13 September 2008

Romance languages before the Romans

It is generally accepted that, at around 1000 BC, the geographic distribution of Italic languages (among them Latin, Faliscan and Osco-Umbran) was restricted to some areas of central and southern Italy. Later on, the expansion of the Romans involved a massive process of language substitution whereby large populations, especially in western Europe, abandoned their languages (Celtic, Etruscan, Iberian, or in some cases obscure languages with no name) and adopted Latin, which was the origin of the subsequent Romance languages that are still spoken in those areas today (Spanish, Portuguese, French, Catalan, Italian, etc.). - This traditional explanation looks quite clear and reasonable, but in fact there are many good reasons to question it, as I’ll try to show here.

The readers of this blog are already familiar with Mario Alinei and the Continuity Theory. One of the main features of this new approach is the fact that the linguistic data are always analysed in connection with archaeological and anthropological data. Another important aspect is the type of linguistic analysis that is carried out. As I have already noted, Mario Alinei is a dialectologist. For many years he was the president of the Atlas Linguarum Europae project and he is considered one of the most important experts in Italian and European dialectology. One of the things he found out is that the main differences between the various Italian dialects had been established at a very archaic period, and not in the Middle Ages, as is generally assumed. The distribution of some kinship or agricultural terminology and the diffusion of some phonological traits from one dialect to another point to a pre-Roman chronology (Alinei, 2000:951-978). This can be applied to areas where other Italic dialects have traditionally been attested (central and southern Italy) but also to other areas where it was supposed that other types of languages were spoken: in northern Italy, on the islands of Sicily, Corsica and Sardinia, and also in Etruria, where the presence of the Etruscans has been interpreted, in the light of the CT, as an intrusive elite that ruled over a mainly Italic-speaking population. One by one, Alinei analyses the data from the various areas and comes to the conclusion that the most relevant elements in the formation of these dialects, even at the most archaic stages, are Italic. Otherwise, how is it possible, for example, that the names for the plough, or for the various parts of the plough, were coined on the basis of a vocabulary that was more archaic than the one the Roman conquerors actually brought with them? In many respects Italian dialects (in Sardinia, in Piedmont, in Tuscany and elsewhere) seem to be 'older' than Classical Latin. The only acceptable explanation for this apparently paradoxical fact is that Italic languages were spoken in these areas before the Romans arrived. After the conquest, Latin became the most influential element in these territories, in a process which eventually shaped the local dialects into what they are today.

But Alinei’s proposals go beyond the geographic boundaries of Italy. He finds evidence for pre-Roman Italic languages in other territories, e.g. the Balearic Islands, southern France and the east and south of the Iberian Peninsula, as well as some Adriatic areas. In fact, he coined the term gruppo italide in order to avoid the geographic connotations of the word Italic. Some linguistic data seem to point in this direction, but a great amount of research is still to be done in order to refine this thesis. It seems, however, that the archaeological data support the existence of this “Italid” group, as can be seen in the areal distribution of the Printed-Cardium Pottery culture (c. 5000 BC; see the image on the right as an example; more information about this picture, here), or even in the distribution of the Epigravettian culture (24000 to 10000 BC). Looking at the maps of these cultures (you can find them in Alinei, 2002), it seems that there is correspondence between some present-day Romance languages and the areas that Alinei considers originally Italid. As we saw in the previous post, about the languages of Switzerland (you can read it here), the historical event of the Roman conquest is not relevant in the distribution of dialects in that area. Something similar could be said about the Italid area in general. (Alinei, 2000, p. 582): la romanizzazione avrebbe lasciato le proprie tracce solo là dove i linguemi precedenti erano già affini al Latino, mentre non avrebbe avuto conseguenze linguistiche rilevanti – salvo l’introduzione di prestiti – nelle aree in cui i linguemi autoctoni erano di ceppo diverso (Germanico, Celtico, Slavo, Illirico); (ib., p. 592): "Dal Neolitico Medio in poi, insomma, le principali aree dialettali sono già manifeste". Which is, of course, a revolutionary thing to say in the field of Historical Linguistics or Romance Studies. And I quite agree with him. First, because it offers a rational way of explaining the emergence of modern Romance dialects, avoiding the typical (and easier) explanations based on conquests and invasions. Secondly, because there are other researchers, with no direct connection with Mario Alinei or the CT, who have reached a series of results which, at least partially, point in the same direction.

One of these researchers is the Spanish linguist Francisco Villar, one of the most prestigious experts in pre-Roman languages of the Iberian Peninsula. As I said, he is no adherent to the CT, and his approach and methodology have little to do with it. In one of his studies (Villar, 2000), he analysed the whole corpus of pre-Roman Hispanic names for people and places, especially hydronyms. He found out that there were both Indo-European and non-IE elements in this archaic vocabulary. One would expect something like this, because it has traditionally been assumed that in Pre-Roman Hispania there were both IE languages, belonging to the Celtic group, and also non-IE languages, for example Iberian. But he also found some unexpected results, for example the existence of another IE language, that he called substrato indoeuropeo italoide. This non-Celtic IE lexicon is found in many areas in the Iberian Peninsula, but especially in the south and north-east. (Villar, 2000, p. 442): “El estrato étnico y lingüístico más profundo y abundante tanto en Cataluña y la Cuenca del Ebro, como en Andalucía que nos permite detectar la toponimia lo constituyen unas poblaciones indoeuropeas muy antiguas, que crearon el primer entramado hidrotoponímico de densidad suficiente como para perdurar a través de los sucesivos cambios de lengua y llegar hasta nuestros días”; (ib., p. 414): “La lengua paleohispánica tiene relaciones dialectales particularmente estrechas con las itálicas y, en forma más lejana y menos definida con las bálticas”.

Finally, I would like to summarize and conclude this article with a hypothesis, which is also an invitation for future research: In the Italian Peninsula, on the islands of Sicily, Sardinia and Corsica, in the south of France and the east of the Iberian Peninsula, dialects belonging to the Italid group have been spoken at least from the Neolithic, with no discontinuity. On the other hand, it is also possible that the first Homo Sapiens Sapiens who settled in these territories were speakers of Indo-European languages.

Bibliography:
- ALINEI, Mario (2000). Origini delle Lingue d’Europa. II. Continuità dal Mesolitico all’età del ferro nelle Principali Aree Etnolinguistiche. Bologna, Il Mulino.
- ALINEI, Mario (2002). Towards a generalised continuity model for Uralic and Indo-European languages. In Julku, Kyösti (ed.), The Roots of Peoples and Languages of Northewrn Eurasia IV. Oulu, Societas Historiae Fenno-Ugricae, 2002, 9-33 .
- VILLAR, Francisco (2000).
Indoeuropeos y No-Indoeuropeos en la Hispania Prerromana. Salamanca, Ediciones Universidad de Salamanca.

14 comments:

JoseAngel said...

Well, who knows what pasts the future will bring...
It is clear you favour an extremely radical view on these issues. As you know, my position is more skeptical, although I welcome new ideas; only I wouldn't want them to erase all that went before... a little more CONTINUITY, if you want to put it thus! Revolutions may have their positive aspects, but one should not turn a blind eye to their power to do away with much that was good in the old regime as well.
It seems clear though that we still have much to learn about pre-Roman languages in Spain. One question: do you accept the existence of language substitution in any region of Spain or Europe? Or would you favour continuity everywhere?

Jesús Sanchis said...

Joseangel, I find your comments quite interesting.

Let’s remember that I ended this post with a hypothesis, and that’s what it is: a hypothesis, a proposal that needs to be tested, especially for the non-Italian areas. The good thing about the continuity-hybridization model, as opposed to the traditional (and not very often challenged) substitution model, is that it offers a more reasonable framework for the study of the history of languages. It doesn’t attempt to “erase all that went before”. In fact, Alinei has often stated that most of the linguistic analysis carried out in the comparative tradition is useful, and his proposals, however revolutionary or controversial they might seem, are based on careful analysis. It’s clear that the new theory opens new research horizons in a field study (historical linguistics) which is in need of ideas.

The expansion of Latin, associated with the Roman Conquest, is a complex phenomenon which needs careful study. The application of the continuity model offers an alternative explanation to the established canon based on language substitution by way of conquest. As I see it, there are at least three possible scenarios for romanization:

1.- The presence of Italid languages facilitated the spread of “Latin” in these areas, by means of hybridization and ultimately, and to a relative degree, assimilation of those dialects by the dialects of the conquerors. Obviously, it still not clear to what extent Italid dialects were present outside Italy, and, as a said earlier, that’s an issue that deserves further attention.

2.- The expansion of “Latin” in other areas, like central, northern and western Iberian Peninsula, or to the northern parts of France, must necessarily have been a much slower process, which was not completed during the Roman Empire and which, in fact, was never completed in some areas (let’s think of Basque and the Celtic languages of Bretagne, for example). We might think that this slow development included a higher degree of language substitution, and was conditioned by geographic factors.

3.- The fact is that, for most of the Roman Empire, no Romance languages have survived. Many reasons have been given for this phenomenon, and it’s logical to think that a variety of factors were involved in it. I personally think that the continuity theory offers some interesting proposals that should be taken into account.

It seems that there is still a lot of work to be done out in this field. The Roman Empire is an important moment in history, a link between the old and the modern world. That’s what makes it particularly interesting. Now, one thing is for sure: when we think of earlier times, i.e. prehistoric times, it makes no sense to talk about the history of languages as a process of continuous substitution. The traditional account of the expansion of Indo-European, based on language substitution on a grand scale in prehistory, is no longer tenable. The conditions for language change in the Paleolithic or Neolithic periods were by no means the same as the ones we find in the stratified societies that developed later. Authors such as Colin Renfrew have tried to offer a more plausible explanation, connecting IE expansion with the expansion of agriculture, but I think this explanation is not satisfactory. So far, the Paleolithic Continuity model of IE seems a better option.

JoseAngel said...

Thanks for such a detailed answer, Jesús. I understand your position much better now, and I'm looking forward to further posts on the issue.

Octavià Alexandre said...

Despite his name, Villar's Italoid has nothing to do with Alinei's Italide. It's the same substrate language which Coromines called Sorotaptic and was situated somewhere between Baltic and Italic in the IE dialectal cloud.

Finally, I would like to summarize and conclude this article with a hypothesis, which is also an invitation for future research: In the Italian Peninsula, on the islands of Sicily, Sardinia and Corsica, in the south of France and the east of the Iberian Peninsula, dialects belonging to the Italid group have been spoken at least from the Neolithic, with no discontinuity.

I call this Tyrrhenian, a Vasco-Caucasian (defintely not IE)substrate language family to which Basque and Iberian belong.

In my view, Tyrrhenian arrived to these territories along with Cardial pottery in the Neolithic.

On the other hand, it is also possible that the first Homo Sapiens Sapiens who settled in these territories were speakers of Indo-European languages.

This is highly unlikely, not to said impossible.

Jesús Sanchis said...

I agree with you, the final hypothesis is highly unlikely. But you know, stating hypotheses is in itself an act of freedom!

I can see at least one important problem in your own hypothesis about 'Thyrrenean' and IE. I don't see how such an extensive, intercontinental phenomenon as IE expansion could have happened in a post-Neolithic environment. What kind of explanations can be used to sustain this claim? Let me guess: internal reconstruction, linguistic paleontology, Kurgans, horses. As you know, I have strongly criticised these arguments in this blog.

Magean said...

First, congratulations for your very interesting blog.

But I have some questions I'd like to ask you :

1)Let's consider the Italid languages-Cardium pottery link hypothesis true. How do you explain more precisely, with your language continuity optic, the expansion of the Romance languages in the areas where other tongues were spoken (i.e. central Spain, northern France...), after the Roman conquest (the celtic substrate in Oïl dialects, for instance, attest the presence of non-Italid speakers in northern France) ?

2)I heard that Villar relates Lusitanian with the Italic family. If it is true, do you think that Lusitanians kept speaking their language, but with a huge latin influence (as Latin was easy to learn for these italic speakers), and then created Portuguese ?

3)The French historian and philologist Henri d'Arbois de Jubainville put forward an hypothesis, according to which Ligurian was an Indo-European language and Ligurian country stretched from northern Italy to Holland. For him, Ligurians populated a vast part of Gaul. If Ligurian was an Italic language, or a bridge-language between Celtic and Italic, could this have facilitate the romanization of northern France ?

4)How do you explain, still with your language continuity optic, the expansion of Turkish language area, from Siberia to Anatolia, including central Asia (you won't tell me an Altaic people already lived in Anatolia when the Turks arrived, will you ?) ?

Thank you in advance for your answers.

Octavià Alexandre said...

I can see at least one important problem in your own hypothesis about 'Thyrrenean' and IE. I don't see how such an extensive, intercontinental phenomenon as IE expansion could have happened in a post-Neolithic environment. What kind of explanations can be used to sustain this claim? Let me guess: internal reconstruction, linguistic paleontology, Kurgans, horses. As you know, I have strongly criticised these arguments in this blog.

I think you're attributing me things I didn't say. While my own theory links the spread of Vasco-Caucasian languages like Tyrrhenian with Neolithic à la Renfrew, this doesn't imply IE spread was necessarily post-Neolithic.

In fact, I think the distribution of IE languages isn't the result of a single but several expansions, both (pre-)Neolithic and post-Neolithic ones, with extensive language replacement.

In the course of this process, some areas might have been "Indo-Europeanized" more than once, archaic IE languages being replaced by modern ones.

The Kurgan theory only accounts for the Greek-Indo-Iranian group, which even Indo-Europeanists like Rodríguez Adrados recognize as being the most innovative within IE.

Jesús Sanchis said...

Entesos, Octavià! Now that I understand your point of view beter, I can only say that I agree with you on this too: the expansion of IE cannot be the result of a single event, and PIE is not just a linguistic 'package' that was born somewhere and brought to the rest of locations in a single migration. These are some of the ideas that I have been defending in this blog from the beginning.

The difference between IE and Basque is the for the former we have a lot of evidence, whereas for the latter (and for Iberian), the evidence is very scarce. In general, I have the impression that any hypothetical conjecture about Basque or Iberian, or any attempt at comparing the corresponding 'proto-languages', is bound to yield little in the form of solid conclusions. But this is just an impression, as I'm not an expert in these areas. However, I have had the chance of listening to some experts in Iberian languages, and you know what? The more they know about it, the less inclined they are to try to produce 'translations' of Iberian words or sentences. They are well aware of the extreme difficulties of such a task, and prefer a more careful approach.

Jesús Sanchis said...

Magean, your questions are quite difficult, and I'm not an expert in these things! Let's see what I can do:

1) I think I answered this question in a previous commnent to this post. In northern parts of Gaul we might have a slower process resembling language substitution, in which the affinity between Celtic and Italic was also a determining factor. Let's remember that these two IE groups are generally considered quite similar.

2) As far as I know, Villar has proposed a very old IE substratum for some areas of the Iberian Peninsula, and that would include several features found in the Lusitanian inscriptions. According to Villar, this substratum would have similarities with Italic. In any case, however, we are talking about hypotheses based on very poor evidence, and there's still a lot to be researched in this field of study.

3) I'm afraid I don't know much about Ligurian, but I'm aware that Ligurians are usually mentioned in any attempt at analysing the ancient languages of Western Mediterranean. Was their language IE, was it not? There is so little evidence from this language that it's really difficult to draw firm conclusions.

4) Again, I don't know much about Turkish or the expànsion of Turkic languages, but I have often thought about the presence of Turkish in Anatolia as a rather strange phenomenon. Can the Continuity approach shed any light onto this? Probably, but I don't know the details. In any case, the linguistic situation of Anatolia has always been quite complex. Turkish is now spoken in most of the modern Turkish territory, but just a century ago languages like Armenian or Kurdish were predominant in the east of the country. Kurdish, an Indo-Aryan language, is still spoken in the region nowadays, even though it is not the official language of any state; Armenian, however, was wiped out from many areas when the Ottoman Empire carried out what is generally referred to as the "Armenian genocide". It would be interesting to analyse in detail the history of Turkic languages in Anatolia and Central Asia, but that's something that I'm not in a position to do!

Magean said...

Thank you for your answers.

Concerning the romanization of Gaul, what do you think about that hypothesis :
http://www.proto-english.org/e71.html
(the author calls the Italic family "Occitan", as for him it existed non only in Italy). He also tells about what he calls Occitan here :
http://www.proto-english.org/o3.html
http://www.proto-english.org/l1.html
http://www.proto-english.org/l7.html

I am sorry for asking you all these questions, but I think this website can interest you.

Thank you in advance, one more time.

Jesús Sanchis said...

Thank you for the link, Magean. It looks quite interesting. I'll try to read it as soon as possible.

Octavià Alexandre said...

The difference between IE and Basque is the for the former we have a lot of evidence, whereas for the latter (and for Iberian), the evidence is very scarce. In general, I have the impression that any hypothetical conjecture about Basque or Iberian, or any attempt at comparing the corresponding 'proto-languages', is bound to yield little in the form of solid conclusions. But this is just an impression, as I'm not an expert in these areas.

Well, you can learn some things on my own blog.

However, I have had the chance of listening to some experts in Iberian languages, and you know what? The more they know about it, the less inclined they are to try to produce 'translations' of Iberian words or sentences. They are well aware of the extreme difficulties of such a task, and prefer a more careful approach.

This kind of "approach" sound to me as obscurantism. Have been these "experts" to be competent at comparative linguistics, they already should have produced some positive knowledge.

Octavià Alexandre said...

Concerning the romanization of Gaul, what do you think about that hypothesis :
http://www.proto-english.org/e71.html
(the author calls the Italic family "Occitan", as for him it existed non only in Italy). He also tells about what he calls Occitan here :
http://www.proto-english.org/o3.html
http://www.proto-english.org/l1.html
http://www.proto-english.org/l7.html

It looks like another more **crackpot** theory, although with some interesting aspects. For example, its "Occitan" is roughly the same language as my Tyrrhenian.

Jesús Sanchis said...

I have read the link provided by Magean. Some of the contents are quite interesting. The authors put forward a series of ideas about the languages of the British Isles and Westernm Europe which were already proposed as a hypothesis by Mario Alinei and the PCP many years ago. I would suggest that the authors read these texts.

The most original (and risky) proposal is the one about what they call 'Ligurian'. The authors themselves point out that it is not even a hypothesis, but just a conjecture. I think this is quite honest. There are many pseudo-linguists out there who are not humble enough to admit the limitations of their proposals.

The authors of this blog rely heavily on Oppenheimer's results, based on DNA analysis. I have commented on Oppenheimer's book in my blog. See these posts:

-http://languagecontinuity.blogspot.com/2009/06/language-continuity-in-europe-iii.html

- http://languagecontinuity.blogspot.com/2009/11/migrations-and-languages-in-prehistory.html

I'm sceptical about Oppenheimer's 'spectacular' results. I don't think it can be so easy to trace the ancestry of British populations from the study of mtDNA or Y chromosomes. It's necessary to take population genetics data more critically.

The authors have published an article about the toponymy of Kent in a journal, and I find this quite interesting. The article is also available in their web-page.