The generalizations in Moore (2006) seem to hold with English and many other languages reported hitherto. They need to be tested for their cross-linguistic validity, however. In this paper we investigate the temporal metaphors in two Asian languages namely Marathi and Japanese and demonstrate that Moore's predictions are not fully borne out in these languages and propose modifications. In sum, while Marathi and Japanese data endorses Moore's predictions about positional terms and Moving Ego, it runs counter to his predictions about unmarked non-deictic FRONT/BACK-EARLIER/LATER relations. We propose, therefore, to combine these predictions to construct a more plausible set of hypotheses that can account for spatio-temporal expressions of these Asian languages. (Kazuko Shinohara and Prashant Pardeshi)