Nemec Two Pratyabhijna Theories of Error

Abstract In this essay, it is argued that Abhinavagupta’s theory of error, the

apurn: akhyati theory, synthesizes two distinguishable Pratyabhijna treatments of

error that were developed in three phases prior to him. The first theory was

developed in two stages, initially by Somananda in the �Sivadr: s: t: i (SD) and subse-

quently by Utpaladeva in his I�svarapratyabhijñakarikas (IPK) and his short auto-

commentary thereon, the I�svarapratyabhijñavr: tti (IPVr: ). This theory served to

explain individual acts of misperception, and it was developed with the philosophy

of the Buddhist epistemologists in mind. In a third phase, Utpaladeva developed in

his �Sivadr: s: t: ivr: tti (SDVr: ) a second theory of error, one that involved the

noncognition of non-duality (abhedakhyati) and served to explain both the

appearance and perception of multiplicity, despite the strict monism to which all

Pratyabhijna authors subscribe. Abhinavagupta’s treatment of error, then, is sig-

nificant not only because it was meant to explain all the various theories of error

offered by opposing philosophical schools, as Rastogi has shown, but more

importantly because it synthesized the thinking of his predecessors on the matter in

a single, elegant account of error.

Keywords Pratyabhijna � apurn: akhyati � abhedakhyati � abheda � Error � Buddhist

epistemologists

Introduction

Navjivan Rastogi has argued that the Kashmiri tantric polymath Abhinavagupta

(fl. c. 975–1025) espoused a theory of ‘‘error by commission,’’ one that claims that

perceptual error is constituted by the occurrence of an incomplete cognition

J. Nemec (&)Department of Religious studies, University of Virginia,PO Box 400126, Charlottesville, VA 22904-4126, USAe-mail: [email protected]

123

J Indian Philos

DOI 10.1007/s10781-011-9130-3

The Two Pratyabhijna Theories of Error

John Nemec

� Springer Science+Business Media B.V. 2011

(apurn: akhyati) that leads to an erroneous judgment regarding the nature of what is

perceived. He further argued that Abhinava simultaneously espoused a theory of

‘‘error by omission,’’ for he describes error as involving a noncognition (akhyati).1

By doing so, Abhinavagupta claimed that his own theory of error could accom-

modate those of every contemporaneous philosophical school, this by synthesizing

them with his own.2

Left out of Rastogi’s analysis, however, is any thorough study of the discussions

of error that appear in the writings of the Pratyabhijna authors who preceded

Abhinavagupta, namely Utpaladeva (fl. c. 925–975), Abhinavagupta’s paramaguru,

and Somananda (fl. c. 900–950), his parames: t:higuru,3 despite the fact that both

authors made substantial contributions to the Pratyabhijna’s treatment of the matter.

While an analysis of these authors’ views does not invalidate Rastogi’s conclusions

regarding the telos of Abhinavagupta’s inclusive nomenclature—that he meant to

account for all the various theories of error of all contemporaneous philosophical

schools—, the evidence found in the writings of Somananda and Utpaladeva also

indicates that Abhinavagupta’s apurn: akhyati theory reveals a significant debt to his

predecessors. Specifically, the apurn: akhyati theory synthesizes two distinguishable

Pratyabhijna theories of error that were developed in three phases prior to Abhi-

navagupta.

The earlier theory was developed in two stages, first by Somananda in the fourth

chapter of the �Sivadr: s: t: i (SD) and subsequently by Utpaladeva in his I�svaraprat-yabhijñakarikas (IPK) and his short auto-commentary thereon, the I�svaraprat-yabhijñavr: tti (IPVr: ). This treatment of error is concerned with ‘‘error by

commission,’’ and it serves to explain individual acts of misperception, including,

according to Utpala, both those caused by conceptual error (e.g., the perception of

silver for mother of pearl) and those caused by perceptual error (as in, e.g., the

perception of two moons due to the existence of a defect of the indriya).

In a third phase, Utpaladeva subsequently developed a theory of ‘‘error by

omission,’’ which is found in his �Sivadr: s: t: ivr: tti (SDVr: ). This second theory served to

explain the appearance of the multiplicity of the universe, despite the strict monism

to which all the Pratyabhijna authors subscribe, and it also was intended to explain

the capacity of the individual cognizer to experience multiplicity.

1 See Rastogi (1986, pp. 5–6 and passim). Rastogi (1986, p. 9) states that ‘‘error by omission’’ occurs

when one ‘‘…visualize[s] a particular aspect of the object and overlook[s] or miss[es] another.’’ On the

other hand, ‘‘error by commission’’ requires the agent of cognition to (ibid.) ‘‘…add something to the

object. It is not that we simply miss something of nacre itself but that we also bring something of silver to

it.’’ In this article, I use the term ‘‘error by omission’’ to refer to error resulting from the failure of an agent

of cognition to cognize a particular percept, or to distinguish between two percepts; and I use the term

‘‘error by commission’’ to refer to errors resulting from the agent of cognition’s active perception of one

percept for or as something other than that which appears in the ken.2 See Rastogi (1986, pp. 4, 23–25 and 33, fns. 110 and 111). See also I�svarapratyabhijñavimar�sinı (IPV)

ad I�svarapratyabhijñakarika (IPK) 2.3.13 (vol. 2, p. 113, line 11 to p. 114, line 2 of the KSTS edition);

and I�svarapratyabhijñavivr: tivimar�sinı (IPVV) ad IPK 2.3.13 (vol. 3, from p. 152, line 22 to p. 153, line 2

of the KSTS edition; and p. 154, lines 13–24 of the same).3 Nothing of what might have been written by Laks:man: agupta, Abhinava’s guru and the disciple of

Utpaladeva in the lineage of the Pratyabhijna, is known to us, as Pandey noted some time ago and as is

well known. See Pandey ([1963] 2000, pp. 164–165).

J. Nemec

123

Now, while Utpaladeva does not explain how these two explanations of error

square with one another, Abhinavagupta does, as both were subsequently incor-

porated into his synthetic explanation of error in his I�svarapratyabhijñavimar�sinı(IPV) and I�svarapratyabhijñavivr: tivimar�sinı (IPVV).4 Indeed, Abhinavagupta’s

primary contribution to the treatment of error in the Pratyabhijna is precisely this

synthesis, which explicitly associates the theory of ‘‘error by omission’’ with that of

‘‘error by commission’’ under the ruberic of an apurn: akhyati, or incomplete

cognition, a formulation that is unprecedented in the writings of Somananda and

Utpaladeva, even if it draws on a nomenclature found in the IPK and IPVr: .By examining the development of the Pratyabhijna treatment of error in four

phases (Abhinavagupta’s synthesis being the fourth and last phase here considered),

then, we will illustrate the ways in which the authors of this guru�sis:yaparamparainnovated on the philosophical writings of their predecessors while simultaneously

endeavoring to maintain continuity with them. Thus, the Pratyabhijna treatments of

error may serve as one exemplar of the dynamics associated with the sometimes

competing tensions between the need for philosophical innovation and the desire to

adhere to tradition.

The Treatment of Error in Somananda’s �Sivadr: s: t: i

While Torella has noted in passing the existence of Somananda’s treatment of

error,5 that author’s thoughts on the matter have not yet been examined at any length

in the secondary literature. Similarly, while Somananda’s engagement with the

ideas of the Buddhist Vijnanavadins is well known,6 the degree to which he engaged

with the ideas of the Buddhist epistemologist Dharmakırti in particular has yet to be

fully explored.7 Here, I shall argue that Somananda’s treatment of error is an

argument that is directed against Dharmakırti, and that in making his argument

Somananda defined the basic concepts that remain central to the treatments of error

found in the writings of Utpaladeva and Abhinavagupta, even if those authors

modify Somananda’s view in significant ways. These concepts include the notion

that all objects of cognition are real qua their appearing within consciousness, which

4 When quoting or referring to the IPK and IPVr: , I refer to the edition prepared by Torella. When

referring to the IPV and IPVV, I refer to the KSTS editions of those works. When referring to Bhaskara

Kan: t:ha’s commentary on the IPV, the Bhaskarı, I refer to the Iyer and Pandey edition.In all the passages of the SD and SDVr: quoted in the present essay, I refer to the readings of the KSTS

edition. I have, in addition, consulted six manuscripts of these texts. (The sources consulted for the editionare listed in the bibliography and include two manuscripts of the mula alone [T and C] and four thatinclude both the SD and the SDVr: [G, J, P, and R].) References to variant readings found in these sourcesare reported only where the readings in question require one to emend the KSTS edition (Ked.). When Inote a variant reading, I collate all of the readings of all of the manuscripts in a positive apparatus andinclude the readings of the KSTS edition in doing so.5 See Torella (1994, p. 139, fn. 11).6 See Torella (1994, pp. xv–xvi.)7 Torella (1994, p. xxii, fn. 28) has noted, however, that Somananda makes mention of Dharmakırti’s

conception of the svalaks:an: as (SD 4.81a), apoha (SD 4.76c), svarthanumana (SD 5.55a), and para-numana (SD 5.61c) in the SD, to which list should be added sadr: �sya (SD 4.86a) and arthakriya (SD

4.21a, 6.59a). See also Torella (2008, p. 516).


123

is itself thought to be the very nature of Siva, as well as the notion that it is the will

(iccha) of Siva that defines the nature of the reality that one experiences. Though he

does not explicitly say so, Somananda also appears to argue that it is not possible to

explain how a cognition can be invalidated unless one acknowledges the existence

of a single cognitive agent, Siva, an idea that is central to Utpaladeva’s theory of

error in the IPK and IPVr: . Finally, Somananda’s theory is of necessity one of ‘‘error

by commission,’’ a position adopted by Utpaladeva in his IPK and IPVr: and sub-

sequently incorporated into Abhinava’s inclusive apurn: akhyati theory.

Somananda’s view may be summed up simply with the idea that all appearances

are manifested in/as Siva’s very power of consciousness, for all of existence is

contained in/as the experience of Siva-the-yogin.8 As such, all appearances—all

cognitions—are real, including apparently erroneous cognitions such as the per-

ception of silver where only mother of pearl exists; and thus, such cognitions are

‘‘erroneous’’ only in a conventional sense.9

Assuming a momentary, self-revealing nature of all cognitions, Somananda

suggests that it is impossible for any cognition to be invalidated (badha). This is so,

he argues, because the fact that the cognition in question appeared in consciousness

cannot be denied or undone in a subsequent moment of time: it undoubtedly existed

insofar as it appeared within one’s awareness, and as such it cannot be said to be

unreal, reality being defined as the fact of appearing in/as the contents of a divine

and all-pervasive consciousness. Thus, the act that produces the erroneous

appearance cannot be invalidated, because the act is momentary and ceases to exist

8 See SD 1.44-45ab: yoginam icchaya yadvan nanarupopapattita / na casti sadhanam: kiñcin mr: dadı ccham:vina prabhoh: . tatha bhagavadicchaiva tathatvena prajayate. ‘‘Just as various forms are produced by the will

of yogis, and there is no extrinsic means whatsoever, be it the clay or something else, apart from the Lord’s

will, in the same way, it is the Lord’s will that arises as reality.’’ Cf. IPK 1.5.7.9 See SD 4.7cd-9: mithyajñanavikalpanam: (�vikalpanam: TCGJPR; �vikalpyanam: Ked.) sattvam:cidvyakti �saktata. vidyate tat tad atrapi �sivatvam: kena varyate / iti ced es:u satyatvam: sthitam evacidudgamat. tatha �sivodayad eva bhedo mithyadikah: katham / vyavaharaya satyatvam: na ca vavyava-haragam. ‘‘Vikalpas that are erroneous cognitions are found to be capable of appearing in consciousness.

That(, itself,) is what it is to be real, which is also the nature of Siva here [in our philosophy]. If you object

by asking by what [this siva-nature] is averted, (we reply:) reality absolutely exists in these [vikalpas],

because they arise in consciousness. Thus, since they simply arise as Siva, how could a distinction exist

that is associated with that which is erroneous, and the like? That [distinction] serves the everyday world,

and it is not the case that reality does not reach the everyday world.’’ See also Utpaladeva’s commentary

on the same (where the anvaya here followed is clarified): mithyajñanavikalpanıyanam: rajatasarpadınam:ghat: adınam iva cidvyakti�saktata praka�samanata nama vidyate, tad eva sattvam: paramarthatvam ato ’pitac chivatvam. anivaryam: caitad is:yata evais:u satyatvam: cidabhivyakteh: , tena tena rupen: a �sivasyaprasaro ’yam: yatah: . ghat:ajñanarajatajñanayo�s ca dvayor api satyatve samyan_ mithyatvabhedas tarhikatham. vyavaharaya. vyavaharaprayojano ’sau. na caprarud: hatvad vyavaharasatyatvam: vasatyatvam:bhavati. ‘‘Those that should be doubted, erroneous cognitions, such as (the cognition of ) silver (for

mother of pearl) or (the cognition of ) a snake (for a rope), like (cognitions of ) pots, etc., are found to be

capable of appearing in consciousness, i.e., they really are manifested therein; that itself is what it is to be

real, i.e., that is the nature of the supreme reality, which also means that is the nature of Siva. Moreover,

we simply accept that this [Siva-nature] cannot be averted: the reality in these exists because they are

manifested in consciousness, since this is the flow of Siva in various forms. And if both the (‘‘correct’’)

cognition(, e.g.,) of the pot and the (‘‘erroneous’’) cognition(, e.g.,) of silver are equally real, then how

does a distinction exist in (so-called) erroneousness? That serves the everyday world, i.e., that has a use in

the everyday world, and the reality of the everyday world does not become an unreality as a result of

being less than fully developed.’’ Compare, also, with IPK 4.12.

J. Nemec

123

the moment after it occurs; as such, it is not possible and indeed there is no need for

it to be invalidated in a subsequent moment.10 The same is true of one’s experience

(anubhava) of the object of cognition, Somananda argues, because it is not possible

for an object that has been experienced to become one that was not experienced.11

It is also impossible for the locus (de�sa) of the erroneous cognition to be

invalidated in a subsequent moment of time, according to the SD, because,

immediately following one’s cognition of the object in question, the place where the

cognition in question was seen ceases to exist.12 Similarly, the cognition of silver

cannot be invalidated by another cognition (jñanantara), because the initial

cognition, being self-contained and momentary, no longer exists when the latter

cognition appears.13

Somananda further denies the possibility that the organ of sense (indriya) by

which the cognition is experienced may be invalidated, because there must be some

means of knowing that which appeared in the ken. In addition, that which was

erroneously seen appeared in the context of a cognition that correctly perceived the

entities surrounding the object in question. For example, the surface on which rests a

piece of mother of pearl that was erroneously perceived to be a ball of silver was

itself truly present and was simultaneously cognized properly. Therefore, to say that

the indriya of the erroneous cognition may be invalidated calls into question the

validity of the appearance of such contemporaneously appearing and correctly

10 See SD 4.16b-d: …va janikriya / kr: tva karyam: kriya yata gatayam: kim: prabadhyate. ‘‘Alternatively,(you might argue) it is the act of production (of the erroneous cognition that is invalidated). (We reply:)An action, having produced its effect, elapses. What is invalidated when it is (already) gone?’’ See alsothe SDVr: on the same: janikriya va yadi badhyate, tan na. rajatadikaryam: kr: tva kriya samapta. tasyaasatyah: ko badharthah: svayam eva tasya abhavat. ‘‘Alternatively, if you argue that it is the act ofproduction (of the erroneous cognition) that is invalidated, we reply: not so. The action, having producedits effect, (the cognition of) silver, etc., is completed. What is the purpose of the invalidation of that[action] which no longer exists, since it expires entirely of its own accord?’’11 See SD 4.17ab: athanubhavaga badha nanubhuto ’nyatha bhavet. ‘‘Now, you might argue the

invalidation is associated with the experience. (We reply:) That which has been experienced cannot

become otherwise.’’ Cf. Utpaladeva’s gloss on the same: atha yadi rajatanubhavagata badhes:yate, tanna. na hy anubhuto vis:ayo ’nanubhuto yuktah: . ‘‘Now, if you wish to argue that the invalidation is

associated with the experience of silver, we reply: not so. For, it is not tenable that an object that has been

experienced is one that has not been experienced.’’12 See SD 4.21cd-22: atha ced de�sabadho va tadde�se rajatam: na hi. yatra kale sarajato de�so ’bhut sagatas tada / kalantaren: a de�so ’sau ka badha bhinnakalayoh: . ‘‘Now, if you instead argue that there is an

invalidation of the locus, since silver does not exist in that place, (we reply:) the moment in which the

locus containing the silver existed subsequently expires. That locus [i.e., the one in which the non-

appearance of silver is seen] exists in a different moment of time. What invalidation is there for two [loci]

that exist in different moments of time?’’13 See SD 4.23: jñanantaren: a jñanam: tadvirodhad atha badhyate / na badho bhinnakalatvat prakta-nasyapy abhavatah: . ‘‘Now, if you argue that the cognition (of silver) is invalidated by another cognition

[i.e., by that of mother of pearl], since it is incompatible with it, (we reply:) there is no invalidation since

the preceding [cognition] itself no longer exists given that it occurs in a different [i.e., earlier] moment of

time.’’


123

perceived phenomena, which is unacceptable.14 Neither can one claim that the

connection (sam: bandha) of the indriya to the object is invalidated, he argues, as this

would render it impossible for the cognition in question to have occurred.15 Finally,

Somananda argues that the behavior (vyavahara) of the erroneous cognition cannot

be the source of error, because all activity occurs by the will of Siva, and as such no

dichotomy may be established on the basis of the legitimate or illegitimate behavior

of the contents of any given cognition.16

If these arguments against the possibility of the existence of any invali-

dated-invalidator relationship (badhyabadhakabhava) reveal only Somananda’s

14 See SD 4.17cd-18c: athendriyasya badhyatvam: tatkalam: yadr: g indriyam. tadandhyam: janyate kenatasya kalantarasthiteh: / sarvaih: samatvam: … ‘‘Now, you might argue that the indriya is invalidated. (We

reply:) Whatever kind of indriya that exists during that moment of time [i.e., when the silver was

cognized] would be made blind. By what is this produced for that which is situated in another moment of

time? It would be the same with all [cognitions].’’ The SDVr: explains the present, somewhat laconic

passage as follows: atha rajatadvicandradijñanakaran: am indriyam anindriyam iti badhyate. tan na. na hibadhakenendriyasyandhyam abhavah: kartum: �sakyam: , tadrajatadijñanakale yadr: g indriyam: tasyatadanım: ghat: adınam upalabdheh: . atha kalantare yada badhakotpattih: , tada tasyanavasthanad badhah: .tatkalantarasthiter badhopagame tadanıntanasamyagjñanair api mithyajñananam: samatvam: namabadhah: prapnoti. na caivam: yuktam: , vartamanakalabhavinam: ghat: adijñananam: samyagtvena sam: bha-vat. ‘‘Now, you might argue that the indriya that causes the cognition of silver, of two moons, and the like

is not a valid means of knowledge and therefore is invalidated. We reply: not so. For, that which

invalidates the indriya is not able to make it blind, i.e., make it a non-entity: some kind of indriya exists

for that [cognition of silver] in that moment of time, i.e., in the moment of the cognition of silver, etc.,

because it cognizes pots, etc., [i.e., what is real and is seen along with the erroneous cognition of silver] at

the same time. Now, you might argue that, when that which invalidates it is produced in a subsequent

moment of time, the invalidation occurs because it [i.e., the indriya] is unstable. (We reply:) When you

accept that the invalidation is of a condition that exists in a moment of time subsequent to that [cognition

of silver], the invalidation obtains equally for erroneous cognitions and for the correct cognitions that

exist concurrently, as well. And this is not tenable, because the cognitions of pots, etc., that exist

contemporaneously [i.e., the cognitions of what is real and is seen along with the erroneous cognition of

silver,] correctly appear.’’15 See SD 4.18cd: …badho va sam: bandhe jananam: katham. ‘‘Alternatively, (you might argue) the

invalidation is of the connection. (Reply:) Whence the production (of the cognition in question)?’’ The

Vr: tti explains this passage in part as follows: athendriyavis:ayadınam: sam: bandhe badhah: . tan na.sam: bandhabadhe hi katham: jñanajananam. ‘‘Now, if you argue that the invalidation is of the connection

of the object, etc., to the indriya, then we reply: not so. For, how could the cognition be produced when

the connection is invalidated?’’16 The justification for this statement is made by way of analogy. If a king issues alloyed coins that are

erroneously valued, so to speak, this does not invalidate the trade that occurs with those coins. Trans-

actions occur in the kingdom according to the rules established by the king. By analogy, then, one cannot

say that the behavior of the silver is erroneous while that of mother of pearl is not: each appears as it does

according to Siva’s wishes. See SD 4.19-20ab: vyavaharasya badha ced vyavahare yathes: t:ata / kvacitsatyasuvarn: asya pratyante vyavaharita. kut:akars: apan: adau va vyavaharo ’pi dr: �syate. ‘‘If you argue that

the invalidation is of the behavior, (we reply:) the behavior is as it is wished to be. Somewhere in a

bordering country, real gold is the currency of trade; on the other hand, one also sees trade in alloyed

coins and the like.’’ Cf. Utpaladeva’s Vr: tti: vyavaharo ’pi badhyate rajatavyavaharo na kartavya iti. tadapi na, vyavaharasyes: t:asya pravr: tteh: . na tadva�sena satyasatyavibhagah: . tatha hi pratyantade�se kvacitsatyasuvarn: asya tatha vyavaharo, yadr: gayasah: . kut:akars: apan: adina ca rajña is: t:atvat dr: �syatevyavaharah: . ‘‘You might say: the behavior itself is invalidated, i.e., the behavior of the (appearance of)

silver is not that which ought to done. This, too, is not so, because a wished-for behavior appears. No

distinction of real from unreal exists by dint of [the behavior]. For instance, somewhere in a bordering

country trade occurs appropriately with real gold, i.e, with some kind of metal. And one (also) sees trade

with alloyed coins and the like because the king wishes it.’’

J. Nemec

123

acquaintance with the general features of the theory of cognition of the Buddhist

epistemologists, who argue for precisely the momentary and self-contained form of

consciousness that Somananda here accepts, the SD elsewhere makes it clear that

Somananda had Dharmakırti in mind when he made these arguments. This is so,

because the SD refers twice to ideas that are peculiar to Dharmakırtian philosophy.

The first reference invokes a two-fold classification of contradiction (virodha) that is

uniquely Dharmakırti’s:17 Somananda argues that neither of two types of contra-

diction can describe the relationship between an erroneous cognition and a sub-

sequent, correct cognition. The form of contradiction caused by the co-presence of

mutually distinct entities—sahanavasthana or ‘‘factual contradiction’’18—, he

argues, cannot account for a contradiction between the cognition of silver, on the

one hand, and of mother of pearl, on the other, because the cognition of the former is

destroyed prior to the appearance of the latter.19

It is also impossible for the two cognitions to be contradictory on the basis of

having mutually exclusive forms—for the contradiction to be one due to the

existence of mutual exclusion, what Dharmakırti describes as virodha resulting from

17 As Bandyopadhyay (1988, pp. 229–232) has shown, the distinction here enumerated involves that of

factual from logical contradiction, one that Dharmakırti alone makes in the present fashion. See NB 3.74:

dvividho hi padarthanam: virodhah: . ‘‘Opposition between objects is of a double kind.’’ One form is found

in the impossibility of the co-presence of two things (sahanavasthana), as with heat and cold. (See NB

3.75—avikalakaran: asya bhavato ’nyabhave ’bhavad virodhagatih:—and 3.76: �sıtos:n: aspar�savat: ‘‘When

(one fact that) has duration (as long as) the sum-total of its causes remains unimpaired, and it (then)

vanishes as soon as another, (the opposed,) fact appears, it follows that both are incompatible, just as the

sensations of heat and cold.’’) The other results from the existence of mutually exclusive forms (para-sparaparihara), such as existence and non-existence (bhavabhavavat). See NB 3.77: para-sparapariharasthitalaks:an: ataya va bhavabhavavat. ‘‘There is also (opposition between two facts) when

their own essence consists in mutual exclusion, as between the affirmation and negation (of the same

thing).’’ The translations here quoted are Stcherbatsky’s, for which see Stcherbatsky [1932] 1993 (vol. 2):

187 and 192.18 The contradiction is factual and not logical because it involves the appearance of an element in a

context in which a mutually contradictory one existed previously. As Bandyopadhyay (1988, p. 231),

referring to NB 3.75, argues, ‘‘we get this virodha when one fact is replaced by an opposed fact, e.g.,

cold-sensation is eliminated by hot-sensation, or darkness by light.’’ It is a ‘‘purely factual contradiction,’’

she continues, ‘‘for here we really get two facts one replacing the other, at the same place.’’19 See SD 4.24ab: sahanavasthitir nasti virodhah: pragvina�satah: . ‘‘There is no contradiction caused by

the co-presence of mutually distinct entities, due to the prior destruction (of the earlier cognition, before

the subsequent cognition arises).’’ See the SDVr: on the same, which echoes NB 3.75–76: sahanavasthitirapi virodho nasti. na hi jñanam: ks:an: ikatvena pragvinas: t:am anyena nivartanıyam us:n: ena �sıtam ivaviruddham: bhavati rajata�suktijñanayoh: . ‘‘There is no contradiction caused by the co-presence of

mutually distinct entities, either. For, when it comes to the cognitions of silver and mother of pearl, the

cognition that on account of its being momentary is destroyed in an earlier moment is not suppressed by

another [cognition], as cold is invalidated by heat.’’ Cf. IPVV ad IPK 1.7.6 (vol. 2, p. 373, line 10 to

p. 374, line 4 of the KSTS edition).


123

parasparapariharasthitalaks:an: ata.20 To invalidate the earlier cognition on this

basis would require one to claim that the cognition in question has a self-contra-

dictory nature: it would have to be a form of awareness that is a lack of awareness.21

In another passage found in the fourth chapter of the SD, Somananda challenges

a fundamental tenet of the philosophy of Dharmakırti and those who followed him,

namely the notion that arthakriya or ‘‘causal efficacy’’ (alternatively rendered by

Dunne as ‘‘telic function’’22) is a sufficient criterion for determining what is real.23

The reference in question is found in the context of Somananda’s expression of a

theological claim, wherein he argues that the very existence of any cognition

necessarily involves Siva’s desire to cognize some wished-for object, the power of

cognition in his view being but one in a sequence of powers beginning with and

initiated by Siva’s power of will or desire (iccha).24 Because Somananda conceives

20 Kyuma (1997, p. 225) suggests that with parasparaparihara Dharmakırti refers to the form of

‘‘incompatibility’’ (virodha) that ‘‘stands between a property or concept and its negation, e.g., the concept

of being eternal (nityatva) and the concept of being non-eternal (anityatva).’’ As Bandyopadhyay (1988,

pp. 229–231) has shown, however, Dharmakırti includes both the form of logical contradiction that obeys

the law of the excluded middle (referred to as parasparavirahavyapakatva) and the form that does not

(parasparavirahavyapyatva) under this single heading. The former is akin to the contradiction between X

and not-X, exemplified by Dharmakırti with the contradiction between being and non-being; the latter is

similar to the opposition of the color black to the color while, which does not exclude a middle term,

being red in color, for example. Cf. Kellner (1997, pp. 496–497). Somananda deals only with the former

type, however.21 Unfortunately, Utpaladeva’s commentary on the present passage survives only in a fragment that does

little to illuminate Somananda’s text, but the idea seems to be as follows. The opponent wishes to suggest

that there is a difference between the earlier and later cognitions, as the former does not furnish awareness

of an object in the ken, while the latter does. Thus, there exists a contradiction by the mutual exclusion of

awareness and an absence of awareness of what is truly present in the ken. Somananda replies by

suggesting that the mutual exclusion in question would rather apply to the selfsame cognition, the earlier

cognition of, e.g., silver. This is so because the cognition in question is thought to furnish awareness of its

object when it causes one erroneously to see some object, but is thought not to furnish that awareness

when it is invalidated. See SD 4.24cd-25: anyonyapariharo va jñanajñanatmakah: sthitah: . ajñanatveparijñate tada syat svavirodhita / ajñanatve svabhavena virodhah: kena varyate. ‘‘Alternatively, (you

might argue that) there exists a contradiction by the mutual exclusion of awareness and the absence of

awareness. (Reply:) When an absence of awareness is perceived, it is necessarily self-contradictory at that

time. When the nature (of the cognition) is that of a nonawareness, there is a contradiction (of this nature)

with its own nature. By what could this [contradiction] be averted?’’22 See Dunne (2004, pp. 256–260). Dunne intended with this translation to invoke both the ontological

and epistemological meanings of the term that were first identified by Nagatomi (1967–1968). Cf.

Dreyfus (1997, pp. 65–67).23 On Dharmakırti’s innovation of the concept of arthakriya see Hattori (1968, pp. 14 and 79); and Hayes

(1997, pp. 105–107).24 The sequence of powers is: nirvr: ti, aunmukhya, iccha, jñana, and kriya. Siva is, according to

Somananda, an innately blissful agent, one who creates simply out of a desire to do so. Thus, nirvr: ti is a

form of delight, and aunmukhya is a subtle, first movement of the power of will (see SD 1.7cd-8; cf. SD

1.9-11ab and 1.13cd-17), which precedes the full-blown form, iccha. Will, in turn, directs and shapes

cognition and action. It is that which determines how these powers will be manifested (see SD 1.19-20ab).

The power of will (iccha) is also that which directs both the creation and the appearance of the form of the

universe, for which see SD 3.33cd-47. It is also that which creates the distinction between unity and

multiplicity, for which see SD 3.35-36ab. Finally, will is that which causes beings to be cognizant or

ignorant of their own inherent freedom, for which see SD 3.72. I have explained the functioning of this

sequence of powers in detail in Nemec (forthcoming).

J. Nemec

123

of Siva as the selfsame agent as the individual who cognizes both the appearance of

silver and of mother of pearl,25 every cognition is therefore necessarily real insofar

as it appears merely to satisfy Siva’s desire to experience the object(s) in question.

Therefore, Somananda argues, the primary standard by which Dharmakırti measures

the reality of the contents of a particular cognition—arthakriya—may itself be

measured only by the degree to which the divine agent, Siva, is delighted by the

contents of his cognitions.26 As such, the capacity for an object to complete a certain

action may not serve as a criterion by which to measure its reality or erroneousness.

In addition to this pair of references to Dharmakırtian philosophy, Somananda

also frames the entire discussion of error as a refutation of the same author’s

understanding of the nature of hallucinations. This is so, because the entire treat-

ment of error found in the fourth chapter of the SD serves to answer a promise that

Somananda made in the first, where, quoting the Praman: avarttika, he committed

himself to proving that all cognitions are real, even those produced by one’s intense

emotional states, such as fear.27 Somananda opens the fourth chapter by stating that

he will prove therein that everything is of the nature of Siva and, therefore, is real.

What he means to suggest is that all cognitions are real—the fourth chapter is

devoted to proving the reality of cognitions—including, presumably, the halluci-

nations that Dharmakırti describes.28 Unfortunately, however, Somananda does not

25 Somananda identifies the Siva who creates the universe with the individual, apparently limited agents

who populate it in SD 1.1, 1.2, and 1.34. He also suggests that the same powers that Siva employs to

manifest all of creation (mentioned in SD 1.20cd-21) operate in the same sequence in any individual

action or cognition (SD 1.22; cf. SD 1.6cd-7ab), this because only one agent exists who engages all

human and other activity. On the solipsism of the Pratyabhijna, see also Ratie (2007).26 See SD 4.20cd-21ab: tavata vyavaharo va yad atmahladamatrakam / arthakriyasamarthatvam etadevasya vastavam. ‘‘It is rather that behavior exists to the extent that the self is delighted, which is itself the

real capacity this [object] has for causal efficacy [arthakriya].’’ As Utpaladeva’s gloss explains, actions

have causal efficacy only in the consciousness of the agent who performs them, and as such the per-

formance even of imaginary actions has an effect for the agent who cognizes them. He offers as an

example an instance of one drinking water in a dream. In such an instance the agent experiences delight in

his or her own consciousness as a result of quenching his or her thirst, despite the fact that the action itself

is ephemeral, the water that was drunk being only imagined and being bereft of any potential to nourish

the body. See SDVr: ad SD 4.19-21ab.27 The reference to Dharkakırti was first identified by Gnoli (1957, p. 18). See also SD 1.45cd-46ab and

compare with Praman: avarttika (pratyaks:apariccheda) 3.282, also quoted by Gnoli.28 See SD 4.1ab: athedanım: pravaktavyam: yatha sarvam: �sivatmakam. ‘‘Next, we must now explain how

everything is possessed of Siva-nature.’’ While Somananda nowhere refers to hallucinations in the fourth

chapter, one can surmise that he meant the passages here reviewed to prove the reality of such cognitions,

because the fourth chapter is the only place where Somananda deals with perceptual illusion of any kind.


123

offer any detailed treatment of Dharmakırti’s theory of hallucinations, and so we do

not know how much he wished to object to the specific features of this theory.29

What we are to make, more generally, of Somananda’s litany of arguments

against the possibility of invalidating any apparently erroneous cognition is also

somewhat unclear. Does he wish to suggest that no cognition may be invalidated at

all, the error being merely an idea that serves only mundane purposes? Or, does he

wish instead to suggest that no cognition may be invalidated in the absence of the

existence of a single agent of cognition, something for which Utpaladeva argues in

the IPK and IPVr:? While the SD is not entirely clear on the matter, it seems highly

likely that Somananda wished to make the latter argument, which he invokes

explicitly when addressing the functioning of the karakas, this immediately fol-

lowing his treatment of the nature of virodha and badha here reviewed.30

Somananda is similarly silent when it comes to explaining the nature and

appearance of erroneous cognitions, though Utpaladeva explicitly addresses these

issues in his IPK and IPVr: . For, Somananda never explains what distinguishes real

from erroneous cognitions, but instead merely assures the reader that the distinction

is found in the agent of cognition’s understanding and not in the nature of the

cognitions themselves.31 And he nowhere explains how an erroneous cognition

29 For Dharmakırti, the hallucinations in question are closely related to what he counts as the fourth of

four types of perception, viz., yogic perception or yogipratyaks:a, as is well known. (The four types of

perception recognized by Dharmakırti are: sense-perception [indriyajñana], mental cognition [manasa-jñana], self-consciousness [svasam: vedana], and the cognition of yogins [yogijñana]. See, e.g., Woo 2003,

p. 439. On whether Dignaga meant to refer to this tetradic classification see Franco 1993; cf. Franco

1986.) Dharmakırti understood yogic perception to involve a series of three stages. These stages include:

intensity (prakars:a), termination (paryanta) and, finally, the direct cognition of yogins (yogipratyaks:a).

(See Woo 2003, p. 440.) By forcing a particular object of meditation into his consciousness, Dharmakırti

suggests, the yogin may cultivate the direct, non-conceptual cognition of complex phenomena, such as the

Four Noble Truths, that are rightly counted as conceptual (vikalpaka) phenomena. The experiences in

question, moreover, are vivid: they inspire alert reactions that are absent when one merely thinks about

something. (See Dunne 2006, pp. 499–500.) And such experiences therefore may be counted as both

trustworthy (sam: vadi) and reliable (praman: a) cognitions: they directly reveal truths the knowledge of

which leads one forward on a soteriologically efficacious path. (This is Dunne’s understanding of the

nature of yogipratyaks:a, anyway, for which see Dunne 2006, esp. pp. 514–515.) Thus, while Dharmakırti

understands yogic perception to share the vividness of hallucinations that are inspired by intense emo-

tional states, the perceptions it produces are distinguishable from hallucinations by the fact that they

reveal ‘‘true’’ (bhuta) objects that guide one on the path to spiritual liberation. (This is again Dunne’s

interpretation, contra Hayes, who according to Dunne argues that the effect of yogic perception is to

validate Buddhist doctrine and dogma. See Dunne 2006, p. 515; cf. Hayes 1997, esp. pp. 106–107.)Unfortunately, Somananda nowhere tells the reader what he understands of the nature of the Dhar-

makırti’s hallucinations. He merely suggests that all cognitions are real insofar as they appear within/asSiva’s consciousness in a manner Siva desires, as we have already seen. It is therefore unclear whetherSomananda wished simply to suggest that all cognitions are real, or if he instead wished in addition todispute more specifically the distinction Dharmakırti draws between true yogic perceptions (such as thatof the Four Noble Truths) and the hallucinations in question.30 See SD 4.32a-c: vinaikatvam: ca na bhavet karakatvam: kadacana / �sa�sa�sr: n_gadike... ‘‘And the nature of

the karakas could never exist for (expressions referring to non-existent entites,) ‘the horn of a hare,’ for

example, in the absence of unity.’’31 See SD 4.26: naivam atra svabhavatve virodho badhanatmakah: / sa vivekadr: �sa jñeyo na svabhavenakutracit. ‘‘In this way, there is no contradiction in the form of invalidation here [in our view] in the nature

of the nature (of phenomena); that [i.e., contradiction] may be known by way of a discriminating vision,

nowhere by way of the nature (of a given phenomenon).’’

J. Nemec

123

ceases to appear; that is, he does not explain how an erroneous cognition may be

corrected.

Instead, he offers merely an analogy to explain the appearance of erroneous

cognitions: just as monetary transactions occur in kingdoms where the king has

alloyed the currency with cheap metals, so too does the world continue to function

in an orderly manner, even if the contents of Siva’s consciousness reflect the

existence of an apparently mundane, diverse everyday reality (vyavahara).32 Even

the erroneous ideas one might have about a cognition that lead one wrongly to see,

e.g., mother of pearl as silver, are themselves real vis-a-vis the fact that they appear

within consciousness. For, Somananda suggests that all concepts are as real as any

direct cognition, as he maintains, contra Dharmakırti, that the vikalpas are possessed

of a real form as much as is the object of non-conceptual perception, because they

are equally possessed of Siva’s form insofar as they are manifested in conscious-

ness.33 On the other hand, Somananda simultaneously acknowledges that some

cognitions should be counted as erroneous by the conventions of the everyday

world, as we have already seen. And he therefore allows for some cognitions to be

real without being empirically valid. In other words, Somananda effectively dis-

tinguished the real from the empirical nature of cognitions, a distinction Rastogi

finds in Abhinavagupta’s oeuvre, but one that has its origins in the Pratyabhijna with

the SD.34

To conclude: while having a real legacy in the form of the influence he had on the

authors of the Pratyabhijna who followed him, Somananda’s treatment of error, the

first phase in the development of the first Pratyabhijna treatment of error, offers

something of an unfinished and in places unclear analysis of the problem. What is

certain, however, is that Somananda’s understanding of error is of necessity one

involving an ‘‘error by commission,’’ for the supposedly erroneous cognitions

undoubtedly appear to the perceiver and are real, in Somananda’s view, simply vis-

a-vis the fact that they inevitably appear within/as a moment of consciousness,

which can be nothing other than Siva himself.

Error in Utpaladeva’s Isvarapratyabhijnakarikas and -vr: tti

A second stage in the development of the first Pratyabhijna theory of error is

recorded in the IPK and IPVr: of Somananda’s disciple, Utpaladeva, a pair of texts

32 See SD 4.10-11ab: tatha ca de�se kvacana rajajña jayate yatha / vyavaharo ’stu dınarair etairvyavaharagaih: . pravartate tathabhutair anyatrapi tathanyatha. ‘‘Accordingly, in a given kingdom trade

must occur with the coins associated with trade (therein), according to the king’s order. So, elsewhere,

too, does it [i.e., the vyavahara or behavior] operate, mutatis mutandis, with those of such a kind.’’

Compare with SD 4.19-20ab, quoted in footnote 16, above.33 To be real in Somananda’s view involves the mere fact of appearing in consciousness, as we have

seen, and it is therefore impossible in his view for something to exist that does not appear in con-

sciousness, just as he argues that it is impossible for something that appears in consciousness not to be

real. See SD 4.25-31. As Utpaladeva points out in his commentary, the same idea is expressed in IPK

1.5.2. Cf. Torella (1992).34 I hasten to add that Rastogi does not assert that the idea is original to Abhinavagupta, however. See

Rastogi (1986, pp. 4, 19–22).


123

that was written prior to the composition of the SDVr: .35 Utpaladeva’s treatment of

error therein presupposes the major tenets of the SD’s analysis while also filling

important lacunae in Somananda’s system and adding a degree of sophistication and

nuance to the Pratyabhijna arguments against the Buddhists. Notably, Utpaladeva

repeatedly argues explicitly for the indispensability of the existence of a single,

unifying agent of cognition to any coherent explanation of the possibility of the

invalidation of an erroneous cognition, a position that he maintains when arguing

against Dharmakırti’s understanding of non-perception (anupalabdhi). He also

identifies a new criterion by which to measure the erroneousness of cognitions, just

as he explains how it is possible for erroneous cognitions to be corrected.

Utpaladeva adopts Somananda’s ideas regarding perceptual error in various

ways. First, he reiterates his teacher’s claim that all cognitions are real qua their

appearance in consciousness,36 and makes clear that he assumes that all cognitions

are momentary and self-contained.37 Second, like Somananda, Utpaladeva under-

stands vikalpas to be possessed of the same nature as non-conceptual, phenomenal

appearances that are manifested in the everyday world, a view that of course stands

in contrast to that of their Buddhist interlocutors.38 Finally, Utpaladeva asserts that

arthakriya is a function of Siva’s will and is therefore not simply inherent in the

35 The order of composition of Utpaladeva’s philosophical writings has been known for some time, this

in part on the basis of the frequent references in the SDVr: to both the IPK and, in one place at least, to

Utpaladeva’s longer auto-commentary thereon, the I�svarapratyabhijña-t: ıka or -vivr: ti, which now survives

only in fragments. (See Pandey [1963] 2000, p. 163; cf. Torella 2008, pp. 513–514.) Some of the

surviving fragments of the T:ıka (ad IPK 1.3.6–1.5.3) have recently been published, for which see Torella

(2007a, b, c, d, forthcoming).36 See, e.g., IPK, and IPVr: ad, 1.3.6-7. See also IPK 1.5.1-2, where Utpaladeva suggests that only that

which is identified with the agent’s internal power of consciousness may appear externally. (On the

presence of this idea in the Vijnanavada, see Torella 1994, p. 111, fn. 5.) Utpaladeva further argues that,

unlike the awareness one might have of a pot and its opposite, there exists no opposite to the light of

consciousness (for which see IPK 1.6.2). Indeed, it is not even possible to infer the existence of an object

that is not manifested in/as the light of consciousness, according to Utpaladeva, about which see IPK (and

IPVr: ad) 1.5.9.37 See, e.g., IPK 1.2 and 1.3.38 According to Utpaladeva, a vikalpa is an instance of determinate knowledge of a thing that is known in

relation to its opposite by an agent of knowing in whom both that which is known and its opposite are

manifested (see IPK 1.6.3). This occurs in the context of a universe that is manifested by reflective

awareness (pratyavamar�sa), the reflective light of consciousness itself, which is one, undivided. All

phenomena manifested in the universe, be they conceptual or non-conceptual in nature, are therefore

manifested by the same power of reflective awareness, according to Utpaladeva. The relationship of

vikalpas to the reflective awareness that constitutes the manifested universe is discussed in some detail in

IPK 1.6. See also Torella (1992, 1994, p. 128, fn. 4).

J. Nemec

123

nature of the svalaks:an: as appearing in/as the phenomenal world, as Dharmakırti

argues.39

On the other hand, the treatment of error found in the IPK and IPVr: innovates on

the one found in the SD by offering more extensive and detailed arguments against

the philosophy of the Buddhist epistemologists. Most of Utpaladeva’s novel argu-

ments are related to the central claim of the IPK and IPVr: , namely that the range of

worldly phenomena and the varieties of human experience would be impossible in

the absence of the existence of a single, unifying agent of cognition, Siva.40 Thus,

Utpaladeva argues at length about the functioning of memory, suggesting that there

must be an agent of cognition to unify the diverse experiences involved in an act of

memory, that of the original experience and that of the recollection thereof.41

Similarly, action (kriya), relation (sam: bandha), universals (samanya), substance

(dravya), space (dik), and time (kala) all are said to be complex phenomena, being

simultaneously both multiple and unitary;42 and it is Mahesvara himself who unifies

these complex phenomena.43 In short, contra the Buddhists, to whom the IPK

attributes the argument that nothing may be simultaneously a unitary and diverse

entity, Utpaladeva argues that none of these complex phenomena can exist without

being at the same time manifested in/as a unitary consciousness.44

39 See IPK 2.3.12 and the IPVr: on the same. Cf. IPK 1.8.6. Note, however, that Dharmakırti understands

arthakriya to be a generic feature of both valid forms of direct perception (pratyaks:a) and inference

(anumana), and not merely a feature of the svalaks:an: as that cause perceptual experiences, about which

see Hayes (1997, p. 106); Nagatomi (1967–1968); Katsura (1984); Mikogami (1979) (the latter three of

which were quoted by Hayes).Utpaladeva’s treatment of arthakriya is in fact complex, and substantially more developed than

Somananda’s. The SD merely denies an innate capacity for ‘‘causal efficacy’’ in worldly phenomena thatis not directed by the will of Siva. Utpaladeva reiterates this point, as we have here noted, but goes furtherin defining the existence of a variety of arthakriyas for any given phenomenon. On the one hand, anycomplex entity is comprised of a number of appearances (abhasas), each of which has its own arthakriya.For example, a blue pot is possessed of a color, physical extension in the form of its shape, a texture in theform of the clay with which it was fashioned, etc., each of which is projected into consciousness by aparticular abhasa that is associated with a particular arthakriya. On the other hand, a single abhasadefines the entity in question and unifies the diverse abhasas of which it is comprised. The arthakriya ofthis unifying abhasa trumps, as it were, those of the other, supporting ones, the pot’s capacity to carrywater being the predominant arthakriya of the present example. (See IPK 2.3.5-6, and below.) None ofthese arguments appears in the SD.40 As is well known, the IPK introduces an objection offered by the Buddhist epistemologists

immediately following the prolegomena, in IPK 1.2.1-2, where it is suggested that only two types of

cognition exist (this being the well-known division between pratyaks:a and anuman: a), and neither

requires the existence of a single, permanent self to operate. The remainder of the Jñanadhikara and the

Kriyadhikara are devoted to refuting this very argument in myriad ways.41 Utpaladeva argues that it is the act of recognizing the correspondence between the original cognition

and the memory of it that requires the existence of a single agent of cognition. This comparison is

necessary, because all cognitions are self-contained and cannot become the object of another cognition.

See IPK 1.2.3-6, 1.3, 1.4, 1.6.8-11, and 1.7.5. Cf. Torella (2007b, pp. 540–542).42 See IPK 1.2.9 (on kriya), IPK 1.2.10-11 and 1.7.2 (sam: bandha), IPK 1.7.4 (on causality), and IPK

2.2.1, where the unity and multiplicity of kriya, sam: bandha, samanyas, dravyas, dik, and kala are

declared. Their unity is not only analogous to but is based in the existence of a single agent of cognition

who pervades the apparently multiple universe.43 See IPK 1.3.6-7. Cf. IPK 1.8.10-11.44 The objection is raised in IPK 1.2.9. Utpaladeva’s reply may be found at the opening of the

Kriyadhikara, in IPK 2.1.1. Cf. Torella (1994, p. 153, fn. 1).


123

Utpaladeva’s treatment of the nature of error in the IPK and IPVr: also hinges on

the same argument; for he refers to the isolated nature of the individual moments of

cognition in order to state unequivocally what Somananda seems to suggest, namely

that no invalidated-invalidator relationship (badhyabadhakabhava) between two

self-contained and momentary cognitions is possible in the absence of the existence

of a single agent of knowing.45 And he extends this argument by offering a

sophisticated retort to the theory of abhava and anupalabdhi that is the basis for the

theory of invalidation according to the Buddhist epistemologists.46 The most sig-

nificant innovation in the treatment of error in the IPK and IPVr: , however, is found

with the fact that Utpaladeva offers an explicit definition of the means of reliable

knowledge (praman: a):

The means of knowledge is that thanks to which the object is situated within

its own confines: ‘this thing, with these characteristics.’ This means of

knowledge is an ever freshly arising light related to a subject. This light,

whose essence is the inner reflective awareness of that which is thus mani-

fested, becomes—as regards the object without spatio-temporal differentia-

tions, etc., and expressed by a single name—knowledge, [provided it is] not

invalidated.47

This definition is noteworthy not only for the various ways in which it echoes the

writings of Dignaga and Dharmakırti,48 and not only because no such definition is

found in the SD; it is also noteworthy because Utpaladeva here suggests that

45 See IPK 1.7.6 and the IPVr: on the same. See also IPVr: ad IPK 1.7.13, where Utpaladeva explicitly

declares that individual cognitions can only gain a mutual association through the existence of a unitary,

knowing subject.46 The argument in question, explained in some detail in Torella’s notes to the corresponding passages of

the IPK and IPVr: , is offered at a point where Utpaladeva considers the possibility that the perception of

mother of pearl can invalidate an earlier perception of silver, this in the same manner that the perception

of an empty surface can lead to the knowledge of the absence of a pot that formerly stood in the place in

question (IPK 1.7.7). At stake is the notion that a subsequent cognition can invalidate a previous cognition

all of itself, thereby vitiating the need to posit the existence of a single knowing agent to compare the two

cognitions. As Torella (1994, p. 141, fn. 13) has noted, Utpaladeva responds by suggesting (in IPK 1.7.8)

that the argument in question must fail, because the second cognition can only reveal what is present, not

what is absent. That is, when one has a cognition of the empty surface on which one previously saw a jar,

this subsequent cognition of the surface cognizes only the surface itself, and not the jar, which is absent.

This is to say that the cognition in question proves only that the surface that one sees is not the jar that

once stood on it: it proves, in other words, tadatmyabhava. It does not reveal that the jar is absent from the

surface, as this realization involves the recognition that the surface is not the pot (vyatirekabhava).Similarly, Utpaladeva argues (in IPK 1.7.12), the cognition of mother of pearl reveals that silver is not

present, but it does not invalidate the previous cognition of silver. This may only be proven by arecognition of the relationship of two cognitions—that of silver and that of mother of pearl—which areshown to be contradictory (virodha); and, after this, one must further ascertain the degree of congruenceof each cognition with reality, which again requires the presence of a single agent of cognition who canmake the necessary comparisons. (See IPK, and IPVr: ad, 1.7.13. Cf. Torella 1994, pp. 145-146, esp. fn.26.) Somananda makes none of these arguments in the SD.47 The passage here quoted is IPK 2.3.1-2: idam etadr: g ity evam: yadva�sad vyavatis: t:hate / vastupraman: am: tat so ’pi svabhaso ’bhinavodayah: . so ’ntas tathavimar�satma de�sakaladyabhedini / ekab-hidhanavis:aye mitir vastuny abadhita. The translation is Torella’s, for which see Torella (1994, p. 161).48 On the affinities of Utpaladeva’s definition of the means of knowledge with the writings of Dignaga

and Dharmakırti, see Torella (1994, pp. 161, fn. 2; 162–163, fn. 4).

J. Nemec

123

knowledge (miti = pramiti) only exists on the condition that it is not invalidated

(abadhita), a condition that Somananda essentially dismissed, as we have seen.

What does Utpaladeva mean to suggest by referring to the possibility of knowledge

being invalidated? How does this criterion square with Somananda’s insistence on

the impossibility of any cognition being invalidated?

The answers to these questions may be found by analyzing the criteria Utpala-

deva establishes to judge the validity of a given cognition. And these criteria

indicate that Utpala, apparently following Somananda, distinguished the real from

the empirical in the IPK and IPVr: , allowing for all cognitions to be real quaappearing within consciousness, while simultaneously recognizing that some among

them reveal false information about the everyday world. In fact, Utpaladeva

recognizes two necessary criteria by which to judge the validity of a given cogni-

tion. First, he recognizes that valid cognitions must perceive objects that are useful

or causally efficacious if the object in question is correctly perceived, even if,

following Somananda, he ultimately understands causal efficacy to be a function of

Siva’s will and not a quality that may be attibuted solely to the object in question.

Second, he adds to this a new criterion: real cognitions not only lead to practical

results (upayoga) but they are also stable (sthairya).49

Invoking the example of the perception of silver for mother of pearl, Utpaladeva

suggests that the cognition of silver is incongruent with the locus where it appears,

not because the cognition of silver is unreal, but because the cognition of silver

cannot perdure where mother of pearl is present. When the appearance of silver does

not accord with the place where it appears, the incongruence (asam: vada) with the

locus (de�sa) suggests that the cognition of silver was erroneous, while that of mother

of pearl was correct. Thus, that the ‘‘reflective awareness’’ (vimar�sa) of silver is

temporary results from the fact that the locus of the perception of silver is mis-

placed: it is incongruent with the nacre that appears before the cognizing agent.50

At first glance, then, Utpaladeva’s novel criterion for measuring the non-erro-

neousness or erroneousness of cognitions appears (1) to confirm Somananda’s view

in part, (2) to extend Somananda’s arguments in a new direction, and (3) apparently

to contradict Somananda’s suggestion that the locus (de�sa) of cognition may never

be contradicted. We shall take these up in order.

First, Utpaladeva’s reference to sthairya confirms Somananda’s dogma that all

cognitions are real, because it does not deny the reality of the erroneous cognition

(of, e.g., the silver) as long as it appears. The problem, Utpaladeva suggests, lies not

with the reality of the cognition, which is real vis-a-vis its very appearance in

consciousness; but the erroneousness of the cognition may be found in the fact that

it cannot perdure (and therefore cannot function in the everyday world as one would

expect a real thing to function). Utpaladeva thus has identified a criterion for reality

that allows ‘‘false’’ cognitions to be ontologically real but empirically erroneous.

49 See IPK 2.2.1 for reference to this pair of criteria as a measure of the reality of actions, relations, etc.

Cf. IPVr: ad IPK 2.3.1-2. Note also that Abhinavagupta relates the criterion in question, the stability or

sthairya of the non-erroneous cognition, to Dharmottara’s interpretation of what it means for a cognition

to be non-erroneous. This is so because Abhinavagupta glosses stairya with sam: vada, which echoes

Dharmottara’s Nyayabindut: ıka, about which see Torella (1994, p. 157, fn. 4).50 See IPK 2.3.13 and the IPVr: on the same.


123

And, in addition, he applies this criterion to both conceptual and perceptual forms of

error, as he is explicit in suggesting that both are explained in an identical manner.

Utpaladeva also extends Somananda’s argument by explaining the appearance of

the cognition of silver as involving an error in the way in which the various

appearances (abhasas) that make up the object in question are unified. That is,

Utpaladeva suggests that any appearance involves the confluence of a number of

abhasas that are unified in the process of cognition. For example, the cognition of

mother of pearl involves the manifestation of abhasas associated with its white

color, its brilliant quality, its round shape, its location, etc. Each of these is pos-

sessed of an individual capacity for causal efficacy (arthakriya�sakti), in Utpalade-

va’s view; and all of them are simultaneously unified by a single abhasa (itself

possessed of a single arthakriya�sakti), that of ‘‘mother of pearl’’ in the present

example, which presents the complex phenomena as a single entity. The mis-

perception of silver for mother of pearl, then, involves a mistaken identification of

only one abhasa among many, the combination of which causes the appearance of

the object in question. And, moreover, the majority of the abhasas that appear—

whiteness, roundness, etc.—are abhasas that are integral to the appearance of both

silver and mother of pearl, which share the qualities in question. The misperception

of silver for mother of pearl, then, involves the appearance in one place of a set of

abhasas that contribute to the appearance of both silver and mother of pearl, the

exception being the locus of the silver, which is here misplaced and is therefore

lacking in the capacity to perdure (that is, it is possessed of an asthairya). The

mistaken perception of this one abhasa, then, explains the misperception in ques-

tion. This sophisticated elaboration on and extension of Somananda’s arguments is

entirely absent from the SD.

Finally, Utpaladeva’s new criterion for measuring the non-erroneousness of

cognitions seems to contradict Somananda’s treatment of error by suggesting that

the locus (de�sa) of the cognition of silver is invalidated by the appearance of the

cognition of mother of pearl. For, if one were to understand Somananda to argue

that no cognition may be invalidated under any circumstances, then one would be

compelled to understand Somananda’s position to be incompatible with Utpalade-

va’s. On the other hand, if I am right instead to understand Somananda to suggest

that no cognition may be invalidated in the absence of the existence of a unifying

agent of cognition, then there is no contradiction in suggesting that an invalidation

occurs. In other words, whether or not one understands Utpaladeva’s reference to

the invalidation of the locus of the cognition of silver to contradict Somananda’s

treatment of error depends entirely on whether one understands Somananda to deny

the impossibility of any invalidation under any circumstances, or instead to argue, as

I suggest he did, for the impossibility of any theory of badha in the absence of the

existence of a single agent of cognition.

What cannot be doubted is the fact that the theory of error offered in the IPK and

IPVr: is one of ‘‘error by commission,’’ one that clarifies and further develops the

theory first offered in the SD. For, when one sees, e.g., silver in the locus of mother

of pearl, the cognition of silver involves the appearance of range of abhasas

associated with mother of pearl—whiteness, roundness, etc.—that were correctly

seen when the cognizer believes that it is silver that appears in the ken, the error

J. Nemec

123

being caused only by the appearance (abhasa) of the locus of silver where mother of

pearl exists. This is to say that the cognition in question necessarily involves the

perception of something that is subsequently recognized for what it truly is; but the

error in question is never caused by the failure to cognize something that is truly

present.

Utpaladeva’s Theory of ‘‘error by omission’’: The Noncognition of Non-Duality(abhedakhyati)

In a third stage of development of the Pratyabhijna’s treatment of error, Utpaladeva

adds to the theory surveyed above a second explanation of error, a theory of ‘‘error

by omission.’’ This is described in Utpala’s commentary on Somananda’s SD, the

SDVr: . Using a nomenclature that echoes that of Prabhakara but with apparently no

intention of referring explicitly to it,51 Utpaladeva there makes reference to a theory

of ‘‘error by omission’’ in no fewer than eighteen places. In each instance, he refers

to error as the result of the noncognition of unity, though he expresses this basic idea

in a variety of ways. Most common is reference to the noncognition of non-duality

(abhedakhyati), but he also refers to the noncognition of unity (ekakhyati), the

noncognition of the non-duality of consciousness (cidabhedakhyati), the non-

apprehension of non-duality (abhedaparamar�sa), and other, related formulations.

All of these passages refer to the failure of the agent of cognition to realize the

undivided nature of the universe, which according to the Pratyabhijna is nothing but

Siva himself in the form of a dynamic but unitary consciousness.

The majority of these references serve to explain the very appearance of the

universe in its apparent multiplicity. To begin, Utpaladeva suggests that sam: saraexists as an erroneous condition the form of which is the noncognition of Siva’s

unity.52 Elsewhere, he states that the existence of the multiplicity that is found in the

51 As Rastogi (1986, p. 4) has already noted, Abhinavagupta nowhere deals with Prabhakara’s bheda-khyati theory, despite the similarity in nomenclature of Prabhakara’s and Abhinava’s theories. Utpala-

deva, like Abhinava, is also silent when it comes to the treatment of the Prabhakara School and the

bhedakhyati theory. And, as we shall see, the abhedakhyati theory found in the SDVr: seems to have been

inspired by Somananda’s critique of Bhartr:hari’s conception of avidya (at least as Somananda understood

it) in the SD.52 See SDVr: ad SD 1.3-4: �sivaikyakhyatirupabhrantimayasam: saravastha yavan nonmis:ati, tavad apitavaty evoktarupa�sivata. ‘‘As long as the erroneous condition of sam: sara, the form of which is the

noncognition of Siva’s oneness, does not arise, Siva-nature is ‘so great,’ i.e., is as it is described earlier.’’


123

everyday world is nothing but the noncognition of non-duality.53 That Utpaladeva

understands the very existence of the universe to be the product of this ‘‘error by

omission’’ is most apparent, however, in the various places where he associates the

error in question with the power of maya that causes the unenlightened agent to

cognize an apparent diversity of phenomena. He suggests, for example, that in the

process of creating sam: sara the power of maya produces precisely the ‘‘error by

omission’’ here under consideration, this by making the entities of the universe

appear to be separated from the individual who cognizes them.54 Elsewhere,

Utpaladeva identifies the power of maya with Siva’s noncognition of the identity of

himself with the apparently multiple universe.55 Finally, he defines the limit of the

53 This is mentioned in the context of responding to an opponent’s objection. The opponent suggests that

Somananda cannot account for a diversity of types of causes, because such diversity would require Siva to

be divided into at least three parts, one for each of the three types of causes in question. (See SD 3.29cd:

nimittasamavayyadikaran: es:u samanata. ‘‘The causes—efficient, inherent, etc.—are identical.’’) Utpala-

deva explains Somananda’s reply by suggesting that there is no problem with the apparent existence of a

diversity of causes, because Siva’s tripartite nature is the nature of the noncognition of the non-duality of

the distinct entities that make up the everyday world. See SDVr: ad SD 3.80-82ab: ...athava tasyaikatve ’pitritvam: yathoktaprakaram: vyapade�sad vyavaharat sarvabhedavyavaharasyabhedakhyatirupatvam.

‘‘Alternatively, although he is unitary, that which in the manner already mentioned is by designation, i.e.,

is spoken of as, his tripartite nature is the nature of the noncognition of the non-duality of all of the

distinct entities that are the everyday world.’’ Note that vyapade�sa is a lemma (found in SD 3.82b) that is

here glossed with vyavahara.54 See SDVr: ad SD 1.1: yo ’ham: namaskaromi sa �sivo ’smadrupen: aikyam: praptah: vastusthitya hisarvatattvavigraho vaks:yaman: anıtya �Sivah: . sa sam: sarartham: maya�saktikr: taikyakhyatya bhavan ana-tmasthan abhasayatı�svarapratyabhijñaprapañcitanyayena. ‘‘I, who pay homage, am Siva, who has

attained unity with my form, for in reality Siva’s form, in the manner that will be explained, is that of all

the tattvas. He makes entities appear to be located outside of himself for the sake of the world of

transmigration, in the way explained in greater detail in my I�svarapratyabhijña, by not perceiving (his)

unity (with those entities) as a result of the power of maya.’’ Reference to the form of Siva as that of the

tattvas is found in SD 1.29cd-33. (Cf. IPK 3.1.2-11.) Note also that sarvatattvavigraha, a bahuvrıhicompound, refers in a technical sense to the thirty-six tattvas, though it also could be understood more

idiomatically to suggest the less technical notion that Siva is one who exists ‘‘in the form of all reality.’’

Reference here to the I�svarapratyabhijña possibly refers to IPK 3.1.2-11. See also IPK 4.14 for mention

of Siva as full of the (infinite) tattvas.55 See SDVr: ad SD 1.39-41ab: tatha maya�saktiva�sad atmakhyatita anandatirikta iva tattvatam etijagannirman: eccharupen: a sthatum. ‘‘And, by dint of the power of maya, i.e., by not perceiving himself,

he, as if separated from bliss, becomes the tattvas in order to be established in the form of a desire to

create the world.’’ Note that one might possibly read atmakhyatitas with what follows it, rather than with

what precedes it: ‘‘he, as if separated from bliss because he does not perceive himself [i.e., his true

nature]…’’Compare the present passage with one found in SDVr: ad SD 2.89-91, where Utpaladeva explains

Somananda’s attack on the Grammarians’ conception of the nature of the power of speech. There, hesuggests that speech has Siva-nature, as do all entities, but that it is not supreme, because it is made bySiva’s maya, which he describes as the noncognition of Siva’s non-duality: ...tathapi vagindriyada�sa parana bhavati, maya�saktya �sivabhedakhyatya tadutthanat. ‘‘…even so, the condition of the organ of speechis not supreme, because it arises by means of the power of maya, that is, by the noncognition of Siva’snon-duality.’’ Cf. SDVr: ad SD 1.41cd-43: …svarupabhedakhyatir eva hi maya�saktikr: ta grahya-grahakarupasam: saratmatayavatis: t:hate. ‘‘…for, it is the noncognition of the non-duality of his nature,which is made by the power of maya, that is established as the nature of sam: sara, which takes form as theobjects and agents of cognition.’’

J. Nemec

123

tattvas, the levels of reality created by maya,56 as the point where there exists a pure

or complete noncognition of the unity of the light of consciousness.57

Utpaladeva also describes the creation of the universe as an ‘‘error by omission’’

in various places where he associates abhedakhyati with a number of other technical

terms and expressions that are associated in Pratyabhijna theological formulations

with Siva’s powers of cosmic creation. In one place, he situates the same ‘‘error by

omission’’ at the very beginning of the process of creation, as he identifies it with

the power of ‘‘eagerness’’ (aunmukhya), a power, found primarily, in Somananda’s

formulation of the Pratyabhijna, that is manifested prior to the Trika triad of powers

(the �saktitraya), will (iccha), cognition (jñana), and action (kriya).58 In another,

Utpaladeva identifies abhedakhyati with the apara condition, the lowest of the three

conditions of existence recognized by the Pratyabhijna and the one identified with

the manifestation of the apparently multiple universe.59 Still elsewhere, Utpaladeva

says that it is the noncognition of the non-duality of the Supreme Lord that effects

even the subtle division of the one Siva into the pure tattvas, which are said to be

56 See SD (and SDVr: ad) 1.29cd-33, where maya is said to assume the form the tattvas.57 See SDVr: ad SD 1.29cd-33: parthive ’pi tu yavat tad asti citprakasaikyakhyatimatratvat(�matratvat ]

conj. em.; �matrat GJPRKed.) pr: thak praka�satmatve ’pi. ‘‘On the other hand, as much as this is the case

even at the (level of the) earthen(-tattva), it exists differently (there), even though its nature is praka�sa,

because it is the pure noncognition of the non-duality of the light of consciousness.’’ In other words,

Utpaladeva suggests that the earth tattva is different from the thirty-five tattvas that precede it: it marks

the last stage of manifestation, and as such it involves not merely the noncognition of non-duality, but

rather it constitutes the limit or the most extreme form of nonawareness, which in turn defines the limit of

that which is manifested.58 See SDVr: ad SD 1.13cd-17: yatha jalasya purvam: nistaran_gasyatitaran_gitam: gacchatah: suks:mah:purvah: kampa aunmukhyarupah: , pan: e�s ca mus: t: itam: gacchatah: purvah: susuks:mah: kampo dr: �syate,tatha bodhasya svarupasthasya purn: asya vi�svaracanam: praty abhilas:amatraracanayogyataya yah:prathamo vikasah: pravr: ttyarambhas tad aunmukhyam: pracaks:ate. pravr: ttyarambha�s ca nirvr: tav apyabhedakhyatidharmatvena tasyah: prathanat. ‘‘When water that is initially waveless becomes extremely

rough, its subtle, initial trembling appears as a form of eagerness, and so does the initial, very subtle

trembling of a hand becoming a fist. In the same way, one speaks of eagerness as the first budding, that is,

the first moment of manifestation, of the capacity to create, which is but the mere desire to create the

universe on the part of a self-contained, complete consciousness. Moreover, the first moment of mani-

festation exists even in (the power of) delight, because it unfolds as the nature of the noncognition of non-

duality.’’ On the sequence of powers as formulated by Somananda—nirvr: ti, aunmukhya, iccha, jñana,

and kriya, see SD 1.2 and 1.22. Cf. Nemec (forthcoming); finally, see also footnote 24, above.59 The reference in question comes in the context of contrasting the understanding of the intellect

(buddhi) found in the Nyaya and related schools with the Pratyabhijna understanding of Siva’s power

of cognition. See SDVr: ad SD 1.26-29ab: aparavasthayam abhedakhyatav ekarasayam:puryas: t:akapramatr:n: am: sa buddhih: . ‘‘That is the intellect of cognitive agents who have subtle bodies,

which exists in the one-faceted apara condition, which is the noncognition of non-duality.’’ The other two

conditions of reality are the paravastha and the paraparavastha, as is well known. On the scriptural

sources for the nomenclature of these three conditions, see Sanderson (1990, pp. 31–32).


123

manifested prior to the emergence of mahamaya.60 Simply, Utpaladeva suggests

that the very manifestation of the apparently multiple objects that make up the

universe results from a perception of multiplicity that results, in turn, from the

mayic absence of the cognition of non-duality.61

Not only does the noncognition of non-duality explain the appearance of the

universe, according to Utpaladeva, but it also accounts for the individual agent’s

experience of it. For Utpaladeva says that it is the noncognition of the non-duality of

consciousness (cidabhedakhyati) that permits one to distinguish between one-

self—the agent of cognition—and the object of cognition.62 Similarly, it is one’s

60 See SDVr: ad SD 2.1: mayottırn: e ’py adhvani vidyonmes:e ’pi karmamalanullasat sam: sarasam: parke ’piparame�svarabhedakhyateh: (�abhedakhyateh: ] G; �bhedakhyateh: JPRKed.) sarvathaivavicchedadagamavagate�svarecchava�sotthaparaparada�savyapade�syoktatattvopagamah: . ‘‘Even on the path which is

beyond maya, [Siva] enters into the tattvas that are spoken of as what is referred to as the paraparacondition, which arises by dint of Isvara’s will, which is known from scripture. This is so because the

noncognition of the Supreme Lord’s non-duality is in no way interrupted even when one is not in contact

with the world of transmigration, because karmamala does not increase when even Vidya appears.’’Note that Vidya here refers to the �suddhavidyatattva, the fifth of the thirty-six tattvas, it being the

lowest of the five tattvas of the so-called pure path (�suddhadhvan). Reference to karmamala here servesto explain the reason for the similarity of the arising of the pure tattvas to the arising of the impure ones:in both cases, the presence of one of the three ‘‘impurities’’ (malas) that cloud one’s awareness ofreality—karmamala, which limits one’s sense of what it is one can do—exists equally for those agents ofcognition who reside in the pure tattvas and in the impure ones. On the nature of karmamala see IPK 3.2,esp. 3.2.5 and 3.2.10, and the corresponding passages of the IPVr: .61 See SDVr: ad SD 1.7cd-8, where Utpaladeva quotes and subsequently glosses his own �Sivastotravalı(SSt), there describing the manifestation of the various effects (nanakaryasr: s: t: i) that make up the

universe as that ‘‘…which is perceived by making variegated none other than (Siva,) the one who has a

fixed nature, this by means of the mayic noncognition of non-duality’’ (…sthitisvarupasyaivamayıyabhedakhyativaicitryaracanopalaks: ita).62 See SDVr: ad SD 1.24–25: ghat: adijñanakale yad ghat:am: janati cidabhedakhyativaicitryabhinnag-hat:adevadattatmakavedyavedakabhasanam: nama yat kramikam apurvam: rupam: sa jñanalaks:an: a kriyapurvaparıbhutavayava. ‘‘In the moment one cognizes pots, etc., the cognition of the pot, i.e., the

appearance of the object of knowledge and the knower thereof in the form of a pot (on the one hand,) and

(an agent,) Devadatta, (on the other,) which are distinguished by the variegation that is caused by the

noncognition of the non-duality of consciousness, that [cognition of the pot], being a sequential, new

form, is an action called cognition, which becomes divided into earlier and later parts.’’On the meaning of apurva, cf., e.g., IPVr: ad IPK 2.3.1-2: yadayatta hi vastunah: svarupen: a nity-

atavi�ses:an: air va vyavasthapyamanata sa pramatuh: svatvenapurvavastvavabhasa eva sthitah: praman: am.‘‘The means of knowledge is that on which the setting of the object within its confines depends (vyav-asthapyamanata), both as regards its mere being and the various possible qualifications such as per-manence and so on. It consists precisely in the manifestation to a knowing subject of a new object(apurvavastu�) as related to him (svatvena).’’ The translation is Torella’s, for which see Torella (1994,p. 161).

J. Nemec

123

failure to see the ‘‘sameness of flavor’’ (samarasya) or equilibrium of Siva’s form in

everything that leads one erroneously to understand the world to be divided into

things pure and impure, etc.63 And the degree to which the individual fails to

cognize non-duality and therefore experiences duality is itself regulated by the

degree to which the subtle body is delighted.64 This is to say that the degree to

which an individual is bound or freed from ignorance is a function of the degree to

which he or she suffers from the ‘‘error of omission’’ in question. Thus, the

individual’s salvation is tied to the same form of error, as bondage and liberation are

in Utpaladeva’s view merely functions of, respectively, the noncognition and

cognition of Siva’s non-duality.65 Even the apparent existence of various, contra-

dictory philosophical systems, Utpaladeva suggests, is the result of the noncognition

of the unity of Siva, as this is what constitutes the appearance of the everyday world

of activity, for which such distinctions exist.66

Now, a series of questions arises regarding the nature of such an error. Is it real?

How does Utpaladeva reconcile the existence of error with his non-dualism? That is,

how can he account for the existence of a difference between the true perception of

non-duality and the erroneous perception of apparent multiplicity? Utpaladeva’s

answer to this line of questioning may be known, first of all, by his gloss of a

passage in which Somananda anticipates an opponent asking why Siva, who is pure,

would be inclined toward engaging with an impure universe. Somananda’s reply

is that it is simply Siva’s nature to create the universe, so no question of his

63 See SDVr: ad SD 1.48: evam: ca sarvasya �Sivarupasamarasyat tadakhyatimaya�suddhya�suddhirupa-paraparadibhedo bhavanam uktah: . ‘‘And in this way, because Siva’s form exists in a state of unity in

everything, the division of entities into the parapara (condition), etc., in the form of purity and impurity,

which consist of the noncognition of that [state of unity], is (merely) expressed.’’64 See SDVr: ad SD 3.97cd-99: abhedakhyati�s ca taratamyena bhidyate puryas: t:akabhogodrekat.‘‘Moreover, the noncognition of non-duality is disclosed in varying degrees, depending on the pre-

ponderance of enjoyment in the subtle body.’’ On the nature of the ‘‘subtle body’’ (puryas: t:aka), see IPK

3.2.14, and the IPVr: on the same.65 See SDVr: ad SD 3.70: yada tv eka�sivatvam eva tattvam: tada tasmin paks:e na dard: hyam: bhedasya, apitu �sivabhedapratıtimatram: moks:as tadapratıtis tu bandha iti tavata pratıtimatren: a, na tu vastvanyathatvebandhamoks:ayor bandhamoks:ata. ‘‘When, on the other hand, only the nature of the unitary Siva is reality,

then, i.e., this being our view, no stability exists for the distinction, but rather liberation is the pure

cognition of Siva’s non-duality, while bondage is the noncognition of the same. Thus, the nature of

bondage and liberation exists to the extent that one perceives them, not because bondage and liberation

are mutually distinct things.’’66 See SDVr: ad SD 3.76cd-78ab: lokavyavaharaya va sarvam: �sastravadiprativadyadi. lokavyavahara�sca tadabhedakhyatimayo na vastugatah: . ‘‘However, everything, i.e., the teaching, the proponent of it, the

one propounding an opposing view, etc., serves the behavior of society, and the behavior of society,

consisting of the noncognition of his non-duality, is not a real thing.’’


123

motivations may be raised.67 In elaborating on Somananda’s explanation, Utpala-

deva suggests that nothing may be brought to light that is not made up of Siva’s

consciousness, and thus nothing in the universe may be said to be impure. He further

argues that error exists only in the form of the non-awareness of non-duality

(abhedaparamar�sana), and such an absence of awareness does not exist (tac ca nakiñcit, he says), because it exists in the form of the absence of a cognition, and

because nothing new is manifested by it.68 Presumably, then, Utpaladeva under-

stands the akhyati in question to be a prasajyapratis:edha, a pure negation.

Moreover, Utpaladeva explains the nature of the noncognition of non-duality in

the course of elaborating on Somananda’s arguments against Bhartr:hari, where there

arises a question regarding the ontological nature of that which is merely an

absence. In a passage that analyzes the nature of the grammarians’ avidya, Utpa-

ladeva questions the ontological status of the ‘‘nescience’’ that obscures Brahman’s

true nature, and using the same language that he employs to describe the nature of

the noncognition of non-duality, he suggests that the grammarians conceive of

Brahman as being impeded by that which has no form, i.e., by that which is

akiñcidrupa.69 As Utpaladeva states, something that is unreal, that which has no

form, can nowhere be of use.70 Here, then, he argues that the grammarians cannot

explain the obfuscation of Brahman’s true nature by way of reference to avidya,

because in understanding the meaning of the term to involve the pure negation of

67 See SD 1.11cd-13ab: kutsite ’kutsitasya syat katham unmukhateti cet. rupaprasararasato garhitatvamayuktimat / pañcaprakarakr: tyokti�sivatvan nijakarman: e. pravr: ttasya nimittanam apares: am: kva mar-gan: am. ‘‘If you ask how someone pure could be eager for something impure, (we reply:) because he

savors the expansion of (his own) form, contemptibility is not appropriate (for the universe). Given that it

is said that it is Siva’s nature to perform the five types of activities, what need is there to search for other

motives for one who is engaged in his own conduct?’’ I thank an anonymous reviewer for suggesting that

the present passage suggests a reference to the �Slokavarttika, Sam: bandhaks:epaparihara 53–55:

atha�subhad vina sr: s: t: ih: sthitir va nopapadyate / atmadhınabhyupaye hi bhavet kim: nama dus:karam. tathacapeks:aman: asya svatantryam: pratihanyate / jagac casr: jatas tasya kim: names: t:am: na sidhyati. prayoja-nam anuddi�sya na mando ’pi pravartate / evam eva pravr: tti�s cec caitanyenasya kim: bhavet.68 See SDVr: ad SD 1.11cd-13ab: abhedaparamar�sanam eva bhrantirupam: kutsitam: , tac ca na kiñcidakhyatirupamatratvat. na tv apurvasya kasyacit pratha. ‘‘Only erroneous nonawareness of (Siva’s) non-

duality is impure, and that does not exist, because it consists merely of a nonperception, and nothing new

is manifested.’’ (On the meaning of apurva, see footnote 62, above.)69 See SDVr: ad SD 2.34-35: vidyanam: samyagjñananam abhavah: prasajyapratis:edho ’pi yady avidya,tat sa �sunyata nivr: ttimatram: na kiñcid iti yavat. tata�s ca �sunyaya nırupaya pa�syantı nama samyagjñ-anatmakah: paks:o badhyate iti citram. samyagdar�sanabadhenaiva dvaitabhrantih: , akiñcidrupasya bad-hana�saktyabhavan na sam: bhavaty etad ity arthah: . ‘‘Even if you argue that nescience is a pure negation, is

the nonexistence of knowledge, i.e., of correct congitions, then it is void, i.e., it is pure cessation. To put it

plainly, it is nothing at all. And therefore, your hypothesis, called pa�syantı, the nature of which is correct

cognition, is invalidated by a void, i.e., by that which is without form. This is amazing. The error of

dualism is the result of nothing but the invalidation of correct seeing. What this means is that this is

impossible, because something that has no form does not have the power to invalidate something else.’’70 This Utpaladeva says in the context of questioning how pa�syantı, the power of speech, can be divided

by individual bodies. That is, if pa�syantı is one, and if it devolves to the level of mundane speech, it is

asked, then why don’t all individuals utter precisely the same words at all times? Considering a possible

reply—that the individual bodies of the individual agents divide pa�syantı, Utpaladeva suggests that this is

impossible, because they are unreal, have no form (i.e., they are akiñcidrupa). See SDVr: ad SD 2.43-

44ab: …na casatyam akiñcidrupam: kvacid upayogi bhavatıti yavat. ‘‘…and, to put it plainly, something

unreal, i.e., something that has no form, cannot be of service anywhere.’’

J. Nemec

123

knowledge—in understanding avidya to be a prasajyapratis:edha—they posit that a

truly non-existent entity, avidya, somehow impedes the real Brahman. This, he

suggests, is a fallacious position to hold.71

The preceding, then, makes plain the fact that Utpaladeva twice contemplates the

ontological nature of that which has no real form. Explaining Somananda’s

understanding of Bhartr:hari’s conception of avidya—an understanding that prob-

ably does not accurately reflect Bhartr:hari’s position, it must be added72—, Utpa-

ladeva glosses his teacher’s criticism by suggesting that the grammarians

understand the real Brahman to be impeded by a non-existent entity, by the absence

of knowledge. In contrast to this position, Utpaladeva argues positively that the very

phenomenon that leads one to perceive duality is itself an absence, the absence ofthe cognition of non-duality. Being a mere absence, this noncognition is for

Utpaladeva ultimately nothing at all, meaning that all that truly exists is the unitary

Siva. This is to say, then, that Utpaladeva’s abhedakhyati theory is of necessity a

theory of ‘‘error by omission,’’ because the noncognition has no ontological nature

whatever, and the error is caused by the pure absence of a cognition of the ever-

present non-duality of Siva.

Abhinavagupta’s Treatment of Error: The apurn:akhyati Synthesis

The preceding lays plain the fact that Utpaladeva developed a theory of ‘‘error by

omission’’ to explain both the appearance of the apparently multiple universe and

the individual agent’s experience of it; and, prior to this, he developed a theory of

‘‘error by commission’’ that was informed by Somananda’s and served to explain

the occurrence of conceptual and perceptual forms of misperception. Here, we shall

argue that Abhinavagupta knew Utpaladeva’s abhedakhyati theory, was thoroughly

versed in the treatment of error found in both the SD and the IPK and IPVr: , and, in a

fourth phase in the development of the Pratyabhijna treatment of error, sought to

synthesize these theories by developing an inclusive nomenclature for both types.

That is, he developed the apurn: akhyati theory of error.

71 Utpaladeva, following Somananda, also suggests that to say that one understands avidya to be a

paryudasapratis:edha would result in a dualism that distinguishes Brahman in the form of knowledge or

vidya from that which is different therefrom. Such a dualism would of course contradict Bhartr:hari’s non-

dualism. See SD 2.34-35 and Utpaladeva’s commentary thereon.72 Torella (1994, p. xxvi, fn. 37) has aleady pointed out that Somananda is overly critical of Bhartr:hari’s

understanding of avidya and vivarta, and indeed the Vakyapadıya easily affords an interpretation of the

former as a �sakti of Brahman, or in other words offers an understanding of the relationship of the absolute

to the manifested universe that is quite sympathetic with Somananda’s. Utpaladeva, on the other hand,

reverses Somananda’s strong criticsm of Bhartr:hari and instead adopts many of his ideas, as Torella has

shown (about which see Torella 1994, pp. xxiii–xxvii, Torella 2008.) On the other hand, it is Utpaladeva

and not Somananda who first refers to avidya as being akiñcitrupa, that which has no form, and so in parts

of his SDVr: , at least, he is willing to engage actively in the criticism of Bhartr:hari’s ontology.


123

An examination of a key passage of Abhinavagupta’s IPVV, one that comments

on IPK 2.3.13, shows that Abhinavagupta there refers explicitly to Utpaladeva’s

‘‘error by omission,’’ as he identifies three types of error, suggesting that the first

two involve the noncognition of non-duality (advayakhyati = abhedakhyati).73

Labeling these two as forms of ‘‘great error’’ (mahabhranti), Abhinavagupta

describes them, respectively, as the type of error that involves a noncognition

(akhyati) affecting every agent of cognition in sam: sara (samastasam: saripramat-r: sadharan: ı mahabhranti); and as a type of error that is located separately in each

individual agent of cognition (pratipramatr: niyata mahabhranti). These he associ-

ates with the power of maya. The former is conceived of as the form of error that the

agent conceives of as sam: sara, what is thought to be the means of attaining hap-

piness; the latter exists in the form of the agent’s thoughts or ideas (pra-matr: vikalparupa mahabhranti).74 This formulation, then, echoes Utpaladeva’s in

that it accepts that it is an ‘‘error by omission,’’ the noncognition of non-duality

(advayakhyati), that explains both the appearance of the apparently multiple uni-

verse and the individual agent’s experience of such diversity where only Siva’s

ubiquitous non-dual consciousness exists.

It is the third type of error that Abhinavagupta understands to involve an ‘‘error

by commission.’’ Unlike the two forms of mahabhranti, the third form of error

affects one agent of cognition to the exclusion of others: it occurs, suggests

Abhinava, when one fails to notice that what one erroneously sees is not visible to

73 Rastogi was the first to note that Abhinavagupta conceived of three types of error, but he makes no

mention of Utpaladeva’s abhedakhyati theory, and he does not indicate that the first two forms of error

recognized by Abhinavagupta and labeled as forms of mahabhranti invoke this theory, while the third

type of error, labeled bhranti, constitutes a theory of ‘‘error by commission’’ that invokes the treatment of

error found in the SD and the IPK and IPVr: . See Rastogi (1986, p. 6).74 See IPVV ad IPK 2.3.13 (vol. 3, p. 155, lines 3–10 of the KSTS edition), also quoted by Rastogi (1986,

p. 27, fn. 21): na kevalam atra samastasam: saripramatr: sadharan: yam: mahabhrantav akhyatih: para-marthah: , yavat pratipramatr: niyatayam api pramatr: vikalparupayam: mahabhrantau saiva tattvam itidar�sayati maya iti. mayapramatur vyaparo yo vikalpo mamedam: sukhasadhanam ityadih: , sa rupam: yasyabhranteh: sam: sarasam: matayah: , sapi sarvaiva samastes:u pramatr: s:u pratyekam: bhedena vartamanaadvayakhyatirupeti purven: a sam: bandhah: . ‘‘By referring to maya, [Utpaladeva] shows that it is not only

here, i.e., in the great error held in common by every agent of knowing in sam: sara, that the supreme

reality [paramartha] is a non-cognition, but that very same [noncognition] is the nature that exists in the

great error that is confined to each agent of knowing in the form of the agent’s vikalpas, as well. The

activity that is conceptualization, e.g., ‘this is a means for me to obtain pleasure,’ belonging to the mayic

agent of knowing, is the form of the error that is thought of as sam: sara; that [error], being of many kinds,

also is the one abiding separately in every agent of knowing on an individual basis, the form of which is

the noncognition of non-duality. Thus, it is related with the former.’’Abhinavagupta also refers to that which is common to all agents of cognition (sarvapramatr: sadharan: a)

in the corresponding passage of the IPV, for which see vol. 2, from p. 112, line 11 to p. 113, line 7, of theKSTS edition.

J. Nemec

123

others, as other agents of cognition instead see that which is truly present.75 This

third type Abhinavagupta labels merely as ‘‘error’’ (bhranti), and, as Rastogi noted,

Bhaskara Kan: t:ha suggests that it occurs when, e.g., one sees silver for mother of

pearl.76

That this third type of error involves an ‘‘error by commission’’ may be known

from an examination of the very manner by which Abhinavagupta describes it,

which is again found in a passage of commentary on IPK 2.3.13, this time in his

IPV. In IPK 2.3.13, Utpaladeva offers his two examples of misperception, viz.,

conceptual error exemplified by the cognition of silver where only mother of pearl

exists, and perceptual error caused by a physical fault in the cognitive apparatus,

such as occurs where one whose eye is damaged sees two moons in the sky.

Abhinavagupta describes the nature of such forms of error as follows:

apurn: akhyatirupakhyatir eva bhrantitattvam. It is a noncognition, in the form

of an incomplete cognition, that is the nature of error.77

Bhaskara Kan: t:ha indicates that Abhinava intends the present passage to invoke the

notion of ‘‘error by commission,’’ for he says in his Bhaskarı that the form of the

akhyati in question is one of a partial cognition, not a noncognition.78 And Abhi-

navagupta himself makes the same clear in the passage of the IPV preceding the

excerpt here quoted. There, he explains the nature of the erroneous cognition, the

cognition of silver where only mother of pearl existed. Subsequent to the perception

of silver, he explains, one has the experience ‘‘this was not silver,’’ and he labels

this experience a subsequent moment of reflective awareness (uttara paramar�sa).

He goes on to say that it is not the case, however, that the experience in question is

similar to the one that one has upon witnessing a strike of lightning, where one

witnesses a flash of light appearing momentarily in the dark sky, only to be followed

by darkness. In the case of lightning, what is experienced is only a momentary flash

of light; and following its appearance, the object in question disappears. In the

75 See IPVV ad IPK 2.3.13 (vol. 3, p. 153, lines 20–22 of the KSTS edition, quoted on Rastogi 1986,

p. 28, fn. 43): etaditiprathamanasya pramatrantaravis:ayatvasya viruddho na prathata iti purn: aprathab-havad apurn: akhyatirupa iyam akhyatir eva bhrantih: . ‘‘It is this non-cognition that is the error, the form of

which is an incomplete cognition, because the complete appearance is absent; for, no contradiction of the

nature of that which is an object (of cognition) for (all) other agents of knowing, appearing objectively

[literally: as ‘this’], comes to light.’’ In other words, the error in question must be an ‘‘error by com-

mission,’’ because objects that appear to all other agents of knowing are never contradicted. Note that this

passage was also quoted in part by Rastogi (1986, p. 26, fn. 14).76 See Bhaskara Kan: t:ha’s Bhaskarı (vol. 2, p. 123, lines 14–15 of the commentary in the Iyer and Pandey

edition, quoted on Rastogi 1986, p. 8): apurn: am: jñanam: tv atra tavad asti �sukteh: rajatataya grahan: ad itibhavah: . ‘‘The incomplete cognition, on the other hand, exists here as it does as a result of the appre-

hension of mother of pearl as silver. This is what is meant.’’77 See IPV ad IPK 2.3.13 (p. 113, lines 11–12 of the KSTS edition). The same was quoted by Rastogi

(1986, p. 26, fn. 10).78 This he does by explaining that the alpha-privative of the nañtatpurus:a compound, akhyati, should be

understood to signal a ‘‘slight’’ cognition and not the absence thereof. See Bhaskarı ad IPV on 2.3.13 (vol.

2, p. 123, lines 9–10 of the Bhaskarı of the Iyer and Pandey edition): apurn: akhyatirupapurn: ajñanarupana tu khyatyabhavarupa, akhyatih: ajñanam: ...ıs:adarthe nañ na tv abhave. ‘‘Apurn: akhyatirupa means in

the form of an incomplete cognition, not in the form of the absence of cognition. Akhyati means a non-

cognition [ajñana]…The alpha-privative (in akhyati) is used in the sense of (a) small (cognition), not

(the) absence (of cognition).’’


123

case of the perception of silver for mother of pearl, by contrast, the entity that is

erroneously perceived as silver—the mother of pearl that truly is present—remains

in the ken after the demise of the erroneous perception.79 Indeed, he says that the

error in question is constituted by the cognition of a new object (apurva) by the light

of consciousness (praka�sa) when the light is possessed of a form that is at variance

with what it illumines—it is described as being possessed of a various or manifold

form (vividharupa). As such, it is not an error that fails to illumine the form of the

object in question (it being described as not being rupapraka�sana), but rather one

that sees it only in part.80

Abhinavagupta further says that this third type of error appears within the uni-

verse that is manifested by, or perhaps as, mahabhranti; and as such it involves the

existence of an error within an error, something similar, he says, to a dream in a

dream or a boil on a pimple.81 What he envisions then, is the following: the universe

is itself a product of an error, the ‘‘error by omission’’ (abhedakhyati) first con-

ceived by Utpaladeva and first articulated in the SDVr: . This error is responsible for

the appearance of the universe, of sam: sara, and for the individual agent’s experi-

ence of it. It appears by dint of the failure of the agent to cognize all of reality in its

true nature, as the dynamic but undivided consciousness of Siva. In addition to these

two underlying forms of ‘‘great error’’ Abhinavagupta identifies misperception as a

peculiar type of error distinguishable from the first two insofar as it only occa-

sionally affects individual cognitive agents within the erroneously-appearing world

79 See IPV ad IPK 2.3.13 (vol. 2, p. 113, lines 7–11 of the KSTS edition, parts of which were quoted by

Rastogi 1986, p. 28, fn. 36 and p. 31, fn. 73): tatraiva kale nedam: rajatam: hy abhud iti hy uttarah: para-mar�sah: , na tv uditapratyastamitayam: �satahradayam ivedanım evedam: neti vimar�sah: , tato yavata purn: enarupen: a prakhyatavyam: vimar�saparyantam: tavat na prakhyatıty apurn: akhyatirupa akhyatir eva bhranti-tattvam. ‘‘For, there is a subsequent reflective awareness in that very moment of time [i.e., the moment

when one realizes that one was wrong about what it is that one saw]—‘this, in fact, was not silver.’ But,

unlike lightning, which appears and disappears (all of a sudden), the reflective awareness [vimar�sa] is not

of that which no longer exists in the present moment. [In other words, that which one recognizes as not

silver remains in the ken: it continues to appear as mother of pearl, and the silver was not a mere flash of

an appearance akin to a strike of lightning, but a misapprehension of what truly appears.] Hence, what

should appear in a complete form all the way up to the moment of reflective awareness does not appear as

such. Thus, it is a non-cognition in the form of an incomplete cognition that is the nature of error.’’80 See IPVV ad IPK 2.3.13 (vol. 3, p. 155, lines 9–10 of the KSTS edition), also quoted by Rastogi (1986,

p. 27, fn. 21): na tv apurvasyarthasya ya khyatir vividharupah: praka�sas tad rupapraka�sanam: bhrantih: .

‘‘On the other hand, it is not the case that the cognition of a new object, being praka�sa of a variegated

form, is an error that involves the non-illumination of the form (in question).’’ See footnote 62 regarding

Utpaladeva’s use of the term apurva. But note that it is possible that one should emend with apurn: asyafor apurvasya in the last line of the text here quoted, in which case one might translate as follows: ‘‘On

the other hand, it is not the case that the cognition of the incomplete object [i.e., the object erroneously

perceived, silver for mother of pearl, for example], being praka�sa of a variegated form, is an error that

involves the non-illumination of the form (in question).’’81 See IPV ad IPK 2.3.13 (p. 124, line 1 to p. 125, line 2 of the Iyer and Pandey edition): nanusatyarupyajñanam apy apurn: akhyatih: . tatas tarhi kim. idam atah: sarvam: bhrantir ity agacchet. dis: t:yadr: s: t: ir unmılis:yaty ayus:matah: , mayapadam: hi sarvam: bhrantih: , tatrapi ca svapne svapna iva gan: d: esphot:a ivapareyam: bhrantir ucyate, anuvr: ttyucitasyapi vimar�sasyasthairyat. ‘‘Objection: the cognition of

real silver is also an incomplete cognition. What about that? It follows from this that one learns that all

[cognitions] are error. (Reply:) Thank heaven! The long-lived one’s eyes are opening, for the entire level

of maya is error, and it is also there that a lower [apara] (form of) error is said to exist, one that is like a

dream in a dream, (or) a boil on a pimple, since the reflective awareness, even though it is meant to flow

steadily, is unstable [asthairya].’’

J. Nemec

123

of duality. Moreover, to misperceive an object, Abhinavagupta intimates, requires

one first to see something, something the very appearance of which is itself the

result of the mahabhranti form of error, ‘‘error by omission.’’ An error within an

error, misperception involves the partial cognition of that which itself appears only

as the result of a noncognition of unity.

There is no contradiction, then, between the two theories of error. Indeed, as

Rastogi has indicated, Abhinavagupta refers to all the types of error by the same

nomenclature, as he in one place, at least, refers to the two forms of mahabhranti as

‘‘incomplete cognitions’’ (apurn: akhyati).82 The two theories of error simply explain

the appearance of different types of phenomena. It is ‘‘error by omission’’ that

creates the environment within which the individual agent’s experience may occur.

On the other hand, to see something within the apparently dualistic environment as

other than it truly is requires one not only to see something, to first cognize an

apparently distinct and discrete object of cognition, but it also requires one in

addition to see that object as something other than what truly appears. As Abhi-

navagupta explains it, such an instance of misperception must involve an ‘‘error by

commission’’ within an apparently dualistic environment that in fact appears as such

only in the absence of the cognition of unity.

Affinities of Abhinava’s apurn:akhyati with the IPK and IPVr:

Although Abhinavagupta’s apurn: akhyati theory constitutes a new development in

the way the authors of the Pratyabhijna describe error, Utpaladeva’s writings serve

as the basis for his treatment of error, because the scope and nomenclature of

Abhinavagupta’s formulation have precedent in the IPK and IPVr: . Already noted

was the significant degree to which the two types of mahabhranti identified by

Abhinavagupta correspond with the types of error Utpaladeva describes with his

theory of ‘‘error by omission’’: Utpaladeva’s abhedakhyati theory explains both the

manifestation of the forms of experience common to all agents of cognition and the

nature of the individual agent’s cognition of the apparently multiple universe. That

is, it explains the occurrence of ‘‘cosmic or metaphysical error’’ and ‘‘empirical

error,’’ as Rastogi labeled them.83 That Abhinavagupta preserves Utpaladeva’s

theory untouched in the domains to which his paramaguru applied it, adding only

the label mahabhranti to them, signals an important point of continuity between the

two authors’ views. For, just as the scope of Abhinavagupta’s theory of ‘‘error by

commission’’ is limited only to instances of misperception—to what Abhinavagupta

labels simply as bhranti (Rastogi labels this ‘‘epistemic error’’84)—so too does

Utpaladeva limit his theory of ‘‘error by commission’’ to misperception, just as he

nowhere applies the theory of ‘‘error by omission’’ to such forms of cognitive error.

82 See IPV ad IPK 2.3.13, where he suggests that the cognition of real silver is as much an incomplete

cognition as is the misperception of silver for mother of pearl. Both are ‘‘incomplete’’ (apurn: a) cogni-

tions, presumably, insofar as both involve the failure to cognize the unity of Siva’s consciousness; but

only misperception is an error within an error, one involving also an ‘‘error by commission.’’83 See Rastogi (1986, p. 6).84 ibid.


123

Utpaladeva also implies what Abhinavagupta states explicitly, namely that per-

ceptual error constitutes the existence of an error within an error. This is so because

in the IPK he describes the waking state as a manifestation that is common to all

agents of cognition (sadharan: ı sr: s: t: i), one that is stable insofar as it involves an

external manifestation that is the object of the various sense faculties;85 and he

suggests in the commentary that this is the state in which misperception may

occur.86 This is to say that Utpaladeva understands misperception to occur within a

universe that is manifested by the very ‘‘error of omission’’ already discussed, and

the error, in his view, is experienced by an agent of cognition whose limited agency

is constructed by the same noncognition. Misperception, by implication, therefore

must for Utpaladeva involve the presence of an error within an error.

Regarding Abhinava’s reference to ‘‘incomplete’’ (apurn: a) cognitions, this

nomenclature echoes Utpaladeva’s and reflects his understanding of what, precisely,

is experienced in the apparently multiple universe; for, Utpaladeva frequently

invokes the language of completeness (purn: atva) and the absence thereof in the IPK

and IPVr: , even if he nowhere articulates a theory of error as involving a perception

in the form of an incomplete cognition (apurn: akhyati). First, Utpaladeva describes

the appearance of diversity to the limited subject (mitatman) by way of comparing it

with the awareness of the unlimited (amitatman) subject in IPK 2.1.7. There, he

suggests that the limited subject experiences spatial difference, while the unlimited

subject—Siva—experiences entities as full (purn: a) of his own self.87 Utpaladeva

subsequently describes the supreme reality (paramartha) as consciousness (bodha)

that is free (svatantra), complete (purn: a) because it is possessed of such a nature;

and he suggests that the impurity associated with limited agency (an: avamala)

results from the absence of the same completeness (purn: atva).88 He goes on to state

85 See IPK 3.2.17.86 See IPVr: ad IPK 3.2.17: sarvantarbahis:karan: a�saktya sr: s: t: ir jagara, tatrapi purvavad dvicandr-adibhrantih: . ‘‘The creation which occurs thanks to the powers of all the internal and external causes is the

waking state. Here, too, as in the dream state, illusion may occur, in the form of seeing two moons, etc.’’

(The translation is Torella’s, for which see Torella 1994, p. 206.) Utpaladeva here refers to the existence

of error as described in the preceding verse (IPK 3.2.16), where he labels the dream state as illusion, this

because what one perceives in a dream does not perdure, just as erroneous perceptions do not perdure. See

IPVr: ad IPK 3.2.16: anyapramatr: sadharan: arupadyabhasanuvr: ttitah: kalantarananuvr: tter bhrantir es: a.

‘‘And this is an illusion inasmuch as this perception, e.g., of a form, common to other subjects, after

having lasted as such [for the length of the dream], does not continue later.’’ (The translation is Torella’s,

for which see Torella 1994, p. 206.)87 See IPK 2.1.7: de�sakramo ’pi bhaves:u bhati matur mitatmanah: / svatmeva svatmana purn: a bhavabhanty amitasya tu. ‘‘Also spatial succesion in things is manifested (only) to the limited knowing subject;

on the contrary, to the unlimited subject things appear full of his own self, like his own self.’’ The

translation is Torella’s, for which see Torella (1994, p. 155).88 See IPVr: ad IPK 3.2.4: svatantro bodhah: paramarthah: , tatharupatvad eva purn: ah: . tasya svatantryadeva tathecchaya svatantryam: vina bodhamatranirman: e bodhatam: vina svatantryamatranirman: e vapurn: atvabhavena parimitatvad dvidhan: utvam: tattvikasvarupaviparyasan malatvam. ‘‘The absolute reality

is free consciousness; only insofar as this is its essence is it in perfect fullness. Since it thus wills, in virtue

of its very freedom [there is the an: ava maculation which is twofold]: ‘maculation’ insofar as it involves

error as regards the self’s real nature and ‘twofold atomicity’ insofar as consciousness becomes limited

with the loss of its fullness, creating a subjectivity in which only the light of consciousness subsists

without freedom, or only freedom without consciousness.’’ The translation is Torella’s, for which see

Torella (1994, p. 198).

J. Nemec

123

that the complete (purn: a) consciousness of the divine subject, of Siva, constitutes

the highest level one may attain.89 Completeness (purn: atva), then, is for Utpaladeva

a function of non-duality, and the existence of that which is incomplete involves the

appearance of duality.

Finally, both Utpaladeva and Abhinavagupta distinguish misperception from the

forms of error that lead one to see duality. Utpaladeva does so implicitly by nowhere

applying his theory of ‘‘error by omission’’ to instances of misperception. Abhi-

navagupta does so explicitly by labeling those forms of error associated with

Utpaladeva’s abhedakhyati theory as mahabhranti, as we have seen. Both authors,

moreover, use the language of incompleteness (apurn: atva) to describe the

appearance and experience of the apparently multiple universe. Indeed, Utpaladeva

in his SDVr: even identifies the non-perception of Siva’s complete nature

(purn: asvarupakhyatimaya�) as the nature of the appearance of the multiple uni-

verse, a formulation the nomenclature of which is close to Abhinavagupta’s

apurn: akhyati formulation, although it is of course a theory of ‘‘error by omission,’’

not commission, that Utpaladeva here expressed.90 Utpaladeva is even explicit in

declaring the ubiquitous existence of Siva’s completeness (purn: atva),91 which

89 See IPVr: ad IPK 4.14: sarvatha tv antarlıne prameye ’ham: matau purn: ayam: �sivataiva. ‘‘But when the

cognizable is entirely dissolved within him and there is the full consciousness of the I, the state of Siva is

attained.’’ The translation is Torella’s, for which see Torella (1994, p. 217). Cf. IPK 4.14. (As Torella

[1994, p. 217, fn. 29] notes, it is at this level that Utpaladeva suggests that no cognizable reality exists at

all.) Utpaladeva similarly locates the final form of the breath, vyana, in the state beyond the fourth

(turyatıta), which he describes as both undivided by space and time (dikkalanavaccheda) and complete

(purn: a). See IPVr: ad IPK 3.2.20: turyatıte dikkalanavacchede purn: e pravahatam: hitvatinirbharavasthovyanasam: jñah: . ‘‘The breath ceases to flow in the state beyond the fourth—this being the state of fullness

and as such devoid of spatio-temportal delimitations—and entering a state of absolute fullness takes the

name of vyana.’’ (The translation is Torella’s, for which see Torella 1994, p. 209.)90 See SDVr: ad SD 1.11cd-13ab: maya�saktikr: tapurn: asvarupakhyatimayacitrakaryatapannsvarupapra-saran: arasat prabhor asya tadrupasya karyabhedasya kutsitatvam ayuktam. ‘‘It is not appropriate that the

object in question, which has his form, is impure, because the Lord savors the development of his own

nature, which becomes the various objects, these being made of the noncognition of his complete nature,

(the noncognition being) produced by the power of maya.’’ Note that the fact that Utpaladeva here signals

the existence of an ‘‘error by omission’’ does not constitute a contradiction with Abhinavagupta’s

apurn: akhyati theory: the present serves to explain the appearance of the objects of the universe, which

Abhinava explains by reference to advayakhyati, this being one of the forms of mahabhranti on his view,

as already noted.91 See IPVr: ad IPK 2.1.8: evam: purn: ataya praka�samanasyapi parame�svarasya sais: a sr: s: t: i�saktir yajjñatr: jñeyasvabhava bhavah: svato ’nyonyam: ca vibhagenavasıyante, na ca tathavasayena tasya svar-upasthitis tirodhıyate. ‘‘The fact that the various entities existing as subject and object of knowledge are

determined as differentiated from each other and from Paramesvara (or from the self), is [due to] the

creative power of Paramesvara, who, yet, thus shines in absolute fullness, neither does his own permanent

form ever in any way cease to exist as a consequence of this determination.’’ The translation is Torella’s,

for which see Torella (1994, p. 156).


123

implies that any experience one may have of it is of necessity incomplete, which is

to say divided, multiple, and must therefore involve some error in the perception of

the unity that is in fact always fully present.

Simply, the scope and nomenclature of Abhinavagupta’s apurn: akhyati theory

have significant precedent in Utpaladeva’s IPK and IPVr: , even if Abhinavagupta

takes an additional step in synthesizing Utpaladeva’s various thoughts on the nature

of error by deploying the language of incompleteness in a novel manner.

Conclusion

The preceding makes clear that Abhinavagupta’s treatment of error owes a great deal

to his paramaguru, Utpaladeva, and secondarily to his parames: t:higuru, Somananda.

And although the manner in which Abhinvavagupta formulates his apurn: akhyatitheory has the advantage of allowing him to claim that his system accounts for all the

possible theories of error of any rival philosophical school, it is more immediately a

synthetic expression of two distinguishable theories of error found in the writings of

his predecessors. By referring to an incomplete cognition (apurn: akhyati), Abhina-

vagupta captures the central point of the theory of error first conceived by Somananda

and further developed by Utpaladeva in the IPK and IPVr: . In this view, misperception,

whether it is conceptual or perceptual in nature, involves the cognition of one object as

another. To label such cognitions as incomplete (apurn: a) not only echoes the language

used in other contexts by Utpaladeva in the IPK and IPVr: , but it also insinuates what

Utpala explicitly states therein, that a misperception involves the correct perception of

a number of abhasas involved in the appearance of a given object, coupled simulta-

neously with the wrong perception of one (or perhaps more?) abhasa involved in the

same appearance. To call such a perception ‘‘incomplete’’ is fitting, given that only a

select element is erroneously perceived in the erroneous perception, which is other-

wise appropriately seen.

Abhinava’s nomenclature also captures the idea of a non-perception or non-

cognition of non-duality by insinuating that apurn: a is a synonym of abheda. For, he

suggests that the akhyati in question is in the form of the cognition of that which is

incomplete (apurn: akhyatirupa), as we have seen: apurn: akhyatirupakhyatir evabhrantitattvam. There can be no doubt, then, that Abhinava wished to echo Utpa-

ladeva’s abhedakhyati theory, as he uses a synonym (advayakhyati) to describe

mahabhranti and suggests that it occurs precisely where Utpaladeva says the

abhedakhyati theory applies: in explaining the appearance of the apparently mul-

tiple universe and the individual agent’s experience of the same. By understanding

the bhranti form of error to involve the occurrence of an ‘‘error within an error,’’

Abhinava also makes explicit the layering of error that is only partially expressed in

the IPK and IPVr: , giving the system a simple elegance that appeals to common

sense. But it does not add substantively to the arguments put forward in the IPK and

IPVr: . In short, Abhinavagupta’s contribution is best understood as the work of a

brilliant commentator, who managed subtlely to reframe the ideas of his pre-

decessors in a succinct and elegant manner without altering the substance of their

teachings.

J. Nemec

123

Utpaladeva, too, showed himself to be an adept commentator and disciple, as he

was able to refine his teacher’s treatment of the problem without contradicting it.

Yet, he also proved himself to be the most innovative of the three authors in

developing a complete treatment of error in its various forms. For, while some of

what Utpaladeva offers us serves merely to clarify what Somananda probably meant

to say about ‘‘error by commission,’’ most notably the arguments regarding the

necessity of the existence of a single agent of cognition, Utpaladeva also added

substantially to his teacher’s arguments against their primary opponent, Dharma-

kırti. With the formulation of a new criterion for judging the erroneousness and non-

erroneousness of cognitions, Utpaladeva also answered a pair of philosophical

questions not addressed by Somananda, while simultaneously confirming the

treatment of error found in the SD. For, reference to the capacity of a cognition to

perdure preserved the fundamental tenet of Somananda’s system—that all cogni-

tions, insofar as they are part and parcel with Siva’s very nature as a dynamic

consciousness, are simply real qua the very fact that they appear—while also

explaining both the possibility for cognitions to be erroneous and the manner in

which they are corrected.

Utpaladeva also developed anew the abhedakhyati theory. And while it is true

that he apparently was inspired by his teacher’s critique of Bhartr:hari’s avidya, his

brilliant formulation of ‘‘error by omission’’ was a new development in the history

of Pratyabhijna philosophy, one that compellingly explained the possibility of the

appearance of an apparently multiple universe despite the existence, according to

Pratyabhijna theology, of only a single, unitary agent of cognition, Siva.

Clearly, then, there was ample room for innovation in the development of the

Pratyabhijna treatment of error, innovation that is so thorougly integrated into the

contours of the guru-student tradition that a casual perusal of the texts could hardly

reveal where the commentators introduce their own novel ideas. But these

innovations were developed for the most part in the writings of Utpaladva, while

Abhinavagupta’s treatment of error, by contrast, shows itself to be rather more

thoroughly indebted to the writings of his predecessors. Thus, while it is intriguing

that Abhinava meant his apurn: akhyati theory to account for all of the various

theories of error proposed by all the rival philosophical schools of his day, what is

most significant about his treatment of error is the degree to which it elegantly

synthesizes and faithfully adheres to Utpaladeva’s thinking on the matter. When it

comes to the Pratyabhijna treatment of error, then, Abhinavagupta may have spoken

the most elegantly, but it is Utpaladeva, inspired by Somananda, who most rigor-

ously defined the philosophical content of the Pratyabhijna treatments of error.

Acknowledgment

I thank Carmen Lamas, Karen Lang, and Andrew Nicholson for reading an earlier

draft of this essay.


123

References

Manuscript Sources

�Sivadr: s: t: i of Somananda. (T) Trivandrum University, number 5854-H.�Sivadr: s: t: i of Somananda. (C) Calcutta Sanskrit College, CS 3, 153.�Sivadr: s: t: i of Somananda, along with the �Sivadr: s: t: ivr: tti of Utpaladeva. (G) Gottingen, Niedersachsische

Staats- und Universitatsbibliothek, Cod.Ms.Sanscr.Vish 5 (11).�Sivadr: s: t: i of Somananda, along with the �Sivadr: s: t: ivr: tti of Utpaladeva. (J) Raghunath Mandir Library,

Jammu.�Sivadr: s: t: i of Somananda, along with the �Sivadr: s: t: ivr: tti of Utpaladeva. (P) Bhandarkar Oriental Research

Institute, Pune, number: 805 of 1891-95.�Sivadr: s: t: i of Somananda, along with the �Sivadr: s: t: ivr: tti of Utpaladeva. (R) Rajasthan Oriental Research

Institute, Jodhpur.

Primary Sources

I�svarapratyabhijñakarika (IPK) of Utpaladeva. Ed. and Trans. by Raffaele Torella. 1994. TheI�svarapratyabhijñakarika of Utpaladeva with the Author’s Vr: tti: Critical Edition and AnnotatedTranslation. Roma: IsMEO.

I�svarapratyabhijñakarikavr: tti (IPVr: ) of Utpaladeva. See I�svarapratyabhijñakarika.I�svarapratyabhijñavimar�sinı (IPV) of Abhinavagupta. Ed. by Madhusudan Kaul Shastrı. 1918–1921. The

I�svarapratyabhijña of Utpaladeva with the Vimar�sinı by Abhinavagupta. KSTS 22 and 23. Pune:

Aryabhushan Press.I�svarapratyabhijñavivr: tivimar�sinı (IPVV) of Abhinavagupta. Ed. by Pt. Madhusudan Kaul Shastrı. 1938,

1941, and 1943. The I�svarapratyabhijña Vivr: tivimar�sinı by Abhinavagupta. 3 Vols. KSTS 60, 62 and

65. Pune: Aryabhushan Press.Nyayabindu of Dharmakırti. 1954. Nyayabinduh: �Srıdharmottaracaryakr: tat: ikasametah: . Ed. by Srı Can-

drasekhara Sastri. Srı Kası Sam: skr: tagranthamala 22. Varan: ası: Chowkhamba Sanskrit Book-Depot.�Sivadr: s: t: i (SD ¼ Ked.) of Somananda. 1934. The �Sivadr: s: t: i of Somanandanatha with the Vr: tti by Utpa-

ladeva. Ed. by Madhusudan Kaul Shastri. KSTS 54. Pune: Aryabhushan Press.�Sivadr: s: t: ivr: tti (SDVr: ) of Utpaladeva. See �Sivadr: s: t: i.

Secondary Sources

Bandyopadhyay, N. (1988). The concept of contradiction in Indian philosophy. Journal of Indian Phi-losophy, 16(3), 225–246.

Dreyfus, G. (1997). Recognizing reality: Dharmakırti’s philosophy and its Tibetan interpretations.

Albany, NY: SUNY Press.Dunne, J. D. (2004). Foundations of Dharmakırti’s philosophy. Boston: Wisdom Publications.Dunne, J. D. (2006). Realizing the unreal: Dharmakırti’s theory of yogic perception. Journal of Indian

Philosophy, 34(5), 497–519.Franco, E. (1986). Once again on Dharmakırti’s deviation from Dignaga on Pratyaks: abhasa. Journal of

Indian Philosophy, 14, 79–97.Franco, E. (1993). Did Dignaga accept four types of perception? Journal of Indian Philosophy, 21(3),

295–299.Gnoli, R. (1957). �Sivadr: s: t: i by Somananda: Translation and commentary, chapter I. East and West, 8,

16–22.Hattori, M. (1968). Dignaga: On perception, being the Pratyaks:apariccheda of Dignaga’s Pra-

man: asamuccaya. Harvard Oriental Series 47. Cambridge, MA: Harvard University Press.Hayes, R. (1997). Whose experience validates what for Dharmakırti? In P. Bilimoria & J. N. Mohanty

(Eds.), Relativism, suffering, and beyond: Essays in memory of Bimal K. Matilal (pp. 105–118).

New York: Oxford University Press.Katsura, S. (1984). Dharmakırti’s theory of truth. Journal of Indian Philosophy, 12, 215–235.Kellner, B. (1997). Types of incompatibility (’Gal ba) and types of noncognition (Ma/Mi Dmigs) in early

Tibetan Tshad Ma literature. In H. Krasser, M. T. Much, E. Steinkellner, & H. Tauscher (Eds.),

J. Nemec

123

Tibetan studies: Proceedings of the 7th seminar of the international association of Tibetan studies,Graz, 1995 (Vol. 1, pp. 495–510). Wien: Osterreichische Akademie der Wissenschaften.

Kyuma, T. (1997). Bheda and Virodha. In S. Katsura (Ed.), Dharmakırti’s thought and its impact onIndian and Tibetan philosophy: Proceedings of the third international Dharmakırti conference,

Hiroshima, Nov 4–6, 1997 (pp. 225–232). Wien: Verlag der Osterreichischen Akademie der

Wissenschaften.Mikogami, E. (1979). Some remarks on the concept of Arthakriya. Journal of Indian Philosophy, 7,

79–94.Nagatomi, M. (1967–1968). Arthakriya. The Adyar Library Bulletin, 31–32, 52–72.Nemec, J. (forthcoming). The Ubiquitous �Siva: Somananda’s �Sivadr: s: t: i and his tanric interlocutors.

New York: Oxford University Press.Pandey, K. C. (1963/2000). Abhinavagupta: An historical and philosophical study (2nd ed.). Varanasi:

Chowkhamba Sanskrit Series Office.Rastogi, N. (1986). Theory of error according to Abhinavagupta. Journal of Indian Philosophy, 14(1),

1–34.Ratie, I. (2007). Otherness in the Pratyabhijna philosophy. Journal of Indian Philosophy, 35(4), 313–370.Sanderson, A. (1990). The visualization of the deities of the Trika. In A. Padoux (Ed.), L’Image divine:

Culte et meditation dans l’hinduisme (pp. 31–88). Paris: Editions du CNRS.Stcherbatsky, T. (1932/1993). Buddhist logic, 2 Vols, 1st Indian edition. New Delhi: Motilal Banarsidass.Torella, R. (1992). The Pratyabhijna and the logical-epistemological school of Buddhism. In

T. Goudriaan (Ed.), Ritual and speculation in early tantrism: Studies in honor of André Padoux(pp. 327–346). Albany: SUNY Press.

Torella, R. (2007a). Studies in Utpaladeva’s I�svarapratyabhijñavivr: ti. Part I. Apoha and anupalabdhi in a

Shaiva Garb. In K. Preisendanz (Ed.), Expanding and merging horizons: Contributions to SouthAsian and cross-cultural studies in commemoration of Wilhelm Halbfass (pp. 473–490). Vienna:

Austrian Academy of Sciences Press.Torella, R. (2007b). Studies in Utpaladeva’s I�svarapratyabhijñavivr: ti. Part II. What is memory. In

B. Kellner, et al. (Eds.), Indica et Tibetica: Festschrift für Michael Hahn Zum 65. Geburtstag vonFreunden und Schülern überreicht (pp. 539–563). Wien: Arbeitskreis fur Tibetische und Bud-

dhistische Studien Universitat Wien.Torella, R. (2007c). Studies in Utpaladeva’s I�svarapratyabhijñavivr: ti. Part III. Can a cognition become

the object of another cognition? In D. Goodall & A. Padoux (Eds.), Mélanges tantriques a lamémoire d’Hélène Brunner (pp. 475–484). Pondicherry: Institut Francaise de Pondichery / Ecole

Francaise d’Extreme-Orient.Torella, R. (2007d). Studies in Utpaladeva’s I�svarapratyabhijñavivr: ti. Part IV. Light of the subject-light

of the object. In B. Kellner, et al. (Eds.), Pramanakirtih. papers dedicated to Ernst Steinkellner on theoccasion of his 70th birthday (Vol. 2, pp. 925–939). Wien: Arbeitskreis fur Tibetische und Bud-

dhistische Studien Universitat Wien.Torella, R. (2008). From an adversary to the Main Ally: The place of Bhartr:hari in the Kashmirian Shaiva

Advaita. In M. Kaul & A. Aklujkar (Eds.), Linguistic traditions of Kashmir: Essays in memory ofPandit Dinanath Yaksh (pp. 508–524). New Delhi: D. K. Printworld.

Torella, R. (forthcoming). Studies in Utpaladeva’s I�svarapratyabhijña-vivr: ti. Part V: Selfawareness and

yogic perception.Woo, J. (2003). Dharmakırti and his commentators on Yogipratyaks:a. Journal of Indian Philosophy,

31(5), 439–448.


123

Documents

Nemec Two Pratyabhijna Theories of Error