In the last decade we have seen significant progress toward the broad goal of self-consistently generating tectonic plates in models of mantle convection.  In general we have seen that plate-like motion in the top, cold thermal boundary layer of convection requires rheologies more exotic than we typically use for non-Newtonian mantle flow (i.e., power-law creep with an index of n=3).  Basic plastic lithospheric rheologies appear to go quite far in generating plate-like motion when coupled to a low-viscosity underlying asthenosphere, although they generally have trouble generating concentrated strike-slip (toroidal) motion. The more extreme self-lubricating (sometimes called shear weakening or pseudo-stick-slip) rheologies can generate strike-slip motion, but the weak zones it generates tend to over-focus beneath the grid scale. Moreover, these rheological mechanisms involve instantaneous response and thus the softened plate boundaries they generate vanish as soon as deformation stops.  Time- or history-dependent weakening mechanisms that rely on rheological response to temperature, grain-size or volatile content can (in some cases) yield narrow plate boundaries and possibly long-lived inactive boundaries,  however they also appear to destabilize the plate-like motion.  The most effective of these time-dependent mechanisms involve damage and volatile ingestion of the lithosphere; this approach is also well-motivated by the apparent predominance of ductile-cracking in the lithosphere within the large transition region that lies between the brittle-ductile and brittle-plastic transitions. However, most of these `damage' theories are ad hoc and over-simplified. In an effort to better understand the damage approach, and provide a theory compatible with viscous mantle flow, we propose a new two-phase damage theory. This theory states that 1) a damaged material (i.e., with microcracks and voids) is, in its simplest manifestation, a two phase material (a matrix phase representing solid host, and a fluid phase representing void-filling material such as water or air); and 2) the energy going into making microcracks is the surface energy of the crack wall, which in the two phase model is the surface energy of the interface between the phases.  This model describes a spectrum of rich behavior including damage, weakening, and shear localization.