Abstract

Orthographic tt in Old Khotanese has often been held as always representing single phonemic /t/. In the 1930s Ernst and Manu Leumann suggested that in four categories of words intervocalic tt represents double /tt/. Here I expand these four categories and analyze the position of these words within the metrical structure of The Book of Zambasta, showing that the Leumanns were half right and that in some cases tt=/tt/. In 1981 two theories were offered about the origin of the tt-device. Oskar von Hinüber’s suggestion that the device is derived from Middle Indic practice has perhaps been most generally accepted. R. E. Emmerick’s tentative suggestion that the device is an internal development, spread from words that he knew from meter contained tt=/tt/, has gained less traction. Both ideas are incorporated here. New analytical tools are used which are based on a study of the metrical system in The Book of Zambasta.