231 90 6MB
English Pages 239 [240] Year 2019
Bettina Bock von Wülfingen (Ed.)
SCIENCE IN COLOR Visualizing Achromatic Knowledge
Bettina Bock von Wülfingen (Ed.)
SCIENCE IN COLOR Visualizing Achromatic Knowledge
This publication was made possible by Image Knowledge Gestaltung. An Interdisciplinary Laboratory Cluster of Excellence at the Humboldt-Universität zu Berlin (sponsor number EXC 1027/1) with financial support from the German Research Foundation as a part of the Excellence Initiative.
Copy-editing Rainer Hörmann, Jim Baker Typesetting and design Andreas Eberlein, Berlin Printing and binding Beltz Grafische Betriebe GmbH, Bad Langensalza ISBN 978-3-11-060468-9 e-ISBN (PDF) 978-3-11-060521-1 Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar. © 2019 Walter de Gruyter GmbH Berlin/Boston www.degruyter.com This publication, including all parts thereof, is legally protected by copyright. Any use, exploitation or commercialization outside the narrow limits set by copyright legislation, without the publisher’s consent, is illegal and liable to p rosecution. This applies in particular to photostat reproduction, copying, scanning or duplication of any kind, translation, preparation of microfilms, electronic data processing, and storage such as making this publication available on Internet.
TABLE OF CONTENTS
9 Editorial
COLOR AND ITS MEANING FOR THE SCIENCES 19 Aldo Badano Color in Medical Images 33 Ulrike Boskamp Color as the Other? Absence and Reappearance of Chromophobia in Eighteenth-Century France 51 Alexander Nagel Research on Color Matters: Towards a Modern Archaeology of Ancient Polychromies 65 Esther Ramharter Do Signs Make Logic Colored? Tendencies Around 1900 and Earlier 81 Michael Friedman Coloring the Fourth Dimension? Coloring Polytopes and Complex Curves at the End of the Nineteenth Century 99 Ricardo Cedeño Montaña Encoding Color: Between Perception and Signal
MEANINGFUL COLORS IN THE SCIENCES 117 Michael Rossi Green Is Refreshing: Techniques, Technologies and Epistemologies of Nineteenth-Century Color Therapies 133 Ian Lawson Pigments, Natural History and Primary Qualities: How Orange Became a Color 147 Daniel Baum An Evaluation of Color Maps for Visual Data Exploration 163 Jana Moser, Philipp Meyer The Use of Color in Geographic Maps 181 Jean-François Moreau, Raffaele Pisano, Jean-Michel Correas Historical and Scientific Note of Color Duplex Doppler Ultrasound and Imaging 195 Bettina Bock von Wülfingen Diagrammatic Traditions: Color in Metabolic Maps 219 Dominique Grisard Pink and Blue Science. A Gender History of Color in Psychology
237 Image Credits 239 Authors
Editorial Why color? Three events will illustrate what motivated this volume: At a conference on the Meaning of Color in Berlin in July 2018, a researcher talked about her ethnographic field work in geobiology during a trip to a tropic island. What might have sounded like a fancy holiday destination was in fact a hostile environment to most living things, seeing that it provided no fresh water or shelter from sun. She was there to observe and interact with geobiologists who study the microbial communities that build dense layers in swamps and exist in only a few regions of the world. She watched the geobiologists measure the density of layers and record colors. They also took photographs as documentary evidence. The colors of the layers were key to age and composition. During the conference on the Meaning of Color in Berlin, someone in the audience asked, “So what kind of Munsell-like color scale do they [the geobiologists] use to give the colors their correct and specific names?” To which the ethnographer replied, “There is no such thing. That is tacit knowledge.” Nobody said a word. The Finnish expert on the color green, who had asked the question, looked incredulous. The digital media researcher next to him raised his hand: “So with what kind of standardized camera technology do they take their photos?” The ethnographer: “Um, with their own smartphones, mostly.” The second example concerns diagrams. In an interview, the Harvard professor of Biochemistry Alain Viel was asked about his textbook in the making: “How do you decide which kind of blue to choose for these acid-endings in your diagram?” Viel: “Oh, I like this blue!“.1 he answered pointing at the blue arrows in a diagram on his monitor. Similarly, in an interview, biochemist Gerhard Michal, author of the most widespread biochemical map Biochemical Pathways, first produced in 1965, told me about the choice of colors in his diagram: “I asked the printer to show me the blues he most liked and of these we picked the nicest blue.”2 The third case is from the realm of medicine. Aldo Badano, one of the authors in this volume, is a researcher at the Food and Drug Administration in the United States involved in evaluating the safety of monitors used in clinical contexts. At a conference on color in Berlin in 2016, he presented one of the first studies to investigate the strongly held belief within the medical field that color in images 1 Alain Viel: Interview, transcript: Cambridge, PA, 2018. 2 Gerhard Michal: Interview, transcript: Munich, 2018.
10
Editorial
makes them easier to read. There had been no study thus far to verify the diagnostic performance of medical practitioners including cardiologists and radiologists in relation to the color scale of the monitor used. The reason why, in the first case, the expert on green and the expert on digital visual technology were struck by the procedure used during the expedition in geobiology was the absence of color standardization. On the one hand, we know from the history of science that starting in the seventeenth and extending well into the twentieth century early scientists, particularly in fields such as botany, zoology, geology and meteorology, used color cards for standardization, which helped to measure, transfer and communicate color impressions. Archaeologist Alexander Nagel, also an author in this volume, propagates the use of Munsell’s color charts (developed at the beginning of the twentieth century by the painter and art professor),3 often used to describe soil in geology and archaeology, to describe paintings on antique artifacts. On the other hand, different digital (and analog) camera technologies and screens capture, store and display colors differently. They achieve different results as regards to hue, saturation and value or lightness – terms which computer graphics researchers since the 1970s have used to represent perceptual color models such as Munsell’s.4 The diagrams mentioned in the second example were used in the biochemistry textbooks and wall charts sold until the end of the twentieth century to teachers and students in disciplines ranging from chemistry to medicine. Today, they have been replaced by online versions with similar symbolic and aesthetic features. We learn from the literature that their use in an educational context raises perceptual, pedagogic and psychological questions.5 What we do not learn from the literature 3 Albert H. Munsell: Atlas of the Munsell Color System, Boston: Wadsworth, Howland & Co., Inc, 1915; Albert H. Munsell: A Color Notation, Boston: G. H. Ellis Co, 1905. 4 Edward R. Landa, Mark D. Fairchild: Charting Color from the Eye of the Beholder. In: American Scientist, 93 (5), 2005, pp. 436–443; Steven K. Shevell: The Science of Color, Oxford: Elsevier Science & Technology, 2003; Michael W. Schwarz, William B. Cowan, John C. Beatty: An Experimental Comparison of RGB, YIQ, LAB, HSV, and Opponent Color Models. In: ACM Transactions on Graphics, 6 (2), 1987, pp. 123–158; John Kender: Saturation, Hue and Normalized Color: Calculation, Digitization Effects, and Use, Pittsburgh: Carnegie Mellon University, Computer Science Department, 1976. 5 Marissa Harle, Marcy Towns: A Review of Spatial Ability Literature, Its Connection to Chemistry, and Implications for Instruction. In: Journal of Chemical Education, 88 (3), 2010, pp. 351–360; David Kaiser: Drawing Theories Apart: The Dispersion of Feynman Diagrams in Postwar Physics, Chicago: University of Chicago Press, 2009; Michelle Patrick Cook: Visual Representations in Science Education: The Influence of Prior Knowledge and Cognitive Load Theory on Instructional Design Principles. In: Science Education, 90 (6), 2006, pp. 1073–1091; Hsin-Kai Wu, Priti Shah: Exploring Visuospatial Thinking in Chemistry Learning. In: Science Education, 88 (3), 2004, pp. 465–492; Jorge Trindade, Carlos Fiolhais, Leandro Almeida: Science Learning in Virtual Environments: A Descriptive Study. In: British Journal of Educational Technology, 33 (4), 2002, pp. 471–488; James H. Mathewson: Visual-Spatial Thinking: An Aspect of Science Overlooked by Educators. In: Science Education, 83 (1), 1999, pp. 33–54.
Editorial
11
is information on the color choices in these diagrams, especially regarding intercultural contexts. This also concerns possible intercultural differences in reading the different colors in different maps. In the last case, the results of the 2015 study by Badano et al. on monochrome images on monitors (which are in grayscale mode) versus polychrome monitors in the medical context showed that people trained on monochrome monitors had difficulty interpreting polychrome images and vice versa. This, they were able to show, influences performance in the detection and demarcation of lesions by the study participants. Yet, there was no knowledge of the relevance of color for the physicians’ performance in the respective institutions. To conclude, on the one hand, we have reasons to believe that most natural scientists encounter color and need to employ color in their empirical, technical and educational work and would benefit from systematic studies of the (history) of the use of color in images. As detailed below, on the other hand, the status of research in the social sciences relating to science and technology, as well as in history and philosophy of science and technology, seems to reflect the generally low awareness of the relevance of color in the sciences and technology. Where theoretical reflections on scientific images have been undertaken, the history of science has been the main field producing individual studies on color. We can divide these studies into the history of the ontology of color, which is most often the history of physics,6 studies on the history of color charts for standardization in the sciences and technology,7 studies that analyze the history of color as material substance,8 and those that relate mimetic color use to reproduce the living aspect of zoological or botanical objects,9 as well as the mimetic use of color in other disciplines such as
6 E. g. Klaus Hentschel: Verengte Sichtweise. Folgen der Newtonschen Optik für die Farbwahrnehmung bis ins 19. Jahrhundert. In: Farbstrategien. Bildwelten des Wissens, 4.1, Berlin, Boston: de Gruyter, 2006; and in M. Bushart, Friedrich Steinle: Colour Histories, Science, Art, and Technology in the 17th and 18th Centuries, Berlin, Boston: de Gruyter, 2015; T. Baker, S. Dupré, S. Kusukawa, K. Leonhard: Early Modern Color Worlds. In: Early Science and Medicine, 20 (4-6), 2015, pp. 289–591. 7 See e. g. André Karliczek: Zur Herausbildung von Farbstandards in den frühen Wissenschaften. In: Ferrum, 90 (Nachrichtenblatt der Eisenbibliothek), 2018, pp. 36–49; A. Temkin, B. Fer, M. Ho: Color Chart: Reinventing Color, 1950 to Today, The Museum of Modern Art, 2008; Rolf Kuehni, Andreas Schwarz: Color ordered: a survey of color systems from antiquity to the present. Oxford: Oxford University Press, 2008. 8 See e. g. Jan Altmann: Färbung, Farbgestaltung und früher Farbdruck am Ende der Naturgeschichte. In: Farbstrategien (s. fn. 6), pp. 69–77; Alexandre Métraux: Farbstoffchemie, Farbexperimente und die französische Malerei. In: Farbstrategien (s. fn. 6), pp. 61–68. 9 As in Kärin Nickelsen: The Challenge of Colour: Eighteenth-Century Botanists and the Hand-colouring of Illustrations. In: Annals of Science, 63 (1), 2006, pp. 3–23.
12
Editorial
geology, meteorology or medicine.10 Appreciating all these groundbreaking works on color in science, authors in this volume deal with the meaning of color: what does it mean in specific fields of medicine, philosophy or to a specific scientific discipline to use or not to use color? And to focus on the other meaning of meaning: what do specific colors mean in specific (scientific) contexts? And how do they come to acquire these meanings? How the book is structured The volume is divided in two parts in accordance with the two meanings of meaning of color: The first half of the articles deals with “meaning” in the sense of relevance or irrelevance, with the simultaneous use and neglect of color in the sciences, as something generally disapproved or something useful – as an object to reflect on in the sciences and medicine themselves. Some of them take David Batchelor’s claim of a “Western chromophobia”11 from antiquity to modernity as a point of departure, mainly to contradict the coarseness of this claim and to show the complex contradictions of color use and neglect in different countries, times and contexts. Still, reading all the contributions together, it did seem in our discussions that in specific historic contexts and moments, in times of backlash after phases of more emancipatory democratic social change, chromophobia became part of racialized and gendered discourses that re-aligned politics, sciences and arts. This is at least what we suggest as a field of research that requires further study. This first group of articles begins with the contribution by Aldo Badano, whose article shows that not to reflect on the use of color in the sciences and medicine may become in practice, dramatically enough, a question of life and death. Ulrike Boskamp, Alexander Nagel and Esther Ramharter follow with problematizations and contextualizations of Batchelor’s claim. They distinguish different historic moments, disciplines and uses of color. Michael Friedman demonstrates a field of chromophilia: the use of color on mathematical models. We close the first half with Ricardo Cedeño Montaña’s study, which focuses on the history of studies on the visual apparatus and color perception. It is the bridge to the second group of 10 See e. g. in Baker et al. (s. fn. 6), pp. 289–591; Klaus Hentschel: Visual Cultures in Science and Technology: A Comparative History, Oxford: Oxford University Press, 2014; Farbstrategien (s. fn. 6) ; on different roles of colour in relation to form in the scientific image see Horst Bredekamp, Vera Dünkel, Birgit Schneider (eds.): The technical image: a history of styles in scientific imagery. Chicago: University of Chicago Press, 2019.. 11 David Batchelor: Chromophobia, London: Reaktion books, 2000.
Editorial
13
articles which focus on the meaning of color in the sense of color being a symbol of something else, and on specific colors and their connotations – i. e., their respective cultural meaning or symbolism. Michael Rossi makes a start with the history of knowledge generation on the perception of green. Ian Lawson shows how the concept of primary colors with some delay led to the acceptance of orange in the color terminology in Europe. His contribution is followed by four articles which all study the meaning of color in various sorts of maps: Daniel Baum on color maps as such; Jana Moser and Philipp Meyer on geographic maps; Jean-François Moreau, Jean-Michel Correas and Raffaele Pisano on the medical imaging technique in Color Doppler Ultrasound (which are topographical maps, too), and finally, Bettina Bock von Wülfingen on metabolic maps. Dominique Grisard closes with the history of psychological studies, which result in naturalizing the gendered conception of pink. Shared problems The idea of claiming specific meanings and connotations of color as natural and universal is a theme that runs through all the texts. They demonstrate the opposite: connotations of color change synchronically and diachronically; they are bound into the contexts they appear in. “The meaning of color” in almost all the contributions signifies cultural connotation but often also primarily refers to questions about values – social, cultural, sometimes supposed technical values that turn out to be cultural values shaping color use (e. g., in Ricardo Cedeño Montaña’s case, there are contingent reasons why color use is understood as advantageous in some cases or nonsense in others). Standardization and the lack thereof and the historic moment in which it appeared (if it indeed did so) as well as the timescale in the shift of values reappear as problems in most of the contributions. The articles in this volume deal with a broad range of historic and modern color uses and types. While those articles dealing with late modern forms of visualizations (mostly on screens) work with the opposition of true versus pseudo color, for others the distinction of mimetic and symbolic color use is in the foreground. This volume makes use of the existing rich studies on the history of color in science and tries to pave the way for more studies on the meaning of color. As previously stated, further work on historically contingent moments of chromophobia and chromophilia would be worthwhile. Another suggestion for further research resulting from our discussion is global history and transculturalism in studies on color use. Almost all articles in our compilation analyze the uses of color in scientific contexts of – to put it bluntly – the global North-West. Alexander Nagel’s study is on
14
Editorial
descriptions of color use in architecture in the Near East but focuses on its reception by German and English scientists; Bettina Bock von Wülfingen’s study compares German and Japanese metabolic maps – still, we are far from provincializing Europe and the United States. Nevertheless, this volume’s subject raises questions and issues that can be investigated in and across other contexts, and presents a first harvest of results concerning the way scientific cultures use, adopt and think about (or rather choose to ignore) colors. Bettina Bock von Wülfingen
COLOR AND ITS MEANING FOR THE SCIENCES
Aldo Badano
Color in Medical Images Despite its widespread use in the interpretation of medical images, color is handled primarily in an ad-hoc manner due to the lack of standard approaches for optimal visualization. The variability introduced by varying color treatments leads to reproducibility concerns in quantitative image evaluation and low inter-observer agreement possibly leading to inconsistent diagnostic decisions with a negative impact on patient treatment and prognosis. Medical imaging techniques that rely on pseudo-color presentation include perfusion techniques, diffusion-weighted magnetic resonance and nuclear imaging. Other modalities use absolute color transfer including medical photography and the emerging field of digital whole-slide imaging and digital pathology. Some have suggested that specific color scales for non-contrast computed tomography of arterial function improve diagnostic confidence, diagnostic accuracy and inter-observer agreement with respect to a grayscale presentation. Moreover, recent research using synthetic and patient images in laboratory and clinical settings indicates that benefits from using color scales is modality dependent and is affected by reader training and by variations in the training and interpretation practices across geographical regions and across schools and health institutions. What is the reason color is used in medical imaging? Is it to improve the performance of the clinician or rather to visualize data that would otherwise be imperceptible? In either scenario, one can also ask if it is at all possible to define meaning within the context of a medical and engineering framework. In this chapter, we discuss current understandings and utilization of color in medical imaging applications and present a perspective for the abovementioned questions from the viewpoint of the utility of color images for the advancement of human well-being. Epistemological landscape The meaning of color in medical imaging can be analyzed within the context of the scientific and technical advancements in the field, the processes that lead to their introduction into clinical practice and the learning aspects that are inherent with any new technology. For more than a century, radiological images have been interpreted using grayscale or monochromatic representations primarily driven by traditional single-phosphor radiographic film media. The visualization of images acquired and displayed using radiographic film were consistent with the measurement of the attenuation of x-rays through body parts revealing interior details
20
Aldo Badano
of the human body and helping physicians understand disease conditions. Even today, among ubiquitous and sophisticated color display technologies, the totality of worldwide image interpretation of mammography x-ray images of the breast for the early detection of cancer is performed using grayscale visualization. It can be argued that training on grayscale film over many decades has biased the perception of radiologists in favor of monochromatic image viewing. Alternatively, one can argue that the simplicity of monochromatic representation (at least in terms of color contrast) represents well the underlying scalar quantity being visualized directly associated with the attenuation of x-ray photons by the imaged object. However, any such account is at odds with the reality of other fields where color permeates all visualization approaches. Take for instance an imaging modality currently gaining interest, thermography,1 a technique that uses infrared light to capture changes in metabolic activity and vascular circulation for a range of applications including breast cancer and flu screening. Since the recent birth of this modality at a time when display technology is predominantly digital, the use of color visualization is inherent to thermographic images. What was a powerful imaging approach in the beginnings of medical imaging is now challenged in several ways by technological advances. On one hand, the digital display of data is now primarily performed on color displays. Grayscale monitors, popular in the second part of the last century, are being phased out of production due to market pressures. It is expensive to fabricate a dedicated grayscale monitor for niche applications with small market shares that might not be able to pay the premium cost of a dedicated solution. In addition, imaging technology has become multidimensional in the broad sense of the term, requiring new approaches for visualization that are amenable to the representation of images and data using color scales to amplify the perceptual relationship with the underlying data. A chief example of this is the incipient transition from traditional or attenuation-based computed tomography (CT) to spectral CT imaging.2 In traditional computed tomography, the quantity represented in the reconstructed volumes is the x-ray attenuation coefficient of the relevant structures. Spectral CT systems can differentiate not only anatomical structures but also material properties that are a function of the x-ray spectral differential properties using energy-discriminating x-ray detectors. In 1 Elena V. Petrova, Hans P. Brecht, Massoud Motamedi, Alexander A. Oraevsky, Sergay A. Ermilov: In Vivo Optoacoustic Temperature Imaging For Image-Guided Cryotherapy of Prostate Cancer. In: Physics in Medicine and Biology, 63, 2018, pp. 6–12. 2 Martin J. Willemink, Mats Persson, Amir Pourmorteza, Norbert J. Pelc, Dominik Fleischmann: Photon-counting CT: Technical Principles and Clinical Prospects. In: Radiology, 289, 2018, pp. 293–312.
Color in Medical Images
21
this case, some have made the direct analogy between black-and-white and color photography and the utilization of spectral information in x-ray imaging. The technological evolution of devices used for displaying color in medical imaging involved the transition from decades old (1950s to 1990s) monochromatic offerings to a multichromatic approach with the introduction of color displays throughout the consumer and professional markets. In some cases, the presentation of chromatic shades expands beyond the traditional range described by the reduced color breadth of the device into a “wide” color space with the inclusion of more saturated primary colors to represent colors that are more realistic with a current market focus on the television market. Today, monochromatic display devices are no longer manufactured by device makers because of a lack of significant market demand. Consequently, monochromatic clinical image interpretations are performed today on color display devices in monochromatic mode even when special methods are required to calibrate such devices to maintain a grayscale performance consistent with monochromatic devices. The meaning of color in the context of medical imaging is interlinked with the concepts of performance and efficacy in the clinical setting. On one hand, the meaning of color is related to the underlying physical quantity that is represented in the images. On the other hand, color relates to the standardization of the visualization process that ensures consistent medical decisions. Within the scope of this text, the definition of meaning in different settings and applications is a recurring theme that needs clarification with respect to context. In this text, we analyze the use and meaning of color in medical images within applications with relevance and utility for clinicians but not necessarily associated with preference, stylistic and esthetical considerations as described in other chapters. A further consideration in understanding the meaning of color in medicine is to note that a visualization always involves colors even when monochromatic in nature. In other words, a grayscale display device or representation such as those used for mammographic interpretation portrays the image data using colors that are constrained and follow an approximately monochromatic profile from minimum luminance (or black level) to maximum (or white point). The terms black and white in the previous sentence are at best vague in that they do not refer to a specific chromaticity state but rather to the chromaticity that the display offers at minimum and maximum input levels. So, in a sense, all displayed images are always presented in colors directly representing image data even if a monochromatic scale is chosen for the visualization.
22
Aldo Badano
1: Synthetic pattern obtained by mixing clustered lumpy and white noise scaled into 256 levels and resembling clinical magnetic resonance perfusion images and visualized with three-color representations: hot-iron (left), gray (center) and rainbow (right) scale. The areas indicated by black arrows appear more uniform under the hot and rainbow scale representation than in grayscale mode.
Misperceptions The use of color scales for the interpretation of medical images can misguide conclusions reached by physicians. Figure 1 demonstrates this effect by representing identical image data with three color scales. It is notable that in some areas of the image, the visualization of the gradients in the data are represented in different ways that might have clinical implications if these were images of patient procedures. A comparative visual analysis of figure 1 indicates that some areas might appear relatively more uniform when employing one of the scales. The black arrows in figure 1 show areas that appear more uniform under the h ot and rainbow scale representation than using the grayscale mode. These variations in visualization can be linked to a weak association between the representation of image data and the color perceived by human readers and are primarily the result of engineering choices. Therefore, we claim that the meaning of color in medical imaging varies and is defined primarily by the utility that color provides to the transfer of clinically relevant information. In a broad sense, color usage can be categorized in three classes: as representation of the color appearance of an object of interest, as an aid for the visualization of an object’s property of interest and as annotation, highlighting or augmentation of the information in an image. While color fidelity in the imaging process is paramount for the first class, it is completely irrelevant for the latter two classes where the only possible strategy is a standard way of data representation to ensure consistent learning processes and interpretation outcomes. The emerging importance of the use of color visualization in medicine was addressed in a recent Summit on Color in Medical Imaging held at the U. S. Food
Color in Medical Images
23
and Drug Administration in 2013.3 It was recognized that color in medical imaging specialties is handled, with few exceptions, in an ad hoc manner with little standardization. The report was interested in color visualization beyond grayscales which have been, for many years now, standardized with significant success by the grayscale standard display function included in the universally adopted DICOM standard.4 The participants also recognized that this might introduce challenges to medical professionals who wish to use color images for diagnosis and treatment. Color in medicine For imaging applications, there is always a source of radiation, an object to be imaged, an image capture device and an image processing stage. Images are typically interpreted by humans using visualization hardware. In this context, color standardization is beneficial since the medical decisions of different medical practitioners will be consistent. More recently, the omnipresence of image analysis tools for all types of image content powered by the increasingly inexpensive available computational resources provides renewed interest in the consistency of color approaches since it is likely that algorithms can be more sensitive to the variability introduced by color mismatches. Moreover, color is affected by the properties and compatibility of hardware and software components of the imaging system. Within this diversity, it is important to adopt a clearly defined color architecture that ensures interoperability across devices that address two distinct aspects: accuracy and consistency. Color accuracy is defined as the ability to reproduce an exact color (or a series of colors) from the object space in the image representation. Consistency is defined as the ability across systems to visualize or perceive identical data by humans. For some imaging modalities, accuracy is paramount. In other instances, the use of artificial or pseudo-color scales requires only consistency. Color is applied in medical imaging in different modes: color for image annotations, pseudo-color mode where the color visualized has no physical meaning or correlation with the color in the object, and true-color mode where color reproduction and fidelity are intended. Below, we describe and illustrate examples of both pseudo- and true-color applications. Whole-slide imaging (WSI) for the primary diagnosis of tissue slides is the modality in most need of color accuracy and consistency standardization. In digital pathology, biopsy samples are stained with dyes and imaging systems are used to 3 Aldo Badano, Craig Revie, Andrew Casertano et al.: Consistency and Standardization of Color in Medical Imaging: A Consensus Report. In: Journal of Digital Imaging, 28, 2015, pp. 41–52. 4 National Electrical Manufacturers Association: Digital Imaging and Communications in Medicine (DICOM), Part 3.14, Grayscale Standard Display Function: ACR/NEMA, 1998.
24
Aldo Badano
capture an image of the entire slide. Color differences that might be used in the diagnostic interpretation can vary among laboratories and systems possibly affecting the outcome of image analysis algorithms. In addition to the lack of a standard on color for WSI systems, there is currently no standard method to address the high dynamic range in the image data. Consistent color in WSI requires efforts to characterize variations using test objects. For example, appropriate calibration slides may reveal differences between scanners and software packages. Another area of concern is endoscopy and laparoscopy, widely used techniques in many medical specialties including gastroenterology, gynecology, urology and surgery. Color plays an important role in the visualization of internal parts of the body which require in most cases some form of light delivery as the source, either invasively or non-invasively using a tubular device housing a light delivery conduit and an image acquisition channel. In capsule endoscopy, source and sensors are free to autonomous travel within open circuits of the gastrointestinal tract. Results from surveys have shown that although true color matching appears to be more important, color standardization is beneficial for technological advancement but remains technically challenging. Fundus photography is used in clinical ophthalmology as a mechanism of documenting the visual appearance of a patient’s retina. In this area, color fidelity and consistency among systems are both relevant. In this modality, the implementation of color corrections lacks standardization. A common method inspired on established methodologies based on the Macbeth Color Checker would provide more consistent color across cameras and systems. Similar issues are described in medical photography of large body parts or close-ups of the skin for diagnostic purposes and in dental photography. Inconsistent color accuracy can also affect applications in associated telemedicine solutions, not only in terms of system settings but also on the ambient illumination conditions. Best practices for medical photography along with reference targets has been identified as contributors to consistent color. In dentistry, photographs are not only widely used for diagnosis and treatment planning but also for esthetics treatments. Lastly, display systems are essential components of an imaging system for the visualization of color images. Display hardware is limited in several ways. First, the continuous (analog) nature of color in objects need to be transformed to a discrete (digital) representation before it is finally presented to the human as analog luminous output from the screen. Secondly, the representation of the color is typically achieved using a set of primaries (conventionally the red, green, and blue channels) which, while well-tuned to the perception of the human visual system, is limited to
Color in Medical Images
25
a set of colors given by the saturated primary chromaticity leading to phenomena including gamut shrinkage, color shifts, gray imbalance and ultimately color contrast reduction. Color displays also have large variability of the color at maximum luminance, also known as the white point. For these reasons, color calibration of displays for medical applications is necessary to guarantee stability of display systems over time and to ensure consistency across the medical enterprise. In handheld devices such as tablets and mobile phone displays, calibration is hindered by hardware and software limitations. Hardware calibration is also less practical in telemedicine for remote diagnosis and management of medical conditions to improve access to care. Color as quantitation tool The use of color in medical images has increased significantly in support of sophisticated visualization approaches. However, the ad hoc manner for handling color and the lack of standardization and common methodologies used to display medical images are often cited as contributing to suboptimal medical decisions with direct impact on patient treatment5 and prognosis.6 One medical imaging modality currently under development that relies on color is the assessment of functional images where a metabolic function is assessed quantitatively based on image data. Functional images are visualized in pseudo-color, defined here as the color-coded scalar imaging data that possess no correlation with the actual color of the imaged object. Color consistency across devices and images improves reproducibility of quantitative outcomes. An example of a pseudo-color application is perfusion magnetic resonance imaging (MRI) or computed tomography (CT). Perfusion images have a critical role in the treatment of stroke patients and in noninvasive diagnosis, staging and therapy response assessment of cancerous tumors.7 The usefulness of pseudo-color visualization was noted by Saba et al. in the imaging for possible carotid artery dissection.8 Saba’s findings that the accuracy and agreement were improved 5 Liron Pantanowitz: Digital Images and the Future of Digital Pathology. In: Journal of Pathology Informatics, 1, 2010, pp. 15–19. 6 Elizabeth A. Krupinski: Human Factors and Human-Computer Considerations in Teleradiology and Telepathology. In: Healthcare (Multidisciplinary Digital Publishing Institute, Switzerland), 2, 2014, pp. 94–114. 7 Bum Joon Kim, Hyun Goo Kang, Hye-Jin Kim, Sung-Ho Ahn, Na Young Kim, Steven Warach, DongWha Kang: Magnetic Resonance Imaging in Acute Ischemic Stroke Treatment. In: Stroke, 16, 2014, pp. 131–145. 8 Luca Saba, Giovanni M. Argiolas, Eytan Raz et al.: Carotid Artery Dissection on Noncontrast CT: Does Color Improve the Diagnostic Confidence? In: European Journal of Radiology, 83, 2014, pp. 2288–2293.
26
Aldo Badano
when the traditional grayscale is replaced by color is surprising and novel. The results of our work are somewhat consistent with Saba’s finding. We performed a laboratory study to investigate the effect of color scale and display device hardware on the ability of humans to detect small intensity changes in images resembling functional images. Using synthetic images that mimic the anatomical and functional structures found in perfusion studies, different visualization platforms were compared including a medical-grade display, a consumer-grade monitor, a tablet and a handheld device. With perfusion MRI images as the basis for designing the synthetic patterns, we generated synthetic patterns mimicking brain perfusion patient images. The images were the linear combination of statistical backgrounds in different proportions.9 Synthetic images, such as the ones depicted in figure 1, with a size of 200 × 200 pixels, were shown to humans with two patterns, side by side, in a two-alternative-forced choice (2AFC) paradigm. Each trial displayed two images of the same spatial pattern but slightly different intensities. The right pattern was flipped along the vertical axis to simulate comparing the two hemispheres of a human brain. Humans were asked to select the image with the highest intensity based on a reference color bar which was always displayed. Random white noise was displayed for 500 ms to delete latent images between trial images. We used a grayscale (gray), a heated black-body (hot-iron) and a rainbow palette (rainbow). Experiments were performed under controlled lighting conditions with an illuminance at the face of the devices of less than 5 lx. Experiments began with a 5-min adaptation period before trials. A split-plot design was used for the experiments to reduce the number of interpretations. A total of seventeen subjects participated in the study including two doctoral students, one radiology resident, one pathologist and many computer science students. The group included eleven males and six females with ages ranging between 22 and 78. As expected, performance with all color scales decreased with lower differences in intensity between the two images in the trial. Both rainbow and hot-iron scales performed better than gray across devices. We found that rainbow appears to be the most suitable color scale for functional image data sets. These results are consistent with the popularity of rainbow in medical imaging. We also found that performance was similar between medical-grade and handheld devices. Less saccadic
9 Cyril Castella, Karen Kinkel, François Descombes et al.: Mammographic Texture Synthesis: Second Generation Clustered Lumpy Backgrounds Using a Genetic Algorithm. In: Optics Express, 16, 2008, pp. 7595–7607.
Color in Medical Images
27
eye movements between trial images and a better general impression of the patterns might have contributed to this finding. In addition, the rainbow scale appeared to make differences in intensity more evident than other color scales leading to shorter decision times. Our findings demonstrate that the choice of color scale and display hardware affects the visual comparative analysis of pseudo-color images. Follow-up studies with patient images confirmed that color affects the visualization of medical images and that the effect is dependent on the clinical task. Interestingly, the rainbow scale has been criticized for lacking natural perceptual order forcing readers to refer more frequently to the color bar increasing time of interpretation. In addition, this scale can obscure details in the green–cyan range where humans are challenged and mislead interpretation through the introduction of data-independent gradients and false boundaries.10 Color as clinical tool Color visualizations are the preferred display mode for many imaging techniques. In clinical applications, color visualization is used for quantitation and detection tasks. Color also plays a role in the classification of lesions into benign and malignant, or more broadly, actionable versus non-actionable lesions. The higher level of complexity of the clinical tasks compared to the laboratory setting experiment described in the previous section adds layers of difficulty to the investigation of the effect of color. In addition, because clinical tasks differ significantly from modality to modality, answers obtained in one type of experimental setting do not always transfer to other clinical areas. In this section, we will demonstrate the variability among two clinical tasks: imaging for ischemic lesions with computed tomography and magnetic resonance imaging for prostate cancer.11 Magnetic resonance apparent diffusion coefficient (ADC) maps and myocardial computed tomography perfusion (CTP) are two imaging modalities using color visualization for which reproducibility has been questioned.12 Could color
10 Frank M. Marchak, William S. Cleveland, Bernice E. Rogowitz, Christopher D. Wickens: The Psychology of Visualization. In: Proceedings of the 4th Conference on Visualization (IEEE Computer Society, San Jose, CA, 1993), pp. 351–354. 11 Silvina Zabala-Travers, Brandon Gallas, Simone Busoni, Michelle C. Williams, Linsia Noferini, Luca Fedeli, Silvia Lucarini, Laura Galastri, Sayed Mirsadraee, Aldo Badano: Display colour scale effects on diagnostic performance and reader agreement in cardiac CT and prostate apparent diffusion coefficient assessment. In: Clinical Radiology, 74(1), 2019. 12 Berrend G. Muller, Joanna H. Shih, Sandeep Sankineni et al.: Prostate Cancer: Interobserver Agreement and Accuracy with the Revised Prostate Imaging Reporting and Data System at Multiparametric MR Imaging. In: Radiology, 277, 2015, pp. 741–750.
28
Aldo Badano
2: Sample myocardial computed tomography perfusion (CTP) positive case obtained from a clinical study using short axis-oriented slices representative of the left ventricle base. The same image is visualized with three color representations: hot-iron (left), gray (center) and rainbow (right) scale. The extent of the lesion indicated by black arrows appears enlarged in hot-iron and rainbow scales representation compared to the grayscale version.
visualization be contributing to this variability? Color imaging techniques often require many visual tasks including localization, detection and quantification. Which scale is the most prevalent, the rainbow, the hot-iron or the standard gray scale? How does this choice affect clinical tasks? We can appreciate the influence of color on clinical tasks by inspecting the differences in the three images showed in Figure 2. All three panels are representations of the same image data and yet, the ischemic lesion appears to be more prominent and extensive when using hot-iron or rainbow scales compared to the appearance under grayscale mode. In prostate ADC maps and myocardial perfusion, color is used in the qualitative and quantitative assessment.13 For example, ADC maps measure water movement in tissue in prostate tumors which is lower compared to normal areas. Accurate ADC quantification can reduce false-positive cases and unnecessary biopsies. In addition, precise localization is essential for biopsy and therapeutic procedures. In myocardial CTP (fig. 2), and due to reduced iodine concentration, ischemic and infarcted myocardium appears as hypodense compared to healthy tissue. Accurate localization and quantification of the transmural extent of the ischemic lesion is needed for accurate prognosis.14 13 Michelle Williams, John H. Reid, Graham McKillop et al.: Cardiac and Coronary CT Comprehensive Imaging Approach in the Assessment of Coronary Heart Disease. In: Heart, 97, 2011, pp. 1198– 1205. 14 Edward M. Geltman, Ali Ehsani, Mary Campbell et al.: The Influence of Location and Extent of Myocardial Infarction on Long-Term Ventricular Dysrhythmia and Mortality. In: Circulation, 60, 1979, pp. 805–814.
Color in Medical Images
29
3: Sample prostate positive ADC maps obtained from a clinical study. The same image is visualized with three color representations: gray (left), rainbow (center) and hot-iron (right) scale. Arrows indicate the selected location by clinicians for the lowest ADC value in the tumor showing relative consistency in the determination across different color modes.
In the experiment described here, twelve radiologists and cardiologists with age ranging from 27 to 36 read the CTP images at the University of Edinburgh (Edinburgh, UK). Nine radiologists at Careggi Hospital (Florence, Italy) with ages ranging from 31 to 44 read ADC maps. All readers read all cases with several color modes. Cases were retrospectively selected by experienced clinicians. For the CTP study, 210 short-axis-oriented CT slices representative of the left ventricle base, mid and apex were manually selected from 44 stress and 31 rest CT scans from 48 patients. A total of 105 of these cases were positive and 105 were negative for ischemic lesions in both invasive coronary angiography and fractional flow reserve. For prostate, a total of 165 MRI ADC map slices representative of all prostate levels were used, with 66 positive for tumor and 99 negative. Images were anonymized before selection. Clinicians used gray, hot-iron and rainbow scales. Samples images are presented in figure 3. All three panels are representations of the same ADC map data with arrows pointing to the location of maximum intensity according to a radiologist. In this particular clinical application, localization appears consistent. In other words, if an image-guided biopsy sample were to be taken from the prostate at that location, it would likely contain approximately the same portion of tissue independently of the color scale used for visualizing the image data. In the experiment, computer tablets were used for presenting cases and recording answers. Cases were assessed on a medical display (EONIS MDRC-2224BL, BARCO, Kortrijk, Belgium), sized 512 × 512 pixel on a full-screen black background. Ambient light was kept within the AAPM CT/MR/NM reading room recommendations of 50 lx. ADC maps using an inverted rainbow scale outperformed hot-iron
30
Aldo Badano
and grayscale modes on tumor detection. Due to the inverted scale, prostate tumors are displayed in yellows and reds, contrasting with the green and cyan of normal prostate. The intra-reader agreement across color scales was approximately 70% while inter-reader agreement was between 60 and 70%. Decision time was lower for experienced readers. Overall, agreement was higher with gray, while hot-iron and rainbow modes caused overestimation of lesion size. In this experiment, the clinical setting was simplified in many ways. First, humans were blind to clinical records. Second, cases were read as single slices as opposed to three-dimensional volumes. Finally, no adjustment of the brightness and contrast was allowed and might have been in some cases suboptimal. These findings suggest that a single, universal color scale for myocardial CTP and prostate ADC might not be optimal. Myocardial perfusion CT readings showed improved performance and more consistency when using grayscale visualization. Although small, the study found a measurable effect of color on image interpretation tasks that simulate clinical tasks. Summary The descriptions in this work indicate that the use of color in medical imaging is an area in need of further study not only in as much as the human interpretation of image data, but also in terms of the implementation of color-sensitive algorithms that will eventually replace or augment human abilities in this task. It is also interesting to note that the goals of accuracy of the color calibration and consistency of the representation of image data are also found in other fields of imaging including military reconnaissance and astrophysics applications. The ability of systems to apply consistent image transformations that are meaningful (related to the source and imaged object) and that can be completely understood by other systems is key to rapid progress and provides a solid foundation for the performance assessment of new imaging technologies. Since standards will continue to play a role in ensuring image quality, it is relevant to lead interested readers to the most recent efforts in this area. Although standards take significant commitment from stakeholders, their effect is critical as they accelerate innovation. A well-specified implementation with calibrated systems supports greater diagnostic effectiveness and consistency. The most notable standards and recommendations in this area come from the American Association of Physicists in Medicine (AAPM), the International Colour Consortium (ICC), the International Commission on Illumination (CIE) and the International Electrotechnical Commission (IEC). The AAPM is responsible for recommendations and professional guidelines in radiology. The ICC defines file formats
Color in Medical Images
31
and color transforms (so called look-up tables) and documents the workflows for using the transforms. The CIE is an international standards organization for color and illumination, with interests in lighting and imaging applications. Although it is unrealistic to expect instant universal adoption of any proposed color standard, the lack of a well-defined framework limits the use of modern digital tools. Moreover, user education around color capture, processing and display and system interoperability is necessary to improve the accuracy and consistency of medical decisions based on color visualization of image datasets. Disclosures The mention of commercial products herein is not to be construed as either an actual or implied e ndorsement of such products by the Department of Health and Human Services. This is a contribution of the U. S. Food and Drug Administration and is not subject to copyright.
Ulrike Boskamp
Color as the Other? Absence and Reappearance of Chromophobia in Eighteenth-Century France
“I feel the color in my cheeks rising again. I must be the color of The Communist Manifesto.”1 In E. L. James’s popular novel Fifty Shades of Grey (2012), this is the physical reaction that Anastasia Steele, the female protagonist, shows at her second meeting with Christian Grey. In accordance with his colorless name, he features gray eyes, gray suits, white office furniture. The book’s original title clearly references this “male” side of the story (fig. 1). Grey embodies what David Batchelor has very appropriately named “chromophobia.” Handsome, rich, lonely, neurotically self-controlled, he wishes to control his surroundings, derives sexual pleasure from dominance over his partner, and he never blushes. In the novel, color is reserved for the female’s skin, which turns various shades of red caused by the male, either when she blushes or when she is beaten. The book’s highly conventional setup of male domination and female subordination, an unambiguous hierarchy that is sexually played out in soft porn, uses and reconfirms a clear binarism of gender with great public success at a historical moment in which this binarism is being very seriously contested. It includes what the art historian Abigail Solomon-Godeau, referring to images, has called a “staging of the relationships of dominance and submission, authority and subordination, deeply inscribed into the structures of patriarchy and phallogocentrism.”2 Color and its marked absence play roles in this staging by transmitting their semantics to the subjects they are assigned to. Batchelor has dedicated his book Chromophobia to this binary, hierarchic relationship. He proposes that the case of color and its subordination “is bound up with the fate of Western culture,”3 and accordingly presents the subjugation of a triad of color, the female and a conglomerate of matter, body and nature under its 1 E. L. James: Fifty Shades of Grey, London: Arrow Books, 2012, p. 28. I thank Annette Kranen for her critical reading of this article. 2 Abigail Solomon-Godeau: Ist Endymion schwul? Spannungsgeladene Fragen zwischen Feminismus, Gay und Queer Studies. In: Mechthild Fend, Marianne Koos (eds.): Männlichkeit im Blick. Visuelle Inszenierungen in der Kunst seit der frühen Neuzeit, Cologne: Böhlau, 2004, p. 34. 3 David Batchelor, Chromophobia, London: Reaktion Books, 2000, p. 22.
34
Ulrike Boskamp
1: E. L. James: Fifty Shades of Grey, 2012.
colorless complements drawing/design, the male, and mind/culture as a feature with an extremely longue durée. The author claims that “Color was a permanent internal threat, an ever-present inner other which, if unleashed, would be the ruin of everything, the fall of culture.”4 Published in the year 2000, his diagnosis echoes a similar reasoning in Klaus Theweleit’s classic study Männerphantasien of 1977.5 Theweleit analyses autobiographical texts by members of the German free corps before WWI and demonstrates how these ultra-conservative soldiers’ view of their relationship with their political enemy was thoroughly gendered–and today we can add: clearly chromophobic. In these writings, the fear of communist hords, metaphorically addressed as red floods, rests on an imaginary6 of a dangerous and untameable other, layered with fantasies of the female, and metonymically of menstruation blood (fig. 2). 4 Batchelor (s. fn. 3), p. 23. 5 Klaus Theweleit: Männerphantasien: Frauen, Fluten, Körper, Geschichte, Frankfurt/Main: Roter Stern/ Stroemfeld, 1977. Theweleit does not extend his thesis transhistorically. 6 For the interdisciplinary reader it should be mentioned that the term “the imaginary” here refers to one of the three orders from the psychoanalytic theory of Jacques Lacan, along with the symbolic and the real.
Color as the Other?
35
2: A. Mrawek: Die rote Flut steigt (The red flood is rising), Cover from Der Wahre Jacob, No. 626, 1910.
Its binary opposite are the equally stereotyped self-portraits of authors who perceive themselves as crystalline, dry and colorless fighters, as phallic towers under a continual strain to resist the threat of being mollified by the red deluge. This setup with its merging of color, gender and politics is clearly related to that of Fifty Shades of Grey and Anastasia Steele’s communist cheeks. It could serve as an example for the universality of chromophobia – a notion I would like to challenge in this article. While it is difficult to contradict a continuity of misogyny in Western culture, the diagnosis of a transhistoric presence of a suppressed other, in which the three very broadly defined components – color, femininity and matter – are merged, seems simply not to be applicable to the largest part of literature on color from the French eighteenth century. With regard to the artistic discourse of late nineteenth and early twentieth century, Batchelor’s argument on chromophobia rests on a broad textual basis.7 It is corroborated by the work of other scholars, for example, recently by Helmut Lethen who qualified a preference for black and white, combined with the rejection 7 See the collection of sources in David Batchelor: Color (Documents of Contemporary Art), London: Whitechapel Gallery and Cambridge: MIT Press, 2008.
36
Ulrike Boskamp
of color as “a core of twentieth-century aesthetics.”8 One of Batchelor’s key witnesses for the historic dimension of chromophobia is the French art critic and theorist Charles Blanc (1813–1882) who very explicitly claimed a hierarchic relation between drawing/design as the male, and color as the female part of painting in his Arts du Dessin of 1867: “The union of design and colour is neccessary to beget painting just as is the union of man and woman to beget mankind, but drawing must maintain its preponderance over colour. Otherwise painting speeds to its ruin: It will fall through colour just as mankind fell through Eve.”9 Blanc drew this argument from the Querelle des Poussinistes et Rubenistes (1672– 1678), a dispute on the role of color in painting in the French Academy of arts 200 years earlier.10 For his reasoning for a transhistoric presence of chromophobia, Batchelor refers to Jacqueline Lichtenstein’s classic interpretation of this academic dispute in her book The Eloquence of Color.11 Lichtenstein’s investigation traces the origins of the antagonistic relation of the key concepts dessin and couleur back to the rhetoric of antiquity. In the emerging theory of painting, the line – in analogy to the text in rhetoric – was regarded as the main constituent of painting. On the other hand, color, though necessary for the physical existence of painting, was qualified as a secondary, physical and ornamental phenomenon of lesser value, in analogy with the human voice in rhetoric. Stressing this connection between the seventeenth-century aesthetics and antiquity, Batchelor posits a transhistoric continuity of chromophobic aesthetics and of a gendered binary of color and non-color from antiquity into the twentieth century. At this point, it seems necessary to introduce historic differentiations. Concerning the quote by Charles Blanc, the art historian Monika Wagner suggested that his statement must be seen as part of a conflict that developed in the course of an “emancipation” of color in contemporary art – especially in the work of Delacroix, 8 Helmut Lethen: Einleitung. In: Monika Wagner, Helmut Lethen (eds.): Schwarz-Weiß als Evidenz. “With Black and White You Can Keep More of a Distance,” Frankfurt: Campus, 2015, p. 8: “Schwarz-Weiß, ein Kern der Ästhetik des 20. Jahrhunderts.” 9 Charles Blanc: Grammaire des Arts du Dessin. Architecture, Sculpture, Peinture, Paris: Renouard, 1867, p. 23: “Le dessin est le sexe masculin de l’art; la couleur en est le sexe feminin.” The above translation is taken from Batchelor (s. fn. 3), p. 23. All other translations from French and German into English are by the author. 10 For Blanc’s familiarity with the historic debate see Kristiane Pietsch: Charles Blanc (1813–1882). Der Kunstkritiker und Publizist, 2004, https://docserv.uni-duesseldorf.de/servlets/DerivateServlet/ Derivate-2939/939.pdf , acc. 05–16–2018. 11 Jacqueline Lichtenstein: The Eloquence of Color. Rhetoric and Painting in the French Classical Age, Berkeley: University of California Press, 1993. Lichtensteins structural argument stresses how the application of rhetoric as a model for art theory led to an amalgamation of the concepts of matter and color in late seventeenth-century aesthetics.
Color as the Other?
37
but also in early impressionism – which Blanc rejected.12 She also points out that the definition of painting as sexual intercourse in a naturalized hierarchy of male and female introduces a biologism, which is not universal but very specific to the nineteenth century. This relates to other investigations in feminist and gender studies that have tackled equally persistent gendered binaries, in recognition of the fact that such imaginaries have shaped intellectual fields far removed from questions of human biology – such as Blanc’s definition of painting or the politics in Theweleit’s Männerphantasien.13 These studies have regularly come to the conclusion that the binary constructions they investigated were less universal, and often much more complicated than they looked at first sight, and that the particularly simple, biologist and binary imaginaries of male and female date no further back than the nineteenth century.14 For example, Lorraine Daston’s investigation into the concept of “intelligence” demonstrates how the idea of nature became applied to the feminine mind and was correlated with an unprecedented attribution of masculinity to the intellect around 1860.15 Anne Fausto-Sterling has worked on the binary fixing of gender in a biological sense and has also demonstrated that this was installed in the same era, in the course of the installment of what Foucault has called “a society of normalization.”16 In spite of the earlier existence of the dualism of color and line during the above mentioned Querelle, the origin of the chromophobic subjugation of the triad of color, matter and the female in the strong, biologistic sense that Batchelor has
12 Monika Wagner: Linie-Farbe-Material. Kunsttheorie als Geschlechterkampf. In: Barbara Hüttel, Richard Hüttel, Jeannette Kohl (eds.): Re-Visionen. Zur Aktualität von Kunstgeschichte, Berlin: Akademie Verlag, 2002, pp. 195–208, p. 196. 13 For example Emily Martin: The Egg and the Sperm: How Science Has Constructed a Romance Based on Stereotypical Male-Female Roles. In: Signs, 16, 1991, pp. 485–501. I thank Andrea Behrends for this and other precious hints. 14 It seems to take laborious deconstructions to open history up for a non-gendered structuring. In different studies, other historic concepts than the binary one have been investigated, thus shaping a new prehistory to a non-binary interpretation of the present. An important example is Anne Fausto-Sterling: Sexing the Body. Gender Politics and the Construction of Sexuality, New York: Basic Books, 2000. A convincing, much earlier study is Londa Schiebinger: The Mind Has No Sex? Women in the Origins of Modern Science, Cambridge: Harvard University Press, 1991. This author demonstrates how the Cartesian concept of a strict separation of body and mind around 1700 led to an acceptance of the equal ability of the female mind, and facilitated female participation in philosophy and sciences. 15 Lorraine Daston: The Naturalized Female Intellect. In: Science in Context, 5, 1992, pp. 209–235, see pp. 220–221 and pp. 226–227. 16 Fausto-Sterling (see fn. 14), p. 8 and p. 40. The expression is from Michel Foucault: Two Lectures. In: C. Gordon (ed.): Power/knowledge: Selected Interviews and Other Writings 1972–1977 by Michel Foucault, New York: Pantheon, 1980, pp. 78–108, p. 107.
38
Ulrike Boskamp
suggested, also seems to be a nineteenth-century construction. A major reason for the suspicion that a continuity of gendered chromophobia through the centuries did not exist are the sources on color from eighteenth-century France. Here, color and color theory were very widely and very publicly discussed, but color was not framed in any binary hierarchy or gendering. The physics and aesthetics of color were entangled, and an idea of scientifically justified color harmonies applicable in arts was widespread. This view on color began to be contested around 1750. Roughly speaking, there was a transition from an era fascinated by the connection of art, pleasure and physics to an era fascinated by the connection of art, emotion and nature. It was at the historic moment when this transition was initiated that in two different intellectual debates color was reintroduced into binary, hierarchic relationships. In the following, I will first sum up the eighteenth-century French disputes on color and then present these two debates in close-ups, in order to show that even in chromophobic debates, color was conceived differently than in the nineteenth century, and that therefore a continuity of the triad color-female-matter as a transhistoric constant does not seem to exist. Not the other: color in the era of Newtonianism At the beginning of the seventeenth century, color theories throughout Europe had taken up the Aristotelian ordering of color between black and white, which were seen as poles or extremes, and between which “primary” or other colors were arranged according to their luminosity. Then in the prevailing Cartesianism of the second half of the seventeenth century, color was qualified as a modification of white light, which formed the prime topic of physics.17 As such, color remained a somewhat second-rate phenomenon. But while its subordination to white light can be noted in Cartesian physics, there is no trace of a complementary, binary relationship, in which color was defined as the other, in the way it was introduced into art theory around the same time in the Querelle des Poussinistes et Rubenistes. This means that the binary construction around color in this era was not part of the scientific discourse,18 but only of the art discourse – which stands in accordance with Lichtenstein’s thesis. The status and significance of color, as well as the imaginary surrounding it, changed radically when Isaac Newton’s Opticks (1706/1720) gained recognition in
17 Abdelhamid I. Sabra: Theories of Light from Descartes to Newton, London: Oldbourne, 1967. 18 For the relationship between color and logic in the history of philosophy see Esther Ramharter in this volume.
Color as the Other?
39
3: Charles Nicolas Cochin (fils): Le Coloris, from Antoine Marin Lemierre: La Peinture, [1769].
France during the first half of the eighteenth century.19 Newton’s contention that white light was composed of colored rays introduced a logical and temporal priority of color, i. e., of colored rays over white light. White light was thus considered a secondary, derived phenomenon, a mixture of its immutable antecedents, the seven colored hues. Color became a field of popular interest and activities, was widely discussed and newly conceptualized in an entangled field between physics, aesthetics and the theory of art.20 The manifold color theories from this era have a few common denominators: They all explicitly state their position towards Newton’s Opticks. They all assume that there is a correct sequence of colors, which 19 Isaac Newton: Opticks or a Treatise on the Reflections, Refractions, Inflections and Colours of Light, London: Smith and Walford, 1704. The Latin version of 1706 was immediately read in France, and a French translation appeared in 1720: Traité d’optique sur les reflexions, refractions, inflexions, et les couleurs de la lumiere, translated by Pierre Coste, Amsterdam: Humbert, 1720. 20 For this context see Ulrike Boskamp: Primärfarben und Farbharmonie. Farbe in der französischen Naturwissenschaft, Kunstliteratur und Malerei des 18. Jahrhunderts, Weimar: VDG, 2009.
40
Ulrike Boskamp
is represented by the color spectrum of glass prisms and the rainbow, from violet or blue to red, and that black and white are not part of this range. Their two major topics are primary colors and color harmony. While the number of primary colors varied in different concepts, it was generally believed that a harmony of colors existed that was or would be explicable with recourse to an analogy of colors and tones that Newton had promoted. The notorious Louis Bertrand Castel’s invention of a color clavecin that was to play color music according to harmonic principles embodied this concept.21 This instrument included a modulation of twelve colors through twelve octaves of luminosity, yet there was no complementary, hierarchic or gendered relationship introduced between the two entities. In these French debates on color, in which the scientific approach to color was prevalent and entered deeply into aesthetic questions, there is no trace of gendering. Color was not linked to irrationality, quite the opposite: it was considered the epitome of a natural phenomenon that could be fully understood thanks to the achievements of experimental physics, and whose harmonic effects comprised science and beauty. A very clear expression of this can be seen in Nicolas Cochin’s image of coloris – color in painting – in which the colors from a Newtonian glass prism are projected onto a painter’s canvas (fig. 3).22 Anti-Newtonianism and the reintroduction of binary and gendered relations With the rise of natural history and of the ideas of Jean-Jacques Rousseau in the second half of the eighteenth century, the popularity of experimental physics faded. What followed for color was that in new, anti-Newtonian color theories, the prism came to be regarded as an instrument of deception. The demand for a more “natural” approach to color was raised and included the requirement that the human eye should be used as an instrument of research.23 New color theories replaced the sequence of colors from the prism by the sequence according to luminosity as perceived by the eye (fig. 4). 21 Louis-Bertrand Castel: L’Optique des Couleurs, Fondeé sur les Simples Observations, & Tournée Sur-Tout à la Pratique de la Peinture, de la Teinture & des Autres Arts Coloristes, Paris: Briasson, 1740. See Boskamp (s. fn. 20), pp. 117–138. In the sparse French art literature of the early eighteenth century such an autonomous harmony of colors, derived from physics and the rainbow, clearly carried positive connotations. 22 Antoine Marin Lemierre: La Peinture. Poëme en Trois Chants, Paris: Le Jay, [1769], p. 20. 23 For this context see Ulrike Boskamp: Natur versus Physik: Jean Baptiste de Lamarcks vergessene Theorie der Farben. In: Martin Dönike, Jutta Müller-Tamm, Friedrich Steinle (eds.): Farben der Klassik. Wissenschaft – Ästhetik – Literatur, Göttingen: Wallstein, 2016, pp. 205–229. While the sequence from the prism was red-orange-yellow-green-blue-indigo-violet, the new order went white-yellowred-blue-black.
Color as the Other?
41
4: Prismatic order of colors, true order of colors, from Jean-Baptiste Lamarck: Recherches sur les causes des principaux faits physiques, [1794].
This also had consequences for the concept of color in aesthetics and art theory. At the beginnings of this historic upheaval, around 1750, color was reentered into hierarchic binary relations. In two different debates – one on painting around the Paris Academy of Arts, the other the Querelle des Bouffons (1752–1754), a public debate about music – the idea of an independent harmony of colors legitimated by physics was contested, and color with its scientific connotations presented as an “other.” In both cases, a second non-colored entity was introduced to rule over the harmony of colors. These two examples are the only ones in eighteenth-century France, in which binary constructions similar to those in the Querelle were used in relation to color. The harmony of clair-obscur versus color harmony in painting In his lecture Sur l’harmonie et sur la couleur, held before the Paris Académie Royale de Peinture et de Sculpture in 1747, the prominent and influential art amateur Comte de Caylus reflected on the role of color in painting. The objective of his talk was to refute the “Newtonian” idea of an independent harmony between colors, and to promote a new ruler of harmony in painting, the “clair-obscur.”24 In academic art theory, clair-obscur refers to the structuring of a painting through the distribution of light and shade. Caylus claimed that harmony was not a relation between colors at all, and that harmony in painting could and should only be achieved through degrees of luminosity, i. e., the shading of colors: 24 Anne Claude Philippe de Tubières, Comte de Caylus: Sur l’Harmonie et Sur la Couleur, Conference, 4. November 1747. In: Jacqueline Lichtenstein (ed.): Les Conférences de l’Académie Royale de Peinture et de Sculpture , vol. V, t.1, Paris: ENSBA, 2012, pp. 70–80. For a presentation of this debate see Ulrike Boskamp: Contre l’Harmonie des Couleurs: Caylus, Gautier d’Agoty et le Retour à l’Ordre dans le Coloris. In: Marie-Pauline Martin, Chiara Savettieri (eds.): La Musique Face au Système des Beaux-Arts, ou les Vicissitudes de l’Imitation dans les Arts (1690–1803), Paris: Editions Vrin, 2014, pp. 225–241.
42
Ulrike Boskamp
“It is not without reason that I have distinguished color from harmony. The latter exists independently and finds itself mounted like an instrument on the tone that one chooses, higher or lower; it does not matter, for the accord. […] Color, on the other hand, can be very good in itself but lacks harmony […].”25 Other theorists followed this new idea. The anti-Newtonian color printer Jacques Fabien Gautier d’Agoty took the opportunity to fight against Isaac Newton’s French disciples, re-interpreted the old analogy of music and painting in such a way that the clair-obscur, not colors, would serve as the principle of harmony, and elaborated on the practical side of a harmony based on clair-obscur.26 The art theorist Claude-Henry Watelet added a physiological argument and claimed that while the individual constitution of painter’s eyes shaped their perception of color, the clair-obscur could be learnt and applied independent of such differences.27 It was rationally applicable and controllable and could even facilitate painters with weak eyes to achieve good results in their art through understanding and exercise. These ideas also seem to have been put into practice by contemporary painters where the pointed use of clair-obscur, of subdued colors and an avoidance of the combination of pure primary colors can actually be observed.28 This first discourse in which color was reinserted into a binary relationship actually presents color harmony in painting as something in need of control. Its colorless, dominant counterpart clair-obscur is introduced to provide a more rational way of achieving pleasing effects on the human eye in painting. Passionate line versus pleasurable color in the Querelle des Bouffons The second reappearance of a binary in the field of color took place in the Querelle des Bouffons, a controversy which opened up between the followers of Italian opera and those of French opera in 1752.29 What was at stake for the two main opponents – Jean 25 Caylus (s. fn. 24), p. 75: “ce n’est pas sans raison […] que j’ai distingué la couleur de l’harmonie. Celle-ci existe en ell-même et se trouve montée comme un instrument, sur le ton que l’on choisit, plus haut ou plus bas, il n’importe, pour l’accord. […] La couleur, au contraire, peut être très bonne et manquer d’harmonie […].” 26 Jacques Fabien Gautier d’Agoty: Sur la Musique des Couleurs, Inventée par le Père Castel, Observation I. In: Observations Sur l’Histoire Naturelle, Sur la Physique et Sur la Peinture, 1, 1752, pp. 37–43. 27 Claude-Henri Watelet: L’Art de Peindre, Poème; avec des Réflexions Sur les Différents Parties de la Peinture, Paris: Guerin & Delatour, 1760, p. 119, see Boskamp (s. fn. 20), p. 198. 28 See Boskamp (s. fn. 20), p. 200–203. 29 Catherine Kintzler: Jean Philippe Rameau. Splendeur et Naufrage de l’Esthétique du Plaisir à l’âge Classique, Paris: Minerve, 2. ed. 1988; for further reading see Andrea Fabiano (ed.): La Querelle des Bouffons dans la Vie Culturelle Française du XVIIIe Siècle, Paris: CNRS éditions, 2005.
Color as the Other?
43
Philippe Rameau (1683–1764) on the side of the French, and Jean-Jacques Rousseau (1712–1778) on the side of the Italian opera – was the assessment of the role of harmony, in the framework of the overarching aesthetic and epistemological question to what end, and through which means music and art in general should be produced. Rameau, the protagonist of French opera, firmly believed that music was and should be based on harmonic rules, rooted in physics.30 His adversary Jean-Jacques Rousseau, 30 years younger, passionately contested this. It was the melody that Rousseau proposed as the “natural” base of all music, setting it against a concept of harmony whose pleasures were triggered by physical-mechanical relations, which he regarded as false, artificial and alienated. In his Essai sur l’origine des langues of 1752 Rousseau elaborated on this, claiming that music was an utterance of human passions which naturally divest themselves through the human body, the voice, into the melody of a song. In a lengthy narration, music is represented by painting, and musical harmony by the harmony of color, apparently on the assumption that this would make the argument more plausible. Rousseau describes the painting of an imaginary land where artists work exclusively with color and are completely ignorant of the drawing or line. They produce non-representational color paintings instead of mimetic images, and their discourse skirts around the combination and nuancing of colors. Rousseau presents this – in parallel to a music based only on physically defined harmonies – as an unthinkable absurdity, a mistake of category, because of its lack of line, of drawing, which here becomes synonymous for the expression of human passions, and for nature:31 “What will we say about a painter who is so empty of feeling and taste that he argues in such a way that he stupidly bases the pleasures of painting on the physics of his art? […] As painting is not the art of combining colors in an agreeable way for the eyes, music is not the art of combining tones in an agreeable way for the ears. If there was only that, both would be sciences and not arts. It is only the imitation of nature that elevates them to the ranks of arts. Because what makes painting an imitating art? It is drawing. And what makes music an imitating art? It is melody.”32 30 Rameau himself had provided a physical hypothesis for musical harmony in his treaty “Generation harmonique” in 1737, in which he provided a new explanation of musical harmony, departing from the physics of overtones. See Kintzler (s. fn. 29), pp. 27–29. 31 Jean-Jacques Rousseau: Essai Sur l’Origine des Langues, Ou il Est Parlé de la Mélodie et de l’Imitation Musicale (1753). In: Bernard Gagnebin et al. (eds.): Écrits Sur la Musique, la Langue et le Théâtre (Œuvres Complètes, vol. 5), Paris: Gallimard, 1995, pp. 375–429. 32 Rousseau (s. fn. 31), p. 413.
44
Ulrike Boskamp
In order to establish a new understanding of music, Rousseau here takes the binary opposition from seventeenth-century art theory, using it as a tool, or better: as an armory. In his hierarchic opposition of natural passions – symbolized by the line – versus artificial pleasures – symbolized by the harmony of colors – , gender is clearly, if implicitly, present. “Plaisir,” which until then had denoted the agreeable perception of harmony in music or painting, now evokes the overly and unnecessarily refined female and is coupled with physics as an alienation from the simplicity of nature.33 One of Rousseau’s followers, Antoine Pluche, makes this explicit in his 1755 edition of his voluminous popularization of science, the Spectacle de la Nature, where he comments on the same topic: “To present the pleasure for the sake of the pleasure itself is a reversal: Let us use a clearer term: it is prostitution. How many artists are condemned by this one word!”34 Pluche qualifies the pleasurable effects of harmony on the human eyes or ears – a mechanical relationship between the human organ and a combination of tones or colors – as invalid, and parallels them to the effect of a female prostitute on the body of a male. Like in Rousseau’s text, the pleasures of harmony are thus accepted as the result of relations of cause and effect, only to be morally condemned. In this second example, the rejection of the older Rameauian esthetique du plaisir is carried out by the allocation of a female gender to the combination of color, pleasure and physics. Similar to the first example, the contested other is the mechanic relationship between a physically defined harmony of colors and the human organs. These pleasures of harmony were gendered female and were to be surmounted by the male passions of melody.35 While in the first half of the eighteenth century, color had been much discussed and not gendered, in the two debates presented here, color was reintroduced into 33 See Élisabeth Badinter: Émilie, Émilie. L’Ambition Féminine au XIIIe Siècle, Paris: le Livre de poche, 1984. 34 Antoine Pluche: Le Spectacle de la Nature, ou Entretiens Sur les Particularités de l’Histoire Naturelle qui Ont Paru les Plus Propres à Rendre les Jeunes-Gens Curieux, et à leur Former l’Esprit, 8 vols., Paris: Estienne, 1752–1755, vol. 7, 1755, p. 110: “Présenter le plaisir pour le plaisir même, c’est un renversement: servons-nous d’un terme plus clair: c’est une prostitution.” 35 The art historian Thomas Crow has diagnosed that visual culture underwent a shift towards a “masculinisation,” in the era of neoclassicism and the French revolution. Artists around Jacques Louis David were promoting aesthetic and political ideals taken from antiquity, setting them against the “rococo” art of the Ancien Régime which was feminized in this process. Thomas E. Crow: Emulation. Making Artists for Revolutionary France, New Haven/London: Yale University Press, 1995. This “historic change in the gender of beauty” is also discussed by Solomon-Godeau (s. fn. 2), p. 15. The discussion is summarised in Bettina Uppenkamp: Kunst und Kunstgeschichte. In: Stefan Horlacher, Bettina Jansen, Wieland Schwanebeck (eds.): Männlichkeit. Ein interdisziplinäres Handbuch, Stuttgart: Metzler, 2016, pp. 256–270, p. 263.
Color as the Other?
45
binary relations, gendered female, and the argument was misogynistic. In both cases, the binary structure taken from the art theory of the earlier Querelle was called in order to subordinate color in hierarchies that touch artistic techniques as well as values and morals. However, for an assessment of these in the mid-eighteenth century, the completely different layout of the discursive field as well as the transformed ideas about what art should do and achieve, need to be taken into account: While in the seventeenth century the argument for the subordination of color had been the irrational quality of its seductive beauty, in these debates it was the scientification, i. e., the rationalization of color and of its harmonic effects which led to its rejection and to female gendering.36 So color and the female were here not linked with nature, as they would be in the nineteenth century, but with physics, artificiality and with the all-too refined and alienated culture. On the other side of the binary, the melody in its metonymic relation with the colorless line, stood not for rationality and the mind, but for nature and passion. So while a resemblance in the binary, hierarchic and gendered structure of the respective arguments is undeniable and explicable with the toolkit from the Querelle, the tableau of entities and ideas that were debated was thoroughly different from those around the advent of Impressionism in the nineteenth century. Imaginations of antiquity, colored and white Around the same time, color was also discussed in another context: with regard to antiquity. In this discourse, while references to the physics of color were also present, the main topics were on the one hand the color in antiquarian objects, and on the other hand ideals of taste and beauty derived from antiquity. In France, the protagonist for a new revival of interest in antiquity, triggered by the contemporary excavations in Italy, was again the Comte de Caylus. He collected and published on ancient Roman objects himself, and campaigned for a change in art, detesting the “rococo” taste of his contemporaries. The new art was to take its motifs from antiquity again – a movement which was later called “neoclassicism.” Within this context, the question of color in Roman antiquity came to the fore and was not just discussed from the textual sources, but also in light of the remains of frescoes and sculpture.37 The general opinion was that the painting of antiquity was of lesser value than contemporary French painting, also because of the bluntness of its colors.38 One 36 Boskamp (s. fn. 20), pp. 187–188. 37 Boskamp (s. fn. 20), pp. 93–99. 38 See for example Antoine Coypel: Discours Prononcé dans les Conférences de l’Académie de Peinture et de Sculpture, Paris, 1721, p. iij; Boskamp (s. fn. 20), pp. 93–99.
46
Ulrike Boskamp
French publication, edited by Pierre Jean Mariette together with Caylus, set out to change this notion, and to make antiquity visible as a colored epoch The Recueil de peintures antiques, imitées fidélement pour les couleurs & pour le trait39 contained printed images of frescoes from Roman antiquity. These had been transmitted through colored drawings taken on the spot from the originals, and the Parisian customers could order versions handcolored directly from these drawings, in order to reduce the loss of authenticity (fig. 5). In the preface, the authors praise the color of antique painting: “Our eyes, used to a magic of painting which is all to often far away from the truth, […] would have pains to accustom themselves to this simplicity of composition, to this union of clair-obscur, to these pure and whole colors that were the joy of the Ancients, and that, I dare to say, would also merit to be our joys, if the love of novelty and the desire to show esprit had not led to an insensitive loss of the taste for the beautiful and simple nature.”40 The code words “belle et simple nature” from the Rousseauian repertoire present antiquity as a historic epoch free of overly sophisticated tastes, very much in accordance with Rousseau’s ideas on harmony developed in the Querelle des Bouffons that had been published with great public success a few years before.41 Again applying a very similar vocabulary, Johann Joachim Winckelmann in the 1750s characterized antiquity, with “noble simplicity” and “quiet greatness,”42 39 Pierre Jean Mariette: Recueil de Peintures Antiques, Imitées Fidélement Pour les Couleurs & Pour le Trait, d’Après les Desseins Coloriés Faits par P. Sante Bartoli, Paris, 1757, http://arachne.uni-koeln.de/item/ buch/1421, acc. 10–2018. The original drawings by Pietro Santo Bartoli were deposited in the King’s library for this purpose, in an attempt to reduce the alteration of colors caused by multiple copying. For Mariette’s relation to antiquity see Kristel Smentek: Mariette and the Science of the Connoisseur in Eighteenth-Century Europe, Abingdon/New York: Routledge, 2016, pp. 191–198. 40 Mariette (s. fn. 39), p. 2: “Nos yeux accoutumés à une magie de la Peinture, qui, trop souvent hors du vrai, n’en cause pas moins une sorte d’illusion & de prestige, auroient peine à se faire à cette simplicité de composition, à cette unité de clair-obscur, à ces couleurs pures & entieres qui faisoient les délices des Anciens, & qui, je l’ose dire, mériteroient encore de faire les nôtres, si l’amour de la nouveauté & le desir de montrer de l’esprit, ne nous avoient fait perdre insensiblement le goût de la belle & simple nature.” 41 Mariette is certainly referring to Pliny’s topical critique in book 35 of his Natural History, where he diagnoses the richness of color in contemporary painting as a decline from the simplicity and high quality of color in earlier painting. Pliny the Elder: The Natural History, transl. and ed. by Harris Rackham, Vol. IX (Libri XXXIII-XXXV), Cambridge: Harvard University Press/London: Heinemann, 1952, pp. 298–299. 42 Johann Joachim Winckelmann: Gedanken über die Nachahmung der griechischen Werke in der Malerey und Bildhauerkunst, Zweyte vermehrte Auflage, Dresden/Leipzig: Walther, 1756, p. 21.
Color as the Other?
47
5: Pitture della stanza segniata nu, from Pierre Jean Mariette: Recueil de peintures antiques, 1757.
proceeding not from its painting but from its sculpture. These ideals found widespread acceptance in Europe and particularly in France.43 Winckelmann propagated a white antiquity as an aesthetic ideal.44 This taste was initiated by the whitewashed remains of Roman sculpture and enhanced by copies in plaster spread all over Europe, and somewhat rested on the earlier artistic revivals of antiquity like the sculpture of Michelangelo (1475–1564). But Winckelmann also inserted his argument into the aesthetic discourse, which was still intermingled with physics. Where he puts forth his ideal of white beauty, he references Newton’s Opticks in which white light is composed of rays of all colors: “As white color is that which reflects the
43 Elisabeth Dêcultot: Winckelmann in Frankreich 1750 bis 1800. Ein Politikum. In: Dresdner Hefte. Beiträge zur Kulturgeschichte, 131, 2017, pp. 43–51. 44 Winckelmann set out from a clear rejection of the merits of color in antiquity, but after arriving in Rome, liked the colored frescoes of antiquity in Herculanum and found for himself that sculptures from antiquity showed remains of color. Yet this knowledge did not change his preference for whiteness. See Bernhard Maaz: Ein Jahrhundert Wirkung: Winckelmann und die Skulptur. In: Elisabeth Décultot, Martin Dönike, Wolfgang Holler, Claudia Keller, Thorsten Valk, Bettina Werche (eds.): Winckelmann. Moderne Antike, München 2017, pp. 9–126, p. 110; Helmut Pfotenhauer: Ausdruck. Farbe. Kontur. Winckelmanns Ästhetik und die Moderne. In: Décultot et al., pp. 67–82, pp. 73–76.
48
Ulrike Boskamp
6: Antonio Canova: Napoleon’s Mother, 1808.
most rays of light, and thus makes itself more sensitive, a body will be more beautiful the whiter it is […]”45 While this (pseudo-) scientific justification for the ideal beauty of a white, colorless body seems to have quickly disappeared as an relict of the transition from Newtonian times, the corresponding taste prevailed. Around 1800, the Italian sculptor Antonio Canova pursued it with enormous international success, using its reference to antiquity for the elevation of the present, for example in the portrayal of Napoleon and his family (fig. 6). The deliberately counterfactual idea of a white antiquity, spread through the medium of sculpture, seems to have been a much more influential precursor of the chromophobic aesthetics in nineteenth and twentieth century Europe than the anti-colorist concepts from the theory of painting presented above.
45 Johann Joachim Winckelmann: Geschichte der Kunst des Alterthums, Dresden: Walther, 1764, pp. 147–148: “Da nun die weiße Farbe diejenige ist, welche die mehresten Lichtstrahlen zurückschicket, folglich sich empfindlicher macht, so wird auch ein Körper desto schöner seyn, desto weißer er ist […].” There is a clearly racist side to this ideal which Winckelmann elaborates in the following sentences.
Color as the Other?
49
Conclusion This slightly sketchy tour de force through shifts in notions of color and othering in eighteenth-century France was undertaken to include the eighteenth-century debates into the history of chromophobia, and with this, to question the thesis of its continuity. The very high status of color in the first half of the eighteenth century, both in scientific and in aesthetic discourses, clearly serves as an argument against a continuity of chromophobia. When color was reintroduced into a gendered binary pair, this happened as a measure of warfare against the application of the physics of color into aesthetics at a phase of transition from rococo to neoclassic art, and also from a dominance of physics to a dominance of natural history. Yet, these two examples of chromophobic concepts do not prove a continuity in the implications of color from the seventeenth century to the twentieth either. The chromophobic arguments do not seem to be transhistoric, but are embedded in and defined by the respective historic tableaus in which the facets of the meaning and imaginary of color is defined relationally, also by the varying non-colored complements. The hierarchy of genders, deeply inscribed into the symbolic order, seems to be activated in aesthetics at moments of upheaval as a toolbox, leading to an imaginary reduction of the value and forces of an adversary, defined as colorist. More than the clearly chromophobic and gendered debates on color of the eighteenth century, the highly ambivalent neoclassicist aesthetics seem to qualify as a prehistory to the attribution of masculinity to Fifty Shades of Grey.
Alexander Nagel
Research on Color Matters: Towards a Modern Archaeology of Ancient Polychromies The current excitement for digital polychrome reconstructions of ancient monuments corresponds to new energies and investments made in advanced image technologies and the development of innovative computer graphic tools in the twenty-first century. Until about a generation ago, enthusiastic explorers used traditional tools to construct how they believed the polychrome past of the ancient world looked like. Occasionally, these physical constructions generated harsh debate, yet the reasons for the debates were complex and in general not limited to what the British art historian David Batchelor has defined as Chromophobia.1 Physical constructions such as those shown in the contemporary traveling exhibition Bunte Götter are still important as we practice thinking more about the complex technologies of paint application and the materials used. Though academics understand that pre-modern sculptures and monuments were brightly painted, every new generation will need to be educated about this one aspect of the past. The goal of this article is to illuminate and contextualize some aspects of the complex history of research, documentation and debate on past polychromies, with particular reference to monuments excavated in Egypt, Persia and Mesopotamia. As my article will show, documenting and defending color has never been an easy task. Beginning with a brief introduction on systems of documenting color in archaeological ceramics, I will introduce a cast of key players and individuals involved in the documentation, “measuring” and translation of hues and colors from the field to the museum and beyond, all invested in the early business of reconstructing sculptural painting (polychromies) in the nineteenth century. I will also discuss ways how these participants’ observations and documentation were acknowledged and distributed in European salons and media.2
1 David Batchelor, Chromophobia, London: Reaktion Books, 2000. 2 I confess that this is a very Western biased approach. I mainly question how European scientists thought and wrote about ancient polychromies in the Middle East. I recognize the important new contributions made on antiquarians researching colors in the Islamic world in recent years: Sheila Blair, Jonathan Bloom (eds.): And Diverse are their Hues. Color in Islamic Art and Culture, New Haven; CT, and London: Yale University Press, 2011.
52
Alexander Nagel
1: Color chart to document glazes in the field, as developed by Erich Schmidt at Rayy near Tehran in Iran, ca. 1935.
Documenting surface in the field: color charts in Egypt and Persia During excavations on the site of Ray near Tehran between 1934 and 1936, the German archaeologist Erich Schmidt (1897–1964) developed a more “efficient” system to document the colors of the hundreds of glazed ceramic shards the team would recover in the field (fig. 1). Grouped into categories from A (Red) to F (Green), each category of glaze had its own system of subsections (e. g., Red, Purple to Brown; Buff, Yellow, Cream to Green etc.). Schmidt did not leave behind a manual, and the system seems to have been circulated less extensively as it was not referred to in any of the final excavation reports.3 3 I am grateful to the staff at the American Museum of Natural History in New York City for facilitating access to the papers and research archives of Erich Schmidt. On the history of excavations at Rayy, Rocco Rante: Rayy. From its Origins to the Mongol Invasion: An Archaeological and Historiographical Study. Leiden: Brill, 2015. Many color paintings on ceramic vessels excavated at Rayy were signed by I. Gerassimov, yet the identity of the maker of the color chart is unknown today. All drawings of the vessels excavated have been labeled with “artist unknown” in the exhibition catalog: Tanya Treptow (ed.): Daily Life Ornamented: The Medieval Persian City of Rayy, Chicago: The Oriental Institute
Research on Color Matters
53
2, 3: Pages from a sketchbook of Ernst Herzfeld, 1929, documenting wall paintings at Kuh-e Kwaja. Names of colors (e.g., “rosa”, “gelb,” “blau”) were written while working on the site for easier identification.
The use of color charts in fieldwork was implemented in Egypt, too, where 1930s explorers of the past used the “Ostwald Colour Albums,” manufactured in London.4 Today, these color charts are proof that recording colors as accurately as possible has always been a challenging, yet considered a necessary task in scientific documentation. Such charts may also have been intended to help in finding an Publications, 2007. See, also, Donald Whitcomb: Lusterware jug from Rayy, Iran, Artist unknown, 1936, In: John Green, Emily Teeter, John Larson (eds.): Picturing the Past. Imaging and Imagining the Ancient Middle East. Chicago: Oriental Institute Museum Publications, 2012, pp. 143–144. 4 Nigel Strudwick: An Objective Color-Measuring System for the Recording of Egyptian Tomb Paintings, Journal of Egyptian Archaeology, 77, 1991, pp. 43–56.
54
Alexander Nagel
easier way of communicating between local fieldworkers who were involved in the documentation and their supervisors who were often only familiar with English-language color terms. In many ways, the most widely used commercial system of communicating colors around the same time were the so called Munsell Color Charts: developed and named after professor and painter Albert Munsell (1858–1918), and first patented in 1906 as a “Color Chart or Scale,” the chart emerged as a “grammar of color” in field archaeology, and is being used until this day, despite having met with criticism over the years.5 Only a few years earlier in 1929, another archaeologist active in the archaeology of Persia at this time, Ernst Herzfeld (1879–1948), in documenting the colors of the Sasanian wall paintings on the facades of the site Kuh-e Khwaja in Sistan in Eastern Persia, still used a more traditional technique by simply writing the names of colors while using pen and markers in his fieldwork diary (fig. 2 and 3). This system was also preferred by Norman de Garies Davis (1865–1941) and Nina de Garies Davis (1881–1965) in their documentation of wall paintings in ancient Egyptian tombs.6 The background Color charts were one way of documentation. Yet there were other early forms of documentation, and the arguments brought forward in Batchelor’s Chromophobia rest on a broad, yet very selected textual basis.7 They do not pass a critical evaluation if we look at the many attempts to bring color in, even document color early on in the nineteenth century. Attempts to find a universal description of colors among artists and antiquarians were first published in the eighteenth century. In Vienna, Johann Ferdinand von Schönfeld (1750–1821) published the first edition of the monumental “Wiener Farbenkabinett” in 1794, naming and numbering over 4,600 color samples. The success of the volume initiated a series of further books aiming to assist painters and printers: in 1799, Johann Heinrich Meynier (1764–1825) published his “Farbtabelle” 5 On the history of the Munsell Color Chart System and its application: Sally Cochrane: The Munsell Color System: A Scientific Compromise from the World of Art, Studies in History and Philosophy of Science, 47, 2014, pp. 26–41; Rolf Kuehni: The Early Development of the Munsell System. Color Research and Application, 27, 2002, pp. 20–27; Criticism: Rudolf Richard Gerharz, Renate Lantermann, Dirk Spennemann: Munsell-Farbtafeln, eine Notwendigkeit für den Archaeologen? Acta Praehistorica et Archaeologica, 18, 1986, pp. 177–187. 6 Alexander Nagel: Kuh-e Khwaja. In: Barbara Helwing, Patricia Rahemipour (eds.): Tehran 50. Ein halbes Jahrhundert deutsche Archäologen in Iran, Mainz: Zabern, 2012, pp. 42–44; Nigel Strudwick: Problems of Recording and Publication of Paintings in the Private Tombs of Thebes. In: W. Vivian Davies (ed.): Colour and Painting in Ancient Egypt, London: the British Museum, 2001, pp. 126–140. 7 Batchelor (s. fn. 1). See the article by Ulrike Boskamp in this volume.
Research on Color Matters
55
for popular audiences in Leipzig.8 Pondering ideas of communication and scientific approaches concerning color, Goethe’s influential text “Zur Farbenlehre” (1810) was widely circulated among intellectual elites in Europe and even reached cities in Persia: the Kerman born scholar Mohammad Karim Khan-Kirmani (1810–1871), in reflecting upon Goethe’s work, commented upon the difference between the existence (wujud) and the manifestation (zuhur) of colors.9 I am unaware, however, of any documents comparable to commercially produced nineteenth-century European color charts in cities such as Mosul, Tabriz, Tehran, Isfahan or Shiraz at that time. Yet as European scholars look beyond their own field, it has been recognized that documenting colors has been an inherent problem from the very start. As recently as 2015, Egyptologist Nigel Strudwick stated that “scholars have [yet to] fully come to terms with the documentation of color, whether painted relief or pure painting.”10 How did European explorers document and map paints they observed when traveling to the East and visiting ancient sites in nineteenth-century Mesopotamia and Persia then? The first example will lead us to Persepolis, the heartland capital of the Achaemenid Persian Empire between the late sixth and late fourth centuries BCE in southwestern Iran, where ongoing work on the original color schemes revealed further traces of ancient paints on the monuments.11 Here, it was the French antiquarian Charles Texier (1802–1871), who was the first to explore and document aspects of the rich polychromies on the facades of the buildings in the mid-nineteenth century. Arriving on the Takht of Persepolis and studying both the site and the nearby tombs of the Achaemenid Persian kings at Naqsh-e Rustam in January 1840, the goal of the expedition was to provide more detailed plans of the ruins.12 Yet the observations made of the paintings preserved on the sculptures are very rich:
8 Johann Heinrich Meynier: Die Kunst zu Tuschen und mit Wasserfarben, Leipzig: Graeff, 1799. 9 Zarah Abdollah: Color in Islamic Theosophy: An Analytical Reading of Four Scholars: Kubrā, Rāzī, Simnānī, and Kirmānī, Journal of Islamic Theosophy, 7, 2011, pp. 35–51. 10 Nigel Strudwick: Interpretation. In: Melinda Hartwig (ed.): A Companion to Ancient Egyptian Art, Wiley Blackwell, 2015, pp. 486–503, particularly p. 499. See also Frank Preusser, Michael Schilling: Color Measurements, in: Miguel Angel Corzo (ed.): Wall Paintings from the Tomb of Nefertari: Scientific Studies for their Conservation. Cairo and Los Angeles: The Egyptian Antiquities Organization and the Getty Conservation Institute, 1987, pp. 70–81. 11 Alexander Nagel: Color and Gilding in Achaemenid Architecture and Sculpture. In: Daniel Potts (ed.): The Oxford Handbook of Ancient Iran, Oxford: Oxford University Press, 2013, pp. 596–621. 12 Charles Texier: Description de l’Arménie, la Perse et la Mesopotamie, Vol. 1., 1842, Paris: Didot, pp. xxxvii-iii; Vol. 2, 1852, p. 163.
56
Alexander Nagel
“[E]n effet, le point capital de ma visite à Persepolis est la reconnaissance positive des peintures qui décoraient les bas-reliefs. Si je me suis empresse de communiquer ce fait aux savants et la presse, c’est qu’il ne reposait pas seulement sur une opinion: le temoignange de mes yeux ne m’avait pas suffi et en plusieurs endroits l’application de réactifs avait fait revivre certaines parties coloriées; c’était donc pour moi un fait démontré et incontestable.”13 While this statement introduces the first volume appearing in 1842, it took until the publication of volume two in 1852, before his full observations on the Persepolitan monuments and a first color reconstruction of a stone relief were published. Texier’s words indicate that he arrived at the site of Persepolis with expectations of the potential recovery of evidence of polychromies on the monuments.14 Texier was the very first person to use a chemical experiment to get “behind” the materials employed on the surface: “[J]e grattai la pierre avec soin, et fis dissoudre la poussiere dans de l’acide hydrochlorique, car j’avais avec moi une petite boite de reactifs. J’obtins un residu de couleur grise, qui, jeté à son tour dans un tube contenant de l’ammoniaque, me donna, vingt-quatre heures après, une solution d’une belle couleur bleue. C’etait, à n’en pas douter, une application de cendre bleue dont la base est du cuivre, qui servait de fond à ces sculptures.”15 On an impressive color reconstruction of a Persepolis relief, depicting a king with two attendants, the figures are set against a blue background. All details, including the skin and the hair are deliberately covered with paint. The garments and headdresses are elaborately embellished, and there is no part of the relief where the color of the stone itself shines through (fig. 4). In the accompanying text, Texier stated that his reconstruction was based on observations of a number of reliefs, all depicting the same subject, though he admitted that his reconstructions do not necessarily approximate the original colors.16 Texier’s stunning and original portrayal of this relief sculpture in a chromolithography is today one of the most vivid and iconic reconstructions of translating ancient Persepolitan stone polychromies for modern European audiences. 13 14 15 16
Texier (s. fn. 12), pp. iv-v. Texier (s. fn. 12), pp. 188–189. Texier (s. fn. 12), p. 189. Texier (s. fn. 12), pp. 188–190 and p. 222.
Research on Color Matters
57
4: Charles Texier’s chromolithographic reconstruction of a doorjamb relief from fifth century BCE Persepolis, published in 1852 (Texier 1852: plate CXI ter).
This interesting episode is an attestation of the notion of “scientific” testing of color and polychromy at an early stage of scientific exploration of the polychromies of the ancient Near East, even though many subsequent investigators who dealt with the surface of the decorations on the Takht remained oblivious to the issue. The Frenchman Eugène Flandin (1809–1871), a painter, and his architect-companion Pascal Coste (1787–1879), visited Persepolis only a year after Texier (in December 1840 and early 1841). Their extensive documentation and study of the site did not include any comments on traces of pigments on the standing remains.17 This is all the more remarkable in view of the profession of these two men.
17 Eugène Flandin and Pascal Coste: Voyage en Perse de mm. Eugène Flandin, peintre, et Pascal Coste, architecte […] entrepris par ordre de m. le ministre des affaires étrangères, d’après les instructions dressées par l’Institut, Paris: Gide et J. Baudry, 1851, pp. 134–135.
58
Alexander Nagel
Translating ancient Egyptian and Assyrian polychromies for European audiences Modern archaeological fieldwork in the early first millennium BCE Neo-Assyrian palace complexes at Khorsabad (ancient Dur-Sharrukin), Nineveh (Kuyunjik), and Nimrud (Kalhu) nearby modern Mosul began around the same time as Texier began to communicate the results of his observations at Persepolis, between 1842 and 1855. Paul Èmile Botta (1802–1870), Flandin, Austen Henry Layard (1817–1894), and Victor Place (1818–1875) frequently referred in their reports of seeing traces of paint on the monuments, glazed bricks, painted plaster and other materials related to the ancient painting process.18 In April 1843, Botta announced abundant traces of paint on the monuments at Khorsabad.19 Hand-colored plates accompanied B otta’s earliest publications on Khorsabad in most of which the names of colors were added in pencil.20 Once the stone reliefs were removed and transported to Europe, museum curators soon became aware of the traces of paint, too. In February 1847, Adrien de Longpérier (1816–1882), Conservateur d’Antiquités at the Louvre in Paris, reflected 18 Eugène Flandin: Voyage archéologique à Ninive, 2. Revue des Deux Mondes: 30 Juin 1845, pp. 768–785; Paul-Émile Botta et Eugène Flandin: Monument de Ninive, 5 vols., Paris: Gide and Baudry, 1846–1850; Austen H. Layard. Nineveh and its Remains, Vols. I-II, London: Murray, 1849; Pauline Albenda: The Palace of Sargon, King of Assyria. Monumental Wall Reliefs at Dur-Sharrukin, from Original Drawings made at the Time of their Discovery in 1843–1844 by Botta and Flandin, Paris: Editions Recherche sur les Civilisations, 1986, p. 33; See Nicole Chevalier: Paul-Émile Botta et Eugène Flandin à Khorsabad: un consul et un peintre archéologues. In: Julie Patrier, Philippe Quenet, Pascal Butterlin (eds.): Mille et une empreintes. Un Alsacien en Orient. Mélanges en l’honneur du 65e anniversaire de Dominique Beyer, Turnhout: Brepols, 2016, pp. 99–108; A great number of recent monographs and articles have dealt with the arrival and reception of ancient Near Eastern monuments in nineteenth century Europe and the Americas: Mogens Trolle Larsen: The Conquest of Assyria. Excavations in an Antique Land 1840–1860, New York and London: Routledge, 1996; Mogens Trolle Larsen: The Archaeological Exploration of Assyria. In: Eckhart Frahm (ed.): A Companion to Assyria, Chichester: Wiley-Blackwell, 2017, pp. 583–598; Frederick Bohrer: Orientalism and Visual Culture, Cambridge: Cambridge University Press, 2003; Ada Cohen, Steven Kangas: Assyrian Reliefs from the Palace of Ashurnasirpal II: A Cultural Biography, Hanover: The Hood Museum of Art, 2010; Zainab Bahrani, Zeynep Celik, Edhem Eldem (eds.): Scramble for the Past: A Story of Archaeology in the Ottoman Empire, 1753–1914, Istanbul: SALT, 2011; David Kertai: The Architecture of Late Assyrian Royal Palaces, Oxford: Oxford University Press, 2015; Lucas Petit, Daniele Morandi Bonacossi (eds.): Nineveh. The Great City. Symbol of Beauty and Power. Papers on Archaeology of the Leiden Museum of Antiquities, Leiden: Sidestone Press, 2017. 19 Paul-Émile Botta: Lettres de M. Botta sur ses découvertes près de Ninive, In: Journal Asiatique, 4.2, 1843, pp. 61–72 and pp. 201–214 and Journal Asiatique, 4.3, pp. 91–103 and pp. 424–435; Lettres de M. Botta sur ses découvertes près de Ninive, Journal Asiatique, 4.4, 1844, pp. 301–314. 20 Eleanor Guralnick: New Drawings of Khorsabad sculptures by Paul Émile Botta. In: Revue d’Assyriologie et d’Archéologie Orientale, 95, 2002, pp. 23–56, pp. 28–30; and Eleanor Guralnick: Color at Khorsabad: Palace of Sargon II. In: Paolo Matthiae et al. (eds.): Proceedings of the 6th International Congress of the Archaeology of the Ancient Near East. 5th-10th May 2009, Rome. Vol. 1, Wiesbaden: Harassowitz 2010, pp. 781–792.
Research on Color Matters
59
upon seeing considerable traces of paint on the architectural sculptures from Khorsabad.21 These were the words of Layard on the colors observed on the monuments excavated by the British in the nearby palaces: “[T]he hair, beard, eye brows, eye lids and eyeballs, black; the inner part of the eye, white; the king’s mitre, principally red; the crests of the helmets, blue and red; the heads of arrows, blue; the bows, red; the handles of maces, red; the harnesses of horses, blue and red; sandals, in the oldest monuments, black, edged with red; in those of Khorsabad, striped blue and red; the rosettes in the garlands of winged figures, red; trees at Khorsabad, a blueish green; flowers carried by the winged figures, green, with red flowers occasionally; fire, always red.”22 The modern rediscovery of traces of polychromies on ancient Near Eastern stone monuments itself needs to be read against the debate surrounding ancient Mediterranean and Egyptian polychromies in the first half of the nineteenth century. Antiquarians in London and Paris were rooted in an environment where theoretical discussions of the extent and function of polychromies on Egyptian and Greek sculpture loomed large. Proving the very existence of polychromies itself was a considerable accomplishment, and documenting facts without today’s available color photography or color charts was important, even though not all antiquarians were excited about the original polychromies. Still, in 1845, for instance, in discussing the stone reliefs excavated at Khorsabad, Flandin commented on the rather annoying (“fâcheux”) impression the colorful reliefs would have made on the viewer. Flandin
21 Adrien de Longpérier: Notice des antiquités assyriennes, babyloniennes, perses, hébraïques, esposées dans les galleries Musée du Louvre, Paris: Vinchon, 1854, e. g., p. 29 no. 5: “La barbe, les yeux et les sourcils portent des traces très sensibles de couleur noire et blanche.”; ibid. p. 31 no. 7: “Les yeux et la barbe conservent des traces de couleur”; ibid. p. 32 no. 10: “Les yeux, les cheveux, la barbe, le diadème, la tete de l’ibex et la fleur de lotus sont encore peints.”; ibid. No. 11; p. 33 no. 13; 34 nos. 15 and 16; p. 35 no. 17: “Le diadème est d’un rouge très vif.”; no. 18; no. 19; p. 36 no. 21; p. 37 no. 26; p. 38 no. 29: “La couleur des jambs et des sandals est encore très reconnaissable.”; pp. 38–39 no. 30: “La tête des chevaus est surmontée d’une sorte de crista peinte en rouge, […]”; pp. 40–41 nos. 32–34. 22 Lanyard (s. fn. 18) p. 312. Further comments on paint ibid. p. 306, pp. 309–310; see also the reference by Julian Reade: Nineteenth century Nimrud: Motivation, Orientation, Conservation. In: John Curtis, Henrietta McCall, Dominique Collon, Lamia Al-Gailani-Werr (eds.): New Light on Nimrud: Proceedings of the Nimrud Conference 11th-13th March 2002, London: British Institute for the Study of Iraq and the British Museum, 2008, pp. 1–21, p. 15 n. 2 quoting from an original notebook of Layard: “Bracelets on arms painted black/crossing with red edging/mace handle red/Tiara of king, horse reins and/ornament above red/Handle of dagger below head of/animal-blue/the head a reddish brown/ornament all black pecked with red/the knob or rope near leg, blue/bracelets red. Tassels ditto.”
60
Alexander Nagel
concurred, however, that the “colors” applied on the reliefs in the Assyrian palaces were the same as those applied to the monuments in Egypt.23 Attempts to reconstruct whatever was left of the paints on these ancient palace reliefs in the European cities soon followed. In Nineveh and its Remains, Layard published four lithographic plates, among them one imaginary reconstruction of an Assyrian interior, brightly colored.24 His concept of ancient Near Eastern polychromies was developed together with Owen Jones (1809–1874)25 and James Fergusson (1808–1886), both architects by training. These two men figured prominently in discourse on polychromies at this time. Significantly, the elaborate reconstruction we see in at least one of the reconstructions cannot be taken as a precise illustration of Layard’s direct observations of the archaeological record. The interest in publicizing the color of these great monuments could best be seen in museum and museum-like presentations of the day. Throughout the first half of the 1850’s, Layard, Fergusson and Jones jointly worked in a campaign to influence the British public to think about the use of color and ornament, encouraged for a polychromatic enrichment of built environments. An important part of this campaign was the creation of an installation of painted plaster casts and reconstructions of an ancient Assyrian palace in the Crystal Palace in Sydenham in South London. Though not preserved today, the palace reconstructions at Sydenham serve as a useful illustration of how debate on ancient sculptural polychromies became stuck in a critical moment when documenting the original colors and paints itself would have been crucial. The Crystal Palace, which opened to the public in June 1854, was a private initiative to show recent developments and progress in technology and production.26 The Nineveh Court was an important phenomenon in the history of the reception of ancient Middle Eastern and Egyptian polychromies in the nineteenth century. Located at the foot of two painted casts of two of the four colossal statues of the Nile temple complex of Abu Simbel each measuring 15.5 m and crafted in blocks of plaster “covered in red, yellow and blue house paint ordered by hogshead,” who reminded one 23 Flandin (s. fn. 18), pp. 106–107. 24 Layard (s. fn. 18), vol. 1, p. 2. On this iconic illustration, see: Ada Cohen, Steven Kangas: Inside an Ancient Assyrian Palace: Looking at Austen Henry Layard’s Reconstruction, Lebanon, NH: University Press of New England, 2016. 25 On Jones: Carol Hrvol Flores: Owen Jones: Design, Ornament, Architecture, and Theory in an Age in Transition, New York: Rizzoli, 2006. 26 The literature on the Crystal Palace is vast. See, more recently, Stephanie Moser: Designing Antiquity. Owen Jones, Ancient Egypt and the Crystal Palace, New Haven: Yale University Press, 2012; Kate Nichols: Greece and Rome at the Crystal Palace. Classical Sculpture and Modern Britain, 1854–1936, Oxford: University Press, 2016. For the oriental displays, see Jan Piggott: Palace of the People: The Crystal Palace at Sydenham 1854–1936, London: Hurst, 2004, pp. 109–112.
Research on Color Matters
61
5: Construction of the Nineveh Court, Sydenham, south façade, around 1860.
visitor of “clowns in a circus,”27 it was this Nineveh Court (fig. 5), in which Layard, Fergusson and Jones created their own vision of an ancient Near Eastern palatial polychrome environment. As a side note: it would have been impossible to reconstruct the polychromies of the Abu Simbel sculptures, as they were stripped of their original polychromies, when in October 1837, the Egypt enthusiast Joseph Bonomi (1796–1878), with financial backing from the Englishman Robert Hay (1799–1863), applied two and a half tons of plaster and water in molding onto one of the original heads. The impact of the lime substances used in the early procedures of molding, often prevents us today from learning more about the polychromies of these and other monuments which may have existed back then.28
27 Anon.: School of Sculpture at Sydenham, The Art Journal (London), p. 258. See the perplex comment by Samuel Leigh Sotheby (1805–1861), a Crystal Palace shareholder: “I happened to be at the Crystal Palace when the casts were first being painted […]. At that moment I was pleased with the novelty, and so expressed myself to Mr. Bonomi. The next day, however, I called, telling him I had quite regretted having entertained such an opinion” (Piggott, s. fn. 26, p. 87). The Nineveh Court was situated in the north-western angle of the Crystal Palace; Ian Leith: DeLamotte’s Crystal Palace: A Victorian Pleasure Dome Revealed, London: English Heritage, 2005, figs. 63–65; Piggott (s. fn. 26): esp. pp. 109–112. 28 Alexander Nagel: Colors, Gilding and Painted Motifs in Persepolis: Approaching the Polychromy of Achaemenid Persian Architectural Sculpture, ca. 520–330 BCE, Unpublished Dissertation, University of Michigan, Ann Arbor, 2010. Joseph Bonomi, Nineveh and its Palaces. The Discoveries of Botta and Layard, applied to the Elucidation of Holy Writ. London: Ingram, 1853, pp. 327–328; Bonomi is still a less well studied character in debates on polychromy in the nineteenth century. The Bonomi papers and correspondence are held today at the Cambridge University Library, Department of Manuscripts. A useful summary on practices of plaster casting, without referring to the disastrous consequences on the surface of originals, however: Mari Lending: Plaster Monuments: Architecture and the Power of Reproduction, Princeton: University Press, 2017.
62
Alexander Nagel
The Nineveh Court displayed vibrantly painted casts of architectural reliefs from Nineveh, Nimrud and Khorsabad set into one structure. In this eclectic combination, the exhibition already lacked historical specificity, which was obviously not the aim (as was not the case with the Abu Simbel colossi, either). The colors used in the painted casts were described in detail by contemporary guidebooks. According to one, the “animals […] were in chocolate on buff grounds, or blue on red, or red on blue. The bulls of the portal were deep dull red with black beards and hairs, their mitres blue and yellow.”29 The Nineveh Court stood out as a landmark experiment in the portrayal of a polychrome ancient palatial environment for the general public. Notwithstanding the excitement the polychrome casts generated, they also met with harsh criticism. For some contemporaries, the reconstructions gave the impression of a “marvelous peepshow, a contemporary toy theatre,” in that the coloring seemed “garish to most p eople of refined taste.” The British Historian Jan Piggott assumes that the colors would have made Nineveh “absolutely ferocious,”30 and an anonymous reviewer in the London Quarterly Review in April 1855 noted scathingly, that “[…] one shudders to think of the generations who groaned beneath the yoke of these sanguinary reds, implacable blacks, and cruel blues [...]. There is only one class of visitors who will discover he slabs in the Nineveh Court to be really relieved at all, and that is the blind.”31 While the number of visitors in the Crystal Palace exponentially rose and the number of artists who used the sculpture galleries in the British Museum decreased dramatically between 1849 and 1870, the aesthetic debate focusing on “taste” overlaid discussion about how much was painted in the original reliefs. It is significant that judgments on the “taste” of the Assyrians soon followed. In 1854, Richard Westmacott wrote, referring to Assyrian monuments, that “[…] for the further we go back to barbarism in art, or to the infantry of art, the more surely we meet with colored sculpture.”32 In reviewing the history of polychrome casts of ancient Near Eastern palace reliefs in the second half of the nineteenth century, we can simply state that they were an important feature in museum displays.33 Only very few polychrome casts of 29 Piggott (s. fn. 26), p. 111. 30 Piggott (s. fn. 26), pp. 75–76 and p. 112. Piggott observed that Jones “had stamped his ideas and taste too heavily on the courts and observes a ‘color fever‘, […] ‘polychromatic’ runs like a motif, or even like a party line, throughout the company’s guidebooks to the Courts.” 31 Anon.: Review of the Nineveh Court. In: London Quarterly Review, April 1855, pp. 163–164. 32 Richard Westmacott jr.: On Coloring Statues. In: Archaeological Journal, 12, 1854, pp. 22–46, particularly p. 28 [emphasis in original]. 33 e. g., Ingeborg Kader: ‘Täuschende Spielereien‘. Kolorierte Abgüsse im 19. und 20. Jahrhundert. In: Vinzenz Brinkmann and Raimund Wünsche (eds.): Bunte Götter, Die Farbigkeit antiker Skulptur. Eine
Research on Color Matters
6: Assyrian Room, Glyptothek, Munich, after 1865, unknown photographer.
63
7: Assyrian Room, Halls of the Ancients, Washington, D. C., around 1900.
ancient Near Eastern palace reliefs are preserved today, such as those of the M useum für Vorderasiatische Kunst in Berlin.34 An Assyrian Room (fig. 6), which opened in 1865 in the Glyptothek in Munich, featured polychrome colossal cast Lamassus and painted scenes of Layard’s Assyria alongside seven original limestone slabs from Nimrud King Ludwig had acquired from London in 1863.35 After the death of Leo von Klenze early in 1864, it was Georg von Dollmann (1830–1895), who oversaw Ausstellung der Staatlichen Antikensammlungen und Glyptothek, München, 16. Dezember 2003 bis 29. Februar 2004, 15. Juni – 5. September 2004. Munich: Glyptothek 2004, pp. 235–246. 34 Berlin, Vorderasiatisches Museum, Plaster Cast of original relief VA 962, Ninth century BCE Northwest Palace at Kalhu/Nimrud. 35 Ignaz Gaugengigl: Beschreibung der Merkwürdigkeiten im Assyrischen Saale der Königlichen Glyptothek zu München, München: Fritsch, 1870; Heinrich Brunn: Beschreibung der Glyptothek König Ludwig’s I. zu München (5th ed.), München: Ackermann, 1887; The Glyptothek was already home to a painted miniature plaster version of the pediments of the temple of Aphaia at Aigina, ca. 500 BCE, by the German architects Leo von Klenze (1784–1864) and Joseph Ohlmüller (1791–1839), to which a new version was given by Adolf Furtwängler (1853–1907) in the early twentieth century. Klenze himself passed away in January 1864 before the Assyrian room was officially opened. The harsh criticism of the polychrome reconstruction of the Lamassu casts is still evident in Britta Schwahn: Die Glyptothek in München. Baugeschichte und Ikonologie, München: Kommissionsverlag UNI-Druck, 1983, pp. 200–201.
64
Alexander Nagel
the construction and polychromy of the two lamassu statues in the Assyrian Room. In the Munich case, critics such as archaeologist Paul Wolters (1858–1936), who would later become director of the Glyptothek described the installation as “arg missglueckt.”36 Finally, a late nineteenth century construction of polychrome plaster casts in Washington, D. C., featured entirely painted Lamassus and a partly painted plaster cast of a relief from the Hall of 100 Columns at Persepolis (fig. 7). This relief depicting a jamb of a door in the Hall of 100 Columns of Persepolis became part of a throne installation on which a dummy depicted an imaginary Great King, Xerxes.37 Preliminary conclusions Even though most of the painted plaster casts attempting to reconstruct ancient sculptural polychromies are no longer preserved today, we can appreciate the palette and patterns of polychromatic application from contemporary descriptions and pictorial documentation. It is important to stress, however, that these polychrome reconstructions were artistic creations in their own right. Their contribution to our modern understanding of ancient sculptural polychromies is perhaps rather limited, as a more accurate “measuring” and “mapping” of ancient sculptural polychromies was only possible with the integration of better scientific methods used by chemists and in conservation science since the second half of the twentieth century. Yet, the nineteenth century attempts are fascinating testimony to the challenges in documenting and defending reconstructions of ancient color and polychromy which will continue to inspire future generations of students of the ancient Near East.
36 Paul Wolters: Neues von der Glyptothek. In: Münchner Neueste Nachrichten, August 2, 1921: 321; Schwahn (s. fn. 35), p. 201 uses the phrase “deutlicher Niveauabfall” (“noticeable lapse in standards”). 37 Franklin Webbster Smith: The Halls of the Ancients constructed on Pennsylvania Avenue, Washington, Washington, D. C., 1897.
Esther Ramharter
Do Signs Make Logic Colored? Tendencies Around 1900 and Earlier
Color plays a role in philosophy and in logic as well: What meaning does color have for logic, and has this changed over time? David Batchelor, in his book Chromophobia,1 diagnoses a repression of color in Western culture because color has been considered a permanent threat, in particular to rationality. With respect to the relation of the two dichotomies colors/black-and-white and male/female, Ulrike Boskamp has developed a more differentiated view in her contribution to this collection. Does logic have its own privileged relationship with color, or is this connection merely a special case of the (perhaps hostile) relationship between color and rationality? The main emphasis of this paper, however, is on the role that (written) signs play in the relation between color and logic. Rationality and logic There has been a close connection between rationality and logic since antiquity, with many parallel discussions surrounding this subject: scholars have asked whether there is one universal rationality or many rationalities, and monism versus pluralism has appeared as a topic of ongoing debates in logic. Authors queried whether the field of psychology instead of philosophy is the appropriate place for both studies of rationality and of logic. The respective possibility to include contradictions is a matter that has bothered very different philosophers. And although these issues have – at least partially – been discussed independently for the two different concepts, rationality and logic, the connection between them is ultimately inescapable. According to perhaps the most commonly held conception of the relationship between logic and rationality, logic is viewed as a part of rationality. One early example of the development of such a conception comes from Aristotle, who held that logic, i. e., syllogistics or dialectic, is how we reason correctly. But since reasoning requires a starting point, something must be responsible for the beginning of this ~ process. This is the intellect (νους/nous).2 Hence, intellect and reasoning combine to make what one might term “rationality.” What this understanding of rationality 1 David Batchelor: Chromophobia, London: Reaction Books, 2000. 2 See Aristotle: Analytica Posteriora 100 b, 5–17.
66
Esther Ramharter
shares with the ancient definition of ratio (lat. ratio, gr. λόγος/logos), according to which ratio is regarded as a numerical proportion,3 and other symbolisms is the advantage of a clear separation and order at the expense of the loss of some information: Ratio creates a new entity – the relation between the concepts – and therefore generates a recognizable difference, while at the same time suppressing a difference – the presumed difference in kind (genus) of the concepts. Regardless of whether one views logic as a part of rationality or the whole of rationality, logic can be regarded as rationality in its purest form: indeed, logic is traditionally claimed to be free from empirical influences or ingredients. On the other hand, logic has been a system of written symbols ever since Aristotle. It is therefore the most sensuous part of rationality, which makes it a potential place for colors.4 Color versus rationality Motifs of opposition between color and rationality abound in history. Hermes hands the black-and-white5 plant, Moly, over to Ulysses to save him from being turned into a pig by Circe – in other words, to keep him rational. Starting in the 5th century B. C., theories of colors presented black and white as prior to colors; as being their basis, their ultima ratio. According to Gilles Deleuze, Lucretius describes colors as simulacra of the surface only.6 Color is deemed inferior in several respects. Charles Blanc’s Grammaire des arts du dessin, published in 1867, can serve as a later example: “Intelligent beings have a language represented by articulate sounds. […] Inorganic nature has only the language of colour. It is by colour alone that a certain stone tells us it is a sapphire or an emerald.”7 Sybille Krämer places this degradation of color in the broader context of a contempt of the visual in general: “The European ocularcentrism is based on a dissociation of corporeal and m ental eye: the eye of the mind ‘sees’ the better the more the physical eyes remain closed.”8 3 See Euclid: The Thirteen Books of the Elements. Translated with introduction and commentary by Thomas L. Heath, vol. 2, New York: Dover Publications, 1956, Book V. 4 In this paper, I always proceed from the assumption that there is neither a universal rationality which is not shaped by culture, nor is there a universal (i. e., not culturally molded) way to see color(s). 5 Black and white are not understood as colors in this case, or at least they are granted a special status that allows them to be contrasted with all other colors. 6 Gilles Deleuze: Logique du Sens, Paris: Les Editions de Minuit, 1969, p. 316. 7 Charles Blanc: The Grammar of Painting and Engraving, New York: Hurd and Houghton, 1874, p. 5. 8 Sybille Krämer: Textualität, Visualität und Episteme. Über ihren Zusammenhang in der frühen Neuzeit. In: Renate Lachmann, Stefan Rieger (eds.): Text und Wissen. Technologische und anthropologische Aspekte, Tübingen: Gunter Narr Verlag, 2003, pp. 17–27, p. 25, transl. E. R.
Do Signs Make Logic Colored?
67
Not only are colors themselves disqualified in the literature, but the term “color” or color-related terms are also used to disqualify (particularly to stigmatize aspects of the irrational). The following example comes from the remarks about Frege made by Anthony Kenny: “Frege distinguishes between a thought and what he calls the colouring of its expression. Scientific language as it were presents thoughts in black and white; but in humane disciplines sentences may clothe thoughts in colourful garb […].”9 Remarkably, both Frege and Kenny in his discourse concerning Frege use the same kind of metaphor: Frege calls the unscientific aspects of language “coloring,” but it is Kenny who goes one step further and raises the metaphorical use of color terms to the level of describing the scientific disciplines. By the way, Kenny gives Frege’s view an interesting twist in that the “colourful garb” which “may” clothe thoughts vaguely indicates a positive attitude. In short: it is a common topos to contrast color and rationality, and to use color as a metaphor to contrast rationality and nature, poetry, sensuousness and the like – not only on very different levels, but also in very different circumstances. Logic as visual representation Logic may be regarded as a part of rationality, but it can also be viewed as the written version of (a part of ) rationality. In their book about books, Güntner and Janzin stated: “Without systems of writing, it is unlikely that formal logical thought would have been developed in the first place.”10 This viewpoint has not been without its controversy. To name just one example, the revival of Thomism – as advocated by thinkers such as Jacques Maritain – raised certain objections. Maritain argued that logic is concerned with processes of thought leading to truth, and that it is not concerned with signs.11 However, logic as a symbolic representation of the process of correctly reasoning has always been the mainstream view, and so it remains. If logic is understood as purely abstract (as consisting of thoughts only), it is – by definition – colorless. In the following, I will therefore study logic as a system of signs. Logic, then, is abstract on the one hand and sensuous on the other. The use 9 Anthony Kenny: Frege. An Introduction to the Founder of Modern Analytic Philosophy, Oxford: Blackwell Publishers, 2000, p. 183. 10 Joachim Güntner, Marion Janzin: Das Buch vom Buch. 5000 Jahre Buchgeschichte, Hannover: Schlütersche Verlagsbuchhandlung, 1995, p. 13, transl. E. R. 11 See Jacques Maritain: An Introduction to Logic, London: Sheed & Ward, 1947, pp. 222–224.
68
Esther Ramharter
of colors would indeed be possible given the sensuous nature of logic. But, generally speaking, contemporary books and papers on logic do not contain any colors (didactical explorations may be seen as a small, but perhaps important, group of exceptions). The – partly external – reasons for this absence of color are manifold, and a comprehensive exploration of these reasons would go beyond the scope of this paper.12 In what follows, I will further confine myself to sketching the historical development of the occurrence of color in logic and the discourse about logic (by selecting some important examples). This will, sometimes implicitly, show some of the concrete reasons for the appearance or absence of color in logic, as they were effective at certain times. Highs and lows of colors in pre-modern logic Logic was only formalized to a very minor degree prior to the second half of the nineteenth century, hardly any more so than Aristotle’s syllogistics. Aristotle used capital letters to designate predicates, and this was also the case in a large strand of the logical tradition from the Middle Ages through to the foundation of modern logic by Frege, Boole and others. This kind of approach uses ordinary writing as the medium. A second strand of this tradition views logic as division of genus into species (διαίρεσις/diairesis) and uses trees as representatives of logic. A third strand of the tradition involves the use of combinatorial diagrams. Provided that logical texts are just plain texts, there is some overlap between the question of whether colors are or can be used and the issue of colors in print or handwriting. The history of printing has had its highs and lows in terms of its relation to colors.13 In particular, the use of color to ascribe different roles or functions to parts of the text – as can be found in missal books and breviaries, for instance (fig. 1) – might have provided a basis for using color in logic. But in fact, this did not happen. Trees have been representatives of logic ever since the use of division (διαίρεσις) in antiquity, and we continue to use the method of truth trees (also called semantic tableaux) to the present day.
12 The lack of color in logic may, for instance, be regarded as a special case of the colorlessness of scientific objects. Logic books might simply be seen as an instance of “ordinary” printed books. One can argue that it is the opposition between line and color that makes the systems of symbolic logic colorless. Independence from matter can also be considered as a reason to exclude colors from logic. For more on these and other reasons see Esther Ramharter: Eine Frage der Farbe. Modalitäten des Zeichengebrauchs in der Logik, Berlin: Parerga, 2011, pp. 342–389. 13 See Ramharter (s. fn. 12), pp. 360–373.
Do Signs Make Logic Colored?
1: In this Breviarium Romanum from 1891, the black lines are to be read out loud, whereas the red lines are either explanations or instructions to be carried out.
69
2: Tree of first order calculus.
3: Sefiroth tree.
The truth tree in figure 2 and the tree of life in figure 3 both clearly demonstrate the similarity and the fundamental difference between the abstract structure used in modern logic and the representation of relations of concepts by trees. They share a structure, but the “real tree” also shows parts that we are accustomed to seeing in color. There is thus a twofold connection between the trees of logic and color. Sometimes the logical trees themselves were colored (fig. 4). On the other hand, in periods when the representations of the trees were depicted in a more realistic way, the association with the colors of real trees became stronger, even in cases where the logical trees themselves were not colored (fig. 5).14
14 For the history of these trees see Steffen Siegel: Tabula. Figuren der Ordnung um 1600, Berlin: Akademie-Verlag, 2009.
70
4: Logical Tree, ninth century.
Esther Ramharter
5: Arbor naturalis et logica, original ca. 1305 by Lullus.
There was also a convergence of formal structures and pictures of living trees in the opposite direction: Leonardo da Vinci geometricized trees in the fifteenth century,15 claiming that the branches are structured according to geometrical proportions. Even if the trees in logic are not drawn in color, real trees are colored.16 Although this may be a weak connection between color and logic, it is a connection all the same. Color can only be “wished away by pure thought or washed away by pure form,”17 as David Batchelor says in his book Chromophobia; however, in the case of trees this seems particularly hard to do, as we have such a strong association with green leaves representing life, red apples representing fertility and fruitfulness and so on. The closeness of color and rationality in the concretization of the trees seems at least remarkable if we remember the black-and-white plant, Moly, and, in a more general sense, the hostile relation between rationality and color. 15 See Siegel (s. fn. 14), pp. 57–58. 16 By saying this (and the same applies for other similar statements made here), I am certainly not committing myself to any kind of realist claim concerning colors; I am only referring to the way in which we usually consider and express matters. 17 Batchelor (s. fn. 1), p. 70.
Do Signs Make Logic Colored?
71
6: Diagrams of Lullus.
In the thirteenth century, Lullus and his followers developed the third strand of logical tradition – in the ars inveniendi or the ars demonstrativa – where combinatorial diagrams were used to represent logical connections. These diagrams were also sometimes colored (fig. 6), in order to make them more perspicuous. In sum, while there were highs and lows for colors in the history of logic prior to the dawn of modern logic, there was no discussion whatsoever of the role of colors in logic before the second half of the nineteenth century. Colors in the discussion of logic around 1900 Modern logic was developed and logic was formalized during the second half of the nineteenth century. A discourse, albeit a scattered one, about colors in the context of logic started to emerge around this same time, the precondition being that a process of reflecting the role of signs in logic was already underway. In 1854, the mathematician and logician George Boole wrote: “A sign is an arbitrary mark, having a fixed interpretation, and susceptible of combination with other signs in subjection to fixed laws […].”18 I will present two examples in the following. One is part of the mainstream of logic, to be found in the work of Gottlob Frege; the other one is from Frege’s contemporary Charles S. Peirce, who is an outsider to the science of logic. Whereas in the Frege-example color comes into play in a metaphorical use, Peirce took serious the option of coloring symbols. What they have in common is their willingness to accept color as a topic in the discourse about logic. 18 George Boole: An Investigation of the Laws of Thought, New York: Dover Publications, 1958, p. 25.
72
Esther Ramharter
7: Frege’s notation for “if p then q.”
“Coloring” to control the irrational – Gottlob Frege’s “Färbung” Gottlob Frege is said to be the founding father of modern logic, which makes his views particularly important for the purposes of this paper. Frege probably started a tradition of how to see matters in this context, and he is regarded as the first scholar to have constructed a formal system of first-order logic. However, the form of representation he used differed considerably from those used in the present day. Instead of using formulas in the usual sense – with constants, variables, quantifiers and connectives joining some elements – he used parting lines to represent the conditional “if p then q” (fig. 7). He also employed a small vertical line as the sign of negation. He stressed the advantages of this kind of notation: for instance, to aid lucidity (Übersichtlichkeit), and to make use of the two-dimensions of the sheet.19 Hence, Frege was aware of the possibility, or the need, for visual means to be exploited optimally. Still, colors do not play any role in this form of notation. Color gains importance not within Frege’s logical calculus, but in his considerations about logic and language. In addition to his famous distinction between sense (Sinn) and reference (Bedeutung), Frege picks up a metaphorical use of the word “coloring” (Färbung) and uses it as a technical term to separate that which can be treated within logic from everything else that is mediated via language.20 “We can now recognize three levels of difference between words, expressions, or whole sentences. The difference may concern at most the ideas, or the sense but not the reference, or, finally, the reference as well. With respect to the first level, it is to be noted that, on account of the uncertain connexion of ideas with words, a difference may hold for one person, which another does not find. […] To the 19 See Gottlob Frege: Begriffsschrift (1879), Hildesheim, Zürich, New York: Georg Olms Verlag, 1998, pp. 103–104. 20 See Klaus Speidel: Représentation, Coloration et Éclairage dans la Philosophie de Langage de Gottlob Frege. In: Jocelyn Benoist (ed.): Propositions et États de Choses. Entre Être et Sens, Paris: J. Vrin, 2006, pp. 147–170. There is not a great deal of secondary literature available on the subject of coloring.
Do Signs Make Logic Colored?
73
possible differences here belong also the colouring and shading which poetic eloquence seeks to give to the sense. Such colouring and shading are not objective, and must be evoked by each hearer or reader according to the hints of the poet or the speaker.”21 The active and passive forms of a verb are different colorings. The following propositions have the same Sinn and Bedeutung, yet they differ in terms of their coloring: “All men are mortal.” “Every man is mortal.” “If something is a man, it is mortal.”22 Conjunctions such as “although” are another instance of this.23 Although Frege provides a number of examples, he does not offer much in terms of a general definition of coloring. What seems clear is that coloring is the thing that remains if the thought is identified: “[W]e must not fail to recognize that the same sense, the same thought, may be variously expressed; thus the difference does not here concern the sense, but only the apprehension, shading, or colouring of the thought, and is irrelevant for the logic.”24 It is not clear what coloring entails, nor if – or in what sense – it is subjective.25 It is also unclear which entities are said to be colored (expressions, thoughts, propositions etc.). The only thing that seems to be clear is that coloring is undesirable and must be abandoned if thinking is at issue:
21 Gottlob Frege: On Sense and Reference. In: Peter Geach, Max Black (eds.): Translations from the Philosophical Writings of Gottlob Frege, Oxford: Blackwell, 1952, pp. 56–78, pp. 60–61. 22 Gottlob Frege: Logical Generality [Not before 1923]. In: Posthumous Writings. Edited by Hans Hermes, Friedrich Kambartel, Friedrich Kaulbach with the assistance of Gottfried Gabriel and Walburga Rödding, Oxford: Blackwell, 1979, pp. 258–262, p. 259. 23 Frege (s. fn. 21), pp. 73–74. Frege uses the term “illuminate” here. It is unclear whether coloring and illuminating (beleuchten) are used synonymously or as different properties. See Speidel (s. fn. 20); Ramharter (s. fn. 12), pp. 144–146. 24 Gottlob Frege: On Concept and Object. In: Posthumous Writings (s. fn. 22), pp. 87–117, p. 96. 25 There is a discussion in the secondary literature addressing whether ideas (Vorstellungen) – and coloring as a sort of idea – really are subjective for Frege. See Michael Dummett: Frege. Philosophy of Language, London: Duckworth, 1973, pp. 83–89; Ramharter (s. fn. 12), pp. 156–158; Joan Weiner: Frege in Perspective, Ithaca, London: Cornell University Press, 1990, p. 175.
74
Esther Ramharter
“If the visual idea of a rose is associated with the idea of a delicate scent and to these are added the auditory ideas of the words ‚rose‘ and ‚scent‘, as well as the motor ideas associated with uttering these words, and if we go on and on heaping associations upon associations until the most complex and elaborate idea is formed, what purpose does it serve? Do we really think we should have a thought as a result?”26 This “abandoning” is a process: “A thought which to begin was only suggested by an expression may come to be explicitly asserted by it. And in the period in between different interpretations will be possible. […] In the present context the only essential thing for us is that a different thought does not correspond to every difference in the words used, and that we have a means of deciding what is and what is not part of the thought, even though, with language constantly developing (bei der organischen Natur der Sprache), it may at times be difficult to apply.”27 Frege realized that there is a development of thought (moreover, of science), but that there are also constraints to this development.28 These constraints could in principle be regarded as part of a general conception of rationality. If Frege had a conception of rationality (of which I am not sure),29 it may well entail such methodological issues. For instance, Frege does not take into consideration any kind of practical rationality, yet without doubt, logic and the limitations it imposes would be key to a Fregean conception of rationality.
26 Gottlob Frege: Logic [1897]. In: Posthumous Writings (s. fn. 22), pp. 126–151, pp. 144–145. 27 Frege (s. fn. 26), p. 141. 28 “The natural course of events seems to be as follows: what was originally saturated with thought hardens in time into a mechanism which partly relieves the scientist from having to think. […] I should like to compare this to the process of lignification. Where a tree lives and grows it must be soft and succulent. But if what was succulent did not in time turn into wood, the tree could not reach a significant height. On the other hand, when all that was green has turned into wood, the tree ceases to grow.” (IV/1 [xv/1] Frege to Hilbert 1.10. 1985. In: Gottlob Frege: Philosophical and Mathematical Correspondence. Edited by Gottfried Gabriel, Hans Hermes, Friedrich Kambartel, Christian Thiel, Albert Veraart. Abridged for the English edition by Brian McGuinness and translated by Hans Kaal, Blackwell: Oxford 1980, pp. 32–35, p. 33.) Frege admits that any progress in science needs colored expressions – it needs the growing green parts – , but his logic abandons everything that is “colored.” 29 Frege certainly has a conception of science, logic, language, and thought, but he never explicates a conception of rationality.
Do Signs Make Logic Colored?
75
Hence, as far as Frege is concerned, there is nothing in poetic or otherwise “colored” aspects of language that could count as being somehow bound by rationality. By using the term “thought” to refer to things that are subject to logic, he appropriates a great deal of our intellectual activities to the rational domain, apparently not leaving much room for anything else. Despite Frege’s efforts to separate everything that cannot be dealt with in logic, there is still something that the domain of logic and the colorings share: the dominance of the paradigm of the visual. Logic is a written, and therefore visual, system of signs. In this way, aspects of language that cannot be ruled by logic must at least be controlled by metaphors that come from the domain of colors, not, for instance, from the domain of sounds. It is the visual sense and the mental eye (not the mental ear) that are priviliged epistemic instruments since the European enlightenment. This disposition has at least one of its origins in ancient rhetorical tradition:30 linguistic means used to highlight certain aspects of a thing in order to deceive some~ one about its value were called color (lat. color or gr. χρωμα/chroma).31 Colors make a difference – Peirce’s logical system Unlike Frege, Charles S. Peirce is an outsider with regard to logic. He is also an outsider insofar as he places a special emphasis on the form of representation in logic. Frege also hints at the importance of using good notation, but it does not constitute an essential feature of a logical system for him – which it is, as we will see, for Peirce. Peirce is perhaps the key point of reference when attempting to pinpoint a place for colors in logic. He explicitly calls for colors in logic, suggesting that color should be used to represent possibility and necessity in modal logic. The simple idea is to color the sheet on which the logical formulas – or “graphs,” as Peirce calls them – are written: “[T]he Mode of Tincture of the province […] shows whether the Mode of Being which is to be affirmatively or negatively attributed to the state of things described is to be that of Possibility, when Color will be used; or that of Intention, indicated by Fur; or that of Actuality shown by Metal.”32 30 Another origin can be found in the optical and physiological research carried out in his time, see David Charles McCarthy: Optics of Thought. Logic and Vision in Müller, Helmholtz, and Frege. In: Notre Dame Journal of Formal Logic, 41, 2000, pp. 365–378. 31 See Michael Astroh: Anschauung, Begriff und Sprache in Freges Frühschriften. In: Gottfried Gabriel, Uwe Dathe (eds.): Gottlob Frege. Werk und Wirkung, Paderborn: mentis, 2000, pp. 103–122, p. 104. 32 Charles S. Peirce: Prolegomena to an Apology for Pragmaticism. In: The Monist, 16, 1906, pp. 492– 546, p. 527.
76
Esther Ramharter
8: Fig. 1 in Peirce Prolegomena.
Rather than simply mentioning this in a casual sense, Peirce develops an entire modal logical system on this basis.33 He also goes to great efforts to describe the application of colors: “Every part of the exposed surface shall be tinctured in one or another of twelve tinctures. These are divided into three classes of four tinctures each, the class-characters being called Modes of Tincture, or severally, Color, Fur, and Metal. The tinctures of Color are Azure, Gules, Vert, and Purpure. Those of Fur are Sable, Ermine, Vair, and Potent. Those of Metal are Argent, Or, Fer, and Plomb. The Tinctures will in practice represented as in Fig. I [here: fig. 8].”34 This simultaneously demonstrates a high-point and a low-point in terms of the inclusion of colors in logic: Peirce definitely introduces colors in his Existential Graphs – his most elaborate system of logic – , but at the same time, facing the
33 See Christian Gottschall, Esther Ramharter: Peirce’s Search for a Graphical Modal Logic (Propositional Part). In: History and Philosophy of Logic, 32, 2011, pp. 153–176. 34 Peirce (s. fn. 32), p. 526–527.
Do Signs Make Logic Colored?
77
obvious troubles when trying to publish a paper including colors, he dismisses this very idea; heraldic hatchings have to replace colors “in practice,” yielding a de facto absence of colors in his logic, as is the case in all common systems. Nevertheless, Peirce at least expressed the idea of using colors, and the reasons he offers are interesting. In his – pragmatistic – understanding, scientists always experiment with diagrams,35 which should be mappings of circumstances and as iconic as possible in their form.36 A model can be used to reconstruct the course of a car accident in a court hearing. Similarly, the chemist can draw structural formulas of actual molecules, manipulate the diagrams on a sheet of paper, and then finally predict something about the real molecules. According to Peirce, the iconicity of the diagrams is the basis on which these procedures work. Logic, in particular, deals with diagrams of thoughts. But how are we to understand that these diagrams should map the thoughts iconically – that is, via similarity? It would go beyond the scope of this article to offer a general answer to this question;37 however, we immediately face an obvious problem with respect to color: Thoughts are not colored (one might argue that although thoughts are not colored, they may be thoughts of colored things. But the task of logic is not to represent things, but, as indicated in the first quotation, relations between states of affairs). Worse still, modality is not even “thinkable.” Peirce is well aware of this fact: “Still, I confess I suspect there is in the heraldic representation of modality as set forth […] a defect capable of being remedied. If it be not so, if the lack of ‘pictorialness’ in the representation of modality cannot be remedied, it is, because modality has, in truth, the nature which I opined it has ([…] Modality is not, properly speaking, conceivable at all, but the difference, for example, between possibility and actuality is only recognizable much in the same way as we recognize the difference between a dream and a waking experience […].)”38
35 Charles S. Peirce: Writings of Charles S. Peirce, ed. by Peirce Edition Project, Bloomington (Indiana): Indiana University Press, 1982, 1984, 1986, 1989, 1993, vol. 6, p. 37. 36 See Charles S. Peirce, Collected Papers, ed. by Charles Hartshorne, Paul Weiss, Cambridge (MA): Harvard University Press, 1961–1979, 3rd edition, 4.531. [In the following reference to Peirce’s papers will be designated CP followed by volume and paragraph number.] 37 Diagrams should somehow mirror the way we think, they should – to give a very simple example – not contain any element that has no correspondent in the thoughts. See Ramharter (s. fn. 12), pp. 208–211. 38 Peirce CP (s. fn. 36), 4.553, fn 1.
78
Esther Ramharter
The situation in logic is thus: Colors cannot be iconic representations of thoughts because thoughts are not colored, and colors cannot be iconic representations of modality because modality cannot even be thought. Yet, there is still one function that colors can – iconically – fulfill: the function of distinguishing. We “recognize the difference.” Colors offer a greater variety of immediately discriminable states. Perhaps an echo of this is discernible in Peirce’s rather self-confident evaluation of his work: “My analyses of reasoning surpass in thoroughness all that has ever been done in print, whether in words or in symbols – all that De Morgan, Dedekind, Schröder, Peano, Russell, and others have ever done – to such a degree as to remind one of the difference between a pencil sketch of a scene and a photograph of it.”39 Exceeding Peirce’s considerations, it seems notable that not only are thoughts colorless, but they also represent the only way of eliminating colors. Batchelor writes: “Even night fails to shroud or abolish colour entirely, as for many of us colour seeps into our dreams. Perhaps that is the point: the other that is colour can only be imagined away. And this may be one reason for all the attention given to it in certain types of philosophy or art, in certain theories of art or environments, or in certain kinds of stories. Because it is only in these realms that colour can be fully and finally eliminated.”40 In the absence of light, dreams – and only dreams – may bring in colors. But even within dreams, symbols usually remain uncolored, as Aldous Huxley writes: “[I]n most dreams the symbols are uncoloured. Why should this be the case? The answer, I presume, is that, to be effective, symbols do not require to be coloured. The letters in which we write about roses need not to be red, and we can describe the rainbow by means of ink marks on white paper. […] What is good enough for the waking consciousness is evidently good enough for the personal subconscious, which finds it possible to express its meaning through uncoloured symbols.”41
39 Peirce CP (s. fn. 36), 5.147. 40 Batchelor (s. fn. 1), p. 70. 41 Aldous Huxley: The Doors of Perception and Heaven and Hell, London: Vintage, 2007, p. 60.
Do Signs Make Logic Colored?
79
Symbols, and a fortiori logical symbols, do not need colors; indeed, they provide the only opportunity to keep colors away. We are thus faced with the following situation: symbols are the only place where colors can be eliminated (because they are not needed here), and symbols are the only place where color could be incorporated within rationality. This is exactly the point at which Peirce wants colors to be brought in. Obviously, the crucial question that arises here is whether colors are needed in logic. Even if the difference between modalities is only akin to the difference between waking experience and dream, and should not, therefore, be a difference concerning color, it is still a difference. This is not the same kind of difference as the difference between propositions with distinct meanings; hence, it calls for special representation. Peirce tried to do justice to this idea. It is not the colors, nor the property of being colored that is iconically mapped onto the existential graphs; it is a distinction (in modality) that has as its “image” another distinction (in color). Peirce’s tinctured system of logic is the last appearance of color in logic – and the discourse about logic – that is of some relevance. Conclusion Does logic have its own relation to color or is this relation merely an instance of the (alleged) opposition between rationality and color? Logic is simultaneously the most abstract and – as a system of written signs – the most sensuous part of rationality; it is its sensuous nature that makes it a possible entry point for colors. Hence, logic is a field where the opposites of rationality and color can be seen in a concrete contest. The meaning of color for logic has changed over time. With various kinds of visual representations in use before the second half of the nineteenth century, colors did occasionally appear in texts on logic during this time. The development of modern logic led to a certain amount of reflection on the meaning of signs for logic, especially the meaning of colored signs. This would have provided a chance for colors to be included in logical systems. Peirce attributed a positive value to the use of colors in logic, whereas Frege wanted to exclude anything “colored” from logic – nevertheless, all his efforts stand under the paradigm of the visual. And yet the logical texts by both these philosophers and by the logicians who would succeed them ultimately ended up the same: colorless.
Michael Friedman
Coloring the Fourth Dimension? Coloring Polytopes and Complex Curves at the End of the Nineteenth Century
During the second half of the nineteenth century, various attempts were made to visualize the fourth dimension.1 Starting in the 1850s, four- and n-dimensional spaces were taken into serious mathematical consideration. This prompted questions regarding the appearance of four-dimensional mathematical objects, since – to state the obvious reason – our senses are limited to the three-dimensional world.2 One of the famous books, which popularized four-dimensional geometry, was the 1884 novella Flatland: A Romance of Many Dimensions. The book, published by Edwin A. Abbott, became known not only for its criticism of Victorian society but also for the treatment of higher dimensions. But beyond such popularization, the visualization of four-dimensional objects was problematic for mathematicians. In order to understand this, we need only to consider the case of three-dimensional objects. In the three-dimensional (Euclidean) space, one can find five convex regular polyhedra3 (i. e., the five Platonic solids, such as the cube or the tetrahedron), and present them, as Dürer systematically did in the fourth book of Underweysung der Messung, by unfolding them onto a plane, obtaining a two-dimensional net, which may later be re-folded to obtain once more the three-dimensional polyhedra. In a four-dimensional space, one can find six convex regular four-dimensional polytopes. Ludwig Schläfli proved the existence of these polytopes in 1852,4 but the question remained: how one can visualize them? Several ways were offered to deal with this problem. Schläfli, for example, ignored the question completely, and did not draw a single sketch. Other mathematicians, for example Alicia Boole 1 See: Klaus Volkert: In höheren Räumen: Der Weg der Geometrie in die vierte Dimension, Berlin: Springer, 2018. For the usage of color in mathematics (and in particular logic), see for example: Esther Ramharter: Eine Frage der Farbe. Modalitäten des Zeichengebrauchs in der Logik, Berlin: Parerga, 2011. Although during the first half of the twentieth century, visualization techniques were more marginal due to the rising of formalism and the crisis of “intuition” (“Anschauung”; a crisis which was much more one of visualization), starting the end of this century, one can observe a flourishing of computer-based visualizations of mathematical objects, most of them are colored. See: Georg Glaeser, Konrad Polthier: Bilder der Mathematik, Berlin: Springer Spektrum, 2010. 2 During the nineteenth century, when the fourth dimension was mathematically discussed, one did not mean the dimension of time, but rather a space which has four coordinates (in the same way the “normal” space where we live in has three coordinates: x, y and z, denoting the three dimensions). 3 A convex body is a body in which no line segment between two points on the boundary of it ever goes outside this body. 4 Ludwig Schläfli: Theorie der vielfachen Kontinuität, Basel: Springer, 1901 [1852].
82
Michael Friedman
1A: Hinton’s colored cubes. B: A colored net of a four-dimensional cube; the model was still sold in 1911, as can be seen in the catalog of Schilling.
Stott and Charles Howard Hinton, had a different approach. In 1888, Boole Stott contributed to Hinton’s book A New Era of Thought, describing different visual ways to grasp the fourth dimension.5 In Hinton’s book, one of the ways to visualize a four-dimensional cube was via multicolored cubes (fig. 1a). Another popular way was to visualize and construct material models of four-dimensional regular polytopes. This involved “unfolding” them into a net of polyhedra and then imitating the way it was done for the three-dimensional polyhedra. For example, for the hypercube, the obtained net is a net of eight cubes; the faces of these cubes are rendered in different colors, in order to indicate how they should be re-folded (fig. 1b).6 Though sometimes the way mathematicians referred to color in relation to the fourth dimension during the nineteenth century was metaphorical,7 it was also used in a practical manner, as if color may be indeed thought as a coordinate of the fourth dimension. Looking at Hinton’s 1904 book The Fourth Dimension, when dealing mainly with the presentation of a four-dimensional cube (tesseract) and other more complicated arrangements, he states that: 5 See: Charles Howard Hinton: A New Era of Thought, Bloomsbury: Swan Sonnenschein, 1888, for a preface written by Boole Stott and H. John Falk, noting they wrote themselves parts of the book, and that “the reader will find colours necessary to enable him to grasp and retain the complex series of observations” (p. vii). 6 For other ways to visualize these polytopes, see: Victor Schlegel: Ueber den sogenannten vier dimensionalen Raum, Berlin: H. Riemann, 1888; Washington Irving Stringham: Regular Figures in n-Dimensional Space. In: American Journal of Mathematics, 3, 1880, pp. 1–14. 7 Gustav Theodor Fechner: Vier Paradoxa, Leipzig: Leopold, 1846, p. 25, notes that humans are beings of color in a three-dimensional world (when he makes the analogy to two-dimensional creatures who can deduce the existence of the third dimension). Fechner adds that since human beings are three-dimensional they cannot understand where the source of color emerges.
Coloring the Fourth Dimension?
83
2: Alicia Boole Stott’s sections of the 600-cell, being one of the four-dimensional regular polytopes.
“[…] a method of designating different regions of space by a systematic colour scheme has been adopted. The explanations have been given in such a manner as to involve no reference to models, the diagrams will be found sufficient. But to facilitate the study a description of a set of models is given in an appendix which the reader can either make for himself or obtain.”8 One can already note a tension between the haptic aspect (the “model,” when Hinton means the three-dimensional material models) and the visual one (the “diagram”), a tension I will return to later. Boole Stott also used color to visualize four-dimensional objects: “In practically constructing […] sections [of four-dimensional polytopes] I have found that their symmetry is made more obvious by colouring the faces.”9 She constructed cardboard models of sections of four-dimensional regular polytopes (fig. 2) and “painted each type of two-dimensional face with the same color in each of the sections in order to show how one section develops into another.”10 Even though Hinton and Boole Stott had different approaches to visualization of four-dimensional bodies, both viewed color as essential. 8 Charles Howard Hinton: The Fourth Dimension, Bloomsbury: Swan Sonnenschein, 1904, p. 136. 9 Alicia Boole Stott: On Certain Series of Sections of the Regular Four Dimensional Hypersolids. In: Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, 7, 1900, pp. 1–21, p. 14. 10 Irene Polo-Blanco: Alicia Boole Stott, a Geometer in Higher Dimension. In: Historia Mathematica, 35, 2008, pp. 123–139, p. 133.
84
Michael Friedman
However, not all approaches were sympathetic to the use of color as a fourth coordinate or as what designates how different sections of a four-dimensional polytope relate to each other, or how different faces of these polytopes should be identified. Victor Schlegel, constructing wire and cardboard models of four-dimensional regular polytopes himself, criticized any attempt to use colors as a “surrogate”11 to denote the fourth dimension, since this would fail to give a sensory impression similar to what is given in the third dimension: for the case of third-dimension coloring a figure succeeds in giving an illustrative representation of depth, in contrast to the case of the fourth dimension. Obviously, the above short survey is far from a thorough analysis of how four-dimensional polytopes were researched mathematically and visually during the nineteenth century, but it is already clear that at the end of the century color was at least one possible way to deal with the fourth dimension. However, it is important to note that a discussion regarding the aesthetic properties of coloring these models – and other colored models – did not take place during this period, i. e., the end of the nineteenth century.12 Hence, I would like to explore how color functioned as an indicator towards the mathematical properties of the researched object rather than its aesthetical or artistic ones. Returning to the above survey, the question arises whether four-dimensional polytopes were the only four-dimensional object for which colored visualization was done. The answer is that this was not so, and as I will show in this paper, for complex plane curves mathematicians were also struggling to find a proper visualization. The second part of this article will deal with the various approaches as how to visualize these curves and in particular, special points of them, called branch points. However, before turning to these case studies and in order to further explain the meaning of color in mathematics, I would like to explain first the problems that three-dimensional material models – seen in figures 1b and 2 – posed for the presentation of four-dimensional objects.
11 Schlegel (s. fn. 6), p. 17. 12 This changed in some sense during the twentieth century; see for example the surrealist artist Man Ray and the series of paintings he drew called “Shakespearean Equations,” based on photos he took of models of surfaces at the Institut Henri Poincaré in Paris; cf.: Wendy A. Grossman and Edouard Sebline (eds.): Man Ray. Human Equations: A Journey from Mathematics to Shakespeare, Jerusalem: The Israel Museum, 2015.
Coloring the Fourth Dimension?
85
Visual and tactile: Boltzmann and Poincaré As was noted above, Boole, Hinton and Schlegel all produced (three-dimensional) material models from different materials. These were models that aimed to “present to the senses” the four-dimensional polytope. This expression was the way the physicist Ludwig Boltzmann described mathematical material models in his 1902 article Models published in the Encyclopedia Britannica: “The term model denotes a tangible representation, whether the size be equal, or greater, or smaller […]. In pure mathematics, especially geometry, models constructed of papier-mâché and plaster are chiefly employed to present to the senses the precise form of geometrical figures, surfaces, and curves.”13 Ten years before the Encyclopedia Britannica article was written, Boltzmann describes that material mathematical models served: “to make the results of our calculation illustrative [anschaulich] and that not merely for the imagination [Phantasie], but visible to the eye and at the same time palpable to the touch by means of gypsum and cardboard.”14 What is evident in Boltzmann’s 1892 remark is his differentiation between the senses: the seeing eye vs. the touching hand, when supposedly a merger, or at least an agreement, occurs between the two senses with the help of the material mathematical model. The question that arises is whether this really was the case? The obvious problem, that the touching hand cannot sense any visible color (i. e., it is irrelevant for touch whether the plaster or the wooden surface is red or green), points toward a larger issue when it comes to visualizing four-dimensional objects. This problem and a possible solution to it are described in the writings of Henri Poincaré. In his 1902 book Science and Hypothesis, Poincaré begins chapter IV, “Space and Geometry”, with the following statement: “Perhaps somebody may appear […] who will devote his life to it, and be able to represent to himself the fourth dimension.”15 Immediately afterwards, he presents not one single mathematician who 13 Ludwig Boltzmann: Models. In: Encyclopedia Britannica, 10. ed., Edinburgh, 1902, pp. 788–791, p. 789. 14 Ludwig Boltzmann: Über die Methoden der theoretischen Physik. In: Walther von Dyck (ed.): Katalog mathematischer und mathematischphysikalischer Modelle, Apparate und Instrumente, Munich: Wolf, 1892, pp. 89–99, p. 90. 15 Henri Poincaré: Science and Hypothesis, New York: Walter Scott, 1905 [1902], p. 52.
86
Michael Friedman
indeed might have done exactly that (for example, Boole Stott or Hinton), but rather describes another type of space, the geometrical space, whose essential properties are: continuous, infinite, isotropic, of three dimensions, and homogeneous, “that is to say, all its points are identical one with another.”16 But what does Poincaré mean when he talks about “representing” the fourth dimension? Poincaré does not answer this question immediately, but rather presents other kinds of spaces, among them visual space and tactile space. Regarding visual space he notes: “These are muscular sensations [related to the third dimension] quite different from the visual sensations which have given us the concept of the two first dimensions. The third dimension will therefore not appear to us as playing the same role as the two others. What may be called complete visual space is not therefore an isotropic space,”17 which, it may be added, is hence different from geometrical space.18 He then continues: “Tactile space is more complicated still than visual space, and differs even more widely from geometrical space.”19 What is clear, though implicit in the argument, is that three-dimensional material, haptic models of four-dimensional objects already pose a problem when considered sensuously in the fourth dimension – as these two spaces: visual and tactile, have different modes of operation and certainly do not overlap with geometrical space, which has no relation to the senses. How does Poincaré resolve this tension between the different spaces he presented? Starting out with a discussion of three-dimensional space, he notes it is with this space that the “operations [of changing perspective] are combined, we see that they form a group” – i. e., an algebraic, abstract structure20 – “which has the same structure as that of the movements of invariable solids.” That is, group theory is seen as the key to understanding geometrical space. Poincaré then notes:
16 Poincaré (s. fn. 15), p. 52. 17 Poincaré (s. fn. 15), p. 53. 18 Moreover, Poincaré (s. fn. 15), p. 54, notes that if the “two muscular sensations [of the eyes] both vary independently, we must take into account one more independent variable, and complete visual space will appear to us as a physical continuum of four dimensions.” In his book The Value of Science Poincaré, while reaching the same conclusion, he refers explicitly to the sensation of the color red (Henri Poincaré: The Value of Science, New York: Dover, 1958, p. 54). 19 Poincaré (s. fn. 15), p. 55. 20 A group is a set A together with an associated action ∗, such that the following four conditions are satisfied: closure (for every two elements a, b in A, a∗b is also in A), associativity, the existence of an identity element and of an inverse element. Compare Hinton’s implicit suggestion of an algebraic structure of colors, considering the color “null” as the number zero while “adding” (mixing) new colors to each other; Hinton (s. fn. 8), pp. 136–137.
Coloring the Fourth Dimension?
87
“in the same way that we draw the perspective of a three-dimensional figure on a plane, so we can draw that of a four-dimensional figure on a canvas of three (or two) dimensions […]. There is nothing, then, to prevent us from imagining that these operations are combined according to any law we choose for instance, by forming a group with the same structure as that of the movements of an invariable four-dimensional solid. […] In this sense we may say that we can represent to ourselves the fourth dimension.”21 As a hint to the hope (expressed at the beginning of the chapter) that someone will be found who “will represent to himself the fourth dimension” and with Poincaré’s suggestion that it was algebraic group theory that can “represent” the four-dimensional space, any visual or haptic means are discarded. This is clearly a reference to Felix Klein’s 1872 Erlangen Program, which sought to classify manifolds according to their transformation groups. Nevertheless, the reason for the rejection of the visual and the haptic might be the plurality of spaces that Poincaré surveyed: beyond geometrical, group-theoretical space, there are the different visual and tactile spaces – both dependent on the observer and both also being different from geometrical space. There were for Poincaré, one might say, not only plurality of interpretations, of what space is, but also discordance between visual space and the tactile space. This is what led Poincaré to prefer algebraic space, which is independent of any sense. The question remains whether Poincaré’s solution indeed reflects how mathematicians during the last quarter of the nineteenth century dealt with the tension between the visual and the tactile spaces. Not only do I claim that the solutions to this tension offered before Poincaré’s works were not necessarily a simple turn towards a more symbolic or algebraic approach (e. g., Schläfli symbols or group theory), but that they also used color to ease this tension. I will show this by examining a few case studies of material models of complex curves.
21 Poincaré (s. fn. 15), pp. 69–70.
88
Michael Friedman
Manifolds and colors During the second half of the nineteenth century, Bernhard Riemann provided one of the first examples to introduce the notion of color, as what may constitute a spatial manifold. Complex curves as covering of the complex line – i. e., Riemann surfaces, and more generally, manifolds, were introduced and developed by Riemann during the 1850s. Riemann used the term “Mannigfaltigkeit” [manifold] in 1854 in connection with “magnitude,” when he stated that he set himself “the task of constructing the notion of a multiply extended magnitude,”22 and invoked various motivations when first using the term. When talking about continuous manifolds, the intuitions and examples Riemann provides for choosing the term “Mannigfaltigkeit” are positions of sensuous objects and colors.23 What Riemann meant with manifolds of color is clearly influenced by the philosopher Johann Friedrich Herbart:24 Taking the concept of color, each particular color is a mode of determination [Bestimmungsweise] of the general concept of color;25 the totality of these determinations forms a manifold. Therefore, in the manifold of color, as the colored point changes, it changes continuously and “different colored points are thereby determined.”26 However, Riemann’s color example is no longer developed, neither in his own later writings nor in the writing of other mathematicians during that period. Obviously, Riemann does not consider color as the cause of tension between the abstract and the concrete or the sensuous, but rather as what may constitute a manifold, or at least function as a coordinate of it. This idea that color may serve as an additional coordinate, or at least as a mode of determination, was taken up again with the construction of material models of complex curves, though – and this should be emphasized – without mentioning Riemann’s remark on color.
22 Bernhard Riemann: The Hypotheses on Which Geometry is Based. In: Roger Baker, Charles Christenson, Henry Orde (trans.): Collected Papers Bernhard Riemann (1826–1866), Heber City: Kendrick Press, 2004, pp. 257–272, p. 257. 23 Riemann (s. fn. 22), p. 258. 24 Herbart explained continuity with an example of a triangle with blue, red and yellow at the vertices, and mixed colors in between (José Ferreirós: Labyrinth of Thought. A History of Set Theory and Its Role in Modern Mathematics, Basel: Birkhäuser, 2007, p. 46). See also: Erhard Scholz: Herbart’s Influence on Bernhard Riemann. In: Historia Mathematica, 9, 1982, pp. 413–440. 25 Riemann (s. fn. 22), p. 258; see also: Ferreirós (s. fn. 24), p. 63. 26 Erik C. Banks: Extension and measurement: A constructivist program from Leibniz to Grassmann. In: Studies in History and Philosophy of Science Part A, 44, 2013, pp. 20–31, p. 23.
Coloring the Fourth Dimension?
89
Coloring the branch points Before elaborating how color was used to visualize curves, a mathematical detour is needed. The set of real numbers, denoted as ℝ, contains all the numbers we use in our daily life: −½, √2, 5 etc. Many equations have solutions in this set, for example, the equation x2 = 2 has two solutions: √2 and − √2, both being real numbers. However, not every equation has a real solution; the equation x2 = −1 has none: there is no real number x0 such that (x0 )2 = −1. To solve this problem, during the sixteenth century other numbers were introduced, called “imaginary” or “complex,” by symbolically denoting i = √−1.27 Therefore i is a solution to the above equation, since i2 = −1. The number x = a+bi , when a and b are real numbers, is called a complex number; a is called the real part (denoted as Re(x)) and b the imaginary part (denoted as Im(x)); for example, when x = 2+i, then Re(x) = 2 and Im(x) = 1. Hence, a complex number is represented by two real numbers: a and b. Note that if b = 0, then x is a real number. One denotes the set of all complex numbers as ℂ, sometimes also called the complex line. A curve y = f(x) is called complex valued (or just complex) if x (and hence y) are complex numbers. For example, for the curve y =x2, if x =1+i, then: y = (1+i)2 = (1+i)(1+i) = 1+2i+i2 = 1+2i−1 = 2i The two-dimensional complex plane, denoted as ℂ2, is the set of all pairs (x,y) when x and y are complex numbers.28 For a complex valued curve y = f(x), every point (x0 , y0 ), such that y0 = f(x0 ), can be represented as a point in a four-dimensional real space ℝ4 as the quadruple of real numbers (Re(x0 ), Im(x0 ), Re(y0 ), Im(y0 )). But visualizing a point with four coordinates is impossible with drawing in a three-dimensional space. Hence, visualizing these complex points (x0 , y0 ) as a drawing on paper (for example, only the real points, i. e., the points for which Im(x0 ) = Im(y0 ) = 0, see figure 3a) or as a material model in a three-dimensional space – by constructing models of surfaces whose points are either (Re(x0 ), Im(x0 ), Re(y0 )), usually called R (for real) or (Re(x0 ), Im(x0 ), Im(y0 )), usually called I (for imaginary) – would always risk being insufficient from a mathematical as well as from a visual point of view. 27 By this it is meant that the root of a negative number exists only symbolically, and does not refer, for example, to a length of a segment, as one can assign to the number 2 or √2. Note also that the symbol i was introduced by Euler towards the end of the eighteenth century. 28 Similarly, ℝ3 denotes the real three-dimensional space (which is the set of all triple (x, y, z), when x, y and z are real numbers), and ℝ4 denotes the real four-dimensional space (being the set of all quadruples of real numbers).
90
Michael Friedman
When one tried to produce these three-dimensional models, R and I were usually presented together. Walther von Dyck, who was responsible for constructing several of these pairs of models for various functions, noted in 1886 that in order to differentiate between the models R and I one colored the various layers of the models differently.29 As Gerd Fischer noted, a hundred years later, however, these colors faded to such an extent that they were no longer visible.30 Notwithstanding this disappearance of coloring, during the last quarter of the nineteenth century there were various visualizations of a special phenomenon – branch points – that arose when considering these curves as covering the complex line, when few of the material models visualizing these points were also colored. But what are these points? First, one considers complex curves as covering of the complex line (simply speaking, as projection to it). To give an example, consider the complex curve y2 = x−2 and its projection to the x-axis (fig. 3b): p: {(x, y) ∈ ℂ2 : y2 = x−2 } → ℂ, (x, y) ↦ x Generically, every point x ∈ ℂ has two different pre-images (x ,y1 ), (x ,y2 ) ∈ ℂ, such that (y1 )2 = x−2 and (y2 )2 = x−2 (regarding the notion of “pre-image,” see the caption of figure 3b). However, for x’ = 2, the number of the pre-images is less than two (explicitly, there is only one pre-image: (2,0)). One might say that when considering the points x’’ ∈ ℂ which are close to x’ = 2, the two pre-images of x’’ “come together” when x’’ approaches x’. Considering only smooth curves,31 these points, whose number of pre-images is lower than the expected one, are called branch points. When n pre-images “come together,” one says that the branch point is of order n−1. Visualizing these branch points on the curve posed a challenge, since it was not only evident that drawing the real part is an incomplete image (fig. 3b), but also it was not clear how to visualize the coming together of several pre-images without indicating that the curve intersects itself, i. e., without presenting a misleading visual image. Indeed, when taking either the three-dimensional models R or I of the complex curve, what one sees in the neighborhood of a branch point is a surface, which 29 Walter von Dyck (ed.): Mathematische Modelle angefertigt im mathematischen Institut der k. technischen Hochschule in München. Modelle zur Functionentheorie (Zu Serie XIV), Munich, 1886, p. 2. 30 Gerd Fischer (ed.): Mathematical Models: From the Collections of Universities and Museums. Kommentarband, Braunschweig: Friedr. Vieweg and Sohn, 1986, p. 78. 31 A smooth curve is intuitively a curve which can be drawn without self-intersections or “spikes” (i. e., there are no singular points); for example, the curve y = x2 is smooth (fig. 3a), but the curves y2 = x3 or y2 = x2 are not as the first has a “spike” and the second is composed of two lines intersecting themselves.
Coloring the Fourth Dimension?
91
3A: When the real part of the complex valued function y = x2 is drawn, the points (2,4) or (0,0) on it can be visualized (see the red resp. grey point), but the point (1+i, 2i) cannot, as its coordinates are not real numbers. B: The real part of the curve y2 = x−2 (drawn in red), the branch point (2,0) (the grey point), and the pre-images with respect to the projection p of the point x’’, being close to the point x’ = 2 (the pre-images are the two red points “above” the black point x’’ on the x-axis); one can imagine how these two red (real) points “coincide” into the grey point when x’’ approaches x’. Hence one says that there are two branches, which “come together” at the branch point. The pre-images of a point x ∈ ℂ is the set of all points (x, y) on the curve such that p((x, y)) = x. In the figure, the red points are projected via p to the black point x’’, hence these points are the pre-images of the point x’’. The problem with this visualization is that it does not show how the complex part (or complex pre-images) of the curve looks like.
self-intersects itself (fig. 5). It must be emphasized, however, that self-intersection is a situation that does not occur – since the curve is smooth and self-intersection is considered as a singularity.32 This anomaly is caused by the fact that one dimension is omitted (see below). A possible solution to this problem was with the usage of color. I will begin, however, with an example that does not use color as a solution so that I can show the differences between that version and other visualization techniques that did use color. Though a full survey of how branch points were visualized during the nineteenth and twentieth century is beyond the scope of this article, it is instructive to look at Felix Klein’s 1874 sketch of a branch point of the second order of a complex curve. Klein draws what a branch point of a singular curve of the third degree looks like: “one awards to the [Riemann] surface an […] outgoing branching […], as illustrated [in fig. 4], for example, in a symmetrical manner, by the drawing.”33
32 See von Dyck (s. fn. 29). 33 Felix Klein: Ueber eine neue Art der Riemannschen Flachen (Erste Mitteilung). In: Mathematische Annalen, 7, 1874, pp. 558–566, p. 566.
92
Michael Friedman
4: Klein’s 1874 visualization of an order two branch point.
However, from the sketch alone it is not clear how the different pre-images “interact” with each other, and it seems that all the three branches intersect. Though the different directions of the drawn diagonal lines on each layer might refer to different haptic sensations (when touching a model produced according to this image) or to different visual perceptions, Klein does not elaborate on this. This problem of how the different branches actually “behave” visually is solved during the 1880s and the 1890s. Given the background of the tradition of construction of mathematical material models, promoted mainly by Felix Klein, Alexander von Brill and Walther von Dyck, it is not at all surprising to also find material, three-dimensional models of branch points. I would like to examine one model, which was produced starting in the middle of 1880s. This model might be considered a three-dimensional realization of a Riemann surface in the neighborhood of a branch point of order 2, as Klein depicted in 1874. The model is described simply as a model of a “three-leaves simply connected Riemann surface, which has at its center a branch point of the second order.”34 Comparing the model, shown in figure 5, to the two-dimensional image that Klein drew, one can note two main differences. First, it is clearer how the different pre-images “come together” at the neighborhood of the branch point. The arrows drawn on this model also help us to understand visually what happens on the surface, and they 34 Ludwig Brill: Katalog mathematischer Modelle für den höheren mathematischen Unterricht, 4th ed., Darmstadt: Brill, 1888, p. 48.
Coloring the Fourth Dimension?
93
5: A model of a branch point of second order, preserved in the Göttingen Collection of Mathematical Models and Instruments.
show the movement of the different points on the surface. This model, though it is unknown when it exactly was made, shows visually how these points behave in the neighborhood of a branch point. As can be seen in figure 5 above, the second main difference is the fact that the different sheets are colored: while it might be thought that the curve intersects itself when only looking at the model and ignoring the different colors, this is certainly not the case. The “self intersection” is only due to the fact that for a complex curve, given by the equation f(x, y) = 0 (when x, y ∈ ℂ) only a three-dimensional section is (and can be) presented: i. e., of the points (Re(x), Im(x), Re(y)). This mode of presentation causes points, which only differ in their Im(y) -values, to coincide in the three-dimensional model. The way to show that these points are actually different was to color the model: in the neighborhood of the “intersection line” there are different colorings: i. e., black lines are drawn on one layer in order to differentiate it from the second “intersecting” layer, which is white. This is done in order to make it visually clear that these layers do not really intersect. This shows already what Poincaré referred to as the difference between visual and haptic spaces: on the one hand, only touching the model would give the wrong sensuous picture, as it would indicate that several branches intersect each other, which is obviously not the case. Visual space, on the other hand, tells another story: the different colors show the observer that what may seem to intersect with itself does not intersect with itself at all, due to the presence of different colors at the “intersecting” layers.
94
Michael Friedman
6: Different coloring in the neighborhood of a branch point (to be seen at the center of the image).
This method of coloring the different layers in order to differentiate between them, notwithstanding the fact that with a three-dimensional model or with a two-dimensional drawing one can only draw intersecting layers, was also used in 1897 in the first volume of the book Vorlesungen über die Theorie der automorphen Functionen, written by Klein and Robert Fricke. There, the authors draw “inner branch points,”35 as shown in figure 6. As can be seen from the drawing, from the inner point a line exits, one side of it is darkened, the other side is bright. This shows obviously that the colors – the different shades of grey – differentiate between the different values around the branch point, and that the supposed self-intersection does not take place in the four-dimensional space. Already these three examples show which role color played for the visualization of complex curves. To make it explicit: color was essential and necessary, as it was used to denote properties of the mathematical object (in this case, the fourth dimension), that could not be brought “to the senses” in any other way. But the necessity of color is not only unique for denoting the fourth dimension. It is also to be seen with the different sets of the 27 lines on the plaster models of a cubic surface: these were colored differently, in a way that describes the different families to which these lines belong.36 35 Robert Fricke, Felix Klein: Vorlesungen über die Theorie der automorphen Functionen, Vol. 1, Leipzig: Teubner, 1897, p. 372. 36 Brill (s. fn. 34), pp. 40–41. For example, in one of the models there are six red lines and six green lines so that every line touches five lines of the different color; see: Fischer (s. fn. 30), p. 11, and: ibid. (the accompanying photo volume), p. 13, photo 10.
Coloring the Fourth Dimension?
95
Different curves on plaster models of surfaces were carved and then colored, in order to differentiate between them.37 Coloring the different parts of surfaces or the different lines and curves on surfaces was therefore – as I claim – a tool to create and show differences:38 either between dimensions or between certain sets of curves. These differences – at least at the end of the nineteenth century – were only shown materially through color. Conclusion The discussion above showed that color was essential when trying to visualize four-dimensional objects – at least when considering the tradition of material models at the end of the nineteenth century, and in sharp contrast to the rejection by Frege or Pasch (among others) of the senses as a possible source of mathematical knowledge.39 The importance of color and coloring was due to the fact that it enabled a better understanding of the nature of the spatial object; it was thought and used as the only possible way to show differences. The usage of color was hence meaningful, since it helped to make the mathematical object more accessible, more “tangible,” to borrow the expression from Boltzmann. For Boole Stott’s models, for example, one could better grasp how the different colored three-dimensional solids, being sections of four-dimensional regular polytopes, change when one takes two neighboring sections; not coloring these sections might have led to serious difficulties in understanding how these sections are related. The same line of thought applies to the three-dimensional models of branch points: not coloring each layer 37 Brill (s. fn. 34), p. 57, § 133, §134. The two types of curves were geodesic curves and asymptotic curves. Intuitively, a geodesic curve is the shortest path between two given points in a curved surface. An asymptotic curve is a curve on a surface, which is always tangent to an asymptotic direction of the surface, that is, when the normal curvature is zero on the surface. The two types of curves are usually different from each other. 38 Recall that Hinton says this explicitly: Coloring the different faces is “a method of designating different regions of space by a systematic colour scheme”; Hinton (s. fn. 8), p. 136. Brill’s catalogue does not mention explicitly this epistemological character, although it is clear that coloring the various carved curves points to their being different from each other, since the colors which were used differed from each other. Other models, made of strings, of two surfaces intersecting each other, when the strings of each surface were colored differently, also used color as an essential epistemic property. Anja Sattelmacher: Anschauen, Anfassen, Auffassen. Eine Wissensgeschichte mathematischer Modelle, Phd Thesis, Berlin: Humboldt-Universität zu Berlin, 2018, p. 111, claims that color drew attention to how the surfaces intersect each other, so that one could use the two-colored threads to mark exactly the intersection points of the surfaces. 39 Moritz Pasch: Vorlesungen über neueren Geometrie, Leipzig: Teubner, 1882, p. 17, notes, “after [the axioms of geometry] are established it is no longer necessary to resort to sense perceptions”; Frege: The Foundations of Arithmetic, New York: Harper, 1960, pp. xvii-xviii, argues that mathematical statements are not dependent on sensations, chalk or blackboard.
96
Michael Friedman
differently might have prompted a wrong image of a self-intersecting, singular curve, whereas the curve itself is known to be smooth, i. e., non-singular and without self-intersections. In this sense, while being a sensuous property, coloring and the differences it sensuously prompted pointed towards a non-sensory, abstract knowledge regarding the mathematical, four-dimensional objects. There are two essential aspects to take into account at this point in the discussion. First, it is obvious that it was most likely irrelevant if a layer or a line was colored in red or green40 – that is, color by itself did not carry any meaning. Nevertheless, the relations with other, neighboring colored layers or lines were meaningful – as the color of the other layer or line had to be different. This means that color was a relational property (what matters are the relations between the various colors and the fact that they are different from each other; see also the different direction of the sketched lines in figure 4). Hence, once more this property points to an abstract mathematical knowledge that is not dependent on visualization or coloring, even though this abstract property is shown essentially through sensuous means. Second, the colored three-dimensional models undermine both Boltzmann’s and Poincaré’s views regarding how to “represent” spatial objects. While Boltzmann did not explicitly discuss four-dimensional mathematical objects, he did know and consider three-dimensional material models of complex curves embedded in a four-dimensional real space. Boltzmann saw the three-dimensional material models as a merger between the hand and the eye. However, it is obvious that coloring the various models indicated a preference for the visual aspect. For the case of nets of four-dimensional regular polytopes, no one could fold in real setting these three-dimensional nets to construct a four-dimensional polytope; one could only imagine folding them into the fourth dimension, and this with the help of the different coloring of the faces. For the case of the models of branch points, only touching the model could supply the senses an incorrect image of the mathematical object, in this case, the complex curve. Poincaré, though acknowledging the existence of visual and tactile spaces, and even of a four-dimensional “visual space,”41 eventually pointed out that the way to represent the geometrical space is with algebraic formulation (group theory). Although he did not explicitly reject visual or tactile space and how they may contribute to our (mathematical) knowledge of space as such, his conception neverthe 40 Questions such as why these colors (or other colors) were chosen (for example, Hinton’s peculiar color palette), or who colored these models, unfortunately go beyond the scope of this paper. 41 According to Poincaré, if each eye had a different sensation, then “the whole visual space would have four dimensions.” Poincaré: Value of Science (s. fn. 18), p. 54.
Coloring the Fourth Dimension?
97
less certainly gave these spaces a lower status when compared to geometrical space. It is clear, however, that the colored three-dimensional models show that one should take into consideration the visual aspect – and colors as well as the actions of coloring and seeing colors, as a sensuous practice – as what supplies a complementary knowledge regarding the mathematical object.
Ricardo Cedeño Montaña
Encoding Color: Between Perception and Signal Input Encoding schemes for producing, storing, and transmitting color information in electronic media are based on the additive mixture of red, green, and blue lights, a three-color principle that originated in nineteenth-century physiological studies of vision. During the twentieth century this principle was first standardized and then implemented in several technical media. Similar to many other aspects of media technologies, over the past century there has been a standardizing trend to order, regulate and stabilize the production of color sensations with electronic media. The standardization of color has been an essential part for the creation of ever more precise and predictable sensors, displays and encoding schemes. Two images open this chapter (fig. 1). Both are magnifications of the input and output hardware of electronic visual systems where the three-color principle has been implemented. Both are standard pieces used in consumer imaging devices. The micrograph on the left corresponds to the charge-coupled device image sensor installed in a common webcam. The image on the right is a macro picture of an in-plane-switching liquid-crystal display. The input is a rectangular grid of circular photodiodes and the output is an arrangement of chevron-patterned electrodes. There are only red, green, and blue elements, but their distributions differ. Half of the photodiodes on the image sensor are green, whereas on the display the colors of the electrodes are evenly distributed. The nature of color has been studied in history, science and philosophy of color. However, it is only with the recent material turn that focuses on the media technologies as the concrete link between science, industry and culture that the question of color in electronic media has been addressed by media history.1 This chapter describes how a certain interplay of ideas, instruments, blueprints and specifications gave birth to the trichromatic theory and its implementation in electronic media. This media history focuses on the works and agreements among a network of scientists and technicians that after explaining this human sensation as the mixture of three primary stimuli, fostered a series of technical developments to homogenize the color sensations produced by electronic media. From sensing device to transmission channel, electronic color is a media operation that exploits 1 Susan Murray: Bright Signals: A History of Color Television, Durham: Duke University Press, 2018.
100
Ricardo Cedeño Montaña
1: Magnifications of a digital image sensor and a liquid crystal display.
the physiological characteristics of human vision to cause color sensations using the minimum amount of sensory data. Four media implementations of this theory frame the following description. These are: the Colorimetric Resolution I by the Commission Internationale de l’Eclairage (CIE) from 1931, the implementation of the National Television System Committee (NTSC) color television during the 1940s and 1950s, the Bayer filter that has been used in imaging devices since the 1980s and the International Telecommunication Union BT.601 recommendation for encoding digital video signals from 1981. To approach these implementations, this text will begin in the minds of scientists during the late eighteenth century whose research on color vision centered on the inner workings of the human eye. Three sensations: red, green, blue Throughout the eighteenth century, many researchers thought that color vision depended on a rather small number of sensations, and a trichromatic theory was proposed several times. In 1757, Mikhail Lomonosov claimed that colors arise in the human eye from the mixture of red, yellow and blue sensations; and in 1777, George Palmer proposed a red-yellow-blue system, in which each ray of light consists of three rays that stimulate three retinal particles to cause the sensation of color.2 While eighteenth-century color science was already able to conceive the existence of three 2 David L. MacAdam: Sources of Color Science, Cambridge: MIT Press, 1970, pp. 40–41.
Encoding Color
101
retinal receptors, however, it lacked the optical and electrical media technologies to demonstrate their existence. In response to the question of where the human eye processes and orders light as color sensations, Thomas Young contended that it was impossible that each receptor in the retina had an endless number of particles, “each capable of vibrating in perfect unison with every possible vibration of light.” He reasoned that an infinite number of particles could not fit on a finite retina. Like Palmer, Young suggested on November 12, 1801, that to cause all imaginable varieties of hues in the eye, each receptor would consist of three regions each sensitive to one particular stimulus, “for instance, to the three principal colors, red, yellow, and blue.”3 Since then, nearly any given color sensation has been described in terms of the mixture of three selected portions of the visible electromagnetic spectrum. Color sensations produced by light may be generalized in terms of three properly chosen variables: C=pX+qY+rZ in which C is any color; X, Y and Z are the chosen primary spectral lights; and p, q and r are the coefficients of the primaries required to match the sensation produced by the light. These primary lights are an adequate formal description of the properties of color vision insofar as they relate to color mixing. In the additive theory of color, only one thing is required of a primary: it cannot be produced by the mixture of another two.4 When the primaries are added together using the maximum coefficients they must form a unit, i. e., a white or colorless sensation,5 and white becomes the basis for the proportions of the primaries in the mix and for defining a standard observer. Palmer and Young’s theories reduced the number of primaries to three and contended that an equal number of receptors on the retina has to be responsible for human color sensations. This simplification to three primary stimuli became the basis for understanding the mechanism of color vision and, two centuries later, for implementing color in electronic media. Young formed a simple order to explain all the phenomena described by Isaac Newton with a discrete number of sensations. 3 Thomas Young: The Bakerian Lecture: On the Theory of Light and Colours. In: Philosophical Transactions of the Royal Society of London, 92, 1802, p. 21. 4 C. L. Hardin, Arthur Danto: Color for Philosophers: Unweaving the Rainbow, Indianapolis: Hackett Publishing, 1988, p. 28, about primary colors in painting see Ian Lawson in this volume. 5 William D. Wright: A Re-Determination of the Trichromatic Coefficients of the Spectral Colours. In: Transactions of the Optical Society, 30, 1929, p. 143.
102
Ricardo Cedeño Montaña
By the 1850s, media technologies were already paving the way to the empirical understanding of color vision. In Edinburgh, James Clerk Maxwell used photo graphy and magic lanterns to demonstrate the three-receptor theory of color vision; and in Berlin, Hermann von Helmholtz used the telegraph as a model to explain the processes of perception. Maxwell believed that building an optical apparatus capable of mimicking the trichromatic theory would be enough to validate it as the color mechanism used by the nerves in the human retina.6 Without dissecting a human eye, Maxwell intended to explain the sensation of color via technical media. Thus, the technology of photography not only provided iconic images of the world but also evidence of how the human visual system processes color. First, he took three black-and-white photographs filtered through red, green and violet glasses. Then, using three magic lanterns with similar color filters, he superimposed the three images. Color first became synthetic in Maxwell’s noisy photograph of a tartan ribbon, thereby laying the groundwork for color film and TV. In the 1850s, building on Young’s trichromatic theory, Helmholtz empirically demonstrated that there are three types of sensory receptors in the retina, each primarily sensitive to one range of wavelengths. Whereas Young had only described these receptors theoretically, Helmholtz, with the assistance of a variety of new media technologies, particularly electric and telegraphic inscription devices,7 provided a mathematical description of the responses of the three receptors to the sensory data stream. Helmholtz did not use media technologies in the laboratory in a merely instrumental way. He conceived of human sensory organs as media apparatuses themselves, with the electrical telegraph serving as a generalized model for human perception. On December 13, 1850, in his report to the Physikalisch-Ökonomische Gesellschaft zu Königsberg about the reaction times of the human sensory apparatus, he compared the processes of perception to the transmission of messages through the electrical telegraph. According to him, the nerve fibers worked like the wires of the electrical telegraph, with each transmitting messages from the outer borders to the governing center. Furthermore, he concluded that any sensation in the human body consisted of three moments: first the reception of the signal by the senses, then its transmission to the brain, and finally the processing of the perception by
6 James Clerk Maxwell: Experiments on Colour, as Perceived by the Eye, with Remarks on Colour- Blindness. In: Transactions of the Royal Society of Edinburgh, 21, 1855, p. 284. 7 Timothy Lenoir: Helmholtz and the Materialities of Communication. In: Osiris, 9, 1994, p. 185.
Encoding Color
103
the brain.8 The eye was a sensitive organ with three types of nerves that encoded red, green and violet signals as a multi-dimensional sensation. An image cast on the retina was a set of discrete electrical impulses, symbolic data similar to the Morse code that was transmitted to the brain as a multiplexed signal along the optical nerve. The sensation of yellow resulted from a strong excitation of red and green receptors with a weak excitation of the violet receptors, while white resulted from the homogeneous excitation of all three receptors simultaneously.9 This quantification of sensory data using media technologies at the instrumental and epistemic level was important because it enabled theoretical problems of color vision to be rendered concrete and standardized. Nineteenth-century physiology thus turned the human sensation of color into a problem of technological media. On the one hand, Maxwell’s additive synthesis of colors marked the beginning of color separation and recombination using media technologies. On the other hand, Helmholtz’s work set the blueprint for the principle of color mixing in electronic visual media. As Friedrich Kittler put it, after such “cold and inhuman” analysis, “nothing stood in the way of the construction of real media that deceived and/or simulated visual perception.”10 Standardizing color The trichromatic theory of color has been implemented in three technical media: the three light-sensitive emulsions of film, the cathode-ray tube layer coated with three phosphors and the three color components of digital images. In all of them, a weighted combination of three different color signals suffices to form a full color space for the human visual system. Yet, all media implementations require a technical standard. The standardization of the sensation of color during the 1930s was a turning point in the history of technical media because it transformed color into a system of parameters to provide precise information about how to represent the same color to a standard observer. This systematization of color as a technical standard for media aspired to override subjective tastes and preferences by rendering color as objectively given. 8 Hermann von Helmholtz: VI. Ueber Die Methoden, Kleinste Zeittheile Zu Messen, Und Ihre Anwendung Für Physiologische Zwecke: Gelesen in Der Physikalisch-Ökonomischen Gesellschaft Zu Königsberg, Am 13. Dezember 1850. In: Königsberger Naturwissenschaftliche Unterhaltungen, C1 V2 1848–1852, 1852, pp. 186–187. 9 Hermann von Helmholtz: Handbuch Der Physiologischen Optik, Leipzig: Leopold Voss, 1867, p. 291. 10 Friedrich Kittler: Thinking Colours and/or Machines. In: Theory, Culture & Society, 23, 2006, p. 42.
104
Ricardo Cedeño Montaña
At the outset, these series of standards described color according to the results from a series of matching experiments using red, green and blue spectral lights. First, they precisely set the location of the primaries in the electromagnetic spectrum; and second, they describe how to synthesize any color sensation using a mixture of three monochromatic lights in a given ratio. Whether as abstract color models, or as the output of technical media, these standards are the result of attempting to explain any color sensation under almost any viewing condition. Thus, they established the foundations for all modern color mixing and transmission. Technical media standards do not simply regulate the functioning of machines, they regulate the processing, storage- and transmission of how sensory data reaches the human senses.11 During the 1920s, William David Wright and John Guild conducted experiments and took measurements using 17 different observers in order to determine a set of color relations to homogeneous physical stimuli. The results of their research yielded a standard observer representing an average human’s chromatic response, a standard white matching the temperature of the light that was employed to set the chromatic measurements and a luminance function describing a standard observer’s response to different wavelengths of brightness. Taken together, these results constituted the basis to standardize a trichromatic system that fixed the proportions in which “three definite but arbitrarily chosen stimuli measured in certain units must be mixed to” produce the sensation of a particular color.12 Detached from any subjective variability, color vision became averaged and objectively numerical. Based on Wright and Guild’s research, in 1931 the CIE agreed upon a standard observer. Its Colorimetric Resolution (I) characterized the standard observer’s visibility values as the mixture of three homogeneous monochromatic “stimuli of wave-lengths 700 mμ, 546.1 mμ, and 435.8 mμ,” denoting these wavelengths as R, G and B, respectively.13 Primary colors were thus fixed to those wavelengths, and this precise tuning paved the way for symbol and number to take command of the representation of color sensations in future electronic media, such as cathode-ray tubes, light-emitting diodes and liquid-crystal displays. The second component of this resolution was a standard white to derive the units of magnitude for each primary. These primary stimuli had to meet one key requirement: the mixture of one unit of each of them had to match the white color defined by the National Physical Laboratory in London, where Guild had taken 11 Friedrich Kittler: Optische Medien, Berlin: Merve Verlag, 2002, p. 32. 12 Thomas Smith and John Guild: The C.I.E. Colorimetric Standards and Their Use. In: Transactions of the Optical Society, 33, 1931, p. 74. 13 Smith, Guild (s. fn. 12), pp. 75–81.
Encoding Color
uv
105
ir
2: CIE Luminance Function, 1924.
his measurements. The CIE resolution set the relative luminance of these units for the primary stimuli, R, G and B, in the ratio 1.0000:4.5907:0.0601; where, red was used as the basic photometric unit and green contributed with the largest part. Temperature was the base quantity for establishing this white. To calibrate their matching instruments, Wright and Guild used the illumination provided by a gasfilled lamp at a temperature between 4800º K and 5000º K as their reference. The CIE finally adopted a temperature close to “4800º K, corresponding to one of the yellower phases of daylight.”14 This temperature became thus the basic white stimulus to be matched by the standard observer. The third component of this colorimetric resolution was a function that provided the necessary quantitative data to express the relative luminosity of each monochromatic stimulus, i. e., their perceived brightness. Figure 2 shows the luminosity function based on the data presented by K. S. Gibson and E. P. T. Tyndall to the US National Bureau of Standards in 1924. It shows the brightness ratio of different colors by plotting the visible spectrum on the x axis along a single dimension of luminosity on the y axis. The function distinguishes the relative visibility of various wavelengths based on their spectral luminosity. To draw this function, Gibson and Tyndall used two spectrometers and averaged the responses from 250 pairs of eyes observing wavelengths from 430 nm to 740 nm.15 Their curve has a peak at around 555 nm and decreases sharply below 500 nm and above 600 nm, showing that in terms of relative luminosity, the human retina privileges the green portion of the spectrum.
14 Smith, Guild (s. fn. 12), p. 84. 15 K. S. Gibson, Edward P. T. Tyndall: Visibility of Radiant Energy. In: Scientific Papers of the Bureau of Standards N. 475, 19, 1923, pp. 132–136.
106
Ricardo Cedeño Montaña
Taken together, the standard observer, standard white and luminance function form a scientific and technical description of color sensation. The tables and calculations compiled in the 1931 CIE resolution provided a foundation for media technologies that at the time did not yet exist: color television and digital imaging. The resolution is also at the heart of scanners, displays, sensors and light sources that rely on it for calibration. This standard transferred the Helmholtz-Young model over to a specification that determines how colors are mixed and under which conditions such a mixture reaches human eyes. Two components: luminance and chrominace Coinciding with the standardization of the RGB color space, color vision research began to be applied in television during the 1920s and 1930s. In order to transmit color images, the entire black-and-white TV system, from the pick-up camera and the transmission line to the receiver, needed to deal with three signals instead of just one. At the height of what Siegfried Giedion calls the time of full mechanization, the first wave of color television apparatuses presented complex answers to the problem of transferring the Young-Helmholtz model and the CIE colorimetric resolution to electro-mechanical TV. These early solutions split colors into R, G and B signals with equal ratios. Later, in the era of analog circuitry and vacuum tube technology, the content and ratios of the three signals changed dramatically. Under economic pressure for compatibility with black-and-white TV systems, color data was encoded as the weighted sum of three signals: one major black-and-white signal and two minor color signals. Color was thus effectively reduced to a third of the transmission channel. Such a compression ratio, 3:1, stemmed from a clever exploitation of the human visual system and was aimed at fitting color into the existing TV transmission and reception systems. Before all of this, however, the media history of electronic color began with colored filters of glass developed to mechanically separate and turn light into three electrical currents. After John Logie Baird transmitted gray scale moving images on radio waves across the Atlantic, on May 18, 1929, he applied for a patent for the transmission of colored images. In his mechanical TV system, Baird used two synchronized Nipkow discs, each with three successive spiral curves of holes, each of which was covered with red, green and blue light-filters. In the transmitter, he placed an artificial retina composed of three photoelectric cells, each sensitive to one primary color. The disc rotated at a rate of 10 times per second, and Baird passed white light through its holes, thereby throwing a sequence of red, green and blue lights onto the object and exciting each cell, respectively. At this rotational frequency the transmitter
Encoding Color
107
sent a signal with thirty colored images per second to the receiving station. Such slow mechanical operation to encode colors as an electrical current was enough to surpass the human perceptual times and, according to one reporter of the science journal Nature, to very clearly reproduce in 1928 the vivid colors of pinkish-purple flowers, green leaves and red strawberries.16 During the 1930s, the Farnsworth image dissector and the Zworykin iconoscope tube transformed television into an all-electronic medium. Yet, color television was still designed using revolving color filters to mechanically decompose and recompose a scene. Among these designs are the almost forgotten chromoscopic adapter for television patented by the engineer Guillermo Gonzáles Camarena in Mexico in 1940 and the field-sequential color system that Peter Goldmark devised for Columbia Broadcasting System (CBS) in 1942. Both amplified and transmitted each video signal separately as a series of sequential electrical impulses. The principle is rather simple: the blue filter in front of the camera pickup tube only lets the blue regions of the image pass through; these blue regions are then scanned and transmitted; placing another blue filter in front of the receiving tube only renders the blue regions of the incoming image.17 The biggest problem for color TV, however, was backwards compatibility, i. e., the ability of black-and-white TV receivers to decode the color signal. To keep the same image sharpness of black-and-white TV, Goldmark’s system required reducing the frames per second from 30 to 24 and the scanning lines from 525 to 405. As a result, his system was incompatible with all the black-and-white sets already installed in millions of homes. When, in 1950, CBS demonstrated this solution for color television to the US Federal Communications Commission, compatibility with black-and-white sets was a major financial issue. CBS direct competitor, Radio Corporation of America (RCA), believed that the simplest way to provide color video was to analytically encode a trichromatic signal as one major black-and-white signal and two minor color signals that a color receiver would then decode back into the three spectral lights. In the camera, RCA split color into three spectral lights using three independent orthicon cathode-ray tubes each sensitive to spectral red, green and blue. In the TV set, RCA used three parallel electron guns firing at a screen composed of a regular pattern of small and closely spaced phosphor dots grouped in trios of red, green and blue dots (fig. 3). Between the electron guns and the phosphor screen there was a mask with as many 16 A. R.: Television. In: Nature, 122, 1928, p. 234. 17 Kingdon S. Tyler: Telecasting and Color, New York: Harcourt, Brace and Company, 1946, p. 159.
108
Ricardo Cedeño Montaña
Scanning Beams
Scanning Beams
3: The RCA color system, 1953.
perforations as trios on the screen. Each perforation was located on the axis of its corresponding trio. Once the three primary signals were decoded, every TV field was constructed by exciting in sequence one dot of each trio. Instead of the quick intercalation of successive monochromatic images, the RCA color TV set displayed to the eye a sequence of multiplexed color dots. This encoding system turned the incoming three primary signals into an outgoing composite signal with two components: luminance, i. e., the perceived brightness, and chrominance, i. e., the color information. The luminance component (Y) was a black-and-white signal and the crominance component was made of two color-difference signals that indicated how the red and blue colors differed from the luminance (R−Y and B−Y). Figure 3 shows how RCA solved the compatibility problem by applying the ratio established by the CIE in 1931. RCA’s Y signal was derived by adding the three primary lights in the weighted sum: Y = .30R + .59G+ .11B, where green contributed 59% to the mixture.18 Their analysis concluded that by adding “red, green, and blue spectral sensitivity curves” in such a ratio “the resulting curve 18 John W. Wentworth: RCA Color Television System. In: Broadcast News, January 1954, p. 10.
Encoding Color
109
has the same shape as the luminosity function”19 of black-and-white TV (fig. 2). On the transmission channel, the luminance component used almost two-thirds of the total bandwidth of 6 MHz, while the color data used only 2 MHz. The core of color TV was in fact dominated by a black-and-white component. This encoding system was, however, an old idea. In 1938, in Seine, France, Georges Valensi had already devised and patented a system to enable black-andwhite TV sets to receive color signals.20 A decade ahead of RCA, his invention split a TV color signal into one signal dependent on the luminance of the image scanned and another signal dependent upon the color of the image. Valensi encoded three incoming signals with an electrical differential that compared their voltages and produced two signals. At the receiving end, a decoder would either restore the three signals from the two for display in color sets or only use the first signal for generating a monochromatic image. Valensi’s reduction from three to two channels was aimed at saving costs and congestion in the transmission lines, and RCA’s dot-sequential system was an answer to the backward compatibility issue. But how were these color compression techniques related to the physiology of color vision? Between 1945 and 1946 at RCA Laboratories in Princeton, New Jersey, USA, the electrical engineer Alda Vernon Bedford conducted a series of experiments with slide projectors and color filters to measure the responses of the human visual system while observing color television. He showed first that the human eye’s acuity for differences in hue and saturation is lower than its acuity for brightness differences, and second, that the human eye has less acuity for blue light than for red and green.21 Bedford’s physiological analysis concluded that human vision is better able to resolve changes in brightness than in color at high levels of detail. Additionally, the human’s responses to equally bright red, green and blue spectral lights are very different, especially when combined to match the CIE standard white. His results matched the luminosity function from 1924, in which the human’s sensitivity to brightness peaks at the greenish wavelengths. For producing the sensation of color in TV, this meant one thing: a single brightness signal mixing in different proportions the primary colors red, green and blue is enough to carry all the fine details of an image. Bedford suggested that this signal be made with “the red and green signals [...] added in the proportions of 49 and 100 percent, respectively,”22 and 19 20 21 22
Wentworth (s. fn. 18), p. 9. Georges Valensi: Procédé de Télevision En Couleurs, FR841335, February 6, 1939, p. 1. Alda. V. Bedford: Mixed Highs in Color Television. In: Proceedings of the IRE, 38, 1950, p. 1005. Bedford (s. fn. 21), p. 1008.
110
Ricardo Cedeño Montaña
the blue signal in 3.1%. These proportions became the physiological basis for the luminance formula, Y, where the amount of green is double that of the red, while blue is almost insignificant. Both Valensi’s electric transmission of two-coded signals and Bedford’s techno-scientific analysis of the human visual system reveal the complex operation of transmitting the minimum amount of color information in a way that creates the sensation of a wide range of hues and shades. Thus, electronic color transmission became a technical operation conducted on the human visual system that caused it to perceive colors where there were almost none. During the early 1950s, RCA simplified its system to one single electron gun, and this was finally picked up by the NTSC and the Federal Communications Commission as the standard for color TV broadcasting in the USA.23 Color was thus added to the broadcast service using a third of the available bandwidth, serving as a minor component next to a massive black-and-white signal that contained all the brightness information. Long before computerized encoding techniques turned images into numerical matrices, this electrical compression technique set the blueprint for the digitization of colors during the 1980s. One grid: digital encoding In the digital domain, video became the serial procession of two-dimensional matrices at a specific and constant frequency. By the mid-1980s, solid-state imaging devices replaced the cathode-ray tube in camcorders, thereby turning moving images into a rigid rectangular grid of colored points with addresses in space and time. Cathode-ray tubes such as the iconoscope used in early TV cameras, collapsed into flat self-scanning image sensors, charge-coupled devices, with photosensitive capacitors arranged in columns and rows. While a charge-coupled device still converts light into electric charge, it further turns this electric charge into data, and it does so without moving mechanical parts. Colors are filtered and sensed as voltages corresponding to the amount of light gathered on each pixel. All voltages are sampled and transmitted as digital data via a semi-conductive layer. This chip-that-sees, added a second level of data conversion to the production of electronic colors, first from light to voltages and then to binary data. Atop the image sensor of a digital camera there is a filter. This filter is composed of individual luminance and chrominace sensing elements repeated over the surface of the charge-coupled device in a reticulated pattern. The left part of figure 4 23 Murray (s. fn. 1), p. 80.
Encoding Color
111
4: Color Imaging Array, 1976.
illustrates the structure of the filter patented in 1976 by Bryce E. Bayer, then an employee of Eastman Kodak Co. The Y parallelepipeds on the back dominate the array and represent the sites where luminance is sampled. The C1 and C2 elements “alternate with the luminance pattern in alternate rows, respectively”24 and represent the chrominance sensors. Light is thus split into the three primary color vectors shown in the orthogonal grid on the right side of figure 4. Bayer filters are implemented in digital cameras as circular photodiodes, see the micrograph in figure 5. Like Bedford, Bayer assigned half the total number of pixels to green, but a quarter to both red and blue. The sensing end of a digital imaging device was thus also designed to exploit the human visual system greater ability to discern luminance details, and the green frequencies contributed the highest amount of sensory data. Parallel to this development, and led by the European Broadcasting Union, an agreement between several international bodies of telecommunication engineers was reached in 1982 that brought to life the first standard for digital video signals: the recommendation BT.601. This standard established the rules for the transposition of analog video signals to digitally encoded data. 24 Bryce E. Bayer: Color Imaging Array, US3971065A, July 20th 1976, col. 3.
112
Ricardo Cedeño Montaña
5: Micrograph of the Bayer Filter atop an image sensor.
Bandwidth is also a critical factor in digital transmission channels. Like the engineers of the 1950s, the authors of this recommendation recognized that to transmit color information, it is more efficient to send luminance with full detail and use a “lower number of samples to represent the color-difference signals” with no luminance data.25 BT.601 does not encode an incoming video signal using the simple three-component representation of the RGB space. Instead, it uses the YCbCr component video (with Y for luminance, Cb: Chrominance-Blue; and Cr: Chrominance-Red). YCbCr decomposes the video signal into a luma signal, Y, that results from adding the R, G and B signals in the CIE ratio from 1931, and two color-difference signals, Cb and Cr. RGB requires equal bandwidths for each of its three color components, resulting in 8 bits for each component and three bytes per pixel. Again, exploiting the eye’s lower sensitivity to color than to luminance, in YCbCr “each pixel is made up of 8 bits for luminance and 4 bits for each of the color-difference signals, a total of 16 bits;”26 that means two bytes per pixel (fig. 6). Such a reduction from three to two bytes allows the digital transmission and storage of moving images at a considerably reduced bandwidth, while at the receiver end the human visual system hardly notices the reduction of color data during the fast reconstruction of each picture.
25 Stanley Baron, David Wood: The Foundation of Digital Television: The Origins of the 4:2:2 Component Digital Television Standard. In: SMPTE Motion Imaging Journal, 114, 2005, p. 329. 26 David Strachan: Video Compression. In: SMPTE Journal, 105, 1996, p. 69.
Encoding Color
RGB
113
Components
Bits
YCbCr
Components
Bits
R
8
Y
8
G
8
Cb
4
B
8
Cr
4
6: Comparison between the RGB and the YCbCr encoding systems.
Output At the scientific level, the production of color sensations has been located inside the human visual system and understood as a process of sampling the incoming luminous signal in three dimensions. At the machinic level, the production of color sensations became an encoding operation of voltages that occurs far away from the human eyes. Colors in electronic media are composed of layers of techno-codes and operations, with each adding a level of abstraction to the production of visual sensations. Once color had been analytically and physiologically understood as a trichromatic signal, media to synthesize color sensations were within reach. But this understanding of color vision was only completed when the mathematical and empirical models were turned into electronic media, thereby closing a feedback loop between the human visual system and media techniques. Electronic media sensibly reduces the chroma information in order to cause the sensation of color. Such reduction stems from empirical evidence showing that the human visual system processes details of luminance better than those of color. It is precisely in this technical detail in the history of electronic color that media operations can be best appreciated. To media, color is not a sensation but a routine technical operation performed directly within the human visual system. Thus, this system is treated as a standardized technical site forgetting for a moment that color is a bodily sensation subject to individual, environmental and cultural variability.
MEANINGFUL COLORS IN THE SCIENCES
Michael Rossi
Green Is Refreshing: Techniques, Technologies and Epistemologies of Nineteenth-Century Color Therapies In an early lecture of 1860, philosopher Charles Sanders Peirce reached for a seemingly obscure example in order to explain how things became perceivable to conscious beings. It was distinctness or distinction, Peirce told his audience, that rendered things sensible. In order to be perceived as a thing (rather than nothing, or everything), an entity – a chemical, an object, a person, a thought, an idea, a sensation – had to be distinguishable from that which it was not. By way of example, Peirce directed his audience’s attention to an ostensibly elementary sense perception – that of color. “There is a gentleman in England,” he wrote, “who has shown by an ingenious research that everything appears green to him. Green however, is not a refreshing color to him, because it is undistinguished.”1 On first pass, the point seems conventional, and the example arbitrary. To see something as a thing itself, it must appear distinct from other things. If all things appear to be green (for instance), then green has no meaning as a descriptor – it is indistinct because it describes everything indiscriminately. As with green, so too with all colors, and, indeed, all things in the world. In order to be identifiable as something, a thing (a color, an object, an idea) must be not-other-things. On closer reflection, however, one is struck by the strangely specific wording Peirce uses. It is not simply that green is recognizable when distinguished from other colors. Rather, green is refreshing – in its distinctness, it has effects upon the human body, and these effects are absent when green is indistinct. It’s a striking notion, made all the more striking by the offhandedness with which Peirce offers it. Peirce assumes that his audience would be familiar with the fact that green is “refreshing.” Further still, with his mention of the “ingenious research” of “a gentleman in England,” he signals his familiarity with scientific work on color perception: the “gentleman” is John Dalton, a celebrated chemist famous for writing about his own color blindness.2
1 Charles Sanders Peirce: Views of Chemistry, Sketched for Young Ladies. In: Max Harold Fisch, Peirce Edition Project (eds.): Writings of Charles S. Peirce: A Chronological Edition, Vol. 1, Bloomington: Indiana University Press, 1982, p. 50. 2 Peirce oversimplified the matter. Dalton did not write that everything appeared green to him. Rather, he had difficulty distinguishing reds, purples, oranges and pinks from different shades of green.
118
Michael Rossi
In fact, far from an incidentally chosen example, Peirce’s use of the color green to demonstrate the powers of distinction dramatizes his awareness of a shift in the epistemology of body, sensation and science in the middle of the nineteenth century. As Peirce put it in his lecture, “once, men were contented with facts and names and uses. Now, we always ask ‘what is the meaning of this thing?’”3 This was especially true in the case of the color green. For natural philosophers, physicians and laypeople in the United States a half century prior to Peirce’s writing, green was “refreshing” because it corresponded synoptically to the healthful and balanced qualities of nature.4 However, for Peirce and his presumed audience – interested and attuned to changes in contemporary science – the “meaning” of green could not simply lie in a correspondence between green light, nature and good health. Rather, green was a sign – an abstraction of environmental stimulus into an electrochemical signal with a particular, measurable and largely isolated effect on sensory physiology. As such, Peirce’s example was at once a statement of a commonly known fact about color, and a gesture at the ways in which such a fact could change. In the first half of the nineteenth century, the therapeutic properties of green were a matter of medical and folk knowledge. Physicians, industrial advisors and social commentators (among others) counseled that green light channeled onto diseased, stressed or otherwise overwrought eyes would bring relief and rejuvenation. This process could be effected through the judicious use of various techniques and technologies: green spectacles, most commonly, but also green eyeshades, green curtains and even simple squares of green colored cloth or paper. Thus did readers of the July 12, 1806 issue of the Philadelphia Evening FireSide find in its pages a paean to the salubrious effects of green glasses. Penned by a pseudonymous “Optician,” its opening stanza proclaimed: “Green spectacles! How few your value prize! How few can testify the good you’ve done! To soften rays that fall upon the eyes; And dim the lustre of that spark, the sun.”5
3 Peirce (s. fn. 1), p. 50 4 For the sake of simplicity and space, this essay focuses mainly on science and medicine in the United States. It should be clear from the sources used and read by the dramatis personae in this essay, however, that interest in color therapeutics was hardly limited to the US. 5 Optician: Ode to Green-Spectacles. In: The Evening Fire-Side, July 12, 1806, p. 218.
Green Is Refreshing
119
From this authoritative beginning, the utility of green light was clear: transforming solar light into an optically pleasing green hue was therapeutic, insofar as it admitted of vision – and, indeed, of the life-giving “spark” of the sun – but still ameliorated the harsher effects of solar light on the living eye. Contemporary healers agreed with the “Optician.” Between 1800 and 1860, medical texts on eye disease are replete with advice about the restorative character of green light. In his 1815 treatise, On the Morbid Sensibility of the Eye, British physician John Steven son reported that “it is […] universally admitted, that a green colour is highly grateful to the eye,” and recommended the use of green glasses and green shades in treating patients with “weak” eyes.6 William Lawrence’s Treatise on the D iseases of the Eye (1833) similarly quoted Austrian ophthalmologist Georg Joseph Beer to the effect that green glasses could sooth cases of overstimulated retina, while Richard Middlemore’s Treatise (1835) advised that those suffering from ophthalmalgia – a painful condition of the eyes – should wear a green shade.7 For his part, in 1843, John Walker recommended “shading the eyes with a piece of pasteboard, lined with green or blue silk, and fashioned lightly round the head” in cases of optical sensitivity. This routine was not to be prolonged beyond that which was necessary, but – in conjunction with opium and belladonna (a mild hallucinogen) – green light was soothing.8 In her practical manual, Notes on Nursing (1860), Florence Nightingale advised that a “light white curtain at the head of the bed,” and “a green blind to the window” would provide patients with the best stimulation for recuperating from all manner of sickness and injury.9 Such recourse to the color green – whether through transmitted light or reflected from green objects – was not merely theoretical, but was widely practiced as well.10 Stevenson, for instance, gave case notes on a patient of his who had experi 6 John Stevenson: On the Morbid Sensibility of the Eye, Commonly Called Weakness of Sight, Hartford, 1815, p. 94. All italics in original unless otherwise noted. 7 William Lawrence: A Treatise on the Diseases of the Eye, London: Printed for John Churchill, 1833, p. 508; Richard Middlemore: A Treatise on the Diseases of the Eye and Its Appendages, London: Longman, Rees, Orme, Brown, Green, and Longmans, 1835, pp. 288, 295–296. 8 John Walker: The Oculist’s Vade-Mecum. A Complete Practical System of Ophthalmic Surgery, London: Brown, Green, and Longmans, 1843, p. 76. 9 Florence Nightingale: Notes on Nursing. What It Is, and What It Is Not, Boston: William Carter, 1860, p. 66. 10 Differing opinions have been obtained about the precise therapeutic application of the color green. As Stevenson emphasized, “there is a very essential difference between looking at and through green.” Stevenson (s. fn. 6), p. 94. While some might recommend green glass for reading, others felt that green glass would attenuate the power of the reflected light rays, leading to eyestrain. However, even those who generally advised caution in using green glasses had to admit that, in cases of very weak eyes, transmitted green light was healthful.
120
Michael Rossi
enced great ocular pain in sunlight. The cure that Stevenson devised for the man included a green shade to be worn at night, as well as a routine of leeches on the eyelids, mercury salts and cathartics – a regimen that (Stevenson reported) eventually cured the man.11 During the American Civil War, nurse Mary Stuart Woolsey evidently took Nightingale’s prescription to heart: she recorded that when she and her fellow nurses prepared hospital rooms, “[w]hite curtains appeared in the windows or green where the light needed softening to the sick eyes.”12 For his own part, upon suffering a nervous attack, the famed naturalist Georges-Louis Leclerc, Comte de Buffon, found that he could no longer tolerate the bright, “yellow” light of the sun, and put up green curtains in his study.13 Beyond the realm of the strictly medical cure, green light – or green shading – provided prophylaxis against eyestrain and injury in vocational as well as everyday use. In 1824, chemist Robert Hare wrote a description of his “deflagrator” – a chemical battery used to incinerate materials at high temperatures – in which the light of “sixteen hundred candle flames” was “condensed within a space no larger than that usually occupied by one.” Gazing upon the concentrated illumination from this device caused him, he said, to experience intense eye pain, followed by bloodshot eyes the next day. He thus cautioned that to avoid the “evil” of eye injury, “the deepest green spectacles” should be used to protect the eye when operating the deflagrator – even to the extent that the experimenter should wear two glasses together if one pair was not enough.14 In a much less dramatic example, the Boston Medical Intelligencer recommended that those who spent considerable time reading – e. g., students or musicians – should place sheets of green glass over the pages that they read in order to ease the strain on their eyes.15 Accountants and clerks, meanwhile, would do best to keep a simple piece of green cloth or a circle of green paper nearby, that they might gaze upon it and rest their eyes periodically. Indeed, even in everyday life, green light could provide a salve against harsh illumination: in his 1837 painting, a “Portrait of Nathaniel Olds,” Jeptha Homer Wade depicted his untidily coiffed sitter looking confidently out at the world through green spectacles,
11 Stevenson (s. fn. 6), p. 45. 12 Jane Stuart Woolsey: Hospital Days, New York: Van Nostrand,1868, p. 47. 13 William Mackenzie: A Practical Treatise on the Diseases of the Eye, London: Longman, Brown, Green and Longmans, 1854, p. 966. 14 Robert Hare: Remarks Respecting Mr. Vanuxem’s Memoir on a Fused Product. In: The Philadelphia Journal of the Medical and Physical Sciences, 8, 1824, p. 351. 15 Anonymous: To Preserve the Eye-Sight. In: The Boston Medical Intelligencer, 4, October, 1825, p. 83.
Green Is Refreshing
1: Portrait of Nathaniel Olds, wearing green spectacles to protect his eyes from harsh light.
121
122
Michael Rossi
designed to protect his eyes against the harsh, white light of the whale oil lamps that illuminated his home (fig. 1).16 Moreover, in addition to the soothing effects of green on the eye, the color also held the possibility of moral prophylaxis – protecting its viewers and their bodily integrity from harmful excesses of passion. As the “Optician” wrote, the purely physiological aspects of green glasses were “but an inconsiderable part Of what by timely use you [i. e., green glasses] can perform; You can preserve in purity the heart, And shield the passions from external harm.” Clarifying the “external harms” that green glasses might protect against, the Optician pointed out that “[…] in this mighty innovating age, to go half-naked, or a little more, Among some topping damsels is the rage. Let modest youths ne’er venture out the door; […] Unless indeed, green-glasses are put on, These, tis a fact, will cast a kind of shade, Over the poor miscalculating maid, And screen the visual organs from the sun” With the zeitgeist so described, and the power of green glass clarified, the Optician could only conclude with an admonishment: “How few with patience listen to advice, How few the value of green-glasses prize: But those who try them once will use them twice: Preserver of the morals, and the eyes!”17
16 Jeptha Homer Wade: Nathaniel Olds, 1837, Oil on canvas; unframed: 76.5 × 61.2 cm, The Cleveland Museum of Art, Cleveland, Ohio. Thank you to Christine Zehner for indicating this painting to me. 17 Optician (s. fn. 5), p. 218.
Green Is Refreshing
123
By extending the logic of visual hygiene to moral hygiene, the same technique by which the eyes might be saved could also be used to moderate the appetites. Peirce’s passing reference to the “refreshing” quality of the color green therefore captured in miniature an array of practices and discourses relating color, sensibility, health, physiology, medical knowledge and moral rectitude. To know green as distinct was to understand these many attributes, and the many ways in which perception and physiology could interact in human sensory experience. Nevertheless, by the time of Peirce’s writing, the mechanism by which a color – any color – could become physiologically and psychologically salient sat at the center of two competing ways of understanding the functioning of the body in relation to sensation. On the one hand, green could be understood as a vector of a sort of vital energy – a force that suffused nature, and which was, therefore, elemental in the growth and healing of living things. On the other hand, it could just as easily be viewed as one electrochemical signal among many others that suffused the human body. As a byproduct of the interaction between the peculiarities of human physiology and the environment, green was, in this view, simply an analytical unit, with no necessary referent to anything save itself. For those who prescribed the healing properties of green light, the mechanism by which the color performed its salubrious effects on the human body was obvious. As Boston physician Edward Reynolds put it, green was the “color which nature, who in all her works seems to have provided with much care for the health and comfort of the eye, has so universally painted the world.”18 Since green was the dominant color of nature, and nature was therapeutic, so then was green light health-giving – not least of all in contradistinction to the artificial light in the offices and manufactories where people increasingly labored.19 “Is it not for this principal reason,” asked Stevenson, “that indulgent nature has so bountifully distributed this cheerful colour through the vegetable kingdom?”20 Nature provided its means of healing, and those means were evident by analogy between that which was natural and that, which was healthful. The Boston Medical Intelligencer affirmed that “green is the universal color [Nature] presents to our eyes” and thus that “Nature confirms the propriety of [the] fact” that green was healthful.21 18 Edward Reynolds: Hints to Students on the Use of the Eyes, Edinburgh: Thomas Clark, 1835, p. 17. 19 On medical and popular responses to industrial lighting in the nineteenth century, see e. g., Peter John Brownlee: Ophthalmology, Popular Physiology, and the Market Revolution in Vision, 1800– 1850. In: Journal of the Early Republic, 28, 2008, pp. 597–626. 20 Stevenson (s. fn. 6), p. 94. 21 Anonymous (s. fn. 15), p. 83.
124
Michael Rossi
Conventional understanding of the therapeutic powers of the color green, therefore, rested on a sort of syllogism: nature was healthful, nature was green, and therefore green was healthful. It was further the case that plants were green, and plants – exhibiting growth and reproduction, but lacking sensation or movement – were exemplary of life stripped down to its purest essence. Such a correspondence further cemented the idea that the color green held as its correlate a variety of “vital energy,” or “life force.” As physician Joseph LeConte explained, “it is generally supposed that there is a direct and immediate conversion of light into vital force in the green leaves of plants.”22 This “vital force” could take a number of forms – from a principle of growth as with plants, to the essence of locomotive force in animals, to the key component of vision itself. What was clear was the connection between the color green and some form of life energy. As LeConte wrote elsewhere, “sunlight falling on green leaves, disappears as light and reappears as life.” 23 The equation between sunlight, the color green and life itself was plain to see. This was an appealing line of reasoning, but it gestured at a correspondence between color, sensation and physiology without providing an analysis of their relationship or a mechanism of their connection. In the case of the refreshing qualities of green, this correspondence did not explain “the meaning of the thing”: it did not explain how green came to be refreshing for those who could distinguish it. It only repeated that green was refreshing. For this purpose, “vital force” was as indistinct to the physiologist as the color green was to Dalton. As one essayist writing on the responses of plants to light scoffed, “an unknown and undemonstrated vital principle has ever been the ready refuge of the puzzled physiologist. […] Only mischief ensues from the use of it in the loose sense which it can in any way be employed to the causes of physiological phenomena.”24 Vital force could be used to answer any question about living creatures; it was, therefore, indistinct as an explanation. Thus, it was possible to acknowledge – as did an 1856 article in Scientific American – that while human beings were obviously “framed with natures which are influenced by color,” nevertheless “the manner [by which] we are influenced is not
22 Joseph LeConte: The Correlation of Physical, Chemical and Vital Force, and the Conservation of Force in Vital Phenomena. In: American Journal of Science and Arts, 28, 1859, p. 309. 23 Joseph LeConte: The Relation of Philosophy to Psychology and to Physiology. In: The Bulletin of the Philosophical Society of Washington, 12, 1895, p. 35. 24 Anonymous: Review of The Theory and Practice of Horticulture or an Attempt to explain the Chief Operation Gardening upon Physiological Grounds by John Lindley. In: The Athenaeum, 28, 1885, pp. 1460–1461.
Green Is Refreshing
125
yet sufficiently understood.” 25 Little was known of “the laws relating to health and color,” the article continued, concluding that such laws could “only be discovered by observation and experiment” – not by common sense correspondence. This idea – that the best way to understand physiological interactions was by deductive experiment rather than inductive analogy – boiled down to a difference between understanding the effects of colors on human bodies as part of a system comprising many distinct parts, versus understanding color, health and bodies as part of a synoptic whole. It was one, which, moreover, had vigorous advocates among rapidly professionalizing scientific and medical practitioners of late nineteenth-century America. As a young ophthalmologist, Benjamin Joy Jeffries, told the American Ophthalmological Society in 1865, the imperative of physicians in the upcoming century was “to learn, appreciate, and adopt all advances in ophthalmological science” as practiced by “scientific ophthalmologists.”26 This meant “patient investigation and experiment,” which Jeffries took to include measurement of the responses of the eye to various stimuli; mathematical analysis of the spectrum and optical effects; and the systematic investigation of the physiological mechanism of vision. It also meant discarding old truisms in favor of what Jeffries termed “new truths” – new ways of understanding the relationship between vision, cognition and human bodies. As Jeffries summed up the matter, “the science of philosophy must henceforth give place to the philosophy of science” – that is, conventional wisdom based on traditional ideas about nature must yield to empirical facts based on methodical analysis. Where the color green was concerned, this meant abandoning analogies between green light, health and the natural world, and instead focusing on the ways in which light of different energies acted on the eye. When looked at in this manner, the foundations for believing green to be therapeutic seemed shaky at best. As the celebrated Austrian ophthalmologist, Karl Stellwag von Carion remarked, “we may say that orange-yellow and green excite the sense of color more powerfully and […] irritate more severely than the other colors.” This could be shown, he continued, “not only by means of the subjective sensation which becomes manifest by looking through different colored glasses, but also by a measurement of the reaction of the pupil” to light of different colors.27 It was therefore the case, concluded von Carion, 25 Anonymous: The Effect of Color upon Health. In: Scientific American, 2, August, 1856, p. 376. 26 Benjamin Joy Jeffries: Report on the Progress of Ophthalmology During the Past Year. In: Trans. Am. Ophthalmic Soc., 1, 1865, pp. 2–12, 4. 27 Karl Stellwag von Carion: Treatise on the Diseases of the Eye, Including the Anatomy of the Organ, New York: W. Wood & Company, 1868, p. 19.
126
Michael Rossi
that “[t]he much recommended green curtains are not so good since in transmitted light they show a very light coloring and cause pain in the eye.” The same was true of lampshades and eyeglasses: “green glasses are not advisable,” wrote von Carion, “since in a bright light, they transmit a very intense and very dark green, approaching a yellow, which increases rather than diminishes the irritated condition.”28 Rather than green glass, von Carion recommended blue, since “the blue color exerts a less severe impression on the retina.”29 British physician John S. Wells agreed, explaining that both the logical analysis of color theory and empirical experiment suggested that blue was a more soothing color than green. “It was formerly supposed,” he wrote, “that the red rays of the solar spectrum were the most trying to the eye, and consequently green glasses which exclude the red rays were much in vogue. But it is now a well-known fact that it is not the red but the orange rays which are irritating to the retina and as blue excludes the orange rays this is the proper colour for such spectacles.”30 For his part, Jeffries similarly rejected both green and dark gray spectacles, asserting that “[s]moked glasses and green do not give the needed protection, ease, or rest to the eye.” He, too counseled his patients to use blue glasses, but cautioned that any sort of treatment should be carefully vetted by a physician, rather than the patient or the supplier of colored lenses. “The special shade of blue which should be worn,” he insisted, “must be decided by each individual case, by the surgeon, and not by the buyer or seller.”31 It was no longer the case that color could be simply known and felt by the seer – rather, color and care of the eye became the responsibility of physicians specially trained in the science of color and of the physiology of the seeing body. It was not, therefore, that color became unimportant to scientists, physicians and their clients – far from it. Color perception was an essential way of exploring the electrochemical workings of the human sensorium. To take just one example, in his towering Principles of Psychology (1890), Peirce’s friend William James reprinted the work of Charles Féré – a French physiologist who had meticulously documented the effects of red, green and yellow light on the circulation, respiration and physical
28 von Carion (s. fn. 27), p. 7. 29 von Carion (s. fn. 27), p. 20. 30 John Soelberg Wells: A Treatise on the Disease of the Eye, London: Lindsay and Blakiston, 1869, p. 545. 31 Benjamin Joy Jeffries: The Eye in Health and Disease. Being a Series of Articles on the Anatomy and Physiology of the Human Eye, and Its Surgical and Medical Treatment, Boston: Alexander Moore, 1871, p. 112.
Green Is Refreshing
127
2: Responsiveness of the body to light of different colors.
strength on some of his patients (fig. 2).32 Féré reported that in all cases, red was the most stimulating of colors and yellow the least stimulating, while green was in the middle.33 This certainly suggested that color had some sort of bodily effect on those who experienced it. Nevertheless, the precise nature of the particular effects of colors in all cases required much more research – more analysis – in order for anything decisive to be stated about their therapeutic possibilities. “The task of tracing out all the effects of any one incoming sensation has not been performed by physiologists,” explained James – and therefore Féré’s work had to be taken as suggestive, rather than decisive. Green might have an effect on human physiology – as might other colors – but what those effects were, and whether they were consistent had to be much more carefully researched before they could be declared salient, much less therapeutic.
32 Charles Féré: Note sur les Conditions Physiologique des Emotions. In: Revue Philosophique, 24, 1887, pp. 561–581. 33 William James: Principles of Psychology, Vol. 2, New York: Henry Holt & Co., 1890, pp. 374–377. Thank you to Lily Huang for this reference .
128
Michael Rossi
Perhaps ironically, then, even as the segment of healing practitioners who subscribed to this analytical approach to understanding the human body in health and sickness grew in number and power – eventually becoming the monopolistic medical orthodoxy that exists to the present day – the study of the healing power of color continued and even thrived among unorthodox practitioners around the turn of the century.34 This practice – called “chromopathy” or “chromotherapy” – maintained the speculative correspondences and vital forces of green light therapies, while expanding the types and varieties of color treatments, and the therapeutic value of different colors. In a sprawling treatise entitled The Influence of the Blue Ray of the Sunlight and of The Blue Colour of the Sky (1876), for instance, retired Union general Augustus Pleasonton reported in detail on experiments that he’d undertaken with blue light over the previous fifteen years. These experiments proved blue was healthful and curative for a wide variety of occasions and ailments: among others of his findings, he reported that plants grown in a greenhouse that he’d specially constructed using panels of blue glass grew stronger and heartier; a mule that had become deaf in its old age regained its hearing after twice daily exposure to sunlight filtered through blue glass; a friend’s wife – sick and near death – was restored to full health after daily baths in blue light.35 Offering a theologically inflected rejoinder to Pleasanton’s work, physician Seth Pancoast penned an 1877 monograph on Blue and Red Light, or Light and its Rays as Medicine, which imparted an enthusiastic and rambling recounting of the links between human health, the vital energy of light and color and the spiritual rectitude of a loving, Christian God. Insisting (in the very title of the work!) that “light is the original and sole source of life, as it is the source of all the physical and vital forces of nature,” Pancoast wrote that all colors – rather than just blue or green – had therapeutic effects; it was necessary to look to ancient philosophy as well as modern physiology to understand how best to employ these lights therapeutically.36 In a slightly more measured register, Edwin Babbitt’s Principles of Light and Color (1878) 34 On the professionalization of medicine in America see Paul Starr: The Social Transformation of American Medicine, New York: Basic Books, 1982. 35 Augustus James Pleasonton: The Influence of the Blue Ray of the Sunlight and of the Blue Color of the Sky, in Developing Animal and Vegetable Life. In: Arresting Disease and in Restoring Health in Acute and Chronic Disorders to Human and Domestic Animals, Philadelphia: Claxton, Remsen & Haffelfinger, 1876, pp. 9, 13. 36 Seth Pancoast: Blue and Red Light, or, Light and its Rays as Medicine: Showing that Light is the Original and Sole Source of Life, As it is the Source of all the Physical and Vital Forces of Nature, and that Light is Nature’s Own and Only Remedy for Disease, and Explaining how to Apply the Red and Blue Rays In Curing the Sick and Feeble: Together with a Chapter on Light In the Vegetable Kingdom, Philadelphia: J. M. Stoddart & Company, 1877.
Green Is Refreshing
129
3: Edwin Babbitt’s thermolume. The patient would sit inside the box and be “bathed” with healing light.
gave a digest of Pleasonton and Pancoast’s work, as well as a compendium of Babbitt’s own research. With a correct understanding of the nature of vital forces, Babbitt wrote, color could be used to cure everything from meningitis and “violent hemorrhaging of the lungs,” to neurasthenia and “general nervousness” among many other ailments.37 In addition to writing about color, Babbitt also designed a number of “chromopathic instruments,” including a “chromo lens” for directing colored lights onto particular parts of the body, and a “thermolume” for bathing patients in colored lights (fig. 3).38 As the authors of these works made clear, they valued science and scientific practice, but they wrote and researched in explicit opposition to the analytical epistemology espoused by Jeffries and his peers. Pancoast, for instance, dedicated his book to “true scientists” whose “minds are not clouded by preconceived notions and theories,” but who were “ready and anxious to learn all that maybe be learned from the ancient philosophers as well as from the investigation of the phenomena of light” – in other words, not Jeffries and his peers. Babbitt was even more pointed, 37 Edwin D. Babbitt: The Principles of Light and Color. Including Among Other Things the Harmonic Laws of the Universe, the Etherio-Atomic Philosophy of Force, Chromo Chemistry, Chromo Therapeutics, and the General Philosophy of the Fine Forces, Together with Numerous Discoveries and Practical Applications, East Orange, NJ: College of Fine Forces, 1878. 38 Edwin Babbitt: Sun and Vapor Bath, United States Patent 408204A, filed September 6, 1886, issued August 6, 1889.
130
Michael Rossi
wondering plaintively if he was “laboring under a delusion, when [he asserted] that no definite conceptions of atomic machinery, or the fundamental processes of thermal, electric, chemical, physiological, or psychological action have been attained” by science? 39 He was not deluded, he decided, and mused that while he might be attacked by the “crystallized conservatives” of establishment science who refused to acknowledge the “thousand facts” that bolstered his research, and “clear[ed] up so many mysteries,” he nevertheless welcomed such a confrontation as a necessary step on the path to unlocking the true healing potential of color. In the early twenty-first century, we are inheritors of this split understanding of color and health. On the one hand, work in chromotherapy has continued from the twentieth century to the present: from colored rooms used to cure “nerve cases” after the first World War; to twentieth-century treatises describing the proper colors that prisons and schools ought to be painted in order to keep inmates calm and students alert; to New Age self-help books in the twenty-first century that come complete with colored pages for channeling healing colors to body and soul.40 On the other hand, it might still be said, as lighting specialists Matthew Luckeish and A. J. Pacini complained in 1926, that “too much benefit has been ascribed to the effect of various colors used as therapeutic agents without adequate support from the research laboratory or the clinic.” As a result, “charlatan exploitation of chromotherapy by medical imposters has done much to dampen scientific interest” in the therapeutic effects of colored lights.41 Luckeish and Pacini’s lament focused on the unsustainable claims about the healthful properties of light – and in this, they were partly right. But rather than seeing chromotherapists as imposters, charlatans and exaggerators, the present study suggests that it is more productive to consider their healing practices as based on an idea of color that was incompatible with that of their peers. Green was refreshing for nineteenth-century scientists whose 39 Babbitt (s. fn. 37), p. vii. 40 On color experiments in Maudsley Hospital see E. N. Snowden: Report on the Kemp Prosser Colour Scheme. In: The Lancet, 193, March 29, 1919, pp. 522–23. On prison colors and so-called “Baker-Miller Pink” see. e. g., A. G. Schauss: Tranquilizing Effect of Color Reduces Aggressive Behavior and Potential Violence. In: Journal of Orthomolecular Psychiatry, 8, 1979, pp. 218–220. On schools and other building environments, see Frank H. Mahnke: Color, Environment, and Human Response. An Interdisciplinary Understanding of Color and Its Use as a Beneficial Element in the Design of the Architectural Environment, Hoboken, John Wiley & Sons, 1996; and Joan Treichel-Arehart: School Lights and Problem Pupils. In: Science News, 105, April, 1974, pp. 258–259. For examples of color as self-care, see, e. g., John Diamond: The Diamond Color Meditation: Color Pathway to the Soul, Ridgefield, CT: Enhancement Books, 2006; and Dougall Fraser: Your Life in Color, Carlsbad, CA: Hay House, Inc., 2017. 41 Matthew Luckiesh, A. J. Pacini: Light and Health; a Discussion of Light and Other Radiations in Relationship to Life and to Health, Baltimore: The Williams & Wilkins Company, 1926, p. 182.
Green Is Refreshing
131
understanding of the relationship between color and the body was predicated on the meaning of green as a synoptic analogy for nature on the whole. This idea could not coexist with the new meaning of green – not as a color of nature, but as a measurable and largely non-referential electrochemical response to a particular stimulus. The two meanings of green were, as Peirce might say, irreconcilably distinct.
Ian Lawson
Pigments, Natural History and Primary Qualities: How Orange Became a Color A physiological table of colors In 1686, the naturalist, watercolorist, translator and philosopher Richard Waller († 1715) published a “Physiological Table of Colors.”1 It was an innovative moment in the history of color printing. He had designed a set of glass pipes stopped with spring-loaded plugs, to be filled with colors. When the plugs were moved aside “the color came down so as to make convenient round spots on the paper” (fig. 1).2 As long as the weights of the pigments in the pipes were kept constant, Waller reasoned, different versions of the table would be identical. Each round spot was labelled with Latin, Greek, French and English names, and the table would thus stabilize nomenclature for these specific shades of color. A naturalist could describe a young plant’s “sea green” or “vitreus” leaves and a reader with the same catalogue in front of them would understand the object, Waller hoped, “with less ambiguity, I think, than is usual: A Standard of colors being yet a thing wanting in Philosophy.”3 This table lies at the intersection of natural history, linguistics, painting and philosophy, and represents Waller’s particular way of thinking about color. In this paper I draw connections between this table and contemporary seventeenth-century ideas about primary colors. I argue that Waller’s table visualizes color mixing in a way that gives us an insight into the history of this idea, and of one secondary color particularly: orange. Orange is conspicuous by its absence in Waller’s table, and in other seventeenth-century color theories. Indeed, originally referring to the fruit, the emergence of the word “orange” as a color term has been dated to roughly this time,
1 Richard Waller: A Catalogue of Simple and Mixt Colours, with a Specimen of Each Color Prefixt to Its Proper Name. In: Philosophical Transactions, 16, 1686–1692, pp. 24–32. For more on the history of the table and its contents, see Sachiko Kusukawa: Richard Waller’s Table of Colors (1686). In: Magdalena Bushart, Friedrich Steinle (eds.): Color Histories, Science, Art, and Technology in the 17th and 18th Centuries, Berlin: De Gruyter, 2015, pp. 3–24. 2 Thomas Birch: The History of the Royal Society of London for Improving of Natural Knowledge, vol. 4, London: A. Millar, 1756–1757, p. 480. As far as I know, there are three surviving colored copies, one each in the Trinity College and Cambridge University libraries in Cambridge, England, and one in the Smithsonian Museum, USA. 3 Waller (s. fn. 1), p. 25. The idea of using color reference chips in scientific fieldwork is familiar to us now through the Munsell Color System. Waller belongs to its ancestors. For Munsell see e. g. Rolf G. Kuehni: The Early Development of the Munsell System. In: Color Research and Application, 27, 2002, pp. 20–27.
134
Ian Lawson
1: Richard Waller’s catalogue of pigment colors, labelled in Latin, French and English.
and Waller’s way of approaching color allows us to trace some of this development.4 I begin with some brief context before moving on to the table and the discussion of primary colors in the seventeenth century. Waller was responding to the troubled but crucial place that color held in Renaissance and Early Modern natural history. In this time, at least until Waller’s contemporary John Ray (1627–1705), the task of classifying plants and animals was not thought to be one of developing a merely conventional system but of carving nature at its joints, finding differences between the essences of God’ s creatures based on appearances and hidden qualities. The senses were key diagnostic instruments for the natural historian, who would not only look at but also smell, touch and taste specimens in order to identify and classify them. Color was an outstanding quality: its subtle gradations and immense variation were immediately apparent to the eye.5 But this variation could also be a problem. There was a constant worry 4 Carole Biggam: The Development of the Basic Colour Terms of English. In: Alaric Hall, Olga Timofeeva, Agnes Kiricsi and Bethany Fox (eds.): Interfaces Between Language and Culture in Medieval England. A Festschrift for Matti Kilpiö, Leiden: Brill, 2010, pp. 231–266. 5 For the use of color in classification, see especially David Freedberg: The Failure of Colour. In: John Onians (ed.): Sight and Insight: Essays on Art and Culture in Honour of E. H. Gombrich at 85, London: Phaidon, 1994; Brian Ogilvie: The Science of Describing, Chicago: University of Chicago Press, 2006; Sachiko Kusukawa: Picturing the Book of Nature, Chicago: University of Chicago Press, 2012; Valen-
Pigments, Natural History and Primary Qualities
135
2: Richard Waller’s watercolor of white and black oats.
about whether color was an essential quality of an object or a flickering apparition that only existed in the perception of an observer. This was especially true of changeable colors such as those on doves’ necks, but even more stable plumage and petal colors vary not only across species but sexes, ages and seasons. The practical problems were many. Multiple costly illustrations would have been necessary to depict all the stages of an organism’s life, noted the naturalist Otto Brunfels (1488–1534), and he supplemented the woodcuts in his Contrafayt Kreüterbuch (1532–1537) with descriptions of plants’ colors. But color terms didn’t always translate well across languages or time; a problem that prompted the Italian naturalist Ulisse Aldrovandi (1522–1605) to pore over ancient texts and etymologies in an effort to fix nomenclature for more accurate textual descriptions.6 Others threw their hands up in despair. David Freedberg has traced the expansion of the problem of color to a worry about tina Pugliano: Ulisse Aldrovandi’s Color Sensibility: Natural History, Language and the Lay Color Practices of Renaissance Virtuosi. In: Tawrin Baker, Sven Dupré, Sachiko Kusukawa, Karin Leonhard (eds.): Early Modern Color Worlds, Leiden and Boston: Brill, 2016, pp. 70–108. 6 Pugliano (s. fn. 5).
136
Ian Lawson
3: Isaac Newton’s color mixing circle.
the ability of illustration generally to capture the essence of a thing and has found in Carl Linnaeus (1707–1778), that great linguistic taxonomist, “the roots of one of the great modern paradigms, that of the transformation of general semiosis into a model in which words are the essential bearers of the meaning of things.”7 By the eighteenth century, color – the visual in general – was not essential for classifying the natural world. Ironically, Linnaeus was writing at exactly the time when, following Isaac Newton’s lead, many philosophers, artists and connoisseurs alike were wracking their brains about the classification of color. The concerns for them were not related to its use but rather to its nature. What is color? How many are there? How are they related?8 Waller’s table came before Newton and after Aldrovandi, and addresses both the concerns of the philosopher and the naturalist. He was deeply invested in the use of color in natural history, having produced a series of watercolors of English grasses and wildflowers, possibly intended for a book by John Ray (fig. 2).9 His table was a pragmatic foundation for this work, but also a contribution to the enterprise of classifying color itself; a novel contribution to the growing number 7 Freedberg (s. fn. 5), p. 257. See also his Eye of the Lynx: Galileo, His Friends, and the Beginnings of Modern Natural History. Chicago: Chicago University Press, 2003. 8 See Sarah Lowengard: The Creation of Color in Eighteenth-Century Europe, New York: Columbia University Press (Gutenberg e-series), 2006, chapter 3. http://gutenberg-e.org/lowengard/A_Chap03. html, acc. 10–11–2018. 9 Lawrence R. Griffing: Who Invented the Dichotomous Key? Richard Waller’s Watercolors of the Herbs of Britain. In: American Journal of Botany, 98, 2011, pp. 1911–1923; Sachiko Kusukawa: Picturing Knowledge in the Early Royal Society: The Examples of Richard Waller and Henry Hunt. In: Notes and Records, 65, 2011, pp. 273–294.
Pigments, Natural History and Primary Qualities
137
4: Elias Brenner’s list of pigments useful for miniature painters, labelled in Latin, French and Swedish.
of color diagrams that trickled off European presses in the seventeenth century and burgeoned in the eighteenth. While most authors in the eighteenth century inherited at least some features from Newton’s famous color wheel, first published in 1704 (fig. 3),10 in a sense, the options were more open in the late-seventeenth century. Waller was neither an Aristotelian nor a Newtonian, nor was he the humanist Aldrovandi had been. Waller was self-professedly inspired by a painting manual published by Elias Brenner (1647–1717) (fig. 4). Brenner had grouped specimens of pigments into rough categories; Waller desired to produce a “more Philosophical, and useful one by the addition of some mixt colors.”11 The result was his unique two-axis organization. 10 See also Ulrike Boskamp’s contribution to this volume as well as her book Primärfarben und Farbharmonie: Farbe in der französischen Naturwissenschaft, Kunstliteratur und Malerei des 18. Jahrhunderts, Weimar: VDG, 2009. John Gage’s contention is that color theory following Newton’s Opticks had to grapple with the circular arrangement and complementary colors: John Gage: Colour and Culture: Practice and Meaning from Antiquity to Abstraction, London: Thames & Hudson, 1993, p. 171. 11 Waller (s. fn. 1), p. 24.
138
Ian Lawson
His table thus comes close to depicting the familiar modern primary and secondary color system of red, yellow, blue and purple, green, orange, except it lacks the geometrical axis required to mix red and yellow together to make orange. For this very reason, it is a valuable document of the meaning of color in Early Modern Europe. Drawing on the color knowledge of painters and philosophers to aid naturalists, Waller’s colors are half-abstractions, blurs between the materials in his glass pipes and perceptions in the minds of observers. This ambiguity was a crucial step in the development of the idea of primary colors, and the reason why in Waller’s day orange didn’t appear to exist. The two dimensions of color Waller’s two axes come with complications. Pigments classed as “simple blues” range across the page, beginning with a white and ending with a black. The spots that descend the left edge of the paper are grouped into three categories: a white stands alone at the top followed by “simple yellows” then “simple reds,” the last of which also includes a black. The page is fleshed out with the mixtures of “each of the Simple Yellows and Reds with each of the simple Blews, and these Mixtures give most of the mean colors, viz. Greens, Purples, &c.”12 Several quirks here help uncover what is hidden in key words like “simple” and “mean,” and show that “&c” does not just mean “orange.” The mixtures are divided into five categories. Simple blues and yellows combine to make “mixed greens” and simple blues and reds make “mixed purples.” When simple blues, yellows or reds mix with white, the results are mixed blues, yellows and reds. White had the power to change the purity of a color but not its hue.13 Blacks, on the other hand, Waller simply included with the chromatic colors. The mixture at the bottom right couldn’t appear blacker: a mixture of soot and ink, it is labelled “Niger, Μελας, Black” but categorized as a mixed purple. Also, because one of the two whites is a simple blue, they combine into a mixed blue even though it looks grey and is called “silver color.” More bizarrely, the mixture of the simples massicot (a yellow) and litmus (a blue) makes a color called “Lividus, Black & blew.” This usually described a bruised color but is categorized here as a mixed green and looks pink on the page. In fact, the simple blue “litmose” itself appears pink. If this is the dye extracted from the lichen Rocella, now familiar as a chemical indicator, presumably what was intended as a deep blue has turned red due to the acid in the 12 Waller (s. fn. 1), p. 24. 13 In this sense, Waller seems to be sticking closer to the model of colors described by Avicenna than Aristotle: see Eric Kirchner, Mohammad Bagheri: Color Theory in Medieval Islamic Lapidaries. In: Centaurus, 55, 2013, pp. 1–19.
Pigments, Natural History and Primary Qualities
139
paper.14 It is certainly worth noting here that roughly a third of Waller’s colors are completely unnamed. This is clearly a work in progress and we should not think that all of these foibles reflect a systematic account of color space. Still, some do point to an essential tension underlying the table between color words, the appearances of colors and the materials that produce them. The table captures Waller’s simultaneous desires to theoretically categorize, yet perceptually define mixtures of specific materials. First, he took a range of pigments and sorted them into blues, yellows and reds by appearance. He then mixed them and classed the results into purples and greens due to the appearances of the ingredients rather than the resulting mixtures, and named them all with reference to worldly objects. The result is that appearance and vernacular names can seem to be in conflict with category. But this science was classificatory rather than predictive – Waller did not provide painters with a recipe for a rich purple appearance; he provided naturalists with actual tokens of phenomenal colors. Each color can be seen as an experiment, in the early modern sense of the word,15 and taken together they help to determine what the imposed categories “purple” and “green” actually mean – what a color looks like that is mixed from blues and reds or yellows. In this sense, the pink litmus dot, which from one perspective can be dismissed as a mistake, a whimsy of the passage of time, can also be thought of as capturing the essence of Waller’s chart. For a painter wanting to mix a green, the red spot among the blues is an error, yet for the aim of setting color nomenclature, a naturalist identifying a flower as lividus is a legitimate use of Waller’s table (providing of course that the naturalists he is communicating with have tables of the same age and printed on the same paper). Of course, Waller’s table was and is embedded in a wider color world that broadly agreed on the names of appearances and would not have thought “blue” referred to the color of the litmus spot. Nevertheless, the logic of the table is to determine the meanings of color words by the appearances of pigments, and of “mixed” colors by the “simple” colors from which they are made. What purple looks like depends on which red and which blue are mixed. Where does this leave orange? Waller’s arrangement of reds and yellows precludes them from being mixed together to create a broad category equivalent to his purples or greens. “Orange” appears only in the vernacular description of the simple yellow orpiment (auripigmentum): “Citrinus, Κίτριυος, Orange color.” The table is stuffed with foods: burnt ochre is “carret color,” massicot is “limon color,” and red
14 Robert Boyle, whom we meet below, was working on chemical indicators during this time. 15 See especially Lorraine Daston, Elizabeth Lunbeck (eds.): Histories of Scientific Observation, Chicago: University of Chicago Press, 2011.
140
Ian Lawson
lake mixed with blue bice gives a “peach color.” Orange was no abstract category to be created by mixing; it was a variety of yellow.16 It most likely became abstracted to fill the need for a mixed hue between red and yellow in the new philosophy of the seventeenth century. To bring out this point I will first go back to earlier ideas of primary and simple colors. Are all colors equal? There are various ancient, medieval and early modern ideas that distinguished some colors as “simple,” “primary” or “exquisite”: more basic or privileged than others in some way. Some ideas particularly were important ingredients for seventeenth-century thought. Until roughly the beginning of the seventeenth century, most philosophical works preferred the Aristotelian idea that the only genuinely simple colors were black and white and that all chromatic colors were mixtures of these (for an example see fig. 5).17 This was a strictly metaphysical idea, not intended to explain mixtures of pigments or colored light, and no Aristotelian thought it should be possible to decompose a colored body into monochromatic parts. Chromatic colors were real and irreducible qualities of bodies, but not simple or primary.18 More practically, ancient and medieval painters typically had not mixed pigments together, leading to an elevation of red, yellow and blue as the most important colors as they were the most vibrant ones available.19 As artists in the Renaissance experimented more with mixtures, painters’ manuals began describing their materials as simple or mixed. In 1634, John Bate explained in a recipe and secrets book: “colors are either simple or compounded, merely tinctures of vegetables or substances of minerals, or both: the simple colors are such as of themselves, being tempered with the water or oyle, doe give a color. The compounded are such, whose
16 Waller’s use of the hue names as superordinate terms for specific pigment colors is consistent with the way such terms were used in sample books of dye colors in preceding centuries. See Carole Biggam: Political upheaval and a disturbance in the color vocabulary of early English. In: Carole Biggam, Christian Kay (eds.): Progress in Colour Studies: Language and Culture, Amsterdam: John Benjamins Publishing Company, 2006, pp. 159–179. 17 See Aristotle: On Sense and the Sensible, book 1 parts 3 and 4. Aristotle privileged some chromatic colors over others on the basis of the expressibility of their ratio of light and dark in integers and their pleasingness to the eye, but not on the basis of their simpleness. 18 For more on hierarchies of qualities in Scholastic thought see Robert Pasnau: Scholastic Qualities, Primary and Secondary. In: Lawrence Nolan (ed.): Primary and Secondary Qualities: The Historical and Ongoing Debate, Oxford: OUP, 2011, chapter 2. 19 Marcia Hall: Color and Meaning: Practice and Theory in Renaissance Painting, Cambridge: CUP, 1994; Gage (s. fn. 10).
Pigments, Natural History and Primary Qualities
141
ingredients do exceed the number of one.”20 The description fits Waller’s “simple” and “mixt” colors well. Seventeenth-century philosophers like Waller and his colleagues at the Royal Society of London began paying close attention to artisanal practice and knowledge for clues about the causes of things, including color. Out of a centuries-long interplay of philosophy and practice came the first whisperings of the idea that red, yellow and blue are primaries in the sense we understand today – privileged due to their fundamental place in color mixing. It was the English natural philosopher Robert Boyle (1627–1691) who gave the idea its clearest early modern formulation. In 1664, he wrote that painters “need imploy [no] more than White, and Black, and Red, and Blew, and Yellow” to produce the “almost Numberless differing colours that are to be met with in the Works of Nature, and of Art.”21 Boyle was well aware that painters used many more than five pigments, but offered this practical knowledge as a heuristic to the new generation of non-Aristotelian natural philosophers concerned with finding the causes of colors. If there were only a limited number of “Simple and Primary Colours” from which the others could be mixed, then the philosopher need only seek the same number of differences in nature – in arrangements of atoms or chemical compositions (Boyle remained deliberately agnostic) – in order to explain the causes of all colors.22 Three different causes would produce three different colors that would mix in the eye to allow the perception of any other. Boyle’s friend Robert Hooke (1635–1703) took this approach one year later in his Micrographia. Hooke explained the way in which pulses of light from the sun were disturbed by reflection or refraction, causing changes that resulted in them being perceived as either red or blue.23 His theory allowed for only these two possible modifications to white light, and all other colors he explained as mixtures of these as they struck the retina. Both primaries could be either dark or pale depending on the strength of the light, and he described yellow as a pale red. The dark primaries mixed to make purple, the paler ones to make 20 John Bate: The Mysteryes of Nature and Art: Conteined in Foure Severall Tretises, London: Ralph Mab, 1634, p. 120. 21 Robert Boyle: Experiments and Considerations Touching Colours, London: H. Herringman, 1664, pp. 219–220. 22 Boyle, (s. fn. 21), p. 220. Anna Marie Roos has identified a Paracelsian strand to early modern English thought on this that encompassed Boyle and Hooke: Anna Marie Roos: The Saline C hymistry of Color in Seventeenth-Century English Natural Philosophy. In: Tawrin Baker, Sven Dupré, S achiko Kusukawa, Karin Leonhard (eds.): Early Modern Color Worlds, Leiden, Boston: Brill, 2016, pp. 274– 300. 23 Robert Hooke: Micrographia, or Some Physiological Descriptions of Minute Bodies Made by Magnifying Glasses, with Observations and Inquiries Thereupon, London: J. Martyn and J. Allestry, 1665, observation 9.
142
Ian Lawson
green. Orange was nowhere to be found. Hooke found evidence for his theory in his experiences with certain substances. Saffron, for example, would produce a deep red when applied thickly and a yellow when applied more thinly. Hooke assumed this to be the rule not the exception and sought an explanation of why it was not the case for all red pigments.24 Hooke and Waller were also good friends and colleagues at the Royal Society, and Sachiko Kusukawa has suggested a relationship between Hooke’s theory and Waller’s table on account of their treatments of red and yellow.25 Indeed, it is not only their friendship that makes this plausible but the way they both move back and forth between considering colors as pigments and as abstract hues. In this aspect, they are also consistent with other writers who approached the primary triad in the early seventeenth century. For all his modernity and clarity, it is Boyle who is the odd one out. Earlier notions of primary colors have been identified in the works of several early seventeenth-century thinkers. Louis Savot (1579–1640), Anselm de Boodt (1550–1632) and François d’Aguilon (1567–1617) are the usual suspects.26 All of these thinkers have one foot in Scholastic natural philosophy and one in a more practical field, and all either have some experience in or connection to visual arts. At least two are directly connected to the English mid-century thinkers: Boyle directly referenced de Boodt and Hooke owned copies of both de Boodt’s and d’Aguilon’s books.27 Much as eighteenth-century authors would struggle to reconcile Newton’s color theory with painting practice, all three also discuss color as a mix of Aristotelian metaphysics and pigment mixtures. None of them claim straightforwardly that there are three primary colors that mix to create three secondary. 24 The explanation remained speculative, to do with the size of the particles of pigments and their susceptibility to being ground down: Hooke (s. fn. 23), observation 10. 25 Kusukawa (s. fn. 1), p. 8. 26 Guido Antonio Scarmiglioni (ca. 1555–1620) is sometimes added to this list, but his final idea of colors is almost identical to Aristotle’s. See his Vidi Antonii Scarmilionii fulginatis, de coloribus libri duo, Marburg: Egenolphus, 1601, especially pp. 110–117. Comprehensive investigations of any of these thinkers’ color ideas are still lacking; however, they have all been discussed several times in papers that include valuable details about their contexts and works: Charles Parkhurst: Aguilonius‘ Optics and Rubens‘ Color. In: Nederlands Kunsthistorisch Jaarboek, 12, 1961, pp. 35–49; Charles Parkhurst: A Color Theory from Prague: Anselm de Boodt, 1609. In: Allen Memorial Art Museum Bulletin, 29, 1971, pp. 3–10; Charles Parkhurst: Louis Savot‘s ‚Nova-antiqua‘ Color Theory, 1609. In: J. G. van Gelder, Joshua Bruyn (eds): Album Amicorum J. G. van Gelder, The Hague: Martinus Nijhoff, 1973, pp. 242–247; Gage (s. fn. 10); Alan Shapiro: Artists’ Colors and Newton’s Colors. In: Isis, 85, 1994, pp. 600–630; Rolf G. Kuehni: Development of the Idea of Simple Colors in the 16th and Early 17th Centuries. In: Color Research and Application, 32, 2007, pp. 92–97. 27 The auction catalogue of Hooke’s library at his death is online at www.hookesbooks.com, acc. 10–12–2018.
Pigments, Natural History and Primary Qualities
143
5: François d’Aguilon’s color mixing diagram. Two sorts of mixtures are represented here. The lines above show red, yellow and blue mixing with white and black and are unlabelled, implying the hues remain unchanged. The lines below show the central colors mixing with one another, producing the secondary hues.
Savot, a Parisian architect, added red and blue to Aristotle’s black and white to claim that there are four “simple colors” and that all others are composed of these.28 Red, though, he called a “double” color: one “thick” that gave shades of brownish-red to pinkish flesh when mixed with white, and one “thin” which when similarly mixed ran from saffron, through “orangé” to pale yellow.29 When he discussed mixtures, he explained that yellows, blues and blacks when mixed create various greens, and reds, blues and blacks create violets and purples. Similar to Hooke, he didn’t need to mix yellow and red for orange – all three were simply shades of the same color. De Boodt, physician to Emperor Rudolph II and a naturalist and painter, listed six “exquisite” colors not produced by mixture: white, black, blue, yellow, crimson and scarlet.30 Here, yellow is separate, but again a color we would likely call “red” these days is divided into two: ruber, translating roughly as “ruddy,” tends towards purple, and miniatus, which likely refers to the pigment red lead, tends towards yellow. De Boodt listed several mixtures that produce further colors, but never ruber or miniatus with yellow. Confoundingly, they themselves mix to produce purple.31 Finally, d’Aguilon wrote a textbook on optics for Jesuit education. It stands slightly apart from the other two works. For one thing, he included a diagram (fig. 5). For another, although he repeated the Aristotelian idea that black and white are the fundamental “extremes,” he posited three “central” colors, red, yellow and blue, to make five “simples,” and listed three first-order mixtures: purpureus, viridis and aureus, the last of which is a golden-yellow.32 As he continued, he clarified his terms and also somewhat the issue of the double-red in the other authors: 28 29 30 31 32
Louis Savot: Nova, Seu Verius Nova-Antiqua de Causis Colorum Sentential, Paris: Perier, 1609, p. 6. Savot (s. fn. 28), p. 8. Anselm de Boodt: Gemmarum et Lapidum Historia, Hanover, 1609, p. 25. De Boodt (s. fn. 30), p. 25. François d’Aguilon: Aguilonii Opticorum Libri Sex, Antwerp: Plantin, 1613. p. 38.
144
Ian Lawson
“Only those species of simple colors are to be understood that our mind understands as separated from matter; not indeed those we see in things. […] Indeed, for example, redness is possessed by the lacca factitia [red lake probably from the cochineal insect], cinnabar, and […] minium (red lead). While this quality inheres in them, not all of them possess simple redness. Lacca has in a certain measure a mix of blue, with which it imitates purple; red lead has a bit of yellow, through which it tends towards gold; cinnabar, placed in the middle place, appears to approach closely to an exquisite redness.”33 D’Aguilon had distinct ideas about the phenomena that color words refer to, independent of their instantiation by particular materials. This meant he could call something a mix of red and blue even if it was composed of one ingredient, quite unlike John Bate’s description quoted above. This was the kind of definition that had given rise to the ambiguity that made color such a problem for natural historians, and which Waller’s table was designed to fix. What does “simple redness” look like: a poppy, a pohutukawa, a robin’s breast? How do you know the quality exists if it is not expressed by a certain colorant? D’Aguilon found it in cinnabar, but the words of Savot and de Boodt imply this was not obvious to everyone. Waller thought in the opposite direction. He fixed the meanings of color words by linking pigments with worldly objects, not by describing pigments with abstract qualities. All of his hue categories encompassed a range of appearances rather than referring to a “simple” or “exquisite” shade, so that what purple(s) looked like were determined by the specifics of available red and blue materials (as de Boodt and Savot also realized). This way of fixing meaning privileges the appearances of materials that are found in the natural world, and this is the key to the orange missing from so many theories of the day. In Waller’s nomenclature, “mixt colors” were not metaphysically mixed as Aristotelians thought, or possibly mixed, as we think of secondary colors today. They were actually mixed from colorants. Similarly, “simple colors” were not those that cannot be mixed from other colors; they are those that need not be. This included orange. There are, available from sources without mixing, a mass of yellow, red, pink, rust, russet, gold, brown and orange pigments. There are a few good stable blues. There were no purples and, depending on your interests, no or very few stable greens, especially that could not be re-described as blue or yellow by someone determined to use as few superordinate hue categories
33 D’Aguilon (s. fn. 32), p. 41. Translation by Rodolfo Garau.
Pigments, Natural History and Primary Qualities
145
as possible.34 When colors are thought of this way, orange isn’t a mixed, secondary color, it’s a simple color available directly from a single natural source, but was classed as a type of yellow. Orange Orange took on its secondary color status and came to be a hue category akin to purple or red simultaneously, as one aspect of the displacement of Aristotelian natural philosophy by a mechanical philosophy that elevated the painters’ main pigments of yellow, red and blue to the position of primaries. But the shift took time. In the painters’ color world, orange was a hyponym of yellow – an answer to the question, what type of yellow is that? Most philosophers in the seventeenth century, from Savot and de Boodt to Waller and Hooke, retained the painterly idea that yellow and red needn’t be mixed together. Materially, this color space was filled by any number of orpiments, ochres and lakes. These belonged to one or the other group but the boundary between red and yellow was so fluid that they themselves were not uncommonly merged into two varieties of the same color. Certainly no separate hue lay between them. The elevation of orange to a broad color group covering a range of specific shades seems to have been due initially to philosophers abstracting away from painterly practice and requiring a new color term to describe the mixture of red and yellow their schemes included. Also, as early as the 1670s, Newton had also identified orange as an “original or simple” color lying between red and yellow in the prismatic spectrum.35 His ideas were notoriously unpopular in the seventeenth century but later, in the eighteenth, he succeeded in focusing attention on color mixing ideas that had their basis in his observations of light, not pigment, and this also very likely played a significant role in stabilizing this shift. In fact, to conclude with a brief discussion on words and language, Boyle’s report of an experiment with light mixtures indicates the uneasy status of “orange” in 1664:
34 See Bate (s. fn. 20) and the notes compiled by Theodore de Mayerne from the painterly wisdom of Rubens, van Dyk, Gentelleschi and others. Mayerne noted that verdigris is only useful as a glaze and green earth is actually blue. Green should rather be made by mixing Dutch pink (a yellow pigment), or yellow ochre with ash blue, white and black. Theodore de Mayerne: Sloane MS 2052, British Library. Available online at http://www.bl.uk/manuscripts/Viewer.aspx?ref=sloane_ ms_2052_fs001r, acc. 10–12–2018. An English translation and publication is available: Donald C. Fels: Lost Secrets of Flemish Painting, Hillsville, VA: Alchemist, 2001. 35 H. W. Turnbull (ed.): The Correspondence of Isaac Newton, volume 1, 1661–1675, Cambridge: Royal Society at the University Press, 1959, p. 98.
146
Ian Lawson
“the [light] Beams trajected through Blew and Yellow [glasses] compos’d a Green, […] an intense and moderate Red did with Yellow make differing degrees of Saffron, and Orange Tawny colors, […] Green and Blew made a color partaking of both, such as that which some Latin Writers call Pavonaceus [peacock-like], […] Red and Blew made a Purple, to which we might add other colors, that we produc’d by the Combinations of Glasses differingly Ting’d, but that I want proper Words to express them in our Language.”36 Boyle stretches to Latin to describe the uncommon mixture of blue and green. Red and yellow is only slightly less troublesome, requiring plants, fruits and leather (tawny) as analogies. The experimental philosophers of the Royal Society were great linguistic innovators,37 and orange is perhaps a prime example. The shift from fruit to color was of course a broad social shift, but here we can perhaps see the role of European science in creating what, in Berlin and Kay’s terms, has become a “Basic Color Term” in English – a term all adult native speakers of the language would be expected to know.38 “Orange” began its career as a color term around the late sixteenth century and went through a long phase as a “transparent” term that referred to the hue by direct analogy with the object, the way we still might use “salmon” or “copper.”39 Waller, Boyle and Hooke all wrote in this way – to them it was still more fruit than color.
36 Boyle (s. fn. 21), p. 223. 37 See Felicity Henderson: Door-Mats and Penumbras, https://hookeslondon.com/2014/03/29/hookeand-english/, acc. 10–12–2018. 38 Brent Berlin, Paul Kay: Basic Color Terms: Their Universality and Evolution, Berkeley: UCP, 1969. Basic Color Terms are a controversial and much-debated idea, and it’s not my intention to wade into the literature here. Berlin and Kay noted, though, that the basicness of “orange” would be doubtful even these days had it not passed their four primary criteria (p. 6). In 1686 it hadn’t – “orange tawny” is not monolexemic. For more on this and tracing the historical development of Basic Color Terms, see Carole Biggam: The Semantics of Colour, A Historical Approach, Cambridge: CUP, 2012. 39 Biggam (s. fn. 4).
Daniel Baum
An Evaluation of Color Maps for Visual Data Exploration Introduction Color is often used in data visualization as an additional cue to support the understanding of the data being visualized. However, care needs to be taken when applying color since it influences our visual perception enormously. Over the past sixty years, the use of color in data visualization – a subfield of computer science – has been studied in great detail. This has led to many generally accepted rules for the use of color. An important aspect when using color is the task to be carried out or the goal to be achieved. A different task may lead to different requirements for the usage of color. When feature detection is the main goal, for example, other color schemes may be more appropriate than those that would be beneficial for an overview of the data. The usability of color is further restricted when three-dimensional (3D) objects are the target of the visualization. In this case, for example, luminance should be avoided to represent differences in the data that is visualized on the surface of the 3D object because the luminance parameter is already required to better accentuate the shape of the object (see the last example in this article). Hence, in order to use color most effectively, many considerations need to be made. In this article, we focus on visualizing continuous scalar-valued data. Such data is usually depicted with the help of color maps that assign to each scalar value a single color. The term “color map” originally referred to a color lookup table that was used in computer graphics to map scalar values to a specific color to be depicted on the computer screen.1 While this article focusses on continuous color maps, different aspects of color for data visualization are considered in other articles of this volume. Also note that the term “color map” might be used differently throughout this volume (see, for example, the article by Jana Moser and Philipp Meyer, and the article by Bettina Bock von Wülfingen). Let us consider a first example. The data depicted in figure 1 is a three-dimensional (3D) image of a marine sediment core containing coral and sea shell fragments. This 3D image was acquired using computed tomography (CT). The 2D visualizations depicted in figure 1 show the same cross-section of this CT image.
1 Kenneth R. Sloan Jr., Christopher M. Brown: Color Map Techniques. In: Computer Graphics and Image Processing, 10, 1979, pp. 297–317; Garland Stern: SoftCel – An application of raster scan graphics to conventional cel animation. In: ACM SIGGRAPH Computer Graphics, 13, 1979.
148
Daniel Baum
1: Visualization of a single slice of a 3D CT image of a marine sediment core containing coral and sea shell fragments. The data is visualized using different color maps.
The differences are only due to the varying color mappings used to visualize it. As a result of applying different color maps, the data might be perceived quite differently, although, at least in this example, the general structure of data can be recognized in all images. Yet, subtle differences can be observed, for example, in the close environment of the coral fragments. These become particularly obvious when we compare the images in the top row with those in the bottom row. In the bottom row, the corals seem to be made up of layers of materials whereas in the top row, we can perceive a continuous variation of the data values. This one example already raises some questions: What is the best color map to visualize some data? What has to be considered when using color to visualize some data? And what pitfalls might appear when the “wrong” coloring is used? This is the type of questions that we want to address in this article. Since the ultimate goal of visual data exploration is to support observers in understanding their data, the effective use of color needs to be measured in terms of two performance characteristics: (i) the time that is needed to understand the data; (ii) the accuracy of the answers that are derived from the visualization of the investigated data. We have carefully studied the previous work on this topic and give a concise summary of the most important conclusions. To underline these statements using examples, we mainly resort to artificially created data since this allows us to clearly demonstrate the effects. Furthermore, we have chosen the examples such that they are simple and reproducible.
An Evaluation of Color Maps for Visual Data Exploration
149
The “correct” use of color in computer-generated data visualization has concerned researchers from the beginning of computer graphics and visualization.2 It was even questioned whether color should be used at all for data visualization.3 As early as 1980, Meyer and Greenberg made a clear statement in favor of using color: “Color has been used to encode the variations of parameters such as temperature and stress. The accuracy with which an observer interprets these images has been questioned,4 but they still remain an effective means for showing data trends.”5 Until today, color remains an important means for visualizing data. And despite the discussions about the accuracy of data perception when using color, even in conservative fields such as medical imaging and image analysis, it has been shown that the use of color is effective and improves the performance of data analysis.6 Thus, the question is not so much whether color should be used at all for data visualization but what the most effective way is to use it. And what aspects need to be considered when using color in order to be effective? As result of the research done over the last 60 years, many rules have been proposed for the use of color.7 In this paper, we focus on data represented as scalar fields f: ℝd → ℝ, in particular two-dimensional (d = 2 ) and three-dimensional (d = 3 ) scalar fields, where to each point in a spatial domain a scalar value is assigned. If the domain is a rectangle or a cuboid, we often call these scalar fields gray scale images, but the domain can also be the surface of a three-dimensional object. Color maps are probably the most important means to visualize scalar fields. However, the design of color maps 2 Richard E. Christ: Review and Analysis of Color Coding Research for Visual Displays. In: Human factors, 17, 1975, pp. 542–570; John M. Booth, John B. Schroeder: Design Considerations for Digital Image Processing Systems. In: Computer, 10, 1977, pp. 15–20; Alan Morse: Some Principles for the Effective Display of Data. In: ACM SIGGRAPH Computer Graphics, 13, 1979; Gary W. Meyer, Donald P. Greenberg: Perceptual Color Spaces for Computer Graphics. In: ACM SIGGRAPH Computer Graphics, 14, 1980, pp. 254–261; Colin Ware: Color Sequences for Univariate Maps: Theory, Experiments and Principles. In: IEEE Computer Graphics and Applications, 8, 1988, pp. 41–49. 3 Booth, Schroeder; Morse (s. fn. 2), see also Aldo Badano in this volume. 4 Booth, Schroeder; Morse (s. fn. 2). 5 Meyer, Greenberg (s. fn. 2). 6 Aldo Badano, Craig Revie, Andrew Casertano, Wei-Chung Cheng, Phil Green, Tom Kimpe, Elizabeth Krupinski, et al.: Consistency and Standardization of Color in Medical Imaging: A Consensus Report. In: Journal of Digital Imaging, 28, 2015, pp. 41–52. 7 Kenneth Moreland: Diverging Color Maps for Scientific Visualization. In: George Bebis et al. (eds.): International Symposium on Visual Computing, Berlin: Springer, 2009, pp. 92–103; Gabor T. Herman, Haim Levkowitz: Color Scales for Image Data. In: Computer Graphics and Applications, 12, 1992, pp. 72–80; Adam Light, Patrick J. Bartlein: The End of the Rainbow? Color Schemes for Improved Data Graphics. In: Eos, Transactions American Geophysical Union, 85, 2004, pp. 385–391; Brand Fortner, Theodore E. Meyer: Number by Colors: A Guide to Using Color to Understand Technical Data, New York: Springer Science & Business Media, 2012; Cynthia Brewer: Designing Better Maps: A Guide for GIS Users, Redlands: ESRI press, 2015.
150
Daniel Baum
is by no means trivial. In the next sections, we review some findings concerning the use of color maps and rules that should be considered when designing them. We underline those findings and rules by applying several commonly used color maps to specifically created artificial test datasets and by evaluating the respective visualizations. Definition of color maps A color map is a function x: [a, b] ⊂ ℝ → C , a