Hindu Texts — A Map

There is no single "Hindu Bible." There is a library — roughly 3,500 years deep, in at least four classical languages, compiled by competing lineages, with overlapping authority claims that occasionally contradict each other. The first move in reading it is not picking a text. It's learning the two-tier filing system that decides which text wins when two disagree.

Shruti vs Smriti — the hermeneutic that decides ties

	Shruti ("heard")	Smriti ("remembered")
Origin	Revealed; not human-authored	Composed by named sages and poets
Authority	Highest. Inviolable in principle.	High but interpretable; subordinate to Shruti
Includes	concept vedas — 4 Vedas plus their four internal layers	Puranas, Itihasas, Dharmashastras, Tantras, Agamas
Approximate dating	~1500 BCE – 500 BCE (oral layer older)	~500 BCE – 1500 CE

If two texts contradict, Shruti wins by Mimamsa hermeneutic rule. In practice, Smriti gets most of the daily attention — it's the storytelling layer, the one householders actually read. The Vedas themselves are chanted more often than read.

The Shruti layer — four Vedas, each in four sub-layers

Each Veda is not one book but a stack:

Layer	Function
Samhita	Hymns and mantras — the ritual core
Brahmana	Prose manuals on ritual procedure
Aranyaka	"Forest" texts — transition from ritual to contemplation
Upanishad	Philosophical inquiry — the vedanta, end of the Veda

The four: Rigveda (1,028 hymns, oldest), Samaveda (melodic chant), Yajurveda (sacrificial formulas), Atharvaveda (healing, household, often called the magical Veda).

The Upanishads — the pivot from ritual to inquiry

Of roughly 200 surviving Upanishads, 10–13 are "principal" — the set Shankara commented on around 800 CE: Isha, Kena, Katha, Prashna, Mundaka, Mandukya, Aitareya, Taittiriya, Chandogya, Brihadaranyaka, Shvetashvatara, Kaushitaki, Maitri. The pivot they enact is large: from "do the rite correctly" to "ask what the self is." overview vedanta is downstream of this turn. The Mandukya Upanishad is only 12 verses long and arguably reorients the entire tradition toward non-dualism.

The Smriti layer — where the stories live

Itihasas (the two epics)

Mahabharata — ~1.8 million words, ~100,000 verses, roughly seven times the combined length of the Iliad and Odyssey. The Bhagavad Gita sits inside it at 700 verses: 5% by volume, far more by philosophical reach.
Ramayana — Valmiki's ~24,000-verse Sanskrit poem, with major regional retellings: Tulsidas's Ramcharitmanas (16th-century Awadhi), Kamban's Iramavataram (12th-century Tamil), Krittibas in Bengali, and many more. The text is plural by design.

Puranas

18 Mahapuranas plus 18 Upapuranas. Encyclopedic in scope — cosmology, dynasty lists, deity narratives, ritual instruction, geography. The six most-cited: Vishnu, Shiva, Bhagavata, Padma, Markandeya, Garuda. The Bhagavata Purana alone runs about 18,000 verses and is the source text for most Krishna devotion.

Dharmashastras

Codes of conduct, law, social order. Manusmriti, Yajnavalkya Smriti, Naradasmriti. Manusmriti's outsized modern prominence is largely a colonial artifact — William Jones translated it in 1794 and British administrators treated it as the Hindu legal code, a status it did not historically hold across most regions or castes.

Lineage-specific

Tantras — Shaiva and Shakta esoteric texts; ritual plus meditation, sometimes deliberately transgressive.
Agamas — Vaishnava and Shaiva temple manuals; how to build, consecrate, and run a temple.
Sutras — Patanjali's Yoga Sutras (4 chapters, 196 aphorisms), Badarayana's Brahma Sutras, Gautama's Nyaya Sutras.

The schools and their anchor texts

School	Anchor
overview vedanta	Upanishads + Bhagavad Gita + Brahma Sutras — the prasthana trayi, "three foundations"
Yoga	Patanjali's Yoga Sutras
Tantra	Sect-specific — Kashmir Shaiva texts, Sri Vidya, Kaula lineages

A working Vedantin is expected to read commentaries on all three of the prasthana trayi, not just the Gita. The popularity of the Gita as a standalone book is partly a 19th-century phenomenon, reinforced by translations from Wilkins (1785) onward.

What this map leaves out

This is a Sanskrit-canon view. Equally load-bearing for Hindu practice but governed by different canons:

Tamil Shaiva and Vaishnava canon — Tevaram, Divya Prabandham, the 12-volume Tirumurai. The Alvars and Nayanars (6th–9th century) were composing devotional poetry in Tamil while north Indian Sanskrit commentary was still consolidating.
Bhakti vernacular literature — Tulsidas in Awadhi-Hindi, Tukaram in Marathi, Mira in Rajasthani-Hindi, Kabir's Hindi couplets, the Kannada vachanas of Basavanna.
Sikh canon — the Guru Granth Sahib draws on Hindu bhakti voices (Kabir, Namdev, Ravidas) but is the scripture of a distinct religion.

A wiki page on Hindu texts that stops at Sanskrit is roughly half the building.

What's contested

Dating the Vedas. The Rigveda's composition window is argued from ~1500 BCE (mainstream Indo-European linguistic dating) to dates many centuries earlier in some Indian scholarship that ties it to the Sarasvati river's drying. The text was oral for centuries before being written, which makes carbon dating moot and linguistic dating the main lever.

The canon's edges. There is no central council that fixed the Hindu canon. The list of 18 Puranas varies across manuscripts. Several "Upanishads" were composed as late as the 17th century to attach the prestige of Shruti to sectarian material. What counts as scripture is partly a question of which lineage you ask.

Manusmriti's authority. Within the tradition, it is one Dharmashastra among many, regionally uneven. Outside the tradition, it has become the most-cited Hindu legal text — a colonial-era inheritance that scholars like Wendy Doniger and Patrick Olivelle have written about extensively.

Why this has to do with other realms

The Vedas may be the oldest continuously transmitted text in any human language. They were preserved by an oral mnemonic system — patha recitation, with built-in error-correcting redundancies (jata-patha, ghana-patha) that re-encode the same syllables in different permutations. The result is a transmission protocol with checksum behavior, designed by people who had no abstract concept of a checksum. The parallel to concept error correcting codes is not metaphor — it is the same engineering problem solved in human memory. A page on information theory that doesn't notice this is incomplete.

An open question

If the canon was never officially closed and new Upanishads were still being composed in the 17th century, what would it take — sociologically, not theologically — for a new text to enter the Shruti tier today, or is that door practically shut?

Key sources

The Hindus: An Alternative History — Wendy Doniger (2009). The most readable narrative survey of the text-lineage, including the politics of which texts got elevated.
The Sources of Indian Tradition, Vol. 1 — eds. Embree, Hay, de Bary (Columbia, revised 1988). Standard translated reader covering Shruti and Smriti excerpts.
A History of Indian Literature — Maurits Winternitz (German original 1908, English translation 1927). The classical philological survey; still the reference for textual chronology.
Patrick Olivelle's translations of the principal Upanishads (Oxford World's Classics, 1996) and Manu's Code of Law (2005) — the modern scholarly standard.
To verify: Frits Staal's work on Vedic oral transmission and its formal-system structure, especially Discovering the Vedas (2008).

Abhishek