Substantivalism and Relationism

The oldest and most persistent question in the philosophy of space and time is ontological: does spacetime exist in its own right, or not? The substantivalist says that space (or, after Einstein, spacetime) is a genuine entity — a container of events that would remain even if every material body were annihilated, and whose points and regions have a reality independent of what, if anything, occupies them. The relationist denies this: there is no spatial or temporal stuff, only bodies and events standing in spatiotemporal relations — being three metres apart, lying between, coming after. On the relationist view "empty space" is not a vast transparent container but simply the absence of anything at all, and to speak of "a point of space" where no body is, is to speak of nothing.

This page sets out the debate in its classic and modern forms. The detailed arguments live in the ontology subfolder — the Leibniz–Clarke correspondence, Newton's bucket, Mach's principle, the hole argument, and supersubstantivalism — and this page relates them and states the modern options.

The classical dispute: Newton against Leibniz

The debate crystallised at the birth of modern physics. In the Scholium to the Principia (1687), Newton distinguished absolute from relative space and time:

Absolute space, in its own nature, without relation to anything external, remains always similar and immovable.

Absolute space and absolute time form a fixed, infinite, immovable backdrop against which all real motion is defined. Newton needed this because his physics distinguishes true motion from merely relative motion, and the distinction seemed to require a privileged standard of rest — a use dramatised by the rotating bucket.

Leibniz, in his 1715–16 correspondence with Samuel Clarke (Newton's spokesman), rejected the container. Space is "the order of coexistences" and time "the order of successions": relational systems, abstractions from the ways bodies are arranged and events succeed one another, not things in addition to them. Leibniz wielded two great principles:

• The Principle of Sufficient Reason (PSR) — nothing is so without a reason. If space were an absolute container, God would have had to choose where to place the material world within it; but all locations are perfectly alike, so there could be no reason to prefer one to another. Absolute space thus forces a choice with no sufficient reason, and must be rejected.
• The Principle of the Identity of Indiscernibles (PII) — things with all the same properties are identical. A universe shifted five metres east in absolute space, or created a century earlier in absolute time, would be observationally indiscernible from the actual one; by PII it is not a genuinely distinct possibility, so the absolute space that would distinguish them is a fiction.

These static and kinematic shift arguments are the heart of the relationist case: if absolute positions and velocities make no observable difference, positing them violates both principles. The substantivalist replies are developed on the Leibniz–Clarke page.

The pivot: acceleration and the bucket

The relationist has an elegant answer to position and uniform velocity: they are relational, and physics agrees, since Galilean (and later Lorentz) invariance makes absolute position and absolute velocity undetectable. But acceleration and rotation are another matter. Newton's bucket — water climbing the walls when, and only when, it rotates — and the two-globes thought experiment suggest that rotation has real, detectable effects even in an otherwise empty universe, where there are no relative motions to ground it. Something non-relational — absolute space, or absolute acceleration, or an inertial structure — seems to be doing work.

This is the pressure point of the whole debate. Three broad responses:

• Newtonian substantivalism — accept absolute space (or, more economically, the affine/inertial structure that distinguishes accelerated from unaccelerated motion). Modern versions posit neo-Newtonian (Galilean) spacetime, which has enough structure to define acceleration but not absolute velocity or position — conceding the shift arguments while keeping what the bucket needs.
• Machian relationism — deny that rotation in a truly empty universe would produce inertial effects; inertia is determined by the total distribution of matter (the "fixed stars"), so remove the matter and the effects vanish. This is Mach's principle, partially and controversially realised in general relativity.
• Reject the framing — hold that the very question "relative to what is the bucket rotating?" presupposes the container picture, and that the right level of description is the spacetime structure itself, neither a Newtonian substance nor a pure relational web.

The modern debate: substantivalism after Einstein

General relativity transformed the dispute. Spacetime is no longer a fixed stage: the metric field $g_{\mu \nu }$ is dynamical, curved by matter and energy and acting back on their motion. Several distinctions now matter:

• Manifold substantivalism takes the bare differentiable manifold of spacetime points to be the substance, with the metric a field defined on it. This is the target of the hole argument, which shows that manifold substantivalism plus the diffeomorphism invariance of general relativity yields a radical indeterminism — two models that agree everywhere outside a "hole" but differ within it count as physically distinct worlds. Something has to give.
• Metric-field (or "sophisticated") substantivalism identifies the substance not with the bare manifold but with the manifold-plus-metric, and treats diffeomorphic models as representing the same physical situation (a gauge symmetry, "Leibniz equivalence"). This dodges the hole argument's indeterminism at the price of conceding that spacetime points have no identity independent of the metric — a partial victory for the relationist intuition.
• Structural realism goes further: what is real is the structure — the pattern of metric and causal relations — with neither an underlying substance nor primitive individual points. This is increasingly seen as the natural reading of general relativity, a genuine third way between the classical antagonists.
• Relationism survives, but must now be relationism about a dynamical field. Modern relationists (Barbour, and the tradition after Mach) reconstruct dynamics from relational configurations alone; the difficulty is recovering the full content of general relativity, including the inertial and gravitational effects the bucket first exposed.

A summary of the options

Where this sits

Substantivalism and relationism is the ontological question of the three that organise this section. It cannot be settled a priori: Leibniz's principles push toward relationism, but Newton's bucket and the inertial effects push back, and general relativity reshapes the board through the hole argument and the dynamical metric. The debate also feeds the other two questions — a relationist about time will treat the order of succession as prior to any temporal "container," and the reality of the metric field bears directly on the status of geometry. The individual arguments are developed next, beginning with the classic absolute-versus-relational statement.