What manner of idealist is Kant? Well, a Transcendental idealist. But what does that mean? The best place to get a handle on Kant’s particular flavor of idealism is in his direct engagement with empirical and dogmatic idealism, and how he defines his position contrary to these. This happens when he is discussing the reality of ‘outer objects’ in The Critique of Pure Reason. In the A edition this engagement takes places in “The Fourth Paralogism of Ideality (With Regard to Outer Relations)” (A367). In the B edition this section is removed, and the arguments it contained are then redistributed through “The Refutation of Idealism” (B275), and “The Postulates of Empirical Thought in General” (B266). He then provides further modifications to these arguments in the Preface to the Second Addition, in a footnote at (Bxl). Why such drastic revisions? Kant’s target in the Fourth Paralogism of the A edition is seemingly Descartes. However, through this engagement, Kant’s position was apparently accused of being a dogmatic idealism, of the Berkeleyian sort (Guyer, 2006, p.116). His solution? Delete the section and break the individual arguments down into clearer steps. There are minor differences, but these seem to be in emphasis and structure more than a sign that Kant accepted the criticisms. The alterations seem to be of the kind: “If you think I am arguing for that I must have not expressed my position well enough. Let me slow down and go over the points in a more systematic way”.
Kant’s transcendental idealism, then, can be seen as differentiating itself from two other idealisms: the ‘empirical idealism’ of Descartes (objects of perception are mere mental ideas, representations) and the comprehensive idealism of Berkeley (everything is mental ideas). In this article I hope to explore in what ways Kant achieves this differentiation, through suggesting a schema with which we can draw the distinctions between them. I argue that what Kant’s A edition critics missed in calling his refutation of Cartesian empirical idealism Berkeleyian is Kant’s radical treatment of space. It is this novel approach to space that marks transcendental idealism.
Transcendental. Empirical. Real. Ideal.
We need to follow the distinctions Kant makes in the A edition in order to see how these three positions (Descartes’, Berkeley’s, Kant’s) are distributed. There are two distinctions: one is the distinction idealism/realism (we can call this the metaphysical attitude of the position), the other is the distinction empirical/transcendental (we can call these the necessary domains towards which the metaphysical attitudes are taken). This means that there are two domains, towards each of which we can have two metaphysical attitudes, for a total of four possible positions.
We need to understand these distinctions in order to understand Kant’s claim that “the transcendental idealist … may well be an empirical realist” (A370). It is this claim that embodies all of the features of a true ‘Copernican Revolution’. We can schematise the domains and attitudes as follows:
With the distinction real/ideal Kant seems to have in mind the distinction matter/thought. However, this on its own does not tell us much without a further exposition of what is meant by these terms. For now, I’ll define ‘ideal’ as that which depends upon, or inheres in, the subject of the Cogito, and ‘real’ as that which doesn’t and thus, in a sense, subsists independently. This will become clearer if we attach these attitudes to their domains in the familiar Cartesian manner. Thus, to think that the objects of empirical intuition are ideal is to think of them as inhering in the subject as representations, itself a kind of thinking (Descartes considered empirical ‘sense’ and ‘feeling’ a kind of thinking). To think of the conditions of the possibility of experience as real is to say that time and space, and at least some of the categories (i.e. causality, quantity) exist independently of the subject’s representations, that is, they exist in things-in-themselves. This is why Descartes is an empirical idealist, but on the question of extended substance and its necessary spatial character, as well as the question of quantity that forms through its mode, motion, he is a realist.
Monism
Beyond the ‘dualist’ position of Descartes we can locate two “absolutist”, or monist, positions by reading each row left to right.
The top row, read from left to right, is a full blown materialism that denies that neither conscious perception of the world, nor the world itself (as condition of possibility), is dependent upon the subject of the cogito. This position can be read as arguing that such a subject does not exist. We can actually attribute something very like this position to Dennett, who probably more than anyone gives it its most sustained defence. For example, in Consciousness Explained, Dennett concludes his discussion of qualia with the suggestion that:
“There is another way to address the possibility of zombies, and in some regards I think it is more satisfying. Are zombies possible? They’re not just possible, they’re actual. We’re all zombies. Nobody is conscious — not in the systematically mysterious way that supports such doctrines as epiphenomenalism!” (Dennett, 1991, p.406)
This is to say that the ideal, thought as such, does not even define our immediate perception — there is no ‘class of things called representations’ only ‘modes’ (if we can be forgiven using this scholastic term) of matter itself, or, for Dennett, functional states of an organism. Interestingly, Dennett at times shares this position with Bergson, in Matter and Memory. However, Bergson is attacking the ideal/real distinction itself, rather than prioritizing one over the other, and thus his position is ambivalent to where on the vertical axis it sits.
The bottom row, read from left to right, gives us ‘dogmatic idealism’. All things depend on the subject. In the B edition this view is attributed directly to Berkeley (B274) (recall that the changes in the B edition were to, perhaps, deliberately contrast Kant’s position with Berkeley’s). He finds its motivation stemming from a supposed contradiction in the very idea of matter itself. Downing (2011) gives an excellent overview of Berkeley’s critique of the notion of matter. Beyond the sceptical arguments as to the knowability of matter, Downing points to a section in The Principles of Human Knowledge where Berkeley offers the interaction problem of substance dualism as key inconsistency (PHK 19). Berkeley’s monist motivations, then, are identical to those of Dennett, who also uses the problem of substance interaction to problematize the ‘Cartesian Theatre’ view of consciousness. So, the monist arguments of Berkeley and Dennett are in agreement on one thing — the interaction problem renders dualism inconsistent.
Towards these monist positions in general, Kant offers the shrug argument that in taking everything as one, the vital distinction between appearances and things-in-themselves breaks down:
“But if one wants to broaden the concept of dualism as it is usually applied and take it in the transcendental sense [i.e. as Descartes does], then neither it nor the pneumatism [dogmatic idealism] that is opposed to it on the one side, nor the materialism on the other side, have the least ground, since then one’s concepts would lack determination, and one would take the difference in the mode of representing objects, which are unknown to us as to what they are in themselves, for a difference in these things themselves.” (Kant, CPR, A379)
There are a lot of assumptions in this reply, that objects are ‘represented’ at all chief among them. However, we can read this, in the case of the two monist positions mentioned, as pointing to the common sense incredulity towards collapsing the whole of the world and experience into thought, or matter. It seems Kant is thinking that there is demonstrably a distinction to be made between how things appear to us, and how they are despite us. The mention of the determination of concepts points us to the necessity for our concepts (causality, for example) to be about something such that we can speak of legitimate and illegitimate applications. That is to say, without the distinction between appearance and things-in-themselves, it becomes difficult to account for error. Berkeley uses a theological solution, and Dennett seemingly an evolutionary one. The former requires the Deus ex Machina, which Kant perhaps (silently) sees as a problem, the latter can affirm the reality of causal connections, but has an uphill battle to describe what exactly our ‘understanding’ of these connections consists in, or any ‘understanding’ (phenomenological sense) for that matter.
So, Kant seems to be thinking that overcoming substance interaction by affirming monism is ‘throwing out the baby with the bathwater’. That is, in attempting to avoid the problem, the solution runs roughshod over basic assumptions that are not so easily replaced. There is, after all, an entirely different dualist position to consider.
What is essential to Kant’s entire project is the nature of space and time, not the number of substances. Accordingly, we would do well to examine how the positions Kant rejects are in a sense stemming from various interpretations of space. We will move straight to Berkeley, as it was the distinction between their positions that necessitated so many revisions and explanations from Kant. As for Descartes, Kant’s target in the Fourth Parologism, I’ve written on his position on space (as independent extended substance) last issue.
Berkeleyian Space
Kant begins by diagnosing the source of Berkeley’s position. “Dogmatic idealism is unavoidable if one regards space as a property that is to pertain to things in themselves; for then it, along with everything for which it serves as a condition, is a non-entity” (B274). Kant’s wager here is that if the question of space is resolved by attributing it as an attribute to things-in-themselves (as Descartes’ real extended substance does), then the slide to Berkeleyian dogmatic idealism is “unavoidable”. Why? Because one can always put together a skeptical yet consistent account that instead explains ‘space’ as mere empirical properties of ideal objects. Berkeley offers just such an account.
Berkeley argues in An Essay Towards a New Theory of Vision that because we never perceive distance itself, any basic idea we have of it (a minimal condition for us to attribute it to things in themselves) needs to be derived from what we do perceive; empirical intuitions. His account rests on a kind of Humean habitual association of ideas: we come to understand that confused (blurred) perceptions means ‘close’ and clearer perceptions means ‘far’ when coordinated with variations in tactile sensations (touchable, untouchable). If our ways of seeing were reversed, such that as an object recedes from my face it becomes less clear and more confused, then we would still hold that the object was drawing ‘closer’ to my face (NTV 26). However, over time we would adjust, as people with myopia do to exactly this (NTV 37). Should someone blind be suddenly given sight, they wouldn’t instantly see distance. Everything would be equidistant, and experienced as ‘in the mind’, until they had learnt how the visual associations coordinated with the tactile sensations (NTV 41). Berkeley also held that tactile sensations and visual sensations (and so on) formed distinct objects that were then associated together through habit (NTV 49), this position stemming from the fact that he can’t speak of one ‘object feeling like and looking like such and such’ because that would imply a single ‘real’ object underlying the two registers of sense (Freud advances a very similar view). Once this account has explained where our basic idea of space comes from (as the habitual connection between certain sensory properties), it reveals itself to be sufficient.
Then, in The Principles of Human Knowledge Berkeley’s arguments against the existence of absolute space (of Newton) mirrors that of Leibniz in the Leibniz-Clark correspondence. Paradoxes are provided (what sense does it make to speak of an object completely alone in an infinite empty universe moving?)(PHK, 112) to an appeal to parsimony: space thought of as mere relative relations between bodies provides a fully consistent account without having to appeal to an absolute spatial container existing independently of bodies or empirical intuition (PHK, 114).
Thus, Berkeley believes in the ideality of space because it is simpler and more consistent than believing in its independent reality. Then, the ideality of this space is determined through tactile, aural and visual properties of objects whose entire essence consist in their being perceived in this or that register of sense (or, more accurately, perceivable as determined by a conditional). Thus, there is no more that needs to be said of space than these considerations of the empirical properties of perceived objects.
If space is thought of as an attribute, dogmatic idealism can always claim this attribute for the ideal, by appealing to the myriad attributes whereby the basic idea of space is suggested to us initially. So how does Kant resolve this?
Transcendental Idealism
I’ll start from Kant’s thesis that a transcendental idealist is an empirical realist. In the above graphical schema this is a complete inversion of Descartes position. I’ll examine the transcendentally ideal first.
What’s Kant’s version of having an idealist metaphysical attitude towards the conditions of possibility for the objects of sense (time and space)? Unlike Berkeley, space and time are not ideal properties, but ideal forms of experience in general.
What Descartes discovered when he realized he could not abstract from an object its extension in general, despite being able to abstract its colour, and hardness, and particular shape, and so on, was not a necessary ‘core’ of things in themselves, but the discovery of the formal frame of our experience. The way in which the objects of experience appear to us is as extended (thus ordered in space). Our not being able to abstract this spatiality from them says more about what needs to be the case (the conditions of possibility) of our experiencing them, rather than how they are in themselves.
This may be clearer if we demonstrate the insight in a more limited domain: our visual experience of colour. Now, colour is a ‘dimension’ in the technical sense of that term, as can be evinced by its use to replace a spatial dimension in topographical maps. These maps are obviously only of use to creatures whose have vision (and, to understand a particular coloured topographical map of ours, they would also need a similar structure of vision to ours). Colour is a dimension of vision. In fact, it is arguably the dimension of vision.
When Kant says that such and such is a necessary transcendental form of something, we can read this as saying that such and such is a constitutive dimension of something. If we limit ourselves to vision as such, we realize that the dimension of colour is not something contingently apprehended by creatures with vision, it is the necessary condition of possibility of vision itself. We don’t ‘learn’ to see colours in general by looking at colours, because if we couldn’t see colour in general already then there’d be nothing to look at. Kant uses this exact argument apropos space: we couldn’t gain an understanding of space by experiencing spatial relations, because such an experience presupposes such an understanding (A23|B28).
Now, a dog, a mantis shrimp, and a human, may have their ‘colour spaces’ populated by vastly different diversities of colour, but, insofar as they all possess the ‘faculty of vision’, colour is a necessary dimension around which this vision is structured, even if it’s merely binary-monochromatic. Notice how there being the necessity of colour (given vision) tells us nothing about the particular colours of empirical objects. All we know is that if we can ‘see’ something, then it is registering somewhere in the dimension of ‘colour’, because that what to ‘see’ means, and, indeed, that it is all that colour in general is.
In fact, Berkeley was as close to this transcendental insight concerning colour as Descartes was concerning extension. He argued, against Descartes, that it is not the case that one can abstract colour from extension (Downing, 2011). The notion of an extended thing that did not register along the dimension of colour was inconceivable. When we imagine a ‘pure extended triangle’ we can imagine no such thing without it possessing some colour.
The entire Kantian innovation can be drawn back to this: Descartes and Berkeley both seem to agree that if some property-in-general could not be abstracted from a thing, this thing possessed this property-in-general necessarily and constitutively, it was a part of its fundamental essence. Where they disagreed was just how far the abstractive acid could distil essence (what properties could be abstracted). However, both were beholden to the scholastic-Aristotelian view that properties inhered in substances despite how far they had distanced their positions from other scholastic assumptions. Kant merely takes the final step: the impossibility of abstracting those final few (or single) properties should not be taken as having arrived at the substantial core of the thing, its essence, but, rather, should be taken as those properties being the fundamental dimensions through which the object is given to us in experience. Space and time are not properties, relations, essences, or ideas, they are the conditions our cognition places on the possibility of experience itself. And just like in the example of colour where a dog and human may disagree about a particular case of colour, but must, necessarily, both be seeing in terms of colour in general to see anything at all, likewise in the case of any experiencing creature, it must be necessarily experiencing in terms of time and space. Two vastly different creatures could disagree about particular spatial-temporal cases, but if they are experiencing anything at all, they are experiencing things spatially and temporally.
Scaling up from colour and vision, to time and space and experience in general, is quite mentally taxing. An example may be useful. Imagine pitting two chess playing AI’s against each other. If we are not interested in observing the progress of the game, there is no need to represent the board and its pieces. If the two AI’s shared their information perfectly, and the computer was sufficiently powerful, we could get the results of a thousand games (with a record of all moves ‘played’) at the press of a button. The amalgamated AI (it would in fact be singular and just playing itself), in not having to ‘represent’ the game space to itself (including not representing two sides as ‘opponents’, by sharing all information), can calculate n number of matches purely logically. The only time requirement, then, is the requirement of the hardware we as experiencing beings have constructed. A God, given the source code, could calculate the outcomes of an infinite number of matches instantaneously. If the game need not be ‘experienced’ by some being like us, there is no need for spatial-temporal limitations, considerations, or instantiation.
So, just like how it makes no sense to talk about what colour things are in themselves, beyond the domain of creatures with vision, it also makes no sense to talk about where and when things-in-themselves are, beyond the domain of experiencing things that represent them.
Empirical Realism
So far these considerations could seem to be merely shoring up dogmatic idealism. The second half of Kant’s position remains to be shown — how one can be an empirical realist when the conditions of possibility of empirical experience are themselves ideal.
Kant’s solution begins with a terminological distinction: inner and outer sense. This is an empirical distinction (signified by the ‘sense’), not a transcendental one. So, all it amounts to saying is that of all of the things we are conscious of, that we have representations of, some of them appear to be sensed within us, and others are sensed as being ‘out there’. ‘Out there’ should not be taken to mean “beyond our representations”, but much more simply, as being represented as though they are ‘over there’, ‘on the other side of my eyelids’, and so on. No transcendental claim is being made, just a taxonomy of different kinds of ‘internal’ representations.
Now, what is it that allows our representations to be ‘tagged’ or ‘tokened’ as being inner or outer? That is, what distinguishes this seeming towards our myriad of representations? Well, those that are represented through the form of space (itself ideal) are tagged ‘outer’ and those represented wholly through the form of time (also ideal) are tagged ‘inner’. All of this, so far, is a play of terms. There is some empirical import in defining things this way, (it predicts that what people common refer to as uncontroversial examples of ‘inner’ mental states would be lacking in spatial determination, but have duration) but not much is being asserted metaphysically. Kant is not claiming that transcendentally ‘outer’ objects (objects beyond our representations) are spatial (as Descartes believed), merely we are prone to think of objects we represent as being ‘outside of us’ if they are represented in spatial terms.
From here, Kant then appeals to his earlier re-definition of matter from the opening paragraphs of the CPR:
“The effect of an object on the capacity for representation, insofar as we are affected by it, is sensation. That intuition which is related to the object through sensation is called empirical. The undetermined object of an empirical intuition is called appearance…
I call that in the appearance which corresponds to sensation its matter…” (Kant, CPR, A20|B34)
We need to follow all of these connections through. There are sensations, these are the name for what in our representations we receive completely passively. A representation that possesses some degree of sensation (this passive reception) is called an empirical representation (empirical intuition). These ‘empiricalized’ representations in general are called ‘appearance’. Now, in all of this appearance in general are traces of sensations (by above definitions), these traces are called “matter”.
Collins (1999) even argues that these ‘grains’ of matter that mark our empirical representations are direct cognitive contact with some transcendent object.
“he [Kant] is not saying that when we perceive, we are correctly inferring the existence of an object corresponding to an apprehended inner representation, … That is the insecure causal argument that Kant rejects. For him, when we perceive, the element of sensation is immediate cognitive contact with the outer object and cannot exist without the outer object” (p.77)
This is perhaps most clearly grasped when we think of touch. When I ‘touch’ an object I bring it into immediate contact with my whole biological sensory apparatus. The degree of sensation is in proportion to actual physical contact, mass, resistance, etc. There’s a final jump perhaps, where this contact goes from a physical collision to a ‘feels like’, becomes coated in a representation. However, at the core of that representation is the physical resistance of real objects, translated perhaps, but so minimally it becomes almost impossible for us to describe “resistance” in terms that don’t reference the sensation at all.
Matter is sensation. This is not just an arbitrary stipulation, or a sceptical confession that we cannot go from the existence of sensation to the existence of ‘external’ matter. Kant means it quite literally. If space is an ideal formal frame of experience, ‘external objects’ just become those objects of ‘outer sense’ who are defined by being composed of sensation that is organized within the ideal frame of space. Thus, the spatially organized objects of empirical intuition, as representations marked and composed of sensation, are as real as real can be. They are, by the definition of ‘sensation’ (given completely passively) independent of the cogito, who must undergo them, but they are played out in and experienced through an ideal representation of space. In fact, it is because the transcendental frame is ideal, is a representation, that the reality of empirical objects can be conceived and grasped.
“Every outer perception therefore immediately proves something real in space, or rather is itself the real; to that extent, empirical realism is beyond doubt, i.e., to our outer intuitions there corresponds something real in space.” (Kant, CPR, A375)
“The real in outer appearances is thus actual only in perception, and cannot be actual in any other way.” (A376)
Kant seemed aware of how counter-intuitive this Copernican Turn is, writing in a footnote:
“One must note well this paradoxical but correct proposition, that nothing is in space except what is represented in it. For space itself is nothing other than representation; consequently what is in it must be contained in representation, and nothing at all is in space except insofar as it is really represented in it. A proposition which must of course sound peculiar is that a thing can exist only in the representation of it; but it loses its offensive character here, because the things with which we have to do are not things in themselves but only appearances, i.e., representations.” (A374–375)
What Kant is ‘playing with’ (insofar as this seems like a ‘trick’) is the content of our epistemological demands. What is it we are asking to know when we ask to know what really exists beneath any and all possible representations of a given object encountered in space? Presumably this question makes demands of the nature of the thing-in-itself. However, if we have followed Kant’s arguments in the Transcendental Aesthetic through, and thus grant the ideality of space and time as forms of experience, then it becomes unclear as to what this demand requires for its satisfaction. It is not just that the demand is unreasonable because we are unable to provide an answer (which is common way of reading Kant’s admonitions about the unknowability of the things-in-themselves), its that the demand cannot be satisfied by any solution. It is, in effect, senseless. Bennett (2016) makes a similar point about Kant’s position elegantly:
“If someone says seriously ‘There is no moon’, we may think he has some fancy about a permanent visual hallucination or a trick by the Martians … Suppose, however, that after walking on the moon he says: ‘I grant that the evidence so far — and perhaps all the evidence there could be — supports the lunar hypothesis; but there is no moon.’ We can no longer regard him as fanciful or sceptical, for now we do not understand what he says.” (p.24)
Or, take for example someone who asks what this particular pen is. We then list off every property the pen appears to possess, including its spatial dimensions, coordinates, and its temporal history, the molecular constitution of its plastic and ink, the names of the truck drivers who drove it from factory to warehouse and from warehouse to newsagent, and so on. The person then replies “But all of that is just how it appears in empirical representations to those beings for which things appear. But, what is it really in itself?” Kant’s reply is that this epistemological demand (which is the demand of scepticism) is entirely senseless and contradictory. To be real means to appear in empirical representations to those beings for which things appear. If you want to know what it (the pen) is really, well, there it is, take a look, it’s that thing there. That pen is not really something else, it is just that pen.
· · ·
Resolution
We are now in a position to distinguish Kant’s transcendental idealism from Berkeleyian dogmatic idealism, and Cartesian empirical idealism. Berkeley believed that space was derived from recurring properties in empirical (marked by sensation) representations. What is fundamental, then, is sensation. Esse est Percipi. These spaceless sensations then required theological appeals and layers of sceptical ad hoc hypotheses to account for the ‘worldhood of the world’. To this, Kant sides with Descartes: space is not derived, it is absolutely fundamental. It underpins the givenness of sensations, and the objects they mark. However, Descartes’ belief in the full independent reality of space, extended substance, with his belief that empirical representation was ideal in an epistemologically derogatory sense (but for the I of the cogito) has issues. This situation cannot appease scepticism without similar theological appeals, themselves suspect. That is, the existence of the extended substance, given that ‘we’ only have direct experience of the empirical, can always be doubted.
By inverting Descartes’ position, Kant offers a solution to this deadlock: if space is ideal, yet fundamental (because it is a necessary form, not a property), objects of ‘outer sense’ that are determined by it, and constructed by ‘atoms’ of sensation, can be said to be as real as anything in their mere possibility of appearing to beings who intuit things spatially and temporally.
Works Cited
Bennett, J. (2016). Kant’s Analytic. Cambridge: Cambridge University Press.
Berkeley, G. and Turbayne, C. (1989). Principles, dialogues, and philosophical correspondence. New York: Macmillan.
Berkeley, G. and Wilkins, D. (2018). An Essay towards a New Theory of Vision. [online] David Wilkins, Trinity College, Dublin. Available at: https://www.maths.tcd.ie/~dwilkins/Berkeley/Vision/1732B/Vision.html
Collins, A. (1999). Possible Experience: Understanding Kant’s Critique of Pure Reason. Los Angeles: University of California Press.
Dennett, D. (1991). Consciousness explained. 1st ed. New York: Back Bay Books, Little Brown.
Descartes, R., Haldane, E., Ross, G. and Chávez-Arvizo, E. (1997). Key philosophical writings. Wordsworth.
Downing, L. (2011). George Berkeley. [online] Plato.stanford.edu. Available at: https://plato.stanford.edu/entries/berkeley/
Guyer, P., (2006). Kant. New York: Routledge.
Kant, I., Guyer, P. and Wood, A. (2000). Critique of pure reason. Cambridge: Cambridge University Press.
Kant, I., & Weigelt, M. (2007). Critique of Pure Reason. London: Penguin Books.
Skirry, J. (2018). Descartes, Rene: Mind-Body Distinction | Internet Encyclopedia of Philosophy. [online] Iep.utm.edu. Available at: http://www.iep.utm.edu/descmind/