extended Lagrangian




physics, mathematical physics, philosophy of physics

Surveys, textbooks and lecture notes

theory (physics), model (physics)

experiment, measurement, computable physics

Variational calculus

Geometric quantization

Quantum field theory



The notion of extended Lagrangian is the notion of Lagrangian refined to extended quantum field theory (or “localized” or “multi-tiered” quantum field theory). At the same time is is the prequantum n-bundle of the theory in its higher geometric quantization in top codimension, hence over the point.

Motivation and Survey

  1. Ingredients of fundamental physics

  2. Topological local Lagrangian gauge field theory

  3. Chern-Simons field theory as a toy example

  4. Topological local field theory in higher category theory

  5. Intermezzo: Elements of higher cohesive geometry

  6. Chern-Simons theory as the archetypical example

  7. Topological local Lagrangian field theory in higher geometry

  8. 7-dimensional Chern-Simons theory as the next example

  9. ∞-Chern-Simons theory as the general example

Ingredients of fundamental physics

Ever since Issac Newton, theories of physics are formulated in the language of mathematics. Modern physics is formulated in terms of modern mathematics in the most intimate way. We aim to give here an account at least of central parts of modern physics naturally formulated in the mathematics of higher geometry that we discussed in geometry of physics - Geometry.

The physics that we are concerned with is fundamental physics, that part of the large subject of physics which is concerned with elementary processes and phenomena of physics from which all others are thought to emerge as approximations to collective effects of complex configurations of the elementary entities.

The best theory of fundamental physics as presently understood is known as Einstein-Yang-Mills-Dirac-Higgs theory. While fundamental, this theory has free parameters. These index different flavors of the same general mechanism. Notably the species and the masses of the fundamental fermion particles are such parameters, as is their coupling to the force fields.

Specifying these parameters is called constructing a model in physics and specifically a phenomenological model of fundamental physics, if aimed at making the observations that are predicted by the theory to closely match those that are being made in experiments. The global choice of these parameters that make the predictions match all available data to the currently best possible degree is called the standard model, specifically the standard model of particle physics for the sector in which the large-scale description of the force of gravity is approximately negligible, and the standard model of cosmology which focuses on that large scale structure.

The following table list some key ingredients of the standard theory of fundamental physics, all of which we discuss in detail further below.

standard model of particle physics and cosmology

gravityelectroweak and strong nuclear forcefermionic matterscalar field
field content:vielbein field eeprincipal connection \nablaspinor ψ\psiscalar field HH
Lagrangian:scalar curvature densityfield strength squaredDirac operator component densityfield strength squared + potential density
L=L = R(e)vol(e)+R(e) vol(e) + F eF +\langle F_\nabla \wedge \star_e F_\nabla\rangle + (ψ,D (e,)ψ)vol(e)+ (\psi , D_{(e,\nabla)} \psi) vol(e) + H¯ eH+(λ|H| 4μ 2|H| 2)vol(e) \nabla \bar H \wedge \star_e \nabla H + \left(\lambda {\vert H\vert}^4 - \mu^2 {\vert H\vert}^2 \right) vol(e)

Topological local Lagrangian gauge field theory

While the standard model of fundamental physics is a specification of the general framework of Einstein-Yang-Mills-Dirac-Higgs theory, that theory itself is a particular realization of several deep general principles of theoretical physics: it is

  1. a field theory;
  2. a gauge theory;
  3. a Lagrangian theory;
  4. a local theory;
  5. a topological theory.

We briefly indicate what this means, a formal discussion of these notions will be given in the main part below.

  • A field theory identifies the basic configurations of a physical system with generalized functions on spacetime, such as tensor fields, and more generally with sections ϕ\phi of some fiber bundle over spacetime

    E ϕ fieldbundle X id X \array{ && E \\ & {}^{\mathllap{\phi}}{\nearrow} & \downarrow^{\mathrlap{field\;bundle}} \\ X &\stackrel{id}{\to}& X }

    (This is the traditional formulation. At field (physics) is discussed that the concept of field fiber bundle needs to be refined to that of field fiber ∞-bundle in order to accuratly capture all aspects of field theory: The traditional idea of field bundles and its problems.)

  • A gauge theory in the broad sense has not just a set or smooth space of field configurations but has a groupoid or stack/smooth groupoid of field configurations: there are equivalences between nominally different field configurations

    ϕϕ \phi \stackrel{\simeq}{\to} \phi'

    called gauge transformations.

    A gauge theory in the original sense specifically has a groupoid of principal connections as its configurations. (Notice that gauge transformations are not in general just a redundancy of the description of the theory, but are genuine data: there are in general non-trivial gauge automorphisms of a field and the stack of gauge field configurations is not in general equivalent to a plain sheaf.)

  • A Lagrangian theory is one which is specified by a function on the space of all its configurations, called the (exponentiated) action functional exp(iS)\exp(i S). This function is assumed to be the integral of a differential form LL on spacetime that itself is a function of the field configurations and called the Lagrangian of the theory.

    exp(iS(ϕ))=exp(i XL(ϕ)(x)) \exp(i S(\phi)) = \exp\left(i \int_X L\left(\phi\right)\left(x\right) \right)

    The critical locus

    {ϕFields(Σ)|d varS ϕ0} \left\{ \phi \in \mathbf{Fields}(\Sigma) | d_{\mathrm{var}} S_\phi \simeq 0 \right\}

    of the action functional is called the phase space of the theory, a point of this is said to be a solution to the equations of motion of the theory.

  • A local theory has a Lagrangian form which at each point of spacetime only depends on the values of the fields at that point, and of finitely many derivatives of the fields at that point.

    exp(iS(ϕ))=exp(i XL(ϕ(x),ϕ(x))) \exp(i S(\phi)) = \exp\left(i \int_X L\left(\phi\left(x\right), \nabla \phi\left(x\right)\right)\right)

    This means that a local Lagrangian is a horizontal form on the jet bundle of the field bundle of which the fields are sections.

  • A topological theory finally is one where the field bundle is naturally associated to every smooth manifold, typically of some fixed dimension and possibly equipped with topological G-structure such as an orientation, but not equipped with any further geometric structure: all geometric structure such as metrics on spacetime are themselves regarded as parts of the field content in a topological theory.

Chern-Simons field theory as a toy example

Most of this was fairly well understood decades ago, by the 1960s. But even with all this conceptual understanding of the structure of fundamental physics available, there were and are a lot of implications of topological local Lagrangian gauge field theory which remained elusive.

In the 1980s attention turned to other topological local Lagrangian gauge field theories than that governing the standard model: it turns out that there are many such theories of physics that share crucial properties with the standard model but which otherwise predict, if considered as phenomenological models, physics radically different from that which we actually observe experimentally. For instance there are respectable theories of fundamental physics which describe physics in dimensions other than the three spatial and one time dimension which we clearly observe, theories that contain no force of gravity, others that contain only the force of gravity, theories that contain mirror partners of every species of matter or contain no matter at all, etc.

In short, it was gradually realized, and realized to be important, that, in the platonic world of mathematics and of theoretical physics, there is a whole space of field theories, in which the one underlying the standard model of observed physics is only one single point, even if regarded with its free parameters.

A crucial step forward happened when around 1989 Edward Witten, in this vein, introduced the study of what is now called Chern-Simons theory a topological local Lagrangian gauge field theory which, as such, shares some crucial principles with Einstein-Maxwell-Yang-Mills theory, while at the same time being radically different and most importantly: much simpler to analyze.

The dimension of Chern-Simons theory is 3 instead of 4, but Chern-Simons theory has a gauge field just as Yang-Mills theory does. For GG a connected simply connected gauge group, this is mathematically represented by a Lie algebra-valued differential 1-form AA.

The Lagrangian of Chern-Simons theory is the Chern-Simons 3-form

L(A)CS 3(A)AF A+cA[AA], L(A) \coloneqq \mathrm{CS}_3(A) \coloneqq \langle A \wedge F_A\rangle + c \langle A \wedge [A \wedge A]\rangle \,,

whence the name of the theory, and hence the Chern-Simons action functional is

exp(iS(A))=exp(2πi Σ XCS 3(A)) \exp(i S(A)) = \exp\left( 2 \pi i \int_{\Sigma_X} \mathrm{CS}_3(A) \right)

At the same time Chern-Simons theory is still being rich in phenomena, in fact so rich in phenomena that for instance it helped illuminate deep mathematical problems that had not otherwise been understood as much before (properties called knot invariants).

Topological local field theory in higher category theory

Another pleasant effect of the introduction of the Chern-Simons “toy model” for topological field theory was that, due to its conceptual simplicity, its basic structure could be appreciated also by mathematicians not trained in theoretical physics, who would otherwise fail to see through all the physics jargon involved in the available discussion of realistic models. Accordingly, immediately after the realization of Chern-Simons field theory, the first mathematically precise axiomatizations of an nn-dimensional topological field theory was proposed, by Michael Atiyah: “functorial quantum field theory

This axiomatization demands in particular that to each closed manifold of dimension (n1)(n-1) is assigned a vector space

Z:Σ n1Z(Σ n1)Vect Z \;\colon\; \Sigma_{n-1} \mapsto Z(\Sigma_{n-1}) \in \mathrm{Vect}

thought of as the space of physical states of the theory on the spatial slice Σ n1\Sigma_{n-1} of spacetime.

However, This original axiomatization did not capture the full locality property: the spaces of states are assigned globally to a space Σ n1\Sigma_{n-1} and are not required to be obtainable by “integrating up local data” over Σ n1\Sigma_{n-1}.

In order to refine this, one needs an axiomatization of field theory that assigns data also to manifolds Σ k\Sigma_{k} of dimension 0kn0 \leq k \leq n. This should be such that the data assigned to a torus Σ k×S 1\Sigma_{k} \times S^1 is the trace in a suitable sense, of the data assigned to Σ k\Sigma_{k} itself. It turns out that the way to capture this is to think of Z(Σ k)Z(\Sigma_k) as being a vector space in higher category theory/directed homotopy type theory: an (n-k)-vector space.

Σ kZ(Σ k)(nk)Vect \Sigma_k \mapsto Z(\Sigma_k) \in (n-k)\mathrm{Vect}

That something like this should work was hypothetized by Dan Freed (in Higher Algebraic Structures and Quantization), the precise formulation was later given by Jacob Lurie (in On the Classification of Topological Field Theories). In the literature the resulting axiomatization is called extended topological field theory (with “extended” referring to extending the axioms from codimension 1 to higher codimension) or multi-tiered field theory (thinking of the data in higher codimension as being higher “tiers” of the theory).

But in the physics-community, at least in the algebraic quantum field theory-community, there is also the well-established term of local quantum field theory which refers broadly to what the axioms of “extended TQFT” capture. Hence we should think of this as formalizing topological local field theory.

Topological local Lagrangian field theory in higher geometry

The formalization of extended topological field theory thus captures the local aspect of field theory. But we saw above that the theories of fundamental physics have one more crucial property: they are Lagrangian. Here we indicate how to formulate Lagrangian theories that are also local in the sense of extended field theory.

The step of passing from a Lagrangian to the corresponding quantum field theory is called quantization. There are two broad approaches to formalization of what this means, called algebraic deformation quantization and geometric quantization. The first of these deals explicitly with producing algebras of quantum observables, while the second describes explicitly how to construct the spaces of states that we considered above. Both are at least roughly dual to each other, in the spirit of the general duality between algebra and geometry. For the following discussion we follow the ideas of geometric quantization, which naturally fit a discussion of geometry of physics. The following table surveys some keywords in this context which we will discuss in detail further below.

Isbell duality between algebra and geometry

A\phantom{A}geometryA\phantom{A}A\phantom{A}categoryA\phantom{A}A\phantom{A}dual categoryA\phantom{A}A\phantom{A}algebraA\phantom{A}
A\phantom{A}topologyA\phantom{A}A\phantom{A}NCTopSpaces H,cpt\phantom{NC}TopSpaces_{H,cpt}A\phantom{A}A\phantom{A}Gelfand-KolmogorovAlg op\overset{\text{<a href="">Gelfand-Kolmogorov</a>}}{\hookrightarrow} Alg^{op}_{\mathbb{R}}A\phantom{A}A\phantom{A}commutative algebraA\phantom{A}
A\phantom{A}topologyA\phantom{A}A\phantom{A}NCTopSpaces H,cpt\phantom{NC}TopSpaces_{H,cpt}A\phantom{A}A\phantom{A}Gelfand dualityTopAlg C *,comm op\overset{\text{<a class="existingWikiWord" href="">Gelfand duality</a>}}{\simeq} TopAlg^{op}_{C^\ast, comm}A\phantom{A}A\phantom{A}comm. C-star-algebraA\phantom{A}
A\phantom{A}noncomm. topologyA\phantom{A}A\phantom{A}NCTopSpaces H,cptNCTopSpaces_{H,cpt}A\phantom{A}A\phantom{A}Gelfand dualityTopAlg C * op\overset{\phantom{\text{Gelfand duality}}}{\coloneqq} TopAlg^{op}_{C^\ast}A\phantom{A}A\phantom{A}general C-star-algebraA\phantom{A}
A\phantom{A}algebraic geometryA\phantom{A}A\phantom{A}NCSchemes Aff\phantom{NC}Schemes_{Aff}A\phantom{A}A\phantom{A}almost by def.TopAlg fin op\overset{\text{<a href="">almost by def.</a>}}{\hookrightarrow} \phantom{Top}Alg^{op}_{fin} A\phantom{A}A\phantom{A}fin. gen.A\phantom{A}
A\phantom{A}commutative algebraA\phantom{A}
A\phantom{A}noncomm. algebraicA\phantom{A}
A\phantom{A}NCSchemes AffNCSchemes_{Aff}A\phantom{A}A\phantom{A}Gelfand dualityTopAlg fin,red op\overset{\phantom{\text{Gelfand duality}}}{\coloneqq} \phantom{Top}Alg^{op}_{fin, red}A\phantom{A}A\phantom{A}fin. gen.
A\phantom{A}associative algebraA\phantom{A}A\phantom{A}
A\phantom{A}differential geometryA\phantom{A}A\phantom{A}SmoothManifoldsSmoothManifoldsA\phantom{A}A\phantom{A}Milnor's exerciseTopAlg comm op\overset{\text{<a href="">Milnor's exercise</a>}}{\hookrightarrow} \phantom{Top}Alg^{op}_{comm}A\phantom{A}A\phantom{A}commutative algebraA\phantom{A}
A\phantom{A}supergeometryA\phantom{A}A\phantom{A}SuperSpaces Cart n|q\array{SuperSpaces_{Cart} \\ \\ \mathbb{R}^{n\vert q}}A\phantom{A}A\phantom{A}Milnor's exercise Alg 2AAAA op C ( n) q\array{ \overset{\phantom{\text{Milnor's exercise}}}{\hookrightarrow} & Alg^{op}_{\mathbb{Z}_2 \phantom{AAAA}} \\ \mapsto & C^\infty(\mathbb{R}^n) \otimes \wedge^\bullet \mathbb{R}^q }A\phantom{A}A\phantom{A}supercommutativeA\phantom{A}
A\phantom{A}formal higherA\phantom{A}
A\phantom{A}(super Lie theory)A\phantom{A}
ASuperL Alg fin 𝔤A\phantom{A}\array{ Super L_\infty Alg_{fin} \\ \mathfrak{g} }\phantom{A}AALada-MarklA sdgcAlg op CE(𝔤)A\phantom{A}\array{ \overset{ \phantom{A}\text{<a href="">Lada-Markl</a>}\phantom{A} }{\hookrightarrow} & sdgcAlg^{op} \\ \mapsto & CE(\mathfrak{g}) }\phantom{A}A\phantom{A}differential graded-commutativeA\phantom{A}
A\phantom{A} (“FDAs”)

in physics:

A\phantom{A}Poisson algebraA\phantom{A}A\phantom{A}Poisson manifoldA\phantom{A}
A\phantom{A}deformation quantizationA\phantom{A}A\phantom{A}geometric quantizationA\phantom{A}
A\phantom{A}algebra of observablesA\phantom{A}space of statesA\phantom{A}
A\phantom{A}Heisenberg pictureA\phantom{A}Schrödinger pictureA\phantom{A}
A\phantom{A}higher algebraA\phantom{A}A\phantom{A}higher geometryA\phantom{A}
A\phantom{A}Poisson n-algebraA\phantom{A}A\phantom{A}n-plectic manifoldA\phantom{A}
A\phantom{A}En-algebrasA\phantom{A}A\phantom{A}higher symplectic geometryA\phantom{A}
A\phantom{A}BD-BV quantizationA\phantom{A}A\phantom{A}higher geometric quantizationA\phantom{A}
A\phantom{A}factorization algebra of observablesA\phantom{A}A\phantom{A}extended quantum field theoryA\phantom{A}
A\phantom{A}factorization homologyA\phantom{A}A\phantom{A}cobordism representationA\phantom{A}

In its traditional formulation, one considers the symplectic form ω\omega which is naturally induced by a local Lagrangian on its (reduced) covariant phase space and says that a geometric prequantization of this form is the choice of a U(1)U(1)-principal connection \nabla on the phase space whose curvature F F_\nabla is ω\omega, hence a lift in

BU(1) conn prequantumcirclebundle F () curvature PhaseSpace ω Ω cl 2 sympl.form. \array{ &&& \mathbf{B}U(1)_{conn} \\ {{prequantum} \atop {circle\;bundle}} && {}^{\mathllap{\nabla}}\nearrow & \downarrow^{\mathrlap{F_{(-)}}} & curvature \\ &PhaseSpace &\stackrel{\omega}{\to}& \Omega^2_{cl} \\ && sympl.\;form } \,.

A section of the associated complex line bundle

P× U(1) Ψ X id X \array{ && P \times_{U(1)}\mathbb{C} \\ & {}^{\mathllap{\Psi}}\nearrow & \downarrow \\ X &\stackrel{id}{\to}& X }

is called a wave function on phase space. One is to choose a polarization of phase space, equipping it locally with “canonical coordinates” and “canonical momenta”. Then a quantum state in geometric quantization is a wave function Ψ\Psi which only depends on the canonical coordinates.

Z(Σ n1){wavefunctionsofcanonicalcoordinates}{polarizedsectionsofprequantumlinebundle}. Z(\Sigma_{n-1}) \coloneqq \{wave\; functions\;of\;canonical\;coordinates\} \coloneqq \left\{ polarized\;sections\;of\;prequantum\;line\;bundle \right\} \,.

From this it is clear that in order to refine this to local (extended, multi-tiered) prequantum field theory we are to assign higher line bundles in higher codimension, namely a prequantum (n-k)-bundle in codimension kk.

As a first indication that this makes good sense, notice that the natural assignment to a field in codimension 0, hence to a field configuration on spacetime, is its action functional. Being a U(1)U(1)-valued function, this is naturally thought of as a 0-bundle, the prequantum 0-bundle.

So we are after a picture as indicated in the following table.

local/extended Lagrangian field theorylocal/extended quantum field theory
closed piece of spacetime of dimension 0kn0 \leq k \leq nprequantum field theoryprequantum (n-k)-bundlegeometric quantizationspace of (nk)(n-k)-wave functions
closed manifold of dimension 0kn0 \leq k \leq nlocalprequantumfieldtheory\stackrel{local\;prequantum\;field\;theory}{\mapsto}circle (n-k)-bundle with connection on moduli stack of field configurationspolarizedsections\stackrel{polarized\; sections}{\mapsto}(n-k)-vector space of states
Σ k\Sigma_k\mapstoexp(2πi Σ k[Σ k,L]):[Σ k,Fields]B nkU(1) conn\exp(2 \pi i \int_{\Sigma_k}[\Sigma_k,\mathbf{L}]) : [\Sigma_k, \mathbf{Fields}] \to \mathbf{B}^{n-k} U(1)_{conn}\mapstoZ(Σ nk)Z(\Sigma_{n-k})

Intermezzo: Recalling elements of higher cohesive geometry

The idea of extended prequantum field theory indicated above is best illustrated by the example of Chern-Simons field theory. We talk about this in a moment below, but since in order to do so we need now at least some basic notions from cohesive higher geometry, we very briefly recall these first, just for reference. For more discussion see at geometry of physics - smooth homotopy types and geometry of physics - Maurer-Cartan forms and geometry of physics - principal connections

Fix some site 𝒞\mathcal{C} of test spaces which encode some notion of geometry. The running example to keep in mind is the site SmthMfd of smooth manifolds, or equivalently its full dense subsite of CartSp on manifolds of the form n\mathbb{R}^n. Assume for simplicity that the site has enough points (which SmthMfd does).

Then the (2,1)-category of (2,1)-sheaves (stacks) on 𝒞\mathcal{C} is the simplicial localization

Sh (2,1)(𝒞)L leFunc(𝒞 op,Grpd). Sh_{(2,1)}(\mathcal{C}) \simeq L_{le} Func(\mathcal{C}^{op}, Grpd) \,.

of that of groupoid-valued presheaves on C\mathbf{C} at the local equivalences. This means that Sh (2,1)(𝒞)Sh_{(2,1)}(\mathcal{C}) is the result of turning stalk-wise (“local”) equivalence of groupoids into actual equivalences.

For the running example of 𝒞\mathcal{C} \simeq SmthMfd we may think of an object in Sh (2,1)(𝒞)Sh_{(2,1)}(\mathcal{C}) as a smooth groupoid. Examples include smooth manifolds, diffeological space, orbifolds, Lie groupoids/differentiable stacks, sheaves of smooth differential forms, etc.

The higher refinement of groupoids which we need are ∞-groupoids. The collection of these is equivalently that of simplicial sets or (compactly generated weakly Hausdorff) topological spaces with the weak homotopy equivalences universally turned into actual homotopy equivalences:

GrpdL whesSetL wheTop \infty Grpd \simeq L_{whe} sSet \simeq L_{whe} Top

The combination of these two aspects, the geometry and the homotopy theory, is that of (∞,1)-sheaves/∞-stacks, here we write H\mathbf{H} for the collection of presheaves of simplicial sets (simplicial presheaves) with the local (stalk-wise) weak homotopy equivalences universally turned into actual homotopy equivalences:

H Sh (𝒞) stackif. PSh (𝒞) L lwheFunc(𝒞 op,sSet) id𝕃id L wheFunc(𝒞 op,sSet). \array{ \mathbf{H} \coloneqq & Sh_\infty(\mathcal{C}) & \stackrel{ \overset{\infty-stackif.}{\leftarrow} }{ \hookrightarrow } & PSh_\infty(\mathcal{C}) \\ & \uparrow^{\mathrlap{\simeq}} && \uparrow^{\mathrlap{\simeq}} \\ & L_{lwhe} Func(\mathcal{C}^{op}, sSet) & \stackrel{ \overset{\mathbb{L} id}{\leftarrow} }{ \underset{\mathbb{R} id}{\rightarrow} }& L_{whe} Func(\mathcal{C}^{op}, sSet) } \,.

This H\mathbf{H} is the (∞,1)-topos over 𝒞\mathcal{C}. This is the kind of context for the higher geometry encoded by 𝒞\mathcal{C} in which all of our constructions take place.

In order to formulate not just kinematics but also dynamics in physics, we need to require that H\mathbf{H} has an intrinsic notion of what it means to have a path or trajectory inside one of its objects. This is axiomatized by the notion of cohesion:

if 𝒞\mathcal{C} satisfies some properties that make it an an infinity-cohesive site, the the global section geometric morphism of H\mathbf{H} extended to an adjoint quadruple of (∞,1)-functors:

H coDiscΓDiscΠ Grpd L lwheFunc(𝒞 op,sSet) XX(*)consthocolim 𝒞 L whesSet \array{ \mathbf{H} & \stackrel{\overset{\Pi}{\to}}{\stackrel{}{\stackrel{\overset{Disc}{\leftarrow}}{\stackrel{\overset{\Gamma}{\to}}{\underset{coDisc}{\leftarrow}}}}} & \infty Grpd \\ \uparrow^{\mathrlap{\simeq}} && \uparrow^{\mathrlap{\simeq}} \\ L_{lwhe} Func(\mathcal{C}^{op}, sSet) &\stackrel{\overset{hocolim_{\mathcal{C}}}{\to}}{\stackrel{\overset{const}{\leftarrow}}{\stackrel{X \mapsto X(*)}{\stackrel{\overset{}{\to}}{\underset{}{\leftarrow}}}}}& L_{whe} sSet }

with some mild extra condition on Π\Pi and coDisccoDisc. This is the case in particular for 𝒞=\mathcal{C} = SmthMfd in which case we say that H\mathbf{H} is the cohesive (∞,1)-topos of smooth ∞-groupoids.

The defining adjoint quadruple of (∞,1)-functors induces an adjoint triple of co(∞,1)-monads:

shape modality\dashvflat modality\dashvsharp modality
ΠDiscΠ\mathbf{\Pi} \coloneqq Disc \circ \PiflatDiscΓflat \coloneqq Disc \circ \GammacoDiscΓ\sharp \coloneqq coDisc \circ \Gamma

For GGrp(H)G \in Grp(\mathbf{H}) a group object in H\mathbf{H}, hence a geometric ∞-group, we have by these axioms of cohesion a pasting diagram of (∞,1)-pullbacks

G * θ G dRBG BG * BGgroup * MCform deRhamcoefficients flatcoefficients * principalcoefficients. \array{ G &\to& * \\ \downarrow^{\mathrlap{\theta_G}} && \downarrow \\ \flat_{dR} \mathbf{B}G &\to& \flat \mathbf{B}G \\ \downarrow && \downarrow \\ * &\to& \mathbf{B}G } \;\;\;\;\;\;\;\;\;\;\;\;\;\;\; \array{ \infty-group &\to& * \\ {}^{\mathllap{MC\;form}}\downarrow && \downarrow \\ deRham\, coefficients &\to& flat\;coefficients \\ \downarrow && \downarrow \\ * &\to& principal\,coefficients } \,.

Here the delooping BG\mathbf{B}G is the moduli of geometric GG-principal ∞-bundles, dRBG\flat_{dR} \mathbf{B}G is that of flat L-∞ algebra valued differential forms and θ G\theta_G is the Maurer-Cartan form on the smooth ∞-group GG.

If 𝔾\mathbb{G} is “slightly abelian” in that it is equipped with the structure of a braided ∞-group, then we write

curv 𝔾:B𝔾θ B𝔾 dRB 2𝔾 curv_{\mathbb{G}} \;\colon\; \mathbf{B}\mathbb{G} \stackrel{\theta_{\mathbf{B}\mathbb{G}}}{\to} \flat_{dR}\mathbf{B}^2 \mathbb{G}

for the Maurer-Cartan form of its delooping ∞-group and call this the universal curvature characteristic.

Then 𝔾\mathbb{G}-differential cohomology is curv B𝔾curv_{\mathbf{B}\mathbb{G}}-twisted cohomology. In particular, if we consider globally defined curvature forms

Ω cl(,𝔾) dRB 2𝔾 \Omega_{cl}(-,\mathbb{G}) \to \flat_{dR}\mathbf{B}^2\mathbb{G}

then the (∞,1)-pullback

B𝔾 conn Ω(,𝔾) F () B𝔾 curv B𝔾 dRB 2𝔾 \array{ \mathbf{B}\mathbb{G}_{conn} &\to& \Omega(-,\mathbb{G}) \\ \downarrow && \downarrow^{\mathrlap{F_{(-)}}} \\ \mathbf{B}\mathbb{G} &\stackrel{curv_{\mathbf{B}\mathbb{G}}}{\to}& \flat_{dR}\mathbf{B}^2 \mathbb{G} }

is the moduli for principal ∞-connections.

Now for

c:BGB𝔾 \mathbf{c} \;\colon\; \mathbf{B}G \to \mathbf{B}\mathbb{G}

a universal characteristic map we say that a differential refinement is a choice of BG conn\mathbf{B}G_{conn} and c^\hat \mathbf{c} in

BG c B𝔾 BG conn c^ B𝔾 conn BG c B𝔾. \array{ \flat \mathbf{B}G &\stackrel{\flat \mathbf{c}}{\to}& \mathbf{B}\mathbb{G} \\ \downarrow && \downarrow^{} \\ \mathbf{B}G_{conn} &\stackrel{\hat \mathbf{c}}{\to}& \mathbf{B}\mathbb{G}_{conn} \\ \downarrow && \downarrow \\ \mathbf{B}G &\stackrel{\mathbf{c}}{\to}& \mathbf{B}\mathbb{G} } \,.

These are the extended Lagrangians in our discussion of extended prequantum field theory.

In particular if, in the context of smooth ∞-groupoids, we consider 𝔾=B n1U(1)\mathbb{G} = \mathbf{B}^{n-1}U(1) the circle n-group then

B𝔾 connB nU(1) connDK(U(1)Ω 1Ω n) \mathbf{B}\mathbb{G}_{conn} \simeq \mathbf{B}^n U(1)_{conn} \simeq DK( U(1) \to \Omega^1 \to \cdots \to \Omega^n)

is given as the ∞-stackification of the image under the Dold-Kan correspondence of the Deligne complex for ordinary differential cohomology.

In this case there is fiber integration in ordinary differential cohomology refined to moduli \infty-stacks: for Σ k\Sigma_k an oriented closed manifold of dimension kk we have a map of moduli ∞-stacks

exp(2πi Σ()):[Σ k,B nU(1) conn]B nkU(1) conn. \exp\left( 2 \pi i \int_{\Sigma} (-) \right) \;\colon\; [\Sigma_k, \mathbf{B}^n U(1)_{conn}] \to \mathbf{B}^{n-k} U(1)_{conn} \,.

Composed with the internal hom this is transgression of differential cocycles. The transgressions of an extended Lagrangian as above are (the moduli for) the higher prequantum n-bundles that we consider here.

Chern-Simons theory as the archetypical example

For illustration purposes, we survey how the Chern-Simons field theory considered as a toy model for topological local Lagrangian gauge field theory considered above has in higher geometry a prequantum (n-k)-bundle associated in codimension kk.

First consider again GG to be a connected and simply connected simple Lie group. Then the integral cohomology of the classifying space of GG is

H 4(BG,). H^4(BG , \mathbb{Z}) \simeq \mathbb{Z} \,.

An element in here is a universal characteristic class of GG-principal bundles. In Chern-Simons theory this is going to be what traditionally is called the level of the theory.

We write

c:BGB 3U(1)K(,4) c : B G \to B^3 U(1) \simeq K(\mathbb{Z},4)

for a representative universal characteristic map of the corresponding degree-4 universal characteristic class.

Under geometric realization as just recalled this has an essentially unique lift to smooth cohomology, hence to a morphism of moduli ∞-stacks

c:BGB 3U(1) \mathbf{c} \;\colon\; \mathbf{B}G \to \mathbf{B}^3 U(1)

which, in a precise sense, is the Lie integration of the canonical 3-cocycle

,[,]:𝔤B 2 \langle -, [-,-]\rangle \;\colon\; \mathfrak{g} \to \mathbf{B}^2 \mathbb{R}

in the Lie algebra cohomology of the Lie algebra 𝔤\mathfrak{g} of GG in that

cτ 1exp(μ). \mathbf{c} \simeq \tau_1 \exp(\mu) \,.

Moreover, this smooth universal characteristic map has a further lift to differential cohomology

c^:BG connB 3U(1) conn. \hat \mathbf{c} \;\colon\; \mathbf{B}G_{conn} \to \mathbf{B}^3 U(1)_{conn} \,.

This modulates a circle 3-bundle with connection – the “Chern-Simons circle 3-bundle with connection” – on the smooth moduli stack of smooth GG-principal connections.

BG conn c^ B 3U(1) conn H BG c B 3U(1) |Π()| BG c B 3U(1) L wheTop \array{ \mathbf{B}G_{conn} &\stackrel{\hat \mathbf{c}}{\to}& \mathbf{B}^3 U(1)_{conn} \\ \downarrow && \downarrow &&&&& \in \mathbf{H} \\ \mathbf{B}G &\stackrel{\mathbf{c}}{\to}& \mathbf{B}^3 U(1) \\ && &&& \downarrow^{\mathrlap{{\vert \Pi(-)\vert}}} \\ B G &\stackrel{c}{\to}& B^3 U(1) &&&&& \in L_{whe} Top }

One finds that the transgression of this to codimension 0 is the Chern-Simons action functional discussed above. The other transgressions are

Here the WZW B-field is the topological part of a 2-dimensional local Lagrangian field theory called the Wess-Zumino-Witten model. One says that the WZW model is related by the holographic principle with 3d Chern-Simons theory.

If the gauge group GG is not simple and simply connected, then there are other canonical universal characteristic classes in H 4(BG,)H^4(B G, \mathbb{Z}). If these have a differential refinement, they give rise to the corresponding Chern-Simons theory.

Notably in the case that G=U(1)G = U(1) is the circle group (hence not simply connected) there is the cup product of the universal first Chern class with itself

c 1c 1:BU(1)B 3U(1). c_1 \cup c_1 \;\colon\; B U(1) \to B^3 U(1) \,.

The smooth and differential refinement of the first Chern class itself c 1:BU(1)BU(1)K(,2)c_1 \;\colon\; B U(1) \to B U(1) \simeq K(\mathbb{Z},2) is tautological: this is simply the identity

c^ 1id:BU(1) connBU(1) conn. \widehat \mathbf{c}_1 \coloneqq id \;\colon\; \mathbf{B}U(1)_{conn} \to \mathbf{B}U(1)_{conn} \,.

The nontrivial part is that, as one can see, the cup product on differential cohomology refines to moduli ∞-stacks of circle n-bundle with connection as the ∞-stackification of what is called the Beilinson-Deligne cup product, and so we have

c^ 1c^ 1:BU(1) connB 3U(1) conn. \widehat \mathbf{c}_1 \cup \widehat \mathbf{c}_1 \;\colon\; \mathbf{B}U(1)_{conn} \to \mathbf{B}^3 U(1)_{conn} \,.

This is the extended Lagrangian for abelian Chern-Simons theory. On field configuration which happens to be given by a globally defined 1-form AA its value is

exp(2πi Σ 3AdA)U(1). \exp\left( 2\pi i \int_{\Sigma_3} A \wedge d A \right) \in U(1) \,.

But now not every field configuration is necessarily given by a globally defined 1-form anymore, and more generally the action functional has a more sophisticated expression.

7-dimensional Chern-Simons theory as the next example

So far we have considered general motivation for extended prequantum field theory, and indication how its application to the well-known case of 3d Chern-Simons theory naturally captures the characteristics of the theory. Now we consider further examples whose higher geometry has not necessarily been considered traditionally.

An immediate class of further examples is obtained by abelian higher Chern-Simons theory induced by the cup product of the first Chern class with itself

[DD 2kDD 2k]H 4(k+1)(B 2k+1U(1),). [DD_{2k} \cup DD_{2k}] \in H^{4(k+1)}(B^{2k+1} U(1), \mathbb{Z}) \,.

This has a refinement (FSS cup) to a differential characteristic class by the Beilinson-Deligne cup product of the differential refinement of the higher Dixmier-Douady class with itself:

DD^ 2kDD^ 2k:B 2k+1U(1) connB 4k+3U(1) conn. \widehat \mathbf{DD}_{2k} \cup \widehat \mathbf{DD}_{2k} \;\colon\; \mathbf{B}^{2k+1} U(1)_{conn} \to \mathbf{B}^{4k+3} U(1)_{conn} \,.

In particular there is the abelian 7d Chern-Simons theory

DD^ 2DD^ 2:B 3U(1) connB 7U(1) conn. \widehat \mathbf{DD}_2 \cup \widehat \mathbf{DD}_2 \;\colon\; \mathbf{B}^3 U(1)_{conn} \to \mathbf{B}^7 U(1)_{conn} \,.

An argument in (Witten 98) says that this abelian 7d Chern-Simons theory (or more precisely a quadratic refinement of it) is related to the self-dual higher gauge theory sector of the 6d (2,0)-supersymmetric QFT on the worldvolume of a single M5-brane_ in analogy to how 3d Chern-Simons theory is related to the WZW model:

∞-Chern-Simons theory\leftarrowholographic principle \rightarrow∞-Wess-Zumino-Witten theoryKaluza-Klein reduction \to
3d Chern-Simons theory3dCS-2dCFTWZW model
7d Chern-Simons theoryAdS7/CFT6 duality6d (2,0)-supersymmetric QFTN=4 D=4 super Yang-Mills theory

The argument in (Witten 98) also says that in view of AdS7/CFT6 duality the field of the 7d theory here is to be identified withe the supergravity C-field in 11-dimensional supergravity/M-theory.

But the supergravity C-field is known to have moduli that are in general more complicated than just B 3U(1) conn\mathbf{B}^3 U(1)_{conn} (due to a Green-Schwarz anomaly cancellation mechanism in 11-dimensional supergravity and a “flux quantization” condition). In (FSS 7dCD) the corrected moduli are dicussed and found to involve a coupling of the abelian 7d theory with a nonabelian 7d Chern-Simons theory whose fields are twisted higher spin connections: “twisted differential string structures”.

To see what this is, recall that above we interpreted the extended Lagrangian of 3d Chern-Simons theory for a simply connected compact simple gauge group GG as the smooth and moreover differential refinement of the canonical universal characteristic class in H 4(BG,)H^4(B G, \mathbb{Z}). Specifically for GG the spin group this is the first fractional Pontryagin class, represented by a characteristic map

12p 1:BSpinB 3U(1). \tfrac{1}{2}p_1 \;\colon\; B Spin \to B^3 U(1) \,.

This can be understood as a higher analog of the second Stiefel-Whitney class w 2:BSOB 2w_2 \;\colon\; B SO \to B^2 \mathbb{Z} which is the universal obstruction to spin structure. We say that 12p 1\tfrac{1}{2}p_1 is the universal obstruction to a higher spin structure called string structure (this and the motivation for this terminology is discussed in detail in geometry of physics - Fields - Examples - Fields of gravity and G-structure). These higher obstructions are the (dual) k-invariants of the Whitehead tower of BOB O:

smooth ∞-groupWhitehead tower of smooth moduli ∞-stacksG-structure/higher spin structureobstruction
ninebrane 10-groupBNinebrane\mathbf{B}Ninebrane ninebrane structurethird fractional Pontryagin class
fivebrane 6-groupBFivebrane1np 3B 11U(1)\mathbf{B}Fivebrane \stackrel{\tfrac{1}{n} p_3}{\to} \mathbf{B}^{11}U(1)fivebrane structuresecond fractional Pontryagin class
string 2-groupBString16p 2B 7U(1)\mathbf{B}String \stackrel{\tfrac{1}{6}\mathbf{p}_2}{\to} \mathbf{B}^7 U(1)string structurefirst fractional Pontryagin class
spin groupBSpin12p 1B 3U(1)\mathbf{B}Spin \stackrel{\tfrac{1}{2}\mathbf{p}_1}{\to} \mathbf{B}^3 U(1)spin structuresecond Stiefel-Whitney class
special orthogonal groupBSOw 2B 2 2\mathbf{B}SO \stackrel{\mathbf{w_2}}{\to} \mathbf{B}^2 \mathbb{Z}_2orientation structurefirst Stiefel-Whitney class
orthogonal groupBOw 1B 2\mathbf{B}O \stackrel{\mathbf{w}_1}{\to} \mathbf{B}\mathbb{Z}_2orthogonal structure/vielbein/Riemannian metric
general linear groupBGL\mathbf{B}GLsmooth manifold

(all hooks are homotopy fiber sequences)

From the point of view of extended prequantum field theory we are to regard each horizontal map in this tower as classifying the potential underlying instanton sector of an extended Lagrangian/prequantum n-bundle, and we may ask if these lift to actual extended Lagrangian.

The next step up the ladder is the universal second fractional Pontryagin class. Its differential refinement has been constructed in (FSS diff coc).

BString conn 16p 2^ B 7U(1) conn H BG 16p 2 B 7U(1) |Π()| BString 16p 2 B 7U(1) L wheTop. \array{ \mathbf{B}String_{conn} &\stackrel{\widehat {\tfrac{1}{6}\mathbf{p}_2}}{\to}& \mathbf{B}^7 U(1)_{conn} \\ \downarrow && \downarrow &&&&& \in \mathbf{H} \\ \mathbf{B}G &\stackrel{ \tfrac{1}{6}\mathbf{p}_2 }{\to}& \mathbf{B}^7 U(1) \\ && &&& \downarrow^{\mathrlap{{\vert \Pi(-)\vert}}} \\ B String &\stackrel{\tfrac{1}{6}p_2}{\to}& B^7 U(1) &&&&& \in L_{whe} Top } \,.

This now defines an extended Lagrangian for a 7d Chern-Simons theory on String 2-group-2-connections.

Coming back to the above arguments about the relation of the 7d theory to a 6d theory, these suggest that the 6d theory involves prequantum 6-bundle-analog of the WZW 2-bundle.

Indeed, the assignment of a higher Wess-Zumino-Witten-type prequantum (n-1)-bundle to an extended prequantum Lagrangian works very generally: the prequantum (n1)(n-1)-bundle is the WZW-type (n1)(n-1)-bundle of the theory. Its underlying (n1)(n-1)-bundle is simply the looping of the nn-bundle underlying the extended Lagrangian. Specifically, the string 2-group-7d Chern-Simons theory has a WZW- prequantum 6-bundle on the string 2-group which is modulated by the looping Ω(16p 2)\Omega(\tfrac{1}{6}\mathbf{p}_2) of the differential characteristic map 12p 2\tfrac{1}{2}\mathbf{p}_2

String Ω(16p 2) B 6U(1) String// AdString [Π(S 1),BString] [Π(S 1),12p 2] [Π(S 1),B 7U(1)] B 6U(1)//B 6U(1). \array{ String && &\stackrel{\Omega \left( \tfrac{1}{6}\mathbf{p_2}\right)}{\to}& && \mathbf{B}^6 U(1) \\ \downarrow && && && \uparrow \\ String //_{Ad} String & \stackrel{\simeq}{\to} & [\mathbf{\Pi}(S^1), \mathbf{B}String] & \stackrel{[\mathbf{\Pi}(S^1), \tfrac{1}{2}\mathbf{p}_2]}{\to} & [\mathbf{\Pi}(S^1), \mathbf{B}^7 U(1)] &\simeq& \mathbf{B}^6 U(1) // \mathbf{B}^6 U(1) } \,.

Here the lower factorization of this map shows canonically that this 6-bundle is equivariant with respect to the adjoint ∞-action of the string 2-group on itself, the analog of the ordinary adjoint equivariance of the ordinary WZW 2-bundle.

Therefore in summary we find the following identifications of extended prequantum field theory with (semi-)traditional notions:

extended prequantum field theory

0kn0 \leq k \leq n(off-shell) prequantum (n-k)-bundletraditional terminology
00differential universal characteristic maplevel
11prequantum (n-1)-bundleWZW bundle (n-2)-gerbe
kkprequantum (n-k)-bundle
n1n-1prequantum 1-bundle(off-shell) prequantum bundle
nnprequantum 0-bundleaction functional

\infty-Chern-Simons theory as the general example

In the examples of 2d Chern-Simons theory and 7d Chern-Simons theory that we have discussed above, the fully extended Lagrangian is equivalently the prequantum n-bundle on the universal moduli ∞-stack of fields, for n=3n = 3 and n=7n = 7, respectively.

But in fact this is the general situation for fully extended Lagrangians:

if FieldsH\mathbf{Fields} \in \mathbf{H} is the object that serves as the moduli ∞-stack of fields for some nn-dimensional theory in that for any spacetime/worldvolume Σ n\Sigma_n of dimension nn the space of fields is the mapping stack [Σ n,Fields][\Sigma_n, \mathbf{Fields}], then the corresponding prequantum n-bundle has to be over this object Fields\mathbf{Fields} and hence to be modulated by a map of moduli \infty-stacks of the form

L:FieldsB nU(1) conn \mathbf{L} \;\colon\; \mathbf{Fields} \to \mathbf{B}^n U(1)_{conn}

hence by a fully extended Lagrangian.

Above we considered the case that FieldsBG conn\mathbf{Fields} \simeq \mathbf{B}G_{conn} for a Lie group GG or FieldsBString conn\mathbf{Fields} \simeq \mathbf{B}String_{conn} for the string 2-group. In general for GGrpd(H)G \in Grpd(\mathbf{H}) a smooth ∞-group an extended Lagrangian as above encodes an extended prequantum higher gauge theory of Chern-Simons type:

the curvature (n+1)(n+1)-form of L\mathbf{L} (i.e. the pre n-plectic form of the n-plectic ∞-stack Fields\mathbf{Fields} )

F L:BG connLB nU(1) connF ()Ω cl n+1 F_{\mathbf{L}} \;\colon\; \mathbf{B}G_{conn} \stackrel{\mathbf{L}}{\to} \mathbf{B}^n U(1)_{conn} \stackrel{F_{(-)}}{\to} \Omega^{n+1}_{cl}

is – by ∞-Chern-Weil theory – an invariant polynomial of the ∞-Lie algebra 𝔤\mathfrak{g} of GG. Accordingly, the action functional

exp(2πi Σ n[Σ n,L]):[Σ n,BG conn]U(1) \exp\left( 2 \pi i \int_{\Sigma_n} [\Sigma_n, \mathbf{L}] \right) \;\colon\; [\Sigma_n, \mathbf{B}G_{conn}] \to U(1)

locally takes an L-∞ algebra valued differential form AA (the higher gauge fields) to a Lagrangian CS(A)CS(A) which locally is a differential form with the property that dCS(A)=F A d CS(A) = F_A. This means that exp(2πi Σ n[Σ n,L])\exp\left( 2 \pi i \int_{\Sigma_n} [\Sigma_n, \mathbf{L}] \right) is a higher gauge theory of Chern-Simons type. We say that it is an ∞-Chern-Simons theory.

This remains true if Fields\mathbf{Fields} is of more general form, so that fields are L-∞ algebroid valued differential form.

This means that fully extended prequantum field theory is considerably more restrictive than the traditional notion of local Lagrangian field theory, just as extended topological quantum field theory is much more restrictive than non-extended TFT.

  1. fewer choices in prequantization Traditional geometric quantization involves making considerable choices. Notably the choice of prequantum bundle lifting the symplectic form. In extended prequantum field theory the prequantum bundles in codimension 1 are constrained to all arise by transgression from a prequantum n-bundle in codimension nn. This is a strong coherence condition that drastically reduces the available choices.

    For instance for GG a compact Lie group, there is an isomorphism

    H n(BG,)π 0H(BG,B nU(1)). H^n(B G, \mathbb{Z}) \simeq \pi_0 \mathbf{H}(\mathbf{B}G, \mathbf{B}^n U(1)) \,.

    This means that if H n(BG,)H^n(B G, \mathbb{Z}) has no torsion, then there is at most one extended prequantization of a given invariant polynomial.

  2. selecting physically natural theories While the traditional formalization of local Lagrangian field theory does capture crucial aspects of those theories that one is interested in in physics, it still admits many “unnatural theories”. The generic local Lagrangian (in the sense of a horizontal form on the jet bundle of the field bundle) contains many higher order derivatives of fields, non-polynomial interactions etc., properties that the typical theories of relevance in theoretical physics do no share.

The remaining question then is how accurately extended prequantum field theory constrains the space of possible local Lagrangians. In other words: which theories of interest can not be thought of as ∞-Chern-Simons theories or ∞-Wess-Zumino-Witten theory?

As a hint to the answer of this question, here is a list of names examples of ∞-Chern-Simons theories. We discuss the items of the list in detail in the following.

The last example here brings us back full circle to where we started this motivation with above: by the central property of string theory, the action functional for string field theory (or some KK-reduction thereof) contains Einstein-Yang-Mills theory.



Let H\mathbf{H} be an ambient (∞,1)-topos, typically a cohesive (∞,1)-topos such as Smooth∞Grpd or one of its slices. For nn \in \mathbb{N} let B nU(1) connH\mathbf{B}^n U(1)_{conn} \in \mathbf{H} be the moduli ∞-stack of circle n-bundles with connection. Let moreover FieldsH\mathbf{Fields} \in \mathbf{H} be the given moduli ∞-stack of field configurations, so that for Σ\Sigma some worldvolume manifold a field configuration on Σ\Sigma is a map in H\mathbf{H} of the form

ϕ:ΣFields \phi \colon \Sigma \to \mathbf{Fields}

a gauge transformation ϕϕ\phi \Rightarrow \phi' is a homotopy between two such maps, and a higher gauge transformation is a higher homotopy.

Then an extended Lagrangian or prequantum n-bundle for an nn-dimensional quantum field theory with fields classified by Fields\mathbf{Fields} is a map in H\mathbf{H} of the form

L:FieldsB nU(1) conn. \mathbf{L} \;\colon\; \mathbf{Fields} \to \mathbf{B}^n U(1)_{conn} \,.

By post-composition this sends a field pho:Σ nFields\pho \colon \Sigma_n \to \mathbf{Fields} to the circle n-bundle with connection on Σ n\Sigma_n modulated by

L(ϕ):Σ nϕFieldsLB nU(1) conn. \mathbf{L}(\phi) \;\colon\; \Sigma_n \stackrel{\phi}{\to} \mathbf{Fields} \stackrel{\mathbf{L}}{\to} \mathbf{B}^n U(1)_{conn} \,.

This is locally – restricted to some good cover UΣ nU \to \Sigma_n of Σ b\Sigma_b) given by a differential n-form

L(ϕ)ωΩ n(U), L(\phi) \omega \in \Omega^n(U) \,,

where L(ϕ)C (U)L(\phi) \in C^\infty(U) is a smooth function and where ωΩ n(U)\omega \in \Omega^n(U) is some volume form or some other prespecified measure. This function L(ϕ)L(\phi) is the ordinary Lagrangian.

Over an oriented closed manifold Σ n\Sigma_n of dimension nn the corresponding action functional is the transgression of L\mathbf{L} to the mapping stack/internal hom out of Σ n\Sigma_n, hence the composite

exp(2πi Σ n[Σ n,L]):[Σ n,Fields][Σ n,L][Σ n,B nU(1) conn]exp(2πin Σ n)()U(1), \exp\left( 2\pi i \int_{\Sigma_n} [\Sigma_n, \mathbf{L}] \right) \;\colon\; [\Sigma_n, \mathbf{Fields}] \stackrel{[\Sigma_n, \mathbf{L}]}{\to} [\Sigma_n, \mathbf{B}^n U(1)_{conn}] \stackrel{\exp(2 \pi in \int_{\Sigma_n})(-) }{\to} U(1) \,,

where the first map is the internal hom on morphisms and the second is the higher holonomy map of a circle n-bundle with connection over Σ n\Sigma_n, hence the fiber integration in ordinary differential cohomology.

For ϕ:Σ nFields\phi : \Sigma_n \to \mathbf{Fields} a field such that the circle (n-1)-group principal ∞-bundle underlying L(ϕ)\mathbf{L}(\phi) is trivializable and trivialized, the differential form L(ϕ)ωL(\phi)\omega discussed above descends to XX itself and the action functional is simply the integration of differential forms

ϕexp(2πi Σ nL(ϕ)ω). \phi \mapsto \exp\left(2 \pi i \int_{\Sigma_n} L(\phi) \omega \right) \,.

This special case is the relation between Lagrangians and action functionals as traditionally considered.

Generally, for 0kn0 \leq k \leq n and Σ k\Sigma_{k} an oriented closed manifold of dimension kk, transgression as above yields a map in H\mathbf{H} of the form

exp(2πi Σ k[Σ n,L]):[Σ k,Fields][Σ n,L][Σ l,B nU(1) conn]exp(2πin Σ k)()B nkU(1) conn \exp\left( 2\pi i \int_{\Sigma_k} [\Sigma_n, \mathbf{L}] \right) \;\colon\; [\Sigma_k, \mathbf{Fields}] \stackrel{[\Sigma_n, \mathbf{L}]}{\to} [\Sigma_l, \mathbf{B}^n U(1)_{conn}] \stackrel{\exp(2 \pi in \int_{\Sigma_k})(-) }{\to} \mathbf{B}^{n-k}U(1)_{conn}

which modulates a circle (n-k)-bundle with connection on the moduli infinity-stack of fields on Σ k\Sigma_k. At least for large classes of theories, this is the (off-shell) prequantum (n-k)-bundle of the theory.


Relation to universal characteristic classes

By the discussion at field (physics) in the section Properties - Relation to twisted cohomology a field configuration in any theory of physics may equivalently be thought of as a cocycle in twisted general cohomology. Here we discuss that evaluating an extended Lagrangian on such a field configuration is equivalently producing a characteristic class. In other words, under this identification with cohomology an extended Lagrangian is equivalently a universal characteristic map and its equivalence class is the corresponding universal characteristic map.

gauge field: models and components

physicsdifferential geometrydifferential cohomology
gauge fieldconnection on a bundlecocycle in differential cohomology
instanton/charge sectorprincipal bundlecocycle in underlying cohomology
gauge potentiallocal connection differential formlocal connection differential form
field strengthcurvatureunderlying cocycle in de Rham cohomology
gauge transformationequivalencecoboundary
minimal couplingcovariant derivativetwisted cohomology
BRST complexLie algebroid of moduli stackLie algebroid of moduli stack
extended Lagrangianuniversal Chern-Simons n-bundleuniversal characteristic map


For discrete higher gauge groups: \infty-Dijkgraaf-Witten theory

GG a discrete ∞-group

extended Lagrangians are cocycles in group cohomology:

π 0H(BG,B nU(1))π 0Top(BG,B nU(1) disc)H Grp n(G,U(1)) \pi_0 \mathbf{H}(\mathbf{B}G, \mathbf{B}^n U(1)) \simeq \pi_0 Top(B G, B^n U(1)_{disc}) \simeq H_{Grp}^n(G,U(1))

these are the extendd Lagrangians of infinity-Dijkgraaf-Witten theory.

For ordinary gauge theory with

We consider GG a Lie group, to be regarded as the gauge group of a gauge theory. We write BG connH\mathbf{B}G_{conn} \in \mathbf{H} for the smooth moduli stack of GG-principal connections.

Setting FieldsBG conn\mathbf{Fields} \coloneqq \mathbf{B}G_{conn} this is the moduli stack for fields in a pure GG-gauge theory. An extended Lagrangian is hence to be a map

L:BG connB nU(1) conn. \mathbf{L} \;\colon\; \mathbf{B}G_{conn} \stackrel{}{\to} \mathbf{B}^n U(1)_{conn} \,.

For GHG \in \mathbf{H} \coloneqq Smooth∞Grpd a compact Lie group and nn \in \mathbb{N}, then geometric realization of cohesive ∞-groupoids ||:HGrpdL wheTop{\vert - \vert} \colon \mathbf{H} \to \infty Grpd \simeq L_{whe} Top induces a natural isomorphism

||:π 0H(BG,B nU(1))π 0L wheTop(BG,B nU(1))H n+1(BG,). {\vert- \vert} \;\colon\; \pi_0 \mathbf{H}(\mathbf{B}G, \mathbf{B}^n U(1)) \stackrel{\simeq}{\to} \pi_0 L_{whe} Top(B G, B^n U(1)) \simeq H^{n+1}(B G, \mathbb{Z}) \,.

This is discussed at Lie group cohomology. The proof is in (dcct).


This means that the instanton sectors of extended Lagrangians for pure ordinary gauge theory are labeled precisely by the integral cohomology of the classifying space BGB G of the gauge group. In ordinary Chern-Simons theory this is called the level of the theory.

Higher gauge theory

If GGrp(H)G \in Grp(\mathbf{H}) is an ∞-group (a smooth ∞-group if H\mathbf{H} is Smooth∞Grpd) and

FieldsBG conn \mathbf{Fields} \simeq \mathbf{B}G_{conn}

is a moduli ∞-stack of GG-principal ∞-connections, then the theory in question is a (higher) gauge theory. In this case the map

c:BGB nU(1) \mathbf{c} \;\colon\; \mathbf{B}G \to \mathbf{B}^n U(1)

underlying the extended Lagrangian is a cocycle that induces an extension of ∞-groups, fitting a homotopy fiber sequence

B n1U(1)BG^BGcB nU(1). \mathbf{B}^{n-1}U(1) \to \mathbf{B}\hat G \stackrel{}{\to} \mathbf{B}G \stackrel{\mathbf{c}}{\to} \mathbf{B}^n U(1) \,.

In this case the extended Lagrangian also plays the role as the universal differential refinement of the obstruction class to lift of structure groups through G^G\hat G \to G. More generally, its homotopy fiber products are moduli ∞-stacks of twisted differential c-structures. See there for more examples.

Construction of extended Lagrangians by Lie integration

Lie integration:

for 𝕘\mathbb{g} an L-∞ algebra and

μ:𝔤b n1 \mu \;\colon\; \mathfrak{g} \to b^{n-1}\mathbb{R}

a homomorphism of L L_\infty-algebras into the line Lie n-algebra, hence a cocycle in L-∞ algebra cohomology, the operation of differentially refined Lie integration (FSS diff coc) produces a morphism of moduli ∞-stacks

exp(μ)^:exp(𝔤) connB n conn \widehat {\exp(\mu)} \;\colon\; \exp(\mathfrak{g})_{conn} \to \mathbf{B}^n \mathbb{R}_{conn}

Under truncation this yields

BG connB n(/Γ), \mathbf{B}G_{conn} \to \mathbf{B}^n (\mathbb{R}/\Gamma) \,,

where GΩτ nexp(𝔤)G \coloneqq \Omega \tau_n \exp(\mathfrak{g}) and where BG conn\mathbf{B}G_{conn} is the moduli of GG-∞-connections. (FSS)

Higher Chern-Simons theory

Various examples are discussed at

extended prequantum field theory

0kn0 \leq k \leq n(off-shell) prequantum (n-k)-bundletraditional terminology
00differential universal characteristic maplevel
11prequantum (n-1)-bundleWZW bundle (n-2)-gerbe
kkprequantum (n-k)-bundle
n1n-1prequantum 1-bundle(off-shell) prequantum bundle
nnprequantum 0-bundleaction functional


Expositions, examples and further pointers are in

Lecture notes are in geometry of physics, and section 1.2 of

A general abstract discussion is in section 3 there, further discussion of examples in section 5.

The construction of extended Lagrangian by Lie integration of L-∞ algebra cocycles is due to

Construction of differential cup-product theories is in

The discussion of the abelian 7d Chern-Simons theory involved in AdS7/CFT6 duality is due to (Witten 98). The discussion of the non-abelian quantum-corrected version is due to

Last revised on January 17, 2013 at 04:26:17. See the history of this page for a list of all contributions to it.