Multi-switches and representations of braid groups
Valeriy Bardakov, Timur Nasybullov

TL;DR
This paper introduces the concept of multi-switches to construct new algebraic representations of virtual braid groups, extending previous methods and broadening the understanding of braid group representations.
Contribution
It generalizes the notion of switches to multi-switches and provides a new framework for constructing braid group representations using automorphisms.
Findings
New representations of virtual braid groups introduced
Generalized several known braid group representations
Provided a systematic approach for constructing algebraic representations
Abstract
In the paper, we introduce the notion of a (virtual) multi-switch which generalizes the notion of a (virtual) switch. Using (virtual) multi-switches we introduce a general approach on how to construct representations of (virtual) braid groups by automorphisms of algebraic systems. As a corollary, we introduce new representations of virtual braid groups which generalize several previously known representations.
Peer 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
TopicsGeometric and Algebraic Topology · Nonlinear Waves and Solitons · Homotopy and Cohomology in Algebraic Topology
Multi-switches and representations of braid groups
Valeriy Bardakov, Timur Nasybullov
Abstract.
In the paper we introduce the notion of a (virtual) multi-switch which generalizes the notion of a (virtual) switch. Using (virtual) multi-switches we introduce a general approach how to construct representations of (virtual) braid groups by automorphisms of algebraic systems. As a corollary we introduce new representations of virtual braid groups which generalize several previously known representations.
Keywords: braid group, virtual braid group, quandle, representation by automorphisms, linear representation.
Mathematics Subject Classification: 20F36, 20F29, 20N02, 16T25.
1. Introduction
A set-theoretical solution of the Yang-Baxter equation is a pair , where is a set and is a bijective map such that
[TABLE]
The problem of studying set-theoretical solutions of the Yang-Baxter equation was formulated by Drinfel’d in [17]. If is a set-theoretical solution of the Yang-Baxter equation, then the map is called a switch on (see [19]). A pair of switches on is called a virtual switch on if and the equality
[TABLE]
holds.
Switches and virtual switches are strongly connected with virtual braid groups and virtual links. Using a (virtual) switch on a set it is possible to construct a representation of the (virtual) braid group on strands by permutations of (see [19, Section 2]). If is an algebraic system, then under additional conditions it is possible to construct a representation of the (virtual) braid group by automorphisms of . The Artin representation (see [13, Section 1.4]), the Burau representation (see [13, Section 3]) and their extensions to the virtual braid groups , (see [7, 40]) can be obtained on this way.
Despite the fact that virtual switches can be used for constructing representations of virtual braid groups, there are representations , where is some group, which cannot be obtained using any virtual switch on . For example, the Silver-Williams representation (see [39]), the Boden-Dies-Gaudreau-Gerlings-Harper-Nicas representation (see [14]), the Kamada representation (see [10]) and the representations , of Bardakov-Mikhalchishina-Neshchadim (see [7, 8]) cannot be obtained using any virtual switch.
In the present paper we introduce the notion of a (virtual) multi-switch which generalizes the notion of a (virtual) switch. Using (virtual) multi-switches we introduce a general approach how to construct representations of (virtual) braid groups by automorphisms of algebraic systems. As a corollary, we introduce new representations of virtual braid groups which generalize several previously known representations.
The paper is organized as follows. In Section 2, we give necessary preliminaries. In particular, we recall the notion of a (virtual) switch, and provide examples and applications of (virtual) switches. In Section 3, we introduce the notion of a (virtual) multi-switch and give examples of (virtual) multi-switches. In Section 4, we describe a general construction of how a (virtual) multi-switch on an algebraic system can be used to construct a representation of the (virtual) braid group by automorphisms of (Theorem 1 and Theorem 2). As a corollary, we construct a representation of the virtual braid group by automorphisms of the free product of the free quandle with generators and the trivial quandle on elements. Finally, in Section 5, we introduce representations of the (virtual) braid groups by automorphisms of certain infinitely generated abelian groups, which extend the Burau representation and the Gassner representation (Theorem 4 and Theorem 5).
Acknowledgement
The authors thank Professor Kauffman for useful discussions.
2. Preliminaries
In this section we give necessary preliminaries. We recall the notion of a switch from [19], and show some examples and applications of switches. We use classical notations. If is a group, and , then we denote by the conjugate of by , and by the conjugate of by . If is a set, then we denote by the set of all bijections from to . All actions are supposed to be right, i. e. if , then for we denote .
2.1. Switches and the Yang-Baxter equation
A set-theoretical solution of the Yang-Baxter equation is a pair , where is a set, and is a bijective map such that
[TABLE]
If is a set theoretical solution of the Yang-Baxter equation, then the map from is called a switch on (see [19, Section 2]). A switch is called involutive if . If is not just a set but an algebraic system: group, module etc, then every switch on is called a group switch, a module switch etc. Examples of switches on different algebraic systems follow.
Example 1**.**
Let be an arbitrary set, and for all . The map is a switch which is called the twist. It is clear that is involutive.
Example 2**.**
Let be a group. Then the map defined by for is a switch which is called the Artin switch on a group .
Example 3**.**
Let be a module over an integral domain , and be an invertible element of . The linear isomorphisms given by
[TABLE]
is a module switch on which is called the Burau switch. If we look on as on abelian group (without module structure), then the Burau switch is also a group switch.
In order to introduce the next example of a switch, let us recall that a quandle is an algebraic system with one binary algebraic operation which satisfies the following axioms:
- (1)
for all , 2. (2)
the map is a bijection of for all , 3. (3)
for all .
A quandle is called trivial if for all , the trivial quandle with elements is denoted by . Quandles were introduced in [24, 34] as an invariant for links. For more details about quandles see [15, 18, 37]. If is a quandle, then for we denote by . The following example gives a quandle switch.
Example 4**.**
Let be a quandle. Then the map for is a quandle switch.
The following example gives a switch on a skew brace (see, for example, [22, 35] for details about skew braces).
Example 5**.**
Let be a skew brace with the operations . Then the map given by
[TABLE]
for is a skew brace switch.
For more examples of switches see [19, Section 2].
2.2. Switches and representations of braid groups
Switches can be used for constructing representations of braid groups. Let us recall the definitions. The braid group on strands is the group with generators and defining relations
[TABLE]
There exists a homomorphism from the braid group onto the symmetric group on letters. This homomorphism maps the generator to the transposition for . The kernel of this homomorphism is called the pure braid group on strands and is denoted by .
The virtual braid group is a group obtained from adding new generators and additional relations
[TABLE]
It is easy to verify that the elements generate the symmetric group in . Also it is known that the elements generate the braid group in . The homomorphism can be extended to the homomorphism by the rule . The kernel of this homomorphism is called the virtual pure braid group and is denoted by .
Let be a switch on . For denote by
[TABLE]
From the relations of and equality (1) we see that the map which maps to for defines a representation of the braid group into the symmetric group . If is involutive, then the map defines a representation of the symmetric group on letters into the group .
Under additional conditions switches can provide representations of braid groups by automorphisms of algebraic system. If is an algebraic system (for example, a quandle, a group, a module etc) generated by elements , and is a switch on with
[TABLE]
for , then for denote by the map given by
[TABLE]
If for the maps induce automorphisms of , then the map which maps to for induces a representation . Note that if is a free in some variety group with the canonical generators , then the maps always induce automorphisms of . The Artin representation (see [13, Corollary 1.8.3]) can be obtained in this way using the Artin switch on . The Burau representation (see [13, Section 3]) can be obtained in this way using the Burau switch on . It is important here that the algebraic system has exactly generators.
Let be a switch and an involutive switch on , respectively. We say that the pair is a virtual switch on if the equality
[TABLE]
holds. If the pair of switches satisfies equality (2), then we say that this pair is matched.
Example 6**.**
If is a group, then is a virtual group switch on .
For a virtual switch on and an integer denote by
[TABLE]
for . From the relations of and equalities (1), (2) we see that the maps , induce a representation of the virtual braid group on strands into the symmetric group .
If is an algebraic system generated by elements , and is a virtual switch on with
[TABLE]
for , then for denote by the maps given by
[TABLE]
If induce automorphisms of , then the maps , induce a representation .
2.3. Biquandles
Let be a switch on such that
[TABLE]
For denote by , , so, on we have two binary algebraic operations , , which are called the up operation and the down operation defined by , respectively. The Yang-Baxter equation for implies the following equalities
[TABLE]
for all . A switch is called a biquandle switch if the following conditions hold.
- (1)
The maps given by , are bijective. We denote by , . 2. (2)
and for all .
If is a biquandle switch on , then the set with the up and the down operations defined by is called a biquandle and is denoted by . If is an arbitrary set, then the twist on is a biquandle switch. The biquandle is called the trivial biquandle on . In the trivial biquandle we have for all .
Biquandles were introduced in [19] as a tool for constructing invariants for virtual knots and braids. Papers [16, 26, 20, 23] give several application of biquandles in knot theory.
3. Multi-switches and virtual multi-switches
In this section we consider a special kind of (virtual) switches, called (virtual) multi-switches, which help to provide new representations of (virtual) braid groups.
3.1. Multi-switches
Let be a set, and be non-empty subsets of . We say that a map is an -switch, or a multi-switch on (if is not specified) if is a switch on (we identify the sets and , so, for -tuples , from we write ), such that
[TABLE]
for , for , where are the maps
[TABLE]
If is an -switch on defined by the maps , then we write . Note that for the map is a switch on . If is not just a set but an algebraic system: group, module etc, then every multi-switch on is called a group multi-switch, a module multi-switch etc. We do not require here that are subsystems of .
Example 7**.**
Every switch on is a -switch on .
Example 8**.**
If is a switch on , and is a switch on for , then the map is an -switch on .
The multi-switches from Example 7 and Example 8 are in some sense trivial. The following proposition provides a module -switch which generalizes the Burau switch. If is a free module over an integral domain , then we can think about any subset of as about subset of (thinking about as about for a fixed non-zero element from ).
Proposition 1**.**
Let be an integral domain, be a free module over , and be a subset of the multiplicative group of . Then the map given by
[TABLE]
is a -switch on .
Proof.
Using direct calculations we see that the map is given by
[TABLE]
therefore is bijective, and the only moment we have to check is that equality (1) holds for . Denote by , . Using direct calculations for , we have
[TABLE]
i. e. , and is a -switch on . ∎
If in Proposition 1 we look on as on abelian group, then the module -switch is also a group -switch. Under conditions of Proposition 1, let be a fixed invertible element from , and . Then for elements we have . Looking only to the first two components we obtain the Burau switch. In this sense, the switch generalizes the Burau switch.
Since an -switch on is a switch on , the notions of the involutive -switch and the virtual -switch follow from the same notions for switches.
3.2. Virtual multi-switches
Usually, if is a virtual switch on a set , then is the twist, and it is quite difficult to find a virtual switch with . V. Manturov [33] found a virtual quandle switch such that is not the twist (see also [4, 6]). Let be a quandle, be the trivial quandle with one element, and the free product of and (see more about free products of quandles in [9]). Let be the quandle switch (from Example 4) on , and the involutive switch with for . Then is a virtual switch on with . Using this virtual switch Manturov [33] constructed a quandle invariant for virtual links which generalizes the quandle of Kauffman [25].
In this subsection we construct two virtual multi-switches with : the first one is the virtual module -switch which extends the -switch introduced in Proposition 1; the second one is the virtual -switch on a biquandle which leads to a virtual quandle -switch which generalizes the virtual switch of Manturov.
The following proposition provides a virtual module -switch.
Proposition 2**.**
Let be an integral domain, be the free module over , and be subsets of the multiplicative group of . Then the pair of maps from to itself given by
[TABLE]
for , , is a virtual -switch on .
Proof.
It is clear that the maps , are invertible. Since acts as the -switch introduced in Proposition 1 on the first four arguments, and as the twist on the last two arguments, i. e. , the map is a -switch on . So, we need to check that is an involutive -switch on , and that the pair is matched. Denote by , , , . For , , we have
[TABLE]
therefore is a -switch on . The equality
[TABLE]
implies that is involutive. The equalities
[TABLE]
imply that is a matched pair of -switches. Therefore is a virtual -switch on . ∎
In the proof of Proposition 2 we noticed that . In this sense, the virtual -switch extends the -switch . The following proposition provides a -switch on a biquandle.
Proposition 3**.**
Let be a biquandle, be a trivial subbiquandle of , and be the maps from to itself such that
[TABLE]
for , . If for all , the equalities
[TABLE]
hold, then is a virtual -switch on .
Proof.
Denote by , , , . Since acts separately on the first pair of arguments, and separately on the second pair of arguments, the fact that is a -switch on follows from the fact that the map is a switch on , and the map is a switch on . Using direct calculations for , we have
[TABLE]
These equalities together with equalities (3) and the fact that is a trivial biquandle imply that , i. e. is a -switch on . The equalities
[TABLE]
imply that is involutive, and the only fact we have to prove is that the pair is matched. For , we have
[TABLE]
which together with equalities (4) from the conditions of the proposition imply the equality . ∎
Corollary 1**.**
Let be a quandle with a trivial subquandle , and be the maps from to itself defined by
[TABLE]
for , . Then is a virtual -switch on .
Proof.
For denote by , . Then with the operations , is a biquandle, and is a trivial subbiquandle of this biquandle. Equalities (4) obviously hold in this situation, and the result follows from Proposition 3. ∎
Corollary 2**.**
Let be a quandle, be a trivial quandle, and be the free product of and . Let be the maps defined by
[TABLE]
for , . Then is a virtual -switch on .
If in Corollary 2 we take a trivial quandle with only one element, then looking to the first two components of the maps we obtain the switch of Manturov described at the beginning of this section.
4. Multi-switches and representations of braid groups
In Section 2.2 we noticed that a (virtual) switch on an algebraic system with exactly generators can be used to construct a representation of the (virtual) braid group on strands by automorphisms of . However, there are representations (where is some group) which cannot be defined using procedure described in Section 2.2 for any virtual switch on . For example, let be the free group with the free generators , be the free abelian group with canonical generators , , and be the free product of and , then the representation introduced in [7, 8] which acts on the generators of in the following way
[TABLE]
(hereinafter we write only non-trivial actions on the generators, assuming that all other generators are fixed) cannot be defined using procedure described in Section 2.2 for any virtual switch on (since has generators).
The representation , where is the free product of the free group generated by , and the free abelian group generated by introduced in [7] which acts on the generators of by the rule
[TABLE]
gives another example of the representation of which cannot be defined using procedure described in Section 2.2 for any virtual switch on . Note that the representations and given in (5), (6) have the same kernel [7].
In this section we describe a general construction of how a (virtual) multi-switch on an algebraic system can be used to construct a representation of the (virtual) braid group on strands by automorphisms of . We do not require that has exactly generators, so, a lot of known representations of virtual braid groups by automorphisms of groups (in particular, representations , given by formulas (5), (6), respectively) can be constructed using the procedure which we will introduce in this section.
Since a virtual -switch on an algebraic system is a switch on , from the first glance it can seem that one can apply the procedure from Section 2.2 in order to construct a representation of the virtual braid group by automorphisms of . However, it is not true due to the fact that is a virtual switch on (but not on ), and the set doesn’t necessarily have a structure of algebraic system.
4.1. General construction
Let be an algebraic system, and be subsets of such that
- (1)
for the set contains elements , 2. (2)
for , 3. (3)
the set of elements generates .
Let be an -switch on such that
[TABLE]
and for the images of maps are words over its arguments in terms of operations of . For denote by the following map from to
[TABLE]
where all generators which are not explicitly written in are fixed, i. e. for , , and assume that is well defined: since the elements are not necessary all different, some of these elements can coincide. The fact that is well defined means that the images of equal elements are equal. For example, if , then we assume that
[TABLE]
If for the maps , induce automorphisms of , then we say that is an automorphic multi-switch (shortly, AMS) or an automorphic -switch on with respect to the set of generators . An -switch can be AMS with respect to one generating set of , but not AMS with respect to another generating set of .
Theorem 1**.**
Let be an algebraic system, and be an AMS on with respect to the set of generators . Then the map
[TABLE]
which is defined on the generators of as
[TABLE]
where is defined by equality (7), is a representation of .
Proof.
We need to check that the automorphisms satisfy defining relations of . Relations () follow from the fact that is an -switch on . In order to check relations () let be such that . For the map we have
[TABLE]
where is a word over in terms of operations of . Since fixes elements , we have for , . Therefore
[TABLE]
where is a word over in terms of operations of . The last equality can be rewritten in the form
[TABLE]
In a similar way we can prove that can be written in the same form, therefore for . ∎
Similar to Theorem 1 result can be formulated for virtual -switches and representations of virtual braid groups. Let , be a virtual -switch on (i. e. the pair is matched) such that
[TABLE]
and for the images of maps are words over its arguments in terms of operations of . For denote by the map from to given by equality (7), and by the following map from to .
[TABLE]
and suppose that are well defined. If for the maps , induce automorphisms of , then we say that is an automorphic virtual multi-switch (shortly, AVMS) or an automorphic virtual -switch on with respect to the set of generators .
Theorem 2**.**
Let be an AVMS on with respect to the set of generators . Then the map
[TABLE]
which is defined on the generators of as
[TABLE]
where , are defined by equalities (7), (8), is a representation of .
Proof.
The proof repeats the proof of Theorem 1 adding a few details. ∎
A lot of known representations of the virtual braid group by automorphisms of groups can be obtained using Theorem 2. For example, the representation given in equalities (5) is obtained in the following way: denote by , , , , , and for denote by , , , . Then the representation is the representation for the virtual -switch on given by
[TABLE]
for , , , (the fact that is a virtual -switch on follows from the fact that is a representation of the virtual braid group).
The representation given in equality (9), the Silver-Williams representation (see [39]), the Boden-Dies-Gaudreau-Gerlings-Harper-Nicas representation (see [14]), and the Kamada representation (see [10]) can be obtained using Theorem 2 in a similar to way.
In the following subsection using Theorem 2 we construct a representations of the virtual braid groups by automorphisms of quandles.
4.2. Representations by automorphisms of quandles
Let be the free quandle on generators , be the trivial quandle on elements , and be the free product of and . From Corollary 2 we know that the maps defined by
[TABLE]
for , form a virtual -switch on . Due to the fact that is a free product of the free quandle and the trivial quandle, it is easy to see that is an automorphic virtual -switch on with respect to the set of generators , . So, Theorem 2 and Corollary 2 imply the following result.
Theorem 3**.**
Let be the free quandle on the set of generators and be the trivial quandle. Then the map given by
[TABLE]
induces a homomorphism .
Proof.
Denote by , for , , , , and the virtual -switch on defined in Corollary 2. Then the map described in the formulation of the theorem is the map which is a homomorphism by Theorem 2. ∎
The representation given in Theorem 3 generalizes the representation given by formulas (6). Note that has generators, so, the representation given in Theorem 3 cannot be obtained using procedure described in Section 2.2. The following proposition gives a representation of the virtual braid group by automorphisms of the free quandle with generators.
Proposition 4**.**
Let be the free quandle with generators . Then the map given by
[TABLE]
induces a homomorphism .
Proof.
Denote by , for , , , , the virtual -switch on defined in Corollary 2. Since is the free quandle, and the maps from equality (10) are invertible, induce automorphisms of , therefore is AVMS with respect to the set of generators (the maps from (10) are exactly the maps , from (7), (8) for ). Then the map described in the formulation of the proposition is the map which is a representation by Theorem 2. ∎
The representation given in Proposition 4 generalizes the representation introduced in [4].
5. Multi-switches and linear representations of braid groups
The question of whether the group is linear was a long standing problem. In 1936 Burau constructed a linear representation (see, for example, [13, Section 3]) which is given on the generators of by the following equality
[TABLE]
This representation can be obtained from the Burau switch (Example 3) using procedure described in Section 2.2. For the Burau representation in known to be faithful, for the Burau representation is known to have non-trivial kernel [11], for the question on faithfulness of the Burau representation is still open.
The Burau representation was generalized in 1961 by Gassner [21] to a related representation of the pure braid group . Let us recall the definition. Let be the free group with the free generators . For an element and the generator denote by the Fox derivative of by (see [13, Section 3.1]). The Artin representation of the braid group by automorphisms of is defined on the generators of by the following rule
[TABLE]
for . Denote by the abelianization map, and let for . Denote by the same symbol the induced homomorphism . The homomorphism given on the generators
[TABLE]
of by the rule
[TABLE]
is called the Gassner representation . If we put , then we obtain the restriction of the Burau representation to . Denote by the natural basis in , then the Gassner representation is given by the formulas
[TABLE]
(see, for example, [28] and [13, Section 3.2]).
It is not known whether the Gassner representation is faithful for . Some results concerning the faithfulness of the Gassner representation can be found, for example, in [1, 28].
The final answer to the question about the linearity of the braid groups was given independently by Bigelow [12] and Krammer [30] who proved that the representation introduced by Lawrence in [31] is faithful. The representation is called now the Lawrence-Krammer-Bigelow representation.
The situation with virtual braid groups is more complicated. It is not difficult to prove that the braid group on strands is linear.
Proposition 5**.**
* is linear.*
Proof.
The kernel of of the endomorphism , which is defined on the generators of by the rule
[TABLE]
is a normal subgroup of index in . So, is linear if and only if is linear and it is enough to prove that is linear. The presentation of by generators and defining relations was found in master thesis of Rabenda [38] (see also [6]). This group is generated by the elements
[TABLE]
and is defined by the relations
[TABLE]
where distinct letters stand for distinct indices. From this presentation follows (see [6, Remark 20]) that is a free product of groups
[TABLE]
The groups are isomorphic to the circular braid group on strands [2], which can be embedded into the braid group on strands [27]. Since is linear, the groups are linear. Since the free product of two linear groups is linear [36], we conclude that is linear, and therefore is linear. ∎
In [5] the Gassner representation was extended to the group of basis-conjugating automorphisms, which is isomorphic to the pure welded braid group . This extension is not faithful for . In [40] the Burau representation was extended to the welded braid group . The question about the linearity of for is formulated in [29, Problem 19.7(b)]. In general, it is not known if the virtual braid groups are linear or not. Also, there are no good linear representation of these groups.
In Section 4.1 we introduced a general construction how a (virtual) multi-switch on an algebraic system with finitely many generators can be used for constructing a representation of the (virtual) braid group by automorphisms of (Theorem 1 and Theorem 2). Sometimes using the same approach it is possible to construct representations of (virtual) braid groups by automorphisms of infinitely generated algebraic systems. In this section we introduce representations of (virtual) braid groups by automorphisms of some infinitely generated abelian groups. These representations lead to the linear representations of pure (virtual) braid groups which are strongly related with the Burau representation and the Gassner representation
5.1. Representations of and
Let be a positive integer, and be the free left module with the free basis over the ring . Denote by the additive group of , by the subgroup of generated by , and by (similarly to Proposition 1 we can assume that ). Let be the -switch on from Proposition 1
[TABLE]
for , . For denote by the following map from to
[TABLE]
Note that the map from (14) is the same as the map from equality (7) if we denote by , for . The map induces an automorphism of by the rule
[TABLE]
where , and denotes the image of under the map given by permutation induced by . Denote by the map from the set of generators of to which maps to for .
Theorem 4**.**
The map induces a representation of the braid group . The restriction of to the pure braid group is a linear representation , which coincides with the Gassner representation .
Proof.
In order to prove that the map induces a representation of the braid group it is necessary to check that the maps satisfy the defining relations of . It can be checked by (a bit massive but not difficult) direct calculations in a similar to Theorem 1 way, and we will not do it here.
Since permutes and and fixes for , it is clear that fixes all . Hence for the automorphism , where
[TABLE]
is the generator of , fixes all . Therefore
[TABLE]
i. e. the restriction of to gives a representation (here we write instead of in order to underline that is a module, while is an abelian group).
In order to prove that the restriction of the representation to coincides with the Gassner representation it is enough to prove that for all , where is the generator of given by equality (16). We will prove this fact using induction on and equalities (11).
The basis of induction () is simple. From equality (14) follows that
[TABLE]
Comparing the last equality with equality (11) we see that , and the basis of induction is proved.
In order to prove the induction step note that acts by the following rule
[TABLE]
Suppose that we proved that for , and let us prove that . By the induction conjecture we have
[TABLE]
Using induction conjecture let us calculate the images for all . If or , then it is clear that
[TABLE]
since both and fix . For we have
[TABLE]
For we have
[TABLE]
For we have
[TABLE]
Note that equalities (21) and (22) coincide. Finally, for we have
[TABLE]
Comparing equalities (19), (20), (21), (22), (23) with equality (11) we see that the equality holds for all . From equality (18) follows that , so, the induction step is proved. ∎
Note that if we put , then the representation is the Burau representation . So, the representation extends both the Burau and the Gassner representations.
Despite the fact that the map from (14) has the same form as the map from (7), Theorem 4 does not follow from Theorem 1: the elements , do not form the generating set of (remember that we think about the set as about a subset of ), and in order to extend the map to the automorphism of we need to use formula (15). At the same time, in Theorem 1 the map was extended to the automorphism of just by action on the generators.
5.2. Representations of and
The representation of the braid group
[TABLE]
from Theorem 4 can be extended to the representation of the virtual braid group using the virtual -switch instead of the -switch . Let be a positive integer, and be the free left module with the free basis over the ring . Denote by the additive group of , by the subgroup of generated by , by , and by . Let be the virtual -switch on from Proposition 2
[TABLE]
for , , . For denote by , the following maps from to
[TABLE]
The maps , from (24) are the same as the maps , from equalities (7), (8) if we denote by , , for . Similarly to equality (15) the maps , induce automorphisms of by the rule
[TABLE]
where , and is the image of given by permutation induced by . Denote by the map from the set of generators of to which maps to , , respectively, for .
Theorem 5**.**
The map induces a representation of the virtual braid group . The restriction of to the pure virtual braid group is a linear representation .
Proof.
The proof is the same as the proof of the first part of Theorem 4. ∎
If in formulas (24) we put , then the map from equality (24) becomes the same as the map from equality (14), therefore the representation extends the representation . Hence, due to Theorem 4, the induced representation
[TABLE]
extends the Gassner representation . It is not known if is faithful or not. In the following proposition we prove that the representation which extends the Gassner representation has non-trivial kernel.
Proposition 6**.**
For the representation
[TABLE]
has non-trivial kernel.
Proof.
It is clear that it is enough to prove that the representation
[TABLE]
has non-trivial kernel, since the elements from the kernel for belong to the kernel for arbitrary . The group has a subgroup which has three generators
[TABLE]
and one defining relation
[TABLE]
(see, for example, [3]). Using direct calculations it is easy to see that
[TABLE]
From these equalities and direct calculations follows that
[TABLE]
From equalities (25), (26), (27) we see that the matrices of the linear transformations , , are upper triangular, therefore the group is solvable. However, has three generators and one relation, therefore by Magnus theorem [32, Section 4.4] any two elements from the set generate non-abelian free group. Therefore the induced representation
[TABLE]
has a non-trivial kernel ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] M. Abdulrahim, A faithfulness criterion for the Gassner representation of the pure braid group, Proc. Amer. Math. Soc., V. 125, N. 5, 1997, 1249–1257.
- 2[2] M. Albar, D. Johnson, The centre of the circular braid group, Math. Jpn., V. 30, 1985, 641–645.
- 3[3] V. Bardakov, The virtual and universal braids, Fund. Math., V. 184, 2004, 1–18.
- 4[4] V. Bardakov, Virtual and welded links and their invariants, Sib. Elektron. Mat. Izv., V. 2, 2005, 196–199.
- 5[5] V. Bardakov, Extending representations of braid groups to the automorphism groups of free groups, J. Knot Theory Ramifications, V. 14, N. 8, 2005, 1087–1098.
- 6[6] V. Bardakov, P. Bellingeri, Combinatorial properties of virtual braids, Topology Appl., V. 156, N. 6, 2009, 1071–1082.
- 7[7] V. Bardakov, Yu. Mikhalchishina, M. Neshchadim, Representations of virtual braids by automorphisms and virtual knot groups, J. Knot Theory Ramifications, V. 26, N. 1, 2017, 1750003.
- 8[8] V. Bardakov, Yu. Mikhalchishina, M. Neshchadim, Virtual link groups, Sib. Math. J., V. 58, N. 5, 2017, 765–777.
