Is ontological revisionism uncharitable?

Abstract

Some philosophers (‘nihilists’) deny the existence of composite material objects. Other philosophers (‘universalists’) hold that whenever there are some things, they compose something. The purpose of this paper is to scrutinize an objection to these revisionary views: the objection that nihilism and universalism are both unacceptably uncharitable because each of them implies that a great deal of what we ordinarily believe is false. Our main business is to show how nihilism and universalism can be defended against the objection. A secondary point is that universalism is harder to defend than nihilism.

1. Introduction

Composite material objects such as tables, ships and statues give rise to well-known paradoxes. Some philosophers argue that the best response to these paradoxes is to deny the existence of composite material objects (e.g. van Inwagen [1990] and Merricks [2001]; van Inwagen and Merricks each except at least some living composita). Others adopt the same conclusion for reasons to do with causal over-determination (e.g. Merricks [2001]). Call the denial that there are composite material objects nihilism. Other philosophers go to the opposite extreme: they hold that whenever there are some things, they compose something. Call this view universalism.

The purpose of this paper is to scrutinize an objection to these revisionary views based on the notion of charity. In mentioning charity, we do not want to give the false impression that either nihilism or universalism is a theory about the interpretation of ordinary English (though their proponents sometimes make claims about the interpretation of ordinary English in the course of defending them). Rather, the objection is that nihilism and universalism are both unacceptably uncharitable because each of them implies that a great deal of what we ordinarily believe is false. We will show that nihilism can be defended against the objection – but also that it is more difficult to defend universalism against it: in fact, it is unclear whether universalism can ultimately escape the objection.

One response to the objection is to argue that nihilism or universalism is consistent with what we ordinarily believe. (For a version of this response, see van Inwagen [1990].) But in this paper, we will concede that these views do imply that a great deal of what we ordinarily believe is false. We will assume, for instance, that nihilism implies the falsity of our ordinary belief that there are tables; and that universalism implies the falsity of our ordinary belief that there is nothing composed of the moon and Barack Obama’s left foot. (Accordingly, we will refer to the disjunction of nihilism and universalism as revisionism.) Our strategy is to argue that even if nihilism implies that a great deal of what we ordinarily believe is false, that does not make nihilism unacceptably uncharitable. In fact, we will argue that it does not make nihilism uncharitable at all, let alone uncharitable to an unacceptable degree. If our argument succeeds, it shows that there is at least one ontological debate where contradicting ordinary belief need not make our theory unacceptable. Much of our discussion applies equally to universalism. The main business of the paper, then, is to show how nihilism and universalism can be defended from the objection. A secondary point is that universalism is harder to defend than nihilism (see Section 4 below).

One of the most vociferous defenders of our ordinary ontological beliefs about composition is Eli Hirsch. The notion of charity plays a prominent role in Hirsch’s defence. For this reason, we will concentrate on the types of charity which Hirsch has distinguished. Many discussions of Hirsch’s work concentrate on his doctrine of quantifier variance and the idea that contemporary debates within ontology might be verbal. In contrast, our focus in this paper is on what charity can teach us about nihilism and universalism. Accordingly, we set aside the question of quantifier variance and the question of whether ontological debates are verbal. These ideas will play no role in the paper; we will simply discuss whether appeal to charity refutes revisionism.¹

2. Three types of charity

The principle of charity constrains both our interpretation of another speaker’s sentences and our attribution of beliefs to her. Various non-equivalent formulations of the principle have been offered by philosophers. Hirsch formulates the principle as follows:

Other things being equal, an interpretation is plausible to the extent that its effect is to make many of the community’s shared assertions come out true or at least reasonable. (Hirsch 2011, 148)

We should note that Hirsch’s formulation of the principle of charity differs from the formulations found in Davidson’s famous work on charity in one very important respect. Davidson makes the very strong claim that charity rules out the possibility of speakers being systematically in error.² Hirsch, by contrast, makes the much more moderate claim that charity gives a presumption that speakers are not in error. According to Hirsch, the significance of the principle he formulates is that it provides ‘a strong (if defeasible) presumption in favour of common sense’ (Hirsch 2011, 99). Hirsch sometimes echoes Davidson (2001, 197) in saying that interpretative charity is ‘constitutive’ of language and meaning (e.g. Hirsch [2011], 230; Hirsch [2013], 438), but this further claim plays no role in his defence of common sense ontology.

One way in which an assertion might be reasonable is if it expresses a reasonable belief. And one way for a belief to be reasonable is for it to be held on the basis of testimony. So a revisionist might seek to reconcile their view with the principle of charity by maintaining that the ordinary beliefs which nihilism and universalism contradict are held on the basis of testimony. It is because of our education we believe that there are tables and that there is no object composed of the moon and Barack Obama’s left foot. We were brought up to join a culture in which it goes without saying that such beliefs are true (cf. Merricks 2001, 74–5).

Hirsch is unimpressed. In his view, the response does not make the common sense beliefs reasonable in the sense he wants. He writes:

Charity, however, requires us to look for an interpretation of these assertions that assigns ‘good reasons’ to them in a more robust sense. (Hirsch 2011, 115, footnote 26)

Unfortunately, Hirsch is not forthcoming about what this ‘more robust sense’ of ‘good reasons’ is. Nor is he forthcoming about why the reasons charity requires must have this up-market quality. In the absence of such explanations, the worry arises that Hirsch’s response is a case of special pleading. Moreover, why should the reasons for believing things which we get from education and our culture be regarded as second-rate, as ipso facto less than the best? In recent years there has been a growing recognition among philosophers that the epistemic credentials of testimony can be, and often are, every bit as good as those of other doxastic sources (Coady 1992; Lipton 1998). Moreover, for a chain of testimony to transmit justification does not require that the originators of the chain should be able to provide a non-testimonial form of justification for their beliefs. As Jennifer Lackey has convincingly argued, testimony does not always transmit knowledge: it sometimes generates knowledge (Lackey 1999).

In requiring ‘good reasons’ in ‘a more robust sense’, Hirsch takes a demanding view of what the principle of charity requires. We regard this view as ad hoc and implausible.

Hirsch writes:

The fact is, I think, that revisionists standardly delude themselves into thinking that they can plausibly explain why people make the mistakes they allege. It is closer to the truth to say that such mistakes would require some highly selective and seemingly arbitrary forms of genetically or culturally transmitted idiocy. (Hirsch 2011, 113–114)

Whatever the merits of the testimony-based explanation offered above, it has nothing to do with idiocy. There is no connection between common sense ontology and learning difficulties. This testimony-based explanation is not, of course, the whole story about why the folk came to believe that there are composita. There remains the question of why the chain of testimony was initiated and why it has persisted. We will address these questions in Section 5.

We suspect that Hirsch’s above point is to do with obviousness: he thinks that the truth of common sense ontology is obvious and so thinks that his opponents deny that people are able to appreciate these trivialities. (If they have to concede that we are able to appreciate many trivialities, then that might make the inability ‘highly selective’ and – in the absence of further argument – ‘seemingly arbitrary’.) Now something might be a priori obvious or obvious on the basis of perception. We will address the former option in the next section, the latter in Sections 4 and 5.

For the rest of the paper, we will focus on three specific aspects of charity which Hirsch distinguishes:

Charity to understanding: ‘the presumption that members of the linguistic community generally understand what they are talking about to the extent at least that they do not make a priori (conceptual) mistakes about seemingly uncomplicated judgments’. (Hirsch 2011, 182)

Charity to perception: ‘the presumption that any language contains sentences used to make perceptual reports, and these reports are generally accurate (to a fair degree of approximation)’. (Hirsch 2011, 185)

Charity to retraction: ‘the presumption that reasonable people are expected to improve the accuracy of their judgments in the face of additional evidence’. (Hirsch 2011, 151; see also 180)

We think that revisionists can make a good case that charity to perception is defeated and that charity to retraction actually tells against Hirsch’s attempted vindication of common sense ontology. Before discussing those types of charity, let us discuss charity to understanding.

3. Charity to understanding

Hirsch invokes this type of charity because he takes his opponents to be only those philosophers who think that it is an a priori conceptual matter that common sense metaphysics is mistaken (Hirsch 2011, 99). Not all philosophers who think that common sense metaphysics is mistaken take that view, but to begin with let us just consider those who do. Does charity to understanding reveal these philosophers to be unacceptably uncharitable?

It will only do so if charity to understanding is a good constraint to place upon philosophical theories (or, if it is a good constraint, if the revisionists violate it to an unacceptable degree). But charity to understanding, as formulated by Hirsch, is a constraint we should reject. It is important to distinguish between genuinely uncomplicated judgments and seemingly uncomplicated ones. In place of Hirsch’s formulation, we claim that there is a presumption that members of the linguistic community generally understand what they are talking about to the extent at least that they do not make a priori (conceptual) mistakes about genuinely uncomplicated judgments. So we agree that there is a presumption that the folk do not make mistakes about trivial matters – matters about which they make genuinely uncomplicated judgements – but what is seemingly uncomplicated need not be genuinely uncomplicated. As John Horden (2014, Section 4) argues, whilst it may well be unreasonable to attribute to us trivial a priori mistakes, it is not always unreasonable to attribute non-trivial a priori mistakes. For instance, to take an example of Horden’s, before one has seen Cantor’s diagonal argument, one may reasonably believe that all infinite numbers are equal. One may also reasonably (if mistakenly) take this belief to be a trivial a priori judgement. Moreover, Hirsch cannot regard the issue of when composition occurs to be a trivial matter: the fact that his opponents offer arguments for their views implies that his opponents do not regard the issue as trivial. To build his objection to nihilism or universalism on the assumption that the issue is trivial would straightforwardly beg the question against his revisionist opponents. That is why charity to understanding, as formulated by Hirsch, does not provide him with a good case against revisionism. (If Hirsch’s view that the matter is trivial were supported by independent argument, matters would be different. But his argument that the issue is trivial is not independent of his charity-based argument against revisionary ontologies: see Hirsch [2011], 102.)

It might be replied, on Hirsch’s behalf, that when the folk understand the objection to taking all infinite numbers to be equal, they will straightforwardly retract their claim, whereas when the folk see the revisionists’ objections, they will become ambivalent and confused but not retract. Even if those claims are correct, however, we do not see that they have any particular argumentative force against revisionism. It is not as though the case for revisionism would cause a characteristic kind of response in the folk which other philosophical arguments fail to. The folk become ambivalent and confused when presented with any moderately sophisticated philosophical arguments, even arguments for conclusions that they already accept. Folk who believe that there is an external world or who believe that it is rational to think that the future will resemble the past are very often disquieted and confused on hearing Putnam’s argument that we are not brains in vats or the reliabilist case for the rationality of induction. In any case, even if the folk were to have such a reaction, the fact that they would is not good evidence that the argument in question is a bad one. Their reaction would not discriminate between good or bad philosophical arguments which are even moderately sophisticated.

So far, we have followed Horden in granting Hirsch his assumption that the question ‘In what circumstances does composition occur?’ is a conceptual question to be answered using a priori methods. While this is not the place to argue against Hirsch’s assumption, we will note that the assumption restricts the scope, and so the interest, of Hirsch’s defence against revisionism: it means that his arguments will not tell against those who do not regard metaphysics as an a priori conceptual discipline. Furthermore, even if Hirsch is correct in thinking that the philosophers who most concern him (Unger, Chisholm, van Inwagen, Lewis) do think that metaphysics is an a priori conceptual discipline, this overlooks the fact that it is open to someone to co-opt their arguments, deny the assumption that those arguments are a priori conceptual in nature, but still draw revisionary metaphysical conclusions from them, e.g. that there are no tables, that there is an object composed of the moon and Obama’s left foot, and so on. (Such a view is explored in Stalnaker [2001] and in Miller [2010], especially 966–968.) Since Hirsch takes his opponents to think that it is an a priori conceptual matter that common sense metaphysics is mistaken, he has nothing to say about these positions.³

For these reasons, we do not think that charity to understanding poses a genuine threat to revisionist ontology. We now turn to another of the types of charity Hirsch has distinguished: charity to perception.

4. Charity to perception (i)

Charity to perception provides prima facie support for common sense ontology. On the basis of what they (take themselves to) perceive, ordinary speakers will report that there is a tree before them, that there is no object composed of the moon and Obama’s left foot and so forth. Charity to perception gives us defeasible reason to assume that these reports are true or at least reasonable.

According to Hirsch, these reasons are undefeated. As we saw above, Hirsch explicitly denies that revisionists can explain why ordinary people are enmeshed in metaphysical error. He suggests that the only way for revisionists to explain these mistakes would be to ascribe to the folk ‘genetically or culturally transmitted idiocy’ (Hirsch 2011, 114).

At this point, Hirsch strays into hyperbole. Some forms of perceptual error may be hard-wired into us. Given optics and the mechanism of the eye, we are irredeemably prone to seeing straight sticks in water as being bent. Motor adaptation makes stationary things appear to move (see Mather, Verstraten, and Anstis [1998]). And so forth. Our susceptibility to these perceptual errors is not idiocy on our part. Of course, we do not believe these erroneous contents, but that is because, although perception presents us with such contents and we have some inclination to believe them, the understanding of the world which our scientific theories provide us with makes us judge such contents to be false. Related, if more controversial, examples are found in philosophy. The folk confidently judge that time passes, and there is a case for claiming that this is a perceptual judgement (Le Poidevin 2011). According to the tenseless theory of time, however, such judgements are false: time does not pass. But the theory is not committed to claiming that the folk are idiots. The theory can be consistently conjoined with the claim that time seems to pass because the folk need to change their tensed statements and beliefs in order to avoid error (Mellor 1998, 66). If anything like this explanation is correct, the presence of widespread and entrenched error on the folk’s behalf need not impute idiocy to the folk.

Here is a more pertinent example of potential perceptual error. Non-philosophers think that they live in a world that contains tables. If asked for their reason for thinking this, they would reply that their senses tell them that there are tables. But consider two scenarios. There is the common sense scenario in which there are tables. There is also the scenario envisaged by van Inwagen in which there are no tables but there are simples arranged tablewise. There is no perceivable difference between these scenarios, because simples arranged tablewise look just like tables (see Merricks [2001], 9; Daly and Liggins [2010], 225). So if the actual world is in fact as van Inwagen describes it to be, the folk would not be able to tell otherwise on the basis of their senses. And since the folk have not even considered van Inwagen’s scenario, they are not led to think that the world is not as they take their senses to tell them how it is. For these reasons, even if the world is as van Inwagen describes it to be, ordinary people would be in error but not be in a position to realize this. Similar accounts can be given for other theories in revisionary metaphysics about why ordinary people should be in persistent and systematic error. The fact that perception cannot discriminate between the scenarios described by common sense and the scenarios described by these theories, in addition to the fact that ordinary people do not entertain those theories, suffices to explain in a simple and general way why ordinary people would be enmeshed in metaphysical error.⁴

It might be replied that we are considering a situation that by Hirsch’s lights is impossible: namely, that there are atoms arranged tablewise but no tables. But, the reply continues, it is hard to know what to make of such counterpossibles. We have three comments to make about this reply.

First, for reasons given by Daniel Nolan (1997), we are not as pessimistic about the evaluation of counterpossibles and the prospects for assigning them non-trivial truth-values. Second, that aside, given the distinction between metaphysical and epistemic possibility, even if nihilism is metaphysically impossible, it can be fruitful to entertain various epistemic possibilities, including one in which nihilism is true, and ask whether perception would be able to distinguish between that situation and our actual situation. So, even if nihilism is metaphysically impossible, the thought experiment still shows that perception cannot distinguish between the situations in question, and it further shows that the objection that we can perceive that there are tables fails as an argument against nihilism. Third, in any case the charge that we are entertaining a counterpossible is question-begging. A situation in which there are atoms arranged tablewise but in which there are no tables is a counterpossible only if nihilism is necessarily false. But nihilism is necessarily false only if nihilism is false; and the truth or falsity of nihilism is at issue here. So it would not be justified to reject our thought experiment on the ground that the thought experiment describes a counterpossible. Whether it does is something that remains to be established. Now, nihilism faces the objection that we can perceive that there are tables. The thought experiment is being offered to defend nihilism against this objection. The thought experiment shows that perception does not discriminate between a situation in which we are perceiving tables and a situation in which there are no tables but in which we are perceiving atoms arranged tablewise. Consequently, the objection that we can perceive that there are tables begs the question against nihilism.

On the basis of these three considerations, then, we conclude that Hirsch’s use of charity from perception against nihilism fails.

Let us turn to Hirsch’s use of charity from perception against universalism. Hirsch stresses that some of the objects posited by universalists would be highly visible if they existed. One example concerns the alleged sum of the daytime parts of a tree and the night-time parts of its trunk. Such an entity has branches during the day, loses them at night, but has branches again the next day. Philosophers such as Lewis, who posit such entities, thereby posit ‘highly visible macroscopic objects’ (2011, 137). According to Hirsch (2011, 136–137) ordinary speakers make the perceptual judgement that there are no such objects in their environment; universalists maintain that ordinary speakers cannot be trusted to make accurate perceptual reports of highly visible objects that are right before their eyes, and in doing so, universalists violate charity to perception.

Horden (2014, 240) argues that Hirsch is wrong about these objects. Consciously echoing Lewis (1986, 213), he suggests that those fusions which universalists posit that are not countenanced by common sense are ‘simply not worth talking or thinking about outside philosophy’. This means that we have good practical reasons for ignoring them: it saves valuable thinking time. For Horden, our failure to notice these objects is an epistemic error that results from our practical rationality:⁵

thus universalists can explain ordinary speakers’ failure to notice strange fusions as a reasonable mistake. Epistemically it is a mistake; but there is a good practical reason for making it. And the practical reason explains the epistemic mistake, even though the mistake is unintentional. (Horden 2014, 240)

We do not find this response to Hirsch persuasive, because we do not find it plausible to think that if an object is of no practical value to someone, then that person fails to notice the object. Footballers notice footballs, but so do those with no interest in soccer; climbers notice mountains, but so do those who have no use for them. Moreover, it is easy to cook up a context where strange fusions, if they exist, are of practical relevance: just offer someone a prize to point out the largest wooden object in a bag of woodchips. We do not think that common sense would countenance such a thing as the fusion of the woodchips, no matter how large the prize. It is important to distinguish Lewis’s proposal that we restrict our quantifiers in order to ignore irrelevant entities from Horden’s more radical proposal that we fail to notice irrelevant entities.⁶ The latter is highly implausible. Consider a standard case of quantifier domain restriction: ‘There’s nothing in the fridge’, where the quantifiers are restricted to edible items. Participants in conversations where that restriction obtains are still able to notice the shelves of the fridge. We do not see why things should be any different with the case of practically irrelevant fusions.

We will now offer the universalist a partial response to Hirsch’s argument. Consider an ordinary rectangular sheet of white paper. It is common sense that we could cut out of it a piece of paper that is pentagonal and large enough to be easily visible. But when one is presented with an ordinary sheet of white paper, it seldom looks like a pentagon with a border. A pentagon that is large enough to be easily visible can be disguised by surrounding it with paper in such a way that the sum of the two is an ordinary sheet of paper. Once so disguised, the pentagon is no longer perceptually salient, and so an observer can be forgiven for not noticing it, even if the pentagon is right before their eyes. Now consider again the sum posited by Lewis. Much the same goes for this. Although the sum has no branches during the night, it is surrounded by branches in such a way as to look like a tree. Displayed in certain surroundings (that is, next to appropriate branches), the sum is no longer perceptually salient and its existence is not obvious on the basis of perception. (This also explains why one does not notice anything losing branches when night falls, or gaining them at sunrise.) Hirsch’s talk of ‘highly visible macroscopic objects’ is therefore misleading: one cannot simply open one’s eyes and notice all the visible macroscopic objects in one’s environment, because an object’s surroundings may prevent it from being perceptually salient. To explain our mistaken perceptual judgements, no appeal to ‘idiocy’ or selective blindness is needed: rather, the universalist need only point out that sometimes we fail to see the wood for the tree.

We have said that this is only a partial response to Hirsch’s argument because there are other examples to which this explanation does not apply. Consider, for instance, a tree–tree: an object composed of two trees which are in contact with each other. The tree–tree is unlike Hirsch’s trunk-tree: it is perceptually salient and there is clear boundary between it and its surrounding environment. The explanatory strategy used for the trunk–tree does not extend to cases such as this one. Universalists require a different strategy, but it is far from clear how it might run.

What should we infer from this? Recall that Hirsch thinks that charity to perception is a general constraint which revisionary theories of ontology fail to meet. We have found that his argument from charity to perception has some force against universalism but no force against nihilism. The lesson we draw from this is not that all forms of revisionism should be rejected, but instead that nihilism stands unrefuted. Whether universalism can ultimately escape the objection remains to be seen. So, at least at present, nihilism is the preferable revisionary ontology of material objects.

5. Charity to perception (ii)

Daniel Z. Korman has recently questioned whether our perceptual experiences can explain why anyone should have reasonably believed that there are composite objects (Korman 2009, 256–258). The revisionist explanation begins by saying that people have reasonably believed that there are composite objects simply because it looks them that there are such things. But what is the sense of ‘looks’ here? There is a phenomenal use of ‘looks’ (hereafter ‘looks_p’), whereby a dog can look_p the same to both a human and a cat. There is also an epistemic use of ‘looks’ (hereafter ‘looks_e’) whereby the dog looks_e to the human to be a show dog (because of how the dog looks_p together with various background beliefs that the human has) although the dog does not look_e to the cat to be a show dog. Suppose that revisionists claim that it looks_e to the folk that there is at least one composite object because it looks_p to them that there is at least one such object. Korman doubts whether ‘there being a single thing before one typically features into [sic] the way things look_p’ (Korman 2009, 257). He illustrates his claim with an imaginary example of van Inwagen’s: the so-called bliger (van Inwagen 1990, 104). We are to suppose that a bliger looks_p like a single animal, whereas close-up it is revealed to be several animals moving as a pack. Korman then argues as follows:

Having now learned the truth about bligers, I spot one from afar. It seems quite natural to say that things are exactly the way they appear to be even though there is another sense in which it still looks to me as though there is a single animal out there … If so, then it cannot be that there looks_p to be a single thing with such and such qualities, for in that case my experience would be nonveridical even at the most fundamental level … [I]t is quite plausible that, at least at some level, things look the same to me now as they did before I learned the truth about bligers. The envisaged eliminativist [i.e. revisionist] would have to insist either that, despite appearances, there is no level at which the bliger experience is veridical or else that, despite appearances, things cannot possibly look the same to me once I learn the truth about bligers. (257)

Korman’s objection can be met largely using the very distinction between two senses of ‘looks’ which he himself employs. Suppose we learn the truth about bligers and then see one at a distance. We agree that ‘things are exactly as they appear to be even though there is another sense in which it still looks to [us] that there is a single animal out there’. First, the way things are is that there are several animals. Knowing what we do about bligers, it also looks_e to us that there are several animals out there. Second, the sense in which it still looks to us as though there is a single animal out there is that it looks_p that there is a single animal. There is nothing untoward in the conjunction of these two claims. Take another example. Knowing what I do about the refraction of light through water, when I partly immerse a straight stick in water, it looks_e to me that there is a straight stick in water. I know something about the refraction of light and how it affects the perception of objects partly immersed in water. So when I partly immerse a straight stick in water, the stick looks to me – looks_e to me – as a straight stick partly immersed in water would look (cf. Austin 1962). It also looks_p to me that there is a bent stick in the water. Korman infers from the claim just quoted that ‘it cannot be that there looks_p to be a single thing with such and such qualities, for in that case my experience would be nonveridical even at the most fundamental level’. Yet, even among those who agree that there is such a thing as ‘the most fundamental level of experience’, it is a commonplace that one’s experience can be non-veridical even at that level. By looking at a white rose though a red filter, the rose looks red to you. That perceptual experience belongs to the most fundamental level of experience if anything does, yet it is not a veridical experience. The revisionists can thereby motivate their taking the first limb of the dilemma which Korman presents at the end of his objection, namely, that there is no level at which the bliger experience is veridical. Alternatively, if it is said that it is a fact that the rose looks_p red under the filter, and that your experience veridically represents that fact, then, by the same reckoning, the revisionist can say that it is a fact that, at a distance, the bliger looks_p like a single animal, and our experience veridically represents that fact.

Korman has a further objection: that ‘even if there does indeed look_p to be a single thing instantiating the qualities given in perception, it is difficult to see why the belief that there is a single thing continues to seem reasonable under reflection’ (Korman 2009, 257). Korman contrasts this situation with the case of the Penrose triangle, a figure where there looks_p to be three straight beams meeting one another at right angles.

But casual inspection of this figure suffices to show that things cannot possibly be as they look to be. No discursive reasoning or calculation is required; one just recognizes that it is impossible for things to be that way … So we cannot in general rely upon the way things look_p to explain the resilience of perceptual beliefs when the contents of those beliefs are a priori impossible, even if it can explain their initial reasonableness. (Korman 2009, 257)⁷

We offer the following reply to Korman’s objection. The falsity of a certain content may be more or less apparent, more or less easy to discover. This holds even in the case of those contents which are (arguably) a priori impossible. The a priori impossibility of the Penrose triangle is at the more apparent end of the spectrum. It is a pleasing feature of the Penrose triangle that casual inspection is enough to reveal that it is impossible. Yet it is not a defining feature of a priori impossible contents that their impossibility should spring out upon the most casual and fleeting inspection. Those revisionists who think that it is a priori impossible that any things compose something will locate this impossibility towards the far end of the spectrum, close (for example) to the content that every consistent axiomatization of number theory is decidable. To identify the a priori impossibility of such contents requires extended reflection. In the absence of such reflection – reflection which only a small minority of people take the time and effort to engage in – the reason which perceptual experience gives for supposing that there are composite objects is undefeated, and so the folks’ perceptual beliefs that composition occurs prove durable.

A final objection which Korman makes is that, by indicting how things look_p with having false content, revisionists are barred from running the strongest arguments for notable revisionist theses (e.g. that composition does not occur) because those arguments contain premises which are justified by how things look_p (Korman 2009, 258). More fully, Korman’s objection is as follows:

[Revisionists think that the way things look_p] fully accounts for the false folk belief that there is a single object before them when they see atoms arranged statuewise. By parity, the fact that there looks_p to be a single thing when one sees some atoms arranged statuewise should suffice to explain their conviction that the statue and the lump of clay are not distinct objects. Likewise, the fact that what shattered the window looked_p to be a single thing should suffice to explain their conviction that the shattering of the window was not overdetermined …

Now, Korman further claims, revisionists take the process of forming beliefs about things on the basis of how they look_p to be ‘wildly unreliable’. So, he concludes, they ‘should be disinclined to trust the premises of their arguments’.

In reply to Korman, we question his assumption that the revisionists’ arguments rely on the results of looking_p. Perhaps modal intuition or counterfactual reasoning, not looking_p, justifies revisionists in believing that statues are distinct from their clay. And perhaps parsimony considerations, not looking_p, justify revisionists in thinking that the shattering of a window is not causally overdetermined by a baseball and by its atoms. Korman’s assumption that revisionist rely on the results of looking_p is therefore suspect. By questioning it, revisionists can block Korman’s argument that their strategy is self-defeating.⁸

6. Charity to retraction

People often assert things only later to retract those assertions given new evidence and time to think. Charity to retraction says that we should not interpret people’s initial assertions as being true, but rather the assertions that they come to subsequent to further evidence and reflection: ‘… we must, other things being equal, favour an interpretation that makes the community’s retractions in the face of additional evidence come out right’ (Hirsch 2011, 152). By this reckoning, of course, an interpretation must, other things being equal, make the community’s original assertions come out false.

What implications does charity to retraction hold for ontological revisionism? Granted, ordinary speakers are willing to assert such sentences as ‘There are tables’ and ‘There is no object composed of the moon and Barack Obama’s left foot’. But, for the most part, they are in considerable ignorance of relevant evidence: the problems, arguments, distinctions and theories offered by revisionists. Nor, for the most part, have ordinary speakers reflected on these matters: they remain in a dogmatic slumber. Charity to retraction then does not enjoin us to interpret ordinary people’s assertions as true. Whether we should interpret them as true depends upon what they would be willing to assert following the input of further evidence and the outcome of reflection. Now, revisionists think that they have a strong case for their iconoclastic claims. It is open for them to say that, even if the folk would not be convinced of those claims even after becoming acquainted with the evidence, many of the folk would withdraw their willingness to assert the sentences they were previously happy to assert. Charity to retraction then tells against, not for, Hirsch’s case for common sense ontology.⁹

Hirsch addresses the above challenge only briefly and we confess we are not entirely sure how his response goes (2011, 182). The key point seems to be that different ontologists’ arguments lead to mutually inconsistent conclusions about what exists. According to Horden, most English speakers who have given the composition question sustained thought have come to reject many ordinary beliefs about composition. If so, then the key point of Hirsch’s response is mistaken (Horden 2014, 236).

Why does Horden think that most English speakers who have given the composition question sustained thought have come to reject many ordinary beliefs about composition? He argues that the only pieces of evidence we have on this matter are the views of those professional philosophers who work on ontology. Thus, the argument moves from the conclusions of the folk to the conclusions of ontologists.

But is it true that most ontologists support revisionism? Even if a majority of the philosophers who have published papers in this area are revisionists, that does not mean that the majority of philosophers who have reflected on this area have that view. Perhaps the conservatives had little to say: for instance, perhaps they are largely swayed by Moorean arguments (about which we will say more below).

Horden’s response is sociological in nature and we have argued that the evidence is not enough to show that most English speakers who have given the composition question sustained thought have come to reject many ordinary beliefs about composition. We will now suggest an alternative response to Hirsch – one which is sociology-free. The response is that, whether or not there is consensus among revisionists, it would be question-begging for Hirsch to assume that the revisionists’ arguments would not sway people to retract their commonsense assertions – even if those arguments would leave them unclear what claims about composition they should endorse in place of common sense. All the argument above relies on is that revisionists’ arguments would sway people to retract their commonsense assertions.

In addition to the passage we have just been discussing, Hirsch makes a number of claims which bear on charity to retraction. To begin with, the above line of response assumes that philosophical argument should, at least among sufficiently rational people, suffice to overturn pre-philosophical beliefs. Hirsch doubts this and thinks that there is something contrived about philosophers’ allegations about cases in which the bar of argument has been met. What he says is worth quoting in full:

When I consider the writings of some of the most prominent deep ontologists of recent years – Chisholm, Lewis, van Inwagen, just to mention three – I can rattle off many cases in which they veto a commensensical judgment in behalf of a philosophical argument, but I would be hard pressed to recall an example in which the reverse happens. My impression is that, in matters of ontology, virtually any theoretical problem, however marginal or flimsy, if it cannot be adequately answered, suffices, by the lights of these philosophers, to trump the most deeply entrenched beliefs of common sense … Given any well entrenched ontological judgment of common sense (about highly visible physical objects), I could not imagine giving it up for the sake of some philosophical argument. If I had nothing more to say about the argument I would simply repeat Moore’s famous point that the force of the common sense judgment shows there must be something wrong with the argument (even if I don’t know what it is). (Hirsch 2011, 91)

We have three comments to make about this passage. First, Hirsch mischaracterizes at least Lewis’s approach. For instance, given a clash between common sense judgements that some things are of value and philosophical theory that says that nothing is of value, Lewis retains the common sense judgements by saying that what philosophy has shown is only that value is not quite what common sense thought (Lewis 2000, 93–94). He has similar things to say about the clash between common sense judgements of simultaneity and physical theory (93–94).

This takes us to the second, and deeper, point. Hirsch’s evaluation of what he calls ‘deep ontologists’ suffers from a sampling error. In considering potential clashes between common sense judgements and philosophical theory, there are three different kinds of case to consider:

• (1) Cases where there is no clash between common sense judgement and philosophical theory.

• (2) Cases where there is a clash between common sense judgement and philosophical theory, and philosophical theory loses out.

• (3) Cases where there is a clash between common sense judgement and philosophical theory, and common sense loses out.

Cases of kind (1) do not call attention to themselves. They may occur frequently, or they may occur infrequently, in a given ontologist’s work, but, when they do occur, they mostly go unremarked. For the most part the fact that your ontological theory does not clash with some common sense judgement is not particularly significant – unless, perhaps, the rival theories clash with that judgement and so there is something distinctive and virtuous about your theory.

Cases of kind (2) are infrequent for a different reason. Philosophers police their own thinking. Suppose you are trying to come up with a theory for some area of metaphysics. Suppose you devise a certain theory but then realise that it clashes with common sense and that common sense wins out. What will do you? You will bin the philosophical theory and go back to the drawing board. Your erstwhile theory will not have made it into print and been publicized. Perhaps you did not spot the clash but someone else did – a listener at a talk, a referee for a journal. Again your theory does not make it into print and again it is not publicized. It might be that philosophers’ strike rate of coming up with theories that lose out to common sense is particularly high. The fact is that their mechanism of self-criticism, peer review and the like will prevent this record coming to light. Philosophical theories that clash with commonsensical thinking but which lack the resources to meet that challenge are not disseminated, just as scientific theories that clash with commonsensical thinking but which lack the resources to meet that clash are not disseminated.

Finally, we have cases of kind (3). These are the cases which Hirsch exclusively considers. But without data about the relative frequency with which cases of each of these three kinds occur in the reflections of revisionists such as Chisholm, Lewis and van Inwagen, Hirsch has no grounds for claiming that these philosophers stack the deck against common sense.

Our third comment on the passage from Hirsch concerns his appeal to Moore. Faced with a clash between philosophical argument and common sense, Moore retained his common sense judgement and rejected the philosophical argument, even if he was unable to diagnose where the argument went wrong. Hirsch takes the same tack. But it is open to everyone to reject any argument whose conclusion they find uncongenial. The harder task is to say, in any given case, why it is more rational to reject the conclusion than to accept it (see Unger [1979], 125 and Daly and Liggins [2010], 225–226). In the passage cited Hirsch merely reports that he is unable to imagine himself accepting the uncongenial conclusions drawn by revisionists. Elsewhere, Hirsch offers objections to specific revisionist arguments. But either those objections both address all of the revisionists’ arguments and are good objections – in which case Hirsch’s Moorean argument is redundant – or they either do not address all of the revisionists’ arguments or some of them are not good objections – in which case Hirsch has to fall back on his Moorean argument and the above criticisms recur.

Furthermore, his appeal to the Moorean move is at odds with his overall argumentative strategy. In his book, Hirsch offers two defences of common sense ontology. The first is Moorean: when a common sense belief contradicts a ‘fancy philosophical argument’ (2011, 98), the former carries greater weight. Rather than the common sense belief being false, it is more likely that the argument involves a fallacy or a false premiss.¹⁰ Hirsch recognizes that this Moorean move fails to convince his opponents, and so he offers a second argument, based on interpretative charity, in the hope that this argument will be ‘harder for [revisionists] to ignore’ (2011, 98). Yet if the defence of the argument from charity turns on an appeal to the Moorean move, as it apparently does in the passage from Hirsch quoted above, then the argument from charity is no improvement on the Moorean move.

In discussing charity to retraction, Hirsch (2011, 159) issues an important clarification:

As Lewis remarks, a stage seems eventually to be reached in ontology when ‘all is said and done, ’ when ‘all the tricky arguments and distinctions and counterexamples have been discovered, ’ so that each position has achieved a state of ‘equilibrium.’¹¹ I’m assuming that in the ontological disputes under discussion the ‘all is said and done’ stage has been reached. [Hence] … charity to retraction plays no role … (Hirsch 2011, 159)

Notice that if we were really at the ‘all is said and done’ stage, then Hirsch would not need to make the Moorean move or launch the charity argument. If we had reached that stage, Hirsch would know what is wrong with each philosophical argument against common sense ontology, and so he would not need to fall back on these alternative vindications of common sense. Rather, for each philosophical argument against a common sense judgement, Hirsch would be able to offer the diagnosis of what is wrong with that argument.

Like the Moorean tack, the above passage has too swift a way of dealing with the arguments of Hirsch’s opponents. It simply declares that they have already been taken into account and that they are not decisive: ‘I’m assuming that in the ontological disputes under discussion the “all is said and done” stage has been reached’. This is reminiscent of Senator George Aiken’s supposed recommendation about the Vietnam War: the U.S. should declare victory and then go home. If only it were so easy. Now, Hirsch can assume what he likes. But that, of course, only raises the question whether his assumption is true. Hirsch’s opponents, the revisionists, will not think that it is. They will deny that we are anywhere near that tranquil stage. You’ve only to look at the raging debates in metaphysics in the most recent issues of your favourite journal to realize that.

Even Lewis cannot be taken to have signed up to Hirsch’s assumption. Lewis thinks that a stage can be reached in philosophy in which ‘all is said and done’ and all of the positions are in equilibrium. But, contrary to the impression Hirsch might have given, when he made those remarks Lewis was not writing specifically about ontology, but philosophy in general. Moreover, Lewis does not claim that each position available at the outset achieves equilibrium at the end of the dispute. He says only that each position at the end of the dispute is in equilibrium. It is consistent with this that certain positions available at the outset of the debate have been found during the course of the debate to be untenable. Hirsch is not entitled to assume that his favoured position, common sense ontology, is one of the positions that reaches equilibrium. To do that would require meeting the objections and arguments of revisionists. Perhaps he can. What has gone by the board, however, is his having any single general reply to his opponents’ use of charity to retraction to block the charity argument. It might be wondered whether it would be possible for us to be in a position in which we see all the objections and counterexamples but remain unsure which of the premises in some argument are the false ones. But if we were in such a state of uncertainty we would not have finished with arguing. Far from it! If we were in that state, we would be lacking some objections or counterexamples – namely, ones which identify which of the premises in the argument are false. Not all would have been said and done and philosophers would still have their work cut out for them.

In his most recent paper, Hirsch modifies his position. He now claims that charity to retraction can be discounted before all is said and done:

It is sufficient that all is almost said and done. That condition is reached when a sufficient number of tricky arguments and distinctions and counterexamples have been discovered, so that each side has reached a stage of equilibrium in the sense of being committed to holding on to the core sentences definitive of its position even in the face of some additional problem. (Hirsch 2016, Section I)

Hirsch claims that we have in fact reached this stage with disputes over the ontology of material objects, ¹² which considerations of charity therefore resolve in favour of common sense ontology. It seems ‘clear’ to Hirsch that he will be able to deal with any problem case or counterargument which his opponents can throw at him without having to retract his common sense position (2016, Section I).

We do not know what justifies this confidence. The justification cannot be the charity argument itself, on pain of circularity. But what then makes Hirsch so sure that revisionists will never be able to refute his common sense ontological position? We worry that Hirsch has under-estimated the ingenuity of his opponents and lapsed into dogmatism.¹³

7. Conclusion

By taking the composition debate as a case study, we have argued that ontological revisionists need not violate plausible principles of charity if they are correctly applied. Charity to retraction actually tells in favour of revisionism about composition. Where composition is concerned, considerations of charity present no obstacle to ontological revisionism.

Hirsch repeatedly stresses (2011, xiii, 100, 126, 184) that the scope of his argument is limited: his intention is to defend the truth of common sense claims provided they are taken to be part of ordinary English. He has no objection to philosophers claiming ‘There are no tables’ if this is to be understood not in its ordinary sense but in a special technical sense, although he requires them to explain what this sense is (2011, 184). This raises the question of the relation between the language of ontology and ordinary English. When ontologists make the claim ‘There are no tables’, are they using the words in their ordinary sense or not? In this paper, we have defended revisionists’ right to deny common-sense claims understood in their ordinary English senses. We leave to another paper (Daly and Liggins forthcoming) the question of whether revisionists ever actually do so.

Funding

This work was supported by the Arts and Humanities Research Council (UK) under [grant number AH/I022465/1], The Foundations of Ontology.