3 FIRST AMONG EQUALS Anatolian
nu ninda an ezzatteni wadar – ma ekutteni
Bedřich Hrozný stared at the passage. Next to the wedge-shaped marks that represented the word ninda was an ideogram – a character representing an idea – that he recognised from Sumerian texts. There it meant ‘bread’, so perhaps it did here too. The repeated suffix –(tt)eni suggested a verb ending, one tantalisingly close to verb endings in Latin and Sanskrit. This gave him the husk of the phrase: ‘bread x-ing, y z-ing’. But Hrozný knew that he could do better than that. In Old High German ezzan meant ‘to eat’ (the word became essen in Modern German). If such echoes were meaningful, then following the theme hinted at by the bread ideogram, wadar might mean ‘water’. Putting it all together he came up with: ‘now you will eat bread, further you will drink water’.
Hrozný had just broken the code of the first Indo-European language ever to be written down: Hittite. Though he didn’t know it at the time, the document that had given him the key was the last will and testament of an early Hittite king, Hattushili I, who had ruled in north-central Anatolia in the seventeenth century BCE. Hattushili was a master of spin, especially when it came to himself: ‘his frame is new, his breast is new, his penis is new, his head is of tin, his teeth are those of a lion, his eyes are those of an eagle, and he sees like an eagle’. But he was also an accomplished warrior who had laid the foundations of one of the great empires of the preclassical world. As he lay dying he dictated his plan for his succession, but in the ancient equivalent of the microphone being left on after the interview has concluded, an over-enthusiastic scribe kept scribbling and captured his last words.1 As death rushed up to meet him, Hattushili the Lion was seized by terror: ‘Wash my corpse well! Hold me to your bosom! Keep me from the earth!’ Three thousand years after its ancestor was first spoken on the shores of the Black Sea, the first Indo-European cry to reach us is heartrending in its humanity.
The story of Hrozný’s decipherment is only a little less so. He cracked the code in 1915 and quickly announced his preliminary conclusion that Hittite was an Indo-European language. The Czech linguist was in a hurry. He had just become a father. The First World War was sixteen months old and, though he was extremely short-sighted, he knew that he could be called up at any moment. He wanted to secure his academic reputation. There was also competition to decipher Hittite, ever since rumours had begun to circulate, more than a decade earlier, that despite being written in the same cuneiform script as the Akkadian language spoken further to the south and east, it belonged to a different family.2
So far the rumours had not been taken seriously. At the turn of the twentieth century, relatively little was known about the Hittite Empire. It was only in 1906, thanks to excavations near Bogazköy, about a hundred and fifty kilometres (ninety miles) east of Ankara, that its impressive capital, Hattusha, had come to light. Conventional opinion held that none of the great Bronze Age civilisations, not Egypt nor Babylonia nor Assyria nor the Hittites, had spoken Indo-European. The Phrygians were thought to have been the first speakers of Indo-European on the Turkish peninsula, starting from the twelfth century BCE. (The rumours had nevertheless filtered out beyond academia. Austrian writer Robert Musil threaded them into his famous novel The Man Without Qualities, which is set in Vienna on the eve of the First World War: ‘… in his eyes even the ladies and gentlemen of the highest society performed a significant if not readily definable office when they chatted with learned experts on the Bogazköy inscriptions …’.3)
The 1906 expedition to Bogazköy (now Boğazkale) had turned up thousands of inscribed clay tablets which the archaeologists realised were the Hittite royal archives. It was these documents, which appeared to have survived a great blaze, that had allowed Hrozný to take his first steps towards deciphering the language and demonstrating its Indo-European credentials. Before the war, he had been sent to Constantinople (Istanbul) to remove them from their packing cases, clean, sort, photograph and copy them. Having returned home to Vienna, he announced his preliminary conclusion to a chilly reception from his peers. He realised that he would have to supply more evidence. Then he was called up.
Luckily for Hrozný, as a bespectacled and somewhat bookish recruit to the Austro-Hungarian army, he was sent back to Constantinople. There he came under the relaxed command of one Lieutenant Kammergruber, who allowed him to spend his days at the Imperial Ottoman Museum, insulated from the chaos of the wartime city by heaps of fire-hardened clay. Hrozný took care to thank Kammergruber when he published his more thorough findings in 1917. This time his fellow Indo-Europeanists took him seriously, but they also expressed their discombobulation before the sheer strangeness of Hittite, a language that is more different from English than Urdu or Albanian, while still, as they would eventually acknowledge, belonging to the same family.
Hittite isn’t alone in the Anatolian branch of the family, though it is the best-documented member of that branch. Other long-dead Anatolian languages would be identified later, including Palaic (once spoken in northern Anatolia), Luwian (thought by some to have been spoken at Troy) and Lydian (thought by others to have been spoken at Troy, but also by ‘rich as’ Croesus, the last king of Lydia). A new Anatolian language was discovered at Hattusha as recently as 2023, on a clay tablet unearthed during that summer’s excavation, but it has yet to be deciphered.4 The main reasons for classifying the Anatolian languages as Indo-European are the ways that they inflect nouns and verbs to signal their role in a sentence (including those verb endings that helped Hrozný to decipher Hattushili’s will). But there are also many ways in which the Anatolian branch differs from the other Indo-European languages. It is the only one to preserve de Saussure’s lost consonants, the laryngeals. Hittite verbs have two tenses, past and present; other ancient Indo-European languages have up to six. Hittite marks nouns according to whether they are animate or inanimate only; other Indo-European languages further divide the animate category into masculine and feminine.
From the outset, the Anatolian branch struck linguists as old; as capturing an archaic, even nascent state of Indo-European. They disputed where it stood in relation to Proto-Indo-European, but the prevailing view was that it was the eldest daughter, the first to split from the Proto-Indo-European trunk. All agreed that it had to have been born very close to the birth of the entire family, meaning that any theory of the origins of Indo-European had to explain the origins of Anatolian too, in all its glorious eccentricity. That task would occupy scholars throughout the twentieth century and into the twenty-first.
Through the sitting-room window I’m watching dusk thicken the cascading branches of an aged cedar tree. Over the wall, from next door’s garden, looms a giant sequoia. I’m in Cambridge, England, at the home of a giant of twentieth-century archaeology, Colin Renfrew. Lord Renfrew of Kaimsthorn, to give him his proper title, is explaining that the exotic flora were imported by a former master of Jesus College with a passion for botany, when suddenly the door opens and in comes Jane, Lady Renfrew, balancing a teapot on her walking frame. She is also an archaeologist, one who in her day specialised in the use of plants in prehistory. Soon we’re sipping tea, eating sponge cake and discussing another titan in their field, Marija Gimbutas.
Even before the twentieth century, archaeologists knew that there had been two major cultural transformations since the last ice age, which had affected both Europe and Asia. The first was the farming revolution. The second was triggered by those nomadic herders who roamed the western Eurasian steppe about five thousand years ago. In neither case could they prove that the cultural shift had been accompanied by migration. The new ways of life could simply have been taken up by existing Europeans and Asians, without the need for any movements of people. But if new cultures could diffuse through populations, so in theory could new languages. There was general agreement that a transformation as dramatic as the spread of the Indo-European languages could not have happened in a vacuum. It had to have piggy-backed on some major social upheaval, and that upheaval must have left a trace in the archaeological record.
Gimbutas was the leading proponent of the theory, discussed in the last chapter, that the steppe nomads had spread those languages. The ‘steppe hypothesis’ had been aired since the nineteenth century, but she did much to develop it after the Second World War when she herself was rootless. Born in Lithuania, she had been forced to flee by the Soviet occupation of her country, eventually settling with her family in the United States. By the time she took up a post as professor of European archaeology at the University of California, Los Angeles (UCLA) in 1964, her theory was fully formed. In her conception of the past, south-eastern Europe had once been home to peaceful, mother-centred farming societies who had practised a goddess cult. These societies she referred to collectively as ‘Old Europe’, because they had been founded by the first farmers to settle the continent.5 Starting around 4500 BCE, several waves of horse-breeding nomads had come west from the steppe, entering the farmers’ orbit. The invaders were male, aggressive and patriarchal, and in the ensuing culture clash Old Europe ceased to exist.
When Gimbutas presented this theory at UCLA, she found that she was out of synch with her fellow archaeologists. A new paradigm had emerged after the war, according to which diffusion, not migration, was the dominant driver of change in prehistory. This ‘new archaeology’ rejected the simplistic equation of flux of people with flux of language, the error, as its proponents saw it, of the discredited archaeologists associated with the Nazi regime. Some of the leading thinkers in the new paradigm were at UCLA, but it fell on fertile ground all around a world traumatised by the memory of the Holocaust. Jane recalls the first time she saw Gimbutas, at a conference of prehistorians in Prague in 1966: ‘She was very elegant and she had wonderful red hair. She was wearing this beautiful green dress, and she was talking about Indo-Europeans. After a few minutes, once she’d really got going, Professor Hawkes [Christopher Hawkes, then professor of European archaeology at Oxford University] got up and banged on the desk. He said, “Don’t take any notice of this woman, she’s talking absolute nonsense,” and sat down again. She said, “Oh my good friend Professor Hawkes and I disagree, but I’m just going to carry on,” and she did.’
Jane sighs happily.
‘She took a great deal of flack,’ says Colin.
Gimbutas was older than Renfrew by a generation, but they came together through their shared interest in Old Europe. He was an authority on radiocarbon dating, she was a polyglot with deep knowledge of the languages and folklore of eastern Europe. They respected each other’s knowledge and energy and collaborated on several projects. In 1967 she invited him to spend a term at UCLA, lending him and Jane a house that she owned in Topanga Canyon. While he was there he came under the influence of the new archaeology, and the two of them had lively discussions. In 1987, by which time he had been named professor of archaeology at Cambridge University, he published his own theory of the Indo-European languages: they had originated in Anatolia nine thousand years ago, spreading east and west with farming. The two leading theories of the Indo-European homeland, and the only two to survive into the twenty-first century, were therefore the brainchildren of two friends who had excavated side by side at the same sites.
Archaeologists were more enthused by Renfrew’s model. They weren’t convinced that marauding bands of steppe nomads could have succeeded in imposing their languages on what was, by five thousand years ago, a large and dense population of farmers, in Europe as in parts of Asia. And they were sceptical of Gimbutas’ ideas about the matriarchal organisation of Old Europe. These remain controversial today. Many archaeologists acknowledge that women played different roles in farming and steppe societies. Women were likely more central to food production in the settled societies, for example. Some archaeologists even think that the hundreds of fat-bottomed female figurines that have been extracted from Old European cemeteries reflect a reverence for the generations of women who coaxed nature into providing new foods. But even if farming societies worshipped female deities, archaeologists then and now felt that those societies were as male-dominated and warlike as any other in prehistory.
Linguists, on the other hand, never really bought Renfrew’s hypothesis. For them the dates didn’t work. The oldest texts in the Hattusha archives date to 1650 BCE, but Hittite names and words crop up in documents that are three hundred years older. There are texts written in the other Anatolian languages that are almost as old, and having compared these with the Hittite texts the linguists concluded that by 2000 BCE Hittite and its sisters were already distinct languages. Extrapolating backwards, allowing time for that divergence, they estimated that the parent language, Proto-Anatolian, was spoken around 3000 BCE. If Renfrew was right, and Indo-European-speaking farmers had moved out from Anatolia around 7000 BCE, the language they left behind, Anatolian, would have had to remain more or less intact for four thousand years before it split. Such stability is unheard of in the history of language. Between five hundred and a thousand years is the rule of thumb for the time it takes a language to evolve into a new one, and it is generally agreed that the interval would if anything have been shorter in preliterate, pre-state societies, where there was no writing or official state language to hold back change.
The linguists had plenty of other objections to Renfrew’s Anatolian hypothesis. For one, they considered the evidence overwhelming that Indo-European languages were intrusive to Anatolia – that they had come into it from outside. The Hittites’ own archives revealed that they were surrounded on all sides by speakers of non-Indo-European languages, including the people thought to have been the indigenous inhabitants of the region: the Hattians. As early as 1920, an American linguist, Carl Darling Buck, noted that although Hittite followed Indo-European rules of word formation, it was awash in ‘alien’ vocabulary. Much of that vocabulary turned out to be Hattian.
The linguists were much better-disposed towards the steppe hypothesis, according to which several waves of nomadic herders had departed the steppe starting in the late fifth millennium BCE, seeding the Indo-European languages throughout Eurasia. In this model, the first to leave had travelled down through the Balkans, crossing the Bosporus and implanting Proto-Anatolian in Anatolia (David Anthony equated these migrants with the mace-wielding chieftains who had toppled Varna). It made sense to the linguists to think of Indo-European-speakers entering from the west because ancient records attested that Anatolian languages were concentrated there, while being noticeably absent from the east of the peninsula. The linguists agreed with Renfrew that the farmers’ languages had probably spread with them, several millennia earlier; they just didn’t think that those languages were Indo-European. In their view, the farmers’ languages were now lost to us, except for relics such as Basque, having been displaced by the Indo-European languages that came later.
By the time Gimbutas died in 1994 she was famous beyond academia and her books had a large, mainly female following. Her writing on the goddess cult of Old Europe had given a new generation of feminists hope that societies could be organised differently, because they had been once before. The male-dominated academy took a dimmer view. They accused her of abandoning scientific rigour, of letting her imagination run away with her. As far as the Indo-European languages were concerned, however, both her and Renfrew’s theories remained on the table. Others had taken up and refined the steppe hypothesis, notably her former student, James Mallory, and Anthony, but the debate had reached an impasse. It took progress in another field to push it on.
Starting in the 2000s, genetic studies (based on modern genes, not yet on ancient DNA) furnished evidence that migration had, after all, been a powerful force in prehistory. It was then that the long-running debate over the farmers was finally cleared up: not just their technology, but they too had moved, and in force. Their ancestry had replaced at least forty per cent of the European gene pool, and in some places close to a hundred per cent. The first hints emerged that they weren’t the last major migration to the region either, because modern Europeans carry the genetic legacy of a third population, besides the farmers and the hunter-gatherers they had displaced.
In 2015, two papers were published in the leading journal Nature that confirmed those suspicions about a third major influx to Europe. Using two different methods to analyse ancient DNA, they came to the same conclusion: migrants had radiated east and west from the steppe around five thousand years ago, and in Europe their ancestry had replaced up to ninety per cent or more of the gene pool. The mainstream press went to town on the discovery, and with justification, because the last piece of the tripartite puzzle that defines modern Europeans genetically had fallen into place. They remain overwhelmingly, to this day, part hunter-gatherer, part farmer and part steppe nomad. It’s hard to overestimate the seismic impact of the advent of first the farmers, and then the nomads. No later movement had anything like their genetic, cultural or linguistic legacies: not the massive migrations set in train by the fall of the Western Roman Empire, nor the displacements that followed the Black Death, the 1918 flu or either of the world wars. Most European men alive today, and millions of their counterparts in Central and South Asia, carry Y chromosomes that came from the steppe.
Some scholars continued to support the idea that the Indo-European languages were born in Anatolia nine thousand years ago, but they were now in the minority and the man who had proposed it was not among them. Giving the inaugural Marija Gimbutas memorial lecture at the University of Chicago in 2017, the by then eighty-year-old Renfrew retracted. In the main thrust of her argument, he said, if not in every detail, his friend and intellectual jousting partner had been right.
When David Reich was a teenager in the late 1980s, he liked to play a game called Civilization. Not the computer game that later took the world by storm, but a quaint old board game that knew some considerable success of its own. It was a game of strategy that, for once, was not all about war. Each player was assigned a population around the Mediterranean Basin at the dawn of agriculture. Trading commodities and calamities, purchasing ‘civilisation cards’ such as philosophy or literacy or medicine, they attempted to expand their empire at the expense of others’. The board was colour-coded according to the boundaries of actual ancient civilisations (Egypt, Crete, Italy …), and the winner was the first to make it all the way through the ‘archaeological succession table’ to 250 BCE.
The game, Reich explained to those of us packing an auditorium in Leiden, in September 2022, was the inspiration for a trio of papers that he and more than two hundred collaborators had just published in another eminent journal, Science. Summarising their findings was a map centred, this time, on the Black Sea. The map’s border was colour-coded to represent the four genetically defined populations that had occupied the western end of Eurasia ten thousand years ago: Iranian farmers, Anatolian and Levantine farmers, Balkan hunter-gatherers and Eastern hunter-gatherers (those who came west from the Caspian Sea after the last ice age, settling around the Dnieper Rapids). Pie charts showed how these populations had mixed over the following millennia, and coloured arrows between the pie charts indicated how migration might have driven the spread of the Indo-European languages.
There’s something a little Chaplinesque about Reich. Diminutive, supple (he frequently perches cross-legged when listening to others), he presents a watchful mask to the world until, every now and then, some emotion wells up and pushes the mask aside. You sense inner steel, which is surely what’s required to coordinate the large, multidisciplinary consortia that his kind of science involves, but also the heavy responsibility that he feels for communicating that science. No one is more alive to the troubled history of Indo-European studies than this grandson of Holocaust survivors, whom the ancient DNA revolution catapulted to scientific stardom. And here in Leiden, at a conference of historical linguists, the forty-eight-year-old Harvard geneticist had big things to say. The linguists had invited him to speak before the publication of the Science papers, and now they wanted to understand those papers’ implications for their field. Beyond the wood-panelled room people were making the most of an Indian summer, swimming in the Dutch city’s enviably clean canals or drinking beer in the sunshine. Inside all eyes were fixed on Reich, and many more people were tuning in from around the world.
To begin with, he explained, the four populations coded in the map were far more genetically distant from each other than the populations inhabiting the Black Sea region today. They were as different from each other, in fact, as modern Europeans and Chinese people. Even the Anatolian and Levantine farmers in the lowlands, and their upland neighbours the Iranian farmers, were clearly distinct in genetic terms. They were likely descended from local populations of hunter-gatherers that physical and cultural boundaries had kept apart. Then around 6000 BCE, if not earlier, Iranian farmers started moving north through the Caucasus, first in search of land, and later lured by metal. They interbred with the people they met to the north of those mountains, and by 4500 BCE individuals were being buried at Khvalynsk, on the Volga, who carried two types of ancestry: Eastern hunter-gatherer and Iranian farmer. These people were the descendants of mixed couples: steppe-dwellers who might have resembled modern Finns, and the more recent Caucasian immigrants who probably looked more like Georgians.
South of the Caucasus, meanwhile, everything was also in flux. At the eastern end of Anatolia, Anatolian and Iranian farmers were mingling at last, sharing genes and technologies. Uruk’s hunger for metal and timber, and the traffic it drove through the mountains, sucked that mingled ancestry up on to the steppe. It was already there by 3700 BCE, when the burial mounds of the fur-clad oligarchs began casting long shadows over the North Caucasus piedmont. And by the time the Yamnaya emerged a few centuries later, they carried a predominantly two-way mix of Eastern hunter-gatherer and Iranian farmer ancestry, with a dash of Anatolian and Levantine farmer.6
The Yamnaya and their descendants had carried that genetic profile east and west, as the landmark studies of 2015 had shown. Mallory, Anthony and many others thought that they had also carried Proto-Indo-European with them, and Reich agreed. But he now considered it unlikely that either the Yamnaya or their steppe-dwelling ancestors had seeded Anatolian in Anatolia. Though his group had detected Eastern hunter-gatherer ancestry in one or two burials in the Balkans, they had found none on the Turkish peninsula. From the ancient bones and teeth sent to them by archaeologists who had excavated there they had extracted plenty of Anatolian and Iranian farmer DNA, but none of the DNA of those hunter-fishers who had once hovered over the Dnieper Rapids. From 4500 BCE to 1300 BCE – from the Neolithic to the dying days of the Hittite Empire – there was virtually no trace of the people who had contributed almost half of the Yamnaya’s genetic complement, south of the Bosporus.
What this meant, Reich explained, was that there was no genetic link between the steppe and Anatolia, to mirror the linguistic link between Proto-Indo-European and Anatolian. It was possible that a very small number of steppe immigrants had managed to impose their language on the indigenous population of Anatolia, without having any impact on the local gene pool, but such cases were exceedingly rare in the history of language (it has been far more common for an incoming elite to abandon their language in favour of the one spoken by the masses, as happened, for instance, when the Huns and Mongols invaded Europe). It was also possible that Reich’s team had not yet sampled enough prehistoric Anatolian remains to detect those steppe immigrants. But given that they had drawn a blank in over a hundred and thirty samples, when each of those samples tapped into many generations of related people and steppe ancestry is unmissable in Europe, it seemed increasingly unlikely that they had overlooked a significant wave of immigration.
If migration really was the main driver of language change in prehistory, as most linguists seemed to believe, Reich felt it was time to revive a theory that predated Renfrew and Gimbutas. In 1926, an American linguist, Edgar Sturtevant, had claimed that Anatolian was not the eldest daughter of Proto-Indo-European, but its sister. They were the two daughters of an older parent language. Sturtevant’s theory was dismissed at the time, but Reich said that the genetic evidence supported it. The ancestral language might have been spoken in the Caucasus between five and seven thousand years ago, by people of predominantly Iranian farmer ancestry. From the Caucasus, his team had traced migrations west into Anatolia and north to the steppe. The first group of migrants could have carried the seed of Proto-Anatolian with them, the second the seed of Proto-Indo-European. On the steppe the northern faction interbred with the local tribes, acquiring Eastern hunter-gatherer ancestry, and in the mouths of their descendants – the Yamnaya – the imported language became Proto-Indo-European.
Between five and seven thousand years ago, in fact roughly in the middle of that window, Varna was thriving. The ancestral language could have been the mother tongue of some of the Balkan coppersmiths’ clients from over the sea. It could have been the lingua obscura we met in Chapter One. In Leiden in 2022, Reich wasn’t able to say exactly where the speakers of that ancestor had lived, whether it was to the north of the Greater Caucasus, in the mountainous region shared by modern Russia and Georgia, or in the more southerly highlands that today fall within the borders of Armenia, Azerbaijan and Georgia (overlapping with Türkiye and Iran). But he believed that further sampling would narrow it down.
Barely had he finished speaking than urgent questions came like a flight of darts from the floor. This was hardly surprising, given that he had just readmitted the possibility that the homeland of the entire family, including Anatolian, lay in Iran, where Oriental Jones had placed it in 1786, or slightly to the north and west of it, in the neighbourhood of Mount Ararat – where Noah’s ark had beached. Most of the linguists seated in that wood-panelled room considered that the roots of Anatolian lay deep in the steppe, and that it had arrived from the west rather than the east. Here was Reich challenging both tenets. But his radical proposal was about to get some support from an unexpected quarter.
At the Max Planck Institute for Evolutionary Anthropology in Leipzig, a global powerhouse of palaeogenetic research to rival Reich’s at Harvard, psychologist Russell Gray was putting the finishing touches to his own Science paper. Reich and Gray both think about language in evolutionary terms, but their approaches couldn’t be more different. Reich works with human DNA. He tracks prehistoric people, then asks how their migration patterns fit – or don’t – the spread of prehistoric languages as proposed by historical linguists. Gray works with linguistic data, tracking languages as if they were biological organisms – as if they evolved like a virus. The data he uses are the words which, through careful sifting (guided by the sound laws), linguists have determined to have been inherited by a family of languages. Gray has trained an algorithm to detect degrees of similarity and difference across this inherited vocabulary, and the algorithm organises the languages into a family tree accordingly (actually a set of trees, ranked by probability). By cross-referencing the trees with historical texts where they are available, and archaeological and genetic data where they are not, he is able to estimate the chronological dates of important events in the family’s life (births, deaths, splits).
By 2022 Gray had been applying this method for twenty years, and the results had consistently confirmed Renfrew’s hypothesis: they pointed to a homeland of all the Indo-European languages in Anatolia around nine thousand years ago. This had earned him harsh, even withering criticism from linguists who found that hypothesis to be incompatible with their understanding of linguistic evolution (and who, in many cases, didn’t like the idea of languages being compared to viruses). It turns out that Gray, a New Zealander with sparkling blue eyes and a penchant for colourful shirts, has a sense of humour – he calls himself a ‘tree-hugger’ – and isn’t easily discouraged. He shrugged off the barbs and went back to work, using the same methods but new data.
There have been many attempts to gather the shared, inherited vocabulary of Indo-European languages in one place, but each of these dictionaries tends to be quite small. Since some of them are also quite old they contain errors that linguists have only identified as their methods have improved (words that shouldn’t have been included, for instance, because they are actually loans masquerading as inheritances). Realising that inconsistencies and poor-quality data could be distorting the trees, Gray’s colleague, linguist Paul Heggarty, set about building a new dictionary – or database – with eighty-odd collaborators. The result was much bigger than its predecessors, containing vocabulary from a hundred and sixty-one Indo-European languages – twice as many as Gray had started out with in the early 2000s. Some languages, such as Nuristani and extinct Gaulish, had never been included in a database before, and the vocabulary was selected according to strict criteria.7 It took Heggarty’s team five years to build, and when it was ready Gray used it to generate a new set of trees. These placed the homeland of the Indo-European languages, including Anatolian, south of the Caucasus – in the region of the Armenian highlands – around eight thousand years ago. In other words, they shifted it east and later than his previous estimates. Many linguists protested, though their protests were a little more muted than in the past, and there was praise for the new database.
When the dust settled, it became clear that a third hypothesis had entered the ring, to compete with Renfrew’s and Gimbutas’. Geneticists and at least some linguists were converging on a narrative according to which the parent of the Indo-European and Anatolian languages, the lingua obscura of the first chapter, was spoken in the Caucasus region between six and eight thousand years ago – between 4000 and 6000 BCE, that is.8 Then in 2024, the Harvard group claimed to have narrowed the birthplace of that obscure ancestor down again (and shifted it slightly northwards). They placed it in southern Russia at a time when Varna was still thriving – about a thousand years, that is, before the Yamnaya emerged.
The debate over the ultimate homeland was far from closed, but even those who disagreed with Reich acknowledged that, for the first time, he had demonstrated a genetic link between the steppe and Anatolia. The genetic, archaeological and linguistic evidence could be reconciled if Proto-Indo-European and Anatolian were sisters, as Sturtevant had argued a century before him. It wasn’t uncommon, in 2024, to hear scholars who had been working on the Indo-European puzzle for decades say how glad they were to be alive at a time when science was finally providing answers.
It’s a myth that you might recognise, since it exists in many Indo-European traditions: a fertility god flies into a rage and goes AWOL, plunging the cosmos into disorder until he or she is found and placated. In the Greek version, the goddess Demeter is so distressed at the abduction of her daughter Persephone by Hades, king of the underworld, that she lets winter descend on the world. In the Greek, Norse and Indic versions, horses or chariots fly over formidable obstacles to reach the errant god, but in the Hittite version the thing that flies is a bee. A goddess sends the bee to track down the aggrieved god Telipinu. Finding him asleep in a meadow, the bee stings him to wake him up. First bleary, then incandescent with rage, Telipinu goes on the rampage until another goddess calms him down.
Some mythologists suspect that the Hittites inherited the story of Telipinu from an older people who knew neither chariots nor horses as modes of transport. If Gray and Reich are right, those people might have herded goats and sheep in the Armenian highlands, or in what is now southern Russia, before finding their way to Anatolia and bringing the myth with them. The Yamnaya might have inherited the same myth, but they updated it with horses – since they had tamed them – before carrying it into Europe and greater Asia. Their descendants added chariots once those had been invented.
Thus language and mythology provide clues to the prehistory of peoples, in this case the Hittites. To take another example, Hittite has a word for ‘wheel’, h˘ūrkis, that has a different origin from either of its counterparts in Proto-Indo-European, *kwékwlos and *roteh2. That makes sense if Proto-Anatolian and Proto-Indo-European split from an older language spoken by people who did not know the wheel and so had no word for it. Each of the two daughter branches acquired their wheel words later, after they had separated. Logically, then, those words bore no relation to each other.
Not all the linguistic evidence supports a Caucasian origin, though. If carriers of Iranian farmer ancestry spoke the ancestral tongue, you’d expect that language to contain a good number of farming-related words, which it would have bequeathed to its daughters. Yet the reconstructed vocabularies of Proto-Anatolian and Proto-Indo-European share only one, possibly two words for an edible seed, neither of which distinguishes between a domesticated and a wild plant, and none at all for those early crops, lentils, peas and chickpeas.9 On the other hand, Proto-Anatolian and Proto-Indo-European do share a rich vocabulary of herding, including words for ‘horse’, ‘cow’, ‘sheep’, ‘yoke’ and ‘to pasture’. It looks more likely that the ancestral tongue was spoken by herders than farmers.
By way of a solution to this apparent discrepancy, it has been suggested that the Iranian farmers’ descendants had lingered so long in the north, in areas unsuitable for cultivation, that they had become herders again and forgotten all their crop-growing knowledge. That is possible. But what about the westerly clustering of the Anatolian languages? Isn’t it strange that they are absent in the east of the Turkish peninsula, if that is where they came from?
Not necessarily. For one thing, there is a possible trace of the first Anatolian-speakers’ passage through those eastern parts. Around 2400 BCE, palace archivists at Ebla in Syria recorded about twenty names, including Duduwashu and Aliwada, that sound distinctly Anatolian to some linguists’ ears. For another, archaeological evidence suggests that non-Indo-European-speaking groups in the Caucasus expanded at about the time that the ancestral tongue would have been fragmenting into its Indo-European and Anatolian daughters. Those non-Indo-Europeans could have pushed the first Anatolian-speakers gradually further west, until they ended up in the places where the Luwian, Lydian and Hittite kingdoms arose later. It is also possible that the first Anatolian-speakers reached the peninsula by boat from Yalta, Sochi or some other northern harbour – perhaps bringing silver to trade. The Roman historian Tacitus, writing around 100 CE, remarked that all ancient migrations came from the sea. He was probably thinking primarily of the Greeks who had colonised the Black Sea in the previous millennium, but his observation might hold for even earlier migrations.
Between 3000 BCE, when linguists say Proto-Anatolian was spoken, and the first written mention of the Hittites, a thousand years yawn blackly. A thousand years is a long time in human affairs; as a French historian once put it, it’s the gap between Gaul and de Gaulle. But one could speculate that in that time, whether they came from east or west, by land or by sea, the first speakers of an Anatolian language dispersed to different parts of the peninsula and adapted to the farming life. There need not have been very many of them to begin with, and each pocket of Anatolian-speakers would have found itself surrounded by other farming groups who spoke non-Indo-European languages such as Hattian or Hurrian (a language recorded in northern Mesopotamia at about this time). Anatolian fragmented into its daughters, and from then on the fortunes of those daughters tracked the fortunes of their speakers.
When we catch our first glimpse of the Hittites in the archives, around 2000 BCE, their stronghold was the city of Kanesh, in the spectacular volcanic landscape of Cappadocia. They never actually called themselves ‘Hittites’; that’s a misnomer we owe to the Bible. They were the people of Kanesh, who spoke Neshili. To the south-east of Kanesh lay the Taurus Mountains and beyond the mountains lay Assyria. Donkey caravans now came winding through those mountains bearing Assyrian merchants and their wares. By 2000 BCE, bronze was being made by adding tin rather than toxic arsenic to copper, and the Assyrians brought copper which they hoped to exchange for tin. Soon they had established a colony at Kanesh and were trading in all kinds of luxury goods, including gold, silver, perfumes and woven textiles. To keep track of their transactions they introduced writing to the region, in the form of cuneiform. The first texts that refer to the Hittites were written by these merchants in their Semitic language, Old Assyrian. Besides Hittite personal names, they contain a sprinkling of Hittite loanwords including ishpattallu (night watch) and ishiullu (contract).
Thanks in large part to the Assyrians, Kanesh grew rich. But the Assyrian trade collapsed around 1800 BCE and a turbulent period ensued, in which neighbouring city-states were constantly at war. To the north-west of Kanesh, inside a large bend in the River Kızılırmak, lay the Hattian-controlled city of Hattush. Five days’ gallop north of that, probably close to where the Kızılırmak tipped into the Black Sea, lay another powerful city called Zalpa where Hattian was also spoken. A Hittite king, Anittas, conquered Zalpa and Hattush and took control of the Kızılırmak Valley. Anittas favoured attacking his targets in the dark, which may be why the Assyrians of Kanesh felt the need to mount a night watch. The oldest surviving document written entirely in Hittite, in cuneiform, describes Anittas’ conquest of Hattush in the first person: ‘At night I took the city by force; I have sown weeds in its place. Should any king after me attempt to resettle Hattush, may the weather god of heaven strike him down.’
Eventually Hattushili I ascended to the throne. Apparently ignoring Anittas’ curse, or feeling that it didn’t apply to him, he moved the Hittite capital to Hattush. He tweaked the city’s name to Hattusha, signalling continuity under the new regime, and changed his own name to the one we know: Hattushili means ‘the one from Hattusha’. The royal archives were established under his reign, and they reveal that the Hittites co-opted the sophisticated political infrastructure that the Hattians had established, along with many Hattian gods and much of their culture. Among the Hattian loanwords that Hittite absorbed were a number that related to that older system of government: labarna (king), tawannanna (queen), halmassuit (throne), halentiu (palace) and sahtarili (singer-priest).
The two peoples fused, much as the Romans and Etruscans would do more than a thousand years later. Some scholars think of the Hittite founding myth, the one about the fertile Queen of Kanesh, as a retelling of the early history of the Hattians and Hittites that glossed the violence and presented them as a family reunited. The queen gave birth to her thirty sons, and the myth went on: ‘She said, “What a monster is this which I have borne?” She filled baskets with fat, put her sons in them, and launched them in the river. The river carried them to the sea to the land of Zalpa. But the gods took them up out of the sea and reared them.’10
The queen went on to have thirty daughters. This time she raised the royal offspring herself, at Kanesh, and when much later her estranged sons returned from Zalpa they met their sisters on the road without recognising them. Against the protestations of the youngest brother, the older boys slept with the older girls. To the eternal frustration of hittitologists, the clay tablet on which the myth is written breaks there and the end of the story is lost. They nevertheless speculate, based on similar myths recounted elsewhere, that the gods rewarded the youngest brother and sister for abstaining from incest, and that the two siblings went on to rule the joint kingdom wisely and well.
Hittite myth presented the two peoples as equal partners, in other words, but there are subtle linguistic clues that the Hittites retained the upper hand. Apart from the Hattian loans that their language absorbed, it changed very little. Hattian, on the other hand, was completely transformed by its entanglement with Hittite. In Hittite the verb was placed at the end of a sentence, in Hattian at the beginning. A comparison of texts written in the two languages shows that Hattian-speakers eventually switched to the Hittite word order, triggering a cascade of other changes in their language.
When you go to Boğazkale, as I did in the summer of 2023, you realise immediately why Hattushili wanted it for his capital. The ruins of the palace where he dictated his last words sit atop a basalt promontory that is protected on all sides by steep slopes or cliffs. Eagles nest in those cliffs, high above a tributary of the Kızılırmak that winds around their base. The lion king’s successors built towards their cumulative vision of an imperial capital which, for as long as it thrived, dwarfed Troy and Athens. Exploiting the fact that water expands when it freezes, Hittite masons split pre- chiselled rock and flattened crags. With the blocks they hewed out of the crags they constructed ramparts that, in their scale and geometrical precision, rivalled the Egyptian pyramids. They sank rock cisterns to a depth of twenty metres (sixty-five feet) and cut near-vertical escape chutes into the palace’s pedestal. The city had inner and outer defence walls, and the gates in the inner wall were shut and sealed each night. One was guarded by a pair of basalt sphinxes, another by lions. Their eyes are said to have flashed fire, perhaps because their now-hollow sockets contained red agates or jaspers. Hattusha was considered impregnable – until it wasn’t.
Hattushili I expanded his domains southwards into what is now Syria. His adopted son Murshili I, whom he named his successor in his will (‘The god will only install a lion in place of a lion’), pushed into Mesopotamia and sacked Babylon. The Hittites’ fortunes waned after that, but in the fourteenth century BCE a canny ruler named Shuppililiuma I ascended the throne and restored them. For the next two centuries the Hittites boasted one of the most powerful empires in the Near East. Their modus operandi seems to have been to treat the subjugated peoples with benevolence, as long as they came quietly, but to show no mercy to those who resisted. The empire absorbed many elements of the cultures of its new subjects, such that the Hittites came to be known as the ‘people of a thousand gods’. Some argue that Hattusha was cosmopolitan from as early as the reign of Hattushili I, with other Anatolian languages – notably Luwian and Palaic – being spoken within its walls.
In 1274 BCE, in what has sometimes been called the first world war because there had never yet been such an impressive turnout of infantry and cavalry, the Hittites and the Egyptians fought each other at Kadesh on the banks of the River Orontes, near the modern Syrian city of Homs. The Hittites brought their souped-up chariots (they had refined an earlier design by shifting the axle from the back to the middle), drawn by horses wearing feathered headdresses. The battle ended in a stalemate, with heavy casualties on both sides, though the Egyptian Pharaoh, Ramses II (Ramses ‘the Great’), went home claiming victory: ‘I slaughtered them. I killed them wherever they were.’
It took sixteen years, but Ramses eventually signed a peace treaty with the Hittite emperor Hattushili III and his queen Puduhepa. In a text inscribed on silver tablets, two of the world’s superpowers promised to ‘settle forever among them a good peace and a good fraternity’. The language of the treaty was Akkadian, the medium of international diplomacy at the time. The original tablets have been lost, but the text was engraved into the walls of Egyptian temples, and a Hittite version, written on clay, was stored in the archives at Hattusha. A replica of the Hittite version hangs in the United Nations headquarters in New York City, outside the Security Council chamber on the second floor.
Around 1200 BCE, not long after Puduhepa’s death, Hattusha fell, the victim of a wave of collapse that reverberated around the Near East and brought the Bronze Age to a close. Traditionally this collapse was blamed on the Sea Peoples, ruthless pirates so named by the Egyptians, whose origins were veiled in mystery. Historical and scientific research has pieced together a more complex tableau. The trouble seems to have started with a drought that caused a severe and lasting famine in the region. A storm of earthquakes rippling along the faultlines that criss-cross the eastern Mediterranean forced starving, desperate migrants to take to the sea, and maritime trade stalled. Eventually, the staggeringly unequal civilisations of Egypt, Anatolia and Greece, heirs of the Uruk model of urbanism, faced plagues and peasant revolts that not even their towering ramparts could keep out. The Hittite capital had only ever been home to the royal family, bureaucrats, soldiers and priests. The peasants who supplied the food (including more than a hundred and eighty kinds of bread or bakery product, according to a museum in nearby Çorum) lived beyond the walls, shut out from the machinery of state and the wealth that oiled it.
Hattusha burned and the Hittite language, soon to pass into oblivion, was baked hard into the clay. Some claim that it was already dead by the time of the crisis, having been replaced as the city’s everyday language by Luwian. Luwian survived for another five hundred years, while Lydian was still being spoken a few centuries before the birth of Christ. At some point in the first millennium of our era, however, the last Anatolian language fell silent, never to be heard again.
