Studies of high-pressure-ultrahigh-pressure (HP-UHP) belts stand to benefit from having a better idea of how subduction constructs them. Their exhumation, too, merits a closer look at epeirogenic processes. Progress in both these matters now appears possible and their relevance tested, using the Alps, the world's most comprehensively studied such belt, as the prime example. Its comparative youth means that less of the potentially pertinent surface detail has been lost by erosion or burial.Much evidence suggests that subduction downbend is actually not a flexure but is a seismically evident, escalator-like, through-plate step-faulting process. Especially where the subducting plate is younger than about 90 Ma, this gives it the property (STE) of incorporating upper-plate material in the step-angles, thus moving forward the position of the downbend surface that it has carved in the upper plate rapidly, and at a shallow angle for hundreds of kilometers. This yields a nominally "flat" interface profile, above which, at the shallow end of the system, the upper-plate crust will have been chilled and perhaps had its lower part removed, thus providing an origin for basement thrust sheets and (as subducted imbricate slices) those seen in HP-UHP belts.Seismology shows that "flat" interface profiles often pass through an inflection nearer the trench. This means that STE is "taking a second cut" and then disgorges the step-angle material along the succeeding flat part of the interface. When a STE-thinned margin imbricates, this diagnostic material appears as a fluid-overpressured tectonic mélange "buttered" (no shearing at the contact) onto the underside of each. Included floaters (up to >5 km in size) exhibit many lithologies, even slickensided, in an ex-oceanic, usually pelitic matrix that protected them from deformation. This "butter" is a focus of attention in HP-UHP belts, especially for its mafic-ultramafic floaters, while the water in its matrix seems important for reaction kinetics, for eventual epeirogenic action, and even to flux partial melting.Subduction of imbricate slices, already cooled by STE beneath them, and their successive lodgement beneath the "roof" and angle of the distant downbend profile carved by STE into the hanging wall, builds a subcreted triangular section extending far down the back wall, much further than hitherto thought possible. So this part exhumes the most, breaking through the "roof" above as a fault-line prone to strike-slip (Insubric, Møre-Trøndelag, Median Tectonic Line).In the Alps, and following extensive STE from the north, successive imbrication marked by Cretaceous flysch sequences transposed the paleogeographic order, so the Piemont ophiolites were from an oceanic-crusted northern forearc, not a separate ocean. Such transpositions cannot happen during collision; they require too much transport. During Eocene collision, the old block-and-basin structure in the European autochthon was epeirogenically rejuvenated (external massifs, Tauern window) by deep-crust heating and petrologic...