The chariot and the bundle: why anattā is not Hume’s theory of the self

In discussion groups at our monastery, and often online, someone will say with real excitement that David Hume discovered what the Buddha taught: that there is no self, only a bundle of perceptions. I understand why. The resemblance is striking, and it has been a door into Buddhist thought for many readers. But I have come to think that the parallel, taken as an identity, hides more than it shows. In this essay I set the two accounts side by side and argue that they differ in method, in content and above all in purpose.

Hume's report

In the section of his Treatise of Human Nature on personal identity, Hume writes that when he enters most intimately into what he calls himself, he always stumbles on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. He can never catch himself without a perception, and can never observe anything but the perception. He concludes that a person is nothing but a bundle or collection of different perceptions, succeeding one another with inconceivable rapidity and in perpetual flux.

Three features of this account matter. It is an introspective report: Hume looked and did not find. It belongs to the philosophy of mind, as an explanation of where our idea of a persisting self comes from (he traces it to the imagination's habit of passing easily along resembling and causally connected perceptions). And it was unstable even for Hume. In the Appendix to the Treatise he confessed that his account of personal identity had left him in a labyrinth, unable to explain what binds the perceptions into one bundle. And at the close of the first book he describes how his sceptical conclusions seemed cold and strained once he had dined and played backgammon with his friends. The bundle theory, like his other sceptical results, did not change how he lived.

The Buddha's teaching

The early teaching of anattā (Sanskrit anātman), not-self, begins elsewhere. In the discourse traditionally counted as the Buddha's second, the Anattalakkhaṇa Sutta, he takes each of the five aggregates in turn: bodily form, feeling, perception, mental formations and consciousness. If bodily form were self, he says, it would not lead to affliction, and one could command it: let my body be thus, let it not be thus. But one cannot. Each aggregate is impermanent; what is impermanent is unsatisfactory (dukkha); and what is impermanent and unsatisfactory is not fit to be regarded as "this is mine, this I am, this is my self". The discourse ends not with a theory but with a change in its listeners: seeing in this way, they grow disenchanted, dispassion arises, and they are released.

Later texts give the teaching memorable images. In the Milindapañha, the monk Nāgasena asks King Milinda whether his chariot is the axle, the wheels, the pole or any other part, or something apart from them all. The king must answer no each time, and concedes that "chariot" is a designation that depends on the parts. Long before that text, a verse attributed to the bhikkhunī Vajirā in the Saṃyutta Nikāya makes the same point: just as the word "chariot" is used when the parts are assembled, so the convention "a being" is used when the aggregates are present. I like to remind readers that one of the classic statements of not-self comes from a woman of the early Saṅgha.

Three differences

Method. Hume's introspection was a philosopher's look inward, made in a study. The Buddhist investigation is a trained practice: sustained, repeated contemplation of the arising and passing away of bodily and mental events, undertaken with concentration and ethical preparation. When a meditator reports not finding a self, the report is the fruit of a discipline with its own standards of care, not a passing glance. One may doubt such reports, but they are not the same kind of evidence.

Content. Hume's bundle is a loose collection whose unity he could not explain. The Buddhist analysis, especially in the Abhidhamma, is precisely an account of how a stream of experience is unified without a self: through causal conditioning, each moment arising in dependence on what went before. Nāgasena's image of a lamp burning through the night, its flame neither the same nor another from watch to watch, captures continuity without identity. Buddhist teachers also keep the conventional person for ethical purposes: actions have results that ripen in the same continuum, and we rightly speak of persons who act and are responsible, while recognising that at the level of ultimate analysis there are only conditioned processes.

It is worth remembering, too, that according to one discourse the Buddha remained silent when the wanderer Vacchagotta asked him directly whether there is a self, and silent again when asked whether there is no self. The teaching is not simply the metaphysical thesis "there is no self", which could be clung to like any other view. It is the instruction to see each thing one might take as self as not-self.

Purpose. This is the decisive difference. Hume drew no practical conclusion. The Buddhist teaching is practical through and through: the conceit "this is my self" is bound up with craving and clinging, and seeing through it is part of the path to the end of suffering. Remove that purpose and you have a different teaching, however similar the argument looks.

A closer neighbour

If any modern Western philosopher stands near the Buddhist view, it may be Derek Parfit. In Reasons and Persons (1984) Parfit argued that personal identity is not what matters; what matters is psychological connectedness and continuity. He noticed the resemblance to Buddhist teaching himself and added an appendix quoting Buddhist texts. And, like Buddhist teachers, he drew a practical conclusion: he wrote that accepting his view made him less concerned about his own death and more concerned about the lives of others. That is far nearer to the spirit of anattā than anything in Hume.

Even here the purposes differ. Parfit was arguing about rationality and ethics, not describing a path to liberation. But the comparison is more instructive because the gap is smaller.

Conclusion

I do not want to discourage anyone from reading Hume and the suttas together. It is a fine way to sharpen one's understanding of both. My point is narrower. The two accounts agree that no unchanging self is found among experiences. They disagree about what that means, how it is discovered and what one should do about it. To call them the same theory is to keep the conclusion and lose the reason for it, and in the Buddhist case the reason is the whole point. When you next meet the parallel, ask the question my own teacher asked me: what is this teaching for?