Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Three Studies of Meal-format Rules

The meal-format level of organization was first recognized by Mary Douglas and the studies analysed below were all influenced by her. They include the work of Douglas (1972), Nicod (1974), and Douglas and Nicod (1974) on British working-class meals, the work of the Russell Sage Project in Gastronomic Categories (Douglas, 1984), and the work of Norge Jerome (1979) on migrant Black households in northern American cities. Meal-format rules include all of the cultural understandings about the relationship between the meal type, the nature of an occasion, and those present at an event. Meal formats are structures for the presentation, distribution, and ingestion of food and are distinguished by the number and types of dishes they include and the way these dishes are presented. Dishes can be presented diachronically - i.e. a series of courses is presented in strict order over time with rules for what can go into each course - or synchronically - i.e. dishes are presented simultaneously in certain spatial arrays. Meal-format rules also specify the manner of serving the dishes - whether it is centrally controlled or not, whether there is a strict order of precedence or not, whether there are second helpings or not, and even the etiquette of ingestion - how food is subdivided and carried to the mouth.

The Douglas and Nicod Study

The first major work on format rules can he found in Douglas and Nicod. The research strategy used in this study was long-term, first-hand observation of eating events in four working-class households. Nicod observed mealtimes but, as quoted above, he did not ask questions about the processes of menu selection. He entered the households as a lodger who openly stated he was interested in the food system. He did not develop a close intimate relationship with the cooks in each household, nor did he ask questions because of his concern with observer interference with the typical pattern. For example, he was afraid that if he asked why peas were chosen for a meal, this question would be interpreted as critical and would influence the way peas were used in the future. His main technique was the observation of all combinations of foods and all types of food events in the four households.

According to Nicod's assumption, he did not need to probe; he could infer pattern from behaviour directly. He could "learn the code by observing the native's application" (1974, p. 5).

Nicod chose four households in four different cities, and lived with the households for one month each and observed their meal-cycles. This kind of co-operation is difficult to find, and Nicod describes the search as one of his greatest problems. He was particularly interested in the daily meal-cycle and the relationship between the major meal as a structure and the structure of Sunday dinner and major feasts (in this case Christmas, New Year's and Christenings). Douglas and Nicod were looking for common structural rules underlying the meal system in the British working class (1974, p. 747).

The difference in food habits between different social classes must not only be judged in terms of the variations of the quality and quantity of the food items in the diet, it must also be seen as the difference between two types of food patterning.

How food items are structured in a meal, and how meals are classified in a typology, are matters which require further investigation. Insofar as there are differences between the social classes, in the patterning of meals in their daily menu. and in the structuring of food items in their standard meals, the social classes have retained different systems of food patterning....

The important dimensions examined in the study were the sensory/aesthetic characteristics of the food and the structural rules for ordering these attributes. The manipulation of paired contrasts was particularly significant: wet/dry, sweet/savoury, whole/part, hot/cold, sculptured and amorphous were some of the contrasts used. The analysis yielded a daily meal-cycle with a progression of increasing segregation of wet and dry elements throughout the day. The daily cycle included a major (potato-based) format, a minor (cereal-based) format, and a tertiary (cake-based) format. There was a definite structural relationship between the major daily meal and the Sunday events, which were structurally parallel but more elaborate (table 4).

As Montgomery (1977) states in his summary of this study:

Table 4. Patterning of meals in the British working class (illustrating the way in which sensory variables were used to define structures)

  Course A Course B Course C
Main meal Hot Hot/cold Cold Hot
  Wet Wet Dry + Liquid
  Savoury Sweet Sweet Sweet
Minor meal Hot/cold Cold Hot  
  Wet Dry + Liquid  
  Savoury Sweet Sweet  
Tertiary meal Cold Hot    
  Dry + Liquid    
  Sweet Sweet    

Table 4a. Main meal

Course Day Dressing/ drink (a) Staple (A) Centre-piece (B) Dressing (b) Trimming (c)
A Xmas Gravy Potato Turkey "roast" Stuffing 2 veg+ Y pud
  Sunday Gravy Potato Other meat Mint/apple " "
  Weekday Gravy Potato   Tomato/


1 or 2 veg
B Xmas White sauce Cereal Fruit Brandy  
  Sunday Cream (Cereal) Fruit (Juice)  
  Weekday Custard (Cereal) Fruit (Juice)  
C Xmas Coffee Cereal Fruit    
  Sunday Coffee        
  Weekday Coffee        

Table 4b. Minor meal

Course Day Dressing/drink (a) Staple (A) Centre-piece (B) Dressing (b/c) Trimming (C)
A Xmas Butter Cereal Meat Pickle + salad cream Salad


  Sunday Butter Cereal Meat " "  
  Weekday Butter Cereal Meat Pickle  
B Xmas (Butter/jam) Cereal Fruit Cream  
  Sunday (Butter/jam) Cereal      
    (Tea) Cereal (cake) Fruit Icing  
  Weekday Tea Cereal (cake) Fruit Icing  
C Xmas Tea Cereal (cake) Fruit Icing  

Source: Nicod. 1974. p. 88.

In the major meal, segregation of liquid foods from solid foods was important, as was the serving of certain foods hot, the changing of dishes and utensils, and giving a central place to potatoes. Nonreversible sequences were found in the order of consumption: potatoes were eaten before cereals, savories before sweets, and wet foods prior to drier items. The shift from wet to dry was accompanied by a change from forks and spoons to fingers. Further, through the structural sequence of the meal, the visual pattern of the food acquired an increasing dominance.

The Russell Sage Project Study

The second example of the analysis of meal formats and cycles is taken from the Russell Sage Project on Gastronomic Categories (Douglas, 1984). Organized by Mary Douglas, this project comprised research on three ethnic food systems in the United States: a community of Oglala Sioux, a rural county in North Carolina, and an Italian-American enclave in suburban Philadelphia. Each community study used a common general methodological framework while pursuing a somewhat different research emphasis. The study of the Oglala Sioux emphasized the ceremonial use of food and the symbolic separation of Sioux food and American food. The North Carolina study was concerned with the relative importance of class or ethnicity in diet (comparing low- and middle-class Blacks to low- and middle-class Whites in the county), and the Italian-American study followed the process of change in the community (Goode, Curtis, and Theophano, 1984).

The common framework of the three studies involved a concern with identifying meal-event types, their occurrence within time-cycles, and their use in different places and for different social contexts. Thus, all three projects identified eating events, temporal cycles, commensal units (social contexts), and eating places. In the following discussion, I will describe the analytical categories used to evaluate each of these components of the framework and then use one community study to illustrate the relative significance of different kinds of data.


In any community, it is necessary to explore the temporal cycles of activity that underlie the differentiation of eating occasions. Meals punctuate temporal activity and ritual cycles in all cultures. In American urban industrial life, there are several overlapping cycles. Days and weeks are patterned by work and leisure, with leisure activity usually occurring around the weekend. Seasonal variations in work activities are punctuated by the patterning of calendrical holidays - both ritual and secular feasts. The life-cycle also generates many occasions: births, marriages, anniversaries, and deaths. The patterning of these cycles in local interacting communities varies greatly in the United States. Identifying such cycles of ordinary and extraordinary feast occasions was the first step in the Russell Sage Project (Douglas, 1984). All communities had daily, weekly, seasonal, and annual cycles of ordinary and feast occasions, and although some were patterned by the national culture, each local cycle was different. In addition, irregular, unpredictable occurrences were also responded to in patterned ways.

Social Context

As it is necessary to identify the temporal framework for differentiating eating occasions in the study of community food-patterning, it is also necessary to delineate the types of social contexts in which eating occurs. These contexts include: (1) the isolated individual; (2) the subhousehold group; (3) the total domestic group; (4) the domestic group plus other groups distinguished by kinship ties, intimacy, sex, age, and generation; (5) large groups of kin; and (6) voluntary associations, churches, and clubs, as well as other possible combinations. The frequency with which particular types of commensal units occur and the occasions for which each type is appropriate or inappropriate must also be considered.


The location of eating events also influences what is eaten. Within the home, different places are used for different occasions and social contexts. Away from home, the location of a meal influences what is eaten not only because of limits of available food but because of notions of what is appropriate for the particular place, which is affected by temporal occasion and social context. Montgomery, Marriott, and Khare have shown the significance of space for differentiating eating events in India.

Time, social context, and place lead to different types of meals and dishes, not only because of socio-cultural notions of appropriateness (signifying honour, prestige, and intimacy), but because of logistical feasibility as well. Some heavy activity periods, the social contexts of a large group, and certain locations result in structural constraints on meal types and food choice.

Methods of the Russell Sage Project Study

The goal of the project was to use long-term intensive observation to understand the structure of meal formats and meal-cycles. Particular attention was given to whether food was presented synchronically and diachronically, in what social order, what sequences of helpings were served, and what etiquette rules in cutting and eating were observed. The use of decoration and the non-eating activities accompanying a meal were also focused on. Figure I is an example of the guide to observations. The study was not limited only to meals within the household: because of the emphasis on location and social context, larger inter-household hospitality events and community-wide eating events were included, as well as eating events in restaurants and workplaces.

The Russell Sage Project (Douglas, 1984) also differed from the research of Douglas and Nicod (1974) in a number of dimensions, including the methodological difference that the researchers asked more questions of their informants and had them describe their own categories of eating events. As with recipes, information about meal formats (menus) can be very different if native categories and classifications are used rather than the observer's frame of reference. Both points of view are important and revealing.

Goode, Curtis, and Theophano (1981) had already studied the Italian-American meal system in Philadelphia using interviews and dietary records. The Russell Sage Project provided the opportunity to add a new dimension of participant observation. Observations were limited to four households over one to two months apiece. However, they were used to complement data collected by surveys (ideal patterns and recalled events) in over 200 households as well as records of actual food behaviour in 35 families.

We hoped to identify broad meal patterns first and then to examine the way in which dishes and items were selected for these patterns. We assumed that recurrent, regular patterns of menu-planning would indicate cultural guidelines that we could demonstrate to be shared and reinforced. The concern was with meal plans for household - controlled eating rather than isolated individual eating away from the domestic context. The same analysis could be applied to regular, patterned extra-domestic eating in schools and workplaces. Eating in domestic groups away from home (restaurants, other homes) was also included, since the domestic group was active in decisions to select this type of meal format for a particular occasion.

We recognized that such a rule system would not operate completely and uniformly within a community. Rules often present ranges of possible alternatives rather than specific demands for a particular dish or meal. Moreover, household variations in activity patterns (work and leisure) and stage in the life-cycle, age, occupation, and income lead to different patterns of rule application. Unexpected constraints on time, personnel, and resources force violations of even broadly stated rules. Moreover, the strength of social transmission and reinforcement varies between domestic units depending on the nature of their embeddedness in social networks and their internal role structures. Finally, different occasions vary in terms of community uniformity: the more public the event (the larger the social context of eating), the more uniformity and social mediation. On the other hand, the smaller and more private the social context, the fewer are the social constraints.

Four research strategies were employed in the following sequence:

  1. Asking people through interviews and surveys about their ideal patterns, away from the context of behaviour.
  2. Collecting diaries and records of actual eating events.
  3. Minimally intrusive observations of food preparation and meal events without any probing or discussion.
  4. Direct probing for explanations of particular choices after a relationship of intimacy and trust was developed over time.

Our initial strategy, based on a desire for speed and limitations on funding, was to use a survey instrument extensively in the community to identify the ideal rules for ordinary and special occasions. We realized that survey questions reflect the limited situational, single-point-in-time conditions understood by the respondent at that moment. Surveys often do not allow for intensive exploration of situational variation (if. . . then . . . ). We were able to get extensive information from over 200 households, and clear patterns were revealed. This survey was followed by the collection of actual dietary intake data in a smaller sample of households (35) to compare ideal patterns to real behaviour.

Later, intensive, continuous participant observation in four households provided us with rich information on social processes. We would suggest intensive ethnography as a first research phase prior to the development of an extensive survey.

Our pilot survey consisted of a 20-page instrument, administered through a one-hour interview by a small team of trained interviewers to 200 households. The important framework for this early stage was identifying the universe of perceived temporal cycles (daily, weekly, seasonal, calendrical, and life-cycles), the universe of possible commensal units (the social context of eating), and the way in which food was used to pattern events. This phase was largely item-centred, e.g. items rather than dishes or menus were assumed to mark events. During this phase a central distinction between types of dish ("gravies" v. "platters") was discovered. The interview also probed the way in which occasions, social groups, social statuses (sex and age), and physical conditions (illness, pregnancy, and lactation) were associated with items, dishes, and menu types. Negative as well as positive rules were elicited.

As we later concluded, one major problem with the pilot surveys was their bias toward patterns experienced during the childhoods of the interviewers. For example, weddings were inevitably described as they used to be celebrated, as if the patterns continued today, and there was a great discrepancy between ideal patterns and the actual behaviour discovered later. "Ideal" was strongly biased toward the traditional.

The interviews in the early phases of research did demonstrate some (but not all) shared ideal statements. As we discovered, early interviews should use open-ended elicitations, since the interviewer has no knowledge of significant folk categories and their meanings.

Contents - Previous - Next