Whether the earliest records of Earth were formed by plate tectonic or non-plate tectonic processes remains debated. One way to explore this is to investigate the preserved early Earth records and compare them with model predictions. For this purpose, it is crucial to consider plate and non-plate tectonic possibilities for geological records of unknown or controversial origins. However, unlike the plate tectonic processes that predict well-known testable signatures, basic mechanisms and geological predictions of non-plate tectonic models are poorly understood. These knowledge gaps have resulted in improper evaluations of the viability of non-plate tectonic models. To address these knowledge gaps, we review the kinematics of representative early tectonic models, explore their testable geological predictions, and evaluate the uniquenesses of these predictions. We also list key contributions that have provided model tests and highlight predictions that remain to be tested. Models discussed here include (proto-)plate tectonics, heat-pipe tectonics characterized by rapid volcanism and downwards lithospheric advection, partial convective overturn tectonics featured by episodic volcanism and gravity-driven crustal (partial) convection, and sluggish-lid tectonics driven by vigorous mantle convection. Many geological predictions that are commonly thought to be unique to plate tectonics are non-unique. For example, heat-pipe tectonics and partial convective overturn tectonics also predict material exchanges between the convecting mantle and surface, subduction zone-like geotherms (globally and locally, respectively), as well as volcanic arc-like detrital zircon ages patterns in some basins. Therefore, preservation of such geological records cannot be used to uniquely identify operation of plate tectonics. In contrast, some geology is limited to certain models. Examples include plate drifting that is only compatible with sluggish-lid tectonics and plate tectonics; and thinning of lithosphere over time that is exclusive to heat-pipe tectonics. Our review shows that (1) plate tectonics predicts diverse crustal settings and is nearly impossible to exclude as a viable early Earth setting; and (2) although non-plate tectonic models predict relatively limited crust-mantle configurations, they remain viable models for Hadean to early Archean time. Additional model tests involving (a) novel analytical methods, (b) new early materials, and/or (c) untested geological predictions are required to narrow the viable range of early Earth tectonic models.