After more than 35 years of TTG (tonalite-trondhjemite-granodiorite) research, we still face many questions about the origin and tectonic significance of these peculiar rocks. What we do know is that TTGs are similar in composition regardless of age, they have high La/Yb, Sr/Y, Sr and Eu/Eu*, they decrease in abundance relative to calc-alkaline granitoids at the end of the Archean, and they are not made in oceanic arcs, shallow levels of oceanic plateaus or at ocean ridges. Furthermore, oxygen isotopes in TTG zircons require interaction of TTG sources with the hydrosphere, and the existence of Hadean continental crust inferred from detrital zircon suites remains problematic. Although we now realize that TTGs require amphibole and garnet fractionation and sources that are at least 50 km deep, what we do not know are the relative roles of (1) melting versus fractional crystallization and (2) melting of slabs versus melting of thickened mafic crust. The mechanisms and rates of slab dehydration control the stability of garnet and amphibole in subduction zones. From what we know about early Archean greenstones, they are more altered than later ones, and thus they would appear to bring more water and fluid-mobile elements into subduction zones, at least by the late Archean when plate tectonics became widespread. Hotter slabs in the Archean should contribute to higher volatile release rates. This may explain the trace element changes we see in TTGs at the end of the Archean. To make continental crust today we need to start at a continental subduction zone where we produce both calc-alkaline (CA) and TTG magmas, and combine the felsic components in a ratio of about 3 parts CA to 2 parts TTG. In contrast, to make an Archean continent, we need nearly 100 % of the TTG component, and may begin, at least before about 3 Ga, by melting the roots of oceanic plateaus.