Active faults accommodate tectonic plate motion through different slip modes, some stable and aseismic, others characterized by the occurrence of large earthquakes after long periods of inactivity. Although the slip mode estimation is of primary importance to improve seismic hazard assessment, this parameter inferred today from geodetic observations needs to be better constrained over many seismic cycles. From an analytical formulation developed for analyzing fault scarp formation and degradation, we show that the final topographic shape generated by one earthquake rupture or by creep (i.e., continuous slip) deviates by as much as 10–20%. This result opens up the theoretical possibility of inverting the number of earthquakes and their sizes from scarp morphologies. Our approach highlights the importance of trade-offs between fault slip history and diffusive processes. An identical topographic profile can be obtained either with stable fault creep together with rapid erosion, or with a single earthquake rupture followed by slow erosion. Our findings reveal that the previously noticed variation of the diffusion coefficient with the age of the scarp may be an artifact related to the history and mechanism of fault slip. These inferences, derived from the simplest possible diffusion model, are likely to be even more pronounced in nature.