again on Khyaati, aalaya in Dharmakiirti, etc.

John Dunne John.Dunne at ORIENT.UNIL.CH
Sun Feb 22 12:59:18 UTC 1998

Birgit noted:

> Schmithausen...suggests a date of
>ca. 700 C.E. for Man.d.ana, and implies...that
>Man.d.ana was active after Dharmakiirti. Take this together with
>Schmithausen's suggestion that Man.d.ana might have "coined" ("praegte")
>the *-khyaati*-terminology.

This is what I suspected, actually. Thanks for supplying the info, Birgit.
I still wonder about the relative chronology of Zaakyabuddhi and Ma.n.dana.
Perhaps it will become obvious at some point.

In my message, I had noted that the absence of the term in Dharmakiirti,
Devendrabuddhi and Zaakyabuddhi may suggest two possibilities: that they
were unaware of the terms, or that they deliberately avoided them. If we
accept that Ma.n.dana was active after all three of these Buddhists, then
the first possibility explains the situation rather nicely. However, if
Zaakyabuddhi was aware of Ma.n.dana's work, or if the terms continue to be
absent in later Buddhist pramaa.anavaadins, then we may need to ask other

Birgit went on to say:

> occurrence or
>non-occurrence of certain terms is not a sufficient condition for the
>acceptance or rejection of concepts which can be associated with these
>terms. Whether Dharmakiirti and his commentators *use* any of the
>"khyaati"-expressions is one issue; whether it is justifiable or useful
>to apply these terms when describing their theories is another. 

Yup. My point, however, is that an absence of these terms in later Buddhist
works would probably not a matter of sheer spite. Indeed, these kinds of
lacunae are exactly the type of clue that can lead to useful results. In
other words, that absence in earlier works may add support to a relative
chronology of authors. But in a more interesting fashion, an absence in
later works may suggest some problem that later Buddhists had with that
particular formulation.

>... widespread usage of *khyaati*-terms might indicate that 
>particular ... theories had gained wide
>acceptance..., but that doesn't tell you anything about whether these
labels >do justice to the philosophical theories they are attached to. And
...the >absence of a technical term in the literature
>is not enough to conclude that no ... concept [for it] exists. 

Ah, but it is one thing for a concept to be interpretively applicable to a
text, it is quite another for that text's author to be aware of (or accept)
that concept. A common example is the problem of the so-called
*praasa.mgika* and *svaatantrika* madhyamaka. The Tibetans (such as
Tsongkha pa in his *lhag mthong chen mo*) explicitly admit that they
invented these terms; they acknowledge that this distinction was not made
in India. Nevertheless, we might choose to use these concepts to interpret
the works of Candrakiirti, Bhaavaviveka, or even Naagaarjuna, but when we
do so, what kind of intepretation are we seeking? Clearly, this mode of
intepretation could easily be called perennialist, and it might well be
accused of blithely creating anachronisms.

Now, I do not mean that it is entirely invalid to apply anachronistic or
author-alien concepts in an attempt to understand a text. I just mean that
the type of understanding derived thereby differs in significant ways from
a hermeneutic that avoids anachronistic and author-alien concepts. And
certainly, from an historical point of view, the latter approach is

> ...applying terms which do
>not occur in certain texts to describe concepts advocated therein is not
>per se problematic.

Well, it all depends on what you're trying to do. If you are trying to do
history, it may well be problematic. And if you are trying to understand an
author within his/her context, it may be even more problematic. 

In any case, as I suggested in the earlier message, the formulation of
*aatmakhyaati* also raises certain issues that do not work well in the
Buddhist system. It places unwanted emphasis on the problem of accounting
for the notion that two entities are *sad.rza* without postulating a
hypostasized *saad.r.zyam*. 

>I am curious about what makes you think that Devendrabuddhi et al. may
>have perceived the *khyaati*-terms as "a strategem to draw them into a
>debate that they could not win". Are you thinking of the argument that
>an acceptance of a Niraakaaraj~naanavaada-type theory entails the
>acceptance of the (unwanted) doctrine of *asatkhyaati*?

No, although this points to one of the reasons why *aatmakhyaati* does not
work well as an intepretive device: it collapses the distinction between a
*satyaakaara* and an *aliikaakaara* reading of Dharmakiirti. I was thinking
more of the issue of simalarity mentioned above. Kar.nakagomin kind of puts
his foot into this one (see his additions to Zaakyabuddhi's comments on p.
207, line 19ff, etc.).

>...Schmithausen voices the opinion that Man.d.ana's representation of
>Madhyamaka arguments have an ontologizing tendency which runs against at
>least Naagaarjuna's intentions. However, Schmithausen does not invoke
>the two-truth-framework in this context - what he considers problematic
>is the high emphasis placed on an "absolute non-existence" in the
>Maadhyamika position as it is presented by Man.d.ana.

Well, I'll have to look more closely at Schmithausen's book one of these
days. As I recall, Vaacaspatimizra's formulation ignores the distinction
between claiming that something is ultimately nonexistent, but
conventionally existent. It thus does not allow for the distinction between
** and ** (or, for Candrakiirti,
** and **). That is, from a conventional
perspective, some things appear real, while others appear illusory/false;
but in ultimate terms, all those things -- whether conventionally real or
conventionally illusory/false-- are equally unreal. Without this
distinction (between tathya- and or between and sa.mvr.ttimaatra), one arrives at Hegel's "night in
which all cows are black," which is the formulation heaped upon those poor
madhyamakas by the nasty *asatkhyaati* analysis.

>Seriously, "atasmi.s tad iti pratyayah." (or "j~naanam.") is
>over-abundant in Nyaaya-Vai'ses.ika literature (Nyaayabhaas.ya,
>Nyaayavaarttika, Padaarthadharmasa.ngraha, Candraananda's
>Vai'ses.ikasuutravr.tti, etc.). There's nothing specifically Buddhist
>about it; it's not too easy an answer, it's simply a non-starter.

Hmmm... I wouldn't say so. Certainly, * tadj~naana* is not just
all over Nyaaya, it's found in nearly every system (and the qualifier
"nearly" is probably not even necessary). The point for Dharmakiirti is
that any elaboration beyond that point necessarily leads to the
construction of a universal or an entity that functions like a universal.
As usual when confronted by the problem of similarity, Dharmakiirti
essentially retreats to *svasa.mvedana*: when an entity is mistaken for an
other, their apparent similarity is accounted for by nature (*prak.rti*) of
those particulars; they just present themselves in that fashion in
perception, and this fact is given by apperception.

In a sense, then, Dharmakiirti is struggling not to elaborate beyond the
simple * tadj~naana*. Of course, he does present a mechanism to
describe how this works, but in ontological terms, he is seeking to base
error on the particulars themselves: the particulars involved are the basis
for the error. This is all comes clear in the passage I cited previously
(PVSV ad PV 1.98-99ab).

So I suppose I would not say that * tadj~naana* is a
"non-starter"; rather, this is the starting place for all Indian theories
of error. The interesting thing about Dharmakiirti is that, as I read him,
he is trying to say that any attempt to move beyond the starting point
carries you into realism of some kind. You might call this an aspect of
Dharmakiirti's "madhyamaka" side.

When I said:
>> ... I am still unconvinced that
>> Dharmakiirti responds to the *Zlokavaartika* in his *Svav.rtti*.

Birgit asked:

>Do you intend that as a general statement, or with specific focus on
>theories of error? In the former case, I'm curious on what prompts your

I was thinking specifically of theories of error, for I doubt that
Ka.rnakgomin's reference to Kumaarila is accurate. But since you mentioned
it, I do think that there is a general problem here: we have only the most
tenuous evidence for fixing a relative chronology for Kumaarila and
Dharmakiirti even in general. Krasser's Hiroshima paper certainly adds
weight to the argument, but it rests on the presumption that certain
arguments against iizvara did not exist in Miimaa.msaka circles before
Dharmakiirti. The history of Miimaa.msaa between Zabara and Kumaarila is so
sketchy that I find this presumption difficult to accept. The other obvious
question is: if Dharmakiirti knew of Kumaarila's
arguments, does this mean that he knew of the whole *Zlokavaarttika*? Were
texts only circulated in their complete form, or were various verses or
sections circulated independently? Was a text such as the *Zlokavaarttika*
composed all at once, or was it composed over several years during which
completed portions were circulated? 

When I read the *pratyak.sapariccheda* of Dharmakiirti's
*pramaa.navaarttika*, I often get the impression that it was meant to stand
on its own -- otherwise, it is simply too repetitive with regard to other
portions of the PV. The same argument has already been made for the
*Svav.rtti*, although there are strong reasons to reject that argument.
Now, I am NOT saying that either the ** or *anumaana* chapters of
the PV were in fact composed and circulated as separate treatises. I am
just saying that this is a possibility worthy of examining. 

Remember there is a precedent for this way of producing texts: legend has
it that the *Pramaa.nasamuccaya* was Dignaaga's attempt to draw together
various scattered works that were already in circulation. 

In any case, the relationship between Kumaarila and Dharmakiirti is
certainly not a straightforward one, in the sense that we have no clear
evidence for fixing a relative chronology. I suspect that we will always be
somewhat in doubt about this issue, perhpas because they were contemporaries.

Birgit also said:

> I must confess that I have not yet
>considered the issue of Dharmakiirti and the *aalayavij~naana*, but I
>was just reminded that a couple of months ago, I was taught that
>Dharmakiirti does not accept an *aalayavij~naana* ...

Ah, yes, the triumph of the Tibetans! As near as I can tell, the ubiquitous
claim that Dharmakiirti did not accept *aalayavij~naana* has entered into
contemporary discourse through various Tibetan sources (perhaps promulgated
by Stcherbatsky?). The dGe lugs pas (who have a severe allergy to *kun
gzhi*) hit this notion particularly hard. I too was first taught that
Dharmakiirti rejects *aalayavij~naana* -- imagine my surprise when I began
to run into passages that are based on the theory.

In the passage I metioned earlier, for example, *vaasanaa* plays a crucial
part. It's a bit lengthy to go into it now (although I could send a
translation, if you would like).

It is worth noting, however, that Zaakyabuddhi (111b=K206, with correction)
does use yogaacaara language when he glosses the various instances of
*vaasanaa* in this passage.For example, in glossing *svaanubhavavaasanaa-*
[Gnoli, 49.22], he gives (derge je, 11b5-6): 

*dngos po de dag gi rang nyams su myong ba sngar skyes pa des ngo shes pa
skye bar byed pa'i bag chags nus pa'i mtshan nyid gzhag pa gang yin pa*
[=K20618-20: *te.saam bhaavaanaa.m ya.h svonabhuaava.h puurva.m utpannas
tena yaa pratyabhij~naanotpattaye vaasanaa zaktilak.sanaahitaa {corrected
from -aaditaa}].

One can argue that *Aalayavij~naana* is assumed as the place where the
*vaasanaa* is "placed" (*aahitaa*). In response, some would argue that
*vaasanaa* can simply be placed in *manovij~aana*. 

A clearer reference appears later in this passage:

*sarve.saa.m [=vikalpapratyayaanaa.m] viplave'pi
pramaa.natadaabhaasavyavasthaa  aa aazrayaparaav.rtter
arthakriyaayogyaabhimatasa.mvaadanaat* [Gnoli, 51.3-4]

By Dharmakiirti's time, any discussion of *aazrayaparaav.rtti* would assume
*aalayavij~naana*. In any case, this is how Zaakyabuddhi understands it. He
glosses (115a1): 

*gnas pa ni kun gzhi rnam par shes pa'i mkhrul pa'i sa bon can no / gnas
gyur pa ni sgrib pa an bral ba'o* [Skt preserved in K(211.8-9): *aazrayo
bhraantibiijam aalayavij~naa.m tasya paraav.rttir aavara.navigama.h*].

Thus, for Zaakyabuddhi (and Kar.nakagomin) at least, Dharmakiirti is
explicitly referring to aalayavij~naana. I tend to agree, especially in
view of the way *vaasanaa* is used in the passage.

There are other, even more obvious places where Dharmakiirti refers to
*aalayavij~aana*. For example, at after rejecting in PV3.334-335 the notion
that a *baahyaartha* is necessary to account for the fact that there are
limits on perception, he remarks: 

*kasyacit ki.mcid evaantarvaasanaayaa.h prabodhakam/ 
tato dhiyaa.m viniyamo na baahyaarthavyapek.sayaa* //PV3.336//

This is classic yogaacaara, and it requires *aalayavij~anaana* as part of
the process. But the most explicit is:

*sak.rdvijaatiiyajaataav apy ekena pa.tiiyasaa/
cittenaahitavaigu.nyaad aalayaan naa.nyasambhava.h //PV3.522//

All the extant Indian commentaries (in Tibetan trans. or Skt.) gloss
*aalayaat* here as *aalayavij~naanaat*; it is also clear that Dharmakiirti
means to defend this position. 

In any case, the Indian commentarial tradition certainly understood
Dharmakiirti to be accepting *aalyavij~naana* (I have seen other references
in Devendrabuddhi, Zaakyabuddhi, Viniitadeva and Praj~naakaragupta). In
part, they understand this position to be linked to the refutation of
*baahyaartha* in the ** chapter of PV. Since the other passage
(PVSV ad PV1.98-99ab) also refers to a yogaacaara position, it is not
surprising to find implicit reference to *aalayavij~aana* there as well.

>BTW, another article which covers theories of error is Eli Franco's
>"Studies in the Tattvopaplavasimha, II. The Theory of Error", Journal of
>Indian Philosophy 12/1984, 105-137. 

Hey, I forgot this one! I read it about ten years ago, which means I
probably had no clue what it meant. Thanks for the reminder.

Well, it's time to stop.




John Dunne
Section de langues et civilisations orientales
Université de Lausanne

More information about the INDOLOGY mailing list