Mineral zoning studies in several granulite terranes have yielded nearly isobaric retrograde pressure-temperature (P-T) paths with significant path dispersion. Explanations include magmatic underplating, incomplete isostatic rebound following thrusting, and cooling after tectonic denudation associated with crustal extension. One-dimensional thermal models help to interpret the P-T-t (time) paths generated by such processes. Basaltic underplating yields P-T paths in good agreement with data for the Grenville province of Ontario, but yields T-t paths that disagree strongly with thermochronometry. In addition, seismic, heat-flow, and gravity data from the Grenville province are not consistent with a mafic lower crust. Calculations based on simple thrust models agree with the T-t data, but yield isothermal rather than isobaric initial P-T paths, and consistently underestimate the pressure at which peak temperature is achieved. The only model we found that fits both the P-T and T-t data incorporates a post-thrusting episode of tectonic denudation near the peak of metamorphism. This model also suggests that the temperature dependence of diffusion in garnet may explain the difference between measured isobaric and isothermal paths in granulites and amphibolites.