Abstraction, Idealization, and Oppression

ABSTRACTION, IDEALIZATION, AND OPPRESSION

LISA H. SCHWARTZMAN

Abstract: Feminists, critical race scholars, and other social-justice theorists some-times object to ‘‘abstraction’’ in liberal normative theory. Arguing that oppressionaffects individual agents in powerful yet subtle ways, they contend that allegedlyabstract theories often reinforce oppressive power structures. Here I criticallyexamine and ultimately reject Onora O’Neill’s ‘‘abstraction without idealization’’as a solution to this problem. Because O’Neill defines abstraction as simply the‘‘bracketing of certain predicates,’’ her methodology fails to guide decisions aboutwhat to bracket and what to include in the theory. Moreover, it may not bepossible to abstract without also relying on some particular ideals. Whileabstraction is unavoidable, I conclude that it must be employed with greaterattention to sociopolitical hierarchies and by using ideals that do not collude withstructures of oppression. I discuss the work of Susan Babbitt and ElizabethAnderson as examples of how nonoppressive ideals might be incorporated intonormative theory.

Keywords: abstraction, Elizabeth Anderson, Susan Babbitt, feminism, idealiza-tion, ideals, luck egalitarianism, Onora O’Neill, oppression, pornography.

1. Introduction

In theorizing about justice and equality, liberal political philosopherstypically engage in various forms of abstraction. For instance, JohnRawls and Ronald Dworkin both construct abstract ideals that theycontend shed light on the nature of justice and equality. In the introduc-tion to A Theory of Justice, Rawls explains that his aim is ‘‘to present aconception of justice which generalizes and carries to a higher level ofabstraction the familiar theory of the social contract’’ (1971, 11). Ratherthan examining the actual functioning of our own society, Rawls insteadexamines the ‘‘principles of justice that would regulate a well-orderedsociety’’ (8).1 Although Rawls’s more recent work in Political Liberalism

1 Defending his emphasis on ‘‘ideal’’ theory, Rawls writes: ‘‘The reason for beginningwith ideal theory is that it provides, I believe, the only basis for the systematic grasp of thesemore pressing problems. . . . I shall assume that a deeper understanding can be gained in noother way, and that the nature and aims of a perfectly just society is the fundamental part ofthe theory of justice’’ (1971, 9).

r 2006 The AuthorJournal compilation r 2006 Metaphilosophy LLC and Blackwell Publishing LtdPublished by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK, and350 Main Street, Malden, MA 02148, USAMETAPHILOSOPHYVol. 37, No. 5, October 20060026-1068

r 2006 The AuthorJournal compilation r 2006 Metaphilosophy LLC and Blackwell Publishing Ltd

may seem less ‘‘abstract,’’ insofar as Rawls more explicitly situates thisproject in the context of contemporary liberal democracies, he none-theless continues to emphasize the role of abstraction in normativetheorizing:

The work of abstraction . . . is not gratuitous: not abstraction for abstraction’ssake. Rather, it is a way of continuing public discussion when shared under-standings of lesser generality have broken down. . .the deeper the conflict, thehigher the level of abstraction to which we must ascend to get a clear anduncluttered view of its roots . . . formulating idealized, which is to say abstract,conceptions of society and person . . . is fundamental to finding a reasonablepolitical conception of justice. (1993, 45–46)

Thus, throughout his writing Rawls stresses the importance of abstrac-tion and the need to devise and justify abstract, idealized models ofsociety and of individuals.

Dworkin’s work differs from Rawls’s in a number of ways, butDworkin too believes that political theories should be abstract; likeRawls, he constructs an abstract ideal to illustrate and defend hisconception of distributional equality, which he calls ‘‘equality of re-sources’’ (Dworkin 1981). In Dworkin’s ideal, immigrants on a desertisland must decide how to distribute the island’s resources. Dworkinargues that people in such a situation would choose to employ an auction,an ‘‘envy test,’’ and a hypothetical insurance market, and he contendsthat this idealized model can guide assessment of our current society’sdistributive mechanisms and social institutions.

Although Rawls, Dworkin, and other liberals celebrate its virtues,abstraction has recently come under attack by a variety of opponents,including communitarians, conservative critics of liberalism, feminists,and critical race theorists. While liberal political philosophers haveresponded at length to conservative and communitarian criticisms, theyhave devoted much less attention to feminist and antiracist objections toliberal ‘‘abstraction.’’ Feminists, critical race scholars, and other social-justice theorists argue that the social context is characterized by varioushierarchies, such as gender, race, sexuality, and class, and that individualagents are affected by these structures in powerful but often invisibleways. They contend that the effects of oppression can be deep seated andyet invisible, and that attempts simply to ‘‘abstract’’ from all knowledgeabout the social structure may backfire and implicitly endorse the effectsof dominant ideologies. Nonetheless, because theories consist of general-ized claims that transcend the specifics of any given situation, theorizingdoes seem to require some form of ‘‘abstraction.’’

In this article, I explore the issue of abstraction by examining the workof a neo-Kantian liberal theorist, Onora O’Neill, who emphasizes thecentral role that abstraction plays in normative theorizing. On the onehand, O’Neill notes that abstraction is itself insufficient; it must be


LISA H. SCHWARTZMAN566

supplemented by a ‘‘plausible conception of practical reason’’ (1996, 44).Nonetheless, O’Neill’s conception of practical reason relies heavily onabstract principles as its ‘‘starting points.’’ Even though she acknowledgesthat such starting points are ‘‘not enough’’ and that we need ‘‘a vindicableconception of how to proceed’’ (44), the procedure of practical reasonthat she constructs and defends rests on a foundation of very abstractprinciples.

I focus here on O’Neill’s discussion of abstraction for a couple ofreasons. First, O’Neill offers a strong defense of the sort of abstractionthat is common in the work of contemporary liberal political philoso-phers, and thus it is useful to examine her work carefully. Second,O’Neill is aware of a number of problems that can arise in employingabstract methods of theorizing. Her work is appealing because sheopposes the use of ‘‘idealized’’ accounts of matters like human agency,rationality, and independence that often appear in allegedly abstracttheories. Nonetheless, O’Neill argues that these idealizations are not anecessary part of abstraction. According to O’Neill, abstraction can andshould proceed in the absence of any idealized account of such matters,and it should be done without employing any particular normative‘‘ideals.’’

After explaining O’Neill’s defense of ‘‘abstraction without idealiza-tion,’’ I present a number of criticisms of her view. Situating myarguments in the context of an analysis of oppression, I contend thatmethods of abstraction often involve the bracketing of information that iscrucial to understanding the nature and sources of oppression. Using thework of Susan Babbitt and Elizabeth Anderson, I argue that normativeidealsFwhich O’Neill rejectsFmay play an important role in under-standing oppressive social structures and in envisioning democratic andnonoppressive alternatives. Rather than setting up unattainable ideals forhow individuals should be, the ideals offered by Babbitt and Anderson areput forth in an effort to transform society. In the final section of thearticle, I offer an illustration of how feminists can use a reconfigured idealof equality to argue against the oppressive effects of certain forms ofpornography. Finally, I conclude by suggesting that normative ideals maybe unavoidable in theorizing and that feminist, antiracist, and other formsof critical theory can benefit from ideals that aim to challenge andcreatively transform social structures of power.

2. Abstraction and Idealization: O’Neill’s Account

At various points in her writing, O’Neill defines abstraction as merely amatter of ‘‘bracketing’’ certain predicates (1996, 40) or ‘‘detaching certainclaims from others’’ (1993, 309); it is ‘‘a matter of selective omission, ofleaving out some predicates from descriptions and theories’’ (1988, 711).In this sense, she notes, abstraction is ‘‘theoretically and practically


ABSTRACTION, IDEALIZATION, AND OPPRESSION 567

unavoidable. . . . All uses of language must be more or less abstract; somust all reasoning’’ (1996, 40). Thus, she claims there is an importantsense in which criticism of abstraction is misdirected.

Examining the attacks on abstraction, O’Neill isolates two criticisms:the charge of empty formalism and the objection to abstract views ofagents. In response to the first charge, which alleges the ‘‘formalism andemptiness of all practical reasoning that invokes principles or rules’’(1987, 67), she writes that while it ‘‘is true that principles underdeterminedecisions’’ (1987, 58), this is not a reason to reject them. Principles andrules must be supplemented by good judgment; they are never sufficientdeterminants of action. In other words, principles need not be understoodin an algorithmic fashion.

The second charge, that abstraction in ethical reasoning is proble-matic because it relies on ‘‘abstract views of agents’’ or on ‘‘idealiza-tions’’ of human agency, is more complex. According to O’Neill,proponents of this critique, which she associates with Hegelian andMarxian attacks on ‘‘abstract individualism,’’ assert that abstract ethicaland political theories make assumptions that actual human agents do notsatisfy. For instance, many abstract theories assume that individualshave capacities to reason, choose, or form preferences that only idealizedagents possess. While O’Neill agrees with the substance of this criti-cismFshe concurs that people do not have ‘‘transparent self-knowl-edge’’ or ‘‘archangelic insight into others’ preferences’’ (1987, 56)Fshedoes not conclude that this is a problem with abstraction. Rather, it is aproblem with idealization:

The objection is not just that much (too much) that is true of human agents isomitted in some accounts of agents, but that much (too much) that is false ofhuman agents is added. Descriptions of agents in much post-enlightenmentethical and political theory are often idealized; they are satisfied only byhypothetical agents whose cognitive and volitional capacities human beingslack. (1987, 56)

In objecting to what she refers to as idealization, O’Neill seems to agreewith at least some aspects of the feminist critique of abstraction. Forinstance, some feminist objections to liberalism can be understood asobjections to the ways that allegedly abstract theories actually makecertain assumptions that are not abstract at all but instead presupposeparticular conceptions of human rationality, independence, preferenceformation, or some other capacity. These assumptions (or ‘‘idealiza-tions’’) tend to privilege men, whites, upper-middle-class people, andothers who have power within sociopolitical structures. Although O’Neillfocuses less on this aspect of the problem, she does note that idealizationtends to favor the more privileged members of society. In a discussion ofjustice, gender, and international boundaries, she suggests that theproblem with idealizations is not simply that they are false but also



that they tend to favor ‘‘specific conceptions of the human agent and ofnational sovereignty which are often more admired and more (nearly)feasible for developed rather than developing societies and in men ratherthan in women’’ (1993, 305). In the same article, she also admits the scopeof this problem: it is not just a couple of theorists who engage in this sortof idealization. Rather, ‘‘[m]uch contemporary moral reasoning, and inparticular ‘abstract liberalism’ (whether ‘deontological’ or utilitarian)’’suffers from this problem (1993, 305).

In addition to this first type of idealization (which posits agents withsuperhuman capacities), O’Neill describes and criticizes a second sort:many moral and political theorists ‘‘treat enhanced versions of certaincapacities as ideals for human action’’ (1987, 56). According to O’Neill,positing ideals of ‘‘[r]ational economic men, ideal moral spectators,’’ and‘‘utilitarian legislators’’ ends up suggesting that these superhuman qual-ities are ones for which actual people should strive (1987, 56). She objectsto this because such ideals are simply unachievable and are often‘‘irrelevant to human choosing.’’ There are many domains where we donot want people to be rational calculators; for instance, she points out theabsurdity of ‘‘ideal rational friends and lovers’’ (1987, 56). In other words,these particular ideals may seem ‘‘ideal’’ for the purpose of certaintheories but may not be ones that we would want people to embody inall realms of human activity. Moreover, O’Neill objects to reliance onideals, since there is no way to justify them without falling back onassumptions about metaphysical truths. As she explains, ‘‘Writers whostart by assuming ‘ideals’ cannot justify them as mere abstractions,and need to supply appropriate justifications: this may be uphill workif they hope to reject metaphysical certainties that might ground demand-ing conceptions of the objectively good, or of the person’’ (1996, 43).Thus, O’Neill concludes, moral theory should not rely on idealizedassumptions that are not really true of actual people (the first sort of‘‘idealization’’), nor should it aspire to particular ideals that it posits asworthy and desirable goals but that are unproven and merely assumedfrom the start.

After describing these two types of idealization, O’Neill proposes whatsounds like a simple solution: moral and political theorists should rejectidealization, in both its forms, but they should not reject abstraction.Referring to her solution as ‘‘abstraction without idealization,’’ sheargues that the problem with the critics of abstraction is that their failureto separate idealization from abstraction leads them, wrongly, to rejectabstraction:

The lack of justification for starting with such ‘ideal’ conceptions of persons,reason or autonomy provides good reasons for querying these positions, butno reason for being suspicious about mere abstraction. . . . The manyconvincing criticisms of the idealized conceptions of persons, reason and



action, on which one or another ethical universalist has in fact relied, whileclaiming only to abstract, do not show that there is anything wrong withabstraction. (1996, 43)

In other words, these theorists highlight problems with idealization andthen make the mistake of assuming that such problems are inherent in anytheory that uses abstraction. However, according to O’Neill, in order todraw this much more drastic conclusion, one would have to show thatsimilar problems arise with ‘‘approaches that abstract, but do not invokeunvindicated ideals’’ (1996, 43). Yet this has not been established. Thus,O’Neill concludes that much criticism of abstraction has been seriouslymisplaced.

3. Critique of O’Neill

One set of problems with O’Neill’s proposal concerns her understandingof the common criticisms of abstraction. While she is right to note that itis difficult to object to abstraction itself when abstraction is defined as themere ‘‘bracketing’’ of certain predicates (or ‘‘detaching certain claimsfrom others’’), it is not clear that this is what the critics of abstraction areattacking. Many of their objections can be understood as objections tocertain types of bracketing, and some seem to be arguing that abstractapproaches attempt to bracket too much. The problems these critics raiseoften pertain to the degree or the form of bracketing, not to the act ofbracketing itself.

Moreover, O’Neill’s definition of abstraction as bracketing or as the‘‘selective omission’’ of certain predicates makes it an odd candidate forpraise. She suggests that so long as idealizations are avoided, ‘‘genuineabstraction’’ (abstraction that does not idealize) can serve as the founda-tion for moral theory. Given her definition of abstraction, however, it isnot at all clear what this would involve. If the omissions are to be‘‘selective’’ and are to focus on certain predicates rather than others,decisions must be made about what should be bracketed and what shouldbe included. Unfortunately, she does not say much about the details ofabstraction; nonetheless, she argues that normative theory needs moreabstraction, not less.

A number of problems also arise with O’Neill’s rejection of what sheconsiders to be the two different types of idealization: (1) idealizedassumptions that are false of actual agents, such as assumptions abouthuman agency, rationality, and relationships; and (2) ideals posited by thetheoristFthat is, ideals that the theorist acknowledges are not yetdescriptive of actual persons but are offered as goals to be achieved.O’Neill rejects both sorts of idealization and suggests that both can beavoided by pursuing a strategy of ‘‘mere abstraction.’’

O’Neill objects most strongly to the first type of idealization shedescribes. When idealized assumptions about such matters as human



agency and human rationality turn out to be false, the models andtheories that arise from these assumptions have no real application in theactual world. Furthermore, these theories are not only false but alsodangerous; attempts to apply them often work in favor of the mostpowerful members of society (upper-middle-class white men, for instance)since the lives of the privileged are most likely to embody these idealizedassumptions.

Although O’Neill’s initial (and most forceful) objection to theseidealized assumptions is that they are false of actual agents, her proposedsolution suggests that it is not merely their falsehood to which she objects.For instance, in a discussion of justice, gender, and internationalboundaries, O’Neill argues that we must ‘‘resist the temptation to relyon idealizing models of human agency or national sovereignty. We shouldinstead consider what sort of theory of justice we would have if weabstracted but refused to idealize any one conception of rationality orindependence, and so avoided marginalizing or excluding those who donot live up to specific ideals of rationality or of independence fromothers’’ (1993, 312; emphasis added). Elsewhere in the same article,O’Neill suggests that we canFand shouldFproceed without any partic-ular conception of gender or national sovereignty.2 Thus, O’Neill is notsuggesting that we simply replace false or oppressive conceptions ofrationality, independence, gender, and national sovereignty with onesthat are more accurate or less oppressive; rather, she is suggesting thatwe attempt to theorize without employing any particular conceptionsat all.

This reading of O’Neill is supported by her discussion of this problemin Constructions of Reason, where she explicitly criticizes Rawls’s con-structivism for relying on a ‘‘certain ideal of the moral person’’ that sheclaims is ‘‘built into the original position’’ (1989, 210). Specifically,O’Neill objects to the way that Rawls constructs agents so that they aremotivated only by the ‘‘highest-order interests in their moral powers’’ andby their ‘‘concern to advance their determinate but unknown final ends’’(1989, 210). For O’Neill, this account of the original position ‘‘explicitlybases principles of justice not on a conception of the person that abstractsfrom the diversity of human agents, but on one that idealizes a certainsort of agent’’ (1989, 210). The problem, she suggests, is that Rawls doesnot abstract sufficiently; he imports some conception of the person when

2 O’Neill argues that, first, we must devise abstract principles that do not idealizeFthatis, that do not assume any specific conception of the human agent. Next, we must apply theseprinciples without relying on any specific ideals. Explaining this second move, the one ofapplication, she states that ‘‘justice can take account of certain differences by applyingabstract principles to determinate cases’’ without ‘‘tacitly reintroducing restricted ideals(e.g., by privileging certain views of gender and sovereignty)’’ (1993, 305). In other words,ideals can be avoided not only in the construction of abstract principles but also in theirapplication.



he should have relied solely on an abstract account of agency. This couldhave been avoided if he had employed ‘‘a more Kantian constructivism,’’one that began ‘‘from the least determinate conceptions both of therationality and of the mutual independence of agents’’ (1989, 212). Whilethis abstract version of constructivism could not guide us in choosingprinciples as concrete as Rawls’s, it could justify our ruling out principlesthat any ‘‘abstractly characterized’’ plurality of agents would reject (1989,218).

For instance, O’Neill explains that her version of constructivism, basedonly on an abstract account of the person, entails the following: ‘‘Noplurality can choose to live by principles that aim to destroy, undercut orerode the agency (of whatever determinate shape) of some of itsmembers’’ (1989, 213). Or, as she puts it elsewhere, ‘‘if a plurality ofagents is to share principles, action on the principles must leave theagency of each member of the plurality intact’’ (1988, 718). What thismeans, more specifically, is that agents must avoid coercion and decep-tion, since these are ways of ‘‘destroying and subverting others’ agency’’(1988, 719). Of course, how one interprets these principles, and under-stands them to apply in any particular case, depends on many features ofthe circumstances, society, and context. O’Neill does not seem to worryabout this problem, and she does not view this as an attack on theprinciples themselves, since it is not the job of principles to include specificrules for their application.

Although one could object to O’Neill’s somewhat hasty dismissal ofthis problem of the interpretation and application of principles, I leavethis matter aside and move on to two other problems with her objectionto idealization (and with her defense of ‘‘mere abstraction’’). In order toshow that normative theorizing must do without any particular concep-tions of matters like human agency, rationality, or dependence, O’Neillwould have to show both (a) that it is the use of these conceptionsthemselves that is problematic and (b) that it is possible to do withoutany such conceptions. However, she fails to establish either of theseclaims.

With respect to the first matter, it is at least possible that the particularproblems with the idealizations she discusses could be solved by replacingthe idealized assumptions with ones that are more accurate and lessideological. For instance, the assumption that agents are fully rational,mutually independent decision makers may be problematic not because itinvolves a particular conception of human agency but rather because it isfalse of most actual human agents. Moreover, it is ‘‘ideological’’ in thesense that it obscures relations of power that exist among human agentsand it often works to bolster the status of oppressive groups. A moredirect solution to the problem of false and ideological conceptions ofhuman agency and functioning would be simply to replace these falseconceptions with more accurately descriptive ones. Likewise, rather than



suggesting that theory should proceed without any particular conceptionof gender (as O’Neill does in ‘‘Justice, Gender, and InternationalBoundaries’’), one might instead address the problem by replacing sexistand ideological assumptions about gender with a more accurate concep-tion, one that does not presume women’s inferiority.3 O’Neill has shownonly that the particular conceptions employed in the theories shediscusses are objectionable. She has not established that all such con-ceptionsFof gender, of national sovereignty, and of the person ingeneralFare problematic.4

Second, it may also be the case that normative theory cannot dowithout some conception of the subjectFfor instance, an account ofhuman rationality or of relations of dependence. As I noted earlier,because abstraction always abstracts from something in particular, thecharacterization that one gives to human agents can only be ‘‘abstract’’ incertain respects. The features that remain in O’Neill’s account cantherefore be said to be forms of idealization, which O’Neill should opposeif her account is to be consistent.

On the surface, O’Neill suggests only that her abstractly characterizedagents are ‘‘diverse,’’ but it may be that a more substantive form ofidealization is lurking in her allegedly abstract account. In focusing onrational agency, rather than on other aspects of personhood, she seems toprivilege a certain ideal of the person. This is not to say that thisidealization is false, it is simply to say that it involves a ‘‘certainconception of the human agent,’’ one that might not be shared byeveryone. Her focus on agency can be seen in her statement that ‘‘if aplurality of agents is to share principles, action on the principles mustleave the agency of each member of the plurality intact’’ (1988, 718).Although this may sound unproblematic to someone who is accustomedto the liberal democratic tradition, and who takes for granted its focus onthe importance of individuality and autonomy, it may not be what definesthe person for others who do not share these background assumptions.

3 As cited above, O’Neill argues that we can apply abstract principles to determinatecases without ‘‘privileging certain views of gender’’ (1993, 305).

4 In a commentary on an essay of O’Neill’s, Martha Nussbaum contrasts O’Neill’sapproach to the issues of justice and gender to another approach (Nussbaum’s own), onethat is based on a particular conception of human functioning that arises out of discussionsin Aristotle (Nussbaum 1993). Nussbaum argues that her own human-functioning approachis superior to O’Neill’s Kantian emphasis on ethical consistency. According to Nussbaum, itis ‘‘essential to focus on content . . . the actual doings and beings of people . . . not just whattraffic rules will police their doings and beings. Indeed, it seem[s] hard to say anythingmeaningful about traffic rules until we know who the parties are and what they are doing’’(1993, 330). Although I do not find Nussbaum’s particular Aristotelian discussion of humanfunctioning entirely convincing, her general point against O’Neill is a powerful one. Oneneed not support the specific Aristotelian approach that Nussbaum advocates to agree thatnormative theory requires holding some particular conception of the person.



In addition to claiming to avoid relying on any particular conceptionof the person, O’Neill denies employing any ideals about what humanbeings should be like: she suggests that one must theorize withoutassuming any particular conception of ‘‘the good.’’ (This is the secondsense of idealization to which O’Neill objects.) However, just as it may beimpossible to avoid the first sort of idealization, it may also be impossibleto theorize without the use of some particular normative ideals. A numberof O’Neill’s critics have suggested that the distinction she draws betweenabstraction and idealization is problematic, and several have pointed toher own reliance on some sort of ideal as part of this problem. Forinstance, in a review of O’Neill’s Constructions of Reason, Allen Woodnotes that, as a part of her alleged abstraction, O’Neill professes to be‘‘agnostic’’ about the human good and about the particular value of theperson. Nonetheless, what she achieves is merely a ‘‘selective agnosticismabout the human good’’ (Wood 1992, 650). Wood explains, ‘‘O’Neill’sKantian constructivism affirms the value for human beings of exercisingtheir rational agency; it seeks to protect that value at all costs, treatingother components of the human good as derivable from that one’’ (650).For Wood, O’Neill’s analysis ‘‘threatens to involve itself in a conceptionof the human good that is not merely an abstraction but also anidealization’’ (650). Although rational agency may be important, wesomehow would need to be able to measure its importance against otherhuman goods in order to determine exactly how important it is. Wecannot do this with O’Neill’s account, however, since she requires that weabstract from all other components of the human good. Wood concludesby noting that although O’Neill aims to criticize liberal theorists whopretend to be agnostic about the human good but nonetheless endorseparticular idealizations, she herself succumbs to this very problem.5

4. Abstraction and Oppression

If it turns out that we cannot simply do away with all conceptions of theperson and all use of normative ideals, we may need to focus more on thequestion of how to avoid employing false idealizations and oppressive orideological normative ideals. On a similar note, because abstractionalways involves the bracketing of certainFbut not allFpredicates,particular attention must be paid to the specific forms of abstractionthat a theorist employs. In this section, I explore the use of abstraction inthe context of circumstances of oppression.

5 Wood actually puts this more strongly: ‘‘This deceptive movement from pretendedagnosticism through abstract formalism to pernicious idealization is so typical of traditionalliberal thinking as to be almost its defining trait. At her best, O’Neill often exposes liberalismon this score, so it is too bad that as a good Kantian she must at last succumb to it herself ’’(1992, 650).



Rather than arguing that abstraction has no role whatsoever innormative theory, some critics of abstraction object to the way thatcertain important features of the social structure are allegedly set aside, orbracketed, while these features nonetheless continue to influence aspectsof the theory. For instance, a number of feminist legal theorists contendthat many allegedly abstract theories actually assume a male norm. Thecontext of male power is important in considering the meaning ofdomestic violence, rape, abortion, and numerous other feminist issues,and yet for many years political and legal theories dealt with these issueswhile allegedly bracketing considerations of gender. The solution thatsome feminists have proposed is not simply to do away with allabstraction (and focus only on the concrete particularities of each andevery individual), nor is it to eliminate all conceptions of the person or alluse of ideals. The best way to approach this problem may be to payattention to the kinds of abstractions, ideals, and conceptions of theperson that the theory employs and to disallow those that are sexist andthat reinforce or perpetuate women’s oppression.

In contrast to this feminist approach, O’Neill does not focus on whichsorts of abstractions she is employing other than to say that it is necessaryto avoid the use of idealizations and ideals. She claims to be engaging in amethod that is ‘‘merely abstract,’’ and she does not seem to acknowledgethat such abstraction necessarily includes certain features and facts andexcludes others. O’Neill’s suggestion, however, that we begin theorizingwith very abstract principles that everyone would agree to (without anyuse of moral ideals or conceptions of the person) could work to reinforcesocial power structures, and it could also end up presupposing falseidealizations. The danger of this occurring seems greatest in cases wheredeeply entrenched hierarchies affect the sense of self that individualagents possess.6

To understand how employing the method of ‘‘abstraction withoutidealization’’ could end up reinforcing oppressive hierarchies, considerwhat happens when we apply O’Neill’s method to a situation in whichsome of the agents possess a diminished sense of self-worth. In the contextof a discussion of servility and self-respect, Thomas Hill describes the caseof what he calls the ‘‘Deferential Wife,’’ a woman who is so devoted toserving her husband that she ‘‘tends not to form her own interests, values,and ideals; and, when she does, she counts them as less important than

6 Although O’Neill suggests that abstraction without idealization could get around thisproblem of ideology, or ‘‘false consciousness,’’ it is not at all clear that it provides the bestmethod for dealing with these deep-seated hierarchies and inequalities. Note that my claim isnot simply that the principles that an abstract method produces could be applied in ways thatare problematic (this is another problem); rather, I argue that the principles devised from amethod that claims to be ‘‘abstract’’ may themselves turn out to be infected with oppressiveand hierarchical thinking. The mere attempt to be abstract, and to avoid false idealizations,does not guarantee that one will achieve this goal.



her husband’s’’ (1991, 5). What would happen if O’Neill’s method ofabstraction were employed by a plurality of agents, some of whom werelike the Deferential Wife?7 It seems that O’Neill has not ruled out this sortof agent, since her method of abstraction is meant to apply to anyminimally rational agent who has an indeterminate amount of mutualindependence from others. The problem is that she assumes that theseagentsFhowever characterizedFwould not choose principles that wouldundercut their own agency: ‘‘No plurality can choose to live by principlesthat aim to destroy, undercut or erode the agency (of whatever determi-nate shape) of some of its members. . . .Victims cannot share the principleson which others destroy or limit their very capacity to act on principles’’(O’Neill 1989, 213). But what if the situation is one of systematicoppression in which some agents lack the self-respect that would motivatethem to reject principles that victimize them or otherwise fail to promotetheir interests?

Rather than extensively discussing this sort of objection to herproposal, O’Neill mentions it only in a footnote and quickly dismissesit. She writes, ‘‘Colluding victims are not a counterexample. If collusionitself is coerced, they are not truly willing; if it is not, they are not trulyvictims’’ (1989, 213). In other words, if one agrees to a principle thatsubordinates one, then one must have been coerced. There is no way thatone can be a willing participant in one’s own victimization. In cases likethat of the Deferential Wife, O’Neill would probably say that to theextent that this woman approves of principles that victimize her, she isbeing ‘‘coerced,’’ and thus her approval is not genuine. Yet this seems abit too quick, since it dismisses the complexity of the very identity that thewoman possesses. Of course, there is something troubling about theidentity of this sort of person. Her very sense of self seems to beimpoverished, and she lacks the autonomous agency that O’Neill’s modelseems to presuppose. Rather than simply concluding that she is beingcoerced or that she is somehow mistaken about what she wants, onemight instead argue (as Susan Babbitt has, in a commentary on Hill’sessay) that it may be necessary for the Deferential Wife to undergo somesort of transformation before she will be able to form her own interests ina way that expresses a fuller sense of self (see Babbitt 1996, 43–46).However, it is precisely this sort of transformation that O’Neill’s analysisprohibits. According to her method of abstraction without idealization,when a plurality of agents chooses principles, ‘‘action on the principlesmust leave the agency of each member of the plurality intact’’ (1988, 718).

7 For a similar discussion of this example of Hill’s see Babbitt, 1996, 43–46. Babbitt’streatment of the case of the Deferential Wife focuses more on questions of personal integrityand on the problems of appealing to idealized accounts of the person (which O’Neill claimsnot to do, since she rejects such idealizations). Nonetheless, many of Babbitt’s points canalso be applied to an analysis like O’Neill’s, despite O’Neill’s rejection of idealization.



This is clearly problematic when some agents resemble the DeferentialWife. Preserving this sort of agency would likely reinforce what is alreadyan oppressive situation.

Although the Deferential Wife may be an exaggerated extreme, thesorts of attitudes and the sense of self that this woman possesses are notentirely foreign to the situation of many women in contemporary society.Because of gender inequality, and the norms and institutions thatstructure and perpetuate it, many women actually do view their owninterests as less important than those of their husbands; thus, they mightagree to principles that subordinate them in certain ways. Furthermore,similar examples could be raised concerning the situation of otheroppressed groups, although the analysis of their particular forms ofoppression would differ from that described here. The general problemconcerns the effects of internalized oppression, which O’Neill’s theoryFwith its emphasis on individual agencyFseems unlikely to combatsuccessfully.

5. Challenging Oppression: The Role of Ideals

To address the problem of impoverished agency in situations of oppres-sion, some feminists and other political theorists argue that specific idealsmust play a role in normative theorizing, and they contend that certainfeatures of the social context cannot be bracketed without running a highrisk of reinforcing oppressive power relations. In this section, I examinethe work of Susan Babbitt and Elizabeth Anderson, two feminist politicalphilosophers who explicitly employ normative ideals. While both theor-ists point to a variety of ways that normative ideals can be used, I focushere on Babbitt’s discussion of the sorts of ideals that individuals employwhen they undergo political transformation, and I describe Anderson’spresentation of an ideal of democratic equalityFan ideal that is positedas something for which society should strive. Both theorists pose achallenge to O’Neill by emphasizing the important role that ideals playin normative theorizing. Not only are these ideals unavoidable, they arealso crucial in challenging false and ideological ‘‘idealizations.’’

In her book Impossible Dreams: Rationality, Integrity, and MoralImagination (1996), Babbitt argues that abstraction itself will not beenough to eliminate injustice. Her point is not simply that abstractprinciples can be misapplied but rather that the understandings peoplehave of these allegedly abstract concepts are themselves skewed byinjustice. As she explains, ‘‘If sexism, say, is embedded in some of theconcepts and terms in which information and events are understood,including conceptions of justice, it cannot simply be taken for grantedthat considerations of justice, in the abstract, will show us how toeliminate it’’ (1996, 90). In other words, the allegedly abstract definitionsof ‘‘justice’’ and ‘‘morality’’ that are often central to liberal theory may



take for granted sexist, racist, and classist power structures and mayinadvertently support and perpetuate these structures. For example, in asexist context, the notion of morality or justice to which people appeal inconsidering an issue like pornography or abortion is somewhat differentfrom the way that morality or justice would be understood if it werealtered or transformed by feminist analysis. Within a sexist social context,people often believe that abstract notions of morality and justice pitquestions of individual privacy against the rights of the fetus in the case ofabortion. Similarly, when it comes to issues of pornography within asexist social context, people typically consider only how the pornogra-pher’s right to ‘‘free speech’’ is to be weighed against the rights of peoplewho are ‘‘offended’’ by sexually explicit material; the rights of women toexercise control over their own bodies and over their sexuality do notfigure prominently in either side of the debate on these issues, as it istypically construed.

In both these cases, however, sexist assumptions are embedded in thevery understandings of the terms: abortion is recognized only as a privacyright and not as something that specifically affects women’s ability tocontrol their own bodies and lives, and pornography is recognized only assomething that has the power to ‘‘offend’’ by depicting sex, rather than assomething that has devastating consequences for women’s status andequality. As a result, when people in a sexist social context think aboutwhat morality or justice requires in the abstract, this is the way the issueswill typically be formulated. As Babbitt notes, ‘‘If sexist assumptions areembedded in the social context of meanings and values, they are notunderstood as sexist assumptions; rather, they are norms’’ (1996, 93).

In order to root out sexist assumptions and change the social under-standings of what is required by ‘‘morality’’ and ‘‘justice,’’ more specificwork is needed. As Babbitt explains,

Discrimination that is built into social norms and values requires work to beidentified. Appeals to principles of justice, however defined, are not going to beenough if such discrimination is not identified as injustice to begin with.Changes to social structures may be required before the sexism that has alwaysbeen there comes to be understood and treated as sexism. (1996, 93)

In other words, it is not always evident what constitutes sexism, racism, orother forms of unjust power. To identify and uproot structures ofoppression, greater attention must be paid to the operation of specificsocial structures. Furthermore, appealing to individual interests orchoices will not be sufficient. Rather, social change must be undertakenas a necessary part of promoting the interests of individuals. Liberaltheory goes wrong when it assumes that we can unproblematically knowwhat these interests are without undertaking an analysis and critique ofsocial power structures. By employing methods of abstraction that fail toquestion the interests that individuals currently express, social power



structures may be reinforced and social change may be impeded. If wewish to transform oppressive structures, suggests Babbitt, we must atleast implicitly also appeal to some ideal or vision of how society could beotherwise. Because of the ways that sexism and racism (for example) arebuilt into many of our assumptions, it may be that identifying thewrongness of these forms of oppression requires ‘‘an alternative visionof what society should be and how it should be divided up, one thatpermits judgments about how the existing society is wrong’’ (1996, 26).Merely abstract principles may not be sufficient to do this, and they mayend up reinforcing the problematic and false idealizations of currenthierarchies.

Although her writing focuses on different issues from Babbitt’s,Elizabeth Anderson also argues for the importance, and the unavoid-ability, of ideals in normative theory. Rather than claiming that theseideals are abstract or that they can be derived from some ahistorical,acontextual, impartial starting point, Anderson suggests that they ariseout of concrete circumstances, that they are subject to change as they arequestioned and challenged, and that they evolve as society itself goesthrough transformations.

In Value in Ethics and Economics, Anderson defends a pluralist theoryof value that emphasizes the importance of normative ideals. These idealsare unavoidable, she says, since we require them to make sense ofourselves: ‘‘Ideals are the self-conceptions through which we try tounderstand ourselves, to make sense of our emotions, attitudes, andconcerns. . . . We make ourselves intelligible to ourselves by cultivatingattitudes that make sense to us, by determining to act in accord withideals we accept that have survived critical scrutiny’’ (1993, 91). Instead ofsuggesting that ideals are merely subjective visions that people hold butthat cannot be criticized by (or justified to) others, Anderson argues thatideals are contestable, that they require empirical support, and that theyare subject to criticism and revision.8 Thus, these ideals are not abstract inthe sense that they bracket all information about actual people, nor arethey merely subjective in the sense that they pertain only to an indivi-dual’s own personal vision and have no wider application.

To illustrate how an ideal that is situated in an understanding of socialrelations of power can be used to challenge problematic idealizations, Iturn now to Anderson’s discussion of the ideal of ‘‘democratic equality.’’In her essay ‘‘What Is the Point of Equality?’’ Anderson uses this ideal tochallenge a variety of liberal theorists who advocate what she calls ‘‘luck

8 For a more extensive discussion of the process of criticizing and justifying normativeideals, see Anderson 1993, 91–116. Anderson argues that ‘‘(1) criticizing an ideal requiresinterpreting its associated attitudes; (2) interpretations can be supported by empiricalevidence; and (3) interpretations of attitudes undermine or support their endorsement’’(1993, 95).



egalitarianism’’ or ‘‘equality of fortune.’’ Rather than simply offeringnegative criticismsFby showing how these theories fail to embodycertain values that they purport to advocateFAnderson proposes analternative ideal, democratic equality. Her strategy of employing an idealthat does not abstract from social relations of power is superior to simplyattempting to abstract from all knowledge about society, rationality, andhuman behavior and agency. While attempting to engage in theorizingthat is ‘‘merely abstract’’ makes it difficult to examine the problematicrelations of power that serve as the background context of such theoriz-ing, Anderson’s method of employing an ideal that is situated in socialanalysis is more likely to succeed in identifying and challenging unjuststructures of power.

According to Anderson’s description, luck egalitarianism encompassesboth equality of resources and equality of welfare theories, and it ‘‘relieson two moral premises: that people should be compensated for unde-served misfortunes and that the compensation should come only fromthat part of others’ good fortune that is undeserved’’ (1999, 290).Although she launches a thorough and convincing attack on this theory,I focus here only on a few significant aspects of her critique. Mostimportant, Anderson identifies the problematic assumption that equalityconcerns the luck or fortune to which individuals are subject once theyhave been awarded that for which they are legitimately ‘‘responsible.’’Two problems that Anderson notes with this conception of equality arethat it assumes that we are in a position to assess individual responsibility,and that it centers more on matters of individual luck or chance than onthe systemic, socially and politically created forms of inequality.9

In contrast to this concentration on individual luck and personalresponsibility, Anderson’s ideal of democratic equality focuses on thesocial structure and aims to create less hierarchical and more genuinelydemocratic relations among people. She explains: ‘‘Egalitarianism oughtto reflect a generous, humane, cosmopolitan vision of a society thatrecognizes individuals as equals in all their diversity. It should promoteinstitutional arrangements that enable the diversity of people’s talents,aspirations, roles, and cultures to benefit everyone and to be recognized asmutually beneficial’’ (1999, 308). Thus, in acknowledging that socialconditions are currently not reflective of this genuine human diversity,Anderson argues that some ideal or vision of a more egalitarian societymust play a role in transforming our current social order.

9 Anderson explains that all luck egalitarians ‘‘place great stress on the distinctionbetween the outcomes for which an individual is responsibleFthat is, those that result fromher voluntary choicesFand the outcomes for which she is not responsibleFgood or badoutcomes that occur independent of her choice or of what she could have reasonablyforeseen’’ (1999, 291). Thus, Anderson emphasizes that these theorists rely on thisdistinctionFand require that we be able to make such a distinctionFin order for thetheory to work.



In addition to demonstrating the positive value of appealing to anormative ideal, Anderson illustrates the importance of examiningFandrefusing to abstract fromFsocial relations of power. Whereas O’Neillattempts to begin theorizing from a starting point of ‘‘merely abstract’’principles, Anderson contends that it is necessary to appeal to anunderstanding of the social order in challenging particular normativetheories. Because democratic equality ‘‘aims to abolish socially createdoppression’’ (1999, 313), our pursuit of this ideal requires us to examinesociety and to acknowledge its hierarchical structures; thus, we mustattempt to construct and apply our normative theories in the context ofthis understanding. One problem with luck egalitarians is that they payinsufficient attention to the extent to which institutional arrangementsaffect people’s choices and limit the options of what they can and willagree to under certain circumstances. In response to the emphasis thatluck egalitarians place on the ‘‘voluntary agreements’’ of individuals,Anderson replies that the ‘‘suffering and subjection generated by people’svoluntary agreements in free markets’’ should not be justified by the merefact that these situations arose due to voluntary choices: ‘‘[F]ree choicewithin a set of options does not justify the set of options’’ (1999, 308–09).Examining the social structures in which individuals make their choices isa crucial part of understanding how ‘‘free’’ these choices really are.Simply by bracketing facts about the social structure, as O’Neill’s methodof abstraction advocates, we will be unlikely to uncover and challengethese problematic assumptions.

In contrast to O’Neill’s unwillingness to endorse any particular values,ideals, or conceptions of the good, Anderson’s theory of democraticequality would work to guarantee certain social goods for everyone.10

Anderson points out that the luck egalitarians seek to compensate peoplefor ‘‘undeserved losses of all kinds within the general space of equalitythey specify (welfare or resources)’’ (1999, 302) and that they wouldoppose any more specific guarantees by society. They rely on individualchoices and preferences to such an extent that they preclude larger,collectively made decisions about goods to which each person is entitled.Even if these goods would not be chosen by individuals, they may still beessential in guaranteeing the conditions of equality. Anderson notes that‘‘egalitarian theory . . . must supply principles for collective willingFthatis, for what citizens should will together, not just for what each can willindividually’’ (310). Although O’Neill does not rely on market decisions,as the luck egalitarians all do (in various ways), she does advocate

10 An important feature of Anderson’s democratic equality is that it emphasizes ‘‘theneed to distinguish between goods that society guarantees to all citizens and goods that maybe entirely lost without generating any claims to compensation’’ (1999, 301). This does notmean that individuals are to be guaranteed certain ‘‘aggregate levels’’ of goodsFeither interms of property or in terms of welfare. Rather, Anderson argues that ‘‘[e]galitarians musttry to secure certain kinds of goods for people’’ (1999, 302).



beginning with principles to which all could agree. She seems toemphasize individuals’ making decisions for themselves in accordancewith the decisions that other individuals make for themselves, and thisdiffers from Anderson’s attention to what could be willed collectively.Anderson specifically notes the possibility that some people could agreeto situations that oppress them, but she argues that this does not justifysuch arrangements: ‘‘If people find happiness in their lives despite beingoppressed by others, this hardly justifies continuing the oppression’’(304). Challenging the oppressiveness of such a situation involves goingbeyond the choices of individual persons, and it may require appealing tosome sort of normative ideal.

6. Application: Employing Feminist Ideals

To illustrate the value of using feminist ideals that are concrete (asopposed to abstract) and nonindividualistic, I will briefly discuss how areconfigured feminist ideal of equality might function in a feministcritique of certain types of pornography. Pornography has been con-troversial among feminists: Some propornography (or ‘‘sex-positive’’)feminists argue that there is nothing wrong with pictures that portraytransgressive sexual acts and activities. They suggest that in the context ofsociety’s puritanical and repressive views about sex, the proliferation ofpornography might even be seen as a challenge to conservatives whoargue that sex outside marriage is always immoral.11 Furthermore, somewomen who work in the ‘‘sex industry’’ as models for pornography, asstrippers, or as prostitutes, argue that their work is a job like any otherand should not be stigmatized by feminists.12 Against these claims,feminist antipornography activists argue that certain forms of pornogra-phy are oppressive to women because they reinforce, perpetuate, andpromote the already prominent view that women are sexual objects

11 Citing the benefits of pornography, Lisa Duggan, Nan Hunter, and Carole Vancecontend that ‘‘the existence of pornography has served to flout conventional sexual mores, toridicule sexual hypocrisy, and to underscore the importance of sexual needs. Pornogra-phy . . . advocates sexual adventure, sex outside of marriage, sex for no reason other thanpleasure, casual sex, anonymous sex, group sex, voyeuristic sex, illegal sex, public sex. Someof these ideas appeal to women reading or seeing pornography, who may interpret someimages as legitimating their own sense of sexual urgency or desire to be sexually aggressive’’(1993, 156).

12 For instance, Debi Sundahl argues that women who work in the sex industry sufferfrom many of the same problems as women working in other jobs. Discussing the work of anerotic dancer, she maintains that ‘‘stripping is traditional women’s work as much aswaitressing, teaching and secretarial work is. Consequently, it suffers from the same lowpay . . . . The working conditions, overall, are also poor; many theaters are run on a quasi-legitimate financial basis, and are not clean or safe’’ (1987, 178). According to Sundahl, the‘‘anti-sex attitudes of the anti-porn movement’’ (179) further stigmatize women who alreadyface many of the same obstacles as other low-income women.



available for male use and consumption. Pornography that portrays maleviolence against women sexualizes this violence and thus contributes to aculture in which rape, domestic violence, sexual harassment, and otherforms of sexist oppression are more likely to occur.13

Although I will not examine all the arguments on each side of thisdebate, I will highlight the benefits of employing an approach that beginsfrom a reconstructed ideal of equality, understood not as abstractsameness but as a feminist principle of ‘‘nondomination’’ that beginsfrom the concrete experiences of women. More specifically, I will claimthat important feminist arguments against pornography cannot be fullyunderstood if one begins theorizing by attempting simply to ‘‘abstract’’without employing any specific ideals.

According to O’Neill, the basic principles of political and moral theorymust be ones to which all agents can agree; as I noted above, she statesthese principles must be ones that ‘‘leave the agency of each member ofthe plurality intact’’ (1988, 718). Applying this criterion to questionsabout the acceptability of pornography, one might find pornographyacceptable if it does not ‘‘violate the agency’’ of either the persons who areused to make it or the persons who view it, sell it, and produce it. In fact,some who argue in favor of pornography contend that it is unproblematicinsofar as women consent to work in the pornography industry andinsofar as some women actually choose to purchase and use pornogra-phy. In other words, this argument claims that although not all womenlike pornography, those who find it offensive do not have to watch orread it, nor do they have to work as models in the industry. Thus, it seemsthat O’Neill’s method could be used to argue that so long as no one is‘‘forced’’ to model for or watch pornography, then its production and saleare unproblematic.

Beginning with O’Neill’s method of abstraction does not necessarilylead one to the conclusion that pornography is unobjectionable. Onecould argue that it actually does affect people who do not explicitly‘‘consent’’ to it. Many pornography models work in the industry becauseof the lack of other well-paying jobs, and some are there due to theforceful coercion of others. Nonetheless, unless the women working in thesex industry were actually forced to take their jobs, assertions about their‘‘lack of other options’’ fail to hold much weight in a society like ours,where well-paying jobs are scarce for those without higher education andvarious forms of privilege. The other argument, that women are subjectedto pornographic images without ‘‘consenting’’ to view them, is also notlikely to work as a powerful argument against pornography. People are

13 For anthologies that include a variety of feminist arguments against pornography, aswell as numerous other feminist perspectives on this issue, see Itzin 1992 and Cornell 2000. Inparticular, see the essays by Catharine MacKinnon and Andrea Dworkin, as well asMacKinnon 1993 and Dworkin 1979.



generally not forced to watch pornographic videos, nor are they forced toread pornographic magazines. Although it is true that sexualized violenceis common in our society, focusing on the issue of consentFas O’Neill’smethod seems to advocateFdoes not facilitate the development of ananalysis of what is wrong with pornography in particular, since much ofwhat people read and view is sexist in some intangible way. We clearly donot specifically ‘‘consent’’ to the existence of many types of culturalmedia, and suggesting that everyone must consent to all forms of culturalrepresentation would be an impossible and unworkable standard. Thus,claims that certain forms of pornography contribute to women’s oppres-sion must be based on something more than the claim that not all womenconsent to the existence of various forms of pornography in our culture.

In contrast to the idea that we must begin from very abstract principlesand ask whether the members of a plurality could each consent to theprinciples in a way that would leave their agency intact, feministarguments against pornography focus on questions of women’s subordi-nation and inequality. Feminism begins not from principles that areentirely abstract and neutral but from the idea that women are subject tovarious forms of sexist oppression and from the commitment that suchoppression must be challenged and changed.14 Guided by a feminist idealof equality, whereby equality is understood to require that no one beoppressed on the basis of membership in a social group, the feministquestion becomes: Do particular social institutions and practices (such aspornography) reinforce, perpetuate, or contribute to women’s inequality?It is from this particular perspective, and from this ideal of equality, thatfeminist critiques of pornography can be seen to emanate.

Thus, rather than asking whether pornography is something to whichindividuals each give their consent, feminists examine the effects thatpornography has on the larger society and on the status of subordinatedsocial groups. Beginning with a concern to end violence against women,and to stop sexual harassment and discrimination, feminists are interestedin questions about whether and how certain forms of pornographycontribute to these problems. As Gail Dines, Robert Jenson, and AnnRusso explain in the introduction to their book Pornography: TheProduction and Consumption of Inequality, ‘‘The feminist analysis of thepornography industry begins with the problem of pervasive violenceagainst women. . . . The compelling issues and questions have to do withmistreatment, harassment, rape, battery, and murder, not sex per se. Theimaginative feminist vision is a society free of inequality and violence,

14 Here I follow Marilyn Frye in defining oppression in the following manner: One isoppressed based on one’s membership in a group (not simply as an individual). An act isoppressive because it functions as part of a system of barriers and must be examined in thecontext of other acts, not as an isolated event. Third, the limitations created and perpetuatedby oppression work to benefit other social groups (see Frye 1983, 1–16).



not a society free of sexual desire and expression’’ (1998, 11–12). Thus,while nonfeminist analyses of pornography tend to focus on questionsof whether or not pornography is ‘‘offensive’’ to persons who mightview it, feminists who oppose pornography highlight questions ofequality and examine the effects of pornography in the context of maledominance.

Unlike liberal theorists who ask questions about individual consentand employ principles that aim to be ‘‘merely abstract,’’ feminist critics ofpornography are concerned to remedy specific forms of social inequal-ityFsexism, as well as racism, classism, heterosexism, and other types ofoppression that are also often manifested in pornography.15 In this sense,feminism offers an analysis that is not ‘‘neutral’’ or ‘‘abstract’’ but insteadaims for a particular social ideal, women’s substantive equality. Achiev-ing this ideal would require somehow reducing various forms of culturaloppression, since some forms of culture (such as sexist pornography andadvertising) promote and encourage practices that are oppressive. Whilethis need not require the banning or censorship of pornography, it doessuggest that people ought to develop some means of reducing theproduction and distribution of such materials and that people shouldthink critically about the effects of various media imagesFand about theconditions of their production and consumption. Such questions are morelikely to be considered when one employs a specific feminist ideal thatunderstands women’s equality to require the construction of a society inwhich women are not oppressed. Approaching questions about theacceptability of pornography by beginning with principles that are‘‘merely abstract’’ is less likely to lead to this sort of contextualizedanalysis.

On the one hand, a defender of O’Neill’s methodology might arguethat it is difficult to apply O’Neill’s method of abstraction withoutidealization to an issue like pornography, since O’Neill does not offerspecific guidance about how her abstract principles ought to be inter-preted and applied in any particular context. Thus, it is possible that onemight be able to use O’Neill’s method of abstraction to devise abstractprinciples that could then be applied to the issue of pornography to yieldthe conclusion that some forms of pornography are oppressive. I donot mean to deny this possibility; rather, my argument suggests thatO’Neill’s method of abstracting from questions about social powerand oppression makes this conclusion less likely. Feminist analyses thattake women’s social inequality as a starting point and employ a visionof a world in which women are not dominated by men are more likely toyield substantive and critical conclusions about how society needs tochange.

15 For an analysis of the ways in which pornography reinforces racism, see Zia 1994.



7. Conclusion

Unlike the allegedly abstract ideals of liberalism, the substantive idealsthat I have discussed do not attempt to abstract entirely from allcontextual values or from all knowledge about the social structure. Inthis sense, they can be said to be more concrete than the abstract ideals ofliberal theory, which purport to be value free and to have arisen fromabstraction alone. As O’Neill points out, however, idealizations oftenappear in places where the theorist claims to be avoiding them: manyallegedly abstract theories actually rely on idealizations of some sort.Unfortunately, she herself falls victim to her own criticism, since sheultimately cannot avoid all reliance on idealizations. Rather than arguingthat such avoidance is possible, or even that it is desirable, I havesuggested that the use of idealizations (and ideals) is problematic whenthese are false and when they assume the inevitability of oppressivehierarchies. As Babbitt and Anderson demonstrate, and as my discussionof feminist analyses of pornography illustrates, it is possible to make useof normative ideals in a way that neither reinforces oppression norendorses false idealizations. Moreover, challenging oppressive socialstructures, such as race and gender hierarchy, may actually requireappeals to ideals that explicitly suggest nonhierarchical alternative modelsfor how society might be structured.

Thus, theories that aim for abstraction without idealization may beinadequate to deal with circumstances of hierarchy and oppression. Notonly are some forms of ideals (and some conception of human subjects)necessary for moral and political theory to proceed, but specific under-standings of social and political circumstances may also be required toaddress inadequacies in accepted moral norms and practices. Whileabstraction is not always problematic, it must be employed with morespecific attention to matters of social and political hierarchy, and it mustmake use of ideals that do not collude with structures of oppression.

Department of Philosophy503 South Kedzie HallEast Lansing, MI [email protected]

References

Anderson, Elizabeth. 1993. Value in Ethics and Economics. Cambridge,Mass.: Harvard University Press.

FFF. 1999. ‘‘What Is the Point of Equality?’’ Ethics 109 (January): 287–337.



Babbitt, Susan. 1996. Impossible Dreams: Rationality, Integrity, andMoral Imagination. Boulder, Colo.: Westview Press.

Cornell, Drucilla, ed. 2000. Feminism and Pornography. New York:Oxford University Press.

Dines, Gail, Robert Jensen, and Ann Russo. 1998. Pornography: TheProduction and Consumption of Inequality. New York: Routledge.

Duggan, Lisa, Nan D. Hunter, and Carole S. Vance. 1993. ‘‘FalsePromises: Feminist Anti-Pornography Legislation.’’ New York LawSchool Review 38:133–63.

Dworkin, Andrea. 1979. Pornography: Men Possessing Women.New York: Perigee Books.

Dworkin, Ronald. 1981. ‘‘What Is Equality? Part 2: Equality of Re-sources.’’ Philosophy and Public Affairs 10, no. 4:283–345.

FFF. 1993. ‘‘Women and Pornography: Review of Only Words, byCatharine A. MacKinnon.’’ New York Review of Books 40, no. 17(October 21): 36–42.

Frye, Marilyn. 1983. The Politics of Reality. Freedom, Calif.: CrossingPress.

Hill, Thomas. 1991. Autonomy and Self-Respect. New York: CambridgeUniversity Press.

Itzin, Catherine. 1992. Pornography: Women, Violence and Civil Liberties.New York: Oxford University Press.

MacKinnon, Catharine A. 1993. Only Words. Cambridge, Mass.: Har-vard University Press.

Nussbaum, Martha. 1993. ‘‘Onora O’Neill: Justice, Gender, andInternational Boundaries.’’ In The Quality of Life, edited by MarthaNussbaum and Amartya Sen, 324–35. Oxford: Clarendon Press.

O’Neill, Onora. 1987. ‘‘Abstraction, Idealization, and Ideology inEthics.’’ In Moral Philosophy and Contemporary Problems, edited byJ. D. G. Evans, 55–69. New York: Cambridge University Press.

FFF. 1988. ‘‘Ethical Reasoning and Ideological Pluralism.’’ Ethics 98,(July): 705–22.

FFF. 1989. Constructions of Reason: Explorations of Kant’s PracticalPhilosophy. New York: Cambridge University Press.

FFF. 1993. ‘‘Justice, Gender, and International Boundaries.’’ In TheQuality of Life, edited by Martha Nussbaum and Amartya Sen, 303–23. Oxford: Clarendon Press.

FFF. 1996. Towards Justice and Virtue: A Constructive Account ofPractical Reasoning. New York: Cambridge University Press.

Rawls, John. 1971. A Theory of Justice. Cambridge, Mass.: HarvardUniversity Press.

FFF. 1993. Political Liberalism. New York: Columbia University Press.Sundahl, Debi. 1987. ‘‘Stripper.’’ In Sex Work: Writings by Women in the

Sex Industry, edited by Frederique Delacoste and Priscilla Alexander,175–80. Pittsburgh: Cleis Press.



Wood, Allen. 1992. ‘‘Review of Constructions of Reason: Explorations ofKant’s Practical Philosophy, by Onora O’Neill.’’ Philosophical Review101, no. 3:647–50.

Zia, Helen. 1994. ‘‘Where Race and Gender Meet: Racism, Hate Crimes,and Pornography.’’ In The Price We Pay: The Case Against RacistSpeech, Hate Propaganda, and Pornography, edited by Laura J.Lederer and Richard Delgado, 233–36. New York: Hill and Wang.



Abstraction, Idealization, and Oppression

Documents

Transcript of Abstraction, Idealization, and Oppression