The Universe Is a ComputerEpisode 12 (extended, the linking episode) / With a Single Spring Network

"A force is a spring vibration passing on to the next lattice site while fading" — that one line drops the whole thing onto a single sheet

With a Single Spring Network A field = a spring network stretched over a lattice. A force is the ripple that travels through it. Local coupling + finite speed = it runs on a computer as is. The phenomena of Episodes 1–11 all drop onto this one sheet.

Core: field = coupled oscillators, force = a propagating wave, local updates so it runs on a computer Hook formula: \(\partial_t^2 u=c^2\nabla^2u-m^2u\)

A reader summed up everything through last time in a single line — "A force is just a spring vibration that keeps passing on to the next lattice site while fading, right? That's why it can be computed." This is not a metaphor; it is the literal identity of field theory. This episode adds no new material. It shows how that one line binds the whole series together, with a single correspondence table and a moving spring network — this is the linking episode.

For those reading only this episode (the one-page premise) This series, "The Universe Is a Computer," is a piece of reading that views the universe not as physics but as a process of computation. Up through last time we built one tool — \(F=1/(Cn)^D\). This is not a new law of force, but a ruler that measures "by what fraction some quantity \(F\) shrinks as you increase the distance or size \(n\)," and the number expressing "how steep that falloff is" is the dimension \(D\) (the slope when drawn on a log–log graph). \(D=2\) means "distance doubles, quantity to 1/4" = the familiar inverse-square law (electric force or gravity). In Episode 10 we saw that \(D\) changes per force among the four; in Episode 11, that atoms and planets can exist stably only when \(D\) is exactly 2. This episode opens up the contents of that "force" all the way, with a spring toy. We write so it works even if you haven't read the earlier episodes.

01A field = a network of weights + springs. Your rephrasing is the textbook itself

First, the word "field." It's the collective name for something that has a value assigned to every nook of space — electric field, magnetic field, gravitational field, and so on. Drop that into a mechanical image: a network of countless weights arranged on a lattice, with each connected to its neighbors by springs. In physics language this is called coupled harmonic oscillators (weights that swing back and forth on springs, tied to one another), and this is the first-principles definition of a field.

In a familiar picture: the taut surface of a trampoline, a water surface, or the springs of a mattress. Poke it anywhere and the dip passes on to the next neighbor and the next, spreading and weakening. This "poke it and a wave travels" is exactly what it means for a force to act. And:

Written as an equation, the motion of the spring network is exactly the single wave equation \(\partial_t^2 u=c^2\nabla^2 u-m^2u\) (in physics this is called the "Klein–Gordon equation." \(u\) is the size of the sway at each point, the left side is the acceleration of the sway, the first term on the right is "being pulled by the neighbors," and the second term \(m^2u\) is the "on-site spring" that shows up later). In Episode 1 we said "\(F=1/(Cn)^D\) is not a force equation but a ruler." So what was the "force" that the ruler measured actually made of — the answer was this single spring network, and that is this episode.

02One correspondence table — which spring part is each earlier phenomenon?

The phenomena we saw separately across the series all drop onto parts of this spring network. The colors are each episode's — this table is the map of how it "binds the other episodes together."

Phenomenon seen so farIts identity in the spring networkEp.
The field itselfweights + nearest-neighbor springs arranged on a lattice (coupled oscillators)all
Mediator particlea ripple traveling through the network (a vibrational mode)10
Finite propagation speed = \(c\), light conelocal coupling that travels only to the neighbor in one step2, 4
Fading \(1/r^{D}\)the ripple disperses over the wavefront \(r^{d-1}\)1, 10, 11
Massless → \(D=2\) (gravity, electromagnetism)nearest-neighbor springs only → honest inverse square10, 11
Weak force \(D\to\infty\)an "on-site spring" = mass term pinning each site to its rest position → exponential decay10
Strong force \(D\to0\)nonlinear spring (a flux tube that stays twisted and taut)10
\(D=2\) is the blade (stable binding)geometry in which the spring network vibrates stably = can store a state11
Complex dimension, log-periodicitya hierarchy of discrete scales in the spring network (a self-similar tick)8
Formula of information → differential equationthe continuum limit as the lattice spacing \(\to0\) (difference → derivative)9
◇ ◇ ◇

03So it can be computed — this was the backbone of the series

"That's why it can be computed" is the deepest line on this sheet. Why can we say so? The spring network is local (each weight is tied only to its neighbors) and finite-speed (one step travels only one lattice's worth). If so, a given weight's "sway at the next instant" is decided by nothing but its own and its neighbors' "sway right now" — this "procedure that computes the next from the now" is called an update rule. Which means:

The core of the linking

Local springs (neighbors only) + finite speed + an update rule = exactly the form a computer can advance one step at a time. There is no need to solve the whole thing at once; you just update each point a little from its neighbors' values — this is exactly the computer's specialty. And this is not just an idea but a practice: lattice QCD (carving spacetime into a fine lattice, arranging spring variables on it, and actually running it on a supercomputer to numerically compute the strong force — a vast decades-old field of computational physics) is a concrete example. Episode 2 (what kind of computer to view the universe as) → Episode 4 (the map of how things travel = the causal graph) → Episode 5 (when there is no shortcut and you can only run it) are linked into one by this single line.

Figure: a living spring network (2D). Press "Send a pulse" to flick the center, and the ripple travels to the next neighbor and the next at finite speed, spreading and fading = this is force. Raise the slider "on-site spring (mass term)" and each site is pulled back to its rest position, and the wave dies exponentially = a short-range force (weak force, Episode 10). At zero it reaches far = a long-range force (\(D=2\)). Red = sway toward you / blue = sway away.
Press "Send a pulse" to flick the center. Raising the mass term shrinks the distance the wave reaches.
weight (lattice point) sway toward you (+) sway away (−)

04The honest line — "runs on a computer" and "solves easily" are different

The honest line ① runs ≠ solves

If the spring is linear (it tries to return in exact proportion to how far you stretch it = an honest spring), then you can decompose the sway into a clean superposition of sine waves and write down the answer in one shot. But that is a world where the waves merely pass through each other and never collide = a "free field" where forces don't act on one another = a world so easy that nothing interesting happens. Real forces (the confinement of the strong force, for instance) are nonlinear springs (when the sway is large the return twists = the waves affect one another), which can't be cleanly decomposed, and you can only actually run the computer from start to finish. We called this "no-shortcut" property computational irreducibility in Episode 5. So the correct statement is "because updates are local, it can be put on a computer," not "so it solves easily" — being writable as a procedure and there being a shortcut are different things.

The honest line ② is the lattice a scaffold or a foundation?

"Passing on to the next lattice site" is the finest computational image, but there is a pitfall. Lay down a grid like a go board naively, and preferred directions along the rows/columns and a preferred rest frame are secretly created. Yet the core of relativity is "the laws of physics are the same seen by anyone moving at any speed" (Lorentz invariance), and the lattice breaks this (= a preferred reference frame appears. This is why the simple lattice model was disqualified in Episode 2). Moreover, the operation of making the lattice ever finer to return to smooth space (the continuum limit) works out only in special situations (critical points) (Episode 9). So is the lattice merely a temporarily laid scaffold for computation, or the true foundation of spacetime — no one has settled this yet. This is the "one hole" running through the whole series. The spring network is real as an image, but whether it is the substrate of the world is a bet. We don't fill this in — never say "solved," that is the rule of the whole series.

Practice problems (solvable from this one sheet)
  1. Which part of the spring network makes the weak force short-range?
    Show the answer
    The presence of, in addition to the nearest-neighbor springs tying weights together, an "on-site spring" that pulls each weight back to its own rest position. In the equation it's the term \(m^2u\), and this \(m\) corresponds to "the weight (mass) of the force-carrying particle." With it, the way the sway travels (the relation between wavelength and frequency = the dispersion relation \(\omega^2=c^2k^2+m^2\)) changes, and the wave dies out rapidly before it can go far — so the distance it reaches is short (= short-range force). This is what makes the wave vanish quickly when you raise the mass slider in the figure, and it is the true identity of the weak force's \(D\) running to infinity in Episode 10.
  2. Why does "local + finite speed" mean "runs on a computer"?
    Show the answer
    Because each lattice's next value is decided by nothing but its own and its neighbors' current values (a local update rule). Instead of solving the whole thing at once, you can advance it lattice-by-lattice, one step at a time = exactly a computational step. This is what lattice QCD actually does (Episodes 2, 4, 5).
  3. Why are "runs on a computer" and "solves easily" different?
    Show the answer
    If linear (free field), it's one-shot with normal modes but there's no interaction and it's boring. Real forces are nonlinear springs = interactions, with no shortcut, so you can only actually run it (computational irreducibility, Episode 5). Runs = writable as steps; solves = there's a shortcut — different things.

SummaryIt was all one spring network

A force is a vibration that travels to the next neighbor while fading across the lattice's spring network. Mediator particle = the ripple, finite speed = \(c\) = the light cone, fading = \(1/r^D\), mass term = the on-site spring (the weak force's short range), nonlinear spring = the strong force's confinement, stable vibration = the blade of \(D=2\), discrete scales = complex dimension, lattice spacing→0 = the continuum limit where the formula of information becomes a derivative. Episodes 1–11 all drop onto this one sheet.

And because it is local + finite speed + an update rule, it runs on a computer as is — this is the backbone of the series. But runs ≠ solves easily (nonlinear must be run, Episode 5), and whether the lattice is scaffold or foundation is undecided (Lorentz, the continuum limit, Episodes 2, 9). The spring network is real as a picture, but a bet as the substrate — the map was drawn, but no flag was planted.

← Episode 11: D=2 Is on the Blade's Edge To the contents Episode 13: There Never Were Any Balls →
This document is Episode 12 (extended, the linking episode) of the series "The Universe Is a Computer." The picture of a field as coupled harmonic oscillators (a spring network on a lattice), the quantum as a ripple of the mediating field (phonon / photon), the wave equation and the mass-term Klein–Gordon equation \(\partial_t^2u=c^2\nabla^2u-m^2u\) (the dispersion relation \(\omega^2=c^2k^2+m^2\) → Yukawa-type decay), lattice gauge theory and lattice QCD (numerical computation of gauge fields on a spacetime lattice), linear = solvable with normal modes / nonlinear = interactions are irreducible, lattice regularization explicitly breaking Lorentz invariance, and the continuum limit holding only at a renormalization-group fixed point (a critical point) are all established physics/mathematics. No new claims are included. The figure is a schematic 2D scalar wave simulation, and the color represents the sign of the displacement. Whether spacetime is literally a lattice (discreteness of the substrate) is a current open problem. — To print, use your browser's "Print" and "Save as PDF" (in the print version the slider, button, and answers are static and hidden).

Print / save as PDF: Ctrl+P (⌘+P on Mac). On screen, "Send a pulse" flicks the spring network, and the mass-term slider shows the distance the wave reaches shrinking. Click "Show the answer" to open each solution.