Many societies have a caste system in which one's social and economic roles and political status are fixed. The caste is usually inherited from one's parents, but may be based on occupation, wealth, or one's ethnic group; over time, as occupations change, ethnic groups intermingle, or lower caste individuals are promoted by reason of wealth or military service, the basis will change but will usually be just an accident of birth. Often there will be just two castes (African examples include the Tutsi and Hutu of Rwanda; in medieval Japan the Samurai and peasants form two castes with Samurai having the right to kill peasants); or there may be several castes (In Hawaii, the four castes were nobility, master craftsmen (incl. priests, healers, boat builders, etc.), commoners, and slaves).
Before the modern era, much of Europe used a three-caste (three-estate) system, comprising clergy, nobility, and commoners. Commoners might be further subdivided based on debts or occupation. Royalty (on top) and serfs or slaves (on the bottom) might be considered to be outside the caste system. In some societies nobility outranked clergy; in other societies clergy outranked nobility. In the latter case, nobles might be subdivided into warriors and lawmakers, with lawmakers in the same highest caste as clergy. England's "caste system" was more flexible: a nobleman's oldest son would be noble, second son perhaps a cleric, remaining sons merely "gentlemen", or promoted to knights if they showed military excellence. Below the gentry, commoners were treated as different castes based largely on wealth or occupation, rather than birth.
It is the Hindu caste system we are concerned with here. The three highest castes are Brahmins (teachers and priests), Kshatriyas (warriors and rulers), Vaishyas (farmers and merchants). A fourth class, the Sudras (laborers and craftsmen), had a lower status; later came the Panchama (e.g. "untouchables") who were termed "out-castes" rather than a fifth caste. Eventually many subcastes developed based on occupation and ethnicity.
Because the original Hindu caste system comprised just three castes, membership in the additional lower castes must have arisen by punishment or conquest. The same three-caste ordering (priests, warriors, farmers) is observed in many ancient European cultures, perhaps most clearly among Romans and Celts. Druids, the highest caste of Celts, were subdivided into prophets, bards and priests. Political authority was sometimes exercised by druids, sometimes by the warrior caste. Even the third caste enjoyed some freedom and religious status, with slaves often from a different ethnic group and having a fourth (out-caste) status.
In the late 18th century, strong similarities were first noted between European languages and Sanskrit, the language of ancient sacred Hindu religious texts. Thus the Indo-European language family was recognized, which became the major impetus to the development of historical linguistics.
All language families with wide ranges result from cultural expansions, many of which have left a mark in archaeological records. The broad range of Arabic results from Islamic conquests 1500 years ago; the Polynesian expansion, dependent on sea-faring and agricultural skills, spread its language all the way from Madagascar to Easter Island; the Roman Empire spread its language throughout much of Europe, and later the conquistadores spread it to Latin America; the Bantu culture had agriculture and metallurgy far more advanced than other African cultures, so its language spread over most of the continent; and so on.
But what about the Indo-European language family? It dominates both Europe and India, yet there's no clear archaeological evidence of any migrations or cultural expansions, let alone clear reason to associate them with a particular language. (The Roman Empire played a role, but many of its conquests had already spoken Celtic, itself an Indo-European language. The development of I-E probably occurred during the Advanced Neolithic period between the Late Stone Age and Bronze Age. Cheese, beer, leather, wagons and plows were among the "Advanced Products" of this important era, which might thus be called Dairy Age, or Copper Age, especially since copper axles were key to wagon building.)
Among many theories of I-E origin, three most heard today are:
It may spoil the suspense to say so at once, but most scholars now accept the third possibility. There is an easy argument for this based on Occam's razor, though I've never seen it in print: Had the language made both an early farmer-to-pastoralist transition and a pastoralist-to-Bronze Age farmer transition, it would have undergone much more change than is seen when Sanskrit is compared with European languages. Instead, the language originated with the semi-nomadic pastoralists and somehow expanded to farming villages across a broad range. There is plenty of linguistic evidence to support this, particularly the large number of reconstructed I-E words related to stockbreeding (especially horses) and wheels. Early neolithic cultures lacked wheels and much experience with cows and horses; I-E couldn't have split up much before 4000 BC, so the split up occurred during what may loosely be called the "Copper Age." (The Balkan Hypothesis may seem compatible with this timing, but the details can't be worked out, while the facts fit the Kurgan Hypothesis like a glove fits a hand.)
Many Indian scholars think the I-E languages originated in India, but European scholars regard them as crackpots. But it is also far-fetched to imagine, as some European linguists do, that the language began in the villages near Greece or in the Danube basin, spread to the horse breeders of the East European steppes, and then Southwards to India. Even if we pretend that a huge nomadic culture could adopt an alien language (for which there are very few, if any, precedents) the language would have made both adoptive village-to-pasture and adoptive pasture-to-village transitions, one of which would have surely involved imperfect learning. Yet Sanskrit has preserved the rich inflectional system of the ancient Indo-European language better than any other living specimen.
It is significant to note that the Cimmerians, Scythians, Slavs, Huns, Magyars, Mongols and Turks are all cultures which projected power from the Eurasian steppes to civilizations neighboring the steppes, but the converse power projection never happened. In none of the invasions just noted did the steppe people change their language. In only two cases -- the Slavs and Magyars -- did the invaders leave their language in Central Europe; however all these historic invasions were by barbarians, with Central Europe already settled with the relics of the Roman Empire. On the other hand, the conjectured "Kurgan wave 2" occurred almost 3000 years before the founding of Rome; that invading culture had more expertise than the Balko-Danubian cultures in stockbreeding, metallurgy, tanning, etc.; and, although we will not develop the argument here, the Kurgan religion and socioeconomic organization easily enabled them to have a prestige status. (That prestige languages can wholly displace an autochthonous language is too well-known to contest; indeed, since I-E now dominates in Western Europe, Russia and Northern India, at least two such major displacements did take place, the only question is where and when.)
The Pontic-Caspian stockbreeders began to dominate its immediate neighbors in East Central Europe and Central Asia by sheer immigration, but its presence was felt in Greece, Central Europe mainly through diluted invasions, imitations and "political" influence.
Thus the earliest Indo-European speakers were stockbreeders in the Pontic-Caspian region, e.g. the the Khvalynsk and/or Sredny Stog cultures. (These "Kurgan" cultures are often described as "semi-nomadic" but the term encompasses dozens of cultures, many predating wagons and horse riding.) By the late 4th millennium BC, the Indo-European languages had split into at least six major dialect groups:
The Pit Grave and Afanasievo cultures of Eurasia were still part of the "Kurgan Horizon"; the four European groups to the west, whose origins must have involved both infiltration and imitation, retained some vestiges of Kurgan material culture. (The extent and significance of such similarities is highly controversial. Balto-Slavic may be a later arrival from Yamnaya, or (less likely?) may be Eastern Corded Ware.)
During the 3rd millennium, Pit Grave developed into various daughter cultures; by the end of that millennium the principal Kurgan cultures included Timber Grave near the Black Sea (prob. the Cimmerians who may have spoken Thracian or an Iranian language), Andronovo near the Caspian Sea (who spoke Indo-Iranian and were soon to invent the war chariot), and the Afanasievo still farther East who had separated much earlier. The Indo-Iranian language split at some point into the Nuristani (now found only in the Hindu Kush mountains), the Indo-Aryans of South Asia and the Iranians who remained in Central Asia (some later followed the Aryans southward). The Indo-Aryans developed the Hinduist religion; the Iranians Zoroasterist, but similarities between these religions and the religions of Greeks, Celts and Norse show that there was an earlier Kurgan religion from which the later religions derive.
The notion that the Indo-European Homeland was among the 5th millennium BC horse breeders of the East European steppes is almost universally accepted by scholars open-minded enough to synthesize all evidence: linguistic and archaeological. Many linguists have a dogmatic aversion to non-linguistic evidence, and often apply Sapir's observation that places a Homeland at the point of greatest diversity. This implies a Balkan homeland for Indo-European. But this over-simplified view ignores geography! Just as a candle fire gives most heat not at its source, but above as hot air rises, so the rapid movements by semi-nomads in the Russian steppes don't lead to diversity there, but to their West, where emigrants encounter sedentary populations.
Linguistic diversity depends on factors like terrain: diversity in the Caucasus mountains, for example, implies persistence, not origin. Diversity is unlikely in open terrain like the steppes. The ultimate origin of Indo-European is an unsettled question, and Sapir's dogma may play a role in settling it when a macro-family including Indo-European is identified, but regardless of its origin, Proto-Indo-European flourished in the Pontic-Caspian steppes 5000 to 7000 years ago, as the "Kurgan" horse-breeding culture prospered.
The ancient Celts, Greeks, Hittites and Latins all had a caste system similar to that of the Hindus; the similarity extended even to the symbolic colors (white for priests, red for warriors, black or blue for other castes). All ancient Indo-European people seem to have sacrificed three different animals at their rituals; originally these may have been a ram for the priests, a horse for the warriors, and a sheep for workers, but substitutions were made in daughter cultures that did not preserve all three of these species. (Greeks, for example, substituted a bull for the horse, and a boar for the sheep. To suggest this means the tripartite rituals are unrelated, however, would be to ignore that there were variations in sacrificial species even between closely related Iranian tribes.)
Beyond the caste system, there are many other very strong connections between Hindu culture and the culture of the ancient Celtic Irish. Druids trained for many years, memorized sacred poems, and practiced meditation -- just as did the Vedic Brahmins. The Celts had a fire sacrifice ritual, with druid priests chanting while spices and other foods were cast into a burning pit; ancient Vedic Brahmin practiced an almost identical ritual. The Hindu rite asvamedha, which culminates when a holy stallion is divided into three sacrificial portions, has similarities to ancient Celtic and Latin rites; Indeed the very word asvamedha suggests an etymology from proto-Indo-European with a meaning like drunk on horse and mead, a meaning far more relevant to the Irish ritual than the Hindu one. And this just scratches the surface of uncanny parallels between the law and cultures of the Vedic Hindus with ancient Indo-European speakers in Europe.
There appear to be several cognates between mythic Hindu names and those of Europe, beginning with the words for god itself: deus (Latin), devas (Sanskrit). The Indic rain god Parjanyas seems to be cognate with the Norse name for Thor's mother Fjorgyn, and so on.
Hindu India is notorious for the practice of suttee (in which a living widow is cremated along with her husband), but archaeological evidence for suttee has been found in the Sredny Stog and Pit Grave cultures of the prehistoric East European steppes, as well as in Central European, Balkan, and even Italian sites (Rinaldone) believed to be related to Kurgan expansion.
Historians often emphasize that the Roman Empire was never really "Christianized", rather the Christian religion became Romanized and may bear little relation to what can be gleaned of the religion of Jesus' early followers. The Romanized Christian religion, like ancient Indo-European religions, emphasizes a "Trinity" although the only clear reference to a Trinity in the Gospels is Matthew 28:19. (This line is at the very end of that book and thus is likely to be a late addition.)
The Y-chromosome, passed strictly father-to-son, can shed much light on prehistory. This is especially true in cultures like India, where a male-inherited caste system has been dominant for more than 3000 years. Note that a population that is 1% Brahmin will become 50% Brahmin after 100 generations assuming that Brahmins have a procreative advantage of just 4.8% per generation.
The distribution of the R1a1 Y-chromosome Haplogroup, and in particular its correlation with the Brahmin caste in India, provides strong confirmation of a theory deriving Hinduism from Central Asia. It is no secret that there are strong racial differences among the different castes in India, but many studies conflict and confuse. Part of the problem is that the castes are classified differently by different authors. (For example, some lump Kshatriya and Vaishya with Brahmin as "high caste"; others lump these with Sudra and Panchama into a non-Brahmin group; either of these mask the strong haplogroup-caste affiliations shown in the table below. Unfortunately, I've seen no study which distinguishes Kshatriya and Vaishya; it would be nice to know if these have different Y Haplogroup affiliations.)
R1a is the dominant Y-chromosome haplogroup among Slavic speakers, the group now dominating the ancient homeland of Indo-European. The only other group among which R1a is particularly prominent is the Hindu Brahmins of India!
Here is a table, based on Table 6 in an often-cited paper by Sengupta, et al, ("Polarity and Temporality of High-Resolution Y-chromosome Distributions in India Identify Both Indigenous and Exogenous and Reveal Minor Genetic Influence of Central Asian Pastoralists"), that breaks down a sample population by caste, language, and Y-chromosome haplogroup. This is the largest such haplogroup/caste study I'm aware of, but please note that Sengupta, et al do not claim it to be statistically valid, and any inferences should be viewed with caution. However the data need not be considered "hand-picked": Sengupta derives conclusions opposite to those I will derive!
Sengupta's paper doesn't distinguish Kshatriya and Vaishya, combining the two into "Middle caste", and doesn't even mention Panchama groups like the "untouchables." To save space I've combined R2, J2a and G2 haplogroups (though broken them out in "Grand Totals"). "Other" haplogroups include O2a (106), O3 (61), F* (38), C (13), misc (11). Numbers in parentheses show counts when only the 205 Indo-European speakers are considered. All "caste" individuals were classified as speaking either an I-E or a Dravidian language. Of the "tribal" individuals, 21 spoke Indo-European, 180 spoke Dravidian, 87 spoke Tibeto-Burman, and 64 spoke Mundari (or Khasian). Other facts of note not visible in my summarized table are: (a) The 9 G2 individuals included 7 Dravidian-speaking Brahmin; (b) O2a is strongest among Mundari speakers in East India; (c) O3 is strongest among Tibeto-Burman speakers in North India.
| - | R1a1 | R2/J2a/G2 | H*/H1/H2 | L1/L3 | J2b | Other | Total | |||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Brahmin ("High caste") | 56 (39) | 44 (22) | 17 (12) | 14 (4) | 6 (4) | 8 (5) | 145 (86)
| Kshatriya & Vaishya
| 15 (5)
| 17 (11)
| 47 (18)
| 19 (3)
| 20 (4)
| 15 (7)
| | 133 (48)
| Sudra ("Low caste")
| 20 (13)
| 13 (8)
| 27 (22)
| 6 (1)
| 4 (1)
| 9 (5)
| | 79 (50)
| Tribal
| 13 (4)
| 24 (0)
| 101 (7)
| 10 (0)
| 7 (0)
| 197 (10)
| | 352 (21)
|
| Total caste
| 91 (57)
| 74 (41)
| 91 (52)
| 39 (8)
| 30 (9)
| 32 (17)
| | 357 (184)
| Grand total
| 104 (61)
| 62/27/9 (28/12/1)
| 192 (59)
| 49 (8)
| 37 (9)
| 229 (27)
| | 709 (205)
| | |||||||||
I did find one study (specific to the Jaunpur District of Northern India) that classifies individuals into the five major castes. "Other" haplogroups include C5 (5), P* (5), misc (5). Note the strong correlations between caste and haplogroup. Also note that the Panchama ("untouchables") are drawn from the same groups as Brahmin, suggesting their ancestors were "demoted" for some reason. (This effect will obscure the caste-haplogroup correlation in studies which use conflated caste assignments.)
| - | R1a1 | R2 | H1 | J/L/F* | Other | Total | |
|---|---|---|---|---|---|---|---|
| Brahmin | 5 | 1 | 13 | 1 | 0 | 20 | |
| Kshatriya | 3 | 41 | 3 | 0 | 0 | 47 | |
| Vaishya | 9 | 7 | 8 | 10 | 5 | 39 | |
| Sudra | 2 | 1 | 0 | 0 | 4 | 7 | |
| Panchama | 9 | 1 | 9 | 3 | 6 | 28 | |
| Total | 28 | 51 | 33 | 8/3/3 | 15 | 141 | |
In the Sengupta sample, 54% of R1a men were Brahmin and 39% of Brahmin were R1a. (These percentages become 68% and 45% when only Indo-European speaking caste members are considered.) An Indo-European-speaking R1a man is eight times as likely to be Brahmin as Middle caste, while an H-haplogroup man is almost three times as likely to be Middle caste as to be Brahmin.
Caste follows the male Y-chromosome (in mixed-caste marriages, children take their father's caste), but still the strong correlation shown in the table seems very significant. The caste system must have been introduced to India fairly recently, the initial partitioning must have been based on obvious ethnic differences; and indeed, as we will see, the caste system must have been imposed at the time of invasion or not much later. It may be interesting that among non-Brahmin R1a men, the lower caste is more common then middle castes. This suggests that the middle castes were a reward for specific ethnic groups and unavailable for R1a's who were punished or otherwise excluded from being Brahmin.
As suggested by the second study, caste-haplogroup correlation may vary greatly from province to province; this is another important source of confusion.
Promiscuity and inter-group marriages will eventually lead to a blending of genetic types in an integrated population unless there are strict social rules. Since the differentiation into haplogroups occurred in the Early Stone Age, it can be deduced that the caste system which separates Brahmin from the lower castes was imposed shortly after these groups came into contact. That H and L are ancient haplogroups native to India is not disputed; the newcomer must have been R1a. (The rarity in India of other West Asian haplogroups like J2a and G (and also E, I, N, etc.) is sometimes taken as evidence against an "Aryan Invasion;" but all it shows is that the Invasion was by a specific culture (presumably sharing a common ancestor with Slavs, who are predominantly R1a).
The inverse migration, northward from India to the East European steppes, is clearly impossible given the Y-chromosome distribution. Not only would we have to imagine India's R1a Brahmin mounting an expedition without their non-R1a warriors (Kshatriya caste), but they went without their own Brahmin cousins with haplogroups like R2 and H. And, after somehow "conquering" the steppes without their warriors and Brahmin cousins, why didn't they subjugate the indigenous non-R1a steppe peoples into lower castes, rather than annihilating them? To me, the notion is too absurd to be anything but a joke; only a political correctness allows it into the technical literature. (Hitler made reference to the Aryan invasion, so serious archaeologists are now called Aryan-supremacists if they deny the possibility of an Indocentric origin of Indo-European or R1a.)
The connection between the Hindu Brahmin caste and the R1a genetic haplogroup is unmistakable; to give it meaning we must briefly review the distribution of relevant haplogroups. The H and L haplogroups are ancient and specific to India. C and F* haplogroups are widespread in Asia and even more ancient. O2a and O3 are somewhat more recent haplogroups, widespread in Asia, with O3 related specifically to recent Sino-Tibetan intrusions. J2 is the haplogroup of the early Fertile-Crescent farmers; details of how it got to India are conjectural, but the caste distinction between J2a and J2b suggests there might have been at least two unrelated migrations. G2 is a slightly mysterious haplogroup, conjectured to have originated near Pakistan about 20,000 years ago, and which is now most common in secluded regions like the Caucasus and Sardinia.
This leaves the R haplogroup, perhaps the most interesting and important of all Y-chromosome types. Its main subtype, R1, is associated with Cro-Magnon Man, who migrated from Central Asia to Europe about 35,000 years ago. There are five subgroups of R; these are R*, a rarish type noted for a (small) occurrence among Australian aborigines; R2, a rarish type with a distribution not unlike G2; R1b*, a rarish type with an African presence strongest in the Cameroons and among Chadic-language speakers; R1b1, which dominates Western Europe; and our focus, R1a. Before continuing, observe that R1 and R2 separated 25-35,000 years ago, traveling North and South respectively. It would be a gross misconception to treat the R1a and R2 groups of South Asia as sibling.
Any map of Y-chromosome haplogroup distributions will show that R1a is now primarily associated with the Slavic people of Eastern and Central Europe, along with some of their immediate neighbors like Balts and Magyars. (There is also an unexplained R1a presence in Norway, which projects to Iceland and northern Britain as the so-called "Viking" gene.) R1a also has a strong presence among Iranian peoples, though doubtless the haplogroup frequency was diluted as Iranians migrated from their homeland (the present Slavic domain). And finally, it has an unusually strong prevalence in South Asia, especially among the Brahmins of Northern India.
R1a is often called "the Indo-European haplogroup" but this is at least somewhat misleading: besides the Slavs and Balts, R1a is quite rare in traditional Indo-European places like Italy, Spain, Ireland and France; yet it is common in some non-IE-speaking East Eurasian tribes. R1a has only modest presence in Germany and Greece is less than 10% R1a. In Greece and Italy the principal haplogroups are J2, E3b, I1b and R1b1. (We've already met J2 and R1b1; E3b is the "Afroasiatic" haplogroup and I1b is native to Southern Europe.)
If the case made above is convincing, you may wonder why other Indo-European peoples, like the Greeks and Celts, do not exhibit high R1a levels. Part of the explanation may be a simple difference in chronology. India's Brahmin caste was founded by steppe people about 1800 BC, when their caste system was well developed and R1a dominant. The Greeks separated from the Indo-European Homeland about 2000 years earlier, so the situation may have been very different.
Another explanation may be more important. It is only in India that the Indo-European caste system survives to the present-day. Perhaps haplogroup study in the ancient Usatovo or Mycenaean culture would have shown the upper class to be largely R1a, but such a differentiation will not survive for long in the presence of any social or political upheaval.
The Celts have almost no R1a at all. Instead they have a huge amount of R1b-L11. We must suppose either that an R1b founder "tagged along" with the Globular Amphora or Corded Ware push, and, by chance, prevailed in the West; or seek the origin of R1b-L11 in some other way.
The Norse people, on the other hand, have a strong R1a component (the "Viking gene"). There is no generally accepted explanation of origin for this Norse R1a, but perhaps there was an ancient migration, similar to the "Aryan invasion" of India, which brought a new religion, along with the R1a haplogroup, to Scandinavia. In one theory, this was as recently as 5th century AD, when Sarmatian or Hunnic warriors accompanied Goths returning to Thule (as reported in 553 AD by the historian Procopius). This theory may be far-fetched, but strong phonetic similarities between the names of Norse and Hindu Gods (as well as the Odin myth itself) suggest that the Norse religion was a relatively recent import from Asia.
it is often observed that there is no "smoking gun" evidence for any theory of Indo-European origin. However, what evidence does exist strongly favors the Gimbutist position. Gandhara Grave and related cultures are in just the right time and place, and have just the right cultural motifs to derive I-E speaking Vedics from the steppes. The eponymous artifact of the Corded Ware culture, almost unanimously accepted as a key component of Western I-E, was anticipated by the use of cords for pottery decorations in the early Pit Grave culture. There are no specific archaeological sites linking Pit Grave with the Afanasievo culture over a thousand miles to its East, nor have any wooden wheels survived from early Afanasievo, but there is strong reason to think that the wheel (possibly a key invention helping I-E dominance) was known in Afanasievo: while wood decays rapidly, toy wheeled wagons made of ceramics have survived from early Afanasievo sites.
Balkanists and Indocentrists dismiss this circumstantial evidence, but offer none of their own in return. A mysterious migration from Central Europe to form the proto-Tocharians of Asia is left unexplained by Balkanists, despite that it must have somehow bypassed the emerging steppe culture. Indocentrists have more severe problems and need to introduce prehistoric northward migrations, several in fact, since they derive I-E subfamilies like Celtic, Anatolian, Baltic from different stages of an Indo-Aryan proto-language!
Despite the clear evidence deriving the Indo-European language and the R1a Haplogroup from the steppes of Eurasia, many "academic" papers claim an opposite conclusion. The Indocentric theory of Indo-European origin has attracted little or no support among competent linguists and archaeologists, but naturally the DNA-typing of Indians is done by Indian researchers, many of whom seem to have a bias which distorts their conclusions and even data analysis methods. In order to deny any "Aryan migration", these Indologists reverse the direction of migration, claiming that both R1a and the Indo-European language originated in India and migrated northward.
If researchers were objective, there would be no need for this webpage: journal articles drawing conclusions about R1a chronology would be objective and draw the same conclusions I've drawn. Instead, however, the articles are full of fallacy and I write this webpage to set the record straight. In the remainder of this section I'll discuss some of the fallacies in the Indian Y-chromosome literature.
Skill at electrophoresis work is irrelevant to the common sense required in interpreting DNA evidence. Skill in the operation of statistical software packages is of no value when inappropriate statistical measures are chosen. To clarify this latter point, imagine a simple hypothetical scenario, in which Population A is all R1a, Population B is all R2, and Population C is a mixture of R1a and L haplogroups. The "statistical distance" between Populations A and B may be identical to the distance between Populations A and C, but to represent that "fact" as a useful summary while suppressing the actual Haplogroup information would be absurd!
Statistical methods like component analysis are of great value in uncovering the patterns when data involves many characters, independent or of uncertain correlation. However, the strict inheritance of Y-chromosome characters means they are fully dependent. When researchers subject such data to methods like component analysis, one can only wonder whether they've lost all sense of perspective in their indulgence of statistical toys, or whether, more simply, they use any method, however illogical, which helps them argue towards their politically desired preconception.
It is well-known that female-inherited (mtDNA) and male inherited (Y-chromosome) genetic characters follow totally different patterns and, as explained above, it was specifically a small band of male adventurers that brought elements of the steppe culture to India, and that it was specifically the procreative advantage of high-prestige Brahmin males that led to the strong presence of R1a Haplogroup in India. Yet, some papers focusing on daughter-inherited mtDNA studies draw broad conclusions denying the possibility of an infiltration by males!
It is easy to imagine that, once R1a was established in India, some R1a individuals (especially bastards) would lose any special status and join out-caste or tribal groups. Thus it is specifically the genetic contrasts between Brahmins, Kshatriyas, and Vaishyas which will be most informative about the caste system origin. Yet many research papers miss this point entirely, either lumping all castes together and comparing with "tribals", or lumping Brahmins and Kshatriyas together to compare groups of castes. Since it is hard to imagine an Indian conflating Brahmin and Kshatriya out of simple ignorance, the peculiar conflation must be a deliberate ploy to obscure the clear genetic signatures in the caste system.
In the absence of a caste system, the genetic mixture of a geographic group will soon become homogeneous. Once one admits the strong Y-chromosome differentiation between Indian castes, it is clear that the impositions of caste systems were coincident with migrations or infiltrations. No one disagrees that Y-chromosome haplogroups H and L originated in India 30,000 years ago or more; it is therefore other groups, especially R1a that came from the outside and into contact with H and L at about the same time the Brahmin distinction arose in India. (This is not to diminish the possibility of a more complicated pre-history: it is likely that Hinduism's caste system was superimposed on a variety of earlier caste systems in the various regions of the subcontinent.)
Given two carafes, one of water, one of diluted wine, only common sense is needed to reconstruct the history and deduce which carafe was poured into which. Similarly, the strong concentration of R1a in the steppes makes it the probable source of other R1a loci. Indocentrists happily point out the presence of haplogroups like H among Brahmin, yet remain strangely silent about the nearly complete absence of H in the West Eurasian cultures they derive from India (as well as the oddity that the Brahmins didn't bring their Kshatriya warriors along to help with their conquests).
Indocentrists make much of the high STR diversity among Indian R1a's. What this shows is that these haplogroups have not suffered recent "bottlenecks." The geographic diversity of India means that relatively isolated R1a pockets were established early in the Vedic culture, and have had a hundred generations to diverge. This is in contrast to the highly mobile terrain of Slavic people; prehistoric and historic Slavic warlords have surely left their genetic signature, reducing measured diversity. (That such diversity reduction is common-place, especially in the steppes, is well-known: it is believed that millions of males carry specifically Genghis Khan's Y-chromosome. Similarly, the Y-chromosome of the 4th-century Irish Emperor Nial Mor Naoighiallach is believed to now be ubiquitous in Ireland.)
The present writer has scoured the Internet for genetic studies of India and has yet to find a single one that presents its data in the only logical way: individual triplets of caste, birthplace or language, and Y-haplogroup, I'd love to be proved wrong and directed to a valid study but, as is, we can only conclude that such data would be too distressing to Indian researchers.
Although the details are a mystery, the R1a haplogroup of the Brahmin caste confirms an "Aryan invasion" which brought Indo-European language and religion to India. Other genes will reveal other facets of human prehistory: for example, Andrew Lancaster has been able to correlate the dispersion of the E1b1b haplogroup, probably from the Horn of Africa, with the dispersion of language families in the Erythraean (Afroasiatic) group.