Science in a Substrateless World

Science Without Building Blocks

For centuries, natural science has often been interpreted through a reductionist, substantialist framework—the assumption that scientific progress consists of peeling back macroscopic appearances to discover fundamental, self-existing 'building blocks' or ultimate physical substrates. Under this view, science searches for intrinsic natures that supposedly ground why physical dependencies obtain and why causal patterns remain stable.

The analysis of this book shows that this interpretation is both unnecessary and incoherent. Intrinsic nature is posited as an independent ground, yet it cannot be identified independently of the very dependencies it supposedly explains. When any physical phenomenon is examined in the way this book proposes, every candidate property is found only in dependence on parts, causes, functional capacities, environmental contexts, and conceptual framing—none of which is itself found from its own side.

Positing an intrinsic nature behind dependencies adds nothing to what science can describe and predict. Nor are the dependencies themselves a final explanation: they too are identified only through further relations, and no last ground is reached. Natural science does not uncover intrinsic essences, for there are none to uncover; it is the rigorous, empirical modelling of regularities within frameworks of measurement and theory—regularities that, like the frameworks, are designated rather than found from their own side.

Causation Without Producers

Classical mechanics pictured causation as one solid body acting across space on another: a cause, complete in itself, producing an effect. Contemporary science rarely speaks this way. It speaks of regularities, correlations, laws and models, and of what can be predicted under specified conditions. That language is already close to the one this book has used. A cause is identified as a cause only in relation to its effect, within a framework that picks out both (the Interlude); no producer is found reaching across to its product, and no underlying substance is found being transformed. Causal regularities hold conventionally, and hold with great precision. Nothing is found to arise ultimately.

The same is true of physical continuity. When water freezes, nothing permanent passes from the liquid to the ice; the continuity is designated on successive configurations that are neither identical nor wholly different, and the functions of each configuration—to flow, to bear weight, to cool a drink—are identified only in relation to what it acts on and the conditions that obtain.

System Boundaries and Scientific Laws as Conventions

Every science must decide where its system ends: the cell and its environment, the climate and the sun, the colliding particles and the detector. Nature supplies no pre-drawn lines. The boundary is drawn by the scientist, according to the question, the scale and the instruments—drawn for a purpose, and kept from arbitrariness by whether the resulting model holds up. A system and its environment are identified only through each other; neither is more fundamental (Chapter 13).

Laws of nature have the same status at a higher level of reliability. The conservation of energy or the equations of gravitation do not describe the behaviour of unconditioned intrinsic entities. They describe invariant regularities, formulated within a theoretical framework and tested within it. They are the most reliable conventional truths there are. They are not the ultimate structure of reality written into things from their own side.

A Lived Example: The Coastline

How long is the coast of Britain? The question sounds as if it must have an answer. Lewis Fry Richardson found, in data published in 1961, that it does not, or rather that it has as many answers as there are rulers. Measure with a ruler a hundred kilometres long and you step over the bays and headlands; measure with a ruler of one kilometre and you follow them, and the coast grows longer; measure with a ruler of a metre and you trace every rock, and it grows longer still. Benoit Mandelbrot made the case famous ('How Long Is the Coast of Britain?', 1967). There is no length of the coastline from its own side.

Yet at any chosen scale the measurement is well defined, repeatable within tolerance and useful: surveyors map by it, coastal authorities plan by it, and two careful measurers using the same ruler will agree. This is framework-dependence that any scientist can see at a glance. The length is not arbitrary, and it is not found apart from the ruler. It is designated together with the scale at which it is measured—and it is entirely reliable within that designation.

Revision Without Relativism

The history of physics shows the same pattern over time. When Einstein's theory replaced Newton's, Newton's laws were not simply discarded; they were reclassified as approximations valid within a domain. Bridges are still designed with them, and mission trajectories are still largely planned with them. What changed was the understanding of their domain: Newtonian mechanics is extraordinarily accurate at low speeds, in weak gravitational fields and at scales well above the atomic, to the precision usually required, and it fails outside those limits. A good theory is not a picture of things as they are in themselves; it is a framework valid where it holds, and correctly said to fail where it does not.

This is Tsongkhapa's account of conventional truth, with its three criteria (Chapter 13), applied to science. Analysis of the ultimate refutes only what is posited as existing by its own nature, so it excludes a primal matter or a creator, but never a convention that claims only conventional validity. Science lives by the second criterion: theories are overturned by further experience, not by fiat. That is why the substrateless view is not relativism. Some hold up and some do not, and the difference is found in experience, not in an essence behind it.

Objection: Doesn't Physics Find Fundamental Particles?

Surely, the objector says, physics has found the substrate after all: elementary particles with fixed, intrinsic properties—an electron's mass and charge, the same everywhere and always.

Contemporary physics points the other way. In quantum field theory, particles are not little things but excitations of fields, and their properties are fixed by how those fields interact. Even mass, the very picture of intrinsic substance, turns out to be relational. The electron's mass arises from its coupling to the Higgs field; it is a measure of interaction, not a quantity of stuff. The proton is more striking still. The masses of its three quarks account for only about one per cent of its mass; nearly all the rest comes from the energy of the strong interaction that binds them. What we weigh as the solid matter of the world is, for the most part, the energy of interaction.

This does not prove the Madhyamaka view, and the book does not rest on physics; physical theories are conventional frameworks, valid where they predict and open to revision. But the direction is instructive. On many readings of modern physics, the more closely it looks for intrinsic natures, the more it finds properties defined by interaction. The search for a substrate, pursued with the most precise instruments we have, does not end in a substrate.

Before Measurement: The Wavefunction and Hilbert Space

There is one place where a modern reader is most likely to bring a substrate back in, often without noticing. In quantum mechanics a system is described by a wavefunction, ψ: a vector in an abstract mathematical space called Hilbert space. Between measurements, for an isolated system, ψ changes smoothly and deterministically, according to Schrödinger's equation. When a measurement is made, a single definite outcome appears, with a probability given by the Born rule. The obvious question follows. What is the system before it is measured?

Two tempting answers

The first answer is that ψ itself is what is really there. On this reading the wave is not a mere instrument of calculation but a physical reality, and the definite outcomes we observe are its surface. David Albert argued that ψ is itself a physical field on a vast configuration space, and that ordinary three-dimensional space is derivative ('Elementary Quantum Metaphysics', 1996); in some many-worlds formulations, everything is ultimately a single vector in Hilbert space. Werner Heisenberg offered a gentler version: before measurement there exists a kind of potentiality, a 'potentia' halfway between possibility and reality, which the measurement actualizes (Physics and Philosophy, 1958).

Whatever their merits as physics, these readings have a familiar philosophical shape. They are the two-tiered picture of Chapter 5 in a new medium: a backstage reality—the wave, the vector, the potentia—doing the real work, and a front stage of clicks and readings where it shows itself. The substance has not been abandoned. It has moved into the mathematics.

The second answer swings to the opposite extreme. Nothing exists until it is observed, or the observer's consciousness brings the outcome into being. Eugene Wigner once proposed that the mind of the observer completes the measurement ('Remarks on the Mind–Body Question', 1961), and popular accounts often say that consciousness creates reality. This either makes mind the new ground or treats the unmeasured world as a void—the extremes of idealism and of non-being that earlier chapters set aside.

Bohr

Niels Bohr refused the question. A quantum object, he argued, cannot be described apart from the whole experimental arrangement in which it is observed. He proposed to use the word 'phenomenon' only 'to refer to the observations obtained under specified circumstances, including an account of the whole experimental arrangement' ('Discussion with Einstein on Epistemological Problems in Atomic Physics', 1949). The wavefunction, for Bohr, was a symbolic instrument for predicting what such arrangements would show, not a picture of something behind them. His assistant Aage Petersen recalled that, when asked whether the formalism mirrors an underlying quantum world, Bohr would answer: 'There is no quantum world. There is only an abstract quantum physical description' ('The Philosophy of Niels Bohr', 1963).

In what it refuses, Bohr's position is close to the Madhyamaka. It declines to posit an in-itself behind the phenomenon. But it leaves one thing standing. Bohr insisted that the apparatus and the results must be described in the concepts of classical physics, and he treated that classical framework as the precondition of any unambiguous description, and so did not subject it to the same analysis. The Madhyamaka analysis does not stop there. The apparatus, the classical concepts and the observer are designated too, as much as the electron they are used to observe.

Barad

Karen Barad takes Bohr a step further, from what can be described to what there is. Before intra-action, she argues, there are no determinate properties at all, not because we do not know them but because there are none to know: the indeterminacy is ontological, not merely epistemic. This is a valuable corrective to the idea that measurement merely reveals values that were already there, hidden. But, as a later section shows, Barad then makes the phenomenon itself the primary ontological unit, and so turns pre-measurement indeterminacy into a positive feature of reality—an indeterminate something.

Rovelli

Carlo Rovelli's relational interpretation goes furthest in dissolving the wavefunction as a thing ('Relational Quantum Mechanics', 1996; Helgoland, 2021). For Rovelli, ψ is not a real object spread through space. It is the bookkeeping one physical system keeps about another, compiled from past interactions and used to predict future ones. Properties exist only relative to another system, in interactions. Between interactions a system has no properties relative to a given observer. As Chapter 4 noted, Rovelli draws the connection to Nāgārjuna himself.

This is the closest of the three to the view of this book. The divergence is subtle. Rovelli often describes the world as a net of interactions, with interaction as what reality is made of. Read that way, the net becomes the new fabric—relation installed as ground, the move Chapter 4 declined. Read more lightly, as a description of how physical properties are designated, it is very nearly a Madhyamaka physics.

The Madhyamaka reading

From the side of the Mūlamadhyamakakārikā, the question 'what is the system before measurement?' has the same form as the question 'does this moment arise or cease?' (Chapter 3). It assumes an in-itself to be asked about. Nāgārjuna's analysis of conditions applies directly: neither an existent thing nor a non-existent thing can be the subject of a condition (MMK 1.6). The unmeasured system is neither a hidden definite something waiting to be revealed nor a nothing waiting to be created. Both answers assume something with a status of its own, and nothing with a status of its own is found.

This does not diminish the wavefunction. Within the theoretical framework of quantum mechanics, ψ and Hilbert space are valid conventional designations—correctable, testable and among the most precise instruments of prediction ever devised. What the analysis refuses is only the further step of treating them as what reality is made of behind the clicks of the detector. System, apparatus, wavefunction, observer and outcome are designated together; none of them is prior, and none is found on its own—not the unmeasured system, not the 'collapse', and not the observer. Meanwhile the outcome is vivid and functional. The detector clicks, and the same theory underwrites the transistors in the phone in your hand.

A substrate for emptiness

There is a deeper reason why the wavefunction is so readily taken as real. Quantum phenomena display, with unusual clarity, what earlier chapters have called the emptiness of phenomena, seen here from the side of its function, and which may be called their plasticity. An electron does not carry a fixed position and a fixed momentum waiting to be read off. What appears depends on the whole arrangement in which it is asked for—ask one way and a particle shows itself, ask another and an interference pattern. The Kochen–Specker theorem made the point formally: for any system whose quantities can take three or more values, no assignment of definite values to all quantities can respect the relations quantum theory requires among them while being independent of the context in which they are measured (Kochen and Specker, 1967). Nothing in the phenomenon is fixed from its own side, and so it can appear in strikingly different ways.

None of this is in dispute. Realist and anti-realist, substantialist and Mādhyamika alike accept the plasticity the experiments display; the difference lies in where it is placed. The substantialist mind cannot let this openness simply be. If appearances are not fixed, there must be some deeper thing that contains all the ways they might appear—a space of possibilities, a wave of tendencies, a vector in which every outcome is already folded up and waiting to be released. Hilbert space and the probability wave are pressed into this role. The very emptiness of phenomena, their lack of any fixed nature, is abstracted into a new level and given a home behind them. It is the old substrate in its most refined form: not a substance with fixed properties, but a substance made of possibilities. Something is quietly lost in the move. Once the possibilities are housed in a structure with fixed laws and fixed weights, the openness has been relocated rather than understood—the fixity Nāgārjuna refuted returning one level down.

Strictly speaking, neither Hilbert space nor the probability wave is anything at all apart from the practice that uses it. They are mathematical abstractions—instruments for articulating how the outcomes of different arrangements are related, and how often each will appear. To ask what the wavefunction is made of, or where Hilbert space is, is like asking where the number seven is kept.

Nāgārjuna answered the underlying worry in the very chapter where his opponent raised it (MMK 24.1–6; Chapter 3): it is not emptiness that would freeze the world, but its opposite (MMK 24.16–20). Then comes the verse that gathers the whole chapter, and in a sense this whole book: for whom emptiness is possible, everything is possible; for whom emptiness is not possible, nothing is possible (MMK 24.14).

The word translated 'possible' (yujyate) does not mean a stock of possibilities waiting somewhere to be realized. It means 'is fitting', 'holds together', 'works'. Nāgārjuna is not describing a space of possibilities behind the world, of the kind Hilbert space is pressed into representing. He is saying that, with emptiness, everything fits—arising and ceasing, cause and effect, the path and its fruit. The plasticity the experiments display is not drawn from a reservoir. Phenomena appear in such varied, dynamic and context-dependent ways because they depend, and their emptiness is not a further reason behind that: it is what their depending amounts to. To place the openness in a mathematical space behind them is to explain emptiness by giving it a substrate—the one explanation emptiness cannot have, and the one it does not need.

One caution belongs here, because it is so often ignored. It is popular to say that a particle in superposition is 'both here and there, and neither', as if physics had confirmed Nāgārjuna's four-cornered logic, the tetralemma (catuṣkoṭi). It has not. A superposition is a perfectly definite state in the mathematics of quantum theory, not a contradiction: every state is a superposition with respect to some quantities and definite with respect to others, and what lacks a definite value is a particular quantity, not the state. The tetralemma negates four positions about something that would have its own nature; it is not a description of vectors. Keeping the two apart protects the Madhyamaka from becoming 'quantum mysticism', and protects the physics from being made to say more than it does.

The table summarizes the readings discussed.

Table 2. Readings of a quantum system before measurement, and where each lets a ground return.

Agential Cuts and Designation

Barad generalizes Bohr's insight into a philosophy. What we call objects, and the boundary between object and instrument, are enacted by 'agential cuts' made within a phenomenon, rather than given in advance (Meeting the Universe Halfway, 2007).

The agential cut is close kin to conceptual designation as this book has described it. A boundary between system and environment, or between observer and observed, is drawn rather than found; yet it is not arbitrary, for it is constrained by the material arrangement in which it is made—just as designation, in Chapter 13, is constrained by what holds up under further experience. Barad calls such cuts material-discursive, refusing to assign them either to mere words or to mere matter. Here the convergence is close enough to be instructive for both sides.

The divergence is equally instructive. Barad names her position a realism and gives it an ontology: the primary ontological unit, she writes, is not independent objects but phenomena (Meeting the Universe Halfway, 139), and intra-action is what the world most fundamentally is. This is the same move in its most sophisticated form. From the Madhyamaka side, phenomena and intra-actions are no more found from their own side than the objects they replace; the phenomenon within which a cut is made is itself designated, by the same framework that does the cutting. Read in this way, agential realism is a powerful conventional account of how boundaries are enacted in scientific practice. Read as an account of what reality ultimately is, it repeats, at the level of relation, the substantialism it set out to overcome.

Structure Without Things?

Barad has a close cousin in the philosophy of science. James Ladyman and Don Ross argue, in a position they call ontic structural realism, that there are no things at the bottom of physics at all—only structure: the relations captured by the mathematics of our best theories (Every Thing Must Go, 2007). Their title is almost a Madhyamaka slogan—they write that the world is not made of anything—and their arguments against self-standing objects run parallel to those of Chapter 2.

They part ways with this book at the same point as Barad. Having dissolved the things, they make the structure what really exists. Structure without relata is relation without relata, and Chapter 4 showed that relation is identified only together with what it relates. The Madhyamaka accepts their negative argument in full and declines the positive conclusion. Structure, like the objects it replaced, is designated within a framework, and is not found from its own side.

What Science Gains

None of this asks science to change its methods. Experiments are run as before, theories are tested as before, and results are trusted exactly as far as they hold. What changes is the story told about what science is doing. It is not digging toward a bottom layer that will one day explain everything, and its success never depended on such a layer. It is refining designations that hold up—drawing boundaries, finding regularities, building frameworks valid within their domains—and revising them when further experience requires.

Something is gained in this. The long search for a final substrate has repeatedly ended in the discovery that the supposed bottom was itself defined by relations: atoms by their bonds, particles by their fields, mass by interaction, properties by measurement. On the substrateless view this is not a disappointment but a confirmation of what science has been doing all along. And it leaves room for something scientists often report but rarely theorize: wonder. The world revealed by the most precise instruments is not a machine of tiny parts. It is vivid, exact, endlessly detailed, reliable in every measurable respect—and nowhere found from its own side. The physicist at the bench and the reader holding a cube of ice are, in the end, meeting the world in the same way.