Saturday, April 5, 2008

Who's Afraid of Idealism?

Today, it is usually enough to refute a theory just to demonstrate that it in any way implies idealism. There was a time, of course, when the reverse was the case, when it was enough to refute a theory just to show that it was close to empiricism, or possibly psychologism. Carnap seemed to regard pyschologism as just another of the basic informal fallacies, a sort of category mistake. One could go straight from, ‘This theory is psychologistic’ to ‘This theory is false’ without needing to go through the premise, ‘Psychologism is false,’ in the same way that one may go from ‘This theory begs the question’ to ‘This theory is false’ without needing to go through the premise ‘Begging the question is false.’

The problem with idealism seems to be that it violates a deep-seated intuition we all have, namely, that most of the universe (outside of fiction and imagination and maybe mathematics) is real, by which we mean, independent of the mind. However, transcendental idealism a la Kant admits this much. The apparently obviously false notion in Kant’s philosophy is that we have no access, period, to this really real world, and thus, in a certain sense, everything that we believe is false. Of course, Kant would reply that truth-predicates only make sense as applied to the phenomenal world, so in fact our basic picture of the world is true—but still, this is hair splitting. We are compelled, as it were, by this deep-seated intuition to believe that not only is there a reality out there independent of our own mind, but that, by some means or other, we must have access to this reality as it is really is an sich. As I mentioned at the start, I’m not convinced that this deep intuition results from anything more profound than a particular cultural milieu, but I admit that it’s there.

Let’s not forget that there are equally compelling, if not as deep-seated, reasons for thinking that some sort of (transcendental) idealism is correct. For however one explains it, it certainly also seems as if we bring pre-fabricated structures already to each receptive experience, and only after these structures have done their work do we have anything like objective perception or thought. Whether these structures are transcendental in the way that Kant suggested, or whether they are genetically encoded neurological protocols, or the existenielles of Dasein, or Gadamer’s tradition, is in some sense beside the point.

Can we do justice to both of these intuitions? Can we be both transcendentalists, in the sense just described, and realists, also in the sense just described? I think we can, so long as we are Husserlians.

Whether our transcendental intuitions can do justice to our realist intuitions depends in no small part on how we understand ‘transcendental.’ Kant explained his concept of transcendental through the concepts of universality, necessity, and a priori synthetic judgments. The categories, while not separable from experience in fact, are notionally separable, and must be deduced from the conditions for the possibility of full, objective apperceptive consciousness. Baldly put, Kant’s transcendental arguments have the form:

Given S
M is a necessary condition for S (If S then M)
Hence, M.

This gives rise to the problem that objective perception results only after (in a transcendentally logical sense) the categories have structured some non-categorically structured stuff, and thus, the objects that result are not the objects that, really, are there. Kantians don’t like this description, but there is no getting around the fact that, for Kant, consciousness in some sense creates the objects it is aware of. This is why Kant is an idealist. Ultimately, Kant’s idealism is incompatible with realism because it comprehends the ideal structures of transcendental consciousness as the condition of possibility for the real world objects of objective perception and thought.

The hardest methodological notion of Husserl’s to make sense of, but also the most promising, is his theory of phenomenology as an eidetic (or ‘formal’) descriptive science. Husserl is not interested in discovering or deducing the a priori conditions of possibility of objective experience; he simply wants to describe that experience as it manifestly is for us. This experience, let it be noted, is directly realist. I do not experience red sensations, or retinal stimulations, or the workings of my cerebellum. I see the red stop sign. Husserl has no interest in denying this obvious feature of phenomenological experience, nor of deducing its possibility; he just wants to describe it.

This notion of describing experience no doubt makes many uneasy, so let me say something about it. Of course, there is a very wooly notion of description, such as if I were to ask you to describe your favorite painting. You would not be communicating to me any fact, but rather your impressions (I like the theme; the colors are pretty; it makes me feel grand). If this were all there were to phenomenological description (as some phenomenologists unfortunately seem to think), then phenomenology would be silly. And yet this does help to expose what I think it is that really causes the unease among many for phenomenological analysis, namely, that we should expect there to be some way, objectively, to police our phenomenological descriptions, and it does not seem like we can do so in any way like one would expect from a proper science: by the use of publicly available evidence, by repeatable experimentation, by direct modeling, and so forth. Now, I cannot make the point often enough, that this same criticism would equally apply to each of the formal sciences. Mathematicians, too, cannot make appeal to public evidence (I cannot observe you observing the square root of two), nor can they perform repeatable experiments (perhaps they could, but what would be the point?). Husserl understands phenomenology as precisely this sort of formal science. Phenomenology is a transcendental science insofar as it seeks to describe the invariant structures of whatever type of experience (perception, memory, imagination, mathematical and logical thinking). Transcendental means for Husserl invariant, rather than condition of possibility. Thus, while Husserl does argue that transcendental structures are a priori, they are not prior to experience; they are a priori in the sense that they are independent of any particular experience, and not built out of any particular elements in experience, in just the same way that our concept of three is independent of any particular set of three objects and is not built up out of any particular experience of sets of three objects. If we apply this lesson globally to the problem of idealism, we see how one might, from a phenomenological perspective, be both a (transcendental) idealist as well as a realist: just as our concepts of arithmetic are invariant among and independent of all sets of actual objects, but in no way are conditions of possibility for those objects nor somehow inconsistent with the belief that these objects exist in full-blooded sense, so too for all the invariant structures of every sort of experience whatever.


  1. Michael, a few thoughts:

    First off, I would have thought that Kant's slogan, "Empirically Real, Transcendentally Ideal" was a way of dealing with the two intuitions concerning reality that you mentioned. Especially since, to my knowledge, Kant took metaphysics to be nothing other than science.

    Second, and more substantively, I'm not convinced that your propositional formalization of Kant's understanding of 'conditions of possibility' is adequate. Although it's true that the forms of space and time, the categories, and the transcendental unity of apperception are necessary conditions for experience, they are not merely necessary conditions. Kant's transcendental arguments, which are not limited to the Transcendental Deduction, are effectively model-theoretic arguments. That is, a Model M validly implies m iff there exists a language L for which there is one and only one valuation s. Beginning with the fact that m, the regressive or abstractive argument structure of Kant's arguments attempt to explicate the underlying structure that accounts for m (i.e. constructs L, which is the metaphysical deduction), and then proceeds to show that there is one unique valuation of it (transcendental deduction proper). Your attempted formalism misses out on this, I think.

    third, Kant's idealism doesn't create -- so it seems to me -- the objects of experience, as much as it determines them. The Reason that Kant keeps some notion of the thing-in-itself is precisely because the spontaneous activity of the subject is determinative, not creative (it's also why there's such a huge literature on affectation, and double-affectation). It's only when you remove the dingansich, like Fichte does, that you can collapse determination into creation. In other words, Kant's conditions of possibility are conditions for determining appearance (of making them meaningful for similarly constituted subjects), not for creating or producing them. And this difference shows up in Kant's work via his distinction between passivity and receptivity. Whereas intuitions are passively given to the subject, our judgments are active syntheses of the manifold given in intuition. The basis of passivity is, in Kant, completely unclear, but he nevertheless asserts the difference.

    Moreover, Kant doesn't assert that the transcendental conditions of possible experience are ontologically prior to experience, since his whole analysis depends upon the fact of empirical experience. They are, after all, Ideal, not real. Without prior experience, Kant cannot get going; without the truths of physics, there's no Kantian Philosophy (indeed, Kant's shift from 'what is?" questions to "How is?" marks precisely this fact). These conditions, however, are logically prior, or prior in explanation (think here of Aristotle's conception of an essence: there is no formal cause of a thing without the thing, even if its essence is philosophically prior, is "explanatorily" prior to the matter it forms)

    Now, moving on to Husserl, I'm a little confused about what you take the difference to be between 'invariant' and 'condition for possibility.' Aren't the pure forms of space and time, and the categories invariant? What do you take to be the precise difference between 'invariant' and condition of possibility?

    So, while there are indeed differences between Kant and Husserl (not least the fact that the latter expands the range of intuition [e.g. one can intuit mathematical and other abstract Gegenstaende]) and specifically situates subjectivity within a historical, intersubjective world, I'm not sure it makes sense to call Husserl a Realist -- especially since he's pretty clear that he understands himself to be an idealist; the epoche, for instance, removes all ontological commitments we might have, in order to focus on the meanings of intuited things for consciousness in order to attend to the spontaneous, meaning-bestowing activities of consciousness.

    Indeed, phenomenological description is nothing other than an attempt to self-reflexively bring forward these meaning bestowing activities, and to distill their underlying logic -- to provide a transcendental model for experience, which can account for every possible experience (in order to provide a normative ground for scientific activity).

    In any event, your description of eidetic analysis seems to miss out on the prior reductions that make it possible (i.e. the epoche, which shifts our attention from the thing in the world, to the meaning of a thing for consciousness, and the phenomenological reduction, which ushers in the noetic-noematic analysis that characterizes eidetic phenomenology), as well as the shift to Husserl's properly transcendental investigations (the normative structure of the spontaneous activity of consciousness underwriting -- as a condition of possibility -- eidetic analysis).

    Anyway, all this said, I would have thought that the quarrel between realists and Idealists really amounts to this: whether one thinks that one can meaningfully discuss 'the really real' outside of conceptual constraints, as if one could maintain a non-conceptual distinction between subject and object. Idealism, as I understand this umbrella term, holds that such an ideal is incoherent. But this doesn't mean that reality is completely mind dependent, only that what we know is dependent upon our subjectivity. Realism, again as I understand it, holds -- and conflates or equivocates -- 3 distinct claims: (1) knowledge and informed action are mind dependent, (2) ontological commitments refer to something non-conceptual (NB that no one knows what 'reference' is), and (3) subject and object can be meaningfully distinguished from one another in a manner that doesn't depend upon minds. To the extent that these 3 claims accurately characterize Realism, it seems clear that Realism is an incoherent position (for [1] implies that any distinction we hold depends upon minds). But then again, "Post-Metaphysical thinking" is just another name for Idealism.

  2. Alexei, I'll try to respond in more detail later. I think you're comments were extremely helpful and insightful. Let me a say a few things now, and then later i'll expand on them.

    Point 1) Kant certainly thought that the distinction between 'empirically real, transcendentally ideal' solved the problem, but I don't think that most people have been satisfied with this claim. Our deep-seated realist intuitions tell us, I believe, that there is a transcenentally real universe.

    Point 2) I would like to hear more about this model-theoretic working of Kant. It seems like a promising way to clarify what Kant is up to, but it doesn't affect the criticism I voiced in the post, because it is still a way of deducing the categories, instead of merely describing them (this I realize is an extremely fraught distinction, and it is this point I will try to elaborate upon later). According to Husserl, deduced propositions do not belong in transcendental analysis, which is one of the important ways his own transcendental idealism differs from Kants.

    Point 3) I like the create/determine distinction, but I don't think that it gets around the problem. The problem I allude to in the post is just the grounds by which Hegel criticized Kant, and many thereafter, namely, if we have objective experience only after the categories have structured that experience, then clearly we are 'distorting' the object of perception, regardless of whether that distortion is 'creating' or merely 'determining.'

    Continuing on, for Kant there is all the difference in the world between the transcendental object x and the noumenal object. This is why his empirically real object is not transcendentally real. So long as this distinction is in place, idealism, by my estimation, cannot be avoided, whatever distinctions and caveats one chooses to introduce.

    Point 4) Yes, I agree, Kant does think that experience must be given, but his project is to take such experience and ask, what must be the case, given such experience? So when I say that the categories are prior to experience, I mean that in a transcendentally-logical manner. Kant does not say that concepts (categories) are not determinate without intuition, only that they are empty. This is one issue where I think he and Husserl have an important disagreement.

    Point 5) I read Husserl precisely as a transcendental realist. I do not think that the intentional object is anything other than a real object, ie the intentional object is the transcendentally real object and also the empirically real object. Of course this needs further defense.

    Point 6) Following up on that, I too believe that ultimately kant and Husserl want to argue that traditional disputes between idealism and realism are either incoherent or just miss the point. That said, Husserl is committed, in a way Kant is not (for whatever reason), to realism--I argue. Husserl, I think, was simply never bothered by skepticism.

    Anyway, I'll try to elaborate on these perfunctory remarks later. I really need to think more about your comments, because I don't think that I've addressed them adequately. I hope at least I've made the lines of contention clearer.

  3. Michael, thanks for the quick clarifications, they do help a bit -- and I'm looking forward to a more elaborate response. Here's why:

    The distinction you want to make between deducing concepts and describing them really needs to be hammered out. For instance, although I'm aware that Husserl is 'against' a certain understanding of 'deduction,' he still deduces. That is, he's against the presupppositional character of deductive arguments, not against making what is implicit (i.e. sedimented in an intentional object or performance) explicit (thematic). Hence the epoche, which 'brackets' such presuppositions. When it comes to phenomenological description, however, there is still something like a 'deduction,' so long as we're willing to use this latter term in the sense of 'foregrounding what's implicit,' 'thematizing,' or 'making explicit'. IN this expanded sense of 'deduction', I don't think that Kant's transcendental arguments and Husserl's phenomenological descriptions differ too much (Strictly speaking, Kant's deductions aren't inferentially valid -- but that's because they're not deductions in the strict, logical sense of premises, inferential rules, and conclusions). For both begin with what is given (not presumed) in intuition, proceed to descriptively uncover
    their invariant conditions, and then thematize the spontaneous activities of a (transcendental) subject.

    To try and bring out the commonality between Kant's transcendental argument and the Husserl's phenomenological analysis, I had tried to point out that Husserl relies -- quite heavily -- on the self-reflexive character of description, which thematizes the spontaneous performances of an agent. There's a sense, then, in which we could talk about phenomenological description being a kind of deduction -- though one in which we do not have to 'assume' anything (It's also why Husserl is so fascinated by those points of where we equivocate between levels of description -- they indicate the slippages between mere 'description' and the activities that make description possible).

    So, I guess I'm wondering (1) if your distinction between deduction and description actually places Kant on the Descriptive side, and (2) whether you can actually draw a hard and fast line between demonstration and description in the first place.

    More simply put, I guess, until the distinction you want to draw between description and deduction is clear, I'm not sure why a model wouldn't be merely description, and not deductive. One is, after all, merely describing how some phenomenon m comes to be understood, according to a syntax L (Kant's forms of judgment) and a semantics s (Kant's categories) that are implicit in performances of a subject that determine the appearance m.

    Lastly, I'm still a little anxious about calling Husserl a Transcendental Realist. For, in the first instance, to take an intentional object to be the thing as it is in-itself, seems to oversimplify the intentional relationships between it, the community for which it 'is', and a given subject. Since Intentionality is neither purely receptive nor purely active, what we aim at and what we actualize aren't the same. So, a given object is never exactly what we take it to be, nor what we want it to become. Rather, it exists somewhere "in-between" the two. The Telos, as we will remember, is the thing-itself; it's not the given.

    And that's a long-winded way of saying, I guess, that a finished transcendental phenomenology might provide the outline for a transcendental realism (it's theme and task being to explicate how we might grasp the thing itself). But it certainly doesn't start off with a Transcendental Realism. Or, even more simply, I worry that you might have undialectically inverted the dialectician's slogan: the goal is the origin.

  4. In characterizing Husserl's thought, I think drawing a distinction between the early and later Husserl aids in mapping the realism one sees in his descriptions pertaining to perception. The move to a form of idealism after Ideas 1 does seem to suggest a different tone in his work. In addition, many scholars actually haven't paid attention to the static phenomenology such realism is easily associated but the genetic phenomenology practiced much later that attached to his idealism. I think these distinctions help clarify how one is to explain Husserl.

    Moreover, J. N. Mohanty defends a type of dualism in which he tries to reconcile the transcendental with the natural attitude in a way that comes off privileging the transcendental as that which makes possible the natural. Any aspiring Husserlian has to ask how to reconcile the 1st and 3rd person viewpoints with the natural attitude. I'm sympathetic to the project, and your plight suggested herein. I just don't know if transcendental realism is a proper way of understanding Husserl.