“I” is experience. And this experience does not arise independently of the external world.

The event of “I” occurs as the body’s sensations and consciousness respond inwardly to stimuli that emerge through contact with the external world. This means that there is no purely independent, self-existing “I.”

Yet we link the stream of events and name it a single subject — “I” — separating it from what is not “I.” This is no different from creating a fixed identity within a constant flux. Seeing flowing water, we say, “Since it always flows through this place, let us call it a river.” But the water itself never stays still; every moment it is replaced by new water from upstream.

If we look closely, consciousness consists of momentary fragments of sensation and thought that arise and vanish. There is no inherent reason these fragments must be experienced as one continuous, unified self. It is only because the brain rapidly integrates and corrects these fragments into a seamless flow of experience that we feel a “continuing me.” The brain’s predictions and corrections produce a coherent narrative, which we then mistake for a real entity — the “self.” In truth, continuity does not exist; it is constructed.

At the heart of this construction lies the brain’s predictive mechanism — what cognitive neuroscience calls predictive processing or active inference.

Consciousness is composed of ever-updating snapshots of sensory input and internal state. The brain swiftly links and calibrates these snapshots, integrating them into the illusion of a continuous subject. It constantly predicts sensory input, minimizes the errors between prediction and reality, and does so dozens of times per second. Through this cycle of prediction and correction, we experience the world as an unbroken flow of the present.

What we take to be continuity is not an external reality but a construct generated by the brain’s predictive machinery.

Action follows the same principle. Every step, every movement of a finger — we believe it begins with our will, with “I.” But actions like walking, speaking, or typing are automatized within the systems of the motor cortex, basal ganglia, and cerebellum. These systems adjust behavior through loops of prediction and feedback, allowing the body to move without conscious command.

Hence, sometimes while walking, we suddenly “come to” and realize the body has been walking by itself — the “I who walks” has disappeared, leaving only walking itself.

In such moments, consciousness is not the cause of action but a post hoc model of what has already occurred. We interpret unconscious neural processes after the fact and create the causal illusion that “I did it.” This is the illusion of agency — the most solid form of the self.

This structure of self can be explained through the concept of the self as a predictive model. To predict its own body and environment, the brain constructs a self-representation — an integrated product of sensation, memory, language, and social context. While this model is maintained, we experience the continuity of “I.”

But this is not a concrete reality; it is a computational device for neural stability. The self is not an existent entity, but a fictional structure created for the purpose of self-prediction.

Attention and metacognition constitute higher layers of this structure. Attention selectively amplifies portions of sensory information, increasing the precision of prediction. Metacognition monitors and regulates the thoughts and emotions generated by the self-model. When these two layers interact, we experience reflexive self-awareness: “I am thinking,” “I am observing myself.”

Yet even this is not an enduring substance, but merely the brain predicting its own model once more — a recursive simulation in which the brain refers back to itself.

Seen in this light, the continuity of self is not a physical reality but a product of cognitive stability.

To efficiently process ever-changing sensory input, the brain maintains the hypothesis of a “persistent subject.” This hypothesis — the “I” — serves as a form of cognitive inertia that binds discontinuous experiences into a coherent narrative. We can speak of “I” only because this model has evolved to trust itself.

Anattā (無我) is insight into this very structure. Non-self does not deny the experience of “I,” but reveals that the sense of “I” is a byproduct of prediction and modeling. Through the brain’s constant weaving of discontinuous events via prediction and correction, through automated programs running in the background, and through attention selecting and amplifying portions of that flow, we experience a “continuous self.” But this continuity is not substance — it is a mode of operation.

Thus, anatta is not a declaration of lack, but the recognition of a pattern of cognition that arises and ceases under certain conditions. It is the seeing of “I” as a conditioned process, not an independent entity.

Our continuous experience exists only in relation to the external world, never apart from it. And when the “doer” disappears and only the act remains, that disappearance is not loss but revelation — a moment of lucid observation.

When nothing is grasped, the flow itself is fully revealed. What is revealed then — that ungraspable unfolding — is all that “I” ever was.