Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Introduction
Overview of factors in human food selection
Economic and ecological factors affecting food
choice
Sensory characteristics affecting food selection
Perception of physiological effects and food
classification
Cultural symbolic dimension
The socio-cultural construction of diets
Summary and conclusions
References
Appendix 1. Obesity and cultural weight
valuations (A variation on the method suggested by Massara, 1980)
Appendix 2. Dietary decision-making: formal
models and ethnographic Qualifications
Appendix 3. Bitter or astringent taste standards
Appendix 4. A field test for estimating
sweetness preferences to improve estimates of sucrose intakes in
individuals
Appendix 5. Methods for describing staple food
classifications
Appendix 6. Symbolic, folkloric, and medicinal
factors
Appendix 7. Gender factors
ELLEN MESSER
Center for Advanced Study in the Behavioral Sciences, Stanford,
California, USA
This chapter will review how anthropologists study socio-cultural factors that influence food intake. There are many different kinds of studies of food intake, including ethnographic, ecologic, economic, biocultural, nutritional, and ethnoscientific approaches. These various kinds of anthropological studies use a wide variety of methods, analyse different aspects of culture and gather different kinds of data.
Ethnographic studies by social and culture anthropologists have always included to a greater or lesser degree some discussion of the food practices of communities. Data on the importance of the food quest in the cultural life-style of a group, the meaning of food for structuring and "realizing" social relations, and the impact of food habits on nutrient intake and health are usually collected by a combination of naturalistic (participant) observations and selective interviews.
More specialized anthropological studies have utilized a wide variety of techniques to collect, analyse, and interpret selected aspects of dietary and nutritional data. Symbolic and cognitive (ethnoscientific) studies, for example, have analysed the internal and external structure of diet in relation to the rest of a culture. The former examine the logical relationships among aspects of a culture; the latter investigate cultural aspects in terms provided by the people themselves. Either approach provides analyses that enable anthropologists to compare the food dimension of a culture to other cultural domains. Common principles for thinking about and behaving in relation to nutrition and health may be arrived at by comparing and contrasting terminologies and activities in both the food and health domains. Such analyses of the food domain of culture may be compared across cultures to illustrate contrasting attitudes toward food, culinary practices, diets, and nutritional well-being.
Ecologic and economic studies within anthropology have considered the relationship of food choices to the foods available in particular environments. Biocultural anthropologists have tried to show how cultural habits, including food beliefs and practices, affect the biological well-being of human populations or social groups in the short run and the evolution of human biological populations in the long run. Conversely, biocultural anthropologists have also attempted to determine the relationships between biological aspects of particular environments or genetic characteristics of particular populations and their cultural beliefs and practices, in order to show interactions and interrelationships between biology and culture over both the short and the long term.
Nutritional anthropologists generally try to explicate the relative importance of these various approaches and factors. Thus, while anthropologists tend to explicate the folkloric factors that contribute to the particular patterns of food acceptance, food preference, and dietary constructions, they are also well aware that material factors play a large part in the selection of foods.
Ecological factors to a large extent determine which foods are available within a culture. Thus, for example, traditional Eskimo societies "choose" a diet consisting largely of fat and meat protein out of ecological necessity, not just for reasons of taste, social symbolism, and similar factors, which are part of their total food culture. Similarly, most low-income people might choose to eat foods other than those that they "select" under their existing economic conditions. Nevertheless, interacting with and beyond these material factors, people tend to select foods for a variety of sensory, cultural, social, symbolic, and health reasons; and anthropologists and other social scientists have developed concepts and methods to study systematically these food habits and their nutritional implications.
This chapter, organized according to categories used by nutritionists, will identify the kinds of information and methods of data collection and analysis relevant to asking and answering questions on human food selection for cross-cultural and intra-cultural comparisons. Methodological sources include studies by anthropologists, nutritionists, and social scientists. The chapter begins with a general overview on human food selection, and then examines ecological and economic factors influencing food choice. The chapter also describes how a better understanding of socio-cultural factors in food selection can contribute to the modelling of, and solutions to, nutrition problems. In conclusion, we will consider some of the problems in moving from individual levels to household and cultural levels of analysis.
Among the initial issues to keep in mind in any study of socio-cultural factors affecting food intake are the interactions between biological and cultural processes in human food selection. Humans accept food items as "edible" or reject them as "inedible" and establish preferences among edible items on the basis of a number of sensory and cultural characteristics. The term "sensory" as used here incorporates psychophysical, cognitive, and affective factors (usually analysed separately by psychologists) that enter into taste discriminations and preferences in food selections. The term "cultural" includes symbolic, social, and economic factors which, in interaction with "sensory" data and preferences, shape food patterns and influence the selection of foods. How different peoples translate biological information about foods (safe versus dangerous; nutritious versus empty calories) into cultural likes and dislikes is a topic that has been treated from almost all anthropological perspectives. So has the related topic of the formation and persistence of a cultural cuisine - a term used to describe "the culturally elaborated and transmitted body of food-related practices of any given culture," which include: (a) the selection of a set of basic (staple or secondary) foods; (b) the frequent use of a characteristic set of flavourings; (c) the characteristic processing (e.g. chopping and cooking) of such foods; and (d) the adoption of a variety of rules dealing with acceptable foods and combinations, festival foods, the social context of eating, and the symbolic uses of foods (Rozin and Rozin, 1981, p. 243; see also Rozin, 1973). These are the terms most commonly used to describe cultural food-related practices by both anthropologists and psychologists, and they offer a standard outline for describing and comparing cuisines cross-culturally.
Within cuisines, people select foods to conform to particular styles of eating, as well as cultural concepts of what items are edible and preferred, and how they should be prepared and served. People eat discontinuously; that is, they structure their food intake throughout the day and also throughout longer periods of time according to cultural convention. In some contexts, they eat whether or not they are physiologically hungry; in others, although they may be hungry and food may be available, they abstain from consumption for social and symbolic reasons. How people in different cultures structure their concepts of hunger, appetite, and consumption has been reviewed in a number of psychological anthropological studies (Malinowski, 1935; Fortune, 1953; Holmberg, 1950; D. N. Shack, 1969; W. Shack, 1971; Young, 1971). In addition, the cultural versus biological bases of hunger and satiety, and implications for corrective diets for both undernutrition and overnutrition, have been debated (Mead, 1964; Ogbu, 1973). This literature includes discussions of the biological and cultural bases of body weight and weight valuations (Belier, 1977; Massara, 1980), and ways to measure body image (especially Massara, 1980). What are considered healthful and beautiful as well as what are desirable body size and proportions vary culturally and can be measured as well as described. Massara (1980) describes one methodology for measuring constructs of appropriate body size and image (Appendix 1).
Furthermore, humans may ingest or feed others items they recognize as harmful. This form of eating behaviour would not be seen in the non-human animal world. Potentially toxic marine food preparations are known, for example, to be delicacies in both traditional Eskimo (Wallace and Ackerman, 1960) and Japanese cuisine. Apart from these deliberate culinary risks are the involuntary instances of food poisoning, heralded in the literature of medieval Europe (Colman, 1976) among others. Each of these cases could be considered examples of maladaptive food habits for the consumers, although, in the last instance, control of food safety might be beyond the grasp of the consumer.
Before evaluating the adaptive or maladaptive nature of food habits, however, one should be careful to consider how food habits are related to short- and long-term nutrition and health, short- and long-term preservation of the food environment, and, finally, short- and long-term perpetuation of a particular society and culture that structures its food environment and ethnobiological understandings of the relationships between food and health in particular ways. Thus, a historical perspective should always be included in any evaluation of the rationality of food habits and this perspective should take into consideration possible changes in ecology, cultural influences, and disease patterns over time and place. Sensory information that once led to "adaptive" dietary pathways may, under other environmental conditions, no longer do so.
Economic and cultural studies have shown how income and food costs determine food selections, and often override considerations of "healthfulness," social desirability, and even taste. Availability of foods in the environment and the relative cost both in time and money of acquiring and then processing particular classes of foods are considerations in any cultural survival strategy. The seasonality of wild and cultivated resources and the problems of scheduling food acquisition activities among a variety of competing activities have been basic factors in the evolution of diets in both the Old and the New World (Flannery, 1973). With the expansion of the modern market system and cash economies, limitations of seasonality have to some extent been overcome. In their place, limitations of cash income and lack of cultural knowledge about nutritional needs and the nutritional contents of foods have become key constraints in the construction of nutritious diets. Similarly, the need to allocate time to activities not directly related to food has influenced food acquisition and preparation.
Where people rely on the market, economic considerations may be foremost in food choices. Even with adequate nutrition knowledge, people may not have the economic means to feed themselves at optimal or even adequate levels. This predicament may be shown in several ways. In their study of food choices on native American reservations Calloway and Gibbs (1976) concluded that inadequate nutrition levels were a result not of lack of nutrition knowledge, but of lack of money. They had respondents describe and rank foods according to nutritiousness and found that the respondents had a clear idea of what a nutritious diet should consist of, but still could not afford such foods. DeWalt and Pelto (1976) asked respondents in a Mexican village to rate foods according to taste, healthfulness, and economic value, and found that people had accurate notions of "nutrition" (or at least, accurate notions of categories analysts identified as a "protein" factor). Their sample, however, chose foods largely according to budgetary considerations. The factors related to food choice in this study were identified through factor analysis from sets of food choices provided by respondents and, therefore, were analytical constructs of the researchers.
The need to pay more attention to economic factors in patterns of food intake also has been noted in general nutrition surveys. Omawale (1980) concluded from a survey done in the Philippines that malnutrition among pre-school children generally was correlated with the ratio of household income to the cost of calories in the diet. Households were likely to eat more expensive calories when they shifted from subsistence agriculture to cash-cropping. He suggested that, with cash-cropping, income must double to maintain "subsistence" levels of calorie intake. Schuftan (1979) has suggested that since the real problem in malnutrition is lack of food-purchasing power, nutrition programmes should address malnutrition in economic rather than nutritional terms; that is, they should measure deficits in the purchasing power of the households and give priority to nutrition programmes that generate income' new employment opportunities in food production, and food-related services.
Certain economists have also suggested that food aid be evaluated as an income transfer. They maintain that nutritional value should be calculated not on the basis of the nutrient composition of the food-aid package, but according to the nutrients that could be purchased for that package's value in programmes where food is not targeted toward particular nutritionally vulnerable groups such as infants and young children. These economists have developed models to predict the additional protein and other calories that should be purchased for given increments in income at particular income levels and commodity prices (Reutlinger, 1982). One might argue that the usefulness of these models is reduced because of the continual price adjustments necessary to use them successfully.
In genera, economists and those anthropologists calculating optimal foraging strategies have used linear programming techniques to predict and evaluate food selections. Anthropologists have often employed energy' protein, and sometimes an additional range of nutrients to understand how people in hunting and gathering economies "optimize" nutritional return, calculated in terms of time and energy rather than money. Johnson (1980) provides an example in Appendix 2 of one methodology for calculating returns for food procurement strategies for a South African horticultural group.
These models have the potential to simulate a range of outcomes with a given set of food resources evaluated according to a given set of objective criteria. Yet they may not prove adequate to predict actual food selection, given that people select foods, not calories or protein units, and given that a wide range of conveniences' taste, and prestige factors enters into the decision-making process (Jochim, 1981). For example, cravings for meat, fat, salt, and condiments may all change the relative ranks of foods in terms of preferences for consumption and the amount of money people are willing to expend for them. Rituals and ritual seasons also command special consideration, since during set times of feasting people consume more nutrient-rich, protein-rich, and expensive foods than usual. On the other hand, ritual seasons of fasting may limit intakes.
Finally, an additional economic factor affecting food selection is the amount of time available to the food provider or preparer- usually the focal female (head decision-maker) of a household. Food providers have been known to calculate the amount of time that must be allocated to procurement and preparation of different foods under various conditions of household organization and cash work. Popkin and colleagues, investigating factors associated with vitamin A deficiency in the Philippines, suggested that children of women traders had lower vitamin A intakes than children of mothers who did not work outside the home because women involved in commerce did not take the time to prepare vegetable sauces rich in vitamin A. This hypothesis, however' remains untested either with ethnographic or dietary data (Popkin and Solon' 1976; Popkin, 1980). Although the nutritional consequences for the household of women's work outside the home may be negative' such valuations must be documented with records of actual intakes.
In summery' ecological and economic factors in human food selection always provide a context within which sensory and socio-cultural factors come into play, but, conversely' sociocultural factors must always be considered in calculating human food selection in economic terms. The following sections describe the social contexts and factors affecting food selection.
At a most basic level, food selection is governed by a number of sensory characteristics' such as taste' smell. texture, colour (and other visual characteristics)' even sound (e.g. crunchiness), and physiologically perceived characteristics, like "fillingness" or "burningness." The exact dimensions of these characteristics, as well as degrees of acceptance or preference for them, differ among and also within cultural populations.
Taste
A liking for sweet taste and a rejection of bitter seem to be innate, perhaps as a result of biological "coding" of "safe" versus "poisonous'' foods. Experiences within particular cultures, however, affect how taste qualities are conceived and labelled, how preferences for degrees of sweet, bitter, and other flavours are formed, and how both of these influence intake. Mexicans, for example' enjoy beverages sugared to a degree that is "too sweet" by most American standards, while Mexicans find the bitterness of unsweetened American coffee unpleasant (Messer, 1986). Some African populations seem to be raised from an early age with a preference for sour tastes, like tamarind (Jerome, 1977), which American children often find puckering.
Most anthropological data on food tastes, preferences, and aversions, while extremely useful for describing the taste component of food preferences and the development of food preferences of a socio-cultural group, are anecdotal. Experimental procedures may consist of little more than observing how children are fed, and then learn to like, the flavours characteristic of a particular area. Alternatively, the experimenter may offer subjects foods identified with a particular flavour or concentration of one flavour, and record whether the subjects eat or reject the experimental food.
In contrast, psychologists, psychophysicists' did chemists, working mainly in laboratory contexts, have developed methodologies to study taste that shoud prove useful to anthropologists as well. Experimental techniques by which to study the taste component of food selection in mammals from rats to humans are described in the literature of taste physiology (Bartoshuk, 1975, 1978; Rozin, 1976; Barker, 1982; Solms and Hall, 1981), and, in some instances, the co-evolution literature within ecology (Swain, 1976). These works outline current psychophysical theory of taste-olfaction, the kinds of experiments developed to test taste function from individual to individual and from group to group, and the procedures that may be used to evaluate taste preferences either within (e.g. by age. sex, residential, or "dietary" groups) or between cultures. Research reports include types and numbers of subjects, preparation of basic taste solutions, administration of experimental materials, analysis, and conclusions, including qualifications.
In addition to providing the nutritionist or anthropologist with a better understanding of the psychophysiology of taste and the kinds of questions and procedures used by psychologists, such articles suggest procedures extremely useful for evaluating the taste component in the formation of food preferences. They also provide the correct standards (in log Molar concentrations) to use in any experimental design. For example, to test for preferences for different sweetness concentrations or for aversions to progressive concentrations of alkaloids or tannins, one should use the standard measures developed by psychophysicists for crosscultural comparability. An example from non-human dietary studies of a methodology to evaluate tolerance for bitter substances is presented in Appendix 3.
Among the biocultural questions that can be posed, if not resolved, regarding taste are those concerning the genetic versus environmental origins of cultural taste preferences and the adaptive value of different taste tolerances and preferences. For example, there seem to be different cultural preferences and tolerances for sweet and salty tastes. It is unclear, however, to what extent they are genetically determined or environmentally developed (Desor, Green, and Maller, 1975). Cultural experience also seems to determine the acceptability of and preference for pungent or "biting" flavours, like chili peppers (Rozin and Schiller, 1980). Of interest to both biologists and psychologists, as well as anthropologists, is the extent to which people are aware of the physiological (nutritional) values of foods, how these values relate to taste (see Rozin, 1982), and how such values affect intake. Messer (1986), measuring preferences for sweetness concentrations within one Mexican town, noted substantial differences within and among households and suggested that such preferences did affect total sugar intake. She first measured sweetness preference by the procedure of "taste matching," then converted sweetness preferences to measures of sucrose intake by multiplying volume of sweet liquid consumed by sweetness (sucrose concentrations). This methodology is described further in Appendix 4.
Tastes for more complex flavours, whether strong like the fermented sauces and pastes of oriental cuisines and the ripe cheeses of Western Europe, or bland like the slightly fermented staple grain porridges of Africa, are also culturally conditioned (and may be judged unpleasant, at least initially' by those unaccustomed to them.)
The development of taste tolerances has enabled people to extend the range of mutritious substances they select from the environment. Among other things, the consumption of microorganisms (yeasts, moulds, fungi), spices, and the products they transform and preserve, have channelled the evolution of particular human populations in particular ecosystems in specific directions (Ritenbaugh, 1978). Yet the mechanisms by which some flavours become acceptable and likes and dislikes are formed have not been well established (Rozin and Fallon, 1981). More precise descriptions within populations of shared tastes, particular subgroups, or groupings by age and sex might help sort out some of these issues.
Also, the cultural component of food likes and dislikes may be clarified by symbolic-linguistic analyses of how flavours of foods (either psychophysical taste and/or culturally extended meanings of flavour) may be linked to concepts of digestion and health. One example is the classical Chinese belief that the five basic flavours have predictable physiological effects by acting in systematic fashion upon particular organs of the body (Veith, 1949). Another example is the use of flavour by certain groups indigenous to lowland South America to classify foods in their environment (Seeger, 1981). Such symbolic studies bring us closer to Levi-Straussian structuralism than biochemical nutrition and psychophysical taste, but they illustrate the range of ways in which nutritional anthropologists can investigate the cultural and biological parameters of cultural taste and their implications for food intake, nutrition, and health, between and within cultural communities.
The use of particular flavours and spices is also one means by which people mark the cultural identity of their foods (Rozin, 1973) and a procedure by which they render unaccustomed foods familiar. The simplest way to analyse such processes is to follow E. Rozin's outline (provided above), and to gather data with observation and/or interviews in order to describe the rules for creating a cultural cuisine. Vietnamese refugees in the United States, for example, experimented with American condiments to create an acceptable substitute for their usual spices (Rozin and Schiller. 1980).
Even within the same region, local cultures distinguish themselves by variations in their spicing patterns; for example, certain Malaysian groups are known for their 'sweet" sauces. At the local level, individual households and household groups may further express their distinctiveness by variations in food preferences and flavouring within the same general pattern of a community. Such information, systematically collected, can lead to understandings of nutritional differences within a community or region, on the one hand, or status differences, on the other.
Additionally, flavouring patterns, including sweetening, are sometimes interpreted as class markers. In American culture, very sweet foods are interpreted by some as a lower-class or foreign phenomenon (Barthes, 1975). In some countries, such as Guatemala, pungent and highly spiced foods are sometimes associated with "less civilized" elements within the nation. The decreasing use of strong tasting or smelling spices like Chenopodium ambrosioides (epazote) and chili peppers among some urban Guatemalans has been interpreted as a way of demonstrating that they are "civilized," since by some these strong flavours are judged suitable only for "less civilized" palates (Messer, field notes from Guatemala, 1979). In nutritional terms, these spices are potentially beneficial, since epazote has effective vermicidal properties and chili peppers provide vitamins A and C.
In similar fashion. the strong odours of certain ethnic cuisines may be rejected as barriers to socializing in the larger culture. Algerian immigrants in Marseilles were reported to have noted that the aroma of their lunches made them conspicuous, and offended other workers, so they arranged to have a food training course for their wives (Autret, 1962).
It should be noted that bland foods characteristic of a "civilized" Western diet may in part be a result of technological factors that encourage efficiency and uniformity in mass production and processing, more than flavour and other subjective qualities (Amerine, 1966). In cases where there has been a move away from traditional foods or varieties of foods to blander, often mass-produced foods, there should be an effort to ascertain the changes in nutrient content, as well as taste.
Texture
Texture is a second essential element in food selection and preferences. In large part, texture determines whether a food is familiar and may influence acceptance of new foods. In Africa, where the basic staple is porridge, ranging from thick to watery, different groups distinguish themselves by the texture they prefer in their staple food. When given a new commodity, they will generally accept it more easily if it can be eaten in forms that are familiar in texture. Rather than record such "textural tastes" in words such as "thin, watery gruel" or "library paste consistency," the researcher should try to record the proportion of flour to water used in the preparation, so as to provide a potential measure of nutrient density should the food portions later be weighed. Such records may also aid in the interpretation of forms of liquid intake by the members of the population, and illustrate - if there are changing proportions of grain to water how homemakers adjust to decreased quantities of food over an annual cycle.
In other cultural cases, the acceptability of the texture of a food may be related to the form of processing. For example, in rice-eating cultures, donated food in the form of grits or bulghur (wheat) may be more acceptable than grains delivered in the form of flours or meals.
The glutinous properties of different varieties of grains such as wheat, maize, and rice may also affect their suitability for producing familiar foods like bread, tortillas, and rice dishes, respectively. Such qualities often affect acceptance of new varieties of foods designed to improve nutrition. For example, the poor adhesive qualities of high Iysine maize developed to help correct the lack of protein in Mexican maize diets was largely rejected by those it was intended to benefit because it did not form a workable tortilla dough. New "miracle" rices have been disparaged by the consumers toward whom they were targeted as having textures that "stick to the throat" (Franke, personal communication). Consumers prefer traditional rice varieties. Even within American culture. whole-grain breads, which have positive "health" connotations, are denser and so may be rejected by some who have developed a taste for the lighter, softer white bread.
For an ethnographic report on such food preferences, it is useful to hold several in-depth interviews with key informants about the typical foods of the culture, their texture and other properties, and their relative ranking from the perspectives of preferability, "healthiness," taste, prestige, and other characteristics. If one has the interest, time, and resources to investigate the distribution of textural preferences within the culture, one can use this preliminary information to set up a systematic interview or rank list. From the scientific perspective, grain breeders or processors may be able to provide evaluations of the adhesive qualities of the particular grains under investigation as well as nutrient analysis of the raw, processed, and digestible portion of the grain as prepared in the culture. With this information one can answer questions concerning the nutritional consequences of particular texture preferences. One can also determine how a particular culture translates the textural properties of grain, as processed within that culture, into a ranked food within the diet.
In addition, certain textural properties are also intrinsic in judgements of qualities like "crispness," "crunchiness," and "freshness," all of which are important in the selection of food items. Barthes (1975) has suggested that there is a general symbolic opposition between "crisp, brisk, and sharp" foods and `'soft, soothing, and sweet" foods in Western cultures. Texture, like flavour, may also be associated with palatability and digestibility. Natives sometimes use these descriptive characteristics in distinguishing among preferred and less preferred foods and those consumed under normal versus starvation conditions (Firth, 1966). Again, the nutritional values of such foods as well as their purported palatability should be determined, where possible, to arrive at a comparison of culturally valued tastes and nutritional value.
Visual Characteristics
Size, shape, and colour are all additional sensory properties of foods that influence food selection and preferences. These aspects of visual appearance often encode information about other taste and textural characteristics by which people judge foods to be more or less appetizing, appealing, or valuable for certain purposes. At the most basic "gut" levels, visual inspection may contribute to acceptance versus rejection of particular items as 'food," as in the usual rejection of items that "look" as well as "smell" rotten, or those that carry symbolic connotations of "abomination," for example, the rejection of foods that resemble genitalia in cultures where such are judged to be "disgusting. "
Visual characteristics, furthermore, often supply the morphological and other criteria for folk classifications for foods such as cereal grains, legumes, tubers, and sauces, and for additional varietal discriminations within these categories. By means of such classifications, people construct diets that incorporate cultural ideas of the nutritional contents of particular food types as well as their other social symbolic meanings.
Colour, as one visual characteristic of foods, often provides a code by which people label and rank varieties within more general "species" of food with particular textural or taste properties. Consequently, colour, because it also encodes other dimensions of cultural value, may influence food selection more than reputed nutritional worth. Throughout much of Mexico, for instance, white maize is preferred for tortillas, since white tortillas are said to look "cleaner," to have softer texture, and to taste better than tortillas made of yellow, red, or purple maize (Messer, 1976b). New, high-nutrient maize varieties have been rejected because of their yellow colour, among other "undesirable" characteristics.
In similar fashion, the combined colour and textural characteristics of rices enter into their varietal classifications, their suitability for particular occasions, and their preferability for particular preparations (e.g. glutinous versus non-glutinous varieties, which may be furthermore coded by colour). A good introduction into the meanings of these terms for biologists is provided in Heiser (1981), although anyone who wants to learn the local values should follow procedures set out by researchers - for example, Conklin (1954) for rice, Messer (1976b, 1978) for maize and maize diet plants, and local ethnographies for other plants (Appendix 5).
Colour is also often a way for cultures to distinguish between varieties of, and indicate preference for, legumes. In folk classifications throughout Latin America, different varieties of beans are thus described, often first by colour, then size, shape, habit of growth, and, at times, place of origin, with different varieties furthermore noted to have distinctive flavour, texture, and cooking qualities (Messer, field notes, Guatemala, 1979). People may also have ideas about the digestibility of different bean varieties coded by colour. Biological terms to describe visual characteristics of beans can be found in Kaplan (1954). Bressani (1983) and his colleagues at the Instituto de Nutricion de Centro America y Panama (INCAP) in Guatemala have been investigating the digestibility factors in beans by colour and type, under different processing regimes. Results of their research can be found abstracted in the Annual Reports of activities of lNCAP.
In addition to providing a typology for food elements, colour may also operate as a kind of "folk index" of purity and refinement, and affect the prestige value of varieties within particular categories of foods. The prestige value of "white" ("refined") versus "black" ("coarse," "unrefined") breads has been a distinction basic to European food habits. White "polished" rices in most parts of the world are preferred because they are more processed and refined, cooked more quickly, more desirable in taste and texture, and consequently, associated with higher cultural status. Yet, white, more refined varieties can, in other cultural contexts, carry lower prestige value, as in the current "health food"-conscious subculture of the United States, which prefers unbleached, less processed grains and breads. Brown eggs are preferred over white by farmers in Ann Arbor, Michigan, by consumers in Boston, Massachusetts, and by Mexican peasants, because they are believed to have yellower yolks and to be tastier. Some American consumers, conversely, prefer white eggs because they are "purer" since they can be more easily candled (screened for impurities). Selective interviewing of persons in different socio-economic divisions within a community will yield valuable information for determining whether such food habits potentially interfere with the introduction of a new food or consumption of a more nutritious older one.
Colour may also provide a sign of expected quality, including ripeness, wholesomeness, and taste in fruits and vegetables. For cosmetic reasons, oranges are coloured orange in the United States. Other fruits are anticipated to taste "green," "ripe," or "overripe" on the basis of, among other qualities, colour. Colour may also be perceived as a sign of decay and inedibility, as in blue or green cheese (unless the cheese is expected to be mouldy), discoloured meats, and spotted or brown fruits and vegetables. But colours that are normally "unappetizing" may not only be accepted but chosen under ceremonial circumstances at certain times of the year. On Saint Patrick's Day in the Boston area, even the beer and the bagels are dyed green! Thus, colours may carry symbolic meanings beyond ordinary health connotations.
Along these symbolic lines, food colourings, like flavourings, may also have customary or ritual values, as is evident in the preference for "red" dishes prepared with achiote, a tasteless "spice" used in Yucatan and Guatemala (which coincidentally increases vitamin A content), or the serving of yellow rice coloured/flavoured with tumeric, an "auspicious" colour, at Javanese weddings (Geertz, 1960). Finally, colour may serve as a code for certain spicing patterns, whether single spices, like yellow tumeric, or combinations of spices, as in Mexican "green," "yellow," "red," and "black" sauces, which may identify at a glance dishes of particular condiment compositions and flavour, and, in certain instances, the special occasions at which they are served. Colour thus provides in various cases a code for "safe" as well as appropriate foods. Nutritional anthropologists try to sort out the terms of this code and its nutritional consequences from cultural and nutritional perspectives.