Sūtrasamuccaya (Compendium of Sutras)

The authorship is genuinely uncertain. In the Tibetan tradition, Tāranātha and other historians ascribe the Sūtrasamuccaya to Śāntideva, alongside the Śikṣāsamuccaya and Bodhicaryāvatāra [citation:10][citation:12]. However, the Tibetan translation itself attributes the text to Nāgārjuna, and modern scholarship often treats this attribution as at least as plausible [citation:3][citation:13]. The confusion is compounded by the fact that Śāntideva's Bodhicaryāvatāra refers readers to a Sūtrasamuccaya 'composed by the noble Nāgārjuna,' which suggests that Śāntideva knew of Nāgārjuna's work and may have composed his own anthology in imitation of it [citation:15]. The Pāsādika edition of the Tibetan text and the Karashima fragment have not resolved the question.

The Sūtrasamuccaya is a more compact anthology of sūtra quotations, while the Śikṣāsamuccaya is a larger work that organizes quotations around twenty-seven root verses with Śāntideva's own prose commentary [citation:9][citation:10]. The Sūtrasamuccaya contains extracts from roughly sixty sūtras, whereas the Śikṣāsamuccaya draws on over 110 works [citation:10]. The Śikṣāsamuccaya is thus the more ambitious and systematic compilation, and it explicitly presents itself as more exhaustive than the earlier Sūtrasamuccaya [citation:10]. Some scholars have suggested that the Sūtrasamuccaya may have served as a model or prototype for the Śikṣāsamuccaya, regardless of whether the same author composed both [citation:13].

The Sūtrasamuccaya survives in full only in Tibetan and Chinese translations. The Tibetan version is Tōhoku 3934, and the Chinese translation is Taishō 1635 [citation:3][citation:11]. For decades, no Sanskrit manuscript was known. Then, in 2009, Seishi Karashima identified two Sanskrit fragments from Khādalik, preserved at the British Library, as belonging to a folio of the Sūtrasamuccaya [citation:11]. These fragments cover quotations from the Ratnarāśisūtra, the Puṣpakūṭasūtra, and the Mahākaruṇāsūtra, and they agree closely with the Tibetan and Chinese translations. The discovery confirms the text's existence in Sanskrit but does not provide a complete manuscript, so scholars still rely on the translations for the full text.

The relationship is textual and biographical. In the Bodhicaryāvatāra (V.106), Śāntideva recommends that those who cannot study the extensive Śikṣāsamuccaya should 'look at the Sūtrasamuccaya composed by the noble Nāgārjuna' [citation:15]. This verse is absent from the earlier Dunhuang recension of the Bodhicaryāvatāra, suggesting it was a later insertion by a reviser who valued the Sūtrasamuccaya [citation:12][citation:15]. The fact that Śāntideva's own text refers to Nāgārjuna's Sūtrasamuccaya has led some scholars to conclude that the Sūtrasamuccaya attributed to Śāntideva in Tibet is a different work, or that the attribution is mistaken [citation:13]. The debate illustrates the difficulty of reconstructing Śāntideva's corpus.

The most common misunderstanding is that the Sūtrasamuccaya is simply a shorter version of the Śikṣāsamuccaya, written by the same author as a kind of preliminary study. In fact, the two works may not share an author at all. The Tibetan tradition ascribes both to Śāntideva, but the internal evidence of the Bodhicaryāvatāra distinguishes between 'the Sūtrasamuccaya composed by Nāgārjuna' and the Śikṣāsamuccaya that Śāntideva himself compiled [citation:15]. Modern scholarship is divided, and the attribution to Śāntideva should be treated as a traditional ascription rather than a settled fact. Readers should approach the text as an independent anthology whose authorship remains an open question [citation:3].