The concept of domination plays a central role both in the relatively new wave of civic republican theorizing, headed by Philip Pettit, Quentin Skinner, and Maurizio Viroli, and in the works of theorists of justice and democracy, such as Iris Marion Young, Ian Shapiro, and James Bohman. Yet, upon close inspection of such work, one will be disappointed to find that while the term is used in various rhetorically decisive junctures, domination is never defined in a cogent manner. Hence, the normative correctness or incorrectness of the various political arguments made by appeal to nondomination is bound to remain vague. So claims Frank Lovett, and his book aims to rectify this situation by developing a theory of domination that defines the term in a cogent manner and explains the normative appeal of nondomination in the context of a theory of social justice, correspondingly dubbed “justice as minimizing domination,” or JMD. As Lovett states, “(JMD) Societies are just to the extent that their basic structure is organized so as to minimize the expected sum total domination experienced by their members, counting the domination of each member equally” (p. 159).

Taking a clue from legal positivist analytical jurisprudence, Lovett divides the book into a “descriptive” part and a “normative” part. The definition of domination adequate for a useful theory, Lovett argues, should be “strictly descriptive,” such that when it's applied, it is capable of identifying real-world instances of domination without reference to the normative features of those instances (p. 19). Lovett points out that the statement “Wilderness areas ought to be preserved for reasons x, y, and z” (p. 18) does not give reasons for action or inaction unless we have a conceptual definition of wilderness that specifies salient, strictly descriptive features of wilderness such that, with those features in mind, we can identify particular tracts of land as instances of wilderness. Yet unlike wilderness, domination is not a natural phenomenon but a social one, and as such, it may very well have features that are simultaneously descriptive and normative. I agree that a cogent definition of domination should specify characteristics; I disagree that all these characteristics have to be thoroughly nonnormative or strictly descriptive in order for the definition to play a practically informative role in a useful theory.

Part 1 of Lovett's book develops his “strictly descriptive” conception of domination, and the normative analysis in Part 2 puts the conception in the context of a theory of justice. Assuming for the sake of argument that his method of separating descriptive analysis and normative analysis is sound, the normative analysis could not really get going in a worthwhile direction if the descriptive analysis is faulty. That alone motivates me to concentrate the rest of my discussion on Part 1. But there are other reasons to concentrate our attention on this part, since one of Lovett's most important contributions is that he delineates a typology of conceptions of domination, which is in itself very useful for understanding, classifying, and arguing through the variety of views about domination that can be gleaned from the plentiful literatures germane to the concept. In chapter 4, Lovett succinctly addresses and draws from vast families of discussions, including arguments derived from Antonio Gramsci, the Frankfurt School, and Michel Foucault, as well as arguments closer to Lovett's own, such as Steven Lukes's interventions on power. The typology he develops enables him to both build on some prior contributions to the enterprise of defining domination and to clearly distinguish his position from those that can be found in or constructed from the extant germane literatures.

So what is Lovett's definition of domination? Put in his own words, “persons or groups are subject to domination to the extent that they are dependent on a social relationship in which some other person wields arbitrary power over them” (p. 119). As he points out (p. 120), this definition excludes the notion that social systems can themselves be considered domination on account of their modus operandi: domination directly involves the intentional behavior of specifiable agents, dominators and dominatees. The dominators and the dominatees must be in a social relationship with each other, the dominators must have more power than the dominatees, and—crucially, to distinguish Lovett's view from prior accounts that focus on differentials in power within social relationships—their social relationship must be such that the more powerful wield “arbitrary” power over the less powerful. People familiar with Philip Pettit's argument regarding freedom as nondomination will at this point probably think there is thus nothing new in Lovett's book, since Pettit also makes a big deal about how domination should be defined in terms of a sort of arbitrariness. But Lovett's account of arbitrariness is markedly different from Pettit's, and he clearly distinguishes this difference (pp. 115–16). While for Pettit, the nonarbitrariness of the power dynamics operative in a social relationship is a function of whether the relationship is somehow made to track or responsively reflect the affected persons' beliefs about their interests as registered in democratic deliberative procedures, for Lovett the democratic-deliberative credentials of the social relationship do not matter. Instead, according to Lovett, the arbitrariness of such power dynamics is a function of whether or not the more powerful persons' behavior is unconstrained “by effective rules, procedures, or goals that are common knowledge to all persons or groups concerned” (p. 96), regardless of whether or not these rules, procedures, and goals are subject to democratic accountability. For Lovett, in other words, the nonarbitrariness of social power dynamics is akin to the way in which many people conceive of what it means for behavior to be constrained by the rule of law.

Lovett criticizes Young and Pettit, among others, for a sort of circularity—for making it seem as if nondomination and democracy are analytically related. Instead, Lovett thinks that “one of the strongest arguments for democracy is that it tends to reduce domination,” and that that argument is “trivialized” if nondomination is conceived as definitional of democracy (pp. 117; 211–12). This is a good point, and Pettit's conception of arbitrariness has always struck me as implausible. But while Lovett's conception of arbitrariness is a substantial improvement, defining nondomination as nonarbitrary power is itself highly problematic. To his credit, in order to register and attempt to lessen the problematic character of his argument, Lovett addresses a key example of systematic, legal discrimination—by his own definition, as repeatedly stated, and as applied in the example, such discrimination is nonarbitrary (pp. 111–19). Consistently, he claims such discrimination does not amount to domination. But he achieves consistency at the cost of cogency. In his example, Lovett focuses on legalized discrimination against Jews in employment. And in the context of considering the merits of keeping the bad of discrimination distinct from the bad of domination, one may at times be persuaded. Yet what do we think of a social relationship—duly characterized by all the regularities of the rule of law—in which some people are systematically imprisoned, starved, tortured, and thrown into a gas chamber to slowly and painfully die? Since Lovett very early in his argument rejects all outcome-based conceptions of domination, and he specifically rejects a very plausible outcome-based conception of substantive nonarbitrariness (p. 114), he would have us believe such people are not dominated, that they are just suffering discrimination, torture, and painful homicide. Surely, something in the analysis has gone astray.

This book is a fine analysis-driven account of a particular conception of domination and attendant theory of social justice based on the good of nondomination, defined correspondingly as the opposite of domination. The book goes a long way toward plausibly specifying what is involved in the interagent dimensions of the phenomenon under study. But it is defective in a way I have not gone into in this review, insofar as the interagency dimension is just one dimension of the phenomenon, and the concept of domination at any rate does have cogent applications in direct reference to entire social systems rather than merely by the mediation of interagent intentional interaction. (In this regard, by the way, the best monograph remains Goran Therborn's What Does the Ruling Class Do When It Rules? [1978].) Even if we confine our perspective to the interagency dimension, however, as my legalized-discrimination-into-the-gas-chamber example demonstrates, Lovett's definition of domination is unconvincing, and the reason it is unconvincing is systematically synergistic with the extreme separation between fact and value to which he attempts to subscribe. Though Lovett doesn't say it, his intuition and reasoning that domination is a purely structural phenomenon—i.e., it has nothing at all to do with outcomes—has a strong elective affinity with his misguided aspiration to keep matters of fact strictly separate from normative analysis. (Normative analysis is necessary at the very least in order to specify what sorts of outcomes are salient in cases of domination that are partly outcome based.) Despite this, however, we may be a lot closer to the truth about the interagent dimension of domination than before Lovett's book appeared. While Lovett's account fails to include at least some cases of domination that are at least in part outcome based, his conception of domination as arbitrary power does pick out significant sets of cases of domination, and it helps to at least partially explain in an analytically elementary fashion what it is that makes them cases of domination.