Large-scale deformation of continents remains poorly understood more than 40 years after the plate tectonic revolution(1). Rock flow strength and mass density variations both contribute to stress, so both are certain to be important, but these depend (somewhat nebulously) on rock type, temperature and whether or not unbound water is present(2). Hence, it is unclear precisely how Earth material properties translate to continental deformation zones ranging from tens to thousands of kilometres in width, why deforming zones are sometimes interspersed with non-deforming blocks and why large earthquakes occasionally rupture in otherwise stable continental interiors. An important clue comes from observations that mountain belts and rift zones cyclically form at the same locations despite separation across vast gulfs of time(3) (dubbed the Wilson tectonic cycle), accompanied by inversion of extensional basins(4) and reactivation of faults and other structures formed in previous deformation events(5). Here we show that the abundance of crustal quartz, the weakest mineral in continental rocks(2), may strongly condition continental temperature and deformation. We use EarthScope seismic receiver functions(6), gravity and surface heat flow measurements(7) to estimate thickness and seismic velocity ratio, v(P)/v(S), of continental crust in the western United States. The ratio v(P)/v(S) is relatively insensitive to temperature but very sensitive to quartz abundance(8,9). Our results demonstrate a surprising correlation of low crustal v(P)/v(S) with both higher lithospheric temperature and deformation of the Cordillera, the mountainous region of the western United States. The most plausible explanation for the relationship to temperature is a robust dynamical feedback, in which ductile strain first localizes in relatively weak, quartz-rich crust, and then initiatesprocesses that promote advective warming, hydration and further weakening. The feedback mechanism-proposed here would not only explain stationarity and spatial distributions of deformation, but also lend insight into the timing and distribution of thermal uplift(10) and observations of deep-derived fluids in springs(11).