Sally Haslanger and Metaphysical Realism

Sally Haslanger claims to be both social constructivist and metaphysical realist. Prima facie, many think that this is a contradiction. There is no doubt that her theory of social structure is in some sense social constructivist, but in what sense can she be metaphysical realist?

Feminism continues to break boundaries in philosophy. Thought by many as first and foremost a political movement, feminist contributions to debates in epistemology and philosophy of science have been controversial. The same goes for feminist metaphysics. One natural reason for this is the dominant view of realism in these branches of philosophy. How can a view that seems to adopt a clear biased standpoint add value to inquiry into what is objectively true, independent of the subjective standpoint of a political movement? Or isn’t this a task metaphysics legitimately can claim to have? Much of feminist contributions in theoretical philosophy have indeed been to cast doubt on the possibility of legitimate authority for philosophers to say what Truth, Reality or Knowledge is. Or even stronger, to hold that Truth, Reality or Knowledge isn’t something that exists independent of human beings at all. That is why feminists in metaphysics automatically get labelled as anti-realists. Indeed, very often, feminists are correctly described as anti-realists, because they are social constructivists. However, recent work by Sally Haslanger claims to be both social constructivist and realist (Haslanger 2012, pp. 198-200).

In what follows, I will try to make sense of how these views might converge, without giving any endorsements. I will first give some reasons as for why thinking that social constructivism and metaphysical realism prima facie seems odd. I then quickly describe Haslanger’s view of social structures, before suggesting that the initial strangeness of her position could be due to meta-metaphysical reasons. Finally, I discuss the meta-metaphysical controversy that I believe underpins the confusion.

Before we look at how Haslanger can be a metaphysical realist, we need to take a closer look at what this view amounts to. One crude and general statement of the view could be that the world is what it is, independent of how humans or other inquiring agents take it to be (Khlentzos 2016). One can see that a motivation for this is that unless this is so, none of our beliefs can be objectively true, since true beliefs are connected to this agent-independent world. It stands in contrast then with metaphysical deflationism. Metaphysical deflationism is the view that there isn’t an objectively privileged description of the world. They deflate our debates that seem to be about our world, to be about how we use our concepts to “carve up” the world (Barnes 2017, p. 3). Metaphysics then consists of debates about our words or concepts, not about the world. Metaphysical realists, on the other hand, believe that debates in metaphysics are about the world, not about our words or concepts. Social constructivists will very often be deflationists because social categories like gender or race are usually taken not to exist independently of human beings. Thus gender and race are concepts we use, and are not about the world as such. They are kinds but not natural kinds – rather they are social kinds. A social constructivist focuses on the fact that these concepts are in a sense made (or constructed) by human beings. Either such a claim is metaphysical or epistemic. If the claim is metaphysical, we construct gender itself. However if the claim is epistemic, the claim is rather that gender is our constructed conception or knowledge of some X in the world (Sveinsdottir 2015, pp. 1-2).

A social-constructivist-metaphysical-realist would then presumably think that social construction isn’t metaphysical but rather epistemic. If feminists would be social constructivists and metaphysical realists about gender, one would have to say that our construction of gender is inextricably connected to something in the world, independent of our minds, something like biology. However, since feminists often reject the biological view of gender, they tend to be metaphysical deflationists, rather than realists.

Sally Haslanger, on the contrary, claims she is both a metaphysical realist and a social constructivist about gender (and race for that matter). Prima facie, this seems to contradict the traditional way we think of metaphysical realism. Because to say that she is a realist and a social constructivist about gender initially seems to put gender ontologically on par with physical objects. How might we construe Haslanger’s view to be both metaphysically realist and social constructivist? To start with, we should mention that Haslanger isn’t really a metaphysical realist about gender per se. Gender and race are social structures, which is what Haslanger considers to be real in this sense. A social structure is for Haslanger that which is universal about being a member of a particular race or gender. Such roles can vary dramatically across time, space, classes, sexualities, cultures and so on. Nonetheless, there are certain hierarchical structural features about them that are stable or constant. We associate specific social roles or attitudes as things that should be had by women, in virtue of perceived features of biological sex. The roles and attitudes vary greatly, but the perceived features of biological sex that justify women occupying a certain social position or other that are subordinate to their male counterpart, stay constant.

These structures are constituted by complex, repeated patterns of interpersonal social interactions. But they are not identical to those patterns (Barnes 2017, p. 8). When we interact with each other in social practices we seem to construct norms, expectations and other social codes, which in turn are sustained over time. This “sustainedness” leaves a rift in our reality, which is the social structure. It is a structure, because it structures our actions and behaviour – we keep falling in line with the rift. Consider the analogy of wheel ruts in the road. They are caused by the repeated patterns of cars etc. driving over the same bit of road over and over. But once the rut is there, it is something different from the individual paths of any particular car, even though its continued existence relies on the continued travel of wagons along the same path. They also explain why cars continue to drive where they do because it’s hard to drive anywhere else on the road than in the ruts (Barnes 2017, p. 8).

But there is one crucial difference between the rut in the road and social structures. The rut seems to have a fundamentally different ontological status than social structures. I can touch and see the rut, and probably even smell the fresh rubber from the wheel of the car going over and sustaining the rut. Surely, this is not something I can do with the social structures – I cannot smell nor see them by any ordinary sense of those terms. Haslanger still wants to say that social structures are equally real nonetheless. They are real in the sense that they are part of our reality. Moreover, since gender and race structures our action and behaviour, they have causal influence on us. In this sense they are real as anything, but they are made. The fact that social structures have causal influence on us makes way for them to be part of causal explanations, which in turn can have strong explanatory value.

As we shall see, I take explanatory value to be important for Haslanger, but to unlock why she calls herself a realist about social structures I think we need to see what motivates her position. It seems to me that what Haslanger is doing is pragmatic metaphysics, rather than to see metaphysics as some kind of “first philosophy”. Her realist view of social structures is part of a larger project of hers called “the ameliorative project”. This project argues that when asking what social kinds like race and gender are, we should consider how an account of race or gender might best help us achieve our legitimate political and social goals, namely feminism (Haslanger 2000). Such an account of race or gender should adequately explain how our human practices works in such ways that it disadvantages non-whites and women. Only when we know how we, through our practices, create the phenomena that give rise to this inequality can we dismantle them. Thus, in choosing theory, we should adopt the assumptions that better explains the evidence for racial and gender subordination. Therefore we use abductive reasoning to infer to the best explanations of the evidence.

Inference to best explanation is something that is commonly accepted in the philosophy of science as ubiquitous to scientific explanation. When inferring to the best explanation “one infers, from the premise that a given hypothesis would provide a “better” explanation for the evidence than would any other hypothesis, to the conclusion that the given hypothesis is true” (Harman 1965). The way I understand Haslanger’s view about structural explanation, it’s a kind of inference to the best explanation.

In modern physics we assume that Einstein’s theories are correct, not because they correspond perfectly with reality, but because they provide us with the best explanation to date. And before that, we could say the same of Newton. We adopt the ontology of those theories – i.e. that wave particles or space-time actually exists – because they provide the best and most adequate explanations of its subject matter. I see Haslanger as pulling a similar move. She wants to postulate the reality of social structures into our ontology, in the same way as we postulate the ontological status of space-time; because it enables us to provide the best explanation given the evidence.

The evidence thus is the evident subordination of women. Especially the fact of the systematic privileging of some and disadvantaging of others based on perceptions of biological sex. The social roles and positions just don’t explain the broader structural feature. In order to properly explain why there are these systematic features of human relations across time, places, cultures etc. we need to be metaphysical realist about social structures. To use philosophy of science jargon, the subordination of women is the explanandum and (unjust) social structures are the explanans. This might sound very Quinean. Quine thought that metaphysics should adopt the ontology of our best scientific theories (Hylton 2014), because presumably, science provides the best explanation of phenomena. However, I do not believe that this is what Haslanger is up to. She rather wants, I think, to adopt the method we use to choose the best scientific theories in metaphysics. Thus, we don’t merely adopt the ontology of the best scientific theories in metaphysics as Quine recommends, but we explicitly use abduction or inference to best explanation as a methodological framework.

Even though we grant Haslanger in approaching metaphysics in this way, there is obviously a big difference between the justification for postulating space-time in physics and social structures in metaphysics. The justificatory gap is evident. For example, theories postulating space-time can be justified demonstrably by the use of instruments. Clearly, theories postulating social structures cannot. It is no secret that successful theories in physics are much more robust than theories in metaphysics, which in turn give their postulations (such as space-time) far higher justificatory status. However, Haslanger don’t need the same accuracy and level of probability that paradigmatic scientific theories enjoy. After all, she is not claiming to adhere to the physicist’s realm, but rather the realm of the metaphysician. The pragmatic use of social construction and metaphysical realism by Haslanger, then, creates tension with the traditional conception of realism perhaps not directly in metaphysics, but rather meta-metaphysically. Should metaphysics be a kind of “first philosophy” that grounds non-fundamentals to use Jonathan Schaffer’s vocabulary, or something that sorts out what kinds of talk are joint carving or not to use Ted Sider’s (Schaffer 2009; Sider 2011). Or can we legitimately claim that metaphysics nonetheless is dependent on pragmatic values?

If it is the former, Haslanger’s view strikes me as contradictory. If metaphysical realism about X is that X is mind-independent, and if X nevertheless is dependent on pragmatic values, it’s not really mind-independent. Therefore, it seems to me that Haslanger’s social-constructivist-metaphysical-realism only makes sense if one meta-metaphysically holds the latter – that pragmatic metaphysics is legitimate. Meta-metaphysics is the study of metaphysics and its foundations (Manley 2009, p. 1). It tries among other things to say what metaphysics is and what it is not. We have seen briefly that Haslanger’s pragmatic metaphysics allow for the social construction of metaphysically mind-independent objects, namely Haslangerian social structures. However, leading metaphysical realists reject this for meta-metaphysical reasons. Let’s take a look at what they think metaphysics is all about before we try to make further sense of pragmatic metaphysics.

Two of the most prominent defenders of metaphysical realism are Theodore Sider and Jonathan Schaffer. Ted Sider thinks that metaphysics is, at bottom, about the fundamental nature of reality. That is, the ultimate goal of metaphysics is insight into what the world is like at the most fundamental level (Sider 2011, p. 1). Sider holds accordingly that there are substantive and nonsubstantive questions in metaphysics. Substantive questions are the questions that are deep, nonconventional and about the world, while nonsubstantive questions are shallow, non-objective, conventional and terminological (Sider 2011, p. 44). Metaphysical deflationists, who think metaphysical questions regard our concepts and not the world, would be nonsubstantive metaphysics on Siders view. He adopts the primitive notion of perfect joint carving and the comparable partial joint carving to ascribe a necessary condition for substantive metaphysics. Only metaphysics that carve up the world at its joints are substantive on this view. To be perfectly joint carving is to be part of reality’s fundamental structure. So it is clear that reality’s joints must be something that is “ready-made” with the world and totally independent of all human activity.

It is clear that Haslanger’s view of social structures is not perfectly joint carving. They do not come ready-made with the world, and they are dependent on human activity. But Sider allows for joint carving to come in degrees – they can be partially joint carving. Things that aren’t perfectly joint carving can nevertheless be partially joint carving. If explanations have relative fundamentality, are lawlike or contain a class of elements to do with causal or explanatory efficacy from the sciences, they can be partially joint carving (Barnes 2014). Relative fundamentality can be seen in light of the “levels hierarchy” picture of science. On this picture physics is most fundamental with chemistry being more or less fundamental and biology somewhat less so (Barnes 2014). Haslangerian social structures seem far from this picture. Social categories also seem far from being lawlike considering that they are socially constructed. Neither do Haslangerian social structures seem to be a kind of scientific explanation in the sense that they involve “thick” normative concepts such as subordination and justice (Barnes 2014).

Jonathan Schaffer has a somewhat different take on metaphysical realism. He wants to revive the former Aristotelian conception of metaphysics. Aristotle’s metaphysics involved also the fundamental. However, instead of dealing with which existing things are fundamentally joint carving or not, Aristotle’s metaphysics was about the primary substances which provide the ground of being. On this view, metaphysics is about what grounds what (Schaffer 2009, p. 347). For Schaffer, we should be less concerned about questions of what exists. Rather, we should be worried about how it exists. That is, whether it is grounded and what, if anything, it is grounded by. That leaves questions of what exists to be merely trivial. Questions of whether properties, meanings, and numbers exist are taken to be obvious. As Schaffer writes, “Of course they do! The question is whether or not they are fundamental” (2009, p.357).

Presumably, Schaffer will have no problem with affirming the existence of genders for example, just as he happily grant the existence of numbers. They exist in our language for example and we can say that they exist in our theories. But it does not follow that their affirmation is therefore the subject of metaphysics. It is simply obvious that chairs and tables exist. These questions are shallow for Schaffer because the existence of everything except the fundamental is what he calls “an ontological free lunch” (Schaffer 2009, p. 361). The deep and substantive questions in metaphysics are rather grounding questions. These are questions about what is fundamental and how the non-fundamental connects to the fundamental. Barnes (2014, p. 7) sums up Schaffer’s point: “What really matters to metaphysics – what does the ultimate explanatory work, what we are primarily interested in qua metaphysicians, etc. – is the fundamental (the basic substances)”. He has no problem with the existence of gender because they have no theoretical cost; they are shallow and cheap. However, the notion of a Haslangerian social structure seems obscure on his view. They are not fundamental, but are neither fully explained by the fundamental (Barnes 2014, p. 8).

Now we can see more clearly, I think, why Haslanger’s approach to metaphysics via feminism breaks with metaphysical realism. Sider and Schaffer have very different views on what metaphysical realism is. But they have one striking thing in common: both think that metaphysics is the study of the most fundamental. They want metaphysics to be what Aristotle baptized it: First Philosophy. They want it to truly live up to its name and go beyond physics and provide explanations of the fundamental nature of reality.

Haslanger wants to say that she is a metaphysical realist, but if I understand her correctly, she is by no means a realist in the meta-metaphysical sense like Sider and Schaffer are. She does not share the same picture of metaphysics as first philosophy that Sider and Schaffer presumably do. Her pragmatic approach to metaphysics cannot be reconciled with a picture of metaphysics as first philosophy, because explaining the fundamental nature of reality cannot be dependent on human interests. Haslanger can still afford to say that there is such a thing called “The Fundamental Nature of Reality”, but her pragmatic metaphysics, dependent on our values as they are, cannot grant that we have access to it.

To Haslanger’s credit, it seems implausible that a subjective human being – the metaphysician – could fundamentally break out of her subjectivity to observe and describe an objective reality – the world ready-made. (The “Gods-eye view”, to use Putnam’s terms.) Even if she could, how would we who don’t hold such powers know that she had done it? The fallibilistic inclinations implied in pragmatic metaphysics seem reasonable. However, is it harmful to pursue speculation about the “Fundamental Nature of Reality”? Why cannot Sider and Schaffer do substantive first philosophy? Couldn’t Haslanger just call her enterprise something else and reserve metaphysics for people like Sider and Schaffer? After all, traditionally, this had mostly been the task of metaphysics. What is the purpose of metaphysics, if we cannot access the Fundamental?

I believe Haslanger wants to say that even though we cannot access the Fundamental Nature of Reality, we cannot really escape the fundamental-talk either. That, I think, is the crucial insight of pragmatic metaphysics; fundamental-talk is inevitable. We cannot help but to categorize the reality we confront as human beings. Unlike the “Gods-eye view”, pragmatic metaphysics emphasize the point that it is only within human practices that reality is, for us, in one way or another. Furthermore, it is only within such practices that entities or objects can be identified and re-identified. Such practices, perhaps analogous to Wittgensteinian “language-games” or “forms of life”, very often carry with them ontological commitments. That is not to say that they construct the World. But that is to say that play the crucial role of constituting the world of experiencable objects and functions as a necessary background for the emergence of meaningfully experiencable objective reality (Pihlström 2007, p. 11-12). Metaphysics then becomes the inquiry into categorizations of reality laden with human practices and thus ultimately human interests and values.

Haslanger’s starting point is exactly human practices. Her social constructivism exploits the fact that such practices create stable patterns of beliefs, behaviour and human action. And if they are created, we can also destroy them when made aware that there is something wrong with them. Haslanger’s approach to ontology thus becomes to categorize the world, not discovering categories that are already there independently of human categorization. However, starting with human practices, there are lots of different ways one can carve up this practice-laden reality. Why think that one of them is objectively privileged over any other? This kind of approach to metaphysics seems fairly deflationary at first. Here we have again reached the confusion we started with. Haslanger view on social structures is not merely that they are concepts we use to carve up the world. She seems to be assuming that there is a way the social world is, and that her theory of social structures as social kinds are trying to adequately describe the way the social world is (Barnes 2014, p.4 my emphasis). She claims some sort of authority that is not on par with the deflationist picture.

What explains this uniqueness of Haslanger’s position? As mentioned above, I think we can understand her innovation as some kind of abductive reasoning approach to metaphysics, which we usually find in epistemology or philosophy of science. This seems like a novel methodological approach for metaphysics that separates her from metaphysical deflationists who only think that we can provide different but equally authoritative metaphysical explanations. Remember that Haslanger’s project is to adequately explain the universal nature of the subordination of women because we need this explanation to ameliorate the situation. If she were a metaphysical deflationist, there is no telling what possible explanation is the correct one. In that case, she would recommend us to pick the explanation that best suits our social and political goals. That would be to carve the world up in the way that best suits the aim of justice. However, Haslanger rather wants to pick the carving that best suits the ordinary patterns of our human social practices. We need to identify what gender is independently of our social and political goals. If not, how could we ever apply such a project like feminism in the first place?

Against this background I think there is some room for a realism of a sort. That is, in evaluating our theories we need to consider whether they are explanatory adequate or not. We need to infer to the best explanation from the evidence (that women are universally subordinated in existing human practices). If our theory of gender is supposed to be successful, it needs to explain the facts about social disadvantage and oppression. As Barnes (2014, p. 7) notes, these are “…fairly standard methods of theory-choice which are perfectly compatible with non-deflationist realism”.

Haslanger’s feminist metaphysics-approach can thus be understood as a kind of social-constructivism-metaphysical-realism after we have seen that metaphysics too can be a pragmatic enterprise. However, it seems like the same goes for a deflationist approach for feminist metaphysicians. So what makes Haslanger’s approach non-deflationist, I think, is her insistence that feminist metaphysics need to successfully explain what gender is independently of the social and political goals of feminism.


Barnes, E. (2014). «Going Beyond the Fundamental: Feminism in Contemporary Metaphysics.» In Proceedings of the Aristotelian Society 114:3, pp. 335-351.
Barnes, E. (2017). «Realism and Social Structures.» In Philosophical Studies 174:10, pp. 2417-2433.
Haslanger, S. (2000). «Gender and Race: (What) Are They? (What) Do We Want Them To Be?» In Nous 34:1, pp. 31-55.
Haslanger, S. (2012). Resisting Reality. Oxford: Oxford University Press.
Herman, G. (1965). «The Inference To Best Explanation.» In Philosophical Review 74:1, pp. 88-95.
Hylton, P. (2014). «Willard van Orman Quine.» Standford Encyclopedia of Philosophy.
Khlentzos, D. (2016). «Challenges to Metaphysical Realism.» Standfor Encyclopedia of Philosophy.
Manley, D. (2009). «Introduction: A Guided Tour of Metametaphysics.» In D. Chalmers, D. Manley, & R. Wasserman, Metametaphysics: New Essays on The Foundations of Ontology (pp. 1-37). New York: Oxford University Press.
Pihlström, S. (2007). «Metaphysics with a Human Face: William James and the Prospects of Pragmatist Metaphysics.» In William James Studies Vol. 2. URL:
Schaffer, J. (2009). «On What Grounds What.» In D. Chalmers, R. Wasserman, & D. Manley, Metametaphysics: New Essays on The Foundations of Ontology (pp. 347-383). New York: Oxford University Press.
Sider, T. (2011). Writing the Book of the World. Oxford: Oxford University Press.
Sveinsdottir, A. (2015). «Social Construction.» In Philosophical Compass 10:5, pp. 884-892.