A reparametrization invariant
2026
Read A reparametrization invariant (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-26853.txt.
Introduction
Bak and Gustavsson introduce a nonabelian surface holonomy built from a one-form gauge potential valued in a loop algebra of the U(N) gauge group. The construction parallel-transports a nonabelian string across a surface. Although invariance is not obvious from the setup, they show the holonomy does not change under reparametrizations of the surface.
Catalog record
| YEAR | 2026 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37183 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.26853
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A reparametrization invariant Abstract We introduce a nonabelian surface holonomy that is constructed from a one-form gauge potential that takes values in a loop algebra of the U ( N ) U(N) gauge group. The surface holonomy parallel transports a nonabelian string. Although it is not manifest in our formulation, we will see that our nonabelian surface holonomy is invariant under reparametrizations of the surface. nonabelian surface holonomy Dongsu Bak, Andreas Gustavsson Physics Department, University of Seoul, Seoul 02504 KOREA 1 Introduction The question whether we can have a nonabelian generalization of the abelian Wilson surface \displaystyle W(\Sigma) \displaystyle= \displaystyle\exp\left(ie\int_{\Sigma}B\right) (1.1) for a closed surface \Sigma was analyzed in [ 1 ] . It was shown, by using the method of [ 2 ] , that a nonabelian generalization can not be foliation invariant. If we foliate the surface by closed loops, then as we change the foliation infinitesimally, it was shown that the Wilson surface will be invariant only if the gauge group is abelian. However, this argument relies on the assumption that the nonabelian generalization involves a two-form gauge field of the form B=B^{a}t_{a} where t_{a} are generators of the gauge group satisfying a Lie algebra \displaystyle[t_{a},t_{b}] \displaystyle= \displaystyle if_{ab}{}^{c}t_{c} (1.2) In this paper we will by-pass the no-go result of [ 1 ] by instead of B using a one-form gauge field A=A^{a}t_{a}(s) , where these generators satisfy a loop algebra without central extension. The fact that A is a one-form might seem strange since we would naively expect that we shall have a two-form since we have a surface. Nevertheless, we will demonstrate that our Wilson surface, and more generally, surface holonomy, is invariant under reparametrizations of the surface. Since gauge symmetry is a central concept in the construction, we begin in section 2 with presenting how the gauge symmetry transforms a wave function of a heavy closed string. Then in section 3 we present our definition of the surface holonomy that transports the string wave function gauge covariantly. In section 4 we show that the surface holonomy is invariant under deformation of the foliation of the surface. In section 5 we combine the foliation symmetry with reparametrization symmetry along each loop, to show that we have full reparametrization symmetry of the surface. 2 The nonabelian string We consider a closed string that is embedded in spacetime as s\mapsto C^{M}(s) for some embedding functions C^{M}(s) in spacetime. We introduce an infinite-dimensional loop algebra extension of the U(N) Lie algebra that is generated by loop algebra generators t_{a}(s) that are parametrized by a continuous parameter s around a closed loop and that satisfy \displaystyle[t_{a}(s),t_{b}(s^{\prime})] \displaystyle= \displaystyle i\delta(s-s^{\prime})f_{ab}{}^{c}t_{c}(s) (2.1) We next introduce a wave function \psi^{i}(s) where i=1,...,N transforms in the fundamental representation of U(N) . We define an action of the generator on the wave function element as follows, \displaystyle t_{a}(s)\psi^{i}(s^{\prime}) \displaystyle= \displaystyle\delta(s-s^{\prime})(t_{a})^{i}{}_{j}\psi^{j}(s) (2.2) where on the right-hand side appears the generators (t_{a})^{i}{}_{j} of the U(N) Lie algebra, satisfying \displaystyle[t_{a},t_{b}] \displaystyle= \displaystyle if_{ab}{}^{c}t_{c} (2.3) One may check that ( 2.1 ) is realized on these wave function elements by using ( 2.2 ). We define a U(N) gauge group element g(C) as \displaystyle g(C) \displaystyle= \displaystyle\exp\left(ie\int ds\Lambda^{a}(C(s))t_{a}(s)\right) (2.4) No path ordering is needed in order to define this exponent unambiguously, that is, in a reparametrization invariant way, since t_{a}(s) commute at any two different points along the loop. The group element acts on the wave function as \displaystyle g(C)\psi^{i}(s) \displaystyle= \displaystyle g^{i}{}_{j}(C(s))\psi^{j}(s) (2.5) where we define \displaystyle g^{i}{}_{j}(C(s)) \displaystyle= \displaystyle\left(e^{ie\Lambda^{a}(C(s))t_{a}}\right){}^{i}{}_{j} (2.6) Now this can be generalized to an arbitrary representation. All that we need to do, is to replace i with an index \alpha that transforms in the given representation. Then starting out from \displaystyle t_{a}(s)\psi^{\alpha}(s^{\prime}) \displaystyle= \displaystyle\delta(s-s^{\prime})(t_{a})^{\alpha}{}_{\beta}\psi^{\beta}(s) (2.7) we derive \displaystyle g(C)\psi^{\alpha}(s) \displaystyle= \displaystyle g^{\alpha}{}_{\beta}(C(s))\psi^{\beta}(s) (2.8) For an integrated wave function over an interval, this implies that its gauge transformation is given by \displaystyle g(C)\int_{I}ds\psi^{\alpha}(s) \displaystyle= \displaystyle\int_{I}dsg^{\alpha}{}_{\beta}(C(s))\psi^{\beta}(s) (2.9) Now we like to generalize in another direction. We consider a wave function that is a product, \displaystyle\psi^{\alpha_{1}\alpha_{2}\cdots\alpha_{K}}(s_{1},s_{2},\dots,s_{K}) \displaystyle= \displaystyle\psi^{\alpha_{1}}(s_{1})\psi^{\alpha_{2}}(s_{2})\cdots\psi^{\alpha_{K}}(s_{K}) (2.10) Its gauge transformation is \displaystyle g(C)\psi^{\alpha_{1}\alpha_{2}\cdots\alpha_{K}}(s_{1},s_{2},\dots,s_{K}) \displaystyle= \displaystyle g^{\alpha_{1}}{}_{\beta_{1}}(C(s_{1}))\cdots g^{\alpha_{K}}{}_{\beta_{K}}(C(s_{K})) (2.12) \displaystyle\psi^{\beta_{1}\dots\beta_{K}}(s_{1},s_{2},\dots,s_{K}) We have introduced K different particles that live on a circle and shown that the group element g(C) has a natural action on the wave function on this system. One might now want to ask whether we can reinterpret this multi-particle wave function as a wave function of a closed nonabelian string, perhaps by taking the large K limit. However, we will not attempt to answer this question here. Instead we will adopt the viewpoint that the wave function on which g(C) and also the surface holonomy U(C,C_{0}) , to be introduced shortly, have a natural action, could very well be the wave function that describes certain degrees of freedom that may or may not exist in the theory with the given surface holonomy. We can now understand from this wave function representation that the group composition rule \displaystyle g_{1}(C)g_{2}(C) \displaystyle= \displaystyle g_{3}(C) (2.13) gets inherited from the group composition rule of corresponding group elements in Yang-Mills theory, \displaystyle g_{1}(x)g_{2}(x) \displaystyle= \displaystyle g_{3}(x) (2.14) where g(x)=e^{ie\Lambda^{a}(x)t_{a}} . Two gauge parameters \Lambda_{1} and \Lambda_{2} are mapped into a third gauge parameter \Lambda_{3} in some rather complicated way. From this computation of \Lambda_{3} from \Lambda_{1,2} in the Yang-Mills case, we immediately obtain the structure of the g_{3}(C) loop group element, \displaystyle e^{ie\int ds\Lambda_{1}^{a}(C(s))t_{a}(s)}e^{ie\int ds^{\prime}\Lambda_{2}^{b}(C(s^{\prime}))t_{b}(s^{\prime})} \displaystyle= \displaystyle e^{ie\int ds\Lambda_{3}^{a}(C(s))t_{a}(s)} (2.15) Namely \Lambda_{1,2,3}(C(s)) are nothing but the gauge parameters in the Yang-Mills case evaluated at a point C(s) on the string, so the mapping from \Lambda_{1,2} to \Lambda_{3} will be also the same as in the Yang-Mills case. As a by-product, we have now essentially shown that g(C) are acting like group elements, where multiplying two such elements gives us a new element of the same type. Strictly speaking, what we have shown is that whenever g(C_{1})g(C_{2}) acts on a wave function of the product form ( 2.10 ) for any K then we have the group property. But does that mean that we always have the group property? In fact it is not very difficult to prove the group property in general using the Baker–Campbell–Hausdorff formula. To second order we find that \displaystyle\Lambda_{3}^{c}(C(s)) \displaystyle= \displaystyle\Lambda_{1}^{c}(C(s))+\Lambda_{2}^{c}(C(s))-\frac{e}{2}f_{ab}{}^{c}\Lambda_{1}^{a}(C(s))\Lambda_{2}^{b}(C(s))+{\cal{O}}(e^{2}) (2.16) with a similar local structure that is extending to all higher orders. In appendix B we show that when a string splits into two strings, the group element factorizes into two factors where each factor is a group element associated to each closed string. 3 The nonabelian surface holonomy A surface holonomy U(C,C_{0}) parallel transports the string from an initial string configuration C_{0} to a final string configuration C by acting on the wave function of the string as \displaystyle\psi(C_{0})\rightarrow\psi(C)=U(C,C_{0})\psi(C_{0}) (3.1) A gauge transformation acts on the wave function and on the surface holonomy together as \displaystyle\psi(C) \displaystyle\rightarrow \displaystyle\psi^{g}(C)=g(C)\psi(C) (3.2) \displaystyle U(C,C_{0}) \displaystyle\rightarrow \displaystyle U^{g}(C,C_{0})=g(C)U(C,C_{0})g(C_{0})^{-1} (3.3) To construct the surface holonomy explicitly, we will assume that we are given a gauge potential A_{M}^{a}(x) on spacetime that takes values in the adjoint representation of the U(N) gauge group. We define a gauge field on a closed string C as \displaystyle{\cal{A}}(C) \displaystyle= \displaystyle\int ds\;A_{M}^{a}(C(s))\delta C^{M}(s)t_{a}(s) (3.4) This gauge field is invariant under a