Some Procedures and Results in the Study of Illative Categories: Tzeltal “Firewood” 1

Abstract

0. This study represents results at one stage in an attempted development of some specifiable techniques of ethnographic investigation and description. In this paper we discuss these techniques, and, by way of an example of their use, show the manner in which they reveal relevant dimensions in terms of which firewood is differentially identified and evaluated by the persons who use it in various Maya Indian parajes2 of the Muncipio of Tenejapa in the highlands of Chiapas, Mexico. The data which underlie an ethnography represent, or should represent, native responses to some set of conditions. It is our feeling that more attention needs to be directed to these conditions than has generally been the case, for insofar as the statements of the conditions governing native responses are present only by implication, the resulting description is replicable only in the most general sense and is thus ambiguous and open to various and personal interpretations. Among the conditions which produce native responses, and which, together with these responses, constitute one sort of data for the ethnog-rapher, are questions put by the ethnographer to members of the population to whom the description pertains. The investigative techniques illustrated here are directed to the formulation of these conditions in the form of linguistic contexts which elicit stable responses and are thus most efficiently replicable with a minimum of ambiguity. In the description which follows, there is illustrated the use of unit linguistic contexts which we term frames3 as they serve as conditions governing responses across some segment of the population under study. It needs to be emphasized here that frame and associated response constitute a descriptive unit (at some level) which rests upon classificatory differences significant to informants rather than to the investigators. This approach to the problem of building a description presupposes some simple analytic and descriptive principles that can be subsumed under the head of "distribution." We seek to discover regularities of distribution of linguistic forms with respect to highly specific environments, that is, with respect to the question-frames with which they occur as regularly elicitable responses, and, among regularly elicitable responses, we seek to discover similarities and differences with respect to frame-sized environments. The procedures of frame-formulation and response elicitation allow us to begin from any hunch or observational point of vantage we may have, and lead, through 389