Tibetan and Himalayan Library - THL

THL Title Text
Tibetan Monastic Education
by Georges Dreyfus
January 1, 2001
Section 4 of 7

Procedures and Rules of Debate

Tibetan debates involve two parties: a defender (damchawadam bca’ ba), who answers, and a questioner (riklamparigs lams pa). The roles of defender and questioner imply very different commitments, as Daniel Perdue explains: “The defender puts forth assertions for which he is held accountable. The challenger raises qualms to the defender’s assertions and is not subject to reprisal for the questions he raises.”37 The responsibility of the defender is to put forth a true thesis and to defend it. Hence, the defender is accountable for the truth of his assertions. The questioner, on the contrary, is responsible only for the questions he puts forth. His questions must be well-articulated, must logically follow from the points already made, and must be relevant to defeating the defender. Their truth content is irrelevant, however, for his task is not to establish a thesis but to oblige the defender to contradict either previous statements or common sense.

The debate starts with a ritual invocation of Mañjuśrī, the celestial bodhisattva patron of wisdom: dhīḥ ji ltar chos can (pronounced “dhi jitar chöchen”). This invocation can be translated as “dhīḥ [the seed syllable of Mañjuśrī]; in just the way the subject.” Obviously, this statement is rather unclear and hence offers ample scope for various creative interpretations, as is often the case with ritual. Some scholars take the statement to mean “dhīḥ; in just the way [Mañjuśrī investigated] the subject.”38 Others, myself included, read it more simply as “dhīḥ; in just the way the subject [is investigated].” This invocation, however, also plays on the homophonic similarities of the syllable dhīḥ and this (di’di). Thus, the statement can also be heard as meaning “This is the way the subject is,” a statement that can be taken as having deeper implications (an explanation offered by Lati RinpochéBla ti rin po che). The subject, this, then refers to conventional objects and the predicate, is the subject, to the empty way in which they exist.

After this ritual invocation, the questioner proposes the topic of the debate in the form of a question, which seeks to elicit the defender’s thesis. The defender answers, stating his position. The questioner may then immediately begin the debate, or he may first seek auxiliary explanations to clarify the position of his adversary. The point of this crucial preparatory phase (jorwasbyor ba) is to establish a starting point for the debate, an area of agreement between the two parties. This may be one of the most delicate phases of an argument, especially if the two parties do not belong to the same tradition or monastery.

Such conflictual situations have long existed in the Buddhist tradition. In ancient India, debates pitched orthodox Hindu thinkers against their heterodox opponents (materialist, Jain, Ājīvika, or Buddhist). These debates, which had political ramifications, were often witnessed by the local authorities, and the stakes could be the conversion of one group to the views of the other. In Tibet, a similar debate is supposed to have taken place at the end of the eighth century when Kamalaśīla is said to have defeated the Chinese Chan monk Mo He Yan, thereby establishing the primacy of the Indian tradition.39 More recently, the young KhedrupMkhas grub is said to have debated and, according to GelukDge lugs accounts, defeated RongtönRong ston. Nowadays, such confrontations between scholars from different traditions are rare; they have been replaced among GelukpaDge lugs pas by debates between monasteries. Most debates take place within a single monastery, where the agreement between parties is easier to establish. Even then, however, fashioning that agreement is crucial and requires great skill. The questioner must dissimulate his real point and the defender must try to guess where his opponent wants to lead him.

Once the two parties believe that they agree on the understanding of the terms of the debate, the main part (ngözhidngos gzhi) can unfold through questions and answers.40 The questions are meant to draw out the consequences of the defender’s statements in order to oblige him to contradict himself or to take a blatantly absurd position. To succeed, the questioner must be able to take apart his opponent’s statements to draw out unwanted consequences. His opponent, the defender, must for his part attempt to block these contradictions by making further distinctions. In doing so, he must give one of the three allowable answers:41

  1. I accept (’dod).
  2. The reason is not established (ta madruprtags ma grub).
  3. There is no pervasion (khyappa majungkhyab pa ma byung).

These three answers derive from the link (or lack thereof) of the reason with the subject and the pseudo-predicate. The defender can say “I accept,” if he thinks that the consequence supports his position. Or he can say “the reason is not established,” when the reason does not correspond to the subject. For example, the consequence “It follows that all dogs are intelligent because they are primates” is faulty because dogs are not primates. Hence, the reason is not established. Or he can say “there is no pervasion” when the reason does not entail the pseudo-predicate as in the consequence “It follows that all dogs are primates because they are mammals.” In such a consequence, the reason is established but does not entail the pseudo-predicate. Or, to put it more literally, there is no pervasion.

As SapenSapaṇ noticed, these three answers differ from the Indian model. Most clearly a Tibetan invention is the third answer (there is no pervasion), which does not exist in Dharmakīrti’s debating tradition. There, defenders must make explicit whether the reason is contrary (gelwa’gal ba) or just uncertain (mangepama nges pa), that is, inconclusive. Thus defenders have four answers to choose from and need to make their response more specific.42 This difference illustrates the originality of Tibetan practices, which go beyond imitating Indian models in responding to the Tibetan context. Moreover, the failure of non-GelukDge lugs institutions to follow SapenSapaṇ’s recommendations regarding these answers shows again the domination of the SangpuGsang phu tradition of debate throughout the Tibetan world.

One of the three allowable answers must be given to all well-formed consequences. In order to be well-formed (like the examples above), a statement should contain three terms: a subject, a pseudo-predicate, and a reason. It should also avoid ambiguity. For example, if the questioner asks whether humans have male or female sexual organs, the defender will not be able to answer without disambiguating the subject (human). At that point, the defender must state his objection. In addition, consequences should not lead defenders into paradoxes (so judged by the rules of conversation). For example, consider the following question: “Did you bring back the computer you stole?” Such a statement cannot be answered straightforwardly without implying an admission of guilt. In all these cases, the defender should point to such a faulty formulation by saying, “the subject is faulty” (chöchen kyönchenchos can skyon can). But absent such faults, the defender has no choice. The ability to give one of the three allowable answers while making meaningful statements is a sign of a good defender. In difficult situations, defenders often try to muddle the situation and break the flow of consequences by saying, “What I really mean is. . .” These explanations are not accepted and reflect poorly on the defender. The questioner may reply, “No need to say much; just give one of the three answers.” Or he may mock the defender, pretending that his adversary is a great teacher about to give a sermon. In most cases, defenders are brought back, more or less gently, to the three answers.

Another element in debate is that defenders must answer quickly. Whenever a defender delays his answer, the questioner urges him on with rhythmic triple handclaps punctuated by the words “chirphyir, chirphyir, chirphyir” (i.e., chi chirci’i phyir, “why”). If this is a formal debate, the audience joins in, thus increasing the pressure on the defender. If the defender still does not answer, the questioner and the audience may start to tease him: “Are you here or are you absorbed in meditation?” If still nothing is forthcoming, the questioner either provides the required information himself or recasts his point in a simpler form. In this way, the debate is made so clear that the defender must answer. If he still cannot, a member of the audience is likely to step in and answer in his place. Such an outcome is humiliating for the defender, a sign that he is not up to the task.

It is in this framework that the debate unfolds strategically. The questioner tries to force his opponent either to contradict himself or to contradict common sense. To do so, he must be able to break down complex arguments into simple elements that can be expressed in a chain of well-formed consequences that follow each other logically. He must also keep track of the position of his adversary and where he wants to take him. The defender must figure out the questioner’s strategy and thwart his efforts, using only the three answers.

Let us take the example of a debate about the definition of impermanence, which is “that which is momentary.”43 The debate starts by delineating the agreement between both parties. The questioner may ask for further clarification, with such questions as “What does moment mean in this definition?” “Does it refer to a brief moment or to a longer one?” The defender may answer that the moment implied by momentary is brief. The questioner then proceeds to draw consequences, thinking that he has enough to go on. He may start, “It follows that things last only for a short moment since they are momentary.” This statement is framed to embody the defender’s answer concerning the meaning of momentariness and is considered the root consequence (tsawé telgyurrtsa ba’i thal ’gyur), which derives from the root thesis (tsawé damchartsa ba’i dam bca’) that the defender must be made to contradict.

The questioner proceeds by drawing out unwanted consequences intended to force the defender to give the no-pervasion answer that contradicts his explanation of the meaning of momentariness. For if the meaning of momentariness is to last only for a short moment, then being momentary must entail lasting for a short moment. To deny this and hold that there is no pervasion is thus tantamount to directly contradicting the thesis. Presented with the root consequence that embodies his view of the meaning of momentariness, the defender must try to thwart the questioner’s attempts by choosing the answer that he can defend and does not contradict his earlier point. In this example, he has one obvious choice: to assent to the consequence. The other possibility, the rejection of the reason as being not established, is less defensible, since it contradicts the fundamental Buddhist view that all things are momentary. And, as noted, saying that there is no pervasion would contradict his thesis concerning the meaning of momentariness. Hence, he will assent to the question, thereby agreeing with a classical interpretation of the Buddhist doctrine of impermanence. The questioner’s task is then to oblige the defender to back off from his acceptance of the root statement, forcing him to make the no-pervasion answer that contradicts his main thesis. To do so, the questioner will draw unwanted consequences from the defender’s position, pushing him to make counterintuitive statements until he reaches the point of absurdity. For example, the questioner will point to a mountain and state: “It follows that this mountain also lasts for a short moment since it is momentary.” If the defender still agrees, the questioner may point to the fact that it cannot exist just for a short moment since the mountain has been there for millions of years. He may also try to oblige his opponent to agree with blatantly counterintuitive statements. For example, he can ask: “Have you never seen any object lasting more than a moment? Have you never seen any object older than a moment? Have you never seen any person older than a moment?”44

Obliging the defender to make ridiculous statements is one of the ways for the questioner to gain the upper hand, forcing a reductio ad absurdum that can be seen as tantamount to a defeat of the defender. But although this way of ending the debate can be quite fun, it is not favored, for it is difficult to distinguish a blatant absurdity from an apparently counterintuitive but valid point. Hence, debaters prefer to end with their opponents’ self-contradiction. In our example, the defender may try to back away from the counterintuitive consequences that the questioner has drawn. He may agree that there are objects older than a moment. At this point, the questioner must take the defender back to his root statement and oblige him to contradict himself. He may state: “It follows that the mountain lasts for more than one moment, since it is older than a moment.” The questioner will try to resist; for example, he might attempt to make distinctions between “being older than” and “lasting for more than.” The defender must then try to block these attempts. If he succeeds, he will be able to take the defender back to his root statement, which he will restate: “It follows that things last only for a short moment since they are momentary.” At that point, the defender has an unenviable choice between two answers: reason not established and no-pervasion. The former implies the rejection of the fundamental Buddhistview that things are momentary. If he chooses this answer, a whole new debate starts. The questioner may try a rhetorical jab: “I thought you were a Buddhist!” But such a move can also backfire. The defender can turn the tables on the questioner, taunting him to establish this fundamental tenet: “I know what great masters such as Dharmakīrti would say. But let us see what you can do!” The questioner will then have to mount a new attack to oblige the defender to retract his rejection of this Buddhist tenet. If the questioner succeeds, the defender will have no other choice than to give the answer that here dooms him: “There is no pervasion.” This is the moment of triumph for the questioner, who will express his victory by saying: “the root thesis is finished (tsawé damcha tsarrtsa ba’i dam bca’ tshar)” or, more briefly, “oh, it’s finished (o tsaro’ tshar)”.45

This is the end of this debate, with the clear victory of the questioner. This victory is due to a direct contradiction between two statements and hence is easily detectable. Such clarity of outcome may explain why there is no formal role for a witness in Tibetan debates. Unlike Indian debates, which proceeded according to formal argumentative criteria on which a witness could adjudicate, Tibetan debates proceed through consequences aimed at exposing direct contradictions in the views of the defender. Detecting such contradictions does not require any special skills and hence the presence of a witness is not necessary.

Not all debates end in a defeat for the defender. Sometimes the questioner is unable to force the defender into contradicting himself and the debate ends in a stalemate. At other times, the defender gives an answer that establishes his view as being well-founded. In our example, the defender may succeed in maintaining that not all things are impermanent and hence escape contradicting himself. He might end his successful defense with a little rhetorical dig, marking his understanding that he has contradicted a basic Buddhist tenet for the sake of argument: “Fortunately, Dharmakīrti was smarter than you. If all Buddhists were like you, we would have long ago ceased to be Buddhist!” This is a clear victory for the defender, especially if he succeeds in making some good points in the process. Sometimes the debate ends abruptly when the questioner’s debate breaks down (taksel chértags bsal chad) and he is left without anything to say. When this happens in an individual debate, the embarrassment is minimal. But in a formal debate (damchadam bca’) the experience can be quite humiliating. The questioner may be left standing speechless in the midst of a large audience for a couple of extraordinarily painful minutes, until the abbot or the disciplinarian rescues him by bringing the debate to a merciful end. The defender may then make matters worse with a few unpleasant comments—for example, “You used to brag so much! Where is your debate now?” Most questioners manage to assert, often stammeringly, a few random consequences. However, it is clear to everyone that their debate has broken down and that they are just trying to avoid humiliation.

In the cases of such a victory or of a stalemate, there are no formal criteria according to which the debate can be adjudicated. For example, if the defender is ridiculed, there are no formal ways to determine what is ridiculous and what is not. The same is also true when the questioner’s debate simply fades away. In these cases, the outcome cannot be determined formally and hence there is little role for a witness. The outcome is left to the often conflicting opinions of participants and listeners. It is only in the case of a direct contradiction on the defender’s part that the outcome can be formally decided, a remarkable feature of Tibetan debates.


[37] Perdue, Debate in Tibetan Buddhism, 28.
[38] So Perdue puts it: “dhīḥ! The subject, in just the way [mañjuśrī debated]” (Debate in Tibetan Buddhism, 103).
[39] See D. S. Ruegg, Buddha-nature, Mind, and the Problem of Gradualism in a Comparative Perspective: On the Transmission and Reception of Buddhism in India and Tibet (London: School of Oriental and African Studies, University of London, 1989).
[40] SapenSapaṇ differentiates three phases in debate: a preparation (jorwasbyor ba) in which the two proponents seek a common basis, the main part (ngözhidngos gzhi) of the debate, and the conclusion (jukmjug) during which the witness summarizes the argument and establishes the winner (C. Beckwith, “The Medieval Scholastic Method in Tibet and in the West,” in Reflections on Tibetan Culture: Essays in Memory of Turrell V. Wylie, ed. L. Epstein and R. Sherburne [Lewiston, N.Y.: Mellon Press, 1990], 307-13). In actual Tibetan debates, the third element is absent since there is no witness.
[41] The three answers concern complete statements containing three terms (subject, pseudo-predicate, and reason). In the GelukDge lugs tradition, an incomplete consequence is allowable. Consider this statement: “It follows that the subject, a mountain, lasts only for a short instant.” Such a statement has only two parts, but is nonetheless considered well-formed since it allows a straightforward answer. An incomplete statement can be answered in two ways: either by accepting (’dod; lit., “[I] accept”) or by refusing (ci’i phyirci’i phyir; lit., “why”) it. In our example, these answers mean: I accept that the mountain lasts only for a short instant; or, I do not accept that the mountain lasts only for a short instant. If the latter answer is given, the rigs lams pa can immediately state the incomplete consequence: “It follows that the subject, the mountain, does not last only for a short instant.” The only possible answer to such a consequence is positive; any other is considered a breach of rules of rational discourse. Because other traditions have reconstituted their practice of debate by imitating GelukDge lugs practices, they follow this usage. Shākya ChokdenShākya mchog ldan seems to disagree with this practice, however, arguing that only three-part consequences should be used.
[42] David P. Jackson, Entrance Gate for the Wise (Section III) (Wien: Arbeitskreis für tibetische und buddhistische Studien Universität Wien, 1987), 361.
[43] Perdue, Debate in Tibetan Buddhism, 275.
[44] The defender is in a difficult position in this debate. Non-GelukDge lugs scholars would answer such a claim, which is often presented by GelukDge lugs scholars as a refutation of their views, by making a quasi-Humean distinction between the domain of reality, where duration does not exist, and the conceptual domain, where duration is necessary. In reality, there is no duration—only a succession of similar moments that create the false but useful impression that things last. True to their moderate realism, the GelukDge lugs tradition understands momentariness in a more commonsensical way. For them, a phenomenon is momentary not because it lasts only a moment but because it is composed of temporal parts and hence is in constant transformation. On these differences in the understanding of the concepts of impermanence and momentariness, see Dreyfus, Recognizing Reality, 63-65, 106-16.
[45] The exact meaning of the statement o tsaro’ tshar is not easy to establish. I have it as “Oh, it’s finished,” where oo’ is taken as an interjection. It could also be taken as directed at the opponent
#!essay=/dreyfus/drepung/monasticed/
Tibetan Monastic Education, by Georges Dreyfus