The Idea of Dh tu-v da in
Yogacara and Tath gata-garbha Texts
YAMABE Nobuyoshi
AKAMAYA NORIAKI AND Matsumoto Shir� are convinced that
tath gatagarbha theory and the Yogacara school share a com-
H
mon framework that they call dh tu-v da or  locus theory.
The word dh tu-v da itself is a neologism introduced by Matsumoto1
and adopted by Hakamaya.2 They argue that the dh tu-v da idea stands
in direct contradiction to the authentic Buddhist theory of prat! tya-
samutp da or  dependent origination, which in turn leads them to con-
sider tath gata-garbha and Yogacara theories to be non-Buddhist. In
their opinion, not only these Indian theories but also the whole of  orig-
inal enlightenment thought (hongaku shis�) in East Asia fell under the
shadow of the dh tu-v da idea,3 with the result that most of its
Buddhism is dismissed as not Buddhist at all.4
The idea of dh tu-v da is thus an integral part of the Critical
Buddhism critique and as such merits careful examination in any evalua-
tion of the overall standpoint. Since Matsumoto łrst found the dh tu-
v da structure in Indian tath gata-garbha and Yogacara literature, we
need to begin with a look at the texts in question. My approach here will
be purely philological and will limit itself to the theoretical treatises (sastras).
BASIC STRUCTURE OF THE DH�TU-V�DA THEORY
For Hakamaya and Matsumoto, the cardinal tenet of Buddhism is
prat! tyasamutp da, which they understand as consisting of a temporal
sequence of causally linked dharmas or  phenomena that lack any solid
basis in reality. Matsumoto illustrates this by means of the following
chart.5
193
YAMABE NOBUYOSHI
CHART 1
dharma dharma dharma
�' �' �' �'
dh tu
Chart 2, in contrast, illustrates the basic structure of dh tu-v da.6
CHART 2
a
r vaka-dharma pratyekabuddha-dharma tath gata-dharma
= super-locus
dharmadh tu = ekay na
= locus
This second model, in contrast to the prat! tyasamutp da model,
introduces a universal, solid basis under such names as dharmadh tu and
buddhadh tu. In Matsumoto s terminology, this universal dh tu is a
 locus that supports phenomenal dharmas as  super-loci. He enumer-
ates the characteristic features of this dh tu-v da model as follows:7
1.  Locus is the basis for  super-loci.
2.  Locus gives rise to  super-loci.
3.  Locus is one,  super-loci are many.
4.  Locus is real,  super-loci are not real.
5.  Locus is the essential nature of  super-loci.
6.  Super-loci are not ultimately real, but have some reality in that
they have arisen from the  locus and share its nature.8
Thus, the dh tu-v da model is essentially a monism (or, according to
Matsumoto s own terminology, a  generative monism n���s��).9
194
THE IDEA OF DH�TU-V�DA
Hakamaya and Matsumoto take this to be an Upanińadic model and thus
not authentically Buddhist. Particularly problematic for them is the fact
that this dh tu-v da framework is not as egalitarian as it appears. As they
see it, one can classify any number of different elements from the three
vehicles to social castes as  super-loci resting on the universal  locus.
Since the diversity of the  super-loci is an essential element of the dh tu-
v da structure, the distinction among  super-loci remains unaffected.
On the other hand, the apparent equality that obtains on the absolute
level serves at once to justify, obscure, and conłrm the discrimination that
appears on the phenomenal level.10 In Matsumoto s opinion, this essen-
tially discriminatory nature of the dh tu-v da structure is clearly
expressed in verse I.39 of the Abhisamay la�k ra:11
Because the dharmadh tu has no distinction, any distinction among
gotra is unreasonable. Nevertheless, because the dharmas to be posited
[on the  locus of dharmadh tu] are distinct, a distinction [among
gotra] is proclaimed.12
In the same way, the ideas of universal  Buddha-nature and icchan-
tika in the Mahayana Mah parinirv �a Sutra do not contradict each
other but combine to form a harmonious whole. The Mah y nas�tr -
la�k ra asserts, on the one hand, that all sentient beings have tath gata-
garbha (verse IX.37)13 and, on the other hand, admits that some people
will never be able to attain nirvana (verse III.11). Hence the dh tu-v da
structure also represents a principle supporting the discriminatory gotra
theory of the Yogacara school.14
MONISM OR PLURALISM?
Matsumoto s arguments are well prepared, and the coexistence of a univer-
sal  Buddha-nature 15 and unequal attainments is indeed problematic.
Still, it may well be possible to explain this coexistence in somewhat dif-
ferent terms.
A good place to begin is the famous dełnition of gotra in the
Bodhisattvabh�mi section of the Yog c rabh�mi :
What is gotra? In brief, gotra is twofold: the one existing by nature
(prakŁtistha) and the attained one (samud n! ta).
The gotra existing by nature is the distinct state of the six-sense-basis
(ńa^ yatana-via
eńa ) of bodhisattvas. That [distinct state] was naturally
195
YAMABE NOBUYOSHI
acquired in the beginningless past and has been transmitted as such [to
the present].
The attained gotra is what is acquired through the practice of merits
in the past [lives].
In this case, both meanings are intended. Further, this gotra is also
called seed (b! ja), dh tu, and origin (prakŁti ).16
Since Hakamaya himself quotes this last sentence as an example of the
monistic dh tu model,17 it is clear that he considers the gotra theory of
the Bodhisattvabh�mi to be a form of monism. His argument is based on
the fact that all the terms given (gotra, b! ja, dh tu, prakŁti) appear in sin-
gular forms.18 But the argument is not without its weaknesses.
First, the paired terms prakŁtistha-gotra and samud n! ta-gotra have a
close analogue in the Vastusa�graha�! section of the Yog c rabh�mi:
In sum, dh tus are twofold: the ones existing by nature (rang-bzhin gyis
gnas pa, WĄżł
, *prakŁtistha19) and the ones enhanced through habitu-
al practice (goms-pas yongs-su-brtas-pa, H �ł
, *abhy sa-paripuńŚa).
The ones existing by nature are, for example, the eighteen dh tus
(khams), which are seeds (sa-bon, *b! ja) staying in their own respective
continuities.
The dh tus enhanced through habitual practice are enhanced seeds
resting in the body (rten, * a
raya) so that the good or bad dharmas
habitually practiced in other, former lives might arise [easily]& .20
In these two passages, it is clear that the prakŁtistha-gotra of the Bodhi-
sattvabh�mi corresponds to the *prakŁtistha-dh tu of the Vastusa�-
graha�! . The correspondence between the samud n! ta-gotra,  attained
gotra, of the Bodhisattvabh�mi and the *abhy sa-paripuńŚa-dh tu,  the
dh tu enhanced through habitual practice, of the Vastusa�graha�! may
not be immediately evident, but is conłrmed by the Mah y na-
s�tr la�k ra-bh ńya, which equates samud n! tam [gotram],  attained
gotra, and paripuńŚa� [gotra�],  enhanced gotra. 21 Consequently, a
correspondence between this portion of the Vastusa�graha�! and the
aforementioned portion of the Bodhisattvabh�mi seems beyond dispute.
The basic message of the passage of the Vastusa�graha�! is that
there are innumerable good and bad elements (dh tu) in sentient beings
that correspond to good and bad mental functions, and that one must
accordingly cultivate the good elements in order to realize good mental
states.22 In other words, here the dh tu theory is clearly of a pluralistic
sort. Furthermore, it is worth noting that the pluralistic structure of this
196
THE IDEA OF DH�TU-V�DA
passage was recognized by Hakamaya himself in an earlier essay.23 At the
same time, we have established that the gotra theory of the Bodhisattva-
bh�mi was closely related to the dh tu theory of the Vastusa�graha�! .
This being the case, it is likely that the gotra theory of the Bodhisattva-
bh�mi itself was pluralistic in structure.24
The pluralistic dh tu model and the gotra theory are intrinsically
related to one another. In the same way that the dh tu of desire is inca-
pable of generating hatred, the gotra of sravakas is unable to generate the
supreme wisdom of the Buddha. Without such distinct gotras, therefore,
it would not be possible to establish a distinction among the three vehi-
cles.25 Accordingly, at least as far as these passages are concerned, the only
chart we are able to draw is the one below (Chart 3). Obviously the plu-
ralism it presents is not the same as the  generative monism that
Matsumoto offers.
CHART 3
sravaka pratyekabuddha bodhisattva
a
r vaka-dh tu pratyekabuddha-dh tu bodhisattva-dh tu
(gotra) (gotra) (gotra)
SAóSKRTA OR ASAóSKRTA?
. .
There is more involved in what has been discussed above than merely
whether dh tu is singular or plural. It has to do with the foundation for
supramundane attainment. If there is any possibility at all for us to acquire
supramundane wisdom, on what does such a possibility rest? The
Bodhisattvabh�mi responds by referring to inherent gotra and dełnes the
foundation as ńa^ yatana-via
eńa, or  the distinct state of the six-sense-
basis. But just what does this ńa^ yatana-via
eńa mean? We get a clue, I
believe, from the following passage of the Abhidharmakoa
a-bh ńya:
The [distinction between noble ones and ordinary ones] is made in terms
of the distinct states of bodies ( a
raya-via
eńa). [This distinction is possi-
197
YAMABE NOBUYOSHI
ble] because the body ( a
raya) of noble ones is transformed (par vŁtta)
by the power of the paths of seeing and practice (dara
ana-bh van -
m rga), so that [the body] is no more capable of generating the
dełlements to be abandoned by the [paths of seeing and practice].
Therefore, when one s body has ceased to be the seed of dełlements
(ab! j! -bh�te  a
raye klea
 n �), like a grain of rice consumed by the łre,
one is called [a noble person] who has abandoned dełlements.26
In the body of literature to which this text belongs,  a
raya usually means,
in the absence of further contextual speciłcation,  body, or perhaps
more precisely, our personal existence centered on the body. This makes
it virtually synonymous with ńa^ yatana, which makes ńa^ yatana-via
eńa
and  a
raya-via
eńa equivalent.27
In the passage cited, the body of a noble person is distinguished from
the body of an unenlightened person in that the former no longer pro-
duces dełlements. In other words, the bodies of the unenlightened
remain in a state that produces dełlements and such bodies are conceived
of as the b! ja of dełlements.28
In the passage from the Bodhisattvabh�mi, the six-sense-basis (ńa^-
 yatana) of a bodhisattva is distinguished from that of a sravaka and a
pratyekabuddha in that the ńa^ yatana of a bodhisattva is capable, even-
tually, of generating the supreme wisdom of the Buddha. The six-sense-
basis of a bodhisattva who has the potentiality to give rise to the highest
wisdom of the Buddha is considered to be the bodhisattva-gotra.
The important point here is that the gotra is taken to represent a par-
ticular state of concrete physical-mental existence. In other words, the
basis for deliverance is posited on the phenomenal level namely,
sa�skŁta and not on the absolute level of asa�skŁta.29
We are reminded here of an important theoretical requirement of
abhidharma Buddhism, namely that an asa�skŁta or  unconditioned
dharma cannot be a generative cause of anything30 and therefore cannot
directly generate supramundane wisdom.31 This same pattern is followed
by the Yog c rabh�mi.32
Tathat is synonymous with dharmadh tu, which, according to
Hakamaya and Matsumoto, gives rise to all the mundane and supramun-
dane elements. It should be noted, however, that the role of tathat in
soteriological contexts of the Yogacara system (especially in the Yog c ra-
bh�mi ) is rather limited. If it is sometimes called the cause of holy dharmas
(namely supramundane wisdom), it is because tathat assists the arising of
198
THE IDEA OF DH�TU-V�DA
supramundane wisdom by becoming its cognitive object or  lambana-
pratyaya. In other words, meditation on tathat in the preliminary stages
eventually induces supramundane wisdom.33 But tathat , in principle,
does not become a generative cause,34 which means that once again the
suitability of the dh tu-v da model as  generative monism to the
Yogacara literature is suspect.35
Classical Yogacara theory regarding gotra maintains that the diver-
gent attainments of the three vehicles are each based on their own dh tu
or gotra. Needless to say, this is a highly problematic position, and one
needs to consider carefully why they were driven to take it. Merely in
terms of the formal logic involved, however, the argument is rather
straightforward: divergent effects must have divergent causes. This is
much easier to understand than the idea of divergent effects based on a
single cause, as the dh tu-v da model suggests.36 I do not believe that
the universal dharmadh tu was the leading principle that supported the
gotra theory of the Yogacara school.37
GOTRA AND DHARMADH�TU
The theory does not, however, hold universally. If we consider a typical
tath gata-garbha text, the Ratnagotravibh ga, we łnd that such a clear
distinction between sa�skŁta and asa�skŁta is not strictly observed.38 The
text does not hesitate to assert that the actions of the Buddha arise from
asa�skŁta.39 Even the prakŁtistha-gotra of the Bodhisattvabh�mi is linked
to the tath gata-garbha, which is equivalent to tathat .40 But if
prakŁtistha-gotra is equated with the omnipresent tathat , no theoretical
basis remains to support the absolute distinction among the three vehi-
cles. Thus, the Ratnagotravibh ga states:
Eventually the rays from the sun-disk of the Tath gata fall even on the
bodies of sentient beings łxed in state of evil (mithy tva-niyata-
sa�t na)& . The statement that an icchantika never attains nirvana was
made to remove the hatred against the teachings of Mahayana, because
the hatred against the teachings of Mahayana is the cause of one s being
icchantika. [In other words, this statement has] a hidden intention [that
even icchantikas will attain nirvana] at another, [later] time [if they aban-
don their hatred of Mahayana]. Indeed, because the originally pure gotra
exists (prakŁti-via
uddha-gotra-sa�bhav d), no one can be ultimately
impure by nature.41
199
YAMABE NOBUYOSHI
The dh tu-v da model of ultimate discrimination based on universal
buddhadh tu does not seem to work very well on this end either. Logic
requires that divergent phenomena must have divergent bases. If the basis
is universal, there is no logical reason to maintain an ultimate discrimina-
tion among the  super-loci. This structure is presented graphically in
Chart 4. Note that the  super-loci here are no longer discriminatory:
CHART 4
attainment of nirvana attainment of nirvana attainment of nirvana
prakŁtistha-gotra = dharmadh tu
DH�TU-V�DA
On the basis of the above arguments, I am persuaded that the classical
gotra theory of the Yogacara school was based on pluralistic dh tus or
gotras. Once the gotra is reinterpreted as universal dharmadh tu or
tathat , it loses the theoretical basis that supports the absolute distinc-
tion among the three vehicles.42
There is yet another factor to be taken into account. India is a coun-
try in which tradition holds great authority. Old theories are not directly
discarded when new ones come along, but are often retained and attempts
are made to reconcile the old with the new. Something like this seems to
have happened in the case of the gotra theory of the Yogacara school.
Actually, even within the Yog c rabh�mi the idea of tathat was
coming to play an ever greater role in the soteriological context. For
example, a portion of the Vinia
cayasa�graha�! section of the Yog c ra-
bh�mi says that the cause of supramundane dharmas does not lie in ordi-
nary seeds but in tathat as a cognitive object.43 The theoretical limitation
of tathat to the realm of a cognitive object, and hence its exclusion from
the realm of generative causes, seems still to be in force. And yet at the
same time one senses a dissatisfaction with the gotra theory of the
Bodhisattvabh�mi, which grounds supramundane attainment on the phe-
200
THE IDEA OF DH�TU-V�DA
nomenal ńa^ yatana-via
eńa. This appears to indicate an expansion of the
role of tathat into the realm of the soteriological.
Now if the source of supramundane attainment is not individual seeds
but universal tathat , there seems no reason to sustain the idea of indi-
viduals being predestined to a particular level of attainment.44 Never-
theless, the idea of distinct gotras is not abandoned, and the attempt is
made to defend distinct, predestined goals by claiming that some people
are faced with an ultimate obstacle that blocks the way into tathat , while
others are not.
At this point the argument, it seems to me, has ceased to be rational
and is simply concerned with preserving tradition. The Yog c rabh�mi is
not a coherent text but a composite of heterogeneous elements, some
more traditional, others more progressive. In general the material in the
Vinia
cayasa�graha�! is more developed than what we łnd in the basic
seventeen sections of the Yog c rabh�mi. Still, the Vinia
cayasa�graha�!
is part of the Yog c rabh�mi and presupposes what is contained in its
basic sections. And since the basic sections clearly present the traditional
gotra theory, the authors of the Vinia
cayasa�graha�! were not at liberty
directly to contradict or ignore that theory. This would appear to be the
historical dynamic at work behind the inconsistency between the monis-
tic tathat and the pluralistic three vehicles.
True, certain passages do suggest a dh tu-v da-type structure. A
good example is the verse of the Abhisamay la�k ra that we referred to
at the outset:
Because the dharmadh tu has no distinction, any distinction among
gotras is unreasonable. Nevertheless, because the dharmas to be posited
[on the  locus of dharmadh tu] are distinct, a distinction [among
gotras] is proclaimed.
The earliest extant commentary on the work, the Abhisamay la�k ra-
vŁtti of �rya-Vimuktisena, explains the meaning of the latter half of the
verse as follows:
[Gotras are differentiated, just as the vessels] made from the same clay
and baked in the same łre [are differentiated] by the distinct names
 honey pot,  candy pot, and so forth [according to what is put in
them].45
In Matsumoto s phraseology, the VŁtti would clearly hold that the
manifold  super-loci are posited on a single  locus, and that this
201
YAMABE NOBUYOSHI
homogeneous  locus is differentiated only in the sense that the hetero-
geneous  super-loci are differentiated one from another. It is precisely
this that he has chosen to name dh tu-v da. But if the distinction among
gotra amounts to no more than the superłcial temporary distinction
between a  honey pot and a  candy pot, it has ceased to characterize
anything essential. This is clearly different from the stringent gotra theory
of the Yogacara school.46 The great possibility for the conversion of
a
r vaka allowed for in the VŁtti seems to justify my suspicions.47 It seems
certain that the distinction among the three vehicles in the VŁtti is not
the absolute predestination of the Yogacara school.48
In addition, we might mention the following arguments:
Therefore, [the k rik I.5cd of the Abhisamay la�k ra]49 teaches that
the dharmadh tu itself, [namely] the gotra existing by nature
(prakŁtistha� gotra�), is the basis for practice, because [the dharma-
dh tu is] the cause of the holy dharmas& .
Others hold that gotra is the distinct state of the six-sense-basis
(ńa^ yatana-via
eńa), which is twofold: one attained by conditions
(pratyaya-samud n! ta) and one existing by nature (prakŁty-avasthita)& .
[This opinion is rejected.]50
[Objection:] If dharmadh tu is gotra, does it not follow that all [the
sentient beings] are [equally] established in the gotra (gotra-stha),51 since
the [dharmadh tu] is omnipresent?
[Answer: The dharmadh tu] is called gotra [only] to the extent that
it is recognized ( lambyam na) and becomes the cause of holy dhar-
mas& .52
The context here is very similar to that of the Vinia
cayasa�graha�!
on a number of points. First, the entire argument is necessitated by the
rejection of the classical Yogacara theory of gotra and the adoption of
tathat or dharmadh tu as the basis for supramundane attainment.
Second, tathat or dharmadh tu thus adopted aids the arising of supra-
mundane wisdom by serving as its cognitive object. Third, in spite of the
adoption of such a universal basis, the author does not discard the tradi-
tional distinction among the three vehicles outright, but strives somehow
to retain it.
Most likely the crucial factor once again was the authority of the pre-
ceding tradition. The Abhisamay la�k ra had close ties with certain
Yogacara texts,53 which makes it easy to understand how the Abhisamay -
la�k ra and its VŁtti could not completely neglect the traditional gotra
202
THE IDEA OF DH�TU-V�DA
distinction of the Yogacara school. At the same time, the Yogacara tradi-
tion seems not to have been so closely binding on the VŁtti as the Vinia
-
cayasa�graha�! . The logic of the VŁtti that supports the differentiation
of gotras is very weak. Hence, as we saw earlier, the distinction among
gotras in this text cannot be considered łnal.
CONCLUSION
In the compass of this essay I have only been able to cover a small portion
of the large number of texts analyzed by Hakamaya and Matsumoto, and
have had to exclude mention of any of the sutras. Obviously so limited an
examination cannot claim to have taken into account all the aspects of
tath gata-garbha thought. However, we can say that at least as far as the
sastra texts we have discussed are concerned, the applicability of the
dh tu-v da model of discrimination based on monism is questionable at
best. The Yogacara theory of gotra is indeed discriminatory, but it is not
based on monism. Tath gata-garbha thought is clearly monistic, but the
gotra distinction does not seem to signify anything essential. Some texts,
such as the Abhisamay la�k ra-vŁtti, show an apparent dh tu-v da
structure, but that structure would appear to be an inconsistency brought
about by a reinterpretation of the gotra theory. That is, replacing the plu-
ralistic gotras with monistic dharmadh tu did not entirely do away with
the traditional distinction among the three vehicles.
I have tried to illustrate this in Chart 5, which presents my under-
standing of the dh tu-v da structure. While it may look very similar to
Matsumoto s scheme, it is really a clumsy patchwork of the super-structure
of Chart 3 with the infra-structure of Chart 4, aimed at showing how the
infrastructure and superstructure of Chart 5 contradict one another. I
CHART 5
superstructure of
sravaka pratyekabuddha bodhisattva
Yogacara
inconsistent with
infrastructure of
dharmadh tu = prakŁtistha-gotra
tath gatagarbha
203
�'
�'
YAMABE NOBUYOSHI
remain rather skeptical of the position that the dh tu-v da had anything
like the solid structure that Hakamaya and Matsumoto seem to accord it.
I wish respectfully to acknowledge their contribution to Buddhist studies
in pointing out the problematics of the tath gata-garbha thought so far
overlooked. At the same time, I have tried to indicate the possibility of
alternative interpretations based on the philological arguments presented
in the foregoing.
204