Parsi Nanis Are Gujarati

Part I of III: Hypergamy, Endogamy, and the Terminal Phase of the Parsi Model

With a nod to the Y. M. Hodiwalla thread on the Parsi Zoroastrians Worldwide Facebook page, 19 April 2026.

“The Parsis are a quarter Gujarati (genetically). Essentially their Nanis are Gujjus and interestingly enough culture is almost always transmitted via the mothers and their mothers.”

I. The Language Is the Mother

On 19 April, Y. M. Hodiwalla published a long lament on the Parsi Zoroastrians Worldwide page. He called the community’s shift from Gujarati to English a “cultural suicide,” a “refined, English-speaking whimper,” the surrender of a thousand-year linguistic inheritance for the “glittering tinsel of modern fashion.” The comments beneath agreed with him almost to a person. One commenter, KCR, asked the most interesting question in the thread, and everyone ignored it:

Did Jadi Rana tell Parsis to change their “mother tongue”? Was the language of Jadi Rana’s era called Gujarati? And no one retains MT after a couple of generations especially if their numbers are small after migration.

That single comment contains the argument Hodiwalla’s essay cannot bring itself to make. The Parsis did not “adopt” Gujarati in some conscious civilisational pact with Jadi Rana. They speak Gujarati because their founding mothers were Gujarati. The Qissa-i Sanjan gives us the romance. The genome gives us the picture.

II. The Parsi Genome

The landmark paper is Chaubey, López, and Metspalu’s 2017 study in Genome Biology, titled, with a borrowed Parsi metaphor, “Like Sugar in Milk.” The team analysed Indian and Pakistani Parsis across the full genome, the Y chromosome, the mitochondrial line, and ancient DNA recovered from burials at Sanjan, the original landing site in Gujarat.

On the Y chromosome, modern Parsis cluster with Iranians and with Neolithic Iranians, not with Gujaratis. The paternal line is almost entirely Persian. The refugees who landed at Sanjan around the eighth century brought their fathers with them and their fathers’ fathers up the male line, and that line has not been meaningfully mingled in twelve hundred years.

On the mitochondrial line, the picture flips. Modern Parsi mtDNA is roughly half to sixty per cent South Asian. Haplogroup M, which sits at around fifty-five per cent in Parsis, is almost absent in Iranians, at under two per cent. Critically, the ancient Sanjan samples already showed forty-eight per cent South-Asian-specific mtDNA. The absorption of local mothers happened at or near arrival, not as a slow leak across centuries.

On the autosome, which averages the whole genome, Parsis come out Iranian-leaning but with around a quarter South Asian admixture. This is the “a quarter Gujarati” number that circulates on social media, and it is correct as a genome-wide statement while being the most misleading of the three figures if you treat it as representative of either parental line.

This is the classic signature of female-biased founding admixture followed by endogamy. Persian refugee men. Local Gujarati women. Then a millennium of marrying within the group. The Y stayed Iranian because no new fathers came in after the first few generations, by which point the founding cohort had probably begun calling kin from Iran to marry their mixed-ancestry daughters. The mtDNA stayed Indian because the founding mothers’ line transmitted through every generation of daughters since. The autosomal average is the arithmetic middle, compressed further by drift and inbreeding in a small closed population.

If you walk up any Parsi’s Nani’s Nani’s Nani line, you hit a Gujarati founding mother roughly half the time. Walk up a male Parsi’s Y line, you hit a Persian founding father always. The language they speak at home, the food they eat, the way they swear in their kitchens, the Gujarati-inflected English their children now speak: all of this travelled down the matriline because the matriline was the only one available to carry it.

Hodiwalla’s essay treats Gujarati as a borrowed garment the Parsis put on in Sanjan and must now refuse to take off. That frame is wrong. Gujarati was not borrowed. It was inherited from the women who became the mothers of the community when the boats arrived.

III. There Was Never a Pure Pars

There is no pure Parsi community waiting to be restored. There never was one. The community began as a Persian-Gujarati hybrid in the eighth century and has been consolidating that hybrid ever since. The Zoroastrian liturgy is Avestan. The daily language is Gujarati. The cuisine runs from dhansak to patrani machchhi. The names are Persian, the idioms are Hindustani, the religion refuses conversion. This is not a contamination of a pure thing. This is the pure thing.

Hodiwalla’s lament, sincere and well-written as it is, rests on a category error. He writes as if the loss of Gujarati would return the community to some Persian essence underneath, when in fact the loss of Gujarati would just delete the Indian half of what the community has been for twelve hundred years. The Nanis were Gujarati. The kitchens were Gujarati. The lullabies were Gujarati. The prayers remained Avestan because the rule was patrilineal descent and father’s religion, but the texture of ordinary life ran through the matriline, and the matriline was Gujarati from day one.

The Parsis are not a Persian community that acquired a Gujarati accent. They are a Persian-Gujarati community now considering whether to acquire an English one on top. That is a separate decision. It may or may not be survivable. It is not a civilisational suicide in the sense Hodiwalla means. The civilisational synthesis was completed in Sanjan a thousand years ago.

KCR was right. No group retains its mother tongue after a couple of generations if the numbers are small. The Parsis held on to Gujarati for forty generations, which is itself astonishing. They held on because they had the endogamous depth to sustain a household economy in that language. The question of what happens when that depth runs out is a separate question, and it is the one Hodiwalla should have been asking. That question is also best approached sideways, by looking first at the communities that faced it before the Parsis did.

Coming in Part II

The Parsi pattern is not unique. It recurs wherever a small group of migrating, typically high-status, men meets a resident population of women. The genetic signature looks similar in every case: foreign Y chromosomes, local mitochondria, intermediate autosome. But the social engines producing that signature are not interchangeable, and the Abrahamic and Islamic worlds ran two very different versions of the same arithmetic over the last two thousand years. The Ashkenazi Jews built a synthesis that mirrors the Parsi one almost exactly: Levantine fathers, European mothers, religious closure, a thousand years of endogamy. The Gulf and North Africa ran the same machinery through a much darker legal architecture, the Quranic concubine system, producing genetic outcomes that look like the Parsi signature on a chromosome plot but read very differently on the page.

Part II: “Sugar in Many Milks: Jews, Arabs, and the Islamic Concubine System.” Coming soon.