Why does Hindi have so many words for “uncle”?

Each word tells you which side and which branch. Chacha: father's younger brother. Tau: father's elder brother. Mama: mother's brother. Phupha: husband of your father's sister (your bua). Mausa: husband of your mother's sister (your mausi).

Look closely and something interesting shows up. Your father's brothers are split by age, but your mother's brothers aren't: older or younger, they're all mama. One reading is that in a traditional north Indian joint family, your father's brothers might live under one roof with you, and the elder one might hold a distinct place in the household. Your mother's brother, by contrast, belongs to the family she came from, a relationship with its own warmth and ceremonial roles. It's a persuasive picture, but it is a reading, not a proven cause, and many families don't match it.

This isn't an Eastern oddity. Latin had patruus for a father's brother and avunculus for a mother's brother. Old English also split them, into fædera and ēam. The modern English word “uncle” comes through French from avunculus, and it was stretched to cover everyone. So English used to make this kind of distinction, and lost it.

Other Indian languages draw their own lines. Tamil has periyappa and chittappa for father's elder and younger brothers, and māmā, which can also mean your father-in-law, because marrying a cross-cousin is an old and respected pattern there. Anthropologists since Lewis Henry Morgan in the nineteenth century have tried to link kin words with family and marriage patterns. Thomas Trautmann's work on Dravidian kinship is one careful example. The links are real but they don't let you read a family's life straight off its vocabulary.

Charles Kemp and Terry Regier compared kin words from 566 languages and argued in Science in 2012 that these systems strike a balance between being simple and being useful: a language can have few words or more precise ones, and most sit near the best trade-offs. Gunjan Anand and Regier then looked at Hindi and English together, and found that kin terms are used more often in Hindi text than in English text and that each language's system fits the way its own speakers talk about family.

That's a pattern about communication, and it's also fairly new work. It suggests that a language's words track what its speakers often need to say. It doesn't tell you why that need arose in the first place.