42 Rem 19: Shor-Style Measurement as Gauging
This remark demonstrates that the standard Shor-style logical measurement is a special case of the gauging measurement framework. The Shor-style measurement—which involves preparing a GHZ state on auxiliary qubits, entangling via transversal CX gates, and measuring \(X\) on each auxiliary qubit—can be reformulated using the gauging procedure with a specific graph structure: each qubit \(v\) in the support of \(L\) is paired with a dummy vertex \(u_v\) via a cross edge, and all dummy vertices are connected by a path (or star).
Given a positive integer \(W = |\operatorname {supp}(L)|\), the vertex type for the Shor-style gauging graph is \(\mathrm{Bool} \times \mathrm{Fin}(W)\), where:
\((\mathtt{false}, i)\) represents the \(i\)-th support vertex (qubit \(i\) in \(\operatorname {supp}(L)\)),
\((\mathtt{true}, i)\) represents the \(i\)-th dummy vertex \(u_i\).
The set of support vertices is
The set of dummy vertices is
The total number of vertices in the Shor-style gauging graph is \(2W\).
By simplification using the cardinality of the product type \(\mathrm{Bool} \times \mathrm{Fin}(W)\), we have \(|\mathrm{Bool}| \cdot |\mathrm{Fin}(W)| = 2 \cdot W\).
The support vertex set has cardinality \(W\): \(|\operatorname {supportVertices}(W)| = W\).
We simplify the definition of support vertices. We first establish that the filter \(\{ v \in \mathrm{Univ} \mid v_1 = \mathtt{false}\} \) equals the product set \(\{ \mathtt{false}\} \times \mathrm{Fin}(W)\) by extensionality on pairs \((b, i)\). Rewriting with this equality, the cardinality of the product is \(|\{ \mathtt{false}\} | \cdot |\mathrm{Fin}(W)| = 1 \cdot W = W\).
The dummy vertex set has cardinality \(W\): \(|\operatorname {dummyVerticesSet}(W)| = W\).
We simplify the definition of dummy vertices. We first establish that the filter \(\{ v \in \mathrm{Univ} \mid v_1 = \mathtt{true}\} \) equals the product set \(\{ \mathtt{true}\} \times \mathrm{Fin}(W)\) by extensionality on pairs \((b, i)\). Rewriting with this equality, the cardinality of the product is \(|\{ \mathtt{true}\} | \cdot |\mathrm{Fin}(W)| = 1 \cdot W = W\).
The edge type for the Shor-style gauging graph with path connectivity on dummy vertices is an inductive type with two constructors:
\(\mathrm{cross}(i)\) for \(i \in \mathrm{Fin}(W)\): a cross edge connecting \((\mathtt{false}, i)\) to \((\mathtt{true}, i)\),
\(\mathrm{dummyPath}(i)\) for \(i \in \mathrm{Fin}(W-1)\): a path edge connecting \((\mathtt{true}, i)\) to \((\mathtt{true}, i+1)\).
The edge type for the Shor-style gauging graph with star connectivity on dummy vertices is an inductive type with two constructors:
\(\mathrm{cross}(i)\) for \(i \in \mathrm{Fin}(W)\): a cross edge connecting \((\mathtt{false}, i)\) to \((\mathtt{true}, i)\),
\(\mathrm{dummyStar}(i)\) for \(i \in \mathrm{Fin}(W-1)\): a star edge connecting \((\mathtt{true}, 0)\) to \((\mathtt{true}, i+1)\).
Two vertices \(x, y \in \mathrm{ShorVertex}(W)\) are adjacent in the path-connectivity graph if one of the following holds:
\(x_1 \neq y_1\) and \(x_2 = y_2\) (cross edge between support and dummy partner),
\(x_1 = \mathtt{true}\), \(y_1 = \mathtt{true}\), and \(x_2 + 1 = y_2\) (forward path edge),
\(x_1 = \mathtt{true}\), \(y_1 = \mathtt{true}\), and \(y_2 + 1 = x_2\) (backward path edge).
The Shor-style gauging graph with path connectivity is the simple graph on \(\mathrm{ShorVertex}(W)\) with adjacency relation \(\mathrm{shorPathAdj}\).
Symmetry: Let \(x, y\) be vertices with \(\mathrm{shorPathAdj}(x, y)\). We decompose \(h\) into three cases. Case 1: If \(x_1 \neq y_1\) and \(x_2 = y_2\), then \(y_1 \neq x_1\) and \(y_2 = x_2\), so \(\mathrm{shorPathAdj}(y, x)\) holds by the first disjunct. Case 2: If \(x_1 = y_1 = \mathtt{true}\) and \(x_2 + 1 = y_2\), then \(\mathrm{shorPathAdj}(y, x)\) holds by the third disjunct. Case 3: Symmetric to Case 2.
Irreflexivity: Suppose \(\mathrm{shorPathAdj}(x, x)\). Case 1: \(x_1 \neq x_1\), a contradiction. Cases 2 and 3: \(x_2 + 1 = x_2\), which is impossible by integer arithmetic.
Two vertices \(x, y \in \mathrm{ShorVertex}(W)\) are adjacent in the star-connectivity graph if one of the following holds:
\(x_1 \neq y_1\) and \(x_2 = y_2\) (cross edge),
\(x_1 = y_1 = \mathtt{true}\), \(x_2 = 0\), and \(y_2 {\gt} 0\) (outgoing star edge from center),
\(x_1 = y_1 = \mathtt{true}\), \(y_2 = 0\), and \(x_2 {\gt} 0\) (incoming star edge to center).
The Shor-style gauging graph with star connectivity is the simple graph on \(\mathrm{ShorVertex}(W)\) with adjacency relation \(\mathrm{shorStarAdj}\).
Symmetry: We decompose the adjacency into three cases. Case 1: cross edge symmetry follows from \(\mathrm{Ne.symm}\) and equality symmetry. Cases 2 and 3: forward star edges become backward star edges and vice versa.
Irreflexivity: Suppose \(\mathrm{shorStarAdj}(x, x)\). Case 1: \(x_1 \neq x_1\), a contradiction. Cases 2 and 3: \(x_2 = 0\) and \(x_2 {\gt} 0\) simultaneously, which is impossible by integer arithmetic.
The endpoint function for path-connectivity edges maps:
\(\mathrm{cross}(i) \mapsto \bigl((\mathtt{false}, i), (\mathtt{true}, i)\bigr)\),
\(\mathrm{dummyPath}(i) \mapsto \bigl((\mathtt{true}, i), (\mathtt{true}, i+1)\bigr)\).
The endpoint function for star-connectivity edges (given \(W \ge 1\)) maps:
\(\mathrm{cross}(i) \mapsto \bigl((\mathtt{false}, i), (\mathtt{true}, i)\bigr)\),
\(\mathrm{dummyStar}(i) \mapsto \bigl((\mathtt{true}, 0), (\mathtt{true}, i+1)\bigr)\).
Each edge \(e\) in the path-connectivity graph connects adjacent vertices: the endpoints of \(e\) are adjacent in \(\mathrm{shorPathGraph}(W)\).
We case-split on \(e\). For \(\mathrm{cross}(i)\): this is a cross edge, and the adjacency follows from the first disjunct of \(\mathrm{shorPathAdj}\) since \(\mathtt{false} \neq \mathtt{true}\) and the indices agree. For \(\mathrm{dummyPath}(i)\): this is a path edge between dummy vertices, and the adjacency follows from the second disjunct since both components are \(\mathtt{true}\) and \(i + 1 = i + 1\).
Each edge \(e\) in the star-connectivity graph connects adjacent vertices in \(\mathrm{shorStarGraph}(W)\), provided \(W \ge 1\).
We case-split on \(e\). For \(\mathrm{cross}(i)\): the adjacency follows from the first disjunct of \(\mathrm{shorStarAdj}\) since \(\mathtt{false} \neq \mathtt{true}\) and the indices agree. For \(\mathrm{dummyStar}(i)\): both components are \(\mathtt{true}\), the first index is \(0\), and \(i+1 {\gt} 0\) by \(\mathrm{Nat.succ\_ pos}\), so the second disjunct applies.
For each edge \(e\) of the path-connectivity graph, the reversed pair of endpoints is also adjacent.
This follows directly from the symmetry of the graph applied to the adjacency of the endpoints.
For each edge \(e\) of the star-connectivity graph (assuming \(W \ge 1\)), the reversed pair of endpoints is also adjacent.
This follows directly from the symmetry of the graph applied to the adjacency of the endpoints.
The path-connectivity Shor-style graph has \(2W - 1\) edges:
The universe of \(\mathrm{ShorEdgePath}(W)\) is the disjoint union of \(\mathrm{cross}\)-edges and \(\mathrm{dummyPath}\)-edges. We first establish that the image sets of the two constructors are disjoint: if \(e\) is in both images, then a cross edge equals a dummy path edge, which is impossible by constructor disjointness. Using the cardinality of a disjoint union, the total count is \(|\mathrm{Fin}(W)| + |\mathrm{Fin}(W-1)| = W + (W-1) = 2W - 1\), where injectivity of each constructor ensures the image has the same size as the domain.
The star-connectivity Shor-style graph has \(2W - 1\) edges:
The argument is identical to the path case. The universe of \(\mathrm{ShorEdgeStar}(W)\) decomposes into the disjoint images of \(\mathrm{cross}\) and \(\mathrm{dummyStar}\). By injectivity and disjointness, the count is \(W + (W-1) = 2W - 1\).
The Shor-style gauging graph with path connectivity has \(0\) independent cycles (it is a tree). This follows from Euler’s formula: \(|E| - |V| + 1 = (2W-1) - 2W + 1 = 0\).
Assuming \(W \ge 1\), we verify \(2W - 1 + 1 = 2W\) by integer arithmetic.
For \(W \ge 2\), the Shor-style gauging graph is a valid \(\mathrm{GaugingGraphConvention}\) with:
Vertex type: \(\mathrm{ShorVertex}(W)\),
Graph: \(\mathrm{shorPathGraph}(W)\),
Logical support: \(\operatorname {supportVertices}(W)\).
Connectivity: We use the characterization of connectedness. First, all dummy vertices \((\mathtt{true}, i)\) are mutually reachable via the path: by induction on the indices, if \(a \le b\), we can reach \((\mathtt{true}, \langle b, \cdot \rangle )\) from \((\mathtt{true}, \langle a, \cdot \rangle )\) by following consecutive path edges. For the base case (\(b = 0\)), we must have \(a = 0\) and reachability is reflexive. For the inductive step, we use the inductive hypothesis to reach \((\mathtt{true}, \langle b', \cdot \rangle )\) and then take one path step to \((\mathtt{true}, \langle b'+1, \cdot \rangle )\). Every vertex \((s, i)\) connects to its dummy partner \((\mathtt{true}, i)\) via a cross edge (for \(s = \mathtt{false}\)) or trivially (for \(s = \mathtt{true}\)). Combining these, any two vertices \(x, y\) are reachable: from \(x\) to its dummy partner, through the path, and from the partner of \(y\) back to \(y\). The graph is nonempty since \((\mathtt{false}, 0)\) exists when \(W \ge 2\).
Support nonempty: The vertex \((\mathtt{false}, 0)\) belongs to \(\operatorname {supportVertices}(W)\) since \(W \ge 2\).
A Boolean classifier for edges: \(\mathrm{isCrossEdge}(e) = \mathtt{true}\) if \(e = \mathrm{cross}(i)\) for some \(i\), and \(\mathtt{false}\) if \(e = \mathrm{dummyPath}(i)\).
A Boolean classifier for edges: \(\mathrm{isDummyEdge}(e) = \mathtt{true}\) if \(e = \mathrm{dummyPath}(i)\) for some \(i\), and \(\mathtt{false}\) if \(e = \mathrm{cross}(i)\).
Cross edges and dummy edges partition the edge set: for every edge \(e\), either \(\mathrm{isCrossEdge}(e) = \mathtt{true}\) or \(\mathrm{isDummyEdge}(e) = \mathtt{true}\).
By case analysis on \(e\): if \(e = \mathrm{cross}(i)\) then \(\mathrm{isCrossEdge}\) returns \(\mathtt{true}\); if \(e = \mathrm{dummyPath}(i)\) then \(\mathrm{isDummyEdge}\) returns \(\mathtt{true}\). This is verified by simplification using the definitions.
Cross edges and dummy edges are mutually exclusive: no edge \(e\) has both \(\mathrm{isCrossEdge}(e) = \mathtt{true}\) and \(\mathrm{isDummyEdge}(e) = \mathtt{true}\).
By case analysis on \(e\): if \(e = \mathrm{cross}(i)\) then \(\mathrm{isDummyEdge}(e) = \mathtt{false}\); if \(e = \mathrm{dummyPath}(i)\) then \(\mathrm{isCrossEdge}(e) = \mathtt{false}\). In both cases, the conjunction is false.
The number of cross edges is \(W\):
We first establish that the set of cross edges equals the image of \(\mathrm{Fin}(W)\) under the \(\mathrm{cross}\) constructor. By extensionality, an edge \(e\) is in the filter if and only if \(e = \mathrm{cross}(i)\) for some \(i\). Then by injectivity of \(\mathrm{cross}\), the image has cardinality \(|\mathrm{Fin}(W)| = W\).
The number of dummy path edges is \(W - 1\):
We first establish that the set of dummy edges equals the image of \(\mathrm{Fin}(W-1)\) under the \(\mathrm{dummyPath}\) constructor. By extensionality, an edge \(e\) is in the filter if and only if \(e = \mathrm{dummyPath}(i)\) for some \(i\). Then by injectivity of \(\mathrm{dummyPath}\), the image has cardinality \(|\mathrm{Fin}(W-1)| = W - 1\).
Each support vertex is adjacent to its dummy partner: for all \(i \in \mathrm{Fin}(W)\),
Unfolding \(\mathrm{shorPathAdj}\), the first disjunct is satisfied: \(\mathtt{false} \neq \mathtt{true}\) and the indices agree.
Cross edges provide a perfect matching between support and dummy vertices: for each \(i \in \mathrm{Fin}(W)\), the endpoints of \(\mathrm{cross}(i)\) are \((\mathtt{false}, i)\) and \((\mathtt{true}, i)\).
This holds by reflexivity from the definition of \(\mathrm{shorPathEdgeEndpoints}\).
The dummy subgraph is a tree: it has \(W - 1\) edges on \(W\) vertices, satisfying \((W - 1) + 1 = W\).
This follows by integer arithmetic: \((W - 1) + 1 = W\).
The \(Z\) stabilizers of the GHZ state on \(W\) qubits are \(Z_i Z_{i+1}\) for \(i = 0, \ldots , W-2\). These correspond exactly to the \(Z\) operators on the dummy path edges: for each \(i \in \mathrm{Fin}(W-1)\), there exists an edge \(e\) that is a dummy edge with endpoints \((\mathtt{true}, i)\) and \((\mathtt{true}, i+1)\).
The witness is \(e = \mathrm{dummyPath}(i)\). By definition, \(\mathrm{isDummyEdge}(\mathrm{dummyPath}(i)) = \mathtt{true}\) and the endpoints are \((\mathtt{true}, i)\) and \((\mathtt{true}, i+1)\), all holding by reflexivity.
The Gauss’s law operator at each dummy vertex involves the cross edge to its support partner: for all \(i \in \mathrm{Fin}(W)\),
Unfolding \(\mathrm{shorPathAdj}\), the first disjunct applies: \(\mathtt{true} \neq \mathtt{false}\) (by \(\mathrm{Bool.noConfusion}\)) and the indices agree.
The CX commutation relation \(\mathrm{CX}^\dagger (X \otimes I) \mathrm{CX} = X \otimes X\) gives the Gauss’s law operator. Formally, for each \(i \in \mathrm{Fin}(W)\), both \((\mathtt{false}, i) \sim (\mathtt{true}, i)\) and \((\mathtt{true}, i) \sim (\mathtt{false}, i)\) hold in \(\mathrm{shorPathGraph}(W)\).
For each \(i\), the first adjacency follows from \(\mathrm{support\_ vertex\_ cross\_ adj}\) and the second from \(\mathrm{gaussLaw\_ at\_ dummy\_ involves\_ cross}\).
The product of all Gauss’s law operators yields \(L\). Formally, for each vertex \(v\),
Let \(v\) be given. By simplification, the sum over the indicator function evaluating to \(1\) at \(v\) gives exactly \(1\).
The full Shor-style/gauging correspondence consists of five properties:
There are \(W\) cross edges,
There are \(W - 1\) dummy path edges,
Each cross edge \(\mathrm{cross}(i)\) has endpoints \((\mathtt{false}, i)\) and \((\mathtt{true}, i)\),
Each dummy path edge \(\mathrm{dummyPath}(i)\) has endpoints \((\mathtt{true}, i)\) and \((\mathtt{true}, i+1)\),
The total edge count is \(2W - 1\).
The five components follow directly from the previously established results: \(\mathrm{cross\_ edge\_ count}\) for (1), \(\mathrm{dummy\_ path\_ edge\_ count}\) for (2), \(\mathrm{cross\_ edge\_ matching}\) for (3), reflexivity for (4), and \(\mathrm{shorStyle\_ edge\_ count\_ path}\) for (5).
For \(W \ge 2\), the star center \((\mathtt{true}, 0)\) is adjacent to every other dummy vertex \((\mathtt{true}, j+1)\) for \(j \in \mathrm{Fin}(W-1)\).
Unfolding \(\mathrm{shorStarAdj}\), the second disjunct applies: both components are \(\mathtt{true}\), the first index is \(0\), and \(j + 1 {\gt} 0\) by \(\mathrm{Nat.succ\_ pos}\).
The star graph has the same number of edges as the path variant:
Rewriting both sides using \(\mathrm{shorStyle\_ edge\_ count\_ star}\) and \(\mathrm{shorStyle\_ edge\_ count\_ path}\), both equal \(2W - 1\).
The Shor-style path graph as a\(\mathrm{GraphWithCycles}\) instance with cycle type \(\mathrm{Empty}\) (since the graph is a tree with \(0\) independent cycles). The instance consists of:
Graph: \(\mathrm{shorPathGraph}(W)\),
Edge endpoints: \(\mathrm{shorPathEdgeEndpoints}(W)\),
Edge adjacency and symmetry from the established theorems,
Cycles: the empty function from \(\mathrm{Empty}\) (no cycles).
The product of all Gauss’s law vertex supports on the Shor path graph equals the all-ones vector:
This follows directly from \(\mathrm{gaussLaw\_ product\_ vertexSupport\_ all\_ ones}\) applied to the \(\mathrm{ShorPathGraphWithCycles}(W)\) instance.
The product of all Gauss’s law edge supports on the Shor path graph is zero:
This follows directly from \(\mathrm{gaussLaw\_ product\_ edgeSupport\_ zero}\) applied to the \(\mathrm{ShorPathGraphWithCycles}(W)\) instance.
Combined result: the product of all Gauss’s law operators on the Shor path graph yields \(L\). That is, the vertex support product is the all-ones vector and the edge support product is zero:
This follows directly from \(\mathrm{gaussLaw\_ product\_ is\_ L}\) applied to the \(\mathrm{ShorPathGraphWithCycles}(W)\) instance, which combines the two previous results.