Monday, May 14, 2012

Why Care That A Dead Language Was Discovered?

Last week, I posted at some length about a newly discovered ancient language, which may never have even been a written language for its native speakers at any time. Today, I want to step back for a moment to explain why this might be worth caring about.

There are lots of dead languages which we know exist but which we have no written text documenting, not even second hand documents like the Assyrian tablet with the women's names transliterated from this newly discovered ancient language.

If this language turns out to be, for example, a divergent dialect of the Hittite language previously spoken in that part of Anatolia for a while that had grown into its own language in the four hundred years since the collapse of the Hittite empire, it would still be a little bit interesting, but would really not add much to what we already known, especially since would only add some proper nouns to the existing lexicon in that family of languages.

This is a new language family we didn't know we were missing at the cradle of civilization long after writing was invented.

But, this find is so much more.  It is not just a newly discovered ancient language, but one that is a language isolate that isn't part of any known modern or ancient language family.  Indeed, while there is a good chance that this particular lost language is one that was attested to have existed historically, it appears to come from a language family that we didn't even know existed.  We had assumed that the attested languages mentioned in ancient historical accounts from around the 8th century BCE in this general region were all members of the West Semitic, Indo-European, Hurrian-Uratian, Caucasian, Elamite language families, or perhaps a relict community of Sumerian language speakers, or perhaps even Dravidian language family or Etrustcan language family members.  A none of the above conclusion for a language in the vicinity of a place where we have a longer historical record than anyplace else in the world is pretty remarkable.

The prevailing assumption was that Indo-European and West Semitic language family expansions had wiped out pretty much all language families other than the Elamite language and Caucasian languages in this region thousands of years earlier.  The fact that an entire unknown language family could have survived this long so close to areas where we have a historical record suggests that the process of language extinction as farming societies expanded may have taken place much later than conventional wisdom had suggested.

Many factors should have led us to expect an undiscovered language family in this region.

Of course, there is also, in hindsight, good reason to suspect that the Zargos mountains where this language family appears to have persisted, at least until the Assyrian king conquered them, could be an exception to the general trend.  Relict languages and populations tend to survive in places that are unsuitable for conventional horticulture - the Caucasus mountains, Nuba Mountain, the Kalahari desert, the arid American Southwest, the depth of the Congo and Amazon and Mesoamerican jungles, the Himalayas, the Andaman Islands, the Highlands of Papua New Guinea (which are mountains surrounded by jungles at lower elevations), the Australian Outback.

We know, in particular, that after the Sumerian empire discarded Sumerian in favor of the Semitic Akkadian language, that the non-Indo-European, non-Sumerian language speaking Kassites from the Zargos Mountains swept in when the Akkadian empire was laid low by a major drought and imposed their own language on the part of the Sumerian empire that they conquered, which had escaped extinction experienced by nearby lowland languages including Sumerian.

Linguistic and cultural regions also tend to line up with drainage basins, probably because pre-modern peoples migrated and traveled mostly along river basins that provided reliable access to fresh water and fish, and along coastlines.  Any pattern of exploration based upon those connections, however, omits endorheic basins that aren't connected to any oceans.

The eastern side of the Zargos Mountains marks a break between the Tigris-Euphrates River Valley and an endorheic basin (i.e. one that does not drain to an ocean) in Iran called the Zaytandeh river basin, which was home to an archaeological culture contemporaneous with that of the Sumerians and the Harappans, to the West and the East respectively.  We also know that the Zargos mountains were advanced enough to be home to trade partners of the Sumerians when their civilization was in existence, and the Mediterranean climate forest steppe that prevails in that basin before it becomes a desert could have provided a friendly enough region for a civilization self-sustaining enough to develop its own distinct language family to persist.

Implications for linguistic models

A strong confirmation of the existence of a previously undiscovered language family in this region significantly impacts how we go about modeling the course of linguistic prehistory. It suggests that large swaths of Iran previously assumed to have spoken Indo-Iranian languages may have not done so until the Medes and the Persians conquered that territory in the 6th century BCE, rather than the 15th century BCE or earlier when the Mittani were encountered by the Hittites. It also, by discounting the effectiveness of Bronze Age civilizations at wiping out pre-existing language families, at least in endorheic basins and refugia, suggests that the substrate language environment encounted by expanding Indo-European, Afro-Asiatic, Bantu, Austronesian, Austroasiatic and other language families may have been more of a patchwork than Occam's razor driven assumptions would suggest. Such a patchwork of whole language families prior to the expansion of the main language families known today also suggests that local cultures may have been more persistent to greater time depths than previously supposed, at least prior to the Iron Age, and that post-Neolithic society may have been comprised of fairly modest sized, insular, culturally isolated polities to a greater extent than previously supposed, at least in some kinds of environments. For example, this find tends to support a supposition that any common ancestor of the various Caucuasian language families may have been quite ancient.

In contrast, this find tends to discredit theories of a common origin of living language families that simply fits the language families that survive today into a neat branching structure with few major branches pruned off in the historic era or late pre-historic era. Any distinctive element of a newly discovered language family (for example, phonemic features that might be implied by a list of names from women who spoke that language), has the potential to upset "grand unified theories" of language evolution by providing an additional, and ancient data point.

Finally, the evidence involved in this tablet may provide something close to being one of the most direct, first hand accounts, of the ultimate extinction of an entire language family at the tail end of the period where much of this pruning of linguistic diversity is believed to have taken place. It provides a much uglier story than the relatively peaceful ones by which the Akkadian language replaced Sumerian, and the Hittite language replaced the Hattic language, for example.

1 comment:

Maju said...

Coming from you particularly and on the field of historical linguistics, I am gladly surprised to be in quite strong agreement with what you wrote here.

Certainly this new language reinforces the notion of a very diverse linguistic reality of West Asia with no trace of Indoeuropean languages until the groups you mention arrived (first the Anatolian-Hittite group, then the Mittani/Indo-Aryan shallow layer and finally the Iranian conquest of several successive layers).