From the ALWS archives: A selection of papers from the International Wittgenstein Symposia in Kirchberg am Wechsel

1. A Linguistic Turn with a Transcendental Twist

According to David Pears, "The simplest general characterization of [Wittgenstein's] philosophy is that it is critical in the Kantian sense of that word. Kant offered a critique of thought and Wittgenstein offered a critique of the expression of thought in language." [Pears; 3]

To get the full analogy, we must take a step back to see the motivation for their projects.

From a high enough altitude it seems that Wittgenstein and Kant begin at the same place. They both wish to set limits for knowledge. They both wish to use an understanding of those bounds to debunk dogmatic metaphysics-- and yet to make room for freedom, if not faith. The question is whether or not, after they perform their investigations into the limits of knowledge and speech, they still maintain a conviction that "noumena" or "ineffables" exist.

It is clear that Kant, employing the transcendental method he develops in the first Critique, finds a way to re-affirm the "transcendental Ideas" of God, Freedom and Soul via a conviction in the experience of the moral law. Hence Kant finds a circuitous route back to some conception of his "ineffables" as postulates. It is also clear that Wittgenstein in the Tractatus believes he finds no way to talk about his ineffable domain.

There are several "ineffables" at issue here. Wittgenstein primarily talks about the say/show distinction with regard to "logical form", that which makes our sentences capable of conveying meaningful information. I will call this the "transcendental" ineffable. But there is also an ethical connotation to the ineffable. This I will call the "transcendent" ineffable, what Kant called "noumenal" being, which included God, Freedom and Soul.

In approaching the question of the ineffable in Wittgenstein, we might first ask what is speakable. With regard to what can be objects for the understanding, and hence what can be spoken, the Tractatus can be considered a linguistic extension of Kantian epistemology. The appropriate application of language for Wittgenstein is the appropriate application Kant's categories. The speakable is the sensible world.

According to Wittgenstein, the only legitimate propositions were the propositions of empirical science; once we stray from facts about the world, we pulled by the syntax of language into nonsense. According to Kant, concepts without intuitions are blind. When we don't have sensible intuitions we are pulled towards empty metaphysics, driven by Reason's regulative search for unconditioned unities. For both, cognitions are the proper objects of language and are eminently sayable. If we can see, smell, feel, taste, or hear something we can name it and describe relations it has with other things.

Once Kant gave his ideas with regard to the proper application of concepts a paradox arose which ate away at the viability of any statement that purported to explain how knowledge in general is possible: If all we can legitimately understand or speak of are the sensible objects of experience, how is it that we can understand or express anything about that which produces knowledge or these objects for experience?

For Kant there were a whole class of propositions that weren't about the empirical world but which he still purported to make sense. These were statements about the transcendental realm, i.e. statements about the conditions for the possibility of knowledge. For Kant, we could use what we knew about the world as a clue to the transcendental, and ultimately we could know necessary conditions for the possibility of the world, understood empirically. Wittgenstein, on the other hand, with the final lines of the Tractatus, banishes everything transcendental into the domain of the unspeakable. Pears says, "Kant felt no scruples about presenting his philosophy in factual language, but Wittgenstein was more rigorous." [Pears; 7] In fact, from Frege to Wittgenstein, the idea that it is possible to say anything about general conditions for the possibility of knowledge progressively loses credibility.

Frege believed that it was impossible to speak literally about that which allowed us to cognize the world and speak with sense. In "The Thought" Frege simply says "something non-sensible" is responsible for our ability to experience the world and grasp the thoughts we express in language. And in "On Concept and Object" it's clear we can't define the basic categories of understanding but we have to assume them in our definition in order to be speaking coherently at all. All we can provide are metaphorical elucidations like "saturated" and "unsaturated" or "function" and "argument" to hint at irreducible, ineffable terms. And if this is so with the general logical categories in our language that allow us to represent the world, how much more so when we attempt to talk about categories and the structures that present the world?

Frege acknowledged the objectivity of concepts and their application to the world of experience, but with regard to any transcendental activity he was very vague. Although heavily steeped in the Kantian tradition, there was a distance between Frege and Kant, according to J. Coffa, "underscored by the paucity of remarks on that most disturbing of Kantian problems, the character of objects of knowledge and their constitution through the categories." [Coffa; 67]

Wittgenstein continues the trend Frege began, pushing the transcendental further into the ineffable transcendent. Cora Diamond mentions the similarities, as discussed by Peter Geach, with regard to the ineffability of "concept" and "object" and, according to Diamond, Wittgenstein adds other words to the list such as "'possible' and 'necessary' and 'impossible'." [Diamond; 184] All logical concepts that work to provide meaning are inexpressible.

This repudiation of transcendental conditions strictly abides by the restrictions Kant put on the understanding. Wittgenstein says we can only speak of objects. [Tractatus, 3.221] That which presents us with objects are not objects themselves. To speak of "logical form" is already illegitimate, because "it" is not an it. "It" would fall into the category of being "neither a something or a nothing". [Philosophical Investigations, #304] We can only talk about its in language and its are the things we make scientific propositions about. We cannot get to any proto-reality. All that we can talk about we can see on the surface. We cannot go behind the scenes to see how knowledge is --or must be-- produced in the way Kant had in mind. But is it all nonsense?

2. The "Ineffability of Semantics" as the Ineffability of the Transcendental

It is possible that Wittgenstein, with Frege, adheres to what Coffa calls the "factuality of meaning". Although we cannot speak of transcendental conditions for knowledge that might provide some sort of justification for our use of logical language, there is a right and a wrong logical symbolism; there is something to show. Pears, an "arch-realist", attributes this to an acknowledgment of a "deep structure" which all factual languages have but which cannot legitimately be expressed in any factual language. [Pears; 6] He explains Wittgenstein's reasoning for saying this structure is ineffable:

"If factual language could contain an analysis of the conditions of its own application, the language in which it analyzed them would itself depend on further conditions, which would still remain to be analyzed, and so on to infinity... Factual sentences, like pictures, present a view of the world but they do not present a view of what made the original view possible." [Pears; 7]

This is what Jaakko Hintikka would call a presuppositional motivation for the "universal medium" view of language. In this view, shared by Frege and Wittgenstein, we cannot get outside of language to examine the basic concepts that give it intelligibility. For Hintikka this results in the "ineffability of semantics" or the inability to say anything about the relations between language and the world, but it does not preclude the idea that strong views on semantics are held. [Hintikka; 24]

Diamond, a "dissenter", seems to agree with Pears' principle, at least with regard to Frege, [Diamond; 1] but, for her, taking Wittgenstein seriously makes this merely a "transitional" statement in the "before-you-throw-away-the-ladder mode of speaking". [Diamond; 185] In the end, nonsense is nonsense. The only thing that is shown is that a form of words we thought had sense, actually does not.

What Hintikka stresses that Pears does not is that semantics is ineffable because any expression of the meaning at a fundamental level is redundant, both of the form that gives it sense and of the explicit information it provides, if any. The conditions for sense have to be presupposed in any statement that makes sense, so we never really get to say those conditions. Hintikka thus says that, for Wittgenstein, when we try to express what gives our sentences meaning, "at worst, the result is nonsense; at best (viz., if we simply repeat what is already presupposed) the result is vacuous." [Hintikka; 23]

In addition to our susceptibility to put its where there aren't any, Hintikka thus points to another way in which Wittgenstein's "theses" can sound intelligible and yet be nonsense. Hintikka cites Wittgenstein in Culture and Value: "The limit of language is shown by its being impossible to describe the fact which corresponds to a sentence...without simply repeating the sentence. (What we are dealing with here is the Kantian solution to the problem of philosophy.)" [Hintikka; 24]

This prime statement about the ineffability of semantics is also a statement about the ineffability of the transcendental. The Kantian "solution" here is being criticized. Wittgenstein is saying that Kant's attempt to lay out the necessary conditions for the possibility of knowledge does not provide us with any new information. It merely repeats in disguise what we already knew and what we perceived as a problem. (Does Kant's "transcendental logic" really tell us any more than we know by general logic?) And if such statements are at best redundant (tautological) then they are at base nonsense when taken as an explanation. Nonsense does not get resurrected into sense once we've explained why its nonsense, because when it seems to make sense it is relying on that which it purports to explain. Nonsense gets eternally buried as redundancy.

So much for the transcendental ineffables, what of the transcendent?

3. "to arrive at where we started and know the place for the first time" (T.S. Eliot in Pears)

It is easy to see how a transcendental connection between our words and the world might be shown but not said, since we use language meaningfully to describe ordinary experience. It is harder to see how any transcendent entities or acts can be shown. From the point of view of the understanding it is impossible. No fact of the world can tell you that an action was ethical-- for Kant or for Wittgenstein. Neither God nor the soul is anything that we can see, touch, feel or sensibly experience. We speak only of that which can be understood, the empirical ego and world. The opacity of the will --even to ourselves-makes it impossible to say whether the world is waxing good or waning evil.

But aren't the arch-realists simply taking Wittgenstein at his word? He says "The book's point is an ethical one" [in Pears; 191] and "There is indeed the inexpressible. This shows itself; it is the mystical." [Tractatus, 6.522] Aren't we getting somewhere by means of his propositions?

The problem comes in understanding how metaphor worked for Wittgenstein. "A simile must be the simile for something." [at least by 1929, Wittgenstein; 42,43] There has to be a way of cashing out a metaphor in literal language. To say we can't literally speak of something is to say we can't speak of it even metaphorically. Once you see philosophical propositions as nonsense, and not even as metaphorical elucidations, it is hard to hold on to the idea that there is anything there that those propositions were getting at. Diamond thus says: "The final step in the philosophical journey... leaves us without that description or any supposedly unspeakable understanding corresponding to it." [Diamond; 3]

The way of cashing out the transcendental is to show it's redundant. But there is no way to cash out the transcendent. One might try to say that the ethical is limned by limiting the speakable --a via negativa-- but this doesn't work. How would this be cashed out? The analogy fails to convey the sense of the ethical, and the person, as the limit of language and thought. There is no way at all to characterize what is on the other side of the speakable or thinkable. And a charitable reading of the Tractatus still seems to side with the dissenters.

So is there nothing to see? Is Diamond right and Pears wrong? Yes and No. A clearer way of saying what we are left with is to go back to our Kantian framework. Kant, in employing his transcendental method, begins with convictions that he cannot doubt. Pears sees Wittgenstein as circling back to the beginning, but for Pears this involves a platonic "uncritical realism". [Pears; 10] Wittgenstein's convictions, however, may start smaller than Plato or Kant's. The transcendent ineffables were transcendental postulates that we forgot were postulates. What we were actually convinced of, before we were misled by language, is our ability to communicate meaningfully and behave ethically. This is the substance archrealists look for: the ethical before it gets built up into a metaphysical notion. But any expression of this immediately pulls us towards a philosophical claim. We get pulled from "this leads to that" and "this was a free act" to a completely determinate physical world and an absolutely free moral subject. This pull Wittgenstein resists.

God, Freedom and the Soul can't be spoken about, they are the metaphysical nonsense that gets thrown away with the ladder... and the ladder is the transcendental structure that supposedly got us to where we started from. We are left with the aspect the world takes on when the questions disappear.

But it may be that we all try to make more sense of Wittgenstein than he could make of himself, and the dialectic Goldfarb points out, between "theses" and their undermining, continues.