Spacetime Foliation — How Gravity Bends the Boundary of Now

If the universe is a single block of history, then by what authority does anyone decide where now is? This is the question gravity refuses to answer the same way twice.

In 1908, Hermann Minkowski told a German audience that space and time were doomed to fade into mere shadows. The two, he said, were one fabric. Half a century later, three Americans, Richard Arnowitt, Stanley Deser, and Charles Misner, tried to cut that fabric. To turn the four-dimensional whole back into something resembling the world we experience, they sliced it into three-dimensional sheets stacked like the pages of a book. Each page was space at a single instant. They called it foliation.

The cuts can be made in infinitely many ways. There is no preferred page. There is no engraved edge separating your past from your future. The line is drawn by geometry, geometry is drawn by gravity, and gravity does not produce one line. It produces a shifting, twisting boundary that depends on where the observer stands, how fast they move, and how much mass sits between them and the rest of existence.

This is the mathematics of spacetime foliation. The ADM formalism that made it precise. The experiments that turned a gauge choice into engineering. And the open question both the theory and the engineering postpone.

The problem the word “now” cannot survive

There is a thought that occurs to almost everyone, eventually, and then gets dismissed because it cannot be reconciled with the day that follows it. The thought is this. Right now, on this side of the world, it is whatever hour it is for you. On the opposite side of the planet, it is a different hour. On Mars, the local clocks read something else entirely. And out beyond the orbit of Neptune, at the edge of the Kuiper belt, the very meaning of the word now has begun to drift. A signal from there takes hours to reach us. By the time we hear it, the moment it described has aged. We agree, by convention, to call all of these instants the same instant. We say it is now everywhere. But the word now does not survive the conversion.

In ordinary life this is a curiosity. In physics it is the central problem of the twentieth century. And in general relativity, where gravity itself shapes the cloth of space and time, it has a precise technical name. The problem is the problem of spacetime foliation. The procedure for naming a now, anywhere and everywhere across the universe, is the procedure for choosing a foliation. There are infinitely many such procedures. None of them is the correct one. They are all, in the language of the field, gauge choices.

What foliation actually is

Spacetime foliation is the operation that takes a four-dimensional manifold, three dimensions of space, one dimension of time, all bound together by a single metric, and decomposes it into a one-parameter family of three-dimensional surfaces. Each surface represents space at a single moment. The parameter is the time coordinate. The family fills the manifold completely. The leaves do not intersect.

The picture to hold in mind is a loaf of bread, sliced not yet by any knife. The slices exist in the abstract. The geometry of the loaf permits them. But the knife has not chosen which direction to cut. And the knife, in general relativity, is the observer. Any worldline through the manifold can supply its own slicing. None has priority.

Hermann Minkowski was the first to argue, formally, that this was the right way to think. In 1908, in a lecture at Cologne titled simply Space and Time, he opened with words that have outlived almost everything else physics said that year. Space by itself, and time by itself, were doomed to fade away into mere shadows, and only a kind of union of the two would preserve an independent reality. He was speaking about special relativity. The point was structural. Two observers in relative motion will disagree about which events are simultaneous. The disagreement is not a measurement error. If we boost an observer to half the speed of light, the events they call simultaneous slope across the spacetime diagram at an angle. They are not the events the stationary observer calls simultaneous. Both are right. Both are wrong. The correct picture is not the slices. The correct picture is the four-dimensional union the slices cut through.

This is the origin of what philosophers, and later physicists, came to call the block universe. The claim is straightforward. The past, the present, and the future are all equally real. They coexist as features of a single four-dimensional structure. The flow of time, the sense that one moment becomes another, is not a property of the block. It is a property of how a conscious observer scans through the block.

Einstein bends the block

Special relativity had handed physics a four-dimensional block. The block was flat. The metric was simple. Slicing it into nows was a matter of choosing a velocity. But Einstein had not yet finished. In 1915, he published the field equations of general relativity. The block now had curvature. Mass and energy could bend it. Light, which had previously traveled in straight lines through Minkowski space, now followed geodesics that curved around stars. And the act of slicing, the act of foliation, became dramatically more delicate.

The Einstein field equations, written compactly, are:

Gμν + Λgμν = (8πG/c⁴) Tμν

The left side encodes the curvature of spacetime. The right side encodes the matter and energy contained within it. Matter tells spacetime how to curve, and curved spacetime tells matter how to move. That sentence is general relativity.

In a curved spacetime, the question of which spatial slice corresponds to the present loses any clean answer. The slices must respect the underlying causal structure, must not cut through their own future, must remain spacelike everywhere. But within those constraints, infinitely many foliations are permitted. A foliation is a way of presenting the geometry. It is not a feature of the geometry itself.

The ADM formalism

By the late 1950s, general relativity had been around for forty years, but most physicists treated it as a theory of static or near-static geometries. The dynamical evolution of spacetime, how one slice of geometry produces the next, was understood in principle but not in practice. To compute it, you needed a way to break the four-dimensional Einstein equations back into a form that looked like ordinary evolution in time.

In 1959, three American physicists at the Institute for Defense Analyses and the University of North Carolina, Richard Arnowitt, Stanley Deser, and Charles Misner, published a paper that did exactly this. They reviewed the framework comprehensively in 1962 in Gravitation: An Introduction to Current Research, edited by Louis Witten. The framework is now called the ADM formalism, and it is the foundation of every modern numerical simulation in general relativity.

The ADM formalism assumes a foliation of spacetime into spacelike slices. It introduces two key objects.

The lapse function, denoted N, measures the rate at which proper time flows along the normal to each slice. It tells you how much physical time elapses between one slice and the next.

The shift vector, denoted Nⁱ, describes how spatial coordinates on one slice are mapped to spatial coordinates on the next. It tells you how the coordinate system shifts as you move from slice to slice.

With these two objects, plus the spatial metric γᵢⱼ on each slice and its conjugate momenta, the ADM formalism rewrites Einstein’s equations as a set of constraints and evolution equations. The constraints must be satisfied on every slice. The evolution equations tell you how the spatial geometry on one slice produces the spatial geometry on the next.

The lapse and shift are not determined by the theory. They are inputs the user provides. Different choices of lapse and shift correspond to different foliations of the same spacetime. The physics is the same. The bookkeeping is different. This is refoliation invariance, the symmetry that the physical content of a solution does not depend on the choice of foliation.

Cauchy surfaces and the condition of being sliceable

Not every spacetime admits a foliation. The technical condition that does is called global hyperbolicity, and it requires the existence of a Cauchy surface. A Cauchy surface is a three-dimensional spacelike hypersurface that every causal trajectory, every possible path of any signal or object, crosses exactly once.

This is the geometric condition that makes the slicing work. If a spacetime is globally hyperbolic, then specifying initial data on a Cauchy surface uniquely determines the future evolution everywhere. The ADM formalism takes this initial data and steps it forward, producing the geometry on each subsequent slice.

Picture a stage manager preparing a long, complex theatrical performance. The script has thousands of actors moving through hundreds of scenes. For the play to be performed at all, the stage manager must construct a sequence of cues. Each cue is a snapshot of the stage at one moment. Every actor must be somewhere when the cue fires. No actor may be in two places. No actor may be missing. The play is the sum of the snapshots. Without a coherent cue list, the play cannot exist.

Spacetime, in general relativity, is the play. The cue list is the foliation. The condition that the cue list can be constructed at all is global hyperbolicity. Every spacetime physicists use for time-evolution problems is assumed to be globally hyperbolic. The condition is not automatic. It must be assumed, or proven, or, in practice, checked.

Foliation in real engineering

The ADM formalism, with its lapse and shift, with its constraints and its evolution equations, is the practical workhorse of numerical relativity. Every gravitational wave signal that LIGO has ever detected has been compared against a template generated by numerical simulation of binary black hole or neutron star mergers. Those simulations split spacetime using the ADM formalism, or one of its modern descendants, the most common being the BSSN reformulation developed in the 1990s by Thomas Baumgarte, Stuart Shapiro, Masaru Shibata, and Takashi Nakamura.

The slicings used in numerical relativity have technical names like one-plus-log slicing for the lapse and gamma-driver shift for the shift. The choice is not arbitrary in practice. Some slicings handle singularities gracefully. Others crash the simulation. But the choice is arbitrary in principle. The physics does not care.

On September 14, 2015, LIGO detected a gravitational wave signal designated GW150914, the merger of two black holes, roughly 29 and 36 times the mass of the Sun, at a distance of about 1.3 billion light-years. The waveform was matched against numerical templates computed across the previous decade by groups in Pretoria, Caltech, Rochester, and Göttingen. The match was within experimental error. The slicing those simulations used was a gauge choice. The wave the detector received was a physical fact. Both descriptions of the same event are correct. The slicing was the bookkeeping. The wave was the universe.

Foliation in your pocket

You carry, almost certainly, a device with a Global Positioning System chip. That chip computes your location by comparing the times stamped on signals arriving from a constellation of orbiting satellites. Each satellite carries an atomic clock. The satellites orbit at an altitude of approximately 20,200 kilometers. They move at roughly 4 kilometers per second.

According to general relativity, their atomic clocks tick faster than identical clocks on the ground, because they sit higher up in Earth’s gravitational potential. The relativistic correction is approximately +45 microseconds per day. According to special relativity, their atomic clocks tick slower than identical clocks on the ground, because they are moving. That correction is approximately −7 microseconds per day. The net result is that the satellite clocks gain about 38 microseconds every twenty-four hours, relative to a clock at sea level.

If GPS engineers had not corrected for this, if they had simply assumed that all clocks tick at the same rate everywhere, your position fix would degrade by approximately 10 kilometers per day. Within a week, navigation would be useless. The system works only because every GPS satellite is operating with a slicing of spacetime imposed by hand. The slicing is called Earth-centered inertial time. It is a particular foliation. It is one of an infinite number of foliations that the Earth-satellite system could be using. The other foliations would yield equivalent physics, but only this one yields navigation that lines up with where you actually are.

Every car, every plane, every fishing boat that uses GPS is enacting a choice of foliation. The choice is hidden inside the software. The driver does not know. The pilot does not know. The captain does not know. They simply trust that the position is correct.

The Hafele-Keating experiment

In 1971, two physicists named Joseph Hafele and Richard Keating flew four cesium atomic clocks around the world on commercial airliners. They flew them eastward. Then westward. They compared the readings against reference clocks held stationary at the United States Naval Observatory.

The eastward clocks were predicted by general relativity and special relativity, combined, to lose approximately 59 nanoseconds. The westward clocks were predicted to gain approximately 273 nanoseconds. The measurements were within the experimental uncertainty of the prediction.

The experiment cost almost nothing, a few plane tickets and some clock rental, and it established, in the most ordinary possible setting, that there is no universal clock. The eastward and westward airliners passed through different foliations of the same planet. They returned to the same place, having aged differently. The differences were nanoseconds. The principle was absolute. Two world lines through the same spacetime experienced different elapsed proper times. The block universe has no problem with this. It says, simply, the world lines were different paths through the same block.

Frame-dragging and the twisting of slices

Gravity Probe B, launched by NASA in 2004 and operated until 2011, tested something more subtle still. The mission carried four gyroscopes, each a polished quartz sphere about the size of a ping-pong ball, machined to be one of the most spherical objects ever produced. The deviations from perfect roundness were less than ten nanometers. The gyroscopes were placed in low Earth orbit, oriented at a guide star, and allowed to drift.

According to general relativity, the gyroscopes should precess by two effects. One is the geodetic precession, about 6,600 milliarcseconds per year, caused by the curvature of spacetime around Earth’s mass. The other is frame-dragging, the Lense-Thirring effect named after the Austrian physicists Josef Lense and Hans Thirring who predicted it in 1918, about 39 milliarcseconds per year, caused by the rotation of Earth literally twisting the spacetime around it. Both predictions were verified to within experimental error.

Earth is not just a mass. Earth is a rotating mass. Its rotation drags the local foliation along with it. The slices, near Earth, tilt and twist around the rotation axis. The boundary between your past and your future, the slice you would draw to define your now, is being dragged sideways by the spinning mass of the planet you stand on. The drag is tiny. The principle is enormous. Mass does not just curve space. Spinning mass spins space. And spin space is also tilted time.

Refoliation invariance: the deepest claim

We are now ready to state the principle that runs through everything. In general relativity, there is no preferred slicing. There are infinitely many slicings, and any of them can be used, and the equations of the theory do not select one over another. Different observers in different states of motion, in different gravitational potentials, sit on different slices. None of them is right. None of them is wrong. They are all gauge choices.

The slice that is now for you is past for somebody else and future for somebody after that. The universe does not adjudicate.

This is the modern, technical version of what philosophers have argued since at least the 1960s. The position called eternalism, the view that past, present, and future are all equally real, is the natural reading of refoliation invariance. The position called presentism, the view that only the present exists, requires a preferred slice that the equations do not provide. The intermediate position called the growing block requires a privileged leading edge that the equations also do not provide. Among contemporary philosophers of physics, the block universe reading is dominant precisely because it is the only one that respects the symmetry general relativity exhibits.

The position is contested. The work of Howard Stein and others has argued that the inference from refoliation invariance to the block universe smuggles in metaphysical assumptions that the geometry does not strictly require. But the structural fact, that no global now is invariant under the symmetries of general relativity, is not contested. The geometry permits any slicing. It privileges none.

The deeper layer: Wheeler-DeWitt

At the deepest level of canonical quantum gravity, the absence of a preferred foliation becomes more dramatic. The Wheeler-DeWitt equation, formulated by John Wheeler and Bryce DeWitt in the late 1960s and published in DeWitt’s 1967 Physical Review paper, applies the ADM formalism to the quantum state of the entire universe.

Written compactly, it is Ĥψ = 0. The Hamiltonian operator, acting on the wavefunction of the universe, gives zero. The equation contains no time variable. Compare this to ordinary quantum mechanics, where the Schrödinger equation describes a wavefunction that evolves with respect to time. The Wheeler-DeWitt equation has no such evolution. It describes a wavefunction that simply is.

This is sometimes called the problem of time in quantum gravity. At the deepest level of the theory, when you quantize the ADM formalism for the universe as a whole, the foliation drops out entirely. There is no external time parameter for the equation to evolve in. Time, in this picture, is not a fundamental feature of reality. It is a relational concept that has to be recovered, perhaps by choosing some part of the universe to serve as an internal clock for the rest.

The question the foliation cannot answer

The foliation of spacetime, in general relativity, is a precise mathematical operation with no preferred answer. The geometry permits any slicing. The physics does not adjudicate. The block is fixed. The slices are the bookkeeping.

But notice what this leaves unsaid. The block is fixed. The slices are infinite. The mathematics is precise. The constants of nature, the structure of the equations, the conditions of global hyperbolicity that allow the foliation to exist at all, are all features of a single coherent geometry that the universe somehow exhibits. Why this geometry rather than any other is a question the foliation does not address. Why this set of equations rather than any other is a question the ADM formalism does not address. Why the universe is sliceable at all, why the cue list can be constructed, is a question the theory assumes rather than answers.

The mathematics describes what the universe does, with stunning precision. The mathematics does not say why anything is doing it. That question, the question of what grounds the entire framework, is one that physics is structurally not equipped to answer.

This is the question the companion documentary on the Sleepy Joe Space YouTube channel takes up at length.

What stays

What stays from spacetime foliation is a precise vocabulary for one of the strangest features of our best theory of gravity. The universe, on this account, does not come with a built-in now. There is no preferred slicing of spacetime. There are infinitely many ways to cut the four-dimensional whole into three-dimensional pages, and all of them are equally valid descriptions of the same reality. The block universe is the natural reading. Refoliation invariance is its mathematical heart.

The ADM formalism, developed in 1959 and refined in 1962, is the technical machinery that makes the slicing work. It introduces the lapse function and the shift vector to describe how slices are stacked, produces a Hamiltonian for general relativity, and provides the computational foundation for every numerical simulation of black hole mergers and gravitational wave signals. Without the ADM formalism, LIGO could not match its data. Without the foliation choice it requires, GPS would not work. The gauge choice no one outside the engineering teams ever sees is the gauge choice that lets the engineering teams know where you are.

What also stays is the discipline. The mathematics is silent on which slicing the universe really uses, because the universe really uses no slicing in particular. Every foliation is a description, not a description of a different reality. The boundary between your past and your future is not a property of the world. It is a property of the path you happen to be following through it.

The block is fixed. The pages are infinite. The knife belongs to the observer, not to the loaf.