In lieu of an abstract, here is a brief excerpt:The nominalization-relativization syncretism is characteristic of languages in Tibeto-Burman areas (e.g. Noonan 1997; DeLancey 1999), whilst the diachronic process of the phenomenon is now pursued in East and Southeast Asian languages as well (e.g. Yap and Wrona forthcoming). These preceding works propose two different directions of change. One is that nominalization has developed into relativization (e.g. Yap and Matthews 2008) albeit with a lack of syntactic explicitness in some cases by the want of historical documents. The other concerns the inverse direction from relativization to nominalization, which is proposed by researchers such as LaPolla (2003 with Huang). Although they are opposing against each other, each survey result remains and raises an intriguing possibility, which is worth reconsideration through the analysis of other languages. However, it should be pointed out that preceding research cannot give a full account of the directions with a paucity of crucial historical evidence. Genetti (2008) embarks on a reconstructing research into the diachronic process in which relativization and nominalization each give rise to the other based on five Tibeto-Burman languages; she provides a good syntactic analysis, but any synchronic study seems to have its own limits. Building on these preceding works, this study addresses the historical development of the Okinawan nominalizer si, which used to be sï (see the next section), out of its earlier usage as the head of relative clause i.e. from relativization to nominalization.