Seminar CLI: Spain and Africa’s earliest Romance

Let me make clear straight away, this post is about the Romance languages, not the literary genre. In fact, it is specifically about the birth of Romance in Spain, and with work on that of course comes indelibly associated the name of Professor Roger Wright, and so it will not surprise you to gather that this post is because on 21st November 2012 he was presenting to the Institute of Historical Research’s Earlier Middle Ages Seminar with the title, “African Invaders and Very Old Spanish”.

This paper was, as Roger admitted straight up, based on published work, and it would only be new to us if we hadn’t read enough of his stuff, but nonetheless, since his thinking has in fact moved on since the works for which he is best known, not least because of dogged opposition from certain quarters, it was new to me and the questions suggested I wasn’t the only one.1 The starting premise, dear to my heart, is that Castilian is weird in Romance terms, having many features that other Romance languages don’t, and the basic question was whether this can be put down to influence from Africa.

Haplogroup Distributons in Iberian, North African, and Sephardic Jewish Populations (Adams et al. 2008)

How does one picture a language? Here is another way, then, in which African influence in Spain has been tracked; click through if you missed that post…

Now, the obvious conduit for African influence on Spain is of course the Muslim conquest, but since here we’re talking about Latin usage, unless one accepts Richard Hitchcock’s argument that much of the Muslim army that mounted the 711-714 campaigns that felled the Visigothic kingdom of Spain would have been Berber and North African recruits who, since they couldn’t yet have really learnt Arabic, must have had only Latin as a lingua franca, we need to look further back, and indeed even if Hitchcock is right the influences could still be older. Augustine of Hippo apparently reports being teased for his provincial Latin, and Isidore of Seville, Visigothic knowledge collector par excellence, reckons there are several peculiar things about the African Latin of his day. Several of these symptoms (betacism, the swapping of ‘b’ and ‘v’, much older than the QWERTY keyboard layout as my documents quickly made clear to me) also appear in the Visigothic slates.2 And, when one considers the respective difficulties of travel across the Pyrenees and across the Straits of Gibraltar, without considering modern state boundaries, obviously that makes sense.

Subsequent additions to the Africa of the Roman Empire would have been unlikely to have dented this African Latin, argued Roger: the Vandals, by the time they hit Africa at least and probably from much further back, are unlikely to have been a linguistic unity and their only common language must also have been various versions of Latin; they would have relied on Latin to deal with the locals, anyway.3 The Byzantine reconquest of Africa from the Vandals would also have had the administrative need to work in Latin as indeed it still partly did even at Constantinople; and the Visigoths meanwhile connected Spain and Africa by their grasp on what is now Ceuta (and still part of Spain, often forgotten except by Morocco). Berber languages, hardly an addition but arguably stronger after the loss of Africa as a Roman province, nonetheless seem significant only inland in this period. There is, in any case, no sign of any Berber influence on Spanish (and only one word in Portuguese) and no mention of Berbers (as opposed to the much vaguer Mauri, Moors) in the texts that describe the Muslim conquest such as the Chronicle of 754, which also doesn’t mention interpreters, Roger pointed out.

Section of handout from Roger Wright, &quo;African Invaders and Very Old Spanish&quo;

Professor Wright’s handout where it gives examples of African symptoms in Latin shared by Castilian

Nonetheless, although reconstructing African Latin’s distinctive characteristics is hard, it does seem hard to find them in Spanish Latin before 711. Isidore, as we say, sees a difference; Paul Alvarus of Córdoba, writing c. 860, does not. Betacism is rare before the seventh century, much more common later. Weirdly, and significantly, Arabic in Andalusia, the most heavily-settled area of course, also shows this symptom. Similar things can be said of the distinction between long and short vowels, the African difficulty Augustine describes: Roger pointed at Castilian ‘montes’ and ‘fuentes’, mountains and springs, from Latin ‘montes’ and ‘fontes’ respectively, to show this lack of distinction in action, and another symptom is the lack of a simple past tense in Castilian, where the past can only be formed by using the verb ‘to have’ an a participle. This doesn’t occur in Catalonia in the period of my documents, and modern Catalan retains as does French a preterite, even if neither are usually used in speech; I noticed the compound tense with excitement in the Beaulieu cartulary towards the close of the ninth century just the other day; but this was already settling in in Africa before the conquest, apparently, and now survives in Castilian. And there were a number of other cases of phonetic, syntactic and vocabulary resemblance that cumulatively seemed hard to argue with, though if you’d like to try I give the relevant section of Roger’s excellent handout as a scan above.

Thus, although the gap between say, 600 and 840, is still hard to fill in terms of linguistic development, in Spain it seems reasonably clear that the 711 invasion is one of the branches, with the consequent implication that its armies and settlers were many of them Latin-speaking. The further implications of that had, as I say, already been somewhat explored by Richard Hitchcock in 2007, but as far as I know Professor Hitchcock has never published that, and though what I’ve said here is as far as Roger went it’s still plenty to think about…

1. The obvious works of Roger’s to refer to are his Late Latin and Early Romance in Spain and Carolingian France (Liverpool 1982) or R. Wright (ed.), Latin and the Romance Languages in the Early Middle Ages (London 1991), but by now he might prefer that we checked his A Sociophilological Study of Late Latin (Turnhout 2003) or specifically on this question R. Wright, “Late and Vulgar Latin in Muslim Spain: the African Connection” in Frédérique Biville, Marie-Karine Lhommé & Daniel Vallat (edd.), Latin vulgaire, latin tardif IX : Actes du IXe Colloque International sur le Latin Vulgaire et Tardif, Lyon, 2 – 6 septembre 2009, Collection de la Maison de l’Orient et de la Méditerranée : Série linguistique et philologique 8 (Lyon 2012), pp. 35-54. For opposition, I suppose one would look most obviously to Michel Banniard, Viva voce : communication érite et communication orale du IVe aui IXe siècle en Occident latin (Paris 2002) but more anciently Rosamond McKitterick, “Latin and Romance: an historian’s perspective” in Wright, Latin and the Romance Languages, pp. 130-145 or Michael Richter, The Oral Tradition in the Early Middle Ages (Turnhout 1994).

2. For an edition of the slates, aimed at just this question, see Isabel Velázquez Soriano, Las pizarras visigodas: entre el latín y su disgregación. La lengua hablada en Hispania, siglos VI-VII (Madrid 2004). I observed in questions, largely on the basis of this post at Magistra et Mater I admit, that this also happens in Lombard Italy, to which Roger’s response was to suggest bad spelling and to observe that almost everything that can happen to Latin happens in Italy. Well, OK, but…

3. Here Roger cited Jonathan Conant, Staying Roman: conquest and identity in Africa and the Mediterranean, 439-700 (Cambridge 2012) and Yitzhak Hen, Roman Barbarians: the royal court and culture in the early medieval West (Basingstoke 2007), neither of which I’ve read but both of which, and perhaps especially teh former of which I had no knowledge before this, I really should.


13 responses to “Seminar CLI: Spain and Africa’s earliest Romance

  1. Allan McKinley

    Did anyone consider whether Punic could have had a legacy – the Cartheginians were in Africa and Spain after all? It may be the way Latin was spoken in formerly-Punic areas was different from elsewhere (in the west mainly Celtic).

    • This was touched in as much as Roger diverted very briefly to scotch the idea that Augustine might have spoken Punic, and that be what was affecting his vowels. Whether that would rule out a legacy local accent or something, I don’t know, but my very limited grasp of linguistics suggests to me that a lack of distinction between long and short vowels must be a development, not an introduction problem. When Latin was introduced to Punic-speakers, presumably it was introduced vowel-sounds intact…

  2. Or the whole thing is much more complex. What about pre-latin substrates? Mediterranean latin is quite different from central or atlantic variants. An african influence? Quite sure, but probably not the only (or the most important) one?

  3. Allan McKinley

    For the vowel sounds of an introduced language to remain intact those it was introduced to would have to have had those vowel sounds (or at least closely similiar ones) in their language. I’m not sure that is the case here.

    I did go to check Punic vowel systems, but ran into the technical issue that Punic is an Arameic language, and therefore unless written in Latin there are no vowels in inscriptions (so no-one seems to have reconstructed them).

    But, inspired by Joan’s substrata comment, I checked the vowels for Iberian. Although I suspect this is provisional (experts are still working out the details of the language), the current state of play seems to be the normal western European five vowels but, crucially, with no distiction in length. So there is a case for (on the Mediteranean coastal region at least) a long-standing lack of vowel-length differentiation in Spain. Would need some seriously expert work to show as probable though.

  4. I hadn’t thought in this direction myself, and I don’t know that you could overwhelm all Roger’s phenomena in background here but it’s worth thinking about. Certainly, Iberia being so heavily regional in geomorphology, it would be tempting to see more divisions between lingustic zones than just the Pyrenees. But with so scanty a source base it’s hard enough to get at regionalised, rather than individual, habits of Latinity, let alone going back to Iberian! And I certainly don’t know anything much about Iberian and its vowel lengths, though I do wonder if the position you outline is a kind of starting assumption rather than an actual state of knowledge. This is more Joan’s province than mine…

  5. Not sure about being only five vocals, just a few weeks ago I heard an expert on north-oriental iberian saying that quantitative analysis suggests that there were no less than seven vocals. The articles on wikipedia about ancient iberian languages and scripts seems to be quite ok. They give a glimpse on the complexity of the problem, not to mention that the Pirinees were more a bridge than a wall, so the geographic area has to be expanded even further…
    Latinization and romance developed upon this landscape, so the complexity has to be even greater, that’s for sure, but I am not a philologist, and the more I heard them debating the less I know… :) A clear article imo about the birth of iberian romance languages could be: Bastardas i Parera, Joan : 1995 : “Quan el llatí esdevingué català?” : La Llengua catalana mil anys enrere : pgs. 73-103, just my 0.02c.

    • Indeed! My pet cite for this is another of his, “El català vers l’any 1000” in Immaculada Ollich i Castanyer (ed.), Actes del Congrès Internacional Gerbert d’Orlhac i el seu Temps: Catalunya i Europa a la Fi del 1r Mil·lenni, Vic-Ripoll, 10-13 de Novembre de 1999 (Vic 1999), pp. 495-513 with English abstract p. 514. He is the man with whose work to start!

  6. Can thoroughly recommend Conant’s book. Truly excellent.
    Betacism is indeed very common in Africa, and may perhaps have been the norm for pronunciation for it to have been able to override standard Latin spellings and produce the likes of Bictor for Victor or Quodbultdeus for Quodvultdeus. It is, as you rightly pointed out, also common in Italy. Then again we also find it in Gaul and indeed in Spain. Isn’t this argument also putting an awful lot of weight on Isidore’s knowledge of regional Spanish latin pronunciation? I mean was the bishop of Seville really able to say that this linguistic feature was not to be found already in Visigothic Spain?

  7. Well, hang on, I mean these are all good points but, if the symptom that we’re trying to explain here is modern Castilian betacism, well, there’s plenty of betacism in the early medieval Catalan documents, but it’s not a feature of the modern language. It’s not very marked in Galician or Portuguese either as far as I know (though I’d guess the Algarve version of Portuguese shows a lot more of it). So we’re looking at one zone (a big zone, admittedly) where betacism doesn’t just stick around but is presumably reinforced. This is also the area most strongly affected (as far as the texts and place-names and indeed Arabic vocabulary in the language) by Islamic settlement. But Arabic shouldn’t reinforce betacism. Therefore… And if you accept that for any of the symptoms Roger was talking about Occam’s Razor can be applied. It needn’t rule out any of the factors you’re all bringing up, but if Iberian is the operative factor Catalonia ought to be affected, if other factors other regions, and so on.

    • Allan McKinley

      I’m not sure you can say betacism sticks around or not. If evidenced in Catalan documents then betacism must have applied there to the extent that scribes overrode existing spelling conventions. That it is not a feature of modern Catalan indicates simply that there has been further linguistic change for some reason, reversing the earlier betacism (or perhaps more accurately the effects of betacism). It may be that betacism was never a Catalan norm but if it is well evidenced in the sources, which are essentially conservative in form and presumably language, this is a case that has to be made.

      That said, I am not sure that a feature as common in the (relative linguistic) environs of Catalonia as betacism is diagnostically helpful. Although we do not know why, there was a clear tendency for this to occur in Latin-derived dialects, so I would suggest this was primarily a potential feature of Latin as a whole rather than a diagnostic feature. Lack of expertise does mean that none of this is much more than speculation though.

    • No, iberian it’s only mediterranean (that is, excluding tartesic or southern iberian in modern terminology), but maybe celtic? (ie: betacism seems to be centered on peninsular celtic areas).

  8. Graham Barrett

    Why do you say that Spanish has no preterite?

    • Ignorance and misunderstanding, it would seem! Thankyou for the correction; I’ve deleted that part, which seems from the handout to be a misunderstanding of what Roger would have been saying about the compound future.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.