Channel Irreducibility and the Discrete Handle

provisional

Overview

This lemma supplies the load-bearing structural result for the non-AdS extension of the ER=EPR derivation: Theorem 3.2 Step 2 establishes that the coherence channel $\gamma_{12}$ is irreducibly connected and non-pinching, and the lemma extends this to the conclusion that the ambient Cauchy slice acquires a non-contractible 1-cycle (a spatial handle). The framework’s irreducibility property forces the handle reading, ruling out the otherwise-compatible flux-tube interpretation in flat $\mathbb{R}^3$.

The argument has four logical steps:

1. Define ambient embedding of a sub-causal-set: every element of the sub-causal-set has a causal-set neighbor outside the sub-causal-set within a Planck-scale region.

2. Embedding implies decoherence. If the channel $\gamma_{12}$ is ambient-embedded, then ambient causal-set elements adjacent to channel elements participate in relational invariants with them, siphoning coherence from $I_{12}$ via Coherence Conservation. This is the framework-internal definition of decoherence.

3. Irreducibility implies non-embedding. Strict irreducibility (ER=EPR Proposition 1.2c, derived from Relational Invariants Theorem 4.1) requires $\mathcal{C}(I_{12})$ to be fully conserved on every Cauchy slice intersecting $\gamma_{12}$, which is incompatible with any positive coherence drain to ambient. By contraposition of Step 2, $\gamma_{12}$ is not ambient-embedded — its elements are causally insulated from ambient except at the observer-region endpoints.

4. Non-embedding lifts to handle topology in the continuum. The argument splits into existence, exclusion, and the explicit antichain construction the Major–Rideout–Surya machinery requires:

   • (4a) Existence. An explicit globally-hyperbolic wormhole manifold (e.g., the eternal Schwarzschild bridge or a Morris–Thorne traversable wormhole with appropriate matter content) has spatial slices with $H_1 = \mathbb{Z}$ (a handle), and admits sprinklings whose combinatorial structure matches the framework’s (ambient ∪ channel) pattern. By Major–Rideout–Surya, 2007 (Theorem 2 and Corollary 2; see Causal Set Statistics), the thickened-antichain construction recovers the wormhole’s spatial homology — including $H_1$ — from this causal set.
   • (4b) Exclusion. A faithful embedding of (ambient ∪ channel) into flat $\mathbb{R}^3 \times \mathbb{R}$ requires the channel elements to participate in the Poisson neighbor statistics of flat-space sprinkling everywhere along the channel’s length. The non-embedding result of Step 3 forbids this. Therefore flat space is excluded as a continuum approximation of (ambient ∪ channel).
   • (4c) Spanning antichain construction (Theorem 4.7). The Major–Rideout–Surya construction invoked in (4a) requires an inextendible antichain spanning both regions, with cardinality $N_{\min}$ on the channel side. The smooth-Cauchy-hypersurface theorem of Bernal–Sánchez, 2003 supplies this: extend the bottleneck 2-surface to a Cauchy hypersurface of $(M_W, g_W)$, sprinkle it via the standard thickened-antichain construction, and the result has all three required properties (maximality, ambient-restriction inextendibility, channel restriction = bottleneck cut). The high-probability nature of the construction is the standard CST caveat from Major–Rideout–Surya, 2007, not a derivation-specific gap.

Together, Steps 4a, 4b, and 4c establish: the handle interpretation is consistent (4a + 4c), the flux-tube interpretation is excluded (4b). This is the content needed for the ER=EPR argument. Hauptvermutung uniqueness (which remains conjectural in causal-set theory) would upgrade this to “the handle is the unique continuum class,” but the argument does not require the upgrade.

Structure of the argument. Steps 1–3 are framework-internal logic with full proofs below. Steps 4a (existence) and 4c (spanning antichain) cite published results, with the framework-internal portion being the consistency of the embedding and the bottleneck saturation. Step 4b (exclusion) follows from Step 3 plus the definition of faithful embedding. Section 5 establishes the quantitative matching of channel element density to wormhole-bridge sprinkling density: the integer cardinality $N_{\min} = \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$ (sourced by the level-stratified quantization of Coherence Conservation Corollary 5.5a) matches the Poisson sprinkling at the throat of an explicit Morris–Thorne wormhole with area $4\ell_P^2 N_{\min}$. The only open item is Gap 1 (Hauptvermutung uniqueness), a wider-field limitation of causal-set theory rather than a framework-internal gap.

Note on novelty. The chain of reasoning “operationally-defined irreducibility of a sub-causal-set ⇒ non-embedding in ambient ⇒ handle topology in the continuum sprinkling limit” appears to be new. To the best of this derivation’s literature scout (2026-05-27), no published CST result establishes this implication. The framework’s contribution is the irreducibility-to-non-embedding link (Steps 2–3); the non-embedding-to-handle link (Step 4) is novel but constructible from existing CST machinery (Major–Rideout–Surya + topology-change literature).

Statement

Lemma (Channel Handle). Let $\mathcal{O}_1, \mathcal{O}_2$ be two spatially separated observers sharing a coherence channel $\gamma_{12} \subset \mathcal{C}$ generated by a relational invariant $I_{12}$ (ER=EPR Definition 1.1). Assume:

(i) Strict irreducibility: $\gamma_{12}$ satisfies ER=EPR Proposition 1.2c — it cannot be decomposed into channels via any intermediary observer.

(ii) Manifold-like ambient: The ambient causal set $\mathcal{C} \setminus \gamma_{12}$ faithfully embeds into a globally hyperbolic spacetime $(M_{\text{amb}}, g_{\text{amb}})$ at sprinkling density $V_c^{-1}$ satisfying the scale separation $V_c \ll v_0 \ll v \ll \tilde{v}$ of Major–Rideout–Surya, 2007.

Then the combined causal set $\mathcal{C} = (\mathcal{C} \setminus \gamma_{12}) \cup \gamma_{12}$ has continuum approximation $(M, g)$ whose Cauchy slice $\Sigma$ satisfies $H_1(\Sigma; \mathbb{Z}) \neq 0$. In particular, there exists a non-contractible loop on $\Sigma$ that traverses the channel in one direction and ambient space in the other.

Furthermore, $(M, g)$ does not faithfully embed into $(M_{\text{amb}}, g_{\text{amb}})$ at the same density. The flux-tube reading — $\gamma_{12}$ as a localized coherence density region within the ambient manifold — is excluded.

1. Channel-Ambient Embedding

Definition 1.1 (Causal-set neighbor). Two distinct elements $x, y \in \mathcal{C}$ are causally adjacent if they are causally related ($x \prec y$ or $y \prec x$) and the Alexandrov interval between them has spacetime volume $\leq V_c$, the discreteness scale.

Remark. This is the standard CST definition of a “link” — an irreducible causal relation not factored through any intermediate element. Equivalently, $x$ and $y$ are causally adjacent iff there is no $z$ with $x \prec z \prec y$ (or $y \prec z \prec x$).

Definition 1.2 (Ambient embedding of a sub-causal-set). A sub-causal-set $X \subset \mathcal{C}$ is ambient-embedded if every element $x \in X$ has at least one causal-set neighbor in $\mathcal{C} \setminus X$ that lies in the Planck-ball causal-future or causal-past of $x$ (i.e., in a region of Alexandrov volume $\leq V_c$ from $x$, in either causal direction).

Equivalently: $X$ is ambient-embedded iff the boundary set $\partial X \equiv \{(x, y) : x \in X, y \in \mathcal{C} \setminus X, x \text{ and } y \text{ are causally adjacent}\}$ projects onto every element of $X$.

Definition 1.3 (Endpoint regions). The observer neighborhoods $\mathcal{N}_1, \mathcal{N}_2$ are the causal-set elements corresponding (in the continuum approximation) to small open regions of $M_{\text{amb}}$ around the two observers’ worldlines, of spatial extent much larger than the discreteness scale.

Definition 1.4 (Endpoint-only embedding). A channel $\gamma_{12}$ is endpoint-only embedded if all elements of $\partial \gamma_{12}$ (boundary pairs from Definition 1.2) project from $\gamma_{12}$ into $\mathcal{N}_1 \cup \mathcal{N}_2$ — i.e., the only ambient neighbors of channel elements are within the two observer regions.

Proposition 1.5 (Mutually exclusive). A channel $\gamma_{12}$ that is endpoint-only embedded is not ambient-embedded (in the sense of Definition 1.2), since elements of $\gamma_{12}$ in the channel interior have no ambient neighbors at all (their only neighbors are other channel elements). Conversely, a channel that is ambient-embedded has interior elements with ambient neighbors, hence is not endpoint-only embedded.

Proof. Immediate from Definitions 1.2 and 1.4: ambient embedding requires every element of the channel to have ambient neighbors; endpoint-only embedding requires no interior element to have ambient neighbors. These are negations restricted to the channel interior. $\square$

2. Embedding Implies Decoherence

Proposition 2.1 (Adjacent ambient elements form relational invariants). Let $x \in \gamma_{12}$ be an interior channel element (not in $\mathcal{N}_1 \cup \mathcal{N}_2$), and let $y \in \mathcal{C} \setminus \gamma_{12}$ be causally adjacent to $x$ (in the sense of Definition 1.1). Then there exists a relational invariant $I_{x,y}$ between the local degrees of freedom at $x$ and the local degrees of freedom at $y$, with positive coherence content $\mathcal{C}(I_{x,y}) > 0$.

Proof. Causal adjacency means $x$ and $y$ are connected by a causal-set link — a direct causal relation not factored through any intermediary. By construction $x \in \gamma_{12}$ belongs to the channel-carrying observer system (the joint structure of $\mathcal{O}_1, \mathcal{O}_2$ with the relational invariant $I_{12}$), and $y \in \mathcal{C} \setminus \gamma_{12}$ belongs to a distinct ambient observer $\mathcal{O}_y$. The link $x \prec y$ therefore constitutes an observer-observer interaction event between $\mathcal{O}_y$ and the channel-carrying side.

By Three Interaction Types Theorem 5.1 (derived), every such observer-observer interaction event is exhaustively classified into exactly one of Type I (Passage, phase exchange), Type II (Fusion, composite formation), Type III (Resonance, new relational invariant), or dissolution (Case D of three-types Step 2).

Dissolution at $(x, y)$ is precluded by the lemma’s persistence hypothesis on $\gamma_{12}$. The strict-irreducibility hypothesis requires $\gamma_{12}$ to be a coherent persistent channel carrying $I_{12}$ undiminished through every Cauchy slice (ER=EPR Proposition 1.2c). Case D1 (mutual annihilation; Three Interaction Types Proposition 2.1, sub-case D1) would remove $x$ from $\gamma_{12}$, contradicting persistence. Case D2 (reorganization; same Proposition, sub-case D2) produces new observers whose new invariants are “indistinguishable from Case A” by three-types’ own analysis — i.e., they reduce to one of Types I, II, or III. The direct possibilities are therefore Type I, II, or III.

Type III generates the relational invariant directly. If the interaction at $(x, y)$ is Type III, the relational invariant $I_{x, y}$ between the local degrees of freedom at $x$ and $y$ is generated by definition (Three Interaction Types, Definition 4.4) with $\mathcal{C}(I_{x, y}) > 0$ (Relational Invariants, Definition 2.1).

Type I and Type II reduce to a Type III paired ledger. Three Interaction Types Step 6, Remark on “Type I as currency, Type III as accounting,” establishes the structural identification: most physically realized Type III correlations are produced through Type I-mediated traffic, and the paired ledger entry on either side of a Type I transfer IS a Type III relational invariant on the joint state space of the two observers post-transfer. Applied to the link $(x, y)$: a Type I event leaves a Type III invariant $I_{x, y}$ at the channel-ambient interface, indexed by the link itself or its immediate causal successor. For Type II, the binding-coherence accounting (three-types Proposition 7.3, reverse direction) requires a coherence source for any composite formed at $(x, y)$; the source can only be $\mathcal{C}(I_{12})$ (the channel’s coherence) feeding into the binding, which is recorded as a Type III binding-ledger entry on the joint state space. In either case, a Type III relational invariant $I_{x, y}$ exists with $\mathcal{C}(I_{x, y}) > 0$. $\square$

Remark 2.2. The proof above rests on Three Interaction Types Theorem 5.1, which is at full derived rigor (see that derivation’s Rigor Assessment) and explicitly establishes the exhaustive four-case classification (Types I, II, III, or dissolution) for every observer-observer interaction event. The mapping from causal-set links between distinct observers to such interaction events is the standard CST reading of links as elementary interaction loci. The framework does not admit “ghost” causal links between distinct observers with no associated interaction type — three-types Theorem 5.1’s exhaustiveness (its Step 5) excludes this by construction. The reduction of Type I and Type II to a Type III paired ledger at the next causal step uses three-types Step 6’s currency-vs-accounting structural identification, which is also at full derived rigor.

Proposition 2.3 (Boundary invariants drain channel coherence). If $\gamma_{12}$ is ambient-embedded, then the family of boundary relational invariants $\{I_{x, y(x)} : x \in \gamma_{12} \text{ interior}, y(x) \in \mathcal{C} \setminus \gamma_{12} \text{ adjacent}\}$ contributes strictly positive total coherence $\Delta \mathcal{C} > 0$ that must be subtracted from $\mathcal{C}(I_{12})$ in any Cauchy-slice accounting.

Proof. By Coherence Conservation (Axiom 1), the total coherence on any Cauchy slice $\Sigma$ intersecting both $\gamma_{12}$ and the boundary region equals the sum of contributions from all relational invariants whose carriers pass through $\Sigma$. The invariants $\{I_{x, y(x)}\}$ from Proposition 2.1 have carriers that pass through every Cauchy slice intersecting the channel interior (since the boundary pairs are local to the channel-ambient interface). Each such invariant contributes $\mathcal{C}(I_{x, y(x)}) > 0$ to the total. By subadditivity of the coherence measure (Coherence Conservation, C1), the contribution of $I_{12}$ alone is bounded above:

$\mathcal{C}(I_{12}) \leq \mathcal{C}_{\text{total}}(\Sigma) - \sum_{x \in \gamma_{12}^{\text{int}}} \mathcal{C}(I_{x, y(x)}) = \mathcal{C}_{\text{total}}(\Sigma) - \Delta \mathcal{C}.$

In particular, if $\Delta \mathcal{C} > 0$, then $\mathcal{C}(I_{12}) < \mathcal{C}_{\text{total}}(\Sigma)$ — strictly less than the full Cauchy-slice coherence flowing through the channel region. This is the framework-internal definition of decoherence of $I_{12}$: a positive fraction of its potential coherence content is siphoned into ambient-channel invariants and is no longer available to $I_{12}$ for observer-pair correlation purposes. See Substrate Noise and Profile Coupling for the related notion of profile-mediated coherence loss; the present argument is a structural counterpart at the channel-ambient interface. $\square$

Corollary 2.4 (Embedding ⇒ partial decoherence). Any ambient-embedded channel is decohered to a degree proportional to the number of interior channel elements with ambient neighbors.

3. Irreducibility Implies Non-Embedding

Theorem 3.1 (Strict irreducibility forbids ambient embedding). Let $\gamma_{12}$ satisfy the strict irreducibility hypothesis of ER=EPR Proposition 1.2c — namely, $\mathcal{C}(\gamma_{12}) = \mathcal{C}(I_{12})$ is conserved on every Cauchy slice that intersects $\gamma_{12}$, and $I_{12}$ does not decompose through any intermediary observer. Then $\gamma_{12}$ is endpoint-only embedded (Definition 1.4) — its interior elements have no ambient causal-set neighbors.

Proof. By contradiction. Suppose $\gamma_{12}$ has at least one interior element $x \in \gamma_{12}^{\text{int}}$ with an ambient neighbor $y \in \mathcal{C} \setminus \gamma_{12}$. By Proposition 2.1, there exists a relational invariant $I_{x,y}$ with $\mathcal{C}(I_{x,y}) > 0$. By Proposition 2.3, this contributes $\Delta \mathcal{C} \geq \mathcal{C}(I_{x,y}) > 0$ to the boundary-invariant total, draining $\mathcal{C}(I_{12})$ below the full Cauchy-slice coherence: $\mathcal{C}(I_{12}) < \mathcal{C}_{\text{total}}(\Sigma)$.

But the strict irreducibility hypothesis requires $\mathcal{C}(\gamma_{12}) = \mathcal{C}(I_{12})$ to be the full conserved coherence content of the channel, with no positive coherence routed elsewhere on the same slice. The ambient-channel invariant $I_{x,y}$ provides exactly such a positive coherence routed elsewhere — specifically, into ambient. Contradiction.

Therefore $\gamma_{12}$ has no interior element with an ambient neighbor. By Proposition 1.5, $\gamma_{12}$ is endpoint-only embedded. $\square$

Remark 3.2 (Operational picture). Theorem 3.1 captures the framework’s existing position that perfect entanglement is incompatible with decoherence. In the standard quantum-mechanics picture, a Bell pair perfectly isolated from the environment retains its full entanglement entropy; once the pair interacts with environment degrees of freedom, the entanglement is dispersed across the system + environment, leaving the pair in a mixed state with reduced bipartite entanglement. The framework’s channel formulation makes this precise: full irreducibility of $I_{12}$ ⇔ no ambient interaction ⇔ endpoint-only embedding of $\gamma_{12}$. Real Bell pairs in practice are partially decohered, which corresponds to partial ambient embedding; the framework’s ER=EPR claim refers to the idealized perfect-isolation limit, which is also the standard Maldacena–Susskind setting.

Corollary 3.3 (Partial irreducibility ⇒ partial handle). A channel with partial decoherence has an intermediate topological character between perfect handle (perfectly insulated) and pure flux tube (fully ambient-embedded). This provides a continuous interpolation between the two pictures, with the “handle fraction” measurable in principle as the channel’s entanglement preservation fidelity. This corollary will not be used in the main argument; it is noted as a framework-distinctive structural prediction.

4. Non-Embedding Implies Handle Topology

This section establishes Step 4 of the program: that an endpoint-only-embedded channel forces a spatial handle in the continuum sprinkling limit. The argument has two parts: existence of a consistent handle interpretation (Section 4.1) and exclusion of the flat-space (flux-tube) interpretation (Section 4.2).

4.1 Handle Existence (Constructive)

Proposition 4.1 (Wormhole manifold construction). There exists a globally hyperbolic Lorentzian manifold $(M_W, g_W)$ with the following properties:

(a) Asymptotic structure: $(M_W, g_W)$ contains a region asymptotically isometric to two disjoint copies of flat $\mathbb{R}^3 \times \mathbb{R}$, each containing one of the observer worldlines.

(b) Spatial handle: Cauchy slices of $(M_W, g_W)$ have spatial topology $\mathbb{R}^3 \# (\mathbb{R} \times S^2)$, the connected sum of flat space with the “handle” $\mathbb{R} \times S^2$. The first homology is $H_1(\Sigma; \mathbb{Z}) = \mathbb{Z}$, generated by a loop traversing the wormhole throat in one direction and ambient space in the other.

(c) Global hyperbolicity: $(M_W, g_W)$ is globally hyperbolic in the bifurcate-horizon sense (admits a global Cauchy surface).

Existence proof. Two canonical examples:

• Eternal Schwarzschild–Kruskal extension. The maximal extension of Schwarzschild has spatial slices $\mathbb{R} \times S^2$ (with one $S^2$ pinch at the bifurcation surface). This satisfies (a) with the two asymptotic regions joined through the Einstein–Rosen bridge, (b) with the handle being the bridge, and (c) by standard arguments (the Kruskal extension is globally hyperbolic; see Hawking–Ellis, 1973). However, the Schwarzschild throat is non-traversable and contains a horizon — not all framework applications require this.

• Morris–Thorne traversable wormhole (Morris–Thorne, 1988). For appropriate (exotic-matter-violating-the-NEC) stress-energy, this is a globally hyperbolic spacetime with two asymptotically flat regions connected by a throat of finite area. Spatial slices have the connected-sum topology required by (b). $\square$

Proposition 4.2 (Wormhole sprinkling matches the framework’s combinatorial pattern). A Poisson sprinkling of $(M_W, g_W)$ at density $V_c^{-1}$ produces a causal set $\mathcal{C}_W$ that decomposes into:

• An ambient sprinkling of the two asymptotic regions (faithful embedding into the two disjoint $\mathbb{R}^3 \times \mathbb{R}$ copies), and
• A throat sub-causal-set $\gamma_W$ of elements sprinkled into the wormhole bridge region.

The throat sub-causal-set $\gamma_W$ is endpoint-only embedded in the sense of Definition 1.4: its interior elements (those sprinkled into the bridge interior, not its mouths) have causal-set neighbors only with other throat elements, not with elements sprinkled into the asymptotic regions. This is because the bridge interior is spatially separated from both asymptotic regions by the topology of the manifold itself — there is no Planck-scale region of $M_W$ that contains both a bridge-interior point and an asymptotic-region point.

Proof. The Poisson sprinkling of $(M_W, g_W)$ assigns each spacetime point an independent probability of being a causal-set element, proportional to spacetime volume. Causal adjacency in the resulting causal set $\mathcal{C}_W$ is determined by the manifold’s causal structure: $x$ and $y$ are adjacent iff they are causally related in $(M_W, g_W)$ and lie within Alexandrov volume $\leq V_c$ of each other.

By the topology of $M_W$ (connected sum, not simply-connected), the bridge interior and the asymptotic regions are spatially separated: a Planck-ball at any bridge-interior point is contained entirely within the bridge region (does not extend into the asymptotic regions), and vice versa. The only points of the manifold where bridge and asymptotic regions meet are the mouth regions $\mathcal{N}_1, \mathcal{N}_2$.

Therefore, an element $x$ sprinkled into the bridge interior has Planck-ball causal neighbors only at other bridge interior points or at mouth points. By Definition 1.4, $\gamma_W$ is endpoint-only embedded. $\square$

Theorem 4.3 (Handle existence for the framework’s channel). Let $\gamma_{12}$ be the framework’s strictly irreducible channel (satisfying Theorem 3.1, hence endpoint-only embedded). There exists a globally hyperbolic wormhole manifold $(M_W, g_W)$ (Proposition 4.1) such that the framework’s combined causal set $\mathcal{C} = (\mathcal{C} \setminus \gamma_{12}) \cup \gamma_{12}$ is consistent with a faithful embedding into $(M_W, g_W)$: the ambient portion embeds into the asymptotic regions, and the channel $\gamma_{12}$ embeds into the bridge.

By Major–Rideout–Surya, 2007 Theorem 2 and Corollary 2 (also reviewed in Causal Set Statistics), the thickened-antichain construction $\mathsf{T}_n(A)$ on $\mathcal{C}$ recovers the spatial homology of $(M_W, g_W)$ with high probability under the scale separation $V_c \ll v_0 \ll v \ll \tilde{v}$. In particular, $H_1(\Sigma; \mathbb{Z}) = \mathbb{Z}$ is recoverable: there is a non-contractible 1-cycle in the antichain nerve that corresponds to the manifold’s handle.

Proof. The consistency of the embedding follows directly from Propositions 4.1 and 4.2: the manifold exists, its sprinkling has the right combinatorial structure, and the framework’s channel maps to the bridge sub-causal-set. The homology recovery is the content of the cited Major–Rideout–Surya theorem, whose hypotheses (globally hyperbolic, scale separation, inextendible antichain through both ambient and channel regions) are all satisfied by the present setup. $\square$

4.2 Flat-Space Exclusion

Theorem 4.4 (No faithful flat-space embedding). The framework’s causal set $\mathcal{C} = (\mathcal{C} \setminus \gamma_{12}) \cup \gamma_{12}$, with $\gamma_{12}$ strictly irreducible (hence endpoint-only embedded by Theorem 3.1), does not faithfully embed into flat Minkowski space $\mathbb{R}^3 \times \mathbb{R}$ at the ambient sprinkling density.

Proof. A faithful embedding $\Phi: \mathcal{C} \to (M_{\text{flat}}, \eta)$ at density $V_c^{-1}$ requires:

(i) The number of elements of $\mathcal{C}$ in any spacetime region of volume $V$ is Poisson-distributed with mean $V / V_c$.

(ii) Order relations in $\mathcal{C}$ correspond to causal relations in $(M_{\text{flat}}, \eta)$.

Consider an interior channel element $x \in \gamma_{12}^{\text{int}}$. Under the hypothesized embedding $\Phi$, $\Phi(x)$ is a point in $\mathbb{R}^3 \times \mathbb{R}$. Consider a small spacetime ball around $\Phi(x)$ of volume $V$ with $V_c \ll V \ll v_0$ (Planck-scale region). By the Poisson condition (i), this ball contains on average $V / V_c \gg 1$ elements of $\mathcal{C}$.

But by Theorem 3.1, $x$ has no ambient causal-set neighbors — its only causal-set neighbors are other elements of $\gamma_{12}$. The local density of $\gamma_{12}$ elements near $x$ is at most the “cross-section count” $N_{\min} = S_{\text{ent}}(I_{12})$ per cross-section (ER=EPR Section 3.4 discrete-throat picture). For a microscopic Bell pair, $S_{\text{ent}} = \ln 2$, and the local channel density is at most $\sim \ln 2$ elements per Planck cross-section.

The ambient density required by (i) is $V / V_c \gg \ln 2$ for any region above the discreteness scale. Therefore the local element count near $\Phi(x)$ is far below the Poisson mean required by (i). This violates faithful embedding.

Concretely: the channel has $O(1)$ elements per Planck cross-section (set by the irreducible coherence count); flat-space sprinkling at density $V_c^{-1}$ has $V / V_c$ elements in any volume-$V$ region. For $V \gg V_c$, the deficit is enormous (factor of $V / V_c$). This is the “void cutting through” the causal set that the Major–Rideout–Surya, 2007 paper (page 19, $P_0 = \exp(-V/V_c)$) identifies as the principal obstacle to faithful embedding for under-stuffed regions.

Therefore $\Phi$ is not faithful, and no faithful embedding into flat space exists. $\square$

Remark 4.5 (The geometric meaning of the exclusion). Flat-space sprinkling fills space densely with causal-set elements; the framework’s channel is a sparse, irreducible thread of $O(1)$ elements per cross-section. The two are not topologically compatible at the discreteness scale, even though they might appear compatible at large scales (where a thin tube of high-coherence-density region in flat space is a legitimate continuum object). The framework’s irreducibility property is what makes the channel sparse-not-dense — it fixes the channel cross-section to the minimum coherence count $S_{\text{ent}}$, not to the ambient sprinkling density.

4.3 Combined Conclusion

Corollary 4.6 (Handle topology forced). Combining Theorems 4.3 (handle existence) and 4.4 (flat-space exclusion):

• The handle interpretation is consistent: there exists a wormhole manifold whose sprinkling matches the framework’s combinatorial structure, with $H_1 = \mathbb{Z}$ recoverable by the Major–Rideout–Surya construction.

• The flux-tube interpretation is excluded: the framework’s causal set cannot faithfully embed into flat space.

The conjunction rules out flux-tube and exhibits a consistent handle interpretation. This is the content needed for ER=EPR Theorem 3.2 Step 2: the ambient Cauchy slice acquires a non-contractible 1-cycle as a result of the channel’s irreducibility.

Hauptvermutung note. If the Hauptvermutung of causal-set theory (uniqueness of continuum approximation) is proved, Corollary 4.6 upgrades from “handle is consistent, flat space is excluded” to “handle is the unique continuum class compatible with the framework’s combinatorial structure.” The framework’s ER=EPR claim does not require this upgrade — uniqueness is a stronger result than needed.

4.4 Spanning Inextendible Antichain

Theorem 4.3 invoked the Major–Rideout–Surya construction, which requires an inextendible antichain $A$ spanning both the ambient region and the channel, with $A \cap \gamma_{12}$ realising the bottleneck cross-section. This subsection constructs such an antichain explicitly, completing the formal infrastructure on which Theorem 4.3 implicitly relied.

Theorem 4.7 (Spanning inextendible antichain). Under the hypotheses of the Lemma (strict irreducibility of $\gamma_{12}$, manifold-like ambient with the scale separation $V_c \ll v_0 \ll v \ll \tilde{v}$), there exists an antichain $A \subset \mathcal{C}$ with all three of the following properties:

(a) Maximality in $\mathcal{C}$: no element of $\mathcal{C} \setminus A$ can be added to $A$ while preserving mutual incomparability — equivalently, the partition $\mathcal{C} = \mathrm{Past}(A) \cup A \cup \mathrm{Fut}(A)$ is exhaustive.

(b) Ambient restriction: $A_{\text{amb}} := A \cap (\mathcal{C} \setminus \gamma_{12})$ is an inextendible antichain in the ambient sub-causal-set $\mathcal{C} \setminus \gamma_{12}$.

(c) Channel restriction: $A_{\text{chan}} := A \cap \gamma_{12}$ is the minimal antichain cut $\Sigma_{\min}^{\mathcal{C}}$ of the channel graph, with $|A_{\text{chan}}| = N_{\min} = S_{\text{ent}}(I_{12})$.

Proof. By Proposition 4.1, the combined causal set $\mathcal{C}$ is consistent with a faithful embedding $\Phi: \mathcal{C} \to (M_W, g_W)$ into a globally hyperbolic wormhole manifold whose Cauchy slices have spatial topology $\mathbb{R}^3 \, \# \, (\mathbb{R} \times S^2)$. The proof constructs $A$ as the sprinkling of a particular Cauchy hypersurface of $M_W$.

Step 1: A Cauchy hypersurface containing the bottleneck exists. The throat of $M_W$ has a minimum-area 2-surface $\Sigma_{\min} \subset M_W$ — the wormhole bottleneck (ER=EPR Theorem 3.2 Step 3, extreme value theorem on the throat cross-sections). $\Sigma_{\min}$ is a compact spacelike 2-submanifold of $M_W$ (hence achronal).

By the smooth-splitting refinement of Geroch’s theorem (Bernal–Sánchez, 2003), every globally hyperbolic spacetime admits a smooth foliation by spacelike Cauchy hypersurfaces, parameterised by a smooth time function $t: M_W \to \mathbb{R}$ with $\{t = \tau\}$ Cauchy for each $\tau$. Since $\Sigma_{\min}$ is a compact achronal spacelike submanifold, one can choose the smooth time function $t$ so that $\Sigma_{\min} \subset \{t = 0\}$. Concrete constructions are immediate for the two canonical wormhole geometries used in Proposition 4.1:

• Eternal Schwarzschild–Kruskal. The bifurcate 2-sphere at $T = X = 0$ in Kruskal coordinates lies on the $T = 0$ spatial slice, which is a Cauchy hypersurface with topology $\mathbb{R} \times S^2$.
• Symmetric Morris–Thorne. The throat 2-sphere at proper-radial-distance $\ell = 0$ lies on any constant-$t$ slice of the static metric; each such slice is a Cauchy hypersurface with topology $\mathbb{R}^3 \, \# \, (\mathbb{R} \times S^2)$.

In either case, denote the chosen Cauchy hypersurface $\Sigma_W \subset M_W$, with $\Sigma_{\min} \subset \Sigma_W$. For asymmetric wormhole geometries (e.g., dynamical throats), the Bernal–Sánchez existence statement still applies, though the time function may be less geometrically transparent.

Step 2: Sprinkling $\Sigma_W$ gives an antichain. Following Major–Rideout–Surya, 2007 Section 3, define the thickened antichain on $\Sigma_W$:

$A \equiv \{x \in \mathcal{C} \mid \Phi(x) \text{ lies in a } V_c\text{-thin spacetime neighbourhood of } \Sigma_W\}.$

Under the scale separation $V_c \ll v_0 \ll v \ll \tilde{v}$, the set $A$ is an antichain in $\mathcal{C}$ with high probability. The justification is that elements of $A$ have $\Phi$-images on a spacelike slice with $O(V_c)$ thickness, and the probability that two such elements are causally related decreases with the slice thickness — precisely the void-probability argument of Major–Rideout–Surya, 2007 (their $\mathfrak{N} P_1 \ll 1$ in equation 12).

Step 3: $A$ is inextendible in $\mathcal{C}$. $\Sigma_W$ is a Cauchy hypersurface of $(M_W, g_W)$, so by definition every inextendible causal curve in $M_W$ crosses $\Sigma_W$ exactly once. For any $x \in \mathcal{C}$ with $\Phi(x) \notin \Sigma_W$, the continuum causal curve from $\Phi(x)$ to the future or past crosses $\Sigma_W$, so $\Phi(x)$ is in either the causal past or the causal future of $\Sigma_W$. Under the sprinkling correspondence, $x$ is therefore in $\mathrm{Past}(A)$ or $\mathrm{Fut}(A)$. The partition

$\mathcal{C} = \mathrm{Past}(A) \cup A \cup \mathrm{Fut}(A)$

is exhaustive, which is the discrete characterisation of an inextendible antichain. Property (a) is established.

Step 4: Ambient restriction $A_{\text{amb}}$ is inextendible in $\mathcal{C} \setminus \gamma_{12}$. Restricting the partition argument of Step 3 to the ambient sub-causal-set: every $x \in \mathcal{C} \setminus \gamma_{12}$ is either in $A_{\text{amb}}$, or in the ambient causal past of $A_{\text{amb}}$, or in the ambient causal future. Property (b) is established.

Step 5: Channel restriction $A_{\text{chan}}$ is the bottleneck cut with $N_{\min}$ elements. $A_{\text{chan}}$ is the sprinkling of $\Sigma_W \cap M_{\text{throat}}$, the bottleneck portion of the Cauchy hypersurface. By the choice of $\Sigma_W$ to contain $\Sigma_{\min}$, $A_{\text{chan}}$ is exactly the discrete realisation of $\Sigma_{\min}$ in the channel sub-causal-set:

• Every causal-set path in $\gamma_{12}$ from $\mathcal{N}_1$ to $\mathcal{N}_2$ must cross $A_{\text{chan}}$, since the continuum analog (every causal curve from $\mathcal{N}_1$ to $\mathcal{N}_2$ through the throat) must cross $\Sigma_W$.

• $A_{\text{chan}}$ is a minimal cut (no proper subset disconnects $\gamma_{12}$), since $\Sigma_{\min}$ is the minimum-area cross-section of the throat (ER=EPR Theorem 3.2 Step 3) and any sub-collection of its sprinkled elements would fail to cover the full 2-surface.

• The cardinality $|A_{\text{chan}}| = N_{\min} = S_{\text{ent}}(I_{12})$ by the saturation argument of ER=EPR Proposition 3.3 (area-scaling lower bound is tight under irreducibility) combined with the Section 3.4 discrete-throat picture ($N_{\min}$ elements per Planck cross-section, each contributing $4\ell_P^2$).

Property (c) is established. $\square$

Remark 4.8 (Symmetric vs. asymmetric wormholes). For the symmetric wormhole geometries used in Proposition 4.1’s existence argument (eternal Schwarzschild, symmetric Morris–Thorne), Step 1’s construction is immediate: the time-reflection-symmetric slice $\{t = 0\}$ is automatically a Cauchy hypersurface containing the bottleneck. Asymmetric wormhole geometries (dynamical throats, non-static matter content) require the more general Bernal–Sánchez argument to construct the time function, but the existence statement is unchanged. The lemma’s downstream conclusions are insensitive to which case applies.

Remark 4.9 (Discrete-continuum correspondence at the antichain level). Theorem 4.7’s construction follows the same discrete-continuum correspondence used throughout the lemma: a continuum object (the Cauchy hypersurface $\Sigma_W$) is mapped to a discrete object (the thickened antichain $A$) via Poisson sprinkling, and the discrete object inherits structural properties (here: inextendibility) from the continuum source. The high-probability nature of the correspondence (Step 2’s $\mathfrak{N} P_1 \ll 1$ bound) is the standard CST caveat — the construction works for almost all sprinkling realisations under the scale separation, not deterministically. This is the same caveat that applies to Major–Rideout–Surya, 2007 Theorem 2 in Section 4.1; the present construction does not strengthen or weaken that caveat.

5. Quantitative Channel-Sprinkling Match

Sections 1–4 establish the qualitative content of the channel-handle correspondence — the framework’s irreducible channel embeds into a wormhole-handle topology (Corollary 4.6) and cannot embed into flat space (Theorem 4.4). This section establishes the quantitative match: for every channel coherence content $S_{\text{ent}}$, an explicit wormhole geometry has a Poisson sprinkling that produces exactly $N_{\min}$ elements per Planck cross-section at the bottleneck.

The argument has two components:

1. Integer-cardinality vs real-entropy identification (§5.1). The framework’s identification "$N_{\min} = S_{\text{ent}}$" in ER=EPR Section 3.4 holds in the macroscopic-entanglement limit. At the substrate level, the level-stratification of Coherence Conservation Corollary 5.5a gives the exact form $N_{\min} = \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$: level-$n$ vertex coherence is integer-quantized in $\hbar\omega_0$, while the level-$(n+1)$ relational invariant $I_{12}$ carries the real-valued $S_{\text{ent}}$. The continuum identification is the $S_{\text{ent}} \gg \hbar\omega_0$ limit of the discrete one.

2. Wormhole existence at integer $N_{\min}$ (§5.2). For every positive integer $N_{\min}$, an explicit Morris–Thorne wormhole geometry has throat area $4\ell_P^2 N_{\min}$, and its Poisson sprinkling at the standard CST density produces an antichain bottleneck of expected cardinality $N_{\min}$ with $O(\sqrt{N_{\min}})$ fluctuations.

5.1 Integer cardinality from level-stratified quantization

Proposition 5.1 (Discrete cardinality identification). Let $\gamma_{12}$ be a strictly irreducible channel carrying a relational invariant $I_{12}$ with coherence content $\mathcal{C}(I_{12}) = S_{\text{ent}}$. The cardinality $N_{\min}$ of the minimal antichain cut $\Sigma_{\min}^{\mathcal{C}}$ at the bottleneck satisfies

$N_{\min} = \lceil S_{\text{ent}} / \hbar\omega_0 \rceil$

(equivalently, $N_{\min} = \lceil S_{\text{ent}} \rceil$ in units where $\hbar\omega_0 = 1$).

Proof. The cardinality bound has two parts, each anchored in an existing framework result.

Upper-bound side: capacity per element. Each element of the antichain cut is a level-$n$ substrate vertex. The proof of Coherence Conservation Corollary 5.5a establishes that at level $n$, every vertex contributes a non-negative integer multiple of $\hbar\omega_0$ to the level-$n$ slice total (via Bootstrap Corollary 2.3: each participating observer carries an integer-quantized coherence content). The minimum non-zero contribution from a single vertex is $1\,\hbar\omega_0$. So the level-$n$ substrate capacity at the cut is bounded by

$\text{Capacity}_{\text{level }n}(A_{\text{chan}}) \leq |A_{\text{chan}}| \cdot 1\,\hbar\omega_0.$

(The vertex capacity is exactly $1\,\hbar\omega_0$ at the smallest non-trivial fill; greater per-vertex contributions correspond to multi-quantum vertices, which are decomposable into more fundamental vertices contradicting the minimal-cut property.)

Lower-bound side: capacity demand. The relational invariant $I_{12}$ is a level-$(n+1)$ object whose coherence content $S_{\text{ent}}$ is sourced by the level-$n$ substrate at the cut (since every coherence path from $\mathcal{N}_1$ to $\mathcal{N}_2$ traverses an element of $A_{\text{chan}}$, by Theorem 4.7 property (c)). The level-$n$ capacity must therefore meet the level-$(n+1)$ demand:

$|A_{\text{chan}}| \cdot 1\,\hbar\omega_0 \geq S_{\text{ent}}.$

The smallest integer satisfying this is $\lceil S_{\text{ent}}/\hbar\omega_0 \rceil$.

Saturation by irreducibility. If $|A_{\text{chan}}| > \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$, the cut has excess level-$n$ capacity beyond what $I_{12}$ demands. The cut could then be partitioned into two disjoint sub-cuts of cardinalities $\lceil S_{\text{ent}}/\hbar\omega_0 \rceil$ and $|A_{\text{chan}}| - \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$, each carrying its own sub-channel — contradicting strict irreducibility (ER=EPR Proposition 1.2c, which forbids any decomposition of $\gamma_{12}$ through intermediary observers). Therefore the bound is saturated: $|A_{\text{chan}}| = N_{\min} = \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$. $\square$

Remark 5.2 (Continuum reading is the coarse-grained limit). ER=EPR Section 3.4’s continuum identification $N_{\min} = S_{\text{ent}}$ is the coarse-grained limit of Proposition 5.1 for $S_{\text{ent}} \gg \hbar\omega_0$ (where $\lceil S_{\text{ent}}/\hbar\omega_0 \rceil / (S_{\text{ent}}/\hbar\omega_0) \to 1$). For macroscopic entanglement the two readings agree to relative precision $O(\hbar\omega_0/S_{\text{ent}})$. For a single Bell pair ($S_{\text{ent}} = \ln 2$ in units of $\hbar\omega_0$), the discrete reading gives $N_{\min} = 1$ exactly, while the continuum reading assigns a fractional cardinality — a Planck-resolution discrepancy intrinsic to the framework’s discrete substrate, consistent with Area Scaling’s tile-counting interpretation (Theorem 5.2 Remark: “There is no continuous degree of freedom in the bound itself — it is a discrete count of indistinguishable units”).

Remark 5.3 (Area formula: discrete vs continuum). The continuum area formula $A(\Sigma_{\min}) = 4\ell_P^2 S_{\text{ent}}$ (ER=EPR Proposition 3.3) is the coarse-grained limit of the discrete area $A_{\text{discrete}} = 4\ell_P^2 N_{\min}$ established here. For the Bell pair, $A_{\text{discrete}} = 4\ell_P^2$ (one tile-equivalent), while $A_{\text{continuum}} = 4\ell_P^2 \ln 2 \approx 2.77\ell_P^2$. The discrepancy $\approx 1.23\,\ell_P^2$ is within the framework’s Planck-scale resolution and reflects the integer-quantization of the level-$n$ substrate. The two readings converge in the macroscopic limit.

Remark 5.4 (No conflict with Cor 5.5a). The proof above invokes Cor 5.5a’s level-$n$ integer-quantization clause, not its level-$(n+1)$ Cauchy-slice total. The level-$(n+1)$ slice total is still integer-quantized in $\hbar\omega_0$ when summed over all relational invariants on the slice; individual relational invariants $I_{12}$ may carry irrational coherence ($\ln 2$, etc.), with integer-quantization recovered only after summation. The framework’s existing position is preserved.

5.2 Wormhole existence at integer $N_{\min}$

Theorem 5.5 (Sprinkling realization for each $N_{\min}$). For every positive integer $N_{\min}$, there exists a globally hyperbolic Morris–Thorne wormhole manifold $(M_W^{(N_{\min})}, g_W^{(N_{\min})})$ such that:

(a) Throat area matches discrete count: $A(\Sigma_{\min}^W) = 4\ell_P^2 N_{\min}$.

(b) Sprinkling cardinality matches: Under Poisson sprinkling at the standard CST density $V_c^{-1}$ with $V_c = \ell_P^4$ (in natural units $c = \hbar = 1$), the thickened-antichain construction at the throat (Theorem 4.7) produces an antichain restriction $A_{\text{chan}} = A \cap \gamma_W$ with $|A_{\text{chan}}| = N_{\min}$ in expectation, with Poisson fluctuations of order $\sqrt{N_{\min}}$.

(c) Endpoint-only embedding: The throat sub-causal-set is endpoint-only embedded in the sense of Definition 1.4, matching the framework channel’s structure under the lemma’s irreducibility hypothesis.

Proof. Use the Morris–Thorne metric (Morris–Thorne, 1988)

$ds^2 = -e^{2\Phi(\ell)}\,dt^2 + d\ell^2 + r(\ell)^2\,d\Omega^2,$

with proper-radial coordinate $\ell \in (-\infty, \infty)$, redshift function $\Phi(\ell)$ regular at $\ell = 0$, and shape function $r(\ell)$ satisfying flare-out $r(0) = r_0$, $r'(0) = 0$, $r''(0) \geq 0$, and asymptotic flatness $r(\ell)/|\ell| \to 1$ as $\ell \to \pm\infty$. For each positive integer $N_{\min}$, choose

$r_0 = \ell_P\sqrt{N_{\min}/\pi}, \qquad A(\Sigma_{\min}^W) = 4\pi r_0^2 = 4\ell_P^2 N_{\min},$

establishing (a). Morris–Thorne stress-energy violates the null energy condition at the throat (an exotic-matter requirement irrelevant here — the framework’s coherence-Lagrangian source provides the required negative-energy contribution; see ER=EPR Section 3.2). Global hyperbolicity holds by construction of the symmetric Morris–Thorne family with regular metric components.

For (b), apply the thickened-antichain construction of Major–Rideout–Surya, 2007 to a constant-$t$ Cauchy hypersurface $\Sigma_W$ of $(M_W, g_W)$ containing the throat 2-surface $\Sigma_{\min}^W$. The standard sprinkling cardinality in a $V_c^{1/4}$-thick collar around a 2-surface of area $A$ is $A/\ell_P^2$ raw Planck cells (one per cross-sectional Planck cell, by Poisson density). With $A = 4\ell_P^2 N_{\min}$, the raw cardinality is $4 N_{\min}$.

The lemma’s per-element convention (Section 4.4 Theorem 4.7 Step 5, citing ER=EPR Section 3.4) absorbs the area-scaling 1/4 coefficient into the per-element effective area $4\ell_P^2$: one “element of $A_{\text{chan}}$” corresponds to one coherence tile, not one raw Planck cell. (This identification is forced by Area Scaling Theorem 5.1, which fixes the effective area per bit at $4\ell_P^2$; the four raw Planck cells per tile are the geometric packing factor in Argument 1.) After this rescaling, the expected cardinality of $A_{\text{chan}}$ is exactly $N_{\min}$, with Poisson fluctuations of order $\sqrt{N_{\min}}$.

For (c), apply Proposition 4.2 directly: every Poisson sprinkling of a wormhole manifold has endpoint-only-embedded throat sub-causal-set by the topological separation of bridge interior from asymptotic regions. $\square$

Remark 5.6 (Bell pair sprinkling). For $N_{\min} = 1$ (Bell pair), the Morris–Thorne throat has $r_0 = \ell_P/\sqrt{\pi} \approx 0.56\,\ell_P$. The sub-Planckian spherical radius is benign — the throat area $A = 4\ell_P^2$ is one tile-equivalent at the Planck-tile scale, and the Planck-resolution restriction of structural postulate S1 (Area Scaling) is on area, not radius. The expected sprinkling cardinality at the bottleneck is 1 element, matching $N_{\min}$.

Remark 5.7 (Poisson fluctuations are not framework noise). The $O(\sqrt{N_{\min}})$ fluctuations in (b) are intrinsic to CST Poisson sprinkling, not framework-specific noise. For macroscopic entanglement ($N_{\min} \gg 1$), relative fluctuation $1/\sqrt{N_{\min}} \to 0$ and the throat sprinkling concentrates around the mean. For the Bell pair, the fluctuation is $O(1)$ relative to $N_{\min} = 1$, consistent with the Planck-scale resolution at which the framework operates and with the discrete-substrate ontology.

5.3 Synthesis

Theorem 5.8 (Quantitative channel-sprinkling match). The discrete identification $N_{\min} = \lceil S_{\text{ent}}/\hbar\omega_0 \rceil$ (Proposition 5.1) and the Morris–Thorne sprinkling construction (Theorem 5.5) jointly give: for every channel coherence content $S_{\text{ent}}$, there exists a wormhole geometry whose Poisson sprinkling at the standard CST density produces exactly $N_{\min}$ elements at the bottleneck cross-section in expectation, with $O(\sqrt{N_{\min}})$ fluctuations.

The continuum limit recovers ER=EPR Section 3.4’s identification $N_{\min} = S_{\text{ent}}$ as the coarse-grained reading for $S_{\text{ent}} \gg \hbar\omega_0$.

Proof. Direct combination of Propositions 5.1 (discrete cardinality identification) and Theorem 5.5 (sprinkling realization). $\square$

Consistency Check: Bell Pair

Model: A perfectly-isolated Bell pair $|\Phi^+\rangle = \frac{1}{\sqrt{2}}(|0\rangle|0\rangle + |1\rangle|1\rangle)$ between two qubit observers at proper distance $L$, with $L$ much larger than the Planck scale.

Channel coherence (level-(n+1)): $\mathcal{C}(I_{12}) = S_{\text{ent}} = \ln 2$ (in units of $\hbar\omega_0$).

Channel cross-section cardinality (level-n, discrete): $N_{\min} = \lceil \ln 2 \rceil = 1$ causal-set element at the bottleneck cut (Proposition 5.1).

Endpoint-only embedded? Yes, by hypothesis of perfect isolation (no decoherence). Theorem 3.1 applies.

Handle topology? Yes, by Corollary 4.6. The channel forces a Planck-scale wormhole connecting the two observer regions.

Throat area:

• Discrete (substrate-level): $A_{\text{discrete}} = 4\ell_P^2 N_{\min} = 4\ell_P^2$ (one tile-equivalent).
• Continuum (coarse-grained): $A_{\text{continuum}} = 4\ell_P^2 \ln 2 \approx 2.77\ell_P^2$ (ER=EPR Proposition 3.3).
• The $\approx 1.23\,\ell_P^2$ discrepancy is within Planck-scale resolution (Remark 5.3).

Wormhole realization: Morris–Thorne wormhole with throat radius $r_0 = \ell_P/\sqrt{\pi} \approx 0.56\,\ell_P$ (so throat area $= 4\ell_P^2$). Poisson sprinkling at density $V_c^{-1}$ produces 1 element at the bottleneck cross-section in expectation (Theorem 5.5), matching $N_{\min}$.

Topologically: Spatial slice acquires a Planck-scale handle. The non-contractible loop is: leave $\mathcal{O}_1$, traverse flat space to $\mathcal{O}_2$ (length $L$), enter the wormhole, traverse the handle back to $\mathcal{O}_1$ (Planck-scale length). This loop generates $H_1 = \mathbb{Z}$.

Decoherence effect: If the Bell pair partially decoheres (interacts with environment), some channel elements acquire ambient neighbors. By Corollary 3.3, the channel transitions from pure handle to handle-plus-flux-tube hybrid. In the limit of complete decoherence, the channel is fully ambient-embedded and the handle disappears — the entangled state has become a product state. This is consistent with the experimental fact that decoherent quantum systems do not exhibit wormhole-like correlations.

Open Gaps

Gap 1 (Connection to Hauptvermutung). The lemma’s conclusion is the weaker “handle is consistent, flat space is excluded.” A full uniqueness statement (“handle is the unique class”) would require the Hauptvermutung. The argument provides one of the strongest currently-available specific instances supporting the Hauptvermutung: the framework’s irreducibility condition picks out a specific topology class from the causal-set structure. If the Hauptvermutung is later proved, this lemma’s conclusion strengthens automatically. This is a wider-field limitation of causal-set theory, not a framework-internal gap.

Rigor Assessment

Addresses Gaps In

• ER=EPR, Open Gap 3 (flux tube vs. spatial handle): closed. Theorem 3.2 Step 2 cites this lemma to establish that the channel induces a spatial handle (not merely an irreducibly-connected, non-pinching structure). Proposition 3.3 and Theorem 5.1 carry full Derived rigor in non-AdS settings on the strength of this lemma.

References

• Major–Rideout–Surya, 2007 — provides the thickened-antichain construction and the discrete-continuum homology correspondence.
• Surya, 2008 — review of CST topology results, including the Hauptvermutung status.
• Morris–Thorne, 1988 — explicit globally-hyperbolic traversable wormhole construction used in Proposition 4.1.
• Hawking–Ellis, 1973 — standard reference for the Kruskal extension of Schwarzschild and its global hyperbolicity.
• Cunningham–Surya, 2019 — worked examples of CST with non-trivial spatial topology ($S^1$, $T^2$); template for the Gap 1 simulation work.
• Bernal–Sánchez, 2003 — smooth-Cauchy-hypersurface refinement of Geroch’s splitting theorem, used in Theorem 4.7 Step 1 to construct a Cauchy hypersurface containing the bottleneck.