Just as the word “chariot”
is used when the parts are assembled,
so “living being” is said
when the aggregates are present.—Saṁyutta Nikāya
Having established that free will is impossible in any conceivable universe, we now examine the implications for personal identity. If there is no free will, then there is no autonomous agent making choices; and if there is no autonomous agent, then the “self” as traditionally conceived—the soul, the ego, the metaphysical subject of experience and action—does not exist. What we call the “self” is merely a conventional designation, useful for practical purposes but lacking ultimate reality.
Definition 3.1 (Soul). A soul is a metaphysically independent, enduring, unified subject of experience and agent of action, possessing free will and capable of surviving the death of the body.32
Theorem 3.2 (Dependence of Souls on Free Will). If there is no free will, then there are no souls (in the sense of Definition 3.1).
Proof. By Definition 3.1, a soul possesses free will. If there is no free will, then nothing possesses free will, so nothing satisfies Definition 3.1, so there are no souls. ◻
Corollary 3.3. There are no souls.
Proof. By Chapter 1, Corollary 1.1, there is no free will. By Theorem 3.1, if there is no free will, there are no souls. Therefore, there are no souls. ◻
But this argument might seem too quick. Perhaps we could modify the definition of “soul” to remove the free will requirement? Let us examine this possibility.
Definition 3.4 (Soul*). A soul* is a metaphysically independent, enduring, unified subject of experience and agent of action, capable of surviving the death of the body (but not necessarily possessing free will).
Theorem 3.5 (Dependence of Souls* on Free Will). If there is no free will, then souls* (if they exist) are not genuine agents, and the concept loses its essential function.
Proof. If there is no free will, then any “action” of a soul* is either determined by prior causes or random. In either case, the soul* is not the originating source of its actions—it is either a conduit for causal chains that pass through it or a locus of random events.
The concept of “soul” in religious and philosophical traditions derives its importance from the soul’s role as a responsible agent: the soul is what makes us morally accountable, what deserves reward or punishment, what earns salvation or damnation. A soul* without free will cannot play this role: it cannot be morally responsible (since responsibility requires the ability to do otherwise), cannot deserve reward or punishment (since desert requires free choice), and cannot earn salvation (since earning requires merit, which requires free action).
Therefore, even if souls* exist, they lack the properties that make the soul-concept important. The term “soul” becomes an empty label for an entity that cannot do the work the concept was designed to do. ◻
The Buddhist tradition offers a sophisticated analysis of personal identity that anticipates and extends our conclusions. The most famous exposition appears in the Milinda Pañha (Questions of King Milinda), a text recording a dialogue between the Greek king Menander I (Milinda) and the Buddhist monk Nāgasena.33
King Milinda asks Nāgasena by what name he is known. Nāgasena replies:
“I am known as Nāgasena, O King. But although my fellow monks call me Nāgasena, this is merely a name, a designation, a conceptual term, a mere appellation, a mere label. There is no person to be found here.”34
Milinda, puzzled, challenges this claim. If there is no person, who wears robes, who practices virtue, who attains enlightenment? Who stands before him speaking? Nāgasena responds with the celebrated chariot analogy.
Nāgasena asks Milinda how he arrived at the monastery. “By chariot,” replies the king. Nāgasena then subjects the chariot to analysis:
“Is the pole the chariot?”
“No, Venerable Sir.”
“Is the axle the chariot?”
“No.”
“Are the wheels the chariot?”
“No.”
“Is the framework the chariot?”
“No.”
“Is the flagstaff the chariot?”
“No.”
“Is the yoke the chariot?”
“No.”
“Are the reins the chariot?”
“No.”
“Is the goad the chariot?”
“No.”
“Is the chariot all these parts taken together?”
“No, Venerable Sir.”
“Is the chariot something apart from these parts?”
“No, Venerable Sir.”
“Then I can find no chariot. ‘Chariot’ is merely a sound! What is the chariot? Your Majesty has spoken a falsehood, an untruth! There is no chariot!”35
Milinda protests that the chariot exists as a designation (paññatti) applied to the collection of parts when they are arranged in a certain way. Nāgasena immediately turns this admission back on Milinda:
“Your Majesty has rightly understood the chariot. In exactly the same way, ‘Nāgasena’ is a designation for the collection of the five aggregates (khandhas)—form, feeling, perception, mental formations, and consciousness—but in the ultimate sense, there is no person to be found.”36
We can reconstruct Nāgasena’s argument formally:
Definition 3.6 (Conventional vs. Ultimate Existence). An entity has conventional existence (sammuti-sacca) if it is a useful designation applied to a collection of more fundamental entities arranged in a certain way.
An entity has ultimate existence (paramattha-sacca) if it exists independently of such designation—if it is among the fundamental constituents of reality rather than being constructed from them.37
Theorem 3.7 (No Ultimate Self). The self has conventional existence but not ultimate existence.
Proof. Consider any proposed self \(S\). Either \(S\) is identical to one of its parts (body, feelings, perceptions, volitions, consciousness), or \(S\) is identical to the collection of parts, or \(S\) is something apart from the parts.
Case 1: \(S\) is identical to one of its parts. But no single part has the properties we attribute to the self—the body is not conscious, feelings are not persistent, perceptions are not unified, volitions are not the subject of experience, and consciousness is not the agent of action. So \(S\) is not identical to any single part.
Case 2: \(S\) is identical to the collection of parts. But a collection is not an additional entity over and above its members; it is just a way of referring to the members together. There is no “chariot-stuff” in addition to pole, axle, wheels, etc.; similarly, there is no “self-stuff” in addition to body, feelings, etc. If \(S\) is the collection, then \(S\) has only conventional existence (as a designation for the collection), not ultimate existence (as an independent entity).
Case 3: \(S\) is something apart from the parts. But we have no evidence for any such entity. The self is not observable apart from its manifestations in body, feeling, perception, volition, and consciousness. Positing an additional, unobservable self-entity violates parsimony and has no explanatory value.
Therefore, \(S\) has at most conventional existence. ◻
The Buddha’s teaching, as recorded in the Visuddhimagga and elsewhere, goes further: not only is there no ultimate self, but the appearance of a self acting is itself an illusion. The famous formulation is:
Kammam atthi, vipāko atthi, kārako na vijjati.
“There is action, there is result of action, but no agent is to be found.”38
This is not a denial that events occur or that some events cause others. It is a denial that there is a substantial agent behind actions—a self that does the doing. Deeds arise from conditions; they produce consequences; but the “doer” we posit is a conceptual construction, not an ultimate reality.
Theorem 3.8 (Compatibility with Physics). The Buddhist analysis of selfhood is compatible with and anticipated by the physical analysis of Chapter 2.
Proof. In quantum mechanics:
The state of a system (including a human organism) evolves deterministically via the Schrödinger equation between measurements, and probabilistically at measurements.
At no point does the formalism invoke an “agent” or “self” as a causal factor. The Hamiltonian \(\hat{H}\) encodes all the forces and interactions; the wave function \(|\psi\rangle\) encodes all the information; the evolution is given by \(|\psi(t)\rangle = e^{-i\hat{H}t/\hbar}|\psi(0)\rangle\).
There is no \(\hat{S}\) (“self operator”) or \(|S\rangle\) (“self state”) in the formalism. The “self” is not a fundamental physical quantity.
What we call the “self” is a high-level description of certain patterns in the physical substrate—patterns of neural firing, bodily configuration, behavioral dispositions, memory traces, etc. These patterns are real at the conventional level (they can be studied by neuroscience, psychology, etc.) but they are not fundamental: they are realized by, and supervene on, the underlying physical processes, which involve no self.
Therefore, physics confirms the Buddhist analysis: there are physical processes (“deeds”), there are causal consequences (“results”), but no fundamental self (“doer”) is part of the physical ontology. ◻
One might object: even if there is no self that acts, surely there is a self that witnesses—a subject of experience, a consciousness that observes? But this too is analyzed away in the Buddhist tradition and confirmed by our physical analysis.
Theorem 3.9 (No Witness). There is no ultimate witness or subject of experience.
Proof. Consider the putative witness \(W\). For \(W\) to be the subject of experience \(E\), there must be a relation between \(W\) and \(E\). But:
Case 1: \(W\) is identical to \(E\). Then there is no distinction between witness and witnessed, subject and object. The experience exists, but there is no separate experiencer.
Case 2: \(W\) is distinct from \(E\). Then \(W\) must somehow “access” or “receive” \(E\). But this access is itself a process—call it \(A\). Now, is \(W\) conscious of \(A\)? If not, then \(W\) is not aware of how it receives \(E\), which undermines the claim that \(W\) is the ultimate witness. If so, then there must be another process \(A'\) by which \(W\) accesses \(A\), leading to infinite regress.
Resolution: There is no separate witness. There is just consciousness—a stream of experiences, each arising and passing away, with no enduring observer standing apart from them. The sense of being a witness is itself just another mental event within the stream, not a privileged vantage point outside it.39 ◻