Balancing the Scales: Dharmakīrti Inside and Out

Recent years have witnessed much debate about the best way to make sense of Dharmakīrti’s apparent diversity of philosophical positions regarding the status of mind-independent objects: at some times he appears to argue in ways that presuppose the existence of such objects, while at others he argues against them. Several approaches have been proposed to account for this apparent contradiction in Dharmakīrti’s various statements regarding ontology. Most famously John Dunne (2004), building on earlier interpretive strategies suggested by Georges Dreyfus (1997) and Sara McClintock (2003), has suggested that we should see such positions as ranged along a “sliding scale” of hierarchically arranged stances, in which “more accurate descriptions of what we perceive and think supersede less accurate ones.” He finds in the Pramāṇavārttika four such levels of analysis: “The Views of Ordinary Persons,” “the Abhidharma Typology,” “External Realism” and “Epistemic Idealism.” Each of these levels is said to give way to the next through a specific kind of “mereological” transition argument, which shows that certain entities accepted as real on a given ontological level cannot be accounted for either as unitary or as multiform–as “one or many”–forcing one to abandon this ontology and ascend to the next, higher level of analysis.

Dunne’s views have provoked much discussion and several significant critiques. I will mention here specifically only Kellner (2011) Arnold (2008). Kellner challenges the uniformity of the “mereological” model for ascending the levels, among others discussing a specific idealist argument of Dharmakīrti, showing that there is no way to satisfactorily distinguish definitionally between the supposedly external object of a cognition and the immediately preceding cognition (samanantara-pratyaya) that gives rise to it, since both are causes of the awareness, and resemble it in form. Arnold questions whether the third, external realist level is philosophically viable, even as a theoretical alternative to the “epistemic idealist” level, to which he believes it inevitably reduces. But neither of them, or any other critic of the sliding scale so far as I know, questions the hierarchy itself–in particular the idea that the idealist or Yogācāra position set forth at certain points in the Pramāṇavārttika and Pramāṇaviniścaya is meant to be presented as clearly and unambiguously preferable to the external realist, “Sautrāntika” or bāhyārthavāda level. It is this widespread assumption I wish here to call into question.

On Dunne’s view, the external realist or “Sautrāntika” arguments Dharmakīrti often relies upon are there only as preliminary, conditional positions—as stepping stones which invariably give way through a specific kind of “transition argument” to the more accurate and more soteriologically beneficial idealist or Yogācāra position which represents Dhar-makīrti’s real view. Dunne characterizes the interpretive strategy that leads him to this conclusion as one of a “hermeneutics of charity.”^[1]

As Dunne explains this,

If… a well articulated passage or position clearly contradicts some other, equally well articulated passage or position, then the general principle of hermeneutical charity we have adopted will prompt us to rank these positions along a hierarchical scale; the lower position, while contextually expedient, is superseded by the higher position that, by virtue of passing a rational test that the lower position fails, conforms more closely to Dharmakīrti’s version of ultimate truth (Dunne 2004: 239, n. 24).

But should we take it as a given that there can be no reason why a philosophical author might choose to present multiple, genuinely alternative points of view, or alternative explanations of the phenomena he is attempting to account for, without falling prey to charges of incoherence? It is not in fact difficult to imagine motives that might prompt such an approach. For example, Vincent Eltschinger (2010, 2014) has recently stressed the importance of the Buddhist epistemological tradition as foregrounding for apologetic purposes a “supersectarian” identity “aimed at defending Buddhism as a whole against Brahmanical hostility” (2010: 399; see also Eltschinger 2014: 174ff.). Taking this into account, one can easily see how it might have seemed desirable to formulate a broad-based, “big tent” defense of Buddhism which could accommodate defenders of somewhat different Buddhist positions within a largely common epistemological framework. I will later mention one more way (apart from the “sliding scale”) in which we might make sense of Dharmakīrti’s apparent accommodation of multiple ontological stances.

At this point I would like, instead, to point out one obvious counterexample to Dunne’s line of thinking on this question—one which has not, so far as I know, been discussed in connection with the vexed question of Dharmakīrti’s ontology and his argumentative method and goals. I am thinking of Dharmakīrti’s own primer on Buddhist logic and epistemology, the Nyāyabindu. It has long been a commonplace among commentators and subcommentators on this work (stretching back at least to Vinītadeva in the 8^th century) to view the Nyāyabindu as a work designed to conform to both a realist and an idealist ontology (usually labelled as “Sautrāntika” and “Yogācāra,” respectively).^[2] Certain commentators observe that, at a few specific points in the text, these two contrary ontological perspectives necessarily diverge, and the question is sometimes raised as to whether some point advanced in the text is incompatible with either the Yogācāra or the Sautrāntika positions. But the idea that the text on the whole is meant to accommodate both ontologies is widely shared, and nowhere, I think, seriously disputed.

What I would like to suggest, simply put, is that both of Dharmakīrti’s major systemic treatises on Buddhist logic and epistemology, the Pramāṇavārttika and the Pramāṇavini-ścaya, are, like the Nyāyabindu, designed to accommodate and to accord, as seamlessly as possible, both with an idealist, Yogācāra, “mind-only” ontology and with that specific variety of external realism which later doxographers designate as “Sautrāntika”–what is sometimes known as the anumeya-bāhyārthavāda, the “inferred external object” view. There may well be–indeed there certainly is–reason to believe that, between these two views, Dharmakīrti regarded the Yogācāra position as philosophically (and perhaps soteriologically) preferable, but, as we shall see, he nevertheless goes out of his way on multiple occasions to present these two positions as viable alternative philosophical paths. Once we abandon the presupposition that discerning Dharmakīrti’s own preferred position on the question of objects and discerning what he intends to argue for in his works are necessarily equivalent, it becomes much easier to account for certain aspects of the way he juxtaposes the vijñānavāda and bāhyārthavāda positions at several key points in the Pramāṇavārttika and the Pramāṇaviniścaya. One important reason it is difficult to read Dharmakīrti’s key statements of the vijñānavāda position as an unambiguous ontological step up from the bāhyārthavāda position is that it is often after having developed what seem to be, and have generally been accepted as, his definitive vijñānavādin arguments that Dharmakīrti advances his most fully developed bāhyārthavādin arguments. Indeed, it seems to be Dharmakīrti’s regular, and perhaps even invariable, practice (as I will demon-strate), to follow his fullest and seemingly most conclusive arguments for the vijñānavādin position with a corresponding and parallel argument defending a specific version of the bāhyārthavādin position. It is difficult to see why he would adopt this approach if his ultimate philosophical and soteriological strategy were designed to bring his readers up to the highest, idealist, level and leave them there.

Pramāṇaviniścaya 1.58

The most striking passage of this kind, and the one that seems to have garnered the most attention from later chroniclers and critics of the Sautrāntika position, is one for which the original Sanskrit text has only recently become fully available, and has not yet, I think, received the attention it deserves–that is, Pramāṇaviniścaya 1.58 (the final kārikā in the chapter on perception), along with its accompanying vṛtti. The passage occurs just after Dharmakīrti has presented his famous sahopalambha-niyama (“necessary co-apprehension”) argument–that “there is no distinction between blue and the awareness of blue, because they are necessarily co-apprehended.”^[3]

Having advanced this claim as part of his defense of the vijñānavāda position, Dharmakīrti now turns to consider its relevance to the bāhyārthavādin position:

Because of this, even if there is an external object, there is no difference between the appearing object and the awareness of it. (1.58ab)

Even if an external object does exist, the non-differentiation between the appearing object (“blue,” or the like) and the awareness of it is established by the co-apprehension and co-awareness of the two.

Therefore, the awareness has a double form.

Therefore it is established that awareness has a twofold form, since it bears the form of both object and awareness. [Pūrvapakṣin:] If the appearing object is not differentiated from the awareness, then how can one think that it is something different?

There can be proof of an external object on the basis of disjunction [vya-tireka]. (1.58cd)

The fact that awareness [sometimes] does not arise when all other causes capable of producing it exist indicates the lack of some other cause; this could be the external object.^[4]

Even if we accept that all that is phenomenally apparent to us in any awareness is the awareness’s own form, we may still be able to infer an external object as a cause of what appears to us. We do not, for example, see “blue” at all times and, in the absence of any evident factor in our immediately prior awareness that determines whether we do or not see blue at any particular moment, we may legitimately infer an additional, extra-mental cause for the occasional appearance of, e.g., blue in our visual awareness. This is clearly presented as a positive argument in favor of the bāhyārthavādin position, and has invariably been interpreted as such by later authors, both Buddhist and non-Buddhist (for whom Pramāṇaviniścaya 1.58 and its attendant vṛtti become, indeed, the classic statement of the Buddhist bāhyārthavādin or Sautrāntika position).^[5]

It is hard to see why, if the defense of bāhyārthavāda (conditional or otherwise) is meant to be a stepping stone on the path to Dharmakīrti’s ultimate stance of vijñānavāda, he should choose, after directly presenting his most fully developed argument in defense of this vijñānavāda, to then deliberately step backwards, or downwards, to the “lower,” supposedly philosophically less tenable and soteriologically less useful level of external realism. Moreover, even if we were able to find some motive for this return to defense of the bāhyārthavādin position after the supposed transition to “epistemic idealism” has already been effected, there is no way what we see in Pramāṇaviniścaya 1.58 can be read as a step backward to a more primitive, more naive, or less fully developed position than the vijñānavādin view that precedes it. On the contrary, it is presented explicitly as a sequel to, and an outgrowth of, the purportedly vijñānavādin analysis that immediately precedes it. It builds on, and entirely presupposes, the dvairūpya of perceptual cognitions that Dharmakīrti has just argued for, and the validity of the sahopalambhaniyama argument that formed its centerpiece. Indeed, in the light of the way Dharmakīrti constructs the bāhyārthavādin argument in 1.58, and the way he connects it with the preceding argument, one must in fact recognize that the sahopalambhaniyama argument, despite the way it has so often been characterized by both premodern and modern expositors of Dharmakīrti’s thought, is in fact not an idealist argument at all. That is to say, it is fully consistent with the final versions of both the vijñānavādin and bāhyārthavādin positions as Dharmakīrti develops them in both the Pramāṇaviniścaya 1.58 and the parallel passage of the Pramāṇavārttika discussed below. Rather than an argument for the non-reality of mind independent objects, the sahopalambhaniyama argument is designed to establish only that the phenomenal content of perception, its grasping and grasped aspects, are artificially abstracted parts of a single cognitive image. Establishing this is a necessary prerequisite for defending a vijñānavādin position, but does not require one to adopt such a position, as Dharmakīrti’s own discussion in Pramāṇaviniścaya 1.58 explicitly shows.^[6]

While the final kārikā of the Pramāṇaviniścaya chapter on perception thus presents us with a viable avenue for the defense of the existence of extra-mental objects, Dharmakīrti does not leave us with a definitive endorsement of this position over the vijñānavādin stance developed earlier.

This new sahopalambhaniyama-consistent argument for extra-mental objects is immediately followed by a further remark that rehabilitates the vijñānavādin posi-tion but, again, as a viable alternative path, not as a decisive victor over the bāhyārthavādin position:

Unless one can say that this disjunction of effect is produced by the absence of a specific material cause [upādāna, i.e. the samanantara-pratyaya].^[7]

The idealist might try to account for the occasional appearance of images such as blue in our awareness by positing, instead of an extra-mental object, a specific (but unperceived) factor in the immediately prior awareness (samanantara-pratyaya –which is the material cause of any given awareness-moment) that causes us to see blue on some occasions and not others. So, we are left with two possible ways of explaining the occasional appearance of specific images in our perceptual awareness, both of which require us to postulate something not itself phenomenally evident–either a putative extra-mental object, or an unperceived causal factor in our prior awareness–whose presence or absence accounts for our perception containing a given image at some times but not others. But there is, notably, no explicit indication of whether either of these views is to be preferred over the other, nor even the suggestion of an argument or a criterion that would enable one to choose between them. Far from attempting to compel us to move “up” from the external realist to the epistemic idealist level, Dharmakīrti seems almost to be going out of his way to display ambivalence: there “can be proof” of the external object, “unless someone were able to say” that the disjunction in effect is produced by a difference in the samanantara-pratyaya. He does not say explicitly whether the object is proven in this way, or whether the vijñānavādin is or is not able to explain the occasional appearance of the effect in this way. We are simply left with two ways of accounting causally for the appearance of specific images in our perceptual awareness, with no indication of how (or whether) we should decide between them.

Having set forth this pair of parallel accounts, the Pratyakṣa chapter of the Pramā-ṇaviniścaya concludes with one further objection against and response in support of the vijñānavādin view:

[Objection:] But how then could this person [i.e. the vijñānavādin], while denying objects for all awarenesses, nevertheless say that a confused aware-ness is not a pramāṇa while an unconfused one is, since there would be no difference between them? [Reply:] Seeing that, even for someone who is not fully awakened, activity sometimes proves to be untrustworthy, due to the fault of being connected with karmic traces of confusion, one can say that one sort of awareness is not a pramāṇa; seeing that, in other cases, this activity has dependable consequences which persist as long as saṃsāra does, being based on firmly established karmic traces, one can say that this other sort of awareness is a pramāṇa, in consideration of its non-disparity with [other] worldly activity. What is stated here is the form of conventional pramāṇas. Others are confused even about this, and draw people into disagreements. Those who devote themselves to the wisdom consisting in thought (cintāmayī prajñā), however, turn their attention toward the ultimate pramāṇa, which is faultless through the discrimination of error and free from decay. This too is slightly hinted at.^[8]

This explanation of how pramāṇa and non-pramāṇa can be distinguished in an object-free world does redress the opponent’s objection, and rehabilitates vijñānavāda, but only as a possibility, not as a position to which we are compelled. It leaves in place the unresolved ambiguity of Dharmakīrti’s prior treatment of the two as alternatives. Furthermore, it rehabilitates the idealist position even as a possible stance only by bracketing both the vijñānavādin and bāhyārthavādin positions as less than the ultimate truth. Again, at least within the realm of “conventional pramāṇas,” we are left not with a decisive argument that forces us to move from an external realist to an idealist stance (or vice versa); these two basic ontological stances are both left on the table as Dharmakīrti ends his discussion of perception.

Pramāṇavārttika, Pratyakṣa 391–398

The basic features of this ambiguous back and forth between the vijñānavādin and bāhyā-rthavādin positions in PVin 1.58 and its accompanying vṛtti are all in fact prefigured in what is clearly the parallel passage in the Perception chapter of the Pramāṇavārttika:

Therefore it is unavoidable that the object, which appears at the same time as the awareness, is not distinguished from the awareness. There could be an inference of a difference in the cause [of one’s awareness], due to the absence of [specific] perceptual awarenesses even when all the other causes are present. Unless one can state a restriction based on the immediately preceding awareness.^[9]

Here too we have, immediately following an argument for the nondifferentiation of the awareness from its object, the claim that one may nevertheless infer an extra-mental object as a cause for the occasional occurrence of specific images in our awareness. And here again we have the possibility of an unrecognized causal factor in our immediately prior awareness suggested as an alternate explanation of these occasional images. The phrasing is extremely close to that of Pramāṇaviniścaya 1.58, and again we have the same juxtaposed optative construction: “There could be an inference… unless one can say…,” leaving us with a similar ambiguity on the external object question. But there is one significant difference between this passage and the (later) vṛtti on Pramāṇaviniścaya 1.58. There Dharmakīrti presented the vijñānavādin and bāhyārthavādin as hypothetical alternative stances without ultimately indicating a preference for one over the other.

Here, on the other hand, Dharmakīrti does at least indirectly indicate such a preference (PV, Pratyakṣa 398):^[10]

This [vijñānamātra view] is the doctrine of the wise; but this dual nature of awareness [as grāhya and grāhaka] is [also] established in reliance on an external object, due to the rule of co-awareness.^[11]

Labelling the vijñānavādin view as “the doctrine of the wise” does certainly suggest that Dharmakīrti considers this view to be preferable (whether philosophically or soteriologi-cally). But it is important to note that this offhand indication of preference does not come coupled with any sort of argument that would justify it. Here too, as in Pramāṇaviniścaya 1.58, we are left with two philosophically viable accounts of perceptual images, one which posits the reality of extra-mental objects, and one which denies them, but with no clear argument for preferring one over the other.

If, then, as it seems, Dharmakīrti does not wish to present any compelling argument that forces his readers to abandon the external realist position in favor of an idealist one, what is he trying to do here? If, as the label “doctrine of the wise” would at least seem to suggest, the vijñānavādin position is Dharmakīrti’s preferred view, why should he present it and the parallel bāhyārthavādin account in such careful equipoise (as he later does, without a similar indication of preference, in the Pramāṇaviniścaya)? Both available Sanskrit commentaries on this passage of the Pramāṇavārttika address the question directly.

Prajñākaragupta comments on Pratyakṣa 398 as follows:

[Pūrvapakṣin:] Why then did the Teacher [Dignāga] explain that [the aware-ness has] a dual form even if the external object exists? Vijñānavāda alone is the correct position. In response to this, he [Dharmakīrti] says:

This [vijñānamātra view] is the doctrine of the wise; but this dual nature of awareness [as grāhya and grāhaka] is [also] established in reliance on an external object, due to the rule of co-awareness.

The dual form [of awareness] is explained with provisional acceptance of the external object, in consideration of others [who hold opposing views]. But wise people accept vijñānavāda alone.^[12]

Manorathanandin, whose comment on the verse closely mirrors that of Prajñākaragupta, explains further:

This is the doctrine “of the wise,” i.e. the Yogācārins who understand reasoning, which explains that there is nothing but consciousness in all conditions. The teacher [Dignāga] has [also] explained the duality of form of awareness relying on the external object accepted by the Sautrāntikas. And this duality of form is established “due to the rule of co-awareness,” i.e. due to the rule of co-apprehension [sahopalambha-niyama] since, even if there is a difference [between the awareness and the object], this [rule of co-apprehension] still exists.^[13]

As Prajñākaragupta and Manorathanandin both rightly point out, Dignāga in his investiga-tion of the objects of perception in his Pramāṇasamuccaya (on which the Pramāṇavārttika is ostensibly a commentary) similarly offers parallel vijñānavādin and bāhyārthavādin accounts without indicating a preference for one over the other:

If the awareness together with its content is the object, then what one appre-hends is an object in the form of one’s own self-awareness, which may be either desirable or not. If, however, the thing to be known is an external object, then

The pramāṇa for this is simply that it has the form of the object.

For then, setting aside the [awareness’ own] form, even though this is what is cognized by the awareness, the pramāṇa for this [awareness] is simply that it has the form of the object, since that object

is known through this. 9 (emphasis mine)^[14]

The connection these commentators posit between Dharmakīrti’s practice and Dignāga’s thus makes sense,^[15] though it simply pushes the basic question back one step further: It is plausible to suggest that Dharmakīrti’s parallel presentation of the vijñānavāda and bāhyārthavāda as viable alternatives is meant to conform to Dignāga’s practice, but this does not explain why Dignāga himself would have adopted this approach in the first place. A basic gesture of inclusiveness towards others in the Buddhist camp is presumably intended by Prajñākaragupta’s “in consideration of others” (parāpekṣayā) and Manorathanandin’s reference to “the external object accepted by the Sautrāntikas.” In any case, both clearly acknowledge that both Dharmakīrti and Dignāga, while themselves personally inclined toward the vijñānavādin stance, are seeking in these passages to accommodate (and precisely thereby not to supersede) the bāhyārthavādin/Sautrāntika position.

Pramāṇavārttika, Pratyakṣa 223–224

I will close by briefly examining one further passage in which Dharmakīrti appears to balance the vijñānavādin and bāhyārthavādin positions in this way: verses 194–224 of the Pratyakṣa chapter of the Pramāṇavārttika. This passage deals with the problem of citratā –“multicoloredness,” or more generally “multiplicity” of any kind in a cognitive image. When we perceive a multicolored object, or any object that appears to have multiple properties of any kind (such as a spatially extended object that appears to exist at more than one point), how is it that we have an impression of the “oneness”? Does it exist only in our minds, or is there some way that external, atomic objects can be aggregated into what is in some sense a “single” object? This is an absolutely key passage for both Dunne’s and Dreyfus’s analyses, as both see it as marking Dharmakīrti’s crucial transition from the external realist to the idealist level of analysis (i.e. from the third to the fourth level in Dunne’s “sliding scale”).

It is with reference to this passage that Dunne says:

Finally, in moving from the third to the fourth level of analysis, the inability to specify whether the image in perception is single or multiple is the primary argument against the existence of extra-mental objects (Dunne 2004: 63).

In response to the intractable problem of the cognitive image’s variegated singularity, Dharmakīrti abandons External Realism in favor of Epistemic Idealism (Dunne 2004: 112).

Dreyfus similarly sees Dharmakīrti’s exploration of the problems posed by the multiformity of perceptual images as setting the stage for a definitive transition from external realism to idealism.

As he says:

This denial of the reality of external objects is where Dharmakīrti finds a solution to the dilemma created by an impression of a solid extended object produced by atoms, which do not have any extension by themselves. The problem is solved by rooting out its source, the assumption that objects exist external to consciousness as a result of atomic aggregation. The Yogācāra view that objects exist only as reflections of consciousness is Dharmakīrti’s answer to the problem created by extended objects (Dreyfus 1997: 103).

The problem with both Dunne’s and Dreyfus’s reading of this passage of the Pramāṇa-vārttika is that here too, as in Pramāṇaviniścaya 1.58, we find, as a coda to what they take to be Dharmakīrti’s conclusive argument driving the shift from the “external realist” to the “epistemic idealist” stance, a return to the defense of the external realist (or, at any rate, an external realist) position. Dharmakīrti’s response to the problem of citratā (Pramāṇavārttika, Pratyakṣa 209–224) is too complex to analyze in detail here. He offers in fact there three different ways to account for multiform appearance in awareness, apparently presented as parallel, alternative lines of argument: one which denies that multiformity can really exist either externally or within awareness (209–219), a second that accepts the existence of multiformity in cognitive images (ākāras) but denies that it can exist externally (220–222), and a third that, accepting the stricture that multiformity can exist only in awareness, nevertheless asserts that multiple, real external objects can give rise to a single, multiform awareness (223–224).^[16] Obviously it is the third position here that is of most interest for my argument, as it represents yet another retrospective rehabilitation of the external realist position.

The relevant verses run as follows:

On the other hand, what contradiction is there if many [atoms] which have this special causal capacity [collectively] do not separately cause the awareness, just as is the case with the self, the sense organ, and so on.^[17]

To be “grasped” is nothing other than to be a cause. When the awareness has the same appearance as something [that causes it], one says that this thing is grasped by the awareness.^[18]

That Dharmakīrti is here transitioning to an alternative line of argument is clearly marked in the text (“On the other hand”) and, as in Pramāṇaviniścaya 1.58, there is nothing to indicate whether this line of argument or the one that precedes it is to be preferred for any reason. On Dharmakīrti’s own explicit analysis, then, the problem of citratā explored in this passage does not compel us, or even forcefully nudge us, toward an idealist stance. He offers the reader answers to the citratā problem that work just as well whether one accepts external objects or rejects them. It is true that the alternative resolution of the citratā question presented in the final two verses of this passage only opens the door to bāhyārthavāda as a possibility; the claim is that multiple atoms can act as external causes for a single multicolored mental image, not that such external causes must exist. Nevertheless, taken seriously, these verses undo much of the supposed work done by the earlier portion of the passage, on both Dunne’s and Dreyfus’s accounts. The external realist position is certainly not definitively abandoned here, and the move from realism to idealism cannot reasonably be described as “Dharmakīrti’s answer” to the citratā problem, but (at best) one of several possible answers.

Conclusion

In the light of the treatment of the bāhyārthavādin position in all of the passages discussed above, it seems difficult to maintain that Dharmakīrti was seeking in his major epistemolog-ical works to push readers to ultimately adopt a vijñānavādin stance. Whatever his personal position on the status of extra-mental objects, he seems deliberately to have written in such a way that either vijñānavādin or bāhyārthavādin readers could comfortably adopt his epistemology. He in fact developed what came to be some of the most influential arguments in support of Buddhist bāhyārthavāda. This helps to explain why later doxographers, when seeking to delineate the “Sautrāntika” view, so often turned precisely to Dharmakīrti’s statements, and specifically to PVin 1.58, as the classic textual formulation of this posi-tion. The specific variety of “external realism” set forth in PVin 1.58 is not a preexisting philosophical position that Dharmakīrti (or Dignāga) adopted as a heuristic device or a stepping stone to draw readers along to his “real” idealist position; it is, rather, explicitly presented as an outgrowth of his own analysis of the phenomenology of perception. It is not only that one can be a “Sautrāntika” in the sense recognized by the later doxographers and still be a Dharmakīrtian. Rather, one can be a Sautrāntika, in the doxographers’ sense of this term, only if one is a Dharmakīrtian. That is, the externalist argument as set forth in Pramāṇaviniścaya 1.58 and, more briefly, in Pramāṇavārttika, Pratyakṣa 223–224 and 398 grows specifically out of the analysis of the dvirūpatā of all perceptual cognitions–the artificial division what is really a unitary cognitive image into a “grasping” and “grasped” form (grāhaka and grāhya ākāra)–and fully presupposes the validity of the arguments on which this analysis is based. That is, the epistemology and phenomenology of perception elaborated by both Dharmakīrti and Dignāga is entirely shared between the vijñānavādin and the bāhyārthavādin position as they elaborate them.

Arnold is entirely right to stress this in his discussion of Dunne’s “sliding scale.”^[19] Whether he is right in suggesting that such a phenomenology (which takes the only “object” directly accessible to awareness to be its own image) tends by its nature to force one into an idealist rather than a bāhyārthavādin position is less obvious. This is in fact precisely the question at issue between the bāhyārthavādin and the vijñānavādin as Dharmakīrti finally represents them: given that we have immediate cognitive access only to the ākāras of our own awareness, can we or can we not legitimately infer an extra-mental cause for the occasional appearance of particular ākāras? Obviously if Arnold were right in this claim, and if Dharmakīrti held a similar view of the question, this would tend to support Dunne’s model of the sliding scale, as it would leave us with a Sautrāntika position deliberately designed to collapse into a Yogācāra one. But, as I have already said, this makes Dharmakīrti’s regular return to defense of the bāhyārthavādin position after presenting his seemingly ultimate vijñānavādin arguments hard to explain.

So, the same general pattern recurs in Dharmakīrti’s works, seemingly whenever he presents his ultimate versions of the vijñānavāda and bāhyārthavāda positions in explicit contrast to one another. We first find an argument for how the phenomenology and episte-mology of perception function on the vijñānavāda view (in one case only, PV, Pratyakṣa 398, accompanied by a fairly clear suggestion that this is Dharmakīrti’s preferred view), fol-lowed by a defense of the bāhyārtha position, adopting precisely the same phenomenology of perception, but taking it to imply the existence of an extra-mental cause for the images that appear in our awareness.

Is Dharmakīrti, then, contradicting himself? Is there anything problematic in developing a set of views regarding perception that accommodates and accords with two conflicting ontologies? Must we, in charity, seek to forcibly extract from his works one definitive stance on the external object question, even when he himself seems to be trying carefully to avoid doing so? I think Dreyfus himself makes a very important observation, when considering the question of Dharmakīrti’s possible self-contradiction in his seemingly inconsistent position on bāhyārthavāda.

To quote:

Suffice to say that I take Dharmakīrti’s essential preoccupation to be episte-mology not metaphysics. His interest in defining the nature of reality is to ontologically ground his epistemology… The exact nature of real entities (whether, for example, they are ultimately mind independent or not) is a lesser concern (Dreyfus 1997: 105).

I think this quite nicely sums up and makes sense of Dharmakīrti’s method here. The real matter Dharmakīrti wishes to stress, and does stress, is precisely the account of the epistemology and phenomenology of perception that is common to his final versions of both the vijñānavādin and the bāhyārthavādin positions. This is what he argues for consistently, forcefully, and without ambiguity.

An interesting side-question here is whether, and if so how, Dharmakīrti sees the final bāhyārthavādin argument in PVin 1.58 as escaping the earlier samanantara-pratyaya argument already discussed by Kellner. She notes that this argument is directed against what she describes as “the theory that Dharmakīrti first adopts and then abandons: that external objects produce a perception which has their form (ākāra), or which resembles them” (Kellner 2011: 294–emphasis mine). The samanantara-pratyaya argument basically hinges on the problem of discriminating exactly what part of the causal complex producing an awareness would count as the “object.” If one defines the object as that which both causes an awareness and bears the same or a similar form, then one will in many cases have to consider the samanantara-pratyaya, the awareness immediately preceding the one in question, to be its object; if one has a continuous awareness of, say, blue, then any later moment of this stream will have as part of its causal complex its samanantara-pratyaya, which also contains a blue image, and hence will meet the test for being an “object.” The consequent inability to define an “object” in a way that distinguishes it from the samanantara-pratyaya is presented as a crippling flaw in this version of the external realist theory. In connection with this, it is interesting to note that the positive arguments for the bāhyārtha position presented in Pramāṇaviniścaya 1.58 (after the samanantara-pratyaya argument has been presented) and in the parallel passage of the Pramāṇavārttika make no reference to any resemblance of form between the inferred extra-mental object and the image contained in the cognition it produces. In PV, Pratyakṣa 224, Dharmakīrti still refers to the awareness having the appearance (ābhāsa) of the aggregated atoms which are “grasped” by it, and it seems that this position should therefore fall prey to the samanantara-pratyaya argument. But the later treatments of the bāhyārtha position say nothing about a shared image or appearance between the atomic external objects and the macroscopic image they produce in our awareness (which, perhaps intentionally, renders the samanantara-pratyaya argument moot). Instead, the external object is postulated purely as an extra-mental cause which explains the appearance of an image such as “blue” at some times but not at others. This is then something like a Kantian Ding an sich, the “thing in itself,” something we can know of as the cause of particular manifestations in our own awareness, but of which we can know nothing more that it causes these manifestations. On this understanding blue atoms would be “blue” only in the sense that they cause in us an awareness containing a blue image. There is no sense in which the atoms can be thought of as “blue” in and of themselves, when not producing an awareness, and consequently there could be no question or resemblance between the atoms and the image they produce in our awareness.

The avoidance of this question of resemblance in the presentation of the Sautrāntika position in Pramāṇaviniścaya 1.58 and the parallel portion of Pramāṇavārttika may or may not represent a deliberate departure from the discussion in Perception 223–224 in the Pramāṇavārttika, but in any case we can still see a basic continuity of approach to the question of external objects in both of Dharmakīrti’s major works, and for that matter in Dignāga’s Pramāṇasamuccaya as well. In all of these works, as in the Nyāyabindu, it seems that, even though Dignāga clearly held vijñānavādin views, and Dharmakīrti seems to indicate (at least in Pramāṇavārttika, Pratyakṣa 398) that this was his preferred position as well, both authors sought to produce an account of Buddhist epistemology that was consistent with either a vijñānavādin or a bāhyārthavādin stance, and hence could be adopted by people of varied ontological persuasions.

There is a larger context to the claim I am making here about Dharmakīrti’s treatment of the Sautrāntika or bāhyārthavādin view in the Pramāṇavārttika and Pramāṇaviniścaya. As I have argued elsewhere,^[20] there is actually quite strong evidence to indicate that some followers of Dharmakīrti were in fact bāhyārthavādins, most notably the highly influential commentator Dharmottara. It was precisely Dharmakīrti’s consistently guarded and careful balancing of the vijñānavādin and bāhyārthavādin positions that made it possible for his followers to adopt either of these ontological stances while still presenting themselves as loyal adherents of the Dharmakīrtian tradition.

In the light of Dharmakīrti’s own seemingly ambivalent treatment of the status of mind-independent objects in his major epistemological works, a full understanding of the Buddhist pramāṇa tradition as it developed in India in the wake of Dharmakīrti’s work would seem to call for a significant reevaluation of the role and status of the bāhyārthavādin position within it.

References and abbreviations

Arnold 2008 D. Arnold, Buddhist Idealism, Epistemic and Otherwise: Thoughts on the Alternating Perspectives of Dharmakīrti. Sophia 47 (2008) 3–28.

Dharmottarapradīpa Durvekamiśra, Dharmottarapradīpa, ed. D. Malvania. Patna^[21] 1971.

Dreyfus 1997 G. Dreyfus, Recognizing Reality: Dharmakīrti’s Philosophy and its Tibetan Interpretations. Albany 1997.

Dunne 2004 J. Dunne, Foundations of Dharmakīrti’s Philosophy. Boston 2004.

Eltschinger 2010 V. Eltschinger, Dharmakīrti. Revue Internationale de Philosophie 253 (2010) 397–440.

Eltschinger 2014 V. Eltschinger, Buddhist Epistemology as Apologetics. Vienna 2014.

ĪPVV Abhinavagupta, Īśvarapratyabhijñāvivṛtivimarśinī, ed. M. Kaul Shastri. 3 vols. Bom-bay 1938–1943.

Kāśikā Sucaritamiśra, Kāśikā. See Śāstrī 1927–1943.

Kellner 2011 B. Kellner, Dharmakīrti’s Criticism of External Realism and the Sliding Scale of Analysis. In: Religion and Logic in Buddhist Philosophical Analysis. Pro-ceedings of the Fourth International Dharmakīrti Conference, ed. Helmut Krasser, Horst Lasic, Eli Franco, and Birgit Kellner. Wien 2011, 291–298.

Matsumoto 1980 S. Matsumoto, Sahôpalambha-Niyama. Sōtōshū Kenkyūin Kenkyū Kiyō (Journal of Sōtō Sect Research Fellows) 12 (1980) 1–34.

McClintock 2003 S. McClintock, The Role of the “Given” in the Classification of Śāntara-kṣita and Kamalaśīla as Svātantrika-Mādhyamikas. In: The Svātantrika-Prāsaṅgika Distinction: What Difference Does a Difference make?, ed. G. Dreyfus and S. Mc-Clintock. Boston 2003, 125–171.

McCrea 2016 L. McCrea, Abhinavagupta as Intellectual Historian of Buddhism. In: Around Abhinavagupta: Aspects of the Intellectual History of Kashmir from the Ninth to the Eleventh Century, ed. Eli Franco and Isabelle Ratié. Berlin 2016, 263–286.

McCrea and Patil 2006 L. McCrea and P. Patil, Traditionalism and Innovation: Philoso-phy, Exegesis, and Intellectual History in Jñānaśrīmitra’s Apohaprakaraṇa. Journal of Indian Philosophy 34 (2006) 303–366.

McCrea and Patil 2010 L. McCrea and P. Patil, Buddhist Philosophy of Language in India: Jñānaśrīmitra on Exclusion. New York 2010.

Nyāyabinduṭīkā Vinītadeva’s Nyāyabindu-ṭīkā: Sanskrit original reconstructed from the extant Tibetan version, with English translation and annotations, tr. Mrinalkanti Gangopadhyaya. Calcutta 1971.

NM Jayantabhaṭṭa, Nyāyamañjarī, ed. K. Varadacharya. 2 vols. Mysore 1969–1983.

PS Dignāga’s Pramāṇasamuccaya: A hypothetical reconstruction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Ṭīkā, ed. E. Steinkellner. Wien 2005. https://www.oeaw.ac.at/fileadmin/Institute/IKGA/PDF/fo rschung/buddhismuskunde/dignaga_PS_1.pdf, last visited 13-08-2019.

PSṬ Jinendrabuddhi, Viśālāmalavatī Pramāṇasamuccayaṭīkā, Chapter 1, Part 1: Critical Edition, ed. E. Steinkellner, H. Krasser, and H. Lasic. Beijing/Vienna 2005.

PV Dharmakīrti, Pramāṇavārttika, ed. Ram Chandra Pandeya. Delhi 1989.

PVA Prajñākaragupta, Pramāṇavārttikabhāṣyam or Varttikālaṅkāra, ed. Rahula Sankri-tyayana. Patna 1943.

PVin Dharmakīrti’s Pramāṇaviniścaya: Chapters 1 and 2, ed. E. Steinkellner. Beijing/ Vienna 2007.

PVV Manorathanandin, Pramāṇavārttikavṛtti. In PV.

Śāstrī 1927–1943 Mīmāṃsāślokavārttika with the Commentary Kāśikā of Sucaritamiśra, ed. K. Sāmbaśiva Śāstrī. 3 vols. Trivandrum 1927–1943.

Ślokavārttika See Śāstrī 1927–1943.

Steinkellner 1990 E. Steinkellner, Is Dharmakīrti a Mādhyamika. In: Earliest Buddhism and Madhyamaka, ed. D.S. Ruegg and L. Schmithausen. Leiden 1990, 72–90.

TBh Tarkabhāṣā of Mokṣākaragupta, ed. E. Krishnamacharya. Baroda 1942.

Watson 2010 A. Watson, Bhaṭṭa Rāmakaṇṭha’s Elaboration of Self-Awareness (svasaṃve-dana), and How it Differs from Dharmakīrti’s Exposition of the Concept. Journal of Indian Philosophy 38 (2010) 297–321.

Footnotes and references:

[1]:

See Dunne 2004: 11, 68–69, 239.

[2]:

See Vinītadeva ad Nyāyabindu 1.1 and 1.3 (Nyāyabinduṭīkā, pp. 4 and 7): iṣṭaś ca prakaraṇārambhaḥ sautrāntikayogācārobhayanayānudhāvanārtham. See Dharmottarapradīpa, pp. 42–44 and 61 for addi-tional references.

[3]:

PVin 1.54 (p. 39): sahopalambhaniyamād abhedo nīlataddhiyoḥ /

[4]:

PVin 1.58 (p. 43): bāhye’py arthe tato’bhedo bhāsamānārthatadvidoḥ / (1.58ab) saty api bāhye’rthe sahopalambhavedanābhyāṃ bhāsamānasya nīlādes tatsaṃvidaś cāvivekaḥ siddhaḥ. dvairūpyaṃ tad dhiyaḥ (1.58c) viṣayajñānarūpābhyām ato dvirūpā buddhiḥ siddhā bhavati. yadi bhāsamāno viṣayākāro buddher abhinnaḥ, tato bhinnam astīti kutaḥ? bāhyasiddhiḥ syād vyatirekataḥ // (1.58cd) satsu samartheṣu anyeṣu hetuṣu jñānakāryāniṣpattiḥ kāraṇāntaravaikalyaṃ sūcayati. sa bāhyo’rthaḥ syāt.

[5]:

For quotations of the passage, see for example the following: bāhyasiddhiḥ syād vyatirekataḥ: NM, vol. 2, p. 492; ĪPVV, vol. 2, p. 129; TBh, p. 35 (as “Sautrāntika” view); Jinendrabuddhi (PSṬ, p. 68). satsu samartheṣu… sa bāhyo’rthaḥ syāt: Kāśikā ad Ślokavārttika, Śūnyavāda 20 (Śāstrī 1927–1943, vol. 2, p. 100), Jinendrabuddhi (PSṬ, p. 68).

[6]:

I am not in fact the first to take note of the fact the sahopalambha-niyama argument is consistent with both the vijñānavādin and bāhyārthavādin positions as Dharmakīrti ultimately develops them. Matsumoto (1980: 26) makes the same point and Alex Watson (2010: 311, n. 46) has recently drawn attention to this point as well.

[7]:

PVin ad 1.58cd (p. 43): yady atra kaścid upādānaviśeṣābhāvakṛtaṃ kāryavyatirekaṃ na bruyāt.

[8]:

PVin ad 1.58 (pp. 43–44): so’pi kathaṃ sarvajñānānāṃ viṣayaṃ vyatirecayann upaplavetarayoḥ pramāṇetaratāṃ brūyād, viśeṣābhāvāt? upaplavavāsanāvisandhidoṣād aprabuddhasyāpy anāśvāsikaṃ vyavahāram utpaśyann ekam apramāṇam ācakṣīta, aparam āsaṃsāram aviśiṣṭānubandhaṃ dṛḍhavāsa-natvād iha vyavahārāvisaṃvādāpekṣayā pramāṇām. sāṃvyavahārikasya caitat pramāṇasya rūpam uktam. atrāpi pare mūḍhā visaṃvādayanti lokam iti. cintāmāyīm eva tu prajñām anuśīlayanto vi-bhramavivekanirmalam anapāyi pāramārthikapramāṇam abhimukhīkurvanti. tad api leśataḥ sūcitam eveti.

[9]:

PV, Pratyakṣa 391–392: tasmād arthasya durvāraṃ jñānakālāvabhāsinaḥ / jñānād avyatirekitvaṃ hetubhedānumā bhavet // 391 abhāvād akṣabuddhīnāṃ satsv apy anyeṣu hetuṣu / niyamaṃ yadi na brūyāt pratyayāt samanantarāt // 392

[10]:

The intervening verses (PV, Pratyakṣa 393–397) parallel the objection against the vijñānavādin account of prāmāṇya and the vijñānavādin response given in the vṛtti on PVin 1.58, but in more detail.

[11]:

PV, Pratyakṣa 398: asty eṣa viduṣāṃ vādo bāhyaṃ tv āśritya varṇyate / dvairūpyaṃ sahasaṃvittiniyamāt tac ca sidhyate // 398

[12]:

Pramāṇavārttikabhāṣya ad Pratyakṣa 398 (PVA, p. 416): kathaṃ tarhi bāhye’py arthe ācāryeṇa dvirūpatā vijñānasyoktā? vijñānavāda eva yuktaḥ. atrocyate. asty eṣa viduṣām vādo bāhyaṃ tv āśri-tya varṇyate / dvairūpyaṃ sahasaṃvittiniyamāt tac ca sidhyati // 398 // bāhyārthābhyupagamena parāpekṣayā dvirūpatā varṇyate, viduṣāṃ tu vijñānavāda eva.

[13]:

Pramāṇavārttikavṛtti on Pratyakṣa 398 (PVV, p. 142): asty eṣa sarvavyavasthāsu vijñaptimātratāpra- tipādako viduṣāṃ nyāyadarśināṃ yogācārāṇāṃ vādaḥ. sautrāntikair iṣṭaṃ bāhyam artham āśritya jñānasya dvairūpyam ācāryeṇa varṇyate. tac ca dvairūpyaṃ sahasaṃvedaniyamāt sahopalambhaniya-māt sidhyati bhede’pi sati tadbhāvāt [corr. tadabhāvāt].

[14]:

Pramāṇasamuccaya 1.9cd (PS, p. 4): yadā hi saviṣayaṃ jñānam arthaḥ, tadā svasaṃvedanānurūpam arthaṃ pratipadyata iṣṭam aniṣṭaṃ vā. yadā tu bāhya evārthaḥ prameyaḥ, tadā viṣayābhāsataivāsya pramāṇaṃ tadā hi jñānasvasaṃvedyam api svarūpam anapekṣyārthābhāsataivāsya pramāṇam. yasmāt so’rthaḥ tena mīyate // 9 //

[15]:

The link between Dharmakīrti’s treatment of the external object question and Dignāga’s is reinforced by Jinendrabuddhi, who, in explaining the version of the bāhyārthavādin view set forth by Dignāga in the Pramāṇasamuccaya, quotes PVin 1.58cd and closely paraphrases the following vṛtti (PS, p. 68).

[16]:

The commentator Prajñākaragupta says that these three alternative lines of argument are offered “in accordance with the reasoning of the Mādhyamikas, the Yogācāras, and the Bāhyārthavādins” (mādhyamikayogācārabāhyārthavādinayena –PVA, p. 288). Steinkellner, with good reason, questions Prajñākaragupta’s use of the label “Mādhyamika” here (1990: 76–78 and p. 86, n. 35). The first position seems closer to what later becomes known as the nirākāravādin position (that cognitions are in reality “without form;” or, in Dreyfus’s terms, the “False Aspectarian” position)–Dreyfus in fact straightforwardly labels it as such, plainly recognizing the divide between the nirākāravādin view argued for in Pramāṇavārttika, Pratyakṣa 210–219 and the sākāravādin view advanced in 220–222 (but not, as far as I can see, taking heed of the third, bāhyārthavādin, position on offer in 223–224–see Dreyfus 1997: 103–104 and 490).

[17]:

On the Nyāya view of perception a group of factors–self, sense organ, mind, object, etc.–are said to collectively cause perceptual awareness, while any of these factors individually will not do so. In the same way, there is no contradiction in asserting that multiple “blue” atoms in appropriate spatial proximity can collectively cause an awareness of blue, even though no single atom could cause such an awareness by itself.

[18]:

Pramāṇavārttika, Pratyakṣa 223–224: ko vā virodho bahavaḥ saṃjātātiśayāḥ pṛthak / bhaveyuḥ kā- raṇaṃ buddher yadi nātmendriyādivat // 223 hetubhāvād ṛte nānyā grāhyatā nāma kācana / tatra buddhir yadābhāsā tasyās tad grāhyam ucyate // 224 Dunne does recognize that these verses form part of Dharmakīrti’s discussion of the citratā problem, and translates them (2004: 411), but does not appear to consider their implications for his “sliding scale” analysis.

[19]:

See Arnold 2008, especially pp. 5–6.

[20]:

See McCrea and Patil 2006: 332–333; McCrea and Patil 2010: 19–20, 139, 142–143; and McCrea 2016.

[21]:

See Vinītadeva ad Nyāyabindu 1.1 and 1.3 (Nyāyabinduṭīkā, pp. 4 and 7): iṣṭaś ca prakaraṇārambhaḥ sautrāntikayogācārobhayanayānudhāvanārtham. See Dharmottarapradīpa, pp. 42–44 and 61 for addi-tional references.