1. The Antirealist Intuition
In his essay ‘England Your England’, written during the Second World War, George Orwell asks, ‘Are there really such things as nations?’ (Reference Orwell, Orwell and Angus2004: 57) The Norwegian adventurer Thor Heyerdahl questions the reality of borders: ‘Borders? I have never seen one. But I have heard they exist in the minds of some people’ (Stein Reference Stein2015: 33). And several years ago, The New Yorker ran a cartoon with the caption, ‘Sure, money may be imaginary—but at least it's got everybody imagining it’ (Hamilton Reference Hamilton2013). These quotations convey a widespread sentiment to the effect that social entities such as nation-states, political borders, and pieces of money are just figments of our collective imagination and not really ‘out there’ in the world. Let us call spatially located, particular social entities such as nation-states, borders, and dollars social objects.Footnote 1 And let us call the sentiment in question the antirealist intuition.
Some positions on the metaphysics of social objects can easily account for the antirealist intuition. In particular, the eliminativist who says that social objects do not exist can maintain that this intuition is getting at the fact that there are no social objects. And the reductive materialist who says that every token social object is identical to a token ordinary material object can maintain that the intuition is getting at the fact that there are no social objects ‘over and above’ the ordinary material objects that we already recognize. However, both eliminativism and reductive materialism face serious difficulties. Eliminativism conflicts with apparent truths (cf. Fine [Reference Fine2001: 2] and Schaffer [Reference Schaffer, Chalmers, Manley and Wasserman2009: §2.1]; cf. also Korman [Reference Korman2015: ch. 4] and Thomasson [Reference Thomasson2019: §1]). For instance, it is a historical fact that the United States of America was founded in 1776 and not dissolved thereafter. Given that to found something is to bring it into existence, it follows that nation-states exist. Likewise, it is a geopolitical fact that there is a border between Vermont and New Hampshire. It follows from this that there are borders. Reductive materialism, on the other hand, is plagued by the problem from Leibniz's Law (cf. Hindriks [Reference Hindriks2013: 419] on organizations and Korman [Reference Korman2019: §2.1] on stores). Consider, for example, the dollar bill in my wallet. The reductive materialist says that this dollar bill is identical to the coincident piece of paper (cf. Searle Reference Searle1995: 57). Yet the dollar bill apparently does not share all properties with the piece of paper. For instance, the piece of paper could have existed even if the United States had lost the Revolutionary War and never introduced the currency of the US dollar, whereas the dollar bill could not have existed under these circumstances. So by Leibniz's Law, which says that a and b are identical only if they share all properties, the dollar bill is not identical to the piece of paper. Or consider river borders, such as the border between Vermont and New Hampshire. The reductive materialist might identify the Vermont-New Hampshire border with the Connecticut River. Yet the Vermont-New Hampshire border could continue to exist even if the Connecticut River dried up. So by Leibniz's Law, the border is not identical to the river.
Other positions on the metaphysics of social objects may be able to explain away the antirealist intuition. In particular, the immaterialist who says that social objects are immaterial or abstract may maintain that the intuition is based on a mistaken association between concreteness and reality.Footnote 2 That is, the immaterialist may contend that we correctly perceive that social objects are not concrete, but we mistakenly associate the real with the concrete. This leads us to mistakenly intuit that social objects are somehow unreal. But immaterialism also faces a serious challenge—namely, accounting for the spatial properties of social objects. Thus, for example, the United States is located in the Western Hemisphere, the border between Vermont and New Hampshire lies to the east of New York, and a particular dollar bill is in my wallet. But if these social objects are immaterial or abstract, it is hard to see how they could have such spatial properties.
If eliminativism, reductive materialism, and immaterialism are rejected, a natural alternative position is nonreductive materialism. The nonreductive materialist maintains that social objects exist (contra eliminativism), that token social objects are not identical to token ordinary material objects (contra reductive materialism), and that social objects are nevertheless material entities (contra immaterialism). Nonreductive materialism avoids the problems raised for the preceding views. It does not conflict with apparent truths in the way that eliminativism does, as it countenances the existence of social objects. It avoids the problem from Leibniz's Law that plagues reductive materialism, since it does not identify social objects with ordinary material objects. And it has an easier time accounting for the spatial properties of social objects than does immaterialism, as it upholds the materiality of social objects. Nonreductive materialism is thus a promising alternative.
But how do we make sense of the antirealist intuition, as nonreductive materialists? This is far from clear. We cannot make sense of it by pointing to the fact that social objects are nonfundamental. For ordinary material objects such as trees, rocks, tables, and chairs are also nonfundamental. Yet, the antirealist intuition does not encompass these objects—there is no widespread sentiment that trees, rocks, tables, and chairs are just figments of our imagination and not really ‘out there’ in the world. Nor can we fully and adequately make sense of the intuition by pointing to the fact that social objects are less fundamental than ordinary material objects. For the intuition has something to do with the mind-dependent character of social objects, whereas relative fundamentality need not have anything to do with mind-dependence—the facts of physics are more fundamental than the facts of chemistry, but neither set of facts is mind-dependent. Finally, we cannot make sense of the intuition by merely pointing to the mind-dependence of social objects. For ordinary material artifacts such as tables and chairs are also mind-dependent insofar as they are essentially the products of human intentional activity (see Baker [Reference Baker2004: 102–103], Korman [Reference Korman2015: §8.4.1], and Evnine [Reference Evnine2016: 69–70]; cf. Hilpinen [Reference Hilpinen1993: §2–3] and Thomasson [Reference Thomasson2003b: 581 n. 2, 592]). Yet, the antirealist intuition does not encompass such artifacts.
My aim in the rest of this paper is to show how nonreductive materialists can make sense of the antirealist intuition. I do this by developing a version of nonreductive materialism that has the resources to meet this explanatory demand. The central idea is that social objects are materially constituted objects that are response-dependent in a certain sense (section 2). I go on to offer an independent argument in favor of this response-dependent view of social objects (section 3). I then suggest that a proponent of this view can appeal to the response-dependent nature of social objects to explain, or explain away, the antirealist intuition (section 4).
2. The Response-Dependent View
There are two basic ideas behind my proposed version of nonreductive materialism. The first is that social objects are materially constituted, much like ordinary material objects such as rocks, rivers, tables, and chairs. The second is that social objects, unlike these ordinary material objects, are response-dependent in a certain sense.Footnote 3 Roughly, they exist partly in virtue of our attitudes towards them (cf. Hayek [Reference Hayek1943: 2–3], Searle [Reference Searle1995: ch. 2], Tuomela [Reference Tuomela2002: ch. 5], Thomasson [Reference Thomasson2003a: §2], Hindriks [Reference Hindriks2006: §3], Khalidi [Reference Khalidi2015: §2], and Ásta [2018: ch. 1] on the attitude-dependence of various other social items including institutional facts, social kinds, social properties, and social concepts). In this section I develop this response-dependent view of social objects more rigorously using the toolkit of analytic metaphysics and ideas from the social ontology literature.
2.1 Preliminary statement
As a first pass, let us say that on the response-dependent view, any social object z satisfies the following two conditions by its very nature:
(MC) At any time at which z exists: (∃x)(x constitutes z & x is material).
(RD) (i) For any y: y constitutes z if and only if y is taken to constitute z; and (ii) for any y such that y constitutes z: [y constitutes z] holds in virtue of [y is taken to constitute z].Footnote 4
The material constitution condition (MC) says that whenever z exists, there is something material that constitutes it. The response-dependence condition (RD) says that something constitutes z just in case it is taken to constitute z (first clause), and that anything that constitutes z does so in virtue of being taken to constitute z (second clause).
A few remarks are in order. As I understand the relation of constitution in (MC) and (RD), it is a relation distinct from identity (for a defense of the view that constitution is not identity, see, e.g., Johnston [Reference Johnston1992], Baker [Reference Baker1997], and Fine [Reference Fine2003]). Specifically, I take constitution to be an irreflexive and asymmetric dependence relation that holds between an object and either another object (i.e., a thing) or some stuff. A statue and the clay from which it is made provide a paradigm example of constitution, with the clay constituting the statue. Note that an object may be constituted by different objects or stuff at different times, for instance, a statue may be constituted by different clay at different times.Footnote 5
I construe the in virtue of relation in the second clause of (RD) as the relation of metaphysical grounding. Metaphysical grounding is a distinctively metaphysical, noncausal determination relation (Fine Reference Fine2001: §5; Schaffer Reference Schaffer, Chalmers, Manley and Wasserman2009: §3.2; Rosen Reference Rosen, Hale and Hoffmann2010). A nice illustrative example of grounding is provided by the Euthyphro dilemma. Euthyphro and Socrates, we may suppose, agree that all and only pious acts are loved by the gods. But whereas Euthyphro maintains that pious acts are pious because they are loved by the gods, Socrates demurs. This because relation is grounding. On the Euthyphronic view, the piety of an act is grounded in that act's being loved by the gods, whereas on the Socratic view that is not the case. The most important features of grounding, for our purposes, are that the grounds metaphysically determine and explain the grounded, and the grounded metaphysically depends upon its grounds. Grounding theorists commonly distinguish full and partial grounding. In the case of full grounding, the grounds fully determine and explain the grounded, whereas in the case of partial grounding, the grounds only partially determine and explain the grounded. The relevant notion for (RD) is that of full grounding.
Finally, in saying that any social object z satisfies (MC) and (RD) by its very nature, I am making an essentialist claim (cf. Fine Reference Fine1994). That is, I am claiming that any social object z essentially satisfies these two conditions as opposed to accidentally satisfying them, just as a triangle essentially has three sides as opposed to accidentally having three sides.
I will now refine this preliminary statement of the response-dependent view of social objects by examining two issues that arise in connection with (RD): who the relevant subjects are and how the response of ‘taking’ something to constitute something else should be understood.
2.2 The subjects
Who are the relevant subjects whose responses partially ground the existence of social objects? A natural proposal is that the relevant subjects are the members of the relevant population or social group. This proposal works well for some cases, especially simple imaginary ones. For instance, suppose that two people on a desert island agree that a certain river is to be the border between their respective territories. In this case, it is plausible to suppose that the river then constitutes the border in virtue of the responses of these two islanders, who are the members of the relevant population. The proposal also works well for some small-scale real-world cases, for example, a case in which two housemates agree that a certain partition is to be the boundary between their respective parts of the living room.
However, the proposal does not fare so well when it comes to most large-scale real-world cases. For example, consider again the border between Vermont and New Hampshire. The official location of this border is the west bank of the Connecticut River at the mean low-water mark. This is the result of an initial agreement between Vermont and the Continental Congress, and a later ruling by the United States Supreme Court. In this case, the relevant subjects are not plausibly taken to be the members of the relevant population, viz. the residents of Vermont and New Hampshire. They are more plausibly taken to be the relevant practical authorities, which include the governments of Vermont and New Hampshire as well as the government of the United States. More generally, I propose that in many large-scale cases, the relevant subjects are the relevant practical authorities, who are taken to be authorities on the matter at hand by members of the relevant population (cf. Hindriks [Reference Hindriks2006: 490–91] on authority-involving social concepts and Ásta [2018: §1.4] on institutional properties). In some cases these authorities are individuals, but in many cases they are social entities such as nation-states, governments, or organizations. And in some cases the authorities are legitimate, whereas in other cases they are merely de facto authorities that are obeyed by sufficiently many members of the relevant population. Thus, for example, when the British government partitioned colonial British India into India and Pakistan in 1947, it was arguably acting as a de facto authority but not a legitimate authority.
Still, there are some large-scale cases in which the relevant subjects are not plausibly taken to be practical authorities. For example, consider the boundaries of a neighborhood such as the West Village in New York City. Unlike the border between Vermont and New Hampshire, the boundaries of the West Village are not drawn by government officials, courts, or other authorities. To account for such cases, we may draw on Ásta's idea of social standing. Ásta (2018: 18–20) suggests that some people who do not have authority nevertheless have social standing in virtue of other people's attitudes or sentiments towards them, for instance, in virtue of being feared, respected, or admired. This social standing gives them some of the same powers that authorities have. For example, the school bully can order another kid to give her his lunch money, much as an authority can order members of the population to perform various acts. So those with social standing can play some of the same roles as authorities. My suggestion is that in the cases of interest to us where the relevant subjects are not practical authorities, they are individuals or entities with social standing. For instance, in the case of neighborhoods, we may take the relevant subjects to include publicly prominent members of the community (e.g., business owners, journalists, and activists) as well as influential figures in the housing market (e.g., real estate developers and realtors).
By taking the relevant subjects to be either practical authorities or individuals/entities with social standing, we can give a plausible account of all the cases of interest to us. The small-scale cases discussed above, in which the relevant subjects are the members of the relevant population, may be regarded as a limiting case in which all the members of the relevant population have social standing. There may also be some cases in which the relevant subjects include both practical authorities and individuals/entities with social standing. Consider, for example, the United States of America. The relevant subjects in this case are plausibly taken to include the fifty states of the United States, which may be regarded as practical authorities, as well as other nation-states with standing in the international community.
2.3 The response
How should we understand the relevant response of ‘taking’ something to constitute something else? A straightforward alternative is to construe this response as a belief: to ‘take’ y to constitute z is to believe that y constitutes z. But this alternative is problematic for two reasons. First, the relevant response apparently has a voluntary character. Thus, for example, my housemate and I can just decide to take the partition to constitute the boundary between our respective parts of the living room. Belief, however, is not voluntary in this way (cf. Williams Reference Williams1973; Hieronymi Reference Hieronymi2006; Setiya Reference Setiya2008). If I look out of the window and see that it is raining, I cannot just decide to believe that it is sunny. Second, under certain circumstances it may be rational for a subject to take y to constitute z while knowing that y does not constitute z. For example, it may be rational for me to take the partition to constitute the boundary while knowing that it does not yet constitute the boundary if I have good reason to believe that my doing so will prompt my housemate to do so as well. But arguably, it is never rational to believe that p while knowing that not p. For these reasons, we should not construe the relevant response as a belief that y constitutes z.
Following others in the social ontology literature (Searle Reference Searle1995: 37–51; Tuomela Reference Tuomela2002: ch. 5; Thomasson Reference Thomasson2003a: §2; Hindriks Reference Hindriks2006: 491), I suggest that the relevant response be construed as acceptance: to ‘take’ y to constitute z is to accept that y constitutes z. And I propose that we understand the relevant notion of acceptance in terms of the notion of commitment. To accept that y constitutes z is to be committed to acting as if y constitutes z. Importantly, the relevant commitment is not made on the basis of a belief that y constitutes z. Rather, it is made on the basis of pragmatic considerations such as considerations of utility, self-interest, or justice and fairness (and on the basis of other beliefs and expectations). Acceptance is thus a pragmatic attitude. As such, it is not subject to the same requirements of epistemic rationality as belief. In particular, it is rational for a subject S to accept that p while knowing that not p, so long as S has good pragmatic reasons for doing so. Furthermore, commitment in the relevant sense is largely, though not entirely, a matter of choice. It is not entirely a matter of choice because there are constraints on what a subject can be committed to. For example, I cannot be committed to acting as if the number 1 is blue. That is not an intelligible commitment. Likewise, the governments of Vermont and New Hampshire cannot be committed to acting as if a particular snowflake constitutes the border between Vermont and New Hampshire. But within certain constraints, subjects can choose to be committed to acting as if y constitutes z in a way that they cannot choose to believe that y constitutes z. Acceptance is thus voluntary in a way that belief is not. The two problems raised above are thereby avoided.
It is worth distinguishing cases in which acceptance of the relevant proposition is immediate from cases in which this acceptance is mediated by acceptance of a more general proposition. For example, in the case of the border between Vermont and New Hampshire, the relevant authorities immediately accept that the west bank of the Connecticut River at the mean low-water mark constitutes the Vermont-New Hampshire border. In contrast, in the case of dollar bills, what the relevant authorities immediately accept is a general proposition of the form ‘(∀y)(Fy → y constitutes a dollar bill)’, where ‘F’ names the conditions laid down for something to constitute a dollar bill (roughly, being a piece of paper of a certain size and shape, bearing a certain inscription, and having a certain origin). In accepting this general proposition, the relevant authorities are committed to acting as if it were true. This commitment entails a commitment to act as if the piece of paper in my wallet (which bears the relevant features) constitutes a dollar bill. The authorities thus accept, in a mediate way, that this piece of paper constitutes a dollar bill. Note that it is irrelevant whether the authorities are aware of the existence of the piece of paper. Just as someone committed to picking up all the trash on the beach can thereby be committed to picking up a plastic bag whose existence they are unaware of, so too the authorities can be committed to acting as if the piece of paper in my wallet constitutes a dollar bill without being aware of the piece of paper's existence.
Let me conclude this section by noting some important ways in which my account of acceptance differs from John Searle's influential account of collective acceptance in The Construction of Social Reality (Reference Searle1995). First, our proposals differ on what is accepted. On the Searlean proposal the object of acceptance is a constitutive rule of the form ‘X counts as Y in context C’ (Searle Reference Searle1995: 27–29, 43–46), whereas on my proposal it is a proposition of the form ‘y constitutes z’ or ‘(∀y)(Fy → y constitutes a K)’. One advantage of my proposal is that it does not rely upon a controversial distinction between regulative and constitutive rules (see Raz [Reference Raz1999: 108–11] for a critical discussion of this distinction). Furthermore, my proposal does not invoke an obscure primitive relation of counting as (see Epstein [Reference Epstein2015: 74–77] for further concerns about Searle's formula ‘X counts as Y in C’). Second, our proposals differ over the subjects of acceptance. On Searle's view, all intentionality exists in the heads of individual human beings (Searle Reference Searle1995: 25–26), and so it is only individual people who can accept the relevant rules. On my proposal, both individuals and social entities can accept the relevant propositions. My proposal has the advantage that it is able to account straightforwardly for the vast array of real-world cases in which the relevant subjects seem to be social entities such as governments, nation-states, or organizations. Finally, our proposals differ over the collective aspect of acceptance. On the Searlean proposal, when a group of individuals collectively accept a constitutive rule, the form of their acceptance is irreducibly plural. That is, the form that each individual's acceptance takes is ‘We accept that X counts as Y in C’, as opposed to ‘I accept that X counts as Y in C’ (Searle Reference Searle1995: 26, 38). Whereas on my proposal, acceptance of the relevant propositions is a collective phenomenon insofar as the practical authorities and individuals/entities with standing who are doing the accepting have their authority or standing partly in virtue of the attitudes and/or behaviors of members of the relevant social group. An advantage of my proposal is that it does not posit a new irreducibly plural form of intentionality.
2.4 Final statement
Here, then, is the final statement of my proposed response-dependent view of social objects. On the response-dependent view, any social object z satisfies the following two conditions by its very nature:
(MC′) At any time at which z exists: (∃x)(x constitutes z & x is a material object or material stuff).Footnote 6
(RD′) (i) For any y: y constitutes z if and only if the relevant authorities or individuals/entities with standing accept that y constitutes z; and (ii) for any y such that y constitutes z: [y constitutes z] is fully grounded in [the relevant authorities or individuals/entities with standing accept that y constitutes z].
3. Creation by Fiat
Thus far I have put forward a response-dependent view of social objects that satisfies the constraints of nonreductive materialism. In this section I provide an argument in favor of adopting this response-dependent view of social objects together with a response-independent view of ordinary material objects such as trees, rocks, tables, and chairs. The argument takes the form of an inference to the best explanation: I argue that the response-dependent view of social objects, coupled with a response-independent view of ordinary material objects, is able to make sense of a striking apparent contrast between these respective objects.
This striking contrast has to do with the creation of social objects, on the one hand, and the creation of ordinary material objects, on the other hand. Under appropriate circumstances, social objects can apparently be created by acts of agreement, decree, declaration, or the like (cf. Searle [Reference Searle1995: 34] on the use of performative utterances in the creation of institutional facts; cf. also Tuomela [Reference Tuomela2002: 123] on the performative character of social reality). For example, two people on a desert island can create a border by agreeing that a certain river is to be the border between their respective territories. The government of a desert island society can decree that seashells of a certain kind henceforth constitute money, thereby creating pieces of money. And the political leadership of a group of people inhabiting a certain territory can declare the independence of a new sovereign state, thereby founding a new nation-state. In contrast, ordinary experience suggests that ordinary material objects cannot likewise be created by acts of agreement, decree, declaration, or the like. For example, a government cannot just decree into existence more trees; it would have to go and plant them. Nor can two children make a sandcastle by agreeing that some heap of sand constitutes a sandcastle; they would have to go and build the sandcastle by arranging some sand into the shape of a castle. Likewise, a carpenter cannot create a table by just declaring that some hunk of wood constitutes a table. But if ordinary material objects cannot be created by fiat, why is it that social objects can be created in this way? There is a striking contrast here that cries out for explanation.
One might try to explain away this striking contrast by arguing that social objects cannot be created by fiat, appearances notwithstanding. In particular, the eliminativist who denies the existence of social objects can argue that there are no social objects created by fiat in our imaginary scenarios because there are no social objects, period. And the reductive materialist who maintains that every token social object is identical to a token ordinary material object can argue that there are no social objects created by fiat in these scenarios because the social objects in these scenarios (the border, the pieces of money, and the nation-state) are identical to ordinary material objects (e.g., the river, the seashells, and the people, respectively), which are evidently not created by fiat. However, if eliminativism and reductive materialism are rejected for the reasons cited earlier, then the striking contrast cannot be explained away in either of these ways.
Still, one might try to explain away the striking contrast by arguing that ordinary material objects can be created by fiat, appearances notwithstanding. For instance, it may be argued that when the International Astronomical Union officially recognized certain constellations in 1928, it thereby created these constellations by fiat. But this is unconvincing. Constellations are just groups of stars. Since these groups of stars existed prior to 1928, the IAU's recognition did not create them. Rather, the IAU's recognition gave a new classification to these existing groups of stars. Another putative example of creation by fiat of ordinary material objects is readymades in art. For instance, it might be thought that Marcel Duchamp created his famous artwork Fountain by fiat. I agree that in this case, Duchamp did create a new object. However, it is far from clear that this object was created by fiat. Arguably, Duchamp created Fountain by repositioning and inscribing a urinal, with the intention of submitting the resulting product to an art exhibit. But then Fountain was created in much the same way that ordinary material artifacts such as tables are created—namely, through the intentional physical manipulation of preexisting material. Moreover, it is an open question whether Fountain should be classified as an ordinary material object as opposed to a social object. I believe that this question should be answered by a philosophical theory of artworks, and I venture that such a theory would countenance the creation of artworks by fiat only insofar as it construes artworks as social objects.
If the striking contrast between social objects and ordinary material objects is not explained away, how is it to be explained? I contend that the response-dependent view of social objects coupled with a response-independent view of ordinary material objects provides a satisfying explanation.
According to the response-dependent view of social objects, social objects satisfy (MC′) and (RD′) by their very nature. It follows that under appropriate circumstances—namely, circumstances in which a suitable constituting material object or material stuff already exists—these objects can be brought into existence by the relevant authorities or individuals/entities with standing accepting the relevant propositions. And an act of agreement, decree, declaration, or the like can constitute (or otherwise generate) the requisite acceptance.
Let me now articulate the response-independent view of ordinary material objects, using the more open-ended preliminary formulation of the material constitution condition and the response-dependence condition. According to the response-independent view of ordinary material objects, any ordinary material object z satisfies (MC) by its very nature. But no ordinary material object z satisfies the first clause of (RD) by its very nature (though some ordinary material objects might satisfy this clause accidentally). Furthermore, no ordinary material object z satisfies the second clause of (RD). Rather, any ordinary material object z satisfies the following response-independence condition by its very nature:
(RI) For any y such that y constitutes z: it is not the case that [y constitutes z] is fully or partially grounded in [y is taken to constitute z].
It follows that even in circumstances in which a suitable constituting material object or material stuff already exists, we cannot bring into existence a new ordinary material object by merely taking the preexisting material object or stuff to constitute a new object. That is why ordinary material objects, unlike social objects, cannot be created by acts of agreement, decree, declaration, or the like.
4. Objectivity and Reality
If social objects are response-dependent, as I have argued, then they lack objectivity insofar as they fail to exist independently of being taken to exist. For on the response-dependent view, the existence of any social object z is partially grounded in the existence of some constituting material object or stuff, and partially grounded in the fact that this material object or stuff is taken to constitute z. But in taking some object or stuff to constitute z, a subject is eo ipso taking z to exist. Thus, social objects do not exist independently of being taken to exist. In contrast, if ordinary material objects are response-independent, as I have argued, then they do not likewise fail to exist independently of being taken to exist.
Nonreductive materialists can appeal to this difference in the objectivity of social objects and ordinary material objects to explain—or explain away—the antirealist intuition. Whether the intuition is to be explained, as opposed to explained away, will depend on the nonreductive materialist's further commitments, in particular her conception of reality. The nonreductive materialist may embrace a conception of reality on which being objective is a necessary condition for being real or fully real. Such a conception of reality, which links reality with objectivity, is a familiar one (Pettit Reference Pettit1991: 588–90; cf. Johnston Reference Johnston, Haldane and Wright1993: 106). This nonreductive materialist can explain the antirealist intuition by maintaining that it is correctly getting at the fact that the reality of social objects is compromised by their lack of objectivity. Alternatively, the nonreductive materialist may embrace a conception of reality on which being objective is not a necessary condition for being real or fully real. For instance, she may embrace a conception on which to be real is to exist (cf. Haslanger Reference Haslanger2018: 269). Or she may embrace a conception on which to be real is to play an irreducible causal and explanatory role (cf. Barnes Reference Barnes2017: 2432–33). This nonreductive materialist can explain away the antirealist intuition by maintaining that it is based on a mistaken association between objectivity and reality. That is, she may contend that we correctly perceive that social objects fail to be objective, but we mistakenly associate the real with the objective. This leads us to mistakenly intuit that social objects are somehow unreal.
I will not take a stand here on which of these two routes—the realist route or the antirealist route—should be pursued by the nonreductive materialist. That hinges on which conception of reality should be adopted, and that is a deep and vexing issue I cannot hope to resolve here. Suffice it to say that both routes accomplish the task set out in this paper, which was to make sense of the antirealist intuition from within a nonreductive materialist view of social objects.
