Chapter 3 Śāntarakṣita and Kamalaśīla: A Synthesis of Madhyamaka and Yogācāra¶

The preceding two sections examined how Dignāga and Dharmakīrti established Indian Buddhism's system of epistemology. Philosophically, that tradition belongs to Yogācāra, or consciousness-only thought, and concerns the conditions of valid cognition. This section turns to another attempt at synthesis, by Śāntarakṣita (c. 725–788) and his disciple Kamalaśīla (c. 740–795). It examines how they integrated Madhyamaka, especially Svātantrika Madhyamaka, with Yogācāra. It also traces how they brought that system to Tibet and profoundly shaped Tibetan Buddhism's philosophical orientation.

The significance of this synthesis extends beyond philosophy. Śāntarakṣita and Kamalaśīla were among the last great Indian Buddhist thinkers to enter Tibet and exert a lasting influence. The framework of learning they brought established Tibetan Buddhism's earliest intellectual structure. It also gave Tsongkhapa's later reforms a concrete point of reference to engage with and move beyond.

I. Historical Background: Scholarly Life at Eighth-Century Nālandā¶

Śāntarakṣita and Kamalaśīla were active in the eighth century, when Nālandā and Vikramaśīla were centers of learning. Several philosophical traditions operated alongside one another there. Prāsaṅgika Madhyamaka, in Candrakīrti's line, used reductio and rejected all affirmative claims. Svātantrika Madhyamaka, in Bhāvaviveka's line, allowed independent arguments and retained some particular characteristics at the conventional level. The Dharmakīrti–Yogācāra tradition supplied a precise epistemology, while tathāgatagarbha thought centered on the universal presence of Buddha-nature.

In this setting of several interacting traditions, Śāntarakṣita offered an organized way to integrate the divergent positions. Instead of reconciling the traditions, he assigned them to different levels, allowing each to do its own work at a distinct philosophical level.

II. Śāntarakṣita's Life and Works¶

Life

Śāntarakṣita (c. 725–788), whose Tibetan name is rendered in Chinese as "Xibei Sangbo" (希贝桑波), was born in eastern India and trained at Nālandā. He later became one of Vikramaśīla's abbots and was among the most important thinkers of his time.

Around 763, the Tibetan king Khri-srong-lde-btsan (742–797) invited Śāntarakṣita to Tibet to help Buddhism take root. Śāntarakṣita entered Tibet twice and, with Padmasambhava (c. eighth century), participated in founding Samye Monastery (Bsam-yas, c. 779), Tibet's first formally established Buddhist monastery. His teaching of Madhyamaka laid Tibetan Buddhism's earliest philosophical foundations.¹

Major works

The Tattvasaṃgraha is one of the largest single treatises in Indian philosophical history. Kamalaśīla wrote an extensive commentary, the Pañjikā. Together, the two are an important source for studying eighth-century Indian philosophical debate.²

III. The Madhyamakālaṃkāra: A Program for Synthesis¶

A two-step approach

The root verses of the Madhyamakālaṃkāra propose a concise philosophical program. It first uses Yogācāra arguments to deny the real existence of external objects, then Madhyamaka arguments to deny the real existence of consciousness itself, reaching thoroughgoing emptiness. These two steps were noted in 6.5; here we develop their internal logic.

Step one: Yogācāra negates external objects

Śāntarakṣita accepts Yogācāra's central claim that external objects are not entities existing independently of consciousness. The argument rests on the Dignāga–Dharmakīrti epistemological framework. All cognition can encounter only the "object aspect," the appearance of an object manifested by consciousness. If external objects are held to exist independently of consciousness, how cognition can cross the gap between subject and object becomes impossible to explain clearly. At the conventional level, therefore, only consciousness and its manifestations exist; there are no independent external objects.

This is the crucial difference between Śāntarakṣita's Yogācāra Madhyamaka and Bhāvaviveka's Sautrāntika Madhyamaka. Bhāvaviveka accepts the real existence of external objects conventionally; Śāntarakṣita accepts Yogācāra's denial of external objects, leaving only manifestations of consciousness at the conventional level.¹

Step two: Madhyamaka negates consciousness

But if only consciousness exists conventionally, what is its status? Yogācāra, especially Dharmapāla's line, accepts consciousness as "established by its particular characteristics." It treats consciousness as dependently arisen and as having some degree of real existence. Śāntarakṣita introduces Madhyamaka reductio here. The analysis of the four alternatives of arising rejects arising from itself, from another, from both, and without a cause. Consciousness with intrinsic nature therefore cannot be established. At the ultimate level, consciousness also lacks intrinsic nature and is empty. If even consciousness is empty, all dharmas are empty: the Madhyamaka conclusion.

The connection with Dharmakīrti's epistemology

As a Svātantrika Madhyamaka thinker, Śāntarakṣita accepts a limited reality at the conventional level. Things can be "established by linguistic convention," including manifestations of consciousness and the causal relations between them. This conventional establishment preserves some particular characteristics in an epistemological sense, sufficient to support the operation of Dharmakīrti's system of inference conventionally.

This is where Śāntarakṣita connects with Dharmakīrti's epistemological tradition. He assigns Dharmakīrti's epistemology—particular and general characteristics, causal efficacy, and so forth—the status of "epistemology valid at the conventional level." Madhyamaka's analysis of emptiness critiques its metaphysical basis at the ultimate level. Each has its place; neither replaces the other.

IV. The Tattvasaṃgraha: A Systematic Refutation of All Philosophical Schools¶

The Tattvasaṃgraha aims to refute every major Indian philosophical tradition of its time point by point, including both Brahmanical schools and positions within Buddhism, to demonstrate the superiority of the Madhyamaka–Yogācāra synthesis. Its twenty-six chapters address the following principal targets:²

The work's historical importance lies in its comprehensive record of eighth-century Indian philosophical debate. Each school's claims, arguments, and the ways they were refuted are presented in detail.

V. Kamalaśīla's Life and Works¶

Kamalaśīla (c. 740–795), Śāntarakṣita's most important disciple, broadly continued his teacher's Yogācāra–Madhyamaka synthesis. He also made independent contributions to the theory of practice and the Lhasa debate.

He followed Śāntarakṣita to Tibet and taught philosophy there. According to Tibetan sources, he died around 795 after being attacked by several assassins. Legend presents this as revenge by the Chinese Chan side for its defeat in debate. The detail cannot be historically confirmed, but it reflects the tensions surrounding the debate.

Major works

VI. The Bhāvanākrama: The Epistemological Basis of the Gradual Path¶

The three Bhāvanākrama treatises form a complete philosophy of practice, progressing from cultivating compassion and bodhicitta, through meditation on calm (śamatha) and insight (vipaśyanā), to the final stage of directly realizing emptiness.³

Three kinds of wisdom: the progression from learning through reflection to meditation

Kamalaśīla emphasizes that realizing emptiness requires three stages of wisdom:

• Wisdom born of learning (śrutamayī-prajñā): acquiring a conceptual understanding of emptiness through studying scripture and the transmitted teachings.
• Wisdom born of reflection (cintāmayī-prajñā): assimilating that conceptual understanding through reflection and critical analysis, removing misunderstandings and arriving at a definite understanding of emptiness. Logical inference (anumāna) is the central tool at this stage.
• Wisdom born of meditation (bhāvanāmayī-prajñā): through repeated meditative training, turning the definite understanding established by reflection into direct, nonconceptual experience in the present. This is the stage of yogic perception, the final goal of the three kinds of wisdom.

The logic of this three-stage structure is clear. Meditative wisdom is direct nonconceptual experience and must rest on the correct conceptual understanding supplied by wisdom born of learning and reflection. Without a correct conceptual understanding of what emptiness actually is, a "nonconceptual state" produced in meditation may be merely a mental blank, rather than genuine insight into emptiness.

The union of calm and insight

Kamalaśīla's integration of calm (śamatha) and insight (vipaśyanā) is another major theme of the Bhāvanākrama. Calm is the meditative stability in which the mind remains focused without distraction. Insight is the wise discernment of emptiness. Both must be cultivated together, with neither neglected. Calm without insight becomes absorption in meditative stillness without seeing truth. Insight without calm leaves the mind insufficiently stable before its object for insight to deepen. Only the union of calm and insight (yuganaddha) constitutes a complete path of practice.

VII. The Lhasa Debate: The Historical Setting of the Sudden–Gradual Dispute¶

Background

Around 792–794, though the exact dates are disputed, King Khri-srong-lde-btsan presided over a debate near Samye Monastery. The participants were Indian Buddhist thinkers, chiefly Kamalaśīla, and Chinese Chan monks, chiefly Hvashang Mahāyāna. The issue was the relative merits of two paths of practice.

The two positions

The Chinese Chan position, represented by Hvashang Mahāyāna: the key to practice is to relinquish all conceptual activity completely. Whether thoughts are wholesome or unwholesome, whether one analyzes emptiness or accumulates merit, all conditioned activity must be abandoned. The mind should abide in a state without discrimination or thought. This itself is liberation, requiring no gradual accumulation.⁴

This position has an internal philosophical logic. If original awakening is already present, any effortful activity, including positive activity, reinforces the illusion of a self that needs to act and thus becomes an obstacle. Only complete relinquishment allows original purity to be recognized.

The Indian gradualist position, represented by Kamalaśīla: his central objection targets precisely this logic. If "completely relinquishing all discrimination" means stopping all mental activity, he argues, it produces only a meditative blank, an ethically indeterminate state without thought, rather than realization of emptiness. Insight into emptiness requires a correct understanding of what emptiness is, and this must be established through the gradual process of learning, reflection, and meditation. To abandon all conceptual activity and "awaken" directly without first attaining correct conceptual understanding is to build a house without foundations.³

More specifically, mere thought-free meditation, a certain level of śamatha, cannot disclose no-self and emptiness. Genuine liberation requires correct insight, the wise discernment of no-self and emptiness, built upon calm. Equating calm with liberation confuses meditative abiding with the discernment of wisdom.

The outcome and the historical controversies

According to Tibetan historical sources, Khri-srong-lde-btsan declared Kamalaśīla's gradualist position victorious and ordered Hvashang Mahāyāna to leave Tibet. In these accounts, the Indian gradualist tradition became Tibetan Buddhist orthodoxy, and Chinese Chan influence was systematically excluded.⁴

Historians, however, have several important reservations:

First, Chinese sources, documents discovered at Dunhuang, differ considerably from Tibetan accounts. They do not record Hvashang Mahāyāna's "defeat" and present his position as entirely reasonable. This suggests that accounts of the "outcome" may substantially reflect later Tibetan Buddhist views rather than objective historical records.

Second, the actual form of the "debate" is unclear. Some scholars consider it more likely to have been a series of informal discussions than a formal contest. On this view, the account of an "Indian victory" is a retrospective construction serving the establishment of Tibetan Buddhist orthodoxy.⁴

Third, Hvashang Mahāyāna's position was itself more complex than the formula "complete abandonment of all conditioned activity" suggests. His own full texts present a different picture from later Tibetan accounts. More crucially, the sudden–gradual distinction need not coincide with a Chinese–Indian divide. Examining Indian materials, Luis O. Gómez identified an indigenous Indian tradition of sudden awakening and concluded that the sudden teaching refuted by Kamalaśīla came from Indian suddenist thinkers rather than Chinese Chan masters. Vimalamitra may have been the model for Kamalaśīla's Indian opponent.⁴ If so, the two sides of this dispute were not originally divided by nationality.

This book's position is that the Lhasa debate was a real historical event, but its scale, form, and outcome were substantially reconstructed later. The philosophical controversy—gradual practice versus sudden awakening, conceptual analysis versus immediate experience—was real. Tibetan Buddhism's choice of the Indian gradualist path was also real. But the narrative of a formal debate ending in Indian victory and Chinese expulsion has limited historical credibility. Reading it as "Chinese Chan versus Indian Madhyamaka" adds another retrospective interpretation. Both sides of the controversy had Indian origins, and later narratives imposed the national divide.

VIII. The Costs of the Synthesis and Criticisms of It¶

The Śāntarakṣita–Kamalaśīla synthesis was an ambitious attempt at integration in the history of philosophy. It also left several unresolved problems, noted from other angles in 6.5 and considered here from within the system:

The cost of subordinating Yogācāra: Śāntarakṣita assigns Yogācāra the role of an "epistemological tool at the conventional level," reserving the final metaphysical position for Madhyamaka. From a Yogācāra standpoint, the three-nature theory already explains emptiness through the three kinds of absence of intrinsic nature. It does not need Madhyamaka to "negate it all over again."

The residue of conventional particular characteristics: as a Svātantrika Madhyamaka thinker, Śāntarakṣita recognizes particular characteristics within linguistic convention at the conventional level. Candrakīrti's Prāsaṅgika Madhyamaka regards this concession as a residue of realism. Accepting conventional particular characteristics leaves things with some inherent feature beyond linguistic convention (5.4 and 5.5). This became a principal reason for the later Gelug ranking of Prāsaṅgika Madhyamaka as the higher view.

Finding a basis for practice after the two-step synthesis: once external objects are empty and consciousness too is empty, what remains for practitioners to work with? Kamalaśīla's Bhāvanākrama gives a practical answer in the three kinds of wisdom and the union of calm and insight. The metaphysical question remains open, however: "If everything is empty, what is the subject of practice?"

IX. Historical Influence¶

In India

Śāntarakṣita and Kamalaśīla were among the leading figures of the Nālandā–Vikramaśīla tradition in the eighth century. The Tattvasaṃgraha is a valuable record of the full range of Indian philosophical debate in their time. Their synthesis had a wider following in India than Candrakīrti's tradition. At least in the eighth century, their approach was mainstream at Nālandā; Candrakīrti's Prāsaṅgika Madhyamaka was not yet mainstream.

In Tibet

Their influence was deepest in Tibet. Śāntarakṣita's Madhyamakālaṃkāra became required reading for the study of Madhyamaka. Kamalaśīla's three Bhāvanākrama treatises became foundational texts of Tibetan practice philosophy. Their foundation of Samye, Tibet's first formally established Buddhist monastery, marked the formal beginning of a local Tibetan Buddhist transmission.

After the Lhasa debate, the Indian gradualist path became Tibetan Buddhism's official position, while Chinese Chan influence was curtailed. The consequences of this choice continued throughout Tibetan Buddhist history and formed part of the background to Tsongkhapa's Gelug reforms (see Essay 11).

Tsongkhapa of the Gelug school ultimately went beyond Śāntarakṣita in his ranking of teachings, placing Candrakīrti's Prāsaṅgika Madhyamaka highest and Śāntarakṣita's Svātantrika Yogācāra Madhyamaka one level below. Yet this move itself engaged with the philosophical framework Śāntarakṣita had established. Without his groundwork, Tsongkhapa's Gelug classification would have lacked a concrete object of criticism.

References¶