---
title: Whiteley’s Cofactor Matroid in Rigidity
url: https://www.emergentmind.com/topics/whiteley-s-cofactor-matroid
type: topic
---

# Whiteley’s Cofactor Matroid in Rigidity

Whiteley’s cofactor matroid is a matroid on the edge set of a complete graph obtained from the row-dependence structure of a cofactor matrix arising in bivariate spline theory. For generic planar placements, the generic \(C^{d-2}_{d-1}\)-cofactor matroid \(\mathcal{C}_{d-1,n}^{d-2}\) is the row matroid of the corresponding cofactor matrix on \(K_n\); in dimension three this is the generic \(C_2^1\)-cofactor matroid \(\mathcal{C}_{2,n}^1\), which is central in the theory of abstract \(3\)-rigidity matroids [1911.00207]. On special configurations it coincides with bar-and-joint rigidity on the moment curve and with hyperconnectivity on monomial vectors, placing it at the intersection of spline theory, rigidity theory, and matroid theory [2106.08923].

## 1. Matrix construction and generic definition

Let \(G=(V,E)\) be a finite simple graph, let \(p:V\to\mathbb{R}^2\) be a placement with \(p(v_i)=(x_i,y_i)\), and let \(s\ge 0\). Whiteley’s \(C_s^{s-1}\)-cofactor matrix \(C_s^{s-1}(G,p)\) is the \(|E|\times (s+1)|V|\) matrix whose rows are indexed by edges and whose columns come in blocks of \(s+1\) associated to the vertices. For an edge \(e=v_iv_j\) with \(i<j\), the row has the form
\[
\kbordermatrix{ & & v_i & & v_j & \cr
e=v_iv_j & 0\cdots 0 & D_{ij} & 0 \cdots 0 & -D_{ij} & 0\cdots 0 },
\]
where
\[
D_{ij} = \big((x_i-x_j)^s, (x_i-x_j)^{s-1}(y_i-y_j), \dots, (x_i-x_j)(y_i-y_j)^{s-1}, (y_i-y_j)^s\big)\in \mathbb{R}^{s+1}.
\]
The generic \(C_s^{s-1}\)-cofactor matroid \(\mathcal{C}_{s,n}^{s-1}\) is defined as the row matroid of \(C_s^{s-1}(K_n,p)\) for a generic map \(p:V(K_n)\to\mathbb{R}^2\), meaning that the coordinates are algebraically independent over \(\mathbb{Q}\). For \(d\)-dimensional rigidity, Whiteley’s cofactor matroid is the case \(s=d-1\), namely \(\mathcal{C}_{d-1,n}^{d-2}\) [1911.00207].

A closely related notation, used for planar point configurations \(p=(p_1,\dots,p_n)\subset\mathbb{R}^2\), starts from the cofactor vector
\[
c(x,y)=(x^{d-1},x^{d-2}y,\dots,xy^{d-2},y^{d-1})\in\mathbb{R}^d.
\]
The degree-\(d-1\) cofactor rigidity matrix \(C_d(p)\) is then the \(\binom{n}{2}\times nd\) block matrix whose row \((i,j)\) has block \(c(p_i-p_j)\) in position \(i\), block \(-c(p_i-p_j)\) in position \(j\), and zeros elsewhere. The associated cofactor rigidity matroid \(C^{d-2}_{d-1}(p)\) is the linear matroid of the rows of \(C_d(p)\), with ground set \(\binom{[n]}{2}\) [2106.08923].

In the three-dimensional case, the cofactor vector reduces to
\[
D(p_i,p_j)=\bigl((x_i-x_j)^2,\ (x_i-x_j)(y_i-y_j),\ (y_i-y_j)^2\bigr)\in\mathbb{R}^3.
\]
For a generic framework \((G,p)\), the right kernel of \(C_2^1(G,p)\) contains six independent trivial \(C_2^1\)-motions. A \(C_2^1\)-motion is a map \(q:V\to\mathbb{R}^3\) satisfying
\[
D(v_i,v_j)\cdot (q(v_i)-q(v_j))=0 \quad\text{for all } v_iv_j\in E(G).
\]
The framework is \(C_2^1\)-rigid when only the trivial motions occur, and minimally \(C_2^1\)-rigid when it is both rigid and independent [1911.00205].

## 2. Abstract rigidity and maximality in dimension three

Graver’s notion of an abstract \(d\)-rigidity matroid formalizes the closure properties shared by generic rigidity matroids. Nguyen’s characterization states that a matroid \(M\) on \(E(K_n)\), with \(n\ge d+2\), is an abstract \(d\)-rigidity matroid if and only if every copy of \(K_{d+2}\) is a circuit and
\[
r_M(E(K_n)) = d n - \binom{d+1}{2}.
\]
Consequently, abstract \(3\)-rigidity matroids are precisely the matroids on \(E(K_n)\) in which every copy of \(K_5\) is a circuit and the total rank is \(3n-6\). Whiteley proved that the generic \(C_{d-1}^{d-2}\)-cofactor matroid \(\mathcal{C}_{d-1,n}^{d-2}\) is an abstract \(d\)-rigidity matroid; in particular, \(\mathcal{C}_{2,n}^{1}\) is a \(K_5\)-matroid of rank \(3n-6\) [1911.00205].

Whiteley’s maximality conjecture proposed that, for every \(d\ge 2\), the generic cofactor matroid \(\mathcal{C}_{d-1,n}^{d-2}\) is the unique maximal abstract \(d\)-rigidity matroid with respect to the weak order on matroids. Clinch, Jackson, and Tanigawa verified the case \(d=3\): the generic \(C_2^1\)-cofactor matroid \(\mathcal{C}_{2,n}^{1}\) is the unique maximal abstract \(3\)-rigidity matroid, and more strongly the unique maximal \(K_5\)-matroid on \(E(K_n)\) [1911.00205].

The three-dimensional proof is closely tied to matroid-erection theory. In the companion work on combinatorial characterization, \(\mathcal{C}_{2,n}^{1}\) is identified as the free elevation of a rank-\(10\) matroid \(R_n\) whose non-spanning circuits are exactly the edge sets of copies of \(K_5\). A key step in the maximality proof is the verification that the double \(V\)-replacement operation preserves independence in the generic \(C_2^1\)-cofactor matroid, completing an inductive construction of bases analogous to Henneberg-type constructions in rigidity theory [1911.00207].

This resolves Whiteley’s maximality conjecture in dimension three, but it does not resolve Graver’s separate conjecture that the generic three-dimensional bar-and-joint rigidity matroid \(\mathcal{R}_3\) is isomorphic to Whiteley’s cofactor matroid. That comparison remains active in later work.

## 3. Rank, independence, and combinatorial characterization

For \(\mathcal{C}_{2,n}^{1}\), independence admits a purely combinatorial characterization. A proper \(K_5\)-sequence is a finite sequence \((C_1,\dots,C_t)\) of \(K_5\)-circuits such that \(C_i\not\subseteq C_{\le i-1}=\bigcup_{j=1}^{i-1}C_j\) for all \(i\ge 2\). The rank of \(X\subseteq E(K_n)\) is
\[
r(X)
=\min\Big\{\, |X\cup C_{\le t}| - t
\;:\;
(C_1,\dots,C_t)\text{ is a proper \(K_5\)-sequence in }K_n
\Big\}.
\]
This gives a graph-theoretic rank formula for the maximal abstract \(3\)-rigidity matroid and solves the cofactor analogue of the combinatorial characterization problem for generic \(3\)-dimensional bar-joint rigidity [1911.00207].

A second, cover-theoretic description uses \(2\)-thin, \(4\)-shellable covers. For a family \(\mathcal{X}\) of vertex sets, with hinges \(H(\mathcal{X})\), define
\[
\Phi(\mathcal{X}) =
\sum_{X\in\mathcal{X}}(3|X|-6)
\;-\;
\sum_{h\in H(\mathcal{X})}(\deg_{\mathcal{X}}(h)-1).
\]
Then for \(F\subseteq E(K_n)\),
\[
r(F)
=
\min_{F_0,\mathcal{X}}
\bigl\{|F_0|+\Phi(\mathcal{X})\bigr\},
\]
where the minimum is taken over all \(F_0\subseteq F\) and all \(4\)-shellable, \(2\)-thin covers \(\mathcal{X}\) of \(F\setminus F_0\) with sets of size at least \(5\). For a flat \(F\), if \(\mathcal{X}^*\) denotes the maximal cliques of \(K_n[F]\) of size at least \(5\), and \(F_0\) the uncovered edges, then
\[
r(F)=|F_0|+\Phi(\mathcal{X}^*).
\]
Thus the rank of a flat is determined by the maximal cliques in the induced subgraph.

Several structural consequences follow. Every cyclic flat of \(\mathcal{C}_{2,n}^{1}\) is the union of copies of \(K_5\). If \(G=(V,F)\) is \(5\)-connected and \(F\) is a cyclic set in \(\mathcal{C}_{2,n}^{1}\), then \(G\) is \(C_2^1\)-rigid. If \(G\) is \(12\)-connected and \(S\subseteq E\) with \(|S|\le 6\), then \(G-S\) is \(C_2^1\)-rigid. The same work also notes that the rank problem for \(\mathcal{C}_{2,n}^{1}\) lies in NP\(\cap\)co-NP [1911.00207].

## 4. Coincidence with bar-and-joint rigidity and hyperconnectivity

Whiteley’s cofactor matroid is not isolated from other rigidity-type matroids. For a configuration \(p=(p_1,\dots,p_n)\subset\mathbb{R}^d\), the bar-and-joint rigidity matroid \(R_d(p)\) is the row matroid of the standard rigidity matrix \(R(p)\). Kalai’s hyperconnectivity matroid \(H_d(p)\) is the row matroid of the hyperconnectivity matrix \(H(p)\). The cofactor matroid \(C^{d-2}_{d-1}(p)\) is the row matroid of the cofactor matrix \(C_d(p)\). All three are matroids on the ground set \(\binom{[n]}{2}\) [2106.08923].

For distinct real numbers \(t_1,\dots,t_n\), these three constructions coincide in a canonical family of realizations. The \(C^{d-2}_{d-1}\)-cofactor matroid of the points \(\{(t_i,t_i^2)\}_{i=1}^n\) on the standard parabola in \(\mathbb{R}^2\), the bar-and-joint rigidity matroid of the points \(\{(t_i,\dots,t_i^d)\}_{i=1}^n\) on the \(d\)-dimensional moment curve, and the \(d\)-hyperconnectivity matroid of the vectors \(\{(1,t_i,\dots,t_i^{d-1})\}_{i=1}^n\) all coincide. Because bar-and-joint rigidity and cofactor rigidity are projectively invariant, any non-degenerate conic may replace the parabola.

The coincidence is organized by the polynomial rigidity matroid \(P_d(t_1,\dots,t_n)\). If \(R[t]_{<d}\) is the space of univariate polynomials of degree \(<d\), and for each vertex \(i\) one chooses a basis
\[
F^i(t)=\big(f_1^i(t),\dots,f_d^i(t)\big),
\]
then the polynomial rigidity matrix has row \((i,j)\) equal to \(F^i(t_j)\) in block \(i\), \(-F^j(t_i)\) in block \(j\), and zeros elsewhere. Its row matroid is independent of the chosen bases. Appropriate choices recover the three rigidity matrices: the basis \(F^{t_i}(t)=\big(\frac{t_i^k-t^k}{t_i-t}\big)_{k=1,\dots,d}\) gives bar-and-joint rigidity on the moment curve, the basis \(G^{t_i}(t)=\big((t_i+t)^{k-1}\big)_{k=1,\dots,d}\) gives cofactor rigidity on the parabola, and the monomial basis \((1,t,\dots,t^{d-1})\) gives hyperconnectivity.

This coincidence has additional algebraic consequences. In even dimension \(d\), the generic hyperconnectivity matroid \(H_d(n)\) coincides with the algebraic matroid of skew-symmetric \(n\times n\) matrices of rank at most \(d\). For \(d=2\), \(H_2(n)=P_2(n)\) is also the algebraic matroid of the Grassmannian.

## 5. Highly connected graphs and reconstruction from the matroid

A later line of work studies the three-dimensional cofactor matroid on an arbitrary simple graph \(G=(V,E)\), written \(C(G)=\mathcal{C}_2^1(G)\). Because the original matrix definition is technically involved for the combinatorial arguments at hand, the rank characterization by \(4\)-shellable \(2\)-thin covers is taken as a definition. For simple graphs with at least five vertices, applying the rank formula with \(F=\varnothing\) and \(\mathcal{X}=\{V\}\) yields the general upper bound \(r(E)\le 3|V|-6\); \(G\) is called \(C\)-rigid when equality holds [2209.06204].

The same paper derives strong connectivity consequences. For the \(t\)-fold union \(C^t(G)\), if \(G\) is \(12t\)-connected, then
\[
r_t(E)=3t|V|-6t.
\]
If \(G\) is \(k\)-connected with \(k\ge 12t\), then \(C^t(G)\) is vertically \((k-6t+1)\)-connected. In particular, when \(t=1\), every \(12\)-connected graph is \(C\)-rigid and \(C(G)\) is vertically \(7\)-connected.

These connectivity estimates lead to Whitney-type reconstruction theorems. If \(G\) is \((12t+2)\)-connected, \(H\) has no isolated vertices, and \(\psi:E(G)\to E(H)\) is a matroid isomorphism between \(C^t(G)\) and \(C^t(H)\), then \(\psi\) is induced by a graph isomorphism \(G\cong H\). For \(t=1\), a \(14\)-connected graph is therefore uniquely determined by its cofactor matroid. There is also a Servatius-type formulation: if \(C(G)\) is vertically \(35\)-connected and \(H\) has no isolated vertices, then any isomorphism \(C(G)\cong C(H)\) is induced by a graph isomorphism.

The same framework yields packing consequences. Every \(12t\)-connected graph contains \(t\) edge-disjoint spanning subgraphs of cofactor rank \(3|V|-6\), and every \(24\)-connected graph has a spanning tree \(T\) such that \(G-E(T)\) is \(3\)-connected. These results extend the role of the cofactor matroid from a local rigidity model to a tool for graph reconstruction and high-connectivity structure theory.

## 6. Flexible circuits, implied nonedges, and current obstructions

Recent work studies Whiteley’s cofactor matroid in dimension three under the notation \(\mathcal{M}_{K_5}\), the unique maximal matroid on \(E(K_n)\) in which every graph isomorphic to \(K_5\) is a circuit. In that formulation, Whiteley’s cofactor matroid is exactly \(\mathcal{M}_{K_5}\), and the open comparison is Graver’s maximality conjecture that \(\mathcal{R}_3=\mathcal{M}_{K_5}\) [2508.12417].

This work isolates two notions that are especially relevant to the circuit structure. A nucleation of a graph is a rigid subgraph on at least five vertices; a graph is nucleation-free if it has no nucleation. A nonedge \(f\) of \(G\) is implied if \(G\cup f\) contains a circuit that includes \(f\). The resulting focus is on independent, nucleation-free graphs with implied nonedges, which are flexible but already encode nontrivial closure phenomena.

One basic family is the ring of butterflies \(R_m\). Each hinge in \(R_m\) is an implied nonedge; for \(m\ge 7\), \(R_m\) is nucleation-free; for \(m\ge 6\), \(R_m\) is independent and has \(m-6\) independent flexes; and for \(m\le 5\), \(R_m\) is dependent. The same paper proves that several operations preserve independence and nucleation-freeness, including \(k\)-sums for \(k\in\{0,1,2\}\), Henneberg-I, Henneberg-II under stated hypotheses, and \(k\)-vertex splits for \(k\in\{0,1,2\}\). More elaborate split-and-glue constructions, such as safe starting graphs and double-butterfly starting graphs, produce further independent, nucleation-free graphs with implied nonedges.

These constructions can also be combined to produce dependent graphs and circuits. If \(G_1\) and \(G_2\) are edge-disjoint graphs sharing only the endpoints of a nonedge \(f\), and \(f\) is implied in both, then \(G_1\cup G_2\) is dependent and \(f\) is implied in the union. If \(G_1\cup f\) and \(G_2\cup f\) are circuits and are otherwise vertex-disjoint, then \(G_1\cup G_2\) is a circuit.

None of these families refutes the conjecture \(\mathcal{R}_3=\mathcal{M}_{K_5}\), because their properties hold in both matroids. Their significance is different: they show that nucleation-free graphs with implied nonedges are key obstacles to settling Graver’s maximality conjecture and to obtaining a polynomial-time characterization of independence in \(\mathcal{M}_{K_5}\). A related open possibility, stated explicitly in that work, is an \(\mathcal{R}_3\)-implied nonedge not contained in any \(K_5\) in the closure \(CL_{\mathcal{R}_3}(G)\); such an example would be sufficient to refute the maximality conjecture.

Source: https://www.emergentmind.com/topics/whiteley-s-cofactor-matroid