Ideological categories are not bipolar but require structurally distinct explanatory determinants.

Evidence

Thus, in their present forms neither model 2 nor model 3 fits the data very well. Such a relatively poor showing by both of these alternative models bolsters our confidence in our own conceptualization of the nature of the relationship between evaluations of ideological labels and self-identification. Nonetheless, the choice between these various models ultimately must be made on theoretical grounds; no amount of empirical testing can establish the appropriate causal ordering in the absence of a sound theoretical basis (Duncan, 1975). From such a theoretical perspective, any specification of the causal relationship between ideological self-placement and evaluations must be consistent with one’s understanding of the nature of those evaluations and their determinants. In this regard, our conceptualization differs critically from the alternative models in our treatment of the question of bipolarity. Because we posit a lack of bipolarity in the meaning of ideological labels, we necessarily must hypothesize structurally distinct determinants of the evaluations of such labels. — John T. Jost and Jim Sidanius, Political Psychology, 2047

Categorization

• XREF: Connects to broader literature (e.g., Converse, ideological constraint studies) on whether liberal-conservative is a single continuum or multidimensional space. — connects to: The category of the Other precedes and enables material oppression, not vice versa — Simone de Beauvoir, The Second Sex, 147; Contentious political categories like riot are symbolically but not causally coherent. — Charles Tilly, Regimes and Repertoires, 301