Does Kant's theory of knowledge lead to solipsism?
In his Critique of Pure Reason Kant set out to establish a theory of human understanding. His approach was to synthesise the opposing views of empiricism and rationalism. He took the empirical principle that 'all our knowledge begins with experience' [p.1] as a foundation of his philosophy, following Locke and Hume. In contrast to them, however, he also included the rationalist view that posits the existence of an apparatus of human understanding that is prior to experience, and is essential in order that we have experience at all. Thus, for Kant, the human mind does not begin simply as a tabula rasa, as supposed by Locke, but must necessarily have an innate structure in order that we may understand the world.
For Kant, this a priori structure is essential to philosophy. Kant argued that the simple empiricism of Hume and Berkeley inevitably leads to solipsistic idealism. In contrast, by uncovering the a priori structure of human understanding, as the necessary condition for conscious experience, Kant argued that he was able to avoid idealism, since the proof of the existence of an external world follows from this structure.
However, some commentators have pointed out flaws in Kant's theory that demonstrate that he does not necessarily escape the charge of solipsism. As Strawson states: 'Kant, as transcendental idealist, is closer to Berkeley than he acknowledges' [1, p.22]. Russell pointed out that all Kant's immediate successors, Fichte, Hegel, Schelling, were led to develop his philosophy in a subjectivist or idealist direction, and 'fell into something very like solipsism' [2, p.689]. In this essay I shall examine this question, firstly by briefly expounding Kant's defence of his system — that the objects of our perceptions have empirical reality, and then by stating the case put against Kant. Finally, I will outline some revisions that have been proposed that may save Kant's system from solipsism.
Kant makes it clear that all knowledge begins with experience. Specifically, it begins with the phenomena that are presented in our immediate consciousness through our faculty of sensibility. Kant uses the word 'intuition' to indicate our reception of an undetermined object in consciousness, in general ('intuition' being the accepted English translation of the German word 'anschauung' which literally means 'looking at' or 'view' [2, p.681]). But having intuitions is not sufficient if we are to experience our world, for raw phenomena do not give us an understanding of what object is given in the phenomena. In order to understand, we need to have a faculty of mind that can organise and synthesise the raw data given in the phenomena, and thus determine the object given in the phenomena as a whole. The result of this determination is a conception of an object or event. Thus, for Kant, knowledge is only possible when we have an intuition accompanied by a conception. Neither have any meaning on their own. As Kant states, 'It is as necessary for the mind to make its conceptions sensuous (that is, to join to them the object of intuition), as to make its intuition intelligible (that is, to bring them under conceptions) . . . In no other way than from the united operation of both, can knowledge arise' [p. 45].
Thus, when I look at a tree, my phenomenal intuition is the impression of green and brown areas of sensation in my field of vision. My conception of the tree is the synthesis of all these visual stimuli into a categorisation which is of a tree. Thus, without conception, all I would have is a vague visual impression. And without the phenomena, my conception of tree would have no reference point in reality, and would be meaningless.
Thus, we build up an understanding of the world we live in through the conceptions that we have of it. And those conceptions are given meaning by reference to intuitions. Kant argues that our objective world view is built within this framework. The conception of an object, is a conception that holds regardless of the particular phenomenal experience that we may have of it. The phenomena may simply depend on our point of view, whereas the object is independent of our point of view. For example, if I look at a mug of tea, I can look at it from the side and my visual impression is of a rectangle with a loop (i.e. the handle) attached to it; whereas, if I look it from the top I see a brown circle with a white rim; and from various angles, I will have a slightly different phenomenal impression. But nevertheless, behind all these distinct impressions, I know that there is something that is fixed, and that is the mug of tea, the single object of all these experiences. It is from the conception of these objects that we have our sense of an objective world, that is somehow independent of the way we view the world. As explained by Strawson, 'To know something about an object, e.g. that it falls under such-and-such a general concept, is to know something that holds irrespective of the occurrence of any particular state of consciousness, irrespective of the occurrence of any particular experience of awareness of the object as falling under the general conception in question' [1, p.73].
For Kant, the world view we have through our objective conceptions is not simply a model or representation of the world. Our objective world-view, based on phenomena and conception, is the real world. The world of phenomena simply is the world I live in. If I look at a tree, what I immediately see really is the tree. There is not a tree that exists beyond the phenomena; the tree exists through the impression I have of it, and the conception of tree that I derive from the impression. In this sense, Kant deviates from simple empiricism that states that what we actually see is the appearance of the tree, but that the real tree is somehow beyond these mere appearances. Technically, Kant defines reality as objective validity [p.27]. Since our world view is objective in relation to phenomenal impressions, we may describe it as having empirical reality.
Kant's view that our phenomenal understanding constitute reality has two merits: (1) it accords with our common sense experience of living in the world (e.g. when I go to the bakers to buy a loaf of bread, I am not suddenly confronted with an anxious moment of wondering if the bakery is real, or whether the bread is all in my mind: I simply accept that what I see and hear around me is real); (2) it is able to resist the problem of idealism that dogs empiricism; i.e. since empiricism states that we do not have an immediate impression of reality, then how can we be sure that reality is there?
Problems begin to arise in Kant's metaphysics with the introduction of the doctrine of transcendental ideality. Transcendental ideality is an inevitable aspect of Kant's enterprise, which — simply stated — acknowledges the fact that understanding arises within the mind (or more precisely, 'the mode of our cognition' [p.15]), and thus transcends objective reality. Thus, although our world view has empirical reality, we must also acknowledge that it is also based on transcendental ideality. When I have a conception of a mug of tea, I realise that that conception is formed through a process of synthesis, which is part of the apparatus of my mind. The contrasting notions of empirical reality and transcendental ideality are not contradictory; they are one of the key dualities in Kant's theory.
An important element in the doctrine of transcendental ideality is that of things in themselves. For, although Kant argued that our objective world view constitutes reality, he also put forward the idea that there is another world of which our phenomenal impressions are mere appearances of. Thus, when I look at my mug of tea, I can think to myself that there is something beyond the mere sensory impression that I am having, existing quite independently of me. Strawson calls this world of things in themselves, the 'sphere of supersensible reality' (since it is beyond sensory impressions). It has also been referred to as ultimate reality. There are a number of important points to be made about Kant's supersensible reality which I give below.
(1) There is some cross-over between Kant's notion of supersensible reality and the simple empiricist notion of external reality. However, they are quite distinct. Most importantly this is because what the empiricists consider to be external reality — that is, the causes of our perceptions — Kant sees as part of empirical reality. Thus, the empiricist will argue that space is part of external reality, a property of things in themselves, whereas Kant would argue that space is part of our apparatus of understanding the world.
(2) Human understanding only extends to an apprehension of sensory experience, hence we can have no knowledge of things in themselves at all: 'the understanding ... comprehends that it cannot employ its categories for the consideration of things in themselves, because these possess significance only in relation to the unity of intuitions in space and time' [p.163]. The implications of this are quite important: whatever, supersensible reality is we cannot regard it as having structure (at least, as we understand the word 'structure'). When I look at my mug of tea, I cannot think that there is a mug-of-tea-in-itself, creating my visual impression, for that would imply a structure to the supersensible world. Before I know any better, I would be populating the supersensible world with tables-in-themselves, spoons-in-themselves, trees-in-themselves, and so on, until the supersensible appears in my mind as structurally indistinct from my objective world view. For Kant, that will not do, as we are not in a position to say anything about how the supersensible world is. Thus, Kant's supersensible world remains ghostly and enigmatic.
(3) The supposed relationship between things in themselves and phenomena remains uncertain. We cannot say that things in themselves cause phenomena since, for Kant, causality is a category of our understanding and can only be applied to sensory experience. Kant says that things in themselves 'affect' our sensibility. Ultimately, given point (2), we can never know this relationship because it is beyond the possibility of our understanding.
It is from point (2) that the notion of transcendental ideality becomes clear. Kant shows that our phenomenal/conceptual world-view does not refer to the supersensible world, only to the world as constructed through cognition. It is in this that it has transcendental ideality.
Problems with Noumena
Kant uses the word 'noumena' to refer to things in themselves, which he describes as 'intelligible existences' which 'are not objects of our senses' [p.162]. Kant is on difficult territory with his notion of noumena. To begin with, given that we can know nothing about them, in what way does it make any sense to talk about them existing. By Kant's own standard, is it not the case that noumena are devoid of any meaning. Kant argues that although we cannot have any conception of what noumena are, we can nevertheless have a representation of them. But he goes on to state that this representation is a 'limitative conception'. That is, that noumenal representations simply mark the limit of human understanding, and that is their necessary function. Unfortunately, this response does not say a great deal about noumena: it does not give them a positive meaning.
The characteristics that Kant placed on noumena leads ultimately to the question of how we can justify a belief in them. Kant argued that they must exist because phenomena are appearances, thus they must be appearances of something, and that something must be noumena. But surely we are not justified in calling them appearances a priori. We may have a conception of them as appearances empirically, in that we say that such-and-such a phenomena is the appearance of such-and-such an object, but as we have seen the notion of phenomena and of object is accommodated within Kant's model of cognition (as argued by Strawson, pp.40-42), and thus appearance is a relation between them. There is no essential property of phenomena which make them appearances. Ultimately, Kant's noumena are unknowable and, as Russell pointed out, 'the "thing-in-itself" was an awkward element in Kant's philosophy, and was abandoned by his immediate successors' [2, p.689].
The important point here is that in the same way we can be sceptical about the existence of external reality within the empiricist framework, and thus be led down the path of Berkeley's idealism and solipsism, so we can be sceptical about the existence of noumena within the Kantian framework and we can be led to a similarly solipsistic position. Given that we have seen that Kant defends himself against the charge of idealism, it is worth returning back to that argument more critically.
Kant's refutation of idealism is based on the insight that what we take to be the outside world is immediately given. The real world that we live in is immediately revealed by our intuitions, and we make sense of that world through our conceptions. Kant wrote that idealism 'assumed, that the only immediate experience is internal, and that from this we can only infer the existence of external things... But our proof shows that external experience is properly immediate' [p.148]. Hence, because the external world is immediately given, there is no justification in being sceptical about it.
The difficulty with Kant's refutation is not so much with the refutation itself, but with what he is defending. The external world, in the Kantian framework, is ultimately the construct of our faculties of mind (this being the thesis of transcendental ideality). It is built upon the intuitions of sensibility, and the conceptions of the understanding. Given this foundation, it is interesting to speculate what the situation would be if I lacked these faculties. No doubt I would cease to exist as a human consciousness, but would it also follow that the external world would also not exist. It is difficult to see how Kant could avoid this conclusion. Thus, Kant's refutation works so long as I am alive and am able to think about the existence of an external world, yet becomes difficult if I speculate on the consequences of my own non-existence.
Thus it follows that the external world is not objective in the sense that it exists independently of the thinking subject. As has been argued earlier, Kant does describe the external world as objective. However, what is meant by 'objective', for Kant, is that we hold a certain relation with the external world, whereby we conceive ourselves as subjects within an independent framework of objects (i.e. as explained earlier, regardless of how I happen to be looking at a chair, the nature of the chair remains fixed and independent of my perspective). This relational objectivity does not necessary imply an ontological objectivity. Kant can defend his view of the external world against idealism, so long as he frames the charge of idealism at the relational qualities of objectivity.
Kant was aware that the external world was limited by our understanding, and noumena are introduced partly as proof of this (i.e. noumena are the limits of human understanding). It is the noumena that play the role of independent existences beyond the thinking subject. They have ontological objectivity (in contrast, for Kant we do not stand in any relation to them, as the very notion of a relation cannot be applied to them). As such it is noumena that Kant needs to defend against charges of idealism, in addition to empirical reality. As they are 'unknowable', it is difficult to know how this defence can be made.
If noumena are taken away from Kant's theory, or are doubted, then everything I understand, including all objects, places and other people, become mere characters within my active mind. The world can only be appreciated as independently existing within the noumenal theory.
Possible Revisions to the Notion of Noumena
The weakness of the concept of noumena has led to a revision of Kant's framework by post-Kantians. Fitche and Hegel simply removed the problematic things-in-themselves, seeing them as unnecessary. But without any notion of an independently existing world, this seems to plunge them into solipsism, as suggested by Russell [p.689]. There are, however, two approaches to noumena which are more promising. Both of them maintain noumena within the framework, but allow that we can 'know' them in some way: either internally, or externally.
So, firstly, Schopenhauer argued that since we are part of the independently existing world, it implies that we have some noumenal characteristics within us. For Schopenhauer, this internal noumena is identified with the will.
Secondly, Nagel and Strawson disagree with Kant on the very important point that noumena are completely unknowable. Thus, they argue that, although we cannot know the intrinsic nature of noumena, we can nevertheless understand its structure: thus, the structure of the noumenal world would be a reflection of that which we have formed through our conception of the world. Strawson talks of 'rejecting the senseless dogma that our conceptual scheme corresponds at no point with Reality' [1, p.42] and states that in contrast with Kant, 'the scientifically minded philosopher does not deny empirical knowledge of those things, as they are in themselves, which affect us to produce sensible appearances'. Nagel, similarly, argues against Kant that 'our knowledge of the phenomenal world is a partial knowledge of the world as it is in itself' [3, p.101].
Both of these revisions lead to a strengthening of the noumenal concept, because they allow us to have some idea of what they are, and how to find them.
To recap, Kant's theory of the external world and our understanding of it, rests on three foundational concepts:
Intuitions/Phenomena -- the immediate undetermined objects of sensibility.
Conceptions -- with conceptions we organise phenomena, and are able to understand them.
Noumena -- the world of unknowable things in themselves.
The combination of Intuitions and Conceptions constitute our understanding of the real world. The real world is given immediately through them, and as such are immune from the problem of idealism. However, this real world is not sufficient to include the notion of an independently existing reality, and it is this role that noumena fulfil. Unfortunately, by their very unknowable nature, we may be sceptical about noumena.
Thus, Kant's theory of human understanding, based on Intuition and Conception is strong. However, the essential idea of noumena — an independently existing Reality — is weak. The crux of this weakness is Kant's insistence that noumena are simply unknowable. The revisions proposed by Schopenhauer and Nagel break with this insistence, and allow us to have at least a limited 'knowledge' of them, and therefore provide much needed support to the idea of noumena.
All quotes from Kant's Critique of Pure Reason are taken from Meiklejohn's translation. References are given by page number, since this translation does not include A/B numbering. Other references are as follows.
 P.F. Strawson (1966) The Bounds of Sense (Routledge)
 Bertrand Russell (1946) A History of Western Philosophy book 3, chapter XX (Unwin)
 Thomas Nagel 1986 The View from Nowhere (OUP)