Extensive tracts of low-gradient topography in steep mountain ranges, either forming rangetop plateaus or terraced pediments on range flanks, are widely distributed in mountain belts around the world. Before the advent of plate tectonics, such populations of planar landforms were interpreted as vestiges of a post-orogenic raised peneplain, i.e., a low-gradient land surface resulting from the decay, during long intervals of base-level stability, of a previous mountain range that was subsequently raised once again to great elevations—thus forming a new mountain range. This two-stage model has been challenged by theories that advocate continuity in tectonic processes and more gradual changes in base level, and thus expect a more immediate and proportionate response of geomorphic systems. Here we present a global survey of erosion surfaces in mountain ranges and put existing theories and empirical evidence into a broad perspective calling for further research into the rates and regimes of long-term mountain evolution. The resulting library of case studies provides opportunities for comparative analysis and helps to classify the landform mosaics that are likely to arise from the interplay between (i) crustal regimes, which at convergent plate margins need be neither uniform nor steady at all times; (ii) radiation-driven and gravity-driven geomorphic regimes, which are mainly determined by crustal boundary conditions and climate; and (iii) paleogeography, through which clues about base-level changes can be obtained. We examine intracratonic and plate-margin settings, with examples from thin-skinned fold belts, thick-skinned fold belts, island-arc and other subduction-related settings, and bivergent collisional orogens. Results reveal that the existence of erosion surfaces is not a simple function of geodynamic setting. Although some erosion surfaces are pre-orogenic, evidence about their predominantly post-orogenic age is supported by apatite fission-track and helium rock-cooling signatures, stratigraphic age-bracketing, stream channel gradient patterns, and other direct or indirect dating criteria. It follows that many portions of mountain belts undergo unsteady, nonuniform post-orogenic landscape evolution trajectories, with intermittent opportunities for relief reduction. The resulting erosion surfaces remain preserved as signatures of transient landscape evolution regimes. We find that (i) occurrences of planar topography form populations of discrete, insular landscape units, only some of which could be interpreted as fragments of a fluvially dissected, and/or tectonically fragmented, regional peneplain. (ii) The post-orogenic time required for achieving advanced stages of relief reduction is variable, ranging from 3 to 70 Ma. (iii) Partly depending on whether the adjacent sedimentary basins were over- or underfilled, some erosion surfaces may have been controlled by raised base levels and may thus have formed at high elevations; however, in many cases they were disconnected from marine base levels by rapid surface uplift, thus acquiring their elevated positions in recent time. In some cases, subcrustal processes such as asthenospheric anomalies, and/or lithospheric slab tear or breakoff, explain extremely rapid, regional post-orogenic uplift. (iv) Overall, the conditions for achieving surface preservation in steep and tectonically active terrain are predictable but also quite varied and contingent on context.