Evidence-based interpretations of fluvial evolution, and especially of river-terrace formation, have advanced significantly in recent decades, with a notable contribution made by activities of the Fluvial Archives Group. Well-dated river-terrace sequences provide frameworks for the understanding of landscape evolution, since they record valley-floor levels that were higher in the past, attributable, from their patterns of occurrence, to regional uplift. The role of climate fluctuation during the Quaternary is also paramount, since this has been an important driver of the varied fluvial activity that has given rise to the staircases of terraces that characterise the temperate latitudes. This approach is contrasted with a more theoretical methodology for using rivers as recorders of landscape evolution, again with an emphasis on uplift, based on the concept of the formation of knickpoints at particular base levels and their migration upstream. Although different timescales can be explored by the two methods, the concept of headward-migrating knickpoints implies a mechanism for incision that is difficult to reconcile with the formation of the broadly parallel river terraces that are observed in many systems. Knickpoints can frequently be observed to coincide with gorge reaches, where river valleys are constricted as a result of resistant bedrock and/or the effects of localised active crustal deformation. This raises the possibility that knickpoints have generally formed in response to factors of local geology rather than migrating from downstream.