The Book Themes Author Get It
Praveen Maloo

Before
the Word

Ancient India · Artificial Intelligence · The Language Between
What is language before it becomes sound?
Ancient India had an answer. AI is rediscovering the question.
Get the Book Read More
Scroll

What did ancient Indian grammarians understand about language that modern linguists are only now rediscovering?

The question at the centre of this book

About the Book

Two traditions, twenty-five centuries apart, asking the same question

In Before the Word, Praveen Maloo traces the extraordinary parallel between Sanskrit philosophy and the architecture of artificial intelligence — two traditions separated by twenty-five centuries, both asking the same radical question: what is language before it becomes sound?

From Pāṇini's generative grammar to the transformer attention mechanism; from Bhartṛhari's sphota theory to the embedding space of a large language model; from the four levels of Vāk to the layers of a neural network — this book argues that the conversation between ancient India and contemporary AI is not a metaphor. It is a living dialogue. And we are only beginning to hear it.

Fluency is not intelligence. The Indian tradition had a precise vocabulary for the difference. We desperately need that vocabulary right now.

The Parallels

Ancient insight meets modern architecture

Pāṇini's Grammar · 350 BCE
↓
The Transformer Architecture
A finite rule system that generates infinite valid sentences — the oldest formal generative model in human history, rediscovered in silicon 2,300 years later.
Bhartṛhari's Sphota Theory
↓
The Embedding Space
The word as an indivisible unit of meaning that precedes sound — a theory of semantic representation that anticipates the vector embeddings at the heart of modern NLP.
The Four Levels of Vāk
↓
The Layers of a Neural Network
From Parā (pure undifferentiated potential) to Vaikharī (articulated sound) — a Vedic map of speech that illuminates exactly what large language models can and cannot do.
Pramāṇa — Valid Knowledge
↓
AI Epistemology
The Indian theory of testimony as a source of knowledge — and why it matters profoundly when the testimony comes from a machine that has no relationship to truth.
What You'll Find Inside

Six ideas this book will change how you see

01
Why Pāṇini's 4th-century grammar is considered the first formal generative system — and what it shares with the architecture of GPT.
02
The four levels of Vāk and what they reveal about the layers of knowing that a language model cannot access.
03
How Bhartṛhari's theory of the indivisible word anticipates the embedding space of a large language model.
04
What the Indian epistemological tradition tells us about the validity of testimony — and why it matters when the speaker is a machine.
05
Why fluency is not intelligence — and why the Indian tradition had a precise vocabulary for the difference, 2,000 years before we needed it.
06
The living dialogue between the oldest grammar and the newest machine — and what it asks of us right now.
The Author
Praveen Maloo
Praveen Maloo
Entrepreneur · Polyglot · Author

Praveen Maloo grew up in India, where education meant Sanskrit alongside mathematics — language was never merely a tool. He has since built two technology companies in the United States, travelled through 170+ countries, and speaks seven languages — each one a different window onto the world.

That restlessness made him attentive to what falls between languages, and to the ancient traditions that took those gaps seriously. He writes at the intersection of Indian philosophy, cognitive science, and artificial intelligence.

"That conversation — between the oldest grammar and the newest machine — is what this book is about."

Get the Book

Available now

Kindle ebook · Paperback · Available worldwide

Kindle Edition — Coming Soon Paperback — Coming Soon

Links will be live within 72 hours · Sign up below to be notified