Sisters, not strangers
Hindi and Gujarati grew from one parent, and much of their core vocabulary is the same word worn down two different ways. So why can a Hindi speaker not simply follow Gujarati? What “sister” means exactly, why it is not the same as borrowing, why no date marks the parting, and how this app labels each pair.
Hindi and Gujarati are sisters. Both are Indo-Aryan, both descend from Sanskrit through the spoken Prakrits and then through Apabhramsha, the last stage of Middle Indo-Aryan, Gujarati by its western branch and Hindi by the central branch that grew out of Shauraseni. That is the thesis of this app, and it is easy to mishear. It means neither one language written in two scripts nor two languages that happen to share a few words. A Hindi speaker in Ahmedabad does not simply understand what is being said. But it feels almost within reach, because so much of what is said is made of words that speaker already owns, worn a little differently.
“Sister” has a precise meaning, and the place to test it is basic vocabulary, the words too ordinary for anyone to bother borrowing. Sanskrit hasta, hand, became हाथ haath in Hindi and હાથ haath in Gujarati. Sanskrit gṛha, house, became घर ghar and ઘર ghar. Sanskrit akṣi, eye, became आँख aankh and આંખ aankh. Sanskrit dugdha, milk, became दूध doodh and દૂધ doodh. Four matches could be luck; the repeating changes are not: Sanskrit st comes out as th in both languages, kṣ usually as kh in both, and a cluster like gdh flattens to dh in both. When the same old sound comes out the same way in word after word, nobody borrowed anything; the two languages lived through the changes together.
The parting is regular too. Sanskrit ghoṭaka, horse, is घोड़ा ghoda in Hindi and ઘોડો ghodo in Gujarati. Sanskrit kāla, black, is काला kaala and કાળો kaalo. The old masculine ending came out as ā in one house and o in the other, so a Hindi speaker who knows the rule can predict રસ્તો rasto from रास्ता raasta, a Persian loan each language dressed in its own ending. Where Hindi keeps an old l, Gujarati often has the retroflex ળ, as in फल phal against ફળ phal. Sanskrit had three genders; Hindi dropped the neuter and Gujarati, like Marathi, kept it, so कपड़ा kapda is masculine and કપડું kapdu is neuter. The rules are few, and Where Hindi says aa, Gujarati says o lists them.
This is why kinship is not borrowing. A borrowed word arrives whole, and it is usually a word for something new. Inherited words are the opposite: the words for hand and house and mother and the numbers. माँ maa and મા maa. बहन bahan and બહેન bahen, both from Sanskrit bhaginī. तीन teen and ત્રણ tran, both from Sanskrit trīṇi, three. दस das and દસ das. Languages borrow words for books and markets all the time; they very rarely borrow the word for three. When the words for three, ten, hand, eye and milk all match by the same rules, only one explanation is left, and it is a family tree.
Where the tree forks has no date. Gujarati’s nearest sister is not Hindi but Rajasthani: the two share a stage scholars call Old Western Rajasthani, or Old Gujarati, attested from around the twelfth century. Hindi comes down a different, central line, from the speech around Delhi. Around the same time the Jain scholar Hemachandra, at the court in Patan in northern Gujarat, wrote a grammar whose final section describes the Apabhramsha of his day, a literary language rather than any one town’s speech. Jain poets wrote in the emerging language over the following centuries, and Narsinh Mehta, around the fifteenth, is traditionally called its first major poet. What the record shows is two paths out of one parent, drifting apart across the Apabhramsha centuries, and no year in which they parted.
That is enough drift to make two languages, and it did. The verb “to be” is है hai in Hindi and છે chhe in Gujarati, by the usual account from different roots, and it is one of several places where the two part outright: मैं main is હું hu, and the का ka of possession is નો no. The infinitive ending differs: खाना khaana and करना karna are ખાવું khaavu and કરવું karvu. A Hindi ear must also do without the nuqta letters Gujarati spelling never adopted, so बाज़ार bazaar is બજાર bajaar. Some common words simply do not match: बेटा beta and દીકરો deekro, किताब kitaab and ચોપડી chopdi. And the script is Nagari with the headline left off, its exact history debated, as The line that went missing explains. Kinship gives a Hindi speaker a head start, not a free pass.
Every entry in this app therefore carries a verdict. The label the app was built around is the inherited pair, हाथ and હાથ, one Sanskrit word worn down two ways. Shared Sanskrit is for pairs where at least one side took the word whole, as with पुस्तक pustak and પુસ્તક pustak. Shared loans are labeled by source: the Perso-Arabic layer that came with Persian as the language of administration, from the Delhi Sultanate’s conquest of Gujarat around 1300 onward, and later through Hindustani, with બજાર again; a Portuguese layer from Diu, Daman and Bassein, with तंबाकू tambaaku and તમાકુ tamaaku from tabaco; and English, with बस bas and બસ bas. Unrelated is for pairs like बेटा and દીકરો, same meaning, different words, and commoner than the family resemblance suggests, since Hindi often reaches for a Persian word where Gujarati kept a Sanskrit one: आदमी aadmi against માણસ maanas. One side can also stand alone: ઢોકળું dhoklu and ગરબા garba have no Hindi word, Hindi borrows the Gujarati one, and the label for such a pair says so: Gujarati only. And where two sides are the same word and nobody can say where it came from, the label says so rather than picking the tidiest story.
So this is the claim, stated once. Two languages, one parent. Hindi and Gujarati are not strangers who became neighbors, nor one tongue in two costumes. They are sisters who left the same house by different doors, in no hurry, and have kept, in their most ordinary words, the same family face.