Analysing Essentially Evaluative Concepts: The Case of Justice

Conceptual analysis is one of the main instruments in the philosopher’s toolbox. Besides elucidating and disambiguating complex ideas, conceptual analysis helps us to sharpen our thinking by refining and enriching our vocabulary, structuring our theories and guiding moral judgement. This valuable gadget, it has been argued, is unavailable to political theorists. A conceptual, non-normative analysis of political notions would be impossible as any such analysis presupposes ethical evaluation. Studying political concepts would be ethical all the way down (e.g. Lukes 1974; Connolly 1983; Dworkin 2011: 166-70). One consideration cited in support is that political notions like ‘justice’, ‘democracy’, ‘liberty’ and ‘equality’ are ordinarily used to morally commend or disparage. Given widespread moral disagreement and conflicting political interests, how to settle the criteria of application of such terms non-partisanly? Ian Carter (2015) has recently contended that this oft-voiced objection is conceptually confused. For one, it fails to distinguish between essentially and non-essentially evaluative concepts. That is, it conflates concepts the use of which necessarily involves an evaluation (like ‘fair’ and ‘just’) with concepts that are not intrinsically evaluative but whose normal use does have evaluative implications (like ‘freedom’ and ‘democracy’). (Bernard Williams’s (1985: 129-30) distinction between ‘thin’ and ‘thick’ ethical concepts captures roughly the same idea more obscurely.) It is intelligible, albeit uncommon, to call acts of charity morally wrong or to reject freedom categorically. By contrast, anyone sincerely declaring “this just arrangement is wrong” fails to grasp the meaning of justice. Non-essentially evaluative concepts are evaluative not because of any feature of the concept, but because of the normative commitments of the persons using them (Carter 2015: 284). Thus, the concept ‘equality’ expresses a value for egalitarians but not for malign tyrants. The upshot is that we can construct value-free definitions of non-essentially evaluative concepts, i.e. definitions that don’t themselves contain evaluative notions. This paper explores what conceptual analysis can offer to the study of essentially evaluative concepts. I shall focus on ‘justice’. My contention is that if we think of conceptual analysis narrowly, as the search for a concept’s ‘defining conditions’, then this analysis is unavailable for justice. Given intractable moral disagreement about which persons, actions and states of affairs in the world count as (un)just and why, it is impossible to define justice without engaging in moral argument, simply by settling its extension. Conceding the point, I proceed to point out that other kinds of conceptual analysis can still suitably be employed. Drawing on J.S. Mill’s (1863) and Henry Sidgwick’s (1907) acclaimed analyses of ‘justice’, I argue that conceptual analysis can 1) disambiguate ‘justice’ by revealing that the term is systematically used to refer to distinct kinds of objects; and 2) elucidate the notion by exploring its basic conceptual structure through the identification of its constituent elements (e.g. rights, blame) and their conceptual linkages. This paper thus hopes to contribute to our understanding of the nature of conceptual analysis in general and to its use in political theory in particular.