Alaska is the presumed starting point for (at least) three very important migrations that defined the cultural history of the entire Western Hemisphere, but so far the archaeological record within the state has shed virtually no light on two of them, and relatively little on the third. . . .
The first of these migration is, of course, the initial peopling of the Americas in the Late Pleistocene. . . . Recent research in various places has increasingly indicated that the Clovis culture of around 13,000 years ago was not the direct result of the earliest migration into the Americas, but it is still the case that any migrations during the Pleistocene (and it’s increasingly looking like there were at least two) almost certainly would have had to go through Alaska. Unfortunately, despite several decades of looking, no sites have yet been found in Alaska itself that can plausibly be taken to reflect the first immigrants into North America from Asia. An increasing number of early sites have been identified in the past twenty years, but these are all still too late to represent a population ancestral to Clovis or any of the other early cultures found further south. Part of the problem here is that preservation conditions for archaeological sites in most of Alaska are atrocious, and in many areas even finding early sites is extremely difficult. The fact that the state is huge and sparsely populated also means that very little of it has even been surveyed for sites[.]. . .
The second of the migrations I mentioned above is that of speakers of Athapaskan languages [Ed. also known as the Na-Dene languages] to the south, ultimately as far as the Southwestern US and the extreme north of Mexico. As I’ve mentioned before, it’s long been quite obvious that Navajo and the various Apache languages, as well as several languages of California and Oregon coasts, are closely related to a larger number of languages in Alaska and northwestern Canada. The distribution of the languages, as well as some internal evidence in the southern branch, strongly suggests that the direction of the migration that led to this situation was north-to-south, and similar evidence similarly suggests that the start point was somewhere in what is now Alaska. Despite the enormous distance over which Athapaskan languages are now spread, the greatest diversity of the languages grammatically is actually found within Alaska. That is, some Alaskan languages are more closely related to Navajo than they are to other Athapaskan languages in Alaska. While this is all clear linguistically, tracing the actual migration archaeologically has been enormously difficult at both ends. Athapaskan archaeology in Alaska in particular is remarkably poorly understood compared to the archaeology of Eskimo groups, due in part to the fact that Athapaskans have mostly occupied the interior areas that are harder to investigate than the primarily Eskimo coastal areas. . . .
The third migration, and by far the best understood, is that of so-called Thule peoples from northwestern Alaska eastward across the Arctic as far as Greenland.
From here (emphasis and editorial comment added).
While the genetic distinctiveness of the Na-Dene compared to other Native Americans already favored a distinct Na-Dene migration wave to North America from Siberia, and that fact that the Navajo and Apache languages have their roots in Northwestern North America was widely understood, the evidence that the maximum linguistic diversity is in Alaska rather than the Pacific Northwest of the United States, for example, is still significant.
The linguistic evidence tends to coroborate and enhance the genetic evidence by independently disfavoring the possibility that the Na-Dene have origins in a population expansion within some enclave of the pre-existing Native American population that made some economic breakthrough (perhaps some new fishing or hunting technique) and had rare genes that were lost due to serial founder effects in other parts of the New World. It also tends to validate linguistic connections between the Na-Dene and the Yenesians (e.g. the Ket people of Central Siberia) by putting the source population for the North American Na-Dene people pretty much as close to their Old World origins as they could possibly be.
Linguistic evidence of Alaskan origins for the Na-Dene also tell us that the oldest Na-Dene archaeological remains in the interior of Alaska ought to be older than the oldest Na-Dene arcaheological remains in contexts where the conditions to preserve these relics is better, which accompanied by our knowledge that the Na-Dene were likely post-Clovis, helps pinpoint places in Alaska (and particular sedimentary layers at particular digs within Alaska) where is makes sense to look for early Na-Dene archaeology. Archaeologists need to be looking for cites centuries or even millenia earlier than the earliest known Na-Dene sites anywhere else in North America.
But, perhaps most tempting of all is the observation that "some Alaskan languages are more closely related to Navajo than they are to other Athapaskan languages in Alaska." This suggests that it may be possible in the case of these languages, as it was in the case of the Bantu languages and the languages of the people of Madagascar, and the languages of the European Gypsies (aka Roma), to identify the source of an exceptionally long distance prehistoric migration not just to the place where the source language family as a whole was spoken, but to one very geographical specific place within that region that was the source of the particular group of long distance migrants who were the source of the expansion.
Each of these examples also cross validate each other. They suggests that often a long distance migration is going to involve an isolated, "heroic" journal of a single community of people with their origins in a single place, no doubt lead by some visionary leader (he may have seemed crazy at the time), to a new homeland whose links to the larger linguistic and cultural grouping in which they have their origins is filtered entirely through the way it manifests in this particular community, and stands in opposition to migration models that posit bland, almost random walk, diffusions of peoples who are experiencing population expansions.
Narratives like the Biblical Exodus story, rather than being implausibly far fetched, seem to be almost paradigmatic of how a long distance migration of an ethnically distinct population to the new homeland happens. The push and pull factors that motivated these long range migrations at the time, and the events that were critical in making it possible for the new arrivals to displace or culturally dominate the pre-existing residents of their new homes may be forever lost to history, but these kinds of migrations necessarily must have had those aspects to them.
Actually, it isn't quite that simple. Viewed more closely, the linguistic evidence is suggestive of Eastern Athapaskan languages arising from two distinct waves of migration with origins in or around Southern Alberta, and Western Athapaskan languages including Navajo having roots in a different migration (although, in fairness, the three separate branches of migration seem to have been reasonably close in time).
The same author expands upon his conclusions in a follow up post (which happens to reference a Wikipedia page I've put a great deal of work into developing).
it’s generally thought that the Athapaskan migrations which eventually led to the entrance of the Navajos and Apaches into the Southwest began in Alaska. The northern Athapaskan languages are actually spoken over a very large area of northwestern Canada as well, but the linguistic evidence clearly points to Alaska as the original place where Proto-Athapaskan was spoken at the last point before it splintered into the various Athapaskan languages. That is, the Urheimat of the Athapaskans seems to have been somewhere in Alaska.
There are two main pieces of evidence pointing to this conclusion. One is the fact that it has been quite well established at this point that the Athapaskan language family as a whole is related to the recently-extinct language Eyak, which was spoken in south-central Alaska. Eyak was clearly not an Athapaskan language itself, but it had sufficient similarities to reconstructed Proto-Athapaskan to establish a genetic relationship. Since Eyak was spoken in Alaska, it therefore seems most probable that the most recent common ancestor of both Eyak and the Athapaskan languages (Proto-Athapaskan-Eyak) was also spoken in Alaska. . . .
A stronger piece of evidence is the internal diversity of the Athapaskan languages themselves. A general principle in historical linguistics is that the Urheimat of a language family is likely to be found where there is maximal diversity among the languages in the family. . . . When it comes to Athapaskan, this condition obtains most strongly in Alaska. The languages in Canada and the Lower 48 are all relatively closely related to each other within the family as compared to some of the languages in Alaska. Although interior Alaska is overwhelmingly dominated by Athapaskan groups, the linguistic boundaries among these various groups, even those adjacent to each other, are often extremely sharp.
This is particularly the case when it comes to the most divergent of all the Athapaskan languages: Dena’ina. (. . . in many publications this term is spelled “Tanaina,” . . . ”Dena’ina” is the generally preferred form these days[.] . . .) This is the language traditionally spoken around Cook Inlet in south-central Alaska, including the Anchorage area. While it’s clearly Athapaskan, it’s very weird as Athapaskan languages go. It is not mutually intelligible with any other Athapaskan language, although it borders several of them, and it is in turn divided into several internal dialects that are strikingly diverse. . . . there are two main dialects, Upper Inlet and Lower Inlet, and that Lower Inlet is further subdivided into two or three subdialects: Outer Inlet, Inland, and Iliamna. . . the Lower Inlet dialect is more conservative than the Upper Inlet one, which shows extensive influence from the neighboring Ahtna language, which is also Athapaskan but not very similar to Dena’ina. Within the Lower Inlet dialect, the Inland dialect is the most conservative . . . presumably due to the relative isolation of this dialect, which is spoken in the Lake Clark area and further north in Lime Village. . . . this the most likely homeland of Dena’ina speakers . . . [who] moved from the interior to the coast relatively recently. . . .
Despite the relative conservatism of the Lower Inlet dialect, however, all its subdialects do show a certain amount of influence from Yup’ik Eskimo (particularly in the development of the Proto-Athapaskan vowel system). This is unsurprising, as these dialects lie on the boundary of the Yup’ik area to the west and south, and the Dena’ina groups in these areas show extensive Eskimo influence in many aspects of their traditional culture. . . . the main distinctions among the Dena’ina groups were economic, having to do with their subsistence systems, while other social systems were pretty similar across the various groups. The Lower Inlet groups, especially those in the Seldovia area on Kachemak Bay near the outlet of Cook Inlet, showed a much heavier dependence on hunting sea mammals and a correspondingly heavier influence from nearby Yup’ik Eskimo groups with a similar adaptation than their compatriots further north who had a more typically Athapaskan lifestyle based on salmon fishing and hunting of terrestrial animals.
The fact that Dena’ina, the most divergent of the Athapaskan languages and therefore the one that most likely split earliest from Proto-Athapaskan, is spoken in Alaska makes it very likely that Proto-Athapaskan was spoken in Alaska as well. Indeed, if . . . the Lake Clark area was the original homeland of the Dena’ina, this potentially places Proto-Athapaskan quite far west within Alaska and quite close to areas traditionally occupied by Eskimo-speakers. . . . however. . . it’s still very unclear when the breakup of Proto-Athapaskan occurred and who occupied which parts of Alaska at that time.
Thus, the Urheimat of the Athapaskan aka Na-Dene languages in North America was probably somewhere West of Anchorage in areas that are now occupied by Eskimos, but wouldn't have been prior to the arrival of the Thule, sometime within the last fifteen hundred years, who apparently displaced them to inland locations. Also, while the author doesn't expressly address the point, the arrival of the Thule in North American took place close in time to the Na-Dene migration to the Southeast United States, which suggests that this Navajo and Apache exodus story may have its origins in flight from the ancestors of modern Eskimos who displaced them from their previous homeland. It is also roughly coincident with the rise of the Cahokia based Mississippian culture (which despite its name apparently reached at least as far as the Atlantic Coast of North Carolina) and the Viking Vinland colony in far Northeastern North America, and roughly coincides with the Chaco culture of the American Southwest. The interactions between the basically hunter-gatherer Athapaskan cultures and the agriculturalist Pueblo culture are explored here.
The author of the excerpts above provides further detail in a third post on the subject which he concludes with the following observations:
[I]t definitely seems that the Dena’ina most likely spread from west to east, and the Ahtna may have been spreading in the other direction at approximately the same time. Archaeological evidence suggests that the Dena’ina spread into the Kenai peninsula (across the inlet from their apparent homeland) took place less than a thousand years ago. Since it seems very clear that the Athapaskans had been in Alaska for a very long time before that, certainly long enough for Dena’ina in the southwest to diverge markedly from the various languages that form a large dialect continuum in the Tanana and Yukon valleys, this suggests that the story of Athapaskan prehistory is both very complicated and very long-term.