to include the “five topics” in sufficient abundance. They were, he decided, “the
partial and adulterated representatives of an earlier class of Pura ̄n.as” (Wilson
1961: vi), embellished and altered in the interests of sectarianism so that they
promoted the worship of one particular god, usually Vis.n.u or S ́iva.
The idea of sectarianism is one which has been imposed upon the Hindu tra-
dition by Western scholars. They saw Vedic religion as the equivalent of the
Western Catholic church, from which various “sects” had broken away, becom-
ing “cut off ” (the literal meaning of the Latin “sectum”) from the central core
of their ancient tradition. But this model of an ancient rock from which seg-
ments can be separated, is not an appropriate one for Hinduism. Julius Lipner’s
metaphor of “an ancient banyan... an interconnected collection of trees and
branches in which the same life-sap flows” is much more appropriate (Lipner
1994: 5), opening up a perspective in which different branches of the Hindu tra-
dition can be linked together without declaring one to be a later development or
breakaway from another. In other words, the bhaktielement found in the
Pura ̄n.as is as ancient a rooting of this banyan as the Vedic one, even though it
may have taken longer to bear fruit in terms of texts and historically discernible
movements.
Whether or not sectarianism is an appropriate word to use in this context, the
Pura ̄n.as classify themselves as belonging to three groups, each connected with
one of the gun.asof Sa ̄m.khya psychocosmology and thus with one of the divini-
ties of the trimu ̄rti.Vis.n.u,Na ̄rada,Bha ̄gavata,Garud.a,Padma, and Va r a ̄haare
sa ̄ttvic and therefore associated with Vis.n.u;Matsya,Ku ̄rma,Lin.ga,S ́iva,Skanda,
andAgniare ta ̄masic (S ́iva); Brahma ̄n.d.a,Brahmavaivarta,Ma ̄rkan.d.eya,Bhavis.ya,
Va ̄mana, and Brahma ̄are ra ̄jasic (Brahma ̄). This division is found in Padma
5.263.81–4, and the connection with the trimu ̄rtiis given by Matsya53.68–9
(although these verses say that Agni, as well as S ́iva, is praised in the ta ̄masic
Pura ̄n.as, and acknowledges the existence of mixed Pura ̄n.as also, linked with
Sarasvatı ̄). It might appear to strengthen Wilson’s theory that the earlier
pañcalaks.an.amaterial had been displaced by later sectarian texts, but in fact this
neat threefold grouping, assigning exactly one-third of the Eighteen to each
gun.a, is entirely artificial.
To begin with, this is not the only self-classification given by the Pura ̄n.as,
although it is by far the best-known: Rocher quotes “an alternative, fivefold sub-
division, with pura ̄n.as dedicated to five different gods” (Brahma ̄, Su ̄ rya, Agni,
S ́iva, and Vis.n.u), found in the Skanda(Rocher 1986: 21). Furthermore, there are
no Pura ̄n.as which proclaim Brahma ̄ as the Supreme God, since Brahma ̄ does not
appear to have been regarded in this light on any significant scale (cf. Biardeau
1981: 175–8). Even Wilson was not convinced that the Pura ̄n.as listed in the
ra ̄jasic division ought to be connected with Brahma ̄, and pointed out that there
is a S ́a ̄kta tendency in some of those included in this list (e.g. Ma ̄rkan.d.eya,
Brahmavaivarta; Wilson 1961: xii). But even if this traditional division of the
Pura ̄n.as is seen as Vais.n.ava, S ́aiva, and S ́a ̄kta, it is still misleading, since there
are comparatively few Pura ̄n.as which are devoted exclusively to the worship of
one deity. There are some which can be described unambiguously as Vais.n.ava
the pura ̄n.as 137