On the Uniqueness of Embeddings of Causal Sets

Technical Summary: On the Uniqueness of Embeddings of Causal Sets

Problem Statement
The central problem addressed is the Hauptvermutung (main conjecture) of causal set theory: whether a single finite causal set can faithfully embed into two macroscopically distinct Lorentzian manifolds. While Malament's theorem establishes that the causal order of a continuous spacetime determines its conformal structure, and Bombelli–Lee–Meyer–Sorkin argued that volume information fixes the scale, the finite case remains subtle. Recent work by Müller [13] demonstrated that for finite sets, preserving only the causal order is insufficient to guarantee uniqueness; distinct spacetimes can admit order-preserving embeddings of the same finite poset. However, the physically relevant setting in causal set theory involves finite structures obtained via Poisson sprinkling at finite density, where both order and volume information are available. This paper seeks to prove a quantitative version of the uniqueness conjecture for such Poisson-sprinkled causal sets, establishing that if a causal set admits "well-conditioned" embeddings into two spacetimes, those spacetimes must be approximately isometric in the high-density limit.

Methodology
The proof is divided into two distinct parts: a deterministic geometric analysis (Part I) and a probabilistic verification (Part II).

• Part I: Deterministic Geometry (Well-Conditioned $\implies$ Uniqueness)
  The authors introduce the concept of a well-conditioned embedding (Definition 2.6), which augments the standard "order + number" faithfulness with a third condition: the preservation of longest-chain lengths as proper times.

  1. Anchor Scaffolds: Using the volume-faithfulness condition (F2), the authors prove the existence of a "scaffold" of $d+1$ embedded points in the deep interior of the manifold that approximate a fixed, well-conditioned Minkowski configuration (Lemma 4.1).
  2. Lorentzian Trilateration: A key technical innovation is the use of a Lorentzian trilateration identity (Lemma 3.2) and a finite-point analogue of the Alexandrov–Zeeman theorem (Lemma 3.4). This allows the authors to reconstruct the local Lorentz transformation between two embeddings based on the preservation of causal distances to the anchor scaffold.
  3. Moving Karcher Mean: To construct a global map $\Phi: M_1 \to M_2$, the authors define a moving Riemannian center of mass (Karcher mean) of the target points weighted by a Gaussian kernel centered at source points (Construction 4.5).
  4. Rigidity and Isometry: By combining the non-degeneracy of the local point cloud (guaranteed by volume-faithfulness, Lemma 4.3) with the distance preservation (F3), they show that the differential of $\Phi$ is an approximate Lorentz transformation. A degree argument (Proposition 4.12) then upgrades this local diffeomorphism to a global one, proving that the pullback metric $\Phi^* g_2$ approximates $g_1$ with an error $\epsilon$ that vanishes as density increases.

• Part II: Probabilistic Verification (Poisson Sprinkling is Well-Conditioned)
  The authors demonstrate that a Poisson sprinkling at high density $\rho$ satisfies the well-conditioned definition with high probability.

  1. Order and Volume: Order preservation (F1) is intrinsic to the construction. Volume-faithfulness (F2) follows from Chernoff concentration bounds applied to the counts of points in causal diamonds, extended to a mesoscopic net via a union bound (Lemma 5.2).
  2. Longest-Chain Correspondence: The correspondence between the length of the longest causal chain and proper time (F3) is established by adapting the Bollobás–Brightwell theorem for longest chains in partial orders. The authors use a "cone-sandwiching" argument to transfer flat-space estimates to curved spacetimes (Proposition 5.3).
  3. Scaffold Existence: The existence of the anchor scaffold is shown to follow deterministically from the volume condition (F2), requiring no separate probabilistic estimate.

Key Contributions

1. Quantitative Uniqueness: The paper provides the first quantitative proof of the Hauptvermutung for finite causal sets arising from Poisson sprinkling. It establishes that two spacetimes admitting such embeddings are related by an $\epsilon$-approximate isometry, with an explicit error bound $\epsilon(\rho, \lambda)$.
2. Well-Conditioned Embeddings: The definition of a well-conditioned embedding isolates the geometric content of the conjecture from its probabilistic content, separating the deterministic geometric rigidity from the statistical properties of the sprinkling.
3. Moving Karcher Mean Construction: The paper introduces a smooth, global map construction based on the Karcher mean of point correspondences, which serves as a robust interpolation method for noisy point sets on manifolds.
4. Role of Volume vs. Order: The analysis clarifies the structural roles of order and volume. Contrary to the traditional "order fixes conformal structure, volume fixes scale" dichotomy, this work shows that in the presence of longest-chain preservation, the order (via the chain/proper-time correspondence) fixes the scale, while the volume condition (F2) is primarily responsible for ensuring the non-degeneracy and isotropy of the local reconstruction.

Results

• Theorem 4.18: If a finite causal set $C$ admits well-conditioned embeddings into two globally hyperbolic spacetimes $(M_1, g_1)$ and $(M_2, g_2)$, there exists a smooth diffeomorphism $\Phi$ between their deep interiors such that $\|\Phi^* g_2 - g_1\| = O(\rho^{-2/(5d)} \lambda^{-2/5} \log^{3/2}(\rho V))$.
• Corollary 5.6: A single causal set presented as a Poisson sprinkling at common density $\rho$ into two spacetimes cannot be a faithful sprinkling of two macroscopically distinct spacetimes. In the high-density limit ($\rho \to \infty$), the two geometries are forced to agree up to the explicit error bound, with the probability of failure decaying as a power of the density.

Significance and Claims
The paper claims to resolve the "finite-set" formulation of the Hauptvermutung in the physically relevant setting of Poisson sprinkling. It explicitly addresses the counter-examples provided by Müller [13] by showing that while order alone is insufficient, the combination of order, volume (via Poisson statistics), and the asymptotic high-density limit is sufficient to recover the geometry uniquely.

The authors emphasize that their result is a quantitative statement about the convergence of discrete data to continuum geometry. They do not claim to solve the problem for arbitrary finite posets or for embeddings that do not satisfy the "well-conditioned" criteria. The significance lies in providing a rigorous, deterministic geometric framework that validates the causal set program's core assumption: that spacetime geometry emerges uniquely from the causal order and volume information of a discrete structure in the continuum limit. The error bounds are explicit, showing that the approximation improves as the density increases, thereby bridging the gap between the discrete causal set and the continuous Lorentzian manifold.