Comments on Charles Mills'
"Race and the Social Contract Tradition"

Sally Haslanger
MIT, Department of Linguistics and Philosophy
Central APA, April 22, 2000

The framing question of Mills' important and thought-provoking paper is whether there is reason for political progressives and radicals to employ the notion of a social contract for either descriptive or normative purposes. In contrast to the common response that the social contract is a piece of "bourgeois mystification" he argues instead that a reformulated conception of the contract, one which he calls the domination/exclusivist contract, is a valuable both theoretically and politically.

The political progressives and radicals he is addressing are those who are not only undertaking normative inquiry into how society ought to be structured, but also provide descriptive models for understanding how societies are actually structured. On their accounts, the central fact to be modeled is the "reality of group domination" (p. 7). In the contemporary context, white supremacy, patriarchy, capitalism, are the core phenomena through which societies are to be understood, and in response to which normative political philosophy has its point. The task, taken broadly, is to understand how group domination "arises out of social processes" (p. 7), and how it replicates itself over time in society as a whole.

Mills argues that the domination/exclusivist contract, as he understands it, can effectively meet these demands: (i) it is a dominance account: it provides a good descriptive model for recent global history (particularly of race), in which the central and normative role of group domination is "formally registered in the conceptual apparatus" of the theory (p. 18); (ii) it is anti-naturalistic: it provides a basis for understanding the "social construction" of race and gender by providing an account of "the shaping role of human causality" (p. 8), and (iii) it is comprehensive: in unpacking the details of the contract we have the resources to see how domination becomes entrenched in the structure of institutions, cultural practices, and individual psychologies (p. 7).

Simply showing that a particular model meets a set of criteria we endorse doesn't by itself show that it warrants our acceptance, for there might well be better models available. But Mills argues that there are two specific benefits to progressives in employing a version of contractarianism. First, it has important polemical and pedagogical value: "a critique that engages contract on its own terms and shows, given the factual record, how inadequate its prescriptions typically are, is likely to be more polemically effective than one which simply dismisses it altogether." (p. 9) Such a critique, Mills argues, will also foster more effective communication between those engaging the issues from different (racial/gendered) perspectives.[1] (p. 18) Second, it has theoretical value as well. Because contract theory is an important resource for thinking about normative matters, employing a contract model for the descriptive component of the theory "present[s] the problems of social justice in a unified and integrated conceptual framework" (p. 19-20).

A theory that can do all this is certainly a good thing. And Mills makes a good case that his account of the domination/exclusivist contract can do all this. Let me start by saying that I am tremendously sympathetic with Mills' approach and the considerations he raises in favor of it. However, I am not a political philosopher (though I am a political progressive in the sense sketched), and in the interdisciplinary context in which I work, social contract theory is not the dominant discourse. So I will have less to say about the usefulness of a contractarian approach within political philosophy, than about the broader question whether a contract model is useful for progressive feminist/anti-racist theorizing. I agree with Mills that there are contexts in which employing the terms of a theory against itself is the best way to launch an effective critique, and I agree also that these rhetorical advantages matter. But as Mills notes, the weight of these rhetorical advantages may vary with context. Given his goal of offering a contractarian model to progressives as a group (not just progressive philosophers, I assume), the question remains: when we are not speaking to the mainstream in a critical mode, but are trying to understand group domination in our own terms, is the social contract model the best way of understanding what is going on? More specifically, is it a good way of understanding racial domination? Is it a good way of understanding gender domination? (Are there interesting differences between race and gender that make it better for one than for the other?)

[Let me acknowledge that I'm not entirely sure whether Mills is offering the domination/exclusivist contract model primarily as a critical/rhetorical tool, or as an account that we should accept on its own terms. (I.e., he may be willing to grant that when not attempting to engage the philosophical tradition there are other models that are equally or even more useful. Note, for example, that he mainly compares the domination/exclusivist contract with a mainstream interpretation of the contract, rather than with other non-contractarian strategies for understanding group domination.) However, Mills suggests that he offers it on both grounds. In particular, the theoretical advantages that come with having an integrated descriptive/normative framework are not insignificant, especially because, as Mills notes, it is plausible that contractarianism is the best normative strategy available. But if there are theoretical disadvantages to accepting contractarianism as a descriptive account, we should be attentive to them. After all, they may help illuminate some of the disadvantages of contractarianism as a normative account as well.]

To motivate my questions more fully, let me expand a bit on what I take to be a fuller set of desiderata (added to the three mentioned above) for a progressive account of group domination arising out of at least one strand of feminist/anti-racist theorizing. Some of these I take to be in the spirit of (and possibly even elaborations of) the three already mentioned by Mills, but some are quite independent, (and may not even be ones he endorses).

* (Anti-structuralism) The goal is not just a theory that is historical (v. ahistorical), but is sensitive to historical particularity, i.e., that resists grand causal narratives purporting to give an account of how domination has come about and is perpetuated everywhere and at all times. This is important for two reasons: (i) we cannot assume that the conceptual repertoire used to understand our own (Western) history even makes sense in other cultural contexts, and we face a very real danger of enthocentrism in attempting such grand (the usual buzzword is "totalizing") narratives. [This is especially important if we want to provide histories that take into account the conceptual framework in which people articulate their own reasons for action.] And (ii) the forces that cause and sustain domination vary tremendously context by context, and there isn't necessarily a single causal explanation (there may be overdetermination); a theoretical framework that is useful as a basis for political intervention must be highly sensitive to the details of the particular social context and its history.[2]

* (Anti-essentialism) Although an understanding of group domination must employ a meaningful notion of group, we must be ever vigilant in avoiding over-generalizations about the attitudes, experiences, or social position of members of the groups. Given the overlapping nature of social groups, every member of a group will also be a member of other groups which will affect his or her identity and social/economic/political status. Moreover, the meaning of group memberships can vary significantly with context.

[* (Anti-victimization) Those in subordinate positions are not passive victims of domination; nor are those in dominant positions full agents of domination. Both function as agents in being complicit with the systems of domination and in resisting them. Moreover, society "imposes" dominant and subordinate identities on members of both groups; subordinate identities are not more or less "constructed" than the dominant. However, the framework of identities clearly benefits the members of the dominant group and disadvantages the subordinate.]

* (Holism) Group domination cannot be explained by reference only or primarily to the attitudes or psychologies of the dominant group; nor can it be explained by reference only (or primarily) to economic forces; nor can it be explained by reference only (or primarily) to the political structure of society; nor can it be explained by reference only (or primarily) to culture (e.g., binary logic), etc. It is the result of the complex interplay of multiple determinants [whose significance varies depending on context].

Obviously, it is a huge challenge to accommodate all of these desiderata, and in some ways they pull against each other. E.g., the goal of providing a dominance account that places group domination at the core of the analysis can result in some tension with the anti-structuralist and anti-essentialist objectives.[3] Depending on context and the immediate purposes of one's theorizing, it will be important to place greater weight on some of these desiderata over others. But it is useful to see how a particular account balances these various considerations.

A quick reading of Mills' essay might lead one to be concerned that although historical, the account is not adequately attentive to historical particularity insofar as it attempts to impose a single narrative on quite different situations (although clearly the account is not structuralist, the charge would be that it isn't sufficiently anti-structuralist?); one might be concerned that it is essentializing in its characterization of the attitudes of group members and their inter-subjective agreements (can we really speak of a group's "will", of whites "imposing" their will?); one might be concerned that it is insufficiently holist in its giving explanatory priority to contractual agreements in the account of group domination (shouldn't we, in turn, explain these contractual agreements at least partly in terms of cultural and economic forces[4]?).

Although I think there may be interpretations of Mills' views on which these concerns are reasonable, I think there are interesting and plausible interpretations on which they can be seen to miss the point. I will argue, however, that there is some cost to opting for these alternative interpretations, for the more abstract and metaphorical the model becomes, the less it is able to provide, as Mills seems to want, a substantive account of "the shaping role of human causation". (more shortly)

To get a better grip on how the domination/exclusivist contract fares with the broader set of desiderata, it is important to look a bit more at the details. What is the domination/exclusivist contract? What sort of model does it provide? Mills proposes that the domination/exclusivist contract is intended as a model of society as we know it, an "overarching optic for thinking about the socio-political" (p. 6). It is not intended as a literal representation of history, but as an illuminating model that captures "some central truths". (p. 6)

This provides a meta-theoretical guide to interpreting the domination/exclusivist contract--it is to be understood as a story that organizes our thinking; although it is not literally true that there is/was a domination contract, society as we know it (and its history) is structured as if there was. But this does not yet give us much detail as to the content of such contracts. For that it is useful to turn to Mills' book The Racial Contract for an example. There he says:

The Racial Contract is that set of formal or informal agreements or meta-agreements...between the members of one subset of humans, henceforth designated by (shifting) "racial"...criteria C1, C2, "white," and coextensive...with the class of full persons, to categorize the remaining subset of humans as "nonwhite" and of a different and inferior moral status, subpersons, so that they have a subordinate civil standing in the white or white-ruled polities the whites either already inhabit or establish, or in transactions as aliens with these polities, and the moral and juridical rules normally regulating the behavior of whites in their dealings with one another either do not apply at all in dealings with nonwhites or apply only in a qualified form... (p. 11)

The Racial Contract, I assume, is a paradigm of a domination/exclusivist contract. Through it, Mills claims, "a partitioned social ontology is created." (p. 16) On this account, "the shaping role of human causality" works through a background set of agreements defining who counts as a person and who not; these agreements are "imposed" upon the designated "subpersons" by being codified in the institutions and practices of the society; social norms are internalized in such a way that the "person-constituting" agreements become unquestioned and unquestionable.

[It appears that domination/exclusivist contracts are intended both to provide a story about the causal origins of group domination as well as an account of the mechanisms of ongoing domination: background agreements among the dominant group got the domination going in the first place and work to perpetuate it. Or, given the non-literalness of the model, perhaps it is better to say: it is as if background agreements among the dominant group got the domination going in the first place and work to perpetuate it.]

As I read The Racial Contract, Mills there seems more willing to claim the literal rather than simply metaphorical accuracy of the domination contract model and the racial contract in particular (e.g., section beginning p. 19). He says,

Although no single act literally corresponds to the drawing up and signing of a contract, there is a series of acts--papal bulls and other theological pronouncements; European discussions about colonialism, "discovery," and international law; pacts, treaties, and legal decisions; academic and popular debates about the humanity of nonwhites; the establishment of formalized legal structures of differential treatment; and the routinization of informal illegal or quasi-legal practices effectively sanctioned by the complicity of silence and government failure to intervene an punish perpetrators--which collectively can be seen, not just metaphorically but close to literally, as its conceptual, juridical, and normative equivalent." (p. 21)

I take it that the current paper is stepping back from this a bit--even if it is almost literally true that there was a racial contract, the domination/exclusivist contract model need not be literally (historically) accurate to be useful. There are, at least potentially, two different steps back: (i) applying the model to a particular case of group domination does not require that the origins of the domination are based in a (literal or metaphorical) contract, for it may be that the origins of domination can be separated from what sustains it now. (This is helpful when attempting to apply the model to gender, for given the apparently "pre-historical" origins of male domination, it would be unfortunate if employing the model required us to make historically accurate claims about an original gender contract.[5] What matters is how gender domination is currently enforced and sustained.) And (ii) applying the model to a particular case does not require that there be something as close to literally a contract as indicated above for race, either as part of the history or current maintenance of group domination. But what exactly constitutes a contract in the relevant sense, and how far from a literal contract can the case be and have the model still apply?

Unfortunately, I don't have a good grasp of what exactly is required--of the historical record, or of the contemporary social facts--in order to say that the domination/exclusivist contract model applies. In the case of the racial contract, there is no "single" act that constitutes a contract, but there are numerous contract-like acts, e.g., "pacts, treaties, and legal decisions". Does the application of the model require that either there are now or have been some explicit authoritative and formal statements relegating certain groups to the status of subpersons?

On one hand, Mills is keen to emphasize the historical reality of such explicit and formal exclusions--this is an important part of the historical record that is systematically ignored and needs to be brought to light; of race, in particular, he claims,

"Whiteness is a system of domination and exclusion brought into existence by mutual (in-group) agreement. The political character of race is thus made theoretically central: race is politically created and is a form of political domination." (p. 15)

Based on this strand in his work--both in The Racial Contract and in the current paper--one might interpret Mills as claiming that the origin and perpetuation of group domination is a certain pattern of exclusive agreements and stipulations by members of a group that has the power to enforce its "will". There is no literal "contract" between the dominant group members, but their agreements are self-conscious and find expression in the juridico-political structure of the society.

It is this strong interpretation, i.e., closer to literal...there is no single act, but there are multiple expressions of at least an implicit agreement, of the dominance/exclusivist contract that is likely to raise the concerns mentioned above: The first question is whether there are or have been forms of group domination whose origins don't depend on such self-conscious political agreements, or which although did at one time, don't over time. In effect, is this the best model for understanding all forms of group domination? Is it adequately attentive to historical variation in how domination is created and maintained? Does it over-generalize about the attitudes and relationships between members of the dominant group? Does is give explanatory priority to psychology (viz., the attitudes of the dominant group) or the political domain (viz., explicit contracts/agreements) in ways that are problematic? As Mills points out, the self-conscious political exclusions of subordinate groups are ignored and misrepresented in mainstream theorizing, but they happen...they are in the record. And this needs to be recognized as the central society-constituting phenomenon it is.

However, at many points Mills is also keen to distance himself from a "voluntaristic" interpretation of the contract. For example, he seems to endorse Hampton's claim that we do not "either tacitly or explicitly exchange promises with one another to create or support certain governmental structures" (p. 5); rather, patterns of domination and exclusion emerge through certain problematic conventions becoming entrenched. (p. 5, cf. p. 8?) It would seem that these problematic conventions may or may not find expression in the political (i.e., juridico-political) order of things. (Perhaps to interpret the quote from p. 15 about race given above, we should note that Mills is working with a broad notion of the "political" broad? See p. 7) And given the epistemic obfuscation endemic to systems of domination (that Mills so vividly captures), it is plausible that the exclusionary "agreements" are not self-conscious or even hypothetically endorsed (in one sense of 'hypothetical').[6]

My guess is that in offering the model of the domination/exclusivist contract to progressives, Mills is not proposing we prioritize formalized political agreements as the universal basis of group domination. This would be implausible, and would not do justice to his goal of providing the domination/exclusivist contract as a quite general model of how human causality works in the social sphere. But what then is the model of human causality being offered? I think it may be useful to compare the domination/exclusivist contract with certain kinds of teleological explanations of natural phenomena.

Consider certain naturally self-maintaining systems: a marble on the bottom of a bowl, a stick in a stream held against a rock by its own backwash[7], (a human body's regulation of temperature or hydration?). Such homeostatic systems are stable even under certain modifications of the context. (Bump the bowl and the marble will move, but return to the bottom.) In simple cases of this sort, it is possible to provide non-teleological (causal) explanations for the behavior of the system and its parts. Even so, there is something right in saying that it is as if the bowl is designed to keep the marble at the bottom, or it is as if the stick was positioned so it would remain in place. We easily resort to the notion of design or purpose in order to make vivid the systematic working of parts in a larger whole, even when it is clear that literally speaking there is no designing agent or broader purpose.

Societies are extremely complex self-maintaining systems. Their parts are interdependent; they sustain themselves in the face of many external and internal changes. If we want to highlight the fact that a society is hierarchical and that its hierarchy is entrenched and relatively stable, one way of doing it would be to suggest that it is as if it is designed to be that way. It is as if the bowl is designed to keep the marble at the bottom; likewise it is as if American society is (contractually) designed to keep certain groups subordinate. Read in a certain way, with due emphasis on the "as if", the point is that there are complicated causal forces at work that maintain the system in a way that produces a particular notable result. The causal forces in question may be arranged as they in fact are simply by accident (think of the stick against the rock), and the notable result may not actually be desired or intended by anyone. But we learn something about the organization of a system (natural or social) by representing it as intended, i.e., as constructed by design.

If we interpret the domination/exclusivist contract as an explanatory strategy something like the invocation of design or teleology in cases where we are keen to capture a system's self-maintaining capacity, then the domination/exclusivist model need not commit us to substantive generalizations about the detailed workings of dominance structures, e.g., about the specific attitudes of those in power, about how power and authority is achieved, manifested, exercised. This will allow us to be sensitive to historical particularity in the mechanisms for sustaining hierarchy. Just as there are many different ways for self-maintaining "natural" systems to emerge and sustain themselves, there are many different ways for self-maintaining social dominance structures to emerge and sustain themselves. Nevertheless all systems of social domination work as if they were contractual.

It is important to note, however, that an "as if" teleology of this sort does not itself provide a causal account of the phenomenon. The stick is shaped and positioned "as if" it were designed to remain against the rock, but it wasn't. And so we cannot say that it's being so-designed is causally responsible for its behavior. Invoking teleology or design in such natural contexts is a kind of heuristic device to help us understand the systematicity of something produced by ordinary causes. But then the same should be said of the domination/exclusivist contract: it does not provide a causal explanation of how domination arises and is sustained--or better, I think, invoking the domination/exclusivist contract is not to claim that the society is/was designed around domination (though it may in fact have been), but is to claim that there are causes responsible for the system that have the same result as a such a designed system would have had. Dominance is not "designed" by god or by nature, but neither is it in every case designed by us, if "design" or "intention" is assumed to be a causal force responsible for its origins or maintenance. However, our institutions, our practices, and our individual actions sustain societies that have the same structure as ones designed for the purpose of subordinating.

But if this is what the domination/exclusivist contract offers, then it isn't entirely clear to me how much we have gained theoretically by introducing the idea of a contract. In fact, we are still left with all the hard questions about how social causation really works, how domination becomes entrenched, how in particular cases it can be destabilized. For example, if someone asks: how does that stick stay against the rock when it would appear that the current should be pushing it downstream? And you answer: it is positioned as if someone designed it to stay there. The questioner might rightly feel dissatisfied with the answer. We aren't being told what forces actually caused it to be there, or what keeps it there, or what we'd have to do to dislodge it. But this is exactly what I think we want to know about the stick, and about group domination. In short, as I read the domination/exclusivist contract there is something of a dilemma: either we see it as giving us a substantive model of how group domination works across the board, but one which is not entirely plausible as an analysis of all cases and seems to over-generalize from the example of race; or we see it as giving us a kind of metaphorical (quasi-teleological) model of group domination, which is illuminating, but doesn't give us the kind of substantive analysis we need in order to understand how "social causation" works.

This concern allows that domination/exclusivist contract model does actually fit (almost literally) the recent global history of race. So I'm not arguing that the dominance contract model is not informative or useful when applied in particular cases; but I'm worried that the model taken to the next level of abstraction--so it applies to all forms of dominance somewhat metaphorically--loses much of its power, for there it obscures rather than illuminates the fact that dominance happens in many different ways: admittedly through contract in some cases, but not in all, for there are other ways in which human causation works to create and sustain hierarchy.

Note, however, that my comments have assumed that what Mills is aiming for in his "descriptive" project is something like what many (progressive) political theorists and social scientists are trying to produce: a (causal) account of group domination (allowing that such an account must provide descriptions and explanations at the level of reasons not just behavior!). But perhaps this is not what he is aiming for after all. A very plausible interpretation would be he is offering a "picture" or "iconography" that when applied to the actual situation highlights its morally relevant features. Such a description may not provide a (substantive) explanation of how domination came about or how it is sustained--so it may not do some things progressives want and need--but its point is to illuminate the actual structure of society in such a way that our normative model can get a grip on it. Given, as Mills suggests, that some form of contractarianism is the strongest contender for a normative account of justice, this is important. Having an integrated framework in which we represent both how things are and how things ought to be is crucial in order to think about what in the current situation ought to be changed. And this is something that a descriptive social contract model can provide us that other "causal" or "explanatory" models often completely miss.

If this is the best strategy for understanding the dominance/exclusivist contract, it does have many of the clear advantages Mills suggests. My point, then, can be reformulated as follows (and may well be one that Mills was never confused about!): we cannot assume that an account of what causes or sustains domination will at the same time provide us with a model of what is morally relevant (or morally repugnant) in dominance systems. In principle, dominance could be caused and sustained by accidents of nature and history; a dominance society might be something like a stick caught on a rock. What the dominance/exclusivist model does, however, is show us how--regardless of how we got or remain there--being stuck there is a gross violation of the ideal norms we aspire to live by. The dominance/exclusivist contract--understood as an "as if" story--does not give us a causal explanation of domination; but neither do causal accounts provide a description that engages our normative framework. There are two "descriptive" jobs to be done. And the dominance/exclusivist contract is impressive in accomplishing one of them.


[1] It is less obvious to me that the pedagogical advantages will occur in contexts where gender is the form of domination being discussed. In my experience, students of color come to class with a better understanding of racial domination and the political complexity of US traditions than women have of gender domination. So although women are often alienated by political philosophy as presented in philosophy classrooms, they have a less politicized take on their own experience and would be more hesitant, I believe, to engage with the multiple traditions strategy. I'm not completely sure about this though, and it is certainly worth a try!

[2] Note that some feminist theorists now even resist the use of the term 'patriarchy' because it suggests that there is a single form of male domination across history/context. The correlative move would be to suggest that thinking in terms of 'white supremacy' is a mistake because it suggests that there is a single form of racial domination, when in fact there are many. Although I see the point of this move, I myself think it is possible to define these notions at a level of abstraction that they can be useful while still allowing for tremendous substantive diversity in the details. In this I am sympathetic with Mills' proposals.

[3] Consider, e.g., postmodern feminist responses to MacKinnon's work.

[4] Note, however, that I don't want to support an approach like Fields' et al that drops racism as an explanatory factor out of the picture and makes the model primarily economic! The point is about needing multiple factor explanation.

[5] Note that feminist theorists have mostly given up the project of theorizing the "origins" of male domination. This is partly due to the very real scholarly problems that arise in getting good evidence of the origins. The project is much more focused these days on how male domination is sustained. Given the emphasis in Mills' work on the issue of origins of racial domination, I wonder whether there are important differences between gender and race on this. Also, note that feminists have not considered the issue of reparations for women...but if there is and has been a gender contract that has excluded women from wealth, earning potential, etc., should there be a parallel argument? What would reparations to women look like?

[6] Possibly a helpful suggestion: it might work to contrast "hypothetical" readings from "subjunctive" readings of the contract, rather than two senses of hypothetical. Hypothetical could be concerned with the possibility of inexplicit v. explicit contracts: if they had had the opportunity, they would have consented/contracted (or something like that). Subjunctive would be: it is as if it were contractual.

[7] Examples are from Mark Bedau, "Where is the Good in Teleology?" in C. Allen, M. Bekoff, and G. Lauder ed., Nature's Purposes, (MIT Press, 1998), pp. 261-291.