Reflecting on *se-/s(e)we-: From typology to Indo-European and back

Indo-European *se-/s(e)we- is traditionally classified as a reflexive stem (Brugmann and Delbruck 1893). However, the earliest attested Indo-European languages (Old Indic, Gathic Avestan, Hittite) show no clear traces of a reflexive use of *se-. Significantly, Hahn (1963) and Meyer (1997) have proposed that its original function was anaphoric. However, they assume that Gk., epsilon if stressed, had a reflexive function. I will argue that even Gk. epsilon is never used in "prototypical" reflexive situations. On the basis of the reconsideration of *se-/s(e)we-, a general reinterpretation of the problem of the 3(rd) person pronoun in Indo-European will be given.