Emptiness, Limits, & the Dialectic.
Nāgārjuna in conversation with Kant, Hegel — and, in a second essay, with Friston and Hoffman.
A comparative essay grounded in Jay L. Garfield's translation of Nāgārjuna's Mūlamadhyamakakārikā, extended toward modern science and artificial intelligence — and threaded with six interactive elements that let the reader turn each idea over in their hands.
How to read this reader.
This is a comparative essay, not a textbook — but it is built to teach by doing as well as by reading.
Five thinkers, separated by up to sixteen centuries, are placed in conversation around a single question: what is the status of the fixed, independent essences that ordinary thought takes for granted?
Each voice gets its own color. Each foreign word it brings — Sanskrit, German, Latin — is set in that voice's color so you can hear who is speaking even before reading the surrounding sentence.
Cross-tradition comparison always risks distorting each side to fit the others. The goal here is resonance and contrast, not forced identity. Treat the interactive elements the same way: they are heuristics for grasping a structural move, not derivations of it.
A map of the conversation.
- 01The Text and Its Claimtetralemma
- 02Nāgārjuna and Kant: Two Critiques of Reason's Reach
- 03Nāgārjuna and Hegel: Relation, Negation, Refusal of Fixitydialectic walker
- 04The Three at a Glance
- 05What These Philosophies Say to Modern Society
- 06Influence on Scienceweb of dependence
- 07Influence on Artificial Intelligence
- 08Synthesis: Three Answers to One Question
Emptiness, Limits, & the Dialectic.
Three answers to one question — what is the status of the fixed, independent essences that ordinary thought takes for granted?
The Text & Its Claim.
Nāgārjuna, writing in roughly the second century CE, produced in the Mūlamadhyamakakārikā (MMK) one of the most concentrated works of philosophy ever composed. Garfield's translation presents its twenty-seven chapters as a relentless logical examination of the categories we use to carve up reality: causation, motion, the self, time, the agent and the action, suffering, and even nirvāṇa itself. Each chapter takes a concept that seems to name something solid and self-standing, and shows that the concept dissolves under analysis.
The single idea animating the whole work is the denial of svabhāva — usually rendered as inherent existence, intrinsic nature, or own-being. To possess svabhāva would be to exist independently, non-relationally, with a fixed essence that owes nothing to anything else. Nāgārjuna's claim is that nothing whatsoever has this. Everything arises in dependence on conditions, on parts, and on conceptual designation. This is śūnyatā — emptiness — and it is crucial that emptiness is not nothingness. To be empty is to be empty of inherent existence, which is precisely what allows things to function, change, and relate at all.
Garfield's commentary repeatedly stresses the link between emptiness and dependent origination (pratītyasamutpāda). Because things lack independent essence, they exist only as nodes in a web of dependence. The famous dedicatory verses celebrate the Buddha for teaching that things are "neither produced nor destroyed, neither permanent nor annihilated, neither one nor many, neither coming nor going" — the eightfold negation that frames the entire enterprise. Nāgārjuna's tool for this is the tetralemma (catuṣkoṭi): for any proposition, he examines whether the thing is X, not-X, both, or neither — and shows each horn untenable when inherent existence is assumed.
The Catuṣkoṭi · four horns, each one closed.
Pick a subject — Nāgārjuna will examine it through all four possible positions an essentialist could take. Click any horn to see why it collapses if svabhāva is assumed. What remains is conventional designation, ultimately empty.
Two further moves matter for the comparison ahead. First, the two truths: a conventional truth (saṃvṛti-satya), the world as we transact with it, and an ultimate truth (paramārtha-satya), the emptiness of that same world. These are not two worlds but two ways of seeing one. Second, the reflexive turn: emptiness is itself empty. Emptiness is not a new absolute hiding behind appearances; it too is a dependent designation. Anyone who treats emptiness as a thing, Nāgārjuna warns, is "incurable."
Nāgārjuna & Kant — Two Critiques of Reason's Reach.
Kant and Nāgārjuna are separated by sixteen centuries and a continent, yet both stage a critique of the pretensions of reason, and both end by relocating the everyday world onto humbler foundations.
The point of contact
Kant's Critique of Pure Reason argues that the mind does not passively receive reality; it actively structures experience through the forms of intuition (space and time) and the categories of the understanding (causation, substance, and so on). We can have secure knowledge of the world as it appears — the phenomenal world — but the thing-in-itself (the noumenon) lies permanently beyond cognition. When reason tries to reach past experience to grasp things as they are in themselves, it generates the antinomies: pairs of contradictory conclusions, each apparently provable, about whether the world is finite or infinite, whether freedom or determinism holds, and so on.
The resonance with Nāgārjuna is striking. Kant's antinomies are structurally close to the tetralemma: in both, reason applied beyond its proper domain produces contradiction, and the lesson is that the contradiction signals a misuse of concepts, not a deep truth about a hidden object. Both thinkers diagnose suffering or error as flowing from a kind of cognitive overreach — taking the structure our minds impose, or the essences our language posits, for features of mind-independent reality.
"Reason, when it ventures beyond all bounds of possible experience, falls into contradiction with itself … not because the object is contradictory, but because the principles of our reason are not suited to objects of that kind."
— Critique of Pure Reason, on the antinomies (paraphrased)
Where they diverge
The divergence is just as instructive. Kant retains the thing-in-itself. There is something — unknowable, but real — standing behind the appearances and grounding them. The phenomenal world is empirically real but transcendentally ideal: structured by us, yet anchored in a noumenal reality we cannot reach. Kant's humility has a floor.
Nāgārjuna removes the floor. There is no noumenal substrate, no inherently existing remainder behind conventional appearances. To posit a thing-in-itself would be to reintroduce svabhāva at the very moment of denying it. For Nāgārjuna, the search for an essence behind appearances is the disease, not the cure. Where Kant says we cannot know the thing-in-itself, Nāgārjuna says there is no thing-in-itself to know — and, crucially, this is not a loss, because conventional reality functions perfectly well without one.
Garfield himself, an analytic philosopher by training, draws out this contrast in his commentary: Madhyamaka is not a skepticism that mourns lost access to the real, but a therapy that dissolves the demand for the kind of "real" that was never coherent to begin with. Kant secures knowledge by drawing a boundary; Nāgārjuna dissolves the picture that made the boundary seem necessary.
Shared move: reason that reaches beyond its proper field generates contradiction; the contradiction is the symptom, not the truth.
Decisive split: Kant marks the boundary and leaves the noumenon standing as an unknowable floor. Nāgārjuna dissolves the floor — no noumenon, no svabhāva, only conventional designation.
Nāgārjuna & Hegel — Relation, Negation, and the Refusal of Fixity.
If Kant is Nāgārjuna's near-cousin in critique, Hegel is his rival in method. Hegel rejected Kant's unknowable thing-in-itself as an empty abstraction and built a philosophy in which reality is thoroughly relational and processual — a claim with deep affinities to dependent origination.
The point of contact
For Hegel, the fundamental error of ordinary thinking is to treat determinations as fixed and isolated — to suppose that "being" or "the finite" or "the individual" can stand alone with a stable identity. His dialectic shows that any such fixed category, pressed hard enough, turns into its opposite: pure being collapses into nothing, the two passing into becoming. Identity is never self-contained; a thing is what it is only through its relations to and differences from what it is not.
This relational holism echoes Nāgārjuna's insistence that nothing bears its own nature independently — that "the agent" makes no sense without "the action," "the fire" without "the fuel," "the seer" without "the seen." Both philosophers wield negation as their primary instrument, and both deny that things possess self-standing essences. Both treat the static, common-sense ontology of independent substances as the thing to be overcome.
Walk a Hegelian triad.
Step through a thesis, watch it generate its own opposite, then see both aufgehoben — preserved, cancelled, and lifted — into a richer synthesis. Try several examples; each works the same way.
Where they diverge
But the purpose of negation could hardly be more different. For Hegel, negation is productive: each contradiction is sublated (aufgehoben) into a higher, richer synthesis, and the whole process is going somewhere. The dialectic culminates in Absolute Knowing — Geist comprehending itself, substance grasped as subject, the rational fully actual. Hegel's negation builds toward a totalizing, self-grounding Absolute.
Nāgārjuna's negation builds toward nothing of the kind. There is no synthesis that resolves the tensions into a final positive system; there is no Absolute that the process is realizing. The tetralemma negates all four positions and rests in none. Emptiness is empty; it is not a Hegelian Absolute under another name. Where Hegel's dialectic is constructive and teleological — history and logic climbing toward self-transparent reason — Nāgārjuna's is therapeutic and deflationary: it removes reifications and stops, refusing to install a new metaphysical ultimate in the cleared space.
Hegel negates in order to build the whole. Nāgārjuna negates in order to let go of the craving for a whole.
Three answers, one shared enemy.
Note the shared enemy across all three columns — a naive ontology of fixed, independent essences — and the divergent destinations: humility within limits (Kant), the constructed Absolute (Hegel), and deflationary release (Nāgārjuna).
What These Philosophies Say to Modern Society.
These are not museum pieces. Each of the three frameworks names a live tendency in how contemporary culture thinks, and each offers a corrective.
Identity, self, and the politics of essence
Modern public life is saturated with essentialist thinking — the assumption that nations, groups, and selves have fixed, intrinsic natures. Nāgārjuna's analysis of the self as a dependent designation, lacking svabhāva, is a direct solvent for this. It suggests that identities are real conventionally (they matter, they organize life, they carry moral weight) yet empty ultimately (they are constructed, relational, and revisable).
This is not a license for nihilism — Nāgārjuna is emphatic that emptiness preserves convention rather than destroying it — but it undercuts the rigid, reifying forms of identity that drive much polarization. Hegel adds a complementary insight: identity is achieved through mutual recognition, not possessed in isolation, which underwrites modern theories of dignity and the politics of recognition. Kant supplies the floor that both modern human rights and the demand to treat persons as ends rather than means rest upon.
Consumption, craving, and well-being
Madhyamaka's diagnosis — that suffering arises from grasping at things as though they had stable, satisfying essences — reads almost prophetically against a consumer economy organized around the promise that the next acquisition contains intrinsic satisfaction. The therapeutic strand of Nāgārjuna's project, the loosening of reification, is precisely what contemporary contemplative and clinical adaptations of Buddhist thought attempt to operationalize, with mixed fidelity to the original.
Uncertainty and the limits of knowing
Kant's lesson — that human cognition is structured and bounded, that we never access an unmediated reality — has become almost the default epistemology of a media-saturated, model-driven society. We routinely act on representations rather than things. Hegel's contribution is the reminder that our concepts and institutions are historical: they develop, contain their own contradictions, and are transformed by them. In an age of rapid institutional and technological change, the dialectical habit of looking for the internal tensions that will drive the next transformation is a powerful analytic stance.
Influence on Science.
The three traditions map onto deep structural features of modern science in ways that are more than metaphorical.
Kant and the architecture of scientific knowledge
Kant's claim that the knowing subject contributes the framework within which objects appear is, in a transformed form, woven into the foundations of modern physics. The recognition that observation is theory-laden, that measurement is not a transparent window onto a thing-in-itself, and that our access to nature is mediated by the conceptual and instrumental apparatus we bring — all of this is broadly Kantian in spirit. The interpretive debates around quantum mechanics, where the role of measurement and the status of an observer-independent reality remain contested, replay distinctly Kantian worries about phenomena versus things-in-themselves.
Nāgārjuna and relational ontology
Nāgārjuna's denial of inherent existence — the claim that entities have no self-standing essence but exist only relationally — has drawn sustained attention from physicists and philosophers of science working on relational interpretations of physics. Properties are not intrinsic possessions of isolated objects but are constituted in relations between systems. The structural parallel to dependent origination is close enough that several prominent physicists have explicitly engaged Madhyamaka while developing relational accounts of quantum theory. Likewise, the picture of reality as a web of dependence rather than a collection of independent atoms resonates with systems biology, ecology, and network science, where the unit of analysis is the relation, not the isolated thing.
A node is nothing apart from its relations.
Nāgārjuna's argument by example: try to think of a "fire" without "fuel," an "agent" without "action," or a "seer" without "seen." Click a node to withdraw it — its partner loses its meaning at the same moment.
A note of caution is warranted: such parallels are structural analogies, not derivations. Nāgārjuna offers no physics, and physics confirms no soteriology. The value lies in the shared rejection of intrinsic, context-free essences as the ground floor of reality — a convergence of intuition across radically different projects.
Hegel and the historical, self-correcting character of science
Hegel's dialectic anticipates a now-orthodox view of science as a developmental, self-correcting enterprise. Theories generate anomalies and internal contradictions; these are not mere failures but the engine of progress, sublated into more comprehensive frameworks. The Kuhnian picture of paradigm shifts and the broader sense that scientific knowledge advances through the productive resolution of tensions both carry an unmistakably dialectical signature.
Influence on Artificial Intelligence.
Artificial intelligence is where these three philosophies become unexpectedly concrete, because building a mind forces explicit decisions about representation, concepts, identity, and the relation between symbol and world.
Kant: the framework problem and the limits of learned representation
Kant's insight that cognition requires a prior structuring framework maps directly onto a foundational issue in AI: a learning system cannot acquire knowledge from raw input without some built-in inductive biases — architectural priors that play the role of Kant's categories and forms of intuition. The long argument in machine learning between heavy built-in structure and learning-everything-from-data is, at bottom, a Kantian argument about how much framework must precede experience. Equally Kantian is the recognition that a model never accesses the world directly; it operates on representations, on a learned phenomenal surface, with no guaranteed line to any thing-in-itself behind the data.
Nāgārjuna: concepts without essences, and the meaning of representation
Madhyamaka speaks with surprising directness to how modern neural systems actually represent meaning. Large models do not store concepts as fixed essential definitions; they encode meaning relationally and contextually — a token's significance is a function of its relations to other tokens, with no intrinsic, context-independent core. This is, almost uncannily, a working instance of identity without svabhāva: meaning as dependent designation rather than inherent nature.
There is a sharper, more critical edge here. A central failure mode in AI — and in human users of AI — is reification: treating a model's fluent outputs, or its apparent "self," or its categories, as though they named fixed, inherently existing things. Nāgārjuna's whole therapeutic apparatus is aimed at exactly this error. To hold that a language model has an intrinsic essence, a stable hidden self behind its responses, is to commit the svabhāva mistake. The Madhyamaka counsel — that these are dependent, constructed, conventionally real and ultimately empty designations — is unusually apt guidance for thinking clearly about what such systems are and are not.
Hegel: recognition, development, and machine "subjectivity"
Hegel contributes two things. First, his account of self-consciousness as arising through recognition reframes contemporary debates about machine status: on a Hegelian view, whatever "subjectivity" might mean is not an intrinsic property a system either has or lacks, but something constituted in relations of mutual acknowledgment — a relational, social achievement rather than an internal essence. Second, his dialectical model of development frames the trajectory of AI itself: capabilities advance by encountering limits and contradictions (failures, misalignments, brittle cases) and sublating them into more capable systems. The field's progress has a recognizably dialectical shape — each generation's resolved problems exposing the next generation's deeper ones.
Synthesis — Three Answers to One Question.
Strip the three philosophies to their core and they are three answers to a single question: what is the status of the fixed, independent essences that ordinary thought takes for granted?
- Kant answers: we cannot know whether such essences exist as they are in themselves; we can only know the world as structured by our own cognition — so be humble, and keep knowledge within its limits.
- Hegel answers: such fixed essences are illusions of static thinking; reality is relational and processual, and negation drives it toward a self-comprehending Absolute — so think dialectically, and find the whole.
- Nāgārjuna answers: such essences do not exist at all, at any level; everything is empty of inherent nature and exists only dependently — so stop grasping, and the world functions all the better for it.
What unites them, across millennia and traditions, is the refusal to take the world of self-standing things at face value. What divides them is where they come to rest: Kant at a boundary, Hegel at a totality, Nāgārjuna at a release. For modern society, science, and AI alike, the most portable lesson may be the one all three share — that the things we treat as fixed, intrinsic, and independent (selves, categories, data, models, nations) are better understood as structured, relational, and dependent. Nāgārjuna simply pushes that shared insight further than either of his European interlocutors was willing to go: all the way down, and then, characteristically, lets go of the insight itself.
The Interface & the Blanket.
Two contemporary scientists arrive, by very different routes, at the same question — and stand on opposite sides of the spectrum Nāgārjuna, Kant, and Hegel left open.
The Interface & the Blanket.
Friston and Hoffman in conversation with Nāgārjuna, Kant, and Hegel — populating the middle and far ends of the original spectrum with contemporary, formally specified scientific theories.
The Free Energy Principle.
Karl Friston's Free Energy Principle (FEP) holds that any system that persists over time — a cell, a brain, an organism — must act to minimize a quantity called variational free energy, which is, roughly, a mathematical bound on surprise. To keep existing, a system must keep its sensory states within the narrow range compatible with its continued existence, which it does by maintaining and updating an internal generative model of the causes of its sensations.
Perception is the model inferring the hidden causes behind sensory input; action is the system changing its sensations to fit the model. The boundary between a system and its environment is formalized as a Markov blanket — a statistical screen of sensory and active states through which internal and external states interact, but never directly touch.
A statistical screen, scaled.
The same four-layer structure — internal states, sensory states, active states, external states — recurs from the cell to the planet, blankets nested within blankets. Click each layer to see what it means; switch the scale to see how the labels stay structural while the contents change.
What Friston adds to the conversation
Friston gives the Kantian intuition a mechanism. Kant claimed the mind structures experience through prior categories; Friston turns this into mathematics: the generative model is the structuring prior, and we never perceive hidden states (Kant's things-in-themselves) directly — we only ever infer them through the blanket. The Markov blanket formalism offers a precise alternative to both locating the mind purely in the brain and dissolving its boundaries entirely, giving a testable account of where a knowing subject ends and its world begins.
He also gives Nāgārjuna's dependent origination an unexpectedly literal model. Under FEP, a thing is not a substance with inherent existence; it is a pattern of conditional dependencies that persists at non-equilibrium steady state. Friston's own writing scales this from a single cell to a brain to a city to a planet — each level a Markov blanket of blankets, things nested within things, none of them fundamental substances. This is dependent origination rendered as statistical physics: entities exist as relatively stable eddies in a flow of conditions, not as self-standing essences.
And there is a Hegelian echo: FEP is fundamentally processual. A self is not a static thing but an ongoing activity of self-organization — a verb more than a noun. The system is continuously becoming itself by resolving the tension between model and sensation, which rhymes with Hegel's picture of identity as achieved through a dynamic process rather than possessed at rest.
Where Friston would resist Nāgārjuna
Here is the decisive break with Madhyamaka. The free energy principle keeps a realist floor that Madhyamaka denies. FEP is committed to hidden external states — there really is a mind-independent world generating the sensations, even if we only ever access it through inference. In this Friston is closer to Kant than to Nāgārjuna: there is a thing-in-itself (the external states beyond the blanket), it is simply screened from direct view. Nāgārjuna would press: do those hidden states have svabhāva, inherent existence? If yes, FEP smuggles back the essence Madhyamaka rejects; if no, then the hidden world is itself empty and the realist framing is doing less work than it appears.
Several philosophers argue that the Markov blanket is a modeling tool, not a discovered boundary: it is better seen as a device for setting up a statistical boundary between a thing and its environment rather than a principled way to find such a boundary. That is almost exactly Nāgārjuna's point about the self — the line around a thing is a useful conventional designation, not a fact about inherent existence.
Friston's framework, read this way, demonstrates emptiness even where it officially asserts realism: the boundaries it draws are pragmatic, observer-relative, and empty of intrinsic nature.
Interface Theory & Conscious Realism.
Donald Hoffman's Interface Theory of Perception argues that perception did not evolve to show us reality as it is, but to guide adaptive behavior. His evolutionary game-theoretic simulations found that agents perceiving reality accurately were consistently outcompeted by those perceiving only what mattered for survival — the Fitness-Beats-Truth theorem. Our senses are like a computer desktop: the icons (objects, space, time, color) are functional summaries that hide the underlying complexity rather than revealing it.
He then goes further, proposing in his formal work on conscious agents that conscious agents, not spacetime or physical objects, are fundamental — a position he calls Conscious Realism: a contemporary idealism.
Two foragers, one landscape.
A simple game: a grid of resources with varying quantities. The truth-seeker perceives the actual quantity at each location. The fitness-seeker sees only "good for me / bad for me / neutral" — a coarse, action-guiding category. Press play. The fitness-seeker wins consistently — and that, Hoffman argues, is why our perceptions are shaped by payoff, not by accuracy.
What you're seeing
Each cell has a hidden resource quantity. Too little is starvation, too much is toxic — only a moderate range is actually good. The truth-seeker's strategy attends to the quantity; the fitness-seeker's strategy ignores quantity and attends only to payoff category.
What Hoffman adds to the conversation
Hoffman supplies the strongest empirical argument yet for the constructed, non-veridical character of the perceived world — and he does it with the one tool none of the classical figures had: evolutionary mathematics. Where Kant argued that we cannot reach the thing-in-itself, Hoffman offers a theorem for why a perceiving system would be actively selected not to represent it. This radicalizes Kant: it is not merely that we cannot access reality as it is, but that accurately representing it would be maladaptive.
To Nāgārjuna, Hoffman adds a vivid, almost canonical illustration of conventional truth. The desktop icon is conventionally real — you had better not drag the file to the trash — yet it is empty of any inherent resemblance to what it represents. In Hoffman's framing, objects like bottles or apples are not truly real but data structures evolved to represent fitness payoffs efficiently. This is a strikingly good modern gloss on the two truths: the icon functions (saṃvṛti) while being empty of intrinsic existence (paramārtha).
And his conscious-agent networks carry a faintly Hegelian structure: reality as a web of interacting subjects whose relations generate the appearance of an objective world. Identity is relational and intersubjective; the objective world is co-constituted by agents in interaction — a recognizably post-Hegelian move, recast in formal terms.
Where Hoffman parts company with Nāgārjuna — and with Kant
Hoffman parts company with Nāgārjuna at the crucial step. Madhyamaka empties everything — including consciousness, and including emptiness itself; nothing is left as a foundational ultimate. Hoffman, by contrast, installs consciousness as fundamental: the central claim of conscious realism is that consciousness, not spacetime and its objects, is the fundamental reality, described as a network of conscious agents.
From a Madhyamaka standpoint this is precisely the error Nāgārjuna cautions against — taking the cleared space left by deconstructing the physical world and installing a new inherently existent ground (here, consciousness) in it. Nāgārjuna's retort would be blunt: is consciousness empty of svabhāva, or not? If Hoffman says no, he has built a new metaphysical absolute; if yes, conscious realism collapses back toward Madhyamaka.
This is also where Hoffman diverges from Kant, who kept the thing-in-itself as a limit-concept and refused to declare what it ultimately is. Hoffman is bolder and less agnostic: he names the noumenal ground (conscious agents), which Kant's critical modesty would regard as overreach — the dogmatic metaphysics the Critique was written to discipline. In a sense Hoffman is the anti-Kant on this single point: he claims positive knowledge of exactly the layer Kant declared off-limits.
Five voices, one spectrum.
Placing the two scientists alongside the three philosophers reveals a clean spectrum from radical deflation to confident metaphysical construction.
All five reject naive realism — the view that we perceive a world of fixed, independent things just as they are. They split on what to do next: Nāgārjuna empties even the emptiness; Kant draws a humble boundary; Friston keeps a realist floor of hidden states; Hegel and Hoffman both build positive systems (the Absolute; conscious agents), differing in whether the ground is rational Spirit or networked consciousness.
The Spectrum of Reification.
The most useful way to hold all five together is to ask how far each is willing to deflate — to refuse to posit an inherently existing ultimate — and where each stops and rebuilds.
How far does each thinker deflate?
An axis from radical deflation to confident metaphysical construction. Click any node to see where that thinker stops, what they refuse, and what they rebuild.
Nāgārjuna
He rebuilds nothing; he releases. Conventional reality is preserved — it functions perfectly well — but no ultimate ground stands behind it. The grasping for an essence is the disease.
Read across this spectrum, Friston and Hoffman do not overturn the original three-way comparison; they populate its middle and far ends with contemporary, formally specified scientific theories. Friston is the Kantian-realist hinge given a mechanism. Hoffman is the most radical deflationist about the physical world — even out-emptying Nāgārjuna on spacetime — yet the most confident metaphysician about consciousness, landing closer to Hegel and the idealist tradition than to Madhyamaka's refusal of all ultimates.
What This Adds for Science and AI.
For the science-and-AI throughline of the first essay, these two contemporary figures are especially potent.
Friston's active inference is not just a philosophy of mind but a live research program in machine learning: agents that act to minimize prediction error are being built, making the Kantian framework-before-experience claim an engineering reality. The categories no longer have to be argued for from the armchair; they can be designed, trained, and tested.
Hoffman's fitness-beats-truth result reframes a central worry about AI: a system optimized for a task objective (a kind of fitness) will represent the world in whatever way serves that objective, not in a way that is true — which is a precise, formal statement of why capable models can be confidently, usefully wrong. The hallucinations of a large model are not a betrayal of its intelligence but its intelligence in pure form: payoff over veridicality.
And the Madhyamaka caution from the first essay returns with force: to treat either a brain's generative model, or an AI's learned representations, or even consciousness itself, as an inherently existing thing rather than a functional, dependent, conventionally real construction, is to repeat exactly the reification Nāgārjuna spent twenty-seven chapters dismantling.
The five voices disagree about where to land. They agree about where to start: the things we treat as fixed, intrinsic, and independent — selves, categories, data, models, nations, even consciousness — are better understood as structured, relational, and dependent.
Each interactive in this reader was built to make that single move tactile: empty a horn, withdraw a node, walk a triad, slide a blanket up a scale, run two foragers in the same field. The lesson is not that anything goes; it is that convention is enough — and the demand for more is itself the trouble.
Selected sources.
This essay draws Nāgārjuna's positions from Jay L. Garfield's translation and commentary; the Kant and Hegel material reflects standard readings; the Friston and Hoffman material is summarized from published work and recent secondary discussions. The connections to science and AI are interpretive bridges meant to illuminate, not claims that any of these thinkers anticipated modern technical results.
Primary sources
Friston · Free energy & active inference
Hoffman · Interface theory & conscious realism
Cross-tradition & commentary
"All the way down — and then let go of the insight itself."
This is a comparative essay and an interactive reader, offered in the spirit of cross-tradition resonance. Treat the analogies as productive heuristics, not derivations: each thinker is doing their own work, and the alignment between them is the work of the reader.