The social constructionist approach taken by Watson and Goulet in their recent paper on the belief system of the Dene Tha is critically examined. In arguing that Dene Tha reality is "socially constituted," Watson and Goulet's paper displays many of the characteristic assumptions of social constructionist analysis: in particular, the use of a dichotomy between an empiricist and a constructionist conception of knowledge and a claim that knowledge is structured into epistemologically equivalent "socially constituted realities." These assumptions are shown to take their plausibility from basic confusions about the term "reality." It is argued that rather than shedding light on Dene Tha ritual beliefs and their practical uses, the effect of the social constructionist perspective, here as elsewhere, is the imposition of arbitrary and unwarranted designations on the ethnographic facts.