Paleo-European languages

[WeepingElf]

We had the European languages before Indo-European thread on the old ZBB, so I shall hereby start a similar thread here.

I am going to begin with a little review of Theo Vennemann's "Vasconic" hypothesis on the occasion that I have changed my views of it somewhat. Vennemann claims that most of western and central Europe once spoke languages related to Basque, and claims to have found very much evidence of this in geographical names (especially the Old European Hydronymy)and substratum loanwords in the languages of western Europe (often reversing the traditionally assumed direction of borrowing in words Basque seems to share with the Romance languages).

I used to be strongly opposed to this idea, thinking that the Old European Hydronymy was a residue of a sister language of PIE which I call "Aquan" instead, and doubting most of Vennemann's etymologies. Now I think that the idea that languages related to Basque were once widespread in western Europe is not all that implausible. My model of the linguistic prehistory of Europe includes a hypothetical language family I call "Paleo-Atlantic" spreading from northern Spain and southwestern France across France, the British Isles, Germany and adjacent areas all the way to Scandinavia in the Mesolithic as part of the repopulation of northern Europe after the last ice age, and I consider Basque a likely candidate of a residue of this. This is exactly what Vennemann proposed; however, I am still unconvinced of his etymologies, which often attempt to reduce names that can aptly be explained by means of known local languages such as Celtic or Old High German, to Basque. And what regards the Old European Hydronymy, I am no longer sure anymore whether it is a thing at all - it may be a meaningless pattern, a linguistic equivalent of ley lines, falling out of the sheer mass of data!

Also, I am not sure whether Basque is really a Paleo-Atlantic language or not. It may be a language that was brought to its present location later from the east, perhaps by the first Neolithic farmers of southwestern Europe. This is connected to the question whether Iberian is related to Basque. The majority of scholars are doubtful of this connection, and so am I. It has been claimed that the Iberian place name Iliberi was the same as Basque hiri berri 'new town', but this remains speculative, and Basque hiri may itself be a loanword from Iberian (the Iberians were urbanized before the Basque). However, it seems instructive to look at another once enigmatic ancient language which could be rather quickly demystified by means of its relationship to known languages. This is Hittite, which soon turned out to be an Indo-European language, and though it seems to belong to an early divergent branch which is different from the longer-known IE languages in a number of ways, the relationship could be discerned and the language understood only a few decades after its discovery. In contrast, all attempts at understanding Iberian by means of Basque, based on the assumption of a relationship of similar time depth, have remained inconclusive. Granted, the Iberian corpus is much smaller than the Hittite one, but in my opinion, this probably means that if a relationship between Basque and Iberian exists, it is far deeper than that between Hittite and the other IE language. The precursors of Basque and Iberian may have already been different languages in the Neolithic. There have been claims that Iberian was related to Paleo-Sardinian, but the latter, known only from geographical names and perhaps a few odd words in Sardinian, without any inscriptional evidence, is even more mysterious than Iberian. so such a hypothesis seems premature.

[Salmoneus]

To clarify: nothing is known about pre-IE languages of Europe. Except that, presumably, there once were some.


Our most solid lead is the apparent borrowing from substratum languages into IE languages; however, we don't really know when or where this might have happened. Our most concrete evidence, therefore, is the handful of words that are both confined to northwestern european (germanic, celtic and italic) AND appear non-IE in form or do not easily correspond to a single reconstructable proto-form. Depending on your views on early IE dialectology, these may have been borrowed from a language, or family of languages, perhaps in pannonia or the lower danube, or perhaps in the baltic or in poland. They mostly relate to European plants and animals. But these are far too few to really say much. Other substrates confined to single families, like "Pelasgian" in Greek, have the problem that we can't tell when they might have entered the language - so it may not have occurred in Europe (or, at best, in far eastern steppe Europe).


Then there are the non-IE languages that may or may not be pre-IE. Only two are known in any substantial way: Basque and Etruscan. Very little is known about the history of Etruscan - its location suggests it may have arrived in Italy after IE (because IE ended up both north and south of it), but that's far from certain, and even if Etruscan is a late arrival there's no way to know if it's a primary migration into Europe (eg. by ship from the near east), or only a secondary migration within europe (eg. across the adriatic from dalmatia and liburnia, an area with relatively high survival of pre-IE genes).

Similarly, attestation of Basque is later than attestation of IE languages. Vasconic languages are first known as the language of a warlike tribe in southern France, the Aquitanians, who are described as being particularly dependent on horses. The name, however, derives from a little-known tribe in northern Spain, the Vascones. We don't know whether this means the Vascones spoke a Vasconic language already, or just that the Aquitanians (or their friends) later invaded Spain and adopted the name of a local tribe (possibly originally as an exonym - there's only so many times the Romans can call you by the name of the people you just conquered before you just give in and accept it). The latter maybe seems more likely, given geographic evidence of the former locations of Vasconic and Celtic languages in the area - the Vascones themselves may actually have been Celts.

Basque is very unlikely to be WeepingElf's "Palaeo-Atlantic". The mesolithic hunter-gatherer speakers of palaeo-atlantic were almost completely replaced by agricultural and maritime cultures from the near east in the Neolithic, and while it's true that the modern Basque have maintained an above-average level of mesolithic ancestry, they're still mostly post-mesolithic, and there's no reason to believe they would have maintained their language through not one but two population replacements. What's more, the 'bigger' Vasconic was, the more of Europe it covered, the more likely it is to be non-mesolithic - one small hunter-gatherer language in the pyrenees might survive, but a language family across the whole of western europe almost certainly would not.

Instead, the two most likely possibilities are a neolithic Vasconic - in which case it would most likely represent the language of the Cardial culture - or a bronze age Vasconic, in which case it would most likely represent another steppe language. This would help with the mystery of why Basques look so IE genetically (they absolutely stand out from other IE populations, but they still look like an IE population - for one thing, pretty much all Basque men are descended from steppe horsemen), if palaeo-vasconic was actually spoken by the neighbours of the Indo-Europeans.



Of the other languages, it's worth pointing out that Minoan civilisation in the Aegean and on Crete and Cyprus may be associated with an influx of caucasian genes that spread across the middle east, the near east, the aegean, and even as far as Sicily. This suggests its not a pre-IE language family - or at least not 'pre' by very much. If you had to pick a language family to put Minoan into, it would probably be Hurrian or Hattic.

[WeepingElf]

Fair, Basque may be part of a Neolithic language family rather than a Paleo-Atlantic language. It depends on whether the Neolithization of this part of Europe was primarily by demic diffusion or by acculturation. I don't know what is the prevalent opinion for northern Spain and southwestern France. It was quite certainly demic in Central Europe, but less certainly so in the western Mediterranean, as far as I currently know. You can probably enlighten me, Salmoneus, given your knowledge in European human genetics. If Basque was a Neolithic language, one would expect kinship to Iberian (unless one of them arrived later still), but there is little evidence for that - as I wrote yesterday, Basque has not turned as helpful in understanding Iberian as the "classical" IE languages were in understanding Hittite. But on the other hand, the Iberian corpus is much smaller than the Hittite one, and a Neolithic language relationship would be more than 1000 years deeper than that between Hittite and "classical" IE, which would both make things more difficult. At any rate, if Basque was brought in by Neolithic settlers from the east, Vennemann's hypothesis would tank (though the language brought in by the Cardial Neolithic may have been related to that brought in by LBK, of course, but that's still not Vennemann's Vasconic).

[Bob]

I'm really into these.

[Moose-tache]

It occurred to me: when a well-developed neolithic culture moves into an area, how common is it for hunter-gatherer languages to survive? It's common for groups around the periphery to slowly adopt agriculture, but if a bunch of yeoman farmers show up on their doorstep in ox carts and straw hats, what does the score card look like? In East Asia it looks like hunter gatherer languages disappeared almost every time a fully developed agricultural population moved into the area. We have no trace of a "Negrito" language from Malaysia or the Philippines, and in Japan the Ainu only survived in very marginal areas. Other languages like Tungusic and Austro-Asiatic probably had time to adopt some agricultural practices before the Big Boys moved in, but only because they lived in remote areas that were marginal for Han-style farming.

If Europe was the same way, then we would only expect Palaeo-Atlantic languages to survive in refugia unsuitable for near-east farming, or areas that were settled slowly, if ever, by farming populations, giving the locals time to adopt agriculture for themselves. Both of these seem pretty unlikely in southern Europe, where most places are suitable for wheat, beans, sheep, etc., and the genetic evidence suggests a large demic movement into the region.

I'm just speculating, of course. This isn't really an area where I have any formal education. Does this make sense to anyone else?

[Znex]

One relatively recent migration of this sort happened when the Bantu groups moved into southern Africa, displacing the Khoi-San hunter-gatherer populations. European contact with the people of southern Africa was recent enough that when the Dutch colonised South Africa, they took over Khoikhoi (although the Khoikhoi were pastoralists) land and its people before any Bantu people had moved into the area. Modern Khoekhoe speakers now live in the wilderness of the Kalahari and surrounding.

[Salmoneus]

One relatively recent migration of this sort happened when the Bantu groups moved into southern Africa, displacing the Khoi-San hunter-gatherer populations. European contact with the people of southern Africa was recent enough that when the Dutch colonised South Africa, they took over Khoikhoi (although the Khoikhoi were pastoralists) land and its people before any Bantu people had moved into the area. Modern Khoekhoe speakers now live in the wilderness of the Kalahari and surrounding.

Worth pointing out that certainly most, and possibly all, the surviving Khoisan are themselves later migrants. The Khoe languages migrated into the region with agriculture, and subsequently stopped farming due to desertification - they're probably related to Sandawe. The Khoe agriculturalists landed straight in the middle of the Tuu and Juu communities - but the Tuu and the Juu are traditionally considered migrants who replaced the phenotypically different and culturally less advanced populations of the area, now represented only in small communities (Dama, Kwisi, Cimba) who have lost their languages.

[mèþru]

The pre-Aryan Vedda language went extinct, but survived in a mixed language into the present day.

[WeepingElf]

At last, someone replying other than Salmoneus with his ignoramus et ignorabimus ad aeternum routine, which is not really a useful contribution to a discussion like this

It is true that in most cases, when farmers move into forager territory, the forager languages are displaced (either to somewhere else or to extinction). On that count, it is questionable whether Basque, Iberian or Tartessian are surviving forager languages, though Basque at least sits in a place which could have been a "somewhere else" in this sense (while this is definitely not the case with Iberian, and probably not with Tartessian, either). The question remains, did farmers move into the Iberian Peninsula from the east, or did local foragers adopt farming? What do the geneticists say?

At any rate, there are no positive signs of a relationship between the three known Paleo-European languages of the Iberian Peninsula. We cannot entirely rule that out, but, as I already observed, if Iberian was as closely related to Basque as Hittite is to the classical IE languages, this probably would have shown by now, and it didn't. So there definitely was no "Proto-Vasco-Iberian" later than about 2000 BC (Iberian is almost 2000 years later than Hittite, and Basque about 2000 years later than Latin, so the time depth must be adjusted appropriately), though there may have perhaps been one in the Neolithic about 6000-5000 BC.

And Tartessian is yet another tough nut to crack. John T. Koch has attempted to show that it was a Celtic language, but that attempt has been rejected by the majority of reviewers.

[Moose-tache]

Hey, go easy on Sal. He's right to pour some cold water on the discussion. If we're really talking about proto-languages that far in the past we might as well be guessing. We might be able to conclude whether a language represents a Neolithic intrusion or a Bronze Age intrusion, etc. But if two extinct languages don't look related, it's unlikely we'll ever be able to demonstrate that they are, barring any new textual discoveries.

[Salmoneus]

My position is not ignorance unto eternity, but ignorance until evidence, or at the very least ignorance until reason. There's a difference between historical linguistics and just guessing to suit our own fancy. But if I am indeed a stopped clock, I'd rather be stopped on "we don't really know that", than on a constant refrain of "but what does this mean for [noted lunatic] Theo Venneman and for the Old European Hydronymy [that doesn't exist]?" Why not a change, for a change - why not "but what does this mean for the theories of Edo Nylund? Can they really be ruled out?" for once?


On Vasco-Iberian: we don't know that either. It seems to me the main obstacles to analysis of Iberian are the extremely limited supply of data on both sides (the Iberian corpus is much smaller than the Hittite corpus AIUI (there are many inscriptions, but mostly very short), and Hittite could be compared with the whole of proto-indo-european, as reconstructed from dozens of languages, whereas Iberian has to be compared with a single Vasconic language - would the identification of Hittite been so rapid if it could only be compared with Albanian, or Irish?) and the fact that nobody really cares. Contrariwise, the fact that the numeral systems apparently match directly seems hard to explain through borrowing or coincidence.

[WeepingElf]

Points taken, Sal. One must indeed be very careful in this field not to be misinformed by one's own fancies. Of course, the "application" of this research in my case are Paleo-European conlangs, such as Old Albic and a number of others I have ideas for. And of course the main reason why people do research generally - curiosity. Indeed, it cannot be said often enough that when doing such things, one must always be aware which hat one is wearing at the moment.

I also think that Vennemann is misguided (though I wouldn't call him a lunatic of the same sort as, say, Octaviano, let alone Edo Nyland, whose hypothesis is impossible where Vennemann's is merely unlikely - I would rate Vennemann at about 100 millinylands, not more), and have grown very skeptical of the Old European Hydronymy. As I have already said, the latter may just be the linguistic equivalent of constellations, or, more appropriately, ley lines: a pattern that falls out of the sheer mass of data (there are many thousands of watercourses in Europe, all bearing some name) but doesn't mean anything. This also means that the whole Aquan hypothesis is something I am no longer in any way sure of; it may be the case that the language of the Bell Beaker people was an early offshoot of IE, but there is no compelling evidence in favour of that!

[mèþru]

I personally think the Bell Beakers spoke multiple Indo-European (specifically Proto-Celtic and Proto-Italic plus probably some others) and pre-Indo-European languages

On Old European Hydronymy - I think a lot of the coincidences may have been because of the Indo-Europeans - they may have made patterns and analogies from unrelated names that resulted in their modification to resemble each other.

[WeepingElf]

I personally think the Bell Beakers spoke multiple Indo-European (specifically Proto-Celtic and Proto-Italic plus probably some others) and pre-Indo-European languages

Indeed, the Bell Beaker people may have been linguistically heterogenous. There are plenty of culture areas where material culture is more or less the same but languages are vastly different.

On Old European Hydronymy - I think a lot of the coincidences may have been because of the Indo-Europeans - they may have made patterns and analogies from unrelated names that resulted in their modification to resemble each other.

Thinking about it, I have hit upon a problem that I have never met in the literature, hence it was at me to coin a name for it: name protraction. What I mean is that settlers name geographical objects of their new homeland after ones of their old homeland, which may result in names ending up in areas where the language the names are from was never spoken. Think, for instance, of names of Basque (e.g. Durango, Mexico) or Arabic origin (e.g. Alcatraz, California) in Latin America. Or names of Celtic, Slavic and other origins in North America. Something like this may have happened in the case of the Old European Hydronymy, too. Yet, I doubt that name protraction is capable of resulting in the majority of names in an area being protracted this way. So the null hypothesis is still that most geographical names in a given area are from languages that have actually once been spoken there.

Also, reading The History of Basque by R. L. Trask, I realized that Vennemann's Vasconic interpretation of the OEH is patently wrong: he interprets the names by means of Modern Basque rather than the reconstructible Proto-Basque, and many of the OEH roots, such as *kar-, *draw- or *war- have shapes which just can't be Proto-Basque (of course, an even earlier stage of Proto-Basque, beyond the reach of internal reconstruction, may allow such forms, but that kind of speculation is of course utterly baseless - that way, anything could be made to "fit"). These names don't really look like PIE ones, either (especially regarding the dominant vowel /a/), but internal reconstruction in this case points at a pre-stage of PIE which seems to comply.

[Moose-tache]

Something that's always bothered me: If we use the D...N river names as a starting point, we would expect OEH to be very opaque.

D...N river names occur most commonly in the PIE heartland, and less commonly in central and western Europe. As such, it's either an Indo-European root or derived from a language many thousands of years in the past. Luckily, there are Iranian words for river that have D...N, so there's a good chance that *den/don was a PIE word for river that just didn't catch on in most branches. So we have a root that was only preserved in one half of one branch of the modern language family, spoken primarily on a different continent. If the connection between OEH and, say, Basque were that opaque, we would not expect to see any obvious correspondences. The easy solutions and correspondences that people draw between OEH and living languages doesn't match the extremely tenuous relationship that actually exists between hydronomy and modern language. Two or three river names in Poland that match words in Etruscan would be on the same level as the D...N = PIE situation, but only pure speculation would allow us to identify those rivers as Tyrrhenian.

[anteallach]

On Vasco-Iberian: we don't know that either. It seems to me the main obstacles to analysis of Iberian are the extremely limited supply of data on both sides (the Iberian corpus is much smaller than the Hittite corpus AIUI (there are many inscriptions, but mostly very short), and Hittite could be compared with the whole of proto-indo-european, as reconstructed from dozens of languages, whereas Iberian has to be compared with a single Vasconic language - would the identification of Hittite been so rapid if it could only be compared with Albanian, or Irish?) and the fact that nobody really cares. Contrariwise, the fact that the numeral systems apparently match directly seems hard to explain through borrowing or coincidence.

Assuming those actually are the Iberian numbers, they look so similar to the Basque ones that I'd have to suspect borrowing, on the grounds that if the similarity is genetic the languages should be sufficiently closely related that Basque should have been more useful in understanding Iberian than it actually has been.

[Pabappa]

Assuming those actually are the Iberian numbers, they look so similar to the Basque ones that I'd have to suspect borrowing, on the grounds that if the similarity is genetic the languages should be sufficiently closely related that Basque should have been more useful in understanding Iberian than it actually has been.

I agree. It's too good to be true, given how far apart the languages appear in all of their other vocabulary. They could be loans from Basque into Iberian with a specific meaning, perhaps the names of coins or whatever would be appropriate for the context they're in ... which I'm not sure we can figure out just yet.

Most people seem to assume that the Iberians were stronger than the Basques and that therefore if they are loans, they are from Iberian into Basque, and that, since Basque uses them still today as numerals rather than narrowly defined terms for sets, they cannot have been loans and therefore the two langauges are in fact closely related. I admit that if the Basques are known to have been subjects of the Iberians, that it's unlikely the Iberians wouldve borrowed names of coins etc from Basque.

[WeepingElf]

Basque may have borrowed many words from Iberian. The two languages were neighbours in the Ebro valley, and the Iberians were an urban civilization earlier than the Basques. The Basque word for 'city', hiri in Modern Basque and ili in Proto-Basque, is an obvious candidate, considering the many Iberian city names beginning with Ili-. The Basque numerals also could have been borrowed from Iberian.

[WeepingElf]

Meanwhile, I have read up a bit on the genetics of the Neolithic people of the Iberian Peninsula, and the Basques in particular. Apparently, Neolithic agriculture was introduced by immigrants who would thus have brought in a new language; however, they were genetically different from the LBK people north of the Alps, and may have spoken a very different language. The Basques seem to be descendants of these Neolithic farmers.

So, I think we can say good-bye to the "Basque as a surviving Paleo-Atlantic language" idea. Vennemann (and Wiik, who proposed essentially the same idea) is thus completely wrong - not only his "Vasconic" etymologies are bogus, his entire edifice collapses. I'd thus rather expect Basque to be related to Iberian, and indeed there is some evidence in favour of that, even if the matter cannot be considered established yet. The Iberian phonology is very similar to that of Proto-Basque, and some words seem to mean the same in both languages. There is also a set of numerals where we don't know yet for sure which is which, but they seem to match Basque numerals. Sure, this can all be due to contact between the two languages, but there are apparently also similarities in inflectional morphology, and the thing looks at least as good as Indo-Uralic, for instance. Thus, we are quite likely dealing with related languages, with a common ancestor in the Neolithic. (And Paleo-Sardinian may be related to this, too, and probably also the lost pre-IE languages of the Italian mainland. And as Etruscan is clearly unrelated to Basque, this gives credit to the idea that Etruscan came from somewhere else, maybe Anatolia, in later times.)

And what regards Tartessian, it may be that John T. Koch's analysis is basically right and Tartessian a Celtic language. There is no shortage of Celtic names from that region, and the name of the legendary king Arganthonios (probably a dynasty of two or more kings rather than a single king) is about as Celtic as it can be.

[WeepingElf]

Meanwhile, the language spoken in Sardinia at the time of Roman conquest may have been a sister language of Etruscan if:

1. Sardinia has its name from the š3rdn, one of the "Sea Peoples" of Egyptian chronicles, settling in Sardinia.
2. The Etruscans are descendants of the twrš3, another of the "Sea Peoples".
3. Both peoples have the same ethnolinguistic origin (NW Anatolia?).

That's a lot of if, of course, and therefore this idea is most likely bullfrogs, and one would rather expect a language related to Basque and Iberian there. There is also no evidence of a large-scale invasion of Sardinia around 1200 BC in the archaeological record. Also, AFAIK there is no close genetic relationship between Sardinians and Tuscans.