BUDDHIST BOOKS AND TEXTS: EXEGESIS AND HERMENEUTICS

In the Buddhist tradition the practice and theory of scriptural interpretation faced conflicting sources and concepts of authority, a voluminous canon of relatively late compilation, and a complex history of interpretations that may be described as "hermeneutic pluralism." Furthermore, for some Buddhist traditions an emphasis on dharma (the eternal truths discovered by the Buddha) rather than on buddhavacana (the literal content of his message) reduces the significance of textual and historical constraints as part of a method of interpretation.

According to tradition, the Buddha was not the sole preacher of dharma. Even during the Buddha's life his disciples acted as missionaries, and their words were considered part of the "original" message of Buddhism. The texts affirm that at the Buddha's own behest the disciples began each sermon with the words "Evaṃ mayā śrutam ekasmin samaye" ("Thus have I heard on one occasion"). This formula presumably served as a guarantee of authenticity, or rather, of faithfulness to the teaching of the Master. Yet, the same introductory formula came to be used indistinctly for sermons attributed to the Master, to his disciples, or to mythical sages and deities.

It was also believed that there had also been previous buddhas, who had their own disciples, all of whom could have preached the dharma. These "Buddhists" from the mythical past could speak to human beings. Their words, as well as the "inspired speech" (pratibhāṇa ) of ancient and contemporary ṛṣis, gods, and spirits, could be regarded as dharma, and thus be prefaced by the famous formula.

Even traditions that believe that the canon was redacted and closed during the First Council at Rājagṛha, shortly after the Buddha's death (c. 483 bce), concede that not all Buddhist elders were present at that gathering, and that at least one group of "five hundred monks" insisted on keeping their own version of the teachings as they remembered them. All available evidence indicates that most of the canons were never closed. The Theravāda school, proud of its conservatism in scriptural matters, was still debating the content of its canon at least as late as the fifth century ce. Even today there is no complete agreement among Theravādins regarding the Khuddaka Nikāya section of their canon. Thus, it is not always possible to distinguish clearly between canonical, postcanonical, and paracanonical Buddhist literature.

All schools believe that at least some texts have been lost, truncated, or altered, and that a number of false or late texts have been incorporated into the canons of various schools. Even if occasionally these statements are used to bolster the position of one school over another, they probably represent an accurate description of the general state of things by the time the first scriptural collections were formally constituted. It is not difficult to see the impact that such a perception, combined with the mythology of revelation outlined above, would have on the tradition's view of the meaning of the scriptures and on the principles that should guide their interpretation.

Buddhist Exegesis: Methods of Interpretation

The Buddhist canons were the result of a long process of compilation and redaction that we can no longer reconstruct. For many centuries the task of interpretation was complicated by a shifting definition of canonicity. The first steps in understanding scriptural traditions—identifying the limits and forms of scripture—were slow and hesitant.

It took roughly three centuries before the oral texts of specialized schools of reciters (bhāṇaka s) were brought together into collections (piṭakas ). Another century went by before the first canons were committed to writing (an early Theravāda canon was written down under King Vaṭṭagāmaṇī of Sri Lanka, c. 32 bce). Even then the canons were not closed; some of the extant collections (i.e., the Tibetan and Chinese "canons"), were not compiled until more than a millennium had passed since the life of the founder, and they have remained open to the introduction of new literature until recent times.

The canon

Most Buddhists came to accept the theoretical division of scripture into three sections, metaphorically called "baskets" (piṭaka )—Sūtra (Pali, Sutta); Vinaya; and Abhidharma (Pali, Abhidhamma)—hence the name tripiṭaka (Pali, tipiṭaka ), or "three baskets." But in practice the corpus of authoritative Buddhist texts is not always divided into these three categories. This division is in itself of secondary importance for the history of Buddhist exegesis, whereas the variety of canons that seem to have existed in ancient India, and their flexibility, are important factors in the development of Buddhist attitudes toward canonical authority and interpretation.

The earliest system of this type was the classification of Buddhist teachings (and texts) into two main divisions: dharma (instruction on doctrine and meditation) and Vinaya (monastic rules and discipline). This early classification of genres probably was followed closely in time by the introduction of a third type of sacred text—the mātṛkā, or numerical list.

Also ancient, and obviously precanonical, is a system of "genres" (aṅga ). The Theravāda tradition distinguishes nine such genres, whereas the Sanskrit tradition counts twelve (sūtra, geya, vyākaraṇa, gāthā, udāna, nidāna, ityukta, jātaka, vaipulya, adbhutadharma, avadāna, and upadeśa ). Although some of these terms are well known as words for literary genres or forms of canonical literature, the exact meaning of the items in these lists is not always transparent. The list clearly shows, however, an early interest in analyzing scripture by literary forms, themes, and, presumably, speaker and audience.

By the time Buddhists began compiling their "canons," several forms of exegesis had developed within the body of literature transmitted as sacred scripture. Beyond the implicit exegetical work of the redactors, which is more obvious in Buddhist scripture than it is in the Judeo-Christian Bible, important sections of the canons are composed of exegetical material. Some works considered to derive directly from the Master's mouth are, in structure and reference, major acts of interpretation or statements on the nature of interpretation. Such, for instance, are the Mahāpadesa Suttanta, on forms of appealing to authority; the Kālāma Sutta, a critique of authority and an affirmation of the hermeneutical value of meditation; the Alagaddūpama Sutta, on the instrumental value of doctrine; and the Pratisaraṇa Sūtra, on the criteria of interpretation. These texts reflect an early concern with the problems of transmission and interpretation. Other texts included in the tripiṭakas are frankly exegetical in character (although they may be of more recent vintage); these include two commentaries (the Niddesa s) incorporated into the Sutta Piṭaka of the Theravādins, two works of theoretical hermeneutics included in the Sutta sections of the Burmese tipiṭaka, the Sūtravibhaṅga section of the Vinaya Piṭaka (an exegesis of the prātimokṣa ), and of course the totality of the Abhidharma Piṭaka.

The abhidharma as exegesis

The abhidharma played a central role in the development of the practice and theory of exegesis in all schools of Indian Buddhism. The hermeneutical strategy of the abhidharma was itself derived from a practice attested frequently by the sūtras: dogmatic lists known as "matrices" (mātṛkā ). These may appear to be mere catechistic or numerical lists; but more than topical indices or lists defining the limits of canonicity they are digests or exegetical guides. Some, evidently the oldest, are preserved in the Sūtra Piṭaka (e.g., Saṅgīti Suttanta, Daśottara Sūtra ), and were the object of commentaries (e.g., Mahākauṣṭhila's Saṅgītiparyāya ).

The role of mātṛkās as early canons of orthodoxy and interpretation is revealed by a legend, according to which the Buddha's disciple Śāriputra composed the Saṅgīti Suttanta in order to prevent a division in the Buddhist order similar to the one he had seen in the Jain community. The basic list, however, is not only a model for a definition of orthodoxy, it is also a pattern for exegetical coherence. The mātṛkās provide the structure for abhidharmic exegesis; each text must fit one or more of the categories contained in a traditional "matrix" or sets thereof. The "matrices" provided a simple logic of classification; all items of doctrine can be understood by opposites (the duka s of Pali abhidhamma : anything is a or not a ) or by contraries (the ṭīkā: x is a, or x is b, or x is neither a nor b ). Some of the earliest works of abhidharma, organized on this model (for instance, the Pali Dhammasaṅganī ), purport to reveal the underlying logic and structure of the sutta s.

Non-Mahāyāna exegetical literature

The abhidharma can be understood as a series of attempts at an exegesis of the whole body of Buddhist teachings (texts and practices). Some books, therefore, tried to preserve an explicit connection with the sūtras. But the abhidharma was more a work of philosophical hermeneutics than of exegesis. Accordingly, a different genre of literature was developed to carry out the difficult task of preserving, recovering, or eliciting the meaning of individual texts.

Two of the earliest Buddhist works of conscious exegesis have been incorporated into the canon in the Sutta Piṭaka. These are the Mahāniddesa and the Cullaniddesa, commentaries on the fourth and fifth books of the Suttanipāta. They date from approximately the third century ce. However, two other works of early but uncertain date occupy a much more important position in the development of Buddhist exegetical theories: the Nettippakaraṇa and the Peṭakopadesa, both attributed to a (Mahā) Kaccāyana (of uncertain date). These works may have been composed in South India or Sri Lanka. The Nettippakaraṇa formulates the principles of interpretation (netti ) common to both works on the basis of twelve techniques classified under the headings of "interpretation as to sense" (byañjana ) and "interpretation as to meaning" (attha ).

Earlier, at the beginning of the Common Era, the compilation of canonical collections and the explosion of abhidharma literature had created great works of synthesis on the Indian Peninsula. The most famous and influential of these was the Mahāvibhāṣā (c. 150–200 ce), a work of collective scholarship that attempted to make sense of the complex abhidharma literature of the Sarvāstivādins, in particular the Jñānaprasthāna of Kātyāyanīputra (or Kātyāyana; c. first century ce). Although it resulted in an equally abstruse work of doctrinal systematics, the Mahāvibhāṣā became an important source for doctrinal and interpretive categories, even for those who criticized it—especially the Mahāyāna.

Mahāyāna exegetical literature

In addition to playing the more obvious roles of criticism, reform, and systematic construction, the Mahāyāna sūtras fulfilled an exegetical role as well. It is common for a Mahāyāna sūtra to attempt a redefinition or reinterpretation of a classical formula from pre-Mahāyāna literature. The Vajracchedikā-prajñāpāramitā, for instance, presents the Mahāyāna reinterpretation of the "Parable of the Raft." The same text, in fact, the entire body of prajñāpāramitā literature, is devoted to what amounts to an all-out criticism of pre-Mahāyāna abhidharma. The traditional order established by orthodox exegesis is deconstructed in a search for the "ultimate meaning" behind the words of the older doctrines. The Laṅkāvatāra Sūtra and the Tathāgataguhya Sūtra —to quote another example—radically change the meaning of a classical passage by identifying truth with "holy silence." The earlier, canonical passage stated that "from the night of his enlightenment to the night of his parinirvāṇa …every word uttered by the Buddha was true." The two Mahāyāna sūtras changed the phrase to read "from the night of his enlightenment, to the night of his nirvāṇa, the Blessed One did not utter a single word."

These new departures, however, are not wholly the creation of Mahāyāna, for some of them are found in the literature of the Mahāsāṃghikas, one of the early nikāya schools. Some members of this school or community held that buddhas never pronounce a single word, yet living beings hear them preach. It was also claimed by some Mahāsāṃghikas that the Buddha can preach all things with a single word.

Another form of continuity within innovation occurs in Mahāyāna texts that follow the pattern of the abhidharma mātṛkā s as a way of redefining or expanding on earlier doctrine. Some of these sūtras may be rightly called "abhidharmic" Mahāyāna sūtras. Such are, for instance, the Dharmasaṇgīti and the Akṣayamatinirdeśa.

Some texts make explicit pronouncements on the principles of interpretation or evaluation of Buddhist scripture in general. For instance, the Laṅkāvatāra Sūtra 's statement on the silence of the Buddha is extended to mean that all words of the Buddha have only a provisional value. They are pronounced only in response to the needs of living beings who cannot penetrate directly into the mystery of the Tathāgata's silence. The Mahāyāna-mahāparinirvāṇa Sūtra offers a model for a hierarchy in scriptural study and understanding. At the first level, one becomes "learned" in scriptural study by studying all of the twelve genres (aṅga ) of scripture. Subsequently, one may study only one aṅga— the vaipulya (here equivalent to Mahāyāna) sūtras. Then one may study only the most subtle passages of the vaipulya section. But one may also study only one stanza of two lines from these sūtras and still be learned. Last, one becomes learned in scripture by understanding "that the Buddha never taught anything." The last of these is clearly, by implication, the most "learned."

Other statements with obvious implications for the interpretation of texts are those dealing with the relative value of various transmissions. Perhaps the best known of these formulations are those of the Saddharmapuṇḍarīka Sūtra (Lotus Sūtra ), which asserts that the Buddha has been in nirvāṇa eternally and denies that the Buddha ever "entered nirvāṇa," as claimed by earlier scriptural tradition. The Lotus also reduces the meaning of the human lives of the Buddha to a mere teaching device, developing the theory of "skillful means" (upāya ) as an explanation for the competing claims of its own brand of Mahāyāna and those of non-Mahāyāna Buddhists. Other texts establish criteria for authenticity that open doors to the new creative efforts of the Mahāyāna. The Adhyāśayasaṃcodana Sūtra, for example, establishes the well-known principle that "whatever has been well said has been said by the Buddha." Such statements are evident signs of a break with tradition and became seeds for further, perhaps dangerously limitless, innovation.

Śāstras and commentaries

In India the methodology used in the composition of technical treatises (śāstras ) was modeled on the commentarial tradition of Indian linguistics, heavily influenced by Patañjali's Mahābhāṣya. But the disciplines of poetics and logic also played an important role in the creation of standards for the composition of com-mentaries.

As Indian technical literature evolved through commentaries, continuity was preserved by reference to a common "root" text, which could be a scriptural text (sūtra) or the conscious work of an individual (śāstra ). A certain latitude for variation was allowed in the commentaries of each school of thought, but the root text was authoritative. That is, the commentary had to be verbally faithful to the root text and had to recognize its authoritative status. But the śāstra s themselves required commentaries, and some of the śāstra s that became the object of commentaries acquired quasicanonical status almost equivalent to that of the sūtras.

The system of authoritative texts followed by authoritative commentaries also produced a plethora of subcommentaries. The hierarchy was not always well-defined and the terminology not always consistent, but it was normally assumed that a sūtra would be the object of a commentary called a bhāṣya, or a more detailed gloss known as vyākhyā or ṭīkā. The root text of a śāstra, on the other hand, was often a versified treatise written in kārikā s or mnemonic verses (functionally parallel to the Hindu sūtras or prose aphorisms), and explicated in a bhāṣya or vṛtti (sometimes by the author of the root text).

The beginnings of the Mahāyāna commentarial literature seem to be coterminous with the development of independent philosophical or dogmatic treatises (śāstra s). But the exact dates of these events cannot be fixed with any certainty because an exact chronology depends in part on establishing the authorship of what may be the earliest major work in the genre, the commentary on the Pañcaviṃśatisāhasrikā-prajñāpāramitā Sūtra (Perfection of wisdom in twenty-five thousand lines), traditionally attributed to Nāgārjuna, whose dates are equally uncertain. This work, the Mahāpraj-ñāpāramitā-upadeśa Śāstra (preserved only in Chinese translation under the title Ta-chih-tu lun ), set the tone for Chinese exegesis, and defined some of the most important issues of Buddhist exegetical and hermeneutical theory for East Asia.

The genres of commentary and treatise flourished in India beginning (approximately) in the fourth century of the Common Era with Asaṅga and Vasubandhu. Although the tendency was to force the text into established scholastic molds, or to use it as a pretext for the formulation of independent philosophical dogmatics, commentators sometimes showed unusual sensitivity to the forms and structures of the text (e.g., Kamalaśīla's commentary on the Avikalpapraveśa ). There was also room for the development of independent criteria and exegetical guides. A valuable example of this type of work is Vasubandhu's extensive treatise on the mechanics of the commentary, Vyākhyayukti (preserved only in Tibetan translation).

Exegetical categories

Outside of India the exegetical issue perhaps became more critical as communities lost some of the sense of continuity with the authoritative tradition of the land of origins. The need for exegetical and hermeneutical principles was especially acute in China. An exegetical schema attributed to an early scholar of Chinese Buddhism, Dao'an (312–385), was based on three categories that supposedly reflected accurately the structure of all sūtras: the setting (nidāna ), the doctrinal and narrative core, and the transmission (parīndāna ). This basic schema was widely used in China (where it was known as the san-fen k'e-ching and was adopted by Zhiyi (538–597) in his classical analysis of the Lotus Sūtra, the Miaofa lianhua jing su. The schema is not attested in India until later, as, for example, in Bandhuprabha's (sixth century ce) Buddhabhūmi Śāstra, a commentary on the Buddhabhūmi Sūtra. In practice, each of the three parts was itself subdivided to account for obvious and important elements of style, narrative development, and so forth, but also to signal those passages considered core or defining statements. The text was expected to satisfy the traditional requirements for a definition of the audience (in the nidāna section) and for a positive assessment of the value of the faith and practice inspired by it (usually in the transmission section). Other schemas, some evidently inspired by a similar conception, developed one or more of the three parts. For instance, the sixfold division of the introductory formula of the sūtras (evaṃ mayā śrutam …) according to the manner proposed in the Mahāprajñāpāramitā-upadeśa Śāstra was extremely popular among East Asian exegetes. This led to multiple variations in the division of the text of the sūtras, such as Shandao's (613–682) threefold introduction and Zhiyi's five parts of the introduction.

The most fruitful and innovative, though perhaps less rigorous, strategies were those that tried to discover metaphysical meanings in the structure of texts. Such was Zhiyi's "twofold approach"—by way of the deep structure and by way of the surface structure (pen-chi erh men ) —to a text. This doctrine was part of a more complex exegetical plan, the four exegetical methods of the Tiantai tradition (T'ien-t'ai ta-ssu-shih ), which brings us closer to broader hermeneutical issues. According to this doctrine, any scriptural passage can be treated from four perspectives: (1) the passage as expression of a particular relationship between the audience and the Buddha (the circumstances determining the intention); (2) the passage as embodying one of four methods of teaching; (3) classification of the passage into one of the categories of absolute or relative statement; and (4) the introspective readings of the passage (guanxin ). In this methodological schema one can see in outline some of the salient features of Mahāyāna hermeneutics: contextual meaning, levels of meaning, and meditation as a tool of understanding.

Buddhist Hermeneutics: Theories of Interpretation and Canonicity

In traditional terms the fundamental questions of Buddhist hermeneutics can be classified into three broad categories:

1. If enlightenment (bodhi ) is at least theoretically open to all (or most) sentient beings, what is the role of sacred words and authoritative texts? How does one distinguish the exegesis of sacred texts from the actual transmission or realization of the dharma ?
2. Since the Buddha preached in so many different ways, adapting his language, style, and even doctrine to the spiritual disposition and maturity of his audience, did he have a plurality of messages or did he have a single truth to offer? If the latter, what was it, and, if the former, how is one to choose among his many teachings?
3. If Buddhism rejects conventional concepts of substance, self, possession, property, and referentiality, so fundamental to our conception of the world, is there a "higher language" that can be used to describe accurately reality as seen from the point of view of enlightenment?

From a modern perspective, one could characterize these three problems as defining the main subfields of Buddhist hermeneutics: the first statement addresses issues of Buddhist soteriology—the conflict between the ascetic-contemplative ideal and the institutional realities of Buddhism, between orthopraxy and orthodoxy. The second problem is that of Buddhist exegetical hermeneutics. Awareness of the late date and the diversity of "canonical" sources generates a "hermeneutical pluralism" that compounds the problem of determining the meaning of diversity and unity in tradition. The third item summarizes the problems of Buddhist philosophical hermeneutics: what are the relative positions and meanings of conventional language, its Buddhist critique, and "the silence of the sages"?

A better perception of the tone underlying the discussion of these issues can be derived from representative traditional responses to the problems:

1. The Buddhist dharma is not dependent on the historical event of Śākyamuni's enlightenment, ministry, and nirvāṇa. Whether a Tathāgata arises in the world or not, the basic teachings of Buddhism—impermanence, suffering, no-self, and liberation—remain facts of existence. Although the tradition initiated by Śākyamuni is a necessary aid to enlightenment, it is (in the metaphor of the Laṅkāvatāra Sūtra ) only a finger pointing at the moon. The moon is always there, waiting to be seen (whether there is a finger to point at it or not); the finger is not the moon. Nevertheless, some Buddhists, the Theravādins, for instance, would insist on the historical significance of Śākyamuni's life and ministry, and on the close connection between exact literal meaning of doctrinal statements and effective practice.
2. The diversity of teachings is not due to confusion or weakness in the transmission. On the contrary, it is proof of Śākyamuni's wisdom and compassion, of his ability to adapt to the needs, capacities, and dispositions of living beings. According to the Mahāpraj-ñāpāramitā-upadeśa Śāstra, his teachings were of four types, according to their definite purport (siddhānta ): worldly (laukika ), or surface meaning; therapeutic (prātipakṣika ), or meaning intended as an antidote to mental afflictions and passion; personal (prātipauru-ṣika ), or meaning intended for particular individuals; and absolute (pāramārthika ), or ultimate meaning. This second point is further complicated by the Mahāyāna belief in multiple buddhas and in the timeless saṃbhogakāya ("body of bliss") of the buddhas, which eternally preaches in the heavens, and beyond, and is seen and heard in the visions of bodhisattva s and sages. Other Buddhists, however, (the Theravādins, for instance) emphasize canonical integrity, rejecting both the doctrine of multiple meanings and the doctrine of multiple buddhas. The Mahāsāṃghikas seem to have wanted to forestall exegetical pluralism and protect the integrity of scripture by claiming that all sūtras have only one explicit meaning.
3. The above attitudes toward the sacred word are inseparable from Buddhist views on levels of meaning, whatever the historical or causal connections between these three problems may have been. The creation of abhidharmic technical language was the first step in separating two orders of truth and expression. Speculation about the nature of the path and the state of perfect freedom of buddhas further contributed to a theory of levels of meaning, since a progression in spiritual insight was taken to imply an increased capacity to penetrate behind the words of the doctrine.

The Mahāyāna insists that the higher sphere is only embodied in the silence of the āryas. The highest stage in the path, and, therefore, the highest order of meaning, can only be expressed in apophatic statements such as "appeasing all discursive thinking" (sarvaprapañcopaśama ) and "cutting out all doctrines and practices" (sarvavādacaryoccheda ). Still, all traditions, including the Mahāyāna, develop a language of the sacred (whether or not it is directly inspired by abhidharmic path theories), for it is necessary to explain holy silence in order to lead living beings to it. Thus, the culmination of this sort of speculation comes with the recognition that language, with all its limitations, is an important vehicle for salvation: language is upāya (e.g., in the Mādhyamika treatises, the Laṅkāvatāra, and the Tantras).

Criteria of authenticity

Early concepts of orthodoxy were based on doctrines of confirmation or inspiration, rather than on a literal definition of "the word of Buddha" (buddhavacana ). A disciple could preach, then receive the Buddha's approval, or the authority of his words could be implicit in the Buddha's request or inspiration. Although it may seem difficult to have maintained this fiction when the Buddha was no longer living among his followers, Buddhists did not always see things this way. Since the dharma is, after all, the Buddha's true body, and since it exists whether or not there is a human Buddha to preach it, one could assume that the preaching of dharma would continue after his death. This justification formed part of the context for the proliferation of texts and the elasticity of concepts of canonical authenticity. It may also explain in part why the abhidharma and, later, the commentarial literature achieved such a prominent role in the development of Buddhist doctrine.

Such flexibility does not mean, however, that no attempt was made to establish criteria or rules for determining the genuineness of any particular statement of doctrine. The Mahāpadesa Suttanta of the Dīgha Nikāya (and its parallel in the Aṅguttara Nikāya and in the Āgamas) recognizes four possible ways of appealing to or arguing on the basis of authority: one may appeal to the authority of the Buddha, a community of monks, several elders, or a single elder. The validity of these appeals and their potential sources of authority, however, must be confirmed by comparing the doctrinal statements attributed to these persons against the "sūtra" and the "Vinaya."

The Sanskrit recensions of the Mahāpadeśa Sūtra add a third criterion: statements of doctrine must conform "to the reality or nature of things" (dharmatā ). The same expansion is found in Pali literature in the Nettippakaraṇa, where the principles are actually applied to the analysis and evaluation of particular texts, and the three criteria are summarized in very suggestive language: "Which is the sūtra with whose approach [phrases and words] must agree? The four noble truths. Which is the Vinaya in which they must be seen? The Vinaya restraining covetousness, aversion, and delusion. Which is the dharma with which they must conform? The dharma of conditioned arising" (paras. 122–124).

The principle implies, of course, that whatever agrees with sūtra, Vinaya, and dharma (i.e., conditioned arising) carries authority for the Buddhist. If applied to texts this could mean that any new creation that is perceived as a continuation of the tradition (secundum evangelium, as it were) could have canonical authority. Indeed, the Mahāyāna used it in just this way to justify the development and expansion of earlier teachings. Theravādins, on the other hand, would understand the broad definitions of the Nettippakaraṇa as references to the letter of the canon, not to its spirit. Ultimately, then, the issue remained one of setting the limits of the interpretability of scriptural tradition.

What then is buddhavacana ? The Nettippakaraṇa passage epitomizes the Buddhist tendency to use philosophical rather than historical arguments for authority. But the three tests do not form a complete system of criteria for textual authenticity. In the canonical versions they seem to refer to statements of doctrine and appeals to authority, not to texts. Most certainly if they were meant to constitute a system for establishing canonical authenticity their value would be limited, if not totally inexistent, for as tests of canonical authority the first (and original) two criteria would be tautological. One must accept that the teaching on the "four appeals to authority" was originally a method for determining orthodoxy, not a criterion of authenticity. Still, since Buddhist notions of the "word of Buddha" were elastic, the principle must have been ambiguous. That is to say, the circularity of the argument that "a genuine sūtra is one that agrees with the sūtra" may not be so obvious in the context of Buddhist notions of canonicity—at least in the early stages in the formation of the corpus of scriptures.

Some Buddhists, the Mahāyānists in particular, came to consider "agreement with the sūtras," rather than "inclusion in the Sūtra Piṭaka," the ultimate test of authority. Thus, in China a distinction is made between, on the one hand, pseudepigrapha, or "spurious sūtras" (weijing ) that are nevertheless "canonical" (that is, in agreement with the spirit of the dharma and therefore to be accepted in the canon) and, on the other, those sūtras that are "false" (that is, in conflict with established Buddhist teaching) and therefore to be excluded from the canon. Both types of sūtra are not "genuine" only when contrasted to the "original" sūtras composed in India (which were themselves obviously much later than the time of the Buddha). Thus, given the history of the canon and the broad definition of "authoritative sūtra" current among all Buddhists, it must have been difficult to find any good reason for excluding a text only on historical grounds—to say nothing of establishing those grounds.

Interpretation

Three major sets of principles become central to the latter development of Buddhist hermeneutics. These are the "four reliances" or strategies for understanding a text; the "four types of intentional and metaphoric language;" and the "four modes of reasoning." Since the last of these three doctrines falls more under the rubric of philosophy, we shall omit it from this discussion.

Authority and interpretation

The problem of establishing criteria of interpretation cannot be completely separated from that of the hierarchy of authority. The interaction of both spheres of hermeneutics is seen clearly in the doctrine of the four "points of reliance" (pratisaraṇa ). This doctrine is found in several Mahāyāna versions, the most popular of which is the Pratisaraṇa Sūtra, a text no longer preserved in the original Sanskrit except in quotations. According to this text there are four criteria of interpretation: (1) relying on the nature of things (dharma ), not on the opinion of a person; (2) relying on the meaning or purport (artha ) of a text, not its letter; (3) relying on those passages that explicitly express the higher doctrine (nītārtha ), not on those that do not express it explicitly; and (4) understanding by intuitive realization (jñāna ), not by conceptual thought (vijñāna ).

Some of these principles are restatements of the Buddhist tendency to emphasize personal realization as the ultimate source of understanding. But this tendency is not without significant implications for a theory of shared or communicable meaning. A sūtra now lost in the original Sanskrit but preserved in Chinese and Tibetan translations, the Adhyāśayasaṃcodana Sūtra, says, "whatever is well spoken [subhāṣita ], has been spoken by the Buddha." This is perhaps the most extreme formulation of the Mahāyāna's view of the historical roots of its traditions. The system, nevertheless, is not totally open, for implicit in it are the earlier notions of the meaningfulness and appropriateness of the words of the Buddha (and by extension, of scripture). Conversely, then, "whatever is spoken by the Buddha has been well spoken" (as stated, for instance, by King Aśoka in his Bhābrā Rock Edict). Therefore, all passages of scripture must be meaningful—as well as agreeing with the dharma and leading to liberation. But it is not always apparent that scripture meets these standards. It is not evident that scripture speaks with one voice. The interpreter must therefore explain the hidden meanings that reveal the underlying unity of intention in scripture, whatever may be the grounds for its authenticity.

Types of intention

In his major works, Asaṅga mentions several methods of understanding that can be applied to scripture. Among these, the four types of explicit intention and the four types of implicit intention suggest the outlines of a hermeneutical theory. Implicit or contextual meanings (abhiprāya ) appear to be alternatives for decoding a passage—words intending an analogy, words intending another time frame, words intending a shift in referent, and words intended only for a specific individual. The four types of hidden intention (abhisaṃdhi ) show another aspect of the process of decoding the sacred text: introductory hidden intention (where the meaning is relevant only for the beginner), metaphysical hidden intention (where the meaning is a statement on the nature of reality), therapeutic hidden intention (where the meaning is realized by following the instructions in combating unhealthy actions or states of mind), and metaphorical intention (where the meaning is not the literal meaning, and often is paradoxical in character; e.g., referring to a virtue as a vice).

The concept of intentional speech brings to mind the third of the four points of reliance and the third principle of Tiantai exegesis. The common problem in these doctrines is best expressed by the classical theory of the two levels of meaning: the implicit or interpretable meaning (neyārtha ) and the explicit or self-evident meaning (nītārtha ). This is perhaps the most important doctrine of Buddhist hermeneutics. Under this doctrine, passages or complete texts can be taken at face value as statements of the "ultimate" teaching of the Buddha or they can be understood as teachings preached in response to provisional or individual needs, and hiding the core teacher under the textual surface, which then requires interpretation or clarification. If a passage is considered to be of the first type, then it is in no need of further elucidation. Its meaning (artha ) has already been brought out (nīta ) by the text itself. If a passage belongs to the second group, then the higher meaning can be found only through interpretation. It must be brought out (neya ) from underneath the surface meaning, so to speak.

Only the Mahāsāṃghikas seem to have rejected this distinction, claiming that all words of the Buddha mean what they literally mean, and therefore need no interpretation. But this extreme position was rejected by other schools, including the more conservative Sthaviras (Theravādins, Sarvāstivādins, etc.).

This hermeneutical schema is closely related to the doctrine of the two levels of truth, the relative or conventional saṃvṛtti and the absolute or ultimate paramārtha, arguably first developed in the Mādhyamika school. But the fundamental distinction in the case of implicit and explicit meaning is between modes of intention and meaning, rather than between levels of reality. The point at issue is not whether the words of the text are the ultimate truth (they are not), but whether or not they point directly and unambiguously to it.

Upāya

Also central to Mahāyāna understanding of the religious text is the concept that all forms of discourse are ultimately saṃvṛtti-satya (or at best a lower level of paramārtha ). This is based on both a radical critique of language (as in the Mādhyamika doctrine of "emptiness") and a revaluation of language as a means to an end. In the case of religious language, the end is liberation—the ultimate purpose and the ultimate meaning of all religious discourse. On the lips of the enlightened speaker, language becomes a "skillful means" (upāya ), pointing at or eliciting (udbhāvanā ) realization of the goal. The doctrine finds mythological expression when it is stated that the Buddha's preaching always conforms to the aspirations and maturity of his audience and that his pronouncements are instruments to guide sentient beings, not propositions expressing absolute truth. The religious text is upāya in at least three ways: it is a compassionate concession to the diversity, aspirations, and faculties of sentient beings; it is an instrument to be used in attaining the goal; and it is the expression of the liberating techniques of the Buddha.

Hermeneutics in the Tantras

Tantric hermeneutics presupposes Mahāyāna hermeneutics. Among other Mahāyāna principles, it accepts the four reliances and Asaṅga's eight types of intention. Of course, Tantric theories of interpretation also retain the fundamental distinction between implicit and explicit meaning. However, the intention is now clearly determined by the context of Esoteric practices. Thus we find again the close connection between interpretation and religious practice that characterizes much of Buddhist hermeneutics and that is also evident in scholastic speculations on the path (as in the abhidharma ); but here orthopraxy becomes central to textual interpretation. Here too, the traditions consider that the "root" text (mūlatantra ) requires an explanation (ākhyānatantra ), but as each text and school has specific practical contexts it is never assumed that a given hermeneutical scheme can be applied to all texts.

The Tantric Candrakīrti (c. 650 ce) explains in his Pradīpoddyotanā the basic principles of his school's hermeneutical system as applied to the Guhyasamāja Tantra. He explains this "root" text by means of seven analytic procedures called "ornaments" or "preparations" (saptaalaṃkāra ). All seven are used directly or indirectly to bring out the meaning of the text in interpretation or practice, but only two sets appear to bear directly on the issue of interpretation, being at the same time exegetical topics, hermeneutical tools, and levels of meaning. These are the six alternative interpretations of the meaning of words in a passage (ṣaṭkoṭikaṃ vyākhyānam ; Tib., mtha' drug or rgyas bshad mtha' rnam pa drug ), and the "fourfold explanation" (caturvidham ākhyānam ; Tib., tshul bzhi or bshad pa rnam pa bzhi ).

The six alternatives (with slight alterations to the order of the original) are: (1) standard terms used in a literal sense (yathāruta ); (2) nonstandard or nonnatural terms (aruta or na yathāruta ; i.e. esoteric jargon); (3) implicit meaning requiring interpretation (neyārtha ); (4) explicit or evident meaning (nītārtha ); (5) intentional or metaphoric lan-guage (saṃdhyāya bhāṣitam, saṃdhyā bhāṣā ); and (6) non-intentional language (no saṃdhyā ). Western scholars are not in agreement as to the hermeneutical function of this "ornament," but the Tibetan tradition considers the six alternatives an integral part of the interpretation theory unique to Tantra.

The "fourfold explanation," on the other hand, is generally accepted as a hermeneutical schema. In this case Candrakīrti makes an explicit connection between levels of meaning and the stages of spiritual growth (particularly according to the schema of utpattikrama and sampannakrama, the latter being divided into five stages or pañcakrama ). The four explanations are: (1) literal, surface, or natural meaning (akṣarārtha ); (2) shared meaning (samastāṅgam ); (3) hidden (garbhī ) meaning; and (4) ultimate (kolikam ) meaning. The first type of meaning is shared by those on the path and those who have not entered it. The second type is shared by the Mahāyāna and the Tantra (specifically those in the "initial" or utpatti stage). The third type is open only to those in the first three stages of the fivefold higher path. The last level of meaning is open only to those who have advanced to the fourth and fifth stages.

This schema is a subtle application of the basic principle of the explicit and the interpretable meanings. One must note, however, that in Tantric exegesis a single text or passage can be both nīta and neya, depending on the receptor of the message. This entails not only a complex hermeneutics, but also the possibility that the so-called direct meaning (nīta ) of one level requires interpretation (i.e., is neyārtha ) for those who are at another level of the path.

Progressive revelation

The Tantric hermeneutical schema manifests, even more transparently than earlier theories, one of the basic assumptions of Buddhist views of meaning: meanings (and "truths") are a function of the audience as much as, or more than, a function of the intention of the author. If this principle is extended into the cosmic or historical dimension, then new hermeneutical concepts can be derived from the doctrine that the Buddha can preach simultaneously to various assemblies of celestial bodhisattva s, adapt his teaching to the needs and faculties of diverse living beings, or preach only one message (in words or silence) yet be understood in different ways. This doctrine, applied to the historical reality of the conflict of authorities, transmissions, and interpretations, provides reasons, albeit ex post facto, for choosing and justifying any particular version of the many "teachings of the Buddha."

The turnings of the wheel

In context, statements about relative teaching and ultimate teaching are more sectarian and polemical in spirit than their mere abstract formulations suggest. When a text states that all teachings of the Buddha are only skillful means or empty sounds there is always a conflicting claim to the ultimate validity of the text's own interpretation of the "one true teaching" underlying the "provisional teachings." This is evident in all the classical statements—the Lotus Sūtra, the Laṅkāvatāra Sūtra, the Saṃdhinirmocana Sūtra, and others. In some of these texts we see attempts to formulate a historical argument in favor of doctrinal claims. The Buddha, the argument goes, preached in two (or three) major periods that divide his ministry as to location, audience, and depth of the teaching. These major divisions in the Buddha's ministry are called "turnings of the wheel of dharma." According to the most widely accepted doctrine (as presented in the Saṃdhinirmocana Sūtra ), there were three "turnings": the Buddha first preached the Hīnayāna teachings in Deer Park in Banaras. Then he preached the doctrine of "emptiness" (i.e., the Mādhyamika teachings) at Vulture Peak in Rājagṛha. Last, he preached, in the same place but when his disciples were more mature, the doctrine of "mind-only" (i.e., the Yogācāra doctrines).

There is, of course, a Mādhyamika version of this story in which the third turning is in fact that of emptiness and the second the "idealistic" doctrine of the Saṃdhinirmocana Sūtra. There is also a late Tantric version that adds a fourth turning: the revelation of Mantrayāna at Dhānyakaṭaka. Nevertheless, the scriptural weight of the Saṃdhinirmocana was such that the scholastics could not ignore its clear statement. Thus, Tsong kha pa (1357–1419), in his Legs bshad snying po, goes through the most subtle arguments to show that the sūtra's ordering of the turnings does not imply a privileged position for the doctrine of mind-only.

The formulation of pseudohistorical apologetics and hermeneutical strategies became popular in China. The Indian schema of the Three Turnings was adopted in the schools of Sanlun and Faxiang, while others developed autonomous systems: Huayan and Niepan (The Teachings of the Five Periods) and Tiantai (The Eight Teachings in Five Periods). The last of these, a synthesis created by the Tiantai monk Zhanran (711–782), divided and interpreted the scriptures according to the four types of doctrine used in the Buddha's preaching (huafa : the doctrines of the tripiṭaka, the common doctrines, the special doctrines, and the perfect or complete doctrine) and his four teaching styles (huayi : direct, gradual, secret, and indeterminate or variable). These eight forms of teaching were used in different proportions and combinations during five periods (wushi ) in the Buddha's ministry: the Buddhāvataṃsaka cycle, the Deer Park cycle, the Vaipulya cycle, the Prajñāpāramitā cycle, and the cycle of the Lotus and Nirvāṇa sūtras.