This is a controversial matter, capable of raising the blood pressure of archaeologists, population geneticists, linguists and the native peoples of North America. Many Indian groups believe firmly that they have been there since the beginning of time. Traditional archaeologists point towards the close of the last Ice Age, while more recent proposals suggest a preceding stage when the Beringian gateway to the New World was open. The latter view is attractive to linguists aware of the extreme linguistic variety of the Americas.
When the news broke in 1996 that a 9,000-year-old "Caucasoid" skeleton had been unearthed at a place called Kennewick in Washington State, the stunned intake of breath from all parties concerned raised the controversy again. Where did "he" come from? What tongue could he have spoken?
The archaeologists tell us that all native Americans must have originated in Asia, perhaps on the "Mammoth Steppe" that stretched across the north of the continent in Palaeolithic times. Recent investigations have shown that modern man pushed up into this region from the Middle East earlier than he penetrated northern Europe.
The Beringian "bottleneck" between the continents is conveniently placed for those interested in correlating movements of people and languages - it has acted as a variable barrier, allowing people through only at certain times (it was blocked by ice, for example, between about 20,000 and 30,000 years ago). Investigation of the typology of the languages on either side of the gateway suggests at least three "waves" since the last retreat of the ice - and probably several more at a much earlier period. Hence the rather sharp boundary between the Salishan languages of north-west America and the Penutian languages to the south of them.
One intriguing possibility suggested by both archaeology and linguistics is that the antepenultimate wave into the New World specifically linked the Asian coast north of Sakhalin and British Columbia/Washington in an ancient North Pacific Rim area. Various unusual traits are shared by the Salishan languages and Nivkh of the Amur River region and there is "substrate" evidence also on Kamchatka. So perhaps the "Kennewick man" is the descendant of a wandering band of coastal hunters from the vicinity of the Amur. The neighbours of the Nivkh, the "hairy" Ainu, have long been described as "Caucausoid" (read today rather: pre- if not palaeo-Mongoloid). There are moreover certain "Americanoid" linguistic traits like "noun incorporation" in both Nivkh and Ainu. The band would have mingled en route with other, more pronouncedly "Mongoloid" people reflecting earlier "waves", and would eventually have been absorbed into the complex linguistic and cultural area of the American North-West.
It seems that genes, language and culture are always out of step once the initial colonisation of new areas has been established. Language shift and the mixing of small populations of nomadic peoples has been widespread - in the North - as between the Eskimo and Chukchi on the Asian side of Bering Strait in recent times. There is little reason to think that this has not always been so around the bottleneck. Perhaps the clash between the "scientific" and the "native" viewpoint when strange skeletons pop out of the earth would be mitigated somewhat if it were more widely recognised that the origins of people and the origins of the languages they bear are separate issues.
For the time being, the man himself remains under lock and key while the legal battle rages between scientists and the local Indian tribe. And linguists might as well forget it: we shall never know what language he spoke.
Michael Fortescue is the author of `Language Relations Across Bering Strait' (Cassell, pounds 49.99)