Notes on Liouville Theory at c<=1
Will McElgin
TL;DR
The paper investigates continuing Liouville conformal field theory to $c\le 1$, examining a timelike boson interpretation for 2D de Sitter geometries and applying the conformal bootstrap to obtain three-point functions. It shows a unique analytic factor identical to the minimal-model case, but a non-analytic factor is necessary to produce a diagonal two-point function, which exists only for central charges of topological minimal models with $h>(c-1)/24$, while the full spectrum remains continuous and typically non-degenerate. By constructing shift relations and their solutions, the work connects spacelike Liouville amplitudes to a non-rational $c\le 1$ sector (the $C_M$ theory) and discusses their implications for minimal strings, showing that sensible diagonal metrics arise only in special topological cases. The analysis reveals a nuanced landscape: a timelike interpretation is problematic in general, but the $c=13-6(q^{-1}+q)$ topological sector yields well-defined amplitudes and a diagonal inner product, suggesting a constrained but meaningful link between Liouville continuation, non-rational CFTs, and minimal-string theories. Overall, the results delineate the precise conditions under which non-rational Liouville continuations yield consistent CFTs and illuminate the role of dual potentials in 2D gravity contexts.
Abstract
The continuation of the Liouville conformal field theory to c<=1 is considered. The viability of an interpretation involving a timelike boson which is the conformal factor for two-dimensional asymptotically de Sitter geometries is examined. The conformal bootstrap leads to a three-point function with a unique analytic factor which is the same as that which appears along with the fusion coefficients in the minimal models. A corresponding non-analytic factor produces a well-defined metric on fields only when the central charge is restricted to those of the topological minimal models, and when the conformal dimensions satisfy h>(c-1)/24. However, the theories considered here have a continuous spectrum which excludes the degenerate representations appearing in the minimal models. The c=1 theory has been investigated previously using similar techniques, and is identical to a non-rational CFT which arises as a limit of unitary minimal models. When coupled to unitary matter fields, the non-unitary theories with c<=-2 produce string amplitudes which are similar to those of the minimal string.
