Got this from PNAS. I'll just post the abstract. Very interesting.
Kusunda: An Indo-Pacific language in Nepal Paul Whitehouse , Timothy Usher , Merritt Ruhlen and William S.-Y. Wa
The Kusunda people of central Nepal have long been regarded as a relic tribe of South Asia. They are, or were until recently, seminomadic hunter-gatherers, living in jungles and forests, with a language that shows no similarities to surrounding languages. They are often described as shorter and darker than neighboring tribes. Our research indicates that the Kusunda language is a member of the Indo-Pacific family. This is a surprising finding inasmuch as the Indo-Pacific family is located on New Guinea and surrounding islands. The possibility that Kusunda is a remnant of the migration that led to the initial peopling of New Guinea and Australia warrants additional investigation from both a linguistic and genetic perspective.
--------------------------------------------------------------------------------
The Kusunda people of central Nepal are one of the few "relic" tribes found on the Indian subcontinent (the Nahali of India and the Veddas of Sri Lanka are two others). They first appeared in the ethnographic literature in 1848, when they were described by Hodgson as follows: "Amid the dense forests of the central region of N?p?l, to the westward of the great valley, dwell, in scanty numbers and nearly in a state of nature, two broken tribes having no apparent affinity with the civilized races of that country, and seeming like the fragments of an earlier population" (1). The Kusunda were one of these "broken tribes"; the Chepang were the other. Hodgson went on to show, however, that the Chepang were, on linguistic grounds, closely related to the Lhopa of Bhutan and must be presumed to have split off from this group and moved west at some time in the past. Hodgson had been unable to obtain any data on the Kusunda language, so nothing could be said of their possible affinity with other groups. Nine years later Hodgson published an article that contained the first linguistic data on the Kusunda language (2) as well as data on other Nepalese languages, but he offered no specific discussion of Kusunda even though his data showed quite clearly that the Kusunda language bore virtually no resemblance to any of the other languages he examined. No additional information on Kusunda appeared for more than a century until Reinhard and Toba (3) offered a brief description of the language, which provided some additional data. The final source on Kusunda appeared in an article by Reinhard in 1976 (4), but there is very little additional information that is not already found in the article by Reinhard and Toba (3).
Although Hodgson had predicted in 1848 the demise of the Kusunda in a few generations, a few Kusunda have managed to survive to the present day. Until recently they were seminomadic hunter-gatherers living in jungles and forests, and indeed their name for themselves is "people of the forest." They are often described as short in stature and having a darker skin color than surrounding tribes. Today the few remaining Kusunda have intermarried with neighboring tribes and drifted apart, and the language has been moribund for decades, although a few elderly speakers with some knowledge of the language still survive.
The Kusunda language is a linguistic isolate, with no clear genetic connections to any other language or language family (4, 5). Curiously, however, it has often been misclassified as a Tibeto-Burman language for purely accidental reasons. Hodgson's original description of the Kusunda language (2) also included vocabularies of various Indic and Tibeto-Burman languages. In 1909, Grierson classified Kusunda as a Tibeto-Burman language (6), like that of their immediate neighbors, the Chepang, who also were forest dwellers and spoke a Tibeto-Burman language. Later scholars often assumed, without looking at the data collected by Hodgson, that Kusunda was a Tibeto-Burman language. Kusunda was classified essentially on the basis of its neighbor's language, not its own, and this error perpetuated itself similar to a scribal error in a medieval manuscript (7?9).
We have discovered evidence that the Kusunda language is in fact a member of the Indo-Pacific family of languages (10). The Indo-Pacific family historically occupied a vast area from the Andaman Islands in the Indian Ocean to the Solomon Islands in the Pacific. Today most Indo-Pacific languages are found on New Guinea, where there are >700 surviving languages. Most of the western languages have disappeared as a consequence of the Austronesian expansion, but several ancient branches have survived on the Andaman Islands, the North Moluccas (North Halmahera and its smaller neighbors), and the lesser Sundas (Timor, Alor, and Pantar). East of New Guinea, Indo-Pacific languages survive on New Britain, New Ireland, the Solomon Islands, Rossel Island, and the Santa Cruz Islands. They also were spoken in Tasmania until 1876. The distribution of Kusunda and the Indo-Pacific family is shown in Fig. 1. Although it is not possible with present evidence to demonstrate conclusively the direction of the migration that separated Kusunda from the other Indo-Pacific languages, it would seem at least plausible that Kusunda is a remnant of the original migration to New Guinea and Australia rather than a backtracking to Nepal from the region in which other Indo-Pacific languages are spoken currently.
