Saṃdhinirmocana Sūtra

1. See glossary entry “ultimate.”

2. See Brunnhölzl 2018, p. 1590, n. 89 on this point.

3. The numbering of paragraphs of the Saṃdhi­nirmocana­sūtra follows Lamotte’s critical edition.

4. See Radich 2007, p. 1257 on the relationship between āśraya­parivṛtti and dauṣṭhulyakāya. Saṃdh. is the only text in the entire Kangyur in which the term dauṣṭhulyakāya is found.

5. In bold are textual resources I used to translate the text into English.

6. See Powers 2015. Unfortunately, I haven’t been able to consult this reference work at the time of completing this translation.

7. Here is a list of the sigla I used to identify the various witnesses of Saṃdh.: (1) Witnesses of the sūtra found in the available Kangyurs and canonical collections (MsK = manuscript Kangyur, PK = xylograph): Kʙ: Berlin MsK, C: Choné PK, Cz: Chizhi, D: Degé PK, Dd: Dodedrak, Dk: Dongkarla, F: Phukdrag MsK, H: Lhasa PK, Gt: Gangteng, He: Hemis I, J: ’jang sa tham/Lithang PK, L: London (Shelkar) MsK, Lg: Lang mdo, N: Narthang PK, Ng: Namgyal, Np: Neyphug, O: Tawang, Pj: Phajoding I, Pz: Phajoding II, Kǫ: Peking 1737 PK, R: Ragya, S: Stok MsK, T: Tokyo MsK, U: Urga PK, V: Ulaanbaatar MsK, W: Wangli supplement, X: Basgo MsK, Z: Shey Palace MsK. Other canonical collections: Ba: Basgo fragments (Ladakh), Bd: Bardan (Zanskar), Go: Gondhla (Lahaul), Do: Dolpo. Source: http://www.rkts.org (last accessed on July 20, 2020). I am following the typology of Kangyur groups suggested by rKTs (Vienna University). I would like to warmly thank Professor Helmut Tauscher and Bruno Lainé for making available to me the editions I used for this translation project. For a general discussion of some Tibetan sources, see Skilling 1994, p. 775. (2) Xylographs of the Viniścaya­saṃ­grahaṇī of the Yogācāra­bhūmi from the Tengyur: VD Degé, VG Golden, VP Peking. My thanks go to Kojirō Katō for having shared with me the bibliographical detail of these witnesses. The Viniścaya­saṃ­grahaṇī is also available in Chinese under the following title: 瑜伽師地論卷第七十六攝決擇分.

8. For the reference of possible additional folios, see Chayet 2005, p. 67 (n°615‍—1 folio, n°590‍—6 folios).

9. 解深密經疏 (ZZ369) is a text originally composed in Chinese that has been translated into Tibetan. On Wonch’uk’s life and works, see Powers 1992a.

10. I would like to warmly thank Zhuoran Xie (Vienna University) for her assistance in reading this text.

11. See also Katō 2011 for textual variations.

12. See Schmithausen 2014, p. 425ff.

13. See Schmithausen 2014, pp. 419–20, n. 1852. On necessary adjustments to Lamotte’s rendition of the original Sanskrit terms, see Hakayama 1984, p. 180 and Delhey 2013.

14. See Tillemans 1997 for a review of Powers 1995. From a general perspective, it seems that Powers chose to ignore the work and methodological approach of Lamotte and Frauwallner.

15. See Schmithausen 1987 and 2014, Delhey 2013, and Skilling 2013 on the simile of the illusionist (māyākāra), which is also included inter alia in the Māyājāla, a sūtra also quoted in the Yogācāra­bhūmi.

16. On this issue, see Brunnhölzl 2018, pp. 414–18, n. 5.

17. See Steinkellner 1989 and Powers 1992a, 1992b, and 1998. For a review of Powers 1998, refer to Wedemeyer 2003.

18. See for instance Hopkins 1999, 2002, and 2006.

19. See Lamotte 1935, p. 12ff., Ware 1937, Edgerton 1937, Edgerton 1953, p. 558, and later Keenan 1980, p. 126, Powers 1991a, and Powers 1993b, p. 28ff.

20. Research relevant to terminological choices, syntactic reading of complex passages, and interpretation of the meaning includes Powers 1991b, 1991c, 1993b, pp. 41–77 (chapters 1 to 4); Tillemans 1997 (chapter 1); Matsuda 2013 on Sanskrit terms (chapter 2); Wayman 1974, Tillemans 1997, Mathes 2007, Matsuda 2013 (chapter 3); Katō 2002, Lusthaus 2002, Waldron 2003, Buescher 2007, 2008, Muller 2011, Schmithausen 1987 and 2014, Brunnhölzl 2018 (chapter 5); Frauwallner 1969, Takahashi 2006, Schmithausen 2014 (chapter 6); Frauwallner 1969, Tucci 1971, Tillemans 1997, Mathes 2007, Schmithausen 2014, Brunnhölzl 2018 (chapter 7); Takasaki 1966, Lamotte 1970, Lin 2010, Matsuda 2013, Schmithausen 1984, 1987, 2005, and 2014 (chapter 8); Obermiller 1933, Matsuda 1995 (chapter 9); Braarvig 1985, Kapstein 1988, Steinkellner 1989, Sakuma 1990, Yoshimizu 1996, Xing 2005, Lin 2010, Yoshimizu 2010 (chapter 10).

21. For a detailed introduction to this text, see Steinkellner 1989.

22. See Billeter 2014. I would like to thank Professor Tom Tillemans for having drawn my attention to Billeter’s principles of translation.

23. See Edgerton 1953, p. 102.

24. See Schmithausen 2014, p. 400, n. 1770.

25. See Lamotte 1935, p. 174.

26. See Mvyut 4414.

27. See Vinay and Darbelnet 1958.

28. Or even “Yep,” “I’m with you,” “So true,” etc.

29. “Une compilation assez maladroite”; see Lamotte 1935, p. 17. For an extensive discussion on the date and composition of Saṃdh., see Lamotte, pp. 14–25. See also Schmithausen 2014, p. 354ff. regarding the relation between the various chapters of Saṃdh.

30. See Schmithausen 2014, pp. 354–55.

31. See Schmithausen 2014, p. 365.

32. See Schmithausen 2014, p. 359.

33. On the usage and various shades of meaning of vijñapti, see Hall 1986.

34. rin po che sna bdun does not refer to jewels only, as found in Lamotte (1935) and Keenan (2000). I follow here Powers (1995), Cornu (2005), and Cleary (1999).

35. The logical subject of ’jig rten gyi khams dpag tu med pa rgyas par’gengs pa’i’od zer chen po shin tu mnga’ ba is the palace (khang). Cornu (2005) and Keenan (2000) seem to read this phrase as a qualifier for the seven precious substances.

36. The first paragraph of the nidāna is a presentation of the place where the Buddha is dwelling. As already mentioned in the introduction, a succession of compounds, mainly bahuvrīhis, enables the topicalization of the temple (khang). Lamotte’s translation reflects this literary device, contrary to Powers who does not topicalize the palace to the same degree on account of some ambiguities regarding the logical subject of a few clauses describing this palace. To illustrate this point, it seems unclear whether the adjectives “steadfast,” “enduring,” or “free” in Powers’ translation qualify the temple or the beings attending it. Cornu mainly follows Powers here but the grammatical necessity to indicate the gender and number of qualifiers in French limits the risk of confusion, which is obviously not the case in English. Regarding the usage of tenses, Lamotte is the only translator who uses both narrative past and present in this first paragraph. He thus switches from the past tense to the present tense in order to describe the characteristics of the temple, a decision I chose not to follow in the present translation.

37. Lamotte, Cornu, and Powers do not translate the anaphoric pronoun de in ’jig rten las’das pa de’i bla ma’i dge ba’i rtsa ba las byung ba. Powers explains in a footnote (see Power 1995, p. 313, n. 3) that this pronoun refers to gnosis according to Wonch’uk, although his translation does not reflect this interpretation. Since wisdom has not been mentioned earlier in the text and since the pronoun de is anaphoric, I read de as referring to the Buddha. Moreover, the concept of “root of virtue” is usually associated with persons and we have a reference to dbang sgyur ba in the next qualifying phrase.

38. The clause dbang sgyur ba’i rnam par rig pa shin tu rnam par rig pa’i mtshan nyid is problematic. Lamotte translates it in the following way: “très pur, il se caractérise par une pensée maîtresse de soi.” Cornu and Powers follow the reading found in D, folio  2.a; S, folio 4.a; Kǫ, folio 1.a; L, folio 3.a; and H, folio 3.a ( dbang sgyur ba’i rnam par rig pa shin tu rnam par rig pa’i mtshan nyid) and render the two occurrences of rnam par rig pa by an apposition: “It was characterized by perfect knowledge, the knowledge of one who has mastery.” (Powers 1995, p. 5). However, in F, folio 4.b we find a variant reading which, I believe, makes more sense: dbang byed pa’i rnam par rig pa shin tu rnam par dag pa’i mtshan nyid. The Tibetan verbal prefix shin tu rnam par is used to render the upasarga su- in Sanskrit, like in suviśuddha. In Mvyut 351, blo shin tu rnam par dag pa thus translates the Sanskrit suviśuddhabuddhiḥ.

39. nges par ’byung ba. In Skt. niḥsaraṇa or niryāṇa, which have the meaning of setting forth, issue, exit, departure, escape, a road out of town. The analogy here is not about emancipation or renunciation as Powers and Cornu translated it but rather with the metaphor of the journey. In that sense, what is meant here is the departure to reach the palace. Lamotte (1935), Keenan (2000), and Cleary (1999) follow Xuanzang’s translation: 大念慧行以為游路 (Cbeta, Taishō 676). Interestingly enough, F does not have nges par ’byung ba but just ’byung ba.

40. rin po che’i pad ma’i rgyal po chen po yon tan gyi tshogs mtha’ yas pas brgyan pa’i bkod pa la rten pa na bzhugs te. This clause has been translated in various ways depending on how one understands the compound rin po che’i pad ma’i rgyal po chen po yon tan gyi tshogs mtha’ yas pas. Lamotte (1935), Powers (1995), and Cornu (2005) read it as a dvandva: “II est orné de qualités infinies, de joyaux, de lotus et de grands rois” (Lamotte 1935, p. 167); “this pattern was adorned with boundless masses of excellent qualities, and with great kingly jeweled lotuses” (Powers 1995, pp. 5–6); “paré d’infinies qualités et de grands lotus royaux incrustés de pierreries” (Cornu 2005, p. 26). However, it seems to me that it would be better to read this compound as a karmadhāraya. Folio 5.a offers a variant reading that could support this interpretation: yon ten gyi tshogs mtha’ yas pas/ brgyan pa’i rin po che chen po pad mo’i rgyal po’i bkod pa’i gnas na nyan thos kyi dge ’dun tshad med pa dang / thabs gcig tu bzhugs te. In addition to this problem, one should note that Lamotte’s translation of the compound rin po che’i pad ma’i rgyal po chen po as a dvandva is inaccurate here. Powers’ reading of this term is correct.

41. chos in the sense of qualities as understood by Lamotte (1935), Powers (1995), and Cleary (1999).

42. spyod yul; gocara. This term refers here to an object perceived by the six senses, so its semantic field pertains to perception as opposed to meditative practice, in which case it would be close in meaning to ālambana (“referential object”). Translating all these terms with “object” would conflate these various semantic fields in the context of the present text.

43. bsam gyis mi khyab pa rnam par ’jog pa (cf. rnam par gzhag pa bsam gyis mi khyab pa; acintyavyavasthānaḥ, see Mvyut 359). Compare with Lamotte: “ses attributs sont inconcevables” (Lamotte 1935, p. 168); Powers: “positing [doctrines] inconceivably” (Powers 1995, p. 7); Cornu: “il était entré dans l’indicible” (Cornu 2005, p. 26).

44. dus gsum mnyam pa nyid tshar phyin pa; tryadhvasamatāniryātaḥ (Mvyut 360). The term niryāta means here “adept, perfected, perfectly skilled” (see Edgerton 1953, p. 303).

45. byang chub sems dpa’ thams cad kyis ye shes yang dag par blangs pa. See Mvyut 366: ye shes byang chub sems dpa’ thams cad kyis yang dag par mnos pa; sarvabodhisattvasampratīcchitajñānaḥ. One should follow here the translations of Lamotte (1935), Keenan (2000), and Cleary (1999).

46. de bzhin gshegs pa ma ’dres pa’i rnam par thar par mdzad pa’i ye shes kyi mthar phyin pa. See Mvyut 368: de bzhin gshegs pa ma ’dres pa’i rnam par thar pa’i mdzad pa’i ye shes kyi mthar phin pa/ de bzhin gshegs pa ma ’dres pa’i rnam par thar par mdzad pa’i ye shes kyi mthar phin pa; asaṃbhinnatathāgata-vimokṣajñānaniṣṭhāgataḥ. See also Mvyut 5192: dbyer med pa; ma ’dres pa; ma ’dres pa’m dbyer med pa; asaṃbhedaḥ. If we understand ma ’dres pa in the sense of dbyer med pa, or even zung ’jug (yuganaddha), the meaning of the term is “indivisible/in unity,” conveying the notion of nonduality of the sameness mentioned several times in this introduction. Lamotte translates ma ’dres pa with “non diversifié,” Cornu with “distinctement,” Powers with “uniquely,” Keenan with “unified.” I don’t think one should understand ma ’dres pa with the meaning of kevala in the present case since it is associated with ye shes in other contexts where the idea of being exclusive to a particular person (e.g., buddhas) is negated (see Keenan 1980, p. 782ff.).

47. mtha’ dang dpung med pa’i sangs rgyas kyis mnyam pa nyid thugs su chud pa. One should read here instead: mtha’ dang dbus med pa’i sangs rgyas kyi sa mnyam pa nyid bu thugs su chud pa; anantamadhyabuddhabhūmisamatādhigataḥ (see Mvyut 369).

48. D, folio 2.b: nam mkha’i khams kyi mthas gtugs pa, which stands in apposition to chos kyi dbyings kyis klas pa (“the ultimate within the domain of truth”). See Mvyut 6430: nam mkha’i dbyings kyi mtha’ gtugs pa, nam mkha’i khams kyi mthar gtugs for the Sanskrit ākāśadhātuparyavasānaḥ. Compare with Mvyut 371: nam mkha’i khams kyi mtha’ klas pa, nam mkha’i khams kyi mthas klas pa as Tibetan equivalents of ākāśadhātuparyavasānaḥ. In Mvyut 431, don gyi mthar gtugs pa and don gyi mthar thug pa are Tibetan translations of paryavasitārthaḥ.

49. nges par ’byung ba; niryāṇika (?). Powers (1995) and Cornu (2005) translate it as “renunciation.”

50. See Mvyut 7450: nges par rtog pa/nges par rtogs pa; nirūpaṇā. Translated by Lamotte with “pénétrante” and by Keenan with “penetrating,” while Powers and Cornu opted respectively for “certain realization” and “réalisation certaine.”

51. See Mvyut 1113: yon yongs su sbyong ba chen po; mahādakṣiṇāpariṣodhakaḥ. D, 4,5; Kǫ, folio 2.b; and H, 7,2 omit yon.  However, it is present in S, 7,2 and L, folio 5.a: yon tan, while F.5.b reads sbyin pa.

52. Powers (1995) and Cornu (2005) read nges pa, but one should read here instead the graphically very similar des pa (“gentleness”) as in Mvyut 1115 where this expression is also found extensively: bzod pa dang des pa chen po dang ldan pa; mahākṣāntisauratyasamanvāgataḥ.

53. brjod du med pa dang / gnyis su med pa’i mtshan nyid. I read this compound as a bahuvrīhi. The full clause [brjod du med pa dang / gnyis su med pa’i mtshan nyid] + [don dam pa] is a karmadhāraya meaning literally “the ultimate that is that whose defining characteristic is inexpressible and absolute.” Powers’ suggestion is also possible here (“the ultimate whose defining characteristic is inexpressible and non-dual”). Lamotte leaves out mtshan nyid. Cornu somewhat mixes qualifiers and qualified terms in his rendering of this clause.

54. yongs su ma grub pa; apariniṣpanna (?). This paragraph establishes the opposition between the imaginary (parikalpita) and the actual (pariniṣpanna). These two aspects are found in the teaching on the three kinds of essencelessness: see Saṃdh. chapters 6–8.

55. shes pa dang mthong ba; jñānadarśana. D, folio 3.b: shes pa but F, folio 6.b; S, folio 5.a; VD, folio 44.b: ye shes. I emended the text in this way throughout the sūtra since this expression is repeated several times.

56. Xuanzang’s translation reads 謂諸聖者以聖智聖見離名言故現等正覺 (Cbeta, Taishō 676).

57. As noted by Tillemans (1997), Powers reads phyir in the sense of “because” here. Keenan and Cleary’s readings of Xuanzang’s translation (為慾令他現等覺故, Cbeta, Taishō 676) agree with Lamotte’s and Tillemans’ understanding of this passage.

58. ston pas btags pa’i tshig yin te. VD, folio 44.b: om.

59. rnam grangs; paryāya. The Dunhuang manuscript of the sūtra instead has gzhung du ’du shes (n°194 folio 62.a; see Hakayama 1984, p. 187).

60. See 1.5.

61. VD, folio 45.b reads the demonstrative pronoun as a plural (i.e., de dag) throughout this paragraph.

62. ’di ltar don in the sense of yathārtha.

63. Powers’ and Cornu’s translations are inaccurate here: “Subsequently they do not make the conventional designations: ‘This is true, the other is false.’ They make conventional designations because they completely understand the object in this way.” (Powers 1995, p. 17). “Comme ils connaissent parfaitement le sens réel de ces phénomènes…” (Cornu 2005, p. 32). Lamotte seems to have translated 1.4 on the basis of the Chinese. In addition, rjes su should be read as rjes su tha snyad for anuvyavahāra.

64. Lamotte’s rather free translation of 1.5 fails to render the opposition between what does not exist and what does according to the sūtra: ’di snyam du sems te/ ’dus byas dang / ’dus ma byas snang ba gang yin pa ’di ni med kyi/ gang la ’dus byas dang ’dus ma byas kyi ’du shes dang / ’dus byas dang / ’dus ma byas kyi rnam grangs kyi [F.5.a] ’du shes ’byung ba/ rnam par rtog pa las byung ba/ ’du byed kyi mtshan ma sgyu ma lta bu ’di ni yod/ blo rnam par rmongs par byed pa ’di ni yod do. The first chapter gives an introduction to central concepts, such as conditioned/unconditioned, existent/nonexistent, imaginary/actual. It prefigures the treatment of the two truths (bden pa gnyis; satyadvaya) in Saṃdh. chapter 3 as well as that of the three natures/essences and three kinds of nonexistence of nature/essence (ngo bo nyid (med pa) nyid; (niḥ)svabhāvatā) expounded in Saṃdh. (cf. chapters 6–8).

65. brtsams pa; ārabhya with the meaning of “referring to/having to do with,” a frequent occurrence in Saṃdh. See Edgerton 1953, p. 102.

66. rtog ge thams cad las yang dag par ’das pa; sarva­tarka­samati­krānta. Regarding the translation of the term rtog ge (tarka), Powers 1995, p. 25 suggests “argumentation,” but the emphasis in the present context is not on logical reasoning. The term tarka denotes here any kind of assumption, presupposition, representation, or conjecture regarding the absolute that is the product of the intellect (manas).

67. The English translation of this passage should convey the paradox of the situation. Although the ultimate is inexpressible, the Buddha gave countless teachings. Syntactic connective particles between clauses about the inexpressible ultimate and the domain of expression and reasoning have therefore an adversative meaning in the present paragraph: ngas ni don dam pa rtog ge thams cad las yang dag par ’das pa’i mtshan nyid mngon par rdzogs par sangs rgyas te/ mngon par rdzogs par sangs rgyas nas kyang bsnyad cing gsal bar byas/ rnam par phye/ gdags par byas/ rab tu bstan to/ de ci’i phyir zhe na. Lamotte’s and Powers’ translation do not make this point clear.

68. so so(r) rang rig pa; pratyātmavedya/pratyātmavedanīya/pratyātmajñāna/prātyatmam (see Schmithausen 2014, p. 346), in the sense of realizing or understanding for oneself in an intuitive way, as personal experience. Powers’ translation does not reflect the meaning of this term: “I have explained that the ultimate is realized individually by the Aryas, while objects collectively known by ordinary beings [belong to] the realm of argumentation.” In the sūtra, so so(r) rang rig pa is a synonym or a qualifier of ye shes (jñāna).

69. Lamotte translates rtog ge’i spyod yul by “affaire de tradition”; see Lamotte 1935, p. 173. Beyond the fact that this is wrong, it is worth noting that the opposition here is between the intuitive and personal knowledge of the noble beings and the intellectual and transactional knowledge of ordinary beings, namely, between gnosis and mind. Powers 1995, p. 27 reflects the personal aspect of so sor rang gi rig but not its intuitive quality.

70. mtshan ma; nimitta. Although “notion” would fit well here, one should keep in mind that nimitta as a polysemic term denotes phenomenal appearance throughout the sūtra. However, it is clear that “phenomenal appearance” and “notion” are two terms that are joined at the hip from the perspective of the doctrine expounded in Saṃdh.

71. In 2.2, the Buddha mentions a paradox. He gives explanations about the ultimate in speculative terms, although the ultimate is inexpressible. This paradox is best rendered in English or French by reading the particle la in the statements in question as having an adversative meaning. For example: chos ’phags gzhan yang don dam pa ni tha snyad thams cad yang dag par chad pa yin par ngas bshad la/ rtog ge ni tha snyad kyi spyod yul yin te.

72. The Turfan Sanskrit fragment found by Matsuda reads paramārtha [dhar]mā vigatābhilāpaḥ at the end of chapter 2’s closing gāthā (cf. Sanskrit text in Matsuda 2013: p. 940 ad Lamotte VIII.41). In D, folio 6.b and F, folio 10.b, we have instead the Tibetan term rtsod dang bral ba. It is possible that brjod med in line 2 and rtsod dang bral ba were inverted metri causae.

73. Powers reads don dam pa and mtshan nyid as being in apposition in the sentence de [i.e., don dam pa] ni rtog ge kun las ’das mtshan nyid. Lamotte reads rtog ge kun las ’das mtshan nyid (sarva­tarka­samati­krāntalakṣaṇa) as a bahuvrīhi, which is much better. See Powers 1995, p. 31 and Lamotte 1935, p. 174.

74. I am using the adjective “indistinct” here in the sense of the first definition given in the Oxford English Dictionary: “1. Not distinct or distinguished from each other, or from something else; not kept separate or apart in the mind or perception; not clearly defined or marked off.” Oxford English Dictionary Online, s.v. “indistinct,” accessed July 20, 2020, https://www-oed-com.ezproxy.leidenuniv.nl:2443/view/Entry/94602?redirectedFrom=indistinct#eid.

75. mos pa; praṇidhāna. See mos pa spyod pa’i sa. See Mvyut 897: mos pa spyod pa’i sa; adhimukticaryābhūmiḥ.

76. Schmithausen reads don dam pa’i mtshan nyid (paramārthalakṣaṇa) as “the defining characteristic that is the ultimate” in 3.5 (see Schmithausen 2014, p. 558, §512.3). However, Saṃdh. chapter 3 is about conditioned phenomena in relation to the ultimate when their respective defining characteristics are examined. The question here is not to determine whether the ultimate is the defining characteristic of conditioned phenomena. Rather, it is to determine whether the conditioned and the ultimate are different by examining their defining characteristics. Therefore, I read don dam pa’i mtshan nyid as “the defining characteristic of the ultimate,” namely, as a genitive tatpuruṣa and not as a karmadhāraya.

77. To render sha stag.

78. I did not translate the phrase so so’i skye bo kho nar gyur bzhin du in an adversative mode (e.g., “though merely ordinary”) because the same phrase is found in the next paragraph in a parallel construction where the syntax in relation to the meaning cannot be interpreted to express contrast. Here bzhin du stands for yathā in the sense of “as” (i.e., “en tant que” in French).

79. grub pa dang bde ba; yogakṣema (see Edgerton 1953, p. 448a,b). Refer to Tillemans 1997, p. 157ff. for a discussion of Powers’ rendering of the term in his translation of Saṃdh. (Powers 1995). Lamotte translates this term with “de sécurité suprême.” See Lamotte 1935, p. 175.

80. I linked the two clauses with “and” because these two clauses are part of the logical argument on being “not different.” We have here a relative-correlative syntactic structure: gang gi phyir … de’i phyir  (i.e., yasmāt … tasmāt). The first two clauses linked by “and” represent the hetu (“premise” or “reason”) posited by yasmāt. The logical structure of the paragraph is “since (a), (b), and (c), therefore (d)” (the conclusion that is stated in the very next sentence “For this reason…”): “Su­viśuddha­mati, ordinary beings (a) do not realize the truth and, as mere ordinary beings, (b) neither do they attain nirvāṇa, the unsurpassable good, (c) nor do they fully and completely awaken to the unsurpassable complete and perfect awakening. For these reasons (de’i phyir, referring to (a), (b), and (c)), it is not correct to say that the defining characteristic of conditioned phenomena and the defining characteristic of the ultimate are indistinct.” D, folio 7.b: blo gros shin tu rnam dag gang gi phyir so so’i skye bo bden pa mthong ba ma yin/ so so’i skye bo kho nar gyur bzhin du grub pa dang / bde ba bla na med pa’i mya ngan las ’das pa ’thob par yang mi ’gyur/ bla na med pa yang dag par rdzogs pa’i byang chub mngon par rdzogs par ’tshang rgya bar mi ’gyur ba de’i phyir ’du byed kyi mtshan nyid dang / don dam pa’i mtshan nyid tha dad pa ma yin zhes bya bar mi rung ste.

81. This entire paragraph is problematic in D: blo gros shin tu rnam dag gang gi phyir bden pa mthong ba rnams ’du byed kyi mtshan ma dang bral ba ma yin gyi/ bral ba kho na yin pa dang / bden pa mthong ba mtshan ma’i ’ching ba las rnam par grol ba ma yin gyi/ rnam par grol ba yin pa dang / bden pa mthong ba gnas ngan len gyi ’ching ba las rnam par grol ba ma yin gyi/ rnam par grol ba dang / ’ching ba de gnyi ga las rnam par grol na grub pa dang / bde pa bla na [F.8.a] med pa’i phyir mya ngan las ’das pa ’thob par ’gyur ba dang / bla na med pa yang dag par rdzogs pa’i byang chub mngon par rdzogs par ’tshang rgya bar yang ’gyur ba. However, Lamotte, Keenan, and Cleary translate the phrases in bold with a double negation. If we look at the same paragraph in F, folio 10.a, we find an interesting textual variant in which the expected double negation is found: ma bral ba’ang ma yin te and ma grol ba’ang ma yin te, just like in Xuanzang’s translation (由此道理當知一切非如理行 不如正理善清淨慧。由於今時非見諦者。於諸行相不能除遣。然能除遣非見諦者。 Cbeta, Taishō 676). Powers chose to translate this passage without proceeding to any emendation, which makes little sense from the perspective of the argument expounded in this section of the text: “Su­viśuddha­mati, it is not the case that seers of truth are free from the signs of the compounded; they are simply free. Moreover, seers of truth are not liberated from the bondage of signs, but they are liberated. Seers of truth are not liberated from the bondage of errant tendencies, but they are liberated.” (Powers 1995, p. 41).

82. We have here again a complex relative-correlative syntactic structure: gang gi phyir … de’i phyir (i.e., yasmāt … tasmāt). All the clauses between gang gi phyir and de’i phyir represent the hetu (“premise” or “reason”) posited by yasmāt. The logical structure of the paragraph is “since (a), (b), (c), (d), therefore (e)” (the conclusion that is stated in the very next sentence, “For all these reasons,…”): “Su­viśuddha­mati, it is not the case that (a) those who realize the truth are not detached from the phenomenal appearance of conditioned phenomena, for they are indeed detached from it. (b) Neither are they not liberated from the bondage of phenomenal appearance, for they are indeed liberated from it. (c) Nor are they not liberated from the bondage of corruption, for they are indeed liberated from it. (d) Since they are liberated from these two kinds of bondage, [F.8.a] not only do they attain nirvāṇa, the unsurpassable good, but they will also fully and completely awaken to the unsurpassable complete and perfect awakening. (e) For all these reasons (de’i phyir referring to (a), (b), (c), and (d)), it is not correct to say that the defining characteristic of conditioned phenomena and the defining characteristic of the ultimate are distinct.” D, folios 7.b–8.a: blo gros shin tu rnam dag gang gi phyir bden pa mthong ba rnams ’du byed kyi mtshan ma dang bral ba ma yin gyi/ bral ba kho na yin pa dang / bden pa mthong ba mtshan ma’i ’ching ba las rnam par grol ba ma yin gyi/ rnam par grol ba yin pa dang / bden pa mthong ba gnas ngan len gyi ’ching ba las rnam par grol ba ma yin gyi/ rnam par grol ba dang / ’ching ba de gnyi ga las rnam par grol na grub pa dang / bde ba bla na med pa’i phyir mya ngan las ’das pa ’thob par ’gyur ba dang / bla na med pa yang dag par rdzogs pa’i byang chub mngon par rdzogs par ’tshang rgya bar yang ’gyur ba de’i phyir ’du byed kyi mtshan nyid dang / don dam pa’i mtshan nyid tha dad pa zhes byar mi rung ste.

83. On a similar line of thought, see 4.10.

84. gdags pa; prajñapti. Lamotte translates this term with “dire.” This does not convey the meaning of gdags pa, which implies the idea of imputation, intimation, conceptualization, or representation. Here, in the sense of “decide,” see Edgerton 1953, p. 359,1.

85. mi bzod; na kṣamate, na kṣamati (see Edgerton 1953, p. 199,1). Pāli: khamati (see The Pali–English Dictionary (Rhys-Davids 1921), p. 234).

86. D: blo gros shin tu rnam dag ngas ni de ltar don dam pa phra ba mchog tu phrag mchog tu zab pa/ rtogs par dka’ ba/ mchog tu dka’ ba/ tha dad pa dang / tha dad pa ma yin pa nyid las yang dag par ’das pa’i mtshan nyid mngon par rdzogs par sangs rgyas te. I read tha dad pa ma yin pa nyid las yang dag par ’das pa’i mtshan nyid as a bahuvrīhi, not a tatpuruṣa. The terms don dam pa and tshan nyid are not in apposition in chapter 3, since the two terms are repeatedly connected through a genitive particle: don dam pa’i mtshan nyid (which occurs 32 times in folios 5.a, 5.b, 6.b, 7.a, 7.b, 8.a, 8.b, and 9.a). I also understand similar constructions with don dam pa and tshan nyid in the following chapter as bahuvrīhis.

87. This phrase is also found in 2.2, 3.6, and 4.7 (see also folios 5.b, 9.a–b, and 10.b–11.a): mngon par rdzogs par sangs rgyas nas kyang bsnyad cing gsal bar byas/ rnam par phye/ gdags pa byas/ rab tu bstan to.

88. Powers translates mtshan nyid as singular (see Powers 1995, p. 49). But the question of the identity or difference in this chapter is evaluated from the perspective of two defining characteristics, namely, the defining characteristic of the conditioned and the defining characteristic of the unconditioned.

89. The last two lines are quoted in the Bhāvanākrama; see Tucci 1971, p. 1: nimittabandhanāj jantur atho dauṣṭhulabandhanāt/ vipaśyanāṃ bhāvayitvā śamathañ ca vimucyata iti.

90. Lit. “in the world of beings.”

91. F reads here shes pa in agreement with D. See F, folio 14.bff.

92. dmigs pa; ālambana. I think it is important here to read dmigs pa as meaning “object” because in folio 11.a the Buddha contrasts these various objects (aggregates, sense sources, constituents, truths, etc.) with the “object conducive to purification” (rnam par dag pa’i dmigs pa, *viśuddhyālambana; see Schmithausen 2014, p. 362, §306.5 and n. 1644). Translating dmigs pa here as “observing” would weaken the central opposition between (a) the objects taken as a reference point for their practice by those who have not realized the defining characteristic of the ultimate and (b) the object conducive to purification, which is present within all phenomena. The purpose of this chapter is to introduce this fundamental point.

93. This paragraph deals with the thirty-seven branches of awakening (byang chub kyi yan lag; bodhyaṅgāni).

94. ro gcig pa; ekarasa. I read the compound thams cad du ro gcig pa’i mtshan nyid; *sarvatraikarasalakṣaṇa (?) as a bahuvrīhi; see D, folio 12.a: rab ’byor de bzhin du don dam pa yang mtshan nyid tha dad pa’i chos rnams la thams cad du ro gcig pa’i mtshan nyid yin par blta bar bya’o. I understand thams cad du ro gcig pa’i mtshan nyid to refer here to the defining characteristic of the ultimate since this definition presents dharma as having various lakṣaṇa. As a consequence, I read this sentence as stating that the ultimate is that whose defining characteristic is always of a single nature in all phenomena that have diverse defining characteristics. See also 4.8, which supports this interpretation.

95. rnam par dag pa’i dmigs pa; *viśuddhyālambana. Lamotte reads here rnam par dag pa’i dmigs pa with the meaning of viśuddhālambana (“objet pur”), but Schmithausen gives *viśuddhyālambana as the Sanskrit equivalent for the Tibetan (see Schmithausen 2014, p. 362, §306.5 and n. 1644). It is worth noting that this term is also found in chapter 7, where it is again equated with the ultimate (don dam pa). In this context, it is said that paratantra is not an object conducive to purification whereas the actual (pariniṣpanna) is. In this sense, one should make the distinction here between vastu and ālambana. The ultimate is conceived here as the referential object, or support object, of a purification that leads to awakening. The usage of ālambana in reference to the ultimate clearly refers to practice in the present chapter.

96. Compare D, folio 11.a: rab ’byor rnam grangs des na khyod kyis ’di ltar thams cad du ro gcig pa’i mtshan nyid gang yin pa de ni don dam pa yin par rig par bya’o with D, folio 12.a: rab ’byor de bzhin du don dam pa yang mtshan nyid tha dad pa’i chos rnams la thams cad du ro gcig pa’i mtshan nyid yin par blta bar bya’o (passim).

97. de bzhin nyid; tathatā. I translate tathatā with “true reality” (in the sense of the true state or nature of things) instead of the more usual “suchness” or “thusness.” However, when de bzhin nyid is qualified by an adjective such as don dam pa, I simply translate it with “reality” to improve readability. See Schmithausen 2014, p. 356, §303.1 (passim), in which tathatā is translated with “True Reality” or “Suchness.”

98. D reads gnyis med pa’i shes pa, while F, folio 16.b reads gnyis su med pa’i ye shes.

99. kho na; eva. For a comparison between Buescher’s and Schmithausen’s translations of this sentence, see Schmithausen 2014, p. 380, §324.

100. D: de bzhin gshegs pa rnams byung yang rung ma byung yang rung ste/ rtag pa rtag pa’i dus dang / ther zug ther zug gi dus su chos gnas par bya ba’i phyir chos rnams kyi chos nyid dbyings de ni rnam par gnas pa kho na yin pa. Lamotte translates chos gnas par bya ba’i phyir with “pour le maintien des choses,” but phyir does not have a dative function here. Powers’ translation reads, “because phenomena abide in permanent, permanent time and in everlasting, everlasting time, the domain of reality of phenomena alone abides.” This does not make sense either. Conditioned phenomena are impermanent as explained at length in the first four chapters. The argument simply runs as follows: since it is present in all things, this alone is permanent. As for the expression chos nyid dbyings, D, folio 11.b reads chos rnams kyi chos nyid dbyings while F, folio 17.a has chos rnams kyi chos nyid/ chos gnas pa’i dbyings, referring respectively to dharmatā and dharmadhātu (compare with Mvyut 1719: chos gnas pa nyid; dharmasthititā). I am reluctant to translate dbyings (dhātu) as “realm/domain” here because the meaning of dhātu as “constituent” makes so much sense, particularly when reading F, in which dbyings is glossed as chos gnas pa, “that which abides within phenomena,” “that which is the support/source of phenomena,” or “the condition of phenomena.” Xuanzang’s translation confirms the suggested translation: 唯有常常時恒恒時如來出世若不出世諸法法性安立法界安住 (Cbeta, Taishō 676).

101. ji tsam gyis; kiyant. The complete sentence reads, “In what sense are they skilled in the secrets of mind, thought, and cognition?”

102. D: blo gros yangs ba ’gro ba drug gi ’khor ba ’di na sems can gang dang gang dag sems can gyi ris gang dang gang du yang sgo nga nas skye ba’i skye gnas sam/ yang na mngal nas skye ba’am/ yang na drod gsher las skye ba’am/ yang na rdzus te skye ba’i skye gnas su lus mngon par ’grub cing ’byung bar ’gyur ba der. “For whatever sentient beings an individual existence (ātmabhāva) comes about and emerges in this saṃsāra comprising six destinations, in whatever community of beings, be it in [the mode of] egg-born, womb-born, moisture-born or spontaneous generation, there…” (Schmithausen 2014, p. 177, §149). Regarding lus mngon par ’grub cing ’byung bar ’gyur ba, one finds the Sanskrit equivalent ātmabhāvam abhinirvartayati for lus mngon par ’grub ’gyur ba (see Yokoyama, Kōitsu, and Hirosawa Takayuki, eds., Index to the Yogācāra­bhūmi (Chinese-Sanskrit-Tibetan) (Tokyo: Sankibō Busshorin, 1996)). Schmithausen reads here, “For whatever sentient beings an individual existence (ātmabhāva) comes about and emerges” (Schmithausen 2014, p. 177, §149), but it seems to me that we could have had here a causative (ātmabhāvam abhinirvartayanti) with the Literal meaning of producing or bringing about a [new] existence, in a word, being reborn (for an instance of this rendering, see ibid., p. 191, n. 812). Otherwise, one would expect a syntactic particle in Tibetan after sems can gang dang gang dag to match Schmithausen’s reading, “For whatever sentient beings,” particularly if one understands ’gro ba drug gi ’khor ba ’di na sems can gang dang gang dag as not in apposition to sems can gyi ris gang dang gang du, which I think is the correct way to understand the expression.

103. I follow here the second possible interpretation of the correlative-relative syntactic structure of the sentence as suggested by Schmithausen (see Schmithausen 2014, p. 178, §150), in which der (tatra) is read as having a temporal connotation, rendered by the adverb “then.”

104. D: sa bon thams cad pa’i sems rnam par smin cing ’jug la rgyas shing ’phel ba dang yangs par ’gyur ro. See Schmithausen 1987, p. 356, n. 508; and 2014, p. 325, n. 1490 for the Sanskrit reconstruction: *( sarvabījakaṃ cittaṃ) vipacyate saṃmūrcchati vṛddhiṃ virūdhiṃ vipulatām apadyate. See also Waldron 2003, p. 218, n. 13. Lamotte (1973, p. 65ff.) suggests kalalatvena saṃmūrcchati in his translation of the Mahāyānasaṃgraha I,34. The appended commentary describes how the appropriating cognition enters the mother’s womb by uniting with the semen and the blood (śukraśoṇita). The embryo with which, from then on, the appropriating cognition shares a common destiny (ekayogakṣema) is the result of this “coagulation.” On the topic, see Kritzer 2000.

105. See Waldron 2003, pp. 94–95 for a translation of 5.2. See also Brunnhölzl 2018, p. 1305ff. regarding Asaṅga’s and Wonch’uk’s commentaries on Saṃdh. 5.2–7.

106. kun tu bsags pa; ācita. This explanation corresponds to an etymological analysis of the Sanskrit term citta. Xuanzang’s translation reads: 由此識色聲香味觸等積集滋長故 (Cbeta, Taishō 676).

107. See Schmithausen 2014, p. 157ff. for an extensive discussion of the various testimonies and interpretations of this sentence. See also Waldron 2003, pp. 95–96 for a translation of 5.3, and Tillemans 1997, pp. 157–58 for a discussion of Powers’ translation of 5.3. Powers translates the verb in last sentence with an active form “because it collects and accumulates forms, sounds…” (Powers 1995, p. 71). It seems to me that, since this is an etymological explanation of citta, a passive is better in the present case.

108. rnam par shes pa’i tshogs drug po, lit. “the sixfold group of cognitions.”

109. Waldron’s translation (Waldron 2003, p. 97) seems to follow the Sanskrit and Tibetan found in Sthiramati’s Triṃśikāvijñaptibhāṣyam (see TrBh, 33.25–34.4 in Buescher 2007). However, the quote of Saṃdh. in this treatise is drawn from 5.5. It does not quite match the last sentence of 5.4. See D: gal te rnam par shes pa’i tshogs gnyis sam/ gsum mam bzhi lhan cig gam/ lnga car lhan cig ’byung na yang der rnam par shes pa’i tshogs lnga po dag dang / spyod yul mtshungs pa rnam par rtog pa’i yid kyi rnam par shes pa yang gcig kho na lhan cig ’byung ngo. Compare with TrBh, 33.26–27 (Buescher 2007): gal te rnam par shes pa gnyis sam gsum mam lnga ’byung ba’i rkyen nye bar gnas par gyur na yang gnyis sam gsum mam lnga’ bar du ’byung bar ’gyur ro. Sanskrit: saced dvayos trayānāṃ sacet pañcānāṃ vijñānām utpattipratyayaḥ pratyupasthito bhavati sakṛd yāvat pañcānāṃ pravṛttir bhavati.

110. For a comparison with the Dunhuang edition of 5.5, as well as for a general evaluation of the relevance of this edition, see Schmithausen 2014, p. 419, n. 1852.

111. D: gal te rlabs gnyis sam gal te rab tu mang po dag ’byung ba’i rkyen nye bar gnas par gyur na rlabs rab tu mang po dag ’byung but VD, folio 54.a: gal te rlabs gnyis sam gal te rab tu mang po dag ’byung ba’i rkyen nye bar gnas par gyur na rlabs kyang gnyis sam rab tu mang po dag ’byung. S, folio 19.a is in agreement with D here (gnyis sam is omitted).

112. On the metaphor of the mirror in Buddhist texts, see Wayman 1974.

113. D: gal te gzugs brnyan gnyis sam gal te rab tu mang po dag ’byung ba’i rkyen nye bar gnas par gyur na gzugs brnyan rab tu mang po dag ’byung but VD, folio 54.a: gal te gzugs brnyan gnyis sam gal te gzugs brnyan rab tu mang po dag ’byung ba’i rkyen nye bar gnas pa gyur na gzugs brnyan yang gnyis sam rab tu mang po dag ’byung.

114. Lamotte 1935, p. 186 translates yongs su sbyor bar yang mi mngon with “ne souffre aucun dommage.”

115. shes pa, although it is worth noting that the Dunhuang manuscript reads ye shes. See Hakayama 1986, 11 E6 (3).

116. D: blo gros yangs pa gang gi phyir byang chub sems dpa’ rang gi so so nang gi len pa mi mthong / len pa’i rnam par shes pa yang mi mthong la/ de yang dag pa ji lta ba bzhin du yin pa. I am following here Schmithausen’s reading of gang gi phyir (yataḥ) in the sense of “as soon as” (see Schmithausen 2014, p. 346–47, n. 1577). This interpretation is confirmed by F, folio 19.b, which reads nam instead of gang gi phyir.

117. nang gi so sor rang rig pa; adhyātmaṃ prātyatmam (see Schmithausen 2014, p. 346–47, n. 1577). See F, folio 19.b: nang gi so so rang.

118. “Neither the accumulated nor the mind” is a pun on ācita and citta.

119. Sanskrit verses are found in Buescher 2007, see TrBh, 34.2–3. Also translated into English in Waldron 2003, p. 101.

120. yongs su grub pa’i mtshan nyid; pari­niṣpanna­lakṣaṇa. I read these compounds as karmadhārayas, not as genitive tatpuruṣas. For an analysis of these compounds, see Schmithausen 2014, p. 359, n. 1626.

121. rnam par bzhag pa; vyavasthā, which has the connotation of something being agreed upon, represented, arranged, settled, decreed, or established.

122. Frauwallner and Lamotte do not translate ji tsam du (see Frauwallner 1969, p. 285: “um sie im täglichen Sprachgebrauch zu bezeichnen” and Lamotte 1935, p. 188: “permettant de les mentionner dans le langage courant”). The Chinese term 乃至 in Xuanzang’s translation (乃至為令隨起言說, Cbeta, Taishō 676) corresponds to the Tibetan ji tsam du (for yāvat or kiyat), which I understand here in the sense of “at least, only, even just.” Interestingly enough, the Go witness has ci tsam gis (see Go, folio 12.a); F, folio 20.a reads gang ji skad du, and He, folio 77.a returns gang ji snyed du (while rjes su is omitted) for ji tsam du in D.

123. The Tibetan rjes su tha snyad gdags pa and the Chinese 隨起言說 correspond to anuvyavahāra.

124. D: ji tsam du rjes su tha snyad gdags pa’i phyir chos rnams kyi ngo bo nyid dam bye brag tu ming dang brdar rnam par gzhag pa gang yin pa’o. Frauwallner’s translation of this passage is slightly ambiguous and could be potentially misleading on account of the German preposition “nach”: “Es ist jede Festsetzung eines Namens und einer Vereinbarung für die Gegebenheiten nach Wesen oder Besonderheit, um sie im täglichen Sprachgebrauch zu bezeichnen.” (Frauwallner 1969, p. 286). If one reads this sentence as “a determination in terms of being and specific defining characteristic,” then the translation is correct. However, if one reads it as meaning “the determination of … in accordance with/corresponding to [their] being or specific defining characteristic,” then the translation becomes problematic since the point made in the text is precisely that there is no such thing. The determination (Festsetzung) of a name and convention for phenomena (für die Gegebenheiten) is not made according to the essence or defining characteristic of phenomena (nach Wesen oder Besonderheit). Hence the term “imaginary defining characteristic.” If this determination occurred according to the essence of phenomena, it would be illogical to call this defining characteristic “imaginary.” Lamotte’s translation is therefore more accurate in the present case, although his rendering of rnam par bzhags pa (vyavasthāna) as a present participle qualifying ming dang brda slightly modifies the original meaning of the sentence. Sanskrit compounds with vyavasthāna can be read as instrumental tatpuruṣa, for example saṃjñākaraṇavyavasthāna, mātrāvyavasthāna, or aṅgavyavasthāna. Xuanzang translates this passage as 云何諸法遍計所執相。謂一切法名假安立自性差別。乃至為令隨起言說。 (Taishō 676), in which nāmasaṃketa (“names and common references”) seem at first glance to be the grammatical passive subject of the main verb of the clause (安立) or an adverb. In fact, F, folio 20.a and He, folio 77.a read ming dang brdas rnam par bzhag pa instead of ming dang brdar rnam par bzhag pa in D and other witnesses mentioned in the preceding notes. The terminative of the Tibetan ming dang brdar could be read adverbially in the sense of “nominally and conventionally” in the sense of “by means of/in terms of names and common references” for nāmasaṃketa.

125. Regarding the choice of terminology for pariniṣpanna, Edgerton mentions two possible meanings (see Edgerton 1953, p. 325): (1) “completely perfected” in the sense of “accomplished/attained,” which corresponds to the second clause in the present definition (6.6) in which the yong su grub pa is glossed in the sense of yang dag par ’grub pa (D) and yang dag par bsgrub pa (F). (2) The second meaning of pariniṣpanna according to Edgerton is “absolute.” The term is also translated into English as meaning “real.” Understanding pariniṣpanna as “established” or “perfected” is indeed problematic. Reality in the sense of tathatā cannot be referred to as “perfected,” because it is not perfectible. If it were, it would be conditioned (see Saṃdh. chapter 3). Likewise, tathatā cannot be termed “established” because if it were, it would be impermanent, but it is said to be the only permanent reality (see Saṃdh. chapter 4). As a consequence, it appears clearly from the context of the definition given above and Edgerton’s explanations that pari­niṣpanna­lakṣaṇa refers to both the “character accomplished [by bodhisattvas]” in the sense of the accomplishment of the defining characteristic of the ultimate as the ultimate attainment of the bodhisattva’s spiritual path, and the “absolute or real character” in the sense of reality itself (tathatā). As emphasized throughout Saṃdh., pari­niṣpanna­lakṣaṇa is indeed none other than the defining characteristic of the ultimate (don dam pa’i mtshan nyid; paramārthalakṣaṇa), the object conducive to purification, (rnam par dag pa’i dmigs pa; *viśuddhyālambana). In the present translation, I opt for “actual,” a term that fits well with the metaphors found in Saṃdh. chapter 6, to explain the three natures (Lamotte chose “absolu,” and Frauwallner “vollkommen”). The term “actual” obviously induces a clear dichotomy between what is imaginary and what is real, which was most probably at the origin of these terminological choices in the source language. On the opposition between the imaginary and the actual by equating the imaginary with the unreal through the usage of the same terminology (i.e., pariniṣpanna and parikalpa/parikalpita), see 1.2.

126. The syntax of this sentence differs according to the various available editions of the text with little bearing on the meaning of the sentence. For example, compare D; VD, folios 55.a–b: chos rnams kyi de bzhin nyid gang yin pa ste/ byang chub sems dpa’ rnams kyis rtun pa’i rgyu dang / legs par tshul bzhin yid la byas pa’i rgyus de rtogs shing de rtogs pa goms par byas pa yang dag par grub pas kyang bla na med pa yang dag par rdzogs pa’i byang chub kyi bar du yang dag par ’grub pa gang yin pa’o; F, folio 20.a–b: gang chos rnams kyi de bzhin nyid de/ byang chub sems dpa’ rnams kyis brtson ba’i rgyu dang / yang dag par tshul bzhin yid la byed pa’i rgyud gang rtogs par ’gyur ba yin te rtogs dang / bsgoms pa bsgrubs pas/ bla na med pa yang dag par rdzogs pa’i byang chub kyi bar du yang dag par bsgrub pa’o; and S, folio 20.b: gang chos rnams kyi de bzhin nyid/ gang byang chub sems dpa’ rnams kyis brtson ba’i rgyu dang / yang dag par tshul bzhin yid la byed pa’i rgyus rtogs par ’gyur ba yin te/ de rtogs pa goms par byas pa yang dag par grub pas kyang / bla na med pa yang dag par rdzogs pa’i byang chub kyi bar du yang dag par ’grub pa’o. Frauwallner reads this passage in the following way: “Es ist die Soheit der Gegebenheiten, ihr Erschauen durch die Bodhisattva auf Grund ihrer Energie und richtigen Beobachtung (yoniśomanasikāraḥ), und durch das Zustandekommen der Übung dieses Erschauens schließlich das Zustandekommen der hochsten vollkommenen Erleuchtung.” (Frauwallner 1969, p. 286). Lamotte 1935, p. 189 is similar to Frauwallner’s reading. It is plausible that Frauwallner simply followed Lamotte’s translation of this passage. Schmithausen suggests that the relative clause starting with gang byang chub specifies the tathatā, which I agree with, although reading the two clauses defining pari­niṣpanna­lakṣaṇa as being in apposition would not negate the fact that the second one merely specifies the first one (tathatā) as being that which is attained by bodhisattvas: “(Suchness) into which the bodhisattvas, in consequence of their zeal and in consequence of their appropriate contemplation, obtain direct insight (*prativedha, cf. Saṃdh_DH: so sor chud pa), and, through the accomplishment of the repeated practice of this insight, finally attain the Highest Perfect Awakening.” (Schmithausen 2014, p. 540, n. 2241). It is important here to note that rtogs shing de rtogs pa goms par byas pa stands on the same level in relation to yang dag par grub pas, which is not apparent in Schmithausen’s translation. Thus, “obtain direct insight” (de rtogs shing) is part of a larger argument: rtogs shing de rtogs pa goms par byas pa (D) or rtogs par ’gyur ba yin te rtogs dang / bsgoms pa bsgrubs pas (F). This meaning unit should therefore not be put on the same level as yang dag par ’grub pa gang yin pa’o as it is in Schmithausen’s translation, since it is one of the reasons why there is an accomplishment according to the available testimonies. Regarding the translation of bar du (yāvat), I followed Schmithausen’s way of solving the problem. Powers understands it in its usual sense of “up to,” but it does not work here, since this would imply that what are accomplished are also inferior realizations, which Powers refers to between square brackets as “stages”; see Powers 1995, p. 83. Schmithausen’s reading of the second clause as specifying tathatā seems preferable.

127. de lta bur ni; evam eva.

128. Lamotte translates this term with “essentiellement fantaisiste”; see Lamotte 1935, p. 190. This is missing the point that this actually refers to the imaginary defining characteristic.

129. D; VD, folio 56.a: gzhan gyi dbang gi mtshan nyid de/ kun brtags pa’i mtshan nyid der rtag pa rtag pa’i dus dang / ther zug ther zug gi dus su yongs su ma grub cing ngo bo nyid med pa nyid kyis yongs su grub pa’i mtshan nyid blta bar bya’o. F, folio 21.b: gzhan gyi dbang de nyid la/ kun brtag brtags pa’i mtshan nyid de rtag pa rtag pa’i dus dang / ’khor bar ’khor ba’i dus su ma grub cing rang bzhin med pa nyid ni/ yongs su grub pa’i mtshan nyid du blta’o; S, folio 21.b: gzhan gyi dbang de nyid kun brtags pa’i mtshan nyid der rtag pa rtag pa’i dus dang / ther zug ther zug gi dus su yongs su ma grub cing rang bzhin med pa nyid ni/ yongs su grub pa’i mtshan nyid du blta’o. F and S are helpful here to interpret the syntax of D. The following simple emendation would improve the syntax of D: gzhan gyi dbang gi mtshan nyid de la/ kun brtags pa’i mtshan nyid der rtag pa rtag pa’i dus dang / ther zug ther zug gi dus su yongs su ma grub cing ngo bo nyid med pa nyid kyis yongs su grub pa’i mtshan nyid blta bar bya’o. Lamotte’s translation of 6.9 is more intelligible than that of Frauwallner. who is more Literal and strictly follows the syntax found in D. Xuanzang’s translation reads: 即依他起相上由遍計所執相於常常時於恆恆時無有真實無自性性圓成實性當知亦爾 (Cbeta, Taishō 676). In my translation I chose to topicalize yongs su grub pa’i mtshan nyid (in agreement with D) because the purpose of this dialogue is to define the three defining characters.

130. rab tu shes; prajānāti with the meaning of pratijānāti; see Edgerton 1953, p. 357. Lamotte translates this term with “repose,” which does not convey the meaning of prajānāti; see Lamotte 1935, p. 190. In the following paragraph, Lamotte uses the verb “connaître” for the same term.

131. mtshan ma dang ’brel pa’i ming la brten nas. VD, folio 56.a: ming dang ’brel ba’i mtshan ma la brten nas.

132. mngon par zhen pa; abhiniveśa. Usually with a negative connotation; see Edgerton 1953, p. 53. Powers reads it as meaning “strongly adhering”; see Powers 1995, p. 87.

133. See the definition of the imaginary defining characteristic in 6.4: “Guṇākara, what is the imaginary defining characteristic? It is what is nominally and conventionally posited as the essence or the distinctive [characteristic] of phenomena, even just in order to designate [them].” See also 7.4: “Para­mārtha­samud­gata, what is the essencelessness of all phenomena as a defining characteristic? It is the imaginary defining characteristic. Why? Because as much as this defining characteristic is nominally and conventionally posited, it is not posited on the basis of an essence or a distinctive [characteristic]. This is why it is called ‘the essencelessness of all phenomena [that is] a defining characteristic.’ ”

134. See 6.5: “Guṇākara, what is the other-dependent defining characteristic? It is the dependent arising of phenomena. It is like this: ‘When this is, that arises; because this arises, that arises;’ and also: from ‘[in dependence upon ignorance as a condition,] conditioning mental factors [arise]’ up to ‘thus, the whole great mass of suffering comes to be.’ ” And also 7.5: “Para­mārtha­samud­gata, what is the essencelessness of all phenomena [that is] arising? It is the other-dependent defining characteristic of phenomena. Why? Because this is [the defining characteristic] arising on account of causes other [than itself], and not by itself. This is why it is referred to as ‘essencelessness with regard to arising.’ ”

135. I read mtshan nyid med pa dang / kun nas nyon mongs pa’i mtshan nyid dang / rnam par byang ba’i mtshan nyid (D) as in the previous clauses, namely, as bahuvrīhis.

136. Lamotte 1935: 191: “Ils excitent la pitié.” A la rigueur, one could have accepted “ils suscitent la pitié”!

137. This enumeration follows the structure found in 4.2.

138. D: thams cad ngo bo nyid ma mchis pa for chos thams cad ngo bo nyid ma mchis pa (D, folio 16.b passim).

139. For a list of texts including this sentence, see Lamotte 1935, p. 198.

140. rnam par gnas pa; vyavasthita (Chinese: 安立).

141. See Lamotte’s and Frauwallner’s translations of this passage (Lamotte 1935, p. 194 and Frauwallner 1969, p. 291). Both read rnam par gnas pa (“établi,” “beruht”) as the main verb in both clauses, which is syntactically dubious. Xuanzang’s translation concords with D: 善男子云何諸法無自性性謂諸法遍計所執相。何以故。此由假名安立為相非由自性安立為相 (Cbeta, Taishō 676). The complete definition of pari­kalpita­lakṣaṇa in 6.4: D should be kept in mind when translating the definition of the lakṣaṇaniḥsva­bhāvatā: yon tan ’byung gnas de la chos rnams kyi kun brtags pa’i mtshan nyid gang zhe na/ ji tsam du rjes su tha snyad gdags pa’i phyir chos rnams kyi ngo bo nyid dam bye brag tu ming dang brdar rnam par gzhag pa gang yin pa’o.

142. rab tu phye ba; prabhāvita (see Schmithausen 2014, p. 400, n. 1770).

143. See Schmithausen 2014, p. 559. Schmithausen reads paramārthaniḥsva­bhāvatā as “lack of own-being [that is] the ultimate reality.”

144. For a discussion of the syntactic construction gcig … gcig, refer to Tillemans 1997, pp. 161–64.

145. See Schmithausen 2014, p. 560 on the textual material pertaining to this sentence.

146. See Schmithausen 2014, p. 561ff.

147. For a complete comparison of this passage across editions, see Kojirō Katō’s forthcoming edition of the text. As an example, it is interesting to compare the syntax of D and F. D: don dam yang dag ’phags pas ni sems can gyi khams na sems can rnams kyis kun brtags pa’i ngo bo nyid ngo bo nyid kyis tha dad par mthong zhing gzhan gyi dbang gi ngo bo nyid dang / yongs su grub pa’i ngo bo nyid kyang ngo bo nyid kyis tha dad par mthong na/ de’i phyir ngo bo nyid med pa nyid rnam pa gsum mi ’dogs kyi. Interestingly, L, S, and T are in agreement with D, as are C, J, N. P. VD, VG, and VP. Only F offers a variant reading (folio 25.b): don dam yang dag’phags sems can gyi khams ni/ sems can rnams kyis kun brtags pa’i rang bzhin ngo bo nyid kyis tha dad par mi mthong / gzhan gyi dbang gi rang bzhin dang / yongs su grub pa’i rang bzhin yang ngo bo nyid kyis tha dad par yang mi mthong ste/ de’i phyir ngas rang bzhin med pa rnams gsum du gzhag go. F explains why the Buddha taught an essencelessness by referring to beings as not perceiving a distinct essence in the three natures: “Para­mārtha­samud­gata, beings in the world of beings do not consider the imaginary essence as different from an essence. They do not even consider the other-dependent essence and the actual essence as different from an essence. As a consequence, I presented the threefold essencelessness.” This does not make much sense. I therefore follow the reading found in D. Frauwallner chose to follow D very closely here. He takes as the subject of the verb mi ’dogs the Buddha, like Powers and Lamotte: “Ich habe … die dreifache Wesenlosigkeit nicht verkündet, weil die Lebewesen in der Sphäre der Lebewesen das vorgestellte Wesen seinem Wesen nach als etwas Verschiedenes ansehen, und weil sie das abhängige Wesen und das vollkommene Wesen seinem Wesen nach als etwas Verschiedenes ansehen.” Lamotte’s translation (Lamotte 1935, p. 196) reads, “Si j’expose [for ’dogs] la triple Irréalité, ce n’est pas parce que les êtres, dans le monde des êtres, considèrent la nature imaginaire comme une nature distincte, ou les natures dépendante et absolue comme des natures distinctes. Au contraire…” With “au contraire,” Lamotte expresses the adversative function of the particle kyi at the end of the clause de’i phyir ngo bo nyid med pa nyid rnam pa gsum mi ’dogs kyi. To achieve this, Lamotte reads mi ’dogs in an affirmative mode and in the negative the statement regarding beings, which is in agreement with Xuanzang’s translation: 復次勝義生非由有情界中諸有情類別觀遍計所執自性為自性故。亦非由彼別觀依他起自性及圓成實自性為自性故我立三種無自性性。然由有情於依他起自性及圓成實自性上增益遍計所執自性故我立三種無自性性 (Cbeta, Taishō 676).

148. sems; citta. This is one of the synonyms for the subliminal mind (kun gzhi rnam par shes pa; ālayavijñāna) as explained in chapter 5. See chapters 5 and 6 on the latent disposition of the mind through karmic seeds in the sense of conventions.

149. bag la nyal; anuśaya. In the sense of “adhering/sticking” and “being latent/inclined” here (see Schmithausen 2014, p. 687).

150. Powers’ translation does not render the meaning of this sentence: “Due to these causes and conditions, in the future [this view of] the own-being of the other-dependent proliferates.” (Powers 1995, p. 107). The other-dependent in the sense of dependent arising refers here to rebirth and future lives.

151. yang dag par (“perfectly”) is important here, as it echoes the statement above and without it the entire paragraph loses its meaning: to achieve perfect liberation, the teaching on essencelessness with regard to both defining characteristics and the ultimate is necessary.

152. L, S, T, and F (e.g., F, folio 27.a) logically confirm shes pa, which is interesting since it establishes a distinction between occurrences of shes pa and ye shes, which D does not do systematically (see Kojirō Katō’s edition of chapter 7). The term here is a synonym for sems; see the parallel passage above in 7.10.

153. F, folio 27.a, has lam ’di nyid dang ’grod pa ’di nyid kyis instead of D: lam ’di nyid dang sgrub pa ’di nyid kyis grub pa dang. C, H, J, N, and Kǫ also read sgrub; VD, VG, VP: bsgrub; L, S: bgrod; F, T: ’grod. (cf. Kojirō Katō’s edition).

154. don dam yang dag ’phags nyan thos kyi rigs can gang zag zhi ba’i bgrod pa gcig pu pa ni sangs rgyas thams cad brtson pa dang ldan par gyur kyang byang chub kyi snying po la bzhag ste. Brunnhölzl reads the qualifying clause sangs rgyas thams cad brtson pa dang ldan par gyur in quite a different way here: “even if all buddhas with [all] their effort] were [to attempt] to establish persons with the śrāvaka disposition…” (Brunnhölzl 2018, p. 1522).

155. ’du byed mngon par ’du bya ba; saṃskārābhi­saṃskaraṇa. Compare with the Sanskrit sentence in Tucci’s edition of Bhāvanākrama (Tucci 1971, p. 22): ekānta­sattvārtha­vimukhasya ekānta­saṃsārābhi­saṃskāra­vimukhasya [nā]uttarā samyaksaṃbodhir uktā mayeti, which has saṃ­sārābhi­saṃskāra instead of saṃskārābhi­saṃskaraṇa as found in Tibetan.

156. chos; dharma.

157. drang po dang drang po’i rang bzhin can/ rtog pa dang sel mi nus pa. Powers understands rtog pa dang sel as “to remove conceptuality.” (Powers 1995, p. 117). F, folio 28.b reads, brtag pa dang / bzhig pa mi nus pa. The problem is that sel is a transitive verb. It is therefore syntactically difficult to take rtog pa as the object of sel. In the present case, Keenan’s solution based on Chinese is interesting: “to make judgments” (Keenan 2000, p. 42).

158. zhib mo brtags pa’i mkhas pa dang ’dzangs pas rig pa; sūkṣmaṃ nipuṇapaṇḍitavijñavedanīyaḥ (see Mvyut 2918). Lit. “It is to be known.” All Sanskrit synonyms for this sentence are found in Mvyut 2013–20.

159. In accordance with the multiple occurrences of this phrase in chapter 2, shes pa should be read here as ye shes in agreement with F.

160. See Lamotte 1935, p. 201, n. 31: las kyi sgrib chen po refers here to saddharmapratikṣepakarmāvaraṇa.

161. chos ’dul ba; dharmavinaya (read as a dvandva).

162. I read rnam par rtog pa’i spyod yul kun brtags pa’i mtshan nyid kyi gnas ’du byed kyi mtshan ma (D) as a karmadhāraya, which means that the last compound in the series of three should be topicalized. It seems to me that since the opposition between nimitta (“phenomenal appearance”) and svabhāva or svalakṣaṇa (“unique/specific/particular defining characteristic or essence”) is central throughout the text, reading the compound in this way clarifies the meaning of this definition of the parikalpita, which basically results from the operation consisting in attributing an essence to appearance by means of verbal conventions.

163. This definition elaborates on the definition of parikalpitasvabhāva formulated in 6.4. In 6.5 and 6.7; that which has the defining characteristic of dependent arising is equated with phenomenal appearance. On the basis of what is dependent on an other, essence is imputed in the sense of a real entity, independent of any other cause to exist as what it is. This corresponds to the imaginary defining characteristic.

164. Lamotte’s translation is built on the same structure but inverts the main clauses of the sentence: In the nimitta, the parikalpita is unestablished. See Lamotte 1935, p. 204.

165. Lamotte’s translation is built on the same structure but inverts the main clauses of the sentence: In the nimitta, the parikalpita is unestablished. See Lamotte 1935, p. 204. D: kun brtags pa’i mtshan nyid der yongs su ma grub. I read kun brtags pa’i mtshan nyid as a bahuvrīhi: “that which has the imaginary defining characteristic” or “that which consists in/is characterized by the imaginary.”

166. yang dag pa’i ting nge ’dzin; samyaksamādhi.

167. D: bcom ldan ’das nges pa’i don bstan pa ’di should be read bcom ldan ’das kyi (or kyis as in VD) nges pa’i don bstan pa ’di, lit. “of the Blessed One,” rendered here as “[expounded] by the Blessed One.” bcom ldan ’das is omitted in L, S, T, and F; see Katō’s edition of chapter 7.

168. Lamotte and Powers understand the statement “All phenomena are without an essence, unborn, unceasing, primordially in the state of peace, and naturally in the state of nirvāṇa” to be the teaching of definitive meaning (see Lamotte 1935, p. 206 and Powers 1995, pp. 135–37). However, the entire point of this chapter is that there is an underlying intent of definitive meaning to this statement. This is the reason why it is explained in the next paragraphs that a third turning of the wheel of Dharma was necessary.

169. D: de bzhin du chos rnams kyi ngo bo nyid ma mchis pa nyid las brtsams/ skye ba ma mchis pa dang / ’gag pa ma mchis pa dang / gzod ma nas zhi ba dang / rang bzhin gyis yongs su mya ngan las ’das pa nyid las brtsams nas/ bcom ldan’das nges pa’i don bstan pa’di yang drang ba’i don gyi mdo sde thams cad du stsal bar bgyi pa lags so. C, H, J, N, Kǫ, VG, VP, L, S, T, and F also read stsal; VD: bstsal (cf. Kojirō Katō’s edition). Lamotte translates stsal with “se recommande” to create a parallel construction with the analogy of the dried ginger (see Lamotte 1935, p. 205). This somehow does not solve our problem. Powers translates stsal bar gyi with a past tense “placed” (see Powers 1995, p. 137). Keenan 2000, p. 48 offers a Literal rendering of Xuanzang’s translation that is similar to the Tibetan version of the text (如是世尊依此諸法皆無自性無生無滅本來寂靜自性涅槃無自性性了義言教遍於一切不了義經皆應安處, Cbeta, Taishō 676): “just so, World-Honored One, the explicit teaching that all things have no-essence, no arising, and no passing away, are originally quiescent, and are essentially in cessation must be put into all the scriptures of implicit meaning.”

170. D: drang ba’i don rtsod pa’i gzhi’i gnas su gyur pa lags la in the sense of drang ba’i don lags te/ rtsod pa’i gzhi’i gnas su gyur pa lags la as in nges pa’i don lags te/ rtsod pa’i gzhi’i gnas su gyur pa ma lags so (see D, folio 25.a).

171. This is an important statement regarding the intent of the third turning of the wheel, which is to bring together those following the hearers’ and the bodhisattvas’ paths within a single vehicle.

172. By repeating the same statement to describe the second and third turnings, it is made clear that interpreting this statement in terms of emptiness alone is provisional. The underlying intent of the statement corresponds to the teaching found in the third turning.

173. Lit. “Blessed One, what is the name of that which has been taught as a Dharma discourse ascertaining the [Tathāgata’s] intent?”

174. gzung bar bgyi; dhārayāmi (cf. Sanskrit text in Matsuda 2013, p. 940 ad Lamotte VIII.41). I suggest reading dhārayāmi, which is a causative present of dhṛ-, as an optative here.

175. This refers to byang chub kyi sems; bodhicitta.

176. This paragraph follows the pattern according to which prajñā is developed as śrutamayī prajñā, cintāmayī prajñā, bhāvanāmayī prajñā. Once the bodhisattvas have heard and contemplated what has been taught, they proceed with practice.

177. Read thob instead of thos. See F, folio 36.b: lus shin tu sbyangs pa dang / sems shin tu sbyangs pa de dag thob pa de’i bar du and lus shin tu sbyangs pa dang / sems shin tu sbyangs pa de dag thob kyi bar du (8.5).

178. D: de lus shin tu sbyangs pa dang / sems shin tu sbyangs pa de thos nas de nyid la gnas te/ sems kyi rnam pa spangs nas ji ltar bsams pa’i chos de dag nyid nang du ting nge ’dzin gyi spyod yul gzugs brnyan du so sor rtog par byed mos par byed do. S, folio 37.b; F, folio 36.a: des lus shin tu sbyangs pa dang / sems shin tu sbyangs pa de dag thos nas de nyid la gnas te/ ji ltar bsams pa’i chos de dag nyid nang du ting nge ’dzin gyi spyod yul gzugs brnyan du sems kyi rnam pa spangs nas/ so sor rtog cing byed mor byed do.

179. For a detailed discussion of 8.7–9, refer to Schmithausen 2014, p. 391ff.

180. rab tu phye ba; prabhāvita in the compound *ālambanavijñaptimātraprabhāvita (see Schmithausen 1984, p. 436; 2014, pp. 400–1 passim). On the possible meanings of the Sanskrit term prabhāvita in the sense of “consisting of” or “characterized as” as opposed to “characterized by,” refer to Schmithausen 2014, p. 400, n. 1770.

181. The compound dmigs pa rnam par rig pa tsam; *ālambanavijñaptimātra can be read as a tatpuruṣa: “the mere representation of a referential object,” or as a karmadhāraya: “a referential object that is a mere representation” or “a mere representation as a referential object” (see Schmithausen 2014, p. 411). If we add to these possibilities the distinction between “characterized as/consists of” and “characterized by” mentioned by Schmithausen in reference to prabhāvita (see n.181), it is clear that this important statement can be interpreted in various ways that are consistent with the syntax of this sentence. If one wishes to read the compound *ālambanavijñaptimātra as a tatpuruṣa, the phrase reads “the mere representation of a referential object.” However, it seems to me that we should read this compound as a karmadhāraya, since the whole point of this paragraph is to answer Maitreya’s original question about the nature of the image qua object. The answer to this question is that the image (pratibimba) that is the object of concentration is not distinct from mind because mind is constituted by a representation that is the actual object of this cognition. With regard to this issue, Xuanzang’s translation reads, 善男子當言無異何以故由彼影像唯是識故善男子我說識所緣唯識所現故 (Cbeta, Taishō 676), which is compatible with the suggested translation. For a complete analysis of these two sentences across selected Chinese and Tibetan editions, see Schmithausen 2014, p. 392ff. and Brunnhölzl 2018, p. 511, n. 139, which contains a detailed summary of Schmithausen’s analysis.

182. On VIII.7, see Brunnhölzl 2018, p. 512, n. 141.

183. D: bcom ldan ’das sems can rnams kyi gzugs la sogs par snang ba sems kyi gzugs brnyan rang bzhin du gnas pa gang lags pa de yang sems de dang tha dad pa ma lags zhes bgyi’am; S, folio 38.b: bcom ldan ’das sems can rnams kyi sems kyi gzugs brnyan rang bzhin du gnas pa/ gzugs la sogs pa gang lags de yang sems de dang tha dad pa ma lags shes bgyi’am; F, folio 37.a: bcom ldan ’das sems can rnams kyi sems kyi gzugs brnyan rang bzhin du gnas pa/ gzugs la tshogs pa de dag kyang sems de dang tha dad pa ma lags shes bgyi’am.

184. D: mtshan nyid la byed pa. F, folio 37.b: mtshan nyid yid la byed pa.

185. sems kyi mtshan nyid but sems kyi mtshan ma would be better here since Maitreya inquires about mtshan ma right after this. Unfortunately, Xuanzang’s translation does not contribute to solving this quandary since 相 can refer to both lakṣaṇa and nimitta (see 若相續作意唯思维心相, Cbeta, Taishō 676). However, the structure of the paragraph in which questions are asked about definitions of terms found in the Buddha’s previous answer in 8.8 indicates that we should emend sems kyi mtshan nyid to sems kyi mtshan ma.

186. Based on the definitions above, it appears that one-pointedness of mind refers to the state in which appearance (the object of the practice of insight), representation (the object of the practice of mental stillness), and emptiness (the nature of reality as explained in the previous chapters) are in unity.

187. so sor rtog pa in the sense of “comprehension/realization” (pratibodha)

188. Compare D: de dang der shes rab kyis shin tu legs par ma rtogs pa’i chos de dag nyid shin tu legs par rtogs par [F.28.a] bya ba’i phyir yid la byed pa’i lhag mthong gang yin pa’o with S, folio 39.b: gang de dang des shes rab kyis legs par ma rtogs pa’i chos de dag nyid legs par rtogs par bya ba’i phyir de nyid yid la byed pa’i lhag mthong ngo and F, folio 38.a: gang shes rab kyis legs par ma rtogs pa’i chos de dang de dag la legs par rtogs par bya ba’i phyir de nyid yid la byed pa’i lhag mthong ngo.

189. The last four meditation objects represent the four seals of Dharma (phyag rgya bzhi; caturmudrā).

190. ma ’dres pa’i chos. On ma ’dres pa’i chos and ’dres pa’i chos, see Brunnhölzl 2018, p. 561, n. 322. Brunnhölzl translates ’dres pa’i chos as “dharmas in fusion.” Considering dharmas without fusing them means considering them individually. However, 8.14 seems to indicate that dharma is used here in the sense of “teaching” rather than “phenomenon.”