by Michael J. Quirk

The Analytic tradition in contemporary philosophy saw its mission to be that of determining "how our words hook onto the world" by articulating "the logical structure of language", which is the public manifestation of thought. In this way, they sought to evade the problem plaguing earlier, empiricist accounts of language (e.g., Locke), which took ideas to be the primary bearers of meaning, and thus had to account how one could pass from the private inner realm of ideas to publicly shared meanings. Analytic philosophy generally held logic to be the essence of language, through which it comes to represent states of affairs and expresses meaningful propositions (cf. Wittgenstein’s Tractatus Logico-Philosophicus). While Wittgenstein never rejected the broad contours of this tradition, he eventually came to doubt the conviction that language has a singular "logical structure", an "essence" that can be laid bare by reforming the vagaries of ordinary language in favor of a formalized "ideal" language. We cannot make sense of language as a "calculus" of explicit rules, according to Wittgenstein, because we cannot escape language's gravitational pull: we cannot cast free of language to survey it and "how it hooks onto the world" from an extra-linguistic vantage point, and thus survey the necessary conditions for sense, reference, and so on. Yet nothing is lost by this inability --it's not even an "inability", since there's no genuine "ability" (i.e., to speak "transcendentally") to contrast it with. We cannot think outside thought nor speak outside speech --yet, in some way, this unrealizable aim is an integral part of what has been called "philosophy", certainly of analytic philosophy.

Thus the later Wittgenstein is obsessed with is "bringing words back from their metaphysical to their ordinary use": reconstruing philosophy as a form of therapy (rather than a doctrine or a "theory of theories") which seeks to transform us, to bolster our wills in such a way that we do not unwittingly fall into traps of the understanding prepared for us by our faulty metaphysical instincts. "What we do is try to show the fly the way out of the fly bottle." This is accomplished by close attention to particular cases where "the impulse to philosophize [in the standard, foundationalist mode]" leads us astray, and to quell that impulse by clearly describing the actual uses of language in these contexts --uses so familiar to us that we hardly notice them-- whereby we have no cause to overgeneralize and over-regiment things. In a nutshell: the later Wittgenstein's obsession is to acknowledge the infinite, indefinite variety of language-uses and language-contexts, and to use this as a hedge against the "contempt for the particular case", the concrete occasions where our words do make sense and contact with the world.

Like Heidegger, the later Wittgenstein could be said to be a practical holist. Meaning --Frege's sense and reference-- can only be established in the context of use, the practical linguistic life of human beings. Two things should be noted about meaning-as-use: first, that language has myriad uses, that it occurs in countless practical contexts (language-games), and that it is a philosophical trap (a "fly-bottle") to think that all these uses and contexts can be resolved to one master context (this was Frege's inaugural error: to have thought that language's primary purpose was representing states-of-affairs in assertions, which (mis)led him to view propositions and meanings as logical entities rather than linguistic deeds). Second, While language-in-use involves rules and rule-following, it is not to be construed as exclusively a matter of rule-following: "language is not a calculus everywhere bound by rules". This is important for Wittgenstein because it contradicts the earlier analytic emphasis on absolute clarity through absolute precision (e.g., the Tractatus's claim that "every proposition has one and only one complete analysis"). Our ability to use words to mean and refer to things is a skill that ultimately reflects a Form of Life, the ultimate practical "bedrock" which cannot be further explained, validated, or grounded. (cf. Heidegger's idea of "world" as disclosed in and through practices).

Wittgenstein's Tractatus was the culmination of the Fregean theory of language as a) basically a vehicle for making assertions and representing the world, and b) in which the "representative" and "expressive" functions of language could, in principle, be separated. In the Investigations, Wittgenstein disputes this in a twofold way: first, while language is used as the Tractatus says it is used, the representative-assertive function of language does not exhaust the range of linguistic practice, nor are its "cognitive" and "noncognitive" uses rigidly demarcated, and second, in order to "represent" facts in language, there must be a wider context of practices in which these facts, and their recognition, can fit and make sense. What is needed, says Wittgenstein, is "agreements in judgments", which is agreement in "Form of Life".

At the start of the Investigations, Wittgenstein contrasts St. Augustine's representative-assertive account of language with simple (simple-minded) counter-examples of workmen fetching slabs when commanded "Slab!", by their superiors, and so on. The point of this is, first, to show Augustine's representative-assertive model of language-in-use to be one among many such models, and one somewhat out of the ordinary (Wittgenstein is suggesting that this model is more appropriate for someone who already knows one language, trying to learn a second one), and second, to blur the edges between "representative" and "expressive" uses of language. The workers identify slabs by doing what their boss urges them to do, in a shared goal-oriented activity. What "slab" means --as a command and as a noun-- is unintelligible apart from that practical context. Later on, Wittgenstein will challenge the rock-bottom appeal to "ostentive definition", so critical to the Tractatus picture of language, as incomplete: one cannot grasp that I am "pointing to" something "yellow" without a prior grasp of the conventions-for-pointing (do I follow his finger? or go backwards along his arm?) and apart from a context in which I one can discern that it is "yellow" I am pointing-out (rather than a tennis-ball, or roundness, or fuzziness). This context is all-important, for language-in-use is what fixes what I mean ("the meaning is the use"; "Don't ask for the meaning [i.e., as some object present-to-mind a la Moore or Frege], ask for the use [the practical context in which we do things with words]".

If "meaning is use", the picture theory is fundamentally misguided: language does more than depict, and it doesn't depict anything apart from the ways in which it is used, which can not be expressed in precise formulae. Wittgenstein is not denying that language can "represent" states of affairs -- he's not an antirealist or "constructivist" or "linguistic idealist" – but he does insist that whatever representing language does is not a result of its "essence", the "logico-pictorial form" it putatively shares with "the world".

That thesis --that language has an essence, and philosophers can tell you what it is and how it helps language "hook onto the world"-- is a species of philosophical nonsense (in the non-mystical sense of "baloney"), which trades on the possibility, even the necessity, of taking up a position outside language-in-use (logically outside? metaphysically outside?) to explain how it works. Insofar as this can't be done, and needn't be done (since "how language works" is made manifest by describing its uses), Wittgenstein doesn't have a theory of meaning in his later work. What takes its place is description --what he calls "perspicuous presentations" (Ubersichtlich Darstellungen), which draw our attention to plain but little-noticed differences in language-in-use and get us to see what we mean without a dubious philosophical "theory of meaning".

The "problem of universals" has occupied (plagued?) philosophers since the time of Plato, and generally philosophers have fallen into one or two categories: Realists contend that universals are "real beings" either over and above the particulars which "participate" in them (Plato), or metaphysically inherent in them (Aristotle). Nominalists, on the other hand, insist that only particular things are real, and that universals (horse-ness, chair-ness, good-ness) are "just words", verbal shorthand for saying "this thing (or property), and that, and that . . . ). Wittgenstein thinks this problem can be dissolved rather than solved by taking stock of the way in which language is used. Realists presuppose that there is some one characteristic or quality which all X's (horses, chairs, goods, games) share, which makes them games. But a close inventory of, for example, games, shows that there is no single characteristic that makes them games, but a variety of overlapping characteristics, no one of which is shared by each and every game. Games are bound together by showing a loose and open-ended set of characteristics --a "family resemblance". So Realism about universals is bound-up with faulty assumptions concerning language and concepts, assumptions that vanish when language-in-use is made "perspicuous". The same holds true for Nominalism, which assumes that the only thing games have in common is that they are (arbitrarily) called games (that "games" is a "mere word", mere "made-up stuff"). But family resemblances are genuine resemblances, and what games have in common is that they are games, that they genuinely fall into this loose, open-ended collection, bound together by a "family resemblance". And language itself, like games, (horses, chairs, and families) is unified by family resemblances, rather than by one singular essence (Realism) or sheer arbitrary convention (nominalism).

The logical atomism advanced by Russell and the Tractatus insisted that "whatever can be said at all can be said clearly", and that such clarity is safeguarded by each proposition's "having one and only one complete analysis". This hypothesis, Wittgenstein suggested in the Tractatus,was essential "if language is to picture facts": there must be one absolutely determinate logical configuration of names in "logical space" that maps out onto a configuration of objects in a "state-of-affairs" in the world, and this "mapping" is itself made possible by their shared "logico-pictorial form". But Wittgenstein, in the Investigations, rejected this view as baseless --it's not the way things usually work in most "language-games" (e.g., "Slab!"). So, if the grounds for "logical form" vanish, along with the need for an absolutely precise "picturing" of language-independent facts, what becomes of "analysis"?

It becomes relativized to a particular context. It makes no sense to speak of "the absolutely simple parts" of something (e.g., a broom), since what is taken to be "ultimate" depends on the uses to which language is being put in certain circumstances. "Is the broom the handle plus brush?" is, for Wittgenstein, a loaded question apart from the background assumptions guiding language-in-use at the moment. Similarly: "Is the 'common sense', solid table real, or rather the 'scientific' table consisting of atomic particles and hence mostly empty space?" It makes no sense to "decide" such a question apart from whether one is using the table as a piece of furniture or as a scientific research object. (Recall Dewey's relativization of "essence" to the context-of-inquiry: the table-as-atomic-particles is real, but "it is real where it is real", and not "less real" than the workaday table simply by dint of the fact that it is the table-described-by-physical-theory). This rejection of absolute "simplicity" in favor of context-bound simplicity is, in effect, a deflation of all metaphysical or epistemological claims to describe things sub specie aeternitatis (from the standpoint of eternity, or a "God's eye view"). The metaphysical idealist (who dismisses, say, time as somehow "unreal"), and the metaphysical realist (who insists not merely that it is real, but absolutely real, real-in-itself apart from experience of it and speaking of it) both fall prey to the impulse to speak outside an actual, practical, "existential" context --to speak outside language games. Yet, in their actual, everyday modes of speech, realists and idealists alike will use temporal words like "now", "then", "before", etc. (Note the way Wittgenstein uses Moore, not to bolster realism, which is a species of philosophical nonsense, but to render philosophical nonsense "perspicuous").

Why the mania for absolute simplicity and absolute precision (i.e., Frege's and Russell's drive for a "logically perfect language")? Because philosophers -- and most of us, at some time or another-- want to sublime logic, to make it more than a mode of practice which orders and shapes our talk. We want logic to rest on something more than our practices, our Forms of Life, so we turn it into something more than it is --something deeply significant, something essential, something "sublime". This is not to suggest Wittgenstein was "against logic", whatever that could possibly mean: he'd be just as convinced that it is indispensible as Russell or Frege. But it is indispensible for us, for us as language-users, and it's importance derives from its perspicuous role in language-in-use. Logic is not the "essence" of language, but a part of language, a crucial part. It is not the essence of language, because language has no essence: it is a diverse collection of language-games which bear a "family resemblance" to each other.

Rules and Rule-following

Language-in-use involves rule-following, and is unthinkable apart from being a "rule-governed activity". The rules emerge from the various practical contexts in which language-games are "played". This has led more than one commentator to remark that Wittgenstein is a "radical conventionalist", that we decide to set the rules for language, and that once these rules are set, we can "mean" things by our words because their use has been fixed by stipulation. But since the "fixing" of rules is arbitrary --mere convention-- our concepts framed in language do not reflect "real essences" (i.e., our concept "tree" is not validated by the essence tree-hood) but only the self-contained, self-referential rule-structure of our language (i.e., the concept "tree" is meaningful for us in our language game because our conventions make it so --a sort of "linguistic idealism"). And if our concepts are mere "linguistic conventions", so too are our jusgements of truth and falsity: Wittgenstein is thus a radical conceptual relativist and a relativist about truth.

While it is true that Wittgenstein rejects any trans-linguistic "essences", I think this reading is wrongheaded, since Wittgenstein rejects the idea that language-in-use is "everywhere bounded by rules", and therefore that language could be a well-defined convention, and could thus completely arbitrary. Suppose I wish to teach someone hot to understand how to extend a number sequence on a piece of paper. Suppose the sequence goes "2,4,6,8,. . . " and after initial tutoring my pupil continued "...10,12,14..." I'm inclined to say "She understands --she's grasped the rules governing this algebraic sequence". Yet, when she passes "100", she extends the sequence differently "...100,104,108..." If I say "you didn't understand, you didn't grasp the rules", she could contradict me by saying "I thought the rules were: add 2 until you get to 100, then add 4, until . . . " I'd be hard put to say she was wrong about her (mis)understanding if understanding this little bit of language-in-use were solely a matter of recognizing and employing clearly-defined or definable rules. For I could always be unclear about what a rule might imply if I apply it in new and different contexts. If "understanding how the sequence goes" were just a matter of "grasping the rules", I'm caught in a skeptical trap, since no set of rules are absolutely clear and precise in their ability to direct me "how to go on".

Yet there is a right way and a wrong way of "going on", and I'd be right to say that my pupil "did not understand how to continue the sequence". She does not posess the required skills for doing so. It is wrong to think that having a skill is always a matter of "following rules" --skilled behavior, like carving wood, riding a bicycle, or painting a portrait, is precisely the ability to transcend rules, to "project" them into new contexts which no one can anticipate, and to "project" them well or badly (a sort of self-evaluative know-how which comes with posessing the skill in the first place). So if language-in-use is not always a matter of following explicit or explicable rules, but a matter of skilled behavior, the "radical conventionalist" reading of Wittgenstein must be mistaken, since for most language-in-use, there are no step-by-step arbitrary "stipulations" to rely upon.

What is relied upon, then? Wittgenstein speaks of human agreement as not primarily an agreement in opinions (rules, conventions, propositions) but an agreement in judgment, which is agreement in "Form of Life". Again, the "bedrock" for all questioning about language and its relation to reality is the totality of human practices, or do-ings which give us a particularly human fix on reality, and our particular human fix on it. To try to go beyond that is to speak philosophical nonsense, to demand an "absolute", necessary, ahistorical perspective in language which we can't have and really don't need.

Is Wittgenstein endorsing relativism at another level --a relativism of "forms of life"? I don't think so, for two reasons. First, "form-of-life " is ambiguous, meaning both one-form-of-life-among-many (perhaps many incompatible and incommensurable forms), and one-of-life's-many-forms. I think the ambiguity is deliberate: the first meaning suggests that we shouldn't assume that "everbody means the same thing" (has the same concepts, or beliefs, or judgements, etc.) across different cultural or historical divides. This prevents one from slipping into a too-easy "objectivism" about understanding alien meaning-and belief-systems. Yet the second suggests that, as life has many forms, as living humans, we can all at least potentially share in them. This precludes radical relativism. I think this was what Wittgenstein was driving at, since his thought is unremittingly hostile to virtually all philosophical dichotomies (realism/idealism, essentialism/conventionalism, and so on). Second, if a form-of-life is a skilled form of know-how, one can always acquire it if one does not already posess it. It may take hard work, with no guarantee of ultimate success, but one can never rule out a priori the understanding of an alien language-in-use. Simply roll up your sleeves, and do as the anthropologists do in the field.

Wittgenstein contrasts "empirical" investigations, like those of science, with the "grammatical" inquiries proper to philosophy. Philosophy does not discover, through systematic observation and controlled experiment, new "facts" about the world, nor does it contrive theories which can predict worldly events or explain them. Philosophy draws on understandings common to everyone insofar as they know how to speak the language. Hence it "analyzes concepts" --which are simple regularities in linguistic practice-- by describing their different facets. "Grammar tells us what kind of an object anything is" --i.e., the structured patterns of how we speak disclose what we mean when we talk about, say, "universals", or "certainty", or "knowledge", or "sensations". Philosophical Investigations are thus essentially "grammatical" investigations.

Several points of clarification are warranted. First, Wittgenstein is no more saying that Grammar "constructs" the world, or provides the "framework" in which we order the chaos of experienced reality, than he is saying that grammar "mirrors" the structure of the world as it is in itself. Realism and idealism (or antirealism) are ruled out by his metaphilosophical conviction that we cannot transcend our linguistic practices to determine "the secret of their success", whether they "find" or "make" the world of facts and truths. Rather, since we can't escape our epistemic and semantic finitude, we'd better examine the ways we order our finite understandings of the world --i.e., by paying close attention to "what is said and when". Second, it seems as if Wittgenstein is doing a sort of transcendental philosophy --drawing a neat distinction between "the conceptual" and "the empirical", between form and content, and confining philosophy to the former categories. This would put him quite at odds with pragmatists like Dewey, Quine, and Davidson, who took pains to blur the conceptual/empirical, analytic/synthetic, and science/philosophy distinctions. Wittgenstein does insist that philosophy is utterly unlike science, so to this extent, like Heidegger, he is not a pragmatist. But for Wittgenstein, the distinction between what is "conceptual" and what is "empirical" cannot be settled by any sort of transcendental investigation: we must painstakingly describe language-in-use to grasp its grammar. Furthermore, what counts as "grammatical" at one moment in the history of a language-in-use may become "empirical" at another: the analytic/synthetic and conceptual/empirical distinctions are historical and to a degree ad hoc, rather than permanent, a priori transcendental categories of statements. Wittgenstein is thus more of a pragmatist (and less of a Kantian) than he might seem.

Philosophy thus becomes a matter of elucidating "grammar". But this is not as humdrum and drab an enterprise as it seems. The sheer complexity of grammar mesmerizes us, and we often misunderstand just what it commits us to. The grammar of language bewitches us, and we fail to note that that grammar is articulating different meanings and "ontological commitments" than what we reflexively and "unthinkingly" think. Moreover, the impulse to be "bewitched by our language" is a natural one, so that in understanding that we are bewitched and confused we also come to see why we are bewitched and confused. And the natural temptation to become mesmerized by taken-for-granted grammar, and thus to fail to see that which grammar reveals to us (which is never hidden, but so obvious and omnipresent we habitually ignore it), is lessened. Philosophy gets its charter as a form of therapy --not as a builder of theories (since philosophy can't explain anything but merely clarifies what any competent language-user knows in using the language) but as a healer of befuddled souls.

One place where "therapy" is sorely needed is in the Philosophy of Mind. The second half of Part 1 and most of Part 2 of the Investigations concentrates on "untying the mental knots" philosophers have tied concerning "mind", "sensation", "subjectivity", "privacy" and the like, with the net result not of proposing a philosophical theory of mind superior to dualism, materialism, and behaviorism, but --"no knots" (N. Malcolm).

Here is the "mental knot" that Cartesian dualism ties: my first-person reports of my own mental events are incorrigible in a way that third-person reports of others' mental events are not. I have privileged access to the contents of my own mind; I have no direct access to the mental states of others. Therefore I can know my mental states in a way that I cannot know those of others: I know I am in pain immediately and indisputably, whereas I can only infer you are in pain on the basis of your bodily states. But such inferences are only warranted insofar as there is evidence that that bodily states must reflect internal, mental causes. I have no such evidence --I can't crawl into "other minds"-- and I have some evidence to the contrary, insofar as the other mind may, sometimes, confess to feigning a mental state, say, "pain". So I'm locked within the privacy of my own mind (unless, of course, a good God necessarily exists, who would not let me be deceived about such matters.....).

But the Cartesian dualist is misled by the intricacies of grammar, of our concepts of "sensation" and "thought". To be more accurate: he sees these concepts and the grammar in which they are embedded as simpler than they actually are, and as having reference-conditions exactly parallel to the ways in which we refer to other (physical) states or processes. He is misled in to thinking that "There are some marbles in my pocket" is analogous to "There are ideas in my mind". Ideas, thoughts, pains, etc. are inner-states just as marbles are in-pocket objects: they are referred-to in the same way. Yet if, for example, pains are inner states and/or processes, their relation to observable behavior is causal and contingent: I have to observe the pain-behavior of others --the groaning, the grimacing, the writhing-- and infer that they are in pain (that they posess the correlative "inner states/processes"), in a way that I do not do with myself. And if they are inferred, contingent entities, I cannot be said to "know" them, the "3rd person" pains of others, the way I know my own pain in the 1st person. Hence the possibility of skepticism. Likewise with other "mental states/events".

That this is not the case becomes clear when we set the "grammar of pain" before ourselves in a "perspicuous presentation". Pain-language draws upon certain rules which establish conceptual connections between certain concepts, like "pain", "knowing", "feigning", etc. The connections are conceptual because empirical investigation can't really prove them false (or for that matter true): if empirical research showed, for example, that a particular case of pain had by a subject "didn't hurt", we would not discover an anomolous fact about someone's weird pains, but we'd discover that whatever this subject "has" it isn't pain. One such key conceptual connection is this: to know that X is in pain is to find that X satisfies certain criteria --"An inner state stands in need of outward criteria".

We learn the pain-language by being trained to make certain discriminations and employ certain criteria: we are introduced to certain rules-of-use until we "get the hang of it" and can follow the rules and project them into new situations. So here's a language-game in action: "Don't hit your brother on the head with that toy: you'll hurt him!", "See what you did! He's crying --you must have hurt him real bad!", "Now he's hit you, and you're in pain --see how it feels?" Pile up enough paradigm cases like these and certain rules --and more importantly, conceptual skills-- for talking about pain are incorporated by the language-user: a) there are often external, environmental causes of pain (thrown toys, pulled molars, stress on the job), b)to say you know someone is in pain involves seeing if certain behavioral criteria are met by him or her (wriggling, screaming, moaning, utterances of "Pain!" or "Ow!"), and c) I know how to apply the predicate "..... am in pain" to myself by being trained in the language I use.

The Cartesian might have a ready reply to each of these points. Regarding a) and b), she might complain "We Cartesians do not deny physical causes for pain, nor do we deny a causal relationship between pain and pain behavior. It's simply that one pole of the causal relationship, the mental, is epistemologically hidden from us in 3rd person ascriptions of pain. Hence other-minds skepticism is not implausible". Wittgenstein, however, would argue that this counterargument is question-begging. The Cartesian assumes that the relationship between pain and pain behavior is causal rather than criterial, and thus makes the unexamined assumption that the causes of external behavior must be internal, and hence metaphysically distinct from it. The Cartesian presumes that these inner-objects, while perfectly transparent to a 1st person perspective, are opaque to 3rd person analyses: "someday, perhaps [if we develop an adequate, accurate, predictively powerful psychology] we may know more about them". But this is a "conjuring trick", since it assumes, and seduces us all, into thinking that that these "inner causes" must be states and processes with the same identity-conditions as other states and processes.

This actually violates the grammar of mental discourse. The Cartesian insists that one knows 1st person mental states more immediately, and hence accurately, than 3rd person discourse. But "it cannot be said (except as a joke) that I know I am in pain". Why? Because I am in pain. If I am in pain, I need no external criteria for validating my assertion "I'm in pain": the pain is quite enough. Hence I can only be said to literally know someone else is in pain, for there the outward criteria can be applied. Furthermore, it shows how misleading the "inner process/state" analysis of pain is: if "an inner process stands in need of outward criteria", and I need no outward criteria to legitimate my own self-ascriptions of pain, then pain is an "inner process" only in a highly qualified, even "pickwickian" sense.

This sounds like behaviorism, yet on closer inspection, Wittgenstein's grammatical analysis of pain-discourse dissolves behavioristic assumptions as well. The behaviorist wants to deny the reality, or at least the epistemic respectibility, of anything which lacks "outward criteria". But in affirming that 3rd person ascriptions of pain are criteria-guided, Wittgenstein is not thereby denying anything inner. At most, he's denying that the identity- and identification-conditions for that which is "inner" are the same as those for "outer" objects. The behaviorist is methodologically constrained to grant cognitive respectibility only to objects with strict, criterial identity conditions: hence the radical behaviorist (Skinner), will further reduce human behavior (which after all can encompass meaningful actions, like asking, commanding, judging, expressing anger, "cursing, greeting, praying", etc.) to gross bodily movement ("I raised my arm" = "My arm went up"). But this is ridiculous. It denies entire fields of experience which can be made perspicuous in language --i.e., intentional action, imagination, etc. (A joke which makes the ridiculousness of behaviorism "perspicuous": What do behaviorists say when they greet each other? You're okay: how am I?). The behaviorist is in the grip of a seductive, distortive explanatory theory, while Wittgenstein's method is that of gaining clarity through description alone.

The best "image" of Wittgenstein's (non)philosophy of mind is perhaps contained in one of his pithy aphorisms: "The best picture of the human soul is the human body". What Wittgenstein is driving at in this passage, I take it, is that the soul, or the mental, is not something other than the body, or physical behavior, but then again the soul/the mental is not nothing but the body/ the physical either. Think of mental states/processes not as separate and separable causes of bodily behavior: the behavior is the outward manifestation of the mental event --it's "physiognomy", so to speak. Likewise, one can configure the "inner" mental state (of another person) as the interiorization of the outer gestures. My pain is --in the "3rd person" perspective-- the writhing and moaning, which is the 1st person felt-pain as externalized and expressed. So there is no question that I know the pain of another directly: I do not infer pain, I see it, and I see it precisely because grammar supplies me with criteria and "agreements in judgement" for detecting it. And the "immediacy" of my own pain is demystified: it does not supply me with special, indisputable conditions for "knowing I am in pain", since knowing how to apply the predicate "in pain" to myself is likewise a matter of training in a language-game. "Immediacy" carries no epistemological weight --indeed, 1st person immediacy makes it awkward to say "I know I am in pain", except in an equivocal, highly qualified sense. In short, Wittgenstein ventures to bring mind and body back together in the person, not by way of a philosophical theory of the person, but by showing that dualism is radically out of sync with the grammar of mind-talk.

The Cartesian (or the Cartesian in us) is not likely to be satisfied with Wittgenstein's "physiognomic" account of the mental. For meaning, or intentionality, is a mark of the mental, and "what one means" is surely "in the head", hence private, hence privileged. The fact that I can grasp meanings by myself (cf. Moore and Frege) and then clothe them in language is demonstrated by the possibility of a private language: one could keep a diary of, say, a certain feeling, assign a graphic or acoustic sign to it, and write/utter it to oneself whenever it recurs. Here, the Cartesian would like to say, is a case of something mental which has no physiognomy.

Wittgenstein doesn't quite deny that one could keep such a diary: he only denies that it could constitute a lexicon for a language, and hence be a repository of meanings. For to identify an internal state (a pain, a thought) correctly, it must be possible to reidentify it, and to reidentify it, it must be possible to misidentify it. There is no way I can do this within the parameters of a purely private language: how do I know when I've made a mistake? How can I know this, given that all I have to go on is my "feeling" that I've identified it correctly (or, for that matter, misidentified it)? Who'll tell me when I screw up?

The private language, insofar as it remains private, is not a language: its vocalizations/inscriptions mean nothing, since their meanings are not fixed or stabilized by a shared practice. Of course, one could try to teach the private diary-language, in which case it becomes language but ceases to be private. And it is unclear how this might be accomplished, unless one already is involved in public linguistic practices --language games situated in shared forms-of-life.

The implications of the "private language argument" are vast, and far more profound than most commentators suspect. First of all, it disabuses us of the idea that "meanings" are ontologically situated in the mind (Descartes) or in the brain (Materialism). Indeed, it's misleading, from Wittgenstein's point of view, to "locate" meanings at all, except as located "in language". And the "location" of language is just as hazy, except insafar as language "is" in our practices --language as spoken, written, thought, and so on. We mean things by using language, and the meanings are in the language we use in common. Our shared practices open up meanings to us (cf. Heidegger's "Language is the House of Being"), and not some occult, essentially "inner" faculty of recognition. Meanings are woven into public practices before they're present to individual minds.

Secondly, Wittgenstein drives a wedge between privacy and "secrecy", or between the plain (grammatical) fact that no one can have my 1st person mental states as my 1st person states, and the misconception that therefore no one can have access to my 1st person mental states because they're mine. The "beetle-in-a-box" example explodes the notion of privacy-as-inexpressibility: if no one can see the "beetles" in each other's box, "the thing-in-the-box cancels out" of the language altogether. But of course thoughts (pains, etc.) aren't like that: I can think the same thoughts (not numerically the same, but the same-in-kind) as you, not as your 1st person thoughts but as my 1st person thoughts, and hence your 1st person privacy is not inaccessible to me if we mean what we say and we say it in a shared language-in-use. Of course, I can try to make what I mean inaccessible to you by hiding it --being evasive or silent. But this is likewise a commonplace human occurrence, and has no grand philosophical importance. "Privacy" amounts to "I am I, and you are you", and you therefore cannot experience what I experience as the "I" that I am, but only as the you that you are. This, however, is obvious --grammatically obvious. And it's precisely because it is a grammatical commonplace that we gloss over it and lapse into Cartesian subjectivism.

Finally, the private language argument is not so much an argument against other-minds skepticism as it is a deflation, or dissolution of it. The skeptic is on to something: I can only have my pain, for example, as my pain, and I can't know yours in the same way. This amounts to: there is an irreducible difference between 1st person pain-reporting and 3rd person pain ascriptions. Yet the skeptic takes this difference to indicate a deficiency in our knowledge of the pain (the thoughts, emotions, etc.) of others. And this is the "conjuring trick" of which Wittgenstein warns us: it assumes that the standards for "cognitive respectibility" are uniform across 1st and 3rd person descriptions. Once one drops that assumption --and the grammar of "I know she's in pain because..." gives one warrant to drop it-- then one can simply shrug off skeptical doubts since they set an impossible standard and then complain when it cannot be met.

Consider this exchange:

-- "How do you know Delores is in pain?"

-- "She's moaning, and she says her tooth hurts."

-- "She could be faking."

-- "She's honest, and she has no reason to fake, so we all have ample justification to affirm she really is in pain."

-- "But how do you know she's in pain? You can't feel her pain."

-- "Of course I can't feel her pain --that's (grammatically) impossible. But I see she's in pain." – "But this is just evasive: how do you ever know anyone is in pain?"

-- "Because I know what the words 'is in pain' mean: I speak English".

– "You're skirting the issue: isn't there always room for doubt?"

-- "As a sort of parlor-game, perhaps. But in real life, no."

-- "Aren't you shutting your eyes to the ever-present possibility that no one else is ever in pain?"

-- "They are shut. Yours too!"

The skeptic, in the preceding passage, is bent on giving the word/concept "know" a special philosophical sense --i.e., know incorrigibly, immediately, non-criterially. Because of this, he constantly accuses the "ordinary" language-user of being evasive. But the skeptic is the one who's evasive: he evades the ontological and epistemic commitments the ordinary (i.e. standard practical) use of "to know" commits him to, by injecting what the pragmatist philosopher C.S. Pierce called "paper doubt" at precisely the point where his reasonable doubts have been quelled (after the pain-claimant is said to be honest). The skeptic is thus not refuted --there's nothing which prevents one from entertaining free-floating doubts-- but cut off from the language-game upon which he parasitically feeds. "What we do is bring words back from their metaphysical to their ordinary use": by doing this, Wittgenstein marginalizes, rather than rebuts the skeptic, since the skeptic refuses to argue his case on any other terms than his own. Hence the sanest and most effective thing to say to a skeptic is "go away".

What Wittgenstein does to the other-minds skeptic in Philosophical Investigations, he does to the epistemological realist in On Certainty: here, he tries to being the words/concepts "knowledge" and "certainty" back from their epistemological to their ordinary uses.

The epistemological realist Wittgenstein uses as a foil is G.E. Moore. Moore's defense of "common sense" epistemological realism rests on his unwillingness to grant philosophical theories the decisive last word in justifying or legitimating our everyday knowledge-claims. The subjective idealist, for example, denies that time is real, or that the external world exists. Yet, argues Moore, the subjective idealist uses temporal words such as "before" and "after", and distinguishes between his act of perception and the object perceived. The subjective idealist thus must use an epistemological or metaphysical theory as a counterweight against the claims of common sense and ordinary usage. But why should one give credence to philosophical theory as against common sense? If a theory contradicts common sense, why overthrow common sense? Why not displace the theory which contradicts it?

Moore then explicates the epistemological elements of common sense: the epistemology of common sense is realistic. Moore lists "common sense" platitudes which he claims he knows are true, and knows with certainty: that he has hands, that objects exist outside his mind, that the earth existed a long time before he did, that no one has been to the moon (at the time of his writing in the 1930's), and so on.

There is much in Moore with which Wittgenstein can sympathize. Like Moore, Wittgenstein is concerned with reasserting the primacy of "the ordinary" in the face of a form of theorizing that is detached from ordinary concerns and interests. But Wittgenstein believes that Moore does not go far enough, for he presses his appeal to common sense/ordinary language in the service of philosophical discourse. Moore believes that "the ordinary" embodies an epistemology --that of direct, straightforward realism. But this, for Wittgenstein, is to show one's hand as a metaphysican-epistemologist after all: Moore cites "the ordinary" in order to use it in an argument with "traditional" philosophers, yet because he is more than willing to try to defeat traditional philosophers on their own terms, he falls victim to the deep tendency that the appeal to "common sense" was designed to guard against. Namely: the typical philosophical tendency to think that one can rise above everyday practice to determine how it "succeeds" and what its success consists in.

Moore conflates two "epistemic" categories or concepts --certainty as the absence of doubt, and knowledge as that which is warranted or justified or grounded. Yet a "perspicuous presentation" of knowledge- and certitude-talk shows that where doubt cannot possibly arise, one cannot be said to have knowledge --or for that matter, to not have knowledge. To speak of "knowledge" in such contexts is at best equivocal, usually awkward, and often enough nonsensical.

Reflect once again on Wittgenstein's quip in the Investigations that "I cannot be said (except as a joke) to know I am in pain. What else could this mean except perhaps 'I am in pain'?" Here Wittgenstein is denying that one could have possible room for doubt: 1st person pain is incorrigible. If I cannot possibly doubt my own pain, I can never be in a position to be in error about it. But if there's no possibility for error, there is likewise none for knowledge. I can say "I know the cat is on the mat --I'm certain!", but here the certainty is of a different order, since I could, possibly be wrong, and my knowledge-claim could possibly be unwarranted. But if I can't possibly doubt pain --i.e., if its grammatically impossible-- then I can't "know" it either. It's questionable whether I can even be certain I am in pain --I simply am in pain. The contexts in which I can intelligibly say "I know I'm in pain" are exceptional, and exceptions which prove the rule. Consider: "Will you stop saying I'm faking! I know when I am in pain!". Here an interlocutor is raising doubts where no cogent doubts can (grammatically) be raised.

Moore's philosophical mistake is to conflate certitude with knowledge, and to employ this (mis)analyzed concept of "knowledge" to a spirited defense of "common sense realism". Yet most of the propositions Moore cites are not of the straightforward contingent "The cat is on the mat" variety. It is not clear how, for example, I could disconfirm something like "I know I have hands". In certain circumstances one might imagine a legitimate conformation or disconfirmation of this proposition (e.g., "Do you have genuine hands or bionic replacement hands?"). But these cases are exceptions, parasitic upon our standard linguistic use. It is hard to see how the justification of such a claim could be more certain than the proposition "I have hands" itself. And in the face of radical skeptical doubt, there is no adequate justification one can give (e.g., "How do you know you have hands?" "Look, here they are." "No, no, I mean: how do you know you're not dreaming/hallucinating, etc.?"). And of course, the path from skepticism is precisely not that of refutation --of showing that one "does too know one has hands"-- but of showing that the skeptic is misled by grammar into thinking that all certain propositions admit of proof, or have a ground. More, the skeptic's archenemy, thus makes precisely the same grammatical blunder as his archetypal skeptic/idealist.

Propositions such as "I have hands", "Material objects exist", "The earth existed a long time before my birth", etc., are part of our world-picture, Wittgenstein suggests. They form, as it were, the "riverbed" through which the "river" of straightforward empirical propositions flows (e.g., "I have large calloused hands", "Uranium atoms exist"). They are not "grounded" in more basic propositions or certitudes, nor are they susceptible of confirmation or epistemic justification. They "stand fast" for us (as common-language-users). They are not ahistorical synthetic a priori necessities like Kant's categories: the "riverbed" can change --new sediment settles, old sediment is gradually eroded and washed downstream. Yet at any given moment, what is and is not part of one's world-picture can be generally decided (even if there are some random propositions which it may be hard to safely assign to riverbed or river-flow).

Note how Wittgenstein sidesteps the issue of whether he's is a "conceptual analyst" or a pragmatist like Quine. On the one hand, to distinguish the propositions of a world-picture from empirical propositions is to distinguish "matters of language" from "matters of fact", or to admit something like an "analytic/synthetic" or "conceptual/empirical" distinction. But Wittgenstein readily admits such distinctions are themselves ad hoc, changeable, historically and culturally variable (the riverbed isn't eternally fixed: it's a product of contingencies, and subject to them as well). On the other hand, the "river" metaphor is reminiscent of Quine's "web of belief" or "field of force", where the "central" beliefs are less prone to revision in the face of experience because doing so would exact an unacceptable pragmatic cost. Yet Wittgenstein dispels any speculation that he's a pragmatist: that, he quips, would be to be misled by a "certain world-view". Namely: that which wants to explain the contours, the topography of language-in-use by way of an overall theory of agency and cognition-as-a-species-of-human-action. Wittgenstein, like Dewey, acknowledges the primacy of practice; unlike Dewey or Quine, he sees no single, theory-centered role which philosophy might profitably take up. Rather, there are different "therapies" for different "ailments". So it is equally misleading to tag Wittgenstein as a "conceptual analyst" akin to Austin, Strawson, and Grice as it is to assimilate him to neopragmatism in the company of Quine and Davidson. Never underestimate his idiosyncracies.

What of these world-pictures? If they differ, if they change, isn't he a radical relativist after all? Doesn't this place him in the vicinity of, say, Foucault, whose "epistemes" (the conceptual-belief-knowledge-truth systems which punctuate human history) are separated from each other by drastic ruptures which preclude any judgements about their relative truth or rational adequacy? Some enlist Wittgenstein in this cause, either happily or ruefully. But I think that all such attempts to make the charge of relativism stick to Wittgenstein fail, and fail because they miss the radically antitheoretical nature of his thought. My concepts and words are ordered by a world-picture and that there is no world-picture-less form of discourse which might enable me to rise above it. Yet there is nothing in the above which prevents me from criticizing rival, incommensurable world-pictures from within the ambit of my own, nothing which might prevent me from understanding these alternative world-pictures sympathetically, and perhaps questioning the adequacy of my own world-picture as a result. Thus the fear of relativism is the fear that unless world-pictures are grounded in something (more) rational, then they are irrational (arbitrary, nihilistically crazy, oppressive "regiemes of truth" in Foucault's sense). But they are in themselves neither rational nor irrational: "they are there, like our life". To cure the would-be philosopher of the desire for more than that --for a ground-to-end-all-grounds, a reason-stopping rationale-- would be to restore his or her health, and to grant him or her the peace of mind "to be able to stop philosophizing when one wants to".

©  Michael J. Quirk, 2000.  Permission is granted for electronic copying, distribution in print form for educational purposes, and personal use.  If you reduplicate the document, be sure to include the copyright. No permission is granted for commercial use.