Sunday, 30 July 2017
Initial R- in Hittite
INITIAL R- in HITTITE
I had read that Indo-Europeanists knew that Hittites had come to Anatolia from outside. I was unimpressed by this. How could the shape of a language record a past migration?
I was rereading Benjamin Fortson’s masterly short sketch of the Anatolian languages recently and the penny dropped. Indo-European undoubtedly has initial r- and Hittite does not. However, a whole range of early languages of Anatolia are missing initial r-. This is highly compatible with Hittite having arrived from outside with a wave of immigrants who mixed biologically – familiarly, with the locals to give a massively bilingual community which re-normalized the old language with certain features of the local language – such a “negative rule” banning r- in certain positions. Since we know that the language ancestral to Hittite had initial r-, in the REX/RICH word for example, that would make Hittite an immigrant language which had come from outside.
The Net, again, reveals that 3000 tablets were found at the Hittite palace-complex of Sapinuwa (Ortaköy). The dig began in 1990 but apparently only 3 tablets have been published. Something has gone seriously wrong here. The find is quite close to the main tablet archive at Boghazköy and from the same time-span, so we do not have the hope of a variant dialect and information of a new kind. However, the new tablets should complete our information and strengthen the state of several hypotheses. They include bilingual vocabulary lists, surely a treasure.
Anatolia is a big place. However, the range of evidence for languages related to Hittite in the so-named Anatolian group covers a vast area, and it would seem that unlikely that the jump off point for the Hittite language was elsewhere in Anatolia and the journey was from, let’s say, Cilicia to the north-central area around Ankara. We have hieroglyphic Luwian from the Syrian border area (Carchemish) and Lidyan from the Aegean coast. It would appear that this branch of Indo-European immigrated into the region from outside.
Robert Beekes points out that there are only 700 legacy IE words in Hittite. We could hope that wider source material, such as the Sapinuwa archive, would bring a few more. For comparison, Welsh has 800 Latin loan-words dating from the Roman Empire. Anatolian is the first IE family to be recorded (maybe in 1800 BC), but had by then moved farther from the ancestral model than almost any other. We could reconstruct rather little of IE if we only had Hittite and Luwian to work from. This marginal status is hard to combine with a theorised central or source position.
The info on the Net indicates that 600 of the Sapinuwa tablets are in other languages, i.e. non-Hittite. That would include Hurrian, widely used by the Hittite polity in rituals.
As Fortson points out, Greek has no (original) initial r. Rhota only occurs, at the start of words, as aspirated. This goes back to an older s- which was reduced to a breathing. Thus the form rhei “it flows” (as panta rhei) goes back to the sr- root (English stream, Irish sruth). In older Greek there was no initial r-, just sr-. There is a word for darkness, in the Norse Ragnarök, twilight of the gods (ragna “of the gods”, rök “darkness”). This matches Greek Erebos (a dark place), Armenian erek ‘evening’. In each case the older initial r has been covered up by a kind of glide vowel. This is complementary to the Hittite evidence and gives us further knowledge of the geography of this sound-shift. Evidently, Armenian has spent most of its history in the Anatolian area, and it should be Anatolian in areal characteristics. Greek is historically, adjacent to the Anatolian languages. It belongs in the same “south central” square of the Indo-European map. The loss of initial R parallels Hittite/Luwian and would perhaps indicate that the Greeks crossed the Aegean from the east or that the peoples who lived on the Aegean before the Greeks had the same phonology as the people of Anatolia, so the Hurrians, Urartians, and so on. All this has a bearing on a well-known theory whereby IE was found in Anatolia – maybe even the Konya Plain – 7000 years ago and spread through Europe with the first Neolithic farmers, being carried in fact by the same humans, who gradually spread out taking farming skills with them. This does not fit very well with the Anatolian IE languages having come into the region from outside, evidently from the Balkans and probably originally from north of the Black Sea. Nor does it fit with the Anatolian group having the most degraded (! or most evolved/ innovated, works either way) version of the original language – which is preserved so faithfully in Lithuanian and Vedic. The “Neolithic = Indo-European” theory is in deep trouble.
Beekes (again) rejects this theory, pointing out that the slow expansion model would imply a long shared distinctive development of Celtic and Germanic, as adjacent language groups in Western Europe. In fact they have no shared history that we know of. The pattern of the IE families is compatible with a “yeast bubbles in bread” pattern, where pastoral groups spread rapidly and opportunistically through a densely populated peasant landscape, settling mainly where the inhabitants were few or the terrain was very suited to pastoralism. They leapfrogged opposition. So the success of the IE speakers as mobile invaders was also the catastrophe of the IE speech community, which broke up into widely separated enclaves, covering a huge diameter but also parted from each other by the peasant regions which had been bypassed and not swamped.
Renfrew’s theory emphasises slow pace, steadiness, continuity, even tranquillity. This process would have given a dialect continuum within Europe and Anatolia, but since we have gaping gaps between the language families we need rather to explain the discontinuities. India may offer a dialect continuum and may have been Aryanised through a different process. The shatter lines between the “families” may reflect the gaps between the original patches of intrusive steppe pastoralists in the early and mid-3rd millennium BC. Some areas are more suited to herding than others. What we seem to see is the farmer languages disappearing to leave an IE sea. This process is unexplored – those languages disappeared and have no history at all. We know about large “islands” of unrelated speech – Basque, Iberian, Etruscan.
The first written Indo-European language is found where writing already was and so where there was a dense farming population, rich enough to support a state superstructure and a profession of clerks. So it was fore-ordained that the first records of IE would capture a language which had not replaced the local population, numerous and thriving, but been absorbed by them in massive bilingualism, and so damaged. Its original structures had been extensively remodelled, metabolised, broken down. Hittite is a not a good source for the archaic stage of Indo-European. We don’t have a sociology of how Hittite died out, or indeed how Luwian, a related language which seems to have replaced it, died out itself. There was a social dynamic in Anatolia, as in Europe.
I would like to know more details about the fate of the legacy r- words. Did they acquire glide vowels? Were they replaced by local words? Or by near-synonyms? What is the replacement pattern? The scholars were right to say that Hittite came from outside. I just hadn’t known the reasons.