Language dispersal based on subsistence economy and a possible candidate for a Nostratic founder population

When I was a linguistics student one of the most exciting ideas at the time was Colin Renfrew’s hypothesis of a farmer dispersal of the major language families in the world. This idea of agriculture-driven language spread made a lot of sense as many language families have their origin around 10.000 years ago or the subsequent millennia after the agricultural revolution. Many major language families, like Sino-Tibetan, were doubtlessly spread by farming. And the Indo-European languages seemed to fit this pattern.

However, the big surprise came when genetic studies increasingly confirmed that Indeo-European was really originally spread by a small herder population from the Russian steppe. Worldwide Indoeuropean languages are spoken by 45% of all people. How could a small population of pastoralists have been responsible for that? The answer lies in that they were a small population of fierce warriors who hijacked one farming population after another one, contributing extensively to the local gene pool and leaving their language (superstrate) as the dominant one. Of course, the further spread was caused by a mixed population of farmers (dominant) and herders. This surprise hadn’t been that much of a surprise, as the Kurgan (Yamnaya) — hypothesis for the origin of Indoeuropean had actually been put forward earlier by Marija Gimbutas, who also had understood that if her hypothesis was true the spread must have been extremely violent and certainly couldn’t have been peaceful.

Early Anatolian farmers did indeed spread languages to Europe. We know that they hardly mixed with the indigenous hunter-gatherer population in Southern Europe and almost completely replaced them. Indoeuropean languages, on the other hand, completely replaced their languages with the exception of Basque, probably the only remaining language spread by farmers in Europe. Basque is an isolated language nowadays. Its closest relatives are speculated to be Caucasian languages. The Indoeuropean Wusun pastoralists ventured even further east and probably maintained a semi-nomadic lifestyle until the 5th century A.D.

As herders are highly mobile compared to sedentary farmers there should be different dispersal patterns:

One little-known extinct Indoeuorpean language is Tocharian. They settled as far as in western China. Their language shows a very inconsistent origin for farming words. While the word for “plow” is Indoeuropean, the word for “seed” isn’t. They were most likely herders who adapted to the local farming culture.

The Yamnaya had doubtlessly great help from the horse and chariot in spreading their languages. What about other languages that were spread by pastoralism rather than agriculture (pastoralists can’t subsist on herding alone, therefore some relationship with farmers has always been necessary)? Turkic and Mongol are obvious candidates.

Semitic is a language family considered to have been spread by farming. I think this is wrong. Most of the early Semitic peoples were pastoralists rather than farmers. The Amorites, for example, lived in proximity to Sumerian farmers, often in symbiosis. They established a kingdom by capturing Mari, a Sumerian city. It doesn’t seem unreasonable that the original Semitic population were herders who went on to hijack pretty much all original farming populations. This basically means that none of the languages of the original farmers remains today. Indeed, this seems to be true. Sumerian, a language most likely spoken by the early hunter-gatherers turned farmers, went extinct around 2.000 BC. Fortunately, writing was invented prior to its extinction, so we know of its existence.

If the same holds true for the Afro-Asiatic language family in general, it could mean that Ancient Egyptian was not the language of the early Egyptian farmers, but the language of their pastoralist conquerors. As writing appeared later in Egypt, the ancestral farming language would be lost in this case.

Many researchers consider the Afroasiatic Urheimat to be located in Africa rather than the Levant. The Horn of Africa is a potential candidate as it shows high language diversity (and an indication of the age of the split). The Horn also contains the largest population of pastoralists in the world. Piracy is often a means of survival. Pastoralists have always been less likely to take up farming. Most African pastoralists consider farming below their dignity. The truth is probably that nomadic farmers and sedentary pastoralists evolved different temperaments that made changing the subsistence economy difficult. Pastoralists would find farming too cumbersome and boring and farmers would find pastoralism too unstable and adventurous.

This hypothesis sounds pretty wild, but if the Indeuorpeans could do it, why shouldn’t an Afroasiatic founder population of herders have done the same feat?

We get an interesting picture of a southern belt of herders (AA) and a northern belt (IE). And of course, herders would have been anywhere between farming communities where the land was less arable.

This opens up the possibility that southern herders were actually of ancestral stock to northern herders mixed with local farmer populations as well as hunter-gatherer populations (as we know from the Yamnaya). This could represent the historic development of the hypothetical Nostratic language family.

If this scenario is true, southern herders would first have subjugated Egyptian farmers, moved out of Africa, subjugated the farmers of the fertile crescent, and have been the founder population of the Yamnaya. Some languages in the hypothetical Nostratic language family would in this case not be part of the languages family but founded by the original forager to farmer populations. This would most certainly include Sumerian.

The origin of Nostratic is considered to have been around 15.000 years ago. This pre-agriculture date would certainly be too early for my hypothesis. Cattle were domesticated around 10.500 years ago and this could be a very realistic date for Proto-Nostratic, beginning with the first herder population. All cattle are descended from as few as 80 animals that were domesticated from wild ox in the Near East some 10,500 years ago, according to a new genetic study. In this case, the Nostratic Urheimat would be back again in the Near East rather than in Africa.

A 15.000-year origin of Nostratic in a single hunter-gatherer population seems highly unlikely. The languages in question cover a vast area and it could only have happened if a linguistically relatively homogeneous population of hunter-gatherers in that area had started to farm and herd. From all we know about hunter-gatherers they are very reluctant to take up farming and aren’t particularly good at herding either. So, if there is such a thing as a Nostratic founder population it probably started right there with the origin of pastoralism. With the passing of time pastoralists got more and more mixed with farmers and hunter-gatherers (a possible origin for social stratification in early civilizations), but their languages prevailed over forager and farmer languages.

For more information check out my book Foragers, Farmers and Pastoralists : How three tribes have been shaping civilization since the Neolithic

Originally published at on July 16, 2021.