Do deep mantle plumes feed volcanic hotspots such as Hawaii, or do shallower plate processes explain them?

Since its introduction the plume hypothesis has been repeatedly adapted to fit observed hotspots rather than making firm predictions, and critics have developed a rival plate hypothesis; with the growing number of models the plume concept has been criticized as a weakly defined hypothesis that as a general term is currently neither provable nor refutable. Seismic tomography has produced evidence read as favoring a plume beneath Hawaii and, since 2011, beneath Yellowstone, but the general question stays contested.