Pathan/Pashtun Admixture Results

Someone asked for the individual HarappaWorld Admixture results for the Pathans and Pashtuns from HGDP (23) and Harappa Ancestry Project (3). So here they are.

You can check the spreadsheet too.


  1. Thank you very much! I really appreciate this. What I find interesting is that the "South Indian" component does not differentiate us HAP Pashtuns from the HGDP samples. Even HRP0286 is within the range of variation for the HGDP Pathans. The real difference/discontinuity lies in the "Baloch" component (obviously, HGD00220 does not count in this regard). A Pakistani Pashtun friend shared their results for DIY HarappaWorld with me, and their "South Indian" was within range (17%), but their "Baloch" was only 33%. I wonder if this finding would survive further sampling or participation. If so, the high "Baloch" component might perhaps be a local Bangash-Turi phenomenon (largest tribes in the Kurram Agency, although there is a good chance the tribal composition of the HGDP samples is predominately Bangash).

  2. This is awesome Zak, thanks a lot. It would be more interesting to take the Pashtun samples and match them Punjabi, Balochi, Sindhi etc. I've seen the combined admixture, however I am wondering how contrasting each population sample using a scatter graph tool or something to show distances. I may be dreaming here :). However, great job.

    • I think it is a safe bet most Pashtuns will be more similar to the HGDP Sindhi in their admixture proportions, rather than the HAP Punjabi participants. In the previous ChromoPainter/FineStructure analysis, Pashtuns clustered close to the Sindhi, rather than the Punjabis. I don't know why, as Pashtuns don't share a border with the Sindhi people, and have historical ties with Punjab. Perhaps the Balochi admixture amongst the Sindhi is rather consistent, pulling them further to the west in comparison to Punjabis. Whatever the reason, there seems to be a deep genetic similarity between Pashtuns and Sindhis. Also, I was browsing through the supplementary data for the Metspalu et al. paper. Pashtuns are roughly equisdistant to the Sindhi and Tajiks in terms of Fst. Both Tajiks and Sindhis are very close to the HGDP Pathans. Are the Yunusbayev/Metspalu Tajiks from Afghanistan? If not, perhaps Afghan Tajiks are the closest group to Pashtuns.

      • The Yunusbayev/Metspalu Tajiks are exclusively from Tajikistan. I have seen no Afghanistani Tajik ADMIXTURE result so far.

        • Very interesting. I would not have expected Tajiks from Tajikistan to be so genetically close to the HGDP Pashtuns (very low Fst distances). So, this means there is a good chance Afghan Tajiks are the closest population to Pashtuns. But perhaps I shouldn't be too hasty. From a cultural perspective, the Tajik category has no internal coherence. Pashtuns are heterogeneous, but Afghan Tajiks are something else. If asked about their ethnic background, they will respond by giving their native city or region. They are very different in this respect from Pashtuns, who tend to have a strong sense of ethnic identity (education also plays a role here). Perhaps they will exhibit differential affinities to Pashtuns, depending on their region, and depending on the Pashtuns being compared. It is a real shame to not have academic autosomal samples from Afghanistan.

          • I do not know the Fst distances between the Yunusbayev/Metspalu Tajikistani Tajiks and the HGDP Pashtuns, but the Yunusbayev/Metspalu Tajikistani Tajiks possess Mongoloid components in ratios considerably higher to those of Pashtuns in general and similar to those of the Yunusbayev/Metspalu Turkmens from Turkmenistan. In PCA and MDS plots the Yunusbayev/Metspalu Tajiks and Turkmens show up close to each other. But it is clear from their genetics and history that different population dynamics are at play behind the formation of Turkmenistani Turkmens and Tajikistani Tajiks. Turkmenistani Turkmens are the result of the admixture of the Seljuq-era (=early 2nd millennium CE) Turkmen invaders from what is now Kazakhstan with the then Iranic natives of what is now Turkmenistan. Genetically, Turkmenistani Turkmens are a mixture of two population elements: the Khorasani Iranian-like major element and the post-Gokturk Central Asian Turkic minor element; this is clear from the ADMIXTURE analysis results. We know from history that this admixture occurred during the Seljuq era and genetics is concurrent with history here. Tajiks of Tajikistan, on the other hand, have a more eastern Iranic-like population element than the Khorasani Iranian-like population element of Turkmenistani Turkmens, as is clear from the ADMIXTURE analysis results. As for the Mongoloid components of Tajiks, they probably have more historical layers than those of Turkmenistani Turkmens.

            I can only speculate for the genetics of Afghanistani Tajiks due to the absence of samples. I guess they possess Mongoloid components in lesser ratios and the South Asian components in higher ratios than those of Tajikistani Tajiks and thus occupy a position between those of Tajikistani Tajiks and Afghanistani Pashtuns. But this is just my educated guess.

          • I should add that probably on average Afghanistani Pashtuns possess Mongoloid components in higher ratios and South Asian components in lesser ratios than those of Pakistani Pashtuns. But, of course, more sampling of Pashtuns from both Afghanistan and Pakistan will clarify this issue.

          • Do take a look at this.

      • Well, the HAP Punjabi participants are very diverse. There are Muslims, Sikh and Hindu samples.There are 25 Arain Punjabi samples, 8 Punjabi Jatts, 2 Punjabi Brahmins, 2 Punjabi Ramgarhias (traditionally considered lower caste among Sikhs) and 10 generic Punjabis from various caste/tribal backgrounds. The Punjabi Jatts and their neighboring Haryana Jatts together would seem to be marginally closer to the Pashtuns than the Sindhis despite being Sikh and Hindu respectively. However, as a whole, the more presumably homogenous group of Sindhis are closer as a group are closer. Whether this is due to consistent Baloch admixture or not is yet to be seen.

        • I just realized that I was wrong about the ChromoPainter/FineStructure analysis. Pashtuns cluster next to the Bene Israel Jews from Mumbai, but are also next to the HGDP Sindhi. Everything else I said was correct. You make a very good point regarding the HAP Punjabis.

    • Let me see what I can do. Since there are more than 2 significant components, it's not possible to show all variation in a single graph.

  3. It seems most of the Pashtuns are in the 18-26% range for their South Indian component with the average being 23% if I recall for the HGDP Pashtuns. However, there are 5 Pashtuns over the 26% range and surprisingly two that have 35% and 36% of the component. Then, then you have the even more bizarre results for HGDP00220 who only has 11% of the South Indian component but 13% of the Siberian and 17% of the NE Asian components. I think that individual along with the two Pashtuns with 35% and 36% of the South Indian components are 3 of the largest outliers. There could be a variety of different explanations for their results but I'm really curious about them. What caused HGDP00237 and HGDP00239 to have such high South Indian components and does HGDP00220 have some type of Turkic Central Asian or Hazara ancestry?

    Another "trend" I've noticed from the samples we the project has is that the Pashtuns who have less of the South Indian component seem to have significantly more of the Caucasian component and slightly more of the East Eurasian (mostly Siberian and NE Asian) components that offset the lesser South Indian component. In the case of HGDP, it is essentially entirely East Eurasian ancestry that lowers his South Indian and Baloch components.

  4. Pashtuns are extremely liberal in regards to assimilating people of different ethnic backgrounds. The Bannusi are famous for integrating migrant Hindus, Punjabis, and Baloch within their ranks not too far back in the past. If one doubts this, one has only to note the average phenotype of these Pashtuns. Many urban Afghan Pashtuns marry Farsiwan/Tajik women. The same dynamic occurs in urban Pakistan with Punjabis and Urdu speakers. Pashtun identity is traced patrilineally, so your mother can be non-Pashtun, and your paternal grandmother can be non-Pashtun, but you will still be considered a Pashtun, as long as your paternal grandfather was a Pashtun (which means he should have had a Pashtun paternal grandfather), and as long as you follow Pashtunwali (which is irrelevant when dealing with contemporary situations, since few if any urban Pashtuns still observe Pashtunwali, with exception to things like hospitality, as the revenge stuff is too much, a remnant of a darker time. Rural Pashtuns, and people in FATA are a different story). And even here, there are huge genealogical exceptions. The Ghilzai have no mythic paternal link to the greater family of Pashtun peoples. Their mythical ancestor is a Turkic nobleman/prince. Their Pashtun link is the mythical daughter of Batan, Bibi Matao. Nevertheless, nobody can seriously doubt the Pashtun credentials of these people (they are the largest group of Pashtuns in Afghanistan). Many Pashtuns throughout the region have become "Pashtunized" in recent memory (the people of the Korengal valley in Afghanistan, the Baloch living alongside the Khattak, many Hindkowans in Peshawar, Mardan, and Charsadda). The Parachi-Ormuri people are becoming Pashtuns as we speak. It is a robust and credible possibility that most of the Pashtuns of NWFP/Khyber Pakhtunkhwa, FATA, and the Afghan provinces of Khost, eastern Paktika, and southwestern Nangarhar (the famous/infamous "Karlanri", or hill tribes) are assimilated indigenous Dardics, likely descended from people similar culturally and linguistically to the Kalasha and Kohistanis (although there may be an Iranic element, such as the previously mentioned Parachi-Ormuri, and perhaps southerly Pamiri speakers). These Pashtuns have no link with Qais Abdur Rashid, the mythic ancestor of all Pashtuns. Genealogy is not history, but it provides broad historical hints. For example, the Khalil, Yusufzai, and Shinwari are surrounded by Karlanri, yet they are considered Sarbanri (like the southern Pashtuns in Qandahar, Hilmand, and to lesser extent, Pakistani Balochistan). There could be ancient-early modern political reasons for this, but it likely reflects the role these tribes played in the “Pashtunization” of their neighbors, and their older Pashtun identity. To summarize, Pashtuns have a very distinctive and tightly knit culture, but endogamy in regards to women is not important to most Pashtun groups, and large groups of people can easily be integrated into the Pashtun genealogical scheme. Genetically, it’s likely these dynamics are not very important. The heterogeneous groups from which Pashtuns are derived were and are probably very close genetically. In addition to that, persistent intra-ethnic gene flow would make any minor differences fade with time. Geography will be more important than genealogy, but it is still interesting to note these genealogical quirks. Although this is peripheral to the discussion, I think Pashto as a language expanded from and is indigenous to the Pashtun portion of Pakistani Balochistan. The area is home to many archaic dialects. Linguistically, the most divergent forms of Pashto are spoken there, side by side (extremely diverse). Also, the area is completely Gharghasti and Sarbanri (mostly Gharghasti). These groups have an uncomplicated/simple place in the Pashtun genealogy, unlike the Karlanri and many Batanri tribes.

    That was rather long, and I hope I haven't annoyed anyone. The four Pashtuns who were grouped with Punjabis and other northwestern South Asians in the ChromoPainter/FineStructure analysis (this obviously includes the two Pashtuns with extremely high "South Indian" numbers) might be members of the Turi tribe. I don't really know this, just speculation on my part. The HGDP sampling location in the Kurram Agency is actually Bangash territory (as far as I know). Perhaps these individuals are just a part of normal Pashtun variation? Nevertheless, I mention the Turi because that Pashtun tribe was not always Pashtun, and they are important in the Kurram Agency (their relatives in Afghanistan are the Jaji tribe). They used to be "humsahyas" of the Bangash. "Humsahyas" are basically serfs of some dominant Pashtun tribe, but with more friendly and "warm" connotations (some vague notion of "protection"). "Clients" is the best translation, although the term literally means "shadow sharers", and "neighbors" is also a good translation. It is thought the Turi originated in the Punjab (or Sind, there is some ambiguity), although it is also claimed they used to be Hindkowans. They eventually become "Pashtunized", overthrow their Bangash "protectors", and forced them out of the Upper Kurram. They are famed for their ferocity and bravery. They also gave the Pakistani Taliban a rather bloody nose (the Turi are devout Shias, and they had no sympathy for Sunni fanatics). I don't know if these tribesmen still have a predominately Punjabi genetic background. I have never met a Turi. But I have seen pictures of them. They look like rather generic Pashtuns. Here are some photos (the people of Parachinar and its environs are mostly Turi).
    HGD00220 has to have recent Turkic ancestry. Also, what surprised me was that at least one Pashtun has less "South Indian" than HRP0286 (HGDP00214, as HGDP00220 does not count). So, my friend is not a real outlier/unusual, as HGDP00214 tends to cluster with the main Pashtun group in the various ChromoPainter/FineStructure analyses. I can’t wait for the next one. Hopefully, there will be many more participants from across South Asia to make the next batch interesting, and I am encouraging a few Pashtuns to send in their raw-data. Whatever may be the case, I think I wrote too much this time.

    • I know you'e hypothesizing that the Pashtuns with the highest South Indian scores of 35% and 36% are of Turi origin who are said to have originated from Punjab, Sindh or to be of Hindkowan origin. Well, it seems slightly strange that the Turi would have been Pashtunized over time and yet have some individuals with higher South Indian components than all of the Punjabi groups on HAP. The closest being the 4 Brahmin and Ramgarhia samples who have 34% of the South Indian component.

      Also, perhaps it's just me but I noticed that the Pashtuns with the lowest South Indian components seem to have higher East Eurasian ancestry. HGDP0214 is 12% South Indian and 8% combined Siberian + NE Asian while HGDP0220 (presumably Turkic/Hazara admixed) has 30% combined Siberian + NE Asian. Similarly, the Pashtuns with the higher South Indian components seem to have slightly less East Eurasian ancestry. It's not a very consistent trend but something I picked up.

  5. As can be observed above, an average can conceal substantial variation. Some Punjabi Brahmin and Ramgarhia participants have 35% "South Indian" (one of each, I think). And 36% is not that far from 34%. HRP0064 is Punjabi (no caste designation, so he could be a Pakistani), and his "South Indian" component is at 43%. HRP0106 is Punjabi (same note as before), and his "South Indian" is 39%. But most non-Jatt Punjabis seem to be at 30-35% "South Indian". I think it is very interesting that in Punjab, members of the highest caste have the same amount of "South Indian" as members of one of the lowest castes. And a group of agricultural Shudras (Jatts are technically Shudras, and Indian and Arab sources from the early Medieval period describe them as low-status. The "Jatts" of Afghanistan are still tinkerers and musicians, and are on the fringe of society, but I don't quite understand their relation to the actual Jatts of greater Punjab, who very far from being on the fringes of society) are the least South Asian shifted, and have substantial "Northeast European" percentages. The whole worn-out truism that higher caste equals greater West Eurasian affinity does not seem to apply here. I wonder if this could be a function of the caste system's antiquity in the Punjab region. Also, I don't necessarily think these Punjabi-like Pashtun individuals are Turi (as mentioned previously, four HGDP Pashtun samples clustered with Punjabis and other northwestern South Asians in the last ChromoPainter/FineStructure analysis. This seems consistent across such analyses, including MCLUST). I did note that the HGDP samples are from the Bangash portion of the Kurram Agency, and that perhaps these individuals are a part of normal Pashtun genetic variation. Pashtuns and Punjabis are neighbors, and I am sure there has been bidirectional gene-flow. Also, Hindkowans are often regarded as Pashtun by the Pashtuns who live alongside them. It is simply a matter of speaking Pashto, and the rites and rituals of marriage. But I don't think that is the case here, as Hindkowans are not a substantial minority in the Kurram Agency (or anywhere in FATA). That dynamic applies more to Peshawar, Charsadda, Mardan, and the northeastern portions of KP/NWFP (places like Abbotabad, Swabi, Buner, Haripur, and even Swat). I simply mentioned the Turi because this tribe has ultimate origins east of the Indus river (or perhaps are descended from local Lahnda speakers (Hindkowans), whose basic affiliations lie in the greater Punjab region). Also, they happen to be in the neighborhood. They are very important people in the Kurram, where these HGDP samples were collected. Be that as it may, I think you make some very good points.

    • Yes, I fully understood that but I was focusing on the averages because the averages give a better idea of the group. I suppose they are rather meaningless when you have 2-3 individual samples though. In any case, there are individual Punjabi Arains with ASI levels of around 20-22% which would roughly correlate to somewhere around 18-20% South Indian levels (similar to the lowest South Indian component Sindhis). However, I can't seem to find the post with the Punjabi Arain samples but you can always look at the individual ASI estimate spreadsheet here.



      There are also plenty of HGDP Sindhis with South Indian components higher than 30% as the average is 29%.

      HRP0064 is a rather peculiar case. I've been told he was a Muslim Pakistani Punjabi Jatt but his South Indian scores is drastically different from the other Indian Punjabi Jatts and 1-2 individual Pakistani Punjabi Jatts. It could always be variation but at the moment he seems to be an anomaly among the Jatts if he is indeed one.

      When you say one of the highest and lowest castes have similar South Indian scores, are you referring to the Brahmins and Ramgarhias? Well, at the moment there are only 2 samples of each so it is rather difficult to draw accurate conclusions. However, I have discussed the caste dynamics of Punjab with a friend of mine and we agreed it is very different to the rest of South Asia along with Sindh and Kashmir who are different in their own ways. It does not follow the traditional 4 varna system and in Indian Punjab specifically, the supposed "Shudra" Jatts are actually socially and politically dominant. Nor do any of the Punjabi "castes" show any special admiration for Brahmins and the Brahmins have no special "position" culturally like they do in other states. "Caste" in Punjab is in many ways more similar to a tribal system made up of various clans. Especially for Jatts and Khatris in Indian Punjab.

      Anyways, my whole point with the South Indian scores in those two Kurram Pashtuns was if these Pashtuns, whether Turi or not, were Pashtunized over time, why are their South Indian scores still so high? Although, there are some other Pashtuns with scores between 26-29% as well which are rather similar to some of the Sindhis, Punjabi Jatts and Haryanvi Jatts.

      • @ HRP0282, where did you learn that Jatts are shudra when they been always outside of caste system? What you call Jats of afghanistan are same as the Gypsy ,I think 10000 of them were sent into that area by a south indian king long time ago, get their dna results then talk. Jatts are not known as musicians but soldiers.

  6. These results clearly indicate pashtuns are very diverse, thats why I think taking samples from different tribes is a must. There are pashtun tribes who link themselves to jews, iranians, central asian and even indians.

    I find it strange though that pashtuns are being compared to sindhis, it maybe true, but phenotype wise pashtuns are very different from sindhis. Also it depends which sindhis are we talking about. There are sindhis, that are clearly mixed with baloch, but then there are sindhis from lower castes who resemble tamils then baloch.

  7. what about kashmiries? any genetics on them? I would think they might be the closer to pashtuns then sindhis

  8. as for punjab, it is a huge land with many different types of people. some punjabis like jatts maybe closer to pashtun, but as whole they are not. Historically Punjab has always been connected to north india, rather then afghanistan, balochistan etc.... Sindh on the other did see many baloch and pashtun rulers, and this is perhaps the reason they mixed more with sindhis

    • Pakistani, give it a rest, it is clearly you.

      • excuse me? Punjabis are north indians. There is hardly any difference between UP people and punjab, even genetically. It depends on the castes. Pashtuns on the other hand are not Indian, so obviously they will be different genetically

  9. I was wrong about the ChromoPainter/FineStructure analysis. I was going on faulty memory, and I actually checked to see if I was right. The main Pashtun cluster is moderately closer to the Punjabi/Northwest South Asian cluster than it is to the main Sindhi cluster. Paul is right about the Punjabi and Haryanvi Jatts. They will be probably be closer to Pashtuns in their admixture proportions than the HGDP Sindhi. Also, there is a good chance these HGDP samples are from a single tribe. So, as is rather evident, even a single tribe can be incredibly diverse.

    • that is the problem, we dont know anything about which tribe was tested. There are over 60+ pashtun tribes, they will all show different results because it is clear how much diversity there is in pashtuns

  10. also we are comparing 28 pashtun samples to only 5 jatt samples. Mathematically this is not a accurate way to calculate averages. If there were at least 25-30 jatt samples, then we can at least get accurate results and say if they similar or different then pashtun samples

    • Pakistani, if you are going to have an intellectual discussion, at least do your homework. There are now 8 Punjabi Jatt and 5 Haryanvi Jatt samples as well as 1 confirmed Pakistani Punjabi Jatt sample. So, that would be 14 samples.

      • it doesn't matter, still you have to have near equal amounts of samples too have any real conclusion. Jatts overall maybe closer to pashtuns compared to other punjabis. however punjabis as a whole are defiantly not close to pashtuns genetically.

  11. There are significantly more than 60 tribes. It is in the 400-600 range. It all depends on how you count the various "zai", and on how you construe the various tribal "khel". Pashtun tribes have a tendency to bifurcate due to political conflict, or new ecological demands (I am sure most tribal peoples are similar). Also, the Kurram Agency is inhabited by Bangash and Turi. There is also a small minority of closely related "sister" tribes. Also, tribes like the Chamkani and Mangal Pashtuns are to be found. The HGDP Pashtuns are probably Bangash. The samples were collected from the Lower Kurram. Still, as always, you can never be sure. There will always be a chance that the samples are from many tribes. Nevertheless, genetically speaking, variation between tribes shouldn't be too great. Most variation will probably be within large, geographically expansive tribes. Small, isolated groups might be a different story (genetic drift should also be taken into account). If the Y-chromosomal evidence holds up, Pashtuns on both sides of the Durand Line should be very similar. There are three papers on Afghan Pashtun Y-chromosomal variation that I can recall. The main finding (in regards to Pashtuns, as these papers also dealt with other groups) for all of them was the extreme similarity of Afghans to Pakistani Pashtuns (fun fact, the Y-DNA Pakistani samples are also from the Kurram Agency). Look at the PCA plot in this paper.
    The Pakistani and Afghan Pashtuns are virtually identical. They plot on the same exact point. And that makes sense, as the original "Pashtunkhwa", or "Afghanistan proper" ("Afghan" was the Persian term for Pashtun/Pukhtun, just like "Pathan" is still the Indo-Pakistani term) was restricted between the Helmand and Indus rivers (a few hundred years ago, Pashto was not spoken west of the Helmand river, or east of the Indus). The Pashtuns of Helmand Province, and the Pashtuns of Attock, are recent migrants. On the north, the Hindu Kush separated Pashtun areas from Central Asia. The southern wastelands also acted as a barrier. In short, it was not a huge region. It was approximately the size of Kurdistan, but even smaller. The Durand Line split the Pashtun region evenly, right in the middle. FATA, Khyber Pakhtunkhwa, and Northern Balochistan adjoined Pashtun regions approximately their own size and extension. The line was also drawn without regard to tribal affiliation. This has changed over time, as Afghan Pashtuns (technically speaking, "Afghan Pashtun" is redundant) have moved around quite a lot (especially in the north). Simply put, I don't think Afghan and Pakistani Pashtuns will be very different. The differences (if they exist) will probably be of the same magnitude that one sees between the HGDP Brahui, Baloch, and Makrani. And I am sure they will all be significantly closer to each other than they are to any other population. But we can't be certain of anything. We need large autosomal samples from across Afghanistan and Pakistan. I could be wrong in regards to all of this. But I don't think we can expect academic samples from Afghanistan anytime soon. The people of Afghanistan have bigger issues to deal with. Also, more academic Pakistani Pashtun samples are unlikely. As a side note, what happened to the "1000 Genomes" Punjabi samples?

  12. Should be "was bounded by the Helmand and Indus rivers", not "restricted between". I need to use a larger portion of my brain when I write these comments.

  13. HRP0282, I appreciate your contributions. They are very informative and well written. I hope to see Afghan Pashtun samples in the near future once the region is hopefully at peace in the near future. However, I hope for more sampling throughout the South Central Asian and South Asian region. I am particularly interested in Punjabi Khatris and Kashmiri Muslims and Pandits as well as some Himachali and Rajasthani populations which have yet to be sampled. Afghan Tajik samples would be great as well.

  14. Thank you, Paul. I also share your sentiments in regards to sampling.

  15. UP people on average (not counting low castes) are about 41.33% south indian on the spreadsheet. Punjabis are 31.77% south indian, while pashtuns are 21.8%. This shows the difference in number. It is a cline, the more east you go, the South indian rises. so not sure why are punjabis being compared to pashtuns, when it clear that pashtun neighbours like Tajiks, baloch, kalash, and even possibly kashmires are much close to them

    • You are wrong about Kashmiris basically. Kashmiris have slightly more South Indian than Punjabis in general at 32%. Jatts have less ASI than Kashmiris, at around 28%. However,it's pretty much common knowledge that Kashmiris look more West Asian than even the Jatts.
      I attribute that to environmental adaptation.
      It's also possible that the slight east asian in Kashmiris is making them look more West Asian, since the amount of East Asian in them is not enough to affect their overall phenotype.

      • What I mean by the last sentence is that perhaps the slightly elevated East Asian in Kashmiris is oddly enough making them look more West Asian. Now, I don't believe this, but I'm just surmising.
        Anyways, you have fair skin, for example, because the gene you have for fair skin codes for fair skin. You have blue eyes because the genes code for it. Everything physical about you is coded by your genes. A black man won't look white if he heads to Antarctica for example. His genes code his skin color.
        So, Kashmiris look more "West Asian" because the coding regions of their genes is more "West Asian", which for some reason is not reflected in the Harappa results. The million or so nucleotides that have been genotyped by 23andme, and which are used by Harappa, will not tell you what a person looks like physically.
        The fact of the matter is, all billion+ nucleotides in your DNA have not been analyzed. Until we can start analyzing a person's entire genome quickly (which will happen some day), I think the results will always be inaccurate.

        If that makes any sense to anybody.

        • because the kashmiri sample sizes are too small.

          Explain how could kashmiris who are more distanced geography wise to India be more ASI than punjabis then? how could jatts be less ASI than kashmiri!!!

          How could punjabi jatts... incl indian sikhs and hindus... be close to pashtuns when they only ever married other Sikhs and Hindus from their sub-clans. In real life, when ever do jatts and pashtuns ever co-mingle??? tell me because i can't think of ONE EXAMPLE

          even PUNJABIS know PASHTUNS are whiter and physically more Iranic looking from them's a stereotype steeped in truth. pashtuns have always considered punjabis 100% INDIAN 100% DALKHOR... and the word PUNJABI is used for Insult word in Afghanis Tojiks turkmens hazaras....Punjabi = meaning HINDU or BLACK.... not to mention these people were slaves of ghazni and Duranni....afghans on pashtunforums are fed up and mad about these self hating Indian hindu jatts trying to steal their identity & history they had enough

          real life...pashtuns and jatts barely acknowledge the other ones existence but guess what... they are more distanced because of the religion gap! lol this means the similarity between the two groups is one that came about separately through separate evolution and is purely COINCIDENTAL!!!! but here they are trying to use the fact that Punjab was part of the empire to argue some kind of fake link!

          some of the high south indian in pashtuns is from afghans taking slaves from India.....many times throughout history, pashtuns came and took indian punjabi slaves and many died trying to cross the HINDU KUSH mountains... its common knowledge...

          • Because even a separation during the past 1000 or so years has not been able to remove the underlying similarity between Pashtuns and neighboring Jats. I would not be surprised if Pashtuns were originally west of Indus Jats that started to split in the post Arab period (jatan-i-gharbi vs. jatan-i-sharqi). After all their Y-lines are almost identical.

          • I agree with you hamza. once we get more samples we will see the clear truth, unfortunatly with these small samples, no one can get clear 100% truth

          • @ Hamza, Jatts history is much older than your Pashtun history, as Parasar pointed out about Ydna, I am going to tell you that Jatts way of thinking is very close to Pushtunwali. Stop your half truths, Afghanistan been always ruled either by Persians or Indians and yours is the most invaded country in history. All invaders into India had learned the hard way to leave Jatts alone as they were too much trouble for them.

        • Kashmiri Pandits are known to have southern brahmans (Agnihotris) in their Bhanmashi group. They also have number of families of Madhyadesh (eg Bilhana of Gopa Agrahara), Mithila (eg the Kaulas), and Bengal origin (eg Jayanta Bhatta).

          That perhaps is the reason for their elevated southern and eastern components.

          • parasar you are back to you hindu nationalistic tricks, calling pashtuns as jatts lol hahahha. first of all pashtuns have almost no influence from the arabs, even these genetics tests show that. secondly it is mostly likely jatts who actually migrated from afghan region and mixed in to the indian population. so stop talking nonsense like usual

          • If you have some evidence of Pasthuns existing in circa 700ad, please provide. Jatts migrated out of Sindh to the upper Indus region. There is no mention of the Pasthuns by the Arabs of the period who documented the whole region reasonably well. The folk in the Zabul, Zamindawar region are called Ratbil or Zunbil.

          • pashtuns have traditions totally opposite of that off hindu jatts. where did pashtunwali honor codes come from? why is pashtun culture so different from indians? all these things give clear evidence that pashtuns are not indians and never were. Just because arabs didn't record pashtuns, doesn't mean anything. since when did arabs record baloch, kalash, bursho etc..? does that mean they didn't exist? you are just being a typical hindu nationalistic

          • Even now there are just a few thousand Kalasha. Burusho number a little more. The Pastuns on hand are a major group numbering perhaps 50 million people.

            The Arabs did record the Bhatti, Jat, Med, Samah, Brahman, Kashmiri, Lohana, etc, among many others. You could review the Chachnama's index here
            The Pashtun are absent.

            Which means that they are a later cultural and ethnic development. We do see their precursors mentioned in the Ghaznavid period, but even at time the Suri and Ghorids are not referred to as Pasthuns.

            Sure they are different from the Jatts, but the question is when did the Pasthun identity come into being?

            Encyclopedia Iranica calls them "recent settlers" but also finds an ancient derivation possible - "The most plausible derivation of Paҳt’o, as already suggested by Markwart (Untersuchungen zur Geschichte von Eran, Göttingen and Leipzig, 1896-1905, II, p. 177; cf. Morg[enstierne]4, par. 40b), is from *Parsuwā, and of Paҳt’ūn from *Parswāna-, with the basic stem *Parsū-; cf. Skt. (Pāṇini) Parśu- “a (northwestern) warrior tribe.” Tedesco, in a letter, compares Pārsa- (from a vṛddhi from *Pārswa-)."

    • Did you really tally up the results up yourself? Also, there are Tajik samples from Tajikistan are they are very different autosomally than the Pashtuns. Perhaps, Afghan Tajiks are different though. The Baloch are also somewhat different but closer than the Tajiks in terms of clustering. Razib Khan pointed out the Kalash are in many ways like an inbred version of Pashtuns but you would know that if you actually read rather than making pointless generalizations and unsupported claims. As for Kashmiris, it's possible but we don't really know because the Pandits nor two individual Kashmiris (Muslim possibly) don't show any special relationship to the Pashtuns.

      • The Fst distance between the HGDP Pashtuns and Tajiks is very low. Only 0.004. You can check the Metspalu et al. supplementary data. Fst distances try to summarize a lot of information in one statistic. But I think they get the job done.

        • Yes, it probably is and I expected such due to shared linguistics, culture and history but look at some of the Tajik's components. They have 7% of the South Indian component, 32% of the Baloch component, 21% of the Caucasian component, 17% of the NE Euro component and approximately 15% of mixed East Asian components. Essentially, Tajikistani Tajiks are much less South Asian shifted and much more East Asian shifted than the Pashtun samples. Overall, fairly close in the grand scheme of things but still different. What is the fst distance between the Baloch and Pashtuns?

    • I'm sure those averages are off. I added up every individual and got 21.5%. I even included the Pashtun who has 11% of the South Indian component but who has 30% of the Siberian and NE Asian component as well as the slightly admixed HGDP0214 who has 12% of the South Indian component but 8% combined Siberian + NE Asian. If I did not include the 11% individual the number is 21.92%. There is a very large amount of Punjabi samples so I find it hard to believe you could add them up when you couldn't even count a group that Zack had written were 23 HGDP and the 3 HAP participants in the "freaking" first of his post.

      • I added up all the punjabis. there is a clear difference between punjabis and pashtun figures

  16. ^ how many kashmiries are tested? and which ones, pandits or muslim? it is well known that pandits have mixed with other indians from the plains, in fact many pandit groups actually came from places like UP/Bihar and settled in kashmir, I am not talking about them. I am talking about real kashmiries native to the valley, muslim kashmiries have a tendency to look similar to south central asian groups. Yes you are correct that Jatts do not look west asian at all, they are very similar to other indians

    • It looks to me like two Kashmiri Muslims have been tested, as far as I can see, and they are no different from Kashmiri Pandit samples.
      To be honest, if anybody is going to mix, it will be Kashmiri Muslims, and not Hindus, because Islam promotes equality. Case in point, muslim samples from UP and Gujarat who both show elevated mediterranean and SW Asian levels.

      • A Kashmiri Pandit would generally identify as Kashmiri Pandit so it is very much possible that HRP0021 and HRP0250 (Kashmiri Butt) I believe are Muslim Kashmiris. There are some "Pahari" Kashmiris participants as well but they labeled as part of the Pahari group anyways rather than Kashmiri.

  17. UP is a very mixed area, there are muslim families in Delhi and west UP who are most likely outsiders. There are pashtun, afghan, Tajik families who lived there as well and may have mixed with the locals

    Anyways, Kashmiri muslim they did mix, but with their pashtun neighbors, pashtuns ruled kashmir. Perhaps this is the reason muslims in the valley look more west asian then pandits? not sure

    • Please, pakistani, if you're going to lie at least try to use "fake" sources. Kashmiri Muslims are not heavily mixed with Pashtuns to any extent. There are Pashtuns living in Kashmir but they remain in their separate communities. Besides, Kashmiris considered Pashtuns and any other non ethnic Kashmiris as outsiders. Nor has the Pashtun rule of Kashmir been historically perceived as a bright spot in Kashmir's history.

      Also, what proof do you offer that Kashmiri Pandits are mixed with non Kashmiris and some clans/krams have origins in the plains of UP and Bihar? In fact, you have things backwards. Kashmiri Pandits are wholly indigenous to Kashmir and it's highly possible that "some" of the ancestors of some Kashmiri Muslims are foreign to the Valley rather than the other way around. Or do you somehow believe Islam is indigenous to the Valley? The only religions that can be considered native to the region are Hinduism and Buddhism to a lesser extent.

      • then why do pandits look like indians, while muslim in the valley look more west asian?

        • They don't look that different on average. There are some much more West Asian looking ones that pop up among Muslim Kashmiris but that may or may not be due to foreign admixture. As a whole, Pandits and Muslims don't look that different. Now, please stay on topic and don't begin spamming cherry-picked pictures of Pandits and Muslims.

          • out of the all the interviews I have seen of Pandits, none of them looked west eurasian, all of them had very strong indic features. For muslims kashmiries however, there is clear outside look about them

          • Kashmir's Muslims are of two sorts; the native Muslims who converted from Hinduism and the migrants from central Asia. The latter category are prominent but few in number and claim Syed origin. They are referred to by other Muslims as Mallahs, marrying among themselves generally but culturally the same as other Muslims. Native Muslims are also of two types; the descendants of Pandits who converted, often retaining their Pandit surnames, and the non Pandit types with surnames such as Khandey, Parray, Lone, etc, but intermarriage is common. There are a few Afghan families in Srinagar which have by now almost compeletely intermarried with native Kashmiri Muslims, but a village near Ganderbal still comprises Pathans who speak Pushto and do not intermarry with Kashmiris.
            Kashmiri Pandits may resemble North Indians, even South Indians sometimes because there was frequent coming and going over the Pir Panjal passes but you can also find Kashmiri Muslims with similar outlook. In general KPs are a good looking fair skinned people with the usual quota of light coloured eyes.

    Could this explain why the HAP Pashtuns (obviously, myself included) have less of the "Baloch" component then the HGDP Pashtuns?

  19. Baloch is a west asian component. Anyways, Pashtuns are mix West Eurasian and West Asian, they have high components from all those areas, pashtun have very high caucasian and northern euro

      Click on the "Baloch" component. The HGDP Pathans are ranked 10th in terms of that component. The HAP Pashtuns are 58th. Most Pakistanis, Northwestern Indians, West Indians, North Indians, and high caste South Indians have more or equal amounts of the "Baloch" component. I have also seen two Pashtun results for DIY HarappaWorld. One of these individuals originates from Afghanistan, and the other from Pakistan. The Afghan has 35% "Baloch", and the Pakistani has 33%. By contrast, their "South Indian" numbers are 19% and 17%. The "South Indian" percentages for these two individuals are clearly within range (some individual HGDP Pathans have close or identical numbers), but not the "Baloch" percentages. And both of these individuals have very small East Eurasian percentages. I wonder if the lower "Baloch" percentages are a technical issue, or a genuine result. If the latter, why do the HGDP Pathans have such high "Baloch" numbers? The other Pakistani Pashtun whose results I have seen, and I myself, originate from places much closer to Balochistan proper than the HGDP Pathans. HRP00286 has ultimate origins in Kandahar, which is even more close to Balochistan, and actually has a Baloch minority.

      • The Baloch component is simply referred to as Baloch because it peaks in the Baloch individuals. It actually represents ancient West Eurasian ancestry that is hypothesized to have originated in the Neolithic West Asian farmers that migrated to South Asia from somewhere near the Northern Middle East. Perhaps Eastern Anatolia. It does not actually originate in the Baloch themselves but represents a shared West Asian ancestral component between the Baloch, Pashtuns and South Asians. It represents the majority of the West Asian ancestry in South Asians. However, it is found in significant amounts in Iranians as well.

        Not sure if you have picked up on it but I have noticed that although some Pashtuns have lower Baloch components, those specific participants tend to have higher Caucasian scores.

        So, what I think might explain this trend of the HGDP Pashtuns have different Baloch scores than some of the individual participants is interregional variation among Pashtuns. As a whole, I hypothesize most Pashtuns from along the Eastern Afghan border and those living in FATA and Khyber Pakhtunkhwa should be autosomally similar. However, perhaps certain tribes or confederations have higher Caucasian or Baloch scores and vice versa. Also, I don't have the chart but I think the Caucasian and Baloch components are very closely related in terms of genetic distance.

        • The "Baloch" component is a statistical fiction (as are all of these "components" inferred from Admixture). Panmictic populations don't exist in the real world. Concerning the names, Zack sums everything up nicely when he says they are just useful mnemonics. I happened to be using "Baloch" as a useful mnemonic. Nobody is engaging in the "fallacy of misplaced concreteness" here. The "Baloch" component has been named as such because it is modal amongst the populations of greater Balochistan. I understand that. I think I have issues conveying my thoughts in a concise fashion. My writing is on the sloppy side, especially when I comment. Also, my writing can be rather disjointed. As I said before, I need to use a larger portion of my brain when commenting. I also need to proofread. Putting that all aside, I was noting the possibility of inter-regional variation. Citing that as an explanation is redundant. What struck me was the fact that Pashtuns who have lived very close to the Balochi people, and for a very long time at that, have less of their modal component than the HGDP Pashtuns, who are from a location very distant to Balochistan. The HAP Pashtuns have less of the "Baloch" component than most North Indians. Certain South Indian groups have more of that component than us HAP Pashtuns. The "Caucasian", "Southwest Asian", and various east Eurasian components reduce that aspect of our ancestry. This has interesting implications on PCA plots. If you go back to the post where HRP00286 results were being discussed, he posted his McDonald results. He clustered further outside South Asia in comparison to HGDP00214. But he has slightly more of the "South Indian" component than HGDP00214. The "Baloch" component is mapping onto something here. More of the "Baloch" component translates into a more "southern" accent on the PCA plots (but more of a "West Eurasian" shift is also implied, obviously). I might sound confusing here, but this is simply an indication that the "Baloch" component, although important throughout "West Eurasia", is on the South Asian side of things. And that is obvious, as the "Baloch" component is the only substantial "West Eurasian" component amongst peninsular South Asians (obviously, the "South Indian" is a hybrid "component", but definitely biased towards "West Eurasia"). The magnitude of the "Baloch" component declines far more quickly west of Pakistan than east of Pakistan. To make a long story short, HAP Pashtuns are probably slightly more divergent from other South Asians than the HGDP Pashtuns. And the "Baloch" component is mapping on to that. We share slightly less of our "West Eurasian" ancestry with South Asians than the HGDP Pashtuns. That is what I was getting at.

          • Okay, I understand what you're saying. I agree that the Baloch component is more South Asian shifted than the Caucasian component for example. However, I believe that HGDP0214 has more East Asian ancestry than HRP00286 as well which is pushing him more eastward than HRP00286 because East Eurasians are slightly closer to the ASI than West Eurasians are. HGDP0214's East Asian ancestry is at 8% (Siberian + NE Asian) while HRP0286's is at 5%. However, I agree with you. The Baloch component is definitely more Southeastern shifted than the Caucasian component and is playing a role as well. I wish Zack would updated his 3D cluster plot. You would be able to see where all of you clustered among the HGDP Pashtuns and all the other participants in the project.

    • A mix of West Eurasian and West Asian? Those two terms are essentially synonymous and hence you're repeating yourself. What's more accurate is that Pashtuns are a mix of West Asian (majority), South Indian (minority) and to some extent East Asian (minority but varies among individuals). Yes, Pashtuns have higher averages Caucasian scores than neighboring populations including the Baloch but similar and higher Northern Euro scores are found among NW Indians as well.

      • the higher caucasian component in pashtuns is clearly making pashtuns look more west eurasian then Indians

        • Do you even read what you write? Are you saying 5-6% higher levels in the Caucasian component is what makes Pashtuns look different? Not adaptations to their environments and different ancestry as a whole? Besides, no one here is making the claim that Pashtuns look like Indians or NW Indians. They clearly look different on average. So, unless you have something to contribute regarding genetics, kindly leave this page.

          • It's not environment only that effects your looks, for example Himachal people in India live even on higher grounds/cold weather compared to pashtuns, yet generally they clearly resemble indians not west eurasian. Pashtuns are defiantly effected by their west Eurasian genetics much more then Indian groups are. I am not sure how nature works but this is definatly the case. Same could be said about Kalash, Bursho and other south central asian groups

  20. "I wish Zack would updated his 3D cluster plot. You would be able to see where all of you clustered among the HGDP Pashtuns and all the other participants in the project."
    That would be great.

  21. "Because even a separation during the past 1000 or so years has not been able to remove the underlying similarity between Pashtuns and neighboring Jats. I would not be surprised if Pashtuns were originally west of Indus Jats that started to split in the post Arab period (jatan-i-gharbi vs. jatan-i-sharqi). After all their Y-lines are almost identical."
    I don't disagree with the fact that Pashtuns and Jats are genetically very close. As Paul very nicely put it, in the grand scheme of things, most eastern West Asians, South-Central Asians, and northwestern South Asians are very close genetically. But there are good historical reasons for not hypothesizing that Pashtuns are descended from western Jats. Afghanistan has a very old minority of "Jats". To be rather frank, I don't understand the relationship between these people and the Jats of greater Punjab. Technically, Jats are Shudras, but they are very important and powerful in Punjab. In Afghanistan, "Jats" live miserable lives. Their very name is an insult. In this regard, they fit the description of Jats provided in early Arab and Indian sources. Most of these Afghan groups speak Hindko (though, other languages are also spoken). They also tend to be nomadic. As I said before, I don't understand the relationship between these people and Jats east of the Indus.

  22. ^ but why do pashtuns/jats look so different then? there has to be underlying reason for this, it could be possible that jats over the centuries have mixed with indians and thats why they look indian today, while pashtuns being a different cultural group didn't mix with indians and thats why they retained their original features more.

  23. "The Arabs did record the Bhatti, Jat, Med, Samah, Brahman, Kashmiri, Lohana, etc, among many others. You could review the Chachnama's index here
    The Pashtun are absent.

    Which means that they are a later cultural and ethnic development. We do see their precursors mentioned in the Ghaznavid period, but even at time the Suri and Ghorids are not referred to as Pasthuns."

    Parasar, I think a lot of what you are saying is correct. I agree that some aspects of Pashtun culture are rather recent developments. During the early Arab conquests, eastern and southern Afghanistan were extensions of greater India. Brahmanism and Buddhism have very ancient histories in the region. Although you could risk your life telling this to a conservative Pashtun in Pakistan or Afghanistan, the fact is that the region put up fierce resistance to "Islamization". The Arabs suffered heavy defeats at the hands of the Zunbils, and the people of highlands took quite some time to heed the call. In many ways, Pashtuns still problematize the distinction between Iranian and Indian. Pashtuns have at various points in history been construed as the most eastern of Iranians, or the most western of Indians. We have close ties to both "Iran" and "Hind". Also, the "Gar" and "Samil" distinction among the highland Pashtuns is thought to be the legacy of a much more ancient difference between originally Buddhist and Zoroastrian tribes. Nevertheless, the Ghorids were not Pashtuns. The Suri were, and still are, Pashtuns. At this period in history, Pashtuns were referred to as Afghans. The "Afghan" ethnicity of the Suri has never been questioned. The latest we can place the crystallization of Pashtun identity has to be substantially earlier than the writing of the Baburnama. Babur notes many Pashtun tribes, and he finds/fights most of them where they still happen to live. The Afridi, Yusufzai, Wazir, Mohmand, Bangash, Turi, Niazi, Isakhel, Lohani, Muhammadzai (not the Durrani tribe, a different one in what is now Pakistan) are all mentioned. The only omission, and the most important one, are the Abdali, who are now called Durrani. He does not seem to know they exist, and he does not find them where they are today. Nevertheless, he also recognizes that all of these tribes are members of a single "ethnicity" ("ethnicity" is perhaps too much of a contemporary Western notion, but you get what I am saying). They all speak a single language, which he calls "Afghani", and constitute an interrelated group by his time. That happens to be the early 16th century.

    • HRP0282,

      Thanks. My ideas about Pashtuns are quite in line with yours, but I am not clear on the dynamics of how the Afghan identity changed to a Pashtun one. I was wondering if you could clarify this point.

      Baburnama is indeed an interesting read. The diversity of peoples and languages reported by Babar is fascinating. Eg. "In the country of Kābul there are many and various tribes. Its valleys and plains are inhabited by Tūrks, Aimāks, and Arabs. In the city and the greater part of the villages, the population consists of Tājiks. Many other of the villages and districts are occupied by Pashāis, Parāchis, Tājiks, Berekis, and Afghans. In the hill-country to the west, reside the Hazāras and Nukderis. Among the Hazāra and Nukderi tribes, there are some who speak the Moghul language. In the hill-country to the north-east lies Kaferistān, such as Kattor and Gebrek. To the south is Afghanistān. There are eleven or twelve different languages spoken in Kābul: Arabic, Persian, Tūrki, Moghuli, Hindi, Afghani, Pashāi, Parāchi, Geberi, Bereki, and Lamghāni. It is dubious whether so many distinct races, and different languages, could be found in any other country ... There is also the country of Ghazni, which is often denominated a Tumān. Ghazni was the capital of Sabuk­tegīn, of Sultan Mahmūd, and of the dynasty sprung from them. Many call it Ghaznīn. This was also the capital of Shāhāb-ed-dīn Ghūri, who, in the Tabakāt-e-Nāsiri, and many of the histories of Hind, is called Muizzeddīn. It is situated in the third climate. It is also named Zābul, and it is to this country that the term Zābulistān relates; many include Kandahār in Zābulistān ... The inhabitants of the open country are Hazāras and Afghans. Ghazni is a cheap place compared with Kābul. The inhabitants are Moslems of the sect of Hanīfah, and orthodox in their faith ... Alāeddīn Jehān­sōz Ghūri, when he subdued this country, broke down the mound, burned and destroyed many of the tombs of the royal family of the Sultan, ruined and burned the city of Ghazni, and plundered and massacred the inhabitants. In short, there was no act of desolation and destruction from which he refrained."

      Babar also mentions Jats and Gurjars in a similar vein as the Afghans - "who were partly Afghans of the tribes of Dilazāk and Yusefzai, and partly men of the Jāt and Gujer tribes ... In the hill-country between Nilāb and Behreh, but apart from the tribes of Jūd and Janjūheh, and adjoining to the hill-country of Kashmīr, are the Jāts, Gujers, and many other men of similar tribes, who build villages, and settle on every hillock and in every valley ... On Friday, the 14th of the first Rabi, we arrived at Siālkot. Every time that I have entered Hindustān, the Jāts and Gujers have regularly poured down in prodigious numbers, from their hills and wilds, in order to carry off oxen and buffaloes. These were the wretches that really inflicted the chief hardships, and were guilty of the severest oppression on the country."

      Regarding your comment: "Nevertheless, the Ghorids were not Pashtuns. The Suri were, and still are, Pashtuns. At this period in history, Pashtuns were referred to as Afghans. The "Afghan" ethnicity of the Suri has never been questioned."

      Are the Ghorids and Suri not related? Utbi did not include the Suri among the Afghans but among the Hindus. Utbi distinguishes the Afghan from the Hindu - "Sultan was occupied in arranging the means of war and he collected a numerous army of various tribes of Turks, Khalajes and Hindus and Afghans and the Gozz troops and they met at a wide place, four farsangs from Balkh" (please also see his Account of the Affair of the Afgháns - "Sultán occupied himself in repelling the nation of the Afgháns, who made their homes in the acclivities of cliffs and the summits of mountains, and for a long time had been accustomed, with violent success, to stretch out their hand (to attack) the extreme border of his territories."). I would also bring to your attention El-Beruni's comment - "In the western frontier mountains of India there live various tribes of the Afghans, and extend up to the neighborhood of the Sindh Valley." Also in 982AD, prior to Utbi and Beruni, the chief of Nagrahara (Jalalabad) is mentioned as having Muslim, Afghan and Hindu wives, implying that the Afghans at that time were not Hindu.

      • Hi Parasar,

        Afghan has always been the foreign term for Pashtun. Even the British used it for the Pashtuns of what is now Pakistan, but after the drawing of the Durand Line, they started referring to Pashtuns under their jurisdiction as Pathans. The term has never really been used amongst Pashtuns themselves. There are various etymologies for "Afghan", but I am not sure any are reliable. But there is one detail that I should mention. In many cases, the "Afghan" denomination has been restricted to the Sarbanri lineage of Pashtuns. Although even these Pashtuns prefer being called Pashtun amongst themselves, they tend to also like "Afghan", and give that as their ethnicity to non-Pashtuns. By contrast, the Karlanri never called themselves Afghan, even when they owed their allegiance to Kabul. It is also an interesting fact that all of the Sarbanri, whether in Pakistan or Afghanistan, live in fertile open plains or valleys. By contrast, the Karlanri, whether in Afghanistan or Pakistan, are all mountaineers.

        In regards to Ghorids and Suri, it is thought the Suri are descended from a prince of the Ghorids, but by the time they were rulers in India, they were thoroughly Pashtunized.

  24. HRP0282, while I dont know the exact history of pashtuns and the so many tribes, however as a whole they are not indians. They are not Iranians either, but they are what their location is, south central asian people, similar to their neigbhours like baloch, Kalash etc.... Their language is eastern Iranian, quiet different from any indic languages. Afghanistan region forms a different cultural zone altogether from India or Iran.

    Though that said, looks wise pashtuns are defiantly closer to Iranians, then to Indians, even Punjabis the most north west indians. There is no (indic) type look in any pashtuns, even the darker pashtuns, who actually look more semetic. In eastern Pakistan, people equate pashtuns and other northern pakistanis like Kalash, Bursho to (Europeans) because they look so different from eastern pakistanis (Punjabis, SIndhis, Muhajirs) and have a totally different culture then them

  25. Hi Imran,

    There is no doubt that the Pashtun people defy easy cultural categorization as Central Asians, South Asians, or West Asians. But I think the most profitable way of going about this is noting history and geography. Pashtuns have played a large historical role in Khorasan, but are also inextricably linked with Greater India. The Pashto language is Northeast Iranian (same as the Pamiri languages, or Ossete in the Caucasus), but the Kabul river is a tributary of the Indus. The Kabul Valley has been ruled from both "Iran" and "Hind". And its peoples have ruled in both "Iran" and "Hind". But this is what makes Pashto culture so interesting. I agree that contemporary Pashtun culture is very different from the culture of the lowlands of Punjab and Sindh. But the differences should not be exaggerated. As far as contemporary Afghanistan is concerned, it is not a distinct cultural zone. Afghanistan north of the Hindu Kush is unambiguously Central Asian. Herat and the surrounding region was an integral part of Iran/Persia. In the center, you have the Hazarajat. All of eastern Afghanistan south of the Hindu Kush is Pashtun, and is contiguous ethnically and geographically with the Pashtun areas of western Pakistan. And finally, one has the ethnically mixed Kohpauya.

    As far as phenotype is concerned, one is dealing with a lot of subjectivity. People look at these things from very different perspectives, and have different background assumptions. You could measure different groups of people, and compare them, but even that involves a lot of subjectivity. Not to mention phenotypic plasticity, epigenetic dynamics, natural selection, the effects of genetic drift, and unavoidable variability. People just do not fall into easily sortable "types". And anyways, what would be the point of it all? Why does it matter? From a genetic perspective, all living humans are equivalent to a single Chimpanzee or Gorilla subspecies. We all belong to a single lineage, Homo sapiens sapiens. Here is a nice open access paper dealing with great ape genetics.

    Nevertheless, biological anthropologists with a lot of time on their hands did measure people from the region. It is clear from their work that Pashtuns (in terms of phenotype) are metrically and morphologically intermediate between the peoples of northwestern South Asia, and the peoples of the Iranian Plateau, but with a slight Central Asian touch. And that gels with what one can see with one's own eyes. They also recorded their observations on variation in unexposed skin color. The majority of unexposed skin color tends to be "brunette white", with a minority of individuals with "pinkish white" unexposed skin color, and a minority of individuals with "light brown" unexposed skin color (as you can see, the very terms are extremely fuzzy, and open to interpretation. I am not aware of any contemporary study of Pashtuns utilizing more objective skin reflectance measures). That also makes sense. Not a lot more one can say on this topic.

  26. Language is really what binds the Pashtuns.

    Pashto is related closely to Munji and Ishkashemi, languages of the Pamirs.

    Just based on this and the lack of Eastern Iranian languages in Pakistan, here is my hypothesis.

    I would assume that they split off from the Pamiri people and gradually moved into the Khyber Pass region and down into Helmand where they assimilated other peoples.

    If this involved force and a large number of conquered peoples, then that could also explain the aggressiveness of Pashtun tribal customs.

    This probably occurred during the invasion of nomadic Hunnic/Iranian/Tocharian groups from Northern Afghanistan into Southern Afghanistan and West Pakistan.

    • One thing to keep in mind is that the Gathas' geographic horizon is modern Afghanistan and the Punjab. Which means that direct descendants of both neighboring sister languages Gathic and Rg Vedic are still next to each.

    • Reality Check

      According to my understand the genetic data from kalash (an isolated dardic group) and kurram valley pashtun/pakhtun (karlanri) samples resemble so much to each other that we have to believe that most of the karlanri pashtun/pakhtuns are pashtunized kalash like dardic tribes.

      • If they are Karlanri samples, does that mean some are definitely Bangash or possibly some Zazai? Both are Karlanri.

        • Jaji/Zazai are found in Afghanistan (extremely small numbers in Pakistan, although they are identical to the Turi, just Sunni). The lower Kurram, where the HGDP Pashtuns were sampled, is mostly inhabited by Bangash. I just found this excellent map, Based on my prior knowledge, it is perfect. You just can't find a more accurate representation of tribal distribution.

      • Reality Check: What are you talking about, whatever your real name is troll?

  27. Hi HRP0286,

    I think that is a very plausible scenario (although I would note the existence of one other Eastern Iranian in Pakistan, Ormuri, but these people are losing their language, and are now considered Pashtun). The people of the Pamirs are very interesting. Pashtuns who test at the Genographic Project have them as their closest population.

  28. well there are also many pashtun that have semitic features with black hair
    why no one ever mentions that they look different than some of the other pashtun very handsome on 03/10/11 i dreamt of the Prophet PBH talking to Abu Bakr Siddique RA and was asking him(Abu Bakr SiddiqueRA) about the Pashtun people he Abu Bakr RA said I am lighter than them in color but they are the same as us he and the Prophet PBH pashtun and looks semitic not like indo aryan

  29. Well, Harappa Ancestry Project has it's first Ghilzai Pashtun. Zack updated the individual admixture spreadsheet and HRP0326 looks to have some "real" and substantial Turkic/East Eurasian admixture (4% SE Asian, + 6% Siberian + 3% NE Asian = 13%). It's plausible that it is the reason for both lower South Indian and Baloch components. Very interesting results. Despite the low South Indian score, this individual is probably less ANI overall than many of the Kurram Valley Pashtuns due to the probably Turkic-like admixture.

    South Indian: 15%
    Baloch: 32%
    Caucasian: 21%
    NE Euro: 12%
    SE Asian: 4%
    Siberian: 6%
    NE Asian: 3%
    Mediterranean: 2%
    SW Asian: 5%

    Also, HRP0318, a Lahori Punjabi, has similar results to the Punjabi Jatts with a slightly higher pull to the Caucasian than the Baloch component.

    South Indian: 27%
    Baloch: 37%
    Caucasian: 14%
    NE Euro: 12%
    NE Asian: 3%
    SW Asian: 3%

    Finally, HRP0320, an UP Muslim, looks to potentially be of upper caste origins and have real Middle Eastern like admixture due to some slightly exotic components and lowered South Indian component.

    South Indian: 28%
    Baloch: 35%
    Caucasian: 12%
    NE Euro: 11%
    Siberian: 3%
    Papuan: 2%
    American: 2%
    Mediterranean: 2%
    SW Asian: 3%

    • hi paul,

      i am a ghilzai pashtun of the safi tribe from tagab in afghanistan from my father's side. my mother is ethnically mixed tajik and pashtun from qandahar. how can i get myself tested and analysed as per above?


  30. HRP0326 has very fascinating results. But I would note that he is actually our first "official" Ghilzai participant. He is not necessarily our first Ghilzai participant in a general sense. HRP0281 could well be a Ghilzai. Karlanri and Gharghasti are minorities in Afghanistan. It is dominated mostly by Durrani and Ghilzai, and the latter are the largest group of Pashtuns in the country. It would be great if HRP0281 could help us out on this question.

  31. Here are the HGDP Pashtun Fst numbers from Metspalu et al (you can find them in the Supplementary Data). Very hard to read in the actual supplements, so I thought it would be nice if I post this (for anyone who might have been curious, but just could not bear dealing with that monstrosity of a visualization). All caveats that apply to the interpretation and utilization of Fst should be kept in mind (also, please do point out any errors).

    1. HGDP Pashtun=0.000
    2. Sindhi=0.002, UP Brahmins=0.002
    3. Balochi=0.003, UP Kshatriyas=0.003
    4. Tajiks=0.004
    5. Makrani=0.005
    6. Brahui=0.006, Burusho=0.006, UP Muslims=0.006
    7. Iranians=0.007, Gujaratis=0.007
    8. Nogais=0.008, Kumyks=0.008
    9. UP Caste=0.009
    10. Lezgins=0.010, Balkars=0.010
    11. North Ossetians=0.011, Uzbeks=0.011
    12. Armenians=0.012, Chechens=0.012, Turkmen=0.012, UP Kols=0.012
    13. Abhkasians=0.013, UP Kanjars=0.013, MAD CHA Gonds=0.013, AP Velama=0.013
    14. Georgians=0.014, UP Dharkars=0.014
    15. UP Chamar=0.015, TN Kurm+tribe=0.015
    16. Tuscans=0.016, Russians=0.016
    17. Uygurs=0.017
    18. French=0.018, Hazara=0.018, Palestinians=0.018, TN Piramalai Kallars=0.018
    19. North Italians=0.019
    20. Druze=0.020, KAR Halakipikki=0.020
    21. Ctr India mix+Nihali=0.021
    22. AP Chenchu=0.022, KER Sakilli=0.022
    23. Bedouins=0.023
    24. KAR North Kannadigas=0.024
    25. UP Dushadh=0.026
    26. Sardinians=0.031
    27. BIH JHA Munda=0.032
    28. KER tribes mix=0.034
    29. Mozabites=0.038, ORI CHA Munda=0.038
    30. Burmese=0.053
    31. TN Pulliyars=0.059
    32. MEG Garos=0.060
    33. Mongols=0.064, Tu=0.064, Xibo=0.064
    34. Cambodian=0.065
    35. Daur=0.071
    36. Oroqen=0.073
    37. Hezhen=0.074
    38. Yizu=0.075
    39. Naxi=0.076
    40. Tujia=0.078
    41. NAG Nagas=0.079
    42. Dai=0.080
    43. Miaozu=0.081
    44. Han=0.082
    45. Japanese=0.083
    46. She=0.085
    47. Lahu=0.086
    48. Bantu NE =0.123
    49. NAN Melanesians=0.130
    50. Bantu SW=0.134
    51. Mandenka=0.136
    52. Yoruba=0.138
    53. Papuans=0.156
    54. Biaka Pygmies=0.164
    55. San=0.184
    56. Mbuti Pygmies=0.185

    It would be interesting to see how things look if one excludes the four Punjabi-like samples. If one were to do this, the Iranians would probably be closer to the Pashtun samples than the Gujaratis are. But one would actually have to do this to be sure.

  32. I should have posted this the first time. Here are comparable Fst distances for the UP Brahmins. They seem to be one of the closest populations to the HGDP Pashtuns. But their numbers are quite different.

    1. UP Brahmins=0.000
    2. UP Kshatriya=0.001
    3. Gujarati=0.002, Pashtun=0.002, UP Muslims=0.002, Sindhi=0.002
    4. UP Caste=0.003
    5. UP Kols=0.004
    6. Balochi=0.006, MAD CHA Gonds=0.006
    7. UP Kanjars=0.007, Burusho=0.007, UP Chamar=0.007
    8. AP Velama=0.008, Tajiks=0.008
    9. UP Dharkar=0.009, Makrani=0.009, TN Kurm+tribe=0.009
    10. Brahui=0.010
    11. TN Piramalai Kallars=0.011
    12. Ctr India mix+Nihali=0.012, Uzbeks=0.012
    13. Nogais=0.013
    14. KAR Halakipikki=0.014
    15. Iranians=0.015
    16. Kumyks=0.016
    17. Uygurs=0.017
    18. AP Chenchu=0.017, KAR North Kannadiga=0.017, KER Sakilli=0.017
    19. Balkars=0.018, Hazara=0.018, Turkmen=0.018, Lezgins=0.018
    20. North Ossetians=0.019
    21. UP Dushadh=0.021, Chechens=0.021
    22. Armenians=0.022, Abhkasians=0.022, BIH JHA Munda=0.022
    23. Russians=0.023
    24. Georgians=0.024
    25. KER tribes mix=0.025
    26. French=0.026, Tuscans=0.026, Palestinians=0.026
    27. Druze=0.028, North Italians=0.028, ORI CHA Munda=0.028
    28. Bedouins=0.031
    29. Sardinians=0.041
    30. Mozabites=0.044
    31. Burmese=0.048
    32. TN Pulliyar=0.056
    33. MEG Garos=0.057
    34. Cambodians=0.061
    35. Tu=0.062, Mongols=0.062
    36. Xibo=0.063
    37. Daur=0.070
    38. Oroqen=0.072
    39. Hezhen=0.073
    40. Naxi=0.074, Yizu=0.074
    41. Tujia=0.076
    42. Dai=0.078
    43. Han=0.079
    44. Miaozu=0.080, NAG Nagas=0.080
    45. Japanese=0.081
    46. She=0.084
    47. Lahu=0.086
    48. Bantu NE=0.128
    49. NAN Melanesians=0.137
    50. Bantu SW=0.140
    51. Mandenka=0.141
    52. Yoruba=0.143
    53. Papuans=0.166
    54. Biaka Pygmies=0.173
    55. San=0.200
    56. Mbuti Pygmies=0.201

  33. "If Punjabis, had been included my guess is that they would have been the closest to the Pashtuns."

    I think you are absolutely correct. Going by this list, it seems the closest populations to Pashtuns are neighboring populations (Sindhi, Tajik, Balochi, Burusho, etc), upper caste North Indians (only Brahmins and Kshatriyas), and Iranians. After this, populations from the Trans-Caucasus, North Indians, Europeans, Central Asians, and Southwest Asians, all seem to be fairly close and equidistant. Followed by South Indians. East Asians are the most divergent group of Eurasians, but North Eurasians are closer (with exception to the Cambodians, who have West Eurasian admixture). The really interesting thing is that Sub-Saharan African farmers are closer than Papuans.

    It is fascinating to compare the numbers with the Brahmins below. They are very close to the Pashtuns, but their affinities are moderately different. They are closest to neighboring populations (Up Kshatriya, UP Muslim, UP Kol, etc), Pakistanis, and Gujaratis. They seem to be closer to South Indians than to other West Eurasians (the opposite of the HGDP Pashtuns). Also, they are more divergent from Africans. But this makes me wonder about the Pulliyars. Who are they? They seem very divergent from both Pashtuns and UP Brahmins. I looked at their percentages for HarappaWorld. They have 0% of the "Baloch" component, but 9% of the "Caucasian" component! They don't really resemble other South Indians in an autosomal sense.

    I do wonder what the Fst distances for HAP Pashtuns could be. Since we all have less of the "Baloch" component than the HGDP Pashtuns, we should probably be slightly more divergent from other South Asians. Also, two of the HAP Pashtuns have much lower "South Indian" percentages, which would have the same effect.

  34. what about haryana/UP jatts, they seem to the most non south asian population among indic populations

  35. All Jatt samples are found in open projects like HAP. No academic samples.

  36. ^ i just dont understand how is it possible that there is such a huge difference between the Chamar caste of punjab and Haryana Jatts? Chamar are showing as 66% south indian on the spreadsheet, while Jatts are 27% south indian. But if you go to google and check them out, there is not a large difference between them. Something is strange about these samples

  37. ^ still, UP and Punjab are not far from each other, also haryana borders UP. So the Chamars in the two areas shouldn't be that different genetically. The only possibility I think of it is perhaps every Caste and ethnic groups in India are very diversified, so these numbers don't necessarily reflect all the population of that particular group or caste, that's why I earlier said that just because the 5 samples of Haryana jatt are showing high European dna, doesn't mean every jatt in Haryana will have that amount of european dna in them. To get proper conclusion about a group, we would need massive amounts of samplings

  38. I don't know what Imran Khan found in Google. Jatts of Haryana and Punjab are the same people with the same surnames. All jatts or jaats moved up into Punjab Haryana from Rajasthan where they arrived from Sistan. In both states Jatts are the top community. They don't fit in the normal brahminical system, permitting widow remarriage for example, which was absolutely taboo in Hinduism. Nor did they pay homage to Brahmins who are treated as menials generally. Not to be casteist, but Imran Khan seems to think that jatts and chamars are in similar social categories. Actually Jatts are at the top in Punjab and Haryana and everyone else follows in due order. The Brahmins are below even Khatris in Punjabs. May I also add that if Pathans were Hindus then Brahmins may have put them in the shudra or untouchable category as well.because of their customs. Actually as foreigners they would qualify merely as mlechha. Unclean. Caste is not that strongly connected to genes. There is only a few percentage points up and down in any particular category and that includes Pathans notwithstanding their fair skin relatively speaking. Someone spoke of genotypes and phenotypes the other day. That needs study.

    • I guess that he is looking their (Punjab/Haryana Jats and Chamars) phenotypes and not finding them different.

      Under prevalent Manusmriti practice you are correct about widow-remarriage, but there is nothing mandatory or proscribed in Hindusim - as there are 'Hindu' (then term was not known to them) law givers that allow widow re-marriage. One of them happens to be my namesake!
      Parasara - "When her husband is missing, or is dead, or has renounced the world, or is impotent, or has been degraded by sin,—on any of the said five calamities befalling a woman, law has ordained another husband for her." pg 22

  39. Well of course jatts are completely Indian. Whoever suggested anything else. Colour however is deceptive. Look at Obama he is half white isn't he but would you know it to look at him?
    Pathans are not Indians, no one said they were. But there are 20 million Pathans in Pakistan. Are they Pakistani? How about the Rohilla Pathans who settled in large numbers in India in UP. What do their genes show? You will notice that Bihar Syeds have a lot of ASI genes. Many Brahmins in Punjab are dark skinned. Similarly, many Chamars in Punjab are fair skinned. You can draw your own conclusions. Chamars in fact in Hoshiarpur Pathankot and elsewhere could pass for Brahmins. In Jammu I have seen villages of Meghs (chamar equivalent) who are pink skinned, even in old age. .
    Now if Pathans were never Indian, (though that concept can be contested as well, but we'll let it pass) then they would not have ASI genes, nor would the Baluch leave alone Tajiks and Iranians in whatever small number. How about Kashmiris who also say they are not Indian? Mind you Kashmiri rural stock is almost entirely unmixed with foreign genes though Pandits may have reasons for higher ASI.
    Now if upto a quarter of Pathan DNA is ASI then we need a definition of what Indian means. It isn't anyone's case that South Indians conquered KP; and surely the brief Sikh occupation of Pathan territory was not sufficient to change a quarter of Pathan genes, nor was the brief stay of Syed Ahmad Shah Barelvi and his Indian Muslims during Sikh rule sufficient to bring about such alteration.
    So we need different explanations, not only of "Indian' DNA but of the notion of caste. It began as you may know not on the basis of race but of occupation. Caste was mobile till it hardened by endogamy. The Harvard study only confirms what Indian historians have been saying for decades. When Pathans were Hindu, as many were before they were Buddhist they would have had a caste system in which their cobblers no matter how fair would have been been untouchable, not even Shudra. When the Greeks invaded India they were treated as Mlechha, unclean by Brahmins. Both caste and genes need more understanding than available data allows us. Your assumption that equated dark with Indian and fair with foreign may have some truth in it but it is not the whole truth. What can be true is that the fair gene came from outside India (though that is disputed as well by our RSS types) but all those Aryans, Greeks, Kushans, Sakas, Huns and what not were fair skinned types, though all mlechha, till they beca,e Indianized by absorption into the Brahminical system. That happened to the Greeks as well who turned Buddhist mostly and then Hindu.

    • do you have evidence pashtuns were hindus? also having ASI does not mean Indian, ASI is over 20,000 years old, ASI probably lived as far as Iran in ancient times, so pashtuns showing 18-20% ASI means nothing really. Pashtuns have a completly different culture and way of life from Indians and this alone suggest's their origin can not be India

      Also I define Indian as anyone who looks Indian, simple as that

  40. I should have said that all Chinese look alike to non Chinese, and all Afghans look alike to non Afghans. The poor Sikhs keep getting bashed up in the US in the belief that they are Arabs. I know jats, in Haryana and Punjab who could pass for Afghans if they dressed that way, and quite often I see pictures of Afghans who could be jats except for the way they tie their turban. If you were to visit Delhi and stroll around Connaught Place you would not be able to say make out a Bihari from a Haryanvi let alone a Brahmin from a Chamar. Of course there are types of looks, but I wouldn't bet any money on them.

  41. Generally all north indians looks similar, doesn't matter from Punjab or Bihar, the only thing is that Jats tend to be the taller/well build version of north indians like from UP or Bihar. In my opinion most Jats do not look afghan/pashtuns, their facial features are different

    • Well of course Jats don't look like Pathans. Some may but not as a group. No one said they did. Yes all North Indians tend to look alike, but if you are a North Indian you would, generally speaking, be able to make out if a person originated from North west India or just North India. There is various types of Indian physiognomy, but distinguishing on the basis of caste can be difficult.
      To answer your earlier question about Pathans and Hindus.It is not possible to give within this short space detailed evidence about the Hindu period of Pathans but I’ll try, briefly.
      I don’t know about Pathan origins, they genes speak best, but there is no doubt that the invading Greeks, Kushans, Sakas, Huns, Turks, Mongols and Mughals must have left a substantial impress in the genetic signature of Afghans not to speak of the Hindu period which prevailed between perhaps the 5th and 10th centuries. These same invaders travelled into India and no doubt left their imprint there as well. So the genetic similarity is not surprising.
      Charsadda was Pushkalavati when Alexander invaded and Peshawar was Purushpura. Taxila was Takshashila. When Buddhists missionaries arrived they found substantial populations of Hindus and Zoroastarians in these places. The tribe Mohmand is derived from the Sanskrit Madhumant, the Orakzai were Arasakoi (Arrian) and the Utmankhel, Utzoi. Hazara is the Sanskrit Abhisara. Pakhtun is from the Sanskrit Pakht and Afghan from the Ashvakyani republic. A Sanskitist could do a better job at all this transliteration.
      Tantric Buddhism first developed in Swat, and the Tantric arts are well connected to Hinduism. The name Gandhara first occurs in the Rig Veda in which it is identified with the Vedic religion. Hindus ruled over the Kabul region from around the fourth century as is evident from the references of the Gupta emperor Samudragupta. The Chinese traveller Xuanzang found a Hindu king ruling over Kapisa with control over ten principalities of the area. During this time Buddhism was on the wane and Brahminism was making a comeback in India. These rulers may have been of Turkic origin but were Kshatriya and therefore Hindu. Two Hindu dynasties are known to have ruled in Afghanistan; a Kshatriya Turkic dynasty and a Brahminical Hindu one. Details are forthcoming from coins and chronicles but it is believed that the Kings Jayapala who was finally defeated by Ghazni was a Mohyal Brahmin, a kind of martial Brahmin sect who are also called Husseini Brahmins for their supposed role in the battle at Karbala. Hindu Shahi rule extended to the borders of Sistan.
      Now then if there were Hindu kings ruling Afghanistan for near six hundred years it stands to reason the population would have been of the same faith. There is some evidence of a Shiv temple built in North Afghanistan at the time.
      Pathans don't have to be ashamed that they were Hindus and Buddhists once upon a time. Islam came into their lives only 1200 years ago, but their genes go back for ever.
      It is harder to deal with your statement that "having ASI does not mean Indian, ASI is over 20,000 years old, ASI probably lived as far as Iran in ancient times, so pashtuns showing 18-20% ASI means nothing really." What does Indian mean then? It doesnt mean ASI and It cant mean ANI, Does it mean; Hindu? That would please the RSS.
      "Pashtuns have a completely different culture and way of life from Indians and this alone suggest's their origin can not be India." Right and the claim of Indians to be Indians not being based on their genetic structure must be a Hindu cultural construct? And if it is were Pathans Indians when they were Hindu or Buddhist? This kind of reasoning leads nowhere.

  42. (These same invaders travelled into India and no doubt left their imprint there as well. So the genetic similarity is not surprising.
    Charsadda was Pushkalavati when Alexander invaded and Peshawar was Purushpura. Taxila was Takshashila. When Buddhists missionaries arrived they found substantial populations of Hindus and Zoroastarians in these places. The tribe Mohmand is derived from the Sanskrit Madhumant, the Orakzai were Arasakoi (Arrian) and the Utmankhel, Utzoi. Hazara is the Sanskrit Abhisara. Pakhtun is from the Sanskrit Pakht and Afghan from the Ashvakyani republic. A Sanskitist could do a better job at all this transliteration)

    ^ all this above you seem to be reading from your hindu nationalistic dictionary or something. Bring evidence of all these claims before saying them. There is no doubt Buddists and some hindus were living northern pakistan/afghanistan region, however that doesn't mean people from central indian plains lived there, it's only hindu nationalistic propaganda I am sensing from your talk, Also every people in the region seems to have 15-20% ASI in them, doesn't matter if its Pashtuns, Kalash, Bursho, Baloch etc... it is clear that this is a very ancient component which came long time ago to this region, it hasn't arrived from India recently that is for sure

    • Sir, I an not a Hindu nationalist. I am not even a Hindu. If you want I can suggest history books from which the information is taken, including the memoirs of the Chinese traveller himself so you can read for yourself. I am sorry the facts don't suit the popular narrative in your part of the world, but there is a lot of myth making about origins in all cultures including the Pathan culture. I regret that you should resort to ad hominem attacks to avoid the truth
      I did ask you to explain what Indian is if ASI and ANI are not Indian. If you imply that Indian means brown then I have to say that even colour comes from genes which according to you are not Indian at all. Perhaps they are African, but so are everybody's and they arrived in India via Arabia, or they arrived via Iran, take your pick. Whatever nook you want to assign India too, you cannot be too from it yourself. Pathans may look different but they share about 85 percent of their genes with North Indians, and among North Indian they seem closest genetically to Jatts if the data is accurate. You may not like that but there is no helping it. Cultural memes have less value than genes.

  43. Excerpt from
    "This ill-conceived notion is reinforced by the fact that we in the subcontinent believe all western invaders to be Pakhtuns. Mehr Gul, Mahmud Ghaznavi, Temur, Chengez Khan (he’s Khan, isn’t he?) et al are all Pakhtuns in common understanding.
    I told him who Mehr Gul was and that he predated Islam by more than a century. The man was incredulous. How could this ‘Muslim name’ have been taken by a kafir, he asked indignantly. Mehr Gul means either Sun Rose or Sunflower and it comes from the Persian which was spoken long before Islam came into existence. I tried very hard to convince him that names do not have religions.
    There are Arabic names from pre-Islamic history that were not discarded with the dawn of Islam. These are still in use in the Muslim world. Arabic names were all right, said the ignorant storekeeper. But Mehr Gul was a good Islamic name and a kafir could not be called by it.
    In the end, he ran out of arguments to posthumously turn Mehr Gul into a Muslim Pakhtun. But he clearly thought I was a charlatan who had got the better of him because of superior oratory. His pride that this Pakhtun with an ‘Islamic name’ had been discomfited by Hindus was, however, deeply hurt.
    As I was leaving, I told him to be proud that an alien savage was defeated by one who may well have been our common ancestor. This one really got the man’s goat."

    • It is a common misconception because Muslims, at least in our part of the world, carry names that are of Arabic or Persian origin, Khan is of course a Mongol title now used by many Muslim. Persian Jews have Persian names too and Arab Christians have Arab names as well. Many Indonesian Muslims continue to bear Hindu names though there is a switch recently to Arab names. President Sukarno is just one prominent example along with his daughter Meghawati.

  44. I am a pashtun whose ancestors moved to Pubjab in the early 1700s on my fathers sides and mid 1800s in my mothers side. I'll send my results ASAP, surprisingly I am not S.Indian at all.

    • Send me your data and we'll see.

    • Hey Skizz, are you in the Harappa spreadsheet? Which number are you? Interested in seeing admixture of Pashtun who have been in Pakistan for a while.

      Funny thing is my ancestors moved to Punjab after you and my family believed we were totally Punjabi since always.

  45. ^ whats your tribe if I may I ask. I really think Pashtuns results will be different according to the tribes. some tribes mixed with more with indians, many others didn't

    • Instead of just speculating without any facts at all, get your Pashtun friends to test and send me their data. That will clear it up.

    • All of my grandparents were from different tribes, Mohmand, Burki (Ormur), Lodhi, Ahmadzai

  46. Actually had a question, how much does the whole process cost? many people were asking prices, but didn't I have an answer

  47. I am thinking of submitting my own results, my family is from karachi, but grand parents came from a place near Delhi, which was a heavily afghan area at the time. My whole family doesn't look (indian) and I am pretty sure we were very recent arrivals from outside. I am thinking my results would be similar to some West UP samples here, which show clear outside ancestry, perhaps mine would be even more outside considering I have never been called Indian or even pakistani in my life by others, they usually think of me as Arab or middle easterner

    • Submit your data and we'll figure it out.

      BTW a fair number of people figure me as South Asian despite my quarter non-South-Asian ancestry.

  48. Unlikely to have less than 20% South Indian (but I'm guessing your more in the 25-35% range). I am around that percentage, yet people always assume I'm of southern European, Hispanic, or Middle Eastern descent. Desis always tell me that I am clearly not one of them, based on facial features and skin color (they always ignore my fluent Urdu and Hindi). Pakistanis are always shocked when they find out I originate from the same country as themselves. I'm always seen as unusual, or an outsider. But Arabs talk to me in Arabic, Iranians in Persian, and Mexicans in Spanish. In short, phenotype simply does not equal genotype. You should not make predictions based on something as flimsy and subjective as physical appearance.

    • isn't your south indian 21%? that's much lower then most desis, most punjabis have around 30-35% south indian on the spreadsheet. Also your Caucasian component is 18%, again this is pretty high compared with desis except some UP muslim samples which had almost at your level

      also genes must play some role in our appearance, I think every thing counts, genes, environment, diets etc... all make us look the way we are

      as for me, my grand father and mother use to speak urdu in a very foreign accent, almost in a afghan/Persian accent and like I mentioned the phenotypes are very non desi in general, I am not just focusing on the fairskin, we really do look like Middle easterners for the most part. Also I heard from my mom that some of the people in her family were Shia's before, this may be some clue

  49. Yes, my "South Indian" and "Caucasian" components are almost equal. But the shift is still toward the "South Indian" component. Also, my most substantial component is the "Baloch", which is definitely a South Asian specific "West Eurasian" component (it is definitely West Eurasian, but it is clearly more significant in South Asia, rather than in other parts of West Eurasia. This is why I consider the Tajiks South Asians, since the "Baloch" component is also their top component, rather than the "Caucasian" component).

  50. ^ well south asia is just geographical term, some people even include Iran in it. I think the Baloch component is mostly like where it is found the most, Balochistan, Pakistan, Afghanistan / NW india area. The other central indians, bengalis, nepalis, sri lankans etc... have significantly lower amount of this component ( not including higher castes), so I would actually call this component (western south asian) or even south/central asian. South Indian is the proper South Asian component which is found in those areas I mentioned

    anyways, I think we know that this component makes a huge part of ANI and ANI generally is not considered native to south asia, even though, yes it has been around for a while

  51. Any chance HGDP00220 is a Hazara? Looking like he is definitely part Hazar due to his results.

  52. I thought it would be interesting to show Fsts of the HGDP Pashtuns to different regions and populations, in order of distance. The regions are broadly based on the Metspalu et al Fst chart.

    HGDP Pashtuns:
    1. Pakistan=0.004, Tajikistan=0.004
    2. Iran=0.007
    3. North Caucasus=0.008
    4. Caucasus=0.011, North India+Gujarat=0.011
    5. Southwest Caucasus=0.012
    6. Abhkazians=0.013
    7. Central Asia=0.015
    8. Europe=0.017
    9. Near East=0.020
    10. Sardinia=0.031
    11. North Africa (Mozabites)=0.038
    12. South India=0.044
    13. East Asia=0.071
    14. African Farmers=0.136
    15. Melanesia=0.143
    16. African Hunter Gatherers=0.178

    I modified some categories, because some populations are much closer or distant than others. Also, the "East Asia" category conflates "TB spkr of India", "Southeast Asia", and "East Asia". To be specific, the Northeast Asian populations are closest to the HGDP Pashtuns, as are the Cambodians, who have some West Eurasian ancestry (already mentioned this). Most East Eurasians without any West Eurasian admixture were in the 0.080's range (as can be observed above). I excluded the Munda and Central Indian populations because I made a huge mistake in computation (lol), so it was best I exclude them. All the numbers are correct and double checked. What I find interesting is that with exception to the Gujaratis and North Indians, the HGDP Pashtuns are much closer to all West Eurasians than South Indians (and besides Gujaratis, Brahmins, and Kshatriyas, most North Indians are somewhat distant). Also, the Iranians are closer than North Indians. As expected, the closest populations are average Pakistanis, and Tajiks.

  53. Very intersting blog and quite an obsesion with Pashtun and fair skin. In a study in Brazil many years ago,it was found that the white skin brazillians of nowadays were actually of afro origins and the white brazillians who appeared now dark were actually of portugeuse european origins. Colour pigmentation dont tell the whole story. In South africa,my country there is the Coloured Community. Although they are classified non white ,they sometimes look more white than evn Afrikaners who are of north european extract.The Coloured are actually a combination of malay,(javanese,sundanese,bughese,timorese),south asians(bengali,kerala,andra paradesh), khoisan and white people and some east african.Yet they are the most beautiful south africans. They are tan skinned and not white . White does not mean beautiful as Imran khan seems to think.
    Why i am intersted i am also Pashtun who migrated more than a 100 years ago. Pashtun are originally from Xinjiang province in China,also known as East turkestan. They are very late in afghanistan therefore they were never hindu. They are east iranic saka descendent from tarim migrated via gorno badakhshan tajikistan into bakhtar and mixed with dardic people,10 percent of their genes and than into east and south afghanistan crossing the hindu kush at westerly direction. Therefore Pashtun are least affected by hinduism. They were certainly some buddist but in majority they had their pagan pashtunwali cult which survived islam and evn today it is like a religion for pashtun. Pashtun are east iranic,not west iranic like afghanistan they must have incorporated some of the previous indic peoples ways who were probably some hindu ,but on the whole pashtun were never hindu. some tribes in the east did have associations with india and trade with india was certainly known.

    Both pashtun and tajik are the children of bactrians ,moreso the tajik, pashtun are combination of saka,bactrian,hepthalites and kushana. they form the ethnogenesis of the modern pashtun. the reason pashtun are so different and the tribal structure is so tight is because of their eastern turkestan origin before the turks moved into tarim basin of northwest china, or wher the uygurs stay today.

  54. The original land of the Pashtun is UYGURSTAN,East Turkestan ,land of the East iranian people. central asia,uzbekitsan,turkmenistan,kazakhstan and northwest china ,the traditional land of the IRANIC not iranian race. The barabric turks who were late commers to civilization than had civil war among one another and the mongols expelled the kyrgyz and they in turn expelled the uygurs from mongolia where they eventually landed in the land of the Aryans,the tajik,pashtun,persians, sogdians,bactrians etc. Xinjiang the original home of Pashtun. sons of saka ,subtribe of scythians. The khotanese, and tumshuq language of tarim is till very close to Pashto and Wakhi. The mummys found in Uygurstan are like Pashtun people,iranic. the uyghurs later married some of the Pashtun forefathers therefore uygur is cross between turkic and iranic race.

    Pashto culture is different because they are not indian,not because they are fair.many Pashtuna re not fair they are beautifully brown and some red brown. I am fed up of the obsession with pigmentation as it blurs the true hsitory of a people. It is an obsession which some Pashtun have learned from the Brahminical hindus.

  55. @Reza Khan,

    According to the dominant scholarly view, the Pashtun/Afghan ethnic group has been living in its current location (=what is now southern Afghanistan and northern Pakistan) since the Late Bronze Age. To the ancient Indians they were known as "Paktha", "Asvaka" and "Avagana"; to the ancient Greeks they were known as "Pactyans", "Aspasii" and "Assaceni"; to the ancient Persians they were known as "Abgan"; to the Muslim Arabs and other Muslims they were known as "Afghan"; and to the recent Indians they were known as "Pathan". So there is no medieval migration from what is now Xinjiang or elsewhere to the current Pashtun territory, as they were already living in their current land for more than 2000 years.

    • According to the dominant scholarly view, the Pashtun/Afghan ethnic group has been living in its current location (=what is now southern Afghanistan and northern Pakistan) since the Late Bronze Age... as they were already living in their current land for more than 2000 years.

      Actually not their current location/land as a whole, only the part in and around the Sulayman Mountains.

  56. I beg to differ from you when you say the dominant scholarly view. Who are these scholars. With the origin of Pshtun there are a lot of Pashtun that are close minded and dont want the truth out about themself. Imran Khan is one such individual that seem to think denigrating others and racism will deflect others from raising real tough questions about Pashtun origin. The saka ,uygur, kazakhstan origin is is total harmony with dna ,historic and anthropological and linguistic studies of origin of Pashtun people. A combination of these give us a clearer picture of Pashtun. Pashtun are latecomers to afghanistan as well as Pakistan. In Gandharan times they hardly existed. these were buddhist area probably inhabited by hindkowans. The fact that arabs dont mention them as well as the Ghorids did not call themself Pashtun are all instances that tally with the idea Pashtun are latecomers to southern and eastern afghanistan as well as in pakistan they only moved there in the 14 and 15 th centuary ad. This is consistent with Pashtun behaviour as nomadic people ,they tend to move evrywhere,like india and even abroad. The Pashtun also never build any strong land based,sedentary culture. They are fo the most temporary sojourners and mercenaries as their occupation.

  57. Hello Reza,

    I think Onur has described things rather accurately. I couldn't have said it better myself.

    Also, Pashtuns are far from nomadic. With exception to the Kochi, most Pashtun tribes traditionally practice subsistence agriculture. The Kochi Pashtun tribes are not even entirely nomadic, nor do they even constitute exclusive lineage groups, unlike many Bedouin tribal units. One finds Nurzai sub-lineages that have been sedentary as far back as their collective memory goes, and one finds Nurzai sub-lineages that have been nomadic pastoralists as far back as their collective memory goes. Same applies to the other Kochi tribes, most of whom are Ghilzai, and have substantial sedentary populations. Although, it is true that most Pashtun tribes are very mobile in terms of wholesale population movement. Inter-tribal warfare often forces weaker tribes to move into new areas, and ambitious tribes rarely miss the chance to expand into the territories of their enemies. The Mohmand are much further to the east of where Babur encountered them, and their tribal chronicles support a relatively recent migration. The Kurram Valley has seen a lot of dynamism, with the Bangash pushing into the area, and then being shoved aside by the Turi. The Zadran and Mangal borderline has always had a tendency to fluctuate, based on which tribe has the upper hand. And Kandahar has shifted between periods of tenuous Ghilzai and Durrani dominance, on multiple occasions.

    Nevertheless, some tribes are strikingly immobile and deep rooted. I wouldn't be surprised if the Mehsud and Waziri are found to have inhabited that odd tangle of impenetrable highland they call home for well over 2,000 years, without any tribal encroachment by related Pashtuns (or anyone for that matter, out of all Pashtun tribes, the Mehsud and Waziri can genuinely claim to have never been subjugated by any foreign power), or any movement of these tribes outside their own isolated eyries. I'm willing to bet that if any samples from these tribes are obtained, they will break out into their own uninformative clusters in Admixture/Structure, due to the long term isolation of the area. Basically, a repeat of the Kalash. But that's really just speculation, and I wouldn't be particularly shocked if I'm wrong. Genes seem to have a way of spreading, despite the existence of substantial geographic barriers.

    • There were some transitions in the area

      Tajiks went from Sogdian(Eastern Iranian) to Dari(Western Iranian) because of the Samanids.

      Pashtuns went from Gandharan(Indo-Aryan) to Pashto(Eastern Iranian) for a reason that I will allow others to answer.

      The Sogdians and Gandharans never disappeared as genetic groups, they adopted different languages and now people are wondering what happened to them.

      I think the whole naming issue is what is really confusing people. If Pashtuns simply renamed themselves Gandharans then everything would be clear again because there would be fewer gaps in the historical record and the genetics would make much more sense as well. 🙂

      • @HRP0286

        Yours is one of the most cogent explanations I have seen.

        • The most nonsensical explanation one can ever hear. The Khakhi tribal confederacy just recently moved from Kanahar-Ghor region to Peshawar and Swat Valleys (in 15th Century). If you are looking for actual Gandarian probably you look at Potothari people.

          • Not that I agree with you, but I would be interested in seeing a lot more representation from Hindkowans, Kamboj, and Pahari people of all castes (especially a comparison between the Pahari Jatts and Punjabi Jatts).

      • That is ridiculously wrong. When Khakhi Khel tribal confederacy (including tribes like Youafzais, Mohammadzais, Utmankhel, etc.), they still were mostly nomads so much so that the distribution system of the land they conquered in Swat and Peshawar Valley was based on a migratory pattern. This was called "da Shiekh Mali daftar" (register of Shiekh Mali). Every 10 years, the land would be redistributed and tribes would relocate to their newly allotted lands.

        As for Wazirs and Mehsuds, they are a new arrival in the area probably located farther north into Afghanistan. Mehsuds are an offshoot of Wazirs which seems to be an offshoot of Urmar. Khattak lived in Shawal area before they were displaced by Wazris.

        Speculation on the basis of eponymous similarity of Asvagan and Pakhtha would be wrong. The transformation of Pectues into ethnonym Pakhtun is also not supported by linguistic rules (linguists say Pakhtun name was derived from Parthwa).

  58. Also I don't mean to say that all Gandharans became Pashtuns but the ones living in Swat, Waziristan, the surroundings of Jalalabad, Kabul and Peshawar did. This was probably the first instance of "Pashtunization"

  59. I think you make a very interesting point here. What strikes me about the pre-Islamic era is the strong connection between "Afghanistan proper" (the Pashtun areas of Afghanistan, along with northwestern Pakistan), and greater India. One thing worth mentioning is the name of the mountain range that defines this region. "Hindu Kush" is actually a corruption of "Hindu Koh". It is thought these mountains were so named because most people to its south were Hindu-Buddhist (and thus Indians), while most people to it's north were Zoroastrians (although Buddhism did eventually spill over to the north as well). Nevertheless, even before Islam, it was also understood by most that the region currently inhabited by Pashtuns, while intimately tied with greater India, was still not really a part of India proper. In fact, the Persians strongly distinguished Gandhara from "India" (modern Punjab and Sindh). For some reason, it's always been assumed that you're not "really" in India until you get east of the Indus.

    • The Pakhtas were a Rigvedic tribe that is to say they were Indo-Aryan in nature, not Iranian like the modern-day Pashtuns, so there is a pitfall if you want to say that Pakhtas = Pashtuns.

      Its easier to believe that Pashtuns, coming from some northerly place such as Badakhshan and the Pamirs(probably because its the place where Eastern Iranian language diversity peaks), came to inhabit some mountain defiles along the Indus.

      They dwelt for a while and formed tribes before opportunistically expanding into the surrounding sedentary area where they assimilated local Indo-Aryan groups thus making their genetic pool more similar to Indo-Aryans but still preserving some of their Iranian affinities.

      In this way, they were very similar to the Turkic groups who migrated to the plains around Iranian speakers in the Central Asian Steppe before opportunistically taking over the sedentary people.

      I would say the Oghuz model is most analogous because the Oghuz Turk leaders did not destroy the local Tajik population in Turkmenistan, but because of their opposition to local Ghaznavid rule they managed to take over without destroying the cities. So they assimilated the people and left a light genetic impact on most of them while leaving their language behind on the majority of modern day Turkmen.

      That seems to be about right for the Pashtunization of the areas as well: Light genetic impact on most but heavy cultural/linguistic change from Indo-Aryan to Eastern Iranian.

  60. Also, Gandhara was considered India by at least a few. In 240 Mani considered to have been in India when he went to Gandhara for study. Xuanzang in the 600s considered himself in India once he reached Peshawar. What defines India and what doesn't is not very clear but I think that people speaking Indo-Aryan languages and practicing Indo-Aryan religions was probably enough for outsiders in the past to consider a place "Indian."

  61. Frye on Pashtuns in Handbuch der Altertumswissenschaft: Alter Orient:

    "For we must remember that the final movement into the Indo-Iranian borderlands, in Islamic times, is the expansion of Iranian Pathans or Pashtuns from their homeland in the mountains of Afghanistan onto the plains of the northwest Subcontinent."

    On Pamiris/Pashtuns and the question of Pamiri origins Sergei ANDREYEV, D. Phil. (Oxon)
    Civil Affairs Officer:

    "In a strict linguistic sense the term "Pamiri languages" can be used only
    tentatively, since it is still unclear whether these languages are inter-related closely enough in order to constitute a single group. While all linguistic authorities agree that they belong to the East Iranian group, some Pamiri languages have more common traits with Pashto - the most numerous East Iranian language, rather than with each other. Only Shughni and Yazghulami display enough similarities allowing to trace their common origin to a not that distant past and put them into one category. At the same time, other Pamiri languages preserved a number of features setting them apart 65 Recreational and medical consumption of raw opium has always been a characteristic trait of the Pamiri lifestyle. from more closely related Shughni and Yazghulami. Thus, it still remains unclear whether Pamiri languages originate from a single proto-Pamiri source or did they develop from separate branches of a more ancient group of Iranian languages, presumably Saka (Edel'man, 1989, p.330; Morgenstieme, 1938, vol.II, p.XVIII; Sokolova, 1967, pp.3-6, 21, 124-125). What unites the Pamiri languages on the extragenetic level is the preservation of a significant number of archaic features that can be traced back to the ancient Iranian languages and a uniform influence of Tajiki/Dari on all of them.

    From the UN Special Mission to Afghanistan

  62. You know, I have to agree, it is a very tempting idea. Nevertheless, I feel an expansion from the north might be problematic. The people who live in the Pamirs are rather tractable, and their culture is, as far as I'm aware, very peaceful. As a fellow Pashtun, I'm sure you share my experience in this regard, Pashtuns aren't the easiest people to deal with ;-). I just can't see a population from that direction expanding and conquering the large congeries of people who inhabit the borderland between ancient Iran and ancient India.

    Nevertheless, my intuition here is relatively weak tea. Culture is a very dynamic set of varied processes, with the only real constant being change (some people will contest this in regard to culture, but I have my arguments). I can imagine that, under this model, the Indo-Aryan ancestors of Pashtuns were a much more tractable set of peoples than their descendants. And perhaps some combination of socio-ecological dynamics in the southern portion of Central Asia pushed Eastern Iranian speakers into the present Pashtun region. It's a very plausible scenario. But I'd like to add one more data point in this regard. In the new Afghan genetics paper, one northern Afghan Tajik is clustering with the HGDP Pakistani Pashtuns on the PCA plot, and is slightly more South Asian shifted than the Afghan Pashtun samples. Now, there is no doubt that the Tajiks of that area have been Iranian in a cultural sense for a very long time, and the area in question is well outside ancient Gandhara. But the individual is substantially more South Asian than Tajiks from Tajikistan. So, the genetically intermediate nature of Pashtuns versus Iranians and northwestern South Asians, as well as the substantial Turkic input we also see, could be explained by taking geography into account. Also, we don't really have samples from eastern Iran. If we did, I wouldn't be surprised if they showed substantial similarity to Pashtuns.

  63. I just saw this, thank you very much for the pdf! I really appreciate that.

  64. The actual relationship between Pathan and Indian is similar to southerly Mongol/Khitan and Chinese. The northern siberian groups aroubd Baikal were never as aggressive as the Mongols and the northern Pamiri groups havent been either (recently but the past was probably different).

  65. If Pashtuns arrived at the modern-day Pashtun lands during the Islamic times, such a migration would surely be recorded, as migrations in and around Central Asia (e.g., the Turkic and Mongolic migrations) were well recorded by Muslims. Where are the records of such a Pashtun migration?

    • By "Pashtun lands" I meant the historical Pashtun territories in around the Sulayman Mountains, not the ones acquired by Pashtuns after the time of Genghis Khan. So the phrase "modern-day" was an error.

    • So let me rephrase my question: where are the records of a Pashtun migration to the area in and around the Sulayman Mountains?

  66. Biruni mentioned the Afghans as a tribe living to the east of Ghazni as wild barbarians basically. The migration was an expansion from those mountain defiles and Gandhara never survived to record what happened. it was a barren waste after the Ghaznavids who probablyAlso, one group of Pathans sided with the Ghaznavids and another sided with Jayapala in their struggle.

    The movement of Pashtuns into the mountains happened before the Islamic era and their expansion into subsequent Indo Aryan locations happened afterwards. The movement of Eastern Iranians into the subcontinent has been recorded, there was an entire era of Hunnic movement in that direction after the fall of the Greeks and before the Sassanids finally managed to conquer many hundreds of years later.

    • Well, I know that the Islamic-era Pashtun migrations to the non-Pashtun locations are recorded one way or another either through direct references or through reports of the ethnic composition of the respective locations prior to and subsequent to the Pashtun migrations. What I want to see is the records of Pashtun migrations prior to the Islamic era. When taken together, all the available historical references, whether Islamic or pre-Islamic, seem to indicate that Pashtuns have inhabited the region around the Sulayman Mountains since probably the Late Bronze Age.

      • This is probably one of the most ridiculous explanation that one can have of Pashtun ethnogenesis, i.e., Pashtuns (over 50 millions) lived around Suleman mountains since Bronze age and became what they are now. The more plausible explanation is when Turkic expansion pushed East Iranians southward into Afghanistan, whoever the aboriginal people that were living there were farther pushed south into the Gangetic plains. Some might have been assimilated into Pashtuns. But this Suleman mountain and Bronze age theory sounds to naive and simplistic (might be connected to South Asianism tendency).

  67. You wont find the answers you are looking for if you want to find a tribe called Pashtuns but you will find some clues that I am trying to point you toward if you broaden your search to include Eastern Iranians who pushed on and on into the Indus river right after the collapse of the Greek kingdom. There was hundreds of years of Eastern Iranian nomadic invasion into that area. They were pushed into the area by the Turkic invasions from the northeast.

    • If Pashtuns arrived at the region in and around the Sulaiman Mountains after the collapse of the Graeco-Indian Kingdom and before the rise of the Sassanids, then who were the Paktha mentioned in the Rigveda (the 2nd millennium BCE, the Late Bronze Age) and Pactyans mentioned by Herodotus (the mid 1st millennium BCE), all living in the region in and around the Sulaiman Mountains and bearing a name very similar to the self-designation of modern-day Pashtuns (who also call themselves Pakhtun)? Also, the Aparytae tribe mentioned by Herodotus and the Aprit tribe mentioned by Panini (the 4th century BCE) are probably identical to the modern-day Pashtun tribe Afridi (who call themselves Apriday), and the Madhumant tribe mentioned by Panini is probably identical to the modern-day Pashtun tribe Mohmand; they are all located in the region in and around the Sulaiman Mountains.

      • Let me ask you a question. Why dont any of these tribal names have the suffix ZAI? The Arachosians used Zoi and the Sakas used Zai. Ashkenazi is an example of the word zai as inherited from them.

        sco =ruler, chief /scyth ( Sco-lo-ti, Sco-pa-sis, A-sku-zai)

        I have run into your argument before and it reminds me of this.

        "In a nutshell, Bellew's thesis is that all Afghan tribal names can be traced to Greek and Rajput names, which posits the further possibility of a great Greek mixing with the ancient border tribes of India. Some of this survives in Punjab's Jhang district today where local inhabitants are conscious of homophonous similarities between their names and the great Greek tribes.

        Bellew looks at the zai and khel suffixes indicating Pakhtun bloodlines. He thinks that zai is from Persian zaadan (to give birth) which is the same as Sanskrit jan; and khel is clearly Sanskrit kul (family). The Hindu name Kuldip means lamp of the family. The Pakhtun use zai and khel interchangeably.

        Bellew starts with the mythology of the origin of the Afghans - perhaps the most detailed story given anywhere. Then he goes to the great Greek historian Herodotus when he discusses the Greek-Bactrian tribes North of Afghanistan.

        The Lydoi (Greek 'y' is actually 'u') are the Lodis, Maionoi are the Miyanis, Mysoi are the Afghan tribes taking Musa as prefix, Thynoi and Bithynoi are Tanis and Bitanis, the Karoi are Karo, Ionoi are Yunus, Doroi are Dor, and Aioloi are Ali.

        It should be noted that wherever possible the Afghans will try to convert their pagan names to Muslim ones, as Isapzais have become Yusufzais. This also inclines them to trace themselves to Jewish roots. Bellew gives us the other dimension: all these Greek-sounding names are also Rajput, meaning that Greek intermixing was with the Rajput races when they lived in the region now occupied by the Afghans.

        Bellew thinks prefix Suleman is derived from Rajput Solan which is today visible in Solanki. Daud, as it appears in Daudzai and Daudputra among Muslims, is actually Rajput Dadi or Dadika. Utmankhel or Utmanzai (to which the family of Wali Khan belongs) are mentioned by Herodotus as a Greek tribe Utoi. Utmanzais have sub-tribes like Baddo (Rajput Yaddo, the tribe of Krishna), Ballo is Rajput Bhalla khatri, Bura is Bora (Vohra) mercantile Rajput, a name taken by Bohras, the Ismailis of Gujrat, Mandal is the Jat tribe Mada, its version Mandanr, live along Jadun or Gadun tribes (of Hazara which is Sanskrit Abhisara), which names are variant of the Jadu Rajput tribe. These are Yadavas of India.

        Gaduns established Gajni which is today Ghazni. The Afghan Batanis are ancient Bhattis, the elite of the Rajputs serving at the court as ministers. Mahmand actually means 'the great Mand'. They are in Peshawar but their Rajput relatives are now found near Bombay. Pliny calls them Mandriani of Afghanistan; they are the Wends of Austria. A branch of them called the Bai-zai are located in Kohat which was an old Greek city.

        The Suri Pakhtun were people brought from Syria by the son of Seleukus who ruled that part of Alexander's eastern empire. The Afridis are mentioned by Herodotus as Aparytai brought to their present abode by Ghaznavi, but they came from the Afghan province of Maimana.

        Similarly, the Orakzai are mentioned by Arrian as Arasakoi, and their rivals Bangash came originally from Ghazni. The Bangash are also called Bangak which relates to Bangat Chohan Rajputs. Their neighbours the Turis are the same as Tiwari Rajputs of India. Thus the story of Pakhtun tribes goes on.

        Also a lame attempt but perhaps better than the idea that Pashtuns were Eastern Iranian speakers at the time of the Dasarajna and never managed to record anything into their language until the 1300s in spite of being a huge tribal conglomeration and surrounded by capitals of Kharosthi scholarship.

      • It's possible the terms Paktha and Pactyan were region-specific. Just as how "Punjabis" refers to various people who are in Punjab now, but the story of how they got to Punjab differs. The origins of various Pashtun/Pathan groups are probably quite different (I still think the Hephthalites were absorbed into their confederation of tribes) but they settled into the local culture and language which is most likely Eastern Iranian in origin.

  68. HRP0286,

    Could you be more specific? Which populations do you think cotributed to the ethnogenesis of Pashtuns after the collapse of the Graeco-Indian Kingdom?

    By the way, what I wrote above has nothing to do with Bellew's Graeco-Rajput hypothesis of Pashtun origins. Also I gave very specific references that are much more likely to be pointing to Pashtuns than Bellew's references in general are.

  69. As I keep repeating,

    Pakthas were NOT a Eastern Iranian tribe as the Northeastern Iranian group which Pashto split from did NOT exist at that time.

    Eastern Iranian itself only split off from Middle Iranian sometime around the 4th century BC.

    It was not spoken in the Suleiman Mountains when the Pakthas were in the vicinity.

    Actually, Bellow is more accurate than you are because he acknowledges that the Pathans have an Indo-Aryan basis and he attempts to demonstrate that by showing the close relatedness between IndoAryan and Pathan tribal names.

    Who were the cause?

    The Pathans admixture date: 2117 years ago (Around 117 BC)


    Major event in 160 BC: Collapse of the Greco-Bactrians at the hands of the Eastern Iranian tribes known as Sakas

    Seems to be the immediate aftermath of the Greco-Bactrian collapse at the hands of the Eastern Iranians nomadic tribes living in Northern Afghanistan. I like to let people do their own research instead of me having to name everything so I won't name them.

    Anyways, I am finished talking about this because its a waste of time to discuss matters with people who I have to keep repeating myself to.

  70. Sorry I meant Bactrian split off from Eastern Iranian in the 4th century BC. Avestan at about 1000 BC.

    The composition of the Rigveda? 1700–1100 BC

    So the Pakthas, part of the Indo-Aryan Purus confederacy in the Dasarajna, could not have possibly spoken Pashto a language which split off from Eastern Iranian far far later in history judging from its affinities with Khotanese Saka and Ossetic.

  71. Pashto has IndoAryan characteristics. I once lived and studied in China and a Chinese Tajik friend of mine from Tashkurgan(a Sariqoli speaker) said he thought Pathans were speaking a highly Indian-influenced version of his own language when they would come to Xinjiang trading carpets and other things.

  72. I fully expected Pashtuns to show lower ASI, lower Balochi, higher Caucasian, higher NE European.

  73. Interestingly, most of these Indo-Aryan features are actually shared between old Iranian and old Indo-Aryan. Also, the Pamiri languages display these same exact Indo-Aryan features. So, either the Indo-Aryan influence is very ancient, dating to a time when most of these languages weren't highly differentiated, or Eastern Iranian languages simply retain a more archaic structure. In that case, Persian simply lost these features. In fact, the very nice paper you've referenced actually states that the latter is the most plausible scenario. It states that Pashto has been a very conservative language, and that nearly all of these shared features with Indo-Aryan are archaic Iranian. There are recent examples of Indo-Aryan influence, but almost all of it involves lexical material. I have some subjective insight into this. I'm very intimately familiar with Pashto, Urdu, Hindi, and to a far lesser extent, Punjabi and Hindko. I speak the first three languages very fluently, and Pashto is the language of my childhood. I have reasonable comprehension when it comes to the last two languages, but I can't manage a conversation (I just bob my head up and down :-)). With this subjective experience in mind, Pashto does not really have a South Asian cadence. For example, I've heard Balochi a lot, and it sounds like a distinctly South Asian language, despite being Iranian. This is why most early anthropologists thought the Balochi language was a mix of Sindhi, Persian, and Seraiki, before serious linguistic research showed that it was an unambiguously Iranian language. Pashto, by contrast, has always been recognized as being rather distinct from neighboring Indo-Aryan languages. It just doesn't sound South Asian influenced.

    • So Pashto being surrounded by Indo-Aryans managed to escape Indo-Aryan influences in their language even though they adopted some Indo-Aryan vocabulary. And it is more conservative than languages like Sariqoli and Wakhi...Sounds quite romantic to me.

      You only read the parts that said that Pashto was very conservative because it rejected adoption of the Indian aspiration and retaining the vital parts of Eastern Iranian retroflex but you missed the part where it said that it did in fact adopt Indo-Aryan phonetic system.

      "Cerebrals are found, not only in Ind. Loanwords, but also in many words of uncertain origin, containing un -Indian sounds like x or z «
      Thus Psht., and especially the Peshawar d ialect, has been largely Indianized in its phonetic system; but it is worthy of note that it has entirely rejected the [contrastive MMTH] aspiration of consonants The Ind. loan words in Psht. Are generally drawn from modern Hindostani or Lahnda (in contrast with the remarks of Darmesteter «[p. XVI ] the latter source is by far the more abundant)."

      So it adopted Indian phonetic system, rejected the aspiration, retained the Eastern Iranian retroflex sounds, and has many Indo-Aryan loan words. So it is very conservative compared to Persian but its still highly Indian-influenced.

    • I wouldn't place too much importance on the 'sound' of Balochi. The Balochi most commonly heard in Pakistan these days is the Lyari/Karachi flavor of the southern Makrani dialect. This is heavily Urdu/Hindi influenced and very much shows the influence in the accent. By the same token, the Rakhshani dialect as spoken in Iranian Baluchistan is influenced by Persian and to a non Baloch sounds very much like Farsi. As an example, listen to the dialog in this video.

      One of the 3 is a Baloch from Karachi and his accent is clearly distinguishible from the others. Even in Pakistan, the standard/prestigious dialect is the Rakhshani dialect despite the numeric prevalence of the Makrani flavor. I suspect that the Balochi spoken a hundred years ago sounded quite different from that of today.

      • Very interesting. I'm assuming the individual on the right side throughout the video is from Karachi? The cadence of his speech is usually what comes to my mind when thinking of spoken Balochi. The other two individuals do sound rather different.

  74. I agree with HRP0282 that the linguistic paper that HRP0286 gave link to does not support his conclusion of any significant degree of Indo-Aryan influence in Pashto but rather refutes it (that is why the word "Indic" is written with quotation marks in the title of the paper).

    As for HRP0286's other link, the estimation of the ANI-ASI admixture date for Pashtuns (=Pathans) cannot be used to support his conclusion about the arrival time of Pashtuns to their current location around the Sulaiman Mountains for various reasons: 1- Such date estimations of admixtures usually give estimation of the date of the last major admixture event, and very roughly at that. 2- Some of the Indo-Aryan-speaking populations have more recent ANI-ASI admixture date estimates. 3- There is nothing that indicates that the ANI-ASI admixture date estimate of Pashtuns should be identical to the time of their arrival to their current location around the Sulaiman Mountains.

    Ancient Pashtuns did not need to be Indo-Aryan speakers in order to participate in the newly-developing Hindu Vedic culture and religion and the Hindu political events and groupings; the ancient Iranic culture, languages and religious traditions were close enough to the ancient Indo-Aryan ones (there was not even a distinction between the Iranic and Indo-Aryan peoples yet; all were termed as "Arya" and its variants).

  75. @Onur

    You didn't read the paper thoroughly enough and it indicated that it is very conservative compared to Western Iranian, and in fact most Eastern Iranian languages. But it did adopt the IndoAryan phonetic system, Peshawar did so especially.

    You don't have much evidence besides that Pakhtas sounds like Pakhtun. That doesn't indicate anything about the language they spoke.

    Please provide evidence of an Eastern Iranian, or just any Iranian, language being spoken anywhere East of Ghazni before the fall of the Greco-Bactrian kingdom.

  76. I meant to say that most Eastern Iranian languages are conservative compared to Western Iranian.

    And waiting to hear evidence of Iranian and not Indo-Aryan "pashto" being spoken in any form east of Ghazni, somewhere near the Suleiman mountains, on the banks of the Indus, or any other plausible place for Pashtuns to reside in Pakistan.

  77. Here is Pashto splitting off from Khotanese Saka date : Est. 300 BC

    Pashto splitting from Ossetic: 450 BC

    Starostin's Dendogram

    Pakthas: 1700-1100 BC

    Fatal error in logic to say that Pakthas are Pashto speakers.

  78. HRP0286,

    At least I presented positive evidence for my assertions. You did not present any positive evidence for your assertions other than the quite irrelevant ANI-ASI admixture date estimates.

    As for the linguistic paper, it speaks of only a small Indo-Aryan phonetic influence on Pashtun, by no means a large-scale one as it refutes such a large-scale phonetic influence.

  79. HRP0286, I have just seen your newest post, so I will write a new reply now.

  80. As for your newest post, I do not feel qualified to speak about estimations of language split times. Maybe someone more learned in linguistics can chime in.

  81. I think Onur has pointed out something very relevant to this discussion. There is a rather thin line between Iranian and Indo-Aryan culture. The region stretching from eastern Iran to northwestern India has always constituted a rather culturally cohesive region, one with it's own set of individuative characteristics, in direct contrast to western Iran, or Gangetic India. And before Islam, the similarities would have been even stronger and more robust. There is a clear intellectual thread linking the Avesta and the Rigveda. Islam, and then the arrival of Turkic peoples, changed the equation in subtle, and not so subtle, ways.

    In terms of phonetics, "hard" northern Pashto (Pukhto) has seen the most Indo-Aryan influence (this dialect of Pashto encompasses Laghman, Kunar, and Nangarhar in Afghanistan, and the Peshawar valley in Pakistan). That makes perfect sense, as this area of the Pashtun region has always been home to large populations of Indo-Aryan speakers. "Soft" southern Pashto, as epitomized by the dialect of Kandahari Pashtuns, is the least Indo-Aryan influenced, but has some traits co-developed with, or acquired from, Farsi. That also makes sense, as the Kandahar region has only recently been Pashtunized, and was traditionally home to Farsi speakers. Central Pashto (spoken in Paktia, Khost, Kurram), the Pashto spoken by hill tribesman on the Afghanistan-Pakistan border (Afridi, Turi, Orakzai, etc), and the "soft" Pashto of the Waziri and Mehsud, have seen minimal Indo-Aryan and Farsi influence.

  82. HRP0282 can you explain Pashto splitting off from Saka in 300 BC? Onur does not feel qualified to make any statements about it.

    Its obvious that there is an Indo-Aryan influence, especially in the Northeast around Jalalabad, Peshawar, etc.

    The timeline for a Pashto-speaking Pakthas is not possible in my opinion because it would be that an Eastern Iranian language penetrates earlier and farther than Bactrian and Avestan did.

    Sergei Starostin`s monumental study just adds to evidence that it split from Saka. It actually provides dates for the splits. And it is the only major attempt to date the splitting of Iranian languages, other statements like Pashto descends from Avestan and things like that were just scholars making educated guesses.

  83. Well, to be completely honest, I'm not sure it needs any explaining. The questions of Pashtun ethnogenesis are shrouded in such deep mystery, that I personally don't maintain any strong views. With no written sources, no serious archeological work in the region, no substantial ethnographic perspectives, no concerted linguistic analyses, and no serious historians with trans-disciplinary aspirations, it's pretty vague territory. Not a lot we can say with certainty. I wouldn't be surprised if what your claiming turns out to be correct, or what Onur is claiming turns out to be correct.

    Nevertheless, one thing that Onur is arguing for is strongly suggestive. A lot of evidence suggests that the Sulaiman Mountain range is where Pashtun ethnogenesis took place. First off, every Pashtun confederacy is derived from this area. The Abdali/Durrani are an offshoot of the Tarin of the Sulaiman mountains, as are the Pashtuns indigenous to Peshawar. The Safi in the far northwest are derived from the Kakar in the far southeast. The Ghilzai are tied to the Bhittani and Marwat, close to the area in question. Only Karlanri Pashtuns lack a constitutive connection to the Sulaiman mountains. But unsurprisingly, they are the only group of Pashtuns whose origins are often questioned by fellow Pashtuns. Lets be frank here, why would people as geographically distant as Peshawari Pashtuns, Kandahari Pashtuns, and Pashtuns in the hills near Quetta, be regarded as close relatives? What underlying reason could account for tying together the Safi Pashtuns of Laghman and Kunar with the Kakar Pashtuns of the Sulaimani area? I used to think some kind of political dynamic was at play when these genealogies were being produced, but that is just speculation, I can't really find any. The most parsimonious explanation is that the derived groups once lived in the Sulaimani area, and are closely related to these geographically distant tribes. At some point in history, the Tarin split into separate lineages. One moved west, producing the Durrani confederacy, while one moved far to the northeast, producing the Yusufzai, Khalil, and related groups (Sarbanri). The same would apply to the Kakar and Safi (Ghurghasti) , as well as the Batanri lineages. Another thing to consider is the fact that Arabs used to refer to all Pashtuns as "Sulaimani". Besides this, the biggest support for what I'm trying to say here is linguistic diversity. The most divergent dialects of Pashto are found in the Sulaimani area. Some varieties of Pashto spoken there are so distinctive, that linguists have debated whether they are actually Pashto dialects, or closely related languages to Pashto. To be honest, I think both of you may be partly right. Perhaps a synthesis of both of your claims would constitute the most plausible scenario?

  84. Hi. My name is Farouk Khan. I live in NJ,USA. I was born in Bijnor,Uttar Pradesh India. My family has always believed that we have Pathan(Pashtun) ancestry. it is true that my first cousin's mother's family is from Peshawar. I just took my dna test from and they showed me that I am 86% South Asian,9% Central Asian and 5% other. I am very confused because I have heard that so many people from Northern India who have last name Khan and the claims of Pathan ancestry are not Pathan at all but those people are either hindu converts to islam or Rajputs. I think I am Pathan. I have very light skin like middle easterners or white people. Please help.

    • I broke down the new Afghan samples using admixtools. The formula is the Moorjani F4 formula using basque, yoruba, afghan samples, georgians and dai. So the only difference is the Onge, since I don't have onge with me. Results below

      New Pashtun samples : 75.95 ANI/West Eurasian
      Hazaras : 60.11
      Turkmen : 69.48
      Tadjik : 66.17
      Uzbek : 66.026

      I am using around 25k SNPs only to get quicker results.

      Each population has around 5 samples, with the Turkmen at only 4.
      I haven't searched for outliers yet.

      • Can you post the official harappa admixture of the afghan sample please?

        • For some reason, I wasn't able to post here before because there was no "reply" button. Here is what you wanted. This is not harappa result. This is my run.

          Pashtun 1 : 17.5% Caucasian, 0.8% Beringian, 10.7% South Indian, 2.9% East Asian, 2 % East Asian, 3.9% SW Asian, 37% Baloch, 19.8% Northern European , 1.4 % Native American
          Pashtun 2 : 20.7% Caucasian, 12.5% South Indian, 3.7% East Asian, 1.7% East Asian, 1.9% SW Asian, 36.5% Baloch, 1.6% Med, 16.3% Northern European, 1.1% Native American
          Pashtun 3 : 10.9% Caucasian, 9.3% South Indian, 11% East Asian, 11.5% Siberian, 4% SW Asian, 32% Baloch, 3.2% Med, 14% Northern European, 0.7% Native American
          Pashtun 4 : 18.7% Caucasian, 0.7% Beringian, 12% South Indian, 3.6% East Asian, 5% Siberian, 39% Baloch, 3.6% Med, 12.7% Northern European, 0.87% Native
          Pashtun 5 : 19.7% Caucasian, 2% Beringian, 13% South Indian, 1.7% Siberian, 1.6% SW Asian, 39.8% Baloch, 14.7% Northern European, 2.4% Native American

          • Curious, you should put your disclaimer here as well. The results aren't fully comparable to HAP. Otherwise, people might speculating that the HGDP Pashtuns are more differentiated than the Metspalu Pashtuns than they actually are. The HGDP Pashtuns have fairly similar scores to the Metspalu Pashtuns on your run as well including lower South Indian and higher NE Euro in comparison to HAP. Overall, the HGDP and Metspalu Pashtuns are quite similar with the HGDP being more South Asian shifted and the Metspalu being more East Eurasian shifted. At least based on their Eurogenes K13 results.

          • Yup, the results aren't comparable to HAP. I should have mentioned it here as well. HAP uses more SNP's, and also has a lot more samples.

          • Regardless, I appreciate you posting the results of your run here. Maybe you can post some of the HGDP Pashtuns results on your run as well on here? Excluding the 30% East Eurasian individual of course.

          • Here is a link to the Afghan data. The Pathans would be the HGDP Pathans in my dataset. The rest are the samples from the recent Pathan study.

          • Are these the results of this?


            What recent Pathan study?

          • Yeah, that's the one I mean. My bad. I meant the recent Afghan study.

          • How many SNPs are being used here as compared to the normal HarappaWorld calculator with consumer genetic data?

            And can someone link me to the Afghan study? How old are these samples?

          • AD, these are the Afghan Pashtuns from the recent Metspalu study on Afghanistan. They are the same 5 Afghan Pashtuns in Eurogenes K13's oracle.

          • Thank you Curious!

            This just makes me more anxious to see their results for HarappaWorld.

          • I have calculated the HarappaWorld admixture results for the Afghan dataset. As soon as I am done with the web hosting move, I'll post them.

          • Thank you Zack! Should prove very interesting.

  85. Thanks Curious!

    • Thanks! Not surprised on the low South Indian component

      • The HAP official results will surely have higher South Indian. These results aren't official as Curious ran a test with lower SNPs and a much smaller reference population database. The results should be similar to HRP370 overall based on their Eurogenes K 13 scores.

        • So would that mean they would have an even higher baloch and northern euro component?

          • No, they will have lower Northern Euro. Probably somewhere in between 11-14%. Baloch, I can't say with complete confidence but I think lower Baloch as well. Like I said, if you want an idea of their results, look at HRP0370. Most will be slightly more South Asian than that individual. One will be significantly East Eurasian shifted and one will have a fair bit more Siberian/Amerindian than the average.

            You're better off waiting until Zack posts the official results. Curious' run was pretty solid for what he had to work with but it's nowhere near as "accurate" as Zack's due to lower SNP count and a much smaller reference database.

            I guess no one likes to read his disclaimer. IranianR1a and Indo-Aryan populations were getting "excited" over non-official results that were probably quite a bit off the mark.