Disagreement in metametaphysical dispute

Abstract

Recent years have seen several studies of metaphysical disputes as disagreement phenomena employing the resources from the research on disagreement in social epistemology. This paper undertakes an analogous study of the metametaphysical disagreement over the substantiveness of metaphysical disputes between inflationists and deflationists. The paper first considers and questions the skeptical argument that the mere existence of the disagreement mandates the suspension of judgement about the substantiveness of metaphysical disputes. Rather, the paper argues that steadfastness in the face of this disagreement is rational, at least for inflationists. Since inflationists are often metaphysicians who were called to this disagreement due to its apparent threat to their first order debates in metaphysics, they can therefore return to these debates in good faith. In contrast, deflationists have no such alternative occupation and the verdict of steadfastness will not alter their engagement in the inflationist/deflationist disagreement: they will continue their attempt to resolve the disagreement to their advantage. Thus, though the verdict of steadfastness is epistemically symmetric between inflationism and deflationism, it induces an asymmetry in the motivation to pursue the inflationist/deflationist disagreement which places the burden of advancing the dialectic of this disagreement with the deflationists while metaphysicians can continue their work as before.

1 Introduction

Metaphysical deflationists argue that debates in metaphysics are concerned with some non-substantive issue rather than being about ultimate reality. Such a deflationism has recently been supported by the analysis of metaphysical debates as instances of disagreement using the resources from the research in social epistemology on disagreements (e.g. Belleri, 2017; Plunkett, 2015; Thomasson, 2017). The view resulting from this work is that metaphysical disagreement should be regarded as metalinguistic negotiation or verbal dispute in accordance with the general tenet of metaphysical deflationism that metaphysical debates are non-substantive.

This deflationist view of metaphysical debates is refuted by others—often metaphysicians themselves—who insist that metaphysical debates are substantive (e.g. Balcerak Jackson, 2013; Lowe, 1998; Sider, 2011). Thus, out of these competing views of the debates in metaphysics emerge a second order metametaphysical disagreement between metaphysical inflationists and deflationists. While first order metaphysical disagreements have received thorough scrutiny as disagreement phenomena, this second order disagreement has hitherto not been analyzed using the resources of disagreement research. It is such an analysis that will be undertaken here with the purpose of defending inflationism against a skeptical argument coming from the mere existence of a disagreement with deflationists.

The study of disagreements is often cast in terms of the question whether it is rational to be conciliatory or steadfast upon the discovery of a disagreement. That is, whether the discovery of a disagreement mandates (possibly extensive) revision of the disputed belief or whether it is rational to retain one’s belief in such conflicts. As a prominent proponent of the conciliatory view, it is not surprising that Christensen (2007) also favors this view in the context of philosophy. But he finds philosophy to be in such a poor epistemic condition that we should generally withhold belief in philosophical matters. This is disheartening news since such suspension of judgement in philosophy is argued to entail some variant of skepticism (e.g. Machuca, 2012). Recently, however, some authors have questioned the grounds for favoring the conciliatory view in general and suspension of judgement in particular in the context of philosophical disagreement (e.g. Grundmann, 2013; Kelly, 2016). They propose that philosophers are rational to have confidence in their philosophical beliefs despite the systematic disagreement in philosophy; philosophers should be resolute.¹ This paper will investigate the import of these various attitudes to disagreement in philosophy in the context of the inflationist/deflationist dispute. As one might expect, an investigation of the inflationist/deflationist dispute as a disagreement phenomenon will not once and for all settle whether metaphysical debates are substantive or not. However, the paper makes three interrelated finds that are relevant for this question: First, the skeptical argument from disagreement, if sound, is a challenge to metaphysical inflationism and to the continuation of first-order work in metaphysics. Second, the skeptical argument makes an implicit assumption about the distribution of philosophical opinion which in the context of the inflationist/deflationist dispute is likely violated and in a way that favors inflationism. Third, if this is so, then it is rational to continue work on first-order debates in metaphysics, and it is argued that this places the primary burden of advancing the dialectic of the inflationist/deflationist dispute with the deflationists.

The paper proceeds as follows: Sect. 2 explores in more detail the inflationist and deflationist views of metaphysical debates, and Sect. 3 then sketches a skeptical argument that due to the disagreement between inflationists and deflationists mandates both parties to suspend judgement about the substantiveness of debates in metaphysics. While Sect. 4 suggests that this will not force an end to metaphysics, such suspension of judgement is still argued to be the cause of crisis in metaphysics. Section 5, however, questions the soundness of the skeptical argument in the disagreement between inflationists and deflationists. Section 6 goes on to argue that this introduces an asymmetry in the disagreement. Where deflationists’ interest in the inflationist/deflationist dispute is unaltered by the prescription of resoluteness, the metaphysicians on the inflationist side are rational to continue their first order debates in metaphysics as before. Insofar as they find metaphysical debates more important, their interest in the inflationist/deflationist dispute will therefore diminish. As a consequence, it is the deflationists who will have to advance the dialectic of the inflationist/deflationist dispute. Section 7 explores the possibility that the inflationist/deflationist dispute is itself non-substantive and finally a conclusion follows.

Observe that the aim of this paper is not to be an original contribution to the literature on peer disagreement. It will therefore not do justice to all the subtleties found in that (very rich) literature. Rather, the twofold aim is (1) to invite a new way of looking at metametaphysical debates using resources from the peer disagreement literature and (2) to suggest and exemplify this—to my knowledge—novel approach to analyzing such metaphilosophical debates that might be interesting to use elsewhere, metaethics being an immediate candidate.

2 Attitudes towards metaphysical disputes

Explications of the content of metaphysics often bear some resemblance to the characterization of metaphysics as the “systematic study of the most fundamental structure of reality” (Lowe, 1998, p. 2). The view usually associated with metaphysical deflationism is that such explications give a misguided impression of the content of metaphysical debates. One example of such a deflationism is found in the tradition after Carnap (1950) where metaphysical debates are analyzed as a type of verbal dispute concerned with language use and the pragmatic issues related to language choice (e.g. Chalmers, 2009; Hirsch, 2005; Price, 2013; Yablo, 1998). In more general terms, deflationists argue that metaphysics does not carry the kind of significance indicated by explications such as that above.

The deflationist conceptions of metaphysics are resisted by others, often the metaphysicians themselves. Jonathan Lowe, for instance, considers the idea “to understand the aim of metaphysics […] as the attempt to analyse our currently accepted ways of talking” (Lowe, 1998, p. 2) but forcefully dismisses anyone undertaking such a project with the proclamation: “let us not pretend that in doing so we would be doing anything worth dignifying by the name ‘metaphysics’” (Lowe, 1998, p. 2). Thus, Lowe is an exemplar inflationist who argues that metaphysical debates do have some substantive content, whereas Carnap is an exemplar deflationist who denies that metaphysical debates are substantive.

Some have questioned whether the characterization of a debate as a verbal dispute entails that the debate is non-substantive (e.g. Balcerak Jackson, 2014; Belleri, 2018). Inflationists could arguably recognize that metaphysical debates could be characterized as verbal disputes, but add the qualification that a language can consist of “joint-carving terms” (Sider, 2011, p. vii) which imbue the apparently verbal dispute with the type of significance that is here associated with the inflationist view. Analogously, there are deflationists who nevertheless develop ontologies, for instance based on ordinary language (e.g. Thomasson, 2015) or science (e.g. Quine, 1948). They will insist that the elements of these ontologies exist or are real—they will be realists—but at the same time, these ontological commitments and the associated notion of reality are not meant to carry the usual significance. The realism is a deflated one; a lightweight realism in the terminology of Chalmers (2009).

To better capture these subtleties, the inflationist view of metaphysics will here be modelled on Gideon Rosen’s suggestion that “a metaphysical claim is a substantive claim about some allegedly important aspect of reality” (Rosen, 2014, p. 552, emphasis added). In contrast, ‘deflationist’ will be used here for someone who denies this and instead argues that metaphysical debates have some defect such that they cannot carry this kind of significance.² On this very broad construal, deflationists argue that metaphysical debates are shallow or perhaps even trivial; they are not resolved by expansive theorizing. Though different deflationists will explicate the problems of inflationary metaphysics and their origin slightly differently, the shared characteristic of deflationism is the belief that metaphysics is not and cannot be some grand endeavor. While deflationists deliver their analyses of metaphysical debates in a voice that suggests that metaphysicians—regardless of metaphysical convictions—should worry, inflationists—such as Lowe—immediately disclose themselves by their dissatisfaction with deflationists’ analyses of their debates. As such, both deflationists and inflationists recognize that something is at stake. If the deflationists’ worries about metaphysics are true, then both inflationists and deflationists seem to agree that metaphysical debates lack something such that they must either end entirely or only continue with (significant) alteration. In the following, ‘substantiveness’ will be used as placeholder for whatever it is that metaphysical deflationists and inflationists agree is essential for inflationary metaphysics. According to inflationists, metaphysical debates are substantive, whereas deflationists argue that they are non-substantive.

Consequently, a second order dispute emerges. It is a dispute between metaphysical inflationists on one side—often the metaphysicians themselves—who disagree with the characterization of metaphysical debates as non-substantive proposed by the deflationists that occupy the other side of the dispute. Thus, there is a disagreement between inflationists and deflationists over the substantiveness of the content of metaphysical debates. Of course, some care must be taken here. If metaphysicians may be wrong about the content of their first order metaphysical debates as argued by the deflationists, then the same might be the case for this second order dispute between inflationists and deflationists. Without a more direct explication of the content of this second order dispute, the possibility remains that a more thorough third order analysis will reveal this dispute to be misguided or at least misinterpreted. Rather than being a substantive debate about the substantiveness of metaphysical debates, the debate between inflationists and deflationists could perhaps be non-substantive. Some meta-deflationists could propose that this debate is merely a verbal dispute negotiating the use of the word ‘substantive’. In this scenario, inflationists and deflationists would agree about the content of debates in metaphysics and simply disagree over how to describe this content. While this meta-deflationist analysis is a possibility, it will for now be assumed that the second order dispute between inflationists and deflationists is substantive and therefore a genuine philosophical disagreement. In Sect. 7, we will return to this issue.

3 A skeptical argument from disagreement

Peer disagreement is standardly defined as a disagreement between epistemic peers, i.e. a disagreement between two persons that for a given subject “are familiar with all the same evidence and arguments and are equals with respect to the general intellectual virtues” (Feldman & Warfield, 2010, p. 3). In the terminology of Christensen (2007, p. 211), peer disagreements satisfy the principles of “evidential equality” and “cognitive parity”. Peer disagreements are interesting since both sides in the debate, qua peers, must be regarded as equally knowledgeable and competent. By construction, such disagreements are not resolved simply by bringing the participants on epistemic and intellectual equal ground.

Christensen (2007, p. 193) gives the example of two friends who, in their heads, try to divide a restaurant bill evenly between them, but arrive at different results. The two friends have a long history of doing this when eating out together and therefore know they are correct equally often. Finding oneself in such a disagreement, Christensen offers the intuition: “It seems obvious to me that, absent some special reason for thinking that I had some advantage […], I should revise my belief” (Christensen, 2007, p. 194). The revelation of a disagreement with an epistemic peer is evidence of possible error in the establishment of one’s belief, and one should therefore lower one’s credence in that belief accordingly. Both friends in the example should be less confident in the result they arrived at after the discovery of the disagreement than they were before.

Christensen also considers how this conciliatory view applies to philosophical debates:

In philosophy, at least, I think that the parties to disputes are fairly often epistemic peers. And often, if one of the parties were to consider the question of what explained his disagreement with a peer […] he would find that he had good reason for taking the explanation in terms of his own error to be just as good as the explanation in terms of his peer’s error. In such a case, he should not hold his position with the sort of confidence that philosophers often seem to have about so many controversial issues (Christensen, 2007, p. 215).

According to Christensen, philosophers should be conciliatory upon the discovery that they disagree with another philosopher, i.e. this revealed peer disagreement counts as evidence that should (significantly) reduce philosophers’ level of credence in their philosophical beliefs. Indeed, Christensen finds philosophy to be in such a poor epistemic condition that he concludes “that in fields like philosophy, taking account of disagreement in the ways I’ve been defending would lead to general withholding of belief in many cases” (Christensen, 2007, p. 215).³ In many philosophical debates, philosophers should simply suspend judgement and as a consequence “a broad skepticism about philosophical matters threatens” (Kornblith, 2010, p. 33).

Thomas Grundmann reconstructs this skeptical argument in the following way:

(1′) If there is a genuine and roughly symmetric disagreement among epistemic peers (= GSE-disagreement), all parties are rationally required to suspend judgment⁴

(2′) In philosophy, GSE-disagreements are widespread and stable.

Therefore,

(C1) In philosophy, hardly any judgment can be rationally retained (Grundmann, 2013, p. 74, emphasis in original).

A disagreement is genuine if it does not rest on a misunderstanding: if there is in fact no conflict among the disputants, then, Grundmann claims, it cannot be the case that the disputants should suspend judgement. For now, we will assume that this condition is satisfied, but we will return to it in Sect. 7. The skeptical argument also assumes that the disputants are epistemic peers. Grundmann gives the example that if he finds himself in a disagreement with a flatearther about the topology of the earth, then it is hardly rational for him to suspend judgement. Somehow, the flatearther is disqualified as someone who mandates the suspension of judgement in such disagreements. Premise (1′) applies only if the disputants are epistemic peers, i.e. they have the same evidence available (evidential equality) and are equally competent in assessing this evidence (cognitive parity).

Symmetry, Grundmann qualifies, requires that “the opposing parties are of the same size” (Grundmann, 2013, p. 74). While it is rather straightforward that the skeptical argument requires the disagreement to be a genuine disagreement between epistemic peers, the requirement of symmetry may come across as more surprising. In defense of the symmetry condition, Grundmann offers the intuition that if 100 philosophical peers discuss a philosophical question, then if only one person disagrees with me whereas the other 98 persons agree, then I should not suspend judgement, i.e. in such cases, the skeptical argument does not apply. Offering a similar argument, Thomas Kelly gives the more vivid example of a group of people that look at a sign at a distance. Each person reads its first word as ‘our’ or ‘out’. However, if a substantial majority leans towards ‘our’ rather than ‘out’, then this should arguably increase their credence that the sign says ‘our’. Kelly accompanies this by several quantitative estimates of the probability that various majorities are correct. One of them is a group of 50 individuals that each have a 60% chance of being correct. If they split 30 against 20 on some dichotomous question, then the probability of the majority being correct is more than 98% (Kelly, 2016, p. 388). Kelly does not give the formula for these estimates, but they seem to be based on the general expression for the probability that the state of the world (denoted X) is x given that the number of votes for x (denoted N_x) is h in a group of n individuals (List, 2004, p. 528):

$$\begin{array}{*{20}c} {P\left( {X = x{|}N_{x} = h} \right) = \frac{r}{{r + \left( {1 - r} \right)\left( {{\raise0.7ex\hbox{$1$} \!\mathord{\left/ {\vphantom {1 l}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{$l$}} - 1} \right)^{2h - n} }}} \\ \end{array}$$

(1)

where l is the probability that each individual is correct and $r=P(X=x)$, i.e. the prior probability that the state of the world is x. Kelly seems to assume that $r=0.5$ which is the same as saying that one regards both answers to the dichotomous question as equally probable at the outset.⁵

The figures given by Kelly suggest that the majority view has a very high probability of being correct even in cases of relatively small majorities. However, some assumptions go into these estimates that are typically violated in philosophical disagreement. First, it is assumed that all individuals in the group have exactly the same probability, l, of being correct. This seems rather unrealistic, but one can compensate for it by setting a conservative value for l. Making thus sure that most individuals are more competent than indicated by l, one will typically underestimate the probability that the majority is correct. Kelly also raises the related concern that one could “consider a view according to which you cannot reasonably believe [a proposition] p so long as some formidable philosopher earnestly denies p” (Kelly, 2016, p. 389). Though Kelly does not go into this kind of detail, the argument might be that such a “formidable philosopher” has a probability very close to 1 for choosing the correct alternative. If this is so, then even the aggregated probability that the group chooses the correct alternative might be less than that of this formidable philosophy and the group would therefore be rational to adopt the view of the formidable philosopher irrespective of whether it is the majority view. Kelly dismisses such concerns since it “attributes too much undermining power to the views of formidable outliers” (Kelly, 2016, p. 389). Presumably, this is saying that even if there are formidable philosophers who have a higher probability of arriving at the correct view, then this probability will still be far from 1. The group will therefore typically have a higher probability of choosing the correct alternative than the formidable philosophers whereby they lose their potential undermining power. However, for present purposes an alternative response, not considered by Kelly, is that such formidable philosophers would compromise the assumption of the skeptical argument that the disagreement is between epistemic peers. Thus, if the existence of formidable philosophers is invoked to shield the skeptical argument from the symmetry requirement, this consideration would instead defeat another of its premises. For the skeptical argument to apply, the competence among the disagreeing parties must be relatively homogenous, i.e. they can be assumed to have (approximately) the same individual competence, otherwise they would not be epistemic peers.

Formidable philosophers, however, can feature in another way that is consistent with peerness, namely as opinion leaders. Other philosophers might look to them for guidance when they decide which philosophical view to adopt. This relates to another assumption of Kelly’s calculations: that the votes must be independent of each other. If a philosophical disagreement has one such formidable philosopher on either side with the remaining participating philosophers simply following their lead, then “the numbers should drop out entirely” (Kelly, 2016, p. 389); it should not matter which of the formidable philosophers has more blind followers. Kelly insists that nothing like this actually obtains in philosophy, however, more moderate dependence between philosophers does seem likely. Such dependence typically has the effect that Eq. (1) overestimates the probability that the majority is correct. Kelly recognizes this challenge and tries to moderate his conclusion accordingly:

the point that I wish to make is a relatively modest one: given the kinds of lopsided majorities that often obtain in philosophy, there is no reason to think that an adequate egalitarian theory (one which was sensitive to the various kinds of influence and dependencies in the way in which philosophers arrive at their views) would end up recommending agnosticism as the correct general response to philosophical controversy, as opposed to a relatively confident belief that a certain position is true in many cases (Kelly, 2016, p. 388).

Kelly’s reasoning, it seems, is that the probabilities given by Eq. (1) are so high in many cases—as we shall also see in Sect. 5—and the cases where voting becomes entirely irrelevant so fictitious that the probability that the majority is correct remains high even though Eq. (1) overestimates it due to the expected moderate dependence among philosophers. Thus, even though the probabilities obtained by Eq. (1) cannot be treated as fact, they can be considered to suggest that sufficiently “lopsided majorities”⁶ provide some justification for resoluteness for those in that majority (see Jaksland (2022) for further discussion). Since the skeptical argument is meant to show that agnosticism is the rational attitude towards a given philosophical question, it is sufficient to defeat it that those in the majority are rational to be resolute.⁷ Kelly argues, in other words, that if I already believe that p, then I will be rational to continue to believe that p despite my discovery that other people disagree with me, if many more people agree with me and I have no reason to suspect that many of those agreeing with me are simply blind followers. Being in such a majority entails that I do not have to reduce my credence in p or perhaps rather that I do not have to reduce it so much that I become agnostic about p. In such cases, I am rational to be resolute.

Notice that this is very different from arguing that philosophical questions can be decided by vote. Unless belief is only warranted in cases of certainty, I can believe something based on an estimate of the probability that the belief is correct even though this probability is less than 1.⁸ Thus, the probability that the majority is correct can be less than 1 and still serve as part of the justification for my (continued) belief in the majority view. One can therefore appeal to asymmetry to defeat the skeptical argument without assuming that the majority is certain to be correct and thus without committing to the arguably controversial idea that a simple vote among philosophers could decide philosophical questions. The point of Kelly’s argument is that if Eq. (1) estimates the probability that the majority is correct to be well above some threshold, then this is sufficient reason for those in the majority to disregard the skeptical argument even though this threshold value will not suffice to establish that the majority view is true.

Kelly’s argument crucially depends on the assumption that philosophers have strictly more than 50% chance of preferring the correct philosophical view, otherwise the probability that the majority is correct will also be 50% for $l=0.5$ and less than 50% if $l<0.5$. For Kelly’s argument to go through, philosophers must therefore be better judges on philosophical questions than, say, a fair coin is and perhaps they must even be quite a bit better to compensate for dependence. Estimating philosophers’ competence on philosophical questions may, however, prove difficult since so few philosophical questions have been resolved. The competence of philosophers can therefore not be based on their track-record on such questions.

For present purposes, however, the context of the skeptical argument from disagreement provides a way around this difficulty. This skeptical argument is driven by the discovery of a disagreement. It is this discovery that mandates a reduction in credence such that it falls below the threshold for belief whereby agnosticism becomes the rational doxastic attitude. However, if philosophers only had close to 50% chance of adopting the correct view on philosophical questions, then they should arguably suspend judgement anyway. The toss of a fair coin would not suffice to justify philosophical belief, so if the competence of philosophers were close to that of a fair coin, then the same should apply to philosophers. But if philosophers’ competence were such that they should always suspend judgement on philosophical questions, then there would be no reason to invoke the skeptical argument from disagreement. This skeptical argument is only relevant if philosophers in general are rational to have philosophical beliefs because only then is the discovery of the disagreement an occasion to change the doxastic attitude from belief to suspension of judgement. The skeptical argument from disagreement therefore implicitly assumes that philosophers’ competence is high enough at the outset to justify belief. In another context, Grundmann (2013, p. 77) stipulates that 70% chance of being correct is sufficient for belief in a proposition. Under this assumption, the individual competence would, in other words, have to be at least 0.7 in order for the skeptical argument from disagreement to be relevant, otherwise suspension of judgement would be mandated anyway. For purposes of illustration, we shall here follow Grundmann’s stipulation, but nothing in the discussion depends on the exact value of 0.7, only that it is somewhat larger than 0.5. More on this in Sect. 5.

4 The consequences of the skeptical argument

While the skeptical argument concerns all of philosophy, the present interest is in its relevance for the inflationist/deflationist dispute, i.e. whether this is a GSE disagreement, whether suspension of judgement is therefore rationally required in this dispute specifically, and what consequences this has. The former two questions will be the topic of the next section, while the present section explores the latter. If the skeptical argument is sound, i.e. if the inflationist and the deflationist realize they have a GSE-disagreement,⁹ then both parties are rationally required to suspend judgement. Observe that we are here considering the skeptical argument in the context of the second order metametaphysics disagreement between inflationists and deflationists and not applying it to first order metaphysical disagreements. While such an application could serve to defend a type of epistemic metaphysical anti-realism, this is not what is at issue here! The point is not that metaphysicians should suspend judgement about questions in metaphysics, but rather that inflationists and deflationists—including the metaphysicians among them—should suspend judgement about the substantiveness of debates in metaphysics as a consequence of their disagreement about this issue.

Presumably, most metaphysicians pursue questions in metaphysics because they regard them as important questions. Most metaphysicians are, in other words, inflationists. Arguably, few metaphysicians would be interested in metaphysical debates if they turned out to be non-substantive: there are few deflationist metaphysicians.¹⁰ However, the skeptical argument does not establish deflationism. Rather, it requires metaphysicians—and everyone else—to be agnostic with respect to the substantiveness of metaphysical debates; they should neither believe in inflationism nor in deflationism. If the skeptical argument is sound, inflationist metaphysicians must, in other words, change their attitude to agnosticism about the content of these debates.¹¹

The following story might illustrate the consequences of such a change to metametaphysical agnosticism. Imagine that I am involved in the design of a large building. As a specialist on colors, I am tasked to decide the colors on the exterior of the building. Having worked on this project for a while, I get to the color of the balconies. I put much effort into this work weighing various alternatives against each other. One day, however, my manager points out that it is not yet decided whether the building will even have balconies and estimates the probability to be only 50%. This will arguably make me reconsider my priorities. Perhaps I will look around for other parts of the building that need to have their colors decided or, if I have grown very committed to the balconies and their coloring, I might spend my time lobbying for balconies so my efforts would not have been in vain. I could of course continue my work on the balconies and simply hope that they will make it to the final design, but my motivation to do so would likely have diminished significantly. Furthermore, if I simply continue my work on the colors of the balconies with unaltered devotion, my manager would probably suspect that I misunderstood the message. If I instead explain that my continued effort is due to my hope that the building will have balconies, my manager might ultimately order me to finish my other tasks first.

Philosophers rarely have managers like this, but the manager can perhaps be regarded as representing a type of extended rationality. While there is nothing outright contradictory in pursuing a metaphysical question even after one is forced to be agnostic about its significance, such behavior might come across as somewhat peculiar if there are other questions to work on whose significance is more certain. This is especially so, if one’s enthusiasm for the original question is entirely unaltered by the change of view to metametaphysical agnosticism. This might even raise the suspicion that this agnosticism is insincere. As such, the metametaphysical agnosticism due to the skeptical argument from disagreement does not force an end to metaphysics. This, however, is not the same as saying that such agnosticism is compatible with metaphysics simply continuing as before. Indeed, the story indicates that once we are forced to suspend judgement about the significance of a particular occupation, then we will—and arguably should—be less inclined to pursue this occupation. In the story, significance depends on whether the final design includes balconies and for metaphysicians, the significance of metaphysical debates depends on inflationism. Once balconies and inflationism, respectively, become uncertain, one will be prone to look for a task that does not depend on them for its significance or to engage in work that might lift this uncertainty. For a field of inquiry like metaphysics, this would arguably manifest itself as a crisis where only a rather limited number of researchers continue their work—and even fewer with unaltered devotion—while the rest will either enter the second-order debate or find some other alternative occupation.

In this light, metaphysics is worse off if the skeptical argument proves to be sound, and the skeptical argument should therefore worry inflationist metaphysicians. Deflationists in contrast should welcome the skeptical argument even though it entails that they too must be agnostic with respect to the inflationist/deflationist dispute. This is because the deflationists nevertheless succeed with what must at least be a derived aim of their disagreement with inflationists: namely that fewer people waste their time on metaphysical debates. However, the details of this situation will not be discussed further here. Rather, the next section will argue that there are reasons to doubt the soundness of the skeptical argument and instead prescribe resoluteness despite the disagreement between metaphysical inflationists and deflationists.

5 Against the skeptical argument

Both Grundmann and Kelly argue against the view that agnosticism is rationally required by the type of disagreement found in philosophy.¹² In his own words, Kelly “survey[s] some prominent alternative views about the epistemology of disagreement and examine[s] what happens when they are applied to the special case of philosophy” (Kelly, 2016, p. 375) and concludes that “there is no plausible view about the epistemology of disagreement on which philosophical agnosticism is compelling” (Kelly, 2016, p. 375, emphasis in original). Grundmann, investigating the skeptical argument explicated above, concludes: “I defend the view that the skeptical argument is unsound” (Grundmann, 2013, p. 76).

Grundmann finds the requirements for a GSE disagreement—a genuine and roughly symmetric disagreement among epistemic peers—so restrictive that he concludes: “we have no good reason to believe that there is (widespread) GSE-disagreement in philosophy” (Grundmann, 2013, p. 76). Thus, he questions premise (2′) in the skeptical argument above. The primary problem, according to Grundmann, is to establish opponents in philosophical debates as epistemic peers. Since so few debates in philosophy are resolved, peerness cannot be based on a track record method; the method for instance used to establish peerness in Christensen’s (2007, p. 193) paradigmatic restaurant bill example. Kelly offers a similar view when he claims that “strong evidence that another person is your peer with respect to some philosophical question, or some cluster of philosophical questions, might be relatively hard to come by” and continues “in the absence of such evidence, sticking to your guns might seem more defensible” (Kelly, 2016, p. 384).¹³

A more concrete challenge to the skeptical argument in context of the inflationist/deflationist dispute comes from the distribution of opinion on this question among philosophers and its implications for the symmetry condition. The largest survey of philosophical opinion is made by Bourget and Chalmers’ (2014) among 931 faculty members at what they describe as “99 leading departments of philosophy” (Bourget & Chalmers, 2014, p. 468).¹⁴ While the survey does not ask directly about metaphysical inflationism, three other positions could perhaps serve as proxies: Affirmative responses to the question about non-skeptical realism about the external world and to the question about scientific realism might be positively correlated with metaphysical inflationism, and the question concerning deflationism about truth has been argued to be associated with metaphysical deflationism (Thomasson, 2014).¹⁵ In the survey, 81.6% identified as non-skeptical realist about the external world, 75.1% advocated scientific realism, and 24.8% favored deflationism about truth (Bourget & Chalmers, 2014, pp. 394, 398). Taking these results as indicative of the inflationist/deflationist dispute in metaphysics, it seems that we might expect a majority of philosophers to favor metaphysical inflationism.

These proxies, however, are obviously not perfect. As already mentioned, there are realist deflationists who might, therefore, commit to non-skeptical realism and scientific realism but who will nevertheless argue that metaphysical debates are not substantive, Quine being a possible example. Others might report to be non-skeptical realist while being largely disinterested in metaphysics. Their non-skeptical realism could primarily be an epistemic stance—that we can know that there is an external world—and not a metametaphysical stance on whether debates about the existence of the external world are substantive.¹⁶ Their commitment to non-skeptical realism would, in other words, underdetermine their position on the inflationist/deflationist dispute. The resulting worry is therefore that we might overestimate the support for inflationism when the two questions about realism are used as proxies. Similarly, Asay (2022) and Marschall and Schindler (2021), among others, have questioned Thomasson’s association between semantic deflationism and metaphysical deflationism, though the latter concede that the two appear to “make an attractive package” (Marschall & Schindler, 2021, p. 118). Indeed, others have supported Thomasson’s conclusion (see, e.g., Knowles, 2018). Nevertheless, these discussions show that deflationism about truth cannot be assumed to be perfectly correlated with metaphysical deflationism, though this does not in itself tell whether metaphysical deflationism is over- or underestimated when using deflationism about truth as a proxy.

Assuming that these proxies overestimate the support for inflationism, inflationists will still be in the majority even if one fourth of the scientific realists proves to be metaphysical deflationists and likewise, if one fourth of the deflationists about truth proves to be metaphysical inflationists. Assuming this is the case, metaphysical inflationists are still at a 55% majority among the respondents. Without an explicit question about the substantiveness of metaphysics, all of this is, in the end, speculation, but given the distribution on the other related questions, a 55% majority for inflationism seems to be a modest estimate.

One might, however, still question whether these survey data are representative. Bourget and Chalmers make no claim that they have sampled the respondents to comprise a representative group of those who should be considered epistemic peers with respect to the survey questions. Furthermore, the survey can at best give a contemporary snapshot of the distribution of philosophical opinion, but it does not take into account that philosophical opinion has changed over time. Equation (1), however, does not make any assumption that the group is representative. It simply aggregates the probabilities based on the actual distribution of opinion in the group and on the competence of its members. In the example of the group deciding whether the sign says ‘our’ or ‘out,’ it is not an issue whether this group is representative of, say, literate people or not. Closely related to representativeness, however, is the concern that members of a group might tend to opt for the same alternative more often than they should according to chance. This is exactly what dependence among voters means in technical terms. The worry that only contemporary philosophers have been surveyed might therefore be posed as the worry that contemporary “philosophical fashion” might incline philosophers towards inflationism. This, however, does not show that the distribution of opinion is irrelevant until every epistemic peer—living or dead—is included in the survey. Instead, it emphasizes that we cannot ignore dependence among respondents which, as mentioned above, entails that Eq. (1) will overestimate the probability that the majority is correct.

Beside the distribution of opinion, i.e. the estimated 55% majority for inflationism, there are two additional variables of Eq. (1): The prior probability for inflationism and the probability that each voter is correct. For present purposes, the former will be assumed to be 0.5, i.e. inflationism and deflationism are considered equally likely, though more might be said about this. The individual competence was already discussed in Sect. 3 where it was argued that the skeptical argument from disagreement implicitly assumes that the individual competence is high enough for belief before the disagreement is discovered. Following Grundmann’s stipulation, this entails that the individual competence is at least 0.7, i.e. each individual has at least 70% chance of choosing the correct answer. Thus, assuming that $n=931$, $h/n=0.55$, $r=0.5$, and $l=0.7$, Eq. (1) gives a probability of 1 (to 30(!) decimal places) that the majority is correct. If the assumptions of Eq. (1) were satisfied, then the 55% majority for inflationism would almost guarantee that inflationism is correct.

However, any dependence among voters in the form of positive correlations, i.e. a tendency to vote the same, will entail that the results of Eq. (1) are overestimations. While such dependence is arguably present in the inflationist/deflationist dispute, the extreme case considered by Kelly where everyone else blindly followed two formidable philosophers is unrealistic. Among graduate students, one might find a strong tendency to inherit the views of the supervisor. In the present case, however, the voters are faculty at leading departments of philosophy. Arguably, they will be more prone to have formed their own beliefs which should at least reduce their tendency to blindly follow the views of others. They might, however, be correlated through shared schools of thought including contemporary philosophical trends. If all the voters blindly followed one of two disagreeing schools of thought—perhaps the one they were indoctrinated into through their graduate program—then that would bring about a similar circumstance. Again, being faculty might reduce this effect, but so would the existence of many different schools of thought. On the deflationist side, there are manifestly several distinct schools of thought such as quantifier variance (e.g. Hirsch, 2002), easy ontology (e.g. Thomasson, 2015), expressivism (e.g. Price, 2013), anti-realism (e.g. Dummett, 1991), maximalism (e.g. Eklund, 2008), and empiricism (e.g. van Fraassen, 2002) to name some. On the inflationist side, the various approaches to metaphysics can arguably be considered different inflationist schools of thought. Naturalized metaphysics as pioneered by Ladyman and Ross (2007) explicitly endorses metaphysical inflationism—despite its criticism of much analytical metaphysics—and defends this view based on the success of science (see Jaksland 2020 for a discussion). Neo-Aristotelean metaphysics’ view of metaphysics as first philosophy is also inflationist (see for instance contributions to Tahko, 2012). But with its focus on substance, essence, and grounding as investigated by largely a priori methods, neo-Aristotelean metaphysics is exactly the type of metaphysics that is criticized by naturalized metaphysics. Though both endorse inflationism, they do so for very different reasons. Other prominent inflationist approaches also include for instance neo-Quinean metaphysics¹⁷ (e.g. Inwagen, 1998), the Canberra Plan (e.g. Jackson, 1998), and joint-carving metaphysics (e.g. Lewis, 1986; Sider, 2011). Thus, both sides of the inflationist/deflationist dispute feature several distinct schools of thought which at least suggests that the majority of philosophers that favor inflationism are not just blindly following a single school of thought. Thus, in the inflationist/deflationist dispute, we can at least disregard the sort of scenario considered by Kelly where numbers become entirely irrelevant.

The remaining dependence, though it as argued cannot be too severe, will still entail that Eq. (1) gives an overestimation of the probability that the majority is correct. This is particularly so, when Eq. (1) estimates that it is next to certain that inflationism is correct based on the stipulated 55% majority for inflationism. However, since we can all but rule out the most extreme cases of dependence, there is reason to believe that the probability that inflationism is correct will still be relatively high even when dependence is taken into consideration, as also Kelly argues in the general context (Sect. 3). That the probabilities strongly favor inflationism, also in worse epistemic circumstances than those considered above, is supported by the fact that the probability for inflationism remains high even in much smaller groups of less competent philosophers. To give just one example, in a group of 93 philosopher who have only 55% chance of adopting the correct philosophical view, Eq. (1) still estimates the probability that the majority is correct to be 90% if there is a 55% majority. Given these figures, the inflationists are arguably rational to be resolute. While the discovered disagreement with deflationists requires them to reduce their credence in inflationism, this effect is countered by inflationism most likely being the majority view.

One might object that too many uncertainties remain both concerning the size of the majority for inflationism and the precise probability that this assigns to inflationism. The skeptical argument from disagreement, however, is not in a better state. Indeed, just the possibility that there is a clear majority for inflationism also leaves it open that we are in a case similar to that considered by Grundmann—though perhaps not as extreme—where the disagreement is between 1 person (a small group) and a group of 99 people (a large group). Unless we have an extreme case of dependence among the 99 people, Grundmann’s intuition that this allows us to disregard the disagreement seems to be sound. Arguing that inflationists should suspend judgement until they can settle these uncertainties amounts, in other words, to granting the skeptical argument the benefit of the doubt. Since there is evidence that inflationists are in the majority and that the dependence among them is not too problematic, the inflationists, if anyone, should arguably benefit from the doubt. Unless this changes, they can disregard the disagreement and its associated requirement to suspend judgement. It is not thereby argued that inflationism is correct, but only that inflationists—all things considered—are rational to be resolute.¹⁸

6 Disagreement in the inflationist/deflationist dispute

In summary, premise (2′) of the skeptical argument can be resisted in the particular context of the disagreement between metaphysical inflationists and deflationists. The study of the inflationist/deflationist dispute as a disagreement phenomenon instead suggests that at least inflationists are rational to be resolute since they are (likely) in a majority. To simplify the discussion, it will be assumed that also deflationists are rational in being resolute even though they are in the minority. Arguably, this should only strengthen the deflationists’ position as compared to the case where they must adopt agnosticism instead.

If all parties are resolute, then the dispute over the substantiveness of metaphysics will continue. Without further evidence, the disagreement will not be resolved. As such, the status quo is symmetric in that each side retains their belief, and the only way to change this is if new relevant arguments come up. However, in parallel with the situation in Sect. 4 where deflationists would be more satisfied when both parties were subject to the skeptical argument, the epistemic symmetry of resoluteness does not entail that inflationists and deflationists should be equally satisfied or dissatisfied.

To see this, consider the observation already made that many of the participants on the side of metaphysical inflationism in the inflationist/deflationist dispute are metaphysicians forced into the debate out of a concern that deflationism poses a threat to the substantiveness of their first order debates in metaphysics. For these metaphysicians, resoluteness is more welcome than the agnosticism entailed by the skeptical argument. Resoluteness cancels the potential crisis in metaphysics due to the skeptical argument since it makes it rational to remain a metaphysical inflationist despite the revelation that deflationists exist. Metaphysical inflationists can thereby continue the first order debates in metaphysics with the conviction that they are substantive. Hence, the mere existence of deflationists poses no immediate threat to metaphysics assuming the soundness of the arguments of the previous section to the effect that resoluteness and not agnosticism is the rational attitude. Just as the skeptical argument and the entailed agnosticism were good news to the deflationists, so resoluteness is good news for the metaphysicians.

This induces an asymmetry between the inflationist and the deflationist sides of the inflationist/deflationist dispute even though resoluteness is epistemically symmetric between metaphysical inflationism and deflationism. Generally, if it is rational to be resolute about a view, then it is rational to continue any endeavor that presumes the disputed view despite the disagreement. Given resoluteness, the working metaphysicians among the metaphysical inflationists can go back to their primary interest: metaphysical debates. And as a result, their engagement in the inflationist/deflationist dispute will diminish. If agnosticism is not mandated, then the metaphysicians will find less motivation to advance the inflationist/deflationist dispute in an attempt to resolve it in favor of metaphysical inflationism. In this sense, their interest in the dispute is reduced as a consequence of resoluteness.

The metaphysicians can continue their work in metaphysics as before. The same of course goes for any activity that assumes deflationism, but there appear to be no such activities; no first order debates rely on the assumption of deflationism. Deflationists are of course welcome to engage in metaphysical debates while resolute in their belief that such debates are non-substantive, but arguably very few will have an interest in such an activity. If so, deflationists appear to have no first order activity to engage in as an alternative to the inflationist/deflationist dispute. This is where the asymmetry enters. Arguably, most deflationists are engaged in this dispute because they deem it important or perhaps even because they have an ambition to end the—in their view illegitimate—metaphysical theorizing. The former ambition is unaffected by resoluteness while the latter is even countered if resoluteness is the rational attitude. Thus, the deflationists’ interest in the inflationist/deflationist dispute is unaltered or even increased by resoluteness; resoluteness forces the deflationists to develop new arguments for the inflationist/deflationist dispute in order to achieve their aims. This, of course, is no different from metaphysical inflationists whose interest in the inflationist/deflationist dispute is independent of its implications for the first order debates in metaphysics. They will continue their engagement in the inflationist/deflationist dispute regardless of the verdict: agnosticism or resoluteness. The difference arises because the many metaphysicians on the metaphysical inflationist side are primarily interested in metaphysics and are therefore less concerned about the inflationist/deflationist dispute once resoluteness is established as rational. Despite the epistemic symmetry of resoluteness in the inflationist/deflationist dispute, an asymmetry is induced in the motivation to resolve the disagreement. The deflationists, who have this dispute as their primary occupation, have more interest in the resolution of their disagreement as compared with the metaphysical inflationists who have an alternative occupation in studying what they are rational to believe are questions about ultimate reality.

Of course, metaphysicians cannot as a consequence simply disregard the inflationist/deflationist dispute. If new pieces of evidence—presumably taking the form of new arguments—are offered for one side or the other, then this requires the participants in the debate to reassess the held positions. The disagreement must always be genuine such that each party has seriously considered the available evidence and found that they favor their respective positions. Suppose someone in the restaurant bill example did the numbers on a calculator, then this would force the participants in the disagreement to reassess their position and would most probably resolve the disagreement. The same goes for the inflationist/deflationist dispute. Whenever new evidence is disclosed, then both sides of the dispute must seriously assess their position in light of this new evidence. The equivalent of a calculator may be hard to come by in the inflationist/deflationist dispute, but if someone found it, then nothing said here would entitle the metaphysicians to ignore this new evidence. What has been argued here is instead that the deflationists—because of resoluteness—will look for the calculator with more urgency. The lesson from disagreement is that the current state of the inflationist/deflationist dispute is such that the deflationists have more motivation to advance the disagreement with new evidence since the many metaphysicians among the inflationists will most likely pursue their studies of ultimate reality instead.

7 A metametametaphysical dispute?

Hitherto, the dispute between metaphysical inflationists and deflationists has been assumed to be substantive. However, as already alluded to in Sect. 2, one may wonder whether this conception of the dispute could be misguided; just as deflationists argue that debates in metaphysics are misguided. A meta-deflationist might propose that the inflationists and deflationists actually agree about the content of debates in metaphysics and simply disagree over what words to use when describing this content; they might for instance be negotiating the use of the word ‘substantive’. For metaphysicians, meta-deflationism might seem promising. First, it entails that the skeptical argument may not be sound in the case of the inflationist/deflationist dispute since the disagreement—as the meta-deflationist argues—is not genuine. Second, meta-deflationism might seem to imply that metaphysicians can disregard the inflationist/deflationist dispute. This dispute only concerns the labelling of metaphysical debates whereas the content is not under dispute. One may be tempted to add that the metaphysicians should be satisfied with this only if the content of their debate really is substantive. Their satisfaction with meta-deflationism should depend on the outcome of the true inflationist/deflationist dispute, but obviously, this is exactly the dispute that metaphysical inflationists and deflationists proclaim to have and which the meta-deflationist diagnoses as merely verbal. It seems, therefore, that the meta-deflationist must insist that any further interrogation into the substantiveness of metaphysical debates is meaningless. The only meaningful discussion concerns the best use of the word ‘substantive’ echoing Carnap’s (1950) analysis of existence questions and claims. How satisfied working metaphysicians should be about this inscrutability of the substantiveness of their debates is rather unclear. While it might be intriguing to explore this further, it must for now suffice to observe that no one, to my knowledge, has defended meta-deflationism and thus given any arguments to the effect that the inflationist/deflationist dispute is not a genuine disagreement.

8 Conclusion

Studying the inflationist/deflationist dispute as a disagreement phenomenon suggests troubling news for the deflationists. The analysis disfavors suspension of judgement in the inflationist/deflationist dispute and proposes instead that it is rational to be steadfast or at least retain belief (being resolute) in the face of this disagreement. This conclusion is promising for all of those who are engaged in the inflationist/deflationist dispute due to the crisis it might induce for the study of first order debates in metaphysics. Metaphysicians can retain the belief that motivated their engagement in metaphysical debates in the first place; in principle regardless of whether this belief is inflationism or deflationism. However, since few deflationists have an interest in the first order debates, an asymmetry is induced. The many inflationists whose primary concern is first order metaphysical debates can continue their work with these debates, whereas the deflationists have no such alternative to their engagement in the inflationist/deflationist dispute. Despite this situation being epistemically symmetric, the deflationists have more interest in resolving the inflationist/deflationist dispute. Those coming from first order metaphysical debates to the second order inflationist/deflationist dispute because it, in conjunction with the skeptical argument, questions their first order work will in many cases discontinue their engagement in the dispute once the skeptical argument is refuted. Therefore, the current state of the disagreement between metaphysical inflationists and deflationists is such that the deflationists are more motivated to develop the disagreement with new arguments. They must bear the burden of advancing this dialectic.