We consider a two-dimensional system with two order parameters, one with O(2) symmetry and one with O(M ), near a point in parameter space where they couple to become a single O(2 + M ) order. While the O(2) sector supports vortex excitations, these vortices must somehow disappear as the high symmetry point is approached. We develop a variational argument which shows that the size of the vortex cores diverges as 1/ √ ∆ and the Berezinskii-Kosterlitz-Thouless transition temperature of the O(2) order vanishes as 1/ ln(1/∆), where ∆ denotes the distance from the highsymmetry point. Our physical picture is confirmed by a renormalization group analysis which gives further logarithmic corrections, and demonstrates full symmetry restoration within the cores. [7]. In all these cases, at least one of the two competing order parameters has XY, i.e. O (2) symmetry, while the other may in general be O(M ). The cases M = 1, 2, and 3 correspond to the other order parameter being of Ising, XY, or Heisenberg type, respectively. The M = 1 case describes easy-plane magnetism [8,9]; the case M = 2 relates to supersolid phases in cold-atom systems [5] and to competing density-wave and superconducting order in layered materials [10][11][12]; models with M = 3 have been considered in the context of high-temperature superconductivity [13]. Both order parameters interact and the symmetry of the coupled problem is O(M ) × O(2). However, at a certain fine-tuned point in phase space one may expect the symmetry to be enhanced, from O(M ) × O(2) to O(N ) with N = M + 2. This is not the most general scenario for competition between two order parameters, but it has been conjectured to occur in many different microscopic models, including all of the cases mentioned above [5][6][7][8][9][10][11][12][13].This symmetry enhancement acquires a particularly interesting aspect in layered or two-dimensional systems, where long range order is absent for continuous symmetries due to the Hohenberg-Mermin-Wagner theorem [14]. However, the O(2) sector supports non-trivial topological configurations, i.e. vortices. The unbinding of vortex-antivortex pairs converts an algebraically ordered superfluid or crystal to a disordered normal fluid. This Berezinskii-Kosterlitz-Thouless (BKT) transition [15][16][17] occurs at a non-zero temperature T BKT .Suppose the fine tuning to a high-symmetry point in the phase diagram is achieved by varying a dimensionless parameter ∆ > 0 towards ∆ = 0 which corresponds to the O(N ) symmetry point. In each realization of this model, the experimental handle corresponding to our parameter ∆ is different -for example in the context of cuprates/organics it would correspond to doping/pressure [13] while for cold dipolar bosons it may be controlled via the angle of a polarizing field [5]. However, no matter which particular microscopic realization of this model is chosen, if N > 2 then T BKT must vanish for ∆ → 0. Indeed, combining spin-wave based renormalization group calculations with crossover arguments one can estimate that T BKT...