“Madhyamakanising” Tantric Yogācāra: The Reuse of Ratnākaraśānti’s Explanation of maṇḍala Visualisation in the Works of Śūnyasamādhivajra, Abhayākaragupta and Tsong Kha Pa

The eleventh-century Indian Buddhist master Ratnākaraśānti presents a unique Yogācāra interpretation of tantric maṇḍala visualisation in the *Guhyasamājamaṇḍalavidhiṭīkā. In this text, he employs the neither-one-nor-many argument to assert that the qualities of the mind represented by the deities in the maṇḍala are neither the same nor different from the mind itself. He also provides five scenarios of meditation to explain the necessity of practising both the perfection method (pāramitānaya) and the mantra method (mantranaya) together in Mahāyāna. Ratnākaraśānti’s explanation exerts a significant influence on the works of later Buddhist masters in India and Tibet, with parts of it being reused in the *Śrīherukopadeśanāmasvādhiṣṭhānakrama by Śūnyasamādhivajra (c. the eleventh century), the eighteenth chapter of the Āmnāyamañjarī by Abhayākaragupta (from the late eleventh to the twelfth century) and the tantric compendium sNgags rim chen mo by the Tibetan master Tsong kha pa Blo bzang grags pa (1357–1419). This paper explores how Ratnākaraśānti’s explanation has been reused and modified in these subsequent works. While Śūnyasamādhivajra faithfully reproduced Ratnākaraśānti’s Yogācāra explanation, Abhayākaragupta modified it to align with his Madhyamaka view. Abhayākaragupta in turn influenced Tsong kha pa, who accepted Abhayākaragupta’s Madhyamaka modification in his works. This paper also engages with current scholarly discussions on textual reuse and the underlying reasons behind it. While Śūnyasamādhivajra and Abhayākaragupta assimilated Ratnākaraśānti’s explanation without acknowledgement, Tsong kha pa quoted the text by name and acknowledged Abhayākaragupta’s reuse of Ratnākaraśānti’s explanation. This paper concludes by discussing the factors that determine whether the reused text is acknowledged or not, and the possible reasons behind textual reuse.

Ratnākaraśānti). In the case of repeat by Ś ūnyasamādhivajra, the form and content are largely unchanged, while in the case of repeat by Abhayākaragupta, the content is modified.
It is sometimes challenging to detect textual reuse without acknowledgement. However, with the advent of digital humanities, an increasing number of reused passages are being detected by computer databases. This paper uses the BuddhaNexus database, developed by the Khyentse Center for Tibetan Buddhist Textual Scholarship at the University of Hamburg, to detect the reuse of Ratnākaraśānti's texts. 5 This database has allowed us to identify Ś ūnyasamādhivajra's reuse of Ratnākaraśānti's works.
Section one of this paper investigates Ratnākaraśānti's explanation of the true nature of the maṇḍala and the deities in maṇḍala visualisation in the * Guhyasamājamaṇḍalavidhiṭīkā (Tōh. 1871). Ratnākaraśānti also outlines five scenarios of meditation to illustrate the necessity of practising both non-tantric and tantric methods. Section two examines Ś ūnyasamādhivajra's reuse of Ratnākaraśānti's explanation in the *Śrīherukopadeśanāmasvādhiṣṭhānakrama (Tōh. 1262). Section three analyses the reuse of Ratnākaraśānti's explanation in the eighteenth chapter of Abhayākaragupta's Āmanāyamañjarī (Tōh. 1198), and investigates how Abhayākaragupta modified the passage of Ratnākaraśānti to align with his own Madhyamaka doctrine. Section four demonstrates the impact of Abhayākaragupta's adjusted passage on the Tibetan master Tsong kha pa in his tantric compendium sNgags rim chen mo. Section five offers some observations and concludes by addressing the following questions: What are the reasons for textual reuse? What are the factors which determine whether the reused text is acknowledged or not?
Ratnākaraśānti on maṅ ḋ ala Visualisation in the *Guhyasamājamaṅ ḋ alavidhiṫı¯kā Background The renowned paṇḍita Ratnākaraśānti was likely the most well-known teacher during the eleventh century CE at Vikramaśīla monastery. 6 He wrote extensively on major Buddhist topics, covering both non-tantric Mahāyāna doctrine and tantric ritual and meditation. This section examines a passage from Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā, a commentary on the influential tantric ritual manual *Guhyasamājamaṇḍalavidhi (Tōh. 1865) written by the ninth-century CE Indian master Dīpaṁ karabhadra.
A few sentences of introduction to the background of the passage examined are due here. The passage concerns maṇḍala visualisation in the ritual of tantric Buddhist initiation (abhiṣeka). In the system of Dīpaṁ karabhadra and 5 BuddhaNexus database: https://buddhanexus.net/. For the principles behind the database see Nehrdich (2020). 6 For the life and works of Ratnākaraśānti see Isaacson and Sferra (2019, pp. 241-243) and Seton (2019). For the philosophical position of Ratnākaraśānti see Seton (2023). Ratnākaraśānti, a maṇḍala is a representation of consciousness shining forth, expressed symbolically by architectural elements of the maṇḍala palace and deities in the maṇḍala. In the maṇḍalatattva (true nature of the maṇḍala) and devatātattva (true nature of the deities) section of the *Guhyasamājamaṇḍalavidhi, Dīpaṁ karabhadra states that each component of a nineteen-deity Mañjuvajra maṇḍala is purified by a category in the Buddhist path (e.g. the four dhyānas, the five faculties, the ten pāramitās etc.). For example, the true nature (tattva) or purity (viśuddhi) of the four arches in the maṇḍala is the four meditative absorptions (dhyāna), and the true nature of the four raised platforms is the four mental concentrations (samādhi). Similarly, the deities in the maṇḍala are also purified by categories of the Buddhist path. For example, the true nature of the ten goddesses (the six offering goddesses Rūpavajrā etc. and the four goddesses Locanā etc.) is the ten perfections (pāramitā). For a detailed correspondence between the architectural elements or deities and the doctrinal categories see Tables 6 and 7 in the Appendix.

The Deities are Neither the Same nor Different From the Mind
In *Guhyasamājamaṇḍalavidhi 348 and the commentary thereon, Dīpaṁ karabhadra and Ratnākaraśānti explain that the mind has wisdom and means as its nature, and therefore has the maṇḍala deities as its nature too. 7 Dīpaṁ karabhadra further states that: Because it [i.e. the mind] has qualities which are to be experienced by itself, [it] does not abide in differentiation or nondifferentiation (bhedābheda°, dbye yod dbye med) and so on. When [the mind] has been made proliferated in this way, 8 the perfections and so on appear as results. 9 What does it mean to say that the mind "does not abide in differentiation or nondifferentiation and so on"? In explaining this, Ratnākaraśānti employs a special use of the neither-identical-nor-different argument (and also the neither-one-nor-many argument). Below is my paraphrase of Ratnākaraśānti's arguments (the original text 7 Dīpaṁ karabhadra's *Guhyasamājamaṇḍalavidhi 348 (A15r5, B20v4): avikalpāt (avikalpāt ] em.; avikalpā AB) tu gāmbhīryam audāryaṃ svaparodayāt | gāmbhīryaudāryataś cetaḥ prajñopāyātmakaṃ matam. 'And from non-conceptualisation, [the mind has] profundity; [it has] vastness, because of the arising of [the fulfilment of the goals of] oneself and others. Because of profundity and vastness, the mind is considered to consist of wisdom and means.' Dīpaṁ karabhadra's verse is a śloka recast of Jñānapāda's Samantabhadrasādhana (Tōh. 1855, 1856) verse 125 in Ā ryā meter, see Cheung (2020, p. 176). For an explanation of Dīpaṁ karabhadra's verse see also Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 348 (D120a1-2, P428a4): de bas na de'i bdag nyid can gyi sems ni lha thams cad kyi bdag nyid can yin no zhes bya ba'i tha tshig go. 'Therefore it is as much as to say that the mind (sems = *cetaḥ) having the nature of them [i.e. wisdom and means] has the nature of all deities.' 8 Ratnākaraśānti explains in the commentary that "in this way (evam)" means the mind has been made proliferated by means of the deities and by means of the true nature of the deities (lha rnams kyi sgo dang de kho na nyid kyi sgo nas), see D120b2, P428b7. 9 Dīpaṁ karabhadra's *Guhyasamājamaṇḍalavidhi 349 (A15r5, B20v4-5): pratyātmavedyadharmatvād bhedābhedādyasaṃsthitam | evaṃ prapañcite bhānti (bhānti ] B; bhrānti°A) phalāḥ pāramitādayaḥ (°d ayaḥ ] A;°yaḥ B). The word phalāḥ in pāda d is a predicate of the word pāramitādayaḥ and is therefore feminine plural. and a more literal translation can be found in the respective footnotes after each sentence).
Ratnākaraśānti anticipates an objection from an opponent: if the qualities of the mind are the deities, then are those qualities of mind, such as faith, different from the mind, or not different? If they are not different from the mind, is the mind having the nature of them one, or many? 10 Ratnākaraśānti then responds that the qualities of the mind such as cognising and feeling are not different from the mind, because both the mind and awareness (yang dag par rig pa, *samyagjñāna) of the qualities of the mind are characterised by the luminosity (gsal ba, *prakāśa) of the sky. But the qualities of the mind are not undifferentiated from the mind at the same time too, because of the undesirable consequence that the mind and the qualities being one. 11 The opponent further asks: then has the single mind become many? Ratnākaraśānti answers: no, because we are aware of the mind and the qualities of the mind as something not different. If the mind were many, the mind would become individual awarenesses like the minds of many beings. 12 The opponent replies: then in this way, is the single mind endowed with many representational forms (rnam pa, *ākāra)? 13 Ratnākaraśānti answers: no, because it 10 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a2-3, P428a4-6): gal te ji skad du bshad pa'i tshul gyis sems kyi chos rnams nyid lha rnams yin na sems kyi chos dad pa (dad pa ] em.; dang ba P; tha dad pa D) la sogs pa de yang ci sems las tha dad pa yin nam tha mi dad pa yin | tha mi dad pa yin na de'i bdag nyid can gyi (bdag nyid can gyi ] em.; bdag nyid can gyis DP) sems gcig yin nam du ma yin zhe na. 'If by the manner taught in this way, precisely the qualities of the mind are the deities, [then] those qualities of mind such as faith too, are they different from the mind, or not different? If they are not different [from the mind], the mind having the nature of them is one, or many?' 11 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a3-5, P428a7-8): de yang rig cing tshor bzhin pa'i chos thams cad ni sems las logs shig (logs shig ] D; logs shing P) na med de | sems dang yang dag par rig pa gnyis ni nam mkha' gsal ba'i mtshan nyid can yin pa'i phyir ro | dbye ba med pa yang ma yin te | sems gcig las de rnams tha mi dad na gcig pu nyid du thal bar 'gyur ba'i phyir ro. 'Furthermore, all the qualities such as cognising and feeling are not different from the mind, because both the mind and awareness are characterised by the luminosity of the sky. They are not undifferentiated [from the mind] too, because if they are not different from the mind which is one, there will be the undesirable consequence of [the qualities and the mind] being one.' 12 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a5, P428a8-b2): de lta na (P428b) ni sems nyid gcig pa du mar (gcig pa du mar ] em.; gcig pa dang | du mar DP; gcig pa'u mar N (unclear)) 'gyur ro zhe na | ma yin te | tha mi dad par yang dag par rig pa'i phyir ro | du ma yin na ni skyes bu du ma'i sems dang 'dra bar so sor rig par 'gyur ba yin no. 'If you ask: in this way, has the mind itself which is one become many? No, because [we are] aware of [the mind and the qualities of the mind] as [something] not different. If the mind were many, they would become individual awarenesses like the minds of many beings.' 13 I take the word rnam pa in rnam par du ma yin here to be a translation of the Sanskrit term ākāra. However, rnam pa could also be a translation of vidhā or prakāra. Given that the words rnam pa rnams appear in the immediately following sentence (see the Tibetan in footnote 14 below) and the importance of ākāra in Ratnākaraśānti's epistemology, it is very likely that the opponent uses the words "representational forms (ākāra)" here in this argument. For ākāra in Ratnākaraśānti's philosophy see Moriyama (2014), Tomlinson (2019Tomlinson ( , 2023 and Seton (2023). Note that there are different English translations of the term ākāra, and there are problems with each of the translations which I am not going into detail here. I provisionally adopt the translation "representational form" used by Seton (2023). is contradictory that representational forms, which are not different from the single mind, are themselves multiple. 14 The opponent further asks: are they, i.e. the representational forms and the mind, one and many at the same time? Ratnākaraśānti answers: no, because the position of them being different has already been refuted, i.e. both are characterised by luminosity (gsal ba, *prakāśa). 15 To sum up, Ratnākaraśānti explains: Therefore, the mind in this way lacks duality in being free from the duality of having differentiations and not having differentiations, or of being one and being many. And because of [this] non-duality, it has no conceptual proliferation. 16 From the above passages, we can observe that according to Ratnākaraśānti, the representational forms (ākāra) of the deities in the maṇḍala and the qualities of the mind such as faith (śraddhā), energy (vīrya) etc. (representing the true nature of the deities) have a neither-the-same-nor-different (bhedābheda) relationship with the mind. How can the representational forms and the qualities of the mind be neither the same nor different from the mind? Although Ratnākaraśānti does not explain it here, he explains in the Prajñāpāramitopadeśa (Tōh. 4079) and the *Madhyamakālaṃkāropadeśa (Tōh. 4085) that although the representational forms (ākāra) are ultimately unreal, they possess a special identity (tādātmya) relation with the real reflexively aware luminosity (prakāśa). The identity between the representational forms and reflexively aware luminosity is a superimposed identity (*āropitaṃ tādātmyam) which, while imposing an identity, still maintains a difference between the two. 17 It should be noted that here Ratnākaraśānti employs the same type of neither-onenor-many or neither-identical-nor-different argument to establish the non-duality of 14 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a5-6, P428b2): de lta na ni sems gcig pu 'di nyid rnam pa du ma yin no (rnam pa du ma yin no ] em.; rnam par du ma yin no D; rnam pa ma yin no P) zhe na | ma yin te | sems gcig las tha mi dad pa'i rnam pa rnams ni (rnam pa rnams ni ] P; rnam pa rnams D) du ma nyid yin par 'gal ba'i phyir ro. 'If you say: in that case, is the single mind [endowed with] many representational forms (rnam pa = *ākāra)? No, because of the contradiction of the manifoldness of many representational forms which are not different from the single mind.' 15 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a6, P428b2-3): gcig pa dang du mar 'gyur ro zhe na | tha dad pa'i phyogs ni sngar sun phyung ba nyid yin pa'i phyir ro. 'If you ask: are they one and many [at the same time]? [No,] because the position of being something different has been refuted before.' 16 Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349 (D120a7-b1, P428b5-6): des na 'di ltar dbye ba yod pa dang (yod pa dang ] D; yod dang P) dbye ba med (D120b) pa'i bdag nyid can nam | gcig pa nyid dang du ma nyid du 'gyur ba'i ('gyur ba'i ] D; gyur pa'i P) gnyis med pas sems de ni gnyis su med pa nyid yin la | gnyis su med pa'i phyir spros pa med pa yin no. 17 For details of Ratnākaraśānti's arguments on ākāra being unreal and the special relationship between the unreal ākāra and the real prakāśa see the excellent studies of Moriyama (2014, pp. 345-348), Tomlinson (2019, pp. 143-177), Seton (2023, pp. 590-596) and most recently Tomlinson (2023, p. 393, 396-400). mind as he does in defending his specific *alīkākāravāda 18 view in the Prajñāpāramitopadeśa and the *Madhyamakālaṃkāropadeśa. 19 And, as Moriyama has pointed out, this use of argumentation contrasts with Ś āntarakṡita's utilization in his Madhyamakālaṃkāra where he employs the neither-one-nor-many argument to prove that entities are lack of intrinsic nature. 20 In fact, Ratnākaraśānti's specific use of the neither-identical-nor-different argument aligns with the prasaṅga argument found in chapter three of the Saṃdhinirmocanasūtra. In this chapter, the Buddha explains that the character that is ultimate reality (paramārthalakṣaṇa) and the character of conditioned factors (saṃskāralakṣaṇa) are neither identical nor different. 21 Ratnākaraśānti's specific use of this type of argument is rooted in the core Yogācāra texts such as the Saṃdhinirmocanasūtra and deserves further study.

Why Tantric Visualisation is Needed (the Five Scenarios)
Now, if in visualisation the representational forms of the deities and the qualities of the mind (i.e. the true nature of the deities) are neither the same nor different from the mind, is it not enough to meditate on the true nature of the deities alone? Why bother to visualise all these deities? In the next passage, Ratnākaraśānti explains why we have to meditate on both the mind as deities and the true nature (tattva) of the deities at the same time. According to Ratnākaraśānti, there are five scenarios: (1) If one meditates on the mind alone, then one would only obtain mundane mental concentration (ting nge 'dzin, *samādhi) like the stage of the infinity of consciousness (rnam shes mtha' yas skye mched, *vijñānānantyāyatana).
(2) Yet if one meditates on emptiness above all, that [result] too becomes only complete cessation, because of not perfecting the actions of purifying the Buddha qualities.
18 As far as I know, the term *alīkākāravāda is a back-translation from the Tibetan doxographical term rnam brdzun pa or rnam pa brdzun par smra ba and is not attested in Sanskrit. For rnam pa brdzun par smra ba see for example Rong zom pa's classification of the sākāravāda in Almogi (2009, p. 34). For a further discussion of Ratnākaraśānti's epistemological position, see Isaacson and Sferra (2014, p. 64, n. 21) and Seton (2023, pp. 590 and 598). Also, for a discussion of the conflation of nirākāra and alīkākāra by scholars, see Seton (2023, p. 590, see also p. 598, n. 6). 19 See for example the sources quoted in Moriyama (2014, pp. 346-347). It appears that Ratnākaraśānti's proof of the non-duality of the mind was not entirely successful in his time. For instance, his contemporary Jñānaśrīmitra, who upheld the citrādvaita view of non-duality, criticised him. For more details, see Tomlinson (2019, p. 8 and pp. 250-261). 20 The neither-one-nor-many (ekāneka) argument is used by Ā ryadeva, Vasubandhu, Dharmakīrti and other early Mahāyāna thinkers to prove the non-existence of external objects. Ś āntarakṡita further uses the neither-one-nor-many argument to demonstrate the emptiness of intrinsic nature for all dharmas. In the Madhyamakālaṃkāra, he argues that if an entity lacks both singular nature (ekatva) and multiple nature (anekatva), then that entity lacks intrinsic nature, see Moriyama (2014, p. 340, 345-348). 21 The Sanskrit original of the Saṃdhinirmocanasūtra is not extant, but we have Sanskrit fragments and quotations, a Tibetan translation, two full Chinese translations and two partial Chinese translations. The terms paramārthalakṣaṇa and saṃskāralakṣaṇa are attested in a Sanskrit fragment of the Saṃdhinirmocanasūtra found in Central Asia, see Waldschmidt (1971, p. 180, Nr. 923) and also the reconstruction of the sentence containing the terms in Schmithausen (2014Schmithausen ( , p. 557, n. 2291. For the Tibetan translation of chapter three of the Saṃdhinirmocanasūtra see Lamotte (1935, pp. 42-47).
(3) Or, if one meditates on [the mind] only as having the nature of the deities, in this case, one does not even become awakened at all through that alone because the perfection of actions is incomplete. (4) Or, if one meditates only on the true nature of what the deities stand for and not the deities, then in this case too, one would attain Buddhahood in many countless aeons but not quickly. (5) Therefore, the meditation of both [the mind as deities and the true nature of the deities at the same time], because it is extremely pleasant to the mind and because it is a special kind of empowerment, causes one to obtain the highest perfect awakening very quickly. 22 The five scenarios describe various tantric and non-tantric practitioners, encompassing both Buddhists and non-Buddhists. The first scenario likely disproves non-tantric and non-Buddhist practitioners of mind-focused meditation. The second scenario likely disproves śrāvaka Buddhists who meditate on a specific quality of emptiness (without the aid of Mahāyāna skillful means). The third scenario likely disproves tantric, non-Buddhist practitioners of meditation. The fourth scenario likely disproves Mahāyāna Buddhists following the perfection method. The fifth scenario likely affirms Mahāyāna Buddhists who practise meditation using both the perfection method and the mantra method. 23 Below is a summary of Ratnākaraśānti's explanation of the five scenarios of meditation (Table 1).

The Reuse in Śūnyasamādhivajra's *Śrīherukopadeśanāmasvādhiṡ ṫhānakrama
Ratnākaraśānti's explanation of the mind and the true nature of the deities in maṇḍala visualisation is unique and exerts an influence in the explanation of the true nature (tattva) or purity (viśuddhi) of the maṇḍala elements and the deities in the later generations. An explanation similar to Ratnākaraśānti's is found in the *Śrīherukopadeśanāmasvādhiṣṭhānakrama of Ś ūnyasamādhivajra (who is said to be identical with Divākaracandra, probably a student of Ratnākaraśānti). 25 *Śrīherukopadeśanāmasvādhiṣṭhānakrama is a text comprised of twenty ritual procedures (cho ga, *vidhi). Ś ūnyasamādhivajra reuses (without acknowledgement) passages from Ratnākaraśānti's *Kusumāñjali (Tōh. 1851), Muktāvalī (Tōh. 1189) and *Guhyasamājamaṇḍalavidhiṭīkā in this text. The twentieth (i.e. the last) ritual procedure in the text is called the ritual procedure of practising the purity [of the The name of the author of the *Śrīherukopadeśanāmasvādhiṣṭhānakrama given in the Tibetan translation is *Samādhivajra (Ting nge 'dzin rdo rje). His name is sometimes given as *Ś ūnyatāsamādhivajra or *Ś ūnyatāsamādhi, but his name is most probably Ś ūnyasamādhivajra as attested in the Hevajrasādhanasaṃgraha codex, see Isaacson (2009, p. 121). For the identification of Ś ūnyasamādhivajra with Divākaracandra/Devākaracandra, see Roerich (1995, p. 392). For information on Divākaracandra's floruit and oeuvre, and the sources for his being Ratnākaraśānti's student, see Isaacson & Sferra (2014, pp. 83-84, n. 105). maṇḍala] (rnam par dag pa bsgom pa'i cho ga), which is comprised of two passages reused from Ratnākaraśānti's texts and one small passage written by Ś ūnyasamādhivajra himself. The first passage (D341a4-b3) is a resue of Ratnākaraśānti's explanation of the perfection method as expounded in Muktāvalī ad Hevajratantra I.i.10, 26 while the second passage (D341b3-7) is about the need for practising both the perfection method and the mantra method together as it was expounded by Ratnākaraśānti in *Guhyasamājamaṇḍalavidhiṭīkā ad *Guhyasamājamaṇḍalavidhi 349.
To clearly show the resue of Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā in Ś ūnyasamādhivajra's text, I put the two texts in parallel columns in the following Table (Table 2).
From the first row of Table 2, we can see that Ś ūnyasamādhivajra frames Ratnākaraśānti's five scenarios with an objection from the opponent: "If through this [perfection method] one obtains quickly the bliss which is perfect awakening, then what is the purpose of other fallacious signs, i.e. palaces, emblems, seed syllables, crescent-shaped ornaments, sound and so on of Heruka and yoginī [used in tantric meditation]?" 27 This objection is probably modelled on another sentence in Ratnākaraśānti's Muktāvalī ad Hevajratantra I.i.10. 28 The point here is that the meditation mentioned in the reused passage in the Muktāvalī (just before repeating the five scenarios) centres on prajñāpāramitā and is non-tantric in nature, and somebody might raise the objection: if a non-tantric method is enough to reach Buddhahood, why take the trouble to visualise all these emblems and seed syllables of the deities which are unreal mental proliferations and are fallacious in nature? Ś ūnyasamādhivajra then replies that visualising all these has a purpose. He goes on to repeat Ratnākaraśānti's five scenarios to emphasise that both the perfection method and the mantra method are necessary because, through them, there is speedy attainment of awakening.
From Table 2, we can see that although the wording of Ś ūnyasamādhivajra's may at times differ slightly from Ratnākaraśānti's, the contents of the two texts are more or less the same. Moreover, there are slight differences in the wording in the Tibetan translations of the two texts probably because the Tibetan translations were independently produced by different translators. These two texts may be even more 26 The corresponding Sanskrit of the reused passage of Muktāvalī ad Hevajratantra I.i.10 can be found in Isaacson (2021, pp. 478-479), and the corresponding passages in the Tibetan translation of the Muktāvalī can be found in D227a3-b2. 27 *Śrīherukopadeśanāmasvādhiṣṭhānakrama (D341b3-4, P428a5-7): gal te bsgom pa 'dis myur du mchog gi bde ba yang dag par rdzogs pa'i byang chub thob na | gzhan he ru ka dang | rnal 'byor ma'i gzhal yas khang dang mtshan ma dang sa bon dang | zla ba phyed dang sgra tsam la sogs pa phyin ci log gi mtshan nyid ci dgos (ci dgos ] D; ci gos P) she na. 28 Muktāvalī ad Hevajratantra I.i.10 (Isaacson, 2021, p. 479): tato yeyam ākāravatī herukasya yoginīnāṃ mantracihnāsanakūṭāgārādīnāṃ ca bhāvanā, sā prapañco viparyāsaḥ saṃsārāvahaḥ prayāso na mokṣāvaha iti kasyacid āśaṅkā syāt. 'Somebody might object: "therefore this meditation of Heruka, the yoginīs, [their] mantras, emblems, seats, palace, and so on, involving representational forms, is mental proliferation, a delusion, an effort which causes saṃsāra, not one which causes liberation."' I would like to thank an anonymous reviewer for bringing to my attention that this excerpt from the Muktāvalī implies that the genitive particle in the phrase "he ru ka dang | rnal 'byor ma'i" in the Tibetan translation of Ś ūnyasamādhivajra's text (see footnote 27 above) is likely incorrect. Moreover, the reviewer pointed out that Heruka and the yoginīs should also be included in Ś ūnyasamādhivajra's list. Previous scholarly research has shown that Abhayākaragupta extensively incorporates passages written by other authors into his own works (often without acknowledgement). 31 Among those authors, Ratnākaraśānti seems to be a favourite source, since Abhayākaragupta reproduces in his own work many passages from both the non-tantric and tantric works of Ratnākaraśānti. 32 Scholarly research also has shown that the focus of this paper, Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā, was incorporated without acknowledgement by Abhayākaragupta in his Āmnāyamañjarī. Kano has pointed out that Abhayākaragupta incorporates Ratnākaraśānti's explanation of the Buddha-nature in the *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 94 into chapter one of the Āmnāyamañjarī, where he "Madhyamakanises" 33 Ratnākaraśānti's explanation by inserting a Madhyamaka phrase indicating that the mind is absent of any intrinsic nature. 34 Furthermore, Sakurai has pointed out that Abhayākaragupta reuses Ratnākaraśānti's explanation of the true nature of the deities in the maṇḍala (*Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 349) "almost verbatim" in the eighteenth chapter of the Āmnāyamañjarī, 35 and that Abhayākaragupta 29 The colophons of three of Abhayākaragupta's works contain dates in Rāmapāla's reign, see Bühnemann (1992, p. 122). I follow Hori (2019, p. 51) for the date of Rāmapāla's reign. 30 For a summary of Abhayākaragupta's life see Erb (1997, pp. 27-29) and Isaacson and Sferra (2019, pp. 249-251). For his date see Bühnemann (1992, pp. 121-123). Luo (2020, pp. 59-63) contains a description of his twenty-eight works and a list of self-references in his writings. For the doctrinal position of Abhayākaragupta see Seyfort Ruegg (1981, p. 103 and p. 115) and Kano (2023). 31 See for example the overview in Isaacson and Sferra (2019, p. 251). For a by-no-means-exhaustive table of Abhayākaragupta's reuse of other people's works see Table 5. 32 Including Ratnākaraśānti's non-tantric works *Śuddhimatī (Tōh. 3801) and Sāratamā (Tōh. 3803), and the tantric works Muktāvalī and *Guhyasamājamaṇḍalavidhiṭīkā; see Table 5 for references. 33 To use Kano's expression in Kano (2023, p. 627). 34 Kano (2023, p. 626). 35 Sakurai (1996, p. 147). adjusts the text according to his Madhyamaka philosophical position. 36 However, Sakurai only mentions the reuse of Abhayākaragupta in passing and does not provide further details. In fact, in Āmnāyamañjarī chapter eighteen, when Abhayākaragupta comments on Saṃpuṭatantra V.2.57 with an explanation of the true nature of the deities, he incorporates not one but multiple passages from Ratnākaraśānti. To highlight the reuse, I have included a table with the two texts in parallel columns (Table 3).
As shown in Table 3, Abhayākaragupta repeats (i.e. reuses without acknowledgement) Ratnākaraśānti's passages from the *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 343, 348 and 349 in one continuous passage, adding sentences of his own in between and at the end. Although the wordings between the Tibetan translations of Abhayākaragupta's text and that of Ratnākaraśānti's text are quite close, we can sometimes see (for example in the first sentence of row one of Table 3) differences in the use of Tibetan particles between the corresponding sentences in the two texts sometimes lead to divergence in meaning. As mentioned above, these differences in the two texts may be due to the fact that the two translations were independently produced.
In row four of Table 3, the places where Abhayākaragupta modifies Ratnākaraśānti's text are underlined. The following is the translation of Abhayākaragupta's text in row four (on the five scenarios): And in this way, when [the mind] has been made proliferated by means of the deities and by means of the true nature [of the deities], one obtains the qualities of the Buddhas, such as the perfections, which have become the results [of the mind]; they are not [obatined] in any other way.
(1) If one meditates only on consciousness, one obtains only mundane mental concentration like the stage of the infinity of consciousness.
(2) If one meditates only on emptiness, at that time too, like a śrāvaka (nyan thos bzhin du), one becomes somebody who has the notion of complete cessation only, because of not purifying the Buddha qualities. In the second scenario, Abhayākaragupta introduces the qualification of "like a śrāvaka (nyan thos bzhin du)" to clarify that this is a disproval of non-Mahāyāna Buddhists such as the śrāvakas and the pratyekabuddhas, who are intent on achieving mere cessation. In the fourth scenario, Abhayākaragupta modifies Ratnākaraśānti's phrase "if one meditates only on the true nature of what the deities stand for" 37 to "if one meditates on the knowledge of emptiness (stong pa nyid kyi ye shes), but not the deities." For both Ratnākaraśānti and Abhayākaragupta, the fourth scenario describes the seeing of emptiness by the practitioners while engaging in the perfection method. According to Ratnākaraśānti, the perfection of wisdom (prajñāpāramitā) is defined as the seeing of emptiness (on the fourth of the four stages of yoga). And the true nature of the deities can be described as either "the qualities of the mind" or as "emptiness." This is because, during the exploration in the four stages of yoga, the second stage involves perceiving the true nature of the deities solely as mind only. And in the third and fourth stages, a practitioner further sees the true nature of the deities as emptiness, initially with characteristics (sanimittā) and then without them (nirnimittā). 38 Abhayākaragupta would largely agree with Ratnākaraśānti's interpretation, but there is a crucial difference between the concept of emptiness of Ratnākaraśānti and that of Abhayākaragupta. For Ratnākaraśānti, emptiness is the absence of the duality of the apprehended object (grāhya) and the apprehending subject (grāhaka). 39 The mind itself is not empty, and what is ultimately real is sheer luminosity (prakāśamātra). 40 On the other hand, Abhayākaragupta considers the mind itself to be empty i.e. without intrinsic nature (svabhāva), and what is ultimately real is the absence of intrinsic nature (niḥsvabhāvatā). 41 In other words, Ratnākaraśānti adopts the Yogācāra understanding of emptiness as an implicative negation (paryudāsapratiṣedha), 42 while Abhayākaragupta adopts the Mādhyamika understanding of emptiness as a nonimplicative negation or absolute negation (prasajyapratiṣedha). Abhayākaragupta feels to need to distinguish emptiness from the mind, therefore he changes 37 See section "Why Tantric Visualisation is Needed (the Five Scenarios)". 38 I am grateful to one of the anonymous reviewers for clarifying Ratnākaraśānti's position here. For more information on Ratnākaraśānti's four stages of yoga (yogabhūmi) in the Prajñāpāramitopadeśa, refer to Namai (1991), Bentor (2002, pp. 42-49), Yiannopoulos (2017, p. 240), Katsura (2018) and Seton (2023, pp. 594-595). For Ratnākaraśānti's application of the four stages of yoga in tantric practices, see especially Bentor (2002, pp. 49-50), Yiannopoulos (2017) and Seton (2017, p. 5). 39 See Ratnākaraśānti's Muktāvalī ad Hevajratantra II.viii.9-10 (Isaacson, 2013(Isaacson, , p. 1040: tad api cittaṃ na sad dvayarūpeṇa, nāsad dvayaśūnyena rūpeṇeti. 'The mind too is not real in the form of the two [i.e. in the form of apprehended object (grāhya) and apprehending subject (grāhaka)], and it is not unreal in the form empty of the two [i.e. apprehended object (grāhya) and apprehending subject (grāhaka)].' See also Muktāvalī ad Hevajratantra I.i.12 (Isaacson, 2021, p. 482): ata eva na sat sarvaṃ dvayarūpeṇa dvayākārarūpeṇa ca, nāsat sarvam advayaprakāśamātrarūpeṇeti mādhyamikānāṃ yogācārāṇāṃ ca sadṛśaḥ siddhāntaḥ śreyān. 'For precisely this reason, it is not the case that everything is real (sat) in the form of the two [i.e. in the form of apprehended object (grāhya) and apprehending subject (grāhaka)] and in the form of the representational forms (ākāra) of the two [i.e. in the form of grāhyākāra and grāhakākāra], and it is not the case that everything is unreal in the form of sheer luminosity without the two [apprehended object (grāhya) and apprehending subject (grāhaka)] (advayaprakāśamātra), therefore the equivalent established position of the Mādhyamikas and the Yogācāras is better.' 40 Seton (2023, p. 590). 41 See Abhayākaragupta's Munimatālaṃkāra chapter one (Kano & Li, 2018, p. 130): paramārthatas tu dharmasya vijñānasya dharmatāyāś ca śūnyatāder niḥsvabhāvatvān na bhedābhedau. 'But ultimately, there is neither differentiation nor non-differentiation, because the emptiness (śūnyatā) and so on of factor of existence (dharma), consciousness (vijñāna) and the nature of the factor of existence (dharmatā) are without intrinsic nature (niḥsvabhāva).' 42 Seton (2017, p. 3).
Ratnākaraśānti's description of the fourth scenario to highlight the importance of emptiness. Furthermore, in the fifth scenario, Abhayākaragupta changes Ratnākaraśānti's "if one meditates on both the mind as deities and the true nature of the deities at the same time" 43 to "when one meditates on all three (gsum ka) [i.e. the mind as deities, the true nature of the deities, and emptiness]." By including emptiness as a separate item in the final accepted scenario of meditation, Abhayākaragupta "Madhyamakanises" the Yogācāra explanation of Ratnākaraśānti.
Here, to help clarify, is a summary table of the explanations of the five scenarios by Abhayākaragupta and Ratnākaraśānti (Table 4). A few passages down from the previously discussed passage, Abhayākaragupta quotes verses 126-128, 125 and 124ab of the Sarvarahasyatantra to explain the true nature of some of the architectural components in the maṇḍala. 44 In the *Guhyasamājamaṇḍalavidhiṭīkā ad Guhyasamājamaṇḍalavidhi 335, Ratnākaraśānti also quotes the Sarvarahasyatantra in the same sequence (with two verses more; Ratnākaraśānti quotes verses 126-130, 125 and 124ab). 45 Abhayākaragupta's source is probably Ratnākaraśānti. It is interesting to note in passing that Ś ākyarakṡita's Hevajrābhisamayatilaka (Tōh. 1277) also quotes the  Cheung, 2020, pp. 152-154).
Sarvarahasyatantra in the same number of verses and sequence as Abhayākaragupta's Āmnāyamañjarī, 46 and just before the Sarvarahasyatantra quotation, there is also a large chunk of parallel passage between the two texts. Given that Ś ākyarakṡita was a student of Abhayākaragupta, 47 we probably have here a repeat of Abhayākaragupta's Āmnāyamañjarī in Ś ākyarakṡita's Hevajrābhisamayatilaka.

Tsong Kha Pa's Reuse
Both the writings of Ratnākaraśānti and Abhayākaragupta had a significant impact on Tsong kha pa Blo bzang grags pa (1357-1419), 48 the renowned founder of the dGe lugs school in Tibet. Tsong kha pa frequently quotes Ratnākaraśānti and Abhayākaragupta in his sNgags rim chen mo. 49 The passage under discussion, the five scenarios of meditation in Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā, is not only quoted in Tsong kha pa's sNgags rim chen mo, but also reproduced verbatim in Tsong kha pa's bZhi brgya lnga bcu pa'i skor gyi zin bris gnang ba. 50 In the sNgags rim chen mo, Tsong kha pa discusses the necessity of practising both the perfection method and the mantra method. He first cites the Vajrapañjaratantra and Jñānapāda's Ātmasādhanāvatāra before quoting verbatim from the third to the fifth scenarios of Ratnākaraśānti in the *Guhyasamājamaṇḍalavidhiṭīkā: Ratnākaraśānti says in the Commentary to the Four Hundred and Fifty Verses [i.e. the *Guhyasamājamaṇḍalavidhiṭīkā], "(3) If one meditates on [the mind] only as having the nature of the deities, in this case, one does not even become awakened at all through that alone, because the perfection of actions is incomplete. (4) Or, if one meditates only on the true nature of what the deities stand for and not the deities, then in this case too one would attain Buddhahood in many countless aeons but not quickly. (5) Therefore, the meditation of both [the mind as deities and the true nature of the deities at the same time], because it is extremely pleasant to the mind and because it is a 46 Ś ākyarakṡita's Hevajrābhisamayatilaka (D117b7-118a3, P510a8-b3). A Sanskrit manuscript containing Ś ākyarakṡita's Hevajrābhisamayatilaka is preserved at the Niedersächsische Staats-und Universitätsbibliothek Göttingen, Germany; Dr. Torsten Gerloff (Hamburg) is currently preparing a Sanskrit critical edition of this text. 47 Sanderson (2009, p. 176). 48 For the date of Tsong kha pa I follow the information given on the BDRC website: http://purl.bdrc.io/ resource/P64. For his life see Jinpa (2019) and Repo (2019). 49 In the sNgags rim chen mo, Tsong kha pa usually refers to Ratnākaraśānti, Abhayākaragupta and other Indian masters by name and cites the name of their works. Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā and Abhayākaragupta's Āmnāyamañjarī are among the most quoted tantric treatises in the sNgags rim chen mo, and a rough count shows that Abhayākaragupta is likely the most quoted person in the sNgags rim chen mo. 50 The bZhi brgya lnga bcu pa'i skor gyi zin bris gnang ba consists of excerpts from four passages from Ratnākaraśānti's *Guhyasamājamaṇḍalavidhiṭīkā without any added explanation. This text is very likely some kind of study notes or working notes of Tsong kha pa. special kind of empowerment, causes one to obtain the highest perfect awakening very quickly." 51 While Tsong kha pa quotes Ratnākaraśānti, his explanation is based on the Madhyamaka explanation of Abhayākaragupta in the Āmnāyamañjarī. Tsong kha pa continues: [Here Ratnākaraśānti] says that if one meditates only on deity yoga, one is not able to be awakened at all. And if one does not meditate on the deities, through meditating [only] on emptiness together with other means, one would attain awakening after many countless aeons. And if one meditates on both the deities and emptiness, the path [to awakening] is speedy. Therefore, this ācārya [i.e. Ratnākaraśānti] also accepts that because the view of emptiness is common to both [types of] Mahāyāna [i.e. perfection method and mantra method], if there is no deity yoga, then there is a delay in the path like the Perfection Vehicle (phar phyin gyi theg pa, *pāramitāyāna), and by connecting deity yoga with the view of emptiness, the path is speedy. [His view] follows what has been discussed earlier [in the Vajrapañjaratantra and by Jñānapāda in the Ātmasādhanāvatāra]. 52 We can see that, in contrast to Ratnākaraśānti, who expresses the Yogācāra view that the mind and the qualities of the mind are the true nature of the deities in a tantric visualisation, Tsong kha pa, himself a Mādhyamika, interprets the true nature of the deities as the view of emptiness according to Abhayākaragupta's Madhyamaka modification but without mentioning Abhayākaragupta. 53 He only mentions Abhayākaragupta and his Āmnāyamañjarī by name after explaining Ratnākaraśānti's passage: 51 Tsong kha pa, sNgags rim chen mo (22a4-6): bzhi brgya lnga bcu pa'i 'grel pa las | yang na lha'i bdag nyid can 'ba' zhig tsam bsgoms na de lta na ni de tsam gyis 'tshang rgya ba nyid du mi 'gyur te | las rdzogs pa ma tshang ba'i phyir ro || yang na lha rnams kyi de kho na nyid bsgom gyi lha rnams ma yin na ni | de lta na yang bskal pa grangs med pa mang por sangs rgyas nyid thob par 'gyur gyi myur du ni ma yin no || de bas na gnyis ka sgom pa ni shin tu yid du 'ong ba yin pa'i phyir dang | byin gyis brlabs kyi khyad par gyis mchog tu myur bar bla na med pa yang dag par rdzogs pa'i byang chub thob par 'gyur ro || zhes shānti pas gsungs te. 52 Tsong kha pa, sNgags rim chen mo (22a6-b3): lha'i rnal 'byor tsam zhig sgom (22b) na gtan 'tshang rgya mi nus pa dang | lha ma bsgoms na stong nyid thabs gzhan dang ldan par bsgoms pas bskal pa grangs med mang pos 'tshang rgya ba dang lha dang stong nyid gnyis ka bsgoms na lam myur bar gsungs so || des na slob dpon 'di yang stong nyid kyi lta ba theg chen pa gnyis ka'i thun mong ba yin pas lha'i rnal 'byor med na phar phyin gyi theg pa ltar lam 'gyang ba dang | lha'i rnal 'byor stong nyid kyi lta ba dang sbrel bas lam myur bar bzhed pa yin te sngar bshad pa'i rjes su 'brang ba'o. 53 I am grateful to one of the anonymous reviewers for pointing out that Tsong kha pa's understanding of emptiness underwent significant change following his vision of the deity Mañjuśrī and that there is the possibility that in the sNgags rim chen mo he is, in turn, reinterpreting Abhayakararagupta's Madhyamaka modification. This is indeed possible, but I do not think Tsong kha pa's and Abhayākaragupta's interpretations diverge significantly here. My impression is that Abhayākaragupta makes an effort to adhere more closely to Ratnākarśānti's wordings, but Tsong kha pa paraphrases more freely and makes the Madhyamaka elements clearer. The sNgags rim chen mo was written when Tsong kha pa was 49 (i.e. in 1405. For the date of composition see Seyfort Ruegg's introduction to the English translation of the Lam rim chen mo, Tsong-kha-pa, 2000, p. 28), certainly after Tsong kha pa's vision of Mañjurśrī during a retreat in 1393 (Jinpa, 2019, pp. 129-130), but further research is needed before we can draw any definite conclusions on this point.
In chapter eighteen of the Āmnāyamañjarī too, Abhaya, after explaining in accordance with Ś āntipa [i.e. Ratnākaraśānti], cites the scriptural source of that from chapter fourteen of the Vajrapañjaratantra: "For the purpose of overcoming ordinary pride, [this] meditation is correctly proclaimed" and "furthermore, in order to purify the impure body, one should meditate on the body of the Buddha." 54 Here, Tsong kha pa points out that Abhayākaragupta's source is Ratnākaraśānti and reproduces Abhayākaragupta's citation of the Vajrapañjaratantra. 55

Some Observations
Both Ś ūnyasamādhivajra and Abhayākaragupta silently incorporate Ratnākaraśānti's explanation of the five scenarios of meditation, so their reuse fall under the category of "repeat (i.e. unacknowledged textual reuse)" and not under "quotation (i.e. acknowledged textual reuse)." Tsong kha pa's reuse, on the other hand, is a quotation where he acknowledges Ratnākaraśānti. Specifically, his quotation falls into the category of citation (i.e. acknowledged textual reuse, the same in both form and content). In contrast, Tsong kha pa only says that Abhayākaragupta's text follows Ratnākaraśānti's but does not quote Abhayākaragupta.
While Ś ūnyasamādhivajra follows faithfully Ratnākāraśānti's Yogācāra explanation, Abhayākaragupta modifies Ratnākāraśānti's text in accordance with his own Madhyamaka philosophy. Tsong kha pa quotes Ratnākaraśānti's text verbatim but follows Abhayākaragupta's modification in his own explanation. As a Mādhyamaka himself, Tsong kha pa regards Abhayākaragupta's Madhyamaka modification as the correct interpretation of Ratnākaraśānti's teaching. In fact, he intentionally reinterprets Ratnākaraśānti through the lens of Abhayākaragupta.
What can we glean from these cases of repeat and quotation? According to Hugon, there are two main functions of quotations: (1) to present an opponent's view, or (2) to support one's own interpretation or explanation. 56 Tsong kha pa's quotation has the second function; he quotes Ratnākaraśānti to support his argument that it is necessary to have both types of Mahāyāna meditation together.
But what are the possible reasons for unacknowledged repeats, as in the cases of Ś ūnyasamādhivajra and Abhayākaragupta? Scholars already have pointed out that the modern concept of plagiarism does not apply to the intellectual world of medieval India and Tibet. 57 In the literary culture of medieval India or Tibet, a master operates within a tradition of lineages and regards himself as an agent transmitting traditional knowledge rather than as an innovator composing something 54 Tsong kha pa, sNgags rim chen mo (22b3-4): man ngag snye ma'i snye ma bco brgyad pa las kyang | a bhayas shānti pa ji ltar bzhed pa bzhin bshad nas de'i shes byed du | tha mal nga rgyal gzhom don du || bsgom pa yang dag rab tu bsgrags || zhes dang | slar yang 'on kyang mi gtsang ba'i lus sbyang ba'i slad du sangs rgyas kyi sku bsgom par bgyi'o || zhes gur gyi le'u bcu bzhi pa las gsungs pa drangs so. 55 For Abhayākaragupta's quotation of the Vajrapañjaratantra see row five of Table 3. 56 Hugon (2015, p. 483). 57 Freschi (2012, pp. 171-172, 174-176), Hugon (2015, p. 482) and Vergiani (2015, pp. 208-209). entirely new. And in that literary culture, as the previous scholarship has shown, to silently appropriate (i.e. repeating without acknowledgement) a predecessor's view indicates a master's endorsement of and respect towards this predecessor. 58 And lastly, as also has been studied, in India and Tibet to repeat something from another text, even without acknowledgement, would be recognised by the intellectual community in the author's time, as the source was probably widely known at that time. 59 However, if the source of a quotation was separated in time and space, by a long time or by great geographic distance, from an author and his audience, the author would be more likely to acknowledge that source by name. Let me use Abhayākaragupta and Tsong kha pa to illustrate. Below is a table of Abhayākaragupta's reuse of other people's work which is by no means exhaustive (Table 5).
From Table 5, it appears that the distance in time between Abhayākaragupta and the source texts might be a factor affecting the identification of source texts. The farther away the authors are from Abhayākaragupta in time, the more likely they are quoted by name as an authority. From the There are exceptions to the above observations. Abhayākaragupta sometimes cites and sometimes silently borrows from masters from the same period. He cites Ś āntarakṡita's Tattvasaṃgraha by name but silently incorporates Kamalaśīla's Madhyamakāloka. Both Sthiramati and Candrakīrti lived hundreds of years before Abhayākaragupta, but Abhayākaragupta assimilates Sthiramati's Pañcaskandhakavibhāṣā and Candrakīrti's Pañcaskandhaka instead of quoting them. Sometimes Abhayākaragupta not only borrows from but also quotes the same master. He quotes Candrakīrti's Madhyamakāvatāra, Madhyamakāvatārabhāṣya and Triśaraṇasaptati, citing the text or author by name, but borrows without acknowledging the source from Candrakīrti's Pañcaskandhaka. And he borrows silently from and also quotes three times by name Ratnākaraśānti's Sāratamā in the Munimatālaṃkāra. 60 With these exceptional cases, another possible factor for the silent borrowing of other master's texts might be at play here. The borrowed master's explanations on certain matters might have become the standard in Buddhist monastic colleges and were 58 Cantwell and Mayer (2013, p. 196), Hugon (2015, p. 483) and Vergiani (2015, pp. 208-209). 59 Freschi (2012, p. 171), Hugon (2015, p. 483) and Vergiani (2015, p. 209). 60 Isoda (1988, p. 76-77). The Rājaratnāvalīvivṛti quoted by Abhayākaragupta in the Munimatālaṃkāra is different from Ajitamitra's *Ratnāvalīṭīkā. The author of this Rājaratnāvalīvivṛti is yet to be identified. See Kano and Li (2020, pp. 55-56) being handed down by tradition. 61 The fact that Ratnākaraśānti is repeated the most by Abhayākaragupta might indicate that Ratnākaraśānti's explanations were widely accepted at Abhayākaragupta's time. Although it is not plausible that Abhayākaragupta was Ratnākarasānti's student, 62 he certainly had access to Ratnākaraśānti's works in the Vikramaśīla monastery. 63 Distance in space is probably also a factor affecting the identification of source texts. In the sNgags rim chen mo, the Tibetan master Tsong kha pa quotes Indian masters by name, but only refers collectively to the views of "some Tibetan masters" when he does not agree with other Tibetan interpretations. 64 Tsong kha pa regards Indian masters as authoritative and quotes them to lend weight to his arguments in refuting wrong views from other Tibetan masters.

Conclusion
In this paper, I provide a few examples of textual reuse in tantric texts, primarily focusing on a line of reuse of Ratnākaraśānti by Ś ūnyasamādhivajra, Abhayākaragupta and Tsong kha pa. I produce comparative tables of the works of these authors to show that, in contrast to Ś ūnyasamādhivajra, who retains the Yogācāra flavour of Ratnākaraśānti, Abhayākaragupta modifies Ratnākaraśānti's text to suit his Madhyamaka philosophy. On the other hand, Tsong kha pa accurately quotes Ratnākaraśānti but utilises Abhayākaragupta's justification to offer Madhyamaka viewpoints in his commentary on Ratnākaraśānti. After applying the insights of earlier studies, I provide some remarks regarding these authors' compositional processes, employing a variety of sources. In contrast to the accepted customs nowadays, in medieval India and Tibet, a master's endorsement of and respect for a predecessor is demonstrated when he silently appropriates (i.e., repeat without acknowledgement) that predecessor's viewpoint. A master would be more likely to give credit to the source of a quotation if there was a significant period of time or distance between that source and his audience. I hope future research on tantric Buddhist commentaries will further improve our understanding of the nature of textual reuse in late medieval Indian and Tibetan Buddhism.
for their helpful corrections, comments and suggestions. Needless to say, I alone am responsible for any remaining mistakes. All English translations of Sanskrit and Tibetan passages are mine.
Funding Open Access funding enabled and organized by Projekt DEAL.

Conflict of interest
The author has no conflicts of interest to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creative commons.org/licenses/by/4.0/.

Appendix
See Tables 6 and 7.   The six offering goddesses (Rūpavajrā etc.) The first six perfections of the ten perfections 344cd The four goddesses (Locanā etc.) The other four perfections of the ten perfections 345 The five Buddhas The five wisdoms (jñāna) = the wisdom of the destruction of defilements (kṣayajñāna) and the wisdom of non-arising of abandoned defilements (anutpādajñāna)

346
The lord of the maṇḍala The mind which has the nature of the wisdom of the destruction of defilements (kṣayajñāna) and the wisdom of non-arising of abandoned defilements (anutpādajñāna)

347
The ten goddesses (Rūpavajrā etc. and Locanā etc.) The ten perfections 347 The first gatekeeper (Yamāntaka) The first two bodhisattva levels 347 The other three gatekeepers The third to the fifth bodhisattva levels 347 The five Buddhas The sixth to the tenth bodhisattva levels 347 The lord of the maṇḍala The eleventh bodhisattva level Note: My understanding of Dīpaṁ karabhadra's verses as interpreted by Ratnākaraśānti is slightly different from that of Sakurai (1996, pp. 139-143) N Narthang edition of the Tibetan canon om. Omitted Ō ta. Ō tani Catalogue number of the Peking edition, see Suzuki (1961) P Peking edition of the Tibetan canon Tōh. Tōhoku Catalogue number of the Derge edition, see Ui et al. (1934)