Transcendental Philosophy and Mind-Body Reductionism
Transcendental Philosophy and Mind-Body Reductionism


The notion of “representation” is central to Kant’s transcendental philosophy. But projects of naturalization and mind-body reductionism tend to reduce talk of representation to stories of causality and evolution. How does Kant fare in this context?

Table of contents

    In Wittgenstein on Language and Thought, Thornton gives an account of naturalization that he calls “representationalism”: “Representationalism attempts to explain linguistic content as resulting from mental content and then to give a reductionist account of the latter. Mental content is ‘naturalised’ through the provision of a causal explanation of content” (p. vii). Thus we have a two-step reduction, first from linguistic content to mental content, then from the mental to the physical. Mental content is seen in representations, “internal mental representations that stand in causal relations to things in the world” (viii). Fodor’s “descriptive causal theory” and Millikan’s “teleological, or natural selective” account are given as examples of such representationalism. Wittgenstein, on the other hand, Thornton shows, opposes such reductionist theories already in the first step: Representations and mental content so understood would be too isolated and internal, too detached from the outside world. Linguistic content and meaning cannot be understood this way. Instead, they should be seen as being more “outside” from the start, making sense only within language and its use in society.

    How would Kant fare in such current discussions? Certainly he has much to say about representations, Vorstellungen, Latin representationes. He also has read Locke and Hume and is aware of their empiricist accounts of impressions and ideas, as well as of Descartes’ res cogitans. Yet Kant does not take the same route they do. He is usually not mentioned in current discussions of naturalization and reductionism of the mental to the physical. Nevertheless, although he does not – returning to Thornton – talk about linguistic content, he has much to say about judgments and representations. Certainly representations must have meaning, and they often arise from perceptions. Objects appear to us, and we don’t make them up. Kant was not an idealist like Berkeley. He even distanced himself from Descartes, whom he also saw as an idealist (A 226/B 274). Unlike them, he never doubted the existence of the outside world. He saw himself as an “empirical realist” instead. So how would Kant react to current physicalist-reductionist accounts of representations and meanings?

    When looking at his early writings, such as his “General Natural History and Theory of the Heavens, or An Attempt to Understand the Structure and Mechanical Origin of the Whole Universe According to Newton’s Principles”, one might think he has a liking for naturalization. But when thinking of his Critique of Pure Reason, one starts to have doubts. Why is that?

    Central to Kant’s transcendental philosophy from his Critique of Pure Reason are the categories, imagination, understanding, schemata, and these can easily appear to be mental in some way. Not without reason, Kantian faculty-talk is sometimes seen as psychological. In any case, one might want to call Kant a representationalist of some kind, simply because the notion of Vorstellung, Latin representatio, holds a central place in his theoretical transcendental philosophy.

    But somehow Kant cannot be a representationalist of the kind Thornton has in mind. He does not understand his theory as giving an account of “mental” content in an individual person’s head, and he certainly does not try to reduce mental representations to causal stories.

    Such an undertaking would undercut his transcendental project from the start, or, rather, it would not touch it, but miss it altogether. Kant is not interested in an individual person’s head and in how empirical concepts arise and are acquired, as Locke was. To the contrary, he wants to establish a priori concepts that make such experience possible. These concepts, the categories, are not understood as mental in opposition to the physical. They are very special concepts. They make the physical as such possible, and the naturalists do not talk about them at all. In a sense, transcendental philosophy undercuts the mind-body naturalist’s project. Kant does not start with a mind-body dualism, with categories in the head and objects out there to be schematized. He also does not go in for the Cartesian res extensa – res cogitans distinction.

    For Kant, the categories underlie the world we experience, because objects are nothing but appearances brought under schematized categories. Not only objects, but even time and space are not out there, independently of us. The forms of time and space are subjective and make objectivity possible. If there is a “head” in the sense of transcendental philosophy, then the world has to be in it – at least the a priori aspects of it.

    Kant distinguishes between an inner and an outer sense, but not between an inner world of representations in the head and an outer world next to it.

    In particular, it is empirical causality that is seen, within transcendental philosophy, to depend on a priori causality, and therefore it would not make any sense to try to reduce the transcendentally mental to the empirically causal. Empirical concepts and representations might be naturalized, but not a priori ones. Transcendental philosophy and the Copernican revolution go the other way around. They are independent of any philosophical project in which causality is taken for granted, as Fodor, Millikan and others do. It is not that such projects do not make any sense. The point here is that even if they succeed, they will not answer Kant’s question about the possibility of experience and objectivity. Naturalizing projects take objectivity and the physical world for granted.

    It is as Barry Stroud says, in contrast to Jay Rosenberg’s historicizing, evolutionary and naturalizing accounts (Rosenberg, 616-20) of the Kantian minimalist “conceptual core” (615): “The absence of any interesting necessary conditions of thought and experience must be established, and not simply asserted as likely on general historical or ‘evolutionary’ grounds. Even the most uncompromising ‘evolutionary’ attitude would not preclude us from asking what it is that makes thought or experience possible – how it is possible for thought and experience to have ‘objects’, or be ‘of’ or ‘about’ something. It remains to be seen that that very general question itself must be given an historical or ‘evolutionary’ answer, even if an historical or ‘evolutionary’ answer must be given to the quite different question of who and why in the development of homo sapiens those conditions ever in fact came to be fulfilled.” (Stroud 1977, 81-82)

    Doubting causality in the way Hume did is mistaken in Kant’s eyes. We need an a priori concept of causality from the start to have any of the coherent experiences that we as a matter of fact do have. Particular empirical causalities can be learned about in experience, but not causality in general, universally, as such, which we need in order to have any meaningful experiences to start with.

    To have an apparently simple experience such as the perception of a ship, we need to see the ship as a unit, and for that we need the category of substance. That the parts of the ship stay where they are and do not float around chaotically and dissolve, and that I distinguish the ship from my perceiving it, presupposes a priori causality. Thus just to perceive a ship, without even invoking the question of empirical causality by asking whether it is going downstream or upstream, the categories are needed.

    Even deeper, it is I who perceives the ship and I am conscious of this act. It is I who gives it unity and meaning in perception and judgment. Already here I do something that requires a priori concepts. (For accounts of Kant on the I and the soul, the ‘unity of thought argument’ and the ‘inner sense argument’, and on the complexities of various kinds of immaterialism of the soul, see Ameriks 2000, especially pp. 27-47. For a defense of the view that the categories go “all the way out”, see Wenzel 2005.)

    We might be tempted to see even these a priori concepts and their application as being something mental again. After all, Kant thinks of the categories as subjective. But the Kantian subject is not a mind-brain that is causally affected. In the framework of his transcendental epistemology, the subject even comprises time and space, as forms of all appearances. If we think of ourselves as being affected by things “outside of us”, außer uns, then these things are understood, transcendentally, as nothing but simply different and distinct from (logically außer) us and, empirically, as objects that are always already subjected to those subjective conditions of time and space and the categories (spatially außer uns).

    The Kantian transcendental subject is not a mere res cogitans. It is more. It comprises time and space as forms of intuition. Kant holds this against Descartes. A pure science of the res cogitans, the “I think”, would not get us anywhere. No rational knowledge of the outside world, even of ourselves, could be obtained from it. Also no limits of our empirical knowledge could be pointed out in this way, which is something important for Kant, but not for the naturalist today.

    Kant distinguishes his transcendental idealism from what he calls “transcendental realism”, which is the view that time and space are things in themselves, independent of us. Common sense takes this view. If one starts in this way, one can depict oneself as some kind of mind-brain-body at one location and the tree one perceives as being “outside”, ten meters away. Then one can start to give a causal story of sense perception, even look into the brain and try to give a causal account of consciousness and our having representations as well, maybe with the addition of evolutionary and social aspects. Reductionism lives here, and Kripke and Putnam for instance have given accounts of what we mean by “water” and H2O in this way. Transcendental realism starts with a picture of the world that is independent of us, with water as H2O already out there. But if we then place ourselves in this world, how can we be sure that this is how it really is? How can we avoid skepticism? Thus with Putnam we run into a problem similar to Descartes’ doubt. We might be a brain in a vat, nay even the whole world might not exist.

    But according to transcendental idealism, neither my brain, nor the tree, nor time and space are independent of representational conditions. It is only as appearances that they are in time and space, and it is only as being subject to the categories that they are objects. This is an instance of the general view that any third-person account presupposes a first-person perspective. Cassam’s criticism that “Kant’s mistake was to conclude … that the unity of consciousness does not involve being presented to oneself as an object at all” might still be within this view (p. 198).

    Experience requires an act of synthesis, which in turn requires unity. It must be my experience. For the materialist it might be the brain or the object that gives this unity. For Kant it is the act that must provide it. In meaningful perception and in judgment we take something as something and the “taking” itself must have unity (Allison 1996, pp. 95, 102). For Kant it is transcendental consciousness (Reflexion 5661, AA 18, 318-9) and the original synthetic unity of apperception (CpR, B 134) that provide this unity, and they do this a priori, that is, prior to experience. When the materialist points to the brain, our sense organs, and their evolutionary adaptations to their functions and the environment, and the socio-linguist points to our language and society, Kant will point out that they take time and space and empirical objects for granted, as things in themselves, and thereby beg the question. If they also want to naturalize the act of taking something as something, we may respond with Allison that “taken in an investigation of its causal conditions, any token of the act of thinking is itself something represented, an object for an I, which, considered as such, is not itself an object in the world. In short, we return in the end to the ineliminability and systematic elusiveness of this ubiquitous ‘I think’” (Allison, 1996, p. 104). Furthermore, we can add that the materialists will run into the problem of skepticism, because they cannot be sure that the objects, which for them exist independently of us, are correctly represented by us whenever we have representations of them, that is, when they appear to us. In Kant’s words, the transcendental realist then “plays the empirical idealist” (A 369).

    In Kant’s picture the object is nothing but its appearance, and so the correspondence problem does not arise. Truth is in judgment, not in appearance. Ironically, one may also say that in the view of transcendental philosophy appearance already gives truth, a-letheia, as Heidegger wanted it, insofar as appearance and its object are not two separate things (contrary to the transcendental realist’s view). The object does not need to be “deduced” from its appearance (A 372). It exists only as appearance. It is its appearance. Appearance is not something extra.

    Imagine the following conversation between a transcendental realist (TR) and a transcendental idealist (TI):

    • TR: “I think representations are generated in the brain.”
    • TI: “You mean processes happening in the brain? Well, they happen in time and space. You imagine them as appearances.”
    • TR: “But are they not caused? Are not our representations, imaginations, perceptions all caused?”
    • TI: “Well, according to their matter, materialiter, yes. But according to their form, formaliter, no. That I see a hand with five fingers, of a certain size, with a certain color, hue and shade, in a certain light, and under a certain angle, yes, there is a causal story to be told for this. But that the hand appears in time and space at all, and that it has parts, for these facts there is no causal story to be told. You can reduce material properties to their causes, but not the formal ones (for which you need the categories and time and space). Furthermore, it is these formal aspects that make your causal stories possible.


    For helpful discussion and comments I wish to thank Karel Mom.


    1. Allison, Henry E. 1996 “Kant’s Refutation of Materialism”, in Idealism and freedom: Essays on Kant’s theoretical and practical philosophy, Cambridge University Press, pp. 92-106; originally in The Monist 1989.
    2. Ameriks, Karl. 2000 Kant’s Theory of Mind. An Analysis of the Paralogisms of Pure Reason, new edition (first edition 1982), Oxford University Press.
    3. Cassam, Quassim 1997 Self and World, Oxford University Press.
    4. Rosenberg, Jay 1975 “Transcendental Arguments Revisited”, The Journal of Philosophy, vol. 72, no. 18, Oct. 23, pp. 611-624.
    5. Stroud, Barry 1977 “Transcendental Arguments and ‘Epistemological Naturalism’, Philosophical Studies 31; in Barry Stroud Understanding Human Knowledge, Oxford University Press 2000, pp. 71-82.
    6. Thornton, Tim 1998 Wittgenstein on Language and Thought. The Philosophy of Content, Edinburgh University Press.
    7. Wenzel, Christian Helmut 2005 “Spielen nach Kant die Kategorien schon bei der Wahrnehmung eine Rolle? Peter Rohs und John McDowell”, Kant-Studien 96, 4/2005, pp. 407-426.
    Christian Helmut Wenzel. Date: XML TEI markup by WAB (Rune J. Falch, Heinz W. Krüger, Alois Pichler, Deirdre C.P. Smith) 2011-13. Last change 18.12.2013.
    This page is made available under the Creative Commons General Public License "Attribution, Non-Commercial, Share-Alike", version 3.0 (CCPL BY-NC-SA)


    • There are currently no refbacks.