1: \begin{abstract}

2: \noindent

3: According to the standard classification of Conformal Quantum Field Theory

4: (CQFT) in two dimensions, the massless continuum limit of the $O(2)$

5: model at the Kosterlitz-Thouless (KT) transition point should be given

6: by the massless free scalar field; in particular the Noether current of

7: the model should be proportional to (the dual of) the gradient of the

8: massless free scalar field, reflecting a symmetry enhanced from $O(2)$

9: to $O(2)\times O(2)$. More generally, the massless continuum limit

10: of a spin model with a symmetry given by a Lie group $G$ should have an

11: enhanced symmetry $G\times G$. We point out that the arguments leading

12: to this conclusion contain two serious gaps: i) the possibility of

13: `nontrivial local cohomology' and ii) the possibility that the current

14: is an ultralocal field. For the $2D$ $O(2)$ model we give analytic

15: arguments which rule out the first possibility and use numerical

16: methods to dispose of the second one. We conclude that the standard

17: CQFT predictions appear to be borne out in the $O(2)$ model, but give

18: an example where they would fail. We also point out that all our arguments

19: apply equally well to any $G$ symmetric spin model, provided it has a 

20: critical point at a finite temperature.

21: \end{abstract}

22: