ProbXiv
Explorersign in
Problem archiveProblem record

Statement

(T. Kerler) [Length = 1 TQFT’s] (1) Describe and construct algebraic L = 1 -extensions of Γ g -representations to TQFT’s, preferably as “simple” generalizations of the Fr ohman-Nicas- U (1)-theory. (2) Produce a classification of L = 1 -TQFT’s in the sense of an extension theory of Q ∗. (3) Identify the Γ g -representations on relative SU (2)-moduli space from [77] with these TQFT’s, and find similar, higher rank theories. (4) Identify the V [0] p as L = 1 -theories, if possible. The conceivable generalizations of the TQFT construction o f Frohman and Nicas described in (1) include using different, possibly non -compact gauge groups instead of U (1) and using more refined versions of intersection homolo- gies for stratified moduli spaces. Given the theory for Q ∗ the solution to item (2) will lead to well defined problems in sp-invariant theory. Constructions of L = 1-theories follow the schemes from (1) and (3). The identifi cation in (4) is carried out for p = 5 in [211]. The notion of finite length can be refined into the notion of q/l -solvable intro- duced in [212], indicating a TQFT over R = M[y]/yl+1 such that the constant order TQFT over the ground ring M is of length q. This, clearly, defines a special case of a TQFT of length ≤ (q·l + q + l). Murakami’s result [291] can be restated as saying that the Reshetikhin-Turaev theor y gives rise to a 1/1-solvable TQFT V [≤1] p with ground ring Fp (i.e. a TQFT of length 3 over Fp[y]/y2 ) such that V [≤1] p (M ) = 1 + y 1 6 λCW L(M ) (47) for any closed homology sphere M. Following Ohtsuki’s work Murakami’s iden- tity (with some extra renormalizations by the order of H1(M )) extends also to rational homology spheres. Let us call a theory with this pro perty a TQFT of Casson type. Recall, that the similar relation (46) for λCW L for manifolds with b1(M )≥ 1 is already contained in the information of a homological ( L = 0) TQFT, and is indeed a special evaluation of the Turaev-Milnor Torsion, s ee [212]. Given the richer structure of a 1/1-solvable TQFT we will expect new in variants Ξ that are refinements of λCW L and the torsion invariants. To be more precise, note that for a pair ( M, ϕ), where ϕ: π1(M )→ →Z defines a cyclic cover, any TQFT V yields an invariantV(M, ϕ) = trace(V(CΣ )) where CΣ = M− Σ: Σ → Σ and Σ ⊂ M is any surface dual to ϕ. In this way the Frohman Nicas theoriesV (j) yields the coefficients of the Alexander Polynomial, and, as shown in [212], thus also λCW L. A more refined invariant, which, roughly speaking, generali zes the Alexander module, is the Turaev-Viro module MT V (M, ϕ). It is described by Gilmer in [143]. MT V (M, ϕ) is given, up to conjugacy, by V(Σ) / ker(V(CΣ )N ) (with N large enough) together with the action of V(CΣ ) on it. The traces of V(CΣ ) or its powers are the most obvious well defined numerical inv ariants ofMT V (M, ϕ). The dimension of the module is yet another such invariant. For a 1/1-solvable theory V the invariantV(M, ϕ) takes values in M[y]/y2 and can hence be written as V(M, ϕ) = λV ϕ(M ) + y·Ξ V ϕ(M ), where λV and Ξ V are now M-valued invariants. If y coincides with the half projective parameter λV does not depend on ϕ, and we expect it to be some function of λCW L. Moreover, if V descends from a 1/2-solvable TQFT with the same property also Ξ V would be independent of ϕ. For the modular TQFT over F5[y]/y2 obtained from the Reshetikhin Turaev theory this invariant has already been defined in [212], and w e may expect it to lift, similarly, to an invariant Ξ Z over Z. For p > 5 we expect, as in the case of λCW L, the next order terms in the expansions (46) of the Reshetikh in Turaev theories to be polynomial expressions in λCW L and Ξ Z.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.