Chapter 6 Alchemical Metaphors for Spiritual Transformation in Abhinavagupta’s Īśvarapratyabhijñāvimarśinī and Īśvarapratyabhijñāvivṛtivimarśinī

In: Śaivism and the Tantric Traditions
Christopher D. Wallis
Search for other papers by Christopher D. Wallis in
Current site
Google Scholar
Open Access

1 Introduction

It is a great privilege to be able to present a paper in honour of Alexis G.J.S. Sanderson. I was fortunate enough to be his pupil for two years at Oxford, and to study with him subsequently in Leipzig and Kyoto. In my view, if one were to accord Professor Sanderson the praise that he in fact merits, it would sound (to those who do not know him) like embarrassingly unrestrained hyperbole. Suffice it to say here that his example formed my ideal of intellectual integrity, an ideal which entails relentless pursuit of the truth as part of a community of scholars engaged in the kind of longitudinal study that prioritizes the field as a whole over personal glory. Professor Sanderson taught me the value of admitting when I don’t know, of sacrificing my own agenda in deference to the truth, and of striving to be as transparent a mediator as possible in the act of transmitting the words and ideas of the ancient Sanskrit thinkers to students of the present day. It is with enormous gratitude to his unstinting scholarly labours (I estimate he has logged well over a hundred thousand hours of research so far) that I offer this paper in his honour.

The oeuvre of the Kashmirian Tantric master Abhinavagupta (fl. c. 975–1015) is one of the many areas of research Professor Sanderson has mastered, and it is this author which the present paper treats. Specifically, we here focus on a trope found in Abhinavagupta’s two commentaries on the Īśvara-pratyabhijñā-kārikā (ĪPK) of Utpaladeva, viz., that of an alchemical metaphor for spiritual transformation. These passages provide no small number of difficulties, because the text as we have it is not secure, and because some knowledge of Indian alchemy (rasāyana, dhātu-śāstra) is needed in order to translate it correctly. While I do not claim to have solved these problems, this paper may certainly contribute to our understanding of how Abhinavagupta thought about the process of spiritual transformation conferred by the uniquely potent insight (jñāna) and yoga offered by initiatory Śaivism. Specifically, we learn much about his usage of the key terms samāveśa, turya, and turyātīta, and it is hoped that this paper advances our understanding of these topics, which are significant within Śaiva theology.*

2 The Pratyabhijñā Doctrine of the Fivefold Self

Some readers are no doubt aware that in Utpaladeva’s ĪPK we find a teaching on the “layers” of the individuated self (see, e.g., III.1.8), parallel to the later Vedāntic teaching of five kośas based on Taittirīya Upaniṣad 2.3–5.1 This teaching, formed as Sanderson says on “slight scriptural precedent,” is adopted by subsequent gurus of Utpala’s lineage; for example, it has a prominent place in Kṣemarāja’s Pratyabhijñāhṛdaya.2 In this model, the self is said to be fourfold: void (śūnya), life-force (prāṇa), the subtle body consisting of the mind and its faculties (puryaṣṭaka, i.e. the antaḥkaraṇa plus tanmātras), and the physical body (śarīra). It is fivefold with the transindividual Power of Awareness (cit, saṃvit) that permeates the whole. In fact, it is not only cit that permeates the other levels: Kṣemarāja tells us that “it is clear that the very essence of each of these levels is the fact of its pervasion by all the loci of perception prior to it,”3 where “loci of perception” (pramātṛ) refers to these levels of embodiment as those realities with which contracted souls identify, and “prior to” means “more fundamental than.”

Abhinavagupta adds to this teaching a homology implied but not spelled out in the ĪPK itself, one that assimilates these five levels to the five “phases of lucidity,” as Vasudeva (2004) calls them: the states of waking, dreaming, deep sleep, the transcendental “fourth” state, and the state “beyond the fourth” (jāgrat, svapna, suṣupta, turya, and turyātīta). We will come to understand the last two terms as we proceed.

Our texts in this study are Abhinavagupta’s two commentaries on the ĪPK, his -vimarśinī (hereafter ĪPV) and his -vivṛti-vimarśinī (ĪPVV). The former is his commentary on the kārikās themselves, the latter is his commentary on Utpaladeva’s lost Vivṛti or longer auto-commentary. For both texts, we will use the Kashmir Series of Texts and Studies (KSTS) edition. First, though, we will consider the two verses of the ĪPK that Abhinavagupta is commenting on, using Torella’s critical edition (2002), and summarize Abhinava’s initial remarks thereon. All translations are my own unless otherwise noted.

kalodbalitam etac ca cit-tattvaṃ kartṛtā-mayam |
acid-rūpasya śūnyāder mitaṃ guṇatayā sthitam ||
ĪPK III.2.11

And this Awareness-principle, consisting of [unlimited] Agency, [becomes] limited—[though] it is strengthened by partial agency (kalā)—abiding as a [mere] attribute in a person whose [habitual] nature is unconscious, [identifying as he does with] the void, [prāṇa, mind,] and [body].

mukhyatvaṃ kartṛtāyās tu bodhasya ca cid-ātmanaḥ |
śūnyādau tad-guṇe jñānaṃ tat-samāveśa-lakṣaṇam ||

By contrast, the characteristic of “immersion into That” is realization of the primacy of the Self-that-is-awareness as the [only] true Knower and Agent, and [a concomitant] insight regarding [the other layers of individuality,] the void, [prāṇa, mind] and [body], as mere attributes of it.4

Explaining the first of these verses in his ĪPV, Abhinava first describes how consciousness—which in its real nature is primordial, a priori, unlimited and free—comes to be in the degraded state we consider as normal. Through the power of his māyā expressed as the three malas, Śiva contracts himself into a limited form (aṇu, the individual soul), then equips himself with the five kañcukas beginning with kalā (cf. ĪPK III.1.9), resulting in a being that identifies itself with what is actually objective, that is, the body, mind, prāṇa, and void (cf. ĪPK III.1.8).5 Identification with the void (śūnya) can be identification with the state of deep dreamless sleep (as Abhinava states it here) but also, and more importantly, the void is the considered the primary locus of the limited “I” (see ĪPK III.2.13), which, being in reality empty (śūnya), vainly seeks to reify itself through identification with the body, mind, and prāṇa. This identification persists in all three states of ordinary consciousness (waking, dreaming, and deep sleep).6

Note that the real “I” is not here the core of an individual being as in Sāṅkhya, but the one transindividual Self of all beings. The individual soul (aṇu) only exists as a particular phase of that transindividual Consciousness, specifically, an expression of the contracted state of bondage. Thus, one may argue, the nondual Śaiva’s “I” is closer to the view of the Vijñānavāda Buddhists than it is to the ātman of Vedānta. (Even the dualistic Śaivas, who did posit a separate and eternal soul, distanced their view of the ātman from those of the brāhmanical schools (Watson 2006).)

3 The Exegesis of the Īśvarapratyabhijñāvimarśinī

Now let us look more closely at our first text as Abhinavagupta charts the trajectory from bondage towards liberation, commenting on ĪPK III.2.12 (KSTS vol. 33, p. 230–231):

yadā tūkta-gurūpadeśādi-diśā tenaivāhaṃ-bhāvena svātantryātmanā vyāpakatva-nityatvādi-dharma-parāmarśam ātmani vidadhatā tataḥ śūnyādeḥ prameyād unmajjya iva āsyate tadā turyāvasthā7 |

But when, through realizing [that the divine] qualities such as all-pervasiveness and eternality apply to oneself, by having the experience of the [real] “I” whose nature is [unqualified] freedom—[an experience] pointed out by the guru’s instruction and other methods that I have explained—[and] having therefore emerged as it were from [identification with] the objective knowables of the Void etc., and [as a result] abiding [in one’s real nature], then that is the [transcendent] state [called] the Fourth.

yadāpi parāmṛṣṭa-tathābhūta-vaibhava-nityatva-aiśvaryādi-dharma-saṃbhedena8 eva ahaṃ-bhāvena śūnyādi-deha-dhātv-antaṃ siddharasa-yogena vidhyate, tadāsyāṃ turyātīta9-daśāyāṃ tad api prameyatām ujjhatīva |

When further [the layers of the objective “self”] from the Void to the [very] tissues of the body are transmuted10 by means of the “alchemical elixir,” i.e. by the [fundamental] “I”-sense which is certainly conjoined with the qualities of magnificent power (vaibhava), eternality, sovereignty, [and others] of such nature that are cognized [as aspects of that “I”], then in this state [called] Beyond the Fourth they abandon (as it were) their objectivity.

Having introduced the three states of waking, dreaming, and deep sleep, Abhinavagupta now discusses turya and turyātīta, which complete the set of the “phases of lucidity.” Now, in this passage, I take Abhinava to be reconciling two modes of realization: one gnostic, rapid, transcendent, and liberating, and the other yogic, gradual, immanent, and siddhi-conferring. Here, the Fourth state is the gnostic realization that one has wrongly taken objective realities to be the self; it is waking up out of the trance of believing “I am the body,” etc. Such a realization can be sudden because it requires no transformation, only a recognition of what is already the case, including a reflective awareness (parāmarśa) of the qualities (dharmas) of one’s real self. As Torella puts it, “the adept, after becoming aware of the supreme nature of the I, becomes as though withdrawn from the knowable which formed his fictitious identity” (2002, xxxiv). Turya is then an exclusive kind of realization. By contrast, the process of turyātīta (“Beyond the Fourth,” but not actually a fifth state)—here described in terms of penetrating the layers of that constructed identity with this deeper awareness or transcendent I-sense—is inclusive and gradual, requiring yogic practice. In the turyātīta experience, the objective layers of the limited self are seen as expressions of the transindividual divine consciousness, and thus are recovered as part of a greater “I” than the one they were excluded from in the previous turya state. This process by which the cidātman penetrates the layers of body, etc., is likened to alchemical transformation, whereby the elixir called siddha-rasa transforms a base metal into gold (or extracts the gold from the base metal).11 The use of the word iva (last word of the passage just cited) denotes that the body etc. do not actually cease to be knowables when they come to be seen as nothing but crystallizations of the dynamic “liquid” essence of consciousness in the turyātīta state,12 just as the previous iva denoted that emerging from identification with knowables does not mean completely leaving them behind (which would entail physical death).

Now, the coherence of this passage only emerges after the emendation to the edition suggested by Torella and adopted here, that of exchanging the words turya and turyātīta. This may seem a dramatic emendation, but it would make little sense for Abhinava to list the three states of ordinary consciousness, then proceed to turyātīta when what is obviously called for is turya, “the Fourth.” Further, it would make no sense to argue that turya is an extension of the turyātīta state without completely ignoring the meaning of those two words; but the other way around exactly matches the meaning of the words.

But what would occasion such a confusion in the edition? It may well be that later scribes (for we do find the edition’s reading in the manuscripts), influenced by the more transcendentalist mainstream Indian philosophies, simply could not imagine that turya could denote the transcendent state while turyātīta, which is obviously intended as the higher attainment, embraced immanence. But this is precisely in line with Abhinavagupta’s Kaula view, for with the text emended as Torella suggests, we have here a model that is central to the Kaula Kālīkula, which Sanderson characterizes as “transcendence followed by an expansion that causes the state of enlightenment to pervade the transcended” (Sanderson 2007, 402–403). The Śivasūtra (well known to Abhinava) inherits this model, teaching the “establishing of this realization first through withdrawal into the heart of consciousness and then through its expansion into the states that constitute the mundane awareness of the bound” (ibid.), which precisely characterizes our ĪPV passage. For example, in the Śivasūtra (1.7) we find the teaching that the Fourth state can spread to the ordinary states of jāgrat, svapna, and suṣupta, imbuing them with awakened consciousness, which the Śivasūtra calls turyābhoga but which is simply turyātīta under another name.13 Of course, an examination of all the extant ĪPV manuscripts, preferably after forming a critical stemma, would be necessary to make a final ruling on the reading of the passage.

To return to our text, Abhinava concludes his ĪPV discussion of ĪPK III.2.12 by informing us that turya and turyātīta are forms of samāveśa, which, when it becomes continuous and stable (āsyate), is itself liberation.

seyaṃ dvayy api jīvanmuktāvasthā samāveśa ity uktā śāstre, samyag-āveśanam eva hi tatra tatra pradhānam, tat-siddhaye tūpadeśāntarāṇi |

This twofold state of one who is liberated while living is called samāveśa in the scriptures. For complete entering14 is itself primary in each of these; other teachings are [only] for its attainment.

This is a surprising statement, perhaps, for I know of no scriptural passage in which these two states are called samāveśa. What Abhinava wants us to understand, I think, is that when the scriptures use the term samāveśa, they are always referring to one of these two states. In turya, then, one fully and directly penetrates into one’s true nature, while in turyātīta, one causes that nature to fully and gradually penetrate the objective levels of one’s limited selfhood; for this reason they can both be appropriately referred to with the word samāveśa (from ā√viś, to penetrate). Abhinava continues:

dehapāte tu parameśvara evaikarasaḥ, iti kaḥ kutra kathaṃ samāviśet

But at the fall of the body, there is only one essence: the Supreme Lord. Thus, who could enter (/immerse), where and how?

In other words, it is only meaningful to speak of samāveśa in the context of embodiment, for only in that context are there apparently differentiated layers of selfhood such that there can be an “entry” of the locus of identity (ahaṃbhāva) from the body, etc., into cit, or an “entry” of cit into one of the layers of limited selfhood (dehādi)—the former entry being turya and the latter turyātīta.

4 The Parallel Passage in the ĪPVV

The corresponding ĪPVV passage (KSTS vol. 65, 327–331) is similar but sheds more light on some important points while simultaneously greatly complicating the issue. Abhinava elaborates further on the alchemical metaphor briefly introduced in the ĪPV; here, though, if we do not emend the published text, he appears to have changed his view from that seen in the ĪPV. There the alchemical metaphor was reserved for the turyātīta state, while here we see two stages of the alchemical metaphor, corresponding to both turya and turyātīta.15 Furthermore, it seems that he now posits two different modalities for attaining both states, one gnostic and one yogic. (Here I differ from Torella’s 1994 hypothesis that a single turya state bifurcates into two kinds of turyātīta.) Abhinava writes:

etad ajñāna-rūpa-mala-pratidvandvitayā samāveśa-lakṣaṇaṃ satya-svarūpe samyag āsamantāt praveśa-lakṣaṇaṃ jñānaṃ, yal-lābhena jñānī, yad-abhyāsena ca deha-prāṇādāv ananta-saṃvid-dharmātmaka-vibhava-samāsādanāt yogī bhavati |

[Utpala teaches that] the “distinguishing mark of samāveśa” is “insight,” since it is opposed to the Impurity that is ignorance, being characterized by a perfect (samyag), that is to say complete (ā samantāt), entry into one’s true nature,16 obtaining which one becomes a gnostic (jñānī), and practicing which, on the levels of body, prāṇa, etc., one becomes a yogī, due to attaining the glory (vibhava) that is an intrinsic quality of infinite Consciousness.

etad uktaṃ bhavati—yadā ahaṃbhāvaḥ svātantrya-diśaiva vyāpitva-nityatvādi-parāmarśa-balāt śūnyādeḥ prameyīkṛtād unmajjya iva āste, tadā turyatā;17 tadāpi ca śūnyādi-saṃskāro ’pi asti,—iti vyatireka18-turyātīta-samatā eva |

This is said [already in the ĪPV]: when the [true] I-sense, due to the power of the realization of its all-pervasiveness, eternality, etc., through the [scriptural] indication of its [innate] autonomy, emerges as it were from the objectified [levels of limited selfhood]—Void etc.—and abides [in its real nature], then that is the state [called] the Fourth. Nevertheless [in that state] the impressions of the Void, etc., still remain. Thus this has exactly the same [nature] as [that which is called] the “separated turyātīta.”

Though I am not aware of another usage of the technical term vyatireka-turyātīta, the meaning here is clear enough (after applying Torella’s suggested emendation): the gnostic who does no yoga enters into a transcendental turya state in which he is authentically immersed in his essence, but the impressions of limited selfhood from which he has successfully separated (vyatireka) himself from remain undissolved (thus his social self might exhibit little to no change). Thus, Abhinava argues, the attainment of turyātīta of the vyatireka variety is in fact no different from the turya state itself. This obviously sets up the possibility of a higher attainment, an avyatireka-turyātīta in which one dissolves those impressions through practice, allowing the practitioner to be not-separated (avyatireka) from his body, mind, etc., yet still liberated; i.e. an immanentist state of liberation.

It is hard to see what Abhinavagupta has gained here, because in the simpler ĪPV model, turya was the transcendental state and turyātīta the immanent (and therefore higher) attainment. Perhaps he simply wants to indicate here that either state can be attained by either gnostic or yogic means. But there is more evidence to examine before drawing conclusions.

Now we see the yogic version of the turya → turyātīta progression. In the following paragraph (continuing directly from the previous ĪPVV citation), note that the first part closely parallels the ĪPV passage we have seen above (pp. 147–149), while the second part is new data.

yadā tu parāmṛṣṭa-nityatva-vyāpitvādi-dharmakaiśvarya-ghanātmanā ahambhāva-siddharasena śūnyādi-deha-dhātv-antaṃ19 vidhyate yena prameyatvāt tat cyavata iva, tadā turya-daśā;

But when [all the layers of limited selfhood] from the Void to the tissues of the body are penetrated by the “alchemical elixir” that is the [true] I-sense—replete with the sovereignty in which the qualities of eternality, all-pervasiveness, etc., are cognized [as aspects of that “I”]—through which [penetration] they abandon (as it were) their objectivity, then that [too] is [called] the Fourth State.

yadāpi viddho ’sau prāṇadehādi-dhātuḥ saṃvid-rasena abhiniviṣṭo ’tyantaṃ kanaka-dhātur iva jīrṇaḥ kriyate yena sa druta-rasa iva ābhāti kevalaṃ tat-saṃskāraḥ, tadāpi turyātīta-daśā sā bhavati |

When, further, these elements of prāṇa, body, etc., [already] penetrated by the elixir of Awareness, are thoroughly permeated [by it], they are [then] “digested” like the element of gold [is by mercury], by which [process] their purifier, the “liquefied essence” [of Awareness] as it were, alone remains—then that too is the state Beyond the Fourth.

Here we have a clear progression of turya → turyātīta without the necessity of emendation. Or do we? According to the earlier ĪPV passage, in the Fourth state, one simply transcends the objective layers of the self, rather than those layers losing their objectivity. Thus, either Abhinavagupta has changed his view since writing the ĪPV, or an emendation is indeed necessary here. If the latter, we could either emend tat to sa (“one leaves behind their objectivity”) or we could emend turya-daśā (in the first paragraph above) to turyātīta-daśā. The latter solution, tentatively proposed by Torella in an email communication (July 2014), seems to me to ignore the grammar that suggests two stages here (the first structured around the relative/correlative yadā tu … tadā, and the second around yadāpi … tadāpi); or rather, more correctly, he sees the grammar (after his emendation) as referring not to two successive stages but to two kinds of turyātīta, the api in tadāpi informing us that “thus, this too is turyātīta.” However, then we have the problem that two apparently distinct stages of the same process are denoted by the very same word, turyātīta. That they are distinct stages is evidenced by the fact that in the first phase (the paragraph ending with the compound turya-daśā), we see the verb √vidh (penetrate20), and in the second (turyātīta) phase we have √vidh followed by abhiniviś, which here denotes a further development of the same process (as also indicated by the adverb atyantam, construing with abhiniviṣṭa). Furthermore, Abhinava’s recapitulation of this discussion (KSTS vol. 65, 348) would seem to argue against Torella’s conjecture here. On the other hand, if we did adopt the emendation, it would allow us to preserve the notion seen in the ĪPV that turya is the gnostic attainment and turyātīta the yogic one. It seems to me, however, that this creates more problems than it solves.

In summary, by not adopting the emendation, we see here a yogic21 version of the Fourth state that can be developed into the state Beyond the Fourth, thus indicating a change in Abhinavagupta’s thought since the ĪPV. To explain in more detail my understanding of this rather difficult passage, the process goes like this: having inundated/penetrated (viddha) the objective layers of selfhood (body etc.) with the “elixir” of one’s ultimate nature (i.e., saṃvid-rasa, autonomous dynamic consciousness), the “gold” hidden within them is extracted, i.e. their dependence on consciousness as their substrate is revealed.22 When those layers have become completely permeated (abhiniviṣṭo ’tyantaṃ), through, one presumes, further spiritual practice,23 all trace of their objectivity (and the saṃskāras thereof) is “worn away” or “digested” (jīrṇa) by the elixir of consciousness—as mercury eats up gold flakes—which thus becomes a single unitary mass of awareness (prakāśa-ghanam eva saṃvid-rūpam, cited infra).

Our understanding of Abhinava’s vision of this process depends in part on grasping his use of an alchemical metaphor rooted in the complex and often ambiguous rasāyana (alchemy, proto-chemistry) theories of medieval India. In this matter I was fortunate to receive the helpful comments of Professors Wujastyk and Houben (of the University of Alberta and the Sorbonne, respectively), who clarified that jīrṇa here stands in for jāraṇa, one of the sixteen rasa-saṃskāras (alchemical processes). Jāraṇa can mean digestion, assimilation, or swallowing (in much the same sense that we speak of an acid “eating away” at a metal). Here the alchemical elixir of the metaphor is of course prepared mercury (siddha-rasa), which can indeed “digest” gold (the modern term is amalgamate).24 That Abhinavagupta was aware of the basics of alchemy is confirmed by his use of the compound druta-rasa, for according to Houben, “initially the mercury remains as fluid as before it started to ‘eat’ the gold etc. but at a certain point its viscosity increases significantly … [it] remains fluid or druta [only] as long as it is not saturated.”25 Clearly, Abhinava wishes to emphasize that here this saturation does not occur, that Consciousness must be understood purely as a catalyst (something that effects change but is not itself affected): it remains as it is, a dynamic “fluid” essence (druta-rasa = cid-rasa).

First, then, in this alchemical vision, the mercury transmutes the base metal into gold,26 then “digests” or absorbs it without a (perceptible) trace (as can be seen in the chemistry video cited in note 24).27 That is to say, if we follow the terms of the metaphor strictly, first the layers of body, etc., are experienced as expressions of the dynamic essence of awareness, then all the saṃskāras implanted in those layers through one’s earlier experience of them as other than awareness are dissolved or “digested.”

To summarize, if we are constituting and interpreting the text correctly, Abhinava has changed his view as follows: in the ĪPV, turya is an exclusive, gnostic, transcendental state and turyātīta an inclusive, yogic, immanent one (the progression from one to the other exemplifying the typical Kaula model of transcendence followed by pervasion), with the alchemical metaphor denoting only the turyātīta stage; whereas in the ĪPVV, there is a rapid gnostic version of turya progressing to turyātīta (in which saṃskāras are not dissolved) and a gradualist yogic version of the same (in which they are dissolved in the turyātīta phase), both stages (of the latter) being described in terms of the alchemical metaphor. In the second text, then, we have a fork in the road, giving us four stages, only two of which a given practitioner is likely to traverse.

Before we move on to examine the last version of the alchemical metaphor, we have one more problem with the present passage: how to interpret the final compound of the phrase sa druta-rasa iva ābhāti kevalaṃ tat-saṃskāraḥ. Here I differ from Torella (2002, 209 n. 35), who seems to interpret it to mean that only the saṃskāras (impressions) of śūnyādi-dehāntam remain. However, that case was already specified for the first, gnostic turyātīta (tadāpi ca śūnyādi-saṃskāro ’py asti, above), and if that were intended here we would have nothing to differentiate the two turyātītas described. Thus I take tat-saṃskāraḥ in apposition to druta-rasaḥ, in the meaning “the purification (or refinement) of that,” or, as translated above on page 10, as a bahuvrīhi meaning “their purifier,” the antecedent of the neuter pronoun tat being śūnyādi-dehāntam in either case. Torella argues (email communication, 10 July 2014) that the saṃskāras of śūnyādi-dehāntam cannot be entirely dissolved here because then there would be no possibility of samāveśa, since, consciousness having become a single unitary mass, there would be nothing that could enter or be entered (cf. p. 151 supra). However, while Abhinava unambiguously does say this with regard to the after-death state (dehapāte tu eka-ghanā eva śivateti tadā samāveśādi-vyavahāro na kaścid, KSTS vol. 65, 328), I am not at all sure that he thought it impossible to go beyond samāveśa, as generally understood, before death; after all, in a continuous nondual state of “complete immersion” (a new sense of samāveśa starting with Utpala’s usage; see Wallis 2014) there will no longer be any kind of “entry” or “penetration” (ā√viś) per se.

We need not speculate overmuch on this question, however, for we can find evidence to suggest that Abhinava did regard such supervention of samāveśa as possible. That evidence is found almost twenty pages further on in the ĪPVV (KSTS vol. 65, 348, commentary on ĪPK III.2.19), where he recaps his earlier discussion (our most recent passage above) but also adds new information:

aham ity eka-rasena anuvedhe tu, yadā idantā ācchāditā bhavati, bhāvanā-sātmyād īśvara-sadāśiva-saṃvidi iva turya-daśāyāṃ rasa-viddha-tāmra-kanaka-nyāyena, yadā vā sarvathaiva pradhvaṃsitā vidrāvitā vā bhavati turyātīta-daśāyāṃ śākta-saṃvidi iva tan-nija-rūpa-samyag-viddha-kanaka-rūpatātyanta-jaraṇāpādita-tat-saṃskāra-vaśa-pītatā28-avaśeṣa-vidruta-rasa-nyāyena; tadā pūrṇa-svātantryollāsa eva deha eva sati api

In the [process of] transmutation by the “one taste” that is [the fundamental] “I,” when

  1. objectivity is covered, i.e. in the Fourth state [that arises] due to becoming habituated to meditative contemplation [on reality], in which one possesses the consciousness of Īśvara or Sadāśiva as it were, according to the maxim of gold [being extracted] from copper due to being penetrated by mercury,

or when

then [in either case]

This passage features a piling on of parenthetical phrases that is rather easier to understand in the Sanskrit than in literal English translation;30 I have illustrated the basic structure here as “when a) or b), then c).” Several things become clear from this paragraph, despite its density. First, if my translation is correct, it corroborates my reading of the previous alchemical passage. Second, it confirms that the saṃskāras are indeed “thoroughly digested” in the state Beyond the Fourth (atyanta-jaraṇāpādita-tat-saṃskāra-), and that this can occur with the body still existing (deha eva sati api). However, having said this, we must note that “thoroughly digested” does not mean “entirely destroyed” if Abhinavagupta is holding strictly to the terms of his metaphor; for when mercury absorbs gold leaf such that the gold is entirely dissolved and thus completely invisible, it is in fact still present in the mercury and can be retrieved by evaporating the latter in a retort. We have no way of knowing if Abhinava knew this, but if so, Torella could well be correct in arguing that a subtle trace of the saṃskāras (which are themselves subtle traces) can remain in the turyātīta-daśā. What certainly is entirely dispelled or dissolved (sarvathaiva pradhvaṃsitā vidrāvitā vā) in that state is objectivity, which was only “covered” (ācchāditā) by subjectivity in the turya state.

We find similar language in chapter three of the Pratyabhijñā-hṛdaya, authored by Abhinavagupta’s disciple Kṣemarāja, where the Sadāśiva-tattva is described in these terms: “[a level of consciousness] in which an implicit and indistinct objectivity is covered by [the predominant] subjectivity [literally, ‘I-ness’]” (sadāśiva-tattve ahantācchāditāsphuṭedantā-mayaṃ).31 Thus, according to the Pratyabhijñā schema, abiding in turya means achieving the īśvara- or sadāśiva-tattva and abiding in turyātīta means reaching the śakti-tattva. In either case the result is the “delightful blossoming of full autonomy” (pūrṇa-svātantryollāsa), i.e. liberation. But if the yogic/alchemical turyātīta is equivalent to reaching śakti-tattva, are we supposed to understand that the gnostic turyātīta reaches śiva-tattva, despite the fact that the latter turyātīta does not dissolve the saṃskāras and the former does? If so, does Abhinava mean to imply subtly that Śakti is in reality higher than Śiva? This would contradict the mainstream doctrines of Śaivism, but not of the Krama, the sect of Abhinava’s first initiation. We know from the Mālinī-vijayottara-tantra, Abhinava’s root-text, that the tattvas of Sadāśiva and Īśvara are indeed associated with the Fourth state and that “Śiva and Śakti exist in the state Beyond the Fourth” (2.28c–29b32). So we would expect that the two turyātītas are associated with tattvas 1 and 2. But which is which?

The reader will recall that we looked ahead eighteen pages in our primary source to see how Abhinavagupta recapitulated his alchemical metaphor. Now we return to our main passage (KSTS vol. 65, 330) to address the questions just raised. First we see that the distinction we have posited between the gnostic and yogic paths to liberation is not as clear-cut as it would seem:

yadā tu dehādeḥ kiṃ tattvam iti cintopakramaṃ prakāśa-ghanam eva saṃvid-rūpam iti, tadā bodha-svarūpīkṛtaṃ tad-rasānuviddham iva33 śūnyādi-dehāntam avabhāti. iti abhyāsāt tasya saṃvid-dharmāḥ śakti-viśeṣāḥ samyag āviśanto vibhūtīr utthāpayanti. anabhyāse ’pi tu tat-kṣaṇāveśa eva ānandodbhava-kampa-nidrā-vyāpti-rūpa-ghūrṇy-āvirbhāvana-krameṇa jīvanmuktatā-lābhaḥ |

When one begins to contemplate “What is the reality of the body, etc.?” [and subsequently realizes] “it is simply a form of awareness, replete with the Light of Consciousness,” then those [levels] from the Void to the body manifest as [they really are,] of one essence with Awareness, as if transmuted by its elixir. Thus, due to practicing [this insight], the qualities of His consciousness, which are aspects of śakti, fully penetrate [those various levels], causing the [various] powers (vibhūti) to arise. But even without practice, in the [rare] case of an instantaneous immersion into That, one obtains the state of liberation-in-life through the process of the direct experience of [the Five Mystic States]: Bliss, Ascent, Trembling, Sleep, and “Whirling,” which means Pervasion.

This passage serves as an explanation of tat-śakti-samāveśa in Utpala’s Vṛtti on III.2.12. The passage confirms for us that Abhinava sees samā√viś (= atyantam abhini√viś in the previous iteration of the alchemical metaphor) as denoting the further development of the process first denoted by √vidh. More importantly, here we find crucial evidence that the distinction between the yogic and gnostic in Abhinavagupta’s thought is not as clear-cut as we have been led to expect: in this passage, a contemplation (cintā) on the nature of reality leads to a realization that entails a spiritual transformation metaphorically described as alchemical transmutation (= turya stage), which then may be stabilized and enhanced with yogic practice such that the qualities of this deeper awareness (e.g., svātantrya-śakti) come to fully penetrate or infuse (samā√viś) all the layers of limited selfhood (= turyātīta stage). Gnostic realization is here inseparably wedded to the pañcāvasthāḥ or Five Mystic States that we see repeatedly in the Kaula scriptures. This emphasis on direct experience (āvirbhāvana) demonstrates that Abhinava’s understanding of the path of the jñānī is not one of intellectual or conceptual realization, but rather one of insights into the nature of reality so powerful that they spontaneously bring on psychophysical experiences.

Curiously, he uses the phrase “instantaneous immersion” or kṣaṇāveśa in describing gnostic realization but then immediately follows it with the term krama, denoting a sequential process of passing through the Five States. I would hypothesize that Abhinava is saying that each of the Five States is (or rather can be) an example of kṣaṇāveśa; even though there is a process, it may unfold spontaneously and in sudden leaps, in connection with the jñānī’s deep contemplation of the nature of reality.

At any rate, now our reading of two distinct tracks (or two distinct turyātītas), one gnostic and one yogic, is problematized. Yet we cannot abandon it, for on the very next page of our text (KSTS vol. 65, 331) we find the following:

turya iti etad-daśā-samāpatti-paryanta-rūpā api turyātītatā tatraiva uktā, vyatirekeṇa tu śūnyāder avasthāpane bodhasya turyātītatā tatraiva uktā—śuddhātmatā niṣkalatā śuddha-caitanyam iti sāmānya-darśaneṣu sarvottīrṇaika-tadrūpa-vedakeṣu darśiteti sūcayati |

The state of turyātīta taught [above] with reference to that [blossoming of insight34] is simply the [further] extension of the realization of the state called turya. But that state of turyātīta was taught there as a state of awareness in which Void etc. remain [as objective knowables], but is separated (vyatireka) [from them].35 This is the state referred to as “the pure Self,” “the Formless,” and “pure Consciousness” in the Saiddhāntika scriptures. It is taught with reference to those who know the Deity36 solely as [being] all-transcendent; so [Utpaladeva] indicates [in his Vivṛti].

Again we are presented with the notion of a vyatireka-turyātīta, though here its transcendental as opposed to its gnostic character is stressed. Since this state is associated with the transcendent deity (sarvottīrṇaika-tadrūpa, a kenning for Anāśrita-śiva, usually considered tattva 1), we are inclined to conclude that, as suggested previously, Abhinava wishes us to understand that the turyātīta which accesses the śakti-tattva is in fact the superior one. Perhaps this is not stated explicitly because it is a subversive view relative to the mainstream of the tradition.

Here Abhinava criticizes the exclusively transcendentalist view of the Saiddhāntikas, since as a nondualist Śākta he wishes to privilege the Kaula “immanentist” view. Likewise, his disciple Kṣemarāja argues that the defining feature of their lineage (trikādi-darśana, meaning the view of the Trika-Krama synthesis uniquely expounded by Abhinavagupta) is the view that the Divine Self is simultaneously transcendent of all and immanent in all.37 Now, since the transcendental turyātīta is identified with the teachings of the sāmānya-śāstra (i.e., right-current Śaiva Siddhānta), we would expect the yogic-cum-alchemical turyātīta to be identified with the viśeṣa-śāstra (i.e., left-current Śākta streams), and just such identification is found by Torella (2002, xxxiv note 52) on the same page of the ĪPVV as the passage just quoted, but to me the text is not so clear. Apparently glossing viśeṣa-darśaneṣu in Utpala’s lost Vivṛti, Abhinava says (KSTS vol. 65, 331): “The specialized views are those which predominantly teach the nonduality of Power [and the Power-holder, i.e. Consciousness]; they are superior (viśeṣa) because they teach the Power-characteristics of the Self, [and thereby] lead one to the direct experience [of that]” (viśeṣa-darśaneṣu iti śāktādvaita-pradhāneṣu, viśeṣaḥ śakti-lakṣaṇa ātmano darśyate sākṣātkāraṃ nīyate yeṣu). There is no explicit correlation with the second, yogic turyātīta, but perhaps we can assume it because the power element of the self (vaibhavādi) is repeatedly mentioned in the passages we have examined in connection with that turyātīta.

5 Summary of the Data

Now to summarize what we have learned about these two states of immersion.38 The primary distinction Abhinava wishes to make is that the first turya → turyātīta constitutes an “active” entry/immersion into one’s essence-nature (one’s satya-svarūpa or ahaṃbhāva, which is cidānanda and prakāśa-ghana), and the second turya → turyātīta denotes a “passive” process by which one is entered; i.e., that in which the various layers of selfhood are permeated by that ultimate I-sense (note that it is passive only in the grammatical sense, for the yogic method involves considerably more work). This distinction is summed up as āveśyāveśaka-bhāvaḥ (p. 331). Since the yogic process is a gradual one, differences are noted between turya and turyātīta, whereas the gnostic turyātīta is said to be identical in nature (though presumably not in degree) to the turya that precedes it (turyatā … turyātīta-samatā).

Table 6.1 summarizes the data in tabular form (items in parentheses are only implicit in the primary source text).

Table 6.1

Two types of Turyātīta in the ĪPVV

Turyātīta (1)

Turyātīta (2)

gnostic (realization)

yogic (practice)





knowables transcended

knowables transmuted

(objectivity transcended)

objectivity dissolved





(īśvara- or sadāśiva-tattva in turya39)

qualities of the Self:

qualities of the Self:

vyāpaka, nitya

vaibhava, aiśvarya





6 Problems of Interpretation

To close, I will briefly discuss some interesting ambiguities and difficulties of these sections of the ĪPV and ĪPVV, apart from those already discussed above and in the notes. The first problem is the one raised (but not discussed) by Torella (2002, xxxv note 52), who tells us that Abhinavagupta’s inclinations lie towards the second turyātīta, the yogic-alchemical one. This seems evident from the passages we have translated here, but on the other hand it is surprising, because in general Abhinava tends to privilege the gnostic over the yogic in his work. Perhaps the context of a clearly liberation-focused yoga outlined in ĪPK III.2.19–20 allowed him to endorse a term that otherwise so commonly denotes the pursuit of siddhi and bhoga in the Śaiva literature.

But we have another problem with Torella’s hypothesis, one briefly discussed already above: the implication in these passages is that the gnostic turyātīta—(1) in Table 6.1—reaches the śiva-tattva, while (2) explicitly only reaches the śakti-tattva. My reasons for concluding that turyātīta (1) reaches the śiva-tattva are as follows: since the term turyātīta refers to the highest liberation where only the absolute “I” remains, and in the tattva-system that attainment is explicitly identified with the top two tattvas, turyātīta (1) must correspond to the śiva-tattva, which also matches its transcendent nature (the Śiva of tattva 1 is often called Anāśrita-śiva40). And we have seen that turyātīta (2) aligns with the śakti-tattva. Now, it doesn’t seem altogether likely that Abhinavagupta would favor an attainment that reached only to tattva 2. However, he may well have held the view explicitly articulated by his successor Kṣemarāja, i.e. that there is no reality to hierarchy with regard to Śiva/Śakti, they being two aspects of one reality, one or the other of them being more prominent in the liberated experience at any given moment (see Vasudeva 2004). In other words, Abhinava may have held the view that since Śiva and Śakti are in fact inseparable except heuristically, to attain one is to attain the other. Or perhaps, as already posited above, we are to understand that Abhinava is allusively suggesting that in this system, Śakti is in fact tattva 1. Such a hidden doctrine would be in keeping with the rest of his esoteric theology, which constitutes a Trika doctrine with a Krama core.

The other problem of interpretation that arises in connection with these passages is a philosophical one, not yet to my knowledge addressed in the secondary literature. This issue centers on the question of who is the agent of the various verbs used here, most especially √viś. For example, when it is said that “there is an entry into one’s true nature,” who or what is the agent of that entry, since the satya-svarūpa that is entered is the only real source of both awareness and agency? This problem is not glaringly obvious because of the nature of the Sanskrit language, in which passive voice is so common, and nouns or pronouns denoting the agent can be omitted entirely, the verb conjugation itself communicating a generic unspecified third-person agent. When a first-person active verb is used, for example when Kṣemarāja glosses namas or naumi as samāviśāmi (in his commentaries on stotras41), the problem is made evident: what exactly is the “I” that enters? Obviously it cannot be mind, body, etc. (since they have no agency of their own), nor can it be cit, for it would make no sense to say that consciousness, which is undivided, enters into itself.

Two possible solutions occur. First, that what enters from the mind, etc., into the Self is a kind of “locus of subjective identity” or ahaṃbhāva. But Abhinava seems to use this term in the passages we have considered to mean the true I-sense, the Self-that-is-awareness. The second and more likely solution is that this language of entry is purely metaphorical, derived from the phenomenology of the experience it denotes (i.e., what it feels like to have that experience), and that in actuality there is no entry at any time: the true “I” simply realizes itself or wakes up to itself, clearly apprehending itself. It may be the case that Abhinava addresses this problem of agency in relation to √viś verbs somewhere in his vast body of work, and I simply have not yet come across it.

7 Postscript: Samāveśa and Turya in Tantrāloka 10

Abhinavagupta also discusses turya and turyātīta briefly in Tantrāloka 10.264–284. Though this passage is not directly a commentary on the ĪPK (being instead a commentary on the phenomenological categories of Mālinī-vijayottara-tantra, chapter 2 and passim), it clearly has in mind Utpaladeva’s phrase tat-śakti-samāveśa in the Vṛtti on ĪPK III.2.12 and very likely also alludes to his lost Vivṛti on the same.42 For Tantrāloka 10.265 informs us that turya is in fact śakti-samāveśa!43 This seems to imply that turyātīta is śiva-samāveśa, which could match the ĪPV account but not the ĪPVV; again, perhaps Abhinava changed his view between the two, the latter having been composed after the Tantrāloka. Abhinava goes on (in 10) to subdivide this śakti-samāveśa into four stages, corresponding to the four epistemological categories of knower, knowing, known and the autonomous pure awareness (para-pramātṛ) which is the source and ground of the previous three (Vasudeva 2004, 230).44 The four stages are as follows. When immersion into this parā saṃvit is only proximate (tat-samāveśa-naikaṭyāt), object-consciousness is dominant ( 10.270d–271a). When there is contact45 with this immersion, the process or faculties of knowing are dominant (tat-samāveśoparāgān mānatvam, 270cd). In full identification with this immersion, the state of the knower becomes clear (tat-samāveśa-tādātmye mātṛtvaṃ bhavati sphuṭam, 270ab). Beyond this, in the state of the so-called pure awareness in which all three previous categories have perfectly fused, in which there is perception but no sense of a separate perceiver or perceived, the Light of Consciousness is self-manifest and we cannot speak of an immersion, except metaphorically (10.269). This helps us understand that the movement into the transcendent Fourth state can indeed happen in stages, the first three of which (prameya, pramāṇa, pramātṛ) are an expression of divine grace (trayaṃ tat tad-anugrahāt, 270b). We can infer that speaking of grace is meaningless in the fourth stage, where there is no duality (though the ĪPV and ĪPVV accounts tell us that there are still saṃskāras of duality at this stage).

Abhinava then defines turyātīta at 10.278: “that [state] whose beautiful nature is full and undivided, overflowing with joy, is called Beyond the Fourth; that alone is the supreme state” (yat tu pūrṇānavacchinna-vapur ānanda-nirbharam turyātītaṃ tu tat prāhus tad eva paramaṃ padam46). His discussion goes on, but it reaches beyond the purview of the present work.

8 Conclusion

The present paper does not, of course, entirely solve the complex textual and exegetical problems we discovered in the comparison of these passages of the ĪPV and the ĪPVV, but perhaps it contributes towards an understanding of their significance for the study of Tantric Śaiva theology. Provisionally, I propose that we see Abhinavagupta changing, developing, and nuancing his view in the time between the ĪPV and the ĪPVV (with Tantrāloka falling between the two). If I am reading the texts correctly, the ĪPV features a simpler model of a gnostic transcendentalist turya succeeded by a “immanentist” turyātīta (the latter being marked by the transcendent element’s pervasion of all that was previously transcended), while the ĪPVV proposes two distinct versions of both turya and turyātīta, gnostic and yogic respectively (giving us four categories in total), where the yogic is to be preferred despite being more gradual because in it the saṃskāras of dualistic experience are finally dissolved.


conj. em.

conjectural emendation




Īśvarapratyabhijñākārikā of Utpaladeva


Īśvarapratyabhijñāvimarśinī of Abhinavagupta


Īśvarapratyabhijñāvivṛtivimarśinī of Abhinavagupta


Kashmir Series of Texts and Studies

Tantrāloka of Abhinavagupta


Tantrālokaviveka of Jayaratha


An earlier version of this work is found in my unpublished doctoral dissertation (Wallis 2014).


“The kośas, mediated through the Pañcīkaraṇa system ascribed to Śaṅkara, had become part of the [Deccani] vernacular tradition by the end of the twelfth [century, and proceeded from there into the pan-Indian Sanskrit tradition].” Jason Schwartz, personal communication, April 2018.


See, e.g., Pratyabhijñāhṛdaya, chapter 7: śūnya-prāṇa-puryaṣṭaka-śarīra-svabhāvatvāt catur-ātmā.


Pratyabhijñāhṛdaya, chapter 8: … dehādiṣu bhūmiṣu pūrva-pūrva-pramātṛ-vyāpti-sāratā-prathāyām.


While these verses have been translated a number of times (cf. Torella 2002, 202–203), they are not easy to translate in such a way that the reader clearly understands what is being said. Here I capitalize words that are equivalent, on this view, to the Deity. An unobtrusive but important word here is tat-, which I have translated as That but could also have been rendered Him. Assuming that it is to be taken as compounded with what follows, then it must denote what one is immersing in. The use of a gender-neutral pronoun that could just as well denote neuter tattvam as masculine Him (= Śiva) is exemplary of the decreased theism of the Pratyabhijñā phase of the tradition.


For example, Abhinava writes idantāpanna-dehādi-śūnyānta-prameya-bhāga-nimagnatvāt prameyam, yo gauro, yaḥ sukhī, yas tṛṣito, yaḥ sarva-rūpa-rahitaḥ so ’ham: “The levels from body to the Void are objects of awareness, [but] because of the submerging of a portion of that objectivity, [there arise the erroneous cognitions] ‘I am the one who is pale’ (body), ‘who is happy’ (mind), ‘who is thirsty’ (prāṇa), ‘who was devoid of all appearances’ (void)” (KSTS vol. 33, p. 230).


seyaṃ jāgrat-svapna-suṣupta-rūpā saṃsārāvasthā (KSTS vol. 33, p. 230). The illusion of separate individuality persists even in the deep sleep state because of the presence of the saṃskāras (cf. ĪPK III.2.13).


turyāvasthā ] conj. em. Torella (email communication, July 2014); turyātītāvasthā Ed.


-vaibhava-nityatva-aiśvaryādi-dharma-saṃbhedena ] conj. em.; vaibhava-nityaiśvaryādi-dharma-saṃbhedena KSTS ed.; vaibhavādi-nityaiśvarya-saṃbhedena Iyer & Pandey ed.


turyātīta- ] conj. em. Torella (email communication, July 2014); turya- Ed.


Āyurveda scholar Dominik Wujastyk (of the University of Alberta) recommended this translation of vidhyate as “transmuted” (over that of “penetrated”) based on his reading of the rasāyana literature, especially the Rasa-ratna-samuccaya 8.94–95 and the Bodhinī thereon (email communications, 7 and 9 July 2014). Ashok Aklujkar also contributed a citation from the same text (5.11: vedhajaṃ suvarṇam—pārada-vedhena saṃjātaṃ suvarṇam), which I believe verifies that vedha must mean transmutation (or similar), not piercing or penetration, though he would wish to retain the latter translation (email, 7 July 2014). (See also n. 27 below.) The alchemical metaphor here (elaborated further in the ĪPVV, infra) is of course not original to Abhinavagupta; we find it earlier in the well-known eighth century Buddhist text, the Bodhicaryāvatāra, 1.10cd: rasajātam atīva vedhanīyaṃ sudṛḍhaṃ gṛhṇata bodhicitta-saṃjñaṃ, which I translate as “Firmly take hold of the alchemical elixir called Intent to Awaken (bodhicitta), which must be thoroughly transmuted.” Vesna Wallace (1997, 19) translates almost identically; in this verse, it appears, rasajāta is unmodified mercury that must be properly transmuted to be safe for consumption, implying that some refinement of the initial raw bodhicitta is necessary. (However, Matthew Kapstein [email communication, 9 July 2014] points out that both the Sanskrit commentator and the Tibetan translation do not take vedhanīyaṃ as translated above, but rather in the active sense, “able to transform [this aśuci-pratimā to a jina-ratna-pratimā, 1.10ab]”—e.g., Prajñākaramati glosses atyuccavedhakāritvād—atīva vedhanīyam.) Our passage does support this latter reading, for here we certainly have siddha-rasa denoting a mercury preparation that can transmute base metals into gold (or more accurately, extract gold from base metals). Thus, the pure dynamic power of awareness called cit is here compared to a chemical catalyst: it needs no refinement or alteration, but can alter that which it contacts.


Cf. Kulārṇava-tantra 14.89: rasendreṇa yathā viddham ayaḥ suvarṇatāṃ vrajet | dīkṣā-viddhas tathā hy ātmā śivatvaṃ labhate priye ||, “Just as iron penetrated by mercury becomes gold, even so a soul penetrated by initiation becomes divine.” Torella writes, “[here] the various components of the levels of the limited subject are gradually penetrated by the elixir of the I, until they become, so to speak, transfigured, removed from their nature of [being merely] knowable realities” (2002, xxxiv).


Kṣemarāja’s phrase, cidrasāśyānatā-prathanātmā samāveśaḥ, in the context of a parallel discussion, in Pratyabhijñāhṛdaya, chapter 19.


jāgrat-svapna-suṣupta-bhede turyābhoga-saṃbhavaḥ.


Abhinava is telling us that the sam- in samāveśa is in the sense of samyañc.


Though Torella proposes an emendation which would bring the ĪPVV in line with the ĪPV (see below).


We have here an implicit analysis of the word samāveśa: samyag and/or āsamantāt + praveśa = samāveśa.


turyatā ] conj. em. Torella (email communication, 16 July 2014); turyātītatā Ed. Without this emendation, the following comment turyātīta-samatā eva makes little sense.


iti vyatireka ] conj. em. Torella; iti avyatireka Ed. Following this emendation (proposed in an email, 15 July 2014) we can take vyatireka in the sense of kevala or kaivalya, i.e., a spiritual state which is separated from the saṃskāras but does not dissolve them. Even if we do not emend, we can still argue for the same meaning: avyatireka- could indicate that he is “unseparated” from his saṃskāras in the sense of still having them, though they are now powerless to obscure his real nature. However, the emendation makes for a clearer meaning.


Oriental Research Library manuscript no. 2403 has śūnyādi-deha-dhāturtvaṃ here.


Note that the verb is here being used in a more precise sense than in the ĪPV passage, in which (I argue) it means “transmute.” See also the extended discussion on vedha-dīkṣā in 29, translated in an appendix to my doctoral dissertation (Wallis 2014).


Besides the alchemical metaphor, Abhinava signals to us that this is a yogic process with the word aiśvarya, which, like v(a)ibhava, often relates to yogic power (siddhi).


Cf. Sarvajñānottara 1.5: tāmrasyaiva tu hematvam antarlīnaṃ yathā sthitam | antarlīnaṃ tathā jñeyaṃ śivatvaṃ pudgalasya tu (“Just as gold is hidden within copper, in the same way the Divinity which a man seeks to know is hidden within [him].”).


In the present context, the nature of the yogic practice alluded to is very likely the proto-kuṇḍalinīyoga that Utpaladeva outlines at ĪPK III.2.19–20 and which was presumably elaborated in his Vivṛti. For Abhinava, such practice must be animated by bhāvanā (contemplative insight) to be truly effective.


This can be clearly seen in a video made by Dr. Andrea Sella (Department of Chemistry, University College London), which includes the following comment: “[In ancient times] mercury was absolutely essential … in extracting gold and purifying it … what gold can do is, it can actually dissolve in mercury” (


Email communication from Dr. Jan E.M. Houben, 7 July 2014.


Even though professional alchemists must have known that mercury actually extracts gold from a base metal, rather than magically transmutes that metal into gold, vedha is certainly used in the sense of transmute or transform—see the citations in note 10, in one of which √vidh is glossed with pari√ṇam (Rasa-ratna-samuccaya-bodhinī ad 8.95). See also the relevant statement in Roşu 1982, 366: “la transsubstantiation [alchimique] du corps (deha-vedha) étant calquée sur la transmutation des métaux vils (loha-vedha) …” He cites Rasārṇava 12.165–166 in support (ibid., note 21).


The reader who has German and wishes to know more about this arcane world of Indian alchemy is referred to Hellwig 2009.


pītatā ] conj. em. Isabelle Ratié (email, July 2014); pītalatā Ed., though pītatā as “gold” is problematic. Another possibility is to not emend the text, and take it instead to be speaking of the digestion/dissolution of the brass or copper (pītala) that remains after gold has been extracted from it. This has not been adopted on the assumption that the present passage recaps the one on p. 153 supra. Also, I presume that Abhinavagupta, not being himself an alchemist, viewed the process of vedha as one of transmutation more than extraction—and if so, there would be no brass (or copper) left to digest. Further, the metaphor of transmutation suits his purposes better.


Some parallel passages: cf. 14.12, 5.151 (svayaṃbhāsātmanānena tādātmyaṃ yāty ananyadhīḥ | śivena hematāṃ yadvat tāmraṃ sūtena vedhitam ||), and Yogarāja’s commentary ad Paramārthasāra 96, a verse on anupāya and atitīvra-śaktipāta, the effect of which is compared by the commentator to alchemical transformation by means of siddharasa (yathā tāmra-dravyaṃ siddharasa-pātāt suvarṇībhavati). Note that later in the same passage the aspirant is referred to as anugraha-śakti-viddha-hṛdayasya, “one whose heart has been penetrated/transmuted by the power of divine grace (i.e., śaktipāta).” We find the same terminology used with reference to dīkṣā, e.g. in the Kulārṇava-tantra (14.89): rasendreṇa yathā viddha-mayaḥ suvarṇatāṃ vrajet | dīkṣā-viddhas tathā hy ātmā śivatvaṃ labhate priye ||, “Just as [a metal] penetrated by mercury becomes gold, even so a soul penetrated by initiation becomes divine.” Cf. Goodall 2004, 402 note 904:

The conception that gold can be created out of copper with an alchemical preparation is, as Prof. Isaacson has pointed out to me, commonly used in tantras as an image for the irreversible transformation that takes place in dīkṣā. See, e.g., Kiraṇa 59.36c–38b … And cf. Sarvajñānottara 1.6 (Devakoṭṭai ed.): rasa-viddhaṃ yathā tāmraṃ hematvaṃ pratipadyate | tathātmā jñāna-sambandhāt śivatvaṃ pratipadyate ||. Cf. also Haravijaya 6.137 … In his commentary thereon Alaka cites the following verse: rasa-ghṛṣṭaṃ yathā tāmraṃ na bhūyas tāmratāṃ vrajet | evaṃ yuktaḥ śivatvena na bhūyaḥ paśutām vrajet.

I render the last verse cited as: “Just as copper rubbed with mercury [becomes gold and] does not again become copper, in the same way one united with Divinity does not again become a bound soul.” Goodall informs me (email communication, 13 July 2014) that the Sarvajñānottara verse cited in his footnote (viz., 1.6) is the most typical form of the maxim; the verse immediately preceding it is cited above in note 22.


And I wonder if the largely redundant first part of the long compound (tan-nija-rūpa-samyag-viddha-kanaka-rūpatā), which fits awkwardly, might have been a marginal annotation in a manuscript that got incorporated into the main text. However, it may be more likely that Abhinava is here incorporating language that Utpala used in his lost Vivṛti, as he tends to repeat Utpala’s compounds with added glosses.


However, Kṣemarāja loc. cit. posits īśvara-tattva as a level in which subjectivity and objectivity are equal and opposite, whereas in our present passage Sadāśiva and Īśvara are not differentiated, both being described as a level at which objectivity is “covered.”


See Vasudeva 2004, 209–210.


iva ] conj. em.; eva Ed.


Inferring the antecedent of tatra from the previous line: jñānollāsa iti ajñāna-vigamād advaya-bodha-prasaraṇād ullāsa[ḥ], “ ‘the blossoming of insight is a blossoming that results from the departure of ignorance and the [concomitant] spread of nondual awareness.”


Exactly how to construe the grammar of this sentence is not clear to me, nor am I convinced that the text is secure.


For Abhinava’s use of tadrūpa to mean the Deity, see, e.g., 1.173c–174b.


From chapter 8 of his Pratyabhijñāhṛdaya: “The Tāntrikas (= Saiddhāntikas and other ritualists) hold that the reality of the Self is all-transcending. Those attached to traditions such as the Kula say it is all-embodying. Those who hold [our] viewpoint of the Trika and [the Krama] hold that it is [simultaneously] all-transcending and all-embodying” (viśvottīrṇam ātma-tattvam iti tāntrikāḥ, viśva-mayam iti kulādy-āmnāya-niviṣṭāḥ, viśvottīrṇaṃ viśvamayaṃ ca iti trikādi-darśana-vidaḥ).


sā iyaṃ dvayī api daśā samāveśo, KSTS vol. 65, 328.


Note that the ĪPK itself (III.2.20) lists the levels of turya-attainment as those of the Vijñānākalas (= level of Mahāmāyā, just outside the śuddhādhvan and therefore not yet liberated), Mantras (= śuddhavidyā-tattva, lowest level of liberation), and Īśvara; but Abhinava takes mantreśa in that verse to refer to the Mantra-lords of īśvara-tattva, then reads -īśa a second time, taking it to refer Lord Sadāśiva (Torella 2002, 208 note 33). Here he is making the correlation correspond to what is found in Trika scripture, for the Mālinī-vijayottara-tantra teaches that “the Mantras, Mantreśas and Mantramaheśvaras occupy the Fourth state” and “Śakti and Śiva exist in the state Beyond the Fourth” (2.28c–29b, trans. Vasudeva 2004, 209–210).


E.g., in Pratyabhijñā-hṛdaya chapters 4 and 5.


See Stainton 2013 or Stainton 2019.


This supposition receives support from the fact that Abhinava explicitly comments on ĪPK III.2.15–17 further on in the same chapter (viz., 10).


pūrṇatāgamanaunmukhyam audāsīnyāt paricyutiḥ | tat turyam ucyate śakti-samāveśo hy asau mataḥ ||, paraphrased by Vasudeva (2004, 229) as “In the fourth state … knowable entities appear as awareness on the verge of reaching plentitude because [the] indifference [that characterized the third state of deep sleep] is abating. Abhinavagupta further identifies this state as an immersion into Śakti.”


For the fourth category of pure awareness, see 10.269: pramātṛtā svatantratva-rūpā seyaṃ prakāśate | saṃvit turīya-rūpaivaṃ prakāśātmā svayaṃ ca sā ||, and Jayaratha ad loc.: parā saṃvid evam aṃśa-trayottīrṇā … svātantrya-mayī para-pramātṛtā … sā hi para-pramātṛ-rūpā śuddhā saṃvit svayam prakāśate na tu paśyāmītyādi-vikalpollekha-bhūmiḥ. That there are four epistemological categories, not three, is due to the influence of the Krama, in which Kālī is identified with the fourth.


Uparāga seems a strange word to use here; its commonest use is “eclipse” or “affliction”—perhaps we should emend to upayoga.


Or we could take pūrṇānavacchinna-vapur as a bahuvrīhi meaning “in which every beautiful embodiment is full and unlimited,” the implication being that every part is now experienced as containing the whole (akhaṇḍa-maṇḍalākāram).


The text editions cited are available online at the Muktabodha Indological Research Institute’s Digital Library:


Primary Sources47

  • Īśvarapratyabhijñākārikā. Raffaele Torella, ed. The Īśvarapratyabhijñākārikā of Utpaladeva with the Author’s Vṛtti: Critical Edition and Annotated Translation. Rome: IsMEO, 1994.

  • Īśvarapratyabhijñākārikāvṛtti. See Īśvarapratyabhijñākārikā.

  • Īśvarapratyabhijñāvimarśinī. Madhusudan Kaul Shāstrī, ed. The Īśvarapratyabhijñā of Utpaladeva with the Vimarśinī by Abhinavagupta. KSTS, nos. 22 and 33. Pune: Aryabhushan Press, 1918–1921.

  • Īśvarapratyabhijñāvimarśinī. K.A.S. Iyer and K.C. Pandey, eds. Īśvara-pratyabhijñā-vimarśinī of Abhinavagupta: Doctrine of Divine Recognition. Vol. 2. Reprint, Delhi: Motilal Banarsidass, 1986 [1950].

  • Īśvarapratyabhijñāvivṛtivimarśinī. Madhusudan Kaul Shāstrī, ed. The Īśvarapratyabhijñā Vivṛtivimarśinī by Abhinavagupta. 3 volumes. KSTS, nos. 60, 62 and 65. Pune: Aryabhushan Press, 1938, 1941, and 1943.

  • Tantrāloka. Paṇḍit Mukund Rām Shāstrī (vol. 1) and Paṇḍit Madhusudan Kaul Shāstrī (vols. 2–12), eds. The Tantrāloka of Abhinava Gupta, With Commentary by Rajānaka Jayaratha. 12 vols. KSTS, nos. 23, 38, 30, 36, 35, 29, 41, 47, 59, 52, 57 and 58. Allahabad: The Indian Press Ltd. (vols. 1, 5–6); Bombay: Nirnaya Sagar Press (vols. 7, 10–12), Tatva-Vivechaka Press (vols. 3–4, 8–9), and Shri Venkateshvar Steam Press (vol. 2), 1921–1938.

  • Tantrālokaviveka. See Tantrāloka.

  • Pratyabhijñāhṛdaya. Paṇḍit Mukund Rām Shāstrī, ed. KSTS, no. 3. Pune: Aryabhushan Press, 1911.

Secondary Sources

  • Hellwig, Oliver. 2009. Wörterbuch der mittelalterlichen indischen Alchemie. Supplements to eJIM, vol. 2 Eelde: Barkhuis & University of Groningen.

    • Crossref
    • Export Citation
  • Roşu, Arion. 1982. “Yoga et alchimie.” Zeitschrift der Deutschen Morgenländischen Gesellschaft (bd. 132): 363–379.

  • Sanderson, Alexis. 2007. “The Śaiva Exegesis of Kashmir.” In Mélanges tantriques à la mémoire d’ Hélène Brunner / Tantric Studies in Memory of Hélène Brunner, edited by Dominic Goodall and André Padoux, 231–442. Collection Indologie, no. 106. Pondicherry: Institut Français d’ Indologie/École française d’ Extrême-Orient.

  • Sella, Andrea. YouTube video, 6 April 2011.

  • Stainton, Hamsa. 2013. “Poetry and Prayer: Stotras in the Religious and Literary History of Kashmir.” Ph.D. dissertation, Columbia University.

    • Crossref
    • Export Citation
  • Stainton, Hamsa. 2019. Poetry as Prayer in the Sanskrit Hymns of Kashmir. Oxford: Oxford University Press.

  • Torella, Raffaele. 2002 [1994]. The Īśvarapratyabhijñākārikā of Utpaladeva with the Author’s Vṛtti: critical edition and annotated translation. 2nd ed. Delhi: Motilal Banarsidass.

  • Vasudeva, Somadeva. 2004. The Yoga of the Mālinīvijayottaratantra. Critical Edition, Translation and Notes. Pondicherry: Institut Français de Pondichéry/École française d’ Extrême-Orient.

  • Wallace, Vesna A. and B. Alan Wallace, trans. 1997. A Guide to the Bodhisattva Way of Life: Bodhicāryāvatāra. Ithaca, NY: Snow Lion Publications.

  • Wallis, Christopher. 2014. “To Enter, to be Entered, to Merge: The Role of Religious Experience in the Traditions of Tantric Shaivism.” Ph.D. dissertation, University of California at Berkeley.

  • Watson, Alex. 2006. The Self’s Awareness of Itself: Bhaṭṭa Rāmakaṇṭha's Arguments against the Buddhist Doctrine of No-Self. Wien: De Nobili.

  • Collapse
  • Expand

Śaivism and the Tantric Traditions

Essays in Honour of Alexis G.J.S. Sanderson

Series:  Gonda Indological Studies, Volume: 22


All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 196 111 7
PDF Views & Downloads 231 115 9