Inferences about plate motions over the past 40 Ma, based on structural and geodetic data, suggest that there is a coherent movement direction on a global scale. Moreover a ‘westward’ drift of the lithosphere relative to the asthenosphere is indicated by plate motions in the hot-spot reference frame. Independent geological observations such as the different styles of deformation in thrust belts associated with subduction following or opposing the relative ‘eastward or northeastward’ mantle flow also support this global tectonic polarity. Unfortunately most hot-spot reference frames are based on assumptions which strongly limit their validity. For example, many combine magmatic sources originating from different depths in the mantle, including hot-spots located along oceanic ridges which are not fixed relative to one another. Using a hot-spot reference frame filtered to exclude shallow hot-spots and anomalous ‘wet-spots’, we would predict a much higher relative ‘westward’ drift of the lithosphere. Plate motions are not random, rather that they are controlled by the relative motion between lithosphere and asthenosphere, and by gradients in lithospheric thickness and composition.Triangular shaped plates may experience rotations resulting from differential drag at the base of the lithosphere with respect to the asthenosphere (e.g. the clockwise rotation of South America). Plates appear to follow an undulatory pattern of motion on a global scale and their poles of rotation fall into a main cone located near the Earth’s rotation poles. Plate rotations induced by triangular shape, differential basal drag or lithospheric anisotropies along subduction zones may generate the scattered distribution of Euler poles. This might have inhibited the recognition of a coherent global tectonic pattern.