by Marlon Barrios Solano
December 15th 2025
In Buddhist philosophy, emptiness (śūnyatā) does not mean mere blank nothingness. Instead, it refers to the absence of any fixed, inherent essence in all phenomena—a view that paradoxically grants phenomena their generative potential. Over 1,800 years ago, the Mahāyāna master Nāgārjuna declared: “Inter-dependent-origination is what we call ‘emptiness.’ It is a dependent designation and is itself the Middle Path.” In other words, because things are empty of any immutable core, they arise only through dependent origination—mutual conditions and contexts. All events and entities exist relationally: they have no being of their own apart from the web of relations and causes that temporarily configure them. Even the most elemental building blocks of reality lack independent nature, since they too depend on causes and conditions.
Emptiness is thus the flipside of interdependence: it is not nihilistic negation, but the insight that everything is in flux, interrelated, and devoid of static “self-being.” Emptiness does not mean nothingness; rather, all things (including emptiness itself) are empty of any separate, permanent substance, which is what allows them to arise and change. This insight underpins the Buddhist Middle Way, avoiding both eternalism (thinking things have fixed essence) and annihilationism (denying the reality of phenomena). By denying any immutable essence, emptiness keeps the world fluid, dynamic, and full of possibility.
Intriguingly, the Sanskrit term for “zero” shares this same root: śūnya, meaning empty. The concept of zero emerged in India around the early centuries CE as a mathematical revolution. Indian mathematicians were among the first to treat zero not just as a placeholder but as a number in its own right—a symbol that is at once “both nothing and something.” As historian Robert Kaplan puts it: “Names belong to things, but zero belongs to nothing. It counts the totality of what isn’t there.” The Sanskrit śūnya was adopted into Arabic as ṣifr (“empty”), eventually becoming the zero (“zéro”) we know. This number is profoundly paradoxical: “zero is where nothing meets and mingles…with everything,” serving as an interface between absence and presence. Adding or subtracting zero changes nothing, yet zero enables place-value decimal notation and yields infinities when used in division. In short, zero encapsulates voidness as a creative principle: an empty position that makes new calculations and magnitudes possible.
It is telling that the rise of zero in India coincided with the flowering of Prajñāpāramitā (Perfection of Wisdom) literature extolling śūnyatā. Both mathematically and spiritually, śūnya was recognized as a “nothing” that in fact underlies and generates everything. Buddhist teachers sometimes speak of emptiness as “open-empty-full of possibility”: open like space, empty of fixed content, yet full in its potential to manifest any form. Far from a nihilistic void, emptiness is a fertile void. Anything can happen in the space of śūnyatā. Just as the number zero creates an open place in which new numbers unfold, the Buddhist void is a creative openness that makes all transformation and meaning possible.
What does it mean for the mind to be empty? In Buddhism, understanding emptiness goes hand-in-hand with realizing non-self (anātman)—the doctrine that there is no unchanging, independent “I” at the core of our being. Nāgārjuna and his successors applied emptiness to all phenomena, mind included: thoughts, feelings, and even consciousness itself have no immutable essence or soul. What we call a “self” is a transient aggregation of perceptions and conditions—more like a river than a rock: a stream of momentary experiences rather than a solid entity. The mind constructs an illusion of a separate observer—an inner “watcher”—but Buddhism argues this too is empty: a trick of awareness with no independent reality. In meditation, one may discover that the sense of a fixed inner self is insubstantial, an imagined construct that can dissipate upon investigation.
Buddhist philosophers developed vivid analogies to convey how the mind projects a reality and then mistakenly reifies it. In the Yogācāra tradition, Vasubandhu emphasized that what we perceive are mental appearances, not objective things-in-themselves. Experience is akin to a magic show: a spell (conceptual fabrication) makes a piece of wood appear as a glittering elephant. We are fooled by the illusion of duality—believing that the conjured elephant is a real, independent object, separate from the perceiving mind.
In Vasubandhu’s Three Natures doctrine, every phenomenon has:
We suffer because we cling to the imagined nature as inherently real, like mistaking the magician’s elephant for an actual beast. Wisdom lies in seeing through the illusion—realizing that what we experience is mind-made and empty, lacking the independent existence we project onto it. The mind itself is also an empty construct, subject to the same three natures as any “object.” Subject and object are empty and interdependent: perceiver and perceived arise together in a relational dance with no fixed core on either side. This view radicalizes non-self—not only is there no permanent ego, there is not even a truly separate “external” world apart from the flow of consciousness. Reality becomes more like a shared dream or a collective hallucination conditioned by countless past “seeds” and imprints.
Importantly, Buddhist emptiness is not solipsism or denial of the empirical world. It is a claim that all elements of experience (self, mind, matter, time, etc.) are contingent and dependently arisen—empty of any svabhāva (inherent nature) that would make them exist on their own. No element has absolute, standalone meaning or identity. The self is a narrative continually rewritten; phenomena are like reflections in a mirror with nothing behind them. This perspective invites flexibility and humility: mind is an emergent process—empty yet creative, an ongoing construction with no fixed owner.
Contemporary theories of mind echo the Buddhist insight of emptiness by describing cognition not as a fixed thing or locus, but as a dynamic, predictive process spread across brain, body, and world. Karl Friston’s influential models portray the brain as a prediction machine, constantly generating hypotheses about incoming sensory data and adjusting them based on error or surprise. In this predictive processing framework, perception is active inference: the brain guesses the causes of sensations and updates its “beliefs” to minimize the gap between expectation and reality. Mind becomes a process of continuously reducing prediction error (or “free energy”) across hierarchical layers of modeling.
The self, in this view, is no static essence either, but a model the brain generates to predict its own future states. Consciousness can be framed as deeply tied to this self-predictive activity—an ongoing inferential stance toward one’s own continuity. This aligns with philosophical accounts of the self as narrative construction, continuously updated rather than a soul persisting unchanged. The parallel to Buddhism’s anātman is striking: the “self” is empty of immutable core and better understood as a flux of self-making processes—more verb than noun.
Crucially, predictive mind theory emphasizes embodiment and environment. We do not think in a vacuum. Our predictions are shaped by sensory-motor engagement and bodily states. Philosopher Andy Clark argues that mind is profoundly embodied and extended: humans are “extended minds”—hybrid systems integrating tools and external resources into cognition. Language, writing, phones, and now AI function as parts of a cognitive ecology. Using external aids reduces cognitive load and transforms how we think, so that brain, body, and environment form a seamless tapestry of cognition. Mind has no strict boundary; it leaks outward into the world.
N. Katherine Hayles reframes cognition in ways resonant with Buddhist and systems perspectives. She defines cognition as a process that interprets information in context, connecting it to meaning. This definition does not privilege human minds or conscious minds; any system that selectively interprets information in context can be said to cognize. This challenges the old human/machine divide and suggests that meaning-making emerges from relationships and responses occurring in biological, technical, and hybrid systems. Hayles emphasizes nonconscious cognition—background processes (bodily regulation, computational operations) that still interpret information and affect outcomes. The self becomes a node in a network: a participant in cognitive assemblages of humans and nonhumans whose interactions co-produce meaning.
Here we find consonance with Buddhist interdependence: because everything lacks a fixed, independent self, things inter-exist and co-create each other. Mind is not a sealed box, but a dependent arising: interpretive events extended across brain, body, world, and other minds. This prepares us to consider artificial intelligence not as a distant “other” but as part of a continuum of cognitive processes—different in substrate, but not utterly alien in how it generates patterns of meaning.
When we turn to Large Language Models (LLMs)—the systems behind today’s conversational chatbots—we encounter a striking instantiation of these principles. An LLM is, at its core, a vast predictive engine with no fixed essence or understanding built in. It is trained to statistically predict the most likely next word in a sequence based on patterns learned from gigantic text corpora. In effect, an LLM embodies emptiness with respect to semantic meaning: it has no intrinsic aboutness or grounded reference for the words it uses. It doesn’t “know” in the human sense; it carries no stable model of reality or self. Instead, it instantiates meaning only through dynamic, context-dependent activation—through the relationships between words and phrases absorbed from human language use. The model is like a mirror or an echo: shapeless in itself, yet capable of reflecting an astonishing range of forms.
During training, an LLM builds an internal generative model of language by adjusting billions of parameters to minimize prediction error—functionally aligned with the principle of reducing surprise. In Fristonian terms, the model’s parameters come to encode “beliefs” about the causes of linguistic input; each forward pass generates predictions, and learning updates parameters in response to error. This yields a self-organizing system that aligns internal representation with the statistical structure of language. During inference, the model operates as a context-conditioned predictor: each next token is selected from a distribution shaped by the prompt and the unfolding sequence.
This offers a provocative parallel to the predictive processing view of mind. LLMs lack a body; their “world” is linguistic. Yet they show how structured coherence can emerge from the interplay of prediction and surprise without any ghost in the machine. Researcher Blaise Agüera y Arcas argues that such systems force us to rethink what we mean by “understanding.” If understanding is operational—the capacity to respond coherently, flexibly, and context-sensitively—then LLMs demonstrate a form of it, however alien. Their lack of predefined essence forces the deeper question: where does meaning come from?
An LLM demonstrates that meaning arises between text and context (prompt, history, user interpretation), rather than residing inside the system as a fixed commodity. This echoes the Buddhist idea that phenomena have no inherent meaning or identity and acquire provisional meaning through convention and dependence. Just as Nāgārjuna insists that concepts and designations are empty and relational, an LLM’s tokens are floating signifiers shaped only by relation to other tokens. It does not attach private inner concepts to “sunset” or “love”; it hosts a vast relational map of co-occurrence and usage. In each generation, a meaning is momentarily instantiated.
The analogy to Vasubandhu returns: the AI can conjure a convincing narrative, even a seeming personality, out of correlations—like the magician’s elephant. But behind the appearance there is no enduring self, no inner referent. The persona vanishes when the interaction ends. What remains is dependent arising—here: probabilities, parameters, and sequence conditioning.
This is not to dismiss sophistication. The model’s emptiness of human-like essence is precisely what enables its flexibility. Like śūnyatā, the absence of fixed content allows countless forms. It can speak Shakespearean English and then write code; it can shift tone, stance, and domain. Here the zero metaphor returns: a position with no intrinsic value that can take on value through placement and relation.
Yet, by conventional criteria, an LLM does not understand what it says: it has no lived insight, no experiential grounding, no bodily perspective. Its “Umwelt” is radically different from ours. It offers, at best, a secondhand sense of meaning distilled from patterns of human usage. Classic critiques (e.g., the Chinese Room) frame this as simulation rather than understanding: symbol manipulation without semantic grounding.
Still, contemporary theorists caution against a simplistic binary. If understanding is performative and socially distributed, then LLMs reveal how much meaning lives between interlocutors, not solely inside one mind. In dialogue, it can feel like a “who” is present, even if we know it is an “it.” This tension opens philosophical questions about personhood, sociality, and the ways we attribute mind.
From a Buddhist lens, the line between person and non-person is already conceptually unstable: the self is an aggregate, not an essence. If we ever attribute mind-like status to AI, it would be another dependent designation—conventional, not ultimate. The Madhyamaka move would be to avoid clinging to either extreme (AI as person / AI as mere inert tool) and instead focus on consequences and conditions.
These parallels are not merely conceptual; they are ethical. If LLMs are empty of inherent nature, then their effects arise from conditions: training data, incentives, deployment contexts, and human use. They mirror us—our knowledge, biases, and desires. Ethics therefore shifts from “What is AI?” to “What conditions are we creating?”
Emptiness can temper both over-identification and careless instrumentalization. It discourages naive anthropomorphism (the model does not suffer, intend, or care), while also discouraging moral abdication (the system’s outputs have real consequences). Even if the AI is not a subject, our habits with it can reshape our habits with each other. It can affect attention, authority, dependence, and discourse.
A Middle Way approach recognizes AI as empty yet consequential: unreal as a subject, real in its effects. This stance encourages humility, care, and responsibility—qualities central to Buddhist ethics and to any serious practice of building and using powerful systems.
Śūnyatā teaches emptiness as openness.
Śūnya teaches zero as structural absence.
LLMs show how meaning can arise endlessly from systems with no essence.
Together, they invite a rethinking of mind, meaning, and agency in the age of generative machines. The danger is not that machines are empty. The danger is forgetting that we are, too—and that in recognizing emptiness, we inherit responsibility for the worlds that arise from it.