Words, said the logicians of the Nyāya school, have meaning because they point to things that exist outside the mind. The Buddhists answered that meaning rests on nothing firmer than the conventions of a community of speakers. Bhartṛhari, a Sanskrit grammarian of the fifth century, rejected both answers. Language, for him, is not a set of labels laid over a world that is already there. No cognition comes without words, he held, and all knowledge appears as if shot through with them. In its strongest form the claim says that the word makes the thing: the world shows up only for a mind already shaped by speech, and the reality behind the world, Brahman, is itself of the nature of the word.
He came out of the oldest science in Sanskrit learning. Pāṇini's rules had fixed the language centuries before, and Patañjali, around 150 BCE, had explained and defended them in the Mahābhāṣya, the "great commentary". Grammar counted as one of the auxiliary sciences of the Veda, a tool for reciting and understanding scripture correctly. Bhartṛhari wrote a commentary on the Mahābhāṣya himself, which survives only in fragments, and in his main work he credits some of his ideas to his teacher, Vasurāta. With this inheritance he turned the grammarian's questions about words and endings into questions about how anything is known at all.
That work is the Vākyapadīya, some two thousand verses in three books, whose name means roughly "on sentences and words". Its best-known argument concerns how a listener understands. Sounds reach the ear one after another, yet the meaning arrives whole, in a single flash he called pratibhā, intuition. What carries that meaning is the sphoṭa, from a root meaning "to burst forth": the sentence taken as one indivisible unit. Grammarians cut sentences into words and words into stems and endings, and he granted that the cutting is useful. Yet the pieces, he argued, are abstractions that analysis makes afterwards; only the whole sentence means anything. Speech itself he described on three levels, from an undivided inner "seeing", through the ordering of thought, to the sounds the voice produces. Correct speech clears the mind, and so the study of grammar becomes, in the first book of the Vākyapadīya, a path to liberation in its own right.
The book found readers well outside the grammarians, many of them hostile. Dignāga, the Buddhist logician and a near contemporary, built a short treatise on time out of its verses. Kumārila, the great exegete of Mīmāṃsā, the school that interprets the Vedas as commands, attacked the sphoṭa in a chapter of his Ślokavārttika: the sounds of a word, he argued, are enough to carry its meaning, and nothing beyond them is needed. Maṇḍana Miśra, an Advaita Vedānta author of the eighth century, gave the sphoṭa a central place in his version of that school. The Śaiva philosophers of Kashmir, Abhinavagupta among them, took over the three levels of speech and added a fourth above them. Modern scholars compare his view that meaning lives in the whole sentence with the philosophy of language of Wittgenstein and Austin.
About the man himself almost nothing is on record. Yijing, a Chinese Buddhist pilgrim in India in the late seventh century, described a famous grammarian named Bhartṛhari as a Buddhist who wavered seven times between the monastery and the life of a layman; researchers have found the account mistaken about both his religion and his time. Indian tradition makes him a king of Ujjain, elder brother of King Vikramāditya, who left a life of pleasure for renunciation and wrote the Śatakatraya, three hundred verses on conduct, love and detachment. Modern scholars, K. A. Subramania Iyer and Bimal Krishna Matilal among them, generally treat the grammarian and the poet as two different people, though some leave the question open. Two of the three centuries, the Nītiśataka on worldly conduct and the Vairāgyaśataka on renunciation, are held here under his name because the tradition ascribes them to him; the century on love is not among them. Historians who build the chronology from his teacher before him and from Dignāga after him place Bhartṛhari at about 450 to 510. Reference works that follow Yijing's account, and with it the identification with the poet, still put his death in the middle of the seventh century.