On Going On: Rules, Inferences and Literary Conditions

Both in and out of the game and watching and wondering at it.
—Walt Whitman

At §293 in Philosophical Investigations, Ludwig Wittgenstein imagines the following scenario:

Suppose everyone has a box with something in it: we call it a ‘beetle’. No one can look into anyone else’s box, and everyone says he knows what a beetle is by looking at his beetle.—Here it would be quite possible for everyone to have something different in his box. One might even imagine the thing constantly changing.—But suppose the word ‘beetle’ had a use in these people’s language?—If so it would not be used as the name of a thing. The thing in the box has no place in the language-game at all; not even as a something: for the box might even be empty.—No, one can ‘divide through’ by the thing in the box; it cancels out, whatever it is. That is to say: if we construe the grammar of the expression of sensation on the model of ‘object and designation’ the object drops out of consideration as irrelevant.1

Richard Rorty says Wittgenstein’s thought experiment boils down to the question: “Can a descriptive term have a sense if its application is regulated by no public criteria”; a question to which Wittgenstein—or at least the voice at §293—appears to give the answer no. Since the beetle example is about the “grammar of the expression of sensation,” Rorty says that for Wittgenstein words like “pain” have a sense only as long as they “are not treated as a name of something whose presence or absence swings free of all differences in environment or behavior.”2 To talk about pain as “qualia” or as a raw feel is just to add a nuance to a language game; one that would allow us, as Rorty puts it, to “disjoin pain from pain behavior” linguistically.  But adding nuances to a language game in this way does not mean the word “pain” (or “beetle”) points to some inner quale; i.e., its use is not a matter of private ostension.

The question of whether a term can have a meaning if its application is regulated by no public criteria, though, arises even more clearly from a different set of examples in the Philosophical Investigations: examples pertaining to the notorious problem of “rule following.”  In these examples, questions about what it means to follow a rule become questions about how words in general are subject to criteria of correct and incorrect application.  And such questions, I want to suggest here, invite comparison with some of the basic arguments of what is sometimes called analytic pragmatism.  Impressed by Wilfrid Sellars’ insight that in order properly to be saying something we have to take ourselves as being in the “logical space of reasons,” analytic pragmatists like Rorty and his student Robert Brandom, (despite their acknowledged debt to thinkers like William James and John Dewey) mount a fundamental criticism of the “classical” phase of pragmatism. In his essay “Dewey Between Hegel and Darwin,” Rorty says Dewey’s naturalism amounts to a blurring of the “distinction between the question ‘What causes our beliefs?’ and the question ‘What justifies our beliefs?’”3 As Rorty cautions in a later essay, we should not “treat the causal ability of certain events to produce non-inferential beliefs…as a justification for [the] holding [of] those beliefs.”4 If Dewey thought the human animal moved on a smooth continuum from causes to justifications—from nature to norms, we might say—analytic pragmatists inherit from Sellars an apprehension about blurring the line between sensory events and justified belief.  What Wittgenstein says at §293 about the public criteria by which language use can be taken as meaningful is comparable, in ways that matter, to the analytic pragmatists’ telling us that in order to be saying anything at all we must be able to justify our claims.

In wanting to identify some affinities between Wittgenstein’s later thought and analytic pragmatism, I will focus here on sections of the Philosophical Investigations dealing with rule-following and private language.  The first thing I want to claim is that Wittgenstein and Brandom each remain committed to versions of what Kant called the quid juris question: by what “right” are we entitled to take experience, knowledge, belief etc. as meaningful—an issue Kant tried to satisfy through transcendental argument.5 The second, related thing I want to claim is that the sort of inferentialism Brandom is concerned to elaborate, as well as Wittgenstein’s exploration of the rule-following paradox, are each bound up with what it means to “go on in the same way”: on the one hand, the role of ‘etc’ or ‘…’ in Wittgenstein’s number series examples; on the other, Brandom’s distinction between, say, parrots that respond reliably and differentially to discrete stimuli and discursive creatures that can go on in the game of giving and asking for reasons.  I think Brandom and Wittgenstein basically agree that, as Brandom puts it, “intentionally contentful states and acts have an essentially normative pragmatic significance.”6

But is Wittgenstein’s idea of the criteria by which we take our talk to be meaningful a matter of “inference”? And if it isn’t, then what exactly are the normative criteria in relation to which we recognize what it means to “go on” in a given language game?  In a third section I consider these questions in light of Stanley Cavell’s claim that in the Philosophical Investigations Wittgenstein provides the literary conditions of [his] philosophical aims.”7 What exactly does Cavell mean by “literary conditions”? Do literary conditions have their own forms of entitlement?  Would such conditions—say, Wittgenstein’s particular scene-setting, thought experiments, aphorisms, and dialogues—amount to an alternative form of justification?  Could a tactful or artful (or beguiling or captivating or worrisome) ordering of words—what we might simply call a style—itself generate the criteria for claiming?  How exactly can, as Cavell puts it, “an ordering of words [be] its own bottom line, [and] see to its own ground?”8 Considered next to Brandom’s normative pragmatics of inference, Cavell, I try to show, finds in the Investigations a kind of claiming inseparable from a way of writing.

I.

At §143 we find Wittgenstein wanting to “investigate the language game” in which a pupil is taught to follow the rule of adding 2. The investigator will be confident the pupil understands the rule when the pupil starts “going on to write down [the series] independently.” But at §146 the investigator begins to wonder just what is to count as the pupil’s having “got the system.”  For example, what if the pupil were consistently to add 2 and then, upon reaching 1000, start adding 4 (996, 998, 1000, 1004, 1008, 1012); and in doing so thinks she has gone on in accordance with the rule?  Here we run into something that looks like a paradox. As Wittgenstein puts it, the paradox involves the way “we can think of more than one application of an algebraic formula and every type of application can in turn be formulated algebraically.” The rule itself, in other words, is not enough to determine what it will mean to go on in the same way for an indefinite number of future applications. The dilemma is neatly expressed at §85 in the aphorism “a rule stands there like a signpost,” which I take to mean that there is nothing about the rule that dictates how it is to be interpreted, and so there is room for doubt about what it would mean to apply the rule correctly.

A little further on, we find Wittgenstein setting up a similar problem. At §193 he imagines a

machine symbolizing its action [such that] the action of a machine…seems to be there in it from the start. What does that mean?—If we know the machine, everything else, that is its movement, seems to be already completely determined. We talk as if these parts could only move in this way, as if they could not do anything else [and in doing so] we use a machine, or the drawing of a machine, to symbolize a particular action of the machine. For instance, we give someone such a drawing and assume that he will derive the movement of the parts from it (just as we can give someone a number by telling him that it is the twenty-fifth in the series 1, 4, 9, 16, . . . .).

As with the number series example, there is here a question about what it means to move from a formula to the derivation of a series.  But here the move occurs more explicitly as a conflation of causal and logical registers.9 Why does the example of the machine-as-symbol involve a conflation of cause and logic? Because it mistakenly construes a description of the machine’s causally coordinated physical components as a warrant for saying that therefore the machine will continue to function in this way in the future. The treating of the machine as a “symbol” of its functioning is turned into an inference about its future states.  This conflation of a causal with a logical determination involves the invention of a special kind of entity: what Wittgenstein calls, at §194, the “ideally rigid machine.” As with the rule-following example, there is here the supposition that a causal event—encountering a signpost, a teacher’s pointing, a machine’s symbolizing in the picture of its components the totality of its applications—could rigidly determine what it means to go on in the same way, indefinitely. This ideal of super-rigidity, which Wittgenstein imagines at §218 as being guided by “rails invisibly laid to infinity,” results from what Brandom calls an “assimilation of normative compulsion and causal compulsion [such that] if the normative ‘must’ were a kind of causal ‘must,’ it would have to be a puzzling, super-rigid sort” (MIE, 14). Linking the examples of rule-following and the machine-as-symbol, John McDowell explains how Wittgenstein gets us to picture “following a rule as the operation of a super-rigid yet (or perhaps we should say ‘hence’) ethereal machine”10 The reason McDowell considers swapping in “hence” for “yet,” I take it, is because the sort of super-rigidity imagined as impervious to unforeseen malfunction would have to be an ideal (“ethereal”) one.11

Shortly after the example of the machine Wittgenstein arrives at the well-known statement at §201 of “our paradox”:

This was our paradox: no course of action could be determined by a rule, because every course of action can be made out to accord with the rule. The answer was: if everything can be made out to accord with the rule, then it can also be made out to conflict with it. And so there would be neither accord nor conflict.

It can be seen that there is a misunderstanding here from the mere fact that in the course of our argument we give one interpretation after another; as if each one contented us at least for a moment, until we thought of yet another standing behind it. What this shows is that there is a way of grasping a rule which it not an interpretation, but which is exhibited in what we call ‘obeying the rule’ and ‘going against it’ in actual cases.

The “paradox” Wittgenstein elaborates here has a number of distinct aspects. The fact that “no course of action could be determined by a rule, because every course of action can be made out to accord [Übereinstimmung] with the rule” is what Brandom calls the “gerrymandering” aspect of the paradox; that is, the way no finite stretch of behavior can generate the distinction between correct and incorrect performance in relation to a rule (MIE, 20-21).  At the same time, to “give one interpretation after another; as if each one contented us at least for a moment, until we thought of yet another standing behind it,” generates a related but distinct “regress” problem. Anandi Hattiangadi describes the regress problem like this: “if understanding the content of an expression requires grasping a rule for its use, we can always wonder what understanding the rule consists in.”12 But there is another aspect to the paradox elaborated at §201: the way Wittgenstein seems to want to get out of the illusion that there is a paradox at all, while at the same time convincingly imagining the (however fictive) problems of gerrymandering and regress.  If the second paragraph of §201 states that there is a way of following a rule that is not an interpretation, then this means that there is a way of following a rule that simultaneously provides a criterion for understanding what it means to follow the rule—that a norm is somehow implicit in the practice of following a rule—and so would be immune both to regress (rules for interpreting rules for interpreting rules) and to gerrymandering (no finite stretch of behavior can generate the distinction between correct and incorrect performance in relation to a rule).  The change in tone between paragraph one and paragraph two at §201—the way the first imagines a paradox that the second diagnoses as a “misunderstanding” at work in the setting up of the problem—seems designed to show both how we can become bewitched by worries like regress and gerrymandering, and at the same time show the “paradox” isn’t a paradox at all.

II.

If the illusion that there is a paradox in Wittgenstein’s rule-following example arises from a conflation of the causal and the logical, it invites comparison with what Sellars called the “myth of the given,” so important for later pragmatists like Rorty and Brandom.  What is the myth of the given?  It is the myth that causal determinations (what the empiricist might call “impressions”) could serve as a foundation for discourse in the space of reasons.  The critique of the myth of the given follows from the view that brute deliverances of the senses alone are not enough to get a language game going; indeed, that the capacity even to make non-inferential observation reports already depends on the normative ability to subsume experienced particulars under general concepts.13 While, say, parrots can be trained to be reliable differential responders, parrots (or ravens, to use an example dear to both Edgar Allan Poe and Nelson Goodman) do not then go on to keep playing the game of giving and asking for reasons.14 Parrots are in this sense like a piece of iron that “responds” to rain by rusting, or a photo-electric cell wired up to tape recorder which says “red!” when we shine red light on it.15 “Merely reliably responding differentially to red things,” as Brandom puts it

is not yet being aware of them as red. Discrimination by producing reliable responses (as a machine or a pidgin might do) sorts the eliciting stimuli, and in that sense classifies them. But it is not yet a conceptual classification…[genuine discursive practice arises] when that responsive capacity or skill is put into the larger context that includes treating the responses as inferentially significant: As providing reasons for making other moves in the language game, and as themselves potentially standing in need of reasons that could be provided by making still other moves.16

Brandom’s description of the criteria by which we recognize that a language game is being played—that there is discourse, that things are being said—depends on an ability to make reliable responses that are inferentially significant.  Beyond the mere “sorting” of  stimuli, to say something is to endorse other inferences one’s claim commits one to.  To be able to say things in this sense—to be able to take responsibility for one’s claims—is for Brandom what is means to “go on” in a language game.

Brandom’s distinction between reliable differential responses and inferentially significant claimings sounds close to the distinction made vivid in the “beetle-in-the-box” thought experiment, particularly where Wittgenstein writes: “if we construe the grammar of the expression of sensation on the model of ‘object and designation’ the object drops out of consideration as irrelevant.”  May we then say that Wittgenstein here describes, and criticizes, a version of the myth of the given? That is, in criticizing the view that the word “beetle” points to whatever is in the box, are you offering a version of the claim that causal occurrences are not themselves enough to get a language game going? Can we take the voice at §293 to be saying that you cannot share an observation if it does not already comport to features of public criteria—let’s call these justifications in the logical space of reasons—and so therefore it is wrong to think that you are pointing to something private when you use the word “beetle” (or “pain”)?17

While I think there’s a genuine point of affinity here, I do not think the beetle example asks us to think of meaningfulness as arising from a giving and asking for reasons. That is, I don’t think Wittgenstein treats meaning as necessarily a matter of inference.  But then what is Wittgenstein’s understanding of the public criteria in relation to which one is able to go on in the same way in a language game?  Consider the question in light of the other examples discussed in the last section. The error Wittgenstein diagnoses in the “ideally rigid machine” is the tendency to treat a picture of the machine’s causal functioning as a symbol for its going on in the same way, indefinitely. To treat the determinism implicit in the image of the cogs and gears as a warrant for saying “therefore” the machine will go on in this way in the future is to imagine a set of rails extending a pattern unwaveringly into the future in accord with instructions specified in a rule.  Like the confusion brought on by the myth of the given, here the causal relations that seem “symbolically” communicated are imagined to provide a logical determination. Consider this now in relation to rule-following: the teacher asks the student to continue the series “2,4,6,8…” and the pupil “follows” the rule such that, at 1000, he starts to write “1000, 1004, 1008…” , claiming that he is going on in the same way. So the teacher thinks it means one thing to follow the rule and the pupil another. That they both take their interpretations of “add 2” to be in accord with the rule means that the rule’s “standing there like a sign post” opens up the possibility for a number of competing interpretations of what it means to go on: rules for interpreting rules for interpreting rules…(regress); and a complimentary dilemma in which any finite stretch of observed performance can be made to accord with any rule (gerrymandering).  But for Wittgenstein the point of this apparent paradox is not, as Brandom would have it, to provide fresh theorizing for making explicit how norms might be implicit in a practice, but rather to say simply that there “must be a way of obeying a rule that is not an interpretation.”18 So while both Brandom and Wittgenstein, in their different ways, link meaning to use, and both are mindful that, as Brandom puts it, “the ‘must’ of justification or good inference is not the ‘must’ of causal compulsion,” (MIE, 12) and while each seems to think there needs to be some public criteria by which moves in a language game can be treated as meaningful, there is nevertheless a crucial difference between them.

This difference comes down to their different ways of understanding the idea that a norm could be implicit in a practice.  Brandom describes his own approach as a

rationalist pragmatism [that] gives pride of place to practices of giving and asking for reasons, understanding them as conferring conceptual content on…states suitably caught up in those practices [and] it is a rationalist expressivism in that it understands expressing something, making it explicit, as putting it in a form in which it can serve as both premise and conclusion in inferences. Saying and thinking that things are thus-and-so is undertaking a distinctive kind of inferentially articulated commitment…and undertaking responsibility to entitle oneself to that commitment.” (AR, 11)

Again, for Brandom an instance of saying is meaningful (that is, conceptually contentful) just insofar as it can “serve as both premise and conclusion in inferences,” and this claim is part of Brandom’s rationalism.  But his rationalism is also a pragmatism because our entitlement to our sayings is matter of “mastering [their] inferential use,” such that one’s know-how includes knowing “what else one would be committing oneself to by applying the concept” (AR, 11).  So our practices implicitly harbor norms which logical vocabulary then expresses and makes explicit.19 Given the way Brandom enlists the later Wittgenstein for his project of inferentialism—the way the account of rule-following in the Investigations involves an effort to understand how a norm can be implicit in a practice—it is tempting to see Wittgenstein as holding the view that an inferential relation remains implicit in the practice of following rules in actual cases.  But while Wittgenstein is, like Brandom, concerned with the public criteria by which our sayings are taken as meaningful (the question of the conditions under which one is entitled to say something), nothing in his examples indicate that he thinks inference is the particular form such criteria take.  But if not the relations of inferential compatibility that confer entitlement through what a given claim commits one to—of what one becomes responsible for in saying something—then what exactly is Wittgenstein’s account of criteria?  If Brandom thinks that “going on in the same way” means knowing how to make moves in the language game of giving and asking for reasons (to be able to play that game is to be able to go on), what exactly is Wittgenstein’s account of how it is that we are able to go on in a language game?

III.

In what remains, I want to try to unpack a bit more the idea that Wittgenstein’s understanding of the conditions under which one is able to go on in a language are something other than Brandom’s “inferentialist” account.20 Wondering how and to what extent “Wittgenstein’s writing is essential to his philosophy,” Stanley Cavell suggests that in the Philosophical Investigations Wittgenstein provides “literary conditions for its philosophical aims.”21 While not explicitly responding to Brandom’s way of reading the Investigations, Cavell’s notion of literary conditions might be thought of as an alternate approach to understanding how Wittgenstein imagines norms as implicit in practice. That is, the way he shows how certain philosophical problems take on a life of their own even as they are being dismissed as false problems; and the way the captivating imagining of the problem is itself the condition under which some particular form of saying becomes possible.  Getting one’s readers to grasp what seems to be a paradox about rule following, only to switch gears and show that the “paradox” is really a kind of philosophical mirage, is to do philosophy (or perhaps to try to get out of the compulsion to do philosophy) in a way such that an entitlement to one’s claims is established on, as it were, literary grounds.  While not quite making Cavell’s claim for specifically literary conditions, John McDowell says something similar when he writes that in passages like the second paragraph of §201 a “mythology is wrung from us, in our need to avoid the paradox of the first paragraph, only because we fall into [a] misunderstanding,” since the paradox imagined at §201 is “not compulsory”; that is, it “starts with something Wittgenstein aims to show up as a mistake.”22 Instead of treating the paradox at §201 as the discovery of a philosophical problem about rule following, we should treat it as a convincingly imagined mythology, and one that is dispelled in the idea that there must be a way of following a rule that is not an interpretation.

In an “Introductory Note” added to “The Investigations’ Everyday Aesthetics of Itself,” Cavell describes his appeal to literary conditions as part of his “learning to ask for further conditions of the [Investigations’] existence—its form as fragments, its palette of terms of criticism, its sparseness of theoretical terms, as if every term of ordinary language can be shown to harbor the power of a theoretical term.”23 The finding of a theoretical term “harbored” in a term of ordinary language is imagined as the way the form and order of the Investigations is a function of its “patently, all but ostentatiously, literary gestures”; gestures that are, Cavell says, the work’s “grounds.”  This merger of the “ostentatious” and the “ordinary” in Wittgenstein’s style is a matter for Cavell, not only of not fixing what isn’t broken (of introducing a theoretical vocabulary where one is not needed), but of unslotting words from theoretic trajectories with the aim of restoring to them their full range of expression.  The series of situations or moods brought out in the Investigations—say, the moment when it appears as though any sequence of numbers could be made out to accord with the rule; or the way an infinite regress of rules for interpreting rules opens up when we raise the question of what it means to apply a rule; or the way the image of the machine, as a picture or symbol of its functioning, is thought to count as a determination of its future states; or the way the thing we call “beetle” could be continually changing, or even be absent—each of these literary gestures lure us into entertaining the Investigations’ peculiar fictions, and might be treated as so many conditions of possibility for Wittgenstein’s philosophical aims. We can then begin to hear the paradox at §201 as very particularly voiced, as if it were a matter of conjuring a symptom that something had gone wrong in our way of wording things.  To refuse to treat the “paradox” at §201 as requiring of us an effort to provide a theory of how norms are implicit in practices is to discover that what was taken to be a genuine problem is really a kind of fantasy, brought on by the way we’re using words like “rule,” “course,” “determine,” “action,” and “accord.”

The rhythm or interval between the first and second paragraphs of §201 might then be generalized as a way of imagining what it means to go on in a language game; not as an inferential game of giving and asking for reasons, but as a method for exemplifying the way certain “problems arise when language goes on holiday” (PI, §38).  Wittgenstein’s method involves arranging words such that the staging and diagnosing of certain philosophical mythologies are at the same time experiments in ways of going on; ways of moving from sentence to sentence.  Later in the Investigations, Wittgenstein seems himself to want to generalize the movement of his writing: “Phrased like this, emphasized like this, heard in this way, this sentence is the first of a series in which a transition is made to these sentences” [PI §534].24 The transition occurs, not so much as a licensed inference, but as a logic the order of the sentences makes possible.

At a different point in his “Introductory Note,” Cavell asks us to think of the “perspicuousness” of the Wittgensteinian aphorism in relation to the feeling of necessity we have in the experience of formal proofs.  Following the aphorism in the Remarks on the Foundations of Mathematics, that “Proof must be perspicuous” (say, the feeling of necessity that arises in showing that the interior angles of a triangle add up to 180 degrees), Cavell thinks of perspicuousness in relation to style.  This is imagined as a “certain unity or reordering of ordinary words [as] (non-formal) moments of complete clarity, ordinary words, that is, which are not meant to line up as premises to a conclusion…Is there perhaps an ordering of words that is its own bottom line, sees to its own ground? Would we, I mean, be prepared to describe any such ordering in this way?”25 Cavell wonders what one would have to do to words to get out of them a certain kind of clarity; to ground their meaning in an order. A kind of literary tact—the sound of these words in this order—would then serve as the condition under which we are entitled to mean in our own, and find meaning in another’s, words. The sort of perspicacity striven for here is not a matter of lining up reasons (it would not be “formal” in the way that a proof is formal), but of an attunement to arrangements of words in specific contexts.26

We might say this is a matter of arriving at a method that is both attuned to the conditions of a language game and experimental in the sense of trying out new arrangements that push against the edges of a language game.  To borrow a phrase from Walt Whitman’s own extravagant experiment in ordinary language, we could think of this as being “both in and out of the game and watching and wondering at it.”27 A method both beholden to the conditions under which meaning can occur (the new saying can be shared with others), and yet at the same time watching and wondering at those conditions as they stretch to accommodate other orders, is strikingly close to what Cavell calls, in a different essay, “speaking inside and outside language games” as “a case of speaking without justification but not wrongfully.”28 Being both in and out of the game such that one’s “rights” to one’s claims are ensured in ways other than the narrowly justificatory is exactly what seems to be expressed in the famous lines at §217: “ ‘How am I able to obey a rule?’—if this is not a question about causes, then it is about the justification for my following the rule in the way that I do. If I have exhausted the justifications I have reached bedrock, and my spade is turned. Then I am inclined to say: ‘This is simply what I do’.” Fusing the rule-following paradox, the conflation of cause and logic made vivid in the machine-as-symbol example, and the relation of “private” language to public criteria, the exhaustion of justifications—the spade reaching bedrock—wrings from the speaker at §217 an inclination to say something that cannot be made any more explicit.  The exhaustion of justifications bottoms out in a way of putting things.

In a section called “The Private Language Argument,” one of forty-six such subheadings in the last section of The Claim of Reason, Cavell says that Wittgenstein imagines the possibility of a private language in order to “release the fantasy expressed in the denial that language is something essentially shared”; and that these sections from the Investigations are “peculiarly colored by the tone of someone allowing a fantasy to be voiced.”29 I take it that what Cavell calls “voicing” is the condition underwriting the claim that language is something essentially shared, and that this occurs not only in the fantasy scenes Cavell finds in the sections on private language, but throughout the Investigations.  It is as if the fantasy of private ostension arose as an effect of something done at the level of composition, such that what is voiced—what is made perspicuous—is the distinction between a language that has lapsed into default settings and a language kept alive in its ordinariness.  A compelling dream like the beetle vignette, in which the thing to which we take ourselves to be pointing were itself, potentially, continually changing, is then not so much the final word on the erroneousness of the idea of private language, as it is a striking way imagining of what can go wrong when we allow certain philosophical habits to lure us into believing a “problem” is a problem at all.

How exactly, on this construal of conditions, are we to understand the relation of philosophy to literature?  Or as Cavell more dramatically puts it: “Can philosophy become literature and still know itself?” (CR 496).  The placing of the word “know” in this closing sentence of The Claim of Reason makes all the difference in taking Cavell’s own philosophical aims to be what they are: namely, a re-aligning of the question of criteria, away from its presumed proximity to knowing (what are the conditions under which this cognition or experience becomes possible?), and toward something Cavell calls philosophy’s “becoming literature.”  “Know” is here given back to philosophy as a word it must do something new with, and “literature” perhaps now names an attunement to the tension between our words and their future applications.  Only by a relentless re-attunement to how we’re using our words, Cavell seems to be saying, is philosophy able to go on.  And this occurs, for Cavell and for Wittgenstein (and, I would argue, also for writers like Emerson and Whitman) through an experimenting with the different kinds of intelligibility that can be arrived at from within a shared linguistic inheritance.

With Cavell’s idea of literary conditions in mind, we might aspire to mean—in our ongoing interdisciplinary language game between literature and philosophy—something new by the “critical” in literary criticism: namely, criticism as an investigation into the conditions under which literary or philosophical works can be taken as meaningful, such that those conditions allow for a shared accounting of—genuine agreement and disagreement about—such works.  For Wittgenstein, the conditions under which such claims become intelligible do not come down to a game of giving and asking for reasons, but are a methodical heightening of the problem of taking responsibility for what we say, understood as a uniquely literary responsibility. Going on means listening closely to the conditions under which our sayings become shareable, and doing so at every turn.

Notes

1. All quotations from Wittgenstein are from Philosophical Investigations 3rd. edition, trans. G.E.M. Anscombe (Edgewood Cliffs: Prentice Hall, 1958), and indicated by section number. My thanks to Rob Chodat, Ian Cornelius, Richard Deming, Jay Elliott, Juliet Floyd, John Gibson, Oren Izenberg, Ross Posnock, David Possen, and Ken Winkler for helping me to think about Wittgenstein, pragmatism, Cavell and other aspects of this essay.
2. Richard Rorty, “Cultural Politics and the Question of the Existence of God,” in Philosophy As Cultural Politics, (Cambridge: Cambridge University Press, 2007), 11-12.
3.Richard Rorty, “Dewey Between Hegel and Darwin,” in Truth and Progress (Cambridge: Cambridge University Press, 1998),  295-296. For Sellars on the relation of sense data to knowledge claims, see Wilfrid Sellars, Empiricism and the Philosophy of Mind, with an Introduction by Richard Rorty and a study guide by Robert Brandom (Cambridge: Harvard University Press, 2007 ) 13-25 [hereafter cited as EPM]); and Willem A. deVries and Timm Triplett, Knowledge, Mind and the Given: Reading Wilfrid Sellars’s “Empiricism and the Philosophy of Mind” (Indianapolis: Hackett, 2000) xxx-xxxii, and 76-77.
4.Rorty, Richard. “Cultural Politics and God,” 10-11. In an earlier essay called “Dewey’s Metaphysics,” Rorty writes that Dewey makes the mistake of “crossing the line …between causal [relations] and the self-conscious beliefs and inferences they make possible,” cautioning that “nothing is to be gained for an understanding of human knowledge by running together the vocabularies in which we describe the causal antecedents of knowledge with those in which we offer justifications of our claims to knowledge.” See Richard Rorty, Consequences of Pragmatism, Minneapolis: Minnesota University Press, 1982, 82.
5.Explaining the need for a deduction of the categories of the Understanding from logical functions, Kant appeals in the Critique of Pure Reason to an explicitly jurisprudential language: “Jurists, when they speak of entitlements and claims, distinguish in a legal matter between the question about what is lawful [quid juris] and that which concerns the fact [quid facti], and since they demand proof of both , they call the first, that which is to establish the entitlement of the legal claim, the deduction” (CPR A 84 / B 116).
6.Robert Brandom, Making It Explicit: Reasoning, Representing & Discursive Commitment. (Cambridge: Harvard University Press, 1994), 13. (Henceforth cited parenthetically as MIE).
7.Stanley Cavell, “The Investigations’ Everyday Aesthetics of Itself,” in The Literary Wittgenstein. eds. John Gibson and Wolfgang Hummer. (London: Routledge, 2004), 20.
8.Cavell, “The Investigations’ Everyday Aesthetics,” 21.
9.For more on the conflation of the causal and the logical in Wittgenstein’s “ideally rigid machine,” see Meredith Williams, Blind Obedience (London: Routledge,  2010), 155-160.
10.John McDowell, “Wittgenstein on Following a Rule,” in Mind, Value & Reality. (Cambridge: Harvard University Press, 1998), 230-231.
11.In his powerfully imaginative interpretation of Wittgenstein’s rule-following examples, Saul Kripke highlights the distinction between the “machine as abstract program…and the actual physical machine, which is subject to breakdown,” comparing Wittgenstein’s machine-as-symbol to Turing machines. Turing machines are “ideal” in the sense that they stand for all a machine needs to be able to do in principle to carry out program-like instructions. See Kripke, Wittgenstein on Rules and Private Language, 35-36 n.24 [henceforth cited as WRP]. For a detailed explanation of Turing machines, see Charles Petzold, The Annotated Turing: A Guided Tour through Alan Turing’s Historic Paper on Computability and the Turing Machine (Indianapolis: Wiley Publishing, 2008).
12.Anandi Hattiangadi. “Making It Implicit: Brandom on Rule Following,” in Philosophy and Phenomenological Research, (Vol.66, No.2, 2003), 427.
13.For Brandom’s unpacking of this Sellarsian point see EPM, 123-130; 139-144. For a helpful account of how Brandom (glossing Sellars) construes the myth of the given as hinging specifically on the move from sense data to noninferential beliefs, see Danielle Macbeth, “Inference, Meaning and Truth in Brandom, Sellars and Frege,” in Reading Brandom: On Making It Explicit, eds. Bernhard Weiss and Jeremy Wanderer (London: Routledge, 2010), 198-201.
14.For Goodman’s arguments about how the predicate “grue” radicalizes Hume’s skeptical solution to the problem of justifying induction, see “The New Riddle of Induction” in Fact Fiction and Forecast (Cambridge: Harvard University Press, 1983). For the relevance of grue to the gerrymandering problem in Wittgenstein’s rule-following paradox, see Kripke, WRP 20-21; 58-59. In the case of Poe’s poem “The Raven,” a confusion of reliable responses with genuine sayings is precisely the narrator’s predicament: the bird repeats by rote a sound the narrator takes to be a series of replies to his increasingly desperate and tormented questions about his “lost Lenore.” For more on how “The Raven” animates the problem of “going on” in this way see my Experience and Experimental Writing From Emerson to the Jameses (Oxford University Press, forthcoming).
15.The iron example is taken from MIE, 87-88. The example of the photoelectric cell hitched up to a tape recorder—the point of which is that we “cannot readily imagine continuing a conversation with [such a] gadget”—is from Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1979) 189.
16.Robert Brandom, Articulating Reasons: An Introduction to Inferentialism. (Cambridge: Harvard University Press, 2000), 17. (Henceforth cited parenthetically as AR)
17.Commenting on John McDowell’s construal of the “myth of the given,” Avner Baz makes just this point when he writes: “justification by an appeal to an unarticulated something [the ‘myth’ that our words could normatively relate to something essentially foreign to them] is no justification at all [and so] the unarticulated something…drops out of consideration as irrelevant, just like Wittgenstein’s ‘private object,’ and can make no contribution to the intelligibility…of our thoughts (judgments, beliefs).” Avner Baz, “On When Words Are Called For: Cavell, McDowell, and the Wording of the World,” Inquiry, 46: 4, 473-500, 2003, 477-78. See also John McDowell, Mind and World (Cambridge :Harvard University Press, 1994), 3-23.
18.Cautioning against Brandom’s belief that Wittgenstein’s way of imagining “norms implicit in practice [requires] work on the part of philosophers to uncover and make explicit,” John McDowell thinks that “there is no reason to suppose there must be a level of normativity below the level at which linguistic practice is described in terms of explicitly using this or that concept, and it is no concern of Wittgenstein’s to suggest that there is.” (John McDowell, “How Not To Read Philosophical Investigations,” in The Engaged Intellect [Cambridge: Harvard University Press, 2009], 98; 108; my emphasis).
19.Brandom sometimes uses the example of conditionals to illustrate what he means by expressivism: “Before introducing the [logical vocabulary of conditionals] one can do something, namely, endorse an inference. After introducing the conditional, one can now say that the inference is a good one. The expressive role of the conditional is to make explicit, in the form of a claim, what before was implicit in our practice of distinguishing some inferences as good.” (AR 81) For an account of conditionals as introduced into inferential practice as an “algorithmic elaboration,” see Brandom, Between Saying and Doing: Toward an Analytic Pragmatism (Oxford: Oxford University Press, 2008) 31-55.
20.Nor is it a matter, in my view, of radicalizing the problem of induction along the lines of Kripke’s claim that Wittgenstein’s rule-following paradox stands in need of a Hume-like “skeptical solution.” See Kripke, WRP, 66-69. For a compelling, and somewhat different, account of Kripke’s construal of the rule-following paradox—as a question of the “right” by which fellow rule-followers “taken into the community” [WRP, 110] and the “normative silence” of persons—see Oren Izenberg, Being Numerous: Poetry and the Ground of Social Life (Princeton: Princeton University Press, 2010) 94-95.
21.Stanley Cavell, “Introductory Note,” in The Literary Wittgenstein, 20.
22.John McDowell,  “Wittgenstein on Following a Rule,” 231; 236.
23.Cavell, “Introductory Note,” in The Literary Wittgenstein, 20.
24.The context of the discussion at §534 is that of considering two ways of understanding the meaning of a sentence. On the one hand, understanding a sentence can be thought of on the model of one sentence being “replaced by another which says the same”; on the other, “only by these words in these positions” [§531]. For Wittgenstein, what it means to understand a sentence includes aspects of both [§532]. I appeal to §531-534 in conjunction with §201 because the way Wittgenstein shows the “paradox” to be a mirage depends, I argue, on his ordering sentences in such a way that transition from one to the other is not a matter of inferential compatibility relations, but a function of what Cavell calls “literary conditions.”
25.Stanley Cavell,  ““The Investigations Everyday Aesthetics,” in The Literary Wittgenstein, 21.
26.For a nuanced account of the relation between conditions, criteria and context in Cavell’s thought, see Steven G. Affeldt, “The Ground of Mutuality: Criteria, Judgment, and Intelligibility in Stephen Mulhall and Stanley Cavell,” in European Journal of Philosophy, 6:1, 1-31.
27.Walt Whitman,  “Song of Myself,” in Leaves of Grass. (New York: Library of America), 30.
28.Stanley Cavell,  “The Argument of the Ordinary,” in Conditions Handsome and Unhandsome 83; 69; henceforth cited as CHU. Cavell makes this point specifically in relation to the way the “interlocutor” in the Investigations “poses a great task, the continuous task of Wittgenstein’s prose, oscillating between vanity and humility. Skepticism appears in [the PI] as one of the voices locked in this argument, not as a solution or conclusion” (Cavell, CHU, 83).
29.Stanley Cavell, The Claim of Reason: Wittgenstein, Skepticism, Morality and Tragedy. (New York: Oxford, 1979), 344. (Henceforth cited as CR in the text). Anyone who has read even a little Cavell knows that he remains interested (to put it mildly) in such fantasies: why and how they persist, what their persistence tells us about human life and the self, and what they tell us about the desire for something like philosophy (issues Cavell takes up, in all of his work, under the theme of skepticism). By placing my emphasis here on how Cavell describes Wittgenstein’s way of writing as indispensible to his philosophical aims, I certainly do not mean to suggest that Cavell is concerned to read Wittgenstein as dissolving, once and for all, such fantasies. For Cavell’s positioning of his reading of Wittgenstein on rules in relation to Kripke’s claim that §201 sets up a problem in need of a “skeptical solution,” and how this is matter of what it means to “go on,” see Cavell, CHU, 69-96.

 

On Going On: Rules, Inferences and Literary Conditions

Both in and out of the game and watching and wondering at it.
—Walt Whitman

At §293 in Philosophical Investigations, Ludwig Wittgenstein imagines the following scenario:

Suppose everyone has a box with something in it: we call it a ‘beetle’. No one can look into anyone else’s box, and everyone says he knows what a beetle is by looking at his beetle.—Here it would be quite possible for everyone to have something different in his box. One might even imagine the thing constantly changing.—But suppose the word ‘beetle’ had a use in these people’s language?—If so it would not be used as the name of a thing. The thing in the box has no place in the language-game at all; not even as a something: for the box might even be empty.—No, one can ‘divide through’ by the thing in the box; it cancels out, whatever it is. That is to say: if we construe the grammar of the expression of sensation on the model of ‘object and designation’ the object drops out of consideration as irrelevant.1

Richard Rorty says Wittgenstein’s thought experiment boils down to the question: “Can a descriptive term have a sense if its application is regulated by no public criteria”; a question to which Wittgenstein—or at least the voice at §293—appears to give the answer no. Since the beetle example is about the “grammar of the expression of sensation,” Rorty says that for Wittgenstein words like “pain” have a sense only as long as they “are not treated as a name of something whose presence or absence swings free of all differences in environment or behavior.”2 To talk about pain as “qualia” or as a raw feel is just to add a nuance to a language game; one that would allow us, as Rorty puts it, to “disjoin pain from pain behavior” linguistically.  But adding nuances to a language game in this way does not mean the word “pain” (or “beetle”) points to some inner quale; i.e., its use is not a matter of private ostension.

The question of whether a term can have a meaning if its application is regulated by no public criteria, though, arises even more clearly from a different set of examples in the Philosophical Investigations: examples pertaining to the notorious problem of “rule following.”  In these examples, questions about what it means to follow a rule become questions about how words in general are subject to criteria of correct and incorrect application.  And such questions, I want to suggest here, invite comparison with some of the basic arguments of what is sometimes called analytic pragmatism.  Impressed by Wilfrid Sellars’ insight that in order properly to be saying something we have to take ourselves as being in the “logical space of reasons,” analytic pragmatists like Rorty and his student Robert Brandom, (despite their acknowledged debt to thinkers like William James and John Dewey) mount a fundamental criticism of the “classical” phase of pragmatism. In his essay “Dewey Between Hegel and Darwin,” Rorty says Dewey’s naturalism amounts to a blurring of the “distinction between the question ‘What causes our beliefs?’ and the question ‘What justifies our beliefs?’”3 As Rorty cautions in a later essay, we should not “treat the causal ability of certain events to produce non-inferential beliefs…as a justification for [the] holding [of] those beliefs.”4 If Dewey thought the human animal moved on a smooth continuum from causes to justifications—from nature to norms, we might say—analytic pragmatists inherit from Sellars an apprehension about blurring the line between sensory events and justified belief.  What Wittgenstein says at §293 about the public criteria by which language use can be taken as meaningful is comparable, in ways that matter, to the analytic pragmatists’ telling us that in order to be saying anything at all we must be able to justify our claims.

In wanting to identify some affinities between Wittgenstein’s later thought and analytic pragmatism, I will focus here on sections of the Philosophical Investigations dealing with rule-following and private language.  The first thing I want to claim is that Wittgenstein and Brandom each remain committed to versions of what Kant called the quid juris question: by what “right” are we entitled to take experience, knowledge, belief etc. as meaningful—an issue Kant tried to satisfy through transcendental argument.5 The second, related thing I want to claim is that the sort of inferentialism Brandom is concerned to elaborate, as well as Wittgenstein’s exploration of the rule-following paradox, are each bound up with what it means to “go on in the same way”: on the one hand, the role of ‘etc’ or ‘…’ in Wittgenstein’s number series examples; on the other, Brandom’s distinction between, say, parrots that respond reliably and differentially to discrete stimuli and discursive creatures that can go on in the game of giving and asking for reasons.  I think Brandom and Wittgenstein basically agree that, as Brandom puts it, “intentionally contentful states and acts have an essentially normative pragmatic significance.”6

But is Wittgenstein’s idea of the criteria by which we take our talk to be meaningful a matter of “inference”? And if it isn’t, then what exactly are the normative criteria in relation to which we recognize what it means to “go on” in a given language game?  In a third section I consider these questions in light of Stanley Cavell’s claim that in the Philosophical Investigations Wittgenstein provides the literary conditions of [his] philosophical aims.”7 What exactly does Cavell mean by “literary conditions”? Do literary conditions have their own forms of entitlement?  Would such conditions—say, Wittgenstein’s particular scene-setting, thought experiments, aphorisms, and dialogues—amount to an alternative form of justification?  Could a tactful or artful (or beguiling or captivating or worrisome) ordering of words—what we might simply call a style—itself generate the criteria for claiming?  How exactly can, as Cavell puts it, “an ordering of words [be] its own bottom line, [and] see to its own ground?”8 Considered next to Brandom’s normative pragmatics of inference, Cavell, I try to show, finds in the Investigations a kind of claiming inseparable from a way of writing.

I.

At §143 we find Wittgenstein wanting to “investigate the language game” in which a pupil is taught to follow the rule of adding 2. The investigator will be confident the pupil understands the rule when the pupil starts “going on to write down [the series] independently.” But at §146 the investigator begins to wonder just what is to count as the pupil’s having “got the system.”  For example, what if the pupil were consistently to add 2 and then, upon reaching 1000, start adding 4 (996, 998, 1000, 1004, 1008, 1012); and in doing so thinks she has gone on in accordance with the rule?  Here we run into something that looks like a paradox. As Wittgenstein puts it, the paradox involves the way “we can think of more than one application of an algebraic formula and every type of application can in turn be formulated algebraically.” The rule itself, in other words, is not enough to determine what it will mean to go on in the same way for an indefinite number of future applications. The dilemma is neatly expressed at §85 in the aphorism “a rule stands there like a signpost,” which I take to mean that there is nothing about the rule that dictates how it is to be interpreted, and so there is room for doubt about what it would mean to apply the rule correctly.

A little further on, we find Wittgenstein setting up a similar problem. At §193 he imagines a

machine symbolizing its action [such that] the action of a machine…seems to be there in it from the start. What does that mean?—If we know the machine, everything else, that is its movement, seems to be already completely determined. We talk as if these parts could only move in this way, as if they could not do anything else [and in doing so] we use a machine, or the drawing of a machine, to symbolize a particular action of the machine. For instance, we give someone such a drawing and assume that he will derive the movement of the parts from it (just as we can give someone a number by telling him that it is the twenty-fifth in the series 1, 4, 9, 16, . . . .).

As with the number series example, there is here a question about what it means to move from a formula to the derivation of a series.  But here the move occurs more explicitly as a conflation of causal and logical registers.9 Why does the example of the machine-as-symbol involve a conflation of cause and logic? Because it mistakenly construes a description of the machine’s causally coordinated physical components as a warrant for saying that therefore the machine will continue to function in this way in the future. The treating of the machine as a “symbol” of its functioning is turned into an inference about its future states.  This conflation of a causal with a logical determination involves the invention of a special kind of entity: what Wittgenstein calls, at §194, the “ideally rigid machine.” As with the rule-following example, there is here the supposition that a causal event—encountering a signpost, a teacher’s pointing, a machine’s symbolizing in the picture of its components the totality of its applications—could rigidly determine what it means to go on in the same way, indefinitely. This ideal of super-rigidity, which Wittgenstein imagines at §218 as being guided by “rails invisibly laid to infinity,” results from what Brandom calls an “assimilation of normative compulsion and causal compulsion [such that] if the normative ‘must’ were a kind of causal ‘must,’ it would have to be a puzzling, super-rigid sort” (MIE, 14). Linking the examples of rule-following and the machine-as-symbol, John McDowell explains how Wittgenstein gets us to picture “following a rule as the operation of a super-rigid yet (or perhaps we should say ‘hence’) ethereal machine”10 The reason McDowell considers swapping in “hence” for “yet,” I take it, is because the sort of super-rigidity imagined as impervious to unforeseen malfunction would have to be an ideal (“ethereal”) one.11

Shortly after the example of the machine Wittgenstein arrives at the well-known statement at §201 of “our paradox”:

This was our paradox: no course of action could be determined by a rule, because every course of action can be made out to accord with the rule. The answer was: if everything can be made out to accord with the rule, then it can also be made out to conflict with it. And so there would be neither accord nor conflict.

It can be seen that there is a misunderstanding here from the mere fact that in the course of our argument we give one interpretation after another; as if each one contented us at least for a moment, until we thought of yet another standing behind it. What this shows is that there is a way of grasping a rule which it not an interpretation, but which is exhibited in what we call ‘obeying the rule’ and ‘going against it’ in actual cases.

The “paradox” Wittgenstein elaborates here has a number of distinct aspects. The fact that “no course of action could be determined by a rule, because every course of action can be made out to accord [Übereinstimmung] with the rule” is what Brandom calls the “gerrymandering” aspect of the paradox; that is, the way no finite stretch of behavior can generate the distinction between correct and incorrect performance in relation to a rule (MIE, 20-21).  At the same time, to “give one interpretation after another; as if each one contented us at least for a moment, until we thought of yet another standing behind it,” generates a related but distinct “regress” problem. Anandi Hattiangadi describes the regress problem like this: “if understanding the content of an expression requires grasping a rule for its use, we can always wonder what understanding the rule consists in.”12 But there is another aspect to the paradox elaborated at §201: the way Wittgenstein seems to want to get out of the illusion that there is a paradox at all, while at the same time convincingly imagining the (however fictive) problems of gerrymandering and regress.  If the second paragraph of §201 states that there is a way of following a rule that is not an interpretation, then this means that there is a way of following a rule that simultaneously provides a criterion for understanding what it means to follow the rule—that a norm is somehow implicit in the practice of following a rule—and so would be immune both to regress (rules for interpreting rules for interpreting rules) and to gerrymandering (no finite stretch of behavior can generate the distinction between correct and incorrect performance in relation to a rule).  The change in tone between paragraph one and paragraph two at §201—the way the first imagines a paradox that the second diagnoses as a “misunderstanding” at work in the setting up of the problem—seems designed to show both how we can become bewitched by worries like regress and gerrymandering, and at the same time show the “paradox” isn’t a paradox at all.

II.

If the illusion that there is a paradox in Wittgenstein’s rule-following example arises from a conflation of the causal and the logical, it invites comparison with what Sellars called the “myth of the given,” so important for later pragmatists like Rorty and Brandom.  What is the myth of the given?  It is the myth that causal determinations (what the empiricist might call “impressions”) could serve as a foundation for discourse in the space of reasons.  The critique of the myth of the given follows from the view that brute deliverances of the senses alone are not enough to get a language game going; indeed, that the capacity even to make non-inferential observation reports already depends on the normative ability to subsume experienced particulars under general concepts.13 While, say, parrots can be trained to be reliable differential responders, parrots (or ravens, to use an example dear to both Edgar Allan Poe and Nelson Goodman) do not then go on to keep playing the game of giving and asking for reasons.14 Parrots are in this sense like a piece of iron that “responds” to rain by rusting, or a photo-electric cell wired up to tape recorder which says “red!” when we shine red light on it.15 “Merely reliably responding differentially to red things,” as Brandom puts it

is not yet being aware of them as red. Discrimination by producing reliable responses (as a machine or a pidgin might do) sorts the eliciting stimuli, and in that sense classifies them. But it is not yet a conceptual classification…[genuine discursive practice arises] when that responsive capacity or skill is put into the larger context that includes treating the responses as inferentially significant: As providing reasons for making other moves in the language game, and as themselves potentially standing in need of reasons that could be provided by making still other moves.16

Brandom’s description of the criteria by which we recognize that a language game is being played—that there is discourse, that things are being said—depends on an ability to make reliable responses that are inferentially significant.  Beyond the mere “sorting” of  stimuli, to say something is to endorse other inferences one’s claim commits one to.  To be able to say things in this sense—to be able to take responsibility for one’s claims—is for Brandom what is means to “go on” in a language game.

Brandom’s distinction between reliable differential responses and inferentially significant claimings sounds close to the distinction made vivid in the “beetle-in-the-box” thought experiment, particularly where Wittgenstein writes: “if we construe the grammar of the expression of sensation on the model of ‘object and designation’ the object drops out of consideration as irrelevant.”  May we then say that Wittgenstein here describes, and criticizes, a version of the myth of the given? That is, in criticizing the view that the word “beetle” points to whatever is in the box, are you offering a version of the claim that causal occurrences are not themselves enough to get a language game going? Can we take the voice at §293 to be saying that you cannot share an observation if it does not already comport to features of public criteria—let’s call these justifications in the logical space of reasons—and so therefore it is wrong to think that you are pointing to something private when you use the word “beetle” (or “pain”)?17

While I think there’s a genuine point of affinity here, I do not think the beetle example asks us to think of meaningfulness as arising from a giving and asking for reasons. That is, I don’t think Wittgenstein treats meaning as necessarily a matter of inference.  But then what is Wittgenstein’s understanding of the public criteria in relation to which one is able to go on in the same way in a language game?  Consider the question in light of the other examples discussed in the last section. The error Wittgenstein diagnoses in the “ideally rigid machine” is the tendency to treat a picture of the machine’s causal functioning as a symbol for its going on in the same way, indefinitely. To treat the determinism implicit in the image of the cogs and gears as a warrant for saying “therefore” the machine will go on in this way in the future is to imagine a set of rails extending a pattern unwaveringly into the future in accord with instructions specified in a rule.  Like the confusion brought on by the myth of the given, here the causal relations that seem “symbolically” communicated are imagined to provide a logical determination. Consider this now in relation to rule-following: the teacher asks the student to continue the series “2,4,6,8…” and the pupil “follows” the rule such that, at 1000, he starts to write “1000, 1004, 1008…” , claiming that he is going on in the same way. So the teacher thinks it means one thing to follow the rule and the pupil another. That they both take their interpretations of “add 2” to be in accord with the rule means that the rule’s “standing there like a sign post” opens up the possibility for a number of competing interpretations of what it means to go on: rules for interpreting rules for interpreting rules…(regress); and a complimentary dilemma in which any finite stretch of observed performance can be made to accord with any rule (gerrymandering).  But for Wittgenstein the point of this apparent paradox is not, as Brandom would have it, to provide fresh theorizing for making explicit how norms might be implicit in a practice, but rather to say simply that there “must be a way of obeying a rule that is not an interpretation.”18 So while both Brandom and Wittgenstein, in their different ways, link meaning to use, and both are mindful that, as Brandom puts it, “the ‘must’ of justification or good inference is not the ‘must’ of causal compulsion,” (MIE, 12) and while each seems to think there needs to be some public criteria by which moves in a language game can be treated as meaningful, there is nevertheless a crucial difference between them.

This difference comes down to their different ways of understanding the idea that a norm could be implicit in a practice.  Brandom describes his own approach as a

rationalist pragmatism [that] gives pride of place to practices of giving and asking for reasons, understanding them as conferring conceptual content on…states suitably caught up in those practices [and] it is a rationalist expressivism in that it understands expressing something, making it explicit, as putting it in a form in which it can serve as both premise and conclusion in inferences. Saying and thinking that things are thus-and-so is undertaking a distinctive kind of inferentially articulated commitment…and undertaking responsibility to entitle oneself to that commitment.” (AR, 11)

Again, for Brandom an instance of saying is meaningful (that is, conceptually contentful) just insofar as it can “serve as both premise and conclusion in inferences,” and this claim is part of Brandom’s rationalism.  But his rationalism is also a pragmatism because our entitlement to our sayings is matter of “mastering [their] inferential use,” such that one’s know-how includes knowing “what else one would be committing oneself to by applying the concept” (AR, 11).  So our practices implicitly harbor norms which logical vocabulary then expresses and makes explicit.19 Given the way Brandom enlists the later Wittgenstein for his project of inferentialism—the way the account of rule-following in the Investigations involves an effort to understand how a norm can be implicit in a practice—it is tempting to see Wittgenstein as holding the view that an inferential relation remains implicit in the practice of following rules in actual cases.  But while Wittgenstein is, like Brandom, concerned with the public criteria by which our sayings are taken as meaningful (the question of the conditions under which one is entitled to say something), nothing in his examples indicate that he thinks inference is the particular form such criteria take.  But if not the relations of inferential compatibility that confer entitlement through what a given claim commits one to—of what one becomes responsible for in saying something—then what exactly is Wittgenstein’s account of criteria?  If Brandom thinks that “going on in the same way” means knowing how to make moves in the language game of giving and asking for reasons (to be able to play that game is to be able to go on), what exactly is Wittgenstein’s account of how it is that we are able to go on in a language game?

III.

In what remains, I want to try to unpack a bit more the idea that Wittgenstein’s understanding of the conditions under which one is able to go on in a language are something other than Brandom’s “inferentialist” account.20 Wondering how and to what extent “Wittgenstein’s writing is essential to his philosophy,” Stanley Cavell suggests that in the Philosophical Investigations Wittgenstein provides “literary conditions for its philosophical aims.”21 While not explicitly responding to Brandom’s way of reading the Investigations, Cavell’s notion of literary conditions might be thought of as an alternate approach to understanding how Wittgenstein imagines norms as implicit in practice. That is, the way he shows how certain philosophical problems take on a life of their own even as they are being dismissed as false problems; and the way the captivating imagining of the problem is itself the condition under which some particular form of saying becomes possible.  Getting one’s readers to grasp what seems to be a paradox about rule following, only to switch gears and show that the “paradox” is really a kind of philosophical mirage, is to do philosophy (or perhaps to try to get out of the compulsion to do philosophy) in a way such that an entitlement to one’s claims is established on, as it were, literary grounds.  While not quite making Cavell’s claim for specifically literary conditions, John McDowell says something similar when he writes that in passages like the second paragraph of §201 a “mythology is wrung from us, in our need to avoid the paradox of the first paragraph, only because we fall into [a] misunderstanding,” since the paradox imagined at §201 is “not compulsory”; that is, it “starts with something Wittgenstein aims to show up as a mistake.”22 Instead of treating the paradox at §201 as the discovery of a philosophical problem about rule following, we should treat it as a convincingly imagined mythology, and one that is dispelled in the idea that there must be a way of following a rule that is not an interpretation.

In an “Introductory Note” added to “The Investigations’ Everyday Aesthetics of Itself,” Cavell describes his appeal to literary conditions as part of his “learning to ask for further conditions of the [Investigations’] existence—its form as fragments, its palette of terms of criticism, its sparseness of theoretical terms, as if every term of ordinary language can be shown to harbor the power of a theoretical term.”23 The finding of a theoretical term “harbored” in a term of ordinary language is imagined as the way the form and order of the Investigations is a function of its “patently, all but ostentatiously, literary gestures”; gestures that are, Cavell says, the work’s “grounds.”  This merger of the “ostentatious” and the “ordinary” in Wittgenstein’s style is a matter for Cavell, not only of not fixing what isn’t broken (of introducing a theoretical vocabulary where one is not needed), but of unslotting words from theoretic trajectories with the aim of restoring to them their full range of expression.  The series of situations or moods brought out in the Investigations—say, the moment when it appears as though any sequence of numbers could be made out to accord with the rule; or the way an infinite regress of rules for interpreting rules opens up when we raise the question of what it means to apply a rule; or the way the image of the machine, as a picture or symbol of its functioning, is thought to count as a determination of its future states; or the way the thing we call “beetle” could be continually changing, or even be absent—each of these literary gestures lure us into entertaining the Investigations’ peculiar fictions, and might be treated as so many conditions of possibility for Wittgenstein’s philosophical aims. We can then begin to hear the paradox at §201 as very particularly voiced, as if it were a matter of conjuring a symptom that something had gone wrong in our way of wording things.  To refuse to treat the “paradox” at §201 as requiring of us an effort to provide a theory of how norms are implicit in practices is to discover that what was taken to be a genuine problem is really a kind of fantasy, brought on by the way we’re using words like “rule,” “course,” “determine,” “action,” and “accord.”

The rhythm or interval between the first and second paragraphs of §201 might then be generalized as a way of imagining what it means to go on in a language game; not as an inferential game of giving and asking for reasons, but as a method for exemplifying the way certain “problems arise when language goes on holiday” (PI, §38).  Wittgenstein’s method involves arranging words such that the staging and diagnosing of certain philosophical mythologies are at the same time experiments in ways of going on; ways of moving from sentence to sentence.  Later in the Investigations, Wittgenstein seems himself to want to generalize the movement of his writing: “Phrased like this, emphasized like this, heard in this way, this sentence is the first of a series in which a transition is made to these sentences” [PI §534].24 The transition occurs, not so much as a licensed inference, but as a logic the order of the sentences makes possible.

At a different point in his “Introductory Note,” Cavell asks us to think of the “perspicuousness” of the Wittgensteinian aphorism in relation to the feeling of necessity we have in the experience of formal proofs.  Following the aphorism in the Remarks on the Foundations of Mathematics, that “Proof must be perspicuous” (say, the feeling of necessity that arises in showing that the interior angles of a triangle add up to 180 degrees), Cavell thinks of perspicuousness in relation to style.  This is imagined as a “certain unity or reordering of ordinary words [as] (non-formal) moments of complete clarity, ordinary words, that is, which are not meant to line up as premises to a conclusion…Is there perhaps an ordering of words that is its own bottom line, sees to its own ground? Would we, I mean, be prepared to describe any such ordering in this way?”25 Cavell wonders what one would have to do to words to get out of them a certain kind of clarity; to ground their meaning in an order. A kind of literary tact—the sound of these words in this order—would then serve as the condition under which we are entitled to mean in our own, and find meaning in another’s, words. The sort of perspicacity striven for here is not a matter of lining up reasons (it would not be “formal” in the way that a proof is formal), but of an attunement to arrangements of words in specific contexts.26

We might say this is a matter of arriving at a method that is both attuned to the conditions of a language game and experimental in the sense of trying out new arrangements that push against the edges of a language game.  To borrow a phrase from Walt Whitman’s own extravagant experiment in ordinary language, we could think of this as being “both in and out of the game and watching and wondering at it.”27 A method both beholden to the conditions under which meaning can occur (the new saying can be shared with others), and yet at the same time watching and wondering at those conditions as they stretch to accommodate other orders, is strikingly close to what Cavell calls, in a different essay, “speaking inside and outside language games” as “a case of speaking without justification but not wrongfully.”28 Being both in and out of the game such that one’s “rights” to one’s claims are ensured in ways other than the narrowly justificatory is exactly what seems to be expressed in the famous lines at §217: “ ‘How am I able to obey a rule?’—if this is not a question about causes, then it is about the justification for my following the rule in the way that I do. If I have exhausted the justifications I have reached bedrock, and my spade is turned. Then I am inclined to say: ‘This is simply what I do’.” Fusing the rule-following paradox, the conflation of cause and logic made vivid in the machine-as-symbol example, and the relation of “private” language to public criteria, the exhaustion of justifications—the spade reaching bedrock—wrings from the speaker at §217 an inclination to say something that cannot be made any more explicit.  The exhaustion of justifications bottoms out in a way of putting things.

In a section called “The Private Language Argument,” one of forty-six such subheadings in the last section of The Claim of Reason, Cavell says that Wittgenstein imagines the possibility of a private language in order to “release the fantasy expressed in the denial that language is something essentially shared”; and that these sections from the Investigations are “peculiarly colored by the tone of someone allowing a fantasy to be voiced.”29 I take it that what Cavell calls “voicing” is the condition underwriting the claim that language is something essentially shared, and that this occurs not only in the fantasy scenes Cavell finds in the sections on private language, but throughout the Investigations.  It is as if the fantasy of private ostension arose as an effect of something done at the level of composition, such that what is voiced—what is made perspicuous—is the distinction between a language that has lapsed into default settings and a language kept alive in its ordinariness.  A compelling dream like the beetle vignette, in which the thing to which we take ourselves to be pointing were itself, potentially, continually changing, is then not so much the final word on the erroneousness of the idea of private language, as it is a striking way imagining of what can go wrong when we allow certain philosophical habits to lure us into believing a “problem” is a problem at all.

How exactly, on this construal of conditions, are we to understand the relation of philosophy to literature?  Or as Cavell more dramatically puts it: “Can philosophy become literature and still know itself?” (CR 496).  The placing of the word “know” in this closing sentence of The Claim of Reason makes all the difference in taking Cavell’s own philosophical aims to be what they are: namely, a re-aligning of the question of criteria, away from its presumed proximity to knowing (what are the conditions under which this cognition or experience becomes possible?), and toward something Cavell calls philosophy’s “becoming literature.”  “Know” is here given back to philosophy as a word it must do something new with, and “literature” perhaps now names an attunement to the tension between our words and their future applications.  Only by a relentless re-attunement to how we’re using our words, Cavell seems to be saying, is philosophy able to go on.  And this occurs, for Cavell and for Wittgenstein (and, I would argue, also for writers like Emerson and Whitman) through an experimenting with the different kinds of intelligibility that can be arrived at from within a shared linguistic inheritance.

With Cavell’s idea of literary conditions in mind, we might aspire to mean—in our ongoing interdisciplinary language game between literature and philosophy—something new by the “critical” in literary criticism: namely, criticism as an investigation into the conditions under which literary or philosophical works can be taken as meaningful, such that those conditions allow for a shared accounting of—genuine agreement and disagreement about—such works.  For Wittgenstein, the conditions under which such claims become intelligible do not come down to a game of giving and asking for reasons, but are a methodical heightening of the problem of taking responsibility for what we say, understood as a uniquely literary responsibility. Going on means listening closely to the conditions under which our sayings become shareable, and doing so at every turn.

Notes

1. All quotations from Wittgenstein are from Philosophical Investigations 3rd. edition, trans. G.E.M. Anscombe (Edgewood Cliffs: Prentice Hall, 1958), and indicated by section number. My thanks to Rob Chodat, Ian Cornelius, Richard Deming, Jay Elliott, Juliet Floyd, John Gibson, Oren Izenberg, Ross Posnock, David Possen, and Ken Winkler for helping me to think about Wittgenstein, pragmatism, Cavell and other aspects of this essay.
2. Richard Rorty, “Cultural Politics and the Question of the Existence of God,” in Philosophy As Cultural Politics, (Cambridge: Cambridge University Press, 2007), 11-12.
3.Richard Rorty, “Dewey Between Hegel and Darwin,” in Truth and Progress (Cambridge: Cambridge University Press, 1998),  295-296. For Sellars on the relation of sense data to knowledge claims, see Wilfrid Sellars, Empiricism and the Philosophy of Mind, with an Introduction by Richard Rorty and a study guide by Robert Brandom (Cambridge: Harvard University Press, 2007 ) 13-25 [hereafter cited as EPM]); and Willem A. deVries and Timm Triplett, Knowledge, Mind and the Given: Reading Wilfrid Sellars’s “Empiricism and the Philosophy of Mind” (Indianapolis: Hackett, 2000) xxx-xxxii, and 76-77.
4.Rorty, Richard. “Cultural Politics and God,” 10-11. In an earlier essay called “Dewey’s Metaphysics,” Rorty writes that Dewey makes the mistake of “crossing the line …between causal [relations] and the self-conscious beliefs and inferences they make possible,” cautioning that “nothing is to be gained for an understanding of human knowledge by running together the vocabularies in which we describe the causal antecedents of knowledge with those in which we offer justifications of our claims to knowledge.” See Richard Rorty, Consequences of Pragmatism, Minneapolis: Minnesota University Press, 1982, 82.
5.Explaining the need for a deduction of the categories of the Understanding from logical functions, Kant appeals in the Critique of Pure Reason to an explicitly jurisprudential language: “Jurists, when they speak of entitlements and claims, distinguish in a legal matter between the question about what is lawful [quid juris] and that which concerns the fact [quid facti], and since they demand proof of both , they call the first, that which is to establish the entitlement of the legal claim, the deduction” (CPR A 84 / B 116).
6.Robert Brandom, Making It Explicit: Reasoning, Representing & Discursive Commitment. (Cambridge: Harvard University Press, 1994), 13. (Henceforth cited parenthetically as MIE).
7.Stanley Cavell, “The Investigations’ Everyday Aesthetics of Itself,” in The Literary Wittgenstein. eds. John Gibson and Wolfgang Hummer. (London: Routledge, 2004), 20.
8.Cavell, “The Investigations’ Everyday Aesthetics,” 21.
9.For more on the conflation of the causal and the logical in Wittgenstein’s “ideally rigid machine,” see Meredith Williams, Blind Obedience (London: Routledge,  2010), 155-160.
10.John McDowell, “Wittgenstein on Following a Rule,” in Mind, Value & Reality. (Cambridge: Harvard University Press, 1998), 230-231.
11.In his powerfully imaginative interpretation of Wittgenstein’s rule-following examples, Saul Kripke highlights the distinction between the “machine as abstract program…and the actual physical machine, which is subject to breakdown,” comparing Wittgenstein’s machine-as-symbol to Turing machines. Turing machines are “ideal” in the sense that they stand for all a machine needs to be able to do in principle to carry out program-like instructions. See Kripke, Wittgenstein on Rules and Private Language, 35-36 n.24 [henceforth cited as WRP]. For a detailed explanation of Turing machines, see Charles Petzold, The Annotated Turing: A Guided Tour through Alan Turing’s Historic Paper on Computability and the Turing Machine (Indianapolis: Wiley Publishing, 2008).
12.Anandi Hattiangadi. “Making It Implicit: Brandom on Rule Following,” in Philosophy and Phenomenological Research, (Vol.66, No.2, 2003), 427.
13.For Brandom’s unpacking of this Sellarsian point see EPM, 123-130; 139-144. For a helpful account of how Brandom (glossing Sellars) construes the myth of the given as hinging specifically on the move from sense data to noninferential beliefs, see Danielle Macbeth, “Inference, Meaning and Truth in Brandom, Sellars and Frege,” in Reading Brandom: On Making It Explicit, eds. Bernhard Weiss and Jeremy Wanderer (London: Routledge, 2010), 198-201.
14.For Goodman’s arguments about how the predicate “grue” radicalizes Hume’s skeptical solution to the problem of justifying induction, see “The New Riddle of Induction” in Fact Fiction and Forecast (Cambridge: Harvard University Press, 1983). For the relevance of grue to the gerrymandering problem in Wittgenstein’s rule-following paradox, see Kripke, WRP 20-21; 58-59. In the case of Poe’s poem “The Raven,” a confusion of reliable responses with genuine sayings is precisely the narrator’s predicament: the bird repeats by rote a sound the narrator takes to be a series of replies to his increasingly desperate and tormented questions about his “lost Lenore.” For more on how “The Raven” animates the problem of “going on” in this way see my Experience and Experimental Writing From Emerson to the Jameses (Oxford University Press, forthcoming).
15.The iron example is taken from MIE, 87-88. The example of the photoelectric cell hitched up to a tape recorder—the point of which is that we “cannot readily imagine continuing a conversation with [such a] gadget”—is from Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1979) 189.
16.Robert Brandom, Articulating Reasons: An Introduction to Inferentialism. (Cambridge: Harvard University Press, 2000), 17. (Henceforth cited parenthetically as AR)
17.Commenting on John McDowell’s construal of the “myth of the given,” Avner Baz makes just this point when he writes: “justification by an appeal to an unarticulated something [the ‘myth’ that our words could normatively relate to something essentially foreign to them] is no justification at all [and so] the unarticulated something…drops out of consideration as irrelevant, just like Wittgenstein’s ‘private object,’ and can make no contribution to the intelligibility…of our thoughts (judgments, beliefs).” Avner Baz, “On When Words Are Called For: Cavell, McDowell, and the Wording of the World,” Inquiry, 46: 4, 473-500, 2003, 477-78. See also John McDowell, Mind and World (Cambridge :Harvard University Press, 1994), 3-23.
18.Cautioning against Brandom’s belief that Wittgenstein’s way of imagining “norms implicit in practice [requires] work on the part of philosophers to uncover and make explicit,” John McDowell thinks that “there is no reason to suppose there must be a level of normativity below the level at which linguistic practice is described in terms of explicitly using this or that concept, and it is no concern of Wittgenstein’s to suggest that there is.” (John McDowell, “How Not To Read Philosophical Investigations,” in The Engaged Intellect [Cambridge: Harvard University Press, 2009], 98; 108; my emphasis).
19.Brandom sometimes uses the example of conditionals to illustrate what he means by expressivism: “Before introducing the [logical vocabulary of conditionals] one can do something, namely, endorse an inference. After introducing the conditional, one can now say that the inference is a good one. The expressive role of the conditional is to make explicit, in the form of a claim, what before was implicit in our practice of distinguishing some inferences as good.” (AR 81) For an account of conditionals as introduced into inferential practice as an “algorithmic elaboration,” see Brandom, Between Saying and Doing: Toward an Analytic Pragmatism (Oxford: Oxford University Press, 2008) 31-55.
20.Nor is it a matter, in my view, of radicalizing the problem of induction along the lines of Kripke’s claim that Wittgenstein’s rule-following paradox stands in need of a Hume-like “skeptical solution.” See Kripke, WRP, 66-69. For a compelling, and somewhat different, account of Kripke’s construal of the rule-following paradox—as a question of the “right” by which fellow rule-followers “taken into the community” [WRP, 110] and the “normative silence” of persons—see Oren Izenberg, Being Numerous: Poetry and the Ground of Social Life (Princeton: Princeton University Press, 2010) 94-95.
21.Stanley Cavell, “Introductory Note,” in The Literary Wittgenstein, 20.
22.John McDowell,  “Wittgenstein on Following a Rule,” 231; 236.
23.Cavell, “Introductory Note,” in The Literary Wittgenstein, 20.
24.The context of the discussion at §534 is that of considering two ways of understanding the meaning of a sentence. On the one hand, understanding a sentence can be thought of on the model of one sentence being “replaced by another which says the same”; on the other, “only by these words in these positions” [§531]. For Wittgenstein, what it means to understand a sentence includes aspects of both [§532]. I appeal to §531-534 in conjunction with §201 because the way Wittgenstein shows the “paradox” to be a mirage depends, I argue, on his ordering sentences in such a way that transition from one to the other is not a matter of inferential compatibility relations, but a function of what Cavell calls “literary conditions.”
25.Stanley Cavell,  ““The Investigations Everyday Aesthetics,” in The Literary Wittgenstein, 21.
26.For a nuanced account of the relation between conditions, criteria and context in Cavell’s thought, see Steven G. Affeldt, “The Ground of Mutuality: Criteria, Judgment, and Intelligibility in Stephen Mulhall and Stanley Cavell,” in European Journal of Philosophy, 6:1, 1-31.
27.Walt Whitman,  “Song of Myself,” in Leaves of Grass. (New York: Library of America), 30.
28.Stanley Cavell,  “The Argument of the Ordinary,” in Conditions Handsome and Unhandsome 83; 69; henceforth cited as CHU. Cavell makes this point specifically in relation to the way the “interlocutor” in the Investigations “poses a great task, the continuous task of Wittgenstein’s prose, oscillating between vanity and humility. Skepticism appears in [the PI] as one of the voices locked in this argument, not as a solution or conclusion” (Cavell, CHU, 83).
29.Stanley Cavell, The Claim of Reason: Wittgenstein, Skepticism, Morality and Tragedy. (New York: Oxford, 1979), 344. (Henceforth cited as CR in the text). Anyone who has read even a little Cavell knows that he remains interested (to put it mildly) in such fantasies: why and how they persist, what their persistence tells us about human life and the self, and what they tell us about the desire for something like philosophy (issues Cavell takes up, in all of his work, under the theme of skepticism). By placing my emphasis here on how Cavell describes Wittgenstein’s way of writing as indispensible to his philosophical aims, I certainly do not mean to suggest that Cavell is concerned to read Wittgenstein as dissolving, once and for all, such fantasies. For Cavell’s positioning of his reading of Wittgenstein on rules in relation to Kripke’s claim that §201 sets up a problem in need of a “skeptical solution,” and how this is matter of what it means to “go on,” see Cavell, CHU, 69-96.