ABSTRACT
To illustrate this reality before a college class, in the early 1990s in New York City, I placed side-by-side on a table two salads: one made with iceberg lettuce and bottled creamy French dressing and the other with assorted baby greens dressed in a light mustard vinaigrette. I then asked the class to tell me as much as they could about the people who were likely to eat each salad. Remarkably, before too many minutes had passed, the board in front of the class was fi lled with two groups of lists-one for each salad-suggesting not only where these people lived and their socioeconomic level, but also the books they were likely to read, the cars they were likely to drive, the fi lms they were likely to view, and so forth. All those participating in the exercise recognized that, as with all stereotypes, there were limits to the accuracy of what we were proposing, particularly when applied at the level of the individual. But we also understood that a great many of the generalizations we offered were relatively accurate, that is, they were likely to be true for a signifi cant majority (70%?, 80%?) of the people who preferred each salad. If this is what one can say on the basis of a salad, imagine what be said on the foundation of large-scale patterns of the foods a people eats and the ways they eat them.
