Nāgārjuna is traditionally credited with retrieving several Perfection of Wisdom Sutras, upon which the Mādhyamika system is based, from the land of the nāgas. He wrote numerous treatises, such as the Five Collections of Reasoning,1 commenting on and explaining the meaning of the Perfection of Wisdom Sutras. For these reasons, he is revered as the founder of Mādhyamika. His works and those of his student, Āryadeva, served as models and foundations for all subsequent Mādhyamika authors, such as Buddhapālita, Bhāvaviveka, Candrakīrti, Śāntarakṣita, Kamalaśīla, Śāntideva, and Atīśa. Nāgārjuna and Āryadeva were therefore labelled “Mādhyamikas of the Model Texts” (gZhung phyi mo’i dbu ma pa) by early Tibetan scholars. Later Mādhyamikas who disagreed on how the meaning of Nāgārjuna’s writings should be interpreted and took positions distinguishing themselves from each other were labelled “Partisan Mādhyamikas” (Phyogs ’dzin pa’i dbu ma pa, literally “Mādhyamikas who hold positions”).
When one seeks to discover among the Partisan Mādhyamikas exactly what positions were held and by whom, difficulties are encountered. In his Great Exposition of Special Insight, Tsong-kha-pa says, “I will delineate how the system of the great Mādhyamikas follows the master Nāgārjuna.” He then adds rhetorically, “Who could explain the minor ones?”2 He reports that some early Tibetan scholars classified the Indian Mādhyamikas according to their position concerning ultimates. Those who asserted that a composite of emptiness and appearance is an ultimate truth were called Establishers of Illusion through Reasoning (sgyu ma rigs grub pa, māyopamādvayavādin). Those who held that a mere elimination of superimpositions with respect to appearance is an ultimate truth were called Proponents of Thorough Non-Abiding (rab tu mi gnas par smra ba, apratiṣṭhānavādin). Tsong-kha-pa concurs with the assessment of the translator Lo-den-shay-rap (bLo ldan shes rab), who said in his Drop of Ambrosia (bDud rtsi’ thigs pa) that it is deluded to divide Mādhyamika from the point of view of assertions on the ultimate.3
Tsong-kha-pa explains that other early Tibetans classified the Indian Mādhyamikas in terms of their position on the conventional status of external objects. Those who asserted that external objects exist as entities separate from consciousness were labelled Sautrāntika-Mādhyamikas because their assertion accorded with that of the Sautrāntika school on that point. Those who asserted that external objects do not exist in that way were labelled Yogācāra-Mādhyamikas because they accorded with the Yogācārins on that point. This system of classifying the Mādhyamika schools was current during the eighth century for we find that the translator Ye-shay-day (Ye-shes-sde, c. 800), who was a student of both Padmasambhava and Śāntarakṣita, explains in his Difference Between the Views (lTa ba khyad par) that:
In the Mādhyamika treatises composed by the master, the Superior [Nāgārjuna], and his spiritual son [Āryadeva], the mode of existence or non-existence of external objects is not clarified.4 After that, the master Bhāvaviveka refuted the system of the Vijñāptikas [Yogācārins] and posited a system of the conventional existence of external objects. Then the master Śāntarakṣita, relying on Yogācāra texts, constructed a different type of Mādhyamika [system] which taught that external objects do not exist conventionally and that the mind lacks inherent existence ultimately. Thus, the Mādhyamikas arose as two types, the former called the Sautrāntika-Mādhyamikas and the latter, the Yogācāra-Mādhyamikas.5
Such a classification provides an accurate chronology6 and accounts for the assertions of Bhāvaviveka and Śāntarakṣita, and their followers, but does not provide a category for Candrakīrti because, although he upheld the conventional existence of external objects, his position on the nature of those objects does not accord with those of the Vaibhāṣikas, Sautrāntikas, or Yogācārins7 (see pp. 72-73).
It is only with the second dissemination of Buddhism to Tibet (marked by the translations of Rin-chen-sang-bo [Rin-chen-bzang-po, 958-1055] and the arrival of Atīśa in 1042) that the division of Mādhyamika which is considered standard today gained prevalence. Tsong-kha-pa reports that at the time of the second dissemination Tibetan scholars coined the terms Prāsaṅgika and Svātantrika, dividing Mādhyamika in terms of how the understanding of emptiness is to be produced in the mind of an opponent. This distinction was based on statements made by Candrakīrti in his Clear Words (Prasannapadā).8 It was during this period that Ba-tsap Nyi-ma-drak (Pa-tshab Nyi-ma-grags, died 1158) collaborated with the Kashmiri scholar Jayānanda to translate the major works of Candrakīrti including the Clear Words, the Supplement to the Middle Way (Madhyamakāvatāra), and his commentary to Nāgārjuna’s Sixty Stanzas of Reasoning (Yuktiṣaṣṭikāvṛtti).9
Thus, it is significant to note that the designations “Svātantrika” and “Prāsaṅgika” did not gain currency in India but in Tibet, and not until the eleventh century, not long before the eclipse of Buddhism in India. Although Bhāvaviveka is credited with founding the Svātantrika school, nowhere in his writings do we find him saying, “I am a Svātantrika.” Although Candrakīrti uses the adjective svatantra four times in his Clear Words,10 he never accuses Bhāvaviveka of being a “Svātantrika” or defends Buddhapālita as a “Prāsaṅgika.”
Furthermore, not only were the names unknown to the principals, the implications of those names seem to have escaped all but one, Candrakīrti. Tsong-kha-pa writes in his Essence of the Good Explanations (Legs bshad snying po) that although Bhāvaviveka found many faults with Buddhapālita’s commentary to Nāgārjuna’s Treatise on the Middle Way, he did not disagree with him on the selflessness of persons and phenomena.11 Bhāvaviveka’s commentator, Avalokitavrata, also finds no disagreement between Buddhapālita and Bhāvaviveka concerning the conventional and ultimate nature of phenomena. He writes:
The knowledge that external and internal dependent arisings conventionally exist as mere illusions capable of performing functions and that ultimately they are without entityness12 is the mode of the perfection of wisdom taught by the proponents of the middle path—the Superior [Nāgārjuna] and his spiritual son [Āryadeva], Bhāvaviveka, Buddhapālita, and so forth.13
Tsong-kha-pa goes on to note that none of the great Svātantrikas—Jñānagarbha, Śāntarakṣita, Kamalaśīla—found any difference between the selflessness of their own system and that of the system of Buddhapālita and Candrakīrti.14
Candrakīrti did not assert that there was any difference between himself and Buddhapālita concerning the ultimate and conventional. However he, and he alone among the great Mādhyamikas, distinguishes the system that he shares with Buddhapālita from that of the other Mādhyamikas. He writes in the autocommentary to his Supplement to the Middle Way (Madhyamakāvatāra):
I implore the wise to ascertain that just as, except for [Nāgārjuna’s] Treatise on the Middle Way, other treatises do not express this doctrine called “emptiness” without perversion, so other systems do not have anything like the doctrine of emptiness that appears in this system in which we state challenges and responses.15
Candrakīrti thus contends that his understanding of the doctrine of emptiness is unique and is not shared by others, such as Bhāvaviveka. Exactly how Candrakīrti and Bhāvaviveka differ on the meaning of emptiness will be considered later (see pp. 148–153). The important conclusion to be drawn here is that the difference between Prāsaṅgika and Svātantrika was by no means evident to the Svātantrika masters Bhāvaviveka, Avalokitavrata, Jñānagarbha, Śāntarakṣita, and Kamalaśīla. This is most surprising in the cases of Avalokitavrata, Jñānagarbha, Śāntarakṣita, and Kamalaśīla in that they postdated Candrakīrti16 and would have had access to his Clear Words and Supplement to the Middle Way in which Candrakīrti criticizes what came to be known as the Svātantrika position, at times mentioning Bhāvaviveka by name.17
It is noteworthy that the significance of the division of Mādhyamika into Svātantrika and Prāsaṅgika has proved confounding to Western Buddhologists. In the early part of this century, La Vallée Poussin wrote:
There are two branches of Mādhyamika, but the difference between them has not been studied, and seems to consist in a mere divergence in the method of demonstration.18
The implications of this “divergence in the method of demonstration” will be considered later in this chapter. In 1962 Edward Conze noted:
We still have no clear idea of Bhāvaviveka’s Svātantrika system, which can only be studied in Tibetan translations. ... Likewise we continue to be puzzled by the teachings and affiliations of the Yogācāra-Mādhyamikas who were responsible for the final synthesis of the Mahāyāna in India.19
More recently, K. Venkata Ramanan has written in his Nāgārjuna’s Philosophy that, “Buddhapālita and Bhāvaviveka belong to two different traditions (the Prāsaṅgika and Svātantrika) of the Mādhyamika School, although the difference between them is still far from clear.”20
If the Indian masters of the Svātantrika school did not perceive themselves as different from Buddhapālita and Candrakīrti, on what grounds did the eleventh century Tibetan scholars make the distinction between Svātantrika and Prāsaṅgika? From their vantage point, what essential differences were evident in the literature between Buddhapālita and Candrakīrti on the one hand and Bhāvaviveka, Śāntarakṣita, et al. on the other that would cause them to perform an ex post facto bifurcation of Mādhyamika? Is the distinction between Svātantrika and Prāsaṅgika innately evident from the Indian sources or was it an artificial distinction imposed from hindsight? The most articulate elaboration of the distinction between the two Mādhyamikas derives from Tsong-kha-pa and his Ge-luk followers, the most important of whom for our purposes are the tenet chroniclers Jam-yang-shay-ba and Jang-gya. Let us examine their arguments in an effort to determine the efficacy of the Svātantrika-Prāsaṅgika division .
Jang-gya’s ostensibly intimidating definition of Svātantrika provides a suitable starting point. The definition will be dissected and examined piece by piece and then reconstructed in an effort to gain a clear perception of the meaning and implications of the term “Svātantrika”:
Autonomous (svatantra, rang rgyud) means that an inferential consciousness (anumāna, rjes dpag) realizing the thesis (sādhya, bsgrub bya) is generated without taking the lead merely from the opponent’s assertions, but by his having ascertained the establishment of the modes (rūpa, tshul) of the sign (liṅga, rtags) with respect to a subject that is established as appearing commonly to non-mistaken valid cog-nizers (pramāṇa, tshad ma) of both parties in the debate through the force of an objective mode of subsistence from the side of the basis of designation. Mādhyamikas who assert the correctness of the necessity for such are Svatantrika-Mādhyamikas.21
The Instructions on Name and Gender (Nāmaliṅgānuśāsana) better known as the Amarakośa) by the sixth-century Buddhist lexicographer Amarasiṃha22 lists svatantra (own continuum) and svairi (own power) as synonyms.23 The ninth-century Sanskrit-Tibetan encyclopedia, the Mahāvyutpatti, glosses the Sanskrit term svatantra with the Tibetan rang dbang (“independent” or “own-powered”).24 Jam-yang-shay-ba states that the tantra of svatantra means “establish” (grub) or “prove” (sgrub) and concludes that:
Here, taking it as “establish,” svatantra (autonomous, own continuum) and svairi (own-powered) mean established from its own side, established from the side of its mode of subsistence, and established under its own power.25
Therefore, Jam-yang-shay-ba and Jang-gya, following Tsong-khapa, explain that the term svatantra means own-powered or autonomous. What are the philosophical implications of svatantra’s meaning autonomous, and do these implications apply to Bhāvaviveka and his followers, that is, in what sense are they “Svātantrikas”?
Jang-gya’s definition refers to an “inferential consciousness (anumāna, rjes dpag) realizing the thesis (sādhya, bsgrub bya)” and to “the modes (rūpa, tshul) of the sign (liṅga, rtags),” thereby indicating that svatantra has a special significance in the context of Buddhist logic. In order to elucidate that significance, it is necessary to digress for the moment to a consideration of proof statements (sādhanavakya, sgrub ngag).
Dharmakīrti’s Commentary on (Dignāga’s) “Compendium of Valid Cognition” (Pramāṇavarttika) says, “Because there are two objects of comprehension (prameya, gzhal bya), there are two valid cognizers (pramāṇa, tshad ma)”26 Buddhists hold that there are two valid types of awareness—valid cognizers—because there are two types of objects to be comprehended, manifest phenomena (abhimukhī, mngon gyur) and hidden phenomena (parokṣa, lkog gyur), which are known by direct perception (pratyakṣa, mngon sum) and inference (anumāna, rje dpag) respectively.
The definition of a hidden phenomenon is a phenomenon that the short-sighted must necessarily realize initially through relying on a sign.27 A short-sighted person (literally, “one who looks nearby”) is a common being (pṛthagjana, so so skye bo), a person who has not achieved the path of seeing (darśanamārga, mthong lam). A hidden phenomenon is one which cannot be perceived directly by persons who are not Superiors (ārya, ’phags pa), but which must be known through the use of a sign or reason. To cite a favorite example of the Buddhist logicians, although a person looking at a smoky mountain pass from a distance cannot see fire, the presence of fire there can be inferred from the presence of the sign, smoke, which is the effect of fire. Many of the most important phenomena in Buddhist philosophy fall into the category of hidden phenomena, such as subtle impermanence, the existence of former and later lifetimes, liberation from cyclic existence, the existence of a Buddha’s omniscient consciousness, and emptiness. The definition states that such phenomena must initially be realized in dependence on a sign. This refers to the important tenet that phenomena which are ordinarily inaccessible to direct perception can eventually be directly perceived by gradually transforming the imagistic perception by thought into yogic direct perception (yogipratyakṣa, rnal ’byor mngon sum).28 It is this yogic direct perception which, in the Mādhyamika school, directly realizes emptiness, a hidden phenomenon, and destroys the seeds for rebirth.
A manifest phenomenon is a phenomenon that the short-sighted do not necessarily realize through relying on a sign.29 Objects ordinarily within the purview of sense perception, such as colors, shapes, sounds, odors, tastes, and objects of touch, are manifest phenomena. Such phenomena may be realized by direct perception without resorting to inference. Direct perception is defined by the non-Prāsaṅgika schools as a non-mistaken knower which is free from conceptuality.30 Thought consciousnesses are, therefore, necessarily not direct perceivers because they are conceptual. Thought is also eliminated by the qualification “non-mistaken” in that thought is mistaken with respect to its appearing object (*pratibhāsaviṣaya, snang yul), mistaking the image of the object that appears to it—a meaning-generality (arthasāmānya, don spyi)—for the object which it is comprehending (prameya, gzhal bya). Direct perception, on the other hand, perceives the object directly, without the medium of a generic image.
The other form of valid cognition, inference, is defined as a determinative knower which, depending on its basis, a correct sign, is incontrovertible with regard to its object of comprehension, a hidden phenomenon.31 A determinative knower is one which explicitly identifies its object. It is incontrovertible in that it comprehends an object with respect to which superimpositions or misconceptions have been removed.32
A correct sign (samyakliṅga, rtags yang dag) is a correct reason (samyaknimitta, rgyu mtshan yang dag) or a correct proof (samyaksādhana, sgrub byed yang dag) and is the basis upon which an inferential valid cognizer is created. A sign is one component of a syllogism (prayoga, sbyor ba), the other being the probandum (sādhya, bsgrub bya) or that which one is seeking to prove. The probandum in turn is composed of two parts, the subject (dharmin, chos can) about which something is sought to be known (also known as the basis of debate) and the predicate of the probandum (sādhyadharma, bsgrub bya’i chos). An example of a syllogism is, “The subject, sound, is impermanent, because of being a product.” The probandum, also referred to as the thesis (pratijñā, dam bca’) or position (pakṣa, phyogs), is that sound is an impermanent phenomenon. Sound is the subject and impermanent phenomenon is the predicate of the probandum. Product is the sign or reason.
In order for the syllogism to be correct, thereby yielding an inferential valid cognizer, the sign must be correct. A correct sign in a particular proof must be the three modes (trirūpa, tshul gsum) of that proof, in fact, a correct sign is defined as “that which is the three modes.”33 The three modes are three relationships of the sign to other elements of the syllogism.
The first mode is the property of the subject (pakṣadharma, phyogs chos), which refers to the presence of the reason in the subject; in order to be a correct reason, that reason must be a property of the subject. In the Sanskrit term pakṣadharma, pakṣa refers only to the subject and not to the entire probandum. This is one of the four meanings of pakṣa according to Dharmakīrti.34 In the syllogism, “The subject, sound, is impermanent because of being a product,” the sign is a property of the subject because product is a property of sound, that is, sound is a product.
The second of the three modes is the forward pervasion (anvayavyāpti, rjes khyab) which is defined as that ascertained by valid cognition as existing only in the similar class in accordance with the mode of statement.35 The similar class (sapakṣa, mthun phyogs) is that class of phenomena of which the sign must be a member in order for the sign to be correct. In the proof that sound is impermanent because of being a product, all phenomena that are impermanent constitute the similar class. In order for product to be a correct sign, it must be ascertained as existent in only the similar class. To exist only in the similar class means that the sign must be pervaded by the predicate, that is, it must be coextensive with the similar class or a subclass of the similar class; whatever is the sign A must necessarily be the predicate B. Product exists in the similar class of impermanent phenomena because product and impermanent phenomenon are synonyms (ekārtha, don gcig).36 “Object of knowledge”, (jñeya, shes bya) would not serve as a correct sign in the proof that sound is impermanent because, although “object of knowledge” exists in the similar class of impermanent phenomena, it does not exist only in the similar class; there are objects of knowledge, such as nirvana, which are not impermanent.
The third of the three modes is the counterpervasion (vyatireka-vyāpti, ldog khyab) which is that ascertained as only non-existent in the dissimilar class in accordance with the mode of statement.37 The dissimilar class (vipakṣa, mi mthun phyogs) is that class of which the negative of the sign must be a member in order for the sign to be correct. In other words, it is that class of which the sign must not be a member. That the sign is only non-existent in the dissimilar class means that the negative of the predicate is pervaded by the negative of the sign, that is, whatever is not the predicate A must necessarily not be the sign B. In the proof that sound is impermanent because of being a product, the dissimilar class is the not impermanent, or permanent phenomena. Since product and impermanent phenomenon are synonyms, product is only non-existent in the dissimilar class of permanent phenomena, and whatever is a permanent phenomenon is necessarily a non-product.38
Based on a syllogism of which the sign has the three modes, an incontrovertible inferential valid cognizer realizing the thesis of that syllogism is produced. Syllogisms may be used to generate an inferential consciousness realizing the thesis in oneself or in others. In distinguishing Svātantrika from Prāsaṅgika, the use of the syllogism in generating a correctly inferring consciousness in an opponent is of primary importance. For example, the syllogism, “The subject, sound, is impermanent because of being a product,” might be stated to an opponent who believes that sound is permanent. According to the logical system of Dignāga and Dharmakīrti, the stater of the syllogism must accept the three modes of the sign.39
This is not the case with the other form of logical statement used by Buddhists in debate, the consequence (prāsaṅga, thal ’gyur). A consequence, like a syllogism, is used to bring about an inferential understanding in the opponent in a debate, but unlike the syllogism, it does so by using the opponent’s own assertions against him so that the opponent is unable to respond. A correct consequence is stated in such a way that the opponent cannot give a factually concordant answer without contradicting his own assertions.40 For example, to an opponent who holds that a sound is permanent, that sound is also a product, and that whatever is a product is impermanent, the following consequence might be stated, “It follows that the subject, sound, is not a product, because of being permanent.” In this case, the Buddhist stating the consequence does not accept all three modes of the sign; he accepts the forward pervasion and the counterpervasion, that whatever is permanent is necessarily a non-product and that whatever is not a non-product is necessarily not permanent, but he does not accept the property of the subject, that sound is permanent. The opponent, on the other hand, does accept the three modes. He accepts that the sign is a property of the subject—that sound is permanent—as well as the forward and counter-pervasions. He is thus unable to make any response to the Buddhist. He cannot respond that the reason is not established (asiddha, ma sgrub pa) as a property of the subject because he accepts that it is. He cannot respond that there is no pervasion because he accepts that whatever is permanent is necessarily not a product. He cannot even accept the thesis that sound is not a product because that contradicts his own assertions. He is left logically speechless. Such a consequence is called a consequence implying a proof, which is defined as a consequential statement capable of implying the three modes.41
The three modes implied by a consequence implying a proof are those of the syllogism of the correct position. That is, the consequence, “It follows that the subject, sound, is not a product, because of being permanent,” stated to an opponent who believes that sound is a product, sound is permanent, and whatever is a product is impermanent, should cause the person to reassess his position so that he may escape the unwanted consequence that sound is not a product. He thus arrives at the implied proof, “The subject, sound, is impermanent, because of being a product,” the three modes of which were implied by the consequence.
From this perspective, T.R.V. Murti is clearly wrong when he says in The Central Philosophy of Buddhism:
Prasaṅga is not to be understood as an apagogic proof in which we prove an assertion indirectly by disproving the opposite. Prasaṅga is disproof simply, without the least intent to prove any thesis.42
Murti here seems to be confusing the Mādhyamikas’ mode of proof with what they are proving. When a Mādhyamika states the consequence, “It follows that the subject, a pot, is not a dependent arising because of being truly existent,” to a Proponent of True Existence (*bhāvavādin, dngos smra ba), the Mādhyamika is implying that and wishes the opponent to understand that, “The subject, a pot, does not truly exist because of being a dependent arising.” The Mādhyamika thus employs consequences with the full intention of proving a thesis. The thesis that he wishes to prove in this case is that a pot does not truly exist. The non-true existence or emptiness of true existence of the pot is a non-affirming negative (prasajyaprati ṣedha, med ’gag), implying no positive phenomenon in its place. The emptiness of the pot is the mere vacuity of the true existence of the pot.43 Thus, in proofs of emptiness, the Mādhyamika uses consequences to imply a syllogism, but the thesis of that syllogism is a non-affirming negative and implies nothing positive about the subject.
In a consequence that implies a proof, the reason must be accepted by the opponent but must not be validly established; the opponent accepts that sound is permanent, but that is not correct. The opposite of the sign of the consequence will become the predicate of the implied proof. The forward pervasion of the consequence must be validly established. It becomes the counterpervasion of the implied proof. The thesis of the consequence must be refutable by valid cognition; the opposite of the predicate of the consequence becomes the reason of the implied proof, thereby establishing the property of the subject of the syllogism.44
Both Svātantrikas and Prāsaṅgikas use syllogisms and consequences. Consequences are used by both schools to break down the opponent’s adherence to this wrong view. However, a disagreement that rises out of the commentaries to Nāgārjuna’s Treatise on the Middle Way is Bhāvaviveka’s insistence on the ultimate necessity of explicitly stating the implied syllogism. Candrakīrti objects, saying that the statement of the consequence is sufficient and that the further statement of the syllogism is redundant for an intelligent opponent; the consequence alone is capable of producing an understanding of the implied syllogism.45
Thus, Prāsaṅgikas (Consequentialists) are so-called because they assert that an inferential consciousness that understands non-true existence is produced through using a sign the three modes of which are renowned to the opponent.46 The Svātantrikas hold that this alone is not enough, that the implied syllogism, the three modes of which are accepted by the Svātantrika himself, must eventually be stated to the opponent. This is one difference between Svātantrika and Prāsaṅgika. However, this difference is not merely formal, it is not confined merely to techniques of logical disputation. If this were the case, the Svātantrikas should be called the Prāyogikas, the Syllogists. Instead, they are called Svātantrikas because they use autonomous syllogisms (svatantraprayoga, rang rgyud kyi sbyor ba). The meaning of autonomous in this context and its implications become evident in examining the meaning of valid cognition (pramāṇa, tshad ma) in Svātantrika and Prāsaṅgika.
For the Svātantrikas, a direct perceiver is a consciousness that is free from conceptualization and is non-mistaken. Non-mistaken refers to apprehending the object’s own character just as it is. Since all five sense direct perceivers apprehend this “own-character” (svalakṣaṇa, rang mtshan), their objects’ own character is the object comprehended by direct perceivers. Therefore, the Svātantrikas assert that sense direct perceivers are valid cognizers in terms of the existence by way of the own character of their respective objects.47
Candrakīrti, however, argues that establishment by way of the object’s own character or by way of the object’s own entity does not exist even conventionally; therefore, he does not hold, as do the Svātantrikas, that sense consciousnesses are valid cognizers in terms of their perception of own-character. He does not deny that sense consciousnesses perceive objects as existing by way of their own character, but since he holds that that own-character does not exist, he rejects the Svātantrika assertion that the sense consciousnesses are non-mistaken. The Prāsaṅgikas assert that things exist one way but appear in another; specifically, the objects of the sense consciousnesses are not established by way of their own character but appear to be so established. Therefore, the sense consciousnesses cannot be valid with respect to the own-character that they perceive; they are deceived by the false appearance of the inherent existence of the five sense objects.48
The Svātantrikas respond that if forms, sounds, and so forth lacked natures that were established by way of their own character, they would lack any capability to perform functions. Therefore, if direct perceivers are not valid in terms of the own-character of the five sense objects, it is impossible for sense direct perceivers to be valid cognizers at all; if these consciousnesses are valid with respect to the five objects, they must be valid in terms of the own-character of the five objects.49
For Candrakīrti, if phenomena were established by way of their own character or their own entity, they would be truly established (satyasiddha, bden par grub pa). In that case, objects would exist as they appear, and consciousnesses which posit those objects would have to perceive correctly that true existence or existence by way of the object’s own character in order to be valid cognizers. However, since the objects are deceptive, appearing one way and existing another, the valid cognizer that posits them need not be valid in terms of own-character. Candrakīrti refutes that sense direct perceivers are valid cognizers with respect to own-character, but in doing so he does not deny that conventional consciousnesses can be valid cognizers in general.50
Here, the difference between the Svātantrikas and Prāsaṅgikas rests on two important points: what they understand “non-mistaken” to mean when they define sense direct perceivers as non-mistaken and, secondly, their understanding of the conventional status of phenomena. For the Svātantrikas, a sense consciousness is non-mistaken if it is free from four superficial causes of error:
1 a cause of error existing in the object (viṣaya, yul), such as seeing a whirling firebrand as a circle of fire,
2 a cause of error existing in the basis (āśraya, gzhi), such as an eye consciousness that perceives a double moon,
3 a cause of error existing in the abode (sthāna, gnas), such as seeing trees move when passing them in a boat, and
4 a cause of error existing in the immediately preceding condition (samanantarapratyaya, de ma thag rkyen), such as seeing red due to being angry.51
According to the Svātantrikas, if these four superficial causes of error are absent, a sense consciousness perceives the object as it exists. For the Prāsaṅgikas, all sense consciousnesses except those of a Buddha are mistaken because they perceive objects as existing by way of their own character. However, in order not to contradict the world, they uphold the existence of sense direct perceivers which are conventionally non-mistaken. A sense consciousness, due to being affected by ignorance, is mistaken in that objects that do not inherently exist appear to it to inherently exist. However, if that sense consciousness is not polluted by some external or other internal cause of error, it is non-mistaken conventionally.52 Such a sense direct perceiver is a valid cognizer for the Prāsaṅgikas because, although it is mistaken about the inherent existence of the object, it is valid in terms of the mere existence of the object.
The fundamental difference between the Svātantrikas and Prāsaṅgikas, however, and the reason why they differ on the meaning of valid cognition, is that the Svātantrikas assert that phenomena are established by way of their own character conventionally whereas the Prāsaṅgikas negate the existence of own-character even conventionally. Existence by way of the object’s own character is upheld by the Svātantrikas and rejected by the Prāsaṅgikas. This distinction was not made by any of the Indian Svātantrika masters, and it is not accepted by the Sa-gya (Sa-skya) order of Tibetan Buddhism.53 The distinction is made by Tsong-kha-pa and his followers based on careful scrutiny of the writings of Bhāvaviveka, Avalokitavrata, Jñānagarbha, Śāntarakṣita, Kamalaśīla, and Candrakīrti. What evidence does Tsong-kha-pa present to prove that the Svātantrikas assert own-character? How do we know that Bhāvaviveka believed that phenomena inherently exist conventionally?
Tsong-kha-pa says in his Essence of the Good Explanations:
There are many explanations in the Prāsaṅgika texts regarding the conventional existence of its [i.e., the object’s own] entityness, its nature, its own character and so forth and there are many [explanations that things] are not established by way of their own entityness, not produced by way of their own entityness, not substantially established, and so forth in the texts of this master [Bhāvaviveka]. Thus, it seems difficult to distinguish [Prāsaṅgika and Svātantrika]. However, this explanation of the meaning of the existence and non-existence of the entityness set forth in the Sutra Unravelling the Thought (Saṃdhinirmocana) is the clearest source [proving that] this master [Bhāvaviveka] asserts that things are established by way of their own character conventionally.54
The Perfection of Wisdom Sutras say, “This and that phenomena, from forms to omniscient consciousnesses, lack entityness.”55 The Sutra Unravelling the Thought explains this lack of entityness in terms of the three natures (trisvabhāva, rang bzhin gsum): the imaginary (parikalpita, kun btags), the dependent (paratantra, gzhan dbang), and the consummate (pariniṣpanna, yongs grub). Imaginaries are said to lack entityness in terms of character, the dependent lack entityness in terms of production, and the consummate are ultimate and lack entityness.
The Sutra Unravelling the Thought says, “Since imaginaries do not abide by way of their own character, they lack entityness of character.”56 In the twenty-fifth chapter of his Lamp for (Nāgārjuna’s) “Wisdom” (Prajñāpradīpa), Bhāvaviveka says:
If you [Yogācārins] say that the entityness of imaginaries, which are mental expressions and verbal expressions, does not exist, then you are deprecating things because of deprecating mental expressions and verbal expressions.57
In commenting on this, Avalokitavrata says:
This [statement by Bhāvaviveka] indicates that, with respect to the Yogācārin statement that imaginary natures lack entityness because of lacking entityness of character, it is unsuitable if they say that the entityness of “form” and specifically the entityness of mental expressions—thought consciousnesses—and verbal imputations—verbal expressions—lack entityness due to lacking an entityness of character because it is a deprecation of things that are conventionally dependent.58
The Yogācārins assert that because imaginaries, the imputation of names and attributes, do not have an entity that is established by way of its own character, they lack an entityness of character. Bhāvaviveka refutes this. He takes the category of imaginary phenomena to apply to both imputations and imputers, and regarding imputers, he says that if the Yogācārins assert that names, which are sounds, and thought consciousnesses, which impute entities and qualities, lack an entityness that does not exist by way of its own character conventionally, they are denying dependent phenomena. Therefore, Tsong-kha-pa concludes that it is clear that Bhāvaviveka asserts that dependent phenomena, that is, impermanent phenomena, are entities which are established by way of their own character.59 Bhāvaviveka apparently upholds this conventional nature but refutes that phenomena exist by way of their character ultimately because he says in his Lamp for (Nāgārjuna’s) “Wisdom”, “If [the Yogācārins] are refuting things ultimately, they are proponents who follow Mādhyamika.”60
By including imputers—terms and thought consciousnesses—in the category of the imaginary and saying that if the Yogācārins are holding that these lack an entityness of character, they are denying dependent phenomena, such as terms and thought consciousnesses, Bhāvaviveka implies that terms and thought consciousnesses and, by extension, all impermanent phenomena are established by way of their own character. This seems clear. Whether Bhāvaviveka is correct is suggesting that the Yogācārins classify imputers in the class of imaginaries is another question.
Although Bhāvaviveka rejects the Yogācārin interpretation of the three natures, he does not dismiss the three natures. For him, the factor of true existence that is falsely imputed to phenomena is an imaginary nature because it is superimposed on phenomena and does not accord with their mode of being. In this sense, the imaginary—true existence which is utterly non-existent—can be said to lack entityness of character in Bhāvaviveka’s system. Dependent arisings are dependent phenomena and lack entityness of production because they are produced and arise in dependence on another; they ultimately lack inherently existent or own-powered production. Emptinesses of true existence are consummate natures for Bhāvaviveka. Emptiness is an ultimate and a lack of entityness because it is the object of the ultimate consciousness and because it is the absence of the entity of true existence.61
Returning to the question of what constitutes valid cognition for Bhāvaviveka and the implications of his position for the differentiation of Svātantrika and Prāsaṅgika, it is important to scrutinize his understanding of the composition and status of sense objects. In the fifth chapter of the Blaze of Reasoning (Tarkajvālā), Bhāvaviveka presents his position on the nature of sense objects in response to a Yogācārin attack on those who assert external objects.
According to Bhāvaviveka, the Yogācārin argument is as follows. Individual minute particle (paramāṇu, rdul phra rab) of form are not the objects of sense consciousnesses because they are too subtle to be perceived by them. Furthermore, an aggregation (saṃghāta, ’dus pa) of minute particles is not the object of a sense consciousness because a mere aggregation, like a double moon, is not a substantial existent (dravyasat, rdzas yod). If individual particles are not sense objects and aggregations of many particles are not sense objects, sense perception of external objects composed of minute particles is impossible. Therefore, the Yogācārins conclude that there are no external objects.62
Bhāvaviveka responds that the Yogācārins’ argument that an individual minute particle is not a sense object is irrelevant because he also asserts that; the Yogācārins are needlessly proving what is already established. To the Yogācārin contention that a composite (saṃcita, bsags pa) of minute particles of the same type is not the observed-object-condition (alambanapratyaya, dmigs rkyen) producing a sense consciousness because of not being substantially existent, Bhāvaviveka responds that the reason is not established; he disagrees with the Yogācārins, asserting that such a composite is substantially existent. He argues that the entities of the minute particles of one type that aggregate as the entity of the object are connected to each other such that each minute particle contributes to the entity of the object. In that sense, each of the minute particles acts as a part of the entity of the sense object so that an awareness having the aspect of the appearance of an aggregation of subtle particles is produced.63
The Yogācārins agree that minute particles are substantially established but hold that they are too subtle to serve as sense objects. However, they also assert that a minute particle is an aggregation of the eight substances: earth, water, fire, wind, form, odor, taste, and tangibility. Bhāvaviveka argues that since it is accepted by the Yogācārins that a minute particle is itself an aggregation while being substantially established, it should be equally true that aggregations such as pots are substantially established.64
Tsong-kha-pa concludes that since Bhāvaviveka asserts that each particle in a composite of particles serves as a cause of a sense consciousness and that each of the particles is substantially existent and seems to assert that each particle is final (that is, most subtle), Bhāvaviveka asserts that partless particles are observed-object-conditions (ālambana-pratyaya, dmigs rkyen) of sense consciousnesses. Bhāvaviveka holds that objects composed of minute particles are established by way of their own character and substantially established and that the sense consciousnesses which perceive those objects are valid cognizers. Therefore, he asserts that a sense consciousness which is not affected by one of the internal or external causes of error is non-mistaken because it perceives an object which is objectively established as a composite of particles.65 Bhāvaviveka’s position on this point accords with that of the Sautrāntikas, except for the fact that the Sautrāntikas assert that sense objects ultimately exist whereas, for Bhāvaviveka, they conventionally exist. This general agreement with the Sautrāntikas on the nature of the observed-object-condition is the basis for calling Bhāvaviveka’s branch of Svātantrika the Sautrāntika-Svātantrika-Mādhyamika.66
It is important here to distinguish this position from that of Candrakīrti. Bhāvaviveka and Candrakīrti concur in their rejection of the Yogācārins, refutation of external objects. But despite the fact that they agree that external objects exist, they do not agree on the conventional mode of being of those objects. Bhāvaviveka holds that at the conclusion of searching for the object designated among its parts, one finds something which is objectively established.67
Candrakīrti asserts that because objects appear falsely to sense consciousnesses, when the object designated is searched for with analysis, it is not found. Therefore, he asserts that minute particles are mere appearances to the awareness designating them and that the objects of sense consciousnesses are imputedly existent phenomena which do not exist from the side of the object.68 Apparently referring to Bhāvaviveka, Candrakīrti says in his Supplement to the Middle Way:
Some [Mādhyamikas] propound that that which the Sautrantika system propounds ultimately, the Mādhyamikas assert conventionally. They propound [this] due to completely misunderstanding the suchness [taught] in the Treatise on the Middle Way.69
Again, in apparent reference to Bhāvaviveka, he says in his Commentary to (Āryadeva’s) “Four Hundred” (Catuḥśatakaṭīkā), “Those of our own school who, like the Vaiśeṣikas, assert that subtle particles are substantially [existent] are not correct.”70
Thus far, we have examined Tsong-kha-pa’s evidence for his assertion that Bhāvaviveka believes that phenomena are established by way of their own character conventionally. There are other statements in Bhāvaviveka’s writings that support this claim, and these will be considered in the next chapter in the discussion of the nature of the person and the selflessness of the person. It is necessary at this point to examine the implications of Bhāvaviveka’s understanding of the conventional status of objects and the valid cognizers which perceive them in the context of the syllogism. If Bhāvaviveka asserts that phenomena are established by way of their own character (svalakṣaṇa, rang mtshan, rang gi mtshan nyid), by way of their own continuum (svatantra, rang rgyud), what is the significance of an autonomous syllogism (svatantraprayoga, rang rgyud kyi sbyor ba), and how does it differ from a consequence?
That Bhāvaviveka supports the use of autonomous syllogisms is evident to Tsong-kha-pa, at least in part, from a statement Bhāvaviveka makes in his Lamp for (Nāgārjuna’s) “Wisdom”:
The thirteenth chapter was composed for the purpose of setting forth the lack of entityness of conditioned phenomena in terms of another aspect, through the force of answers to refutations [by opponents] and own-powered inferences.71
Avalokitavrata’s commentary says:
Here, regarding that, the purpose of the chapter is to indicate that conditioned phenomena are just without entityness [through] giving answers to the refutations by opponents and [through] flawless own-powered inferences.72
Here, Bhāvaviveka explains that the thirteenth chapter of Nāgārjuna’s Treatise on the Middle Way teaches emptiness by answering the refutations of opponents and by using autonomous or own-powered inferences. Inference in this context refers not to the consciousness but to the proof statement, a syllogism, intended to produce in the opponent a consciousness inferring the thesis. The fact that Bhāvaviveka refers to such inferences as “own-powered” may indicate that such proof statements are not merely stated in terms of the opponent’s position but that they have an autonomous status which is accepted by both parties. This interpretation of the term “own-powered” is strongly supported by two statements from Jñānagarbha’s Differentiation of the Two Truths (Satyadvayavibhaṅga), “Just as what is seen by cowherds and above abides as conventional truths ... ” and “No one has any quarrel concerning the factor that appears to the consciousnesses of the two parties [in a debate].”73 According to Jam-yang-shay-ba’s annotator, Nga-wang-bel-den (Ngag-dbang-dpal-ldan), the first passage is the clearest source proving that the Svātantrikas assert that phenomena are established from their own side because it says that things exist conventionally in accordance with how they are seen by cowherds and above (those of greater intelligence), and it is undeniable that whatever object appears to ordinary, common beings such as cowherds appears to be established by way of its own character, by way of its own mode of subsistence.74 It is important to note that Nga-wang-bel-den sees no reason to prove that objects appear to be established by way of their own character; it is something everyone would agree on. The disagreement arises from the fact that the Svātantrikas hold that this appearance is correct, that is, the inherent existence that appears to cowherds conventionally exists, and the Prāsaṅgikas hold that it is false—nothing inherently exists even conventionally.
In the context of debate, this problem manifests itself in the question of a commonly appearing subject (chos can mthun snang). A commonly appearing subject refers to a subject that is established within appearing commonly to both parties by a valid cognizer which is non-mistaken with respect to the subject’s being established by way of its own character. To be established as commonly appearing means that the subject is established for the challenger by the kind of valid cognizer by which that subject is established for the opponent.75 That is, the Svātantrikas assert that phenomena are established by way of their own character conventionally; therefore, this type of existence pertains to the components of a syllogism and their relationships, the three modes, as well. Thus, the subject, predicate, and reason exist by way of their own character. The presence of the reason in the subject and the presence of the reason in only the similar class of the predicate exist by way of their own character. Since the Svātantrikas hold that non-conceptual sense consciousnesses which are not defective due to not being polluted by a cause of error are non-mistaken, the objects with respect to which such subjects are non-mistaken must be established by way of their own character because they appear to be.76 That is, since phenomena appear to be established by way of their own character and also exist that way, according to the Svātantrikas, consciousnesses which correctly perceive that own-character are valid cognizers. Thus, when a phenomena, such as a form, is posited as the subject of a debate, that form appears to exist by way of its own character to the valid cognizers of both parties and they assent to that appearance; it appears commonly.77
Therefore, an autonomous syllogism is a syllogism which generates an inference realizing the probandum, without taking the lead from the opponent’s assertions, after having ascertained the subject and three modes of the sign with a valid cognizer which is non-mistaken with respect to their existence by way of their own character. Not taking the lead from the assertions of the opponent means that it is not sufficient that the subject and three modes be established by a valid cognizer that is renowned to the opponent alone. Once one uses autonomous syllogisms, one is admitting that the subject appears in the same way to both parties.78 Jñānagarbha says, “No one has any quarrel concerning the factor which appears to the consciousnesses of the two parties.”
There are other statements that can be identified that corroborate the Ge-luk position that the Svātantrikas assert that phenomena exist in the way that they appear, as established by way of their own character. Jñānagarbha says in the autocommentary to his Differentiation of the Two Truths: “Real production and so forth do not appear when a thing appears.”79 In other words, when something appears to a sense consciousness, its ultimate production, which is conceived by ignorance, does not appear to the sense consciousness. He says in the root text:
If you assert that [things]
As they appear are not [analyzed with] reasoning,
We agree with that.
If they were [analyzed with] reasoning, everything would be disturbed.80
Because of being entities [which exist] as they appear,
This [factor] is not analyzed.
If they were analyzed,
Since they would be other, that would be damaged [by direct perception].81
Nga-wang-bel-den glosses these two stanzas as follows:
If the Proponents of True Existence assert that phenomena exist as they appear to an awareness which is not damaged by another valid cognizer and assert that phenomena are not investigated by a reasoning consciousness which analyzes the ultimate, then Jñānagarbha agrees with their assertions. If phenomena as they appear to sense consciousnesses were analyzed with ultimate analysis, all presentations of conventionalities would be confused. Because phenomena are established as they appear to non-defective awareness, ultimate analysis is not applied to this existence by way of the object’s own character. If they were investigated with ultimate analysis, they would become other than they are, that is, non-existent. Therefore, that non-existence would be contradicted by direct perception.82
Jñānagarbha’s position is that ultimate existence is to be analyzed and refuted by reasoning, but existence by way of the object’s own character—the way that objects appear to non-defective awarenesses—is not to be analyzed. If it were the object of analysis, since it is a central tenet of Mādhyamika that the object of analysis is not found by the reasoning consciousness investigating the ultimate, the conventional status of phenomena would be negated such that presentations of the world would be impossible. Such things as cause and effect could not be posited for him if objects did not exist in the way that they appear. Such an appearance is certified by direct perception which precludes the possibility of its being scrutinized by ultimate analysis. Thus, for Jñānagarbha, objects must exist in the way that they appear; otherwise, they would not exist at all. He cannot conceive of the possibility of positing conventionalities within the emptiness of inherent existence.
Śāntarakṣita also seems to agree that the unanalyzed perceptions of non-defective awareness exist as they appear when he writes in his Ornament for the Middle Way, “From the context of children up to omniscient wisdom, I do not refute the pleasant nature of a single unanalyzed experience.”83
Thus, although the Svātantrikas assert that phenomena are not truly established (satyasiddha, bden par grub pa) ultimately established (paramārthasiddha, don dam par grub pa), or established in reality (samyaksiddha, yang dag par grub pa), they uphold that phenomena are established by way of their own character (svalakṣaṇasiddha, rang gi mtshan nyid kyis grub pa), from their own side (svarūpasiddha, rang ngos nas grub pa), and are inherently established (svabhāvasiddha, rang bzhin gyis grub pa). For the Prāsaṅgikas, true existence, ultimate existence, existence in reality, existence by way of its own character, existence from the object’s own side, and inherent existence are synonyms and are equally utterly non-existent.84 Therefore, a consciousness that establishes the subject and the three modes of the sign in terms of their existence by way of their own character is not a valid cognizer, it is a wrong consciousness (viparyayajñāna, log shes), according to Prāsaṅgika. Therefore, Prāsaṅgikas, who are seeking to prove that phenomena do not exist by way of their own character, reject the use of autonomous syllogisms because an autonomous syllogism requires a commonly appearing subject. Acceptance of a commonly appearing subject implies consent to the appearance of phenomena as being established by way of their own character. The Prāsaṅgikas reject such establishment and, therefore, in the light of their tenets, it is impossible for the subject of the debate to appear in the same way to a Prāsaṅgika and non-Prāsaṅgika.85 This appears to be Nāgārjuna’s point when he savs in his Refutation of Objections (Vigrahavyāvartinī, 30):
If something were observed,
Such as an object of direct perception,
It would be proven or disproven.
Because it does not exist, I am blameless.86
If an object of comprehension which was inherently established and established by way of its own entity were an object observed by a valid cognizer such as a direct perceiver, then it would be suitable that the object observed by such a valid cognizer be either proven or refuted in the opponent’s system. However, since in Nāgārjuna’s system, such an object of comprehension does not exist even conventionally, Nāgārjuna cannot be blamed for not having an inherently existent thesis.87 Thus, this passage seems to indicate that inherently existent objects of observation and observers do not exist; it is not suggesting that dependently arisen valid cognizers and objects of comprehension do not exist.88
Does this mean that the Prāsaṅgikas are unable to prove the emptiness of non-inherent existence? No, although they are prohibited from using autonomous syllogisms, they use consequences or other-approved syllogisms which do not entail any of the dangers of the autonomous syllogism. If autonomous syllogisms are not asserted, it is sufficient that the subject and three modes be established by a “valid cognizer”of a non-Prāsaṅgika opponent; these do not have to be established as inherently existent for the Prāsaṅgika, in fact, they cannot be.89
For the Svātantrikas, when two parties are debating about an attribute of an object such as permanence or true existence, the substratum of that attribute—the subject of the debate—must be established as commonly appearing, it must be demonstrable that the object upon which the attribute is based is objectively established; without an objectively established subject, it is impossible to prove the modes of the sign because the Svātantrikas hold that an attribute which lacks an inherently existent substratum is infeasible. This is the meaning of a commonly appearing subject. For the Prāsaṅgikas, the existence of a subject which is demonstrably established from its own side implies that it is established by way of its own entity which means that it would be ultimately established. Thus, for the Prāsaṅgikas, it is incorrect to assert the existence of a common appearance to both parties when proving qualities in terms of a substratum that exists by way of its own entity.90
For the Prāsaṅgika, since existence by way of the object’s own entity does not exist, such a commonly appearing subject does not exist. As long as the opponent has not generated the Prāsaṅgika view in his continuum, he is not able to differentiate existence by way of the object’s own entity and mere existence. He is not able to understand the modes of the sign in the Prāsaṅgika’s terms. He cannot be shown how phenomena are merely conventionally existent until he realizes the view of emptiness. Until this happens, the Prāsaṅgika is unable effectively to state reasons in his own terms because the opponent will not understand them, continuing to confuse inherent existence and conventional existence. Therefore, the Prāsaṅgikas assert that a mode of existence which is objectively established that appears commonly to both parties is not to be found; they assert that a commonly appearing subject does not exist.
Nonetheless, although the Prāsaṅgikas do not assert autonomous reasons, in order for reasoning to provide means of generating valid understanding, they must assert that the three modes of the sign can be ascertained and that the three modes are validly established. They are able to uphold this by using reasons which are renowned to others, taking the lead from the opponent’s own assertions. They state reasons that are renowned to the opponent himself or, from the point of view of the Prāsaṅgika, renowned to others; the terms “renowned to oneself’ (svaprasiddha, rang la grags pa) and “renowned to others” (paraprasiddha, gzhan la grags pa) are synonyms, with “oneself’ and “other” both referring to the opponent. The Prāsaṅgika is able to generate the view in his opponent by stating reasons, the three signs of which are established in terms of the opponent’s assertions, without the Prāsaṅgika having to assert the subject and the three modes himself.91
With all of the foregoing discussion in mind, Jang-gya’s definition of Svātantrika may be restated in the hope that its meaning will be clear, or at least clearer than it was upon first reading.
Autonomous means that an inferential consciousness realizing the thesis is generated, without taking the lead from the opponent’s assertions, but by his having ascertained the modes of the sign with respect to a subject that is established as commonly appearing to non-mistaken valid cognizers of both parties in the debate through the force of an objective mode of subsistence from the side of its basis of designation. Mādhyamikas who assert the correctness of the necessity for such are Svātantrika-Mādhyamikas.
If the Svātantrikas and Prāsaṅgikas differ so radically on the meaning of valid cognition, the means of generating the view in an opponent, and the conventional status of phenomena, in what sense may they both be called Mādhyamikas? Why are Svātantrika and Prāsaṅgika divisions of the same school? Candrakīrti remarks in his Clear Words that, “If one is a Mādhyamika, it is unsuitable to make use of autonomous inferences because another [inherently existent] position is not asserted.”92 Tsong-kha-pa says in his Great Exposition of Special Insight:
In general [Candrakīrti’s opponents] are both the Proponents of True Existence, who assert that things have an ultimate nature, and the Svātantrikas, who assert that they have a nature which is established by way of its own character conventionally. The Svātantrikas are also called Proponents of No Entityness (niḥsvabhāvavādin, ngo bo nyid med par smra ba) but here, for the sake of brevity, the proponents of non-inherent existence should be understood to be the Prāsaṅgikas, and the proponents of inherent existence should be understood to be both the Proponents of True Existence and the Svātantrikas.93
The twentieth century Tibetan scholar Sha-mar-gen-dun-den-dzin-gya-tso (Zhwa-dmar-dge-’dun-bstan-’dzin-rgya-mtsho) raises the argument that the Svātantrikas are not Mādhyamikas because they assert that emptiness—the mere, absence of true existence—exists by way of its own character and that all phenomena are established by way of their own character, thereby falling to both extremes.94
Despite these statements suggesting the contrary, the Svātantrikas are considered to be members of the Mādhyamika school. Tsong-kha-pa explains Candrakīrti’s comment that it is unsuitable for Mādhyamikas to use autonomous syllogisms, contending that it is indeed unsuitable for Mādhyamikas to assert that phenomena exist by way of their own character but it does not follow from that that anyone who asserts that phenomena so exist is not a Mādhyamika, that is, there may be Mādhyamikas who assert inherent existence. For example, he says that although it is unsuitable for a monk to break vows of their formulated code of conduct, there are monks who go against that code but who are nonetheless monks.95
Still, just as there must be a certain level of conformance with the code of conduct for a person to be considered a monk, there must also be certain assertions which classify the Svātantrikas as Mādhyamikas. Jang-gya explains what they are. He concedes that the Svātantrikas assert that phenomena exist by way of their own character conventionally and that they hold that the object designated is findable among its bases of designation. However, they assert that all phenomena in the universe are empty of true existence, which they define as being established by way of a mode of subsistence which does not depend on appearing to a non-defective awareness. (This will be explained in chapter 4). This emptiness is accepted by the Prāsaṅgikas (although they would say that it is a coarse emptiness); it is not false as is the Yogācārin emptiness of subject and object being different substantial entities. It is also not a trifling emptiness in that it applies to all phenomena equally; everything lacks true existence. It is compatible with dependent arising, allowing the Svātantrikas to abandon the extremes of existence and non-existence and to establish the viability of conventionalities such as cause and effect within emptiness.96 Tsong-kha-pa notes that, “These scholars also refute the tenet of the true existence of phenomena through many doors of reasoning and assert non-true existence well. Hence, they are Mādhyamikas.”97 Finally, Nga-wang-bel-den says in his Annotations (mChan ’grel) to Jam-yang-shay-ba’s Great Exposition of Tenets:
Because they propound a middle way free from the extreme of permanence—that phenomena ultimately exist—and the extreme of annihilation—[that phenomena] do not exist conventionally, and because they assert that an entityness which is truly established does not exist, they are called Mādhyamikas and Proponents of No Entityness. The divisions are two, Prāsaṅgika and Svātantrika; those who claim to be Mādhyamikas, other than those two, are merely designated as Mādhyamikas; it is unsuitable that they be Mādhyamikas in fact. Because they are similar in refuting the extreme of non-existence—[that phenomena] do not exist even conventionally—and the extreme of permanence—that things truly exist—and assert a middle which is a non-affirming negative, the mere elimination of true establishment in dependently arisen phenomena, both the Prāsaṅgikas and Svātantrikas are Mādhyamikas. However, that [school] which is free from all subtle extremes of permanence, such as [phenomena] being established from their own side, and extremes of annihilation, such as forms not being established as external objects, is the system of the Prāsaṅgikas.98