Multi-Virtual Semi-Pure Braid Groups

• Multi-virtual semi-pure braid groups are kernel subgroups of multi-virtual braid groups where only virtual crossings determine the permutation, resulting in a normal subgroup of index n!.
• They are defined via an explicit Reidemeister–Schreier presentation that separates classical (type 0) from higher virtual types, underscoring distinct algebraic behaviors.
• Their semidirect-product structure with the symmetric group and links to flat and universal braid groups make them pivotal in advancing virtual braid theory and related quotient studies.

Multi-virtual semi-pure braid groups are kernel subgroups of multi-virtual braid groups in which the strand permutation is computed from the virtual crossings alone, while the classical crossings are ignored in the projection to the symmetric group. In the framework introduced after Kauffman’s multi-virtual braid theory, the group $M_kVH_n$ is defined as a normal subgroup of index $n!$ in the $k$-multi-virtual braid group $M_kVB_n$, admits an explicit presentation by Reidemeister–Schreier generators, and sits in a canonical semidirect-product decomposition with $S_n$ (Bardakov et al., 19 Jul 2025).

1. Position within multi-virtual braid theory

Kauffman’s $k$-multi-virtual braid group $M_kVB_n$ extends the virtual braid group by allowing one classical crossing generator $\sigma_i$ for each adjacent pair of strands and $k$ virtual types

$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$

The type $n!$0 plays a distinguished role: only crossings of type $n!$1 participate in detour moves, while detouring across types $n!$2 is forbidden. This produces the hierarchy

$n!$3

and for $n!$4 the flat virtual braid group $n!$5 embeds in $n!$6 (Bardakov et al., 19 Jul 2025).

The defining relations of $n!$7 combine braid relations for the $n!$8, Coxeter relations for each fixed virtual type, commutation of distant generators, and mixed relations involving the distinguished type $n!$9. In particular,

$k$0

$k$1

and

$k$2

The deliberate omission of certain mixed relations between different virtual types is part of the definition; these omissions encode the forbidden detour moves (Bardakov et al., 19 Jul 2025).

Within this ambient group, Bardakov–Kozlovskaya–Negi–Prabhakar define two distinct kernel subgroups. The multi-virtual pure braid group $k$3 kills the total permutation contributed by both classical and virtual crossings, whereas the multi-virtual semi-pure braid group $k$4 kills only the permutation recorded by the virtual crossings. The two groups are therefore parallel rather than identical constructions, and the semi-pure terminology refers to this asymmetric treatment of classical and virtual generators (Bardakov et al., 19 Jul 2025).

2. Definition by epimorphism to the symmetric group

Writing $k$5, the subgroup $k$6 is isomorphic to $k$7. The semi-pure structure is defined through the epimorphism

$k$8

for all $k$9 and all $M_kVB_n$0. The multi-virtual semi-pure braid group is

$M_kVB_n$1

Its image is $M_kVB_n$2, so $M_kVB_n$3 is a normal subgroup of index $M_kVB_n$4 (Bardakov et al., 19 Jul 2025).

This definition has a clear algebraic meaning. In the pure case,

$M_kVB_n$5

all crossings contribute to the endpoint permutation. In the semi-pure case, the classical generators $M_kVB_n$6 are sent to the identity, so only the virtual generators control the permutation. The paper summarizes the distinction by saying that in $M_kVB_n$7 the classical crossings are “ignored” in the permutation map, whereas in $M_kVB_n$8 they are identified with the same elementary transpositions as the virtual generators (Bardakov et al., 19 Jul 2025).

A common misconception is to read “semi-pure” as a weaker version of purity in the sense of an inclusion $M_kVB_n$9 or vice versa. What is established is more precise: the two groups are kernels of different surjections onto the same symmetric group. Their presentations coincide in some formal features but differ in the type-$S_n$0 relations, and this difference governs the algebraic behavior of the classical part (Bardakov et al., 19 Jul 2025).

3. Presentations and internal algebra

The semi-pure kernel $S_n$1 is described by generators $S_n$2 indexed by ordered strand pairs and virtual type. For type $S_n$3,

$S_n$4

and for $S_n$5,

$S_n$6

$S_n$7

For each higher virtual type $S_n$8,

$S_n$9

and $k$0 is obtained by the same conjugation pattern. Conjugation by the embedded symmetric group permutes the strand indices of these generators (Bardakov et al., 19 Jul 2025).

The resulting presentation of $k$1 has generators

$k$2

with three families of defining relations. First, there is commutativity for disjoint pairs and types: $k$3 whenever $k$4 are pairwise distinct and $k$5. Second, type $k$6 satisfies the semi-pure braid relation

$k$7

for distinct $k$8. Third, for each $k$9 one has flat-type pure relations

$M_kVB_n$0

for $M_kVB_n$1 (Bardakov et al., 19 Jul 2025).

This presentation isolates the algebraic asymmetry between type $M_kVB_n$2 and the higher virtual types. The subgroup generated by the $M_kVB_n$3 is isomorphic to the classical virtual semi-pure group $M_kVB_n$4, while for each $M_kVB_n$5 the subgroup generated by the $M_kVB_n$6 is isomorphic to the flat virtual pure braid group $M_kVB_n$7. The type-$M_kVB_n$8 sector is therefore Artin-like in the semi-pure sense, whereas each higher virtual sector is flat-virtual in character (Bardakov et al., 19 Jul 2025).

4. Semidirect-product structure and symmetric variants

The ambient multi-virtual braid group splits over both kernel constructions: $M_kVB_n$9 In the semi-pure case, the complement $\sigma_i$0 is realized by $\sigma_i$1, and its action on $\sigma_i$2 is the permutation action on the strand indices of the generators $\sigma_i$3. This supplies a canonical splitting of $\sigma_i$4 and makes $\sigma_i$5 a normal subgroup of finite index with explicitly controlled conjugation action (Bardakov et al., 19 Jul 2025).

The same pattern persists for the symmetric multi-virtual braid group $\sigma_i$6, the quotient in which detour moves are allowed for all virtual types. Its semi-pure kernel

$\sigma_i$7

again has index $\sigma_i$8, but its presentation is obtained from that of $\sigma_i$9 by adjoining extra mixed relations. For $k$0,

$k$1

and for $k$2,

$k$3

These additional relations strengthen the interaction among virtual types and formalize the “symmetrizing” of detour behavior (Bardakov et al., 19 Jul 2025).

A useful conceptual summary is that non-symmetric multi-virtual semi-pure groups encode a hierarchy of virtualities, while symmetric multi-virtual semi-pure groups collapse much of that hierarchy by permitting type-independent detouring. The distinction is algebraically visible in the presence or absence of the additional mixed relations above (Bardakov et al., 19 Jul 2025).

5. Low-rank structure and neighboring quotient theories

The case $k$4 is especially explicit. Bardakov–Kozlovskaya–Negi–Prabhakar show

$k$5

Thus the three-strand semi-pure group decomposes as a free product of the ordinary virtual semi-pure group and $k$6 flat virtual pure factors. This gives a concrete combinatorial model of how the higher virtual types enlarge the classical semi-pure structure (Bardakov et al., 19 Jul 2025).

For the $k$7-virtual three-strand group $k$8, with $k$9, the semi-pure kernel $\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$0 is generated by a classical sector

$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$1

together with the type-$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$2 flat-virtual sector

$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$3

and their oppositely oriented analogues. The relation

$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$4

captures the flat-type component, while the type-$\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$5 generators satisfy the six semi-pure Artin relations inherited from $\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$6 (Bardakov et al., 19 Jul 2025).

The multi-virtual theory is accompanied by related quotient families. Bardakov–Kozlovskaya–Negi–Prabhakar define multi-welded and multi-unrestricted braid groups, together with their pure and semi-pure kernels obtained from induced maps to $\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$7 (Bardakov et al., 19 Jul 2025). In the universal framework, the two one-forbidden quotients of $\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$8 are isomorphic, and the corresponding one-forbidden quotients of $\rho_i^{(0)},\rho_i^{(1)},\dots,\rho_i^{(k-1)}.$9 are likewise isomorphic; at the unrestricted level, the universal quotient $n!$00 and its descendants retain a semidirect-product structure over $n!$01 and, for $n!$02, have perfect commutator subgroup and $n!$03 as smallest non-abelian finite quotient (Ocampo et al., 4 Jun 2026). These results concern the ambient multi-virtual or unrestricted groups rather than $n!$04 directly, but they delineate the quotient landscape in which multi-virtual semi-pure groups sit.

6. Universal frameworks, analogues, and open directions

A later universal construction places multi-virtual braid groups inside the family $n!$05 of universal virtual braid groups. For $n!$06, $n!$07 is obtained as a quotient of $n!$08 by identifying

$n!$09

and then imposing braid relations in each fixed type $n!$10 (Ocampo, 2 Apr 2026). The universal theory supplies a finite-index right-angled Artin subgroup $n!$11 and proves for $n!$12 linearity, residual finiteness, solvable word and conjugacy problems, the Tits alternative, perfect commutator subgroup for $n!$13, and rigidity of finite quotients. The same paper treats these properties explicitly for $n!$14; for semi-pure multi-virtual groups it presents the kernel-pattern as the relevant model, but does not develop $n!$15 intrinsically there (Ocampo, 2 Apr 2026). A plausible implication is that the universal quotient description should strongly constrain the algorithmic and finite-quotient behavior of $n!$16, though that transfer is not fully spelled out.

Representation theory also points outward from the ambient universal setting. Homogeneous $n!$17-local representations of $n!$18 and of its welded quotient $n!$19 have been classified, with $n!$20 serving as a unifying host for braid-type groups with multiple crossing types (Nasser et al., 21 Apr 2026). This suggests a representation-theoretic route to multi-virtual semi-pure groups by restriction from universal representations, but the explicit representation theory of $n!$21 remains open.

The concluding section of the foundational multi-virtual paper isolates four directions that bear directly on semi-pure groups: constructing representations into automorphism groups of groups or quandles; identifying algebraic structures analogous to the Hecke-algebraic role of classical braid groups; defining groups and quandles for multi-virtual knots and proving invariance under multi-virtual Reidemeister and detour moves; and extending the theory to parametric multi-virtual braid groups in the sense of Loday–Stein (Bardakov et al., 19 Jul 2025). No explicit representation theory or invariant construction is fully developed there.

Multi-virtual semi-pure braid groups therefore occupy a precise but still expanding position in braid-type algebra. They are explicitly defined kernel subgroups of $n!$22, presented by generators $n!$23 and semi-pure/flat relations, decomposed as complements to $n!$24, and sharply understood in low rank. At the same time, universal and quotient-theoretic work shows that they belong to a broader ecosystem of multi-type virtual, welded, and unrestricted braid groups, where outer automorphisms, forbidden relations, and RAAG-based finite-index structures impose additional constraints on future developments (Bardakov et al., 19 Jul 2025).