The Gram matrix of a Temperley-Lieb algebra is similar to the matrix of chromatic joins
Qi Chen, Jozef H. Przytycki

TL;DR
This paper demonstrates that the matrix of chromatic joins and the Gram matrix of the Temperley-Lieb algebra are similar after rescaling, with a basis change given by diagonal matrices, revealing a structural connection.
Contribution
It establishes a similarity relation between the chromatic joins matrix and the Temperley-Lieb Gram matrix through a specific basis transformation.
Findings
Matrices are similar after rescaling
Change of basis is given by diagonal matrices
Reveals structural connection between two algebraic objects
Abstract
In this paper we show that the matrix of chromatic joins and the Gram matrix of the Temperley-Lieb algebra are similar (after rescaling), with the change of basis given by diagonal matrices.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Combinatorial Mathematics · Algebraic structures and combinatorial models · Advanced Topics in Algebra
The Gram matrix
of a Temperley-Lieb algebra is similar to the matrix of chromatic joins
Qi Chen and Jozef H. Przytycki
Dedicated to the memory of Xiao-Song Lin (1957-2007)
To appear in Communications in Contemporary Mathematics
1. Introduction
Rodica Simion noticed experimentally that matrices of chromatic joins (introduced by W. Tutte in [Tu2]) and the Gram matrix of the Temperley-Lieb algebra, have the same determinant, up to renormalization. In the type case, she was able to prove this by comparing the known formulas: by Tutte and R. Dahab [Tu2, Dah], in the case of chromatic joins, and by P. Di Francesco, and B. Westbury [DiF, We] (based on the work by K. H. Ko and L. Smolinsky [KS]) in the Temperley-Lieb case; see [CSS]. She then asked for a direct proof of this fact [CSS], [Sch], Problem 7.
The type analogue was an open problem central to the work of Simion [Sch]. She demonstrated strong evidence that the type Gram determinant of the Temperley-Lieb algebra is equal to the determinant of the matrix of type chromatic joins, after a substitution similar to that in type , cf. [Sch].
In this paper we show that the matrix of chromatic joins and the Gram matrix of the Temperley-Lieb algebra are similar (after rescaling), with the change of basis given by diagonal matrices. More precisely we prove the following two results:
Theorem A**.**
We have , where is a diagonal matrix with ; here denotes the number of blocks in the type non-crossing -partition ; see 2.1 of Section 2 for precise definitions.
Theorem B**.**
We have where is a diagonal matrix with ; here denotes half of the number of non-zero blocks in the type non-crossing -partition ; see 2.2 of Section 2 for precise definitions.
2. Definitions and notation
In the description of the matrix of chromatic joins (of type and ) we follow V. Reiner, R. Simion and F. Schmidt [Rei, Sim, Sch].
2.1. The type case
An -partition of type is a partition of the element set into blocks. The number of blocks is denoted by . To represent pictorially, we place the numbers anti-clockwise around the boundary circle of the unit disk and draw a chord, called a connection chord, in the disk between two numbers if they are in the same block of and there is no in the same block with . We say that is non-crossing if all connection chords can be drawn without crossing each other. Notice that each block is represented by a tree. Denote the set of all non-crossing -partitions of type by . On the other hand if is even, we have bipartitions of points of type , those -partitions of type with every block containing exactly 2 numbers. Denote the set of all non-crossing -bipartitions of type by . We have a bijection realized by considering the boundary arcs of a regular neighborhood of the connection chords (see Fig. 1 and compare Fig. 2, Fig. 3).
2.2. The type case
An -partition of type is a partition of the element set into blocks with the property that for any block of , its opposite is also a block of , and that there is at most one invariant block (called the zero block†††For topologists the term invariant block is more natural than zero block so we use this names interchangeably in the paper.) for which . Since all non-zero blocks occur in pairs one defines as half of the number of all non-zero blocks. To represent pictorially, we place the numbers anti-clockwise around the boundary circle of a disk and draw a connection chord in the disk between two numbers ‡‡‡Here we use the order . if they are in the same block of and there is no in the same block with . Then is said to be non-crossing if all connection chords can be drawn without crossing each other.§§§The non-crossing condition forces a partition to have at most one zero block. Denote the set of all non-crossing -partitions of type by . The set is illustrated in Fig. 2.
On the other hand if is even, we have bipartitions of points of type , those -partitions of type with every block containing exactly 2 numbers. Denote the set of all non-crossing -bipartitions of type by . Similar to the type case we have a bijection realized by considering the boundary arcs of a regular neighborhood of the connection chords (see Fig. 3).
2.3. The matrices
For any -partitions and of type (resp. ), denote by the finest -partition (not necessarily non-crossing) of type (resp. ) that is coarser than both and . The matrix of chromatic joins of type and are respectively:
[TABLE]
For any -bipartitions and of type (resp. ), one can glue them along the boundary circles respecting the labels. The result, denoted , is a collection of disjoint circles on a 2-sphere. The Gram matrix of Temperley-Lieb algebra of type and are respectively¶¶¶The matrix was first used by H. Morton and P. Traczyk to find a basis of the Kauffman bracket skein module of a tangle [MT], and played an important role in Lickorish’s approach to Witten-Reshetikhin-Turaev invariants of 3-manifolds [Li]. The matrix was first considered by Rodica Simion in 1998; compare [Sch].:
[TABLE]
where is the number of circles, is the number of zero (i.e. invariant) circles with , and is the number of pairs of non-zero circles with in .
3. Proof of Theorems A and B
Proof of Theorem A.
For let . Since is also equal to the number of boundary components of the regular neighborhood of the pictorial representation of we have . The formula can be obtained from the expression for the Euler characteristic of a plane graph: Let be a graph corresponding to the non-crossing partition . is a forest of vertices and edges. Similarly, let be the graph corresponding to . We should stress that does not have to be a noncrossing partition and that the graph is a plane graph obtained by putting inside a disk and outside the disk with composed of the points on the unit circle
(e.g.:
, or
).
By construction, is a plane graph of vertices and components. It has edges. Furthermore, if we embed in a disjoint union of 2-spheres (each component of in a different sphere) we can identify with the number of regions of the embedded graph. The Euler characteristic is on the one hand equal to and on the other hand equal to , as needed.
Theorem A follows directly from the formula. ∎
Proof of Theorem B.
For let and be the number of zero blocks of . Recall that is the number of zero (i.e. invariant) components of . As in the type case we have . Furthermore, we have (see Lemma 1):
[TABLE]
(Notice that can be equal to , , or [math].) From these we conclude that:
. Thus Theorem B follows. ∎
Lemma 1**.**
The zero blocks and zero components satisfy the following identity:
[TABLE]
where and .
Proof.
The lemma reflects the basic properties of a 2-sphere with an involution fixing two points and its compact invariant submanifolds.
To demonstrate the formula we consider all cases of blocks of , and divided into four classes:
(1) If is a non-zero (i.e. non-invariant) block of then all its constituent blocks in and are non-zero blocks and the boundary components of a regular neighborhood of the geometric realization of (denoted ), that is circles in corresponding to , are non-invariant (non-zero) curves, i.e. not in .
(2) If is a zero block of but all its constituent blocks in and are non-zero blocks, then exactly two components of are invariant curves.
(3) If is a zero block of and exactly one constituent block is a zero block then exactly one component of is an invariant curve.
(4) If is a zero block of and exactly two constituent blocks are zero-blocks (necessarily one in and one in ) then no component of is an invariant curve.
These conditions taken together prove the formula in Lemma 1. ∎
4. Corollaries
Theorem B and the results of [MS, CP] allow us to answer Problems 1 and 2 of [Sch] about a formula for the determinant of the type- matrix of chromatic joins:
Corollary 2**.**
[TABLE]
where is the Chebyshev polynomial of the first kind:
[TABLE]
The matrix can be generalized to a matrix of two variables as follows:
[TABLE]
It follows from Lemma 1 that
[TABLE]
where is a diagonal matrix with . Furthermore, ******This follows from Proposition 3 of [Rei], which asserts that there exists a fixed-point free involution on such that and .
and thus we have:
Corollary 3**.**
[TABLE]
Remark 4**.**
Consider the Gram matrix of type based on non-crossing connections in an annulus.††††††This interpretation of the Gram matrix of type B Temperley-Lieb algebra is mentioned in [Sch] as an annular skein matrix and utilized in [MS] and [CP]. This matrix is the same as the one considered before in Theorem B via the branched cover described in Fig. 4.**
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[CP] Q. Chen, J. H. Przytycki, The Gram determinant of the type B Temperley-Lieb algebra, e-print: http://arxiv.org/abs/0802.1083
- 2[CSS] A. Copeland, F. Schmidt, R. Simion, Note on two determinants with interesting factorizations, Descrete Mathematics , 256:449–458, 2002.
- 3[Dah] R. Dahab, The Birkhoff-Lewis equation, Ph D dissertation, University of Waterloo, 1993.
- 4[Di F] P. Di Francesco, Meander determinants, Comm. Math. Phys. , 191:543–583, 1998.
- 5[KS] K. H. Ko, L. Smolinsky, A combinatorial matrix in 3-manifold theory, Pacific Journ. Math. , 149(2), 1991, 319-336.
- 6[Li] W. B. R. Lickorish, Invariants for 3-manifolds from the combinatorics of the Jones polynomial, Pacific Journ. Math. , 149(2):337–347, 1991.
- 7[MS] P. P. Martin, H. Saleur, On an Algebraic Approach to Higher Dimensional Statistical Mechanics, Commun. Math. Phys. 158, 1993, 155-190; e-print: http://front.math.ucdavis.edu/9208.3061
- 8[MT] H. R. Morton, P. Traczyk, Knots and algebras, Contribuciones Matematicas en homenaje al profesor D.Antonio Plans Sanz de Bremond , ed. E.Martin-Peinador and A.Rodez Usan, University of Zaragoza, pp. 201–220, 1990.
