What Are the Puranas? -- The Five Marks, the Eighteen Texts, and Why None of Them Had One Author

Purana means, roughly, 'ancient lore' or 'that which is old,' and it names one of the largest and least uniform categories in the entire Hindu scriptural corpus. Readers who arrive at this site's individual Purana articles -- on the Markandeya Purana's talking-birds frame story, the Shiva Purana's disputed place on its own canonical list, the Linga Purana's account of why Shiva is worshipped as a formless pillar, the Bhagavata Purana's theology of devotion, or the Kalika Purana's place among the secondary Upapuranas -- eventually run into the same more basic question each of those articles has to set aside to do its own job: what actually makes a text a Purana in the first place, and how many of them are there? This article is that orienting piece. Read it before or after any specific Purana article on this site; either order works, because the point here is the shape of the genre, not the content of any one text inside it.

The short answer, which the rest of this article unpacks, is that a Purana is traditionally defined not by a single story or deity but by a set of five recurring topics it is expected to cover, that eighteen texts are conventionally recognised as the great or 'Maha' Puranas, that a looser and less agreed-upon set of Upapuranas sits alongside them, and that modern textual scholarship treats every one of these works -- without exception -- as a composite product of many hands across many centuries rather than the output of a single author at a single sitting. None of this makes the Puranas less important to the traditions that hold them sacred. It describes, honestly, what living scripture assembled across a very long span of time actually looks like once anyone goes looking closely at its edges.

The traditional definition of a Purana runs through five characteristics, pancha-lakshana, most often traced to a verse preserved in several Puranas themselves and repeated in the classical lexicon Amarakosha. A text earns the name Purana, in this framework, by covering: sarga, the creation of the universe; pratisarga, its periodic dissolution and re-creation; vamsha, the genealogies of gods, sages, and royal lines; manvantara, the great cosmic ages, each presided over by a Manu, the progenitor of humanity for that era; and vamshanucharita, the continuing histories and deeds of the dynasties descended from those genealogies. Read this list against any specific Purana on this site and the pattern becomes visible immediately -- the Markandeya Purana devotes roughly forty chapters to exactly this kind of manvantara and genealogical material before and after its famous Devi Mahatmya section, precisely because a Maha Purana was never meant to be only a story collection.

It is worth stating plainly, in the same honest spirit this site brings to dating questions, that the five-fold definition describes an ideal template more than a strict checklist every Purana satisfies in full. Some Puranas lean heavily into one or two of the five categories and touch the others only briefly; others, particularly the later and more devotionally focused texts such as the Bhagavata Purana, wrap philosophy, hymn, and detailed narrative theology around the traditional five-part frame rather than organising themselves strictly by it. The pancha-lakshana is best understood as the genre's founding blueprint, present in some form in every text that claims the name, rather than a rigid formula that predicts a given Purana's actual table of contents chapter for chapter.

Eighteen texts are conventionally named the Maha Puranas, the great Puranas, and this list is one of the more stable pieces of the whole tradition, repeated with only minor variation across most later Puranic sources: Brahma, Padma, Vishnu, Vayu (sometimes replaced by or paired with Shiva in some regional lists), Bhagavata, Naradiya, Garuda, Markandeya, Agni, Bhavishya, Brahmavaivarta, Linga, Varaha, Skanda, Vamana, Kurma, Matsya, and Brahmanda. Some medieval lists substitute the Shiva Purana for the Vayu Purana or include both under slightly different names, and the Shiva Purana's own article on this site covers that specific disputed case in detail -- one of several places where the tidy number eighteen turns out to hide real manuscript-level disagreement about which text belongs on it.

Orthodox tradition attributes all eighteen to a single author, the sage Vyasa, in the same way it attributes the Mahabharata to him -- a claim that functions, across Hindu tradition, more as a statement of the Puranas' unified religious authority than as a historical record a manuscript scholar could verify chapter by chapter. Below the eighteen Maha Puranas sits a second, looser tier: the Upapuranas, or secondary Puranas, traditionally also numbered at eighteen but far less stable a set, since different sources naming the Upapuranas frequently disagree with each other about which texts belong on the list. The Kalika Purana, covered in its own article on this site, is a well-known example, generally placed twelfth on the Upapurana list preserved in the Kurma Purana, and valued in its own right as the textual source for much of the ritual detail surrounding Kamakhya worship, regardless of its secondary formal status.

The most important thing to carry away from all of this is the one point every individual Purana article on this site returns to in its own way: none of these eighteen-plus-eighteen texts is a single-authored, single-moment composition, whatever the Vyasa attribution claims. Modern manuscript scholarship, working text by text, has found the same underlying pattern almost everywhere it has looked closely. The Markandeya Purana carries an early narrative core dated by Wendy Doniger to roughly 250 CE, with its famous Devi Mahatmya section layered in some three centuries later. The Skanda Purana's oldest attested core, reconstructed by the Skandapurana Project from early Nepalese manuscripts, is a shorter and substantially different text from the vast regional compilation, including the Kashi Khanda and Sri Saila Khanda, that grew up around its name over the following six to eight centuries. The Shiva Purana's own claim to a place among the eighteen is itself contested in some medieval lists. This is not a pattern of a few damaged or unusually late texts among otherwise pristine ancient originals. It is the ordinary condition of the genre: living scripture, orally rooted and continuously copied by hand across a subcontinent and a couple of thousand years, accumulating material the way every long-lived oral-and-written tradition does.

None of this diminishes the Puranas within the traditions that hold them sacred, any more than the composite, centuries-long history of the Mahabharata diminishes its epic, or the layered compositional history of the Rigveda's tenth mandala diminishes the philosophical force of its creation hymns. It describes, accurately, what a scripture built by many hands actually looks like once anyone goes looking for its edges -- and it is the honest starting point for reading any single Purana on its own terms, rather than as a monolith it never claimed to be.