We welcome this opportunity which Drs. Ricard, Sabadini and Spada (RSS) have offered us to elaborate several points on the relationship between convective circulations and rotational dynamics. First of all, mantle convection is a non-linear fluid dynamical phenomenon with some form of non-linear rheology. The usage of linear viscoelastic modes cannot hope to capture many of the complex features in mantle convection, such as the catastrophic flushing events. The relationship of the linear viscoelastic normal-mode theory to other approaches, such as viscous and the non-linear viscoelastic treatments, can be found in Moser et al. (1993). In this reply we will address the issues raised by RSS by first going over the reference system. This will be followed by a discussion concerning long-term polar wander and relative angular momentum. Then we go over the issues concerning the off-diagonal terms in the total moment of inertia and polar wander. Finally, the limitations of the normal-mode approach used by RSS are pointed out within the context of recent developments in geophysics.