Recognizing Recognition: Utpaladeva’s Defense of Śakti in His “Proof of Relation” (Sambandhasiddhi)

Though one of many possible interpretive orientations, Utpaladeva’s short work, “The Proof of Relation”, may be profitably read in terms of the intention to reveal Śiva via an exposition of His śaktis. This intention, as declared by the author himself in his much more widely studied Verses on the Recognition of the Lord, if carried over into the “Proof of Relation”, does much to account for Utpaladeva’s intention in the latter text, which, although written to convey a rational argument across sectarian lines, may also be read as an exposition of the opening verses of benediction.

What sort of relation exists between thing, word, and meaning?And how might this threefold relation bear upon consciousness, being, or reality?Is the status of such metaphysical relations confined only to the epistemic?Are relations evanescent realities, or mere ascertainments, in the domain of cognitive biases or convenient fictions-quick-thinking shortcuts for the sake of quotidian corporate-world transactions, but, alas, imprecise judgments that break down completely under conditions that fail to match a strictly controlled set of variables?Or, conversely . . .Are there relations that are the underlying conditions for the mutual involvement that all objects evidently have with one another?Are there any relations possessed of inherent causal efficacy, which would allow one to intuit correctly their ontological status under any and all conditions however subtle, concealed, or intrinsic such ontological status might be?In the Indian philosophical tradition, examples of both extremes are to be found concerning attitudes toward metaphysical relations (sambandhas).Both such extreme cases are tied to contrasting assessments of the role that Sanskrit linguistics plays in expressing or embodying cognition.
One answer of the latter kind-relation-as-ontological-is explored by the influential philosopher and theologian of Kashmir, Utpaladeva, one of the paradigmatic early writers in the "Doctrine of Recognition" (Pratyabhijñā) textual tradition, who also called themselves "Autonomists" (Svātantrya-vādins).His lesser-studied short text, the "Proof of Relation" (Sambandhasiddhi) takes up these questions by way of response, in large part, to Dharmakīrti.The latter, a Buddhist philosopher and epistemologist-an intellectual arch-nemesis of the Śaiva Autonomists, though also a worthy exemplar of a rival type of idealism-also wrote a short text, "Examination of Relation" (Sambandhaparīks .ā).For Dharmakīrti, relation, it must be said, does not pass its exam.Not surprisingly, as an influential voice in the Buddhist textual tradition of Idealists (Vijñāna-vādins) as well as a theorist of epistemology (pramān .a), Dharmakīrti propounds a vision of relations of the former kind-that is, relation-as-merely-epistemic.Such Buddhist criticism, philosophically and socially successful as it was, ruffled the Śaiva nondual theologians who shared with the Buddhists a commitment to the deroutinization and elimination of a mindstate characterized by cognitive biases rooted in strategic, binary, choice-based thinking (the vikalpa state) in preference to a state of uncontrived, unbroken, unbiased cognition (the nirvikalpa state).To suggest that relations are themselves in the domain of the contrived (vikalpa or in Dharmakīrti's words, kalpanā) deals an injurious insult to realists, many of whom were dualist Śaivas.This prompted the Śaiva "Autonomists" to come to the rescue of their apparently hapless, philosophically lackluster co-religionists, bolstering the latter's dualistic realism with what contemporary philosophy would regard as a type of objective idealism.At stake in this objective idealist vision, among other things, is the status of relations, which for Utpaladeva are a very subtle type of śakti. 1 The argument of the present work concerns the centrality of śaktis in Utpaladeva's answer to the above philosophical questions.I intend to outline how they form an indispensable aspect of his overall vision as well as his treatment of relations specifically.I wish to recommend also that Utpaladeva's as-it-were essay "Proof of Relation" be read in terms of his intent (nowhere stated explicitly in the text, but carrying over from the Verses, and strongly implied in the benedictory lines of "Proof of Relation") to expound upon the most subtle stratum of śaktis discernable.The "Proof of Relation", I suggest, thereby picks up more or less where Utpaladeva left off in some of the main conclusions drawn by his much-cited as-it-were tenure-book, the Verses on the Recognition of the Lord ( Īśvarapratyabhijñākārikās) (hereafter, Verses).As is perhaps now well known, Utpaladeva displays a strong influence from the Grammarians (mostly Bhartr .hari) in his Verses.As has been shown more recently, much of Utpaladeva's innovation consists in leading his textual tradition in a 180-degree pivot in its attitude toward the Grammarians, generally embracing their philosophy where his guru, Somānanda, repudiated them. 2 Not coincidentally, in writing a short treatise on relations, Utpaladeva is explicitly referencing a "Sambandha-text" that precedes Dharmakīrti: the "Exposition of Relation" of Bhartr .hari, the third of many short works that together comprise the third volume (kān .d .a) of the Grammarian's [Treatise] On Words and Sentences (Vākyapadīya). 3 That is, Vākyapadīya, III.3.Śaktis are central to the way that Utpaladeva seeks to appropriate Bhartr .harian metaphysical categories.In doing so, he seeks to forge an innovative new path (navamārga) conducive to tantric (specifically mantramārgaic) 4 metaphysics, albeit one that claims the sanction of tradition: whether that tradition be Vākyapadīya's status as a non-sectarian (if Brahmanically-oriented) technical treatise (śāstra), or the Śaiva scripture (āgama) itself.There is a progression, then, from one commentary on relation to the next, from the "Exposition of Relation" of Bhartr .hari, to the "Examination of Relation" of Dharmakīrti, to the "Proof of Relation" of Utpaladeva. 5 The present study concerns the status (or in Dharmakīrti's case, non-status!) of the relation-śakti, in each of these texts.
The interpretive orientation of reading Utpaladeva's Verses in terms of its revealing Śiva by means of Śiva's śaktis is an emic one, central to the argument of the Verses itself 6 .Utpaladeva writes: The Ubiquitous Śiva (Nemec 2011) and "Evidence for Somānanda's Pantheism" (Nemec 2012)   3 This article follows the name "Vākyapadīya" to refer to all three of the kān .d .as not, by any means, because I wish to contravene Aklujkar's observation that the name historically refers only to the first two kān .d .as.I do so simply because "Vākyapadīya" is the more widely-recognized name.See, for example, "Number of Kārikās" (Aklujkar 1971) and "Bhartr .hari's Trikān .d .ī"  (Aklujkar 1970).4   For the scriptural distinction between the "Outer Way" (Atimārga) and "Way of Mantras" (Mantrāmarga) as articulated in the early tantra, the Niśvāsamukhasam .hita, see e.g., "Impact of Inscriptions" (Sanderson 2013).

5
My heartfelt thanks to Prof. Parimal Patil at Harvard, who first emphasized to me the importance of these three texts and their relationship to one another.Personal communication, 13 November 2015.For excellent partial treatments of the "Proof of Relation", see Pandey "Bhāskarī", (Pandey 1954) vol. 1, pp. xvii ff., and Allport, "Utpaladeva's Doctrine of Recognition", (Allport 1982), esp.pp.50, 219.The latter is newly available online after a period of relative obscurity.The other texts in the "Three Proofs" (Siddhitrayī) of which "Proof of Relation" is a part have received some recent attention.The "Proof that a Knower is Not Insentient" (Ajad .apramātr .siddhi) has been translated in full in Lawrence (1999), Sentient Knower.The "Proof of the Lord" ( Īśvarasiddhi) was the topic of Utpaladeva's Proof of God (Ratié 2016).6 I note that The Doctrine of Recognition (Kaw 1967) and Pratyabhijna Karika of Utpaladeva (Kaw 1975-1976) also emphasize Utpaladeva's defense of śaktis.
Verses, Knowledge Section (Jñānādhikārah .), 1.2-3. 7e reading I am proposing, then, ought not to be considered controversial.It may, moreover, be but one of many legitimate interpretive lenses through which to read the Verses.I emphasize it as a way of "thinking-with" Utpaladeva and to make plain in a sort of schematic sense what he understands himself to be defending.Of what sort are these śaktis-a term I have until now left, perhaps too cavalierly, without an English gloss?There is much room for clarification here, as the polyvalent "shakti" is by now an established part of the English language.The OED, for example, offers: "In Hindu religion, the female principle, esp.when personified as the wife of a god, as Durgā is the Sakti [sic] of Siva [sic] etc.; supernatural energy embodied in the principle." 8Śakti here no doubt retains some of that more general orthodox connotation, but it must be qualified as a term of art here denoting a heterodox tantric conception of śakti quite close to its etymological root: i.e., Śiva's powers, capabilities, capacities, or abilities (from the root, *śak).In accordance with a now widespread formulation, within Śaiva tantric philosophy as well as without, śaktis are by definition derivative or dependent.They are in relationship to their possessors in what might be called an adjectivally constituent manner-fire and its power to burn, for example.It is also worth emphasizing here that śaktis are by Utpaladeva conceived as fluidly continuous with Śiva and so are not entirely devoid of agency or self-awareness as they would be in a system with a sharp distinction between Śiva and Śakti-Yoga or Classical Sāmkhya, say.To be sure, Śaktis are nowhere in evidence as fully personal agential aspects of cognition or personal deities as they will later be in, say, the Light on the Tantras (Tantrāloka) of Abhinavagupta, which venerates Parā Devī (also in language Bhartr .hari would recognize) as the faculty of intuitive insight (pratibhā). 9In the Verses, śaktis are much more abstracted aspects of the knower and doer in Utpaladeva's opening verses (cited above).It is worth bearing in mind, however, that they are of the nature of agency and not entirely devoid of it.Hence the icchāśakti of Śiva's principal three powers (śaktitraya) might be glossed as the power of "will" (conveying intention, volition, and agency) as well as the more cognitively neutral "desire." The premise of the Verses is as follows, on Utpaladeva's own testimony.His Verses are meant to give a soteriologically catalytic outline of Śiva's saktis.It is by means of these various śaktis that recognition of Śiva is established (I.1.3)and the delusion or distraction (moha) militating against Śiva's otherwise self-evident omnipresence is conveyed, in service of humanity (I.1.1).Although Utpaladeva is bringing this understanding about, 10 he is not so much logically proving anything as illustrating what he considers to be established prior to logical demonstration or refutation by any "not-insensible" (ajad .a) person, even if the emancipatory consequence of that realization is not yet experientially available. 11In this way, the Verses are staged as a series of arguments deploying other textual traditions in order to prove the soundness of nondual Śaiva claims, but ultimately rest upon an illustration of the various śaktis by which Śiva is to be known.The aforementioned śaktis are presented in the Verses in sets of hierarchies from general to particular.At the most general level of the Verses' scheme of organization are the latter two of the śaktis in the famous Śaiva triad of will, knowledge, and action  (Bansat-Boudon and Tripathi 2011, p. 248 n. 1129), and is intended to reflect the ambiguity of jad .a/ajad .a as both active and passive-conveying whether or not an object possesses sentience, as well as whether or not it is capable of being perceived.Much of Pratyabhijñā theory develops around such distinctions, all of which contribute to Utpaladeva's choice of wording.8 I note, not without some amusement, the following comment in red letters on this rather Orientalist-inflected entry for "Shakti" from the online OED: "This entry has not yet been fully updated (first published in 1933)".9 The Goddess is praised in Tantrāloka, I.2 naumi citpratibhām .devīm .parām .bhairavayoginīm | mātr .mānaprameyām .śaś ūlāmbujakr .tāspadām || 10 (upapad-, I.1.1;IV.17; IV.18). 11Beyond logical demonstration or refutation because neither are well-formed concepts when Awareness itself is held to be the Supreme Principle, and the freedom to logically affirm or deny it (either of which freedoms are self-evident) are both equally evidence of its existence.Here are echoes of Anselm's famous proof of God, with the addition of offering Consciousness itself as, by definition, "that than which nothing greater can be thought".
(icchā, jñāna, and kriyā), to which the first two sections of the Verses are devoted (prior to the shift in register in the briefer final two sections concerning scripture and cosmology). 12Lawrence has pointed out that Utpaladeva's division of the text into Knowledge and Action categories roughly corresponds to the contemporary distinction between epistemology and ontology. 13The śakti of knowledge (jñāna), broadly speaking-the topic of the Knowledge Section-is further subdivided into the śaktis of memory (I.4), knowledge (I.5), and exclusion (I.6) (smaran .a, jñāna, and apohan .a) evoking the famous trio from the Bhagavad Gītā. 14Note that these two broad categories refer, in part, back to Utpaladeva's assertion in Jñānādhikārah .verse 1.2, that the one "knower" and "doer" is established from the beginning.In Utpaladeva's cosmology, action śaktis are, in an ultimate sense, nondifferent from those of knowledge, but are constitutive of the process of manifestation of differentiated gross objects.These are detailed in the Action Section as follows: unity-in-difference (bhedābheda) (II.2), perception (pramān .a) (II.3), and causation (kāryakāran .atā) (II.4).Within the above śaktis, there is the possibility of even further subdivision.Section II.2 is one with which we are especially concerned.
A treatment of unity-in-difference, it also comprises a defense of the reality of the range of typical realist categories, locating them as intelligible only in terms of Doctrine of Recognition or Śaiva Autonomist metaphysics.For Utpaladeva, such categories are derived or dependent realities and are thus (if you will) sub-śaktis or secondary śaktis, among which are numbered none other than relations (sam .bandhas) and genera (sāmānyas).
The reader is commended to Torella's excellent translation for the finer details of the above, necessarily summary account. 15For the purpose of this study, only the following need be borne in mind.The universal self, knower and doer that it is, is (for Utpaladeva) intimated by such things as relations and genera (for Utpaladeva these latter are equivalent to universals, though not for the Buddhists), both of which have as their foundation a metaphysics that is neither reducible to multiplicity or unity alone, but a continuous unity-in-multiplicity.Again and again as different possible relations are examined in the Verses, Utpaladeva maintains that such relations can only have as their foundation a perceiver engaged in a process of synthesis (anusam .dhāna).This is in contrast to the caricatured Buddhist opponents (p ūrvapaks .as).They are portrayed, based largely on Dharmakīrti's real position, as: (1) Taking their eliminativism on the offensive with respect to an enduring self-noting cases where there is purportedly no requirement for a perceiver to account for things such as memory, and placing a burden of proof on those who are realist with respect to a perceiver.And, (2) Turning against itself the realist definitional requirement of a relation that it rest on two coeval relata, thus acknowledging the reality of the cause-effect process (important in Buddhist doctrine), but denying its status as a relation, properly so-called (because its would-be relata are not coeval, but sequential).Where the Buddhists countenance processes of cause-effect, however, Utpaladeva exploits this opportunity to refer back to the agential perceiver as the only sensible foundation for any cause-effect.Utpaladeva does so using what he must have considered to be some of the most cross-sectarian tools available: those of logic and Sanskrit grammar.
12 Although it might also be glossed more neutrally as the power of "desire" or "volition," icchāśakti the most common gloss in the secondary literature, "will", reflects the fluid continuity between subtle and gross elements (tattvas) in this system, with a degree of agency inhering in each element. 13An early example occurs in Rediscovering God (Lawrence 1999, p. 54). 14The Kr .s .n .a of the Gītā in this understanding would of course be esoterically known as Śiva or Kālī.See, for example, Abhinavagupta's innovative interpretations in his Gītārthasam .graha, translated into English in Marjanovic (2002), Abhinavagupta's Commentary.See also Ratié (2011), le soi et l'autre pp.169 ff.If it seems odd to the reader that jñāna repeats within the Verses as its own subcategory or, if one likes, sub-śakti within its own śakti, I would argue that this is only Utpala's way of demonstrating that the cycle of knowledge (jñāna in the broad sense) consists of a cycle with three mutually-constitutive moments or movements: (1) memory (smaran .a), (2) the moment of knowing in dependence upon that memory (jñāna in the narrow sense), and (3) the discernment that this entails (apoha).Being united in the One perceiver that these latter three śaktis imply (because a śakti implies a possessor)-Maheśvara or Śiva-they are, in the ultimate sense, mutually entailing or without sequence, though they appear in constantly cycling sequence to the limited perceiver (pramātr .) or living being (jīva) (I.7.1;II.1.2). 15 The Īśvarapratyabhijñākārika of Utpaladeva (Torella 1994).
It should not seem terribly out of place, then, that amongst Utpaladeva's concluding statements to the Action Section, the reader finds grammatically-based ontological claims concerning relations: Based on the foregoing (ata eva), what is called the reality of action and its factors is properly the reality of connection (bhāva samanvayah .) [inhering in] case-ending meaning (vibhakti arthah .), with the unitary perceiver as its basis. 16rses, Action Section (Kriyādhikāra), 4.16 That is, actions such as cooking and so forth are expressed in an impartial and doctrinally neutral way in terms of (in this case, Sanskrit) grammar.What the Buddhists put forth as a process of cause and effect is expressible in terms of the action-factors (kārakas) that go into a process.On the theory of the Grammarians, as famously articulated by the As .t . ādhyāyī of Pān .ini, there are factors of action in sententially-based cognition that are logically prior to declensions and thus case-endings to words.When, in cognizing the process of, say, cooking, objects are arranged around that process-the perception of them all relates to the complex action "to cook".Objects thus fall into perceived categories of agent, object, instrument, patient, origin, and substratum.In an active-sentence construction the case endings-nominative, accusative, instrumental, dative, ablative, and locative-all have the above action-factors as their logical predecessors, the as yet unaccounted-for genitive case being something of a "wild card" and thus a case ending connoting relation quite directly.The relation of nominative case back to agent is not fixed, however, as in the case of a passive sentence construction, contributing to the rationale for seeing action-factors as logically prior to case-endings-in something of a syntactical structure ontologically preceding and underlying the formation of the sentence itself.In any case, what all of the case-endings here do is inflect words in such a way as to order them in webs of relations organized around the performance of a sequential action process.The one indispensable factor in the process, other than action itself, is, for the Grammarians, the agent (kartr .), which is logically prior to the grammatical subject.There are echoes here of Utpaladeva's assertion, at the beginning of his Verses, that the doer (kartari) is in no need of logical demonstration, being as it is presupposed in anything one might wish to express.Here, too, is a technical expression of the root motivation for the Śaiva philosophers' self-designation as "Autonomists" (svātantrya-vādins), the defining characteristic of the agent being that it is the sole action-factor that is not fully or partially dependent (pāratantrya), but rather fully and completely autonomous (svātantrya).
In his Verses, then, Utpaladeva lays out a number of śaktis by means of which God, as Consciousness, is experientially apprehended rather than presumed to be logically proven.The presence of the supreme lord as knower, doer, and agent is in various ways partially intimated by means of that lord's various śaktis, God being the supreme noun-form of which śaktis are various predicates, and thus being apprehensible in part by way of each.When it comes to Utpaladeva's grammatically-based conclusions, however, we have gotten deep into the woods.The fourth part of the action section is near the very conclusion of the polemical arc of the Verses.What exactly does the present work mean by the "lowest perceptible stratum" of śaktis mentioned above, and how does grammar play a role in that lowest stratum?By way of illustrating the response to this question, let us refer back to the chronologically first one of our three "Sambandha Texts"-Bhartr .hari's text, the "Exposition of Relation".In the following, I hope to show a bit how relations and universals factor into Bhartr .hari's exposition, and to briefly touch upon the Śaiva appropriation of this understanding.
Bhartr .hari's work shows, without a doubt, the sort of Brahmanical orientation one might expect during an age when Vedic interpretation was a vital preoccupation for any orthodox person engaged in Grammatical theory.It is certainly no Buddhist work, though it clearly displays an awareness of Buddhist scholarship.Even so, the Vākyapadīya strives for a cross-sectarian comprehensiveness that the very traditions it describes have been keen to exploit after the fact, further developing their orientations toward a philosophical problem along the lines of its description by Bhartr .hari himself, and by reference to Bhartr .hari more generally.In brief, Bhartr .hari has sought to be comprehensive, and the traditions he describes have cited him as an authority or even claimed him as a card-carrying member of their own tradition.Such a situation has led to descriptions of the Grammarian's "perspectivism" or "catholicity" in modern scholarship. 17Certain grammatical dogmas of Bhartr .hari are strongly in evidence and unopen to interpretation-for example, the assertion that the onset of meaning and permanent (nitya) identity of word (śabda) and meaning (artha) occurs solely in the unbroken sentence (akhan .d .a-vākya-sphot .a). Other, more philosophical commitments are less clear.For example, the Bhartr .harian influence on Śaiva doctrine is sometimes illustrated by way of a now-famous quote: Without proceeding from the word, there is no cognition (pratyayah .) in the world.All knowledge (jñānam .) is, as it were, intermixed with the word.If this eternal identity of knowledge and the word were reversed, knowledge (prakāśah .) would cease to be knowledge; it is this identity that makes identification (pratyavamarśinī) possible. 18e resonance with the syntactically-oriented Śaivas, as summarized above, is clear enough.But as B.K. Matilal pointed out some time ago with respect to these same verses, Bhartr .hari's thought can still be spun (at least?) two different ways: The above may imply the rather strong Bhartrhari thesis: (Bl) All cognitive episodes are equivalent to verbal thoughts.Or, it may imply the weaker claim: (B2) Most cognitive episodes with which we deal are invariably verbal thoughts at some implicit level. 19guably, it is the strong claim that is closer to Bhartr .hari's intent.But this does not stop Buddhist readers of Bhartr .hari from developing his theory along different lines.For them, a permanent relation between word and its meaning carries no special metaphysical or ontological significance, for such meaning inheres solely in the empirically evident fitness (yogyatā) of word for its meaning or vice-versa, meaning for its word.Thus, the Buddhists evince a preference for "fitness" as an exhaustive description of the word-meaning relationship, as opposed to the other term Bhartr .hari deploys for such a reciprocal word-meaning relationship: śakti.
What is the prima facie status of śaktis in Bhartr .hari's formulation of word and meaning?They would appear to be subtle realities in hierarchies of dependence, not unlike Utpaladeva's conception in his Verses.It is in the Vākyapadīya generally that a theory of relations and universals takes shape around the Grammarians' metaphysically loaded conception of the indivisibility of Sanskrit word (śabda) and meaning (artha), a sort of primary relationship, which in turn carries implications concerning the relationship between individuals (dravya) and universals (jāti), the particular (viśes .a) and the general (sāmānya).Just as there exists the more neutral "fitness" in relation to the much more ontologically loaded "śakti," so too the more neutral genera (sāmānya) is described in realist grammatical formulations in terms of a universal (jāti), which the Buddhists, naturally, reject, their nominalism favoring only the particular.Śaktis, in Bhartr .hari's "Sambandha-Text", appear in the following way.
The "Exposition of Relation" itself consists, in its first part, of an effort to home in on a precise understanding of relation (sam .bandha), which is something Bhartr .hari intends to acknowledge as an 17 The former designation is Jan Houben's and the latter, Radhika Herzberger's. 18My translation is indebted that of Ratié, le soi et l'autre, pp.161-162, which gives the transliterated Sanskrit as follows: na so 'sti pratyayo loke yah .śabdānugamād r .te/anuviddham iva jñānam .sarvam .śabdena gamyate//vāgr ūpatā ced utkrāmed avabodhasya śāśvatī/na prakāśah .prakāśeta sā hi pratyavamarśinī.Ratié also points out the textual differences between Rau's edition (which I was not yet able to consult at the time of this writing), vs the text as it appears in Abhinavagupta's shorter commentary on Utpala's Īśvarapratyabhijñākārikā-that is, the edition has bhāsate in place of gamyate. 19Word and the World (Matilal 1990, p. 128).
existent despite its subtle status as a highly derived or dependent entity, one that eludes signification, and that in everyday speech may be easily conflated with concepts that ought to be held distinct.Of interest to the present discussion, a relation is described as a śakti, which Houben glosses in English as capacity: Where this [relation] is, because some service is rendered [from one thing to another, or: from signifier to signified and vice versa], there one arrives at a property (viz.dependence) [but not at the relation itself].It is even a capacity (i.e., something dependent) of capacities [which are themselves dependent upon the entity which possesses the capacity]; it is even a quality (i.e., something dependent) of qualities [which are themselves dependent upon the entity which possesses the quality] [so it is extremely dependent]. 20(Vākyapadīya III.3.5)Despite its subtlety, relation is to be reliably detected in cases where one grammatical elementmetaphorically speaking-is in a disposition of "rendering service" (upakāra) to another grammatical element.Such grammatical considerations lead, in the second part of Bhartr .hari's chapter, to an extremely subtle meditation on the pivotal role of the relation as expressive of the nature of Brahman, which itself both encompasses and transcends discursive concepts such as being and nonbeing.Śakti makes a key appearance here once again, in Vākyapadīya III.3.87.Houben comments, "The term śakti 'capacity', as we have seen earlier, is usually something with an indeterminate ontological status . . . .On a cosmological level, it bridges the unity of ultimate reality and the diversity of creation." 21elation is, then, a quality-(gun .a)-of-qualities or a capacity-(śakti)-of-capacities.Śaktis are, for Bhartr .hari, compelling in their indeterminacy and yet the śakti of relation is for him an intimation of the indivisibility of unity and diversity.In relation to this, Houben cites also a chapter from Bhartr .hari's Chapter on Directions (Dik Samuddeśa) of his treatise On Words and Sentences.Houben offers a commentary and translation that is worth quoting at length.Houben writes: Towards the end of the chapter, the indeterminate character of capacities is pointed out: 3.6.24Having understood that the two postulations with regard to these capacities, namely that they are one and that they are many, do not follow the thing-as-it-is, one should not consider them to be in accordance with reality.

3.6.26
There is no unity without diversity, nor is the other (i.e., diversity) there without unity.In the ultimate reality, this division between the two is utterly absent.3.6.27For capacities there is no division in the same way as is the situation for [things] having capacities.Nor is the unity which figures in daily life found in their real nature.3.6.28Unity could not be established if diversity were not postulated; and diversity would be lost if unity were not postulated.The consequence of this indeterminate nature of capacities . . .for the way reality is dealt with in language, was pointed out in the one kārikā which we skipped above: 3.6.25With regard to objects whose reality is beyond speculations, the world is followed in its usages based upon conventions.
Houben concludes, "In other words, since the nature of things cannot be determined in a universally acceptable way, nothing can be done but to adopt the established usages current in daily life." 2220 upakārāt sa yatrāsti dharmas tatrānugamyate / śaktīnām api sā śaktir gun .ānām apy asau gun .ah .Houben's above explication of Bhartr .hari's reasoning with respect to śaktis anticipates that of Utpaladevas's Verses II.2, where the latter describes relations in terms quite a lot like, to use Bhartr .hari's phrase, "śakti-of-śaktis": that is, in dependence upon unity-in-diversity, itself dependent upon action.Moreover, relations and genera are understood as being in close interaction in the cognitive process-if not for Bhartr .hati himself, then certainly for his commentator on the third kān .d .a, Helārāja.For example, Helārāja adjudicates between the substantialist and universalist camps of the Grammarians by speaking in terms of grammatical particles. 23The genus to which a word refers is given by its stem.Genera are then made specific through the inflectional endings of their declensions, which are themselves capacities.A given declined noun, then, represents the intersection of genera and śaktis in webs of relations with other words in the sentence.It is this level of cognition-the level of the individual word-where genera and śaktis intersect.The nodal points, as it were, of the genus indicated by a given noun-stem are, as it were, linked together in this-way or that-way depending upon the inflectional endings.For Bhartr .hari, however, for whom the sentence is the most real level of language in which word-and-meaning relationships are permanent (nitya), there is a realness with respect to the power a sentence has in expressing webs of relations.The divisions of a sentence into words and word particles, for their part, represent only an artificial, analytical division.Utpaladeva, then, makes the case that reality is intelligible not solely in terms of unity or division, but as a continuum of both (bhedābheda).This ontological unity-in-diversity is visible at the level of the complete sentence.It is in this way that, for Utpaladeva, the metaphysical truth of unity-in-diversity is syntactically intelligible at the level of the sentence, of which śaktis, genera, and relations are derived and extremely subtle realities, though not insubstantial, due to their causal efficacy (arthakriyā) in faithfully patterning cognition after manifest phenomena.In offering this linguistically-informed metaphysical account of Bhartr .hari, Utpaladeva pushes back against a nominalist interpretation of Bhartr .thari by the Buddhist epistemologists, for whom genera are not to be understood as universals at all.Such Buddhist critique is evident in the second of the "Sambandha texts."Whereas for Bhartr .hari the unbroken sentence is the only meaningful unit-of which words and particles are derivatives-in his "Examination of Relation", Dharmakīrti takes a given object of manifestation as the primary reality.A given object's attributes and its classification alike are not just derived realities but artificial ones, in the domain of constructed, binary thinking (vikalpa).Thus, the "mixing" or grouping of objects with one another according to category is not due to some substantial universal inhering in those grouped objects, but due to a process of exclusion (apoha) whereby particulars (in contrast to inexistent genera) are artificially grouped together.So too, in inherence relations such as "blue lotus", it is not the case that there is some essence of "blueness" that is separate from the lotus itself.Dharmakīrti offers a critique of different ways of conceptualizing relation that is meant to be exhaustive.Relation is not to be found in either of its two possible meanings: not in (1) the temporal sense, in cause-and-effect and words indicating as much, such as "expectation" (apeks .a) or "dependence," (paratantrya), nor is it to be found in (2) the structural sense, in "mergence" (r ūpaśles .a).For Dharmakīrti, following an analysis based on the Buddhist emphasis on momentariness (ks .an .ika), cause-effect expressions such as "dependence" are self-evidently mere conventions (sam .keta) or figures-of-speech (v.15).It is nonsensical, colloquial, anthropomorphizing talk to suggest that a seed "expects" a sprout, or sprout "depends upon" a seed.But it is likewise foolish to consecrate some ontological status to so-called relation in structural terms when, in a given moment, two supposed relata are either (1) separate and distinct, thus actually unrelated by virtue of not yet having merged; or (2) altogether unitary, thus unrelated by violating the definition of relation wherein relation must obtain between two relata, not one only.That is, something cannot be self-related, nor is there any relation between not-yet-merged "relata" apart from what the imagination attributes to them.
It is this emphasis on the object of manifestation, and not its classification or attributes, that motivates Dharmakīrti to say: Sense objects are in themselves unmixed/ It is the imagination that connects them. 24shing to defend the substantiality of relations broadly-between word and meaning; general and particular; a particular and its attribute-Utpaladeva's thought parts with Dharmakīrti by this, only the fifth verse of the latter's "Examination of Relation".They were, however, largely in agreement in Dharmakīrti's first verse.This first verse of Dharmakīrti's dismisses "dependence" as an appropriate gloss for relation, on the grounds that, as already mentioned above, cause-effect processes cannot be relations properly-so-called.In this way, both Dharmakīrti and Utpaladeva reject the language of cause and effect as expressive of a relation.But, whereas Dharmakīrti goes on to critique "mergence of form" (r ūpaśles .a), Utpaladeva defends it, on grammatical grounds.
In this way, Utpaladeva in his "Proof of Relation" is partially in agreement with Dharmakīrti, at least concerning the unsuitability of concepts like "expectation" or "dependence" as synonyms for relation.This type of tactic of course echoes Utpaladeva's strategy in the Verses of agreeing with his Buddhist interlocutor up to a certain point.But, more significantly, I would argue, it shows Utpaladeva challenging the Buddhist Epistemologist's notions of linguistic convention (sam .keta)-notions that were, by then, well-developed in Buddhist discourse-by appealing to Bhartr .hari's earlier assertions concerning linguistic convention in the context of the subtlety of śaktis-namely, that figurative speaking proliferates in situations where there is some subtle reality that eludes speculation.In this case, precisely because relation (sam .bandha) is for Utpala a highly dependent śakti, it would seem natural according to Bhartr .hari that figurative speech and, thus, imprecise terminology arise with respect to it.Utpaladeva is thus in agreement with Dharmakīrti around terms like expectation and dependence, but parts from Dharmakīrti by locating a real and uncontrived relation in the other option: "mergence".We both have read Bhartr .hari-Utpaladeva seems to tell his Buddhist opponent-so we both know subtle realities elude description, but that doesn't make them unreal.
In sum, relations are, by all accounts, subtle realities.For Dharmakīrti, such subtlety is evidence of an insubstantiality and evanescence.Relation is nothing other than the contrivance (kalpanā) of an impersonal thought-continuum that constantly superimposes relations on objects where there really are none.Utpaladeva, by contrast, builds part of his case around that very subtlety of the relation-śakti as articulated by Bhartr .hari as evidence of its real, highly-derived, status.It is a very fine level of granularity at which to engage the śaktis that are experiential evidence of Śiva's existence.The relationśakti is, however, due to its subtlety, of such rarefied and subtle nature as to be easily mistaken for, and conflated with, the unreal.Packing into his benedictory verses of the "Proof of Relation" much of his metaphysical arguments that are to come, Utpaladeva then declares: We praise that One-by whose will is the creation of the ordinary world, accomplishing the ends of all 25 as relation, whose essence is difference-nondifference-Shiva! (1) 24 That a given sense object is neither-reducible-nor-otherwise to its qualities or categories, however, places its substantiality under suspicion, for Dharmakīrti.See the account of Dharmakīrti's "neither-one-nor-many" strategy in his "Examination of Relation" in Foundations of Dharmakīrti's Philosophy (Dunne 2004) p. 43 ff. 25 ityamiśrā svayam .bhāvāstānyojayati kalpanā in "Examination of Relation" v. 5cd.My translation is indebted to V.N.Jah, Philosophy of Relations, pp.12-13.A related passage from the Svārthānumāna section of Pramān .avārttika of Dharmakīrti, also quoted in the "Proof of Relation," is translated in "Nyāyamañjarī" (Watson and Kataoka 2017) p. 57 n. 117.See also the critique of "distributed entities (such as a "universal" or "whole")" in Foundations of Dharmakīrti's Philosophy (Dunne 2004) passim, esp.p. 104 and 95 ff.
That awareness of the Supreme Lord 26 excels, whose nature, though it is One, marvelously unfolding, is essentially all things [as] Relation, [which is] Being, Difference, etc. (2). 27 Śiva is here praised, as in the Verses, as the doer.Here, explicitly, as the creator of the world.
Accomplishing the ends or goals of all (that is, of all devotees) is surely one meaning of saha sarva-artha sādhitā though, given the context, it is also true that Śiva accomplishes what has meaning or semantic intelligibility.How is this done?That meaning (artha) is implicated in various ontological relations visible within corresponding grammatical relations, the most primary of which is the relation of word and meaning.Relations are, as we have seen, one of the sub-śaktis that are in turn derived from the śakti of difference-nondifference.Moreover, that doer, as was also said in the Verses, is the supreme lord and, thus, supreme agent.As awareness (prakāśah .) itself, it possesses a unitary nature, though it also embodies a manifest world that fluidly encompasses sharp distinctions, apparent solidity.How is this done?Again, as relation which plays an important role in the śakti of expansion (unmes .a) of Śiva's form (vapus .a)-that is, as all things.
Having here outlined in the most general terms what I hope has been a suggestive argument concerning the value of reading-with-Utpaladeva his exposition of śaktis, I note a few of the central concerns of my forthcoming translation of this text, currently under development.To begin with, there is the question of Bhartr .hari's prima facie outlook vs the Śaiva updating and appropriation of it.If, as Matilal has characterized it, the Śaiva philosophical approach to grammar represents a "Linguistic-superimposition" outlook, as opposed to Bhartr .hari's "Linguistic-nondualism", 28 what distinctions are to be made diachronically between Utpaladeva, Helārāja, and Abhinavagupta in the development of such a theory of superimposition?What are their important differences?Equally compelling, and at least as intricate, is the question of Dharmakīrti's Treatise on Valid Cognition (Pramān .avarttika), Utpaladeva's direct citations of which, though only two, are far from superficial and would lead the reader to believe that Utpaladeva has gained significant purchase on portions of this difficult text.Given Utpaladeva's purported depth of reading of Dharmakīrti, could it be said that Utpaladeva represents the perspective of his opponent fairly?How does Utpaladeva's reading of Dharmakīrti square with contemporary readings of Dharmakīrti, which by turns defend or critique an understanding of his ontology in terms of a "sliding scale" of analysis? 29In brief, does Utpaladeva attribute to Dharmakīrti a perspective that the latter philosopher would understand and recognize?And, finally, if taken seriously in the context of modern philosophy and not as just a historical artifact-how might the dialogue between Utpaladeva and Dharmakīrti be supplemented by a consideration of the pivotal 20th century dispute between Russell and Bradley, much of which concerned the status of metaphysical relations as well? 30o conclude the present work, then, it should not pass without mention that to read Utpaladeva's "Proof of Relation" in terms of one of his own broader intellectual horizons-the defense of śaktis-is to follow an emic line of thinking that sympathetically inhabits, to the degree that it is possible for a modern scholar, Utpaladeva's overarching motivations.It is intended as one step toward a strong-objectivity.That is, an objectivity that seeks to sympathetically account for a number of different perspectives.It is, however, one step only.Reading-with-Utpaladeva here represents an important stage in a longer process, not an attempt to valorize or privilege Utpaladeva's work at the expense of his interlocutors.Of vital importance, also, are Bhartr .hari and Dharmakīrti's respective motivations in their sam .bandha-works.These have here been accounted for only in passing, but are important topics for the next phase of this study, still to come.
Acknowledgments: An annotated English translation and philosophical exploration of the "Proof of Relation" is the topic of my forthcoming dissertation.I wish to thank John Nemec for his active interest and support for this topic when I first presented it to him in Spring 2015 during his supervision of my review of literature.John's example may be seen at every level of this work in terms of content and methodology, as his writing helped shaped my direction even prior to our brief work together at UVa.In going against the grain of North American Indology by seeking a short text to translate in its entirety, I was inspired by Nemec (2009), "Translation."I also wish to thank my dissertation advisor at CIIS, Jim Ryan, and my svādhyāya instructor at the American Institute of Indian Studies (AIIS) in Pune, Jyotsna Khare, each of whom read with me the grammar of the Sambandhasiddhi and offered their own invaluable insights.Dr. Khare also expressed enthusiasm for my wish to produce an English translation of the text, for which I am grateful.Much of the very impetus for this project, including its consideration of the importance of śaktis, comes from a long-standing dialogue with Mark Dyczkowski, whose guidance has been incalculably valuable.In the early stages of this work, Sonam Kachru was a tremendous resource and conversation partner.Last but certainly not least, I am indebted to an anonymous editor without whose comments the work in its present form would not have been possible.This work bears the imprint of each of these mentors.Mistakes are my own.

Conflicts of Interest:
The author declares no conflict of interest.