In introductory textbooks & lecture notes on conformal field theory, it is usually stated that solving the highly nontrivial dimensional quantum field theory in 2 dimensions is possible due to the existence of the powerful infinite-dimensional symmetry. The symmetry under consideration is, of course, the Virasoro symmetry.

Now consider a completely different class of 2-dimensional field theories – those invariant under the group of area-preserving diffeomorphisms. Here the Lie algebra of the symmetry group consists of vector fields of vanishing divergence (on the flat $\mathbb{R}^2$, this means $\partial_{\mu} v^{\mu} = 0$). This is also an infinite-dimensional symmetry algebra, so naively, I expect it to be powerful enough to allow us to solve/classify 2-dimensional field theories invariant under area-preserving diffeomorphisms.

My questions is: to what extent can similar methods be applied to such theories? What are the most noteworthy results?

Finally, some motivation: at least classically, the 2-dimensional Yang-Mills theory is invariant under the area-preserving diffeomorphisms. We already know that there exists a non-abelian conformal gauge theory in 2 dimensions (the Wess-Zumino-Witten model) which is exactly solvable. Maybe the 2-dimensional Yang-Mills theory admits a similar definition, only its "blocks" are built by using "primary fields" of a different infinite-dimensional algebra rather than the Kac-Moody algebra?

I've heard that the 2-dimensional quantum Yang-Mills theory has been constructed rigorously by solving for certain heat kernels on the gauge group $G$. I wonder if there's an equivalent formulation which is more "Wightmanian" in its nature, specifically, one that utilizes explicitly specified $n$-point blocks that are used to build the correlation functions.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – Chris
    Aug 5 '19 at 22:57
  • $\begingroup$ Have a look at Iliopoulos 2018. $\endgroup$ Aug 6 '19 at 16:47

To be concrete let spacetime be 2D with Minkowski signature.

  1. On one hand, the infinite-dimensional algebra $${\rm locconf}(1,1)~=~{\rm Vect}(\mathbb{S}^1)\oplus {\rm Vect}(\mathbb{S}^1)\tag{1a}$$ of locally defined conformal transformations in 2D is the algebra of conformal Killing vector fields (CKVF), which is 2 copies of the real Witt algebra, cf. e.g. this Phys.SE post. In light-cone coordinates, a CKVF satisfies $$ \partial_{\mp}X^{\pm}~=~0\qquad\Leftrightarrow\qquad X^{\pm}=X^{\pm}(x^{\pm}). \tag{1b}$$

  2. On the other hand, the infinite-dimensional algebra $${\rm sdiff}(1,1)\tag{2a}$$ of 2D area-preserving transformations is the algebra of symplectic vector fields (SVF), which includes the algebra of Hamiltonian vector fields (HVF). In light-cone coordinates, a HVF is of the form $$X^{\pm}~=~\pm\partial_{\pm}H, \tag{2b}$$ where $H=H(x^+,x^-)$ is an arbitrary function.

  3. The two infinite-dimensional algebras (1) & (2) behave naturally like oil & water, i.e. they almost don't mix: Their intersection is quite small, namely it is the 2D Poincare algebra $${\rm iso}(1,1),\tag{3}$$ which is generated by 1 Lorentz boost and 2 translations.

  4. However, the algebra (2) is much bigger than the algebra (1) in the sense that it is often possible to embed a Virasoro/$W$-algebra inside (2), cf. Refs. 1-2. So in that sense the conformal methods apply to theories invariant under area-preserving diffeomorphisms, cf. OP's question.


  1. E. Sezgin, Area-Preserving Diffeomorphisms, $w_\infty$ Algebras and $w_\infty$ Gravity, arXiv:hep-th/9202086.

  2. D.V. Artamonov, Introduction to finite $W$-algebras, arXiv:1607.01697.

  • $\begingroup$ Awesome, thanks for writing this up. So because $sdiff$ is bigger than $locconf \sim Vir$, there should be additional constraints on the Virasoro representations? How hard is it to solve these constraints e.g. for the minimal models? Has Yang-Mills in 2d been constructed in this way and if it has, what is the corresponding CFT? $\endgroup$ Aug 6 '19 at 20:52
  • $\begingroup$ There is an error in the labelling of items in point 3, and possibly point 4. In particular, 3. seems like it should say "The infinite-dimensional algebras (1) and (2) behave..." and so I infer that 4. should say "However, the algebra (2) is much bigger than the algebra (1) in the sense..." I'm leaving this as a comment rather than editing, though, in case it is clear to someone with more expertise that 4., as written, is correct. $\endgroup$ Aug 7 '19 at 15:48
  • $\begingroup$ Hi @MatthewTitsworth. Thanks for the feedback. I tried to relabel for clarity. $\endgroup$
    – Qmechanic
    Aug 7 '19 at 16:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.