When Germanic was first caught in the candle flame of writing, in the second century CE, there was only one runic script and one Germanic language. That language was spoken in a relatively compact area centred on the Jutland peninsula, extending southwards towards the Alps, and it had yet to fragment. Linguists consider that it was still very close to Proto-Germanic. They disagree as to when Proto-Germanic was born (most say 500 BCE, a few put it as early as 2000 BCE), but the prevailing view is that it developed out of dialects that arrived in the region with the first Corded Ware warbands.
Italic and Celtic looked quite different at the moment they were first written down. The Italic languages had already diverged into Latin, Oscan, Umbrian and possibly Venetic – a language which, as its name suggests, was spoken in the north-eastern corner of the modern country, around Venice. Since no early Italic inscriptions have been found outside that country, the parent language, Proto-Italic, is thought to have been born in or close to it. Its date of birth is usually fixed, rather loosely, at sometime before 1000 BCE.
Celtic was also mature by the time it was etched into stone, but it was spoken across a much larger swathe of the continent. Early Celtic inscriptions record three distinct languages: Gaulish in Gaul, Lepontic in northern Italy, Celtiberian in present-day Spain and Portugal. Celtic is presumed to have been spoken further east as well, in part because the Greeks and Romans – Europe’s first historians – said it was. Linguists consider that the common ancestor of all the Celtic languages was spoken at roughly the same time as Proto-Italic, but where it was spoken is a more difficult question to answer, because of the huge territory that Celtic-speakers filled when their languages hove into view.
Proto-Celtic has been pinned on the Atlantic seawall in the west, in the Austrian Alps in the east, and on the border of modern France and Germany in between. The last theory is the leading one today, in part because the French–German borderlands boast the highest density of Celtic place and river names. On the grounds that such names resist change because they serve a valuable function as signposts, some linguists consider them useful indicators of where ancient languages were spoken. The Rivers Main and Meuse were both named for Celtic deities, while the Neckar prob- ably took its name from a Celtic root nik, meaning “wild water”.
Another reason to place Proto-Celtic there, and not closer to the sea, is that its reconstructed vocabulary contains very few words related to maritime technology. Linguist David Stifter reports that Proto-Celtic had to borrow words for ‘ship’ and ‘sail’, suggesting that its speakers were landlubbers.
Germanic, Celtic and Italic are related by common descent. This is evident from their grammar, their pronunciation and their core vocabulary (English father – mother – brother; Old Irish athir – máthir – bráthir; Latin pater – māter – frāter). But the relationships between them aren’t equal. Celtic and Italic are generally considered to be closer to each other than either is to Germanic, like twins with a third sibling. The first two form superlatives in the same way, while Germanic does it differently. Germanic also has a whole class of verbs, the so-called modal verbs, that the other two lack. These are verbs that get placed before another verb, in its infinitive form, to express possibility, intent, ability or necessity (English examples are “must”, “shall” and “could”). Some linguists suspect that Italic and Celtic arose as a single, possibly short-lived language, Italo-Celtic, while Germanic arose separately. And they think that all three split from Proto-Indo-European early on, before the centum–satem (hard k to soft s before certain vowels) switch.
Loanwords help to place these prehistoric languages in time and space too. These fall into three categories: loans that Italic, Celtic and Germanic made to each other; loans that they received from sister branches that expired before they could be written down; and loans from the non-Indo-European languages that dominated the continent when the first Indo-European-speakers arrived.
Of the loans that the three surviving branches made to each other, the borrowings are more obvious between Germanic and Celtic than they are between Germanic and Italic, as if the first two had been closer in space. Relatively early on, for instance, Celtic donated its “king” word, rīg-, to Germanic, where it became rīk- (the root of German Reich and Dutch rijk, meaning “empire”).
We know this word came from Celtic because, of the three branches, only Celtic converted the ē in Proto-Indo-European h3rēg’s to an ī (the Italic branch kept the ē, as in Latin rēx). Think of Vercingetorix, the Gaulish king or “supreme king”, to translate his name literally, who led an unsuccessful revolt against Julius Caesar in 52 BCE.
The dead Indo-European sister languages left their ghostly mark, curiously, in a clutch of words for animals with big feet: Dutch pad (toad), Irish pata (hare), Welsh pathew (dormouse). These loans also hint at the early proximity of Germanic and Celtic, which must have been close enough to borrow from the same lost language. Then there are the loans that came from the Neolithic farmers of Europe. From these natives, most likely from the women they abducted or took as wives, the Indo-Europeans absorbed a host of words for plants and animals that were rare if not absent on the steppe. They included “lark”, “blackbird”, “turnip” and possibly a word meaning “bull” (tauro-, the root of Minotaur and toreador). The vocabulary tracked the distribution of the flora and fauna it described. In the balmy south of Europe, the immigrants took up words that would become cupressus (cyprus), ficus (fig), lilium (lily) and rosa (rose) in Latin. From the foragers who haunted the wind-blown dunes of the north Jutland coast they acquired a word that became Proto-Germanic selhaz, and eventually English “seal”.
These loans, which are still in use today, are older than the Indo-European languages they enrich, since the farmers brought them to Europe thousands of years earlier, from the Near East (many may have been Hattian in origin). A few may be older still, if the farmers borrowed them in turn from the hunter-gatherers who inhabited Europe when they arrived. And along with the words came knowledge. Now, at long last, the Indo-Europeans acquired a word for “bee” (bhi-), probably because their indigenous wives taught their children the art of sylvestrian beekeeping – how to lure a swarm to a hollow tree or other crevice and periodically harvest honey from it.
Bringing the linguistic clues together with the archaeological and genetic evidence, some linguists propose that the immediate ancestor of all three branches – Italic, Celtic and Germanic – was spoken during the mining boom that launched Europe’s Bronze Age. On the modern Czech–German border, which might have roughly coincided with the frontier between the Corded Ware and Bell Beaker worlds, stand the Erzgebirge or “Ore Mountains” – the only place in continental Europe where copper and tin occur together. From 2000 BCE, this region was home to a number of towns that doubled as important centres of metal production and ritual. They would have attracted people from both worlds, who brought a kaleidoscope of still mutually intelligible dialects. (A famous relic of those prehistoric towns is the Nebra sky disc, a bronze disc with a blue-green patina, inlaid with gold symbols, that depicts a solar boat sailing across the celestial ocean. A replica of it floats high above our heads on the International Space Station.) Around 1600 BCE, boom turned to bust in the Ore Mountains, and people left the region in search of better prospects. This could have been the moment of schism between Germanic dialects, which stayed in the north, and Italo-Celtic ones, which moved away to the south-east. Some of the émigrés settled on the Hungarian Plain, along the old trade routes connecting the Baltic to the Aegean (the roads along which amber and bronze were carted, and much else besides). There they built a series of large and well-defended settlements lining a corridor formed by the Rivers Tisza and Danube. Like their Bell Beaker ancestors, these “Hungarians” worshipped the sun, but unlike the people of Nebra, who had buried their dead, they cremated theirs, packing the ashes into urns which they then buried in fields. This distinctive death rite now began to spread in two directions – north-west across Austria, and south-west towards Slovenia and Italy. If an Italo-Celtic language spread with it, the thinking goes, then where the path forked, the nascent Proto-Italic and Proto-Celtic languages parted ways. It’s just one theory, but the enigmatic and long-extinct Venetic language lends some support to it. Around four hundred Venetic inscriptions are known, and some linguists conclude from these that Venetic was not Italic, but something older – a relic of the ephemeral Italo-Celtic tongue (meaning that if you heard it spoken, you’d have an inkling of what the languages of the Roman emperor Nero and his nemesis, the Celtic queen Boudica, sounded like when they were one). Some Venetic inscriptions were found in Austria and Slovenia, exactly where Italic and Celtic might have taken leave of each other.
At about the same time that the Hungarian settlements came to be, very similar towns developed in northern Italy. Archaeologists see enough similarities between the two – notably the “Urnfield” cremation style – to convince them that they were connected by trade and human traffic. Steppe ancestry had already reached northern Italy by 2000 BCE (from where it diffused gradually southwards, towards the future city of Rome), but if the people who first brought that ancestry spoke an Indo-European language, it was probably one of the lost ones. Linguists suspect that the forerunner of Latin arrived later, with the migrants from Hungary. By 1600 BCE, those migrants were settling in the Po Plain, close to the modern city of Parma, and from that time on the population of the region grew. The people who frequented its thriving markets, who also carried steppe ancestry, might have bartered in Proto-Italic.
The markets thrived until 1200 BCE – the ominous date that sounded the death knell for so many Mediterranean civilisations, including the Hittites and Homer’s Greeks – but then both the northern Italian and the Hungarian civilisations vanished from the archaeological record. They might have suffered from the wider economic downturn, or perhaps a new pulse of migration out of Hungary triggered a crisis in Italy. Tens of thousands of people fled the Po Valley, scattering with their pottery and dialects to other parts of the Italian peninsula. As they went, linguists think that Proto-Italic split into Latin, Oscan and Umbrian. All three languages lived long enough to be written down. Oscan graffiti on the walls at Pompeii guided its inhabitants towards mustering points in times of siege.
The second stream of migrants from Hungary headed north-west across Austria, plausibly carrying the dialects that would become Proto-Celtic and lending a “king” word to the early Germanic-speakers into whose orbit they now strayed. From its Rhine cradle, Proto-Celtic then expanded, fragmenting as it went into Gaulish, Lepontic and Celtiberian in the west and undocumented sister languages in the east.
For every scenario I’ve sketched here, at least one alternative exists. As with all the early branchings of the Indo-European family, uncertainty reigns – even if that is less true than it was in Marija Gimbutas’ day. But perhaps the greatest outstanding mystery regarding Celtic is when it reached Britain and Ireland – the only places where, besides Britanny in France, it is still spoken today.
After the Beaker people came to Britain around 2450 BCE, their DNA replaced around ninety per cent of the local gene pool, and all of the Y chromosomes. The turnover was similarly dramatic in Ireland when they crossed the Irish Sea around two hundred years later. Such a dramatic genetic rupture was almost certainly accompanied by a linguistic one, and linguists are pretty sure that the Beakers introduced an Indo-European language to those islands, but they don’t think that language was Celtic. The dates don’t add up. The Beakers were gone from Britain and Ireland by 1800 BCE (though their genetic legacy lived on), and Proto-Celtic was born no earlier than 1500 BCE. Somebody else brought Celtic in, at which point the language of the Beakers atrophied and died. One suggestion is that it was farmers from what is now France who crossed the Channel in large numbers around 1200 BCE, but some linguists think that date is still too early. They suspect that a later group of immigrants brought Celtic in, whom geneticists have yet to detect.
Excerpted with permission from Proto: A New History of Our Ancient Past, Laura Spinney, HarperCollins India.
