Logic or History? The Political Stakes of Marxist-Feminist Theory

Hannah Hoch, “Industrial Landscape” (1967)

This article is the final installment of a dossier entitled Gender and Capitalism: Debating Cinzia Arruzza’s “Remarks on Gender.”

When I first started writing this series of remarks in Italian (Riflessioni degeneri), subsequently collected into a single piece for the English version, my aim was twofold. The first was to make a complex debate – one that has unfolded over the course of several decades – accessible to a public of activists and people interested in gender, race, and class politics. The second was to contribute toward reopening this crucial debate about how we should conceptualize the structural relationship between gender oppression and capitalism. This is why I am deeply grateful to Oksala, Farris, and Manning for accepting to respond to my piece and for articulating powerful and illuminating critiques, and which have helped me think through these complicated matters more carefully and rigorously. Specifying the relationship between the logical and historical dimensions of capitalism is one of the most controversial problems in Marxist theory, and one about which I am very uncertain. But, as this is the point of contention between Oksala, Farris, Manning and myself, I will address a set of concerns pertaining to this problem which is relevant to the central issues at stake: whether or not we can claim that gender oppression is a necessary feature of capitalism and, if so, at what level of abstraction can we make that claim. While there is an array of further criticisms in their responses to my essay, this issue is the focus of all three. Hence I will spend most of the limited space at my disposal addressing it. In the appendix, I will respond to two of Manning’s misrepresentations of my position. It is likely that these misrepresentations are misunderstandings caused by the ambiguity of some of my initial formulations. However, as they are connected to political issues, it is important to clarify them for the sake of advancing our discussion and marking the real points of dissent.

From altogether different perspectives, Oksala, Farris, and Manning share a similar objection to my critique of Ellen Meiksins Wood’s 1988 New Left Review article, “Capitalism and Human Emancipation.” 1 All three observe that I have conceded to Meiksins Wood both the distinction between the logical and historical dimensions of capitalism, and her claim that gender and racial oppression cannot be shown to be necessary to capitalism in a logical sense. Crucially, all three conclude that these concessions vitiate any compelling basis for distinguishing my account of the unitary theory from what I have called an “indifferent capitalism” approach. Thus, all three authors assume that the failure to show that gender oppression is a logical precondition for capitalism entails that the relationship between capitalism and gender oppression is merely contingent and opportunistic. 2

Manning and Farris conclude that we must demonstrate the logical necessity of gender oppression and of racial oppression for capitalism. Conversely, Oksala reaches the opposite conclusion, that we should be wary of totalities and epistemic certainties and fully endorse historical configurations as contingent, variable and opportunistic combinations of distinct fragments. While their conclusions pull me in opposite directions, this dilemma is based on a shared presupposition: all three seem to presuppose that there is nothing between logical necessity and arbitrary contingency; that one either demonstrates the first or fully endorses the second; and that gender oppression and racial oppression are either logical preconditions for capitalism or that their relationship to it is opportunistic, highly variable and, ultimately, dispensable. In other words, all three accept the presupposition of a theory of “indifferent capitalism” endorsed by Meiksins Woods and other Marxists. It is precisely this presupposition that I reject and that I tried to question in my piece.

Farris’s and Manning’s preoccupation with proving the logical necessity of gender oppression for capitalism arises from an understandable political concern, i.e., the worry that failing to demonstrate this logical dependence would mean that gender oppression is politically secondary to class politics. Unfortunately, this concern is motivated by the widespread (and, in my view, mistaken) tendency among Marxist authors to directly derive political conclusions or theses from theoretical arguments developed at a high level of abstraction. This tendency lies behind the political conclusions Meiksins Wood draws in her own piece. Faced with this tendency, it is perfectly understandable that Farris, Manning, and other Marxist feminist authors have been preoccupied with showing the requisite logical necessity so as to grant women’s struggle the appropriate attention and centrality. A similar preoccupation arises regarding the logical necessity of racial oppression. While I understand and share this concern, I think that we should reject rather than answer this interpellation, given that it rests on a mistaken presupposition: the dynamics of political struggle cannot be directly deduced from theoretical observations at this level of abstraction.

In fact, this tendency has led some Marxists or activists within class politics to elaborate hierarchies of processes of political subject formation, struggles, and practices based on abstract logic – hierarchies within which racial and gender oppression would be secondary due to capitalism’s logical independence from them. This approach, however,  only reveals a bookish understanding of political struggle. It is an approach that neglects the lived experience of exploited and oppressed people, the concrete and conjunctural processes through which they come to acquire political agency and subjectivity, the way they perceive themselves and their conditions of existence, the sedimentation of their past struggles, the actual history of the country they live in or they come from, and a number of other factors that are crucial to the effective construction of political strategies and projects. 3 Once I have decoupled the issue at stake from this underlying political preoccupation, I can address it for what it is: a theoretical and analytical concern. While this concern may or may not have some political consequences, we certainly cannot directly deduce the kind of relevant political consequences that have been drawn in the past by a number of authors and political organizations.

Gender Oppression and Historical Causality

Before I restate my main arguments against “indifferent capitalism” more clearly,  I think some points raised by Manning and Farris need to be clarified. Manning claims that I (like many other Marxist feminist authors) “do not even feel the need to argue against”… the possibility that we can “locate gender or race in the essential logical structure of capitalism.” While my main claim – i.e., that capitalism is not indifferent to gender oppression – does not, in my view, depend on demonstrating that gender oppression is a logical precondition to capitalism, it also does not depend on arguing for the impossibility of such a demonstration. Thus, I do not argue that such a demonstration is impossible because I am actually agnostic regarding its possibility, and because my central claim does not require that I take a position on this issue. Despite my agnosticism in regards to its possibility, it is clear to me that I have not managed to actually demonstrate it myself, and that I find the other attempts so far, (including both Farris’ tentative attempt, and Maya Gonzalez and Jeanne Neton’s effort cited by Manning) unpersuasive.4 While I am open to the possibility of locating gender and race as necessary preconditions of capitalism at the logical level, this position is very difficult to prove. Unlike Manning, I think that the attitude of several Marxist feminist authors is a symptom of a real theoretical difficulty rather than an out-of-hand or prejudicial dismissal.

In light of these concerns, I wish to clarify a second point, raised by Farris. Farris disagrees with my claim that logic and history are separate, and then suggests that they have a dialectical relationship. I suspect that Farris took my claim to suggest that logic and history are separate in the sense that they are unrelated. But, of course, I reject this claim as it entails that logical analyses of capitalism tell us nothing about concrete capitalist societies. Such a position would make our theoretical efforts meaningless and sterile exercises of intelligence. My claim about logic and history is much more banal than the one Farris attributes to me, as I only meant that they are distinct. Although she questions this distinction, the non-identity of logic and history is consistent with Farris’ view that logic and history have a dialectical relationship, given that such a relationship precludes their identity: a “dialectical relationship” can entail opposition, but certainly not unmediated identity and, consequently, holding that logic and history are identical (i.e. denying the distinction between them) is equivalent to denying the dialectical relationship between them.

While I may agree with Farris that there is a dialectical relationship between logic and history, it is still unclear to me what we actually mean by “dialectical.” The quote provided by Farris from Mészáros, according to whom the social structure is unthinkable without its historical dimension and vice versa, simply re-articulates the problem without clarifying it. While I obviously agree with this point, it still leaves open the theoretical task of specifying how we concretely think the two aspects together and the exact relationship between them. As I have already said, I cannot offer a complete solution to this larger problem. However, by addressing the more relevant distinction between logical possibilities and practical possibilities, I will provide some elements towards clarifying the relationship between logic and history.

Since I have clarified these two points – first, that I am not prejudicially dismissing attempts to demonstrate that gender is a logical precondition to capitalism and, second, that I do not deny the relationship between the logical dimension and historical dimension of capitalism – I can now reformulate my main argument against the “indifferent capitalism” thesis. My argument is formulated as an “immanent critique” of Meiksins Wood’s argument: I provisionally accept her presupposition that demonstrating the logical dependence of capitalism on gender oppression is not possible (a point I am myself agnostic about), and I then proceed to show that this premise does not license the conclusions she draws. My argument has two parts: the first methodological part argues that we must distinguish between logical possibilities and historical (or, alternatively, practical) possibilities; the second concerns the distinction between a logical precondition to capitalism and a “necessary” consequence of capitalist accumulation. In articulating my argument again, I will also answer some of Oksala’s objections by restating the necessity to think of capitalist societies as moving totalities; in contrast to expressive totalities in which each part reflects and corresponds to the others, or where each part is “functional” to the whole, this moving totality is a set of social practices, relationships, and institutions which are all subject to the determining constraints and pressures posed by the logic of capitalist accumulation.

As I will expand upon further, such a conceptualization entails refining and distinguishing the different senses in which we use the words “necessity,” “necessary,” as well as “laws.” When speaking of historically situated social practices and phenomena, terms such as “necessity” and “laws” should always be taken in a looser sense than what is generally understood. Even the so-called capitalist “laws of motion” are “tendencies” in that, qua social practices and relations, they do not operate with the necessity pertaining to certain physical laws. In other words, we can speak of “necessity” and “laws” in a weaker sense than our usage of those terms in the natural sciences. They are also weaker than logical necessity and logical laws, and perhaps part of the confusion on this point lies in the conflation between the necessity pertaining to logic (to our logical thinking) and that pertaining to history, social practices, and social relations – which is a different and weaker kind. This applies to both the relationship between gender and racial oppression and capitalism as well as to the unfolding of capitalist accumulation itself. In using the term “necessary consequences,” I am referring to the consequences that will tend to be produced by capitalist accumulation, which should not provoke the theoretical anxieties evidenced by my respondents.

Practical Plausibility and Concrete History

When working with logical categories or with the analysis of capitalism at a high level of abstraction, we should recognize that we are referring to a theoretical model used to describe some aspect of reality. While I do not deny that this theoretical model has undoubtedly been constructed through the formalization of actual, historically-produced constraints, it is still a theoretical model that refers to – while remaining non-identical with – the reality it describes. When thinking at this level of abstraction we operate with specific kinds of possibility, which I would define as either logical or nomological. Logical possibilities denote what is thinkable without contradicting the given definition of the object at stake. If, in my definition of capitalism, the core of capitalist accumulation is the valorization of value (and hence the extraction of surplus-value through exploitation), then I cannot conceive of a capitalist society sans exploitation without logically contradicting my given definition. But I can think any configuration that does not contradict it as being possible.

Put simply, logical possibilities concern the coherence of our thoughts toward objects. As such, the range of logical possibilities is as a general rule both wider and more rigid than that of real possibilities, which have to account for constraints other than logical coherence and thus, to a limited extent, allow for the existence of contradictory processes. A nomological possibility has to do with sets of laws: it circumscribes as possible everything that does not contradict those given laws. Now, coherence in thought does not exhaust the range of constraints that limit the set of possibilities on a practical or historical level. Nomological coherence, in its turn, would be able to express the entire range of constraints that either limit or rank possibilities on a historical or practical level, if and only if we assumed that all practical necessitating constraints can be formalized into laws. I am deeply skeptical that this is the case. This is not to suggest an irrationalist approach to parts of reality; I only mean to suggest that logical formalization is not the sole rational means of grasping reality at our disposal, and that not all necessitating constraints are grasped in this way or formalized at that level. The cognitive mapping of certain constraints demands concrete historical analysis or other heuristic tools.

To make this point clearer, the concrete history of capitalism involves several practical constraints. For example, given the origin and development of capitalism in the United States, I think that we can make a plausible case that racial oppression, in varying and historically specific forms, not only is but will likely remain a constitutive part of American capitalism and American society. It is likely that this practical constraint cannot be formalized in the same way that the extraction of surplus-value can. Nonetheless, it is a necessitating constraint that significantly qualifies possibilities according to degrees of probability. We may, of course, conceive of a version of American capitalism without racism, as such a possibility does not contradict either a given definition of capitalism or its formalized laws of motion. However, this scenario is historically implausible. Practical plausibility is what should really matter for political concerns, because we do not do politics, or at least we should not do it, by means of mere thought experiments whose only rule is  logical coherence. To sum up, the unmediated jump between different forms of possibility is a mistake; we should reject this mistake rather than taking it as the guiding premise of our theoretical efforts.

Precondition or Consequence?

To continue with the second part of this re-articulation of my critique of “indifferent capitalism,” proponents of this approach such as Meiksins Wood fail to distinguish between the status of determinants that are logical preconditions to capitalism, and the status of determinants that are necessary consequences of capitalism. Farris observes that my distinction between these two terms is unclear, arguing that “if we say that capitalism produces oppression by necessity, we are in fact still putting forward an argument that requires explanation at the logical structural level, and not only at the historical level.” Farris is certainly correct in claiming that this necessity cannot be proved only at the historical level without some explanation at the logical structural level. However, as I have hoped to show in the preceding paragraphs, I have not denied this. Farris’s critique is fair given that the formulation and structure of my argument was unclear and potentially misleading. Hence her criticism has, thankfully, allowed me to reformulate my point in more robust terms.

The distinction between a logical precondition and a necessary consequence of a social dynamic is warranted given that a set of social phenomena can be necessarily and constantly produced by the logic of capitalist accumulation without being a logical precondition for it. So, even if we were compelled to concede that gender oppression or racial oppression are not logical preconditions for capitalism, this concession would still not entail the conclusion that the relationship between capitalism and these forms of oppression is only an opportunistic and contingent one, and that capitalism is “indifferent” to them. In order to demonstrate the indifference of capitalism to gender and racial oppression, a proponent of the “indifferent capitalism” approach would have to argue that gender and racial oppression are not “necessarily” produced and reproduced by the logic of capitalist accumulation. In other words, the definition of capitalism as essentially “indifferent” to women’s and racialized people’s oppression – exploiting them in merely opportunistic terms – fails to take into account the fact that capitalism does not just “use” pre-existing oppressions, but also produces them as a byproduct of accumulation. To quote Martha Gimenez on gender inequality:

Gender inequality thus conceptualized, as a structural characteristic of capitalist social formations, is irreducible to microfoundations; i.e., it cannot be solely or primarily explained on the basis of either men’s or women’s intentions, biology, psychosexual development, etc. because it is the structural effect of a complex network of macrolevel processes through which production and reproduction are inextricably connected. This network sets limits to the opportunity structures of propertyless men and women, allocating women primarily to the sphere of domestic/reproductive labor and only secondarily to paid (waged or salaried) labor, thus establishing the objective basis for differences in their relative economic, social and political power. However, analysis of concrete or specific instances of gender inequality within households, enterprises, bureaucracies, etc. is not only amenable to study at the level of microfoundations, but requires this. We cannot fully explain oppressive practices in a given institution without taking into account the agency of the major social actors; these actors’ intentions, attitudes, beliefs, and practices have to be explained in terms of the structural conditions that made them possible. 5

Claiming that gender oppression is a necessary consequence of capitalism entails relating an analysis of the logic of capitalist accumulation to historical considerations concerning the role of gender differentiation in reproductive activities. As it should appear clear by now, there is no real disagreement between Farris and myself on this point: the only real disagreement concerns both her insistence on the necessity to demonstrate that gender and racial oppression are logical preconditions for capitalism, and her attempt to prove this point by means of an analysis of the necessity of the nation-state for capitalism.

To summarize my above response, I challenge the notion that capitalism has a merely opportunistic and contingent relationship to gender oppression on two counts: the first is that the “indifferent capitalism” approach is based on a conflation between different kinds of possibility and fails to take into account forms of practical necessity or constraints that cannot be formalized into laws of motion; the second is that it fails to take into account that sets of social phenomena can be necessary consequences of the logic of capitalist accumulation, even if they are not a logical preconditions for it, and that, moreover, these necessary consequences can be analyzed, at least to some extent, in abstract terms.   

What is Unitary in “Unitary Theory?”

These two points allow me to answer part of Oksala’s objections. Some of her acute objections stem from my ill-advised use of the term “organizing principle” to discuss the link between gender oppression and capitalism. “Organizing principle” is a misleading term for the social dynamics I am describing, as I recognized only too late. I am willing to drop it altogether, as it fails to properly account for my claim that capitalist accumulation is the shared framework of various forms of oppression, and because it may imply functionalist conceptualizations of the dynamics of capital or suggest the existence of a “plan of capital.”

Another difficulty has to do with the term “unitary theory,” which raised some problems from two opposing viewpoints: both for Oksala and for Manning. The fault here is not entirely mine, as I inherited the label from Vogel’s book, Marxism and the Oppression of Women, where, in my interpretation, the term “unitary” was only meant to demarcate her position from dual or triple systems theories. 6 (The term “unitary,” however, may once again be misleading since it can be interpreted as indicating that all sets of social phenomena examined – hence the various forms of oppression at stake – are organized according to a single logic that can be studied with a single theoretical tool or mode of critical analysis. Understood in this way, it clearly raises a serious problem for my own position, as I want to avoid a functionalist account as much as I want to avoid a merely contingent account of the relationship between various forms of oppression and capitalism. Given that dropping this label may be more complicated because of the history of the debate on these issues, I wish to clarify once again what my position on these points is.

I argue that capitalist accumulation produces, or contributes to the production of, varying forms of social hierarchy and oppressions as its necessary consequences. Moreover, I argue that it has a greater consequential and determining power than other forms of social hierarchy, and that it poses necessitating constraints that determine all other forms of social relations. Thus, my claims are more robust than simply stating that in a total social formation something “is connected to something else.” However, my claim is weaker than arguing that capitalist accumulation organizes other social hierarchies according to a single logic. Moreover, it is my contention that the logic of capitalist accumulation is pervasive (that is, that it has the capacity of coloring all other social relationships), which is one of the grounds for speaking of a contradictory and articulated moving totality.

I am aware that this restatement of my position can and probably will be unsatisfying to many. The difficulty lies in the fact that I am trying to carve a different path between the opposite directions in which Oksala, on the one hand, and Manning and Farris, on the other, would like to pull my argument. Making this path compelling necessitates additional work on articulating the notions of historical necessity and possibility, of totality and of determination, as well as reconsiderations of the relationship between logic and history, and of the nature of Marx’s method of exposition. Clearly, being an even more difficult task, such work would require a combination of robust empirical, historical reflection and theoretical elaboration. It is work to be done in the years to come.


In Manning’s piece there are two misrepresentations of some of my arguments that seem theoretically or politically relevant enough to require a clarification. This clarification is not intended as polemical, but as a necessary step to clear the ground from possible misunderstandings and further the debate. Moreover, it is quite possible that I bear some responsibility for these misunderstandings.

The first of these misrepresentations concerns the issue of the benefits to men afforded by gender oppression. My concern with this misrepresentation of my argument is both theoretical and political, because whether we consider men and women as two antagonistic classes obviously has relevant political consequences. Manning writes that “it is important to challenge Arruzza’s statements about the negligible benefits of patriarchy to men”; then, to counter my argument, she goes on making a list of activities and practices that cannot be socialized, including sexual abuse and violence. 7

I am puzzled by Manning’s interpretation of my position, and I must admit that I am frustrated by her implication that I deny that men significantly benefit from women’s oppression. Such a position is basically an antifeminist one that I have always been antipathetic to. My argument was quite simple and concerned the definition of class relationships as based on exploitation, understood as the extraction of surplus. My intent was to show that men and women cannot be considered as two classes in this technical sense due to the absence of extraction and appropriation of a surplus: in this sense, and only in this sense, men would lose nothing from the socialization of care work. Moreover, I have never claimed that gender inequality’s benefits to men are negligible. I have only claimed that they are not an adequate ground for speaking of a class antagonism between men and women or for defining men and women as two classes. This does not make these benefits negligible: occupying a higher hierarchical place in the social order, working less than those who are in a subaltern position, having greater social recognition, and having a privileged access to violence and domination over others are not negligible benefits at all. As a matter of fact they are a powerful device of division and conflict among exploited and oppressed people and within a class. They would not be so powerful, if they were negligible. Still, within the theoretical framework I tried to elucidate, these elements alone do not define what a class relationship is, and cannot be taken as a ground for defining gender oppression as a class antagonism involving two classes - men and women. As my argument concerned the class status of gender inequality, it is unclear to me how Manning’s reminder of sexual abuse and violence, for example, may challenge it, unless what is a stake is the suggestion of a different definition of class and class relationship. If this is the case, then there is indeed a theoretical disagreement between Manning and me.

Finally, Manning’s misinterpretation of my argument may be connected to another substantive theoretical disagreement with my position: Manning claims that there always will be a sphere of unsocialized work, and that women or feminized people will do it for the most part. Based on these premises, Manning concludes that there is as little reason for men to let go of the benefits they draw from women’s oppression as there is for capitalists to socialize their profits. Since the theoretical bases for these observations – whether Manning’s argument is a psychological, anthropological or ontological one, whether it refers to only capitalist societies or is a more general claim – are unclear to me, I will forego further debate on this issue.

The reason why I deemed necessary to clarify this point is the identification of the gender oppression of women by men with class antagonism would entail that mixed-gender politics is a self-undermining political strategy for women insofar as such inter-class cooperation would, ultimately, be detrimental to women. Antagonistic class interests, indeed, cannot be reconciled in the long term. As a matter of fact, the definition of gender oppression as a class antagonism has led, more often than not, to separatism. Perhaps one of the reasons for conceptualizing gender oppression as a class relationship is that we want to consider unpaid reproductive work as exploited work. However, we do not really need to define women as a class in order to achieve this outcome. Workerist feminists, for example, have conceptualized unpaid care work as exploited productive work without conceptualizing the relationship between men and women as a class relationship; others, while denying that reproductive work produces value, have argued that it contributes to both the reproduction of capitalist society and the production and extraction of value, hence it is exploited.

The second point concerns race. I have never written that the category of social reproduction circumscribes the essential racializing processes within capital. As a matter of fact, it does not. We cannot adequately analyze race without addressing numerous determinants including the international division of labor, colonialism, imperialism, migration, and combined and uneven development; these determinants cannot be exhaustively conceptualized through the notion of social reproduction as used by Marxist feminists. While there may and should be important overlapping aspects, social reproduction theory cannot adequately address these matters on its own. Consequently, I am skeptical of the fashionable tendency to theoretically conflate discussions on race and gender: while it is true that racial and gender oppression are strongly intertwined at some level of analysis, and while it is often the case that one cannot understand the one without the other, at the level of the theoretical analysis of the consequences of capitalist accumulation and their determination of the reproduction of forms of hierarchy and oppression, there are several factors that play a fundamental role for one form of oppression but not for another.


1 Ellen Meiksins Wood, “Capitalism and Human Emancipation,” New Left Review I/88 (1988), 3-20.
2 This point of criticism, that I don’t have any persuasive theoretical basis for a distinct unitary theory left, seems to me more explicitly articulated in Oksala and Manning, but I suspect that Farris may agree with them given her insistence on the necessity to have a logical demonstration of the fact that gender and racial oppression are preconditions to capitalism.
3 From this viewpoint, Meiksins Wood’s political considerations on gender are quite at odds with the more sophisticated stance she takes on Thompson’s theory of class experience. See Ellen Meiksins Wood, “The Politics of Theory and the Concept of Class: E. P. Thompson and His Critics,” Studies in Political Economy 9 (1982), 45-75.
4 Farris’s attempt draws on Davidson’s theory of the nation-state: this is an extremely complex and controversial issue that I cannot fully address here. I will confine myself to note that Davidson’s explanation of nationalism leans toward a functionalist explanation that raises an array of serious difficulties: see Neil Davidson, “The Necessity of Multiple Nation-States for Capitalism,” Rethinking Marxism 24.1 (2011), 26-46. This is why, while I find Farris’s considerations on intersectionality illuminating, I am more skeptical about her outline concerning the logical necessity of gender and race based on capital’s need for a multiplicity of nation states. As for the other attempt cited, I agree with much of Gonzalez’s and Neton’s piece. However, a key point of their argument is the assumption that “for labour to exist and serve as the measure of value, there must be an exterior to labour” and that “for labour-power to have a value, some of these activities [the ones that reproduce labor-power] have to be cut off or dissociated from the sphere of value production.” Unfortunately, in my view their piece does not really succeed in demonstrating on a logical level that this is the case, hence I do not think that they have demonstrated that gender is logical precondition to capitalism. I will write about this point more on a different occasion. See Maya Gonzalez and Jeanne Neton, “The Logic of Gender,” Endnotes, Vol. 3 (2013).
5 Martha Gimenez, “Capitalism and the Oppression of Women: Marx Revisited,” Science & Society, 69.1 (2005), pp. 11-32, 24.
6 Lise Vogel, Marxism and the Oppression of Women (Chicago: Haymarket, 2014 [1983]).
7 This is the contentious sentence: “A man would lose nothing, in terms of workload, if the distribution of care work were completely socialized instead of being performed by his wife.”

Author of the article

is a member of the editorial collective at Viewpoint Magazine and an Associate Professor of Philosophy at the New School for Social Research in New York and a feminist and socialist activist. She is the author of the author of Dangerous Liaisons: The Marriages and Divorces of Marxism and Feminism.