1. Introduction
How does the idea of a merely verbal dispute stand from a broadly Quinean perspective? Let such a perspective include W.V.O. Quine’s general scepticism about the distinction between questions of fact and questions of language (as exemplified in the analytic/synthetic distinction), without necessarily including Quine’s views on other topics.Footnote 1 It seems clear, prima facie, that from such a perspective, the distinction between verbal and factual disputes is not in good standing at all. Yet scepticism about verbal disputes also seems harder to maintain than scepticism about analyticity, for the idea of verbal disputes has a rich life outside of philosophy, and we are often, even in ordinary circumstances, confronted with disagreements that seem quite clearly verbal.
In this paper, I have two aims. The first is to show that Quinean scepticism about analyticity does, in fact, generalize to the idea of verbal disputes as well. Although I am sympathetic to this Quinean scepticism, I will not defend it in detail; rather, my aim is to take it as a starting point and see where it leads. So, my conclusion should be understood as a conditional claim, namely that if one is already sceptical of the analytic/synthetic distinction for certain kinds of reasons, then one should also be sceptical of the verbal/factual disputes distinction for the same kinds of reasons.Footnote 2 This reveals additional commitments that this scepticism regarding analyticity involves. My second aim is to explain more carefully what such scepticism consists in, or what attitude a Quinean should ultimately adopt towards the idea of verbal disputes. This explication is useful because, first, it addresses the concern raised above by reconciling this scepticism with our ordinary ability to recognize verbal disputes and, second, because discussions of the broadly Quinean scepticism of notions like analyticity often leave it unclear exactly what that scepticism consists in. The case of verbal disputes will provide a helpful case study for clarifying the Quinean position more generally.
I will proceed by first clarifying the notion of verbal disputes that is in question. Following that, I will present an objection to the idea of verbal disputes that should be compelling to those who already accept a familiar objection to the idea of analytic truths. Finally, I will consider a variety of attitudes that these objections might inspire, and consider the advantages of some over others, leaving us with a clearer picture of how one who accepts these objections should understand the ideas being objected to. This result is particularly important for recent attempts to employ the verbal/factual distinction in recent debates in metametaphysics.
2. Verbal Disputes
The distinction between verbal and factual (or non-verbal) disputes has a long history. It can be traced back at least to Plato.Footnote 3 David Hume held that the distinction played an important role in understanding the intractability of various philosophical disagreements:
From this circumstance alone, that a controversy has been long kept on foot, and remains still undecided, we may presume, that there is some ambiguity in the expression, and that the disputants affix different ideas to the terms employed in the controversy.Footnote 4
Hume’s discussion captures the underlying idea of what makes a dispute verbal: if a dispute is merely verbal, then the disputants mean something different by some term in the disputed sentences without realizing it. That condition seems to account for both the apparent defectiveness of verbal disputes along with the fact that the defectiveness is somehow a matter of language. This idea has been spelled out more carefully by David Chalmers:
A dispute over S is (broadly) verbal when, for some expression T in S, the parties disagree about the meaning of T, and the dispute over S arises wholly in virtue of this disagreement regarding T.Footnote 5
Although there may be problems with taking a difference in meaning (or in beliefs about meaning) as a sufficient condition for verbality, it is plausible that it is a necessary condition.Footnote 6 Various other philosophers who have recently discussed verbal disputes (such as by Eli Hirsch and David Manley) are also committed to this necessary condition, either explicitly or because their characterizations entail it.
This conception of verbal disputes will be my focus in what follows. In particular, I will argue that when characterized by the necessary condition above, the verbal/factual distinction is open to objections that mirror some objections traditionally given against the analytic/synthetic distinction.
3. The Problem with Verbal Disputes
The analytic/synthetic distinction (ASD) is the distinction between sentences true purely in virtue of their meaning, and those true partly in virtue of meaning and partly in virtue of the way the world is.Footnote 7 For example, ‘All bachelors are unmarried,’ is often regarded as a paradigmatic analytic sentence, while ‘All bachelors eat pancakes,’ would count as synthetic. A number of arguments against the ASD can be found in the literature, but typically they share the same spirit, inspired primarily by Quine. I will present one particularly plausible version of these arguments, drawing on material from a variety of other philosophers (including Keith Donnellan, Hilary Putnam, Jerry Fodor, and Edward Shirley). The argument goes as follows. Sometimes we change our attitude toward a sentence: at one time we accept it and then later we reject it. If the sentence in question is analytic (before the change), then we must regard the change in attitude as due to a change in the meaning of the sentence: if the sentence is true in virtue of meaning, then we can come to regard it as false only if we no longer take it to have the same meaning.Footnote 8 On the other hand, if the sentence is synthetic, then such a change could reflect either a change in meaning or a change in our beliefs about the subject matter.
The problem, so the objection goes, is that there is an inescapable element of arbitrariness in whether some change in attitude should be regarded as reflecting a change in meaning or a change in belief. That is because in most cases we could interpret it either way: generally when we change our attitude toward a sentence, that change could be accounted for by appealing to changes in our beliefs, without appealing to changes in meaning. Furthermore, in most cases the reasons we have for changing our attitude in the first place do not provide principled reasons for preferring one interpretation over the other. So it’s often unsettled whether a change in attitude reflects a change in meaning rather than a change in belief.
Examples can be found throughout the history of science. It is plausible that at some time ‘Water is an element’ was regarded as analytic, but due to the development of modern chemistry, that sentence is now regarded as false. However, there seems to be no principled reason to think that such a change reflects a difference in the meaning of ‘water’ or ‘element,’ especially since the reasons for the change were the discovery that water consists of complex molecules—an object-level, empirical fact of the sort that is generally a reason to change our beliefs about water, rather than the meaning of ‘water’. That’s not to say that we could not explain this change in terms of a change in meaning, but it shows that there is nothing about the case in question that forces us to do so. Other examples include sentences like ‘Atoms are indivisible,’ ‘I can hear you only if you are within earshot,’ ‘Whales are fish,’ and ‘One cannot reach the place from which one came by traveling away from it in a straight line and continuing to move in a constant sense’.Footnote 9 All of these may have been regarded as analytic at some point and all of them are now regarded as false. According to defenders of the ASD, the meanings of these sentences must have changed. However, contrary to that view, there are no principled reasons for thinking that there was a change in meaning rather than a change in belief in these cases.
The verbal/factual disputes distinction (VFD) is open to a similar objection. According to those who accept the VFD, we can distinguish between disputes that arise from differences in what the disputants mean by their terms and disputes that arise from a difference in their factual beliefs (that is, their beliefs that are not about the meanings of their terms). Notice, however, that this distinction is simply a synchronic, interpersonal version of the diachronic, intrapersonal distinction between changes in meaning and changes in belief. In the case of analyticity, we considered a single person changing her attitudes towards a sentence over time. In the case of verbality, we have two people with conflicting attitudes towards a sentence at the same time, and the distinction that the defender of verbality wants to draw is essentially the same as the one we criticized above, transposed into this new setting.
Hence the same objection applies. These disputes can generally be construed as resting on underlying disagreements about the facts, rather than divergences in meaning. Generally there is no independent reason to think that they mean something different rather than just hold different beliefs.
For instance, a case where two people agree about Jane’s height in inches but disagree about whether she is tall might be thought to be a paradigmatic verbal dispute (hinging on discrepancies in their interpretation of the word ‘tall’), but even it can be construed as factual. Perhaps both participants have a property in mind, tallness, and disagree about its extension. The reasons they offer in defence of their positions might be of the form ‘Bill is definitely tall, and he is shorter than Jane, so she must be tall’; namely, factual claims about the extension of tallness (not sociolinguistic claims about the use of ‘tall’)—precisely the kinds of things our beliefs about a property’s extension should be sensitive to. Why not, then, regard it as a factual disagreement about the extension of a property?Footnote 10
A philosophical example with a similar upshot can be found in certain disagreements over personal identity. In some disputes over whether, for instance, a certain entity is a person, or whether two temporally distinct person-like entities are stages of the same person, the participants agree about many of the underlying physical (and perhaps metaphysical) facts, like when and where the entities exist, what they are composed of, their causal history, and so on. Nonetheless, the disputants still disagree about whether they are persons, or the same person. In such a case, their agreement on the underlying facts generates some temptation to construe the dispute as merely verbal, resting on a difference in the disputants’ usage of ‘person’.Footnote 11 However, there is another way of construing the dispute, according to which they both have some property, personhood, in mind, and they disagree about whether certain entities instantiate that property. Clearly they disagree about its extension and perhaps the conditions necessary for possessing the property, but that needn’t prevent them from having the same property in mind and then having a factual disagreement about it (otherwise, factual disagreement will be much less common than we ordinarily suppose).
The upshot of these considerations is that those who accept a standard objection to the ASD should also object to the VFD. Even setting aside particular examples, it is clear that the two distinctions depend on a general distinction between changes or conflicts at the level of factual beliefs and changes or conflicts at the linguistic level. Since those of a broadly Quinean view are sceptical of this general distinction, for the kinds of reasons given, they should be sceptical of the VFD as well as the ASD.
4. How to be Sceptical about Verbal Disputes
But what, exactly, do these considerations show? First, it is not immediately obvious what scepticism about verbal disputes even consists in. Second, it’s difficult to see how such scepticism can be maintained, given the role that verbal disputes play in ordinary contexts, where it often happens that we come to regard some disagreement as resting on discrepancies in our linguistic usage.
To answer this question, it is useful to begin by noticing that, even regarding the ASD, there are various different conclusions one might draw from the standard sceptical arguments and that these conclusions are not always clearly distinguished (perhaps for good reason, since it’s not totally clear what follows from the problem I’ve described). So, it is worthwhile to start by articulating some of the different conclusions that might follow. This will lead us to a better understanding of what a plausible form of scepticism regarding the VFD looks like.
I’ll begin by mentioning two possible views just to set them aside, before discussing some others that are more plausible. The first is that there simply is no ASD and no VFD, perhaps because the distinctions are unintelligible. This is, perhaps, the view that Quine himself held regarding the ASD. However, it is plausible that this view is too strong. As Grice and Strawson point out, “the existence of statements … about which it is pointless to press the question whether they are analytic or synthetic does not entail the non-existence of statements which are clearly classifiable in one or other of these ways,” and there is still non-collusive agreement about whether certain sentences are analytic, and certain disputes verbal, in some straightforward cases. Many of us can find examples of disagreements from our own experience that strike us as clearly verbal. A familiar example from the philosophical literature is given by William James’ case of a person walking around a tree while a squirrel crawls around the trunk of the tree in order to remain unseen by the person. The tree always remains between the person and the squirrel. A dispute over whether the person is going around the squirrel could easily be based on confusion over two different senses of ‘around’—the person does go around in the sense that her path encloses the squirrel, but the person does not go around in the sense that she never moves behind the squirrel. As the arguments in the previous section suggest, it is difficult to give principled reasons for thinking that this dispute really is verbal, if one already accepts certain views about meaning. Nonetheless, our judgments in these and other ordinary cases are not always based on carefully considered reasons. Instead, one might simply have a gut feeling that the word ‘around’ is being used differently; more generally, some disputes simply strike one as verbal. These reactions put pressure on the idea that the notion of verbal disputes is simply unintelligible (just as Grice and Strawson argue that complete rejection of the ASD is rendered implausible by clear examples of analytic truths). If the claim that verbality is unintelligible were the only conclusion to draw from the previous arguments, one might simply bite the bullet and think that these case judgments are unjustified. There is, however, an alternative route where the sceptic can agree that our judgments about whether some dispute is verbal (or some statement analytic) are not entirely misguided but think instead that they do not reveal an explanatorily interesting or philosophically useful feature of those disputes. We will consider this alternative more carefully below.
The second view I want to set aside is that there really is an ASD and a VFD, but they can only be drawn on the basis of empirical evidence from linguistics, psychology, and other sciences. On this view it is objectively settled, for each sentence, whether that sentence is analytic or synthetic (for instance), but whether it is so can only be known on the basis of empirical inquiry. Such a view would avoid the complete rejection of the distinctions considered above: they are real distinctions, but they lack certain features traditionally associated with them by philosophers. For example, on this view whether some sentence is analytic is not reflectively accessible to speakers just in virtue of their linguistic competence. Hence this ASD would be unsuitable as an explanation of the possibility of a priori knowledge, a role that traditional empiricists often thought the ASD could play.Footnote 12 Similarly, applying this view to the VFD, one could hold that in each dispute, there is a fact of the matter concerning whether the disputants are merely talking past one another or disagreeing about the facts, but that ordinary linguistic competence is incapable of determining which is the case. Only when supplied with the right kinds of empirical evidence, informed by the proper theories in linguistics and psychology, can one settle the question. Though similar to the sort of view that I will defend below, I don’t think it is a plausible view for the sceptic. I think the problem with the distinctions in question goes deeper. In particular, this view suggests that the proper form of scepticism regarding the ASD and VFD is mere epistemological scepticism: that the evidence we have in ordinary circumstances is insufficient to justify us in thinking that a particular sentence is analytic or a dispute verbal (but that through the proper kind of inquiry, we could discover evidence that would lead to justified beliefs about these matters). In contrast, the problem of the previous section does not seem merely epistemological, or to depend merely on our lack of empirical evidence. Rather, the arguments point to a problem with the very notions of analyticity and verbality; in particular, they point to an intrinsic and ineliminable arbitrariness in the distinctions themselves. This suggests that the problem is not merely the meagreness of our evidence when applying the ASD and VFD, but something having to do with the distinctions themselves, and the nature of our thought and language more generally.
So, rather than reject the distinctions entirely, or accept an epistemological scepticism regarding them, we should look for a different kind of view. In light of this, I think the best sceptical view is one along the lines of that defended by Hilary Putnam.Footnote 13 The view, in a nutshell, is that there may well be an ASD and a VFD, but they lack the kind of depth, clarity and rigor needed to perform the kinds of philosophical tasks that their proponents hoped they could perform (I will elaborate on this in a moment).
There are at least three ways to flesh this view out in more detail, each of which is fairly plausible (I remain neutral on which is ultimately preferable, though I find the first the least plausible and the last the most plausible). The first is that although there is an ASD and a VFD, as a matter of fact there are no, or almost no, analytic sentences or verbal disputes. The distinctions are intelligible, in principle, but fail to mark any real distinction between actual cases, because, as a matter of fact, sentences are (almost) always synthetic and disputes (almost) always factual. In this case, I will call the distinctions empty, like the distinction between flying horses and non-flying horses. A second possibility is that the distinctions are vague. Although we can make sense of them, and although there are instances on both sides, there is a large set of cases where it is indefinite how they apply. Thus although there might be some disputes that are clearly verbal, there will be many borderline cases where the dispute is not clearly verbal and not clearly factual. In that case, the distinctions are like that between red and orange. A third possibility is that the distinctions are arbitrary, by which I mean that (focusing on VFD), outside of some clear cases at the extremes, disputes can be understood as either factual or verbal and it is objectively unsettled which construal is correct. In other words, they can be equally well construed as factual or verbal.Footnote 14 An example of this sort of arbitrariness is whether a sentence like ‘The referent of ‘Halifax’ is among the set of things picked out by the predicate ‘is a city’’ is about language or about the world. Given the equivalence between that sentence and ‘Halifax is a city,’ both characterizations are defensible and neither is solely correct.Footnote 15
These three views all lead to the conclusion that there is a distinction, but it cannot play certain roles that philosophers hoped it could. This naturally inspires the question: exactly what work are these distinctions unable to do? Set aside the ASD in order to focus on the VFD, and recall the quotation from Hume above. Hume seems to hold that various features of philosophical disputes—such as their intractability—can be explained by the fact that they are merely verbal. This view naturally gives rise to the further claim that if we had a principled and rigorous way of determining which disputes are verbal and which factual, we would have a tool that could guide our decisions about which disputes to take seriously and which to disregard.Footnote 16
So, following these two suggestions, there are at least two philosophical applications of the VFD, and both are undermined by the kind of scepticism we are considering. One of these two applications is methodological: we might want a practical guide to which disputes are worth taking seriously, one that will inform our decisions about which disputes to engage in. The VFD is sometimes held to provide a methodological tool that can help us make those decisions. The second application is explanatory: we might want a constitutive account of what makes disputes worthwhile or defective. The VFD is sometimes held to provide such an account. I will discuss both of these in turn, and explain why scepticism about VFD poses a challenge for them.
Methodological applications of a distinction impose epistemological requirements on our ability to tell whether and how the distinction applies to the objects being sorted. If we are to use the distinction between As and Bs to determine whether certain objects are F or G, we must be better, or more reliable, at determining whether the objects in question are A or B than we are at determining whether they are F or G. Otherwise, the A/B distinction would hinder rather than help us, since we could simply check whether the thing we’re investigating is F or G directly. For instance, if we want to use the distinction between brown and yellow bananas as a guide to which bananas we should eat, we need to be better at determining whether bananas are brown or yellow than we are at telling which bananas are worth eating and which are not, independently of their colour. Otherwise the distinction will not, in practice, help guide our decisions about which to eat.
One upshot of the scepticism we are considering is that we are not very reliable at judging when a dispute is verbal, even when it’s clear that something isn’t quite right with the dispute in question. How does the scepticism undermine our reliability? The thought is that, in general, our beliefs about whether some dispute is verbal or factual are based largely on how naturally we can characterize the dispute in the formal or material mode, either mentioning or using the key terms of the dispute (e.g., characterizing the dispute as one about whether ‘tall’ applies to Jane vs. about whether Jane is tall). The prima facie plausible thought is that these characterizations correlate with whether the dispute actually is verbal or factual, making us reliable judges. However, that thought is undermined by the objection we’ve discussed. Even if we are strongly inclined to characterize some dispute in the formal mode, as with the tallness case or the personal identity case, that does not rule out the opposite characterization and leaves open the possibility that the dispute is factual. Hence we are unreliable judges of whether a dispute is verbal or factual, so the VFD should not be relied on to do methodological work.Footnote 17 From the sceptical perspective, it may well be that we are strongly inclined to regard some dispute as verbal, but that will not be because we have a clear, precise and unproblematic conception of the VFD which we apply to the case in question. In many cases, it will not be obvious why we regard it as verbal, while in others, it will be because we are antecedently confident that the dispute is seriously defective (in which case the VFD is not a guide to the defectiveness of certain disputes, but rather a response to it).
For explanatory applications, the crucial point is that a distinction is only as explanatory as the basis on which it is drawn, because the distinction will inherit the features (including any emptiness, vagueness, or arbitrariness) of that basis. It’s not enough that the extension of the distinction is relatively clear or that there is widespread agreement about certain cases. It must also be that the distinction is a suitable one for the purposes at hand, or that it matches the relevant features of whatever it is meant to explain. An example might make this clearer: suppose we want to explain why bees are more attracted to some flowers than to others. The flowers can be roughly categorized into red and orange, with a lot of borderline cases. If the bees’ behaviour exhibits a similar kind of vagueness (i.e., most of the bees visit the clearly red flowers but some also visit those closer to orange, with their numbers trailing off somewhere in the middle), then we could explain the bees’ preferences in terms of the distinction between red and orange flowers. If, however, there is a sharp cut off between the flowers that the bees visit and those they don’t, one that is not mirrored in the distinction between red and orange flowers, then that distinction is not a suitable explanation. A more suitable explanation will match the precision in the bees’ behaviour (it might still involve vague notions like colours, including red and orange, but they will need to be supplemented or rendered more precise in some way). Vagueness, however, is not the only feature that is relevant; there are other ways that an explanans must match its explanandum. For instance, if a distinction is arbitrary, in the sense of being objectively unsettled or left open by the facts in a large number of cases, then it is unsuited as an explanation of a less arbitrary difference between the relevant objects. So, by showing that the VFD is empty, vague, or arbitrary, the objection shows that the VFD is unsuitable as an explanation for the difference between defective disputes and worthwhile disputes.Footnote 18 Again, the sceptic in this case does not doubt that there are some verbal disputes, but she will doubt that being verbal provides an informative explanation for the defectiveness of certain disputes, due to the problems that we’ve considered for that distinction.
To draw these points together, recall that the VFD is the distinction between disputes that are due to the participants meaning different things by their terms and those that are due to their believing different things about the facts. For the distinction to be methodologically and explanatorily useful, we must be able to reliably track whether disputes rest on a difference in meaning (compared to what we’re using it as a guide for) and it must be non-arbitrary whether disputes rest on a difference in meaning (compared to what we’re trying to explain). The objection shows that neither is the case.
I will close with an analogy that should help illustrate the view I’m recommending for the Quinean sceptic about the VFD. Psychologists have observed that there is widespread agreement in how people apply the distinction between bright and dark to coughs and sneezes (when asked, people generally classify coughs as dark and sneezes as bright).Footnote 19 This widespread agreement suggests that there is some bright/dark distinction that applies to coughs and sneezes, just as agreement over (some) cases might support the claim that there is an ASD and a VFD (as Grice and Strawson argue). However, the agreement does not show that the bright/dark distinction can do certain kinds of work. In particular, the distinction would not be a useful guide, in practice, to whether some event is a cough or a sneeze, at least for most people, since most of us are much better at determining whether something is a cough or a sneeze by other means. It would be difficult to first figure out whether something is bright or dark (in the relevant sense), and then decide whether it is a cough or a sneeze on that basis. Similarly, the distinction does not provide a good explanation of what the difference between coughs and sneezes consists in, for even if it is a relevant difference, it is not robust enough: even if there is agreement over how to apply the distinction, there is still less agreement concerning that than there is concerning whether the events in question are coughs or sneezes. The distinction might be useful for other purposes (it might tell us something interesting about the psychological basis of metaphor, for instance, or about our understanding and use of language more generally), but not for these.Footnote 20 The sceptic should think the same of the VFD: there may well be a distinction, and it may well be important for some purposes, but it cannot do the methodological or explanatory work that philosophers have hoped it can do.
5. Conclusion
Quineans who are suspicious of the analytic/synthetic distinction should be equally suspicious of the verbal/factual distinction. This scepticism about verbal disputes is, on the face of it, more problematic, because there seem to be clear cases in ordinary contexts of verbal disputes. So, rather than a blanket rejection of the verbal/factual distinction, Quineans should adopt a more nuanced view, according to which there is such a distinction, but it is empty, vague or arbitrary. The result is that the distinction is unsuited to perform certain philosophical tasks. We can, perhaps, recognize verbal disputes in straightforward cases, but the distinction is too problematic to be relied upon for rigorous methodological or explanatory purposes.
Acknowledgements:
Many thanks to Susanna Rinard, Stephen Yablo, Agustin Rayo, Robert Stalnaker, Matti Eklund, Stephen Maitzen, Mahrad Almotahari, several anonymous reviewers, Meghan DeSouza, and Steve Pelletier.
