Social Inconsistency

1.1. The Nature of Social Facts

The social world is made up of such things as organisations, social groups, artefacts, conventions, practices and laws. It is metaphysically distinctive, because its nature is conditioned by human thought, talk and behaviour, in particular thought, talk and behaviour of a collective sort. Whereas the chemical facts, for example, are indifferent to what we think or expect of them, it seems the social facts wouldn’t be as they are if we didn’t think about them and react to them as we do. To label this pre-theoretic contrast, we might say that the social world is humanly malleable in a way that much of the world is not.

One desideratum for a theory of social reality is that it make sense of this malleability, and part of that task is to distinguish it from other ways in which we shape the world around us. We also change the world by building things in it, landscaping it, polluting it, and so forth. These are profound ways in which we can change things, but they seem metaphysically different. These are causal, rather than constitutive ways of determining the facts.¹ While we have causal influence over many parts of reality, it seems distinctive to social reality that we in addition have constitutive influence over it.² When, for instance, a state passes a law to expand the franchise to the over-15s, 17-year-old Jane is thereby changed into a voter—there is no causal chain that runs between these facts, just a direct relation of metaphysical determination. The state could not, however, change the value of pi to 3 by passing a law to that effect.³ Once we’ve noted this power that we have to constitutively shape the social world, the challenge is to explain how it works, and to do this in a way that respects its limits. Most social facts are not subject to arbitrary stipulation, either by individuals or by collectives. We cannot, for instance, dictate that there is not a recession on; social facts are clearly not so malleable that they are whatever we want them to be. There is an epistemic dimension to this too: our relation to the social facts is not such that we can know them just by inspecting our own attitudes toward them. By and large, we have to find out what the social facts are, often through pretty onerous research. One way to put this is that the social facts seem to enjoy some form of objectivity.⁴ Making sense of this objectivity, whatever precisely it amounts to, is a second desideratum for our theories of social facts, which has to be squared with the malleability desideratum.

This, it turns out, is somewhat tricky to do. Some theories of social facts, particular earlier ones, have sought to make sense of the malleability of social facts by characterizing them as dependent on, or even consisting in, our attitudes.⁵ Hayek, for instance, writes that “in the social sciences the things are what people think they are. Money is money, a word is a word, a cosmetic is a cosmetic, if and because somebody thinks they are” (Hayek 1943: 3). This has the virtue of giving us a very concrete proposal for how we constitutively shape the social facts: for a social fact p, we make it the case that p by believing that p. However, it seems to put us in a bad position to account for the objectivity of social facts: how would we not have automatic epistemic access to social facts, if they are as Hayek describes them here, and how would we not change them by changing our beliefs about them?

So prima facie, a mind-dependence thesis accounts for the malleability of social facts at the expense of their objectivity. But there is room to manoeuvre. Searle (1995), for instance, manages to strike something of a balance between mind-dependence and objectivity by making the mind-dependence indirect: the general conditions that have to be met for social facts to obtain are mind-dependent, but whether things meet those conditions is up to the world, and not up to us. Thus we may have privileged epistemic access to what it takes for someone to be a voter, for example, but no privileged epistemic access to whether Jane is a voter—we’d have to find out Jane’s age first.

Even with such manoeuvres, though, Searle seems badly placed to deal with certain features of the social world—for instance, how it can be possible for there to be a recession on in a society in which no-one has the concept of recession (Thomasson 2003a). For such and other reasons, various authors (Guala 2014; Epstein 2015; Torregno 2017) have argued that we should let go of the idea that social facts are mind-dependent. Even without ruling on that question directly, we can say the following: that it might be best not to tie our understanding of the malleability of the social world directly to mind-dependence. We had best start by approaching the combination of our desiderata at a slightly higher level of abstraction, so that we can decide at a later point in the story how much mind-dependence, if any, we should countenance.⁶ I am going to suggest that theories that do a good job of reconciling malleability and objectivity do so by discerning a two-dimensional structure in social reality. Let me explain what I mean by that.

1.2. Two-Dimensionalism about Social Reality

Suppose that Jimmy wants to go on the roller-coaster, but the sign says that one has to be at least five feet tall to go on. It’s a social fact that Jimmy can’t go on the roller-coaster. If we want to explain that social fact, the most immediate and natural thing to say is that it is because Jimmy is only 4’6’’. That is kind of striking: it seems we can account for a social fact by citing a plain old physical fact. And indeed, relative to that being the rule, Jimmy’s height is what is salient in explaining why he is roller-coaster-ineligible.

Jimmy may nevertheless demand more explanation. Why can’t he go on, just because he’s 4’6’’? Just telling Jimmy that that’s the rule is unhelpful. The rule could have been different—the cut-off point could have been 4 foot instead of 5 foot—so we can ask why it is as it is. The explanation might be, for example, that the person who runs the roller-coaster wrote ‘5 foot’ on the sign, and they have the authority to make the rules. That is a distinct but complementary explanation of Jimmy’s predicament. It tells us how and why the plain old physical fact that Jimmy is 4’6’’ has the bearing that it does on the social fact that Jimmy can’t go on the roller-coaster.

We can typically distinguish two complementary explanations of this sort with respect to social facts. One concerns how the world is, and one concerns conditions that we require of that world. I call this dual explanatory structure a two-dimensional one because it gives social facts a certain sophisticated modal profile. We can imagine the social world to have been different in two separate ways. On the one hand, the plain old physical facts that underpin and determine the social world could have been different. On the other hand, the conditions that we require those physical facts to fulfil could have been different. Jimmy could have gone on the roller-coaster if he’d been taller, and Jimmy could have gone on the roller-coaster if the rule had been different.

This allows for two distinct kinds of counterfactual reasoning about the social world, which we can usefully think of as involving different modal dimensions. We can hold our world fixed and consider what it would be like under the imposition of different conditions, or we could hold our conditions fixed and consider what different worlds would be like under the imposition of those conditions. As an analogy, think of the former as looking at the same planet through a range of different telescopes, and the latter as using the same telescope to look at a range of different planets.

This nicely reconciles the desiderata, albeit only at a certain level of abstraction, with many details left to fill in. One the one hand, this approach does justice to the objectivity of the social world: social facts are explained and determined by underlying physical facts, and are just as objective as them, not mere figments of our collective imagination. On the other hand, it does justice to the malleability of social reality: we human beings—since we’re the ones who put the conditions in place—are crucially involved in determining how the social world comes about.

As it will turn out, this two-dimensional structure is also what allows us to imagine a consistent physical world giving rise to an inconsistent social world. To make the case for that (and, later on, some cases against it) it will help to have somewhat more precise tools to work with—for that purpose I borrow Brian Epstein’s (2015; 2016a) anchoring-grounding model.

1.3. The Anchoring-Grounding Model

Two-dimensional models of social facts have been around for a while, if not always under that description. Searle’s (1995; 2010) theory of institutional facts arguably has this sort of structure, or an approximation of it, under its hood, though Searle himself did not draw out this aspect and sometimes glosses his theory in ways that go against it. The same goes for theories that either build upon or resemble Searle’s theory, such as Hindriks (2009), Thomasson (2003b) or Tuomela (2013). A more explicitly two-dimensional approach was proposed by Einheuser (2003; 2006)—she adapts a Kaplan-style semantics for context-sensitive terms to give a metaphysics of convention-dependent facts, with special attention given to modelling counterfactual reasoning about such facts. But the most developed two-dimensional theory is arguably Brian Epstein’s (2015; 2016a) anchoring-grounding model.⁷ On Epstein’s theory social facts are first of all associated with grounds, in the standard metaphysical sense: facts that underlie and metaphysically determine them.⁸ The grounds associated with a given social fact may themselves be social facts, but by looking at the grounds of those facts and so on and so forth, we eventually get to some non-social facts, because the social world is not metaphysically basic. The facts that can play the grounding role with respect to social facts are not of any specific type: what they look like depends on the kind of fact that’s being grounded. That something is a dollar bill is grounded in how and where it was manufactured (a fact about its history); that someone is popular is grounded in how other people regard them (a fact about attitudes) that Jimmy can’t go on the roller-coaster is grounded in his height (a fact about physical properties).

When we ask why some social facts are as they are, we are typically asking about their grounds, in the first instance. If it’s a fact that Fred is guilty of bicycle theft, that’s explained by his taking a bicycle that wasn’t his. The latter fact grounds the former. But as in Jimmy’s case, there are further questions to be asked. Why is that what it takes to be guilty of bicycle theft? Now we’re asking why the grounding relations run as they do. In Epstein’s theory the grounding principles that govern a given case are explained and determined by different facts which he calls anchors. In this particular case, the anchoring facts will be facts about the legislative process which led to some law governing theft being adopted. Anchors are often social facts themselves, and what they anchor are grounding principles of the general form ‘if A, then A grounds B’ (where B is some social fact). The facts that show up among the grounds and the anchors are not metaphysically different kinds of facts, and one and the same fact could feature in either role on occasion. But the explanatory roles are different.⁹¹⁰ The grounding principles together compose what Epstein calls the frame (hence he also calls them ‘frame principles’). One could think of a frame simply as a set of all the grounding principles that are in force. Another, more model-theoretic and coarse-grained way of thinking of it is as a function from worlds (without social facts) to worlds (with social facts). The space of possible frames and the space of possible worlds-not-yet-endowed-with-social-facts (‘pre-social worlds’, for short) constitute the two dimensions of a two-dimensional model; we can think counterfactually about the social world in one way by holding the frame fixed and varying the world, and in another way by holding the world fixed and varying the frame. We could vary them in concert for a ‘diagonal’ counterfactual.¹¹ For our purposes it is worth saying a little more about the kinds of facts that crop up in the anchoring role. When we have done the work of uncovering the grounding principles that govern some aspect of social reality, and we then ask why those grounding principles are as they are, a variety of facts could come into view.¹²

• Facts concerning speech acts or inscriptions. The fact that a crime of robbery needs to involve a threat of violence is explained (in part) by the fact that a certain form of words defining the act of robbery appears in the law code.

• Facts about the functional roles of social kinds or institutions. The fact that waving your arm can constitute a greeting but thinking happy thoughts cannot is explained (in part) by the fact that the latter is not outwardly perceptible and hence not apt to play the role of a greeting.

• Facts about paradigm instances of a social kind. That dancing the Macarena involves such-and-such moves is explained by the appearance of those very moves in the 1996 music video.

• Stable patterns in behaviour. The fact that deer musk-markings constitute territorial claims is explained by the fact that deer are stably disposed to react to them in a certain way.

What this is meant to get across is that a variety of anchoring facts of different kinds can simultaneously be involved in explaining why the social world works as it does. Unlike what one might want of, say, an explanation of why the laws of chemistry work as they do, it’s quite reasonable to think of social reality as being ‘cobbled together’ by the distributed operation of many such facts. There is not likely to be a neat, unified explanation on offer of the social world as a whole—rather, there is a patchwork of local explanations. And harking back to our earlier discussion of mind-dependence, we can see that within Epstein’s approach, it becomes a more local question whether either the grounds or the anchors relevant to a given social fact are facts of a mental nature—as opposed to Searle (1995), mind-dependence is not an architectural feature of Epstein’s model.

I will rely on Epstein’s theory to impose some clarity and precision on what follows. But it will be possible to recapture the gist of what follows in other approaches that admit of some two-dimensional structure.

4.1. A Simple Argument

One quick way to argue against social dialetheias is to argue against the possibility of dialetheias generally, and then extend that to the social case. For instance, one could (a) argue that one ought to accept classical logic, (b) demonstrate that dialetheias entail everything on such a logic, (c) note that the social world is clearly not trivial, and then (d) modus tollens one’s way to rejecting social dialetheias. If one could conclusively argue that classical logic (or some other explosive logic) is correct, or that dialetheias are for some other reason disastrous, rejecting social dialetheias on this basis would be fair play. It would not be fruitful to enter that general debate here; I point the interested reader to Priest, Beall, and Armour-Garb (2004).

The main thing to note about this argument is that it doesn’t do anything to explain why the social world doesn’t permit dialetheias, or give us any handle on how the mechanisms that generate social facts prevent contradictions from arising. Insofar as the mechanism I’ve sketched above casts some measure of doubt on the impossibility of real contradictions, invoking the impossibility of real contradictions against it only begs the question. For that reason I’ll move on to lines of argument that engage more specifically with the metaphysics of the case.

4.2. Social Anti-Realism

One response to the prima facie case is to take it at face value, but conclude from it that we ought not accord social facts the same kind of serious ontological status as, say, chemical or astronomical facts. From dialetheism about social facts, we infer anti-realism about social facts, the thought being that contradictions are tolerable only if they aren’t real. In this section, I explore this idea. What I will argue is that while some form of anti-realism about the social may be viable, it is not ultimately an effective way to deal with social inconsistency.

Anti-realism comes in varieties, and I will first note one type of social anti-realism so that I can set it aside for present purposes. A line of thought one sometimes encounters is that social facts are mind-dependent, and therefore not real.²³ As we’ve seen in §1.1 above, it is debatable whether we should regard social facts in general as mind-dependent, and as various authors have pointed out in recent years (Khalidi 2016; Mason 2020), the inference from mind-dependence to anti-realism is a questionable one. I will say this: if there is a sense of the term ‘real’ in which one can reasonably conclude that something is not real from the fact that it is mind-dependent, it will be a rather attenuated sense which will not placate those worrying about social inconsistency. Some (e.g., Fine 2001) have proposed reserving the status of reality for the metaphysically fundamental facts, and since what is mind-dependent is thereby not fundamental, this would indeed let us validly infer that mind-dependent social facts are not real. However, this would do no more than put social facts on the same plane of reality as, for example, chemical, biological and astronomical facts—all standardly regarded as not metaphysically fundamental—and to find an inconsistency in the social facts would then be the same, as far as badness goes, as finding one among any of those facts. So I set the kind of anti-realism that is motivated by the (supposed) mind-dependence of social facts aside, and focus on more full-blooded anti-realisms.²⁴

This still leaves us with options; it will help to have a somewhat specific proposal to grapple with, so for concreteness’s sake let us explore a fictionalist proposal. We construe Epstein-style frames as capturing the rules of an elaborate and useful pretence, one that we all participate in every time we engage in social interactions. Organisations, artefacts, laws and the like do not really exist, but it serves our practical purposes to act as if they do, and this pretence is pervasive, in that there aren’t any occasions where we have cause to drop it.

Instead of grounding principles of the form ‘if A, then A grounds B’, we would take the frame to consist in principles of the form ‘if A, then in the fiction, B’.²⁵ If the frame is inconsistency-permitting, then depending on what the non-social world that we take our cues from is like, we may end up with a fiction that is inconsistent.²⁶ But—the thought would go—this is not metaphysically problematic. Fictions, like beliefs, utterances, and theories, can be inconsistent.

Fictionalism in general is a well-explored approach, and the social fictionalist would have resources to draw on in articulating the details of their view.²⁷ There are challenges, though. Thomasson (2003a) mounts an argument against treating social facts as fictional, which turns on the idea that for us to be able to pretend that a thing has some property, the property in question must already be associated with instantiation conditions that we take the thing to not really satisfy. To pretend, for instance, that a tree-stump is a bear is to have in mind what it really takes to be a bear, and pretend that the tree-stump satisfies those conditions.²⁸ Our social practices couldn’t be a pretence of this sort, Thomasson argues. When we treat an object as having some social property, we do not act as if some object satisfies that property’s instantiation conditions when really it doesn’t; though we might have stipulated the relevant instantiation conditions out of thin air, in typical cases the object really does satisfy them, and as a result, really does have the property.

I suspect the social fictionalist could thread the needle here, and perhaps insist that the conditions we get to put in place for social facts aren’t really instantiation conditions for properties at all, but really just conditions for certain pretence-behaviour involving that object to be apt, whatever that might precisely entail psychologically and behaviourally. More generally, they could draw upon the example of other fictionalisms, for instance mathematical fictionalism (Field 1980) in which there is likewise no contrast to be drawn between what fictionally satisfies the instantiation conditions of a mathematical property and what really does.

There are other more general challenges for the fictionalist (see Eklund 2019 for an overview) but I think that at the end of the day, a coherent social fictionalism could probably be articulated. And if one is specifically concerned not to countenance anything inconsistent in reality, it will do that trick. Let us suppose for the sake of argument that this is done; I suspect that the fictionalist would not thereby have achieved quite what they hoped.

Here is the real trouble, as I see it. Let’s say that one’s misgivings about contradictions relate to explosion and triviality. Note, now, that a social fiction in which everything is true is not really any better than a social reality in which everything is true. For even if the social world is a fiction, it is a fiction we continually participate in and which we cannot practically opt out of—the social world, fictional or otherwise, is the means by which we collectively get along together and get things done. If, in the presence of a contradiction, we are committed to a social fiction which is entirely trivial, that fiction would not serve its purpose.

One could deal with this problem by adopting a non-explosive logic, but that option was already available without going fictionalist. And although I won’t argue the details here, I think that other types of general misgivings about contradictions, whatever their merits,²⁹ are also unlikely to be assuaged simply by going fictionalist about social dialetheias, for the same reason: the social world is such that, even it isn’t real, we are condemned to treat it for virtually all intents and purposes as if it is, warts and all. This type of trouble would thus generalize to any form of social anti-realism which, like fictionalism, manages to classify the social as unreal but leaves us to deal with it for practical purposes as if it is real. The motivation for anti-realism falls away: an anti-realist would still have to take some additional, different measures to assuage their worries about social inconsistency, and there is no reason to think that they are in any better place to implement those measures in virtue of having gone anti-realist.³⁰

The alternative that this makes salient is a ‘revolutionary’ anti-realism about the social which tells us not only that the social is unreal but also that we ought to stop treating it as real. This would be an error theory about social reality: there are no social facts, and thus no inconsistent ones. Error theories have been defended about specific social discourses; for instance, eliminativists about race (e.g., Appiah 1995) argue that, as there are no such things as human races, claims about race are erroneous across the board. But as a general approach to social facts, there seems to be very limited mileage in a revolutionary error theory; assuming it could be implemented at all, it would thereby effectively deprive us of the social world altogether, and of all that it does for us. We would cut off our noses to spite our faces.³¹

Thus social anti-realism seems, if not necessarily unviable as a general approach in social ontology, not to be an effective response to the prima facie case for social inconsistency set out above.

4.3. Appeals to Institutional Functions

Theories in social ontology sometimes accord an important role to institutions, as a central type of social entity used to explain various other kinds of social entities and social facts. Let’s take this idea on board for the sake of argument. Institutions typically serve a purpose, and for a given institution to serve its purpose it has to be put together a certain way. Take the aforementioned institution of greeting people. There are many possible conventions for greeting people, but the possibilities are constrained by the function of greeting. Waving your hand or clicking your heels could be a way of greeting. Mentally adding 5 and 7 together is not a possible way of greeting, because it isn’t a public event and therefore simply won’t do the job. Thus there are functional constraints on what principles we can anchor to govern institutions.

Perhaps, then, functional constraints are what stop us from having an inconsistency-permitting frame. We just cannot anchor the grounding principles that determine the facts about some social institution in such a way that the facts about it could end up being inconsistent, for then the social institution wouldn’t be able to fulfil its purpose. So if the institution does exist and does serve its function, then it cannot have been so anchored.

This initial thought, however intuitive, faces challenges:

1. There would have to be a suitably close connection between (a) the facts about some social institution being inconsistent and (b) the institution not fulfilling its function. Setting aside the idea that dialetheias as such are world-ending disasters, it is not obvious that they always entail bad consequences for institutional functioning. Some might have no consequences for institutional functioning; some might even be beneficial. When they do give rise to problems in the functioning of an institution, these might be malfunctions rather than a general failure to function.

2. Because this strategy invokes the idea of an institution’s purpose, it only applies when institutions have purposes. But some (e.g., the hereditary nobility) might not, and it would be odd to hold that institutions can be inconsistent, but only when they serve no purpose.

3. Suppose that, in some case, inconsistency would stop some actually existing institution i from fulfilling its function. Suppose that some anchoring facts (A₁, A₂, A₃, …) are in place that would, on the face of it, anchor some inconsistency-permitting frame f_i governing i. On the present proposal, the operative frame for i cannot be f_i—so some different frame f_i* must instead be in place. But given that there typically would not be a unique way to make f_i consistent, what is the frame f_i*, and what facts determine that it is the operative frame?

Thus, while it is plausible that there are ways our institutions couldn’t have been, this idea is hard to shape into an explanation of the consistency of social facts.

4.4. Anchor Holism

We have assumed above that anchoring happens in a piecemeal way: we anchor a principle here, a principle there, and the frame is what we get by adding up everything that’s been anchored. This is why inconsistency seems easy: if we anchor principles without coordination, why expect the result to always make sense? But perhaps this is not how anchoring should be understood. Perhaps anchoring is not piecemeal; perhaps entire frames are put in place at once. Call this the holistic option.

We can articulate this further. Let’s say that there is a space of candidate frames f₁, f₂, f₃, …, best viewed in this case as the space of functions from pre-social worlds to social worlds. What determines which of these is our frame is which one of these frames fits best with our overall anchoring-related behaviours. These candidate frames are not just any old list of principles, in the same way that a possible world isn’t just any old set of propositions. A candidate frame has to fit a certain formal profile, and that includes (we propose) not being inconsistency-permitting. As a result, whatever is anchored is guaranteed to force consistency upon the social world.

This is just a technical proposal for how to build consistency into the Epstein model; it needs motivating. But a reasonable motivation can be given. Picture a frame as a theoretical posit: something that an anthropologist studying the social interactions within some population (say) would hypothesize to explain the behaviour they observe. They observe what people say and do, and they propose a way of making all that add up to a social reality, one in which people are responding in reasonable ways to each other and to the situations they’re in. A frame, viewed in this way, has a particular kind of explanatory function: it is meant to rationalize people’s social behaviour.³² If this is the point of positing a frame, then it makes sense that frames might have to obey certain general conditions of coherence, for that could be part and parcel of their rationalizing explanatory role.³³

We’re not quite there yet. If this is to constitute a response to the issue of social inconsistency, we need to argue specifically that one of the conditions that candidate frames must obey is consistency. To do that we must show that consistency contributes to making the explanation yielded by the frame an appropriately rationalizing one. Furthermore, we cannot merely argue that by inducing consistency, a frame is ceteris paribus a better explanation. That would only motivate attributing a consistency-inducing frame on the specific occasions where it would yield a better explanation. If consistency is a general requirement on candidate frames, it must be the case that, in each case where we might explain a population’s behaviour with an inconsistency-permitting frame, there is a consistency-inducing candidate frame that does a better job of explaining that behaviour.

In effect we need something like a dominance argument for consistency as a condition on frames, somewhat like the dominance arguments that formal epistemologists offer for probabilism as a condition on rational credences (e.g., Joyce 1998). But it is not clear that such an argument is available.

Imagine what a case might look like in which we would consider attributing an inconsistency-permitting frame to a population. Imagine we are documenting the social structures of population P, the inhabitants of an isolated village. For population P, imagine that grounding principles for social facts are typically put in place via the pronouncements of a local authority figure—call them the Reeve. The Reeve comes up with principles and declares them to the group, and the villagers treat these pronouncements as authoritative. Now distinguish two cases.

1. The Reeve has put in place a principle to the effect that sheep are not for eating, and later, perhaps in a lapse of judgement, puts in place a principle to the effect that mutton is to be served on Fridays. This presents the villagers with a conundrum, and they end up quietly ignoring the second pronouncement, treating it as some kind of anomaly. If we as theorists now have to choose between attributing an inconsistent frame to P on which sheep are both for eating and not for eating and a consistent frame which excludes the mutton-on-Fridays principle, the choice is easy; the latter explains the villagers’ behaviour much better than the former.

2. The Reeve makes many pronouncements, and often doesn’t remember what principles they have put in place. Fortunately, their principles often concern matters of little practical import. They put in place a principle to the effect that lambs born on a Sunday are noble creatures, and a principle to the effect that lambs born on a weekend are not noble creatures. Nothing practical follows from this either way, so villagers are happy to affirm the truth of both of these principles if prompted. In this case, it seems there is no explanatory advantage to be gained by attributing to P a consistent frame which edits out the inconsistency, as opposed to an inconsistent one that leaves it in. A consistent frame, in this case, would have both a lesser degree of fit with the apparent anchoring facts and less explanatory value with regard to villagers’ behaviour (specifically, their utterances).

It does not seem, then, that consistent frames dominate inconsistent ones when it comes to rationalizing behaviour; it is not such a universal panacea that we need to build it into the very idea of a frame.

The defender of consistency has a reply to offer here, though. For if our imaginary anthropologist is a classical logician, it may not seem to them that interpreting case (2) through the lens of an inconsistent frame is a good idea at all. If it is true in (2) that sheep are and are not noble creatures, then by explosion everything is true in the social world of (2), including many things that do not chime at all with the behaviour in the group. The same applies to any situation in which an inconsistent frame is applied. So by the lights of such an anthropologist, consistent frames do dominate inconsistent ones.

This leaves the defender of consistency in a somewhat ambiguous position. On the one hand, they now have a reasonable way to maintain their position against the prima facie case for social dialetheias given in §4. Rather than just denying the possibility of such things flat out, they can tell a story about anchoring which explains, on their principles, how it is that dialetheias fail to arise. On the other hand, this story only works given a background commitment to consistency, for only on those principles does it make dialectical sense to appeal to the principle of explosion in the dominance argument. Thus if we were trying to sway an audience that’s unopinionated about the possibility of inconsistency, and which perhaps has been given some reason to doubt consistency in the form of the argument of §3, this story would do no good.

Independently of how it bears on the issue of inconsistency, anchor holism is an interesting refinement of the Epstein model, and may be worth exploring further for its own sake. It is also possible that anchor holism can be paired with motivations other than the ‘interpretationist’ one I’ve sketched here, and that might provide other ways for the dialetheist and consistentist to skirmish. But I will leave the matter here.