Speed of Light from Loop Closure

provisional

Overview

This derivation answers a deceptively simple question: why is there a maximum speed in the universe, and why is it the speed of light?

The speed of light is one of the most precisely measured constants in physics and one of the strangest. Nothing with mass can reach it, massless things must travel at exactly that speed, and it is the same for all observers regardless of their motion. In standard physics, this is postulated (Einstein’s second postulate of special relativity) and confirmed by experiment, but not explained.

The argument. The framework derives the speed of light from the loop closure condition:

• Every observer’s cycle must close in both time and space simultaneously — the same phase loop that advances the observer through one temporal period also propagates through space.
• This simultaneous closure fixes a ratio between spatial extent and temporal period: the distance covered in one cycle divided by the time of one cycle. That ratio is the speed of light.
• The ratio must be the same for all observers, because if two observers had different phase propagation speeds, phase exchange between them would require the underlying geometry to be multivalued — a contradiction.
• The speed is finite because zero speed would mean no spatial extent (contradicting the requirement for boundaries) and infinite speed would mean zero cycle time (contradicting the positive cost of each cycle).
• The Minkowski metric — the geometry of special relativity with its characteristic sign difference between space and time — emerges because increasing spatial traversal at fixed coherence cost necessarily decreases temporal traversal. Space and time compete for the same budget.

The result. The speed of light is the universal phase propagation speed through the coherence geometry, structurally fixed by loop closure. It is not an arbitrary speed limit but the rate at which coherence information propagates through the observer network. The Minkowski metric of special relativity follows as a consequence.

Why this matters. Special relativity’s foundational constant is derived rather than postulated. The derivation also reveals that space and time are not independent dimensions but two projections of a single loop closure geometry, connected by the speed of light.

An honest caveat. The derivation requires a structural postulate that the coherence geometry is pseudo-Riemannian (a smooth metric with a well-behaved quadratic form). The continuum limit from the discrete interaction graph to continuous spacetime relies on a standard conjecture in causal set theory.

Note on status. This derivation is provisional because it contains speed-of-light S1 (pseudo-Riemannian structure). If that postulate is promoted to a theorem, this derivation would be upgraded to rigorous.

Statement

Theorem. The loop closure condition requires an observer’s cycle to close in both spatial and temporal projections simultaneously. This constraint fixes a universal ratio $c = L/T$ between spatial extent and temporal period for all observer loops, where $c$ is the phase propagation speed through the coherence geometry. The Minkowski metric signature $(+,-,-,-)$ emerges from the conjugate relationship between these projections. The speed of light is not an empirical constant — it is structurally determined by the coherence geometry.

Derivation

Structural Postulate

Structural Postulate S1 (Smooth ambient geometry). The ambient coherence geometry $(\mathcal{H}, g)$ — the multi-observer spacetime — is a smooth manifold.

Remark (Tightened content). The original postulate assumed a smooth pseudo-Riemannian manifold with a non-degenerate quadratic form. Most of this content is now derived:

1. Individual state spaces are smooth manifolds — derived as Loop Closure, Theorem 0.2 (state spaces constructed from $U(1)$ Lie group orbits via the bootstrap mechanism).

2. Quadratic form (not Finsler) — derived from the local isotropy of $U(1)$ loop closure. The coherence cost of a spatial displacement depends only on its magnitude, not its direction. An $SO(d)$-invariant metric is necessarily quadratic — this is a standard mathematical result, not a physical assumption. A Finslerian metric would require a preferred spatial direction at each point, breaking the rotational symmetry inherent in the $U(1)$ structure.

3. Lorentzian signature $(-,+,+,+)$ — derived from the interaction graph’s partial order (Time as Phase Ordering), which distinguishes a temporal direction, combined with the loop closure constraint $L = cT$ (Theorem 3.1 below) which relates temporal and spatial projections with opposite signs.

The irreducible content of S1 is now specifically: the ambient multi-observer geometry assembles into a smooth manifold. Individual state spaces are smooth manifolds (derived). The geometric connections between them are smooth (relational invariants via ER=EPR create smooth throat geometries). But whether the global assembly — the network of observers connected by relational invariants — produces a smooth manifold (rather than something more exotic) is the compatibility condition identified by the Continuous-Discrete Duality (Conjecture 4.1). If that conjecture is proved, this postulate becomes a theorem.

Step 1: Dual Projections of the Observer Loop

Definition 1.1. Let $\mathcal{O} = (\Sigma, I, \mathcal{B})$ be an observer satisfying loop closure (Loop Closure, Definition 4.1). The observer loop $\gamma: S^1 \to \Sigma$ lives in the coherence geometry $(\mathcal{H}, g)$. The interaction graph $\mathcal{G}$ (Time as Phase Ordering) provides a partial order $\prec$ (the temporal direction).

Definition 1.2. The temporal projection $\gamma_T$ of the observer loop is its image under the ordering map $\prec$ — the advancement through one complete phase cycle. This costs coherence $\hbar$ (Action and Planck’s Constant, Definition 3.2) and takes duration $T$ in the interaction graph ordering.

Definition 1.3. The spatial projection $\gamma_L$ of the observer loop is its image in the coherence geometry $\mathcal{H}$ transverse to the temporal direction. The observer’s boundary $\mathcal{B}$ encloses a region of spatial extent $L$ during each cycle.

Proposition 1.4. The temporal and spatial projections are not independent: the same $U(1)$ phase that advances temporally also propagates spatially through $\mathcal{H}$. The loop must close in both projections simultaneously.

Proof. The observer loop $\gamma$ is a single closed curve in $\Sigma \subset \mathcal{H}$. Its closure $\gamma(0) = \gamma(T)$ is a single condition. The temporal and spatial projections are projections of the same loop — they do not close independently. If $\gamma_T$ closes (one complete phase cycle) but $\gamma_L$ does not (the spatial configuration doesn’t return), then $\gamma$ itself is not closed. By loop closure (Axiom 3), $\gamma$ must close, so both projections must close together. $\square$

Step 2: Phase Propagation Speed

Definition 2.1. The phase propagation speed is the ratio of spatial extent to temporal period for an observer loop:

$c_\mathcal{O} = \frac{L_\mathcal{O}}{T_\mathcal{O}}$

Theorem 2.2 (Universality of $c$). The phase propagation speed is the same for all observers: $c_\mathcal{O} = c$ for all $\mathcal{O}$.

Proof. We show $c_\mathcal{O}$ is constant within each connected component of the interaction graph $\mathcal{G}$, and therefore universal for the physical universe.

Consider two observers $\mathcal{O}_1, \mathcal{O}_2$ in the same connected component, interacting via a Type I interaction (Three Interaction Types). Phase transfer conserves total phase: $\delta\theta_1 + \delta\theta_2 = 0$ (Relational Invariants).

Suppose $c_1 \neq c_2$. Then one unit of temporal phase for $\mathcal{O}_1$ corresponds to spatial extent $c_1 \delta t$, while for $\mathcal{O}_2$ it corresponds to $c_2 \delta t$. Phase transfer $\delta\theta_1 = -\delta\theta_2$ would require a spatial conversion factor $c_1/c_2 \neq 1$, meaning the same phase unit carries different spatial content for different observers.

But phase is the $U(1)$ variable conjugate to the Noether charge (Loop Closure, Theorem 5.1). The Noether charge is the coherence content $\mathcal{C}(\Sigma)$ (Minimal Observer Structure, Proposition 4.2), defined by the coherence geometry $(\mathcal{H}, g)$ (Structural Postulate S1). The metric $g$ assigns a unique spatial length to each displacement at each point. If one unit of phase corresponds to spatial extent $c_1/\omega_1$ for $\mathcal{O}_1$ and $c_2/\omega_2$ for $\mathcal{O}_2$ at the same point in $\mathcal{H}$, with $c_1 \neq c_2$, then $g$ is multivalued — contradicting S1 (non-degenerate quadratic form).

Therefore $c_1 = c_2$ for any pair of observers that can interact. Since all physically relevant observers lie in the same connected component of $\mathcal{G}$, $c$ is universal: $c_\mathcal{O} = c$ for all $\mathcal{O}$, and $c$ is a property of $(\mathcal{H}, g)$ alone. $\square$

Remark. The universality is established within each connected component of $\mathcal{G}$. For causally disconnected regions, the phase propagation speeds need not be compared — the argument strictly applies to the observable universe as a single connected component.

Step 3: The Constraint $L = cT$

Theorem 3.1 (Loop closure in spacetime). For any observer $\mathcal{O}$ with spatial extent $L$ and temporal period $T$:

$L = cT$

This is a constraint, not a dynamical equation — it follows from the simultaneous closure requirement.

Proof. By Definition 2.1, $L_\mathcal{O}/T_\mathcal{O} = c_\mathcal{O} = c$ (Theorem 2.2). Rearranging: $L = cT$. $\square$

Corollary 3.2 (Clock-rod equivalence). An observer of period $T$ has spatial extent $L = cT$. Every clock is simultaneously a rod, and vice versa:

• A clock of period $T$ is a rod of length $L = cT$
• A rod of length $L$ is a clock of period $T = L/c$

Space and time are two projections of the single loop closure geometry, related by $c$.

Step 4: Finiteness of $c$

Proposition 4.1 (Finiteness). The phase propagation speed $c$ is finite: $0 < c < \infty$.

Proof. $c > 0$: If $c = 0$, then $L = 0$ for all observers — no observer has spatial extent. But by Observer Definition (N2), every observer has a non-trivial boundary $\mathcal{B}$, which requires $L > 0$. Contradiction.

$c < \infty$: If $c = \infty$, then $T = 0$ for all observers with finite $L$ — every observer completes its cycle instantaneously. But the temporal ordering $\prec$ from Time assigns positive duration to every non-trivial cycle (Theorem 4.2: the ordering is a partial order, and a single cycle $\mathcal{O} \prec \mathcal{O}'$ always has $T > 0$ by the positive coherence cost, Proposition 2.1 of Action). Contradiction. $\square$

Proposition 4.2 (Maximal signaling speed). No physical process propagates faster than $c$.

Proof. Any signal between two events $A, B$ must travel along a directed path in the interaction graph $\mathcal{G}$ (Time as Phase Ordering, Definition 1.1). Each edge $v_i \to v_{i+1}$ of this path is mediated by an observer $\mathcal{O}_{k_i}$ connecting the two events through its internal phase advance. The spatial displacement along each edge is bounded by $c \cdot \Delta\tau_i$ (the phase propagation speed times the proper time elapsed), since $L = cT$ is the maximal spatial extent per cycle (Theorem 3.1).

For a path of $n$ edges, the total spatial displacement $\Delta x$ satisfies:

$|\Delta x| \leq \sum_{i=1}^n c \cdot \Delta\tau_i = c \cdot \Delta\tau_{\text{total}}$

The total temporal displacement $\Delta t \geq \Delta\tau_{\text{total}}$ (with equality for a single observer; in general $\Delta t \geq \Delta\tau$ by the Minkowski metric, Theorem 5.1 below). Therefore $|\Delta x|/\Delta t \leq c$. Equality is attained by massless observers whose loops saturate $L = cT$ with zero spatial confinement (Proposition 6.2). $\square$

Step 5: The Minkowski Metric

Theorem 5.1 (Minkowski signature from conjugate projections). The coherence geometry of observer loops has an indefinite metric of signature $(+,-,-,-)$.

Proof. Consider an observer $\mathcal{O}$ at rest in the coherence geometry. Its cycle has coherence cost:

$\mathcal{S} = \hbar = \int_0^T \mathcal{L} \, dt$

Now consider the same observer in spatial motion with velocity $v$. The loop must still close — the same phase must complete one cycle. But the loop now traverses both temporal and spatial directions. The total coherence cost of one cycle is still $\hbar$ (the minimum cycle cost is universal).

In the rest frame: $\mathcal{S}_{\text{rest}} = \hbar$, with temporal path length $cT$ and spatial path length $0$.

In the moving frame: the temporal path length is $cT'$ and spatial path length is $vT'$. The total coherence cost must still be $\hbar$, but the path is longer in the combined space-time geometry.

By Structural Postulate S1, $ds^2$ is a quadratic form. The unique such form satisfying:

1. $ds^2 = c^2 dT^2$ for a stationary observer ($dL = 0$)
2. $ds^2 = 0$ for a massless observer ($L = cT$, so $c^2 dT^2 = dL^2$)
3. $ds^2$ is invariant under change of observer (universality of $c$)

is:

$\boxed{ds^2 = c^2 \, dT^2 - dL^2}$

The minus sign arises because spatial and temporal extents are conjugate aspects of the same loop: increasing spatial traversal at fixed total coherence cost necessarily decreases temporal traversal. This is the Minkowski metric. $\square$

Corollary 5.2. The proper time $d\tau$ of an observer satisfies $c^2 d\tau^2 = c^2 dt^2 - dx^2 - dy^2 - dz^2$. In $d = 3$ spatial dimensions (Three Spatial Dimensions):

$ds^2 = c^2 dt^2 - dx^2 - dy^2 - dz^2$

Step 6: $c$ Is Structurally Determined

Proposition 6.1. In the framework, $c$ is not a free parameter. Its value is fixed by the coherence geometry:

$c = \frac{\text{coherence distance per cycle}}{\text{coherence time per cycle}}$

Together with $\hbar$, $c$ establishes the conversion between spatial and temporal coherence units. The numerical value $c \approx 3 \times 10^8$ m/s in SI units reflects the human choice of measurement standards.

Proposition 6.2 (Massless observers). An observer whose loop closure saturates $L = cT$ with zero spatial confinement — whose entire coherence cost is in phase propagation — is a massless observer (photon). Its rest frame does not exist because setting $v = c$ in the Minkowski metric gives $d\tau = 0$. Massless observers are the limiting case of the loop closure geometry.

Consistency Model

Theorem 7.1. The $S^1$ minimal observer model of Loop Closure (Theorem 9.1) extends to a $(1+1)$-dimensional Minkowski spacetime model satisfying all results of this derivation.

Proof. Let $\mathcal{O} = S^1$ with period $T_0$ and coherence cost $\hbar$ per cycle. Embed in $\mathbb{R}^{1,1}$ with metric $ds^2 = c^2 dt^2 - dx^2$ (S1).

• Spatial extent: $L_0 = cT_0$ (Theorem 3.1). $\checkmark$
• Universality: All $S^1$ loops share the same ratio $L/T = c$ because $c$ is a property of the ambient metric. $\checkmark$
• Finiteness: $c > 0$ (non-trivial boundary $L_0 > 0$) and $c < \infty$ (non-zero period $T_0 > 0$). $\checkmark$
• Minkowski signature: The metric $ds^2 = c^2 dt^2 - dx^2$ has signature $(+,-)$, with the minus sign on the spatial term reflecting the conjugacy of spatial and temporal projections. $\checkmark$
• Maximal speed: A signal along a null path ($ds^2 = 0$) travels at $dx/dt = c$. No timelike path exceeds this. $\checkmark$
• Massless limit: An observer with $T_0 \to 0$, $L_0 \to 0$ at fixed $L_0/T_0 = c$ traces a null worldline ($ds^2 = 0$). $\checkmark$ $\square$

Physical Interpretation

Rigor Assessment

Fully rigorous:

• Proposition 1.4: Simultaneous closure of spatial and temporal projections (follows from the loop being a single closed curve)
• Theorem 2.2: Universality of $c$ within connected components (follows from S1 non-degeneracy + phase transfer via Type I interactions). The connected-component scope is honestly flagged.
• Theorem 3.1: $L = cT$ constraint (direct consequence of Definitions 1.2–1.3 and Theorem 2.2)
• Corollary 3.2: Clock-rod equivalence (immediate from $L = cT$)
• Proposition 4.1: Finiteness of $c$ ($c > 0$ from non-trivial boundary, $c < \infty$ from positive cycle cost)
• Proposition 4.2: Maximal signaling speed (signal paths in $\mathcal{G}$ bounded by $c$ per edge, with proper inequality chain)
• Theorem 5.1: Uniqueness of Minkowski signature from S1 (quadratic form) + three conditions (standard metric geometry)
• Theorem 7.1: Consistency model verified on $S^1$ in $\mathbb{R}^{1,1}$

Structural postulate (tightened):

• S1 (smooth ambient geometry): The quadratic form is derived from $U(1)$ isotropy, the Lorentzian signature from the interaction graph’s causal structure, and individual state space smoothness from Loop Closure Theorem 0.2. The remaining postulated content is: the ambient multi-observer geometry assembles into a smooth manifold. This reduces to the continuous-discrete duality’s compatibility condition (Conjecture 4.1).

Deferred / conjectural:

• The continuum limit from the discrete interaction graph $\mathcal{G}$ to the continuous Minkowski metric requires the Hauptvermutung of causal set theory (Time as Phase Ordering, Proposition 7.1), which is an external conjecture. This is the same deferred element as in the time derivation.
• Propositions 6.1–6.2: The structural determination of $c$ and massless observers are interpretive (no new mathematical content beyond $L = cT$ and its limiting case).

Assessment: The derivation of $c$ as a universal phase propagation speed, the $L = cT$ constraint, the maximal signaling speed, and the Minkowski signature are rigorously established from the axioms and S1. The connected-component scope and continuum-limit dependence are clearly flagged.

Open Gaps

1. $c$ from $\hbar$ and $G$: Whether $c$ is independently determined or derivable from the other fundamental constants is a key open question. The three constants $\hbar$, $c$, $G$ may be reducible to fewer independent structural parameters of $\mathcal{H}$.
2. Massless observer spectrum: The framework should derive which massless observers exist and their properties (spin-1 for photons, spin-2 for gravitons). This requires the spin-statistics connection applied to the $L = cT$ limiting case.
3. Causal structure: The Minkowski metric determines the causal structure (light cones). The framework should show that this causal structure is equivalent to the partial order $\prec$ on $\mathcal{G}$ in the continuum limit — connecting the microscopic (graph) and macroscopic (metric) descriptions of causality.