Theory Development

5. Meaning and Background

Considerations of meaning, the way in which agents understand the situations they are in, and more generally the kinds of motives that animate people in particular social groups and settings, are often consigned to the category of background knowledge, and it is indeed the case that local knowledge of various largely uncodifiable kinds is needed to interpret the behavior of individuals and their interactions. This background knowledge is often sufficient to account plausibly for the behavior in question, such as the behavior of adolescents, without reference to any “social science” knowledge at all, at least to those who share this local knowledge.

Limits of Local Knowledge

  • Very vague (e.g., teenagers lack sufficient knowledge)
  • Stretch too easily to cover all cases (e.g., teenagers are hormonal)
  • Simply insufficient even on their own terms ( e.g., teenage girls have self-esteem problems)
  • Not generalizable (e.g., there was a situation at the prom where people got carried away)

The limitations of local knowledge– folk or local background knowledge– are nevertheless acute. Often the explanations it provides are very vague, or stretch too easily to cover all cases, or are simply insufficient even on their own terms. Some important behavior, such as suicide, is incomprehensible as a rational act. And theories of suicide often add little to either comprehension or prediction even in those cases close at hand and most amenable to “local knowledge” explanations.

When we turn to aggregate patterns, these difficulties become more serious. This kind of knowledge is rarely sufficient to account for, much less predict, the statistical patterns involving adolescent pregnancy. And local knowledge often does not generalize well to other populations and circumstances. Thus, to the extent that the “mechanisms” in questions under consideration involve the meanings attributed to behavior by people, it must be accepted that there is a great deal of variation and complexity beyond anything that might be thought of as a “mechanism.” The same correlation may conceal or depend on a wide variety of local and complex behaviors and circumstances, for which any simple model will be inadequate. And some statistical patterns can be given only the most vague “sense.”