Was Wittgenstein a Realist or an Antirealist?

(A while back, in my post “The Scholar, the Prophet, the Monk, and the Healer”, promised a series of philosophically oriented posts that would explore “whether Rorty, Heidegger, Wittgenstein, or Dewey successfully evaded the dualisms they struggled with, or how well they manage to avoid the extremes of nihilism and dogmatism, or what kind of ethics and politics their thought implies.” This is the first of this series; as I indicated before, I welcome commentary and criticism. – MJQ) 


Was Wittgenstein a Realist or an Antirealist?

The short, correct, uninformative answer to this question is “no”.

There is nothing glib or superficial about this answer, since Wittgenstein’s philosophical work was, from beginning to end, directed in opposition to a certain conception of the task of philosophy: that of philosophy as a body of doctrine that articulates, with an aura of necessity, the origin and nature of reality, what we know about it, and how we can meaningfully communicate about it. That is: his thought is an all-encompassing case against metaphysics, epistemology, and semantics/philosophy of language, as it had been understood and practiced. As an alternative, Wittgenstein, both early and late, re-conceived philosophy to be an activity, a form of belief-therapy which advanced no definitive theses, metaphysical or epistemological or semantic, an activity which sought only to cure his readership of “diseases” of the intellect. Whether Wittgenstein succeeded in doing philosophy without lapsing into “theory” or was successful at reconfiguring philosophy as therapy is open to debate. But this issue swings free from what I take to be his signal accomplishment in bypassing the whole realism/antirealism conundrum, a mainstay in metaphysics and epistemology. I think he managed this feat ambiguously in the Tractatus, but definitively in the Philosophical Investigations. And that, I think, is one of Wittgenstein’s chief innovations and accomplishments as a philosopher in the analytic tradition.

Think of the philosophical realist/antirealist debate in terms of Plato’s metaphor in the Phaedrus: do our concepts “carve Nature at the joints” or not? Do our judgments reflect these intrinsically “natural” slices of reality, and our true assertions reflect or mirror them? The realist would argue that genuine concepts can only carve nature at the joints: the job of our concepts is to correspond to the way reality is articulated in itself, quite apart from how it is described by us.  Concepts that do not “carve nature by the joints” are either related to secondary qualities (a la Locke: concepts that correspond to qualities in us that are caused by external realities exhibiting primary qualities), or are pseudo-concepts, mere projections that represent nothing themselves. When we have justified true beliefs – knowledge – our beliefs accurately represent the way in which reality is itself articulated or structured, because genuine concepts are themselves representations of these structures or articulations. Epistemological realism thus follows from, and depends upon, some kind of metaphysical realism.

The philosophical antirealist counters that the realist’s “correspondence” between belief or proposition and reality is an occult relation, an uncashable metaphor. Our concepts divide the world the way one would divide a bowl of Jell-O with a knife – there are no “joints” to carve, and our concepts are our means of ordering and structuring being, projections of our own interests or predilections or sensible-perceptual makeup. . Of course, one can be a realist about one kind of thing (say, “facts”, or the entities of physical science, or the “middle-sized dry goods” of common sense perception) and an antirealist about another (say, “values”, mystical experiences, religious sensibilities, the gods) – which may be the majority metaphysical viewpoint. Still, arguments about which kinds of entities about which one should be realist or antirealist generally trade on the argumentative and imaginative framework outlined above.

The dialectic between realist and antirealist, as Richard Rorty, Charles Taylor, and Donald Davidson[1] have argued in very different ways, presupposes the cogency of representationalism, the idea that propositions/sentences represent reality-as-such and that their truth or falsity is a function of whether that representation obtains or not. (“Concepts”, after the Linguistic Turn, make sense only in the context of sentences, which are the candidates for truth or falsity, for contemporary realists and antirealists alike.) “Truth”, on this realist reading, is an explanatory term. For the realist, it explains the microstructure of the correspondence relation that “makes” sentences true: sentences are true because they mirror the way reality is in-itself. For the antirealist, “truth” is a kind of convenient fiction (cf. Nietzsche), a term of convenience indicating that the way we slice reality up in accord with our conventions coheres well with other such slices, or how our projections line up with other sets of projections. But this entire dialectical dance depends, for its cogency, on representationalism. If representationalism is rejected as incoherent or empty, the stakes in affirming either realism or antirealism go down considerably, if not completely.

Wittgenstein’s entire philosophical career was dedicated toward lowering these stakes. In the Tractatus, he pushed representationalism to its utmost limit, but when he does so, one can no longer say what exactly is being represented (what are the atomic facts that atomic propositions picture?), and the realist/antirealist dialectic melts into air. In the Tractatus, a radically representationalist program is put to use as a kind of self-consuming artifact, a provisional semantic argument to undermine all semantic argument, including itself.

The ontology of the Tractatus exists only to make sense of semantics: the way in which words “hook onto the world”. For this to take place sentences – or more precisely, the propositions expressed by sentences – must be analyzable into atomic propositions that are, essentially, configurations of logically proper names in logical space. These in turn correspond to objects, agglomerations of which are called states of affairs or atomic facts, which correspond to the atomic propositions. This semantic ontology is an extreme form of representationalism: the “picture theory” of meaning.  Complex propositions are truth-functions of simple/atomic propositions, whose truth or falsity is determined by the ways in which they are logical pictures of states of affairs. And the core of this semantic ontology is that of “logico-pictorial form” which must be shared by propositions and the facts they represent.

However once you get to this point, it becomes clear that it does not matter exactly what is depicted. What is a Tractarian Object? It does not matter. What matters is that there must be objects in atomic facts that share with names in atomic propositions that which Wittgenstein calls logico-pictorial form. What objects are is not determinable, not sayable, and quite irrelevant. Therefore metaphysics – whether realist or idealist or solipsist – is meaningless, or better yet, beside the point.

The realist/antirealist conundrum rests on at least one assumption that the Tractatus challenges: that we can say what representations are ultimately representations of. The Tractatus sees that as meaningless, not just in the sense of sinnlos (without sense) but of unsinnig (nonsense, pointless, gibberish). We need to know that true propositions represent, that they say how things are and show their sense, and speculation about what their objects “ultimately are” is a nonsensical attempt to say what can only be shown. Any attempt to do this, to say the showable, is to try to represent the conditions for the possibility of representation, and thus to try to represent outside all representation. This is as impossible as it is unnecessary: a meaningful proposition shows its sense, and does not have to say it. So think of the Tractatus of drawing quasi-Kantian conclusions from the premises of Logical Empiricism: metaphysics of any kind, realist or antirealist, materialist or solipsist, tries to say that which is only showable. The rest is silence.

It gets worse – or better, if you want to look at it another way. The antimetaphysical conclusion of the Tractatus undermines the Tractatus itself: what look like semantic theories being tested and conclusions about logic and representation being drawn emerge, at the end, as further specimens of philosophical gibberish. But if you follow (and buy into) the argument of the Tractatus to its end in proposition 7, the nonsense has been instructive. What looks initially like a more rigorous and plausible formulation of Logical Atomism is a mere placeholder, a provisional procedure for first dismantling the lurking metaphysical pseudo-questions of “What exactly are objects?” and “What exactly is a thought?”, and then going on to dismantle itself. Wittgenstein could have used any number of self-destructing methods to get one to Tractatus 7: perhaps he could have elaborated and reformed Parmenides or Heraclitus instead of Russell or Frege. As Wittgenstein was working within the analytic tradition, and grappling with problems then current like the Theory of Types, he used the method that was most expedient at the time. But the result would have been the same: whereof one cannot speak, thereof one must remain silent.

In the Philosophical Investigations, however, Wittgenstein abandons the trope of the picture theory of propositions, and with it, the whole idea of representationalism, even as a self-destroying provisional method. His path to this viewpoint was long and meandering: his dawning conviction that color propositions are not truth-functionally independent (e.g., if “x is blue” is true, then “x is red” must be false), his being influenced by his Cambridge colleague Piero Sraffa’s view of “logico-pictorial form” as pointless (e.g., what is the logico-pictorial form of the Italian “up yours” street gesture), his strengthening belief that Russell’s Theory of Types needed a stronger form of refutation than that given in the Tractatus, and so on. In the Investigations, not only the picture theory is jettisoned, but the very idea of the centrality, indeed the very possibility, of a general theory of meaning is rejected, even if such theories is a species of nonsense to be thrown away like a ladder when no longer needed. When Wittgenstein came to write what was to become the Philosophical Investigations, the strategy that informed the Tractatus was rejected as unnecessary, a needless flourish. No methodological or rhetorical ladders are needed to climb our way out of nonsense. The leap is not that far.

Instead, what is appealed to is sound linguistic practice, the “perspicuous presentation” (übersichtliche Darstellungen) of which dissolves philosophical problems as our very human inability to see what is right in front of us, the ordinary workings of language-as-we-use-it. Generalized theories of meaning (which inevitably must address the “representationalism” issue), are as useless as they are impossible. What is needed is condensing the “cloud of philosophy” into a “drop of grammar.”

Wittgenstein, in the Investigations, is not interested in taking sides in any philosophical debates (realism v. antirealism, essentialism v. conventionalism, empiricism v. rationalism, etc.); he is interested in getting his readers to put aside these debates less as “nonsense” than as as diseases of the intellect. When asked what time it is, both the realist and idealist will look at their watches and respond “It’s 12 o’clock” when it’s noon, and the empiricist and the rationalist physician will claim “the virus caused his cold” when that is the case. To continue taking seriously metaphysical debates about the reality of time or causality is not so much illogical as it is sick.  It is only a certain mind-set that would see intractable epistemological and metaphysical conundrums raised by these essentially “grammatical” issues: “Grammar tells us what kind of thing anything is”, such as “time” in “what time is it?”, and so on. “Grammar” should, I think, be glossed alongside Donald Davidson’s quip, that “truth of sentences is relative to language, but that is as objective as can be.” To aspire to an objectivity stronger than this is to be held in the grip of a picture that distorts our relation to the world and ourselves. It is a matter not for sounder theorizing, but Zen-like therapy.

Think, for example, of the classic Zen “Mu koan.”[2] A monk asks the sage Joshu whether a dog has a Buddha nature – a question whose answer seems a no-brainer, since everything according to Buddhism has a Buddha-nature. Joshu emphatically answers “mu!” — literally “no”, but implying something more globally negative, like “nothingness!” or “not applicable!” This is an important nuance, since were the monk to retort “So, are you saying that a dog does not have a Buddha-nature?” Joshu would also have responded “mu!” Any attempt to verbalize a “metaphysics of the Buddha-nature” on the part of the monk merits a reply of “mu!” – which is to say his theoretical stance cannot be answered by another “better” theoretical stance, but by abjuring theory altogether. “Mu!”, like a blow to the head with a stick, or an irrelevant comment like “three pounds of flax”, is therapy against any impulse to metaphysics or epistemology. (The affinities between Wittgenstein – and for that matter Heidegger – and Zen are numerous, and deserve to be articulated, but in such a way that they do not fall prey to Nietzsche’s withering critique of being yet another species of facile “Western Buddhism.”)

For the later Wittgenstein, our true statements neither represent nor fail to represent reality-as-such and in-itself, for there is no way to wedge ourselves between language and the world to either specify “the correspondence relationship” or dismiss it as “a mere projection.” That is the metaphysical-epistemological construal of the world (roughly congruent with Heidegger’s “onto-theology”), and overcoming this is perhaps the chief motivation behind his view of philosophy-as-therapy and how it overcomes philosophy-as-theory. The realist and the antirealist are both held captive to bad metaphors: Plato’s “carving reality at the joints”, the Cartesian “inner-theatre”, Logic as “the ideal language” that really represents things (or, at the other extreme, the “viscousness” of Sartre’s l’etre en soi). There is no global way in which what we say represents “the real”. But if this is so, there likewise is no way to say that we can determine which propositions don’t represent, which are true by “convention” only, either. Joshu’s response fits here too: “Mu!” Neither is applicable.

It is not that words don’t represent things – that “dog” doesn’t mean dogs, that “Paris” doesn’t refer to or represent Paris, or that “Bernie Sanders” doesn’t stand for Bernie Sanders, and so on. Ordinary, everyday representation is in perfect order as it is, as part and parcel of our sound human language-practice of talking about things, people, and places. It is only when our “impulse to philosophize”, as Stanley Cavell puts it, our ineluctable need to transcend the ordinary, the human, as the final court of appeal, kicks in that we get ourselves into trouble, become epistemologists and/or metaphysicians, and make ourselves spiritually sick.

There are many examples in Philosophical Investigations, On Certainty, and Wittgenstein’s later writings that support my claims here, which I shall explore in later posts. Here is one from Zettel that is often quoted, often misunderstood, which I think beautifully illustrates how Wittgenstein was neither a realist nor an antirealist.


“We have a color system as we have a number system. Do the systems reside in our nature, or in the nature of things? How are we to put it? Not in the nature of numbers or colors.” (Zettel, 357)


This has been invoked (by Bernard Williams, for instance[3]) in support of the claim that Wittgenstein was an antirealist or even a “linguistic idealist.” But listen to what Wittgenstein is actually saying. The philosophical realist stakes her claim on these systems being rooted in “the nature of things” as opposed to our nature; the antirealist reverses the order and claims that “we make” numbers and colors. But numbers and colors are already articulated in our grammar, which tells us what kind of thing anything is. So “not in the nature of numbers or colors” undermines both the idea that numbers or colors are self-identifying entities/classes/systems that are what they are because our noetic radar is picking them up as natural kinds “really out there”, and the idea that “numbers and colors” are simply “our projections”, classes and categories we make up at will and whim, like Borges’s “Chinese Encyclopedia” quoted by Foucault in The Order of Things.[4]

One could say that the number and color systems emerge both because of the nature of things and our nature: in fact, John Dewey did say something like this in Experience and Nature, when he understood the subject/object distinction to have emerged from intelligent transactions between organism and environment, where organism and environment are reciprocally, functionally emergent. While Dewey and Wittgenstein both work in the same pragmatic neighborhood, Wittgenstein did not engage in a Deweyan “descriptive metaphysics” because he had no interest in detailing “the generic traits of experience.” He was instead interested in the perspicuous presentation of Grammar (which, as an aside, looks a lot like Heidegger’s “Being” in Sein und Zeit part 1 division 1), so that we won’t see things like numbers or colors and their systems as necessarily bound up in the question of whether they are “really out there or just made up by us in our heads” – that is, taking this classic metaphysical/epistemological conundrum as meaningful. We have a number system and a color system. We grasp what sound linguistic practice guides us toward when we speak about numbers and colors. We know what inferences to make, we know which to avoid, and we know how to talk about truth and falsity in these domains. We thereby and therefore know what numbers and colors are. Yes, getting clear on numbers and colors has something to do both with “the nature of things” and something to do with “our nature.” But we should not take it as imperative to parse through and definitively specify what is “out there” versus “in us”, or which parts or aspects are “out there” and which “in us”, because that would be a pointless attempt to speak outside speech or think outside thought. It would be another symptom of the disease of the intellect which traditional, theoretical, non-therapeutic philosophy is.

As Stanley Cavell often put it, while there is something very human in ramping up problems to this philosophical (i.e., metaphysical/epistemological/onto-theological) level, it amounts to the very human tendency to aspire to the super-, trans-, or even in-human, a tendency that is impossible to carry out, and ultimately detrimental to our humanity. Thus the Investigations should be viewed as, essentially, a moral polemic against entering into the “realism v. antirealism” debates. For Wittgenstein, it is wrong to be a metaphysician, or an epistemologist, because it is cognitively misguided and pointless, yes, but more importantly because it is bad, a turn away from a healthy, flourishing human life.


[1] See Richard Rorty, “Introduction” to Objectivity, Relativism, and Truth: Philosophical Papers vol. 1(Cambridge, 1990); Charles Taylor, “Overcoming Epistemology” in Philosophical Arguments (Harvard, 1997); Donald Davidson, “On the Very Idea of a Conceptual Scheme” in Inquiries Into Truth and Interpretation (Oxford, 1984).

[2] See https://www.ibiblio.org/zen/gateless-gate/1.html

[3] Bernard Williams, “Wittgenstein and Idealism” in Moral Luck (Cambridge, 1982)

[4] See http://www.multicians.org/thvv/borges-animals.html

Also for you:

Michael Quirk

Previous post

#BlackLivesMatter and #Fightfor15

Next post

In Defense of Gravitas