Unlike our historian friends, anthropologists do not have the luxury of drawing a line in the sands of time and declaring a closure date for our research. Ethnography never ends. Even the demise of the original field–worker does not conclude the enterprise, given the inevitability of re–studies (usually conducted by younger scholars eager to overthrow past paradigms). This article is a product of contemporary ethnography; it describes a project that has a beginning but no clear end. At some point during the 1990s, the research took on a life of its own and, if anyone is in charge, it is certainly not the ethnographer. In many respects, therefore, the project parallels the digital revolution: decentered, unpredictable, and "out of control."
My address focuses on the long–term consequences of international migration and the historical dynamics of diaspora formation. The research is longitudinal in the sense that it tracks a single, tightly bound kinship group during thirty–five years of field research. From its inception, this has been what anthropologists refer to as a multisited ethnography, even though that term had not been invented when the research began (see Marcus 1995). The project started in 1969 as a "typical" (for that era) village study, focusing on a Cantonese community of two thousand people.