representation in memory. This is independent of whatever label (right or wrong) a subject is spontaneously giving to a target odor. Thus the cognitive effort in such a case is lower than for the identification task. In a typical cued-identification task each participant is required to sniff for about 4 seconds a target odor, while the examiner read aloud four alternative verbal labels. Each participant has to identify the correct label for the odor. In a typical recognition task each trial may comprise the presentation of a target odor and a recognition set of four odorants. Each participant is required to smell for about 4 seconds the target (e.g., rose). A few seconds after this he is presented



JB[v.20020404] Prn:29/01/2007; 16:34



F: CELCR809.tex / p.4 (313-373)

Gesualdo M. Zucco

1989; Schab 1991). Usually, instead of providing a label, they remember specific life-episodes (e.g.: “Oh, this reminds me of the odor of my grand-mother’s farm”) or they use interjections (e.g.: “Ohh, Uhmm, Yes...”) as if they had the name of the odor on the ‘tip of their nose’ (Lawless & Engen 1977). A similar condition does exist in the verbal domain and it is known as “tip of the tongue state” (Brown & McNeill 1966). Furthermore, the subjects may label the odorants hedonistically (“I like – I dislike”; “it is good – it is bad”). Thus, the description of odorants seems to be based on an emotional and perceptual code system, while the description of other stimuli (e.g.: verbal and visual) is strictly linked to a lexical system well organized in semantic memory (see the models on the representation of knowledge based on semantic associative networks, e.g.: Harley 1995). – The correct identification scores vary from 16 % to 40–50 %. However, in the latter a case only for very familiar odorants (as the odor of the coffee). Despite both sexes are impaired in the identification of odorants, the females perform better than the males; and the young people better than the elderly. Also, the same subject frequently encodes the same odor differently on successive presentations (see Cain 1986 quoted in Schab 1991), while different subjects may give very different labels for the same odor. These phenomena demonstrate that verbal encoding for odorants is quite inefficient, and that the associations odor-name are based on personal idiosyncratic experiences. Then, it is not surprising whether all the attempts to classify the odorants through a lexical system have been unsatisfactory (see for instance, among the others, the classic model devised by Henning 1916). – In terms of limitations of encoding, it was observed that both common complex odorants and simple chemicals are encoded and remembered as well as meaningless geometrical shapes. In fact, such stimuli are difficult to name and it is supposed that they also are unitary perceptual events with little structural information (see Lawless 1978; Engen 1991). – Counting backwards (that is, a verbal interference task) has no effect on odor recognition, as firstly noted by Engen, Kuisma and Eimas (1973). – Odor labeling, whether self-generated or experimentally provided has no effect on subsequent recognition memory (Herz & Engen 1996). Thus, if a group of subjects is given the names of the presented odorants, but a second group is not, then the performance between the two groups at recognition will not differ. – Verbal rehearsal has no effect on the recognition of odorants (as noted by Gabassi & Zanuttini 1983 and several other authors). one at time four test tubes one of which contains the previously sniffed odor, and he is asked to recognize the target.

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.5 (373-446)

Odor memory 

– Studies carried out by Danthiir et al. (2001) have shown that olfactory memory ability is independent from other higher-order abilities and that it has very weak connections with verbal tests. – Finally, the non-verbal right hemisphere seems to be involved in the elaboration of odorants. This was observed with brain injured patients (Abraham & Mathai 1983; Zatorre, Jones-Gotman, Evans & Meyer 1992) and normal subjects (Zucco & Tressoldi 1989); see Royet & Plaibly 2004, for a review on the lateralization of olfactory processes. A last peculiarity of odor memory is that it seems scarcely sensitive (or no sensitive at all) to the effect of learning strategies and interference tasks. However, these effects have not been examined carefully in the literature, and usually the tasks adopted in the experimental studies were too easy (as counting backwards tasks to produce interference, and odor labeling tasks, or rehearsal, to improve recognition). Recently, Zucco (2003) carried out two experiments on the effects of both strategies and interferences on the recognition of odorants. The author adopted learning strategies and interferences tasks with a well established high improving and disrupting effect in the verbal and visual domains. The rationale for the experiments was the following: if the memory for odorants shares commonalities with the other memory systems, then it should be affected by interferences and strategies. The first experiment was a replication and extension of a study by Lyman and McDaniel (1986) on the relationship between strategies and odor memory. Four groups of participants were asked to smell 30 odorants and to perform different elaborative tasks for each of them as: – – – –

providing a name or a short definition, creating an image, describing a specific life episode, simply smelling the odorants (control group).

One week later participants were presented 60 odorants (30 targets and 30 distractors) and they were asked to try to recognize those presented at acquisition. Results showed no significant effects of encoding tasks on the correct recognition of odorants. The mean scores of hits were the following: Control .82

Imagery .79

Episode .83

Label .75

The second experiment examined the relationship between memory for odorants and interferences. Three groups of participants were exposed to either 15 olfac-

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.6 (446-563)

 Gesualdo M. Zucco

tory or 15 visual (photographs of human faces) or 15 acoustic (environmental sounds) stimuli. In the four sessions of the experiment (one per week) they had to recognize the stimuli whether in a no-interference condition, or in an intra-modality interference condition or in two inter-modalities interference conditions. Let us consider a participant in the olfactory group (the same procedure was followed also for the subjects belonging to the visual and acoustic groups): – no interference: the participant was asked to sniff 15 odorants. Then, he was requested to recognize each of the target odorants among three (i.e., the old odorant, and two new odorants). – intermodality interference: the participant sniffed a new set of 15 odorants. Then he was exposed for 2 minutes and half to an interference task (visual or acoustic), consisting on pleasantness judgements. At the end of such a task, he was asked to try to recognize the 15 target odorants. – intramodality interference: the participant sniffed a new set of 15 odorants. Then he was exposed to an olfactory interference task. At the end of such a task, he was asked to try to recognize the 15 target odorants. Consistently with the literature the interference affects selectively the recognition of visual and acoustic material, but has no effect on the recognition of odorants. In fact, in the no-interference condition, recognition scores for acoustic and visual stimuli were significantly better than those in the intermodality interference condition, while the intramodality interference condition produced significantly worse recognition scores. On the contrary, no significant differences among conditions of interference were observed for the olfactory group (see Table 1). Then, although strategies and interferences represent effective methods to improve or disrupt verbal and visual memory (as observed also in Experiment 2), they failed in showing any effect on odor recognition. As mentioned throughout the paper there are several distinguishing characteristics of olfactory memory (summarized in Table 2) which may suggest its uniqueness in cognition. Now, the question is: why do odorants behave differently from visual and verbal stimuli? Table 1. Mean scores for acoustic, visual and olfactory groups, as regards the four conditions of interference. Inference Groups Acoustic Visual Olfactory

Control

Acoustic

Visual

Olfactory

10 13 11.3

11.8 11 10.3

12 12.7 10.7

13.3 14.3 10.4

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.7 (563-658)

Odor memory

Table 2. Peculiarities of olfactory memory a.

Odor memory is slightly or not at all affected by retention intervals. Forgetting is, in fact, minimal relative to that for pictures and words There is a low initial acquisition level compared to visual and verbal material Odor memory is resistant to retroactive interference No difference between groups were observed in recognition tasks of odorants learned intentionally or incidentally Familiarity and pleasantness of the substances used usually do not influence recognition of odorants Odors and words seem to have a very weak relationship Recognition memory for odorants seem to be not affected by the use of learning strategies and interferences

b. c. d. e. f. g.

My answer is a theoretical proposal that integrates the anomalies here presented in a single interpretation. The main hypothesis lies on the assumption that odorants do not give raise to a conscious representation thereof and could be stored in memory at a level below consciousness. Then, the sequence from encoding to storage of olfactory information could occur through the following steps: (1) subject (S) is presented an odor; (2) during encoding S is conscious that he is smelling an odor (e.g., coffee),

but (3) when the stimulus is taken away, he lacks a conscious representation of it; (4) the odor is, however, automatically stored in memory,

but (5) at an implicit – unconscious – level of knowledge.

Thus, people can be conscious of olfactory stimuli only at the encoding and recognition stages, that is, when they are concretely present. Between these periods, conscious access to the olfactory trace is not possible. The assumption that odorants do not give raise to a conscious representation of them relies mainly on introspective reports: all of us have experienced that recalling an odor is an impossible or particularly hard task, regard how easy it is to conjure up an image of a visual stimulus or to rehearse with the silent language a poem or a song. The difficulty, however, could depend on the fact that people have no conscious representation of the stimuli experienced (then, we could maintain that a conscious recall implies a conscious representation). Of course everyone can remember that in one occasion an odor was encountered and can also name it, but this does not mean that a conscious representation or the recall of odorants is possible. However, speculations based on introspective report are not the



JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.8 (658-685)

 Gesualdo M. Zucco

best scientific demonstrations, and unfortunately we have only indirect methods to observe whether a conscious representation of olfactory stimuli actually exist. So, I would like to start by examining the results of the experiments on the effects of strategies and interfering tasks on memory for visual and verbal stimuli. It is well known – as we have seen also here – that while the former improve their retrieval, the latter have a disruptive effect on them (Baddeley 1990, 1998). I assume that an improvement or disruption of verbal and visual material is possible only because subjects have access to their internal representation of such stimuli. This means that to be effective, interferences and strategies have to act on these conscious representations. In short: if together with – or soon after – the presentation of the word “dog” people are not able to create an image or to rehearse such a stimulus, then neither strategies nor interferences can affect recognition; and the stimulus is recognized on the basis of its own characteristics. This is exactly what happened in the first experiment on the effect of strategies on odor memory: no differences were found among conditions, as none of the strategies used affected recognition of olfactory stimuli. In my opinion, this happened because strategies have to act on the conscious representations of the stimuli in order to be effective, and this in turn seems not possible for the olfactory domain (probably, olfactory representations are mainly perceptual, whereas visual and verbal representations are also conceptual, the latter giving raise to conscious representations, while the former do not). The same is true when adopting interfering tasks: subjects of an experimental group do not recognize odorants worse, than a control group, because they lack a conscious representation of those odorants. Indeed, to create a conscious representation of an odor (e.g., coffee) is a very difficult task, compared to how easy it is to imagine a cup of coffee, or to rehearse the word “coffee”. Storage and access to olfactory stimuli in memory, then, should not imply an effort but be automatic. I think the proposal presented above could integrate the anomalies of olfactory memory (see Table 2) in a single interpretation, and it also satisfies the scientific criterion of parsimony. So, for instance, if you consider the absence of a retroactive interference for odorants, you may explain such an effect on the basis that subsequent learning experiences cannot affect previously learned odorants, for which we have no conscious representation. Time passing has no effect on odor memory, for the same reason. Incidental or intentional tasks do not affect the recognition of stimuli that cannot be rehearsed or elaborated; neither strategies nor interferences affect recognition of odorants; and so forth, for all the other anomalies quoted above.

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.9 (685-756)

Odor memory 

What kind of relationship is there between this proposal and other interpretations on the nature of odor memory?2 Engen (1982) suggests that odorants are represented in memory as unitary and distinctive events with little attribute redundancy. Odors, then, are learned holistically (in an all-or-none fashion), which explains their relatively low initial acquisition level and their resistance in time (the interference caused from other olfactory stimuli is minimal). Fundamentally, I agree with Engen’s point of view that odorants could be stored as distinctive and unitary events, and I think that this interpretation does not contrast with, but could be integrate within the model here presented. The idea that there is no conscious representation for odorants can account for their distinctiveness. In my opinion, however, the latter is not sufficient to explain all peculiarities of odor memory, and has to be better operationally defined. Furthermore, within the discussion of whether odor memory can be conceptually (or semantically) driven or perceptually driven (Herz & Engen 1996; Lyman & McDaniel 1990; Schab 1991) the above analyses, and the results of the experiments on strategies and interferences that we have reviewed, seem to suggest that odor memory is mainly perceptual in nature. Indeed experimental manipulations do not affect recognition of odorants. A further point could concern the possible relationship between the theoretical proposal discussed here and other cognitive theories, as well as the relationship between this proposal and theories on consciousness (Zucco 2003). However, this matter is beyond the purposes of the present contribution.

References Abraham, A., & Mathai, K. (1983). The effect of right temporal lobe lesions on matching of smells. Neuropsychologia, 21, 277–281. Baddeley, A. (1990). Human Memory. Hove: Lawrence Erlbaum Press. Baddeley, A. (1998). Essentials of Human Memory. Hove: Psychology Press. Bromley, S., & Doty, R. (1996). Odor recognition memory is better under bilateral than unilateral test conditions. Cortex, 4, 25–40. Brown, R., & McNeill, D. (1966). The tip of the tongue phenomenon. Journal of Verbal Learning and Verbal Behavior, 5, 325–337. Cain, W. (1986). To know an odor. Proceedings of the 4th Takasago Symposium: The Influence of Smell on Humans (pp. 55–64). Tokyo: Takasago. Cain, W. et al. (1998). Odor identification: Perceptual and semantic dimensions. Chemical Senses, 23, 309–326.

. For a more general model on the perception of odors, please look at the interesting contribution by Stevenson & Boakes 2003.

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.10 (756-877)

 Gesualdo M. Zucco

Danthiir, V., Roberts, R., Pallier G., & Stankov L. (2001). What the nose knows. Olfaction and cognitive abilities. Intelligence, 29, 337–361. Doty, R. (1997). Studies of human olfaction from the University of Pennsylvania Smell and Taste center. Chemical Senses, 22, 565–586. Doty, R. (2001). Olfaction. Annual Review of Psychology, 52, 424–452. Ebbinghaus, H. (1885). Über das Gedachtnis. Leipzig: Dunker. Engen, T. (1982). The Perception for odorants. New York: Academic Press. Engen, T. (1991). Odor Sensation and Memory. New York: Praeger Press. Engen, T., Kuisma, J. E., & Eimas, P. D. (1973). Short-term memory for odorants. Journal of Experimental Psychology, 99, 222–225. Engen, T., & Ross, B. (1973). Long-term memory for odorants with and without verbal descriptions. Journal of Experimental Psychology, 100, 221–227. Gabassi, P., & Zanuttini, L. (1983). Riconoscimento di stimoli olfattivi nella memoria a breve termine. Giornale Italiano di Psicologia, 10, 51–60. Goldman, W., & Seamon, J. (1992). Very long-term memory for odours: Retention of odourname associations. American Journal of Psychology, 105, 549–563. Harley, T. (1995). The Psychology of Language. From Data to Theory. London: Psychology Press. Henning, H. (1916). Der Geruch. Leipzig: Barth Verlag. Herz, R., & Engen T. (1996). Odor memory: Review and analysis. Psychonomic Bulletin and Review, 3, 300–313. Jehl, C., Royet, J., & Holley, A. (1994). Very short-term memory for odours. Perception and Psychophysics, 56, 658–668. Jones, B., Moskowitz, H., & Butters, N. (1975). Olfactory discrimination in alcoholic Korsakoff patients. Neuropsychologia, 13, 387–393. Larsson, M. (1997). Semantic factors in episodic recognition of common odorants in early and late adulthood: A review. Chemical Senses, 22, 623–633. Lawless, H. (1978). Recognition of common odorants, pictures and simple shapes. Perception and Psychophysics, 24, 493–495. Lawless, H., & Cain, W. (1975). Recognition memory for odorants. Chemical Senses, 1, 331–337. Lawless, H., & Engen, T. (1977). Associations to odorants: Interference, mnemonics, and verbal labeling. Journal of Experimental Psychology, Human Learning and Memory, 3, 52–59. Lyman, B., & McDaniel, M. (1986). Effects of encoding strategies on long-term memory for odorants. Quarterly Journal of Experimental Psychology, 38 (A), 753–765. Lyman, B., & McDaniel, M. (1990). Memory for odorants and odorants’ names: Modalities of elaboration and imagery. Journal of Experimental Psychology (LMC), 16, 656–664. Nickerson R., & Adams M. (1979). Long term memory for a common object. Cognitive Psychology, 11, 287–307. Rabin, M., & Cain, W. (1984). Odor recognition: Familiarity, identifiability and encoding consistency. Journal of Experimental Psychology: Learning Memory and Cognition, 10, 316– 325. Richardson, J. T., & Zucco G. M. (1989). Cognition and olfaction: A review. Psychological Bulletin, 105, 352–360. Royet, J. P., & Plaibly, J. (2004). Lateralizatin of olfactory processes. Chemical Senses, 29, 731–745. Schab, F. (1991). Odor memory: Taking stock. Psychological Bulletin, 2, 242–251. Shepard, R. (1967). Recognition memory for words, sentences and pictures. Journal of Verbal Learning and Verbal Behaviour, 6, 156–163. Stevenson, R. J., & Boakes, R. A. (2003). A mnemonic theory of odor perception. Psychological Review, 2, 340–364.

JB[v.20020404] Prn:29/01/2007; 16:34

F: CELCR809.tex / p.11 (877-916)

Odor memory 

Underwood, B. (1969). Attributes of memory. Psychological Review, 76, 559–573. White, T. L. (1998). Olfactory memory: The long and short of it. Chemical Senses, 23, 433–441. Wood, J., & Harkins, S. (1987). Effects of age, stimulus selection and retrieval environment on odor identification. Journal of Gerontology, 42, 584–588. Zatorre, R., Jones-Gotman, M., Evans, A., & Meyer, E. (1992). Functional localization and lateralization of human olfactory cortex. Nature, 360, 339–340. Zucco, G. M. (1983). Esiste vicarianza sensoriale nel sordo? Una ricerca sul ricordo olfattivo in soggetti adulti sordi e udenti (odor memory in deaf and hearing adult people). Paper presented at the third conference of the Italian society of scientific psychology, Bologna. Zucco, G. M. (2003). Anomalies in cognition: Olfactory memory. European Psychologist, 3, 77– 86. Zucco, G. M. (2005). Cognitive aspects of olfaction: Odour memory. Main lecture at the “XIII. Jahrestagung der Arbeitsgemeinschaft Olfaktologie und Gustologie”: Aktueller Stand der Klinischen Olfaktologie/ Gustologie. Dessau, HNO Klinik, 2–3 December. Zucco, G., & Tressoldi, P. (1989). Hemispheric differences in odour recognition. Cortex, 25, 607– 615.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.1 (45-119)

From psychophysics to semiophysics: Categories as acts of meaning A case study from olfaction and audition, back to colors Danièle Dubois Exploring categories of odors from free sorting tasks and their representations in language by means of numerous different linguistic devices invited us to reconsider the cognitive model elaborated from the analysis of lexical forms and visual (and particularly color) categories. The diversity of linguistic resources available to ‘name’ and describe categories in olfaction and in audition as well, allows to clarify the relations between subjective and individual representations (psychology) vs. collective or shared lexical meaning (linguistics). It entails the elaboration of a semiotic conception of categories, accounting for individual and collective experiences and practices (including linguistic practices) for the different senses, shifting apart from the cognitivist conception of categorization as information processing.

.

Introduction

This paper relates the evolution we followed in our exploration of cognitive categories, from classical paradigms mainly concerned with visual objects, to the domains of olfaction and vision. The investigation of olfaction makes it necessary to reconsider the conceptual framework prevalent in cognitive sciences, on theoretical as well as on empirical grounds, and to evaluate new hypotheses in the field of acoustics. It has further led us to advocate a semiotic approach of categories and naming for colors, accounting for individual and collective experiences and practices of colors. Anthropological research has often pointed to the cultural values of the various senses, their differential influence on the elaboration of world representations (Levi-Strauss 1989; Descola 1986). They have stressed how important vision is to our cultures’ ‘worldviews’ (Howes 1991; Classen 1993). Cognitive research, even when facing cross-cultural issues (cf. Medin & Atran 1999, for example) has

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.2 (119-177)

 Danièle Dubois

mainly developed its studies about visual objects, and has done little in analyzing the linguistic resources and the psycholinguistic processes that could mediate between individual perceptual representations and shared conceptual ones (see as counterexamples, the work of Gentner (1982; Imai & Gentner 1993) and Waxman (1999) in developmental psychology). We will report some research which attempts to extend the paradigm originally elaborated to investigate cognitive categories on colors and simple visual objects to other senses, mainly olfaction and audition. This research was conducted within an interdisciplinary program of cognitive sciences.1 Beyond psychology and linguistics, we have been confronted, for the study of odors, with chemistry and neurosciences, and, for the study of noises, to mechanics and sound engineering. This empirical work led us to reconsider the cognitivist information processing model and to shift to an analysis of semiotic processes elicited from the various senses and to their conceptualizations in our cultures and languages. On a more epistemological and philosophical ground, it therefore leads us to both a conception of a more embodied and situated cognition, and correlatively to a more precise analysis of the linguistic and cognitive constraints, on semantic categories. We will first point out some of the difficulties encountered while exploring olfactory categories within the experimental paradigms that have been productive in the visual domain. We will then present the main results that can be obtained from a psychological exploration of cognitive categories of odors (from free sorting tasks) and their representations in language. This will show how we were lead to a more precise analysis of the linguistic devices available and involved in accounting for the subjective experience of this sense (David et al. 1997). The same rationale will be reproduced when presenting the research on audition aiming to answer the following questions: can categories for odors and noises be identified? On what principles of categorization are olfactory and auditory categories constructed? What are the linguistic resources available to ‘name’ and describe these categories? How do the relations between cognitive and linguistic categories vary across the senses? Can the results obtained from paradigms used in vision be generalized to any other mode of perception? The (preliminary) conclusions drawn from the main results presented here will sketch some theoretical and methodological issues that remain to be worked on. . The work on olfaction was part of a research program funded by the Cognitive Science program of the French Ministry of Education (MRT) and of the CNRS (“Cognisciences”). It involved M. Chastrette, N. Godinot, A. Holley, C. Rouby, G. Sicard (Laboratory of Neurosciences, Lyon I), B. Schaal (Institut du goût, Dijon), J. M. Hombert, M. Mouélé, C. Grinevald (DDL, Lyon II). Studies on audition have been conducted in close collaboration with M. Castellengo and J. D. Polack (LAM, Paris) and Ph.D. students (F. Guyot, V. Maffiolo, M. Mzali, C. Vogel and C. Guastavino).

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.3 (177-241)

From psychophysics to semiophysics: Categories as acts of meaning 

. Some historical landmarks . From Psychophysics ... In order to identify what the mental organization given from the different senses is psychophysics designs procedures within which: – –

–

experimental ‘stimuli’ are described as dimensions or parameters established by natural sciences, subjects’ answers are collected within ‘closed’ procedures with a priori categories of potential answers chosen by the experimenter (those given by the natural sciences and/or common sense concepts grounded in obviousness). Such procedures allow quantitative data analysis of qualitative judgments and therefore statistical ‘measurements’.

Within this paradigm, the analysis of the cognitive representations elaborated from the senses is conceived as the ‘subjective’ (i.e., psychological) evaluation of objective (i.e., physical) measurements. It entails an ‘obvious’ empirical sequence that starts from physics to psychology through the physiological analysis of the receptors of the classical five senses (sight, touch, hearing, smell, taste) and the identification of discriminative and absolute thresholds of perception. Subjects are therefore conceived as ‘imperfect’ measurement instruments in comparison with physical instruments. However, in contrast to physical instruments (which are reliable in giving the same value for two measurements of the same phenomenon), psychological phenomena are intrinsically variable both within subjects (who learn the second presentation of the ‘same’ stimulus is necessarily different from the first one) and across subjects (whose personal experiences are also necessarily different across cultures, and also within one culture). Founding psychological measurements in physical ones is therefore overlooking the specificity of psychological processes (Straus 1935; Bruner 1990; Cosnier 1998) and also the fact that the interpretation of instrumental measurements themselves result from human knowledge (physics or chemistry) and technological control on the world. . . . . to Cognitive sciences: Basic color terms In the 1970s, the development of cognitive science suggested a new paradigm starting from the identification of linguistics units (basic words, or basic names) in order to challenge the relativist hypothesis concerning the role of language in the

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.4 (241-279)

 Danièle Dubois

structuring of categorical knowledge. The emblematic issue2 is the linguistic and anthropological work of Berlin & Kay (1969) who attempted to ground the linguistic meaning of simple basic color names in their referential value given by basic (perceptual) categories of colors. The present paper discusses their thesis, and from the comparative work conducted on other modalities reconsiders the field of colors itself. (A more developed presentation of the main arguments is presented in Dubois & Grinevald 1999, 2003.) . Cognitive categories and naming It can be shown that if the psychophysics paradigm ignores semantic considerations, the cognitive one relies upon an implicit conception of lexical semantics deeply grounded in our intellectual tradition. As Rastier (1991) has pointed out, from Aristotle to contemporary cognitivism, word meaning remains interpreted within a triad that relates words (mainly names) to things either directly (formal semantics) or through the mediation of thought or concepts. Although the names of the three elements of the triad have changed throughout history and according to knowledge domains (as summarized in Figure 1 below), word meanings remained defined by their referential values, with the requirement to be adequate and true. concept THOUGHT Representation

sign WORD/NAME symbol

TRUE

thing REFERENT object

Figure 1. The semiotic triad in different domains: bold: the philosophical tradition; CAPITAL: linguistic description; italic: cognitive denomination. The semiotic triad goes back to Odgen and Richards (1949: 11) and is discussed by Rastier (1991) in the context of cognitive sciences.

. ... and obviously unavoidable reference, quoted by most authors concerned with colors, cf. Holz, Graumann, Heeschen, Niemeier, Plümacher, Wildgen, Wyler this volume for example.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.5 (279-341)

From psychophysics to semiophysics: Categories as acts of meaning

In short, within this model, word meaning remaining essentially referential presupposes an ontological transparency of the relation between ‘already made’ objects and dimensions as perceptual primitives, and their names.

. Odors . Odors and psychophysics However, when we turn our attention from colors to odors, the cognitive landscape looks quite different: in our attempt to empirically ground cognitive categories for odors, we first had to face the difficulties of processing olfactory stimuli according to the psychophysics paradigm. –

–

Odors have no description in natural sciences that can give any ‘measurement’ to the presented stimuli: when one takes into account the fact that most familiar odors, such as the odor of a rose or the odor of chocolate, result from the combination of about 400 to 600 molecules, no simple metric can be elaborated, and one can hardly expect to quantify such a complexity (Chastrette 1981). Therefore, if molecular concentration can be related to subjective intensity in perception, it is only possible for monomolecular solutions. Furthermore, even for these simple-molecule odorants, the chemical description does not necessarily fit or map onto either psycho-physiological indicators (olfactory cells responses) (Sicard & Holley 1984) or psycho-physical judgments (Schiffman 1974). The physiology of the olfactory receptors and the receptors themselves remain largely unknown (Sicard, Chastrette, & Godinot 1997).

In other words, the concept of odor is itself not clearly defined in experimental research, since olfactory research hesitates between investigations about responses to odorants (physical substance) and representations of odors (psychological concept) (see also Hudson & Distel 2002 on this point). It is not only an open question to experimentally define the boundaries of the stimulus because of the difficulty in controlling the stimulation related to its vaporous state, it is also conceptually difficult to conceive odors as an ‘entity’ (may be the absence of a spatial shape gives it the feeling of being immaterial). Furthermore, for olfactory stimulations, as for acoustic ones, experiment designers have to face the question of their temporal boundaries. However, in contrast with acoustic stimuli, it is even more difficult to isolate what ‘dimension’ can be considered as defining the olfactory stimulus or belonging to a contextual property of its commonly encountered experience, in our culture and a fortiori in other ones (Howes 1986; Mouélé 1997; Heeschen in this volume). This lack of well-defined autonomy of the stimulus is critical in olfaction,



JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.6 (341-417)

 Danièle Dubois

but can also be considered as a more general issue concerning the ecological validity of psychological experiments (in the case of vision cf. Dubois, Resche-Rigon & Tenin 1997; in audition cf. Guastavino & Cheminée 2003; Guastavino et al. 2005). . Cognitive approach and categories of smells We could expect to escape these difficulties by shifting to a more cognitive approach. Our original project on olfaction was therefore to reproduce Berlin and Kay’s (1969) and Rosch’s (1978) procedures, within the French culture as well as within other cultures and languages, in order to test the universal vs. relativist hypotheses. However, we could not find any basic terms in French (nor in English) to establish any reference points as the basic color terms, and we could not rely on any chart representing the olfactory space. As had already been noticed by scientists in olfactory research (see Engen 1987; Holz; Wildgen; Zucco in this volume) we were faced with the problem of odors having no names, at least in Indo-European languages (see Boisson 1997 for a review). Therefore, using ‘basic odor terms’ as the starting point to explore universals in odor categories, as suggested by Berlin and Kay for colors, was not possible. Furthermore, on the stimulus side, there is actually nothing that would correspond to the Munsell chart that could display odors on a scale – nothing similar to wavelengths as an objective dimension for color description in a unified theory of light. Chemistry, as the relevant scientific domain of objectivity for odors, does not provide such a capacity for ranging olfactory stimuli. In short, Berlin and Kay’s attempt at identifying basic color categories from basic color terms was impossible to transfer to olfaction. As far as knowledge for odors is concerned, and in contrast with other domains in which research on categorization has developed over the recent years (from the seminal work of Rosch 1978 on biological taxinomies to Medin & Atran 1999), we could not find systematic classifications for odors. The classificatory tradition of occidental cultures had – and still has – little interest in odor classifications, in comparison to the long tradition of objects classifications (see Rouby & Sicard 1997; Holley 1999; Candau 2000 for a review). Odor classifications rely mainly on the classification of objects as odorant sources, such as plants in Linnaeus (1756) or medicinal products (Lorry 1788), roughly organized along an ‘hedonistic’ scale for which pleasantness is taken to be the primitive and basic classificatory principle (von Haller 1763; Zwaardemaker 1925). Beside these attempts in natural sciences, we can find expert classifications of perfumes, but as they are restricted to specific sets of odorants and confidential groups, they cannot pretend to be exhaustive (see Dubois 2006 for a review on this point). On a formal ground, it has been shown that such classifications do not present any hierarchical structure as founded in taxonomies, nor as extensional nor as intensional structures (Chastrette 2002). When perfumes are defined by ‘notes’, as

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.7 (417-509)

From psychophysics to semiophysics: Categories as acts of meaning 

property descriptions, no systematic organization, that could mimic the necessary and sufficient conditions of scientific classifications, can be observed neither. However, we could process some free sorting tasks (Dubois 1997; Rouby & Sicard 1997) that allowed us to conclude that odors, as well as the visual objects already explored within the same procedures and data analyses (Dubois 1991), may be adequately represented in memory into categories. Nevertheless, two main differences between odors and visual objects can be seen: –

–

the (intra-individual and inter-individual) variability in categorization criteria as well as the effect of the sequentiality in the presentation of odorants may call into question the stability of the categories when inferred from data compiled across subjects (see David & Rouby 2002). The stability of categories of odors should therefore be further confirmed in experimental paradigms centered on the analysis of individual protocols, inasmuch as there is no consensus and as classifications may result from a wider variety of principles of categorization that vary from one subject to another. the principles of categorization rely on similarities or knowledge of the odorant sources, but also on the pragmatics of human activities and on pleasantness that appears as an even more important factor. (It can be noted that this was however previously observed in categories for complex visual objects; Mazet et al. 1987.)

In short, our work on the categorization of smells allows us to conclude that, if visual objects seem to be processed as stimuli which ‘stand out there’, odors are more likely to be structured and conceptualized as effects of the world on the subject, and therefore to appear as less autonomous from the experiential context. In order to further test this hypothesis, we have processed a more precise analysis of the devices through which the olfactory properties of objects were objectivized in language.

. Linguistic devices for odors Here, we present the main results from larger explorations of linguistic devices used for referring to odors. These results are more exhaustively reported in David et al. (1997) and Dubois and Rouby (2002). . The veridical label revisited When scoring the subjects’ verbal performances in olfactory experiments, words, such as orange, apple, vanilla . . ., are considered as the correct odor names or the veridical labels, in identification tasks, as well as in diagnostic tests (Engen 1987;

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.8 (509-555)

 Danièle Dubois

Doty 1992 among many others). Such scoring actually relies on the common sense evidence that words simply refer to things to be named as it holds quite nicely in vision (Dubois, Resche-Rigon & Tenin 1997). Even if numerous authors have already noted the peculiarity of verbal behavior in olfactory research (Engen 1987; Zucco in this volume), the veridical label or the correct response, remains the scoring criterion, implicitly assuming that there are names that adequately (truthfully) refer to any odor. However, experimenters themselves are not always coherent and remain unclear on how to name the stimuli: either by the name of the chemical substance (e.g., amyl acetate, phenylethyl alcohol, as named in the stimuli descriptions) or by the name of a source (the name of the odorant objects, banana and rose, respectively). This observation suggested us, that it is important to clearly distinguish between the name of the odorant (substance) and the name of the odor (the subjects’ verbal answer to a stimulation). Inasmuch the full instructions given to the subjects are not often explicitly reported in the description of the experimental procedures, we may wonder whether the subjects are questioned about “what is it?” or “what do you feel?” when presented with an olfactory stimulus. In short, in psychological research on olfaction, the so-called veridical label is actually the name (word) that the experimenter expects, as a relevant (obvious) answer to the question within the experimental setting; in other words the noun of the commonly encountered object (source) that produces an odor quite similar to the one produced by the presented odorant. Such a statement not only made it necessary, on a theoretical ground, to revisit the ‘triad’ (Figure 1) but it lead us to examine more closely how odors (as psychological objects) were represented in one language and how such a personal experience is communicated (in French in our case: Dubois & Rouby 2002; see also Mouélé 1997 in Wanzi (Gabon) and Heeschen in this volume in Kwoma, Eipo and Yale). It has to connect further such linguistic issues to explicit hypotheses concerning the structural properties of categories in human memory, still largely unknown for smell, and explored through other non verbal paradigms, such as pair comparisons, free sorting tasks on odorants (Rouby & Sicard 1997; Dubois, Rouby & Sicard 1997; Richardson & Zucco 1989). Empirically, it leads to a reanalysis of the “kinds of words or descriptions used by subjects misidentifying odors” (Engen 1987). However, we will not consider any longer that the lack of ‘correct response’ is a “misidentification” of the odor. Since there is no a priori definite ‘veridical label’ available (neither in the language, nor consequently in the subjects’ memories), it can be considered that any answer given by a subject is an attempt from his/ her part to communicate (to the experimenter) ‘something’ about odor in their memory and their identification of the presented odorants.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.9 (555-619)

From psychophysics to semiophysics: Categories as acts of meaning 

. A brief summary of the data from Dubois and Rouby (2002) The present analysis has been carried out on the wordings that 40 subjects gave in a spontaneous identification task of 16 familiar odorants. The answers to only three of the 16 odorants (hereafter named as lemon, orange and apple) were analyzed, as an example of our rationale. The research also relies upon other pieces of knowledge and ‘measurements’ elaborated from both linguistic hypotheses and from psychological ones on categorization for odors, collected in rating or in free sorting tasks. A first group of answers consists in any answer (simple word, constructed or polylexical forms) that includes some wording of the expected source (in other words any verbal form that would include the ‘veridical label’ wording): lemon, orange, apple. This group accounts for between 32 and 52% of the answers for each stimulus). It includes: – –

–

–

the name of the expected source (the commonly called ‘veridical label’): lemon, orange or apple. the name of the expected source + an adjective that refers to a more specific level of categorization than the ‘veridical label’ one (citron sucré – sweet lemon; pomme verte – green apple). This type of answers indicates that subjects are able to discriminate odors ‘below’ the previous level, or that they can (or want to) specify a property of the ‘smell of x’. It is consistent with the hypothesis that odors could be structured along typicality like other categories of objects (Rosch 1978; Dubois 1991). adjectival forms built on the source noun (citronné – lemony; orangé – orangelike): instead of giving a noun that would refer to an object, subjects suggest that odors are some property or quality of an object. the name of another source or artifact (candies, for example) syntactically constructed with the name of the expected source (as a compound noun: bonbon citron – lemon drops; sirop d’orange – orange syrup; shampooing pomme – apple shampoo): this type of answers mainly concerns food, drugs and cosmetics now flavored with a large variety of aromas. These answers can be considered as relevant as the ‘genuine’ source, inasmuch they belong to subjects’ experience.

To sum up, this first group of answers would mainly refer to the levels of categorization at which the odors are structured in the subjects’ memories, worded through the (basic or derived) names of existing odorants as sources of the odors. The second group of answers includes generic terms of sources specified in the previous group. We would consider here as correct answers the ‘generic terms’ acceptable for the previous specific sources: fruit, flower, food, chemicals ... The amount of answers of this type ranges from 9 to 15% and includes forms, such as

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.10 (619-697)

 Danièle Dubois

– – – –

fruit (fruit), fruité (fruity) (considered as a ‘related odor’ (fruity) in Engen’s “liberal scoring”), more frequent for orange and apple than for lemon, flower, mainly given for apple, as well as generic names of food (or any different wordings of different types of food, for example bonbon – candy), and finally drug, cosmetics and cleaning products that happen to be more and more frequently perfumed or flavored with ‘natural’ scents also mentioned previously under a more specific wording.

The third group includes co-hyponyms of the acceptable sources. The psychological hypotheses on categorization allow us to account for the quite important set of answers that can be linguistically considered as co-hyponyms. That is as nouns referring to objects belonging to the same superordinate category (as the expected (accepted) sources: orange is frequently given as lemon (but not the reverse), and apple is very frequently named by diverse fruit names (and frequently as strawberry, raspberry, lime). This type of answers accounts for 10 to 30% of the total answers). The same rationale can be further applied to the other categories previously listed in the second group. It concerns co-hyponyms of food, or cohyponyms of chemicals that could actually be flavored or odorized ‘real world objects’, experienced by the subjects as smelling like the stimuli, and as such, these answers can also be taken into consideration as ‘partially correct’ in scoring. Such scoring mainly concerns types of food and candies for lemon, orange, and apple, but also soaps or cosmetics for orange and apple, such as cakes, candy, shampoo . . . (The amount of such answers ranges from 10 to 20%.) Finally, the answers left refer –

–

to hedonic judgments (about 4%) that can also be considered as a relevant part of knowledge for odors (it is often considered as the main classificatory feature in olfaction). The scoring of answers, such as agréable (pleasant), increases the acceptable answers of 1% (apple) up to 5% (lemon), to knowledge and familiarity (about 4%) as something that is known or unknown, familiar or undetermined, being in the state of “tip of the nose” (Engen 1987) or ‘in lack of word’ as aphasic patients are (1% up to 4% of answers) or just stated as odeur (odor), or as an other generic terms referring to the subclass of pleasant ones, such as senteur (smell), and parfum (perfume).

In short, the small amount of correct answers, commonly encountered in the psychological literature, relies on a conception of meaning that considers that there is just one ‘veridical’ name (label) for each exemplar, and in which cognition is viewed as logical rather than experiential (Varela, Thompson & Rosch 1991; Lakoff 1987).

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.11 (697-771)

From psychophysics to semiophysics: Categories as acts of meaning 

. Linguistic resources for the description of odors It therefore seemed to us that a further exploration of the diversity of linguistic devices that could account for olfactory experience was worth going into (see David et al. 1997 for more specific data). Briefly, when one analyzes the verbal productions for definitions or types of odors as elicited from classical ‘norm productions’ in categorization research, it is observed that the main common device used was the generic compound noun with the name of the typical source (the so-called veridical label). But we could also notice the following regular forms: –

deverbal adjectives, such as agréable (pleasant), désagréable (unpleasant), insoutenable, insupportable (unbearable), nuisible, nocif (harmful), nociceptif, répulsif (repealant), agressif (agressive), or plaisant (pleasant), piquant, déplaisant (unpleasant), gênant (annoying), polluant (polluting), écoeurant (sickening), répugnant (disgusting), malodorant (ill-smelling), repoussant (repealant), suffocant (suffocating).

Such forms give information about both the meaning carried by the verb stem (to please, to aggress, to annoy . . .) and the nature of the effect thus produced: the suffix -able indicates a possibility, whereas the suffix -ant constructs a past tense form suggesting an actual effect. Therefore, a génante odor is interpreted as having a more unpleasant effect than just a désagréable one. – –

quelque chose + ce que/ qui (something that + that which) . . . ( > 30% of occurrences) and finally personal marks, such as in Il y a donc des odeurs qui nous attirent et des odeurs qui nous répugnent. (Then, some smells attract us while other disgust us.); Les odeurs familières, réconfortantes: celle de mon chien, de l’après-rasage de mon mari, d’un gâteau de ma mère. (Familiar, comforting smells: that of my dog, of my husband’s aftershave, of a cake of my mother baked.)

When scoring these data from definitions or types by contrasting ‘objective’ utterances (sentences asserted with no personal pronoun and no possessive article) which state ‘what it is about’, with ‘subjective’ ones (in which the subject is present in the definition of the odor), we obtain the following ratio, in the two types of corpora (definition and type elicitation) (see Table 1 below):

Table 1. Percents of mentions to the speaker vs. to something (from David et al. 1997: 72).

Speaker involved What is it about

Definition

Type elicitation

62% 38%

83.5% 16.5%

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.12 (771-834)

 Danièle Dubois

This contrasts with what is commonly observed in the production of category norms (see Poitou & Dubois 1999; Dubois & Poitou 2002), or when questioning the same way for noises or sounds, and colors in which responses there is no mention of the subject. Furthermore, if we contrast the linguistic forms which have a collective reference (on, nous, vous – one, we, you) with the forms which refer to individuals (je, moi, ma – I, me, my), we notice that definitions are (not surprisingly) more frequently given as ‘collective’ (92.2%) rather than individual (7.7%), as it is actually required by the definition process itself. But when subjects are asked to enumerate types of odors (type elicitation), they produce utterances referring both to their personal experience and to a collective one (respectively 58.5 and 41.5%; see David 1997: 226). In short, if odors are constructed within one’s subjective experience, this experience is conceived rather as a personal (individual) experience than as a collective one (shared among a group). What is shared is not actually the experience itself but the fact that this experience is (collectively) known as being personal rather than shared. These definitions indeed refer to diverse conceptualizations of odors. If an odor is always defined as ‘something’ (see David 2000) stated in the first part of the definition, the expansions diversely vary from an objectivist description in which the word odor refers to something in the world (therefore rather synonymous of odorant) to a more physiological object or psychological object. The latter can be defined as a sensation related to something from the world (as an emergent result of different processes: triggering, eliciting, identifying). Besides revealing different ontological assumptions, this issue has consequences in designing experimental settings and, as developed here above, in scoring verbal answers. In contrast to experiments in vision and acoustics where the stimuli and the correlated psychological object have the same ‘name’ (color, sound or noise), olfactory research has to account for the fact that the description of the stimulus is concerned with the odorant (the substance) not with the odor it elicits. Being aware that such questions are still open in olfactory research should prevent general and unproductive debates over past cognitive research, contrasting for example universalist vs. relativist conceptions of cognition assuming either that categorization influences language or that language influences categorization. The challenge is rather to avoid opposing natural (and therefore ... ‘veridical’) sciences to human ones and to develop interdisciplinary research that specifies how language (and languages) articulate to categorization at every level of description within the cognitive sciences.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.13 (834-943)

From psychophysics to semiophysics: Categories as acts of meaning 

. Noises and sounds Shortly shifting to noises and sounds, we could notice the following linguistic devices: – –

–

few specific names, such as voix (voice), musique (music), chant (song), all referring to human-produced noises, as for odors, a generic term, bruit (noise), and less frequently, son (sound), compound with a source name, such as voiture (car), vent (wind), pluie (rain), oiseau (bird), nominal constructions, such as claquement (slamming), craquement (creaking), grincement (squeaking), frottement (ruffling), démarrage (starting), freinage (breaking), ouverture (opening).

This systematic analysis of the semantics of the definitions of noises and sounds, as effects, sources or physical properties, further reveals the following pattern (see Table 2 below and for more exhaustive data see David 1997: 235): Table 2. Mentions of different types for noise, sound and odor (from David 1997) Corpus type

Effects

Sources

Physical properties

Noise (bruit) (N=182) Sound (son) (N=193) Odor (odeur)

28% 24.5% 70%

70.5% 52% 46.5%

16.5% 45.5% 8%

These results show that the conceptual representation of bruit (noise) is less abstracted from the object-source than that of son (sound), the latter being more frequently described in terms of physical properties of an objective world (the results for ‘odor’ are given for cross senses comparison purposes). An analysis of the linguistic cues of personal implication similar to the one performed on odors was done on the ‘sound/ noise’ corpora (definitions and types) in order to underline more precisely the objective/ subjective properties of the acoustic phenomena (cf. §4.3 above, Table 3 below and David 1997). Table 3. Percents of mentions to the speaker vs. to something (from David 1997) Noise (Bruit) % of answers Speaker involved What is it about

Def. 64 36

Sound (Son) Type 81.5 18.5

Def. 57 43

Type 73.5 26.5

Furthermore, if, as we had done for odors, we pay attention to the individual vs. collective presentation of the presence of subjects (I, my vs. one, we, you), we noticed that personal pronouns and possessive determiners are slightly more frequent

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.14 (943-1002)

 Danièle Dubois

in the corpus related to noises (37.5%) than in the corpus related to sounds (30%) (David 1997: 230). When there are marks of subjects, they refer more frequently to an individual judgments (je, moi, ma – I, me, my). This last observation could seem contradictory with the preceding analysis that suggests that the expression of noise is more subjective than that of ‘sound’. However, the personal pronoun je (I) is mainly associated with a verb introducing the (un)certainty of the speaker concerning its enunciation or the certainty of his/ her knowledge, such as je pense que (I think that), je crois que (I believe that), il me semble que (it seems to me that)..., and more precisely his/ her enunciation about technical or physical characteristics of the sound, as in je crois que c’est mesuré en hertz (I think it is measured in hertz) or physiquement, il me semble que ça se rapproche d’une onde (physically, it seems to me, this it is close to a wave). In short, the main point to be remembered from this analysis, is that acoustic phenomena can give rise to (at least) two main cognitive objects, noise and sound, noise being more odor-like, whereas sound being more color-like (Dubois 2000).

. Back to colors A similar analysis was processed of the diversity of lexical devices used to describe color chips reproduced on a large diversity of ‘charts’ produced for diverse purposes (measures vs. marketing, precise definition vs. imaginative illustration, etc. . . .) (Dubois & Grinevald 1999). It confirmed that colors too can be conceptualized as sources as we observed for noises and odors, as effects, as well as ‘entities’ of the world. Our study of denomination for chips on color charts for artists, wall painting, as well as for hair dyeing, cosmetics, car paints, wool balls . . . (see also Wyler in this volume), lead us to observe a large diversity of linguistic devices even to refer to the ‘same’ color chips. This diversity can be related to the diversity of the sources that can be colored or to the nature of the sources that constrain the conceptualization of the color itself. Colors can therefore also be seen as semiotic cues of the existence (presence) of objects rather than as objects as such. For example, if a color name, such as ocre (ochre), can refer to an entity of the world inasmuch it is material and defined as a pigment, on a color chart for artist, it becomes the semantic cue referring to something typically colored as such (sand or soil) in decorative painting or cosmetics, or even as a symbolic effect and value given in one culture to a ‘color’, as observed in cosmetics, cars, or more generally any product to be on sale (Dubois & Grinevald 2003). As far as naming is concerned, such a diversity in naming colors confirms the observations made by Wyler, Niemeier, Plümacher (in this volume), which remain diversely interpreted. The reliance on the belief that there is one true denomination of each color (i.e., the one given by scientific terminology that actually aims

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.15 (1002-1058)

From psychophysics to semiophysics: Categories as acts of meaning

at giving an adequate and univocal (or even true) label), imposes to consider that any other formulation is (cognitively) deviant or (linguistically) metaphoric or metonymic. If we rather consider that the linguistic resources contribute to the construction of a concept, therefore the landscape looks different and allows us to connect our research to other fields, such as anthropology or the history of sciences. Therefore Berlin and Kay’s basic color terms are just a very special case that cannot account for the diversity of naming sensory experience. The conceptualization of colors appears largely more constrained by technical practices than by physiological determinism as pointed out by Latour (1993).

. Concluding remarks Our analysis was an attempt to ground the evaluations of the subjects’ verbal answers on a more precise knowledge of the psychological processes that connect subjective individual representations to collective and shared lexical meanings. It remains however “an open question whether these (as yet) unnamed categories share the same status as labeled categories” (Waxman 1999: 274). As numerous linguistic, psycholinguistic and anthropological studies have shown, the hypothesis that the diversity of linguistic forms may constrain the ‘ontology’ given to the ‘entities’ has to be further worked out. We can however already suggest that the diversity of linguistic forms induces (or reveals) different distances between the ‘subject’ and the ‘objects’ of the world: from complex phrasing expressing the effects of the world on the subject, to simple ‘basic’ names as observed for colors, which suggest the idea that things are “crying out to be named” (Berlin 1973) or that words, as labels, could map onto things. Further experimental investigations will therefore have to account for the fact that our culture has diversely lexicalized the different senses (and other cultures eventually differently). If we always perceive ‘something’, through the diversity of our senses, language diversely objectivizes and ‘stabilizes’ our cognitive representations of the world into a large variety of linguistic forms. When categories are seen as acts of meaning (Bruner 1990; Dubois 2000), they rely mainly on the subjects’ relations to the world, through their practices, including verbal ones. In short, the analysis (and eventually scoring) of verbal descriptions as answers to sensory stimuli involves the elaboration of theoretical knowledge on three domains: (1) an adequate semantic theory of lexical meaning (linguistics), (2) a theory of cognitive representations in memory (psychology), (3) an explicit theory of the relations between language and cognition in order to make relevant inferences regarding subjective judgements from linguistic



JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.16 (1058-1160)

 Danièle Dubois

data, allowing to use language as an overt, observable, public representation of mental representations (psycholinguistics and cognitive linguistics).

To conclude, if the proposed topic of this book was “How can language cope with color and smell?”, we would suggest that research could develop in order to better understand how colors, sounds, noises, and odors cope with the diversity of language devices available to account for such sensory domains?

References Berlin, B. (1973). Folk systematics in relation to biological classification and nomenclature. Annual Review of Ecology and Systematics, 4, 259–271. Berlin, B., & Kay, P. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Boisson, C. (1997). Quelques généralités sur la dénomination des odeurs: variations et régularités linguistiques. Intellectica, 24, 29–49. Bruner, J. (1990). Acts of Meaning. Cambridge: Harvard University Press. Candau, J. (2000). Mémoires et expériences olfactives: Anthropologie d’un savoir faire sensoriel. Paris: PUF. Chastrette, M. (1981). An approach to a classification of odours using physico-chemical parameters. Chemical Senses, 6, 157–163. Chastrette, M. (2002). Classification of odors and structure-odor relationships. In C. Rouby et al. (Eds.), Olfaction, Taste and Cognition (pp. 100–116). Cambridge and New York: Cambridge University Press. Classen, C. (1993). Worlds of Senses. London: Routledge. Cosnier, J. (1998). Le retour de psyché: Critique des nouveaux fondements de la psychologie. Paris: Desclée de Brouwer. David, S. (1997). Représentation d’objets sensoriels et marques de la personne. In D. Dubois (Ed.), Catégorisation et cognition: de la perception au discours (pp. 211–242). Paris: Kimé. David, S. (2000). Incertitude dans les représentations olfactives, auditives et gustatives. Cahiers du LCPE, 4. David, S. (2002). About odors and their linguistic expression in French. In C. Rouby et al. (Eds.), Olfaction, Taste and Cognition (pp. 82–99). Cambridge and New York: Cambridge University Press. David, S., & Rouby, C. (2002). Arbres, classes et distances. Cahiers du LCPE, 29–70. David, S., Dubois, D., Rouby, C., & Schaal, B. (1997). L’expression des odeurs en français: analyse lexicale et représentation cognitive. Intellectica, 24, 51–83. Descola, Ph. (1986). La nature domestiquée. Paris: MSH. Doty, R. L. (1992). Diagnostic tests and assessment. The Journal of Head Trauma Rehabilitation, 7, 47–65. Dubois, D. (1991). Les catégories sémantiques naturelles: prototype et typicalité. In Dubois, D. (Ed.), Sémantique et cognition: Catégories, concepts et typicalité (pp. 31–54). Paris: Éditions du CNRS. Dubois, D. (Ed.). (1997). Catégorisation et cognition: de la perception au discours. Paris: Kimé.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.17 (1160-1284)

From psychophysics to semiophysics: Categories as acts of meaning 

Dubois, D. (2000). Categories as acts of meaning: the case of categories in olfaction and audition. Cognitive Science Quarterly, 1, 35–68. Dubois, D. (2006). Des catégories d’odorants à la sémantique des odeurs: une approche cognitive de l’olfaction. Terrain, 47 (sept), 89–106. Dubois, D., & Grinevald, C. (1999). Pratiques de la couleur et dénominations. Faits de langues, 14, 12–25. Dubois, D., & Grinevald, C. (2003). En voir de toutes les couleurs: processus de dénomination des couleurs et constructions cognitives. In C. Vandeloise (Ed.), Langues et cognition. Traité des sciences cognitives (pp. 79–113). Paris: Hermès. Dubois, D., & Poitou, J. (2002). Des ‘normes catégorielles’: structuration cognitive et/ ou linguistique des catégories sémantiques? Intellectica, 35, 217–249. Dubois, D., & Rouby, C. (1997). Olfaction et cognition: du linguistique au neuronal. Intellectica, 24, 9–20. Dubois, D., & Rouby, C. (2002). Names and categories for odors: The veridical label. In C. Rouby, B. Schaal, D. Dubois, R. Gervais & A. Holley (Eds.), Olfaction, Taste and Cognition (pp. 47–66). Cambridge and New York: Cambridge University Press. Dubois, D., Resche-Rigon P., & Tenin, A. (1997). Des couleurs et des formes: catégories perceptives ou constructions cognitives. In D. Dubois (Ed.), Catégorisation et cognition (pp. 7–40). Paris: Kimé. Dubois, D., Rouby, C., & Sicard, G. (1997). Catégories sémantiques et sensorialités: de l’espace visuel à l’espace olfactif. In Enfance. N◦ spécial ‘l’odorat chez l’enfant’, 141–151. Engen, T. (1987). Remembering odors and their names. American Scientist, 75, 497–503. Engen, T., & Engen, E. (1997). Relationship between development of odor perception and language. In Enfance. N◦ spécial ‘l’odorat chez l’enfant’, 92–102. Foley, W. A. (1997). Anthropological Linguistics. Malden: Blackwell. Gentner, D. (1982). Why nouns are learned before verbs: Linguistic relativity versus natural partitioning. In S. Kuczaj (Ed.), Language Development: Language, Thought, and Culture (pp. 301–334). Hillsdale, NJ: Lawrence Erlbaum. Guastavino, C., & Cheminée, P. (2003). Une approche psycholinguistique de la perception des basses fréquences: conceptualisation en langue, représentations cognitives et validité écologique. Psychologie Française, 48 (7), 91–101. Guastavino, C., Katz, B., Polack, J.-D., Levitin, J., & Dubois, D. (2005). Ecological validity of soundscape reproduction. Acta Acoustica & Acustica, 9, 333–341. Haller, A. von. (1763). Liber XIV. Olfactus. Elementa physiologiae-Corporis umani, 5 (pp. 125– 185). Lausanne: Grasset. Holley, A. (1999). Éloge de l’odorat. Paris: O. Jacob. Howes, D. (1986). Le sens sans parole: vers une anthropologie de l’odorat. Anthropologie et sociétés, 10, 29–46. Howes, D. (Ed.). (1991). The Varieties of Sensory Experience: A Source Book in the Anthropology of the Senses. Toronto: Toronto University Press. Hudson, R., & Distel, H. (2002). The individuality of odor perception. In C. Rouby, B. Schaal, D. Dubois, R. Gervais & A. Holley (Eds.), Olfaction, Taste and Cognition (pp. 408–420). Cambridge and New York: Cambridge University Press. Imai, M., & Gentner, D. (1993). What we think, what we mean, and how we say it. Chicago Linguistic Society, 29, 171–186. Lakoff, G. (1987). Women, Fire, and Dangerous Things. Chicago: Chicago University Press. Latour, B. (1993). Le topofil de Boa-Vista: la référence scientifique: montage photophilosophique. Raisons pratiques, 187–216.

JB[v.20020404] Prn:11/04/2007; 13:53

F: CELCR810.tex / p.18 (1284-1389)

 Danièle Dubois

Levi-Strauss, C. (1989). Des symboles de leurs doubles. Paris: Plon. Linnaeus, C. (1756). Odores medicamentorum. In Lars Salvius (Ed.), Amoenitates Academicae, Vol. 3 (pp. 183–201). Stockholm. Lorry, M. (1788). Sur les parties volatiles et odorantes des médicamens tirés des substances végétales et animales. Hist. Soc. Roy. Med. (pp. 306–318). Paris: Librairie de la Société Royale de Médecine, Barrois. Lucy, J. A. (1999). The linguistics of ‘color’. In C. L. Hardin & L. Maffi (Eds.), Color Categories in Thought and Language (pp. 320–346). Cambridge: Cambridge University Press. Mazet, C., Fleury, D., & Dubois, D. (1987). Représentations pour l’action. Toulouse: Octares. Medin, D., & Atran, S. (Eds.). (1999). Folkbiology. Cambridge: MIT Press. Mouélé, M., (1997). L’apprentissage des odeurs chez les Waanzi: note de recherche. Enfance, n◦ thématique: L’odorat chez l’enfant, perspectives croisées, 209–222. Ogden, C. K., & Richards, I. A. (10 1949). The Meaning of Meaning. A Study of the Influence of Language upon Thought and of the Science of Symbolism (1 1923). London: Routledge & Kegan Paul LTD. Poitou, J., & Dubois, D. (1999). Catégories sémantiques et cognitives: une étude expérimentale en sémantique lexicale. Cahiers de lexicologie, 74, 15–27. Rastier, F. (1991). Recherches sémantiques et sciences cognitives. Paris: PUF. Richardson, J. T., & Zucco, G. (1989). Cognition and olfaction: A review. Psychological Bulletin, 105, 331–351. Rosch, E. (1978). Principles of categorization. In E. Rosch & B. Lloyd (Eds.), Categorization and Cognition (pp. 28–47). Hillsdale, NJ: Lawrence Erlaum. Rouby, C., & Sicard, G. (1997). Des catégories d’odeurs? In D. Dubois (Ed.), Catégorisation et cognition (pp. 59–81). Paris: Kimé. Sahlins, M. (1976). Colors and cultures. Semiotica, 16 (1), 1–22. Schiffman, S. S. (1974). Physicochemical correlates of olfactory quality. Science, 185, 112–117. Sicard, G., & Holley, A. (1984). Receptor cell response to odorants: Similarities and differences among odorants. Brain Research, 292, 283–296. Sicard, G., Chastrette, M., & Godinot, N. (1997). Des représentations de l’espace olfactif: des récepteurs à la perception. Intellectica, ‘du linguistique au neuronal’, 24, 85–107. Straus, E. (1935). Vom Sinn der Sinne. Berlin: Springer Verlag. Edition franç.: Du sens des sens. Million 2000. Varela, F. J., Thompson, E., & Rosch, E. (1991). The Embodied Mind: Cognitive Science and Human Experience. Cambridge, MIT Press. Waxman, S. (1999). The dubbing ceremony revisited: Objects naming and categorization in infancy and early childhood. In D. Medin & S. Atran (Eds.), Folkbiology (pp. 233–284). Cambridge: MIT Press. Zwaardemaker, H. (1925). Classification des odeurs. L’odorat, 178–224.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.1 (46-131)

Cognition, olfaction and linguistic creativity* Linguistic synesthesia as poetic device in cologne advertising Peter Holz The starting point of this paper is the observation that there is no consistent and conventionalized ‘lexicon of olfaction’ in any language. This is due in large part to the neurophysiological incongruence of language-processing and smellprocessing structures in the brain. Nonetheless we do in some way speak about smells, odors, scents. The discourse of cologne advertising, which I concentrate on, provides a vast amount of quasi-olfactory expressions. The majority of these adjectives, nouns and noun phrases are generated by synesthesia as a metaphorical process. On the theoretical basis of general semiotics (Peirce) and the poetic function of language (Jakobson), I assert that in cologne discourse poetic means are indispensable, since the referential function cannot cope adequately with describing smells.

.

Sketch of the problem

Reflecting on olfactory qualities and their linguistic representation is certainly not among the most frequent human activities. As a matter of fact, it is quite rarely done in everyday communication. But, nonetheless, it can be a rather interesting topic for research in cognitive linguistics. If one is exposed to a certain smell, one has an immediate perception which just cannot be controlled. And if you are then asked to describe your perception by just one adjective, you will realize that this is quite a hard job. If you happen to find one, let’s say sweet, someone else might describe the same scent as fruity or

* I am tremenduously grateful to Blossom Wrede for her proof-reading and for several competent constructive comments on my train of thought which thereby became much more transparent (even to me).

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.2 (131-204)

 Peter Holz

even warm or erotic or merely pleasant or unpleasant. The descriptions, however, will be very diverse and arbitrary. The problem of describing a smell by just one word does not occur very often. We are even very seldom confronted with the question of describing a smell by linguistic means. But if we do approach the problem linguistically, we can observe quite quickly that conventional lexemes for olfactory categories are missing in virtually any language. Dubois, however, mentions an exception to the rule: “In some African languages [. . .] odors are cognitively constructed as ‘objective realities’ and have names (even nouns): for example, about 11 ‘basic terms’ for odors have been identified in Li-Wanzi” (Dubois 1997: 188). But on the whole, linguistic communities do not have original lexical units concerning olfaction equivalent to ‘basic color terms’, which have been and still are a great topic of research in cognitive linguistics (cf. Berlin & Kay 1969; Rosch 1977; Lakoff 1987; and especially Dubois’ and Niemeier’s contribution in this volume). There is no such thing as ‘basic smell terms’. And this simple fact makes the relation between language and olfaction both quite a difficult task and also an interesting field of linguistic research.

. Hedonistic judgments as basic cognitive categories in olfaction From a psychological standpoint the hedonistic distinction of pleasant vs. unpleasant is the most fundamental way to react to smells. Even neonates respond to certain olfactory stimuli with significant facial expressions corresponding to attraction vs. rejection. But the positive vs. negative reaction to smells is not a stable universal cognitive capacity. Humans can indeed be trained to consider a given smell peasant or unpleasant (cf. Vroon et al. 1996: 99f.). Just as the mental lexicon develops in the maturation of the individual in order to cognitively distinguish entities in the world, also the capacity to verbally distinguish more and more nuances of pleasant vs. unpleasant smells is elaborated. As time passes by, we become able to make various associative lexical choices from the pleasant-vs.-unpleasant-paradigm, such as fascinating and amazing (pleasant) vs. odd and disgusting (unpleasant). Our hedonistic judgments get more and more sophisticated. To put it into the terminological framework of the prototypical conception of categorization, the semantic cores in the gradual bipolar category olfactory perception are simply pleasant respectively unpleasant. All other chosen lexemes must be considered elements of this category deviating more or less from the prototypical centers, due to secondary connotations.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.3 (204-259)

Cognition, olfaction and linguistic creativity 

So, however we refer linguistically to a smell, in the end any chosen lexeme can be traced back to the fundamental cognitive hedonistic distinction of pleasant vs. unpleasant and can be assigned to one of the prototypical poles due to inherent connotations.

. Common linguistic constructions to describe smells . Iconic reference1 A common way to refer linguistically to smells is to employ syntagmatic constructions containing the adverb like which implies a comparison of the external source (i.e., the substance which emits the olfactory stimulus) and the internal olfactory perception. Examples are: (1) It smells like peppermint (2) It smells like rotten bananas

Although in these cases the olfactory stimuli do not necessarily emanate from the substances mentioned, the olfactory impression is compared with the odor of the assumed source. The smell of a rotten banana might be caused by something else, perhaps by a kind of human sweat or whatever. In terms of semiotics, we can say that language which – according to Peirce – in itself is basically symbolic establishes in these cases an additional iconic relationship between – –

a chemical entity, namely the molecules emitting olfactory source, whose precise nature is unknown (i.e., unnamable) and a psychological state, namely the memory of the olfactory quality of a rotten banana.

To speak of iconicity here is legitimate, because some kind of similarity between both is assumed. . Indexical reference The situation is somewhat different in expressions, such as (3) It smells burnt (4) It smells rotten

. For the terminological distinction between iconic, indexical and symbolic sign properties cf. Peirce (CP §§3.362 and 4.531).

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.4 (259-333)

 Peter Holz

Since we can rationalize a causal relationship in space and time between source and perception and not one of similarity, these phrases can semiotically be classified as indexical. The predicate rotten is assigned to a perception whose stimulus certainly emanates from some olfactory source, but it neither identifies the source nor specifies what exactly produces the smell. . Special cases: Metaphorical reference In utterances, such as (5) Your perfume smells hot (6) A sickly sweet scent follows her

the qualifying adjectives are not used in a literal sense. They must be classified as metaphorical expressions. With respect to the semiotic terminology just introduced we may say that the above examples are rather indexical than iconic, since a causal relation between source and perception is implied. Concerning hedonistic evaluation we can say that (due to certain connotations) (5) represents the pleasant- and (6) the unpleasant-domain. Paraphrases to explain the connotations might be: (5a) smells hot = smells sexy = pleasant (usually) (6b) sickly sweet scent = repulsive smell = unpleasant

But the referential domain of hot and sickly sweet is not actually olfaction. Originally hot refers to the tactile perception (thermal resp. pain perception) and sweet to the gustatory one. The process generating the resulting metaphorical expressions is a categorical shift from one sensory modality to another. This process, called linguistic synesthesia, will be discussed in detail below (cf. Shibuya et al. in this volume). Susanne Niemeier (in this volume) emphasizes a certain viewpoint in cognitive linguistics that considers metaphoric expressions that rely in principle on the phenomenon of metonymy to be mostly realized as pars pro toto. But, as she admits, “we can only understand [. . .] metonymies as metonymies if we know how they developed” (Niemeier in this volume). Whereas that approach implies a diachronic procedure, my analysis is by definition synchronic, so I stick to the term metaphorical expressions.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.5 (333-376)

Cognition, olfaction and linguistic creativity 

. Why language cannot properly cope with olfaction The reason for the poverty of the lexicon of olfaction is a neurophysiological one. Olfactory stimuli are basically processed in evolution ary primitive parts of the human brain, namely in sub-cortical areas belonging to an anatomically distinct and functionally related structure, usually referred to as the limbic system. This cerebral subsystem is mainly concerned with the organization of visceral body functions, the processing of emotional states and dynamics, and is also closely involved in long term memory processes as is confirmed by several psychological experiments reported and also carried out by Zucco (cf. in this volume; cf. also Bear et al. 2 2001: 437ff.; Emrich et al. 2002: 25f.). Figure 1 shows the crucial components of the limbic system. The limbic system is not concerned with the abstract symbolic activity of language processing (perception and production). The cerebral processing of language can be roughly located in certain parts of the neo-cortex in the left hemisphere (in most people), usually designated as Broca’s area for language production and Wernicke’s area for language perception (for a more specific functional differentiation cf. Bear et al. 2 2001: 576ff.). Gyrus cinguli Septum

Fornix

Olfactory bulb Hypothalamus Amygdala Corpus mamillare

Hippocampus

Figure 1. Major components of the limbic system (www.driesen.com, 13.12.2003)

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.6 (376-424)

 Peter Holz Primary motor cortex Broca’s area

Primary auditory cortex

Wernicke’s area

Primary visual cortex

Figure 2. The language processing centers of the brain (www.driesen.com, 13.12.2003)

According to our evidence, the neural connections between left-hemisphere cortical areas and the sub-cortical limbic structures are relatively poor. As a consequence of that it is apparently impossible to adequately synchronize the cerebral organization of smell perception with the language processing areas of the brain in such a way that a stable lexicon of olfaction results (cf. Burdach 1988: 22; Gschwind 1998: 22). Chernigovskaya and Zucco (in this volume) also emphasize the neural processing of olfactory stimuli in the right (non-verbal) hemisphere of the neo-cortex and consequently the lack of a proper vocabulary in olfaction.

. Does it make scents? – The poetic function of language in cologne advertising . Theoretical and methodological terminology – six functions of language In order to prepare the empirical analysis I would like to turn to the functional concept of language established by Roman Jakobson (cf. Jakobson 1981: 23ff.). Jakobson postulates six linguistic functions within the process of verbal commu-

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.7 (424-485)

Cognition, olfaction and linguistic creativity

Table 1. Roman Jakobson’s model of verbal communication Factor

↔

Corresponding function

Context Addresser Addressee Contact Code Message

↔ ↔ ↔ ↔ ↔ ↔

Referential function Emotive function Conative function Phatic function Metalingual function Poetic function

nication, each corresponding to an appropriate communicative factor. His model of verbal communication is shown in Table 1. For this paper only three of these functions are relevant, namely the referential, conative and poetic function. . The referential function The referential function can be roughly paraphrased as a communicative focus lying on providing information about extra-linguistic entities, i.e., information about the external world (i.e., context). Jakobson claims that in most cases the referential function takes the governing role in communication. But even though a set (Einstellung) toward the referent, an orientation toward the context [. . .] the so called referential [. . .] function is the leading task of numerous messages, the accessory participation of the other functions [. . .] must be taken into account by the observant linguist. (Jakobson 1981: 22)

. The conative function Of course, we have to assume that any use of language in advertising carries a heavy load of persuasive properties in order to boost a certain product, such as the colognes discussed here. Hence, the conative function whose effect is to make an appeal to the reader, certainly plays the dominant role in advertising. An indication of this is that we will hardly ever find any negative statements about a cologne. But what I claim is that in cologne discourse the conative effect is not achieved by obvious expressions, such as “Buy me, I am good for you!” The persuasive impact is usually concealed and rather encoded by a more or less subtle blend of quasi-referential and poetic means which are intended to persuade. By emphasizing this blend, I argue that in the case of language on colognes (as one domain in which language on smells can be observed) the poetic function necessarily plays a crucial role in modifying both the referential and the conative aspects.



JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.8 (485-553)

 Peter Holz

. The poetic function and the projection principle The main criterion to identify the effect of the poetic function is the focus “on the message for its own sake” (Jakobson 1981: 25), i.e., it draws the reader’s attention to certain formal elements, to the linguistic material the very utterance is built from. But how can the postulated poetic function be observed empirically? As is quite commonly known, in the chain of speech two general types of relations between linguistic units can be observed: – –

syntagmatic relations in praesentia, corresponding to the horizontal axis of combination (defined by spatio-temporal contiguity) and paradigmatic relations in absentia, corresponding to the vertical axis of selection (defined by equivalence).

And, according to Jakobson once more, linguistic evidence for the poetic function can be observed, wherever “the poetic function projects the principle of equivalence from the axis of selection to the axis of combination” (cf. Jakobson 1981: 27). This can be explained in the following way: a propositional sentence, such as (7) I ran into the house because it was raining, i.e., ‘It is the case, that I ran into the house, because it was raining’,

merely provides information about something going on in the external world, thus the referential function is predominant. In contrast, the sentence (8) I darted and dashed, raced and rushed into the house, because it was raining

conveys basically the same referential information, yet it also carries poetic impact. The repetition of verbs from the lexical field human locomotion is to a high degree semantically redundant, since the process of moving into the house can be sufficiently described by just one of the verbs. But the unusual alliteration darted/ dashed (phoneme /d/) as well as raced/ rushed (phoneme /r/) structures the utterance in a somewhat rhythmical way, draws attention to the phonetic form and thus charges the utterance with poetic qualities and engages emotions and imagination of the hearer.

. Hypothesis The specific question in this paper is how Jakobson’s terminology can be applied to the topic at hand, namely the linguistic analysis of advertisement texts for men’s cologne?

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.9 (553-622)

Cognition, olfaction and linguistic creativity 

The underlying idea of my approach is the concept of linguistic synesthesia as poetic device which is the basis for my hypothesis that synesthetic constructions in language can be interpreted as a phenomenon determined by poetic function and that they can be observed frequently and necessarily in cologne advertising.

. Synesthesia: Neuropsychological capacity vs. poetic device If we use the term synesthesia, we have firstly to distinguish between a neuropsychological and a linguistic phenomenon. From the perspective of neuropsychology, synesthesia means that, if some stimulus affects the central nervous system through a certain sensory channel (e.g., audition) some individuals, called synesthetes, have in fact a parallel perception belonging to a different sensory modality (e.g., vision). Cases have been reported where people have visual perceptions (certain colors or shapes) whenever they listen to a certain music. Although the so called colored hearing is the most frequent synesthetic phenomenon, there is a great variety of cases, where other sensory modalities are melted together, e.g., color and smell, taste and vision of shapes, audition and taste/ smell, etc. (cf. Cytowic 1989: 21ff.; Emrich et al. 2002: 33ff.). For genuine synesthetes these cross-modality perceptions are stable throughout their lifetime. We may also use the term pluralistic perception. These perceptions are in fact a highly individual capacity and hence are – according to reports of synesthetes – hardly accessible to verbal communication. From the perspective of linguistics, in contrast, we can define linguistic synesthesia as the co-occurrence of interdependent lexemes originally stemming from different sensory modalities. We may talk of a verbal simulation of synesthetic perception or of a linguistic creation of cross-modality illusions. In contrast to the former, the latter is an attempt to establish lexical units by convention. In German, for example, a synesthetic construction such as klirrende Kälte (English: ‘clinking cold’) is quite common, not very conspicuous and can thus be regarded as lexicalized. However, even odder combinations can be observed that are by no means established as lexical units. And the more unusual they are, the more clearly they still portray the very process of their emergence. As Wildgen points out in this volume very firmly, the apparently unclosable gap between the individuality of perception and the conventional, hence social and communicative character of language is a, perhaps the reason, why we found it necessary to organize the conference in 2002 and why this book has come into being. To put it plainly, linguistic synesthesia is one creative linguistic strategy which may be able to close this gap.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.10 (622-707)

 Peter Holz

. Some remarks on the modalities of sensory perception Since we necessarily talk about sensory perception all the time when engaged with the topic of synesthesia, it is necessary to throw a quick glance at the problem of how many different senses or sensory modalities there actually are in human perception. This is indeed not a trivial subject (cf. Bruggencate 1974: 188). The everyday subdivision of our perceptory apparatus into five senses (sight, hearing, touch, taste and smell) is more a common sense, than a scientific one. Further, it is deficient, somewhat arbitrary and thus misleading (cf. Bruggencate 1974: 188). Although I find the physiological = psychological model suggested by Shibuya et al. (in this volume) quite attractive to explain the relation between sensation, perception and cognitive processing (including synesthetic expressions), it is probably oversimplified, since it takes into account only the five classical senses. I have used a classification system suggested by Renate Zimmer (9 2001: 60) that takes into consideration: – – –

physiological criteria, namely different receptor cells, physical criteria, namely different kinds of stimuli and psychological criteria, namely the information gained

because this was more adequate for my analytical needs. To indicate that the different modalities are integrated and interdependent parts of a greater whole, Zimmer speaks of sensory systems rather than modalities. As we shall see in the next chapter, this classification is a helpful tool in analyzing synesthetic phenomena in cologne advertisements. The table is translated from German.

Table 2. Classification of sensory perception (cf. Zimmer 9 2001: 60) Sensory System

Sensory Organ

Receptors

Stimuli

Visual Auditive Tactile

Eye Ear Skin, hand, mouth

Light waves Acoustic waves Touch (incl. heat, pain)

Kinesthetic Vestibular Gustative

Tendons, muscles, joints Vestibular apparatus Mouth, oral cavity, tongue, palate Nose, nasal cavity

Photoreceptors Mechanoreceptors Tactile-, thermal-, mechanoreceptors Proprioceptors Mechanoreceptors Chemoreceptors, Mechanoreceptors Chemoreceptors

Olfactory

Body movement Acceleration Chemical stimuli Gaseous chemicals

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.11 (707-782)

Cognition, olfaction and linguistic creativity 

. Empirical evidence . The corpus Since everyday language cannot cope sufficiently with the precise description of olfactory qualities, the question arises, whether alternative linguistic strategies can be employed or even invented to compensate somehow for that lack. For my investigation I chose the domain of cologne advertising as a source for linguistic data because in that branch the problem of creating and establishing a language for scents is everyday business. The subject of my empirical analysis is a random choice corpus of 50 German advertisement texts for men’s cologne taken from the Internet (cf. www.douglasbeauty.com, February 19th 2002). The data analyzed in this paper is an outline of my doctoral thesis (Holz 2005), where I have analyzed of the whole corpus. Although cologne advertisements almost always use pictures to promote the product, language is also an important means to communicate messages about scents. Especially in Internet advertisements, we can find – next to attractive pictures – very compact texts conveying information about a scent and trying to describe its olfactory quality. . Synesthetic constructions in cologne advertisements I would like to shed light on three types of linguistic synesthesia which are very prominent in cologne advertisements. They can be observed on different levels of linguistic analysis and can be described in accordance with the poetic function of language. The text samples from the cologne-corpus are necessarily short, following the unwritten advertising law: much information in a small space. As a result of, they are semantically and morpho-syntactically structured in a very compact way. In these texts we can find language about scents in vitro, under laboratory conditions. Maybe they are not wholly representative for a common sense use of language on smells, but they constitute an interesting glance at the basic problem language has with the description of scents.

Table 3. Linguistic synesthesia on different linguistic levels Linguistic level

Synesthetic construction

Lexical Morpho-syntactical Textual

Immediate synesthetic expressions Patterns of lexical recurrence Semantic clustering

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.12 (782-880)

 Peter Holz

. Lexical level: Immediate synesthetic expressions 1 – Ad-hoc adjective compounds There are a lot of ad-hoc synesthetic adjective compounds in the corpus that appear quite odd, since they are neither lexicalized nor commonly used in everyday language. As can be seen in the examples in Table 4, the heads of the compounds refer to a different sensory domain than their specifiyers. They are supposed to describe the scent of a cologne, although none of them originates in the domain of olfaction; this category does not even occur. The perceptive impact of the single constituents is indicated in brackets. Of course, not every adjective occurring denotes a sensory quality and thus cannot be assigned to a certain sensory modality. Table 4. Ad-hoc synesthetic adjective compounds Specifyer

Head

Smooth (non-specific) Bitter (gustatory) Transparent (visual)

powdery (tactile) warm (tactile) frosty (tactile)

So we can observe the following combinations: – non-specific + tactile (mechanoreception) – 1x – gustatory + tactile (thermal) – 1x – visual + tactile (thermal) – 1x The projection principle is active here in the following way: lexemes that are equivalent with respect to the referential domain of sensory perception (i.e., paradigmatic relation) co-occur in the linearity of the text. They are thus projected onto the horizontal axis of combination (i.e., syntagmatic relation). An unusual effect on the reader is caused here by the inherent contrast between different modalities within the super ordinate paradigm of sensory perception. The same applies to the following examples. . Lexical level: Immediate synesthetic expressions 2 – Ad-hoc noun phrases In Table 5, I suggest taking note of the referential origin of the heads of the noun phrases. They refer, oddly enough, to audition and are specified by adjectives from non-auditory modalities. The reader is made to believe that the olfactory character of the cologne is described, but all the nouns refer to (or stem from) the auditory and not from the olfactory modality, although it is not at all likely that the advertisers have been

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.13 (880-915)

Cognition, olfaction and linguistic creativity 

Table 5. Ad-hoc synesthetic noun phrases Specifyer

Head (audition)

Fresh (thermal) Flowery (multi-modal)2 Cool (thermal) Ozony (olfactory)3 Spicy (gustative)

Prelude Chord Accent Chord Notes

listening to the cologne. We find them again in syntagmatic relation to adjectives from different modalities: – tactile (thermal) – 2 x – olfactory – 1 x – gustatory – 1 x – multi modal – 1x Figure 3 models the process by which immediate synesthetic expressions are created according to the projection principle. The sample adjective compound herb-warm (English: bitter warm) stems from the advertisement for the cologne Joop Homme.

Paradigm sensory perception

herb-warm (synesthetic)

herb (gustatory)

warm (tactile)

Figure 3. Modeling of linguistic synesthesia

. The adjective flowery is of non-specific sensory origin, since flowers have at least visual, tactile and olfactory properties. Thus it has been classified multi-modal. . Ozone, which is toxic in high concentration, smells in fact fresh in low concentration (cf. Vroon 1996: 83).

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.14 (915-989)

 Peter Holz

. Morpho-syntactic level: Patterns of lexical recurrence Building up significant patterns of lexical recurrence is another linguistic strategy for crossing modality borders and thus establishing an associative network of linguistic synesthesia. To exemplify what I mean by this, I have chosen the advertisement text for the Eau de Toilette Joop Homme. The focus in this text lies on the words: – aufregend (exciting) – 2x, – geradlinig (straightforward) vs. Geradlinigkeit (straightforwardness), – leidenschaftlich (passionate) vs. Leidenschaftlichkeit (passion). (S1) Ein aufregender (exciting) Duft (scent) für den geradlinigen (straightforward) und leidenschaftlichen (passionate) Mann (man). (S2) Ein Duft von blumigen, holzigen und exotischen Noten in Harmonie mit herb-warmen Akzenten. (S3) Die kühle Frische von Bergamotte vereinigt sich mit dem Feuer der HerzKopf-Note aus Zimt, Orangenblüten und Jasmin. Der exotische Fond von Sandelholz, Vetyver und Patchouli und ein Hauch von Ambra, Tabak, Moschus und Honig verschmelzen harmonisch ineinander. (S5) Der schmal geschnittene und klar gestaltete Flakon (flacon) unterstreicht die Geradlinigkeit (straightforwardness), seine warme und ungewöhnliche Farbgestaltung (color design) die Leidenschaftlichkeit (passion) und aufregende (exciting) Wirkung des Duftes. (www.douglasbeauty.com, 19.02.2002)

The significance of the recurrence of the lexemes emphasized in the text above can be found in the fact that the lexemes isolated are relevant according to recurrence not by chance. They characterize entities that occur in almost any text about cologne: the scent itself, the cologne (as a liquid), the bottle (as a container) and the male body. In the cognitive processing of cologne advertising the cologne (liquid), the bottle (container) and the body can function as ‘sign vehicle’ or ‘representamen’, since they can be pictured and thus made accessible to visual perception. The scent itself must be regarded as the absent ‘object’ of semiosis. Language then is organized around these four central semiotic units.

Table 6. Central semiotic units and their perceptual potential Perceivable entity

Corresponding modality

Scent Perfume (as liquid) Flacon (as container) Male body

Olfactory visual + tactile visual + tactile multi-modal

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.15 (989-1056)

Cognition, olfaction and linguistic creativity 

Interpretant Masculinity +x

Semiosis

Representamen Perfume Flacon Body

Object Scent

Figure 4. Modeling of semiosis in olfaction discourse

I termed the ‘interpretant’ masculinity + x, for the vague concept of what a man is or is supposed to be or is desired to be (by whoever) is the target of semiosis and very often associated with some other sensual (non-olfactory) experience (the semiotic terminology here again follows Peirce CP 1.339 and CP 2.230). Table 6 shows the central semiotic units in cologne advertisement texts as well as their perceptual potency (leaving out the abstract interpretant). The process of semiosis in cologne discourse can then be illustrated by the following model (Figure 4). The qualities excitement, straightforwardness and passion – in the Joop ad – linguistically represented as adjectives respectively nouns – are mixed with the central semiotic units. By this maneuver entities from the paradigm of sensory perception, but from different sensory modalities are linked together by the linguistic principle of contiguity. In S1, exciting, whose sensory domain is non-specific, is used to predicate scent, an olfactory category. In S5, the same adjective describes the effect of the scent which is supposed to be influenced by the design of the bottle and the color of the liquid. A semantic connection is thus created between olfaction and color vision and tactile perception. Whereas in S1 the adjectives straightforward and passionate refer to the noun man (i.e., male body). They are transformed into nouns in S5, where the shape of the bottle is said to emphasize the straightforwardness and its warm color design (bottle and liquid) the passion of the effect of the scent. A semantic connection is thus created between the multi-modal (male) body, and assumed tactile (i.e., thermal) and visual properties of the scent.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.16 (1056-1151)

 Peter Holz

Table 7. Associative semantic network of cross-modality predication Semiotic entity

Sensory modality

Qualifying lexemes

Scent Cologne (= colored liquid) Bottle (= colored container) Body (= man)

Olfactory visual, tactile, thermal visual, tactile, thermal multi-modal

Exciting exciting, passion straightforwardness straightforward, passionate

Table 7 demonstrates the associative semantic network that is built up by what I call cross-modality predication. . Textual level: Semantic clustering I have chosen the advertisement for Hugo Boss – Boss Elements Aqua to explain the concept of semantic clustering as another poetic feature that structures the text as a whole. A cologne certainly smells. Water normally does not. But in the Boss-example the suggested scent is permanently being associated with the sensuous experience of the body’s contact with water: (9) Die Frische und Lebendigkeit des Wasser für einen klaren Duft – Boss Elements Aqua heißt der Herrenduft, der so erfrischend und belebend wie das pure, klare Element Wasser ist. Sprudelnd-frisch und stimulierend sprüht dieser Duft vor maskuliner Vitalität und Energie – wie der Mann, der ihn trägt. Im Auftakt gibt sich der Duft spritzig und meeresfrisch durch Bergamott und einen elegant-ozonigen Akkord. Die Herznote ist würzig und floral bestimmt durch Piment und Jasmin, während der Fond mit der maskulinen Präsenz des Zedern- und Sandelholzes strahlt. Der praktische Zerstäuber sorgt für eine gute Verteilung des Eau de Toilettes und zeigt sich ebenso aquatisch-frisch in der Form, auf der in Glas gefangene Wassertropfen perlen. (www.douglasbeauty.com, 19.02.2002)

Water can in fact be perceived by all sensory modalities: – visual – auditory – tactile – kinesthetic – vestibular – gustatory – olfactory (only to a certain degree) Table 8 displays the isolated lexemes referring to the lexical field water (translation in brackets) in order to make clear what I have termed semantic clustering.

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.17 (1151-1199)

Cognition, olfaction and linguistic creativity 

Table 8. Lexemes from the lexical field water Nouns (5)

Verbs (2)

Adjectives (10)

Boss Elements Aqua Frische (freshness) Wasser (2x) (water) Wassertropfen (water drops)

perlen (to bubble) sprüht (to spray)

aquatisch-frisch (aquatic fresh) belebend (vitalizing) erfrischend (refreshing) klar (2x) (clear) meeresfrisch (oceanic fresh) pur (pure) spritzig (splashing) stimulierend (stimulating) sprudelnd-frisch (bubbling fresh)

By the exhaustive usage of water-related words the text creates – one might also say pretends – a semantic relation between pleasant water experiences and the alleged scent of this very cologne. (If this correspondence is legitimate or not is up to the smellers. Any perceptive evaluation from my standpoint lies beyond this investigation.) The multiple sensuous potential of water however, i.e., its capacity to be perceived through different sensory channels, is exploited verbally in order to create a sensory universe of water in which the apparent olfactory quality of the cologne is embedded. The projection principle as an indicator for the poetic function of language can be explained as follows. The referential overall information of the text can be paraphrased thus: Boss Elements Aqua is as pleasant as all the comfortable water-experiences that your body has ever had. The exhaustive use of lexemes from the paradigm water is actually redundant from the perspective of reference, since the repetition of water-lexemes does not provide new information. But from the perspective of poetics, all the waterlexemes belong to the same semantic paradigm and are from this viewpoint somewhat equivalent. They are projected into the linearity of the text, which is thus semantically structured in a rhythmic way, due to the recurrence of semantically equivalent words. . Summary and conclusions The referential function of language has fundamental problems in taking the dominant role in the olfaction discourse in general and in cologne discourse in particular. Scents evade precise referential access. Since the conative function is an inherent feature in advertisement texts for colognes, the persuasive effect cannot be achieved by mere positive descriptions of the product and its olfactory qualities. The empirical evidence I have provided and interpreted from the perspective of synesthesia as poetic feature shows that po-

JB[v.20020404] Prn:29/01/2007; 16:43

F: CELCR811.tex / p.18 (1199-1303)

 Peter Holz

etic means are deeply involved in supporting the conative function in the cologne advertising. So, if it comes to an analysis of descriptions of scents from a linguistic standpoint, the priority should lie more on the description of poetic features, than on any referential -semantic analysis. Linguistic staging of synesthesia seems to be an elementary constituent of cologne advertisement. The strategies I pointed out, namely synesthetic expressions and phrases, semantic clustering and patterns of lexical recurrence are actually stylistic features to be described in a linguistic analysis of the poetic function of language. I thus conclude that research on language about olfactory qualities necessarily requires the analysis of poetical features. Without the poetic perspective any semantic analysis of olfactory language turns out to be ungraspable and arbitrary. The assumption that linguistic reference can be used properly in communication about smells is simply false – and ex falso quod libet.

References Bear, M. F., Connors, B. W., & Paradiso, M. A. (2 2001). Neuroscience – Exploring the Brain. Baltimore et al.: Williams and Wilkins. Berlin, B., & Kay, P. (1969). Basic color Terms: Their Universality and Evolution. Berkeley: University of California Press. Bruggencate, G. ten. (1974). Allgemeine Sinnesphysiologie. In O. H. Gauer et al. (Eds.), Series: Physiologie des Menschen. Vol. 10: Allgemeine Neurophysiologie. München et al.: Urban & Schwarzenberg. Burdach, K. J. (1988). Geschmack und Geruch – Gustatorische, olfaktorische und trigeminale Wahrnehmung. Bern et al.: Huber. Cytowic, R. (1989). Synesthesia – A Union of the Senses. New York et al.: Springer. Dubois, D. (1997). Cultural beliefs as nontrivial constraints on categorization: Evidence from colors and odors. Behavioral and Brain Sciences, 20 (2), 188. Emrich, H. M., Schneider, U., & Zeidler, M. (2002). Welche Farbe hat der Montag? Synästhesie: Das Leben mit verknüpften Sinnen. Stuttgart et al.: Hirzel. Gschwind, J. (1998). Repräsentation von Düften. Augsburg: Dr. Bernd Wißner. Holz, P. (2005). Die Sprache des Parfums – Eine empirische Untersuchung zur Grammatik, Metaphorik und Poetizität des Parfumwerbetextes. Hamburg: Verlag Dr. Kovac. Jakobson, R. (1981). Linguistics and Poetics. In S. Rudy (Ed.), Selected Writings III (pp. 18–51). The Hague et al.: Mouton. Lakoff, G. (1987). Women, fire, and dangerous things – What categories reveal about the mind. Chicago et al.: University of Chicago Press. Peirce, Ch. S. (1931–1958, CP). Collected Papers, Vol. 1–6, C. Hartshorne and P. Weiss (Eds.), Vol. 7–8, A. W. Burks (Ed.). Cambridge, MA: Harvard University Press. Rosch, E. (1977). Human categorization. In N. Warren (Ed.), Studies in Cross-cultural Psychology. London et al.: Academic Press. Vroon, P., Amerongen, A. van, & Vries, H. de. (1996). Psychology der Düfte. Zürich: Kreuz. Zimmer, R. (9 2001). Handbuch der Sinneswahrnehmung. Freiburg et al.: Herder.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.1 (46-122)

Understanding synesthetic expressions Vision and olfaction with the physiological = psychological model Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru In this article, we introduce a cognitive model called the physiological = psychological model and discuss how one understands synesthetic expressions of language, e.g., warm color. We argue that there are two major types in synesthesia: one is the type whose synesthetic mapping is based on the co-occurrence of the senses, and the other on the emotional similarity of the senses. As for the former, we claim that the strength of sensory relations in daily experiences determines what type of synesthetic combination is possible (interpretable), e.g., warm color vs. ??red temperature. As for the latter, we argue that the interpretation of synesthetic expressions, such as fragrant music, is enabled by the synthesis of different emotional experiences.

.

Introduction

Synesthesia is the phenomenon of mapping from one sense modality onto another. According to Shen, for example, synesthesia is defined as the case of “conveying the perception of, or describing, one sense modality in terms of another” (Shen 1997: 47). An expression such as warm color is a classic example of a synesthetic expression. It involves the mapping from the tactile sense referred to by the adjective warm onto the visual sense referred to by the noun color. On the other hand, warm breeze is not a synesthetic expression, because both warm and breeze refer to the tactile sense, and there is no ‘sensory mismatch’ in this expression as one sees in warm color. This paper introduces a cognitive model called the physiological = psychological model (henceforth, the PP model). We discuss how we understand synesthetic expressions of vision and olfaction. Before presenting our analysis on synesthesia, we will first give an outline of the PP model in the next section. Section 3 discusses how synesthetic expressions, such as warm color and sweet smell, are understood in terms of sensory co-occurrence. The relationship between sensory experiences

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.2 (122-186)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

and synesthesia will be illustrated in Section 4. Synesthesia is motivated not only by sensory experiences, but also by emotional experiences. Section 5 discusses synesthetic expressions motivated by emotional experiences.

. The PP model and its structure Linguists have proposed various kinds of linguistic models. In Generative Grammar, for example, one uses a variety of formal constraints and rules to describe syntactic structures of language (e.g., Chomsky 1957, 1965, 1981, 1995). In Generative Semantics (another form of transformational grammar which derived from the theories of Chomsky from the 1960s), researchers aimed for an integration of syntax, semantics, and pragmatics with the use of derivational and transformational rules (e.g., Lakoff 1971; McCawley 1968). In Cognitive Linguistics, which grew out of Generative Semantics, one explores linguistic phenomena in terms of our general cognitive abilities (e.g., Lakoff 1987; Langacker 1987, 1991). The strategic changes from Generative Grammar to Cognitive Linguistics reflect the attempts by researchers to impose various kinds of constraints on linguistic models. Despite such a strategic change, however, there is one thing that seems to have been held as if it were an implicitly agreed method for the study of language. We could call such a method an ‘introspective approach’. In Generative Grammar, for example, although linguistic forms are what can be externally observable (e.g., utterances and written sentences), they are analyzed as mental representations and described in a highly introspective way. Likewise, it goes without saying that studies in Cognitive Linguistics are largely based on the introspection of the analyzer (e.g., semantic/ constructional networks). Introduction of the PP model into language research is an attempt to impose a different type of constraint onto a language model. The PP model is a cognitive model based on the findings of physiological and anatomical studies in brain science. With this model, we seek to develop a linguistic model which reflects the physiological nature and the anatomical structure of the brain.1 One of the most significant characteristics of the PP model is that in this model psychological and cognitive phenomena are equated with physiological activities of the brain, as brain scientists theorize about the brain (e.g., Beaumont 1983). In neuroscience, it is well acknowledged that particular domains of the brain correspond to particular cognitive functions (e.g., Gazzaniga et al. 1998; Rolls 1999).

. Development of the PP model is still in the process. Comments from readers on the brain scientific validity of this model are welcome.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.3 (186-235)

Understanding synesthetic expressions 

Eye Nose Ear Tongue

Skin & Muscle

Figure 1. Relationship between cerebral cortical areas and sense modalities

Figure 1 provides a rough sketch of the correspondence of various sense modalities with the relevant areas in the brain. The correspondence between the modal senses and the relevant brain areas as shown in Figure 1 has been confirmed through a long course of lesion studies in neuropsychology (e.g., see Finger 1994 for an introduction to the history of neuropsychology). In recent years, a variety of methods to observe brain activities have been developed with technological advances. Included in such modern approaches are, for example, ERP (event-related potentials), PET (positron emission tomography), CT (computed tomography), and fMRI (functional magnetic resonance imaging). Modern technology has helped to make dramatic progress in the studies of the structures and functions of the brain areas. In the PP model, meanings are defined as particular reactions in the brain against linguistic stimulus (e.g., linguistic sound or letter), and a process of language comprehension is described as a series of such reactions. This type of semantic view taken by the PP model follows the view that psychological and cognitive activities are equated with physiological activities in the brain. In a classical feature-based semantic model, meaning is defined as a bundle of binary abstract features (e.g., Katz & Fodor 1963; Chomsky 1965). For example, in a classical semantic analysis, one argues that such instances as infant bachelor are

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.4 (235-263)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

Figure 2. Cognitive network in the PP model

not acceptable, since the binary feature [–ADULT] of infant does not conform to the feature [+ADULT] of bachelor (Taylor 1995: 31). Such a semantic analysis is a typically introspective approach. In such an approach, one describes meanings of linguistic expressions with a symbol-manipulation view, and conducts an analysis without any reference to the properties of the brain. In Cognitive Linguistics, although such a formal symbolic semantic analysis has explicitly been abandoned, semantic analysis in this framework still heavily depends on the introspection of the analyzer. The PP model provides a new type of semantic analysis. The model captures the physiological and anatomical properties of the brain. In this model, language research is based, or ‘grounded’, on the actual brain structures and their functions. In this paper, we will show, through the case study on synesthesia, that employment of the PP model has a potential to lead us to a better understanding of the relationship between language and cognition. In the PP model, types of cognitive domains and interrelations among them are determined according to the anatomical structure of the brain. The present version of the model assumes four major domains: the External Primary Domain, the Internal Primary Domain, the Internal Secondary Domain, and the External Secondary Domain. Each of these basic domains subsumes several sub-domains. Figure 2 gives the overall sketch of the network of the cognitive domains assumed in the PP model. The PP model describes the comprehension process of information as the various activities in these cognitive domains. The External Primary Domain contains five sub-domains: Ear, Eye, Nose, Tongue, and Skin and Muscle. The activities assumed to be observed in the External Primary Domain are physiological reactions in the receptors of the special senses, such as ears, eyes, nose, tongue, and reactions in the somatosensory recep-

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.5 (263-369)

Understanding synesthetic expressions 

Table 1. Relationship between the Internal Primary Domain and brain anatomy Internal Primary Domain Tactile Sense Gustatory Sense Olfactory Sense Visual Sense Auditory Sense Sensory Association

Brodmann’s areas 1, 2, and 3 Brodmann’s area 43 and the insula the pyriform cortex, the entorhinal cortex, and the orbitofrontal cortex Brodmann’s areas 17, 18, and 19 Brodmann’s areas 41 and 42 the temporal association area and the parietal association area

Table 2. Relationship between the Internal Secondary Domain and brain anatomy Internal Secondary Domain Emotion Judgment Situation Cognition Behavior Control

the amygdala and hypothalamus the frontal association area the hippocampus the motor cortex (Brodmann’s area 4), premotor cortex, supplementary motor cortex (Brodmann’s area 6)

tors. Also involved in this domain are compulsory movement and transformations of the body. The Internal Primary Domain contains sub-domains, such as the Tactile Sense, the Gustatory Sense, the Olfactory Sense, the Visual Sense, the Auditory Sense, and the Sensory Association. The sub-domains of the Internal Primary Domain play crucial roles in the semantics of language. Table 1 shows the correspondence between these sub-domains and the relevant anatomical areas in the brain. The Internal Secondary Domain consists of Emotion, Judgment, Situation Cognition, and Behavior Control. These sub-domains and their anatomical counterparts are shown in Table 2. The fourth major domain assumed by the present version of the PP model is the External Secondary Domain which contains the sub-domain, Muscle. The activities observed in this domain involve voluntary movement and transformations of the muscles, including linguistic actions, such as speaking and writing. In the PP model, comprehension is equated with a series of reactions in the various cognitive domains that are in associative relation with given linguistic stimulus (i.e., linguistic sound or letter). For example, in this model the meaning of the word apple is analyzed as particular reactions in the brain caused by the

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.6 (369-420)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

particular linguistic sounds of apple. (1) is a highly schematic description of the reactions assumed to occur following the input of apple.2 (1) Situation: Occurrence of the linguistic sounds of apple Reactions in the hearer: – Vibration of the eardrum (tympanum) [the linguistic sounds of apple] – Reaction in the auditory nerve (inner ear nerve) [the linguistic sounds of apple] – Reaction of the Auditory Sense [the linguistic sounds of apple] – Reaction of the Sensory association [the association between the auditory sense of the linguistic sounds of apple and the senses related to the elements of an apple] – A series of reactions in the sensory areas (caused through the reaction of the Sensory Association following the linguistic input of apple)

In our accounts, activation of the relevant brain areas caused by given linguistic sounds is the meaning of the unit. In the case of apple, the linguistic sounds cause a series of reactions in the sub-domains of the Internal Primary Domain, such as the Tactile Sense, the Gustatory Sense, the Olfactory Sense, and the Visual Sense, through the reactions of the Sensory Association. One perceives various features with an apple. Reactions in the sensory areas caused by apple correspond to such features of an apple. The meaning of apple is a compound of a variety of sensory reactions in the brain of the hearer. Figure 3 shows the comprehension process of apple. The arrows in bold indicate activation. The assumption that the meaning of a linguistic unit (e.g., apple) is equated with a certain set of sensory reactions which take place in our perception of an entity (apple) is by no means novel news to brain science. For example, in Allport’s theory of distributed semantic memory, it is argued that “the same neural elements that are involved in coding the sensory attributes of a (possibly unknown) object presented to eye or hand or ear also make up the elements of the auto-associated activity-patterns that represent familiar object-concepts in ‘semantic memory’” (Allport 1985: 53), and that the elements of activity-patterns representing lexicons are associatively linked with the meanings represented, in a distributed way, over ‘attribute domains’ (e.g., action oriented elements, kinesthetic elements, tactile el-

. Here, we distinguish the description of a process from the description of the content. For example, in the description of “Reaction of the Tactile Sense [warmth]”, “Reaction of the Tactile Sense” should be understood as the description of the process in question and [warmth] as ‘a reaction corresponding to some warmth in the tactile sense’ (mental content).

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.7 (420-420)

Understanding synesthetic expressions 

“apple”

Step 1 Reaction of the Auditory Sense

Step 2 Reaction of the Sensory Association Occurrence of meaning

Figure 3. Meaning of apple

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.8 (420-480)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

ements, visual elements, and auditory elements). The PP model’s assumption as illustrated in Figure 3 is in accordance with such a neural model as Allport’s.3

. Comprehension based on sensory co-occurrence In Section 1, we introduced warm color as a classic example of synesthesia (see example (2)). Let us discuss what motivates the understanding of such a synesthetic expression. (2) Or take a pot you already have and paint it in a warm color, such as bright red or fuscia.4

In daily life, a particular sense tends to co-occur with other senses. The experience of fire, for example, includes warmth, a burning smell, shape and colors of flame, sound of sparks, etc. Likewise, the experience of electric stoves includes warmth, shape and colors of the heater, electric noises, etc. Learning of simultaneous proximity between sense A and sense B forms an associative relation between the two senses. In the case of warm objects, sensory co-occurrence between the tactile sense of warmth and the visual sense of reddishness is high, which we argue leads to the formation of a sensory association of these two senses. Through sensory experiences, associative relations between two senses are formed, and the activation of one sense leads to the co-activation of the other. We argue that such a sensory associative relation motivates the understanding of expressions such as warm color. Comprehension of warm color can be described in the PP model as in (3).5 (3) Situation: Occurrence of the linguistic sounds of warm color Reactions in the hearer: – Reaction of the Auditory Sense [the linguistic sounds of warm] . Allport’s attribute domains roughly correspond to the sub-domains in the PP model, but it should be noted that Allport’s model and the PP model are not identical, but differ from each other in the classification of domains. As already mentioned in note 1, the PP model is still being developed. For example, the model needs to improve on the classification of the domains and the sub-domains, as well as the classification of sense modalities. . www.paintinfo.org/articles/warmup.htm . In this paper, due to the nature of the target phenomenon (i.e., synesthesia), the arguments will mainly concern the first three cognitive domains assumed in the PP model (see Figure 2). It should, however, be emphasized that comprehension of a given linguistic unit (e.g., phrase, construction, context, etc.) essentially involves the reactions in all four domains. See, for instance, Nozawa (2002) for the arguments of the applicability of the PP model to other linguistic phenomena.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.9 (480-573)

Understanding synesthetic expressions

– Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of warm and the tactile sense of warmth] – Reaction of the Tactile Sense [warmth] – Priming in the Sensory Association [the association between the tactile sense of warmth and other senses that are in strong associative relation with warmth] – Reaction of the Auditory Sense [the linguistic sounds of color] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of color and the visual sense of some colors] – Reaction of the Visual Sense [reddishness]

Figure 4 is the depiction of such processing. The zigzag arrows in bold indicate inhibition. Synthesis of successive reactions caused by the linguistic inputs of warm and color enables a particular activation of the Visual Sense about reddishness. The ‘mismatch’ of sense modalities of tactility and vision denoted by warm and color is solved by the co-occurring associative relation between them.6 A similar principle underlies a synesthetic expression, such as sweet smell. Consider the following example: (4) The incense producer’s art comes in with the delicate process of blending essential oils and other gums to produce a sweet smell.7

When eating an ice cream, for example, the gustatory sense of sweetness co-occurs with the olfactory sense of a particular smell, such as vanilla. We argue that such sensory experiences of an entity (e.g., ice cream) give the basis for the production and comprehension of expressions, such as sweet smell. (5) provides an account for the understanding of sweet smell. Figure 5 is the depiction of such processing. (5) Situation: Occurrence of the linguistic sounds of sweet smell Reactions in the hearer: – Reaction of the Auditory Sense [the linguistic sounds of sweet] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of sweet and the gustatory sense of sweetness] – Reaction of the Gustatory Sense [sweetness] – Priming in the Sensory Association [the association between the gustatory sense of sweetness and other senses that are in strong associative relation with sweetness] . Strictly speaking, the strength of associative relations between different sense modalities differs in each speaker. This implies that each speaker differs in her acceptability judgment and the comprehension mode of a given expression. Description of such subtle aspects is an interesting issue, but it is beyond the scope of this paper. . The Times (London, England), Dec 24, 2001: 12.



JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.10 (573-579)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

– Reaction of the Auditory Sense [the linguistic sounds of smell] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of smell and the olfactory sense of some smell] – Reaction of the Olfactory Sense [a particular smell (e.g., vanilla)]

Step 0 Experience of something warm (fire)

“warm”

Step 1 Reaction of the Auditory Sense Reaction of the Sensory Association Occurrence of the meaning (warmth)

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.11 (579-579)

Understanding synesthetic expressions 

Step 2 Priming in the Sensory Association

“color”

Step 3 Reaction of the Auditory Sense Reaction of the Sensory Association Inhibition of the conflicting senses Occurrence of the meaning (reddishness)

Figure 4. Meaning of warm color

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.12 (579-579)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

Step 0 Experience of sweet food (ice cream)

“sweet”

Step 1 Reaction of the Auditory Sense Reaction of the Sensory Association Occurrence of the meaning (sweetness)

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.13 (579-579)

Understanding synesthetic expressions

Step 2 Priming in the Sensory Association

“smell”

Step 3 Reaction of the Auditory Sense Reaction of the Sensory Association Inhibition of the conflicting senses Occurrence of the meaning (smell of vanilla)

Figure 5. Meaning of sweet smell



JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.14 (579-639)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

. Sensory experiences and synesthesia One of the most remarkable aspects of synesthesia is the directionality of possible mappings. (6) shows that mapping from the tactile sense onto the visual sense is possible, but mapping into the opposite direction is not. (7), on the other hand, shows that mapping from the gustatory sense onto the olfactory sense is possible, but not into the opposite direction. This section discusses the underlying motivation of what makes the synesthetic mapping in (6a) and (7a) possible. (6) a. warm color b. ?? red temperature (7) a. sweet smell b. ?? smelly taste

Figure 6 shows Williams’ (1976) generalization that mapping from the tactile and gustatory senses onto the others occurs quite naturally, but mapping into the opposite directions does not (i.e., from the olfactory, visual, and auditory senses onto the others):8 color touch

taste

smell

dimension sound

Figure 6. Generalization by Williams (1976: 463)

As explained in Section 3, we argue that comprehension of warm color is made possible by the co-occurrence of the tactile sense and the visual sense which one learns in daily sensory experiences. The important question regarding the examples in (6) is: What is it that causes a different acceptability between (6a) and (6b)? The tactile and visual senses are likely to co-occur, but what makes mapping from the tactile sense onto the visual sense possible, while the opposite directional mapping is not acceptable? The same question can be found also in (7). Although the gustatory and olfactory senses seem to frequently co-occur in daily sensory experiences, why do the examples in (7) show a different acceptability? The acceptability of the given examples is predicted by Williams’ generalization, but how does the PP model explain such a phenomenon? Sensory co-occurrence between two sense modalities forms a sensory association of the two senses. A sensory association, however, does not only specify what . Dimensional adjectives include flat, deep, etc. The present paper ignores such adjectives.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.15 (639-709)

Understanding synesthetic expressions 

types of senses are involved, but it also specifies the strength of the association. For example, the tactile sense and the visual sense co-occur frequently, but the strength of the sensory association which the tactile sense has with the visual sense differs from what the visual sense has with the tactile sense. We call the specification of a sensory association the ‘structure of sensory experiences’. The sensory association which the tactile sense has with the visual sense is stronger than that of the opposite. In daily sensory experience, the visual sense does not normally presuppose the involvement of the tactile sense. For example, looking at a chair gives one some visual information of the object, but it does not provide any tactile information. In contrast, touching a chair normally involves the visual information of the object as well. Frequency of a sensory co-occurrence determines the strength of the sensory association. In the case of the tactile sense and the visual sense, the pattern of the sensory association from the visual sense to the tactile sense is not as strong as that from the tactile sense to the visual sense. We argue that the strength of sensory association plays an important role in determining the possible synesthetic mapping. Synesthetic expressions as we have been discussing in this paper consist of a modifier (e.g., warm in warm color) and a modified element (e.g., color in warm color). In the semantic structure of synesthetic expressions, it is preferable that the first element (i.e., the modifier) evokes some reaction which is in strong associative relations with the reactions evoked by the second element (i.e., the modified), otherwise the semantic processing of a sequentially given linguistic stimulus would not be performed smoothly. In warm color and red temperature, the former forms a preferable sensory association for successful comprehension. In warm color, the modifier activates the relevant tactile sense which activates the co-occurring relevant visual sense, since as discussed above the tactile sense often involves the use of the visual sense. The visual sense evoked by the following element of color then matches with the priming pattern caused by the modifier, and the comprehension of this expression converges. In contrast, red temperature does not form a preferable sensory association for successful comprehension. In this expression, the activation of the relevant visual sense caused by the modifier red does not cause the co-activation of the tactile sense strongly enough, since as noted above, the visual sense does not often involve the use of the tactile sense. In red temperature, the modifier does not evoke a necessary sensory association (i.e., the tactile information, in this case) for the following element to match with, and hence an anomaly arises.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.16 (709-757)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

To sum up, in synesthesia of vision and tactility, the modifier should be the one that refers to the tactile sense, not the visual sense. In synesthesia, comprehension of a given expression depends heavily upon the structure of sensory experiences.9 The same principle seems to be operating in the examples of (7). We argue that sweet smell (7a) is acceptable, because the modifier evokes some gustatory experience, which involves a particular olfactory experience as well. As well as warm color, in sweet smell the sensory co-occurrence caused by sweet matches with the olfactory sense evoked by smell. On the other hand, smelly taste (7b) sounds anomalous. This is due to the sensory function of the olfactory sense evoked by smelly. Unlike the gustatory sense which is likely to co-occur with the olfactory sense, the olfactory sense does not always involve the activation of the gustatory sense. In synesthesia of gustation and olfaction, the modifier should be something that refers to the gustatory sense (not the olfactory sense), so that the sense evoked by the modified element successfully matches with the priming associative pattern caused by the modifier. When the modifier is an olfactory adjective and the following element is something that evokes some gustatory sense (e.g., smelly taste), the modifier does not evoke a necessary sensory association (i.e., the gustatory information) for the following element to match with.

. Comprehension based on emotional experiences (8a) is not a synesthetic expression, since fragrant and smell both refer to the sense of olfaction. On the other hand, (8b) is a synesthetic expression, because it involves two different sensory words: fragrant (an olfactory adjective) and music (an auditory noun). (8) a.

Plants such as rhododendrons will give a fragrant smell through their blooms, while junipers, pines and spruces will give a forest fragrance through their foliage.10 b. This fragrant music is as rare and precious as the essential oils used in the titular ancient healing art. Created in partnership between one of Europe’s leading beauty therapists and a leading holistic musician, this

. A similar expression to warm color can also be found in other languages (e.g., atatakai iro, in Japanese). We suggest that the cross-linguistic availability of similar synesthetic expressions is ascribed to the physiological function of the brain. Bodily experiences are not language specific, but universal among human beings living in a similar ecology. As long as people share a similar kind of sensory experience of an object (e.g., fire), similar sensory co-occurrence is learned. See Shibuya and Nozawa (2003), where it is argued that adaptive behavior contributes to the formation of the structure of sensory experiences. . http://www.campbellsnursery.com/Spotlight/fragrant_gardening.htm

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.17 (757-873)

Understanding synesthetic expressions 

album creates the perfect mood and atmosphere to complement the experience of pure relaxation brought about by aromatherapy.11

We have discussed examples whose understanding is attained through activation of relevant co-occurring sensory areas in the brain. Examples like (8b) cannot be explained by the sensory constraint proposed in the preceding section, because some type of fragrance and some kind of sound are not likely to co-occur in daily life. Given the discussions so far, lack of sensory co-occurrence (i.e., experience) implies unacceptability of the expression. If so, then how do we explain the synesthetic mapping of (8b)? We suggest that an understanding of examples like (8b) involves association between a sense and an emotional value.12 (9) is a description of the process of the comprehension of (8b). (9) Situation: Occurrence of the linguistic sounds of fragrant music Reactions in the hearer: – Reaction of the Auditory Sense [the linguistic sounds of fragrant] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of fragrant and the olfactory sense of fragrance] – Reaction of the Olfactory Sense [fragrance] – Reaction of the Sensory Association (the association between the olfactory sense of fragrance and the emotion of pleasure) – Reaction of Emotion [emotion of pleasure] – Priming in the Sensory Association [the association between the emotion of pleasure and the senses that are in strong associative relation with such an emotion] – Reaction of the Auditory Sense [the linguistic sounds of music] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of music and the auditory sense of some kind of music] – Reaction of the Auditory Sense [some music that causes emotion of pleasure]

The olfactory sense of fragrance causes some kind of emotion of pleasure. Understanding this expression means matching relevant reactions: particular reactions in the Auditory Sense caused by the linguistic input of music match with the reactions in the Sensory Association with the emotion of pleasure caused by the preceding input of fragrant. Interpretation of this expression is achieved by the . http://www.lightworks.com/MonthlyAspectarian/2000/March/300-13.htm . Other authors also point out emotional experiences as motivations for the description of scents as well as colors. See the chapters by Holz, Wyler, and Graumann in this volume.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.18 (873-874)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

synthesis of pleasant feeling caused by fragrant and a particular pattern of auditory reaction caused by music. The mismatch of different sense modalities is solved by the matching of the reactions in the Internal Primary Domain and the emotional reactions in the Internal Secondary Domain. Figure 7 is a depiction of such an interpretative process. Notice here again that bodily experiences of a hearer play crucial roles in the successful comprehension of this expression (see steps 0a and 0b in the figure). The olfactory sense of fragrance causes an emotion of plea-

Step 0a Experience of fragrant air

Step 0b Experience of pleasant music

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.19 (874-874)

Understanding synesthetic expressions 

“fragrant”

Step 1 Reaction of the Auditory Sense Reaction of the Sensory Association Occurrence of the meaning (fragrance)

Step 2 Reaction of the Sensory Association Reaction of Emotion (pleasure)

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.20 (874-874)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

Step 3 Priming in the Sensory Association

“music”

Step 4 Reaction of the Auditory Sense Reaction of the Sensory Association Inhibition of the conflicting senses

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.21 (874-925)

Understanding synesthetic expressions 

Step 5 Occurrence of the meaning (pleasant music)

Figure 7. Meaning of fragrant music

sure, through which the synthesis of pleasant feeling and the auditory reactions is achieved. There are various types of emotions, such as pleasure, fear, anger, sadness, disgust, etc. The emotional reactions against particular perceptual stimulus affect the interpretation of a given synesthetic expression. Consider the example in (10). (10) A bad fit will make any dress ugly, that’s for sure. But a busy pattern, a toosnug or too-puffy cut, or a loud color will compound the situation.13

As well as fragrant music, loud color in (10) also might appear to be a counterexample of the arguments presented in sections 3 and 4, because loudness and some kind of color are not likely to co-occur in daily life. We suggest that an understanding of loud color is also achieved through the synthesis of the sensory reactions and the emotional reactions caused by the lexical components. The process of comprehension of (10) can be described as in (11): (11) Situation: Occurrence of the linguistic sounds of loud color Reactions in the hearer: – Reaction of the Auditory Sense [the linguistic sounds of loud] . www.bozemandailychronicle.com/articles/2003/06/03/features/lifestyles/bridesmaidsbzlife.txt

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.22 (925-976)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

– Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of loud and the auditory sense of some kind of noise] – Reaction of the Auditory Sense [some kind of noise] – Reaction of the Sensory Association (the association between the auditory sense of some kind of noise and the emotion of disgust) – Reaction of Emotion [emotion of disgust] – Priming of the Sensory Association [the association between the emotion of disgust and the senses that are in strong associative relation with such an emotion] – Reaction of the Auditory Sense [the linguistic sounds of color] – Reaction of the Sensory Association [the association between the auditory sense of the linguistic sounds of color and the visual sense of some colors] – Reactions of the Visual Sense [some color that causes emotion of disgust]

In loud color, particular reactions in the Auditory Sense caused by the linguistic input of color match with the reactions in the Sensory Association with the emotion of disgust caused by the preceding input of loud, through which the synthesis of sensory reaction and emotional reaction becomes possible.14

. Conclusion This paper introduced a cognitive model called the physiological = psychological model. The PP model is theorized based on the physiological and anatomical properties of the brain. With such a model, we introduced a new way of language analysis to understand the relationship between language and cognition. With the PP model, we discussed how one understands synesthesia. There seems to be two major types in synesthesia: one is the type whose synesthetic mapping is based on the co-occurrence of the senses, and the other on the emotional similarity of the senses. In this paper, we did not discuss contextual effects on synesthesia, but it should be noted that comprehension of synesthetic expressions is subject to contexts (just like other linguistic expressions). Consider the following examples: (12) a. I heard the heavy sound of elephant footsteps. b. I heard the heavy sound of the funeral procession. c. I heard the heavy sound of the committee sighs.

. An interesting discussion of the relationship between color expressions and expressions from the musical domain can be found in the chapter by Plümacher (in this volume).

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.23 (976-1055)

Understanding synesthetic expressions 

Example (12) illustrates a case where the same synesthetic phrase heavy sound is understood by a different type of experience. In (12a), it is the co-occurrence of the senses that allows the comprehension of the phrase, while in (12c) it is not the sensory co-occurrence but rather the emotional similarity of the senses that enables the comprehension. (12b) is a case where it is hard to pin down which type of experience is in operation. Examples like (12b) reveal the fact that comprehension of synesthesia is made on the scale between the co-occurrence of the senses and the emotional similarity of the senses which forms a fuzzy boundary. The main concern of this paper was to give an explanation of synesthesia of vision and olfaction. Fine-grained analysis of synesthesia should of course not be limited to these two sense modalities. For more extensive accounts of synesthesia, see Shibuya and Nozawa (2003), where synesthetic expressions of various sense modalities are discussed from a cross-linguistic perspective.

Acknowledgements We would like to thank William Croft for comments on an earlier version of this paper. We are also grateful to Masa-aki Yamanashi for his encouragement. Special thanks go to Stefan Fuhs and Helen Bissett. Any remaining errors are of course our own.

References Allport, D. A. (1985). Distributed memory, modular subsystems and dysphasia. In S. Newman & R. Epstein (Eds.), Current Perspectives in Dysphasia (pp. 32–60). Edinburgh: Churchill Livingstone. Beaumont, G. J. (1983). Introduction to Neuropsychology. New York: Guilford Press. Chomsky, N. (1957). Syntactic Structures. The Hague: Mouton. Chomsky, N. (1965). Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Chomsky, N. (1981). Lectures on Government and Binding. Dordrecht: Foris. Chomsky, N. (1995). The Minimalist Program. Cambridge, MA: MIT Press. Finger, S. (1994). History of neuropsychology. In D. W. Zaidel (Ed.), Neuropsychology (pp. 1–28). San Diego: Academic Press. Gazzaniga, M. S., Ivry, R. B., & Mangun, G. R. (1998). Cognitive Neuroscience: The Biology of the Mind. New York: W.W. Norton. Katz, J. J., & Fodor, J. A. (1963). The structure of a semantic theory. Language, 39, 170–210. Lakoff, G. (1971). On generative semantics. In D. D. Steinberg & L. A. Jakobovits (Eds.), Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology, (pp. 232– 296). Cambridge: Cambridge University Press. Lakoff, G. (1987). Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: The University of Chicago Press.

JB[v.20020404] Prn:29/01/2007; 11:20

F: CELCR812.tex / p.24 (1055-1100)

 Yoshikata Shibuya, Hajime Nozawa and Toshiyuki Kanamaru

Langacker, R. W. (1987). Foundations of Cognitive Grammar, Vol. I: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, R. W. (1991). Foundations of Cognitive Grammar, Vol. II: Descriptive Application. Stanford: Stanford University Press. McCawley, J. D. (1968). The role of semantics in a grammar. In E. Bach & R.T. Harms (Eds.), Universals in Linguistic Theory (pp. 124–169). New York: Holt, Rinehart and Winston. Nozawa, Hajime (2002). Language Theory of Behavior and Cognition. MA thesis, Graduate School of Human and Environmental Studies, Kyoto University. Rolls, E. T. (1999). The Brain and Emotion. Oxford: Oxford University Press. Shen, Y. (1997). Cognitive constraints on poetic figures. Cognitive Linguistics, 8, 33–71. Shibuya, Y., & Nozawa, H. (2003). Constraints on synaesthesia. In Proceedings of the Annual Meeting of the Berkeley Linguistics Society, 29 (pp. 403–414). Berkeley: Berkeley Linguistics Society. Taylor, J. R. (2 1995). Linguistic Categorization: Prototypes in Linguistic Theory. Oxford: Clarendon Press. Williams, J. M. (1976). Synaesthetic adjectives: A possible law of semantic change. Language, 52, 461–478. Yamanashi, M. (1988). Hiyu to rikai (metaphor and understanding). Tokyo: University Press.

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.1 (47-133)

Die Summe hiervon ist diese: Die Sache der Sinne ist, anzuschauen; die des Verstandes, zu denken. Denken aber ist Vorstellungen in einem Bewußtsein vereinigen. (Immanuel Kant, Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten können, 1783)

Olfactory and visual processing and verbalization Cross-cultural and neurosemiotic dimensions Tatiana V. Chernigovskaya and Victor V. Arshavsky The paper discusses the neurological basis for olfactory and visual preferences governing human behavior, with the right cerebral hemisphere (RH) playing the dominant role, both in individuals and in types of culture in which olfaction is an important part of the semiosphere. Subjects with RH reactions showed a reliable cross-correlation of biopotentials in the RH when stimulated by odors preferable for them. Classification and verbalization of colors also demonstrates significant differences in the types of strategies used by RH vs. LH subjects. Most professional testers of odors appear to be RH personalities. The important role of cultural, as well as of linguistic, backgrounds is stressed. Right hemispheric sensory processing correlates with adaptation and resistance to stress and somatopsychic diseases.

There is a poor relationship between language and the olfactory world: identification by name of odors is very difficult, and there is not a vast vocabulary for orders within the human mental lexicon. In fact it is synthetic by nature – we either use vocabularies for taste, or for color, or even for tactile and auditory sensations. The visual – especially color – semiosphere is probably the most thoroughly elaborated by the majority of human languages, while olfactory is the less verbalized of all sensory modalities, probably due to its subconscious nature and cultural prohibitions. Other modalities, such as tactile, auditory, and gustatory, occupy intermediate po-

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.2 (133-205)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky

sitions on this scale. Olfactory information is known to be complex, uncertain (fuzzy) and extremely difficult to verbalize (Harper et al. 1968; Zellner & Kautz 1990). It should be emphasized that we face the difficulty of ‘translating’ olfactory Gestalt messages into discrete linguistic terms. Color is a spatial sign, sound is a temporal sign, odor and taste are probably both (Chernigovskaya 1995). Synesthetic perception is most expected in this context. The iconicity of synesthesia is evident; the iconicity of olfaction can be described as a kind of cognitive synesthesia, as it has no vocabulary of its own. It involves memory, associations with past images, and reveals episodes of personal experiences (cf. Shibuya et al. in this volume: a cognitive psysiological = psychological model of synesthetic expression). In spite of the fact that humans are believed to have lost the signal-values of olfactory stimuli (Hoffmeyer 1996), behavior is still partly regulated by them, even if it is not conscious ly understood what these signs mean to us (Hoppe 1977; Hanisch 1982; Steiner & Neumann 1982; Ugolev & Chernigovskaya 1989; Arshavsky & Goldstein 1994; Schaal et al. 1998; Weinstein 2003). There are gender and age differences in both perception and naming of odors: females and younger individuals do it better than males and older subjects. Categorization of olfactory stimuli is also not an easy task for humans: the grouping of odors might be driven by different multimodal factors, such as personal experience, personal memories and broader Gestalts, emotional backgrounds, and current states (Ugolev & Chernigovskaya 1989; Chernigovskaya & Arshavsky 1994). Moreover, the grouping of odors is different across different situations and across different testing sessions. They are personal, emotional, and very unstable. Associations are also very ‘old’ – carrying the memory of generations, possibly not only human generations. As we know, in other species behavior – not only sexual – is to a great extent guided by olfaction, for many species it is the main language for communication. Recent data show that the amazing ability of animals (much worse in humans) to discriminate an individual olfactory stimulus in the mixture of unfamiliar stimuli might be caused by a large number of independent channels, with elements of binary coding allowing a rough approximation of the level of each channel arousal, and even the possibility of separate transfers of information about stimulus quality and intensity (Minor & Krutova 2001) The question of the neural organization of human olfactory processing is not new. Reminiscences and even déjà vu or vivid olfactory hallucinations caused by temporal lobe epilepsy were first described more than a century ago by John Hughlings Jackson. Even in a normal situation we can find ‘fits’ of anosmia, hypo- or a hyperosmia, which appear to be associated with emotional states, endocrine status, and personal experiences, very often of a subconscious and limbic nature. Hypersomia is also seen in patients with hyper-dopaminergetic states and in patients with Tourette’s syndrome (Sacks 1987). Smells are almost never neutral if the in-

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.3 (205-290)

Olfactory and visual processing and verbalization 

tensity is well above threshold. Still more amazing is when an olfactory stimulus is below the conscious threshold and is as if ‘not perceived’, causing either changes in behavior and/or specific and very personal (often sexual) associations. – Olfactory memories can be so vivid that one could almost talk of controlled hallucinations. Olfactory hallucinations coming with epileptic seizures are most often met in patients with focal right-sided lesions (Whitton 1978; Kupertman 2003; Kuperman & Zislin 2003). As Holz suggests (this volume), smells cannot be categorized in terms of classical logical semantics or in terms of lexemes, rather they should be discussed in prototype theoretical perspective. We also agree with Dubois when she argues that in cases of audition and olfaction the gap between linguistic and cognitive categories is much larger than in visual domain (Dubois 2000 and this volume). One of the most important features of odor recognition memory is that it is only slightly influenced by the length of retention intervals relative to that for pictures and words. Odors are poorly remembered initially and well retained over time, and they seem to be acquired holistically, in a Gestalt fashion. Zucco’s main hypothesis lies on the assumption that odors do not give rise to a conscious representation and could be stored in memory at a subconscious level. Conscious access to the olfactory trace is not possible except for acquisition and non-intentional retrieval. Storage and access to olfactory stimuli in memory, then, should not imply an effort but be automatic. Odor labeling and verbal rehearsal has no effect on subsequent recognition memory (Zucco 2003; Engen 1982, 1991). Danthiir et al. (2001) have shown that an olfactory memory ability is independent from other higher-order abilities. It is established that different smells can stimulate or tranquilize and therefore regulate interpersonal relations and different types of behavior, including sexual. It is explained by direct connections of olfaction with the limbic system, causing evident emotional effects as well as those of autonomic nervous system (Economides 1986; Hanisch 1982; Staubli 1987; Steiner & Neumann 1982). In the process of psychological adaptation odors can have well-expressed signal meaning. On the contrary, damage to this sphere can cause alarm or even depressive states. It was shown that individuals with different types of hemispheric reactions to external stimuli demonstrate various alarm levels and different compensation. Alarm in its turn can cause a whole set of psychosomatic pathologies and neuroses. As a consequence a kind of aromo-correction might be necessary. There are reasons to suggest that opposite cognitive styles are differently distributed in populations of different cultures (Chernigovskaya 1993, 1994, 1999; Rotenberg & Arshavsky 1997). According to Ornstein (1972), it is Western civilization that stimulates the development of left hemispheric functions, while Eastern civilization is more dependent on the abilities of the right hemisphere. This is the reason why the altered states of consciousness (yoga, meditation) which, accord-

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.4 (290-356)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky

ing to Ornstein, are based on the right hemispheric skills, are mostly used in the Eastern civilizations. As an example: the native population of Chukotka lives in a cultural context which is quite different from the average European. There are many hunters and reindeer owners. Although young and middle-aged generations have received education of different levels, the style of life is generally almost the same as centuries ago. All cultural traditions are strong, most representatives of the native population are highly skilled in spatial orientation, in non-verbal communication and in special kinds of art, such as bone carving. The proportion of individuals with low levels of education (fewer than 8 years) engaged in comparatively simple agricultural jobs is significantly higher in the native population than among immigrants. Communication is also rather specific: verbal communication is limited, laconic, and rigorous, with many non-verbal ritualized actions. It was observed that olfaction plays a very important role which is reflected in their behavior – from young children to adults. In spite of the diminishing signaling role of olfaction in behavioral coordinates of modern societies where it occupies mostly hygienic and esthetic niches in comparison with its evident behavioral roles in the wild, in some populations – mostly in traditional, or archaic societies – olfactory functions are still of the primary importance. Individuals of such societies were shown to have predominantly right hemispheric types of information processing. In natural adaptation smells have well expressed signaling value. Illustrations of it were vividly described in Arshavsky’s book (2001). Deprivation and loss of the olfactory factor, characteristic to left-hemispheric individual and populational types can hardly be appraised as a positive tendency: actual alarm that can develop as a result of it most often is at the bottom of psychosomatic disturbances. However, it should be stressed that the adaptive value of olfaction is associated mostly not with ethno-cultural specificity of a certain group but with individual psychophysiological characteristics; such individuals form a predominant right-hemispheric type of mentality in a given population as a result of group selection. Studies of cerebral hemispheric patterns for sensory and cognitive functions indicate that differential processing strategies influence the perception of all kinds of stimuli. In apparent contrast to numerous lines of research in other sensory modalities, the role of hemispheric functional differences in the chemoreception, evaluation, and verbalization of odors is much less known. The right hemisphere is shown by some authors to be involved in processing odors. This was observed in brain injured patients and in normal subjects (cf. Abraham & Mathai 1983; Zucco & Tressoldi 1989; Ugolev & Chernigovskaya 1989; Chernigovskaya & Arshavsky 1994). It has been demonstrated by many investigators that there are cultural differences in the perception of odor, and it is also associated with hemispheric specificity (cf. Zatorre et al. 1992; Zucco & Tressoldi 1989).

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.5 (356-455)

Olfactory and visual processing and verbalization 

Differential characteristics of hemispheric involvement in olfactory perception have been described as a relevant feature for space orientation, especially, in the wild. Similar to binocular vision and binaural hearing, birynal perception is much more effective in speed and accuracy in comparison with monorynal. At the same time higher left-sided sensitivity was revealed in 71 % of adult subjects contrary to almost no asymmetry in children (Toller et al. 1980). Such asymmetry leads to the hypothesis that the right hemisphere should have specialized mechanisms for olfactory analysis, such as the temporal lobe which can discriminate and identify smells (Abraham & Mathai 1983). A systematic psychophysiological investigation of the hemispheric functions in representatives of different cultures has shown types of physiological activation of the brain hemispheres under different functional conditions. It is accepted that the percentage of EEG alpha-waves, the alpha-index, reflects the level of nonspecific activation caused by the brain stem reticular formation: there is a negative correlation between it and the degree of involvement of the corresponding cortical area in mental activity. However, the alpha-index is increased during successful processing of the solution of a creative task performed by a creative person (Butler & Glass 1987; Arshavsky 1988). During altered states of consciousness (meditation) alpha-index and alpha wave amplitude are also increased in comparison with the ordinary state of consciousness (Hirai 1974), although mental activity in meditation is present (Ornstein 1972). It is reasonable to compare the alpha-index and spatial synchronization of brain biopotentials (a cross-correlation analysis of the first EEG derivative) under functional loads addressed predominantly to the one or the other hemisphere. The increase in the spatial synchronization of brain biopotentials recorded from different points of the scalp reflects the contribution of definite cerebral mechanisms in the functional system which ensure the performance of the corresponding functions. Usually there are no differences in the amount of spatial synchronization between the two hemispheres in the baseline state but such differences are obvious if the subject is involved in mental activity: spatial synchronization is increased in the right or in the left hemisphere depending on the quality of the task and characterizes the functional activity of the corresponding part of the brain. To see cerebral involvement in olfactory and color perception and its association with behavioral features we used Lüscher’s color preference test and analogous free choice odor preference test (Amoore 1963). Standard degustatory templates of basic odors were also presented to subjects for 3 minutes each item with airflow velocity of 0.1 m3/ per hour during EEG monitoring. Subjects were clinically normal 73 adults and 20 professional testers. The level of innate alarm level and actual alarm level was evaluated according to Spilberger’s 40-score scale. It was shown earlier (Arshavsky & Goldstein 1994; Chernigovskaya & Arshavsky 1994) that both preference for and rejection of odors and colors are highly

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.6 (455-652)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky

Table 1. Changes in alarm level (Spilberger test) in subjects with different hemispheric types after preferable or rejected odors’ presentation Hemispheric type

n

RH LH

26 47

Baseline level Preferable IAL AAL

after presentation of odors rejected IAL AAL IAL

AAL

38.9±1.3 42.1±2.7*

39.8±1.1 42.3±1.9

46.1±0.8 52.3±2.6**

43.3±1.6 49.9±1.3*

35.9±0.8 49.6±1.7*

40.1±0.8 42.7±1.9

* P < 0.05, ** P < 0.01; RH = right hemisphere; LH = left hemisphere; IAL = innate alarm level; AAL = actual alarm level

correlated (r = 0.6–0.9): if a certain color is chosen in a certain situation a certain odor also tends to be chosen. This suggests internal associations of color and odor classifications. Color and odor rejection was associated with actual alarm level.
Black, brown and gray ( f > 0.834) were rejected together with acid, tart and
neutral ( f > 0.812) in the low alarm level condition, while in high alarm level condition any dominance of odor or color rejection is absent (f = 0.164; P < 0.001 for any stimuli). On the contrary, preferences for colors and odors depend on innate alarm

level. The choice of blue or green ( f > 0.748) or peppermint or musk ( f >
0.834) correlates with high alarm level, while of red, yellow and violet ( f > 0.762)
and floral and ethereal odors ( f > 0.603) – with low alarm level (P < 0.001). Such reactions reflect modes of behavior compensating: alarm vs. passive and avoiding vs. active. We also studied reactions to fragrances – their rejection or preferences – presented to the right and left hemispheres (RH/ LH) in normal adults, including professional tasters. Results showed that RH personalities prevailed at the cost of a decrease in LH performance (correspondingly 0.55 vs. 0.45; P < 0.05), while in the group of non-professional testers we found the opposite pattern of results (0.32 vs. 0.68; P < 0.01). This is not surprising, as testers use a Gestalt type of processing as their main tool. In our experiments subjects with the RH type of reactions (as previously evaluated by special questionnaires) showed a reliable cross-correlation of biopotentials in the RH when stimulated by odors preferable for them. Individuals of the LH type showed a correlation of biopotentials in the LH when stimulated by the odors rejected by them previously. The choice of preferable odors and colors of the Lüscher set was dependent on individual levels of anxiety. Classification and verbalization of colors showed significant differences in the types of strategies used by RH vs. LH subjects. The data suggest that most RH individuals demonstrate specific memory and verbalization of odors and that most professional testers of odors appear to be RH

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.7 (652-664)

Olfactory and visual processing and verbalization 

Table 2. Changes in choice frequency of colors and odors in subjects with different hemispheric types after preferable or rejected odors presentation. Hemispheric type

Preference Color f

odor

f

1 2

0.308 0.346

Ø 3 4 Ø

≤0.076* 0.426 0.426 ≤0.064**

Rejection color f

odor

f

Baseline RH

26

LH

47

3 4 5 Ø 1 2 Ø

0.308 0.272 0.269 ≤0.115* 0.362 0.404 ≤0.064**

Ø

≤0.154

Ø

≤0.154

Ø

≤0.115

Ø

≤0.154

0 6 7 Ø

0.269 0.269 0.385 ≤0.038**

0 6 7 Ø

0.231 0.308 0.346 ≤0.038**

Ø

≤0.115

Ø

≤0.076

Ø

≤0.154

Ø

≤0.165

Ø

≤0.149

Ø

≤0.176

In an hour after previously preferred odor presentation RH

26

LH

47

3 4 5 Ø 1 2 Ø

0.346 0.272 0.308 ≤0.076* 0.340 0.404 ≤0.076**

1 2

0.308 0.346

Ø 3 4 Ø

≤0.076 0.426 0.426 ≤0.064**

In an hour after previously rejected odor presentation RH

26

LH

47

3 4 5 Ø 1 2 Ø

0.308 0.346 0.269 ≤0.076* 0.383 0.447 ≤0.064**

1 2

0.308 0.346

Ø 3 4 Ø

≤0.076** 0.426 0.468 ≤0.043**

* P < 0.05; ** P < 0.01 Colors: 0 – grey, 1 – blue, 2 – green, 3 – red, 4 – yellow, 5 – violet, 6 – brown, 7 – black, Ø – others. Odors: 0 – none, 1 – floral, 2 – etheral, 3 – muscus, 4 –peppermint, 5 – camphora, 6 – acid, 7 – saprogenic, Ø– others.

personalities. They also show the important role of social and cultural, as well as of linguistic, backgrounds. Right hemispheric visual, auditory and olfactory processing seems to correlate with certain behavioral characteristics reflecting successful adaptation and resistance to stress and psychic or somato-psychic diseases, and adaptive behavior in general. The correlation analysis of the EEG first derivative in RH and LH persons showed different patterns of space synchronization of biopotentials, with preferable vs. rejected odors associated with alarm level (Tables 1 and 2, Figures 1–4).

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.8 (664-680)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky I

II

III

IV

0.8 0.7

right hem.

0.6

left hem.

0.5 0.4 0.3

Right-hemispheric personality, n = 26 I

II

III

IV

0.8 0.7

right hem.

0.6

left hem.

0.5 0.4 0.3

Left-hemispheric personality, n = 47 I – baseline, II – preferable smells, III – rejected smells, IV – post-stimulus

Figures 1 and 2. Coefficient correlation change of the EEG first derivative in individuals of different hemispheric types subjected to preferable or rejected olfactory stimuli.

Individuals of RH type use Gestalt processing of all the items presented with picking up the preferable; the rejected in this case will be the last in the row. LH persons on the contrary use a step by step rejecting strategy, so that the preferred odor is left to be the last in the row. Professional testers have lower latency thresholds of odor recognition and appreciation. In another test normal adults coming from different languages and cultures and tested for lateralities and cognitive styles were accessed for voluntary free associations concerning individual memory for odors. Associations were later eval-

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.9 (680-695)

Olfactory and visual processing and verbalization 

Nonprofessionals (n=17)

0.8 0.7 0.6 0.5

1st min

3rd min

Professional testers (n=9)

0.8

1st min

3rd min

0.7

right hem. left hem.

0.6

right hem.

0.5

left hem.

0.4

0.4

0.3

0.3

Right-hemispheric personality, preferable smells

Nonprofessionals (n=36)

0.8 0.7 0.6 0.5

1st min

3rd min

right hem. left hem.

Professional testers (n=11)

0.8

1st min

3rd min

0.7 0.6

right hem.

0.5

left hem.

0.4

0.4

0.3

0.3

Left-hemispheric personality, rejected smells

Figures 3 and 4. Time thresholds in strong correlation of hemispheric bio-potentials in nonprofessionals and professional testers of different hemispheric types subjected to preferable or rejected olfactory stimuli.

uated by the subjects as neutral, negative or positive and classified according to different semantic fields. Similar to Gilbert and Wysocki (1987), our data suggest a high level of universality in the semantic organization of the olfactory semiosphere: the major classification clusters were similar in different groups. At the same time it is evident that cultural specifics and social constraints play a very important role (Ugolev & Chernigovskaya 1989).

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.10 (695-770)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky

Conclusion Our data suggest that there is a neurological basis for olfactory preferences governing human behavior, with the right hemisphere playing a very important role. It is also evident that right-hemispheric populations and individuals tend to specific type of culture in which olfaction is one of the central parts of the semiosphere. As Wildgen (this volume) discusses it in this volume, perception governed by neural architecture (see also Fahle in this volume) and communication based on social structure form the transition between sensibility and sense. Language appears to be the vehicle to label cortical representations of input and to normalize subjective experiences, and thus subserves not only communication but reflection as well and thus helps in orientation in the world and in adaptation to it. This means that language, being a cultural phenomenon though based on genetically developed algorithms, relates natural objects to neurophysiological events via conventional semiotic mechanisms. Our perception could be described in vague terms of objectiveness only because we have an agreement in naming, i.e., ‘boxes’ in which to pack the sensations.

References Abraham, F., & Mathai, K. (1983). The effect of right temporal lobe lesions on matching of smells. Neuropsychologia, 21 (3), 227–281. Amoore, J. (1963). The principle of odor classifications. Nature, 198, 271–275. Arshavsky, V. V. (1988). Interhemispheric Asymmetry in Search Activity System. Vladivostok: Academy of Science (in Russian). Arshavsky, V. V. (2001). Differences that Unite Us. Essays in Populational Mechanisms of Brain Interhemispheric Asymmetry. Riga (in Russian). Arshavsky, V. V., & Goldstein, N. I. (1994). EEG space pattern and changes in anxiety level under olfactory stimulation. Human Physiology, 20 (1), 27–36. Butler, S., & Glass, A. (1987). Individual differences in the asymmetry of alpha activation. In A. Glass (Ed.), Individual Differences in Hemispheric Specialization (pp. 103–120). New York: Plenum Press. Chernigovskaya, T. V. (1993). Die Heterogenität des verbalen Denkens als cerebrale Asymmetrie. In P. Grzybek (Ed.), Psychosemiotik-Neurosemioti (pp. 37–54). Bochum: Dr. N. Brockmeyer. Chernigovskaya, T. V. (1994). Cerebral lateralization for cognitive and linguistic abilities: Neuropsychological and cultural aspects. In J. Wind & A. Jonker (Eds.), Studies in Language Origins (pp. 56–76). Amsterdam, Philadelphia: v. III. Chernigovskaya T. V. (1995). Iconicity in visual and olfactory processing: Perception, memory, verbalization. Annual Meeting of the Language Origins Society, 11. Pecs/ Hungary: Janus Pannonius University. Chernigovskaya T. V. (1999). Neurosemiotic approach to cognitive functions. Journal of the International Association for Semiotic Studies – Semiotica, 127 (1/4), 227–237.

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.11 (770-882)

Olfactory and visual processing and verbalization 

Chernigovskaya, T. V., & Arshavsky V. V. (1994). Hemispheric asymmetry in olfactory processing. Neurophysiological and cognitive aspects. 23th Meeting of International Neuropsychological Society, 38. Angers/ France. Danthiir, V., Roberts, R., Pallier G., & Stankov L. (2001). What the nose knows. Olfaction and cognitive abilities. Intelligence, 29, 337–361. Dubois, D. (2000). Categories as acts of meaning: The case of categories in olfaction and audition. Cognitive Science Quarterly, 1, 35–68. Economides, S. (1986). Integration centrale de l’information olfactive chez l’homme. Bull. Speca, 1 (2), 178–182. Engen T. (1982). The Perception of Odors. New York: Academic Press. Engen T. (1991). Odor Sensation and Memory. New York: Praeger. Gilbert A. N., & Wysocki, C. J. (1987). The national geographic smell survey: Results. National Geographic Magazine, 172, 514–525. Hanisch, E. (1982). The calming effect of fragrances and associated remembrances. The Nose. Drom, 2 36–38. Harper R., Bate-Smith, E. C., & Land, D. C. (1968). Odor Description and Odor Classification. London: Churchill. Hirai, T. (1974). Psychophysiology of Zen. Tokyo: Igako Shoin. Hoffmeyer, J. (1996). Signs of Meaning in the Universe. Bloomington: Indiana University Press. Hoppe, K. D. (1977). Split-brain and psychoanalysis. Psychoanalytic Quarterly, 46, 220–248. Kupertman, V. (2003). Olfaction as an instrument of prophecy in psychotic delusions. (Unpublished manuscript) Kuperman, V., & Zislin, J. (2003). Olfaction as a visioner’s tool. In O. Weinstein (Ed.), Aromas and Odors in Culture. NLO, 1, 206–217 (in Russian). Minor, A. V., & Krutova, V.I. (2001). Discrimination of individual odors and models of sensory coding in olfactory systems of mammals. Symposium ‘Olfactory Coding’, Krakov – Satellite to the VIth International Congress in Neuroethology, 53. Berlin: Ibro. Ornstein, R. (1972). The Psychology of Consciousness. San Francisco: Freeman Co. Rotenberg, V., & Arshavsky, V. V. (1997). Right and left brain hemispheres activation in the representatives of two different cultures. Homeostasis, 38 (2), 49–57. Sacks O. (1987). The Man Who Mistook His Wife for a Hat (and Other Clinical Tales). New York et al.: Perennial Library. Schaal, B., Rouby, C., Marlier, L., Kontar, F., & Tremblay, R. E. (1998). Variabilité et universaux au sein de l’espace perçu des odeurs. Approches interculturelles de l’hédonisme olfactif. In Géographie des Odeurs, sous la dir. de R. Dulau et J.-R. Pitte (pp. 25–47). Montréal: L’Harmattan. Staubli, U. (1987). Olfaction and the ‘Data’ memory system. Behavioral Neuroscience, 101, 757– 765. Steiner, W., & Neumann, P. (1982). Fragrances as stimulators. The Nose. Drom., 2, 12–28. Toller, S., Hendal-Reed, M., & Sleight E. (1980). Hemispheric processing of odours. Biological Psychiatry, 11 (3–4), 262–274. Ugolev, D. A., & Chernigovskaya, T. V. (1989). Semiochemistry and human behavior (sensory, psycholinguistic and instrumental aspects of the formally defined problem). Proceedings of the Xth International Symposium on Olfaction and Taste, 17. Oslo. Weinstein, O. (2003). Aromas and Odors in Culture. NLO, Vol. 1–2. Moscow (in Russian). Whitton, T. (1978). EEG frequency patterns associated with hallucinations in schizophrenics and ‘creativity’ in normals. Biological Psychiatry, 13, 123–133.

JB[v.20020404] Prn:26/06/2007; 10:20

F: CELCR813.tex / p.12 (882-899)

 Tatiana V. Chernigovskaya and Victor V. Arshavsky

Zatorre, R., Jones-Gotman, M., Evans, A., & Meyer, E. (1992). Functional localization and lateralization of human olfactory cortex. Nature, 360, 339–340. Zellner, D., & Kautz, A. (1990). Color effects perceived odor intensity. Journal of Experimental Psychology of Human Perception and Performance, 2, 391–397. Zucco, G. M. (2003). Anomalies in Cognition: Olfactory memory. European Psychologist, 3, 77– 86. Zucco, G., & Tressoldi, P. (1989). Hemispheric differences in odour recognition. Cortex, 25, 607– 615.

JB[v.20020404] Prn:29/01/2007; 12:01

F: CELCR8CN.tex / p.1 (45-222)

Contributors

Viktor V. Arshavsky Balitic Russian Institute School of Psychology Riga, Latvia

Susanne Niemeier University of Koblenz-Landau Institute of English and Romance Studies Koblenz, Germany

Tatiana V. Chernigovskaya St. Petersburg State University Department of General Linguistics St. Petersburg, Russia

Hajime Nozawa National Institute of Information and Communications Technology Knowledge Creation Research Center Kyoto, Japan

Danièle Dubois Centre National de Recherches Scientifique (CNRS) Laboratoire CNRS de linguistique Paris, France Manfred Fahle University of Bremen Department of Human Neurobiology Bremen, Germany Andrea Graumann University of Oldenburg Institute of German Studies Oldenburg, Germany Volker Heeschen University of Munich Institute of Ethnology and African Studies Munich, Germany Peter Holz University of Bremen Department of Literature and Language Bremen, Germany Toshiyuki Kanamaru Graduate School of Human and Enviromental Studies University of Kyoto Kyoto, Japan

Martina Plümacher University of Technology Institute of Philosophy, Philosophie of Science, History of Science and Technology Berlin, Germany Yoshikata Shibuya National Institute of Information and Communications Technology Knowledge Creation Research Center Kyoto, Japan Wolfgang Wildgen University of Bremen Department of Literature and Language Bremen, Germany Siegfried Wyler University of St. Gallen Department of Cultural Studies St. Gallen, Switzerland Gesualdo M. Zucco University of Padua (I) Department of General Psychology and University S. Raffaele, Milan (I) Department of Cognitive Sciences Padua and Milan, Italy

JB[v.20020404] Prn:26/06/2007; 10:53

F: CELCR8IN.tex / p.1 (42-153)

Index

A Abraham , ,  Adams  advertising / advertisement , , , , , , , , , , , , , , , , , , , –, – after-image , ,  Albers , , , ,  Albert  Allen  Allport , ,  Amoore  Aristotle , , ,  Arnheim  Arshavsky , , , – Atran ,  Austin 

Berlin , , , , , , , , , , , , , , ,  Bissett  Boers  Boisson  Bowden , , , ,  Britten  Brock  Bromley  Brown , , , , –,  Bruggencate  Bruner ,  Bruno , ,  Buffon  Burdach , ,  Butler  Butters 

B Bühler  Bach ,  Baddeley  Ballmer  Barcelona  basic color ,  basic color categories  basic color term / name , , , , , , , , , , , , –, , , –, , , ,  basic hue  basic meaning , , –, ,  basic odor term , , ,  basic odors  basic term / name , , , , , , ,  Bear ,  Beaumont  Bennett 

C Cain – Candau ,  Cangelosi  Cassirer  Casson  categorization –, , , , , , , , , , , , , , , –, ,  cerebral hemisphere , , , , , , , –,  Chastrette ,  Cheminée  Chernigovskaya , , , , –,  Chomsky ,  Classen  classify / classification –, , , , , , , , , –, , , , , , , , , 

Cognitive Linguistics , , , , , , , , , ,  cognitive linguistics , , , ,  color constancy , , ,  complementarity , , ,  complementary colors , , , , ,  conceptualization , , , , ,  Condillac , – conscious / consciousness , , , , , , , , , , –, , ,  Cosnier  Croft  cross-cultural  culture / cultural –, –, , , , , , , , , , , , , , , , , , , , , –, , –, –, , , , , , , , , – Cytowic 

D Danthiir ,  David , , , – Davidoff  Dedrick ,  denotation / denotational , , , , , , , , ,  Descola  Detienne  Dirven ,  discrimination , ,  Distel 

JB[v.20020404] Prn:26/06/2007; 10:53

F: CELCR8IN.tex / p.2 (153-277)

 Index distinguish , , , , , , , , , , , , , , , ,  distinguishing capacity of color names  distinguishing marks / marker , ,  Dittmann – Donmall  Doty ,  Douglas  Dubois , , , , , –, , , , –, , , , ,  E Ebbinghaus  Economides  Eibl-Eibesfeldt – Eimas  emotion / emotional , , , , , , , , , , , , , , –, , , , , , , , , , , , , , , –, ,  Emrich ,  Engen –, , –, ,  Evans  evolution , , , , , –, , , , , , ,  Eysel 

G Gabassi  Gage , –, , , ,  Gazzaniga  Gell ,  Gentner  Gerritsen  Gestalt / Gestalt perception , , , , ,  Gestalt law , ,  Gestalt psychology , , ,  gestalt, semantic ,  Gilbert  Gill  Gipper ,  Glass  Goddard ,  Godinot  Goethe , ,  Goldin-Meadow  Goldman  Goldstein , ,  Grüsser-Cornehls  grammar / grammatical , , , , , , , , , ,  Graumann , , , , , , , , ,  Greco  Grinevald , ,  Gschwind , ,  Guastavino 

F Fabre-Vassas  Fahle , , , , , , , , , , ,  Finger  Fischer ,  Flierl  Fludd ,  focal color , , –, , , ,  focal color term / name , ,  Fodor  Forge  Forsius  Friedl  Frieling  Frisch  Fuhs 

H Hölscher  Haiman  Haller  hallucination , , , , , , ,  Hanisch ,  Harkins  Harley  Harnad  Harper ,  Harris  Harrison , ,  Haynes , ,  Healey ,  hedonistic classificatory principle , , , , –

Heeschen –, , , , , , , , , , ,  Heller ,  Henning  Herder  Herz –,  Hess  Hickethier  Hirai  Hockett  Hoffmeyer  Holley , ,  Holz , , , , , , , , , , , ,  Hoppe  Howes ,  Hudson  Humboldt ,  I iconic / iconicity , , , ,  identification , , , , , , , , –,  illusion , , , , , , , –,  image / imagination , , , , , , , , , , , , , , , , , , , , –, , , , ,  imagery  Imai  indexical , ,  intercultural , , , , , ,  intermodality  Itten , , –, , ,  J Jäger  Jackson  Jakobson , , – Jehl  Johnson  Jones  Jones-Gotman  K Küppers ,  Küthe  Kanamaru , , 

JB[v.20020404] Prn:26/06/2007; 10:53

F: CELCR8IN.tex / p.3 (277-390)

Index  Kandinsky , ,  Katz  Kautz  Kay , , , , , , , , , , , , , , ,  Kirchner  Klee , , ,  Kleine-Besten  knowledge , –, , , , , , , , , , , , , , , , , –, , , , –, , , , , , , , , , ,  Koch  Kramsch  Krashen  Kroeber-Riel  Krutova  Kuehni  Kuisma  Kuperman  Kupertman  Kuschel , 

L Lüscher , , ,  Lakoff , , , ,  Landau  Langacker , ,  Larsson  Lashley  Latour  Lawless – Lazar-Meyn  Lehmann , ,  Lenneberg , , , – Levi-Strauss  lexical , , , , , , , , , , , –, , , , , , , , , , , , , ,  lexicalize / lexicalization , , , , ,  lexicon , , , , , , , , , , ,  limbic system , , , , ,  linguistic relativity ,  Linnaeus  Locke 

Lorig  Lorry  Loving  Lowie  Lucy –, , , – Lullus , ,  Lyman , 

M MacLaury ,  Maihofer  Mathai , ,  Mazet  McCawley  McDaniel ,  McNeill  Medin ,  Mella  memory , –, , –, , , , , , –, –, , , , , , , ,  metaphor / metaphorical , , , , , , , , –, , , , , , , , , , , ,  metonymy , , –, –,  Metzger  Meyer  Minor  Minsky  modify / modification of odor / color terms , , , , , , , , , , , , , ,  modified term / name , , , –, , , , , , –, ,  Monberg ,  Morford  Moskowitz  Mouélé ,  multimodality  Munch ,  Munsell , , , ,  Musset 

N Neumann ,  Newsome  Newton , 

Nicholson  Nickerson  Niemeier , , , , , , , , , , ,  Nozawa , , ,  Nussdorf  O Ornstein – Ostwald  P Paramei  Parlier ,  Peirce , , ,  Penfield  perception constancy  Piaget  Pinkal  Plümacher , , , , , , , , , , , ,  poetic description , , , , , –, , – Poitou  Pokorny ,  Porphyry  prototype / prototypical , , , , , , , , , –, , , , , , , , ,  psycho-physiological effect , , ,  psychological effect , , –,  Purkinje  R Rabin ,  radicalized color name ,  Ramsey  Rastier  recognition , , , –, ,  reference / referential , , , , –, , , , , , –, , , –, , , , , , , , , , , –, , , , , , , , –, , , , , , , , , 

JB[v.20020404] Prn:26/06/2007; 10:53

F: CELCR8IN.tex / p.4 (390-507)

 Index relational order , , , –, , , ,  Rembrandt  representation / represenational –, –, , , , , , , , , , –, , , , , , , , , , , , , , , , , , , –, –, , , , , , , ,  Resche-Rigon ,  Richardson , ,  Roberts  Rolls  Rompza  Rosch , , , ,  Rosenthal  Ross , – Rotenberg  Rouby , – Royet  Runge  S Sacks  Saint-Martin ,  Salzman  Sapir  Schaal  Schab , ,  Schiefenhövel  Schiffman  Schleidt ,  Schmitter  Schweppe ,  Scott  Seamon  semiosis , , ,  semiotics / semiotic , , , , , , , , , , , , , , , , , , , , ,  Shen  Shepard  Shevell ,  Shibuya , , , , , , , 

Sicard – Silvestrini  simultaneous contrast , ,  Sloane  Smith , ,  Spang ,  Spilberger ,  Stadler , ,  Staubli  Steiner ,  Stiles  Strathern , ,  Straus  Strauss  subconscious , , , – suggestive description , , ,  symbol / symbolic , , , –, , , , , , , , , –, , , ,  symbolize  synesthesia / synesthetic , , , , , , , , , , –, , –, , , , –, –,  T Taylor  Tenin ,  Thompson  Timm ,  Toller  Trabant  Tressoldi , ,  Turner  U Ugolev , ,  unconscious , , , , , , ,  Underwood  universal meaning , 

V Varela  Venn  verbalization , , , , ,  Verspoor , ,  Victor  vocabulary , , , , , , , , ,  Vroon , ,  W Wassmann  Waxman ,  Weimer ,  Weinberg  Weinstein  Weisgerber  Weskamp  Whitton  Whorf  Wildgen , , , , , , , , ,  Williams  Wittgenstein  Wood  Wundt  Wyler , , , , , , , , , , , , , , , , ,  Wysocki  Wyszecki  Y Yamanashi  Young  Z Zanuttini  Zatorre ,  Zellner  Zimmer ,  Zislin  Zucco , , , , , , , , , , , , , ,  Zwaardemaker 

In the series Converging Evidence in Language and Communication Research the following titles have been published thus far or are scheduled for publication: 9 8 7 6 5 4 3 2 1

Lascaratou, Chryssoula: The Language of Pain. Expression or description. x, 232 pp. + index. Expected October 2007 Plümacher, Martina and Peter Holz (eds.): Speaking of Colors and Odors. 2007. vi, 244 pp. Sharifian, Farzad and Gary B. Palmer (eds.): Applied Cultural Linguistics. Implications for second language learning and intercultural communication. 2007. xiv, 170 pp. Deignan, Alice: Metaphor and Corpus Linguistics. 2005. x, 236 pp. Johansson, Sverker: Origins of Language. Constraints on hypotheses. 2005. xii, 346 pp. Kertész, András: Cognitive Semantics and Scientific Knowledge. Case studies in the cognitive science of science. 2004. viii, 261 pp. Louwerse, Max and Willie van Peer (eds.): Thematics. Interdisciplinary Studies. 2002. x, 448 pp. Albertazzi, Liliana (ed.): Meaning and Cognition. A multidisciplinary approach. 2000. vi, 270 pp. Horie, Kaoru (ed.): Complementation. Cognitive and functional perspectives. 2000. vi, 242 pp.