Metamorphic core complexes (MCCs) are interpreted as domal structures exposing ductile deformed high-grade metamorphic rocks in the core underlying a ductile-to-brittle high-strain detachment that experienced tens of kilometres of normal sense displacement in response to lithospheric extension. Extension is supposedly the driving force that has governed exhumation. However, numerous core complexes, notably Himalayan, Karakoram and Pamir domes, occur in wholly compressional environments and are not related to lithospheric extension. We suggest that many MCCs previously thought to form during extension are instead related to compressional tectonics. Pressures of kyanite-and sillimanite-grade rocks in the cores of many of these domes are c. 10–14 kbar, approximating to exhumation from depths of c. 35–45 km, too great to be accounted for solely by isostatic uplift. The evolution of high-grade metamorphic rocks is driven by crustal thickening, shortening, regional Barrovian metamorphism, isoclinal folding and ductile shear in a compressional tectonic setting prior to regional extension. Extensional fabrics commonly associated with all these core complexes result from reverse flow along an orogenic channel (channel flow) following peak metamorphism beneath a passive roof stretching fault. In Naxos, low-angle normal faults associated with regional Aegean extension cut earlier formed compressional folds and metamorphic fabrics related to crustal shortening and thickening. The fact that low-angle normal faults exist in both extensional and compressional tectonic settings, and can actively slip at low angles (< 30°), suggests that a re-evaluation of the Andersonian mechanical theory that requires normal faults to form and slip only at high angles (c. 60°) is needed.