Unlike the Hertzsprung–Russell diagram for stars, there remains no formal classification for solid exoplanets composed of varying proportions of gas, rock+metals and ice. Still, as with stars, planetary mass and composition – expressed in geochemical and cosmochemical terms – mold bulk physical characteristics. Two physical attributes control terrestrial-type planet interior dynamics: viscosity (η) and intrinsic heat production (A). Viscosity can differ by orders of magnitude between different common mantle silicate minerals (e.g. olivine, pyroxene), so that even small proportional changes yield large differences to η. A key parameter to consider in this context is (Mg:Si:Fe), because this value largely determines which minerals will be present in silicate mantles. Bulk Silicate Earth's (Mg:Si:Fe) is close to solar values, and we can assume that this also holds for terrestrial-type exoplanets in that they follow the compositions of their host stars. Transition between mechanically weak (olivine-dominated at (Mg/Si)≤1, low η) vs. strong (pyroxene-dominated at (Mg/Si)>1, high η) mantle convective regimes occurs over a narrow transitional range of (Mg/Si) values because small volume fractions of a weak phase are sufficient to form an interconnected network that in turn governs the strain response of mantle rocks to deforming stresses acting upon them. Heat production in younger planets ought to be greater from more radioactivity and latent accretionary/gravitational heating vs. older (cooler) ones. This has important consequences for how heat loss is accommodated by interior dynamics and how it is expressed via outgassing to secondary atmospheres. Here I show how combining geodynamics with astrophysical observations provides insights to terrestrial exoplanet η and A vs. age. Younger (≤2 Gyr) stars tend to have low (Mg/Si)≤1. If these stars mirror the silicate mantles of their rocky exoplanet companions, we forecast that such younger low (Mg/Si) pyroxene-rich rocky exo-mantles ought to tend towards both high η and A, with episodic sluggish/rapid convection and thus slow cooling, and low oxygen fugacity that degas H2 and CH4 under near-surface partial melting conditions. Contrariwise, older (>5 Gyr) olivine-rich (high Mg/Si) oxidized (like Earth) exo-mantles should tend towards both low η and A, effectively cool, and degas N₂, CO₂, H₂O. By implication, a fundamental age-composition dichotomy is anticipated to exist between young (hot, reduced, Fe-rich) and old (cold, oxidized, Fe-poor) rocky exoplanets that can already be evaluated by mass-radius density-age data. I test, with recent atmospheric retrieval data, the predictability of such models with an example of an "Ultra-hot Jupiter" hosted by a Sun-like main sequence star.