Virtual Knot Groups

• Virtual Knot Group is a group-valued invariant defined via virtual braid groups and Wirtinger presentations, encoding both classical and virtual crossing data.
• It distinguishes classical knots from genuinely virtual ones by utilizing algebraic techniques such as lower central series and Alexander invariants.
• Extensions including augmented, welded, and parity-based groups provide deeper insights into knot topology and pose open problems in residual finiteness and categorification.

A virtual knot group is a group-valued invariant assigned to a virtual knot or link, generalizing the classical knot group and capturing topological information sensitive to virtual features. The concept encompasses a family of group invariants defined via various diagrammatic or braid-theoretic presentations, with several extensions and refinements, most commonly adapted to operate within the virtual knot category introduced by Kauffman. These invariants are central to the algebraic study of virtual knot theory, distinguish classical from genuinely virtual knots, and connect to deep questions about the structure and classification of virtual links.

1. Foundational Definitions and Presentations

The virtual knot group is most efficiently defined using the virtual braid group $VB_n$, which extends the classical braid group $B_n$ by adding virtual crossing generators $\rho_i$ that satisfy symmetric group and mixed relations with the classical generators $\sigma_i$ (Bardakov et al., 2012). Explicitly, $VB_n$ is generated by

$$\{\sigma_1,\dots,\sigma_{n-1}\} \quad \text{(classical)} \qquad \{\rho_1,\dots,\rho_{n-1}\} \quad \text{(virtual)},$$

with the mixed relation $$\rho_i \rho_{i+1} \sigma_i = \sigma_{i+1} \rho_i \rho_{i+1}$$ ensuring compatibility with virtual and classical crossing operations.

Given a virtual link $vL$ represented as the closure of a virtual braid $\beta_v \in VB_n$, the standard (Bardakov–Bellingeri) virtual knot group $G_v(vL)$ is defined as: $B_n$0 where $B_n$1 is given on generators by: $B_n$2 (Bardakov et al., 2012). This construction generalizes the classical knot group by incorporating the action of virtual crossings and introduces an additional central generator $B_n$3 encoding virtual structure.

An equivalent diagrammatic Wirtinger presentation is given by assigning generators to each (semi-)arc and imposing relations at classical crossings that reflect over/under-crossing relations and, in refined settings, additional automorphism generators to enforce invariance under the virtual detour move (Dye et al., 2021, Boden et al., 2015).

2. Invariance Properties and Reduction to Classical Cases

Virtual knot groups are invariant under virtual isotopy, including all generalized Reidemeister and virtual Markov moves. This is established by verifying that braid relations, conjugations, stabilizations, and virtual exchange moves induce bijective relabelings or Tietze-equivalent presentations (Bardakov et al., 2012). In particular, for a classical knot $B_n$4, the representation $B_n$5 reduces to the Artin action on $B_n$6, and the group specializes to

$B_n$7

where the extra factor $B_n$8 is central and corresponds to the virtual generator (Bardakov et al., 2012).

Kauffman’s original virtual knot group $B_n$9, defined by ignoring virtual crossings, can be recovered as a natural quotient: $\rho_i$0 demonstrating that virtual knot groups encode strictly more information than their classical analogs or Kauffman’s naive group invariant (Bardakov et al., 2012, Dye et al., 2021).

3. Families and Extensions of Virtual Knot Groups

Several generalizations and refinements have been developed:

• Extended and Augmented Virtual Knot Groups: Dye–Kaestner’s family of groups introduces extra central parameters and commuting automorphism relations, resulting in a lattice of invariants including the Boden group (with three central parameters) and the Silver–Williams extended group (with two parameters) (Dye et al., 2021, Boden et al., 2015). These subgroups interpolate between the classical group and the full parameterized virtual knot group.
• Welded and Quandle Groups: Variants obtained by further collapsing parameters, as in the welded group $\rho_i$1 or the quandle group $\rho_i$2, fit into a commutative diamond of quotients and provide knot invariants for welded and flat categories (Boden et al., 2015).
• Alternative Representations: Representations using automorphisms of $\rho_i$3 relate to the Bardakov–Mikhalchishina–Neshchadim construction, unifying previously known group-valued invariants within the same framework and yielding link groups with enhanced sensitivity to virtual features (Bardakov et al., 2016).
• Parity-based and Free-Group Invariants: Manturov and others assign invariants not only in free-abelian and automorphism groups, but also in free products of cyclic groups and groups associated with parity filtrations in Gauss diagrams. These are stable under virtualization and constitute Vassiliev finite-type invariants for long virtual knots (Manturov, 2010, Manturov, 2020).

4. Algebraic and Homological Structure

The algebraic structure of virtual knot groups diverges sharply from classical knot groups. Notably:

• Abelianization: Any virtual knot group (in the Bardakov–Bellingeri sense) abelianizes to $\rho_i$4 (Bardakov et al., 2012).
• Group Decomposition: For classical knots, $\rho_i$5 splits as a free product $\rho_i$6.
• Semi-Direct and HNN Extensions: Groups associated to virtual knots can exhibit semidirect product structure with free groups or be realized as HNN extensions and infinite amalgamated free products, as in Mikhalchishina’s families $\rho_i$7, $\rho_i$8, and $\rho_i$9 (Bardakov et al., 2018).
• Lower Central Series: Unlike classical knot groups (where the lower central series stabilizes early), virtual knot groups may have arbitrarily long lower central series. Explicit examples show torsion appearing in $\sigma_i$0 and nilpotency properties such as residual nilpotence that distinguish nonclassical knots (Bardakov et al., 2018, Bardakov et al., 2018).
• Obstructions via Alexander Invariants: For almost classical knots (those with an Alexander numbering), the reduced group splits as a free product of the classical knot group and $\sigma_i$1, and their Alexander polynomial exhibits a principal ideal property and skein relation (Boden et al., 2015).

5. Examples and Discriminatory Power

The virtual knot group is a powerful but not universal invariant. Key examples include:

Knot Type	Bardakov–Bellingeri Group $\sigma_i$2	Kauffman Group $\sigma_i$3	Discrimination
Unknot (classical/virtual)	$\sigma_i$4	$\sigma_i$5 or $\sigma_i$6	No virtuality detected
Virtual Trefoil	Non-free: $\sigma_i$7	Abelian or cyclic	Virtuality detected
Kishino Knot	Often $\sigma_i$8 (by $\sigma_i$9); $VB_n$0 distinguishes	$VB_n$1	Not detected by $VB_n$2, detected by refined invariants

For the virtual trefoil, all major definitions yield non-free groups, establishing its nontriviality in the virtual category (Bardakov et al., 2012, Dye et al., 2021, Boden et al., 2015, Bardakov et al., 2018). In contrast, for the Kishino knot, many group-valued invariants collapse to the free group; however, refined invariants such as Mikhalchishina’s $VB_n$3 do distinguish it from the trivial knot (Bardakov et al., 2018), and stack group constructions (as per Winter) can detect nontriviality for certain virtual knot stacks but fail for a specific subset (Winter, 2024).

6. Applications and Limitations

Virtual knot groups serve as a computationally accessible and algebraically robust method for distinguishing virtual knots and studying their properties. They provide:

• Group-theoretic detection of nonclassicality and nontriviality.
• Filtrations and quotients (lower central series, nilpotent quotients) as computable invariants more sensitive than many polynomial or combinatorial invariants (Bardakov et al., 2018, Bardakov et al., 2018).
• Obstructions to sliceness and almost classical character via Alexander invariants (Boden et al., 2015).
• Connections to quantum invariants in cases where group invariants fail (e.g., Kishino-type knots not detected by stack groups, requiring Jones polynomial computations) (Winter, 2024).

Limitations are also well-documented: in particular, for classes of virtual knots that are welded-trivial or admit both vertical and ordinary stacks with free group fundamental groups, no amount of group- or quandle-type stack invariants can distinguish them from the unlink, necessitating the use of quantum or biquandle methods (Winter, 2024).

7. Open Problems and Future Directions

Key open directions include:

• Faithfulness of the $VB_n$4 representation for $VB_n$5: this remains unresolved (Bardakov et al., 2012).
• Residual finiteness and nilpotency of general virtual knot groups: although certain cases are established, the question for arbitrary virtual knots is open (Bardakov et al., 2012, Bardakov et al., 2018, Bardakov et al., 2018).
• Uniqueness results: To what extent the sequence of nilpotent quotients or families of virtual knot groups uniquely determines virtual links up to isotopy (Bardakov et al., 2018).
• Extension to higher quantum and categorified invariants in the virtual and welded category.

This landscape highlights the centrality of virtual knot groups as algebraic invariants and their close connection with virtual knot topology, algebraic group theory, and low-dimensional topology (Bardakov et al., 2012, Dye et al., 2021, Boden et al., 2015, Bardakov et al., 2016, Manturov, 2020, Manturov, 2010, Bardakov et al., 2018, Bardakov et al., 2018, Winter, 2024).