The Paradox of Transcendental Reflection

I first read the Critique of Pure Reason seriously over four decades ago. Like many novices reading Kant, I was impressed by the epistemological distinction between the a priori and a posteriori and the semantic distinction between synthetic and analytic judgments. Kant’s question intrigued me: By what right do we claim truth in synthetic a priori judgments? What justifies the assertion that deep reflection allows for an advancement of knowledge about the ultimate features of reality?

Kant claimed that a true analytical judgment is one where the meaning of the predicate is included in the meaning of the subject, while a true synthetic judgment is one wherein the meaning of the predicate is not so included. A synthetic judgment is thus “ampliative,” that is, to say that “all bachelors are happy” is to make an assertion that cannot be known to be true simply by thinking deeply about what the word ‘batchelor’ means.

As is well-known, Kant criticized traditional metaphysics by showing that its claim to extend knowledge “beyond the realm of possible experience” was chimerical. In the absence of intuition – that which is “given” through sensation – concepts simply relate to other concepts analytically or semantically. Since no intuitions “fall under” the concept ‘God’, we cannot know that ‘God loves human beings’, unless, of course, we are able to claim this to be merely an analytical truth, that is, that the concept of ‘God’ includes as part of its very meaning ‘loving human beings’. Putative metaphysical judgments that turn out to be analytical in this way are, for Kant, “regulative judgments.” While incapable of miming the ontological contour of the supersensible world, they are useful in ordering our supersensible concepts, and thus our thinking about the supersensible world. Kant thought his analytic/synthetic distinction exhaustive. Either judgments are analytic or synthetic; tertium non datur.

In reading the metaphysical and transcendental deductions of the Critique, I was struck by the oddity of what Kant was writing and what I was doing in reading. Kant was offering arguments about how it is that knowledge consists in the application of concepts to intuitions such that there is a “synthesis of the manifold of sensation.” I thought that what he wrote was plausible and was even able to grant that what he said was likely true. But with this an uncomfortable argument seems to emerge.

Let us regard as true the Kantian statement, ‘an object is that by concept of which the manifold of sensation is united’. If this statement is true, it must be true either analytically or synthetically. But clearly it is not an analytic truth for no amount of simple reflection upon ‘object’ allows one to conclude by meaning alone the concept ‘that by concept of which the manifold of sensation is united’. Therefore, it must be true synthetically.

But now the discomfort becomes acute because it is unclear what sensible intuitions must be united to make true the judgment ‘an object is that by concept of which the manifold of sensation is united’. Synthetic judgments for Kant are true a posteriori except for arithmetic and geometry which make direct appeal to the pure forms of sensibility. But neither sensibility nor its pure forms are synthesized in judging true the proposition, ‘an object is that by concept of which the manifold of sensation is united’. Accordingly, the sentence seems to be left without justification, and with it a great many of the statements Kant employs in his discussion of the transcendental unity of apperception.

Clearly, I had stumbled upon the paradox of transcendental reflection. Kant asks his readers in the Critique of Pure Reason, “What are the necessary conditions for the possibility of experience as such?” After claiming that all propositions are either analytic or synthetic and that synthetic a priori metaphysical judgments are problematic in extending knowledge beyond the realm of possible experience, he writes hundreds of pages in which he is seemingly using synthetic a priori judgments justifying these claims. Kant’s transcendental reflection apparently did not have to follow the same justificatory practices with respect to knowledge and truth that our reflections on the nature of things must follow. When reflecting upon the conditions of our knowledge of the nature of things, we no longer need to play by the same rules as we do when reflecting simply upon the nature of things.

Transcendental reflection, our thinking about how we think things, is exempt from the rules it prescribes to our thinking of things. Kant had perhaps “done away with knowledge of God to make room for faith,” but in doing so he created a cottage industry for philosophers. They who could not with justification lay out the truth conditions of ‘God created the universe’ -- there are no intuitions united either under the concept ‘God’ or the concept ‘created the universe -- could now claim truly this statement: ‘The judgment that ‘God created the universe’ cannot be regarded as true because there are no sensations falling under ‘God’ and ‘creating of the universe’. While clearly this proposition is a priori, we need no longer worry if it is synthetic or analytic. It is a statement within the field of transcendental reflection after all, and while such reflection sets the rules for meaning and truth for other provinces, like the Politburo of the old Soviet Union, it is wholly exempt from the rules that it prescribes for others.

The problem of transcendental reflection is a problem of grounds: What legitimates claims of transcendental truth? Why can we not ask with sense whether the statement ‘truths divide exhaustively between the analytic and synthetic’ is itself an analytic or synthetic statement? After noticing that true judgments are both “clear and distinct,” Descartes argued that clarity and distinctness form the very criteria of truth. Analogously, we might argue that reflections that philosophers regard as true that do not meet the truth criteria of what they prescribe are transcendental. Accordingly, the claim that we can say truly that there are conditions that do not apply to what is said truly actually constitutes the very criteria of the transcendental.

The problem of the transcendental standpoint and the truths discerned in occupying it has often been overlooked or ignored. The verificationist criterion of meaning asserted that only those propositions are meaningful that are comprised of tautologies or can be checked up upon in experience. But clearly, the statement that ‘only those propositions are meaningful that are comprised of tautologies or can be checked up upon in experience’ is neither a tautology nor can it be checked up upon in experience. Faced with the inability to say with truth the material conclusion of their argument, some retreated to regarding the statement as neither true nor false, but merely a proposal. Of course, this begs the questions of why one would be motivated to adopt the proposal in the first place.

The twentieth century, though often increasingly wary of transcendental reflection, has nonetheless had difficulty avoiding it. After laying out the conditions making possible propositions of sense, Wittgenstein in the Tractatus points out that none of the propositions he has written can be regarded as either true or false. They are like a ladder one climbs that can be thrown away upon reaching the summit. Such propositions might be elucidations, but they themselves have no truth conditions. Wittgenstein famously says that while saying what cannot be said, he nonetheless hopes in this saying that something might be shown. While one cannot state the conditions for the meaningfulness of propositions meaningfully, one can nonetheless show in one’s saying how to use propositions meaningfully. Wittgenstein notes sadly that the most important things of life cannot be said.

Wittgenstein knew that Russell’s paradox had spelled doom to Frege’s logicist program. That program depended upon the unrestricted use of the axiom of comprehension within set theory, the notion that any set of conditions clearly demarcate and distinguish sets from one another. Set theorists make extensive use of sets that have sets as their own members. Given the axiom of comprehension, this condition should uniquely determine sets, that is, for all sets, either they have sets as members of themselves or they don’t have sets as members of themselves: tertium non datur.

Russell then asked us to consider those everyday sets that don’t have sets as members of themselves, e.g., the set of elephants has as its members elephants, not sets of elephants. He directs us to consider the collection or set of all of sets that are not members of themselves. Now since we can ask with sense whether a set is a member of itself or not a member of itself, and tertium non datur, we can ask with sense whether the set of sets that are not members of themselves is itself a member of itself or is itself not a member of itself. A little reflection shows that if the set of all sets is a member of itself, that is, is a member of the set of all sets that is not a member of itself, then it itself must not be a member of itself. Conversely, if the set of all sets that are not a member of themselves is not a member of itself, then it must be a member of itself. That this paradox was not allowed in logic shows that somehow logic was not going to be regarded as a case of transcendental reflection, for it itself must obey its own rules!

That logic must obey its own rules is assumed in the celebrated Incompleteness proof of Gödel. He showed that paradox arises on the assertion that all known mathematical truths (tautologies) can be derived from a finite set of axioms. By ingeniously semantically encoding information into the syntax of arithmetic, it can be proven that there will always be a provable true proposition G from some axiom set that states that it itself cannot be proven on the basis of that axiom set. Adding a new axiom will not solve this problem because a statement can be proved stating that it cannot be proved on the basis of the new axiom set. While logicians carefully distinguish their metalanguage from the object languages about which the metalanguage speaks, they do not countenance theorems in the metalanguage contradicting those of their object languages.

Paradox dooms logic, but not transcendental reflection! How else can we explain the rise of phenomenology with all its fanfare and hopes? Reading the texts of Husserl, Heidegger and others brings us again into the orbit of the transcendentality that Kant had birthed, and Fichte, Shelling, Hegel and others so effectively exploited.

The phenomenological tradition of such reflection differs from the Neo-Kantian tradition in that while the latter is engaged with the principles by which knowledge is legitimately had, the former utilizes evidence.  Husserl realized that truth is itself not something that can be accounted for on naturalistic assumptions, and thus argued that so-called natural truths must rest upon non-natural grounds. Accordingly, the very grounds of the truth of metaphysical truths must be non-metaphysically investigated. One must go zu den Sachen selbst and bracket questions of metaphysics and the natural world in order to apprehend those grounds upon which the natural world and metaphysics rests. These grounds, thought Husserl, were to be found in the direct apperception of that which is immediately given to consciousness.

But phenomenological reflection proved to be no easy task, and reflection on “the things themselves” was soon seen to involve reflecting upon many other things, some of which were not so unambiguously evidence. In fact, the criterion by which to evaluate the nature and strength of evidence was not clearly something one could simply “see” evidentially. Marshaling evidence and relating that evidence to philosophical problems seems to involve principled transcendental reflection. Husserl knew this, and by the publication of Ideas in 1913 adopted the position of transcendental idealism that he once wished to bracket. Transcendental reflection demonstrated the necessity of a transcendental ego related noetically to the Sinn-world of noematic content. This transcendental ego could not be examined phenomenologically without presupposing that very ego under investigation. The problem was that transcendental reflection seemed to require a transcendental ego that was, by definition, not amenable to phenomenological investigation.

It is at this stage that Heidegger enters our story, penning Sein und Zeit and striving mightily therein to avoid the paradoxes to which Husserl’s hidden transcendental ego fell prey. By re- thinking what a transcendental ego really is, Heidegger was able to avert the problem of how the transcendental ego can direct itself upon its world. For Heidegger, the occult ego of Husserl became Being-in-the-world, Dasein. The ego is already embedded in its world and it is this embeddedness. With this step Heidegger would try to do something nobody had yet succeeded in accomplishing. Heidegger wanted phenomenologically not only to access those beings in the world that constitutes the basic experience and structure of Dasein, but he wanted to examine the conditions for the intelligibility of phenomenologically accessible beings in the world; he wanted to coax out of hiding those worldly conditions making possible beings in the world. His interest was in the be-ing (“to-beness”) resident within the horizon of the world itself. He claimed that his investigation was ontological, that it had to do with be-ing, that is, it concerned not primarily beings, but those conditions of intelligibility that made possible the intelligibility of beings as such. 

But Heidegger’s work in Sein und Zeit was beset by transcendental paradox as well. His pointing out of different ways of being seems at times to leave out the very possibility of a way of being doing the pointing. Take, for instance, his distinction between Vorhandensein (present-at-hand be-ing) and Zuhandensein (ready-to-hand be-ing). This distinction is fundamental for Heidegger. Objects appear to us either as “present-at-hand” or “ready-to-hand”, either as objects having properties or as equipment to be used in our everyday pragmatic concerns. But what is the being of the one who distinguishes be-ing-present-at-hand from be-ing-ready-to-hand? Is the distinction between the objective and pragmatic an objective or pragmatic distinction? If neither, then should Heidegger not have distinguished some other category beyond the objective and pragmatic?

Heidegger’s detailed analysis of the be-ing of Dasein in Sein und Zeit seems to push towards theoretical comprehension, a present-at-hand description of those fundamental structures that are not in themselves present-at-hand. But this is exactly what transcendental reflection does: It attempts a theoretical description of a province of being that cannot be theoretically described.  Transcendental phenomenology perhaps has made the most valiant attempt to grant explicit truth conditions for statements of the transcendental. Clearly, Husserl was attempting in his formal ontology to escape the paradox of transcendental reflection.  But as mentioned before, the hope that there could be a stable province of being impassible to its own investigation was quickly extinguished by Heidegger's insight that knowing being is itself an activity of being, that at the foundation of being, there is be-ing, and that there is be-ing all the way down, as it were.  

The paradox of transcendental reflection are encountered by a being, who in its be-ing, has be-ing at issue for it. Such reflection and paradox can sometimes be brought to the surface by the Geisteswissenschaften, who realize profoundly that the Naturwissenschaften proceed so successfully because they exclude what to the human spirit is central: We are not who we are and can never not be who we are.  Difference rules the first set of disciplines and identity the second. 

So what is deeper in human experience, the geistliche paradoci of transcendental reflection, or the tidy coherency of  natural science?