This article is taken from the June 2025 issue of The Critic. To get the full magazine why not subscribe? Right now we’re offering five issues for just £25.
Addressing the Asiatic Society in Calcutta in 1786, Sir William Jones, the pre-eminent Oriental linguist of his age, declared:
“The Sanskrit language, whatever be its antiquity, is of a wonderful structure … bearing a stronger affinity [to Greek and Latin] … than could possibly have been produced by accident … No philologer could examine them all three, without believing them to have sprung from some common source which, perhaps, no longer exists.”
He noted similar derivations across the languages, as in the word “father”: pitar (Sanskrit), patēr (Greek), pater (Latin). Words for foods, animals and numerals also revealed other striking cognates.
Polyglot, poet and puisne judge, Jones had mastered Sanskrit so that he might better understand Indian customs and laws. The prima facie evidence for there being connective tissue between the Classical Western and Indian languages was persuasive.
He was amongst the first to advance in a compelling way the notion of a bond between languages spoken 4,000 miles apart. Errors in his analyses, however, resulted in an overestimation of the number of latter-day vernaculars connected to Sanskrit. As a consequence, he included Egyptian, Japanese and Mandarin, whilst omitting Hindustani and Slavic. Where to begin?
Most widely accepted is the so-called Steppe theory — that Proto-Indo-European (PIE) was spoken by nomadic herders between the Dnieper and Don rivers some 6,500 years ago. As with Sanskrit, its closest living relative, it was highly-inflected (think of man, men, man’s, men’s) and used vowel changes systematically (drink, drank, drunk; song, sing, sang, sung). The pastoralists who spoke it were rich in the vocabulary of goats, sheep and dogs but poor in that of crops and cultivation. In due course, farming, pottery and copperware entered the lexicon. Before writing’s invention, none of it was written down.
Five millennia ago, aridisation prompted PIE speakers to disperse west, east and northwards to the Balkans, Caspian Sea and Baltic coast. Some moved in wobbly ox-drawn carts, weighed down with pelts, salt and ceramics. Over time, waves of migrants pushed on into the Carpathians and Urals, adding branches to the Indo-European language tree. Vulgates and then other languages formed. By 2,000 BC they had made inroads as far west as Ireland.
A disquisition such as Proto: How One Ancient Language Went Global is necessarily hampered by the absence of data that precedes the invention of writing (c. 3,200 BC) and its patchy coverage thereafter. So, genetic science and archaeology are marshalled to support historical linguistics through triangulation: DNA profiling assists researchers in discovering melanin levels in corpses; burial rituals reveal to archaeologists the significance of animals, jewellery and weaponry, as well as associated myths and their vocabulary.
Specks of food from teeth and stomach remains yield data on diet. Ötzi the Iceman from the Tyrol dined on einkorn wheat, dried ibex and cooked deer for his last meal 5,300 years ago. He uttered his last words in pre-Italic Rhaetic.
Alcohol’s role in ritual, social bonding and language dispersal is of ancient standing. Although not detectable in the body three months after its consumption, the presence of grape and grain in archaeologically-retrieved vessels is a strong indicator of its consumption.
Two-handled cups were common in ancient Celtic, Greek, Roman, Baltic and Slavic cultures, with mead, wine, fermented mare’s milk, ale and spirits swilled from them. Toasts accompanied couplings, consecrations, coronations and (avant la lettre) christenings — bibe multis annis! This sharing of alcohol eased the exchange of foreign tongues, to be followed on occasion by the twining of genes between tribes and groups.
A good diet supports population growth, with languages fusing to habitable environments. Another reliable guideline is that the brute facts of geography can incubate or stunt linguistic evolution. Albania and the Basque country, for example, have vertiginous mountain ranges which have limited outsider incursions. Likewise, it was from the Baltic coastline, lakes and dense forests that Lithuanian and Latvian emerged, the former a highly-inflected “frozen vestige” of PIE.
Similarly, distance and isolation play a role in the development of accents. When a sub-group separates, its pronunciation will diverge over time because of a lack of enforcement of the established speech patterns.
How did PIE migrate from its heartland to India? Which cultures might it have influenced and been influenced by? A glance at the atlas suggests an eastward trajectory from the Steppe heartland via the Caucasus and Persia. Yet this exodus may never have happened.
Instead, according to the Loop conjecture, some of the original herders who had migrated north-east to the Urals later took a path south, through the ’stans and into India. Their language, Baltic-Slavic, had 55 farming words found to be shared with Indo-Iranian. Languages spoken on the “direct to India” route had no such correlation. DNA findings appear to support this claim of a more circuitous journey.
Another viewpoint, the “out of India” thesis, contends that the PIE homeland was in the north-west of the country, with the prevailing direction of travel westwards — not from the Steppe at all. Its appeal is predictably strong amongst Hindu nationalists. As the author observes: “At different times in different places, the Indo-European story has been beaten like warm copper to fit a political mould … Today, the story is still being beaten out of shape … but the beating is most energetic on the subcontinent.”
What of PIE itself? Linguists can use big-data harvesting techniques, surveying syntax trees to identify patterns in a prehistoric language and possible core features in its proto-language. Word lists based on “controlled speculation” containing around 2,000 words and stems have been reverse-engineered.
Some items are akin to broken mosaic tiles, of little value, but some appear to be closer to home e.g. swésōr (sister), hérmos (arm), dwo (two), pésk (fish), mēnōt (month/moon). PIE had eight grammatical cases and three genders, as does Sanskrit, and was highly evolved. Readers tempted to venture further still might like to consult Steven Mithen’s The Language Puzzle.
Laura Spinney is a science writer whose last book, on the Spanish flu, covered the pandemic that proliferated amongst the exhausted armies and civilian populations of 1918. Proto is an account of the debated origins of a language that mutated into a multitude of variants and strains now spoken by nearly half of the world’s population.
If you have a science hat, it will help you here. The author has rallied her sources and carried out her own field research. It is a very thorough undertaking. Even so, “the picture is patchy” in places, as she admits. Some of the story we know, then there is postulation, and the rest will remain unknown.
