Skip to navigation – Site map
Symposia. Language or Experience

Radical Pragmatism: An Operator’s Guide

John Capps


Huw Price has recently argued that representationalism – the notion that the primary function of statements is to represent the world – is an utter failure. In its place he proposes a “global expressivism” that instead links the meaning of statements to how they are used. This makes his global expressivism a kind of pragmatism: a linguistic pragmatism because it focuses on linguistic meaning; a radical pragmatism because it rejects representationalism across the board. Price also introduces a distinction between two types of representation: external representation, which he mostly rejects, and internal representation, which he champions. In this article I weigh the strengths of Price’s radical pragmatism. I conclude that radical pragmatism has significant benefits, especially if we supplement external and internal representation with a third variety that I call operational representation.

Top of page

Full text

Introduction: Radical Linguistic Pragmatism

  • 1  To keep these two senses separate, I’ll be capitalizing the second, philosophical, sense of “Pragm (...)

1Pragmatists have a problem. Much as the term “liberal” has both a general, everyday sense and a more precise, philosophical sense – and, worse, these two senses may conflict with each other – “pragmatism” too can take on different meanings depending on the situation. To begin with there is, of course, the common everyday sense of “pragmatism”: here, to be pragmatic is to be moderate, middle-of-the-road, and non-ideological. Then there is the philosophical sense of pragmatism: here, Pragmatism is associated with a particular philosophical orientation, one that emphasizes use and practice over abstractions and abstract theorizing, an orientation that may even involve particular ways of theorizing about such things as truth and meaning. Pushed further, philosophical Pragmatism also stakes out very specific claims in epistemology, the philosophy of language, legal theory, and other areas.1

  • 2  Still, for a strong defense of Rorty’s anti-representationalism see Clough 2003.
  • 3  See, for example, the essays in Kitcher 2012; Brandom 2011, Price 2011 and 2013, as well as Willia (...)
  • 4  Price writes, “I may be somewhere between Rorty and Brandom – though if so, I think, then not very (...)

2Given these two senses, it is natural to ask whether and how a radical Pragmatism is possible, a version of philosophical Pragmatism that is neither moderate nor middle-of-the-road, a version of Pragmatism that is anything but pragmatic. Certainly Rorty-style neo-Pragmatism of the 1980s and 1990s had a reputation for being radical – but it also found many critics among philosophers. Some philosophers castigated Rorty for betraying the Pragmatic tradition, while others chastised him for endorsing cultural relativism.2 More recently, a second wave of neo-Pragmatists – many of them former students of Rorty’s or directly influenced by him – have defended Pragmatic approaches that ostensibly avoid the problems of Rortyan neo-Pragmatism. These second-wave neo-Pragmatists, including Philip Kitcher, Robert Brandom, Michael Williams, and Huw Price, have staked out positions that, while inspired by Rorty, diverge from his position to a greater or lesser degree.3 Of these, Huw Price’s Pragmatism is likely closest to Rorty’s – a point which Price himself concedes.4

  • 5  See Blackburn 1993, and Brandom 2000 for details. Price prefers to call his position a kind of “ex (...)

3Like Rorty, Price defines Pragmatism primarily in terms of an anti-representationalist approach to linguistic meaning. An assertion’s meaning is not, Price argues, a function of what it may or may not represent; an assertion may have meaning, in other words, even if it represents nothing at all. Price defends a “global anti-representationalism” that rejects representationalism across the board. In its place he proposes a global expressivism or inferentialism that goes well beyond the intentions of those – Simon Blackburn and Robert Brandom, in particular – who’ve helped put more local forms of expressivism and inferentialism on the map.5 By taking anti-representationalism to its logical extreme, Price’s Pragmatism is certainly not middle-of-the-road. In fact, it is a most unpragmatic Pragmatism.

4Thus Price takes Pragmatism to be primarily a linguistic thesis: essentially, global anti-representationalism with regard to linguistic meaning. It might appear that, if he is correct, and if Pragmatism equates to global anti-representationalism, then it is possible to have a free-standing radical Pragmatism that doesn’t require additional support from Pragmatist conceptions of truth, justification, experience, or what-have-you. In fact, his position is more subtle. Though he defends global anti-representationalism, he actually proposes replacing one sense of “representation” (“e-representation”) with another (“i-representation”). I’ll argue here that his proposal is only partly successful: while arguing in support of his radically pragmatic global anti-representationalism, I’ll also propose yet another way of thinking about representation. To put it briefly, there is an important sense of representation – what I’ll call “operational representation” – that Price overlooks. Recognizing this third sense of representation permits a Pragmatism that is no less radical in its anti-representationalism but that, nonetheless, presents an opening for Pragmatist accounts of truth, justification, and experience.

Pragmatic Prelude

5Before going further I’d like to say a few words about what’s at stake. The basic question is this: when we make a statement, what does that statement do? Of all the functions that our statements serve – to describe, to convince, to express, to prescribe or proscribe, etc. – which is most basic? This is a theoretical question, but there’s also a practical side: how we answer this question will affect how we understand other things we care deeply about.

6For example, most everyone cares – in some way, shape, or form – about the kind of person they are. We care about being the sort of person who tries to tell the truth, who loves her children, who appreciates good music, and who is concerned about global climate change. But not only do we care about being this sort of person, we also state that it is good to be this sort of person: we state that it is good to tell the truth, to love your children, and to be worried about climate change. We naturally treat these statements, statements about our most basic commitments, as both meaningful and, we hope, true. But what makes these statements meaningful and true? And are they really meaningful and true? Some attempts to answer these questions will strike many of us as clearly inadequate (though it may be hard to say exactly why). Ayer’s (1936) emotivism is one example: most of us find it preposterous that, when we say “it is good to love your children”, we’re only expressing an emotion. There must be more to it than that, we tell ourselves, especially given the importance of these statements. Likewise, Mackie’s (1977) error theory will strike most of us as unacceptable (though again it may be difficult to say why). To propose that ethical statements are universally false just doesn’t do justice to their importance.

7Of course, we care about other statements too – statements about the physical world, or about history, or about politics, or recent movies, or even celebrity gossip. We want to get clear on these as well. But not all of us care about history, or politics, or science in the same way that we all do care, or should care, about the kind of person we are. This makes statements about our character and ethical commitments especially important and interesting. And it makes understanding these statements likewise important and interesting.

8As a result, the issue of representationalism is, in part, the issue of how we make sense of statements, some of which are especially critical to how we understand ourselves and our most basic commitments. Granted, it’s not a practical question in the same sense as choosing the right weight of motor oil. But it is a practical question in the sense of getting our cognitive house in order, of making sure that we can explain and defend the basic commitments that make us who we are.

9Often, working out the details is a pretty drab affair. Knowing there’s this practical pay-off can help us through the dry spots. With that in mind, here’s a Pragmatic approach to representationalism.

Pragmatism, Anti-Representationalism, and Expressivism

10For Price, Pragmatism equates to anti-representationalism. Even more, Price defends a Pragmatism that is not just opposed to representationalism, it is globally opposed to representationalism: that is, opposed to “the idea that the function of statements is to ‘represent’ worldly states of affairs and that true statements succeed in doing so” (2013: 24). The main problem for representationalism, Price argues, is “the placement issue”. This is the problem of distinguishing between those discourses that are easily understood in representationalist terms (science and everyday talk of facts would be two examples) and those that are not. For example, moral discourse is often viewed as non-representational: the assertion “torture is wrong” just doesn’t seem representational in quite the same way that “my house is green” seems to clearly represent my house and its color. The representationalist must then explain how to “place” these discourses so that (a) they still involve making meaningful assertions even though (b) these assertions do not represent anything.

11In other words, representationalism runs into problems because it is far from clear what many apparently meaningful assertions could possibly represent, if they represent anything at all. Again, moral discourse is one such area. In our naturalistic age, few are willing to accept moral “facts” – with the result that, if we want moral discourse to be meaningful, then we must reject the idea that it represents the facts. Modal discourse is another problematic area: assertions about what is necessary or possible seem perfectly meaningful (and these assertions have a central role in good scientific reasoning) but it is difficult, once again, to describe what exactly these assertions represent. Like moral facts, modal facts are the sort of thing that many philosophers are not eager to embrace. As a result, it’s hard to defend representationalism globally, across all types of discourse.

  • 6  See Blackburn 1993 and Gibbard 1992 for the canonical case in support of expressivism.

12Instead, one approach is to propose alternatives to representationalism that explain the meaningfulness of moral and modal discourse, but without relying on representationalist assumptions. Blackburn’s and Gibbard’s expressivism is perhaps the best known alternative with respect to moral discourse.6 According to Blackburn and Gibbard, we should not view moral discourse as attempting (or, as error theorists have it, attempting and inevitably failing) to represent the moral terrain. Better, they argue, is to view moral discourse in “expressivist” terms, as the expression of a certain evaluative attitude. As Gibbard puts it:

Metaethics asks what ethical questions are: what does it mean to call an act wise or foolish, admirable or reprehensible?…The expressivist’s strategy is to change the question. Don’t ask directly how to define “good”, for no correct definition can break out of a normative circle[…]. Instead of seeking a straight definition, expressivists propose, seek a characterization of a different form. Ask what states of mind ethical statements express […]. The term “expressivism” I mean to cover any account of meanings that follows this indirect path: to explain the meaning of a term, explain what states of mind the term can be used to express. (2003: 6-7)

13Or in Blackburn’s words:

The theory I want to defend is one that gives a story about the way in which ethical thought functions. Valuing something, it says, is not to be understood as describing it in certain terms, any more than hoping for or desiring something are describing it in particular terms. Rather, the state of mind of one who values something is distinctive, but nevertheless it is itself a natural, and naturally describable, state. Once we find ethics here, we understand the essential phenomenon, which is that of people valuing things. When they value things, they express themselves in terms of what is good, bad, obligatory, right, justifiable, and so on […]. The ethical proposition gets its identity as a focus for practical thought, as people communicate their certainties, insistences, and doubts about what to value.

14This strategy – that of expressivism – leaves ethical properties and propositions alone with their own specific identities (1998: 49-50).

15If Gibbard and Blackburn are correct, then the meaning of moral terms is a function of what they are used to do – namely, express evaluative attitudes – and is not a function of whether they successfully represent a “worldly state of affairs”.

16Clearly these arguments support only a “local” expressivism. Perhaps moral discourse is best understood expressively while other types of discourse are best understood representationally. Thus, local expressivism is compatible with local representationalism. Moreover, being sensitive to variations among discourses looks like the pragmatic way to go.

17But Price disagrees. Local expressivism still poses the “placement issue” of providing a principled, non-ad hoc reason for treating some discourses as representational and others as expressive. Price’s solution instead is to argue for global expressivism and reject representationalism across the board. The placement issue evaporates along with any role for representationalism.

  • 7  I’ll largely concede Price’s points against representationalism. For the details of his argument s (...)

18Price’s case for global expressivism has two sides. The first is negative: like Rorty, Price claims that the arguments for representationalism are weak, riddled with unjustified assumptions, and generally unconvincing. In short, there are serious difficulties in getting a representational account off the ground, regardless of the area of discourse.7 The second side is positive: Price argues that global expressivism can do a better job than either global representationalism or local expressivism at explaining our linguistic practices. In other words, he claims that global expressivism has the conceptual resources for us to discard representationalism and never look back.

19So before getting into details, it’s worth keeping two related questions in mind. The first question is empirical: does global expressivism correctly describe how we talk? Perhaps expressivism does give the best account of how we use ethical language. But representationalism seems to be the best account of what’s happening when we talk about science or everyday physical objects. When I say “my house is green”, it seems evident that I am representing some worldly state of affairs. How does global expressivism explain this away?

20The second is a question of consistency: is Price being consistently Pragmatic? Because Pragmatists connect meaning with use, they often tell genealogical stories showing how certain theoretical concepts can be traced to common, concrete, interactions with the physical world. These stories describe a physical world that is causally connected to the concepts we use and the ways we talk and, as a result, it’s natural to view them as representing that physical world: they thus seem to take representationalism for granted. If that’s so, then there’s an evident tension between Pragmatism and global expressivism, and Price must show how his global expressivism does not undercut his Pragmatism.

Two Types of Representation

21Price responds to the empirical question – isn’t scientific discourse representational? – by arguing that global expressivism can account for types of discourse that seem obviously representational. This would mean that scientific discourse, say, is different only in degree, not in kind, from moral and modal discourse.

22Price’s argument rests on a distinction between two different senses of “representation”. The first, “e-representation”, is representation as traditionally understood (“e” stands for “environmental” or “externalist”):

On the one hand, we have the environment-tracking paradigm of representation, dependent on such notions as covariation…think of examples such as the position of the needle in the fuel gauge and the level of fuel in the tank, the barometer reading and air pressure and so on. In these cases, the crucial idea is that some feature of the representing system either does, or is (in some sense) “intended to”, vary in parallel with some feature of the represented system […]. In biological cases, for example, this notion gives priority to the idea that the function of a representation is to co-vary with some (typically) external environmental condition: it puts the system-world link on the front foot. (2013: 36, emphasis in original)

23Price raises familiar objections to representation in this sense. First, he argues that representationalism can’t work as a global account of linguistic practice. Then, turning to Blackburn’s quasi-realism and expressivism, and to Brandom’s inferentialism, he argues that these alternative accounts can be generalized to all types of discourse. In order to do so, however, he must explain why some types of discourse seem obviously representational even when they are not. He does this by introducing a second sense of representation: “i-representation” (“i” for “internalist” or “inferential”):

On the other hand, we have a notion that gives priority to the internal functional role of the representation: something counts as a representation in virtue of its position or role in some cognitive or inferential architecture. Here it is an internal role of some kind –perhaps causal-functional, perhaps logic-inferential, perhaps computational – that takes the lead. (2013: 36, emphasis in original)

24Price then concludes:

Once the distinction between these two notions of representation is on the table, it is open to us to regard the two notions as having different applications for different theoretical purposes. In particular, it is open to us to take the view that at least by the time we get to language there isn’t any useful external notion, of a semantic kind – in other words, no useful, general, notion of relations that words and sentences bear to the external world, that we might identify with truth and reference. (2013: 37 emphasis in original)

Price’s point is this: while plenty of things, such as gas gauges, are e-representational, words are not. Instead, words are i-representational. As a result, many discourses – such as science – that appear to be e-representational are actually i-representational. So while they are representational, they are not representational in the traditional e-representational sense. In what sense then are they representational?

  • 8  As noted earlier (see footnote 4) Price could just as easily – and perhaps more accurately – call (...)

25It may help to look first at moral discourse – no surprise, since this is expressivism’s home turf. Take a moral claim: for example, “lying is wrong”. Given its surface structure, one might be tempted to read this claim e-representationally, perhaps as representing a moral fact about lying. But of course we know better. If we agree with Blackburn and Gibbard, then we know that “lying is wrong” doesn’t represent anything at all: it is really the expression of an evaluative attitude. However, and this is Price’s point, “lying is wrong” does function like a representational claim within moral discourse. In particular, like a representational claim, we can use it to draw inferences.8 Still, the fact that this claim functions like a representational claim does not mean that it is one. Or, rather, it means that it is an i-representation, not an e-representation.

26Turning to scientific discourse, I think Price has something like the following in mind (he is sometimes vague when it comes to details). Consider the claim that “the Higgs Boson has a mass of approximately 125GeV”. At first glance (and probably even a second or third) this claim seems to represent the Higgs Boson as having a particular mass. But, just as with “lying is wrong”, appearances can be deceiving. If Price is right, the function of this claim is not to represent a worldly state of affairs. Rather, its function is to allow us to draw further inferences – perhaps further inferences about the Higgs Boson, or about other related areas in high energy physics. It is because this claim has a certain “position or role in some cognitive or inferential architecture” (2013: 36) that we say it is a representation, though as in the case of moral discourse it is only an i-representation, not an e-representation. By preserving the internal function of scientific assertions, Price hopes to make the external function superfluous. When we engage in scientific discourse we are not representing external reality. Instead, among other things, we are expressing the inferential rules that govern scientific discourse. So, according to Price, scientific discourse is different only in degree, not in kind, from moral discourse. The placement issue evaporates.

27If Price is correct, then i-representation gives us all the benefits of representationalism but without the costs that come from trying to solve the placement issue. Language is not e-representational, but it performs the same function as if it were. Price’s focus on the function of language, while side-stepping the metaphysical problems lurking in the shadows, seems rather Pragmatic.

Representation and Pragmatic Methods

28This brings us to the second question: is Price’s Pragmatism consistent? Generally, Pragmatists connect meaning with use. Price agrees. But, it seems, the main use of many discourses is to represent. It would seem to follow, on Pragmatic grounds, that representationalism correctly describes how these discourses function. This means it isn’t clear that Price can be both a Pragmatist and a global anti-representationalist: that he can connect meaning with use yet deny that we use language to represent.

29Similarly, as both Blackburn (2013) and Robert Kraut (2007) have pointed out, Pragmatists like telling genealogical stories that show how theoretical concepts develop from our concrete, practical interactions with the environment. (Dewey is probably best known for doing this, though he’s not the only one. Rorty also dabbled in this genre.) The problem is that these genealogical stories typically assume that representationalism is correct. As Blackburn puts it:

Even genealogical and anthropological stories have to start somewhere. There are things that even pragmatists need to rely upon, as they produce what they regard as their better understanding of the functions of pieces of discourse…Such genealogical stories start with a common-sense background of use, and a world of physical objects, with distinct locations, changing only according to distinct regularities with a distinct speed limit. (2013: 78)

If Blackburn is correct, then a favorite Pragmatist strategy depends on representationalism. Kraut calls this the “no-exit problem”:

The imagery of stepping outside a discourse and surveying it (without distortion) from an external perspective is unclear. It is hard to say “where we stand” in conducting such inquiries, or which resources we can import without circularity, or how to measure success, or even whether we have succeeded in stepping outside the discourse in question. External viewpoints are not always available: certain regions of discourse cannot be encapsulated, climbed out of, and surveyed from the outside […]. Call this the No Exit phenomenon. (2007: 9-10)

30So this is the challenge for Price’s Pragmatism. First, we can all agree that the placement issue poses a real problem for how we understand moral discourse. This is because, although moral facts and properties are difficult to defend, it’s hard to see how moral discourse can be meaningful without these representationalist concepts. Expressivists solve this problem by focusing only on what we do with moral discourse. They “side-step” any metaphysical questions that might come up. Because expressivists avoid moral facts and properties they have to be especially careful that they don’t let them sneak back in. In other words, there has to be a clean exit from representationalist concepts and idioms and this seems possible – just possible – in the case of moral discourse, and then only if we’re careful. But, Blackburn and Kraut argue, scientific discourse doesn’t have a clean exit. Even if we focus only on what we do with scientific discourse, the stories we tell – likely, genealogical stories about how science evolved out of everyday interactions with the physical environment – will invoke the same representational concepts and idioms that expressivism was supposed to avoid. According to Blackburn and Kraut, scientific discourse cannot escape representationalism. Price responds:

But why, precisely, should a global pragmatism need an exit, of the kind Blackburn and Kraut have in mind? The view that it does so seems at least in part a legacy of the cases with which the expressivist began, such as that of ethics. There it was important that the distinctive ontology of the ethical viewpoint – values, moral properties and the like – not be in view from the pragmatist’s external standpoint. By focusing on moral talk, rather than moral properties, the expressivist simply sidesteps the metaphysical conundrums that trouble her representationalist opponents, realists and anti-realists alike […]. The case of science is different. There isn’t a placement problem for scientific language, at least at first pass. So there isn’t any pressure to escape to a theoretical standpoint from which one doesn’t need to mention such things. (2013: 157-158)

31Price’s point is this: Talk of facts and properties isn’t as problematic in scientific discourse as it is in moral discourse. As a result, when talking about science, Pragmatists can help themselves to facts and properties without letting representationalism sneak back in. Pragmatists can continue telling genealogical stories, stories about how scientific discourse evolved from interactions with everyday objects, without becoming committed to representationalism. In other words:

Blackburn seems to be misled by the assumption that the pragmatist needs a standpoint outside any vocabulary on which she wants to turn her gaze […]. [But] there is plenty for the pragmatist to say, both in i-representational and e-representational terms, about the practical foundations of our ordinary ways of talking and thinking about, and hence coping with, the everyday objects themselves. (2013: 159-160)

If Price is right, then Pragmatists can continue drawing connections between meaning and use, and theory and practice, without worrying that representationalism will flare up again. There will, he assures us, be “plenty for the pragmatist to say”.

32Unfortunately, Price doesn’t give examples of what exactly Pragmatists can say. All he says, above, is that Pragmatists will be able to speak “both in i-representational and e-representational terms”. What can that mean? On a first pass, Price is simply saying that Pragmatists are allowed to talk about the physical world, and its connection to scientific discourse, without assuming any controversial metaphysical theories. Thus,

The expressivist’s motto should be that vocabularies should be mentioned but not used – theorized about but not employed, at least in the armchair. As long as the expressivist keeps this in mind – ensures that her initial theoretical perspective only mentions the target vocabulary – there’s no danger that her own casual deflationary use of the metalinguistic semantic vocabulary will lead her into the metaphysical trap. (2011: 314; emphasis in original)

33In addition, Price sometimes suggests that Pragmatists will be doing something like anthropology:

One of the great virtues of expressivism is the way that it replaces metaphysical questions with questions about human thought and language. In place of metaphysical questions about the nature of value, or modality, say, it offers us questions about the role and genealogy of evaluative and modal vocabularies – and these are questions about human behavior, broadly construed, rather than questions about some seemingly puzzling part of the metaphysical realm […]. Expressivism isn’t a way of doing metaphysics in a pragmatist key. It is a way of doing something like anthropology. (2011: 315)

34This means that:

The shift I recommend is not a matter of recasting metaphysical issues as issues about language but of abandoning the metaphysical questions altogether, in favour of the anthropological questions. (2013: 181)

35So Price is saying that Pragmatists can give a quasi-anthropological account of how certain areas of discourse – areas of discourse that seem paradigmatically representational – came about. These accounts show why these areas appear representational even though they really aren’t. And the same goes for these anthropological accounts themselves. They too might appear representational but really are not. Rather, as noted above, “there is plenty for the pragmatist to say, both in i-representational and e-representational terms, about the practical foundations of our ordinary ways of talking and thinking” (2013: 160).

36But we’re not out of the woods yet. The next question is how these anthropological accounts hold up: how do we distinguish a credible, compelling genealogy from a dubious, unconvincing one? The obvious answer is that a credible, compelling genealogy meets a standard that the dubious, unconvincing one simply doesn’t. But what is that standard?

37According to Price, there are two kinds of “external constraints” on specific discourses. One, corresponding to e-representation, is “covariance…between a tokening of a representation and an element of an external environment” (2013: 37). The second, corresponding to i-representation:

Is a kind of “in-game externality” provided by the norms of the game of giving and asking for reasons – the fact that, within the game, players bind themselves, in principle, to standards beyond themselves. (2013: 37)

A gas gauge e-represents the fuel level. Whether it does a good job depends on an external constraint: the actual level of fuel. An area of discourse i-represents a certain subject matter. Whether it does a good job depends on “the norms of the game”.

38But this is likely to strike most people as way too squishy. If the only constraints on an area of discourse are “in-game” then there’s no principled basis for preferring one discourse over another. By its own standards, of course, every game is simply the best. The same would be true of discourses. If Pragmatists choose to give anthropological accounts of science, or politics, or morality, then that’s all well and good – and it would be surprising if these weren’t fantastically successful by whatever the in-game standards happen to be. But these accounts won’t convince anyone who isn’t already a global expressivist. So, actually, that’s not all well and good.

  • 9  In fact, this is one of the points that distinguishes Philip Kitcher’s Pragmatism from Rorty’s – s (...)

39Price’s position is very close to Rorty’s. Like Price, Rorty blurred the distinction between science and other kinds of discourse. And, like Price, Rorty claimed that science (as well as other areas of discourse) was governed only by internal standards of correctness. Rorty’s views, of course, were met with howls of protest: some said that he was a shabby relativist; others claimed that he wasn’t really a Pragmatist. Even though Rorty was often more subtle than his critics acknowledged, his critics did have a point. His claims about science, in particular, often seemed a bit too dismissive and a bit too unaware of actual scientific practice.9

40The problem for Price is similar. As we’ve seen, Price argues that Pragmatists can continue doing genealogy without risk of self-contradiction. Pragmatists, he’s claimed, can continue to give anthropological accounts of different areas of discourse, accounts supporting a global expressivism, without lapsing into representationalism. But if we extend global expressivism to these accounts as well (it is global expressivism, after all) then they apparently lose all credibility: since they at most i-represent their target discourses, and because they must therefore meet only “in-game” standards, there is no external, evidential basis for deciding that some genealogies are more credible than others, or even credible at all.

41To return to the original question: is Price’s Pragmatism consistent? As we saw, Blackburn accuses Price of covertly depending on representationalism. Price responds by claiming that Pragmatism can be expressivist all the way down. While this would answer the consistency question, it comes at a high cost. By rejecting representationalism across the board, Price seems to have undercut his own position, depriving it of any kind of evidential support.

A Third Type of Representation

42This suggests another way of being a global anti-representationalist – and of being a radical Pragmatist. As we saw, Price’s position is more modest than it first appears. On the one hand, his target is primarily “e-representation”: gas gauges e-represent the fuel level in your gas tank, but words do not e-represent worldly states of affairs. On the other hand, Price is perfectly comfortable with representation in the sense of i-representation – though, as I argued above, we shouldn’t be: Price’s i-representation undercuts his Pragmatism. So the question is whether you can be a global anti-representationalist (that is, reject e-representation) without undercutting Pragmatism. I think you can and here’s how to do it.

43First, recall that for Price i-representation is a way of meeting the traditional demands placed on representation while shedding the problems raised by the placement issue. Traditionally, representationalism gives an account of the function statements serve (they represent the world) as well as an account of the constraints on specific discourses (namely, external constraints). The problem, as Price notes, is that many statements either don’t serve a clear representational function or the constraints on them are not clearly external. I-representation is how Price then explains the function of statements (“perhaps causal-functional, perhaps logic-inferential, perhaps computational” (2013: 36)) and the constraints on them (“in-game” constraints). This suggests that there are two requirements on any account of representation: first, that it explain the function that statements serve; second, that it account for the constraints on areas of discourse.

44However there is a third function representation plays in addition to its environmental tracking function (e-representation) and its “perhaps logico-inferential” (2013: 36) function (i-representation). This is its operational function. I’ll call this “o-representation”:

o-Representation: This notion gives priority to the operational functional role of representation. Something counts as a representation in virtue of its ability to facilitate operations between one’s cognitive architecture and the environment.

The idea here is that representation isn’t merely a matter of mirroring reality (e-representation) or of clarifying the inferential connections among statements (i-representation). Rather, a fundamental function of representation is to smooth interactions with the world. From the standpoint of o-representation, statements function primarily as tools that help us come to grips with our world. The constraints on these tools are neither straightforwardly external nor straightforwardly internal.

45Some examples will help. Let’s begin with a very simple claim: “the pan is hot”. This is the sort of thing we might say in a kitchen, perhaps when cooking with another person. E-representation does not adequately capture what is being said (or represented). The e-representational content of “the pan is hot” is simply that the pan is, well, hot. But that content, while true, doesn’t explain why someone would say “the pan is hot”, or the significance of saying “the pan is hot”, or what is gained by saying “the pan is hot”. There are many truths one could also state (“the pan is black”, “the pan is iron”, “that is my pan”) so why say, “the pan is hot”? O-representation explains why. You say “the pan is hot” when you want to alert your partner: most likely you’re either worried she might burn herself or it’s time to add something to the pan. You say “the pan is hot” because something needs to be done: someone needs to grab a hot pad, or add oil, or perhaps both. Saying “the pan is hot” helps facilitate an interaction with the environment. No one gets burned and dinner gets cooked. It’s this dimension – the operational and functional dimension expressed by o-representation – that e-representation neglects. Likewise i-representation doesn’t fully capture what is being represented by “the pan is hot”. From “the pan is hot” one can infer all sort of things, including “don’t touch the pan” and “time to add the oil”. However one can also infer a multitude of other things such as “the pan is not cold”, “the pan is not at room temperature”, “don’t sit on the pan” and “don’t put the cat in the pan”. Only some of these inferences are relevant, because only some of these inferences actually help you do something. It is because these inferences help you interact with your environment, by safely putting dinner on the table, that they give representational content to “the pan is hot”. From the standpoint of o-representation, it is therefore the operational content that matters, not the inferential content. This means that the representational content of “the pan is hot” (as well as myriad other everyday statements) is best understood in terms of o-representation, not e- or i-representation.

46To take a second example, consider the social-scientific concept of race. It’s widely recognized that the concept of race has no real biological basis. In addition, the concept of race has a pretty ugly history, often being used to justify all sorts of despicable policies and acts. At the same time, as Philip Kitcher has noted, the concept of race can serve a useful, beneficial function: if we want to figure out why certain populations die younger, or suffer from some diseases more than other groups, the concept of race can be a useful way of identifying and studying those populations. As he puts it:

The notion of race is likely to continue to straddle the divide between well-motivated science (for example, the quest to trace patterns of human migration) and social applications. Any pragmatic assessment of its value will have to deal in a synthetic and balanced way with both types of context. (2012: 156)

47Likewise, Elizabeth Anderson has argued that the ideal of color blindness is both morally and conceptually confused (2010: 156): color blind policies, while perhaps desirable in theory, tend to over-simplify the concept of race and legitimize continued discrimination and segregation. If we reject the concept of race entirely then we risk fostering more racial discrimination, not less. Or, as Eddie Glaude warns:

Races are as real as football fields, dollar bills, or head-negroes in charge. If they are proven to rest on bad science, like phlogiston, it doesn’t make them any less real […]. Race language or black identity cannot be escorted offstage because of some questionable conception of the real. (2007: 63-64)

48Taken together, this means that a statement such as “African-Americans are nearly twice as likely to have diabetes as non-Hispanic whites” does not simply represent a worldly state of affairs (as e-representation would have it); nor does it merely help us to draw inferences (as i-representation would have it). Of course, these functions are not entirely dispensable: certainly part of the function of this statement is that it does tell us something about our world and that it does allow us to draw certain inferences. (Along these lines, Kitcher proposes that “races are both biologically real and socially constructed”, 2012: 149.) But this is neither the whole nor the most important part of the story. E- and i-representation may not be entirely dispensable, but they are far from being primary or fundamental. Instead, the primary function of statements like this is to lead to actions: here, actions that we hope will lead to a decrease in diabetes. Knowing that African-Americans have a higher rate of diabetes should lead to research, or perhaps a donation to support that research. This is the statement’s o-representational content.

49Here’s a third example. Geologists define minerals as naturally occurring, inorganic, solid substances, with a definite chemical formula and a crystalline structure. Most people are surprised to learn that ice is a mineral by this definition (albeit only ice that is naturally occurring, such as glacier ice; ice cubes from the freezer don’t count). We’re surprised, no doubt, because ice just doesn’t seem like a mineral: unlike most minerals it regularly falls from the sky and can be easily liquefied; we will never see it on a list of recommended “vitamins and minerals”; prospectors searching for minerals will never be satisfied with a handful of snow. So what are we doing when we say that ice is a mineral? Again, it does not seem that we are merely representing a worldly state of affairs (as e-representation would have it): ice being a mineral is at least as much a matter of definition, and of the history of geology, as it is a matter of ice’s natural properties. And it does not seem that we are merely pointing out the inferences we can now draw. After all, ice being a mineral is a function of its natural properties, including its crystalline structure and definite chemical formula. So ask yourself this question: how did you react the first time you heard “ice is a mineral”? If you’re like me, you were probably skeptical for all the reasons given above. Depending on the circumstances, maybe you questioned the person who said it, or maybe you tried to verify it on-line, or maybe you just kept reading, waiting to see where this will go. The statement “ice is a mineral” need not lead to any direct experience of ice (though it could, especially if you’re a geologist). Rather, the important point is that this statement leads to some kind of action that reconciles our prior beliefs (ice? a mineral!?) with the external environment (ice has the same natural properties as many paradigmatic minerals). Those actions reflect its o-representational content, which will be different for different people in different circumstances.

50A fourth example. Consider the case of the Panchen Lama, the second highest ranking lama – after the Dalai Lama – in Tibetan Buddhism. After the Dalai Lama’s death the Panchen Lama plays a critical role in identifying the Dalai Lama’s next incarnation. Today there are two Panchen Lamas. One, Gedhun Choekyi Nyima, was recognized by the Dalai Lama when he was six years old – and was promptly detained by the Chinese authorities. He hasn’t been seen in public since 1995 and his whereabouts are unknown. The other, Gyancain Norbu, is recognized by China, holds a leading position in the Buddhist Association of China, and will most likely figure prominently in choosing the next Dalai Lama. Most observers would say that Gedhun Choekyi Nyima, the Dalai Lama’s choice, is the real Panchen Lama. But there’s also something odd about making this claim. After all, according to Tibetan Buddhism, the current Panchen Lama (whoever that is) is actually the eleventh in a lineage stretching back to the 14th  century, each one a reincarnation of the previous Panchen Lama. If the question is whether Gedhun Choekyi Nyima or Gyancain Norbu is the real reincarnation of the tenth Panchen Lama, then most of us reading this article would probably say that neither is, because there is no such thing as reincarnation. However, at the same time, most of us probably find China’s actions deplorable: Gedhun Choekyi Nyima is the real Panchen Lama, we say, and China’s actions shouldn’t affect how the next Dalai Lama is chosen. So what are we doing when we state that Gedhun Choekyi Nyima is the real Panchen Lama? Well, once again, it does not seem as if we are representing an external, worldly, state of affairs. After all, most of us would deny that anything like reincarnation exists or that Gedhun Choekyi Nyima is really the reincarnation of the tenth Panchen Lama. Nor does it seem as if we’re merely endorsing the inferences this statement allows. Here it isn’t the inferences that matter so much as the fact that China is using illegitimate means to control the process of choosing the next Dalai Lama. Instead, the statement that Gedhun Choekyi Nyima is the real Panchen Lama leads to actions: perhaps to lobbying China for his whereabouts and condition, perhaps to rejecting Gyancain Norbu’s legitimacy as Panchen Lama, perhaps to working out contingency plans for recognizing the next Dalai Lama. Once again, this is what the statement “Gedhun Choekyi Nyima is the real Panchen Lama” o-represents.

51To sum up, o-representation underscores the operational importance of statements: how statements are used to stimulate and guide interactions with the world. As these four examples illustrate, statements can be o-representational even when they are not clearly e- or i-representational. I’ll have more to say about this below.

52An account of representation should also explain the constraints on particular areas of discourse. As the examples above illustrate, the constraints are both external and internal. Whether it’s appropriate to say “the pan is hot” depends on both external factors (the temperature of the pan) as well as internal features of the discourse of cooking (a 65ºc pan may be hot to the touch, but it’s not what a cook means by “hot”). Race, as Kitcher notes, is both “biologically real and socially constructed”, ice being a mineral is both a function of its natural properties and a particular disciplinary definition, and the identity of the Panchen Lama depends both on the actions of the Chinese authorities as well as the traditions of Tibetan Buddhism. The constraints on these particular statements hold as well for the discourses in which they occur.

53There are two typical objections to what I’ve said so far. The first is this: even if the examples above are legitimate cases of o-representation, there are many others that are not. For example, some might claim that “the cat is on the mat”, “snow is white”, and “grass is green” – not to mention other claims favored by philosophers – are best understood e-representationally, as straightforward representations of the external world. There’s some truth to this: after all, it’s hard to imagine what actions “grass is green” would inspire. But that’s exactly the problem, and it’s a problem for e-representation, not o-representation. Statements like “grass is green” are neither interesting nor surprising; in fact, it is hard to imagine a normal real-life situation (at least a situation outside a philosophy paper) where someone would actually say “grass is green”. Perhaps when bemoaning a brown lawn? Or after a long, snow-covered winter? The question then is whether these sorts of statements should drive our theory. Should we treat statements of this sort as the rule, or as the exception? My recommendation is that we treat these as the exception. Although true, they are boring, pointless assertions that we’re unlikely to come across except when doing philosophy. I suspect a large number of philosophical examples fall into this category.

54Another way of putting it is this: if the main function of language is simply to allow us to say things like “grass is green” – if “grass is green” is the paradigm – then it’s hard to see what value language would have or why it would ever evolve. If, on the other hand, language evolved to help us cope with our environment (a plausible assumption to make), then it is practically a truism that o-representation would be more fundamental than e-representation. This would mean that statements with o-representational content – statements about hot pans, race and diabetes, the classification of ice, and the identity of the Panchen Lama, for example – should be viewed as the rule, while statements about trivial properties of snow and grass are the exceptions. Our theories will be better off to the extent that we focus on the former, not the latter.

55The second objection is that this account makes things a lot less tidy. Both e-representation and i-representation presuppose simple, clear-cut constraints on different areas of discourse. In the case of e-representation the constraints are external; in the case of i-representation the constraints are internal (what Price refers to as “in-game externality” which isn’t really external at all). In the case of o-representation the constraints can be both. One might think that o-representation would then take on the weaknesses of both external and internal constraints. To the contrary, I think the account is actually stronger for incorporating both types of constraints. As we’ve seen, the problem with exclusively external constraints is that they seem completely absent in many areas of discourse. The problem with exclusively internal constraints is that they prevent us from taking an external standpoint on the credibility of specific discourses. The common-sense (dare I say Pragmatic?) solution is to combine the two. The constraints on any given area of discourse are a combination of external and internal factors that are necessary to effectively interact with one’s environment. The result may be less tidy, but it has the advantage of being a lot more realistic.

56O-representation therefore has some significant advantages over both e- and i-representation. Because it defines representation in terms of operations and interactions, it avoids the metaphysical problems facing traditional representationalism. Unlike e-representation, it does not face the placement issue. It thus offers a fully generalizable, global account of representation. And, unlike i-representation, o-representation doesn’t depend on only “in-game” constraints. By recognizing both internal and external constraints, o-representation recognizes that specific areas of discourse depend on external factors for their credibility.

57Indeed, it’s possible to view e- and i-representation as special cases (very special cases) of o-representation. If the core function of representation is to facilitate operations and interactions with the external world, then e-representation captures those cases where the interaction is purely a matter of mirroring. I’d argue that we’re not likely to come across many such cases in the wild, though we frequently find them in the artificial surroundings of philosophy. As noted above, a statement such as “grass is green” may come close to being purely e-representational – but it is nearly useless to say in real life. As far as what we use language for, these kinds of statements should be viewed as the exception, not the rule. Likewise, i-representation covers those cases where a discourse has little contact with the external world. Some areas of mathematics may be purely i-representational (though we still expect mathematics to pay practical dividends in terms of our interactions with the world). Other discourses may be i-representational if they don’t facilitate any practical operations or interactions – but when a discourse only follows its own internal logic that is usually a sign of sickness, not of health. Once again, we should treat these as exceptions, not rules.

A Historical Interlude

58This leaves many of the details of o-representation to be worked out, but there are some historical precedents to help point in the right direction. Price frequently alludes to Wittgenstein, Sellars and Dewey in support of his global anti-representationalism. Turning our attention to Dewey, it is surprising how little he had to say about representationalism explicitly. Price quotes from a 1905 letter where Dewey writes about giving the “coup de grâce to representationalism” (2011: 204) but just about the only place where Dewey explicitly addresses representationalism – by name – is in a pair of articles published in The Journal of Philosophy in 1922. There Dewey defends what he calls a “naturalistic realism” where things – ranging from natural objects, to artifacts, to words – function as representations when they “mean or support a certain conclusion” (1922: 352). The connection between the representation and what is represented is a pragmatic connection:

The “pragmatic” feature comes in when it is noted that experiment or action enters to make the connection between the thing signifying and the thing signified so that inference may pass from hypothesis to knowledge. It is then seen that some “consequences”, namely those of the experiment, are an integral part of the completing or fulfilling or leading out of the “representation” into final objects. (1922: 354)

Dewey’s point seems generally consistent with what I’ve argued here: representation is largely a matter of facilitating interactions (or, as Dewey says, “experiment or action”) with the world.

59Dewey would also speak of truth in operational terms. In his debate with Russell over the correspondence theory of truth he wrote:

My own view takes correspondence in the operational sense:…the meaning, namely of answering, as a key answers to conditions imposed by a lock, or as two correspondents “answer” each other; or, in general, as a reply is an adequate answer to a question or a criticism – as, in short, a solution answers the requirements of a problem. On this view, both partners in “correspondence” are open and above board, instead of one of them being forever out of experience and the other in it by way of a “percept” or whatever. (1941: 178, emphasis in original)

Dewey’s first point is that correspondence (we can just as well read “representation”) is not a matter of our minds or language mirroring the external world. Rather, the relationship is much more active (“operational”) than the mirroring metaphor implies. His second point is that correspondence cannot be reduced to a purely internal relation: whether a statement is true depends on more than how well it coheres with other claims. There are objective constraints on whether a statement corresponds to (or represents) “worldly states of affairs” just as not every proposed solution will successfully solve its designated problem.

Conclusion: Radical Pragmatism Redux

60The question is whether one can be a radical Pragmatist and reject representa-tionalism across the board. My answer is a qualified yes: following Price, I agree that representationalism (specifically, e-representationalism) appears unable to solve the placement issue. So: yes to global anti-representationalism. I also agree with Price that there are other senses of representationalism to replace e-representation, but I disagree with him that i-representationalism is the right replacement. I would argue that o-representationalism is the way to go. So: a qualified yes to global anti-representationalism.

61To recap: first, o-representation enjoys the benefits of e-representation but without the metaphysical baggage. Most importantly, it helps explain how scientific statements help us interact with our environment: while e-representation emphasizes mirroring and tracking, o-representation emphasizes operations and interactions. Second, o-representation is also better than i-representation at accounting for scientific discourse. It does not reduce science to a game following purely internal rules; rather, by highlighting the operational function of our statements it helps explain why science is the generally successful enterprise it is. Here again the crucial difference is that while i-representation emphasizes internal relations and inferential connections, o-representation emphasizes the operations that bring us into relation with worldly states of affairs.

62O-representation allows us to be resolutely anti-representational in all the ways that matter. While representationalism raises a host of metaphysical problems (including the placement issue Price identifies), these problems do not arise for o-representation. To the extent that moral and modal forms of discourse facilitate interactions with the surrounding environment, we may safely view these discourses as o-representational; there is thus no need to find grounds for distinguishing between scientific, moral, and modal discourses, among others.

63The radical Pragmatism Price proposes, and which I’ve proposed modifying, is clearly indebted to Rorty’s linguistic neo-pragmatism. Here, pragmatism is treated as a linguistic thesis opposed to traditional forms of representationalism. But there are some important differences. By emphasizing operations and o-representations, this approach makes space for pragmatic approaches to other concepts besides representation: pragmatic concepts of truth, justification, and experience now have a clear point of entry. As I’ve suggested above, there are likely some points of contact between a radical Pragmatism and Dewey’s philosophy.

64Richard Rorty often argued that Dewey was “waiting at the end of the dialectical road which analytic philosophy traveled” (1982: xviii). This prompted Ralph Sleeper to write that, actually, “Dewey was trying to block that road from its beginning” (1987: 5). Whichever is right – and I think they both are – radical Pragmatism shows us the road ahead.

Top of page


Anderson E., (2010), The Imperative of Integration, Princeton: Princeton University Press.

Ayer A., (1952), Language, Truth and Logic, London: Dover Publications.

Blackburn S., (1993), Essays in Quasi-Realism, Oxford: Oxford University Press.

Blackburn S., (1998), Ruling Passions, Oxford: Oxford University Press.

Blackburn S., (2013), “Pragmatism: All or Some?”, in Price (2013), 67-84.

Brandom R., (2000), Articulating Reasons: An Introduction to Inferentialism, Cambridge: Harvard University Press.

Brandom R., (2011), Perspectives on Pragmatism, Cambridge: Harvard University Press.

Clough  S., (2003), Beyond Epistemology: A Pragmatist Approach to Feminist Science Studies, Lanham: Rowman and Littlefield.

Dewey J., (1922), “Realism Without Monism or Dualism”, Journal of Philosophy, 19: 309-17, 351-61.

Dewey J., (1941), “Propositions, Warranted Assertibility, and Truth”, Journal of Philosophy, 38: 169-86.

Gibbard A., (1992), Wise Choices, Apt Feelings, Cambridge: Harvard University Press.

Gibbard A., (2003), Thinking How to Live, Cambridge: Harvard University Press.

Glaude E., (2007), In a Shade of Blue: Pragmatism and the Politics of Black America, Chicago: University of Chicago Press.

Kitcher P., (2012), Preludes to Pragmatism, Oxford: Oxford University Press.

Kraut R., (2007), Artworld Metaphysics, Oxford: Oxford University Press.

Mackie J., (1977), Ethics: Inventing Right and Wrong, London: Penguin.

Price H., (2011), Naturalism without Mirrors, Oxford: Oxford University Press.

Price H., (2013), Expressivism, Pragmatism and Representationalism, Cambridge: Cambridge University Press.

Rorty R., (1982), Consequences of Pragmatism, Minneapolis: University of Minnesota Press.

Sleeper R., (1987), The Necessity of Pragmatism, New Haven: Yale University Press.

Williams M., (2013), “How Pragmatists Can Be Local Expressivists”, in Price 2013, 128-144.

Top of page


1  To keep these two senses separate, I’ll be capitalizing the second, philosophical, sense of “Pragmatism” while reserving the lower-case “pragmatism” for the first, everyday sense of the word.

2  Still, for a strong defense of Rorty’s anti-representationalism see Clough 2003.

3  See, for example, the essays in Kitcher 2012; Brandom 2011, Price 2011 and 2013, as well as Williams 2013.

4  Price writes, “I may be somewhere between Rorty and Brandom – though if so, I think, then not very far from either” (2013: 193).

5  See Blackburn 1993, and Brandom 2000 for details. Price prefers to call his position a kind of “expressivism” but his anti-representationalism could just as well (and perhaps more accurately) be described as a form of inferentialism. Here I will typically follow Price’s lead and refer to his position as “global expressivism”.

6  See Blackburn 1993 and Gibbard 1992 for the canonical case in support of expressivism.

7  I’ll largely concede Price’s points against representationalism. For the details of his argument see the first of his Descartes lectures “Naturalism without Representationalism” (2013: 3-21), as well as the essays “Moving the Mirror Aside” (2011: 3-33), and “One Cheer for Representationalism?” (2011: 304-321).

8  As noted earlier (see footnote 4) Price could just as easily – and perhaps more accurately – call his position a form of global inferentialism. He cites the inferential content of statements every bit as much as their expressive content. In “Pragmatism, Expressivism, and Anti-Representationalism” (2011: 190-219), Brandom argues against Price’s appropriation of inferentialism, raising objections similar to Blackburn’s.

9  In fact, this is one of the points that distinguishes Philip Kitcher’s Pragmatism from Rorty’s – see the essays in Kitcher 2012.

Top of page


Electronic reference

John Capps, « Radical Pragmatism: An Operator’s Guide », European Journal of Pragmatism and American Philosophy [Online], VI-2 | 2014, Online since 24 December 2014, connection on 25 September 2017. URL : ; DOI : 10.4000/ejpap.311

Top of page

About the author

John Capps

Rochester Institute of Technology

Top of page


Licence Creative Commons
Author retains copyright and grants the European Journal of Pragmatism and American Philosophy right of first publication with the work simultaneously licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Top of page
  • Logo Associazone culturale Pragma
  • Logo DOAJ – Directory of Open Access Journal
  • Logo École des Hautes Études en Sciences Sociales – EHESS