The Proto-Indo-European homeland was the place where the Proto-Indo-European language (PIE) was first spoken. This language later split into different dialects, which eventually became the earliest Indo-European languages.
The most widely accepted idea about where PIE was spoken is called the steppe hypothesis. It suggests that PIE was spoken in the Pontic–Caspian steppe region around 4000 BCE. Another possibility, called the Armenian hypothesis, places the homeland of the early PIE language (known as "Indo-Hittite") south of the Caucasus mountains. This idea has gained more attention in recent years because of research on ancient DNA. A third idea, the Anatolian hypothesis, suggests PIE was spoken in Anatolia around 8000 BCE. Other theories, such as the North European hypothesis, the Neolithic creolisation hypothesis, the Paleolithic continuity paradigm, the Arctic theory, and the "indigenous Aryans" (or "out of India") hypothesis, have also been proposed. However, these are not widely accepted and are considered less supported by evidence.
The search for the Proto-Indo-European homeland began in the late 18th century after the discovery of the Indo-European language family. Researchers have used methods from fields such as historical linguistics, archaeology, physical anthropology, and human population genetics to study this topic.
Hypotheses
The steppe model, the Anatolian model, and the Near Eastern (or Armenian) model are the three main theories about where the Proto-Indo-European (PIE) language originated. The steppe model suggests that PIE speakers lived in the Pontic-Caspian steppe around 4000 BCE. Most scholars today support this theory because it is backed by evidence from language studies, archaeology, and genetics.
Linguist Allan R. Bomhard (2019) explains that the steppe model, first proposed by archaeologists Marija Gimbutas and David W. Anthony, is supported by both linguistic and archaeological findings. These findings include cultural groups in the steppe region between 4500 and 3500 BCE. While many other theories, like the Anatolian model, are no longer widely accepted, some questions remain about how early forms of Greek, Armenian, Albanian, Celtic, and Anatolian languages spread. These questions are still studied within the context of the steppe model.
Another theory, the Near Eastern model, also called the Armenian hypothesis, was proposed in the 1980s by linguists Tamaz V. Gamkrelidze and Vyacheslav Ivanov. This theory connects Indo-European languages to Caucasian languages and is based on some disputed linguistic theories and archaeological findings. Recent DNA research has led to new ideas about a possible homeland in the Caucasus or northwest Iran for an earlier version of PIE. However, other studies argue that PIE likely originated in the Eastern European/Eurasian steppe or from a mix of steppe and Caucasus languages. Some linguists believe that the spread of Anatolian languages was more likely through the Balkans than through the Caucasus, based on evidence from language diversity and geography.
The Anatolian model, proposed by archaeologist Colin Renfrew, suggests that the homeland of early PIE was in Anatolia around 8000 BCE, with PIE proper later moving to the Balkans around 5000 BCE. This theory links the spread of Indo-European languages to the movement of farming across Europe. However, this model faces challenges because the timeline it proposes does not match linguistic or genetic evidence. For example, certain words related to wheeled vehicles appear in many Indo-European languages but not in Anatolian languages, suggesting that PIE speakers had vehicles before Anatolian speakers did. Additionally, genetic studies do not find evidence of Anatolian origins in the Indian gene pool.
DNA evidence from ancient and modern people shows that farmers from Anatolia did spread across Europe starting around 6500 BCE, mixing with local hunter-gatherer populations. However, around 2500 BCE, a large group of pastoralists from the steppe region near the Black Sea, linked to the Corded Ware culture, moved into Europe. These people contributed significantly to the ancestry of many Europeans today, including Norwegians, Lithuanians, and Estonians, who have nearly half their ancestry from this group. Genetic studies also show that steppe ancestry is common in speakers of Indo-European languages in India, especially in the Y chromosome.
In general, the spread of a language or dialect often depends on access to valuable resources. For Indo-European speakers, this resource was likely horse-based pastoralism, which gave them an advantage over groups focused on farming.
Other theories about the Indo-European homeland exist, but most are not widely accepted by scholars today.
Theoretical considerations
Traditionally, the homelands of linguistic families are identified using evidence from comparative linguistics, which studies similarities and differences between languages, along with archaeological findings about past populations and migrations. Today, genetic research using DNA samples is increasingly used to study how ancient people moved across regions.
By comparing languages, experts can reconstruct the vocabulary of a proto-language, which helps them understand the culture, technology, and environment of the people who spoke it. This information can then be compared with archaeological evidence. For example, in the case of Proto-Indo-European (PIE), which is based on later languages like Anatolian and Tocharian, reconstructed vocabulary includes words related to daily life and trade.
Zsolt Simon explains that while reconstructed vocabulary can help estimate when PIE was spoken, it may not always accurately show where its speakers lived. This is because people might have used certain words not because they were part of their own environment, but because they learned them from other groups they interacted with.
Proto-Finno-Ugric and PIE share some words, often related to trade, such as terms for "price" and "draw, lead." Similarly, words like "sell" and "wash" were borrowed into Proto-Ugric. While some researchers suggest these languages might have a common ancestor (the hypothetical Indo-Uralic family), most experts believe this is due to frequent borrowing, which implies their homelands were near each other. PIE also shows borrowed words from Caucasian languages, like Proto-Northwest Caucasian and Proto-Kartvelian, suggesting PIE speakers lived near the Caucasus Mountains.
Gamkrelidze and Ivanov once proposed that PIE borrowed words from Semitic languages, which might explain a southern homeland. However, Mallory and Adams argue that some of these borrowings may be uncertain or from later periods. They believe certain words, like táwros ("bull") and wéyh₁on- ("wine; vine"), are more likely examples of borrowing.
Anthony suggests that the few accepted Semitic loanwords in PIE, such as terms for "bull" and "silver," could have been acquired through trade and migration routes rather than direct contact with Semitic-speaking people.
According to Anthony, the Anatolian languages were the first Indo-European group to split from the main language family. Because Anatolian languages preserve many ancient features, they may be considered a "cousin" of PIE rather than a direct descendant, though they are generally seen as an early branch of the Indo-European family.
The Indo-Hittite hypothesis suggests that Anatolian languages and other Indo-European languages share a common ancestor called Indo-Hittite or Indo-Anatolian. However, this theory is not widely accepted, and there is little evidence to support reconstructing a distinct proto-Indo-Hittite stage that differs significantly from PIE.
Anthony (2019) proposes that PIE originated mainly from languages spoken by Eastern European hunter-gatherers in the Volga steppes. It also shows influences from languages spoken by northern Caucasus hunter-gatherers who moved to the lower Volga region. Later, there may have been a smaller influence from the Maikop culture, located to the south, which is believed to have spoken a North Caucasian language. This influence likely occurred during the Neolithic or Bronze Age but had little genetic impact.
Phylogenetic analyses of languages
Lexico-statistical studies, which aim to show how different branches of Indo-European languages are related, began in the late 20th century with research by Dyen et al. (1992) and Ringe et al. (2002). Later, several researchers used a method called Bayesian phylogenetic analysis (a mathematical approach used in evolutionary biology to determine relationships between species) to study Indo-European languages. These studies also tried to estimate when the different branches of these languages separated from one another.
Earlier research suggested that the development of these branches took place over a long period of time. A study by Bouckaert and others, which included geography, strongly supported Anatolia as the origin of the Indo-European languages. This finding supported Colin Renfrew’s idea that Indo-European languages spread from Anatolia along with the development of agriculture between 7500 and 6000 BCE. According to their analysis, the five major groups of Indo-European languages—Celtic, Germanic, Italic, Balto-Slavic, and Indo-Iranian—became distinct between 4000 and 2000 BCE. The researchers noted that this timeline fits with later events, such as the movement of steppe peoples after 3000 BCE, which they believe also helped spread Indo-European languages.
Steppe hypothesis
The Steppe Hypothesis tries to find the origin of the Indo-European language expansion, which led to many migrations from the Pontic–Caspian steppe between the 5th and 3rd millennia BCE. In the early 1980s, most experts agreed with the "Kurgan hypothesis," named after burial mounds called kurgans in the Eurasian steppes. This theory placed the Indo-European homeland in the Pontic–Caspian steppe during the Chalcolithic period.
According to the Kurgan hypothesis, as described by Gimbutas, Indo-European-speaking nomads from Eastern Ukraine and Southern Russia expanded on horseback in several waves during the 3rd millennium BCE. These groups invaded and took control of peaceful European Neolithic farmers in what Gimbutas called "Old Europe." Later versions of her theory focused more on the patriarchal and patrilineal nature of the invading culture, contrasting it with the supposedly egalitarian and matrilineal culture of the people they conquered.
Archaeologist J. P. Mallory dated these migrations to about 4000 BCE and reduced the emphasis on violence, making his version of the theory more compatible with a less gender-focused explanation. David Anthony, focusing on evidence of horse domestication and wheeled vehicles, believed the Yamnaya culture, which replaced the Sredny Stog culture around 3500 BCE, was most likely the Proto-Indo-European speech community.
Anthony described how cattle-raising spread from early farmers in the Danube Valley into the Ukrainian steppes between the 6th and 5th millennia BCE, creating a cultural boundary with hunter-gatherers whose languages may have included early Proto-Indo-European. He noted that domesticated cattle and sheep probably did not come into the steppes from Transcaucasia because farming communities there were not widespread and were separated from the steppes by the glaciated Caucasus Mountains. Later cultures in the area, like the Cucuteni-Trypillian culture, adopted cattle.
Indologist Asko Parpola believed the Cucuteni-Trypillian culture was the birthplace of wheeled vehicles and the homeland for Late Proto-Indo-European, assuming Early Proto-Indo-European was spoken by Skelya pastoralists (early Sredny Stog culture) who took over the Tripillia culture around 4300–4000 BCE. The Sredny Stog culture (4400–3400 BCE) had origins linked to "people from the east, perhaps from the Volga steppes." It played a major role in Gimbutas’s Kurgan hypothesis and coincided with the spread of early Proto-Indo-European across the steppes and into the Danube Valley around 4000 BCE, ending the era of Old Europe. After this, the Maykop culture began, Tripillia towns grew, and people from the eastern steppes migrated to the Altai Mountains, founding the Afanasevo culture (3300–2500 BCE).
A key part of the Steppe Hypothesis is identifying Proto-Indo-European culture as a nomadic pastoralist society that did not practice intensive agriculture. This is supported by the fact that words related to cows, horses, and wheeled vehicles can be reconstructed for all Indo-European branches, while only a few agricultural terms are reconstructable. This suggests agriculture was adopted gradually through contact with non-Indo-Europeans. If this evidence is accepted, finding the Proto-Indo-European culture requires searching for the earliest introduction of domesticated horses and wagons into Europe.
Proponents of the Anatolian Hypothesis, Russell Gray and Quentin Atkinson, argue that different branches of Indo-European languages might have developed similar vocabulary independently, creating the illusion of shared inheritance. They also suggest that words related to wheeled vehicles might have been borrowed later. Supporters of the Steppe Hypothesis disagree, saying this would violate established principles for interpreting linguistic data.
Another piece of evidence for the Steppe Hypothesis is the presence of shared loanwords between Uralic languages and Proto-Indo-European, suggesting these languages were spoken in nearby areas. This would have occurred much farther north than Anatolian or Near Eastern scenarios allow. Kortlandt argues that Indo-Uralic is the common ancestor of both Indo-European and Uralic languages. He claims Indo-European was a branch of Indo-Uralic that changed significantly when its speakers moved from the area north of the Caspian Sea to the area north of the Black Sea. Anthony notes that deep language relationships are hard to prove due to time depth, and similarities may result from borrowings from Proto-Indo-European into Proto-Uralic. He also points out that North Caucasian communities were part of the steppe world.
Kloekhorst argues that Anatolian languages have preserved archaisms also found in Proto-Uralic, supporting a steppe origin for Proto-Indo-European.
The subclade R1a1a (R-M17 or R-M198) is most commonly associated with Indo-European speakers. In 2000, Ornella Semino and others proposed that this haplogroup spread from north of the Black Sea during the Late Glacial Maximum, later expanding with the Kurgan culture into Europe.
In 2015, a study by Haak et al. found evidence of a "massive migration" from the Pontic-Caspian steppe to Central Europe around 4,500 years ago. The study showed that people from the Corded Ware culture (3rd millennium BCE) in Central Europe were genetically similar to those from the Yamnaya culture. The authors concluded this supports the theory that at least some Indo-European languages in Europe originated from the steppe.
Two other 2015 genetic studies supported the Steppe Hypothesis. They found that specific subclades of Y chromosome haplogroups R1b and R1a, common in Yamnaya and other early Indo-European cultures like Sredny Stog and Khvalynsk, spread from the Ukrainian and Russian steppes along with Indo-European languages. These studies also identified an autosomal component in modern Europeans not present in Neolithic Europeans, introduced with R1b and R1a lineages and Indo-European languages.
However, the "folk-migration" model is not the only explanation for all language families. The Yamnaya ancestry component is concentrated in Europe’s northwestern regions, and models for languages like Proto-Greek are still debated. The steppe genetic component is less common in Mycenaean populations, suggesting Proto-Greek speakers may have been a minority among agricultural societies. Some argue Proto-Greek gained prominence through cultural influence by elites. While genetics and language often correlate, archaeologists note that such migrations may not fully explain the spread of Indo-European languages or archaeological cultures.
Anatolian hypothesis
The main competitor to the Kurgan hypothesis is the Anatolian hypothesis, proposed by Colin Renfrew in 1987. This theory connects the spread of Indo-European languages to the movement of farming from the Near East during the Neolithic period. It suggests that Indo-European languages began spreading peacefully into Europe from Asia Minor (modern-day Turkey) around 7000 BCE, along with the spread of farming. The expansion of agriculture from the Middle East is believed to have influenced the spread of three language families: Indo-European to Europe, Dravidian to Pakistan and India, and Afro-Asiatic to Arabia and North Africa.
Renfrew (2004) revised his original idea after facing criticism. He later suggested that the earliest form of the Indo-European language, called Pre-Proto-Indo-European, originated in Anatolia around 7000 BCE. However, he placed the homeland of Proto-Indo-European itself in the Balkans around 5000 BCE, a culture known as the "Old European culture" proposed by Marija Gimbutas. Renfrew still believed that the roots of the Indo-European language family were in Anatolia around 7000 BCE. However, evidence from the Bronze Age, such as words related to "wheel," may not apply to the Anatolian branch of the language, which may have separated from Proto-Indo-European before wheeled vehicles were invented.
After studies on ancient DNA were published in 2015, Renfrew acknowledged that people speaking Indo-European languages likely migrated from the Pontic steppe to Northwestern Europe.
A major criticism of the Anatolian hypothesis is that it requires an unrealistically early date. Linguistic analysis shows that the Proto-Indo-European language includes words for inventions and practices linked to the Secondary Products Revolution, which occurred after the early spread of farming. Based on this, Proto-Indo-European cannot be older than 4000 BCE. Additionally, some scholars argue that certain language similarities, like the verb "to be" in Hittite and Sanskrit, may not have survived over the long time span the Anatolian hypothesis suggests.
The idea that farming spread from Anatolia in one wave has been updated. Instead, farming appears to have moved in multiple waves through different routes, mainly from the Levant. Evidence from domesticated plants suggests an early movement by sea from the Levant, while the overland route through Anatolia was most important for spreading farming into Southeast Europe.
According to Lazaridis et al. (2016), farming developed separately in the Levant and the eastern Fertile Crescent. These regions later interacted, and the chalcolithic population in northwest Iran was a mix of Neolithic farmers from Iran, Levant people, and Caucasus hunter-gatherers. The study notes that farmers from Iran spread northward into the Eurasian steppe, and people related to both Iranian farmers and steppe pastoralists moved eastward into South Asia. They also explain that Ancestral North Indians (ANI) have genetic links to both early Iranian farmers and Bronze Age steppe people, making it unlikely that Indo-European languages in India originated from Anatolia.
Alberto Piazza stated that "genetically, people from the Kurgan steppe descended partly from Middle Eastern Neolithic people who came from Anatolia." Piazza and Cavalli-Sforza suggested that the Yamnaya culture may have developed from Middle Eastern Neolithic farmers who moved to the Pontic steppe and adopted pastoral nomadism.
If farming spread from Anatolia around 9,500 years ago and the Yamnaya culture emerged 6,000 years ago, it would have taken about 3,500 years for people to migrate from Anatolia to the Volga-Don region, likely through the Balkans. In this region, a new pastoral culture developed due to environmental conditions unsuitable for traditional farming but favorable for herding. Their hypothesis is that Indo-European languages spread later from the Yamnaya culture region, possibly after Neolithic farmers from Anatolia settled there and adopted nomadic lifestyles.
Wells agrees with Cavalli-Sforza that there is "some genetic evidence for migration from the Middle East." While there is strong genetic and archaeological evidence for an Indo-European migration from the southern Russian steppes, there is little evidence for a large-scale migration from the Middle East to Europe. One possibility is that earlier migrations (8,000 years ago) may have left genetic signals that have since spread over time. Although there is some genetic evidence for migration from the Middle East, it is not strong enough to fully trace the spread of Neolithic languages across all Indo-European-speaking regions in Europe.
Southern archaic PIE-homeland hypothesis
Different ideas have been suggested about where the ancient Proto-Indo-European (PIE) language originated. Some believe it came from the Eurasian or Eastern European steppe, others think it began in the Caucasus region to the south, and some suggest it had a mix of influences from both areas.
Gamkrelidze and Ivanov proposed that the original PIE homeland was south of the Caucasus, specifically in eastern Anatolia, the southern Caucasus, and northern Mesopotamia during the 5th to 4th millennia BCE. They based this on a theory about certain sounds in PIE called glottal consonants. They also pointed to PIE words that suggest contact with more advanced cultures to the south, the presence of Semitic and Kartvelian language influences in PIE, and possible connections to Sumerian and Elamite languages. However, because the glottal theory was not widely accepted and there was little archaeological evidence, their idea was not widely supported until Renfrew’s Anatolian theory later revived parts of their proposal.
Gamkrelidze and Ivanov also suggested that Greeks moved west across Anatolia to their current location, that some Indo-European speakers moved north and came into contact with Finno-Ugric languages, and that the Kurgan area (or the Black Sea and Volga steppe) was a secondary homeland from which western Indo-European languages spread.
Recent DNA studies show that people from the steppe had a mix of ancestry from Eastern Hunter-Gatherers (EHG) and Caucasus Hunter-Gatherers (CHG). This has led to new ideas that the original PIE homeland might have been in the Caucasus or even in Iran, the common ancestor of both Anatolian languages and all other Indo-European languages. Some researchers argue this could support the Indo-Hittite hypothesis, which suggests that proto-Anatolian and proto-Indo-European languages split from a shared language no later than the 4th millennium BCE. These ideas have been discussed by several scientists, including Haak et al. (2015), Reich (2018), Damgaard (2018), Wang et al. (2019), Grolle (2018), Krause & Trappe (2021), and Lazaridis et al. (2022).
Damgaard et al. (2018) found that people from Anatolia during the Copper and Bronze Ages had CHG ancestry but no EHG ancestry. They concluded that steppe populations did not contribute to the ancestry of Anatolians during this time, suggesting that the spread of Indo-European languages into Anatolia was not linked to large migrations from the steppe. They noted that this could mean Indo-European languages arrived in Anatolia through small-scale movements and trade, rather than major migrations. They also mentioned that many linguists think the Balkan region was a more likely route for Indo-European languages to reach Anatolia than the Caucasus.
Wang et al. (2019) found that the Caucasus and the steppes were genetically separate in the 4th millennium BCE. However, the Caucasus acted as a corridor for genetic exchange between cultures south of the Caucasus and the Maykop culture during the Copper and Bronze Ages. This suggests the possibility that PIE originated south of the Caucasus, which could explain the early split of Anatolian languages from other Indo-European languages. They proposed that the mix of EHG and CHG ancestry found in steppe populations might have come from a natural genetic gradient or from CHG-related ancestry reaching the steppes independently before the arrival of Anatolian farmer ancestry. They argued that this genetic evidence supports the idea that PIE could have originated south of the Caucasus and spread northward with CHG ancestry, explaining the early separation of Anatolian languages.
Lazaridis et al. (2022) stated that genetic evidence supports the possibility that Proto-Indo-European originated either in the steppe or in the south (the "Southern Arc"), but they argue the evidence points to the latter. They described the Southern Arc as an area including Anatolia, North Mesopotamia, Western Iran, Armenia, Azerbaijan, and the Caucasus. They suggest that Proto-Indo-European may have emerged in this region and spread to Anatolia through the movement of people with Caucasus/Levantine ancestry after the Neolithic period. This would have separated Proto-Anatolian from the rest of Indo-European languages. They also proposed that later migrations from the Southern Arc brought Proto-Indo-European to the steppes. They noted that the spread of non-Anatolian Indo-European languages is linked to the Yamnaya pastoralists and related groups, but not to steppe migrations because ancient Anatolians lacked EHG ancestry. They emphasized that further research is needed to identify the exact sources of these movements.
Bomhard’s (2017, 2019) "Caucasian substrate hypothesis" suggests that the original homeland of Indo-Uralic (a proposed common ancestor of Indo-European and Uralic languages) was in Central Asia or the North Caspian region of the steppe. He expanded on earlier ideas, proposing that a Eurasiatic language, which included a Northwest Caucasian language, influenced the development of Proto-Indo-European.
David Anthony (2019), an expert on Indo-European languages and anthropology, criticized the Southern/Caucasian homeland hypothesis, including ideas from Reich, Kristiansen, and Wang. He argued that Proto-Indo-European likely originated mainly from languages spoken by Eastern European hunter-gatherers, with some influence from Caucasus hunter-gatherers. He rejected the idea that the Bronze Age Maykop people of the Caucasus were a source of Indo-European languages or genetics. Anthony noted that the Yamnaya people’s ancestry included European farmer components, not Maykop ancestry, and that their paternal lineages were more closely related to Eastern European hunter-gatherers than to Caucasus populations.
Other hypotheses
Lothar Kilian and Marek Zvelebil suggested that the Indo-European (IE) languages began in Northern Europe around 6000 BCE or later. They believe this happened when early Neolithic farmers moved to northern Europe and blended with original Mesolithic hunter-gatherer groups. The steppe theory agrees with the idea that the PIE homeland was larger, as the "Neolithic creolisation hypothesis" allows the Pontic-Caspian region to have been part of PIE territory.
Fringe theories
The Paleolithic continuity theory, also called the "Paleolithic Continuity Paradigm" by Mario Alinei, its main supporter, is an idea that suggests the Proto-Indo-European language (PIE) existed during the Upper Paleolithic period, which was thousands of years before the Chalcolithic or Neolithic periods, as other theories suggest. This theory is considered unlikely by linguists because it relies on the assumption that there were no major changes in the people living in Europe since the Last Glacial Maximum, as shown by genetic and archaeological evidence.
In 1997, Mallory did not include this theory in his list of widely discussed and accepted ideas about the origins of the Indo-European languages.
Soviet scholar Natalia R. Guseva and Soviet ethnographer S. V. Zharnikova, influenced by Bal Gangadhar Tilak’s 1903 book The Arctic Home in the Vedas, proposed that the Indo-Aryan and Slavic peoples originated in the Arctic region near the northern Urals. These ideas were later promoted by Russian nationalists.
The Indigenous Aryans theory, also known as the "out of India" theory, claims that the Indo-European languages originated in India. Languages such as Hindi and Sanskrit, spoken in northern India and Pakistan, are part of the Indo-Aryan branch of the Indo-European language family. The Steppe model, which describes the movement of people from the Eurasian Steppe as an "Aryan invasion," has been criticized by Hindu revivalists and nationalists. These groups argue that the Aryans were originally from India. Some scholars, including B. B. Lal, Koenraad Elst, and Shrikant Talageri, suggest that the Proto-Indo-European language itself may have originated in northern India, either at the same time as or before the Indus Valley civilization. However, mainstream scholars do not consider this "out of India" theory to be credible.