Who Is the 'I'?: Kant and Hume on the Unity of Self-Consciousness

i.

Hume approaches the question of the self by examining what is present when reflecting inward. Through introspection, Hume finds that only "perceptions" constitute the mind – a term he uses broadly to include sensations, emotions, thoughts, and memories. Hume explains that there is never an enduring "self" that exists upon introspection, observing that only the perceptions exist in his mind. Because Hume holds that anything that is part of the mind must appear in experience in some form, he concludes that there is no substantial self beyond the perceptions that appear in consciousness.

This leads to what is known as Hume's "bundle theory" of the self. Hume argues that the mind is not a single, unified entity but a "bundle" of different perceptions that succeed each other very quickly. For Hume, self-consciousness is simply consciousness of these perceptions themselves, not of an additional subject that exists separate from them.

Although Hume denies the existence of an underlying self, he explains why our perceptions seem connected rather than entirely disjointed. In the Treatise, he argues that perceptions are associated with one another through relations they bear in experience, such as resemblance, contiguity, and causation; and that these relations help explain why perceptions do not appear in complete isolation. Memory also plays a significant role, allowing us to call up past perceptions and connect them with present ones, further reinforcing the sense that they belong together.

These associative connections produce the impression of a continuing sequence of perceptions, even though, according to Hume, there is no single subject that unifies the perceptions themselves. For Hume, self-consciousness is therefore just awareness of the perceptions that happen to occur at any given moment, and our ordinary sense of personal identity arises from the way these perceptions are linked, not from an independently existing self. However, it is important to note that Hume himself acknowledges in the Appendix to the Treatise that his explanation of what binds these perceptions together is "very defective," noting that he cannot fully account for the principle of connection among them.

ii.

Kant agrees with Hume that when we reflect inward, we do not encounter a distinct "self" that appears as an object in experience. For Kant, introspection presents us only with particular "representations" – our thoughts, sensations, and experiences – and never with a distinct subject that exists separate from these representations. By examining one's own consciousness, Kant claims that we find representations themselves, not a separately observable self that these representations explicitly belong to.

Kant also agrees with Hume in that nothing in introspection reveals a simple or enduring substance that could serve as the owner of the representations. Self-consciousness does not involve awareness of a soul or inner substance, but only awareness of the representations that occur within consciousness. In this respect, Kant accepts that inward reflection discloses the contents of consciousness, not a perceivable subject distinct from those contents.

iii.

Although Kant agrees with Hume that introspection reveals only representations and never an observable self, he argues that Hume draws an unwarranted conclusion from this observation. Kant claims that Hume's "bundle" characterization of the mind as a mere succession of perceptions cannot account for the unity our experience exhibits, and therefore, in Kant's view, does not adequately explain what it means to be conscious of representations as "mine."

Kant argues that in ordinary experience, we are aware of many different representations at once, yet, we are conscious of them as belonging to a single point of view. Kant explains that simply having a collection of perceptions is not enough to show that they belong to the same subject compared with merely occurring together. According to Kant, Hume's bundle theory "presupposes the unity of thought but does not have the resources to explain it," since it treats the mind as nothing more than a succession of rapid perceptions. Kant explains that Hume cannot rule out the possibility that these perceptions belong to different subjects, demonstrating that the bundle theory cannot explain why our consciousness has the unity that it does. Kant notes that if each perception were associated with a separate subject, no single subject would be aware of them together, and no unified consciousness would arise.

Kant therefore argues that Hume's account fails to explain the very unity it relies on. Hume's theory assumes that perceptions together constitute a unified sequence of representations, but, according to Kant, it offers no explanation for how such unity is possible. Kant's critique is that without an account of what makes different representations belong to the same subject, Hume's bundle theory cannot explain how we are conscious of representations as "mine" at all.

iv.

Kant explains the possibility of being conscious of oneself as the subject of one's representations by arguing that self-consciousness depends on the combination of these representations into one consciousness. This awareness does not arise from perceiving an inner subject, and instead consists in being aware that the representations are had by the same subject. Kant expresses this through the requirement that "the 'I think' must be able to accompany all my representations," meaning that it must be possible for "me" to recognize "myself" as the subject who thinks of these representations. This awareness does not come from inspecting the contents of consciousness, but from being conscious of the activity by which representations are synthesized into one consciousness – an activity that makes it possible to regard them as "mine", even though the synthesizing subject itself is never perceived.

Kant distinguishes this basic form of self-consciousness, which he calls the analytic unity of apperception, from a deeper condition that makes it possible. The analytic unity consists in the thought that all of my representations belong to me, or, in Kant's terms, that I can attach the "I think" to each of them. Kant argues, however, that this is only possible if the representations have already been combined in one consciousness. This combination does not come from awareness of a substance that owns the representations, but from the activity through which the mind brings them together. Kant calls this underlying condition the synthetic unity of apperception, and he argues that without this prior synthesis, it would not be possible to recognize the representations as belonging to the same subject at all.

Kant emphasizes that this synthesis is not something we perceive, but something we must be conscious of performing whenever we experience representations as belonging together. He explains that recognizing representations as "mine" requires not just that I can attach the "I think" to each of them, but that I am conscious of combining them in a single act of thought. This consciousness of synthesis is what Kant calls the synthetic unity of apperception, and he argues that it is the fundamental condition that makes any self-consciousness possible. Without this unifying activity – which is a condition of experience rather than an object within it – the representations would remain dispersed without any awareness that they belong to a singular subject. In this way, Kant claims that awareness of oneself as the subject of one's representations arises through consciousness of synthesizing them into one experience. This account shows, for Kant, self-consciousness arises not from perceiving a self in experience but from being conscious of the mind's unifying activity of representations.

v.

While I agree with both Hume and Kant that all that exists when looking inward are perceptions or representations, I disagree with Kant's argument that experience is only possible if a single, unified self ("the I think") is already in place to unify these representations together. Kant fails to offer a persuasive explanation of why such a subject must exist prior to experience, treating this unity as a necessary condition rather than showing why it must be true. This assumption that Kant makes is insufficiently supported, overlooking a plausible alternative: that self-consciousness is not an a priori structure built into the mind, but a capacity that gradually develops as cognitive capacities develop. If this is accurate, the unity that Kant takes as given may instead be something that emerges over time with frequent exposure to representations, meaning that this unity could be cultivated through development rather than presupposed from the outset.

Kant's argument depends on the claim that the unity of consciousness must already be in place before experience can occur, but this claim is asserted rather than demonstrated. Kant reasons that because we are conscious of representations as belonging to one subject, there must be a single subject that unifies them. However, this inference assumes that unity in experience requires a pre-existing unifying subject, without considering whether unity might instead develop gradually as cognitive capacities mature. By treating unity as a necessary starting point, rather than something that could emerge through development, Kant does not address the deeper question of how unified consciousness arises at all. This gap leaves his argument vulnerable: the necessity of a transcendental self is claimed, not proven. Kant might reply that any developmental story already presupposes unified experience, and therefore cannot explain the unity of consciousness; but this move simply restates the very necessity claim at issue rather than showing why unity could not itself emerge.

A developmental model of self-consciousness offers a clear alternative that Kant does not consider. Infants do not exhibit full self-consciousness as their capacity to reflect, think of themselves as subjects, and distinguish themselves from the world emerges only gradually. This suggests that the unity Kant describes may itself develop as cognitive capacities mature, rather than existing as an a priori requirement. If self-consciousness can vary in degree across development – and even across adults with severely limited cognitive capacities – then it is natural to think of it as a capacity that can be cultivated, not a fixed transcendental structure that must already be in place whenever experience occurs. With this view, the unity of consciousness may arise from patterns that form among our representations over time: through memory, habits, and an increasing capacity to connect and integrate experiences; rather than from a subject whose unifying role is assumed to be in place from the beginning. Thus, if unity is something that develops in this way, then Kant's insistence that it must be presupposed as a starting point for experience no longer appears compulsory.

Following this alternative theory, Kant's assumption that a unifying subject must already be in place in order for experience to be possible no longer holds as a matter of necessity. The developmental model instead shows that this unity may emerge gradually as cognitive capacities mature. This does not deny that experience exhibits unity; rather, it challenges Kant's claim that such unity must originate from a transcendental subject. By treating unity as a logical requirement rather than a phenomenon that itself needs explaining, Kant appears to presuppose the very point at issue: that unity cannot arise from anything other than an a priori self. If the unity of consciousness can be developed through the gradual formation of connections among representations into one consciousness, then Kant's positing of a transcendental unifier becomes an unnecessary assumption. The burden rests with Kant to demonstrate why the self is innately capable of unity rather than developed, and without such justification, his account of self-consciousness remains incomplete.