205

A Critical Exchange onthe Idea of Dh„tu-v„da

Response

MATSUMOTO Shirõ

IFIND YAMABE’S ESSAY on the idea of dh„tu-v„da both important andfull of valuable information and interpretations, and I am gratefulfor his studied critique of my hypothesis. While I am not prepared

to accept his views and conclusions, I welcome the serious response andwish to respond to it brieµy here.

Yamabe sums up his position succinctly near the end of his essay:

The Yogacara theory of gotra is indeed discriminatory, but it is not basedon monism. Tath„gata-garbha thought is clearly monistic, but the gotradistinction does not seem to signify anything essential. Some texts, suchas the Abhisamay„la½k„ra-v£tti, show an apparent dh„tu-v„da struc-ture, but that structure would appear to be an inconsistency broughtabout by a reinterpretation of the gotra theory.

I cannot, however, accept his conclusion, because I ³nd the Yogacara the-ory of gotra to be based on monism, and because the gotra distinctionseems to be of essential importance to tath„gata-garbha thought. More-over, unlike Yamabe, I ³nd the dh„tu-v„da structure of the Abhi-samay„la½k„ra-v£tti to be completely consistent. To clarify my position,let me take up the de³nition of gotra from the Bodhisattvabhðmi thatYamabe cited in his essay.

Before discussing the de³nition, I should like to express my surpriseat the way he goes about criticizing my hypothesis. I should have thoughthe would have given attention to my essays “On the Ekayana Theory inYogacara” and “The Šr‡m„l„dev‡ Sutra and Ekayana Theory,” on whichmy framing of the hypothesis of dh„tu-v„da is mainly based. The formeressay is especially pertinent to Yamabe’s critique in that its arguments arefocused on the important passage of the Mah„y„nasðtrala½k„ra-bh„¤ya

commenting on K„rik„ XI. 53, which seems to clarify the meaning of theterm dharmadh„tu for the Yogacara thinkers. The passage in questionreads “because the dharmadh„tu of the šr„vaka etc. is undifferentiated”(šr„vak„d‡n„½ dharmadh„tor abhinnatv„t). Here we have a clear state-ment of the uniqueness of dh„tu for the Yogacara thinkers. I would beinterested in knowing how Yamabe would interpret this passage.

Now let us turn to the Bodhisattvabhðmi de³nition of gotra.1

Yamabe’s conclusion draws on the work of other scholars (among them,Hakamaya2) who take the terms gotra, b‡ja, dh„tu, and prak£ti thatappear in the de³nition as all synonymous. Yamabe cites the phrase, “tatpunar gotra½ b‡jam ity apy ucyate dh„tu‹ prak£tir ity api,” which he ren-ders, “Further, this gotra is also called seed (b‡ja), dh„tu, and origin(prak£ti).”3 I do not ³nd the translation inaccurate,4 but I believe it isalso possible to render it, “But that gotra is also called b‡ja, and the dh„tuis also called prak£ti.” The Tibetan translation of the passage, it should berecalled, is “rigs de ni sa bon shes kyaª bya’ khams de ni raª bshin sheskyan bya‹o” (Derge edition, Wi, 2b5). This is clearly consistent with myown reading.

Admittedly there are passages in which Sthiramati states that thewords gotra, b‡ja, and dh„tu are synonymous, as Yamabe points out,5 andeven where he takes gotra and prak£ti as synonyms. But that isSthiramati’s interpretation.6 The Tibetan translation of the sentence citedabove suggests to me the possibility of a slight difference between the twogroups of words. In other words, one must admit the possibility thatdh„tu and prak£ti are taken to be different from gotra.

This interpretation of mine is supported by the appearance of theterm prak£ti-stha½ gotram in the de³nition of gotra cited above. Yamabefollows other scholars in translating the term as [the gotra] “existing bynature,” based on the Tibetan translation, “raª bshin gyis gnas pa”(Derge edition, Wi, 2b4). I ³nd fault with the translation, however, andprefer instead to read it as “the gotra located on prak£ti” or “the gotraexisting on prak£ti.” I further consider this prak£ti to be the unique locusor dh„tu of manifold gotra, giving this gotra theory the structure of adh„tu-v„da.

I admit the interpretation is altogether novel, but this does not makeit wrong. The author of the Bodhisattvabhðmi appears to equate prak£tiwith dharmat„ in the word dharmat„pratilabdha7 (“acquired by dharma-t„”). And this equation is reinforced by the passage of the Abhisamay„-

206

MATSUMOTO SHIRÕ

la½k„ra-v£tti alluded to by Yamabe.8

I believe, then, that the gotra of šr„vaka and the like are positionedon a single locus, which is called prak£ti in the de³nition. Hence myconclusion that Yogacara system has a dh„tu-v„da structure. In thisconnection I would also like to remark on the new interpretation of theword „lambana-pratyaya in the Yogacara texts to which Yamabealludes. He translates the word to mean “cognitive object” and criti-cizes my interpretation of dh„tu-v„da as “generative monism.” But„lambana-pratyaya does not in this context refer to a “cognitiveobject” but to “cause as locus.”9 In other words, I take „lambana thereto mean “locus” or “basis.” This reading is further supported by theterm tad„lambanaprabhava that appears in a section of theMadhy„ntavibh„ga-bh„¤ya in which Vasubandhu says that the dharmasof aryas are produced from “that locus” (tad„lambana)—that is, fromthe dharmadh„tu.10 It seems clear to me, therefore, that the Yogacarasystem has a dh„tu-v„da structure.

I wish to acknowledge Yamabe’s contribution to the elucidation ofthe n„n„dh„tu or “different-dh„tus” theory of Yogacaras, but I fail tosee how this is in any way inconsistent with the basic dh„tu-v„da struc-ture as I propose it. Rather, the manifoldness of dh„tu referred to in thepassages explaining the n„n„dh„tu theory is to be taken as a plurality ofgotras posited on a single locus, which is what I call dh„tu.

As for tath„gata-garbha thought, I prefer to leave my views ofYamabe’s interpretations for another occasion. I would only ask him tohave a second look at the passage of the Ratnagotravibh„ga he cites.According to his translation, it is stated there as follows: “Eventually therays from the sun-disk of the Tath„gata falls even on the bodies of sen-tient beings ³xed in a state of evil (mithy„tva-niyata-sa½t„na).”11 Doesthis really deny the existence of sentient beings ³xed in an evil state? Ithink not. Does the idea that we all, despite our caste, breathe the sameair really provide a basis for equality rather than discrimination?Moveover, it should be noted that the Ratnagotravibh„ga also admitsthe existence of agotra (etad agotr„«„m na vidyate)[I.41].

With this I conclude my remarks.

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

207

Riposte

YAMABE Nobuyoshi

TO BEGIN WITH, I would like to thank Matsumoto for his carefulresponse to my criticisms. I acknowledge that my essay did notgive due attention to certain points he made in two important

essays,1 but the fact that it was an oral presentation and the restrictions oftime made it dif³cult to enter into the necessary technical detail. Myremarks were no more than an attempt to outline my own understandingof dh„tu and gotra in as simple a way as possible. The Abhisamay„lam-k„ra verse I.39, which contains in a single verse the (apparent) dh„tu-v„da structure seemed to offer an ideal pivot around which to set up myargument. Since Matsumoto himself has called on the same verse as a suc-cinct demonstration of the “basic structure of dh„tu-v„da,”2 I assumedthis would provide some common ground for discussion.

In any case, I take it as my responsibility to comment now on thesetwo papers, particularly that dealing with “the Yogacara theory,” and toaddress the points of his response—even if, once again, time and spacemake it dif³cult to meet Matsumoto’s rigorous standards. Since our opin-ions seem to differ more radically in matters concerning Yogacara, I willfocus in what follows on the Yogacara theory of dh„tu and gotra.

Matsumoto’s dh„tu-v„da model is essentially static. He seems tobelieve that the same dh„tu-v„da structure (discrimination based onmonism) can be found throughout the Yogacara and tath„gata-garbhaliterature. In doing so, he bases his idea of dh„tu mainly on passages hav-ing to do with the monistic dh„tu—namely dharmadh„tu and a (ratherobscure) use of dh„tu in the singular from the Abhidharma Sutra. Hesees these singular dh„tu functioning as the universal basis of all phenom-ena, as the ground from which all things arise. Even though dh„tu is fre-quently de³ned as “cause,” his argument runs, this is only a secondary,derivative meaning based on the image of a (universal) locus. He does notdeny that dh„tu frequently appears in the plural form in Yogacara texts,

208

but he claims that all such manifold dh„tu are grounded in the singular(dharma)dh„tu. Therefore, he concludes, the existence of plural dh„tudoes not contradict the dh„tu-v„da model.

For my part, I have tried to delineate the development of the conceptof dh„tu (in most cases interchangeable with another concept, gotra) inmore dynamic terms. I am persuaded that the structure of the dh„tu the-ory in early Yogacara literature—by which I mean chieµy the Yog„c„ra-bhðmi—was essentially pluralistic.3 Dh„tus were primarily manifoldelements, each capable of producing corresponding elements. In otherwords, the primary meaning of the word in this tradition was “(pluralisticand individual) cause,” and the spacial imagery was merely subordinateto this. I ³nd it improbable that these manifold dh„tu were always basedon a single “locus” (Matsumoto’s dh„tu par excellence, as it were).

In particular, I do not believe that the gotra theory of the Bodhi-sattvabhðmi was based on monism in any form. Later, these dh„tu andgotra theories were restructured on a monistic model, and in the course ofthis reinterpretation, certain “hybrid” texts like the Mah„y„nasðtr„-lamk„ra and the Abhisamay„la½k„ra were created with both monisticand pluralistic elements. These texts seem to support Matsumoto’s modelof discrimination based on monism. But if one takes these texts in theirhistorical context, the discriminatory elements show up as remnants offormer pluralistic traditions. In contrast, purely tath„gata-garbha textslike the Ratnagotravibh„ga and the Šr‡m„l„dev‡ Sutra do not exhibitthese discriminatory factors.

I do not wish my remarks to be taken to mean that I deny the exis-tence of a dh„tu-v„da-type structure entirely. I believe that such a hybridstructure did in fact exist at one point in doctrinal development. For thatmatter, our views on the tath„gata-garbha theory are not so radicallydifferent either. I, too, think that the tath„gata-garbha theory had anessentially monistic structure, although I do not see it as intrinsicallydiscriminatory. It is regarding the structure of the dh„tu and gotra theo-ries of the early Yogacara tradition that our opinions diverge most widely.Accordingly, I should like to address this question more directly.

THE MEANING OF DHÃTU

Matsumoto claims that the original meaning of the word dh„tu is “a placeto put [something] on,” a kind of “base” or “locus.”4 This is the crucial

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

209

point around which virtually all dh„tu-v„da arguments revolve. But theimportance of the idea is not matched by its clarity.5

If the claim is an etymological one—that the original meaning ofdh„tu is “a place to put [something] on”—the question is a complex one,since the verbal root dh„- and the suf³x -tu, the two elements that makeup the word dh„tu, are both multivalent, and so the range of theoreticallypossible meanings for the compound is correspondingly great.6 An exam-ination of the usage of the term bears this out.7

According to Louis Renou’s extensive study of the suf³x -tu, theword dh„tu was used in two broad sets of meanings in the R. gveda: “basis,foundation” and “element, layer.”8 In the R. gveda, he argues, the termdh„tu is almost always compounded with a numeral (saptadh„tu,tridh„tu)9 and “³gures as the principle of division.”10 Hence the lattermeanings (element, layer) are both prevalent and well attested. The for-mer meanings (basis, foundation), however, are based on a single occur-rence of dh„tu as an independent (neuter) noun in the R. gveda (V.44.3):

atya½ havi‹ sacate sac ca dh„tu caThe oblation follows the steed [=Agni],it is the essence and the base [of sacri³ce].11

This is an exceedingly “obscure” verse, concerning which Renou notesthat the commentator, S„yana, glosses this dh„tu as dh„raka½ sarvasya(“the support of everything”).12 The gloss would appear to be close toMatsumoto’s interpretation, but it should be remembered that S„yana isa very late ³gure (fourteenth century).

Buddhist texts do occasionally interpret dh„tu as dh„ra«a, “sup-port,”13 which leads me to suppose that “support” (or, for Matsumoto,“locus”) was in fact one of the meanings of dh„tu.14 But this is not thesame as concluding that “support” or “locus” accounts for the word’setymological origins. Except for one obscure verse, the sense of “ele-ment” seems much more predominant in the R. gveda. If Matsumotowishes to make the claim that “locus” was the original meaning of theword dh„tu from which all the other meanings derived, he needs to pro-vide more supporting evidence. His claim concerning the etymologicalmeaning is, at best, not obvious.

THE USE OF DHÃTU IN YOGACARA LITERATURE

Matsumoto’s interpretation of the term dh„tu in Yogacara literature isbased on his reading of a verse of the Abhidharma Sutra and a passage

210

YAMABE NOBUYOSHI

from the Madhy„ntavibh„ga.15 I will leave a discussion of the Madhy„nta-vibh„ga for later. The verse from the Abhidharma Sutra runs:

The dh„tu from the beginningless past is the common basis for all thedharmas. When it exists, all the destinies [in samsara] and the attainmentof nirvana exist.16

Matsumoto asserts that “The expression sam„šraya [the commonbasis] clearly shows that the word dh„tu has the meaning of ‘locus’.”17

He goes on:

In relation to the [two] nominative-case nouns, gati‹ and nirv„«„dhi-gamo, dh„tu is the locus as is shown by the locative case tasmin [a pro-noun that represents dh„tu]. When it is put in the locative-absolutephrase tasmin sati, the meaning “if it exists” is added to the basic mean-ing “at that existing one,” and here the meaning “cause” arises. Thus ifthe word dh„tu has the meaning “cause,” it should be regarded as a sec-ondary meaning derived from the primary meaning of “locus.”18

There are several questionable points in these claims. First, the argu-ment is made that dh„tu means “locus” because the pronoun referring toit (tasmin) is in the locative case. In this verse, as Matsumoto himselfobserves, tasmin constitutes a locative absolute phrase together with thefollowing sati. But since the locative absolute is very common in Sanskrit,if anything that can be put into locative absolute has the meaning “locus,”it would follow that virtually all the Sanskrit nouns have that meaning. Ido not, therefore, see how the appearance of the locative absolute phraseproves anything about the meaning of the word.

Concerning the ³rst point sam„šraya, “the common basis,” weshould note that in the Yogacara school, the word „šraya is used in a verygeneral sense, almost equivalent to pratyaya (condition).19 Although theword „šraya itself would have a spacial connotation, I am not sure howmuch of that connotation is retained in this kind of technical context.

Furthermore, I cannot agree with Matsumoto’s claim that “locus” isthe primary meaning of the word dh„tu, and that the meaning “cause”derives from it.20 As far as I can see, in the early Yogacara literature, themeaning of “cause” is far more predominant than “locus,” and only whenthis primary meaning cannot be applied does the alternative meaning of“support” (or “locus”) come to the fore. Since the verbal root dh„- hasthe meaning “to generate or cause,” and since -tu can signify an agent, itis theoretically possible to deduce the sense of “cause” directly from the

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

211

components of the word without having to make the deduction indirectlythrough another meaning.

I believe it can be shown that the basic image of dh„tu in the veryearly Yogacara literature (the portions of the Yog„c„rabhðmi that do notpresuppose the theory of „layavijñ„na) was that of multiple (typicallyeighteen) elements capable of reproducing themselves from one momentto the next. By way of demonstration, I would cite a few passages, begin-ning with one from the Šr„vakabhðmi:

What are dh„tus? What is the skillfulness in dh„tus?Answer: dh„tus are eighteen, namely the eye-dh„tu, the color-dh„tu,

and the eye-consciousness-dh„tu, …[list of eighteen dh„tus]. The abilityto know, approve, and examine that these eighteen dharmas arise, issue,and become manifest from their respective dh„tus, their respective seeds(b‡ja), and their respective origins (gotra) is called skillfulness in dh„tus.To know the arising of the eighteen dharmas from their own dh„tus, thatis, the comprehension of causes and conditions, is at the same time thecomprehension of dh„tus.21

The meaning of this passage becomes clearer when we read it in con-junction with the following paragraph of the Abhidharmakoša-bh„¤ya:

The meaning of dh„tu is “origin” (gotra). Just as the manifold origins ofiron, copper, silver, and gold in a mountain are called dh„tus (minerals),the eighteen origins in one body or [personal] continuum are called“eighteen dh„tus.” In this case, “origins” refer to mines („kara). Thenwhose origins are the eye and so forth? [Each of the eighteen dh„tus isthe origin] of [the subsequent dh„tu of] its own type (j„ti), because [theformer] is the homogeneous cause [of the latter].22

[Objection:] In that case, an unconditioned [element] (asa½sk£ta)would not be a dh„tu. [Answer:] In this case, [it is a cause] of mind andmental functions.23

The two passages seem to me to be saying essentially the same thing.Each of the eighteen dh„tus arises from the same element in the preced-ing moment that is its generative cause (dh„tu). In other words, theessential nature of dh„tu seems to be the capacity to reproduce itself insuccessive moments.24

This point is con³rmed in the following passage from thePañcavijñ„nak„ya-sa½prayukt„-manobhðmi of the Vinišcayasa½graha«‡:

212

YAMABE NOBUYOSHI

What is eye but not eye-dh„tu is the eye of an arhat in the ³nal[moment]. This is the ³rst item [of the tetralemma].25

When the arhat enters nirvana nothing is left to carry over to the nextmoment, which is why in this ³nal moment the eye no longer reproducesthe same element in the following moment. And since the eye does notfunction here as a generative cause, it is no longer to be called dh„tu. Themodel here is clearly one of successive causality.26

This leads me to believe that in the early Yogacara literature dh„tuswere primarily plural elements reproducing themselves from one momentto the next.27 The primary meaning was thus “generative cause (in plural-istic contexts)” and did not derive from any image of universal “locus.” Isee no reason to argue that these plural dh„tus were supported by anotheruniversal dh„tu.28

On the basis of the actual usage of the words dh„tu and „šraya, then,it is not at all certain that the dh„tu in the verse of the Abhidharma Sutrameant universal locus. It should rather be taken to mean, as indeed mostcommentaries take it to mean, “cause.”29 I conclude that this verse doesnot support Matsumoto’s “locus” model.

Even though I ³nd myself in disagreement with Matsumoto’s rea-soning here, it must be admitted that the verse does give a monisticimpression.30 The point is not entirely beyond dispute,31 but even if onegrants that this verse is indeed monistic, this still needs to be understoodin the context of the transformation of the dh„tu theory from a pluralisticmodel into a monistic one. It is incorrect to assume that the entireYogacara theory of dh„tu was based on monism. The verse is no doubt animportant one, but it does not represent the whole of the Yogacara phil-osophy. As we have seen, there is strong reason to believe that the dh„tutheory of the early Yogacara school was pluralistic, and, in particular, thatsuch a pluralistic dh„tu model was directly related to the classical gotratheory of this school.32

MAHÃYÃNASÐTRÃLA¢KÃRA-BHÃS.YA

As Matsumoto points out, the Mah„y„nasðtr„la½k„ra-bh„¤ya at Mah„-y„nasðtr„la½k„ra XI.53 says:

Because the dharma is similar, the one-vehicle theory [was propounded].The dharmadh„tu is undifferentiated for šr„vakas and the like. [This

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

213

explanation is based on the] interpretation that [the word] y„na meansgoal.33

Matsumoto states that this passage “clearly shows the uniqueness of‘dh„tu’ for the Yogacara thinkers” and asks for my reading of the pas-sage.34

Obviously, in this passage dharmadh„tu refers to something univer-sal. I concede the point. But I question, ³rst, if the passage supports themodel of “generative monism” as Matsumoto claims; and second, if thegotra distinction is indeed based on such a universal dharmadh„tu.

The Model of Generative Monism

In order for this passage to ³t into the dh„tu-v„da model, it must ful³lltwo conditions: dharmadh„tu must be the locus of the three vehicles, andit must be the generative cause of the three vehicles.

As to the ³rst condition, as is clearly indicated in Sthiramati’s glossshes par bya ba’i yul, “the object to be known,”35 dharmadh„tu as y„tavyais a cognitive object, not a “locus.” Matsumoto is probably correct toassociate the passage from the Mah„y„nasðtr„la½k„ra-bh„¤ya cited abovewith the following line of the Madhy„ntavibh„ga-bh„¤ya:36

Dharmadh„tu is called so because it is the cause of the noble dharmas.Here, [the word] dh„tu is used in the sense of “cause.”37

Matsumoto claims that „lambana-pratyaya means “cause as locus.”38

But in reading this passage with the aforementioned gloss by Sthiramati,it is very dif³cult to take „lambana in any sense other than shes par byaba’i yul, “cognitive object.” Since the reading of „lambana as “locus” isnot the usual interpretation of the word, I should like Matsumoto to pro-vide further supporting evidence in support of his claim.

As to the second condition, that dharmadh„tu must be the generativecause of the three vehicles, I have already detailed in my paper my reasonsfor believing this is not the case. For his part, Matsumoto criticizes myargument by citing the Jñ„n„la½k„r„loka Sutra (de-bzhin-ñid ni gzhi’o,“Tathat„ is the basis [*pada]”).39 I wish to refute the criticism.

To begin with, the Jñ„n„la½k„r„loka is not a standard source of theYogacara doctrine, even though it was one of the many sources of theMah„y„nasðtr„la½k„ra.40 If Matsumoto wishes to criticize my under-standing of the Yogacara theory, he should reference his remarks with amainstream Yogacara text.

214

YAMABE NOBUYOSHI

It is true that there is a similar statement in the Vinišcayasa½graha«‡that describes „šraya-pariv£tti, “transformation of [personal] basis,” (inthis context equivalent to tathat„) as *prati¤¦h„-hetu, “supportive cause,”as opposed to *janma-hetu, “generative cause.”41 This *prati¤¦h„ wouldbe more or less equivalent to the *pada of the Jñ„n„la½k„r„loka Sutra,but prati¤¦h„ does not necessarily mean “locus” in the Yogacara literature.What is more, “supportive cause” and “generative cause” are clearly dif-ferent.42 In short, the passage in question does not in any way upholdMatsumoto’s “generative monism” model.

As I have discussed earlier, tathat„ in this kind of soteriological con-text has a rich doctrinal background.43 Reading the Mah„y„nasðtr„la½-k„ra in this light, I see no need to amend the interpretation presented inmy essay. I rather reiterate the point that tathat„ as a cognitive objectdoes not directly generate supramundane wisdom, much less other worldlydharmas.

The Gotra Distinction Based on a Universal Dharmadh„tu

As I discussed at some length,44 the passage in question from the Mah„-y„nasðtr„la½k„ra is not positive proof of the three-vehicle theory. It israther an argument that presupposes an already established three-vehicletheory and attempts to explain away scriptural passages that contradict thedoctrine. In a sense, this is a concession to the one-vehicle scriptures. Thetext is not referring to the universal dharmadh„tu in order to positivelyestablish the gotra discrimination. They are simply making a concession tothe one-vehicle theory, acknowledging that the three vehicles could beconsidered one in the sense that all of them share the same dharma-dh„tu. Diversity among the three vehicles is already presupposed here.

We need to look elsewhere for a principle of diversity. As Matsumotocorrectly observes,45 such a principle is found in Mah„y„nasðtr„la½k„raIII.11, where the existence of hetu-h‡na, “one who lacks the cause,” isaf³rmed. The Mah„y„nasðtr„la½k„ra-bh„¤ya at III.11 is a famous sectionthat divides those who cannot attain nirvana (aparinirv„«a-dharmaka) intotwo classes: temporarily hopeless people (tat-k„lâparinirv„«a-dharma)and eternally hopeless people (atyanta). The explanation given by theverse and commentary, however, is quite simple and clearly presupposesan already well-established gotra theory.

Such detailed arguments are found in the Bodhisattvabhðmi. Judgingfrom the broad af³nities between the Mah„y„nasðtr„la½k„ra and the

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

215

Bodhisattvabhðmi, and judging from the terminological similaritiesbetween the relevant sections, it is evident that the Gotr„dhik„ra of theMah„y„nasðtr„la½k„ra was very closely related to the Gotrapa¦ala of theBodhisattvabhðmi. In order to understand the structure of the Yogacaragotra theory, therefore, we need to understand the Bodhisattvabhðmi cor-rectly, but it is precisely at this point that Matsumoto and I diverge mostradically.

BODHISATTVABHÐMI

In his response, Matsumoto attempts to read prak£tistha-gotra as “thegotra existing on prak£ti [i.e., monistic dh„tu].”46 This is clearly in linewith his view that plural gotra, as the cause of enlightenment, are differentfrom the universal dh„tu or garbha.47 Indeed, his reading makes senseonly when we can differentiate gotra from dh„tu. The problem is thatsuch a differentiation is very dif³cult. Gotra and dh„tu are frequentlytreated synonymously, as Matsumoto himself later admits.48

As for the compound prak£ti-stha-, it is true that when stha- is used asthe second member of a compound, it usually means “existing in [someplace].” Nevertheless, such general usage would not automatically guar-antee that Yogacara writers used the term in the same sense. The sense ofa Yogacara technical term must be determined primarily by its usage inthe Yogacara literature itself. In this regard, we may note in the ³rst place(and again, as Matsumoto himself admits)49 that the Tibetan versiontranslates the compound as raª bzhin gyis gnas pa, reading the instru-mental sense (prak£ty„, “by nature”) into the ³rst member of the com-pound. This interpretation is further supported by the following passagefrom the Bodhisattvabhðmi:

What is the enhancement of dh„tu? Because of the former practice ofwholesome dharmas based on the existence by nature of the seeds ofwholesome dharmas (prak£ty„ kušala-dharma-b‡ja-sampada), the seedsof wholesome dharmas in each subsequent moment become moreenhanced, [then] most enhanced; they arise and they abide. This is calledthe enhancement of dh„tu (dh„tu-pu¤¦i).50

The scheme seems to ³t neatly with the aforementioned gotra theoryaccording to the following pattern:

prak£ty„ kušala-dharma-b‡ja-sa½pada —— prak£tistha-gotradh„tu-pu¤¦i —— samud„n‡ta-gotra

216

YAMABE NOBUYOSHI

Indeed, prak£ty„ kušala-dharma-b‡ja-sampada, “the existence by natureof the seeds of wholesome dharmas,” seems to supply the concrete mean-ing of the notion of prak£tistha-gotra. The interpretation looks still moreplausible when one takes into account the repeated use of the instrumen-tal form prak£ty„ in the Gotrapa¦ala of the Bodhisattvabhðmi.51 In thecase of the dharmat„-pratilabdha, it is probably not even possible to readit in any way other than as an instrumental tatpuru¤a. (Matsumoto’s owntranslation is “acquired by dharmat„.”52) Thus I believe that here bothprak£ti and dharmat„ are used in an instrumental (adverbial) sense (“bynature”) and do not refer to anything transcendental.53

It is of course true that the Abhisamay„la½k„ra-v£tti records anopinion that interprets the term prak£tistha-gotra as dharmadh„tu.54 Twopoints, however, should be noted in this regard.

First, Matsumoto’s claim is that prak£ti is the universal locus, and thatgotras are plural super-loci, but what the Abhisamay„la½k„ra-v£tti says isthat prak£tistha-gotra as a whole is identi³ed with (singular) dhar-madh„tu, not just prak£ti. Such an idea does not ³t into Matsumoto’sdh„tu-v„da model.

Second, this interpretation of the Abhisamay„la½k„ra-v£tti relies ona rejection of the original de³nition of ¤a^„yatana-više¤a given by theBodhisattvabhðmi. It is evident that the Abhisamay„la½k„ra-v£tti isemploying the idea of prak£tistha-gotra for its own agenda; it does notconvey the original meaning of the Bodhisattvabhðmi at all. As for the¤a^„yatana-više¤a itself, I can only direct the reader to the argumentspelled out in my essay.55 This expression may have had something to dowith the tradition that the Bodhisattva had six keen sense-faculties.56 Inany case, ¤a^„yatana clearly refers to concrete psycho-physical elements; itcannot refer to dharmadh„tu.

As already mentioned, the Bodhisattvabhðmi is the locus classicus ofthe gotra theory of the Yogacara school, but in the Bodhisattvabhðmithere is nothing to suggest that the gotra system was based on monism.This is clear proof that the Yogacara gotra system does not theoreticallyrequire a monistic basis. On the contrary, it was actually based on plural-ism at the time of the Bodhisattvabhðmi. The Gotr„dhik„ra of theMah„y„nasðtr„la½k„ra is also very likely based on a pluralism.57 I canonly conclude that pluralism was the principle that accounts for the diver-sity of views among those in the Yogacara school. Monism cannot offer apositive explanation.

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

217

RATNAGOTRAVIBHÃGA

My riposte has already run on too long, and I must therefore refrain fromdelving into the details of the tath„gata-garbha texts. I would only like toreact brieµy to points raised in Matsumoto’s response. In doing so, I mustadmit, in all honesty, that I cannot make much sense of his comments onthe passage I quoted from the Ratnagotravibh„ga.58 The passage clearlydenies the idea of eternally hopeless icchantikas.59

The situation is basically the same with regard to the Šr‡m„l„dev‡Sutra. It seems fair to say, with Matsumoto, that the doctrinal structure ofthis sutra is monistic.60 Nevertheless, the sutra does say:

O Blessed One, šr„vaka-y„na and pratyekabuddha-y„na are all gatheredin Mahayana…. Thus, the three vehicles are counted as one. O Blessed One,realizing the one vehicle, one realizes the unsurpassed enlightenment.61

Although Matsumoto would object to this type of inclusive approach,62 itis hard to read the above quotation as discriminatory. Matsumoto himselfstates that the Šr‡m„l„dev‡ Sutra is based on the one-vehicle theory.63

This being the case, there is no textual basis to claim that the Šr‡m„l„dev‡Sutra, “the most typical dh„tu-v„da text,”64 was discriminatory.

As I state in my essay,65 I do not doubt that tath„gata-garbhathought has a monistic structure. But I do believe that the dh„tu-v„datype of monism becomes discriminatory only when the differences amongthe “super-loci” are ³xed and made unchangeable. Both the Ratnagotra-vibh„ga and the Šr‡m„l„dev‡ Sutra clearly reject such ³xation. Given theteachings of these two representative texts of the tath„gata-garbha tradi-tion, one is hard put to see how a monistic tath„gata-garbha theory canbe intrinsically possessed of discriminatory elements. Quite the contrary, itseems to me that only when such tath„gata-garbha ideas were mixed intoYogacara doctrine, which had a discriminatory and pluralistic gotra theoryas part of its tradition, does the problem of “monistic discrimination”(i.e., dh„tu-v„da) come about. This is why I have claimed that dh„tu-v„da was essentially a “patchwork” of these two different traditions.

CONCLUDING REMARKS

Needless to say, the interaction between the Yogacara and tath„gata-garbha traditions is a complex issue. There is no doubt that certain partsof the Yogacara and the tath„gata-garbha traditions are closely inter-

218

YAMABE NOBUYOSHI

twined, and that there are many aspects in which the two look alike. Inthis sense, I may have overstated the difference between these traditions.

Still, it does not seem to me that the whole of these two traditionscan be said to have shared the kind of clear-cut structure that Matsumotocalls dh„tu-v„da. If there was indeed any coherent structure to their com-monality, one has to wonder why such vehement arguments wereexchanged between the followers of tath„gata-garbha and YogacaraBuddhism in China and Japan. It rather seems to me that, in spite ofmany shared terms, the classical Yogacara dh„tu/ gotra system as is foundin the older parts of the Yog„c„rabhðmi had a radically different structurefrom its counterpart in full-µedged tath„gata-garbha texts such as theRatnagotravibh„ga.

In conclusion, I wish to thank Matsumoto once again for taking mycritique seriously. I am aware that universal and discriminatory elementsoften coexist in a single text, and I do not necessarily doubt that there isa side to the tath„gata-garbha theory that could well function to obscurediscrimination. Although I do not happen to think that discrimination isthe inevitable conclusion of the tath„gata-garbha doctrine, I certainlyacknowledge his contribution in pointing out the potentially dangerousaspects of this theory. If I have not properly understood him on certainpoints, I would like to ask him kindly to address misunderstandings onsome future occasion.

A CRITICAL EXCHANGE ON DHÃTU-VÃDA

219