Endogamy and assimilation: Parsis in India

37 Comments

The Guardian has a long piece about Parsis, The last of the Zoroastrians. The author is the child of a Parsi mother who married a white Briton. Though he brings his own perspective into the piece, I appreciated that he did not let it overwhelm the overall narrative. The star are the Parsis themselves, not his own personal journey and viewpoints.

In relation to the Parsis, there are two aspects in the Indian context that warrant exploration

– high levels of cultural assimilation

– high levels of cultural separateness

These seem strange outside of India, but they make sense within India. The Zoroastrian priests who migrated to what became Gujurat integrated themselves into the local landscape as an endogamous community. From a genetic perspective, the best overview is “Like sugar in milk”: reconstructing the genetic history of the Parsi population:

Among present-day populations, the Parsis are genetically closest to Iranian and the Caucasus populations rather than their South Asian neighbors. They also share the highest number of haplotypes with present-day Iranians and we estimate that the admixture of the Parsis with Indian populations occurred ~1,200 years ago. Enriched homozygosity in the Parsi reflects their recent isolation and inbreeding. We also observed 48% South-Asian-specific mitochondrial lineages among the ancient samples, which might have resulted from the assimilation of local females during the initial settlement. Finally, we show that Parsis are genetically closer to Neolithic Iranians than to modern Iranians, who have witnessed a more recent wave of admixture from the Near East.

The major finding is that most of the ancestry, on the order of 75%, of Indian Parsis is generically Iranian. On the order of 25% is “indigenous”, probably Gujurati. The fact that the proportion of mtDNA, the maternal lineage, is closer to 50% in both modern and ancient samples, indicates that the mixing with Indians occurred through taking local women as brides. Further genetic investigation in the above paper suggests that this was not a recurrent feature. In other words, once the Zoroastrian community in Gujurat was large enough, it became entirely endogamous.

And yet the Parsis speak Gujurati (or in Pakistan Sindhi and Urdu), and in many ways are culturally quite indigenized.

I bring this up because this portion of the above piece I’ve seen elsewhere:

The small community of Iranian Zoroastrians is even more liberal, allowing female priests, and there are also nascent neo-Zoroastrian movements in parts of the Middle East.

I have read repeatedly that Iranian Zoroastrians are more “liberal” when it comes to the issue of religious endogamy. The term “liberal” indicates innovation. But if you look at the historiography it seems clear that though Zoroastrianism was strongly connected to Iranian-speaking people, it was not exclusive to them. Zoroastrianism was cultivated and encouraged by the Sassanians in much of the Caucasus, and it spread into Central Asia, even amongst some Turkic groups, before the rise of Islam.

Zoroastrianism was not aggressively proselytizing but before 600 A.D. neither was Christianity outside of the Roman Empire. This is not a well-known fact due to the religion’s gradual diffusion to non-Roman societies, such as Ireland and Ethiopia, but before Gregory the Great missions to barbarian peoples were not organized from the Metropole but were ad hoc (the Byzantines eventually began centrally organized missionary activities after 600 A.D., but they were never as thorough or enthusiastic as the Western Christians). Our understanding of Zoroastrianism then may not be clear because the religion went into decline during a transformative period in world history when the religious boundaries and formations we see around us were still inchoate.

The major takeaway from all this is that strenuous Parsi arguments for the ethnic character of Zoroastrianism are a reflection not of their Iranian religious background, but their Indian cultural milieu. Zoroastrian “traditionalists” in India are actually Indian traditionalists, who have internalized an innovation to allow for integration into South Asia. And, I would argue some of the same applies to Indian “traditionalists,” whose cultural adaptations to the “shock” of Turco-Muslim domination resulted in the strengthening of particular tendencies within Indian culture that were already preexistent.

2+

37 Replies to “Endogamy and assimilation: Parsis in India”

  1. Dear Razib,

    I had a question regarding Parsis; the literature states that Parsis are not like Modern Day Iranians genetically, even if they were “pure”, not only because of their substantial South Asian Admixture and Indian mtDNA haplogroups, but also because they lack the radical amounts of Near Eastern/Caucasus/Anatolian/Arab admixture that modern-day Iranians have, in addition to any ENA from Turkic/Mongolic/SSA sources. This is why they are closer to ancient Iranian groups that no longer exist.

    I had a question regarding this; given their South Asian admixture and unique genetic background, are Parsis closer to NW Indian groups like the Ror than they are to ancient Neolithic Iranian groups? And, on that note, are groups like the Ror closer to any West Asian group (outside of the subcontinent/South Asia) when compared to the most-AASI communities on the mainland?

    Appreciate your insights. Thanks.

  2. I had a question regarding this; given their South Asian admixture and unique genetic background, are Parsis closer to NW Indian groups like the Ror than they are to ancient Neolithic Iranian groups? And, on that note, are groups like the Ror closer to any West Asian group (outside of the subcontinent/South Asia) when compared to the most-AASI communities on the mainland?

    you are a genius in asking the same question in 1,000 different ways. no. and yes. high confidence on the no, medium confidence about the yes.

    i’m getting confused as to why you keep asking these questions. you know the general answer. what’s the point in constant specific pairwise comparisons?

    it’s like if i told you “yes, your dick is big.” “yes, but is my dick bigger than this specific guy?”

    1. Sorry, my intention is to figure out if certain claims of Parsis being very distinct from all Indian groups (including NW Indian groups) is correct. The implication being, that Parsis are somehow completely alien or “foreign” to the subcontinent and not even as “Indian” as the Rors. This obviously rings false in many ways, not the least of which is their cultural/political connection to the subcontinent (Rajiv Gandhi et. al)

      I said earlier that I consider Rors/Jats to be part of Indian civilization, despite their divergence from most Indian groups; I just get annoyed when I see Parsi claims to the contrary. Almost as if they are “above” all other Indian groups. (its something I’ve seen claimed by certain Iranians) Apologies if that was misconstrued.

  3. the paper is clear. they’re 25% guju and 75% some sort of eastern Iranian. no apples-to-apples comparison to NW indian groups (some suggestion the Iranian source is more like baloch in being low-steppe in that paper and elsewhere).

    the parsis are clearly a synthetic group. their self-identity is their business i guess.

    1. Thats good to hear. I guess that wacko Iranian who liked to go around claiming that Parsis were Modern-day Iranians with minor South Asian admixture was WRONG after all. People with the autosomal profile of Parsis dont exist in Modern-day Iran. And their South Asian admixture further solidifies this. At best, they are a unique ethnic group that can only be found in South Asia. You are right about their synthetic nature.

  4. the problem is that you think about this like leibniz. that we can ‘compute’ what indian/s asian is with some calculator. i don’t think we can.

    i’m way more AASI than you, but you are clearly way more indian than i am judging by your comments (despite your aversion to copulating with most south asians due to their physical appearance).

    1. Thanks for the clarification. And you are right, I am indeed very “Indian” in my thoughts due to reasons that escape me. Perhaps its my past interactions with West Asian chauvinists who like comparing dick sizes and demonizing AASI folk. It is something I am slowly healing from. Apologies.

      1. ~25% autosomal and 50% mtdna isn’t minor to me. but your mileage may vary.

        i think we need to sample iran much better. the papers on Iran I’ve seen haven’t released enough DNA. the steppe % varies a lot i bet, and we know that R1a1a varies a lot too (way more in the east, khorasan, not so much in the west, the core of persian iran around isfahan).

      2. APthk, just a thought, if those west asian chauvinists sensed your need to associate with their group then its not surprising that they may have “abused” your need to belong. Some people are dcks and they probe for vulnerability. See the world from your own jatt-american perspective, not only will it be better for you, but also for your people.

        1. when trads accuse me of being a chandala convert on twitter i don’t respond that i’m not a chandala. that only gives the slur power. if i was a chandala i’m clearly superior to them so who cares? ;=)

          i am sympathetic to warlock’s attitude. own what makes brownz distinctive. don’t try to associate with mlecchas as if their blood and memes are precious. they are as they are.

          (i will explain i’m not Muslim because i am not a big fan of Islam and racists assume i’m hindu anyway so i don’t have to do the ‘well the haters attack me for being muslim’)

      3. That’s interesting you’ve had that interaction with Persian folks. Anytime they’ve asked where I’m from and I’ve told them, their first reaction is usually “Oh, don’t they say we’re both descended from the same people!” (and then they refer to Indo-Aryans). They always mention our connection rather than how we’re different

          1. the average iranian doesnt know anything about india other than the brief history of a small group of indo europeans invading india and passing their language to the native population. but as far as culture goes, these are 2 extremely different groups.

        1. I came across this website by accident while doing research on g25 and read through the comments.
          I just gotta add, Iranians and South Asians are extremely different, both genetically and culturally speaking.
          I am an Iranian Lur living in the west, I have gone to university with a high south asian populaiton, mostly international students from punjab and central india. Very different culture, language and appearance indeed.

          The closest people culturally and genetically to Iranians are armenians, turks and azeris, azeris are essentially iranian/georgian/chechens speaking an east asian language. Even baluchistan in iran is considered to be a very alien region to the rest of iran, let alone south asia.

          1. you don’t need to tell us anything about genetics. i can post and analyze all the raw data in various forms. no need for calculators. i am the calculator 🙂

    2. A side remark re Leibnitz, who can be considered as a genetic cousin of some SAs (oit would say that as a good mathematician, apart from genetics, is an additional prove of his Indian roots). The founder of the German Academy of Science, originated (so as M.Luther) from the large (germanized) Serbian tribe Ljutic (Ljut=hot (as chilli), also = angry, also – the name of the field flower).

    1. I know, and that is why I admire you. A great example of a Brown American that I should try to emulate. I’ve been in awe of your wisdom even since the GeneXP days. I can only hope to reach your level of erudition in my later years. Thanks for all your insights and for giving us a platform to communicate with you. It is a great privilege for all of us, and it really helps me push back against supremacist narratives that are falsely peddled by Chauvinists of all stripes.

      1. to be clear, i think some level of ‘self-pride’ about various ethnic stuff is fine and normal (as in the human norm). i have a naturally weird psychology in that this isn’t a big thing for me except for amusement (LORD INDRA represent). but everything needs balance. too much group-pride makes me suspicious of lack of self-pride. ostentatious individualism makes me wonder about egotism. so everyone follow the golden-mean and be chill

  5. First, let me state that I am not a chauvinist and I strive to be as scientific and objective as it gets in my genetic analyses and interpretations, so please read what I will try to explain with that in mind.

    The modern Iranian populations do not really seem to show any clear Arab admixture except the Iranian Arabs (not covered in Global 25) and Seyyeds.

    https://drive.google.com/file/d/1bEeIIjXaE66JKLAwsDMwYPYlfmiTinIh/view?usp=sharing

    Look at the modern (Muslim) Persian (Fars) Iranian average for instance, what differentiates it from the modern Zoroastrian Iranian average is its higher Turkic and South Asian admixture, not higher Arab or other Semitic admixture.

    One cannot understand Iranian genetics without first understanding the deep-rooted regional genetic distinctions within the territory of what is now Iran. The most major one is east to west, there is a thousands of years of genetic distinction between the western and eastern parts of Iran as proven by the ancient DNA results. Some people, not knowing this very old distinction and not knowing the ancient DNA results, think of the modern genetic structure of the western parts of the country (save the minor Turkic admixture during the last 1000 years) as formed as a consequence of the Muslim Arab migrations rather being much older than that. What makes the Iranian Persians, Zoroastrians, Azeris, Kurds, Lors, Mazandaranis, Gilaks and Qashqai genetically different from the Iranian Khorasanis, Baloch and Bandaris is not any higher Arab admixture in the former, but the higher Iran Neolithic and South Asian ancestry and lower Anatolia Neolithic and Levant Neolithic ancestry in the latter, but none of those populations have Levant Neolithic ancestry high enough to clearly point to any Arab or known Semitic admixture.

    1. First, let me state that I am not a chauvinist and I strive to be as scientific and objective as it gets in my genetic analyses and interpretations, so please read what I will try to explain with that in mind.

      yeah, but you are a sperg who is persistent enough that I’ve banned you several times. you have a weird obsession with middle eastern ethnic purity. perhaps it isn’t an obsession and your asperger’s is getting the better of you.

    2. The closest group to Iranians are Caucasians and Mesopotamians such as Assyrians and Armenians.
      Iranians are essentially an extension of south Caucasians and Mesopotamian mix.

      the distance between south asia and Iran is crazy, like comparing a syrian with a bulgarian.

      1. @ Shaco
        Its nothing new dude, Almost everyone in this website know this already,
        Iranians(Western ones) are even closer to Southern Europeans(Greeks, Bulgarians and Italians(Tuscanis and Sicilians) than they are to Eastern Pakistanis(Punjabis and Sindhis)

  6. I read the title as endgame and assimilation. And looking at the parsee numbers, it surely is endgame and assimilation for them.

  7. @Razib,

    Is there a caste population that matches them well maternally? All I remember is that one MtDNA sequence, M3a1, is strongly represented amongst them, though that was from a study in the early 2000s

  8. aditya, hard to distinguish caste by mtdna. i assume something generic guju. if it was brahmin would be higher steppe % i think? i guess i can try and model it some time.

    1. i did a rudimentary computation based on shear percentages and mercantile vanias are a good fit for other 25%. Makes sense based off class structures with Parsis in same strata

  9. yeah, but you are a sperg who is persistent enough that I’ve banned you several times. you have a weird obsession with middle eastern ethnic purity. perhaps it isn’t an obsession and your asperger’s is getting the better of you.

    I wrote my comment in response to APthk’s speculations about the genetic composition of the Iranian ancestors of the Parsis. He seems to consider modern Iranians too changed genetically since the times of the Parsi migration to use in modeling Parsi ancestry. That is not what I observe in my analyses. Parsis can be modeled as a mixture of modern Iranians and Gujaratis with a high fit.

    https://drive.google.com/file/d/10pZ46QConIiey9gqiyXhnYj_iu9Lq-Nx/view?usp=sharing

    I have no stake in this subject, I am not even Iranian or South Asian, would not fret about being wrong in this.

    1. What are you even talking about? Parsis are like Modern-day Iranians with South Asian ancestry? Says who? The entirety of the published research disagrees with you my “unbiased” friend. BTW, arent you the same Turk who goes around writing all kinds of nationalistic drivel on Eurogenes and other websites about how South Asians cannot be “Caucasian” and that MENA people are “pure”? You’re full of BS.

      Read the literature for once, ALL the published peer-reviewed studies show that Parsis are NOT similar to modern-day Iranians, and that they have South Asian admixture and South Asian mtDNA groups. Did you even bother reading the paper Razib posted above. Here it is linked for your reading pleasure:
      https://genomebiology.biomedcentral.com/articles/10.1186/s13059-017-1244-9

      Here is a concise summary of the research just in case you cant read: “We estimate that the admixture of the Parsis with Indian populations occurred ~1,200 years ago. Enriched homozygosity in the Parsi reflects their recent isolation and inbreeding. We also observed 48% South-Asian-specific mitochondrial lineages among the ancient samples, which might have resulted from the assimilation of local females during the initial settlement. Finally, we show that Parsis are genetically closer to Neolithic Iranians than to modern Iranians, who have witnessed a more recent wave of admixture from the Near East.”

      Read that? Parsis have no modern equivalent population that exists in present-day Iran. They are closer to Neolithic Iranians than to Modern Iranians. They are a unique group with an autosomal profile that only exists within South Asia and the Parsi population that lives there.

      There are also plenty of sources that state that Iranians have picked up additional Arab-mediated Near Eastern admixture ever since the Parsi migration, also found in the literature. Let me dig them up for you so that you can also be proven wrong once more. I will reply to your other BS post in a short while. Also, please spare us your amateur “analyses” of the data. What credentials do you even have? Have you published your findings in a peer-reviewed reputed Journal? NO. Then Shut-up.

      If you don’t have a bone of contention to pick here, why even bother writing false BS? The truth is, you have a hard-on for demonizing AASI admixed-groups. Ironic, considering that Turks all have varying amounts of East Eurasian ancestry and Y-Haplogroups from East Eurasia, which is a sister group of AASI. Don’t hate your own people Onur.

      1. @APthk

        I do not know what you are talking about. I am not someone obsessed with ideas of “purity.” You either confuse me with someone else or do not know what you are talking about.

        I shared with you the Global25-based results of the Parsis. They show that, unlike the vast majority of Eastern Iranians and South Asians, the Parsis have Levant Neolithic input extra Tepecik Neolithic, which points to Iranian ancestry from what is now Western or Central Iran. Further confirming that, they show higher amounts of Anatolia Neolithic ancestry than Eastern Iranians in general. It is normal that the Parsis are closer genetically to Iran Neolithic than modern Iranians are since South Asian admixture draws them closer to Iran Neolithic peoples. BTW, the difference in the proximity to Iran Neolithic between modern Iranians and Parsis is slight, you can check it out yourself using G25.

  10. From an outsider perspective who is interested in genetics, I find discrimination based on physical characteristics (like melanin content, cheekbones, height, shape of nose and lips, eye color) way way more saner than the nonsensical discrimination based on percentage admixture of genetics.

    1. @Chittadhara

      Agreed. But even saner would be discrimination based on level of intelligence, achievement, affluence, etc.

  11. Parsi global25 results with the following very distant and basic inputs: Mesolithic Caucasus, Neolithic western Iran, Mesolithic eastern Iran, Epipaleolithic Levant, a modern Indian substitute for AASI, Steppe_MLBA and Anatolia PPN(A?)

    Target,Distance,GEO_CHG,IRN_Ganj_Dareh_N,IRN_HotuIIIb_Meso,Levant_Natufian,Paniya,RUS_Sintashta_MLBA,TUR_Boncuklu_N
    Parsi_India:Parsi10,0.03495187,2.6,39.0,0.0,9.0,16.4,16.8,16.2
    Parsi_India:Parsi11,0.04960740,6.4,40.0,0.0,6.4,14.4,12.4,20.4
    Parsi_India:Parsi13,0.04659934,2.4,40.4,0.0,9.8,16.2,15.8,15.4
    Parsi_India:Parsi2,0.04073881,7.8,38.2,0.0,9.6,16.4,11.6,16.4
    Parsi_India:Parsi26,0.03146460,5.4,39.0,0.0,9.4,11.4,19.2,15.6
    Parsi_India:Parsi34,0.03618847,9.4,35.8,0.0,9.4,16.0,13.4,16.0
    Parsi_India:Parsi4,0.03802665,8.6,31.2,3.6,10.2,15.6,15.2,15.6
    Parsi_India:Parsi7,0.03754248,5.4,38.6,0.0,4.6,13.8,16.6,21.0
    Parsi_India:Parsi8,0.02573260,4.2,39.8,0.0,2.8,18.2,11.8,23.2
    Parsi_India:Z177,0.03372674,2.8,34.2,13.2,5.6,13.4,9.8,21.0
    Parsi_India:Z178,0.03352398,1.4,40.2,0.0,7.8,17.2,9.6,23.8
    Parsi_India:Z179,0.02821920,10.2,31.0,6.4,11.4,13.0,12.6,15.4
    Parsi_India:Z180,0.02944012,9.4,34.0,0.0,9.6,17.0,7.2,22.8
    Parsi_India:Z181,0.03714209,7.0,38.6,0.0,11.0,14.2,16.0,13.2
    Parsi_India:Z182,0.03211790,5.4,36.0,0.0,11.0,16.0,19.4,12.2
    Parsi_India:Z183,0.03462495,16.6,30.0,0.0,14.6,13.2,11.8,13.8
    Parsi_India:Z185,0.03218626,8.2,37.6,0.0,8.4,13.6,11.6,20.6
    Parsi_India:Z186,0.04410617,7.4,39.0,0.0,7.8,17.6,10.4,17.8
    Average,0.03588553,6.7,36.8,1.3,8.8,15.2,13.4,17.8

    Unfortunately I had to use Paniya instead of simulated AASI because anthrogenica is still down
    .
    For the Pakistani Parsis:

    Target,Distance,GEO_CHG,IRN_Ganj_Dareh_N,IRN_HotuIIIb_Meso,Levant_Natufian,Paniya,RUS_Sintashta_MLBA,TUR_Boncuklu_N
    Parsi_Pakistan:5544382,0.02722531,4.0,39.6,0.0,11.2,17.8,14.0,13.4
    Parsi_Pakistan:5544383,0.03951380,5.0,38.0,0.0,10.2,14.6,16.4,15.8
    Parsi_Pakistan:5544384,0.03287212,8.0,36.4,0.0,12.2,15.6,18.4,9.4
    Parsi_Pakistan:5544385,0.03967836,5.6,40.8,0.0,8.4,8.2,12.2,24.8
    Parsi_Pakistan:5544386,0.03114159,6.8,32.2,9.6,6.8,12.4,13.8,18.4
    Parsi_Pakistan:5544387,0.03661164,6.8,23.2,15.2,10.8,12.8,16.6,14.6
    Parsi_Pakistan:5544388,0.03515931,7.8,35.8,0.0,11.0,10.4,15.2,19.8
    Parsi_Pakistan:5544389,0.03324727,6.8,32.8,3.4,10.0,15.8,14.8,16.4
    Parsi_Pakistan:5544390,0.03813061,3.4,36.4,2.0,5.0,19.8,17.0,16.4
    Parsi_Pakistan:5544391,0.03147197,9.0,36.8,0.0,9.0,15.2,13.0,17.0
    Parsi_Pakistan:5544392,0.03830887,8.4,37.8,0.0,12.2,13.8,10.6,17.2
    Parsi_Pakistan:5544393,0.04480936,15.2,34.4,0.0,8.4,9.4,14.2,18.4
    Parsi_Pakistan:5544394,0.03609961,3.8,41.4,0.0,8.4,9.6,18.2,18.6
    Parsi_Pakistan:5544395,0.03362409,4.4,33.4,6.2,7.6,14.4,14.4,19.6
    Parsi_Pakistan:5544397,0.04669061,4.8,35.2,0.0,6.4,24.2,13.4,16.0
    Parsi_Pakistan:5544398,0.03383940,4.6,40.4,0.8,10.0,13.4,12.6,18.2
    Parsi_Pakistan:5544399,0.03264242,9.2,32.6,0.0,9.6,17.6,17.0,14.0
    Parsi_Pakistan:5544400,0.03894832,6.2,37.4,0.0,7.4,13.0,16.4,19.6
    Parsi_Pakistan:5544401,0.03518554,14.4,31.2,0.0,12.6,13.2,10.6,18.0
    Parsi_Pakistan:5544402,0.03475864,4.6,37.8,1.4,8.6,17.2,14.0,16.4
    Parsi_Pakistan:5544403,0.04209762,7.6,38.0,0.0,11.0,12.2,14.6,16.6
    Parsi_Pakistan:5544404,0.03822070,8.0,37.0,0.0,8.8,13.2,18.0,15.0
    Parsi_Pakistan:5544405,0.03331931,7.8,36.8,0.0,11.0,17.4,12.0,15.0
    Average,0.03624332,7.1,35.9,1.7,9.4,14.4,14.7,16.9

    Not very different on average, expected worse fits given the temporal distance between some sources and the targets.

    1. @DaThang

      Thanks. Similar to my analysis results. The major difference is that you use MLBA rather than EMBA as the steppe source, Natufian rather than Levant Neolithic and no Tepecik Neolithic or a similarly eastern Anatolia Neolithic source.

  12. Though in my analysis I get better fits, probably due to using Levant Neolithic and Tepecik Neolithic.

Comments are closed.