New Genetic Analysis Sheds Light on Origins of Indian Castes


Editor’s note 4/10/2017: This post has been slightly edited from the original after a reader pointed out that it inferred something regarding Dravidian speakers that wasn’t explicitly stated in the paper. For as long as humans have lived in complex communities, cities and civilizations, they have divided and classified their societies. Those divisions have been based on age, gender, appearance or – in many cases – occupation. In many traditional societies artisans would share the same social status; as would soldiers, priests and workers in any number of other occupations.

In antiquity, the status of a family rarely changed. If you were a farmer, your sons would be farmers, and so on. While today social status barriers are crumbling in many societies, in others they remain largely unchanged.

India’s complex social stratification, known as the caste system, has been one of the traditional cornerstones of society. Though urban Indians are shedding the caste labels of their parents and grandparents, many rural Indians – who make up 72% of the entire population – hold steadfast to the system. In small villages and towns, the Brahmin caste – consisting of scholars and priests – is still revered as one of the highest social strata. And members of the Dalit caste – formerly known as “Untouchables” – are still viewed as unclean and remain separated from others.

The rigidity of the system still present in rural India has made many wonder exactly how long castes have existed. Historical records are unclear, as early Hindu scriptures like the Bhagavad Gita are somewhat ambiguous when it comes to the topic. Some historians even propose that the caste system as we know it today is largely a construct of the English Colonial Era, arguing that the development of such a system could have been deemed necessary to instill order.

Genetic analysis has also proven inconclusive, as analysis of small segments of the human genome has yielded different results. But a new study by geneticist David Reich and colleagues, published in the September 24 issue of Nature, takes a new approach to understanding the genetic history of India.

The core difference between Reich’s genetic analysis and previous studies is in the sheer amount of genetic material analyzed. Reich’s team examined more than 550,000 points across all segments of the human genome. In doing so, they hoped to obtain a more complete picture of Indian genetic history.

The research team analyzed the DNA of 132 individuals from India and neighboring regions, dividing them into 25 distinct groups based on geography, caste and language. They calculated how genetically ‘closed’ each of these groups were. In the caste system it is rare to marry someone from another class, making caste societies very closed, or ‘endogamous.’ If this endogamy continues over many generations, it will leave a behind a genetic signature for scientists to discover.

Reich and his team found such a signature, indicating a long history of endogamy in several of the groups. In fact, the research team calculated that the DNA of six of the groups can be traced back to just a few individuals who lived anywhere from 30 to more than 100 generations ago. Assuming a generation time of 25 years, that establishes the existence of the caste system in the range of 750 to more than 2,500 years ago — long before the British colonial era.

In a second analysis, Reich and his team examined how ancient migrations could have influenced the formation of castes. First the researchers divided the Indian groups into language families: Indo-European and Dravidian. Dravidian tongues, like Tamil and Malayalam, are mainly spoken in southern India and are believed to be a remnant of languages spoken by some of the earliest inhabitants of the region. Indo-European languages, like Punjabi and Urdu, are more common in the north. They are believed to have arrived with a migration of farmers from southwestern Asia or the Near East about 9,000 years ago.

Reich and his colleagues then compared the genetics of each of the Dravidian and Indo-European groups to a sample of European DNA. The team reasoned that, if Indo-European groups were really descended from the farmers, they would show more genetic similarity to the Europeans than the Dravidians.

Not surprisingly, the authors’ hypothesis held true. The Indo-European speakers, like the Kashmiri Pandit and Vaish, were more genetically similar to Europeans. And because the majority of the upper castes speak Indo-European languages, there could be a relationship between the arrival of Indo-European people and the formation of caste structure. Further evidence that an ancient caste system has permeated through India for thousands of years.

  • Vasishta

    Regarding what is stated in the last paragraph, “And because the majority of the upper castes speak Indo-European languages, while the lower ones tend to be Dravidian speakers”..

    ..The vast majority of Indians, are Indo-European speakers regardless of caste. This essentially points to the fact that a pre-Indo European population adopted the culture and language of these Indo-European warrior-herders who later became the twice born/dvija OR upper caste Hindus.

    What was exclusive to upper caste Hindus until a while ago was the privilege of Sanskrit and Vedic learning. It was only the Brahmins, Kshatriyas and Vaishyas, the Aryan twice born castes who were historically allowed to carry out their own Homa, chant the Gayatri Mantra (an invocation of sorts) and being admitted to Ashramas, where they were thought the Vedas, Upanishads and other sacred texts alongside training in physical pursuits by credible teachers, mainly sages.

    Varna as in, Dharmashastra is unique to the Indo-Europeans of South Asia. So essentially, Aryans, as a people in Hindu society are those of high castes alone.The twice born castes initially divided themselves into specific castes and professions based on the characteristics of cosmic energy. These Aryans divided themselves based on their qualities. Thus, the intelligent and spiritually oriented became the Brahmins, the chivalrious and strong became Kshatriyas (warriors) and the skillful and industrious became Vaishyas (traders).

    Regardless everyone in India is mixed to a certain extent, since the area was historically a melting pot of several races. Nothing is beyond human nature, not even a form of stratification that makes endogamy a system, can stop humans from mixing. However, most data points to the Upper Castes being the least mixed, alongside having an overwhelming West-Eurasian (“Caucasoid”) predominance.

    As an aside, it is highly unfortunate that 23andMe doesn’t deliver to India, considering that Indians are an under studied population. You’re leaving out 1 Billion potential customers you know, which also bring about avenue for further, non-mainstream research as in projects like Dodecad, Eurogenes, etc.

    • Hi Vasishta,

      Thanks for your comment. Regarding delivery to India, every country must go through legal and regulatory clearance before we can add it to the list of countries we can ship to. When we can add it, we will!

      • Ranjit

        Hi Shwu,

        Kindly investigate the histroy from superior histrorians who has deep study in that, as the information which vasistha had given to you, is not truth may be its his opinion but not a fact. Aryans, Portugeez, Dutch, Mughala’s, paris’s are not origin nativers of india they are outsiders as the same has DNA analysis. kindly confirm more Thanks…..

  • nareeman

    “the upper castes speak Indo-European languages, while the lower ones tend to be Dravidian speakers?” What nonsense! People of the same group, upper or lower castes, speaks the same language.

    • Sudeep

      Agreed. The author is provoking European ideals of complexion. Dravidian people have colors of all and have different caste within the same language. So im calling Bullsh!t on your research upon caste with this! [Apologize for the late comment, but ran across this article only recently 02/2013]

    • Bhasky

      When the author says “the upper castes speak Indo-European languages”, I think the author means the group that migrated into India before the actual mixing. Remember the caste system wasn’t prevalent till this group moved in to India. The “upper & lower” terminology is used in this context. The idea of caste based on “intelligence and spirituality” as proposed by Vasishta in the first comment is baseless as you see people of varying intelligence in all sections of society. After the migration and social stratification and continued mixing inspite of the stratification, we have people of all classes speaking the same language in a particular region. The idea that caste existed for as long a 9000 years is quite astounding. In fact it puts to rest the theory that British introduced it into India.

    • Gunzo

      nareeman, you are incorrect. Brahmins in Tamil Nadu may speak Tamil and practically it is their mother tongue (and I may mind myself, it is also soul for a lot of them); However Sanskrit is also a soul for them. For Non Brahmins in Tamil Nadu, Sanskrit is not a soul definitely. So the author is correct in saying that upper castes speak a different language than lower castes in at least considering Brahmin / No Brahmin as Upper / Lower castes.

      • Steve Bakewell

        That also applies to all non-brahmins in North India too.. and honestly all low-caste people speak is a colloquial version of original tamil.

        • Gunzo

          No. Kodunthamil (Kodum Thamizh) was recognized even in Tholkappiar’s times. Diglossia is allowed in Tamil. Yes, people who are not educated do not speak the language well and this is really true in USA, England as well. Anyway, my aunt used to tell so many proverbs and allegories in Tamil and she had never been to school even one day. Some of them were applicable to the hardships and emotions of hardworking womenfolk! I wish I recorded these!
          Well anyway, I think the author did not convey his sense correctly. I think he meant to say “Upper caste people were associated with Sanskrit while other castes were not”.

        • Steve Bakewell

          Kodunthamil is now officially known as “Malayalam”

          “I think the author did not convey his sense correctly”

          No he completely stopped making sense when he equated dravidian language group with low-caste he shouldn’t have said that! most north indians speak indo-European languages regardless of their caste and most south indians speak dravidian languages regardless of their caste.

          The author seems to be desperately trying to validate his Aryan-invasion theory which is nothing but fiction.

  • Radhakrishnan

    Excellent work. But we Indian are missing. Do U have plans to open in India.

  • DKen0

    Most charitable thing I can say about this article is that it is a piece of garbage. The article claims to be based on the study in Nature which starts out with North vs South: ANI vs ASI. Somehow author AnneH managed to completely miss the basic dichotomy and turned it into a distorted article about castes. She is in “good” company with the likes of Max Muller who fabricated the Aryan Invasion Theory (AIT) to justify British imperial rule.

  • Janice

    Anne H needs to go back and read the article from Nature properly. The point discussed was completely different ANI and ASI. ASI people are native andaman and Indian ocean (islands) inhabitants, who are not the ones called as Dravidian. I am shocked that 23and me would allow such an inaccurate piece of garbage accepted as an article here. Anne H please go back and read the article and then write about your opinions. It looks like you just read the abstract and rewrote based on your ‘opinion’ and not not science. I am truly shocked such inaccuracies and opinions are being passed off as science!!

  • Gunzo

    Man, this article is very correct. This study exactly says the same thing which we knew in reading the olden day Sanskritan literature, customs, laws and regulations etc. This study shows the genetic make up of higher castes to be the new comers into India from the north west.

    Some kind of castes had been established even during Rig Vedic times. The modern day pundits and RSS and Sanskritans argue that there were no “castes” in Rig Vedic times and society was divided as per “qualities” rather than “color”. I would accept this if Rig Veda had specified as to how these “qualities” were measured, at what age, what examinations (for example something like Tofel, GRE, UPSC, SAT kind of qualifying exams!). This rig Vedic system had completely degraded to a very harsh caste situation in the society by the time Manu Smriti was written formalizing what was there in the society. I mind you Manu Smirti did not divide the population, it only codified what was in practice at that time.

    Any intermixing in the eons was due to wars between kingdoms and soldiers letting loose themselves. Also some people in the fringe of the castes will always comingle. Also, it is not impossible for me to know that higher caste males copulated with lower caste females and left their children to grow up in the lower caste to fend for themselves. The aristocrats of course intermixed with other aristocrats even if they were Dravidian aritocrats! The genetic studies definitely prove what we see in the society even today namely higher castes speak or “Chant” Sanskrit while lower castes do not.

    I think that the idea that South Indians are dark and North Indians are fair is incorrect in some sense. I watched all the videos in You tube on un-touchability. Predominantly, the dalit people seem to be dark in color in NORTH India too. My own conclusion is that Vedas and Gita and all the Sanskrit literature and Puranas tell us the same story as the genetics where populations were enslaved and segregated out thoroughly in dehumanizing manner. I am not a Muslim or Christian or a Dalit. I am of South Indian origin born into a mid level caste.

  • Jimmy Smith

    Even the fairest Indian doesn’t look like any middle eastern so you’re lying because middle eastern people have semite features

  • Arcsabre

    Indo-European did not even exist in India (or anywhere else in the world) 9000 years ago what likely came to India in that timeframe was the migration of Dravidian / Proto-Dravidian (around the 6K – 3K B.C. timeframe) from the Near East, with the Indo-Europeans only coming much later after 2000 B.C. (atleast according to mainstream scholarship) and they initially spent the first few centuries as pastoral nomads; later organizing into tribes around 1500 B.C. and forming the Vedic Civilization. According to Reich’s own study the admixture between populations occurred somewhere between 4200 – 1900 years ago this puts it in the bracket of the arrival of the Indo Europeans to India and the Vedic Civilization (1500 – 500 B.C.) so the statement that the caste system has existed in India for ‘thousands of years’ differentiating the Indo Europeans from the Dravidians and other peoples, based on this study is a wonderful stretch of imagination.
    Moreover there is no historical evidence for a caste system in India for ‘thousands of years’ and if you did some googling to look up the wiki article on this you’d see that the viewpoint you are trying to state has been all but dismissed.
    Please research the facts before you post

  • Arcsabre

    The ‘truth’. How would you like if someone BSed about your background?

Return to top