Human societies don’t just come from nowhere, and nor do the languages we speak. And unlike the archaeological record, our languages are still living lineages, inherited directly from how our ancestors spoke. Romanian is a direct linguistic legacy of Rome, for instance — and so too are Italian, Spanish, French, and various others. Whereas English is most definitely not.
If Europe’s multiplicity of ‘peoples’ seems daunting, then try some basic linguistics, and the complexity dissolves. Who are or were the Etruscans, Basques, Bretons, Serbs, Kosovars, Estonians and Hungarians? Their languages immediately give a ready framework of reference points in time and geographical space, of shared or different origins, and past population contacts.
This works worldwide, and does not need any written history to help. Languages themselves open up a whole new window on our past, a ‘second opinion’ to corroborate or challenge the perspectives from archaeology and genetics. Ignore the linguistic dimension to prehistory and you really are missing a trick. If some theory on our origins cannot explain the linguistic record too, then it is incomplete or wrong, and needs correcting until it can.
Most of the world is dominated by just a few great families of languages: Bantu over southern Africa, Indo-European across much of Eurasia, and so on. A recurrent theme is how and why these families came to spread, at the expense of the languages spoken before them. Was it the people who spoke them who moved, and outpopulated other groups? Or did people mostly stay put, and switch language, so it was just the language that spread? Do the lineages of our languages tend to match those of our genes, then, or not?
Many different ‘forces of history’ have shaped the fates of our languages, but the main pattern is how the balance has shifted over time, from the Neolithic to today. Languages tend to mirror the growing disparities between the human groups that spoke them: in subsistence and demographics; in the scale and organisation of human societies; in technologies for transport and warfare. In the great debate on the ultimate vs. proximate causes that shaped our modern world, there is plenty still to learn from the record left by our languages.
Take the Andes. This is one of the world’s few independent hearths of ‘civilisation’, and home to one widespread language family, Quechua. Everyone but the linguists has presumed a simple, uninformative story: that the Inca empire spread all Quechua, out of their capital in Cuzco. In fact, language relationships within Quechua provide evidence of how the nature and impacts of Inca control differed significantly across their empire, and can even track how the Incas moved particular populations, from A to B, around their realm. In any case, Quechua did not first spread out of Cuzco, and did so long before the Incas. So it’s actually a source data on the nature and impacts of their predecessors too, the Wari.
Likewise in Europe: the Romance, Germanic and Slavic languages are a record of the impacts of the Roman Empire and the upheavals that followed, and how they too varied from region to region.
Written records of ancient languages are very valuable, but not indispensable. By comparing even just surviving languages, we can work out which are related to each other, and how closely, and we can partly reconstruct their ancestral stages. Any such ‘language family’ stands as proof of some past cultural and/or population expansion that our prehistory needs to account for. The task, together with archaeology and genetics, is to identify where and when it began, and how and why it spread.
Languages can also be compared on very different criteria, though, in general structural type. Often, clear geographical patterns emerge: ‘language convergence areas’. Unfortunately, outside linguistics these are much less well known than language families, and all too often confused with them. Whereas what language areas tell us about prehistory is actually very different: not an expansion and divergence of people and cultures out of a single source, but contact, interaction and convergence out of multiple independent sources.
Historical linguistics can ‘reconstruct’ the sounds of some individual word roots back to an ancient, ancestral language. Latin rex, for instance, has the same source as Sanskrit raj- and Old Irish rí. Does this not also ‘reconstruct’ a Proto-Indo-European-speaking people, then, who must have had a ‘king’?
In fact, the language data do not mean anything so sure. Reconstruction is reliable in its sound laws; far less certain is what a word meant exactly at an early stage — in this case, presumably just some figure of authority or ‘chief’. Language does not reconstruct some long-lost ethnic identity or cultural complex. There are still some diehard claims out there: that Indo-European can only match an archaeological culture that already knew the wheel, for instance. But that too is interpretation far beyond what the language data really entail: no more than some concept of cyclicality or rotation, whether in life, the seasons or the heavens, long before the wheel.
Many old certainties have evaporated, many language prehistories need to be rewritten. The traditional methods and findings of historical linguistics are now being challenged by radically different quantitative and ‘evolutionary’ analyses, co-opted from the biological sciences and set to work on language data. Sure, there have been some over-enthusiastic excesses here too, but when soundly applied these methods can revolutionise our understanding of language prehistory. Roll on the ‘battle of the methods’…
If you value the great diversity in human languages, the future is very bleak indeed. It took tens of millennia to build up, and now, like many other ways of life and cultures, languages are dying at breakneck speed, one every week or two. Almost all current trends are leading to social units and networks at ever larger scales, with ever more communication and mobility right across them. That inexorably favours just a few juggernaut languages like English, Spanish and Mandarin.
Not that this is new. Globalisation is just the tail-end of a process that has been gathering pace for millennia, and especially in the last few centuries. The nation state, colonialism and imperialism all played their part, and before them the great empires of Antiquity, back even to the ‘Neolithic Revolution’. Bantu languages, for example, over the last three millennia or so, have steamrollered almost all the linguistic diversity that had underlain them across the entire southern half of Africa. Only a precious few click languages survive, most of them already doomed.
This is what these chapters are about: trying to work out the very unequal fortunes of our language lineages, and what they tell us also of the origins and fates of the people who spoke them.
I was twelve when I had in front of me the Latin sumus, estis, sunt. I was a bit nonplussed: it didn’t seem as unfamiliar as I’d been expecting, only I couldn’t work out why. Cue my Latin teacher: “Where do you think the French got nous sommes from?” Nos sumus — nous sommes. From that point on, I think I was hooked.
Ten years on, Ceaușescu still fresh in the mind, I was standing before a Gothic cathedral in a breathtaking walled town in Transylvania. My resurrected Latin was working wonders for learning Romanian, but here, a thousand miles from my starting point in Berlin, the door was answered in native German. And why not? My host was a ‘Saxon’, his home town founded by his forefathers, eight centuries ago. I revelled in Romania, a historian’s and linguist’s wonderland — and one month later, I applied for a Master’s in historical linguistics.
“Not seeing the wood for the trees” is a classic stumbling block in language prehistory. When historical linguists look at language families, their basic model is ‘family tree’. Yes, they do pay lip service to alternatives: more network-like relationships, waves not branches, and ‘dialect continua’. But in practice historical linguistics favours neat, clear-cut branches. They are very tempting: easier to get your head around, and to handle mathematically, so modern ‘evolutionary’ analyses also tend to prefer them (although the best have some useful workarounds).
Where things go awry is when you try to set some nice family tree into the real world. Human populations do not live only in neat binary branching relationships — so on what grounds can we expect their languages to? In some contexts, like distant one-off migrations, the real world may oblige, but for most of the world and most of prehistory, it has done its own thing. Several breakthrough moments for me have come from daring to doubt a traditional tree. Time and again — for Germanic, Romance, Indo-European, Quechua — the language data suddenly re-form to point to a very different scenario for prehistory, and a far more plausible fit, at last, with archaeology and/or genetics.
Perfectly. They are my wider research and interests. If it hadn’t been for Transylvania, I’d have ended up an archaeologist anyway. For me, it was always the historical curiosity that came first. Linguistics can’t by itself explain how any particular languages came to be how they are, where they are, and relate to each other as they do. How did seven towns in Transylvania come to speak their ‘weird’ forms of German? How come Quechua ended up in the highlands and jungle of Ecuador, and the Andean foothills of Argentina? None of that has any explanation unless we put languages in the real-world contexts that made them how they are: the history and prehistory of the people who spoke them.
In our chapters here we’re trying to turn it around, to show how much linguistics can contribute to understanding the past, especially wherever and whenever we do not have the luxury of written history.
I was in Trinity College library in Cambridge, looking for a different book, when I chanced upon Archaeology and Language, by Colin Renfrew. I ended up reading half of it there and then. I’d long been fascinated in both halves of the title, but historical linguists’ attempts to bring them together had generally left me cold, and deeply dubious. From what I knew of history and archaeology, they seemed like unlikely stories. And as a linguist, too, they looked like a house of cards. But nobody seemed to dare challenge the linguistic orthodoxy. Renfrew did, as an outsider to linguistics, and with a richer archaeologist’s perspective on the real-world contexts of prehistory. At last someone was asking what was and what was not actually plausible as an explanation for how, millennia before the modern world, just a few huge language families had come to dominate so much of the world. Twenty years on, I’m delighted to have Colin Renfrew as my co-author of these chapters.
It’s nice to hear the penny drop for my archaeologist and geneticist colleagues, as it dawns on them how much they’ve been missing. But nothing comes close to fieldwork in some remote Andean village. There’s always the beauty and wilderness, but the language adds something more. Quechua, with its seven million speakers, is our greatest surviving link back to the speech of the New World before European conquest. Just listening to it feels so quintessentially original and native to this beautiful part of the world.
‘Enjoy’ is not quite the word, though. For linguistic fieldworkers everywhere, ‘our’ languages are dying before our … ears. And we’re mostly powerless to stop it. Your job turns into seeking out a human ‘endangered species’, the last speakers left in some remote corner of the world. It’s all pretty incongruous, rigging up your high-tech recording kit for a handful of bemused, obliging octogenarians. And your scientific mission, for posterity, is to intrude on their memories, to resurrect in their minds the words and ghosts of their past. Their long-gone childhood friends, their late spouse, mother, grandparents: how would they have said some trivial but linguistically crucial X?
Any language is a continuous lineage, passed on through the generations from as far back as human speech itself. But for this one, now the game is up. And you’re one of the very last people ever to have its final syllables ringing in your ears. To me, it feels like a responsibility, and a privilege.
If we really want a full, coherent story of our own origins, there’s a whole new dimension of data on us that we had better not ignore: the languages we speak.
The next one that I should organise myself?
It would need to be a conference where specialists dedicated to just their own single discipline — linguists, archaeologists, geneticists, anthropologists, historians, whoever — did *not* feel comfortably at home! The greatest obstacle I see to uncovering the past is not the lack of surviving data or the limitations of our methods, it’s the glass walls between our disciplines. When it comes to working out our origins, the cross-disciplinary whole is far greater than the sum of its individual parts.
Am I allowed to say: “Read our chapters”? To judge whether these are the sorts of questions you want to grapple with, try these chapters as a first port of call. The reason we wrote them is because there seems to be no general introduction to what language can tell us of prehistory, and no coherent worldwide overview. There’s plenty of discussion of individual cases, but little consistency in approach from one continent to the next.
Above all, do not suppose that language prehistory is the same thing as historical linguistics. That has always been part of the problem. One cannot understand the linguistic perspective on our past from a position of ignorance about the other realities of prehistory. So make sure you’re interested not just in languages, but just as much in archaeology, history and genetics. If that is your thing — taking all those together to piece together a multi-faceted, holistic vision of our last 13,000 years or so — then read on …