pith. sign in

arxiv: 2512.00190 · v3 · submitted 2025-11-28 · 🧮 math.CO · math.SP

On the nullspace of split graphs

Daniel A. Jaume , Victor N. Schv\"ollner , Cristian Panelo , Kevin Pereyra This is my paper

Pith reviewed 2026-05-17 03:09 UTC · model grok-4.3

classification 🧮 math.CO math.SP
keywords split graphsnullspaceadjacency matrixclique-kernelnullitybiadjacency matrixTyshkevich compositiongraph singularity
0
0 comments X

The pith

The dimension of the clique-kernel of a split graph is at most one, giving an exact formula for its nullity in terms of the biadjacency submatrix R.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out to describe the nullspace of the adjacency matrix for split graphs, graphs whose vertices partition into a clique and an independent set. It defines the clique-kernel as the subspace that tracks whether clique vertices appear in the support of any kernel eigenvector and proves that this subspace has dimension at most one. The resulting formula states that the nullity of the split graph equals the nullity of the biadjacency submatrix R plus the dimension of the clique-kernel. A reader would care because the formula reduces the question of singularity for these graphs to a concrete property of one submatrix and supplies a complete algebraic account of their kernel in terms of the given partition.

Core claim

For a split graph Sp with biadjacency submatrix R between its clique and independent set, the authors introduce the clique-kernel Cker(Sp) and prove that its dimension is at most one. This immediately yields the identity null(Sp) = null(R) + dim(Cker(Sp)), which expresses the nullity of Sp completely in terms of R. The same framework is used to characterize the supports of kernel vectors in unbalanced split graphs via swing vertices, to track how the nullspace changes under Tyshkevich composition, and to obtain a closed formula for the determinant of the adjacency matrix.

What carries the argument

The clique-kernel, a subspace that records whether clique vertices can appear in the support of an eigenvector from the kernel of the adjacency matrix.

If this is right

  • The nullity of any split graph is completely determined by the nullity of its biadjacency submatrix together with a one-dimensional correction term.
  • Kernel supports in unbalanced split graphs are structured around swing vertices.
  • The nullspace transforms in a controlled way under Tyshkevich composition of split graphs.
  • The determinant of the adjacency matrix of a split graph admits a closed formula derived from the same decomposition.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The bound on clique-kernel dimension may simplify algorithmic tests for singularity once a split partition is known.
  • Similar subspace arguments could apply to other graph classes that admit a fixed clique-independent set partition.
  • The explicit formula supplies a direct route to counting the multiplicity of the zero eigenvalue without computing the full spectrum.

Load-bearing premise

The graph must be presented with a fixed partition into a clique and an independent set that defines the biadjacency submatrix R, and the adjacency matrix must be considered over the real numbers.

What would settle it

A concrete split graph whose clique-kernel subspace has dimension two or greater, or for which the stated nullity formula fails to hold when the adjacency matrix is written with respect to the given partition.

Figures

Figures reproduced from arXiv: 2512.00190 by Cristian Panelo, Daniel A. Jaume, Kevin Pereyra, Victor N. Schv\"ollner.

Figure 1
Figure 1. Figure 1: Split graph Sp (clique vertices in black, independent in white). If Sp is a split graph, the notation Sp(K, S) means that we are referring to Sp together with an s-partition (K, S) for it. Unless otherwise specified, we implicitly assume that both K and S are nonempty, and the adjacency matrix of Sp has the form A(Sp) =  J − I R Rt 0  , after applying the ordering (K, S) to V (Sp). Therefore, the blocks … view at source ↗
Figure 2
Figure 2. Figure 2: Split graph Sp(K, S) with Supp(Sp) ⊂ S, but 11 ∈ R/ (R). Direct computations shows that Supp(Sp) = {5, . . . , 9} = S. However, the equation Rx = 11 has no solution because 1 = R∗3x = (x6 + x7) + (x8 + x9) = R2∗x + R1∗x = 2. In the next proposition, we generalize the argument used for the split graph in [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Split graph with supported clique vertices and nullity one. [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Bipartite representation of the split graph [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Bipartite representation of a split graph (clique vertices in black) with nullity [PITH_FULL_IMAGE:figures/full_fig_p020_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Tyshkevich composition of Sp ≈ P4 and G ≈ C4. The Tyshkevich composition preserves (and potentially amplifies) the singularity coming from the biadjacency matrix R of the split graph. Theorem 7.1. Let Sp(K, S) be a split graph, and let G be a graph. Then, N (Sp ◦ G) ⊇ {(0, z, 0)t : z ∈ N (R)}. In particular, if N (R) ̸= {0}, then Sp ◦ G is singular. Proof. Let H = Sp ◦ G. By using the ordering (K, S, V (G)… view at source ↗
read the original abstract

We study the nullspace of the adjacency matrix of split graphs, whose vertex set can be partitioned into a clique and an independent set. We introduce the clique-kernel, a subspace that decides whether clique vertices lie in the support of a kernel eigenvector, and we prove that its dimension is at most one. This yields the formula $null(Sp) = null(R) + \dim(\mathrm{Cker}(Sp))$, which fully describes the nullity of a split graph in terms of the biadjacency submatrix $R$. We also analyze unbalanced split graphs through the concept of swing vertices and characterize the structure of their kernel supports. Furthermore, we study the behavior of the nullspace under Tyshkevich composition and derive a closed formula for the determinant. These results provide a unified algebraic framework for understanding when a split graph is singular and how its combinatorial structure determines its nullspace.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper studies the nullspace of the adjacency matrix of split graphs, which admit a partition of vertices into a clique K and independent set I. It introduces the clique-kernel Cker(Sp) as the subspace of kernel vectors supported on the clique, proves that its dimension is at most one, and derives the formula null(Sp) = null(R) + dim(Cker(Sp)) where R is the biadjacency submatrix between K and I. The work further analyzes unbalanced split graphs via swing vertices, examines the nullspace under Tyshkevich composition, and provides a closed formula for the determinant.

Significance. If the central results hold, the paper supplies a concrete algebraic description of nullity for split graphs in terms of a biadjacency submatrix plus a correction term of dimension at most one. The clique-kernel construction and the composition analysis could serve as useful tools for determining singularity and kernel supports in this graph class.

major comments (1)
  1. The main formula null(Sp) = null(R) + dim(Cker(Sp)) is derived for a fixed partition V = K ∪ I. Split graphs can admit multiple such partitions (for example, when a vertex of I is adjacent to every vertex of K and can be reassigned to K while preserving the split property). The manuscript contains no explicit argument establishing that the right-hand side is invariant under choice of partition. Since null(Sp) is an intrinsic graph invariant, this invariance must be shown for the formula to fully describe the nullity as claimed.
minor comments (2)
  1. The abstract states that the formula 'fully describes' the nullity; this phrasing should be qualified to reflect any dependence on the chosen partition.
  2. In the section introducing swing vertices, provide a precise definition and an example showing how they affect kernel supports in unbalanced cases.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and insightful comments on our manuscript concerning the nullspace of split graphs. We address the major comment point by point below.

read point-by-point responses

  1. Referee: The main formula null(Sp) = null(R) + dim(Cker(Sp)) is derived for a fixed partition V = K ∪ I. Split graphs can admit multiple such partitions (for example, when a vertex of I is adjacent to every vertex of K and can be reassigned to K while preserving the split property). The manuscript contains no explicit argument establishing that the right-hand side is invariant under choice of partition. Since null(Sp) is an intrinsic graph invariant, this invariance must be shown for the formula to fully describe the nullity as claimed.

    Authors: We agree with the referee that the derivation in the manuscript is carried out for a fixed partition V = K ∪ I and that an explicit argument for invariance of the right-hand side under alternative valid partitions is not provided. Because null(Sp) is independent of any particular choice of split partition, such an invariance proof is required to substantiate the claim that the formula fully describes the nullity. In the revised manuscript we will insert a dedicated subsection establishing this invariance. The argument will proceed by considering the possible transitions between partitions (for instance, relocating a vertex of I that is adjacent to every member of K into the clique) and verifying that any resulting change in null(R) is exactly offset by a corresponding adjustment in dim(Cker(Sp)), so that the sum remains constant. This addition will make the formula manifestly intrinsic to the graph. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained linear algebra on defined subspaces

full rationale

The paper fixes a clique-independent set partition for the split graph Sp, defines the biadjacency submatrix R and the clique-kernel Cker(Sp) as the subspace of kernel vectors supported on the clique, proves dim(Cker(Sp)) ≤ 1 via the split graph structure, and obtains the nullity formula as a direct consequence. This is standard block-matrix kernel analysis over the reals with no reduction of the final expression to a fitted quantity, self-referential definition, or load-bearing self-citation. The result is independently verifiable from the adjacency matrix block form and does not rename prior results or smuggle ansatzes.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The work relies on standard facts from linear algebra and graph theory; the only new object is the explicitly defined clique-kernel subspace.

axioms (2)
  • standard math The adjacency matrix of a graph is symmetric with 0-1 entries and the nullspace is taken over the reals.
    Invoked implicitly when discussing kernel eigenvectors and nullity.
  • domain assumption Every split graph admits a partition into a clique and an independent set.
    Used to define the biadjacency submatrix R and the clique-kernel.
invented entities (1)
  • clique-kernel no independent evidence
    purpose: Subspace of the nullspace that captures support on clique vertices.
    Newly defined object whose dimension is proved to be at most one.

pith-pipeline@v0.9.0 · 5454 in / 1420 out tokens · 33629 ms · 2026-05-17T03:09:28.072174+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

  1. [1]

    D. A. Ali, J. B. Gauci, I. Sciriha, and K. R. Sharaf. Coalescing fiedler and core vertices.Czechoslovak Mathematical Journal, 66(3):971–985, 2016

    work page 2016

  2. [2]

    M. D. Barrus and D. B. West. The a_4 structure of a graph.Journal of Graph Theory, 69(2):97–113, 2012

    work page 2012

  3. [3]

    Cheng, K

    C. Cheng, K. L. Collins, and A. N. Trenk. Split graphs and nordhaus– gaddum graphs.Discrete Mathematics, 339(9):2345–2356, 2016

    work page 2016

  4. [4]

    Földes and P

    S. Földes and P. L. Hammer. Split graphs with distinct representative properties.Combinatorics (Proc. Fifth Hungarian Colloq., Keszthely,

    work page

  5. [5]

    Colloq. Math. Soc. János Bolyai18, Vol. II:345–356, 1976

    work page 1976

  6. [6]

    Földes and P

    S. Földes and P. L. Hammer. Split graphs.Proceedings of the Eighth Southeastern Conference on Combinatorics, Graph Theory and Com- puting (Louisiana State Univ., Baton Rouge. Congr. Numer.19, 1977

    work page 1977

  7. [7]

    Foldes and P

    S. Foldes and P. L. Hammer. Split graphs having dilworth number two. Canadian Journal of Mathematics, 29(3):666–672, 1977

    work page 1977

  8. [8]

    Földes and P

    S. Földes and P. L. Hammer. On a class of matroid-producing graphs. Combinatorics (Proc. Fifth Hungarian Colloq., Keszthely, 1976). Colloq. Math. Soc. János Bolyai18, Vol. I:331–352, 1978

    work page 1976

  9. [9]

    M. C. Golumbic.Algorithmic graph theory and perfect graphs, vol- ume 57. Elsevier, 2004

    work page 2004

  10. [10]

    P. L. Hammer and B. Simeone. The splittance of a graph.Combinator- ica, 1(3):275–284, 1981. 27

    work page 1981

  11. [11]

    R. A. Horn and C. R. Johnson.Matrix analysis. Cambridge university press, 2012

    work page 2012

  12. [12]

    D. A. Jaume and G. Molina. Null decomposition of trees.Discrete Mathematics, 341(3):836–850, 2018

    work page 2018

  13. [13]

    Nulldecompositionofbipartite graphs without cycles of length 0 modulo 4.Linear Algebra and its Applications, 614:176–196, 2021

    D.A.Jaume,G.Molina,andA.Pastine. Nulldecompositionofbipartite graphs without cycles of length 0 modulo 4.Linear Algebra and its Applications, 614:176–196, 2021

    work page 2021

  14. [14]

    N. V. Mahadev and U. N. Peled.Threshold graphs and related topics, volume 56. Elsevier, 1995

    work page 1995

  15. [15]

    Panelo, D

    C. Panelo, D. A. Jaume, and M. Machado Toledo. On the core-nilpotent decomposition of threshold graphs.Available at SSRN 5316214

    work page

  16. [16]

    I. Sciriha. On the construction of graphs of nullity one.Discrete Math- ematics, 181(1-3):193–211, 1998

    work page 1998

  17. [17]

    I. Sciriha. A characterization of singular graphs. 2007

    work page 2007

  18. [18]

    I. Sciriha. Maximal core size in singular graphs. 2009

    work page 2009

  19. [19]

    Sciriha and S

    I. Sciriha and S. Farrugia. On the spectrum of threshold graphs.Inter- national Scholarly Research Notices, 2011(1):108509, 2011

    work page 2011

  20. [20]

    Tyshkevich

    R. Tyshkevich. Decomposition of graphical sequences and unigraphs. Discrete Mathematics, 220(1-3):201–238, 2000

    work page 2000

  21. [21]

    Von Collatz and U

    L. Von Collatz and U. Sinogowitz. Spektren endlicher grafen: Wilhelm blaschkezum 70.geburtstag gewidmet. InAbhandlungen aus dem Math- ematischen Seminar der Universität Hamburg, volume 21, pages 63–77. Springer, 1957

    work page 1957

  22. [22]

    Whitman.Split graphs, unigraphs, and Tyshkevich decompositions

    R. Whitman.Split graphs, unigraphs, and Tyshkevich decompositions. PhD thesis, Wellesley College, 2020. 28

    work page 2020