The Spectral Theory of Cographs: A Survey, Organized by Graph Matrix
Abstract
Cographs, the P4-free graphs, are exactly the graphs obtained from a single vertex by repeated disjoint union and complementation. They contain the threshold graphs, π-graphs and form one of the primary areas for spectral graph theory. This survey is organized by matrix: after a preliminary section, there is one section for each of the adjacency, Laplacian, signless Laplacian, distance, eccentricity, Seidel, distance Laplacian and distance signless Laplacian matrices. Within each section we collect what is known for general cographs, for the subclass of π-graphs, and where relevant, for threshold graphs. We highlight the results from the recent years, including the foundational work of Mohammadian et al., Jacobs et al., Ghorbani, Li et al., Niu et al., Lou et al., Wang at al., Mandal et al. Two structural facts are used throughout: every connected cograph has diameter at most $2$, which ties all four distance-type matrices to the adjacency and Laplacian matrices, and the partition of the vertex set into twin classes is equitable for every one of the eight matrices, giving a uniform description of each spectrum (for the Laplacian this is the twin-reduction theory of Barik--Reddy, which we present in detail). Besides surveying the literature, we record several consequences, and mark them as such: an exact decomposition of the eccentricity matrix of a connected cograph, yielding an eigenvalue-free interval for the eccentricity spectrum of all connected cographs; the full distance Laplacian spectrum and its extremal eigenvalues; an eigenvalue-free interval for the distance signless Laplacian valid for every connected graph of diameter at most $2$; an interlacing theorem tying the distance spectrum to the adjacency spectrum. A comparison table across all eight matrices and a list of open problems close the paper.
Related
BibTeX
@misc{manna2026spectral,
title = {{The Spectral Theory of Cographs: A Survey, Organized by Graph Matrix}},
author = {Pallabi Manna and Santanu Mandal and Fernando C. Tura},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.06981},
eprint = {2611.06981},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.06981},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - The Spectral Theory of Cographs: A Survey, Organized by Graph Matrix
AU - Pallabi Manna
AU - Santanu Mandal
AU - Fernando C. Tura
PY - 2026
DA - 2026/11/04
PB - arXiv
DO - 10.48550/arxiv.2611.06981
UR - https://arxiv.org/abs/2611.06981
AB - Cographs, the P_4-free graphs, are exactly the graphs obtained from a single vertex by repeated disjoint union and complementation. They contain the threshold graphs, π-graphs and form one of the primary areas for spectral graph theory. This survey is organized by matrix: after a preliminary section, there is one section for each of the adjacency, Laplacian, signless Laplacian, distance, eccentricity, Seidel, distance Laplacian and distance signless Laplacian matrices. Within each section we collect what is known for general cographs, for the subclass of π-graphs, and where relevant, for threshold graphs. We highlight the results from the recent years, including the foundational work of Mohammadian et al., Jacobs et al., Ghorbani, Li et al., Niu et al., Lou et al., Wang at al., Mandal et al. Two structural facts are used throughout: every connected cograph has diameter at most $2$, which ties all four distance-type matrices to the adjacency and Laplacian matrices, and the partition of the vertex set into twin classes is equitable for every one of the eight matrices, giving a uniform description of each spectrum (for the Laplacian this is the twin-reduction theory of Barik--Reddy, which we present in detail). Besides surveying the literature, we record several consequences, and mark them as such: an exact decomposition of the eccentricity matrix of a connected cograph, yielding an eigenvalue-free interval for the eccentricity spectrum of all connected cographs; the full distance Laplacian spectrum and its extremal eigenvalues; an eigenvalue-free interval for the distance signless Laplacian valid for every connected graph of diameter at most $2$; an interlacing theorem tying the distance spectrum to the adjacency spectrum. A comparison table across all eight matrices and a list of open problems close the paper.
ER -