Empires of the Word: A Language History of the World

by Nicholas Ostler

  Instead we see a steady fall-back, and the unmixed spread across the country of English, a mixture of Angle, Saxon, Frisian and perhaps Jutish varieties of Low German. The only parallel, in fact, to this spread of a Germanic language is what happened when the Germanic invaders encountered virgin territory, in the islands of the North Sea and in Iceland. There of course the Vikings’ language, Old Norse, spread, because it had no competition. Could the Britons of the urbanised lowlands somehow have just melted away? Nothing less is needed to explain the complete walkover within Britain of those Germanic languages, and above all of English.

  A recent theory, from David Keys, says that they may have.49 The mid-sixth century (close to 550) was the time when bubonic plague entered Britain, along trade routes from the Mediterranean. Significantly, it would have been Britain (the west and centre of the island) which it hit, rather than England (the south-east), because only Britain maintained trade links with the empire. And it would be less likely to spread to the Saxons since they did not consort with Britons and, living outside the established Roman towns and cities, may have lived at a lower density. It would have been virtually simultaneous with the mortālitās magna that hit Ireland, according to the Annals of Ulster, devastating the aristocracy (and no doubt every other class). Maelgwn, king of Gwynedd in Wales, also died of plague in 547 or 549, according to the Annales Cambriae. A folk memory of this dreadful disease, and the depopulation it caused, would remain in the Arthurian legend of the Waste Land, combining famine with military defeat, and a mysterious wound (to the king) in the groin area—one of the characteristics of bubonic plague.

  There is even a little genetic evidence that strikingly bears this out. Comparing the pattern of Y-chromosome DNA from samples in a line across from Anglesey to Friesland, a recent study found that the Welshmen were to this day clearly distinct from those in central England, but that the English and Frisian samples were so similar that they pointed to a common origin of 50-100 per cent of the (male) population; this could have resulted from a mass migration from Friesland.50 On the usual assumption that the Roman-period population of the island had reached 3 to 4 million, it seems hardly possible that anything other than an epidemic could have so eliminated the Britons from the ancestry of central England.

  So English supervened. It did not long have the eastern and central regions of the island to itself: in the late eighth century a new force entered the system, a new set of Germanic invaders, the Norsemen or Vikings, from Scandinavia. They progressed from coastal raids to settlement in the west of Scotland and the east of Northumbria to a partition of the island with the Saxons by treaty (c.886), and finally in 1013 to outright conquest of the whole kingdom. This was by Sveinn Forkbeard, succeeded by his son Knútr, better known as Canute.

  Unlike the British-English divide, relations between Anglo-Saxon and Viking, if initially hostile, proved fairly permeable in the longer term. One way of understanding this is to see the Vikings as classic Germanic invaders, military raiders who won most of the battles but lost the peace, in that they settled down—perhaps with English wives—and largely picked up their subjects’ or victims’ language. Nevertheless, since the language into which they were settling was a close-ish relative (though with a good twenty generations of separate development behind it), there was easy scope for bilingualism and a degree of mutual understanding. The result was an abundant infusion of Norse loan words into English, and quite a lot of impact on the grammar too. In modern English, some 7 per cent of the basic vocabulary is of distinctly Norse origin (including such words as take, get, keep, leg, sky, skin and skirt);51 and it is this mix of the two languages which gave rise to the bizarrely unrelated set of third-person pronouns he, it, she and they.*

  The early era of western European conquests thus closed with a kaleidoscopic shifting of Germans westward, and of Slavs southward. The Germans were able to retain their language only when they conquered territory that was largely, or totally, empty—Britain devastated by plague, and Iceland previously uninhabited. Their conquests in the western heartlands of the Roman empire had essentially no linguistic impact. Latin remained strong in the west and south of the continent; there, the linguistic effects of Roman conquest were never undone. The Slavs, perhaps because they were invading less civilised—and hence less highly populated—regions had much greater effect where they settled in the Balkans; but they too were absorbed or eliminated in the areas of ancient civilisation that they overran, parts of Greece and Anatolia.

  The long-term effect was a linguistic partition of Europe that has been familiar ever since: Romance in the south and west, Germanic in the north and centre, Slavic in most of the east, and Greek in the extreme south-east. The main event in the fifth century was in fact the switch of Britain in the northwest from the Romance (or perhaps still Celtic) to the Germanic zone. There was considerably more change to come in this island: the further spread of Germanic into the last redoubts of Celtic over the next thousand years, compounded by a late attempt at a reassertion of Romance over Germanic, and the Norman conquest of England. But the tale of these events must wait until we turn to the growth of English itself.

  * Compare these pronouns in Old English (hē, hit, hēo, hīe) with Old Norse (hann, that, hon, their/thau/thær—using English th for the Norse ō). Mix-ups between rather different systems of endings, well preserved in both Old English and Norse, may also have caused the breakdown of case marking for nouns.

  * (These asterisks show forms that have been reconstructed by linguists, but are not actually found in some text.) This absence of P is not as strange as it might seem. It also seems to have afflicted the indigenous language of Iberia, and even early Basque, and is typical too of modern Arabic. But Celtic did not remain a totally P-less language for long. At least some of its variants, including most dialects of Gaulish, and also British (leading to modern Welsh, Cornish and Breton), later started to pronounce the sound qu- as p. Hence its presence in the words for four and five (pedwar and pump in modern Welsh, probably *petuar and *pinpe in Gaulish, on the evidence of some kiln records, mentioned in note 22 on p. 566). As a result, where initial qu- had been the mark of question words in the original language (cf. Latin’s conservative quis, quid, quando, ‘who, what, when’), initial p- has this role in this variety of Celtic language (cf. Welsh pwy, pa, pam, ‘who, which, why’, and presumably much the same in Gaulish). The other Celtic languages also changed the qu-, but just simplified it to a k- sound. Hence Irish ceilhir, cóic (’four, five’), and cé, cad, cá (’who, what, where’). What evidence there is for Celtiberian suggests it was more like Irish than Gaulish in this respect.

  * The earliest known Etruscan inscriptions date from about a century earlier, c.700 BC. The Etruscans had themselves learnt how to write from the Greeks, though probably through contacts much farther south, round Cumae in the Bay of Naples.

  * Contrast Lusitanian, spoken farther south: we know hardly more than two words of this language, but those two words are enough to disqualify it as Celtic: porcom tavrom, ‘pig bull’. The first has a P; the second has its V and R in the wrong order: compare Gaulish tarvos. Old Irish tarb, Middle Welsh tarw.

  * By contrast, Germanic has the same underlying root for ‘bronze’ as Latin: Gothic aiz, Old English ār, Old High German ēr versus Latin aes, suggesting that this technology was already an established acquisition before the common ancestors of the Italic- and Germanic-speaking tribes went their separate ways.

  8

  The First Death of Latin

  Philosophantem rhetorem intellegunt pauci, loquentem rusticum multi.

  The rhetorician philosophising is understood by few, but the plain man speaking by many.

  Gregory of Tours, Preface to Historia Francorum (C.AD 575)1

  The history of western Europe after the German invasions is the tale of how the kingdoms established by the conquering tribes went on to become distinct nations. Dialectal differences in the Latin that people spoke widened, and wide-ranging travel
became less common, as the road system decayed and public order became unenforceable far from cities. No longer was there a Roman army with a common tradition, and troops that might expect to be transferred anywhere. Where literacy survived, principally in the Church, so did written Latin. But this was not enough to maintain any spoken standard. The gap between spoken and written language widened, but without people having any sense of what was really happening, namely that the spoken language was changing. Little by little Latin spelling came to seem more and more irregular and perverse: but this obscurity was acceptable, even desirable, as reading and writing were the preserve of a small elite, mostly clerics and lawyers.

  This period, the second half of the first millennium AD, gives us our main evidence of what happens to a universal language in the western European, Christian, tradition, when it begins to lose currency, when people, although still speaking it, begin to lose sight of its vast scope, and live above all in their local communities. Three hundred years after the Goths and Germans had divided up the territories of the empire, it had become extremely difficult for the people of Spain, France and Italy, when they did meet, to understand one another’s speech. The learned, the only ones who would be conscious of the problem, came to call anyone’s ordinary speech an idioma, to be contrasted with the universality of grammatica, which was the normal word for Latin in the Middle Ages.*

  Charlemagne’s Europe, 8th Century AD

  From the early fifth to the mid-eighth centuries, the powers in western Europe shifted from generation to generation, allowing the idea to establish itself that universal kingdoms or citizenships could never be of this world. But then, from the late eighth century, the power of the Frankish king grew, in alliance with the papacy, and for a century the areas of France, western Germany and most of Italy were united. The Frankish king who presided over the height of this glory was Charlemagne, who reigned from 768 to 814. His aspirations were cultural as well as political. In 781 he invited Alcuin, the head of the cathedral school at York, to become head of a new academy of scholars at Aachen, his capital. The fruit of this congregation has become known as the Carolingian Renaissance. In the course of it, and along with many other reforms in education, Alcuin established new standards for the spelling† and pronunciation of Latin.

  Alcuin, as a speaker of North-Country English, approached Latin as a foreign language, to be learnt ab initio from books; in this he would have been at one with perhaps the majority of the scholars at Aachen, many of whom would have come from the German-speaking east of Charlemagne’s empire. He succeeded in establishing a common pronunciation for Latin, close to what we now think of as ‘the modern pronunciation’, which was an intelligent attempt to reconstruct the sound of the language on an authentic ancient model; as he entitled his work:

  Mē legat antīquās vult qui prōferre loquēlas;

  Mē quī nōn sequitur vult sine lēge loquī.

  Let him read me who wishes to carry on the ancient modes of speech; He who does not follow me wishes to speak without law.2

  This involved a practical shift which was greatest for the Romance-speaking scholars. When reading out a text, they now had consciously to deviate from their traditional, vernacular pronunciation of the language: for example, viridiārium, ‘orchard’, could no longer come out as verger, as it would when they were speaking naturally.3 The practical shift ultimately led to a conceptual one. Gradually, they began to see this written style differently: grammatica was not just the natural, indeed the only correct, way to write for speakers of a Romance idioma; once given a distinct style of pronunciation, it was a separate language, just as it was for their German-speaking fellow-citizens (and the English- and Irish-speaking scholars across the seas).

  Once written Latin had become established as a distinct, if not yet foreign, language, occasions began to arise when there was a need to write down something that would explicitly record the sounds of a vernacular. The earliest known example of this is the so-called Strasburg Oaths of 842, when two brothers, Ludwig the German and Charles the Bald, grandsons of Charlemagne, had to swear to support each other in the hearing of their respective followers, but in a situation complicated by the fact that their audiences spoke different languages, German and Romance. Their words have been recorded for us verbatim by Nithard, yet another grandson of Charlemagne,4 and the Romance version provides the first surviving text in Romance rather than Latin. It seems that the texts had been set down before they were uttered. It was highly unusual for anything other than proper Latin to be written down, and to explain it, it is assumed that the purpose was to offer each of the two brothers a crib sheet.5 Any Romance speaker could of course read out a Latin text to the common people in a pronunciation that they might understand: he would just come out with the vernacular words suggested by the Latin text. But it was a very different matter if a German speaker were to be asked to do this. And so Ludwig was offered the ninth-century equivalent of a teleprompter.

  The first few phrases will show that speaking Romance was no longer just a matter of changing a few details of regular Latin:

  Pro Deo amur et pro christian poblo et nostro commun salvament, d’ist di in avant, in quant Deus savir et podir me dunat, si salvarai eo cist meon fradre Karlo et in ajudha et in cadhuna cosa, si cum om per dreit son fradra salvar dift…*

  In proper Latin, one cannot get much closer to this than:

  Pro Dei amore et pro christiano populo et nostro communi salvamento, de hoc die in posterum, in quanto Deus sapientiam et potentiam mihi donabit, sic servabo ego hunc meum fratrem Carolum et in adiumento et in re quaque, ut quis iure suum fratrem servare debet…

  This need for transition between written and spoken language was the major problem left unsolved by Alcuin’s reforms. He had provided a common spoken and written form of Latin that would unite the literate across western Christendom, from Donegal to Dalmatia. But the cost was that now ordinary Romance parishioners could not understand their own priests during church services; and in this era, to ensure orthodoxy, not only liturgy but even the sermons tended to be recited or read from a written Latin text, rather than delivered extempore. As a result, at the Council of Tours in central France in 813, as at the Council of Mainz in Germany in 847, an explicit exception is made, to guarantee the continued understanding of the people: ‘…And that each should work to transfer the same homilies into plain Romance or German language [rusticam Romanam linguam aut Thiotiscam], so that all can more easily understand what is said.’6

  Preservation of documents for a thousand years tends not to happen without serious intent, and so not surprisingly there is little record of the vernacular languages when all the serious records were still being kept in Latin. There is a cheese-larder list from a Spanish monastery datable to the late tenth century, preserved because it had been scribbled on the back of a document of donation.7 But in the ninth, tenth and eleventh centuries, phonetic transcriptions of vernacular languages are usually found as little snippets in Latin documents. There are verbatim statements in Italian, recorded as sworn, to validate ownership for lands belonging to Montecassino monasteries. There is a vivid caption to a fresco on the wall of St Clement’s church in Rome from the late eleventh century, illustrating a famous but futile attempt at persecution of St Clement, when his attackers were miraculously deluded into mistaking him for a column. Their leader shouts to his men:

  Filli delle pute, traite. Gosmari, Albertel, traite. Fàlite dereto colo palo, Carvoncelle

  Sons of whores, pull! Gosmario, Albertello, pull! Push back with the stick, Carvoncello!

  while the saint comments in (ungrammatical) Latin:

  Duritiam cordis vestris saxa traere meruistis

  Hardness of heart yours rocks to pull you have deserved.

  Only when serious works of literature started to appear in the vernacular, invading the traditional ground held by the written language, did the real status of the ‘rustic’ languages begin to become clear. And this happened first at the other end of
the Romance-speaking world, in Normandy and England, where the Normans started writing down ballads and lays of the kind that they heard the minstrels sing. The Chanson de Roland, from the late eleventh century, is the oldest and best of these works, telling the tale of a heroic rearguard action fought against the Moors in the time of Charlemagne. It is signed in its last line:

  Ci falt la geste que Turoldus declinet

  Here ends the adventure that Turoldus retold

  and there seems no reason not to identify this Turold with a specially named character who appears in the Bayeux Tapestry, delivering a message to William the Conqueror.

  In the twelfth and thirteenth centuries, poetry in the Romance languages begins to be written down all over western Europe, in Provence, in northern France, in Galicia, Castile and Catalonia, and in Italy. The breakthrough came in areas that Latin had never strongly represented, in the celebration of courtly love—the modern sense of the word ‘romance’ is no coincidence—and in heroic tales of chivalry and war. Latin was increasingly hived off as a learned language for monasteries, schools and universities.

  The first theorist of these new linguistic developments is none other than the leading Italian poet, Dante Alighieri, who lived from 1265 to 1321. In his De vulgari eloquentia he recognised that Latin, grammatica, was in essence the preserved older form of the Romance languages.*

  He seems to have had as much difficulty in convincing his audience that these ancestral differences were the predictable result of gradual change as Darwin was to find, with a different subject matter and timescale, five centuries later.