1. Introduction: From Expressivism to Neopragmatism

Expressivism about a limited area of discourse is—roughly put—the view that we should understand the basic assertions in that area as expressions of attitudes, rather than in terms of states of affairs that they represent.Footnote ¹ It is, therefore, opposed to what can be called representationalism about that area of discourse. Quasi-realism adds to expressivism the goal of vindicating all the forms of talk—within the relevant discourse—that might have seemed to motivate a representationalist view. These forms include talk of truth, belief, propositions, reference, and so on. So one significant motivation for quasi-realism is that it promises to allow us to respect our naturalistic scruples even when talking about such things as numbers and values: things which have seemed hard to place in a purely naturalistic world.

Neopragmatism is what quasi-realism becomes when it is generalized in two ways. The first form of generalization is a widening of its target. That is, neopragmatism includes the idea that representational notions play no role at all in explaining how language works, even when that language concerns such paradigmatically representationalist-friendly things as rocks and trees. Quasi-realism, on the other hand, contrasts its target discourses with discourses that it takes to be well explained in representational terms. The second way in which neopragmatism is a generalization of quasi-realism is that the expressivism underlying quasi-realism is replaced with a more general use-theoretic view (Gert, Reference Gert and Gert2023; Macarthur & Price, Reference Macarthur, Price and Misak2007). That is, neopragmatists think of assertions, most generally, as moves in a social practice, rather than as expressions of attitudes. Of course, some moves might aptly be described as expressions of attitudes. But, the neopragmatist will stress, there are plenty of other sorts of moves that—like a move in a square dance—one can make quite independently of any particular attitude.

Neopragmatism, as characterized so far, can be developed in a number of ways. The focus of this paper is on two of them. The first I will call normative neopragmatism; the second, naturalistic neopragmatism. Naturalistic neopragmatism, well represented in the work of Huw Price, is a project that is meant to fit unproblematically into the framework of the sciences.Footnote ² It might not count as paradigmatic natural science, since it need not be taken as aiming to give explanations that are historically accurate. It is enough for many such neopragmatists that the genealogies they offer are live possibilities. That is enough to make it clear that there is no real pressure to reach beyond the world of the sciences in order to make sense of our talk about numbers, or normative reasons, or whatever else has been generating what they take to be metaphysical extravagances and misguided disputes. So neopragmatist genealogies are often more proof-of-concept than proof.Footnote ³ But even so, they are not merely proof-of-concept; they also often show that it is no surprise at all that we have the sort of talk we do. They do this by showing how the relevant vocabulary serves the interests of (at least) some significant portion of the linguistic community, and how the use of that vocabulary might be passed from generation to generation. If this showing is done well, our sense that there are metaphysical mysteries connected with the vocabulary should dissipate.

In contrast to naturalistic neopragmatism, normative neopragmatism, represented centrally by Robert Brandom, describes linguistic practices as games of giving and asking for normative reasons. These reasons then either commit us or entitle us to certain further moves in the game. Now, it is true that normative neopragmatists will deny that there is any reduction of reasons, commitments, or entitlements to some non-normative basis. But in this respect, they do not differ from their naturalistic counterparts. A metaphysical reduction of normative properties (or numerical ones, or mental ones, or chromatic ones…) is not the aim of any sort of neopragmatist. Rather, the important difference between the two forms of neopragmatism concerns how to explain talk of reasons, commitments, and entitlements. The normative neopragmatist’s explanation of such talk—in contrast to the naturalistic neopragmatist’s—itself appeals to various reasons, commitments, and entitlements.

To sum up: both normative and naturalistic neopragmatists explain our linguistic practices, but the explanations offered by the former use normative notions while those offered by the latter appeal only to notions that are employed in the sciences. This might seem to put them at odds. But in fact the notion of explanation is sufficiently plastic that the two projects are, in principle, perfectly compatible. The explanations at the center of Price’s neopragmatism are, as I have noted, genealogical. They are linguistic-evolutionary stories that make our own use of this or that vocabulary unmysterious by explaining how a practice that makes use of it might have emerged over time. Normative neopragmatists, on the other hand, offer what we might call ahistorical explanations of a practice: descriptions of what speakers must be doing—currently—in order to count as participating in a certain linguistic practice.Footnote ⁴ There is no inconsistency in holding that ahistorical explanations of linguistic practices must be couched in normative terms, while genealogical explanations need not be.

It is not entirely clear how Brandom and Price themselves take their projects to be related. Price often seems to want to co-opt Brandom’s ahistorical inferentialism for certain purposes, and Brandom simply does not say very much about the sorts of genealogies that Price offers.Footnote ⁵ Still, there are some normative neopragmatists who clearly think that naturalistic neopragmatism is doomed—that it cannot succeed even on its own terms—and that the only viable form of neopragmatism must take some normative notions on at the ground level. These include David Macarthur (Reference Macarthur2014, Reference Macarthur2019), Lionel Shapiro (Reference Shapiro2014, Reference Shapiro, De Caro and Macarthur2022), and Brandon Beasley (Reference Beasley2020). I should be understood to be referring to such theorists when I use the label “normative neopragmatist.” I will discuss their arguments below, in order to show that they miss their mark.

To be clear: normative neopragmatists certainly regard themselves as naturalistic in an important sense. But naturalism can take many forms, and the sort of naturalism advocated by normative neopragmatists is a strong form of what is called liberal naturalism. There is not a great deal of consensus as to what, precisely, liberal naturalism is. Still, here is Matteo Morganti’s (Reference Morganti2022, 246) relatively uncontroversial description of it.

Some liberal naturalists will accept reasons and meanings as natural-but-not-scientific. Scientific naturalists, on the other hand, accept an ontology that is based on causal efficacy. The importance of causal powers explains why, for scientific naturalists, “the ontological status of mathematical and abstract entities is somewhat of an embarrassment, and so generally avoided or quietly passed over” (De Caro & David Macarthur, Reference De Caro, Macarthur, De Caro and Macarthur2010, 4). The liberal naturalist will admit that science is one excellent way to learn about the natural world: about atoms, acids, animals, and asteroids. But the doing of science—not to mention other ways of “getting to know what the world is like”—will require appeal to properties that go beyond the causally efficacious entities and properties studied by science. This does not mean that the liberal naturalist accepts supernaturalism. Supernaturalism is the view that certain special entities (God, perhaps, or fate) positively violate the laws of nature. Such a view is rejected by liberal naturalists just as strongly as it is by scientific naturalists.

To illustrate liberal naturalism, consider, as Mario De Caro and Alberto Voltolini (Reference De Caro, Voltolini, De Caro and Macarthur2010) do, modal properties. These are not part of the ontology of the scientific naturalist: the sciences do not study the nature of necessity or possibility, or indeed the nature of scientific law, despite trying to determine what the true scientific laws are. Still, and as Price points out, the actual doing of science is shot through with modal notions: we cannot do science without them. So most liberal naturalists will accept modal properties as just as real as causally efficacious properties and allow them to play their characteristic roles in the doing of science. And liberal naturalists will also accept other entities and properties as perfectly natural, even if they do not play a role in the doing of science. In the words of David Macarthur (Reference Macarthur2015, 574), “Liberal Naturalism makes available the humanly vital conceptual space of everyday life and language.” Macarthur’s version of liberal naturalism will therefore include meanings, rules, art, and so on, as natural-but-not-scientific—and therefore as available for naturalistic-but-not-scientific explanations.

Price has no worries about the use of modal notions in scientific explanations, and he counts as a liberal naturalist. But there are stronger and weaker forms of liberal naturalism. The sort that Price thinks his genealogies depend on is sufficiently weak that it would not be misleading to say that he takes them to be scientific (indeed, he often describes his genealogical project as anthropology, or even biology) (Price, Reference Price2008, 91). This does not mean that he is, overall, a weak liberal naturalist. My point is only that his linguistic genealogies—the bedrock of his neopragmatism—do not, in his view, depend on anything stronger. That is consistent with the acceptance of a very strong form of liberal naturalism as acceptable—indeed, as required—when explaining, say, the influence of art on culture, or the ways that volitional disabilities impact moral responsibility. But those are simply not the focus of Price’s own theorizing.

The form of liberal naturalism involved in normative neopragmatism is much stronger than the form that naturalistic neopragmatism depends on; it includes normative notions such as rules, reasons, correctness, and content. One way of framing the question at the heart of this paper, then, is this: If we push Price’s genealogical theorizing as far as we can in service of demystifying the sort of talk that has seemed to call for metaphysics, what sort of liberal naturalism will we actually be required to rely on? Certainly, it will include modal notions in its naturalistic explanations. But will it need to go so far as to include normative notions?Footnote ⁶ My answer to this last question will be “I think not. And it has, at any rate, not been demonstrated.”

As I will stress later, the naturalistic neopragmatist may be able to agree with the all the positive views of the normative neopragmatist—just as both sorts of neopragmatist can agree with all the positive views of self-described non-naturalist realists. But just as neopragmatists of both sorts think that there is a lot of demystifying theorizing to be done after making claims that involve the existence of various non-natural entities, naturalistic neopragmatists think that there is useful naturalistic theorizing to be done after making the claim that any genuinely linguistic utterance is a move in a game of giving and asking for normative reasons. In particular, naturalistic neopragmatists think that we can offer explanations of such games, and the reasons they involve that (i) does not make basic use of any primitive normative terms, such as “reason,” “commitment,” or “correctness” and (ii) does not leave us with any mysteries that cannot be addressed by the natural or social sciences. Macarthur, Shapiro, and Beasley think this cannot be done.

The structure of the paper is as follows. Section 2 makes the naturalistic neopragmatist’s goals clearer. Section 3 lays out, in extremely general terms, one reason it would be nice if the naturalistic neopragmatist’s goal could be achieved: even the complete success of the normative neopragmatist’s project will leave some important philosophical topics in relative obscurity. Section 3 also sets up a challenge to naturalistic neopragmatism that is then developed in more detail in Section 4. That challenge pushes for the inclusion of normative notions as essential items in any neopragmatist’s theoretical toolkit, and it is the point of the section to resist this. Section 5 makes room for our normative assessments of linguistic utterances by taking those assessments to be one important target of naturalistic neopragmatist theorizing. If that theorizing is successful, we should not have any naturalistic worries about our use of such assessing language when we then go on to do some normative neopragmatist theorizing. Section 6 builds on Section 5 by emphasizing how normative and naturalistic neopragmatism can be perfectly compatible, though normative neopragmatists will start further down the explanatory road than their naturalistic cousins. Section 7 concludes with some remarks about the ways in which—despite my insistence that the two neopragmatist projects are compatible—the arguments of the paper nevertheless favor the naturalistic version as more aptly taken to be a version of pragmatism.

2. The Naturalistic Neopragmatist’s Goals

What precisely is it that the naturalistic neopragmatist is trying to explain? One way to get clearer on this question is to distinguish, as Huw Price (Reference Price2011, 185–187) does, between what he calls “object naturalism” and what he calls “subject naturalism.” Object naturalists try to reduce metaphysically problematic entities or properties to naturalistic ones. That is, they defend the claim that those entities or properties are nothing over and above certain naturalistic entities and properties that are arranged in certain ways—otherwise they are fictions. Naturalistic neopragmatists, on the other hand, are subject naturalists. They do not address metaphysical questions in a reductive, head-on way. That is, they do not take the interesting questions to be “What are numbers—if they really exist?” or “What are values—if they really exist?” Rather, their central explanatory targets are, for example, our use of number words and our assessment of much of that talk as true or false. The reason the position is called “subject naturalism” is that its primary focus is on us—human subjects who use language—and our linguistic behavior.

A typical naturalistic neopragmatist’s question might be “What explains why we use number words as we do?” This question is as naturalistic as the evolutionary biologist’s question “What explains why white-crowned sparrows sing the songs they do?” Brandom seems to misunderstand Price’s goals, conflating explanation with vindication. Here is what Brandom (Reference Brandom and Price2013, 86) thinks Price is telling us to do:

But this description is not right, and for two reasons. First, Brandom puts a normative term (“correctly”) into Price’s theoretical vocabulary.Footnote ⁷ Of course, Price will not deny that we can call certain uses of a particular target vocabulary—say the language of probability—“correct,” and other uses “incorrect.” But as part of a subject-naturalist project, the way to explain this is to explain the use of the word “correct,” not to come up with normative standards that apply to the target vocabulary. Rather, the ultimate goal is to explain use: both of a given target vocabulary, and—as part of a distinct project—of the normative vocabulary we sometimes use in connection with that target vocabulary. And a second reason Brandom’s description is not right is that it does not even mention the notion of the genealogy of a practice. Rather, it focuses on current practice: on the limited task of characterizing the particular skill one imparts to language-learners when one instructs them in the use of a particular bit of vocabulary. But it seems to me that Price’s most significant contributions to neopragmatist theorizing are his genealogical explanations. It is these that do the demystifying work.

There is a very real chance that by leaving genealogy out of the picture, we will be forced to use normative notions to give anything like an informative picture of the particular linguistic practice that is the focus of our theorizing. How would this happen? Well, suppose we grant that to explain the meaning of a term, one must specify how to use it. What could this possibly mean, from an ahistorical perspective, other than how to use it correctly? And if we are thinking about assertions in particular, it is hard to see how such a specification of correct use could avoid amounting to a set of normative claims about what an assertion that uses that term entitles or requires one to infer, and from what other claims one might permissibly infer it, and so on. The content of a claim would then be a matter of its location in an inferential web—a node in the space of reasons—which would mean, to many, that it is irreducible to anything non-normative. And indeed that is precisely how things stand from the normative neopragmatist perspective.

I’ve just explained how an ahistorical take on language use might lead to a normative project, even from a subject naturalist perspective. But there is another project one might pursue. In the words of Wilfrid Sellars—a contributor to both normative and naturalistic neopragmatism—even though the content of an Ought claim (or any other claim about the manifest image) is irreducible to anything non-normative, we can nevertheless claim that Ought is causally reducible to Is. By this, he meant that “one can give a causal explanation of the history of moral agents without making ethical assertions” (Sellars, Reference Sellars and Sicha1980, 188, my emphasis). I read this as the claim that one way in which we might give a non-normative explanation of moral talk (and other sorts of talk that have given rise to placement problems) is to show how the linguistic practices that allow us to engage in that sort of talk are the result of historical processes of linguistic evolution. Theorizing that focuses on this historical fact could be part of an interesting and naturalistic form of neopragmatism.

James O’Shea discusses Sellars’ two-sided view, on which our talk about elements of the manifest image is in one sense reducible but in another sense not. But though he cites Sellars’ claim about the causal reducibility of Ought to Is, he does not draw attention to the word “history” that I myself have highlighted. Rather, O’Shea (Reference O’Shea and DeVries2009, 197) describes ethical discourse, in Sellars’ view, as follows:

As a result, for Sellars’ version of naturalistic neopragmatism to succeed, at least as O’Shea (Reference O’Shea and DeVries2009, 199, my emphasis) puts it, it would need to

But this ahistorical way of putting things can lead to a misinterpretation of Sellars’ project, and Price’s, by raising worries about the place of rules and rule-following in nature. In fact, it is plausibly the source of the challenge I will now present.

3. The Challenge

We are now in a position to address the challenge to naturalistic neopragmatism that is the focus of this paper. The worry is that in seeking to explain our linguistic practices, the would-be naturalistic neopragmatist cannot avoid an appeal to such things as propositional content or a notion of correctness or to normative reasons. As Sellars himself (Reference Sellars1954, 212) put it—though he thought he had a response—“Must we not admit […] that in describing a language game, we must not only mention its elements, positions and moves, but must also mention what its expressions mean?” Against an affirmative answer to Sellars’ question, the view I will defend is that Price’s genealogical project requires only a weakly liberal naturalism (which includes, e.g., modal notions). Normative neopragmatists, on the other hand, argue that it must make use of a comparatively stronger form (which also includes normative notions).Footnote ⁹ If I am right, two further conclusions follow: (i) naturalistic neopragmatism has more explanatory power than normative neopragmatism and (ii) the explanations offered by the two forms of neopragmatism are likely to be perfectly consistent with each other. Both (i) and (ii) are consequences of the fact that naturalistic neopragmatists will offer explanations of some of the normative neopragmatist’s starting points (the ones that characterize a stronger liberal naturalism). Naturalistic neopragmatism, that is, allows a weakly liberal naturalist to engage in normative neopragmatism with a clear conscience.

According to the challenge that I am introducing in this section, it is only a strong liberal naturalism that has the resources to provide adequate explanations of our linguistic practices. For naturalistic neopragmatists, including Price and the present author, this would be very disappointing. Normativity is one of the most central targets of philosophical theorizing. There would be something deeply unsatisfactory in appealing to normativity as part of one’s theoretical bedrock, while at the same time claiming to defend a methodology, rooted in a pragmatic view of language, that is meant to put metaphysical worries to rest.

My worry about the challenge will not be that the normative neopragmatist is begging any questions, or is involved in any sort of vicious circularity, or incoherence. It is perfectly possible for a neopragmatist to offer a non-representational account of, say, our use of normative vocabulary—“reason,” “correct,” and “incorrect,” for example—that makes use of that very vocabulary. There need be nothing viciously circular about this. After all, even a naturalistic neopragmatist will make use of modal vocabulary—in the form of causal vocabulary—in explaining the emergence of modal vocabulary. Of course, that explanation will not appeal to the idea of causally tracking modal facts, but it will still make use of modal vocabulary in other ways. That is why, in answer to the question “Are there causes?” Price’s response is “Yes, but they are not as much part of the furniture as we might have thought” (Price, Reference Price, Galavotti, Dieks, Gonzalez, Hartmann, Uebel and Weber2014, 601).

More generally, there is no way to explain anything without making use of some vocabulary. Any given explanation of our use of some specific bit of language—if we pursue it far enough—will inevitably bottom out in some vocabulary that is not explained in more basic terms. But the further we can get in such explanations before this happens, the more we have explained. My complaint is, therefore, that normative neopragmatists, by insisting that we must use normative notions at the ground floor, significantly and unnecessarily limit the amount they can explain. Another way of putting this is that normative neopragmatists fail to appreciate that while theorizing in science simply cannot avoid modality, it is reasonable to hope that it can avoid relying on normative claims. And, as I will argue, there is more hope for seeing Price’s project as science than normative naturalists have thought.

Suppose—just to fix ideas—that the most basic normative term is “ought,” and that a normative neopragmatist’s non-representational account of ought-claims is cast (partly) in terms of the expression of a specific form of approval. Even given that supposition, such neopragmatists will not be explaining normative notions simply in terms of approval. They will not be saying, for example, simply that ought-claims do in fact typically express the relevant form of approval; that is the sort of thing a naturalistic neopragmatist might say. Rather, they will be characterizing ought-claims in language that itself includes “ought.” For example, they will explain what one is doing with the claim “One ought to floss one’s teeth” in terms of the inferences we ought (and ought not) make, both from that claim and to it. That might not be so bad, if we had a good explanation of what we, as theorists, were doing, when we said that we ought to infer one claim from another. But the normative neopragmatists do not have any such explanation, except one that itself uses the term “ought.” If “ought” is genuinely primitive, that is, it is genuinely primitive. Using it in an inferentialist account does not render it less primitive, even if it does allow us to say some not-trivial things about its own use. An initial understanding of “ought” is presupposed in the account.

To end this section, it is perhaps worth noting that both normative and naturalistic neopragmatism are making fully general claims: either all the relevant explanations of our use of a vocabulary—even the vocabulary of the sciences and of everyday objects—will end up having to make use of normative notions, or none will. So there is an obvious middle position: some will and some will not. I agree that some incarnations of neopragmatism might occupy it. One might even think that this should be the default view. After all, once we reject the representationalist picture of how all truth-apt language works, what we are left with is a functional pluralism of assertive practices. Different vocabularies allow us to do very different things. Why, then, think the explanations of all vocabularies will bottom out in the same place—whether that be in the language of the natural sciences, or the language of reasons, commitments, and entitlements? This is an important question. I will offer, and contest, the normative neopragmatist’s answer to it in the next section and offer my reasons for optimism about the naturalistic neopragmatist’s answer.

4. Making the Challenge Precise: We Need Content If We Want an Explanatory Target

Let me now address three extant versions of the objection to naturalistic neopragmatism: those offered by David Macarthur, Lionel Shapiro, and Brandon Beasley. First, consider an analogy between an explanation of the evolution of eyeballs and an explanation of the evolution of a particular linguistic practice—an analogy to which a naturalistic neopragmatist might well appeal. Shapiro’s objection, in essence, targets this analogy. That is, it stresses an important difference between eyeballs and linguistic practices. In particular, to explain linguistic practices one needs—according to Shapiro—to talk about meaningful words, not just sounds. And talk of meaningful words takes us out of the realm of the weak liberal naturalism that would suffice if we were explaining the evolution of eyeballs. Here is how Shapiro (Reference Shapiro2014, 504) puts it:

Macarthur expresses a very similar concern, in terms of meaning, rather than content:

And Beasley puts the point in very similar ways:

I am not persuaded by these complaints. In order for us to demystify talk of causation, it is simply not necessary that we initially identify the relevant uses of “cause” in such a way as to include all and only those uses that express the content cause. Rather, all we need is a general fix on the relevant patterns of use, based on rough-and-ready judgments that particular uses belong to the same pattern. Armed with this general fix, we have enough to go on in seeking the sort of genealogical explanation that would make it unmysterious that a pattern of use of that sort would evolve. It is true that, given our mastery of words like “content” and “meaning,” we can use those words to focus attention on relevant instances of the use of the vocabulary we are interested in: say, causal vocabulary. We can say, that is, “These seven sample sentences here are using “cause” with the meaning I am after.” But such claims play their role in the context of discovery, not the context of explanation. Naturalistic neopragmatists, once they begin their explanatory project, need not talk about meaning or content.

To make the above point clearer, think about how medical scientists use suggestive patterns of symptoms to get an initial fix on the maladies that are the objects of their study. Such symptoms play their role in the context of discovery, not in the context of explanation. Once they begin the explanatory project of determining what is going wrong in the bodies of (many or most) of the people who exhibit the suggestive pattern, they need not pay much further attention to those symptoms. Certainly, they need not hold that in order to have the relevant malady, one will need to have the symptoms, or that if one has the symptoms, one has the underlying malady. In the same way, we can use our own developed sense about sameness of meaning to focus our attention on instances of the use of a certain vocabulary. Like medical symptoms, this sense of ours helps pick out token instances. Still, as in the medical case, the usefulness of the collection of token instances is that it can suggest hypotheses about the origins of linguistic practices. But in putting forward those hypotheses, and defending them, no mention need be made of meaning or content.

I think the arguments of Shapiro, Macarthur, and Beasley have a common source, at which I have already gestured in my remarks about O’Shea’s interpretation of Sellars. They mistakenly take Price’s subject naturalist strategy to be an ahistorical one. Here, for example, is Beasley (Reference Beasley2020, 3):

From Beasley’s point of view—and Shapiro’s and Macarthur’s—Price’s explanations of linguistic practices are in an important way like explanations of games. To produce an explanation of an actual game, one would observe people playing it and try to come up with a description of what they were doing. Thought of in this sort of way, it seems pretty clear that the “doing” will be explained as following certain rules—not simply exhibiting certain regularities. One will, for example, not mention even commonly observed moves that players actually make, if one has determined that they are errors. But this sort of normatively couched ahistorical presentation of the rules of a game is simply not a genealogy of the sort Price offers. This may be why Beasley (Reference Beasley2020, 2–19) puts Price’s use the word “genealogy” in scare quotes and uses it as a label for expressivist-functionalist accounts. But they are distinct. An expressivist-functional account is an ahistorical description of a going practice. A genealogy is a historically framed explanation of how there came to be a practice for which that account is apt. In Price’s own words,

Beasley (Reference Beasley2020, 12) sees “Price’s pragmatist philosophical anthropology as the project of constructing subject naturalist pragmatic metavocabularies for our most puzzling concepts.” But again, constructing a metavocabulary is an ahistorical matter and is therefore not what a genealogy aims to provide. I agree that if we cut ourselves off from the past, and say nothing about the possible historical processes through which we might have come to use certain words, it may be that we cannot avoid the use of normative vocabulary in giving useful descriptions of a linguistic practice. But we also, therefore, simply turn our back on certain demystifying projects—including demystifying any normative terms that might occur in such a pragmatic metavocabulary.

Now, it is true that Beasley describes Price’s project as having two levels. At one level—the inferential level—assertions that make use of a particular vocabulary will all be characterized in terms of a corresponding inferential structure, often described as a particular “game of giving and asking for reasons.” At another level—the pragmatic level—each vocabulary also gets kind of functional explanation, not in terms of rules of inference, but in terms of the overall purpose served by the vocabulary. But note: both such levels are ahistorical ones. The result of this, and in Beasley’s words, is that.

But this makes it sound as if we should be able to derive the appropriateness conditions from the functions we determine that the vocabulary serves. And this is the source of Beasley’s complaint: he thinks that naturalistic neopragmatists do not have a sufficiently rich vocabulary in which to tell their explanatory stories. Why? Because “no vocabulary so constrained is expressively powerful enough to […] target the conceptual activity it is supposed to explain” (Beasley, Reference Beasley2020, 13).

Contrary what Beasley supposes, the goal of the naturalistic neopragmatist is not to yield the appropriateness conditions for a vocabulary via genealogy. It is only to make it unsurprising that our use of the vocabulary takes the form it does, so that we no longer find such talk mysterious, or worry about where “in the world” to locate the objects or properties we use that vocabulary to talk about. That is enough to undermine traditional metaphysical worries about those objects and properties. Beasley repeatedly claims that Price “thinks naturalistic practical stances determine appropriateness conditions for assertions” (2020, 16).Footnote ¹⁰ But the anti-metaphysical upshot of naturalistic neopragmatism does not require this. Indeed, it does not even require that there be determinate appropriateness conditions, rather than a comparatively heterogeneous set of practices with the same word (thought of as a sound or mark), that end up working out.Footnote ¹¹ These are all reasons for optimism about the naturalistic neopragmatist’s project—even when it is thought of in the fully general way mentioned at the end of Section 3. The “patterns of use” that the naturalistic neopragmatist is aiming to render unmysterious are simply not meant to draw a line between correct and incorrect uses. Where does correctness come in, then? Is not language a rule-governed practice? Let us now turn to this issue.

5. Making Room for Normative Assessments of Use

We can and do assess the inferential moves people make with any given vocabulary and call some of them “correct” and others “incorrect.” This idea is part of what stands behind the normative neopragmatist’s conviction that we must identify the targets of subject naturalistic theorizing in normative terms. But if the naturalistic neopragmatist is right, we should try to understand talk of the correctness of, say, modal or moral claims at least as much by considering our use of “correct” as by considering our use of the modal or moral terms themselves. The naturalistic neopragmatist’s hope is that our theorizing about the use of “correct” will not require us to use the term “correct” (or some other normative notion)—otherwise we cannot claim to have elucidated normative notions without presupposing an understanding of normative notions in those very elucidations. I think that hope can be vindicated.

From an empirical point of view, we can say that a certain vocabulary is in fact used in a certain way (or finite set of ways). That empirical claim could be turned into a bit of conceptual analysis by changing the factual claim into a normative claim or a claim about rules of use. But, if our goal is to demystify language, as opposed to arguing about how language ought to be used, we need not make that change, and we can include in our description many uses that we ourselves would regard as incorrect. And that goes just as well for the words “correct” and “incorrect.” Beasley seems to miss the possibility that the subject naturalist is—among other things—trying to provide a genealogical explanation of our use of the very sort of normative language in which Brandom couches his inferentialist descriptions of linguistic practices. But this genealogical project is, as I have been stressing, quite different from any such inferentialist description. Neglect of this distinction explains why Beasley’s argument is so straightforward. He notes that as a matter of definition, “narrowly naturalistic vocabularies cannot use any normative or intentional vocabulary.” He then points out that inferentialist descriptions of assertive practices—which are the only descriptions that are sufficient to capture the content of the assertions they involve—make ineliminable use of talk of appropriateness conditions in “the full-blown normative sense of ‘appropriate’ of the game of giving and asking for reasons” (Beasley, Reference Beasley2020, 16–17). Given the implausibility of the idea that we can explain, in naturalistic terms, precisely what someone must be doing in order to count as following a certain rule for the use of a term, it is no surprise that he sees Price’s project as doomed.

One source of Beasley’s problem is that he identifies genealogies with the construction of pragmatic metavocabularies (2020, 14). And he takes a pragmatic metavocabulary for a concept to be a very specific thing: a vocabulary, V_m, that is sufficient to specify practices or abilities that are themselves sufficient to deploy the vocabulary, V_o, which expresses that concept.Footnote ¹² That is, as I have been stressing, an ahistorical project. But Price himself, though he sees a great deal of value in this sort of inferentialism, holds that “[t]he most interesting part of the project is then to explain how there come to be statements with particular contents” (2013, 41, emphasis in the original).Footnote ¹³ That is, the most interesting thing—what, in my view, does the really important demystifying, anti-metaphysical work—is the genealogy.

One part of a genealogy of a linguistic practice will advert to the fact is that it is in fact taught by one generation to the next. Such teaching involves assessments—but not on the part of the theorist. Rather, language teachers, who already have dispositions to use words in certain ways, also have dispositions to sanction learners when those learners use words in ways that diverge from their own. These dispositions involve using words like “correct” and “incorrect.” Once we see this, we should see our own normative assessments of various rules for use—or token uses—as expressions of this very same disposition, not as part of our genealogical theorizing. The correctness of those rules or token uses need not, that is, be taken to be part of what we are aiming to explain with our genealogies. Nor need we make use of any such normative assessments as part of our theoretical project.

My point here is that in explaining the use of a particular word, the naturalistic neopragmatist need not be seen as offering any normative claims. Rather, the initial stage of theorizing involves giving a description of the way that word is in fact used. This description of use includes not only a specification of when the word gets uttered but also how other speakers respond to its use on those occasions. If such a description is overly specific, one would be right to suspect that something normative is being smuggled in, since many speakers will deviate from that specific use, and the implicit suggestion will be that they are not using it correctly. But if the use is described in a way that is wide enough to acknowledge a range of more (or less) subtle differences, that can still shed considerable light on what is going on with claims that involve that word. Moreover, such a description of patterns of use is far from the end of the neopragmatist’s story. A crucial part of the project of demystifying our use of the target vocabulary is the production of a plausible naturalistic explanation of the evolution of a practice of using it in the way described. Interestingly, that explanation itself might make the resulting pattern clearer.

Some readers will still be tempted to complain that many of the uses that fall within the bounds of the wide description will be incorrect. But—again—that should not be taken as a criticism of the theorizing the neopragmatist is doing. Indeed, the neopragmatist who has offered the wide description is likely to agree with those readers in many cases. But to do this is not to introduce any normative claims into the theorizing. Rather, it is to use some vocabulary (“correct,” “incorrect”) that is also, itself, open to neopragmatist theorizing. That is, to shed light on what is going on when we find some of the uses of the target vocabulary to be incorrect, the neopragmatist will do some—descriptive—genealogical theorizing about our use of the words “correct” and “incorrect.” Even here, though, the description of our use of such words will not be given in normative terms. Moreover, that description too will be wide, so that some uses of “incorrect” that fall under its umbrella will themselves be assessed by many people as incorrect. But that is no problem: such a second-order use of “incorrect” will itself be susceptible to the very same genealogical explanation. To invert one of Brandom’s phrases, our explanations will be non-normative all the way down.

Given the lack of determinate rules in the descriptions of use that the neopragmatist offers—so that those descriptions will be wide enough to encompass quite a range of more specific patterns exemplified by distinct speakers—another worry may arise at this point: that the naturalistic neopragmatist cannot avoid saying that we are all constantly talking past one another. If people are using the target vocabulary in very different ways—even including “correct” and “incorrect”—are we still talking about the same things or using the same concepts? Surely, we need some common referent (for referring-words), or some shared attitude (for words that are used to express attitudes), or—more generally—some shared rules (whatever those sorts of things might be) if disagreements in assertions using the relevant words are going to count as genuine. This is the specification problem. Footnote ¹⁴

The mistake in the above framing of the specification problem is the assumption that shared referents, attitudes, or rules are required in order to underwrite the linguistic behavior characteristic of taking ourselves to be talking about the same things. And, it is worth stressing, it is only an explanation of the linguistic behavior that is needed—not an explanation of (the metaphysical issue of) what it is actually to be talking about the same thing. To see how the assumption is unwarranted, recall that the naturalistic neopragmatist views language as a social practice involving various marks and sounds. Our practices of agreement and disagreement (which are intimately connected to our views about univocality) can be underwritten simply by the fact that speakers very often take a token mark or sound to be of the same kind or (in the case of pronouns and other such place-holders) to function as if they were of the same kind. This taking-to-be-the-same can be explained primarily in terms of syntax, phonology or morphology, and a little bit of context.Footnote ¹⁵ Of course, sometimes later events undermine our initial impression of univocality. But that presents no problem. We are only trying to describe the verbal behavior indicative of people taking themselves to be talking to each other about a common topic. That explanation will be consistent with them sometimes, later, coming to the conclusion that they were not.

6. Where Should Philosophy Start?

The difference between normative neopragmatists and naturalistic neopragmatists—when it is conceived of as a dispute, and not just the pursuit of distinct projects—concerns the appropriate starting points for philosophical theorizing. Starting too late—that is, assuming premises that might themselves have raised relevant philosophical worries—can be a mistake. But starting too early—that is, assuming too little—can undermine the very possibility of producing anything adequate to one’s purposes. Given Price’s purposes, Shapiro, Macarthur, and Beasley argue that he starts his genealogical projects too early: that Price’s conception of how the functioning of various metaphysically puzzling vocabularies is to be explained must, contrary to Price’s stated views, begin with a wider explanatory base than is provided by the concepts made use of in the sciences. In particular, normative neopragmatists think we will need to start our theorizing with a normative notion: semantic content.

Shapiro’s view is that normatively couched explanations of the sort he favors should still count as naturalistic. After all, he writes, “instances of speech and thinking would not be characterized in a way that requires any supernatural capacities” (Shapiro, Reference Shapiro, De Caro and Macarthur2022, 159). It is true that the normative and intentional concepts will not be bearing any causal weight in the explanations in which they figure. So normative neopragmatists, in offering their explanations, do not have to posit any scientifically suspicious non-causal mechanisms of, say, epistemic access to norms. That is certainly desirable. Any form of liberal naturalism will fall short of supernaturalism. But if we use normative or intentional concepts at the ground level in our explanations, we are, I think, giving up too early in our efforts to demystify that sort of language. That is, we are presupposing an understanding of precisely the sort of things that many philosophers have been deeply puzzled by.

A similar worry arises about David Macarthur’s claims that attention to what he calls “the practical scientific image” causes placement problems to “evaporate” (Macarthur, Reference Macarthur2019, 569–70). The practical scientific image includes not only the posits of whatever passes muster as sufficiently rigorous science—physics, chemistry, biology, and so on—but also what is presupposed in the doing of that sort of science. At a minimum, that includes persons, ordinary artifacts such as chairs, tables and experimental equipment, and meaningful language. Because the practical scientific image includes mind, meaning, and morality in this way, Macarthur holds that there is no genuine conflict between the scientific and manifest images. Perhaps this is right. But to say that there is no conflict of this sort is not at all to say that there is no problem in understanding how mind, meaning, and morality fit together with the posits of the sciences. Just to say “scientists already acknowledge, in a way they cannot get out of, that there are such things as people, and meaningful assertions, and norms of scientific practice” is not even to take the first step in explaining what they, and we, are doing when we talk about such things. So it is false that “to take the step to recognizing the presuppositions of doing science undermines the distinction between scientific and manifest images and, thereby, undermines the placement problem” (Macarthur, Reference Macarthur2019, 572).

For all of the reasons I’ve offered above, I do not think that neopragmatism should start with the strong sort of liberal naturalism that Shapiro, Macarthur, and Beasley advocate, even if it might end up vindicating a strong liberal naturalism for other projects. Strong liberal naturalists hold that “the old conception of a strict dualism of the natural and the normative is […] a metaphysical prejudice rather than a matter of what the sciences, realistically conceived, show us” (Macarthur, Reference Macarthur, De Caro and Macarthur2010, 136). And they have some arguments that make this latter claim plausible, including the idea that any form of naturalism—including the most austerely scientific one—is itself a normative doctrine. But this perfectly consistent with the idea that some our concerns about some phenomena can only be satisfactorily addressed by explanations that avoid reliance on normative notions.

The problem with the attitude defended by Macarthur, Shapiro, and Beasley is not that the ontology and methodology they endorse embodies any sort of mistake. I agree: there are such things as reasons, meanings, and content. And we can appeal to these things in certain kinds of explanations. Similarly, there is no problem with moral philosophers taking it for granted that there are lots of people in the world, that they can suffer pain, and that some of their actions are free while others are involuntary. But still, it is philosophically legitimate to wonder about our knowledge of other minds and about freedom. What philosophers typically want is simply an understanding adequate to resolve the sorts of worries that are of particular concern to them. No one—with the possible exception of my mother—can worry about everything! And the theoretical worries of philosophers will depend on what else they happen to have questions about. So we need not endorse, with any degree of dogmatism, any particular set of restrictions on our methods or on the set of entities to which we appeal in our explanations.

The point that there are different sorts of explanations, and that they yield different sorts of intelligibility, is recognized by many theorists. Indeed, there are as many forms of intelligibility as there are different places people are willing to begin their explanations. So there is nothing wrong with Brandom’s inferentialism—a style of theorizing that appeals, at the ground floor, to normative reasons, commitments, and entitlements. For someone who is simply not worried about such things, Brandom’s theorizing surely yields important insights. But for someone—for example, most metanormative theorists—the appearance of normative notions at the ground floor seriously undermines the interest of the whole project.Footnote ¹⁶

In emphasizing that if the goal is understanding something, one is free to start wherever one feels one already has a satisfactory understanding (for one’s purposes), I am echoing a point that Brandom himself repeatedly emphasizes (Brandom, Reference Brandom2008, 227). But it is also true that the explanations offered by the sciences are typically not philosophically controversial. Philosophers do not get into disputes with evolutionary biologists about the origins of the dances of bees. The naturalistic neopragmatist’s hope is that the same sort of philosophically uncontroversial explanation can be given for those bits of language that have given rise to metaphysics. That is, they hope to pre-empt metaphysical explanations by making it unmysterious that we use the relevant vocabularies as we do. Of course I have not shown that this is always going to be possible. But I do hope to have shown that the arguments of the normative neopragmatists do not establish that it is not.