Notes

n.1dharmāṇāṁ pravicayam antareṇa nāsti kleśānāṁ yata upaśāntaye ’bhyupāyaḥ | kleśaiś ca bhramati bhavārṇave ‘tra lokas taddhetor ata uditaḥ kilaiṣa śāstra (Abhidharma­kośa­kārikā 1.3).

n.2tasyārtha­viniścaya­sūtrasya dharma­pravicayārtham upadeśaḥ | na hi vinā sūtropadeśena śiṣyaḥ śakto dharmaṁ pravicetum | dharma­pravicaya upakleśopaśamanārtham | tadupaśamo ‘pi naiṣṭhika­padāvaptaye bhavati (Artha­viniścaya­sūtra­nibandhana, Samtani 1971, p. 72). We read dharma­pravicaya upakleśopaśamanārtham rather than dharma­pravicayopakleśopaśamanārtham as per Samtani’s printed text.

n.3Skilling 2012.

n.4Ferrari 1944, p. 617; translation ours.

n.5Ferrari 1944, p. 522.

n.6Ferrari 1944, p. 549; translation ours.

n.7Vaidya 1961, p. 319.

n.8asya sarvaṁ vṛttāntaṁ abhiniṣkramaṇa­nāma­sūtre vidyate iti tibbatabhāṣayā kathitam, Vaidya 1961, p. 319. The term employed for “Tibetan” is tibbata.

n.9See Samtani 1971, pp. 8–9.

n.10De Jong 1975, p. 116.

n.11Ferrari 1944, p. 551; translation ours.

n.12“Succoso,” Ferrari 1944, p. 552; translation ours.

n.13The Tibetan opts for one of two possible ways of parsing the syntax of the first paragraph, rather than retaining the ambiguity of the Sanskrit; the original may be understood as taking “at one time” (ekasmin samaye) either with what precedes or with what follows, as commentators including Vīryaśrīdatta point out (see Samtani 1971, pp. 75–76). As Vīryaśrīdatta also points out, saying “this is what I heard at one time” implies that one has heard other teachings at other times, thus indicating the arhat Ānanda’s quality of having heard a lot (bāhuśrutya), i.e., being learned in the Dharma.

n.14The Nibandhana (Samtani 1971, p. 78) explains that it is “eastern” either because it is in the eastern part of Śrāvastī or because it is to the east of the Jeta Grove, the location of the very famous pleasance offered by the foremost of male lay practitioners, Anāthapiṇḍada.

n.15The Nibandhana (Samtani 1971, pp. 80–81) explains that this expression means two things: it refers to the three trainings, i.e., discipline, mental concentration, and wisdom (adhiśīlaśikṣā, adhicittaśikṣā, and adhiprajñā­śikṣā); and it also indicates that the beginning, middle, and end of the Dharma are not mutually contradictory.

n.16It is worth noting that Pāli Suttas read this as “with meaning” and “with expression” (sātthaṁ sabyañjanam); Vasubandhu is also aware that “some” (kha cig) read something along those lines, although it is difficult to say whether he was referring specifically to the Pāli texts or to parallel Sanskrit transmissions (*sārthaṁ savyañjanam): kha cig ni don dang ldan pa dang / tshig ’bru dang ldan pa shes ’don te (Lee 2001, p. 5; see also Nance 2012, p. 131 for a translation).

n.17The Nibandhana explains that this sentence is meant to ensure that the listeners avoid three defects: not listening, listening but understanding the meaning of what was heard in a distorted manner, and not being able to retain even what has been heard and understood correctly, since it was not listened to with the necessary reverence. These three defects are exemplified by a vase upside down, a dirty vase, and vase with holes: such a vase will not be able to profit from the rain of Dharma (Samtani 1971, p. 83). This example is also found in Vasubandhu’s Vyākhyāyukti, in the Pratītya­samutpādādi­vibhaṅga­nirdeśa, in the Gāthārtha­saṁgraha­śāstra, and in Haribhadra’s Abhisamayālaṁkārālokā, where he attributes it to Vasubandhu (see Skilling 2000, p. 301).

n.18In addition, the four ārūpya­samāpattayaḥ (“formless attainments”) are found right after this item in the Sanskrit (Samtani 1971, p. 2; p. 18).

n.19Saṁjñā is not easy to translate: it refers to a concept or idea through which one identifies, correctly or incorrectly, the object of cognition by determining that it has a certain “sign” (nimitta, a term closely connected to perceptual experiences). Its function is associated with naming, since saṁjñā can be verbalized by those who are acquainted with language, as pointed out in Abhidharma texts. It can also refer to an evaluative belief about something, such as it being desirable, etc., or a notion that one should develop through training (such as the idea that the Dharma teacher is the Buddha). In some forms of Abhidharma, it is explained that the consciousnesses based on the five senses have weak saṁjñā, while it is sharp in case of thought-consciousness. Saṁjñā complements the perception of objects (vijñānaskandha) with enough determination to become fit to be put into words, expressing specific referents and their features. It is regarded as a mental state (caitta), accompanying the basic mind (citta) that is defined as the mere perception of objects, incapable of grasping their specificities.

n.20Here there is an additional sentence in Samtani’s edition of the Sanskrit that reads yaduta asmin satīdaṁ bhavati asyotpādād idam utpadyate (Samtani 1971, p. 5). This additional sentence in Samtani’s edition of the Sanskrit could be translated as “It is thus: this being there, this comes about; due to the arising of this, this arises.”

n.21The Nibandhana explains kevalasya as implying that the aggregate of suffering has no self, i.e., it is “nothing more than” an aggregate of suffering; Pāli commentaries often prefer another possible sense of kevala, “entire,” and previous translators have rendered parallel passages accordingly. However, we do find Pāli commentators including the Nibandhana understanding of kevala (suddhassa vā, sattavirahitassāti attho). We could not find a single English word carrying both the sense of “which is no more than” and “in its entirety,” hence we have used a longer expression to translate the single term kevala.

n.22The Sanskrit in Samtani’s edition (Samtani 1971, p. 6) has “the present” (pratyutpanne) instead of “the prior and the following limit” (pūrvāparānte); this agrees with the Chinese translations and, importantly, with some of the Sanskrit manuscripts (F and ANe in Samtani’s edition). “Prior limit” and “following limit” are technical terms referring generally to the past and future, but more specifically to past and future lifetimes, and are often used while describing the twelve parts of dependent arising and its subdivision into three lifetimes and when describing wrong views about identity or difference between lifetimes. Thus, the referent of “prior limit” and “following limit” is occasionally the “past” and “future,” but the meaning of the terms is different (they indicate what marks the border between the present life and other lifetimes, hence the use of -anta, here translated as “limit”).

n.23Samtani’s edition does not contain the word karma (karmaṇi), but he reports that karmaṇi is found in F (Samtani 1971, p. 6, n. 3). The third type of karma, according to the Nibandhana commentary, refers to a mixture of the first two.

n.24Before this phrase, the Sanskrit in Samtani’s edition also has “nonawareness in respect to instances of dependent arising” (pratītya­samutpādeṣv ajñānam, Samtani 1971, p. 6), and Samtani reports ANe as having the alternative “nonawareness in respect to dependent arising” (pratītya­samutpāde ’jñānam, Samtani 1971, p. 6, n. 5).

n.25Here there is an additional portion in Sanskrit in Samtani’s edition that reads sāvadyāna­vadyeṣu dharmeṣu ajñānam sevitavyāsevitavyeṣ dharmeṣu ajñānam hīna­praṇīteṣu kṛṣṇa­śukleṣu dharmeṣu ajñānam (Samtani 1971, p. 6). This could be translated as “nonawareness in respect to blameworthy and blameless dharmas, nonawareness in respect to dharmas that should and should not be practiced, nonawareness in respect to low and foremost, or black and white dharmas.”

n.26Samtani did not consider yang dag pa ji lta ba bzhin as one of the possible renderings of yathābhūta, rather understanding it as matching a hypothetical *samyak yathāvat (Samtani 1971, p. 7, n. 1). However, parallels suggest that here the Tibetan matches the Sanskrit.

n.27Here the Degé version has lus ’di ni lus las byung ba; Ferrari usefully pointed out that here lus ’di ni is probably a printing mistake for chos ’ di ni (Ferrari 1944, p. 557); we agree that here we should read the Tibetan as chos ’ di ni lus las byung ba, matching the Sanskrit kāyiko hy eṣa dharmaḥ, since the transmitted reading would make little sense.

n.28Here the Sanskrit reads ime bhikṣavaḥ trayaḥ saṁskārā ucyante (Samtani 1971, p. 8). This could be translated as “the three assembled factors.”

n.29Here the Sanskrit reads ime ṣaḍ vijñāna­kāyā vijñānam ity ucyante (Samtani 1971, p. 8). This could be translated as “These six collections of consciousness are explained as consciousness .”

n.30Here the Sanskrit reads tatra kataman nāma catvāro’rūpiṇaḥ skandḥāḥ | katame catvāraḥ (Samtani 1971, p. 8). This could be translated as “What, then, is name? It is the four formless aggregates. What are the four?”

n.31Here the Tibetan lacks the expression “and in what depends on the four great elements”; we have nevertheless included it on the basis of Samtani’s Sanskrit (catvāri ca mahābhūtāny upādāya, Samtani 1971, p. 9), as it represents a standard definition of “form,” and thus the omission seems odd. We have also included the rhetorical question that immediately follows, which is absent in the Tibetan.

n.32Samtani’s Sanskrit edition (Samtani 1971, p. 9) does not include “roughness,” nor does the Nibandhana seem to read it. The Tibetan suggests the following hypothetical Sanskrit: *gurutvaṁ khakkhaṭatvaṁ karkaśatvaṁ ca.

n.33Here the Sanskrit reads ṣaḍ ādhyātmikāny āyatanāni | tadyathā (Samtani 1971, p. 9). This could be translated as “the six internal entrances; they are…”

n.34Here the Sanskrit reads cakṣuḥ­saṁsparśaḥ śrotra­saṁsparśaḥ ghrāṇa­śaṁsparśaḥ jihvā­saṁsparśaḥ kāya­saṁsparśaḥ manaḥ­saṁsparśa iti | ayam ucyate sparśaḥ (Samtani 1971, p. 10). This could be translated as follows: “Eye contact, ear contact, nose contact, tongue contact, body contact, thought contact. This is explained as contact.”

n.35Here the Sanskrit reads idam ucyate vedanā (Samtani 1971, p. 10): “This is explained as feeling” (additional sentence not found in the Tibetan translation).

n.36Here the Sanskrit reads tadyathā uṣṇa­narakā aṣṭau | katame ’ṣṭau tadyathā saṁjīvaḥ kāla­sūtraḥ saṅghātaḥ rauravaḥ mahā­rauravaḥ tapanaḥ pratāpanaḥ avīciś ca | śīta­narakā aṣṭau | [katame’ṣṭau] tadyathā arbudaḥ nirarbudaḥ aṭaṭaḥ hahavaḥ huhuvaḥ utpalaḥ padmaḥ mahā­padmaḥ | pretāḥ tiryañcaḥ manuṣyāḥ ṣaṭ kāmāvacarāś ca devāḥ | katame ṣaṭ cātur­mahā­rājikāḥ trāyastriṁśāḥ yāmāḥ tuṣitāḥ nirmāṇa­ratayaḥ paranirmita­vaśavartino devāḥ (Samtani 1971, p. 11). The Sanskrit text could be translated as “It is the eight hot hells. What are the eight? They are Reviving, Black Thread, Crushing, Howling, Great Howling, Burning, Intense Burning, and Unwavering. It is the eight cold hells. What are the eight? They are Swelling, Thorough Swelling, Aṭṭa, Hahava, Huhuva, Blue Lotus, Lotus, and Great Lotus. It is the pretas, animals, humans, and the six deities within the sphere of desire. What are the six? They are the Four Great Kings, the deities of the Thirty-Three, Yāma, Tuṣita, Emanation-Delight, and Control of Others’ Emanations.” Samtani also points out that the order differs in Fe and ANe (Samtani 1971, p. 11, n. 5). See also the Nibandhana (Samtani 1971, pp. 140–41) for useful explanations of the names of these different classes of deities.

n.37Here the Sanskrit reads tatra rūpa­bhavaḥ katamaḥ tadyathā brahma­kāyikāḥ brahma­purohitāḥ mahā­brāhmaṇaḥ parīttābhāḥ apramāṇābhāḥ ābhāsvarāḥ parītta­śubhāḥ śubha­kṛtsnāḥ anabhrakāḥ puṇya­prasavāḥ bṛhat­phalāḥ avṛhāḥ atapāḥ sudṛśāḥ sudarśanāḥ akaniṣṭhāś ceti (Samtani 1971, p. 12). This could be translated as “What, then, is existence with form? It is those in the Brahmā group, the Provosts of Brahmā, the Great Brahmās, Limitedly Splendid, Splendid without Measure, Wholly Good, Unclouded, Merit Increasing, Abundant Result, Not Great, Without Pain, Seeing Well, Good Sight, and Lesser than None.”

n.38Here the Sanskrit reads ime trayo bhavāḥ (Samtani 1971, p. 12): “These are the three existences” (additional sentence found in the Sanskrit but not in F and ANe or the Tibetan; Samtani 1971, p. 12, nn. 9–10). See also the Nibandhana (Samtani 1971, pp. 141–44) for useful explanations of the names of these classes of deities.

n.39Here the Sanskrit reads bhava­pratyayā jātiḥ | jātiḥ katamā yā teṣāṁ teṣāṁ sattvānāṁ tasmiṁs tasmin sattva­nikāye jātiḥ saṁjātiḥ upapattiḥ avakrāntiḥ abhi­nirvṛttiḥ prādur­bhāvaḥ skandha­pratilambhaḥ dhātu­pratilambhaḥ āyatanānāṁ pratilambhaḥ skandhānām abhinirvṛttiḥ jīvitendriyasyodbhavaḥ nikāya­sabhā­gatāyāḥ samavadhānam (Samtani 1971, pp. 12–13). This could be translated as “ ‘As for birth with existence as its condition,’ what is birth? It is the birth of such and such sentient beings in such and such specific classes of sentient beings; it is their thorough birth, descent, coming forth, and manifestation, the obtainment of the aggregates, the obtainment of the bases, the obtainment of the entrances, the proceeding of the aggregates, the coming into being of the life faculty, and the fact of being brought together within the commonality of a specific class. This is explained as birth.”

n.40Here the Sanskrit reads khuru­khuru­niśvāsa­praśvāsa­kaṇṭhatā (Samtani 1971, p. 13): “having a throat that sounds like khuru­khuru when exhaling and inhaling.” The Sanskrit khuru­khuru and the Tibetan ngar ngar resemble the English term “wheezing,” whose etymology is also onomatopoeic.

n.41Samtani points out that ANe and F lack the ca after jarjarībhāvaḥ, and we think this matches the Tibetan and seems altogether preferable.

n.42The commentary explains this as “the separation of the collection of name from the collection of form” (arūpiṇo hi nāma­kāyasya rūpa­kāyād viśleṣo bhedaḥ; Samtani 1971, p. 154).

n.43Here the Sanskrit reads antarhāṇiḥ (Samtani 1971, p. 14), meaning something like “disappearance.” The Tibetan mi rtag par ’gyur ba could correspond to anityabhāva. Samtani (1971, p. 14, n. 4) proposes anityatābhāvaḥ, but we do not think the Tibetan is intended to represent the tal-pratyaya.

n.44The Nibandhana commentary explains this as the loss of the projecting force of karma, which “throws” the assembled factors (like an arrow), at birth, only for a determinate length of time (pūrvopāttāyuḥ­saṁskārāṇām āvedhakṣayaḥ, Samtani 1971, pp. 155–56).

n.45The Tibetan dga’ ba’i ’dod chags may suggest rather a genitive case relationship between dga’ ba and ’dod chags, but this seems highly unlikely and is not supported by the Nibandhana; in general, we would read the usage of particles in Tibetan texts translated from the Sanskrit with some degree of flexibility, and not necessarily in their most idiomatic sense.

n.46The Tibetan here has a plural marker, rnams, but it seems out of place and is not supported by any parallels we could find.

n.47Samtani (1971, p. 17) does not have anything that matches dga’ ba med pa (“without joy”). He proposes that it could translate prītirahitam (Samtani 1971, p. 17, n. 8). We think, on the other hand, that the probable Sanskrit original should be niṣprītikam, which is attested in a sūtra quotation, very close to our passage (indeed we wonder whether this is a quote of Distinctly Ascertaining the Meanings ), that appears in Ratnākaraśānti’s Sāratamā : yat tad āryā ācakṣate upekṣakaḥ smṛtimān sukha­vihārīti niṣprītikaṁ tṛtīyaṁ dhyānam upasampadya viharatīti (Jaini 1979, p. 50). In the same context of the third meditation (dhyāna), the term niṣprītikam also appears in the Saṅgha­bheda­vastu (Gnoli 1978a, p. 144).

n.48Samtani’s Sanskrit edition here includes a section on the four formless attainments that is absent in the Tibetan (Samtani 1971, p. 18).

n.49Tib. ’dod chags does sometimes translate rāga, instead of kāmarāga as we have in the Sanskrit edition. However, ’dod chags is elsewhere attested as a translation of kāmarāga, not only of rāga, and we have translated accordingly.

n.50Here the Sanskrit reads yathāvasthitam yathāpraṇihitam (Samtani 1971, p. 23): “as it is disposed and placed/set/directed.”

n.51Here the Sanskrit reads iha bhikṣavo bhikṣur araṇyagato vā vṛkṣa­mūla­gato vā śūnyāgāra­gato vā imam eva kāyam ūrdvaṁ yāvat pādatalād adhaḥ keśamastakāt tvakparyantaṁ yathāvasthitaṁ yathāpraṇihitaṁ pūrṇaṁ nānāprakārasyāśucer yathābhūtaṁ samyak prajñayā pratyavekṣate (Samtani 1971, p. 23): “Here, monks, a monk, having gone to the forest, or to the root of some tree, or to an empty house, observes with right wisdom this very body, up from the soles of the feet and down from the hairy head, enclosed in skin, just as it is disposed and placed, full of many types of impurities, as it is.” The Tibetan has some small differences; it could be back-translated as *tatra katamā samādhibhāvanā āsevitā bhāvitā bahulīkṛtā kāmarāgaprahāṇāya saṃvartate iha bhikṣavo bhikṣur imam eva kāyam ūrdhvaṃ pādatalād adhaḥ keśamastakāt paryantaṃ pūrṇaṃ nānāprakārasyāśucer yathābhūtaṃ pratyavekṣate.

n.52Samtani remarks, convincingly, that kha spu should be in fact read as ba spu (Samtani 1971, p. 24, n. 10).

n.53These two terms refer to Āyurvedic physiological categories.

n.54Here the Sanskrit reads santi asmin kāye keśā romāṇi nakhā dantā rajo malaṁ tvak māṁsam asthi snāyu śirā vṛkkā hṛdayaṁ plīhā klomakam antrāṇi antra­guṇā āmāśaya pakkāśaya udaryaṁ yakṛt purīṣam aśru svedaḥ kheḍaḥ siṁhāṇako vasā lasikā majjā medaḥ pittaṁ śleṣmā pūyaṁ śoṇitaṁ mastakaṁ mastaka­luṅgam iti pūrṇe nānāprakārasyāśucer yathābhūtaṁ pratyavekṣate (Samtani 1971, pp. 23–24): “In this body there are hair, body hairs, nails, teeth, dirt, filth, skin, flesh, bones, sinews, channels, kidneys, heart, spleen, lungs, entrails, mesentery, stomach for the raw, stomach for the ripe, bowel content, liver, excrement, tears, sweat, spit, snot, grease, fluid, marrow, fat, bile, phlegm, pus, blood, head, and head membrane.” We would here propose to emend āmāśaya to āmāśayaḥ.

n.55Samtani also counts thirteen types of grain in the Tibetan (Samtani 1971, p. 24, n. 5). We are especially uncertain about “corn,” which is a possible translation of nivāpa (assuming that nisvapa in the Tibetan is a misspelling of nivāpa). “Mat bean” is a possible translation of mukuṣṭa/makuṣṭa, assuming that the Tibetan mon sran na gu is a mistake for mon sran nag gu (the Stok Palace Kangyur edition has, in fact, nag gu). The list in Samtani 1971, p. 24 is as follows: dhānya­tila­sarṣapa­mudga­yava­māṣāṇām (“rice, sesame, mustard, lentils, barley, and beans”).

n.56Here the Sanskrit reads imāni śūka­dhānyāni imāni hala­dhānyāni (Samtani 1971, p. 25): “These are grains with awn, and these are grains to be ploughed.” The Tibetan repeats the whole list, but the Sanskrit only presents this abbreviated sentence.

n.57Here the Sanskrit reads evam eva bhikṣavo bhikṣur imam eva kāyaṁ yathāvasthitam yathāpraṇihitam yāvat pratyavekṣate (Samtani 1971, p. 25): “In the same way, bhikṣus, a bhikṣu thus observes this very body … up to … just as it is disposed and placed.” The Sanskrit contains the abbreviation yāvat, which is often difficult to interpret conclusively: it is quite likely that in some cases it is meant as just an abbreviation for saving space in manuscripts, so that the yāvat should then be understood as outside of the main text, not in the voice of the speaker. The Tibetan also has an abbreviation here, snga ma bzhin du (“just as before”), which we think would be translating *pūrvavat.

n.58Here the Sanskrit reads adhyātmaṁ vikeka­jena samādhi­jena (Samtani 1971, p. 25): “born from isolation and born from samādhi,” instead of Tibetan dben pa’i ting nge ’dzin las skyes pa’i.

n.59See n.­58.

n.60vivekajena (Samtani 1971, p. 26): “from isolation” is here added in the Sanskrit, right before “born from samādhi.”

n.61adhyātmaṁ viveka­jena (Samtani 1971, p. 26): “born from inward isolation.”

n.62The Tibetan here has shin tu legs par, while the Sanskrit has sādhu ca suṣṭhu ca (“well and properly”). According to Samtani, what has not been translated in the Tibetan is suṣṭhu; on the other hand, shin tu legs par is an attested translation of suṣṭhu, which suggests that perhaps what was not translated was sādhu. Since, however, sādhu would be legs par, we suggest that shin tu legs par may be meant as a translation of sādhu ca suṣṭhu ca.

n.63Samtani (1971, p. 26, n. 6) suggests that reg par should be corrected to rig par, which would give suviditā and somehow be related to sudṛṣṭā (“seen it well”); he also reports that the manuscript that he calls N3 has a correction to sujuṣṭā. Having found a number of parallels where supratividdha(ā) is preceded by some form of dṛś, Samtani’s suggestion of sudṛṣṭā sounds like a good suggestion. However, the Nibandhana’s gloss to sudṛṣṭā is bhāvanākāle viśeṣotpatti­yogāt sudṛṣṭā | susevitety arthaḥ, which suggests that adopting the reading of the manuscript that Samtani refers to as N, sujuṣṭā, would be more sensible (in fact, sudṛṣṭā there hardly makes any sense). We propose to read sujuṣṭā in the Sanskrit and consider reg par byas as a possible translation of juṣṭā.

n.64Reading divasasaṁjñādhiṣṭhitaṁ for divasasaṁjñādhiṣṭhitā, in analogy with the subsequent section. The commentary does not seem to read adhiṣṭhita.

n.65Here the Sanskrit adds sarvāntam imam lokam (Samtani 1971, p. 27): “throughout this world on all sides.” This additional phrase, however, as Samtani reports, is not in ANe (Samtani 1971, p. 27, nn. 5–6).

n.66See n.­63 on reg par.

n.67Here too reading divasa­saṁjñādhiṣṭhitaṁ for divasa­saṁjñādhiṣṭhitā, in analogy with the previous section; Samtani notices that “the reading is corrupt” in ANe (Samtani 1971, p. 27, n. 10).

n.68Tibetan here has dran pa, but we agree with Samtani (1971, p. 29, n. 5) that this seems most likely a mistake for ’dun pa, since this is a stock expression repeated in this very passage (and in many other texts).

n.69This sentence is lacking in the Tibetan, but given the overall structure of the sūtra, we have decided to include it on the basis of Samtani’s Sanskrit edition: tadyathā | smṛti­sambodhyaṅgaṁ dharma­pravicaya­saṃbodhyaṅgaṁ vīrya­saṁbodhyaṅgaṁ prīti­saṃbodhyaṅgaṁ praśrabdhi­saṃbodhyaṅgaṁ samādhi­saṃbodhyaṅgaṁ upekṣā­saṃbodhyaṅgam (Samtani 1971, p. 33).

n.70The Sanskrit lacks the first liberation. See Samtani 1971, p. 321, n. 3.

n.71The Tibetan for all this section corresponds to F and ANe, as reported in Samtani’s appendix (Samtani 1971, pp. 320–22).

n.72For the matching Sanskrit, see Samtani 1971, p. 321.

n.73Sanskrit has āryamārga (Samtani 1971, p. 322).

n.74For the matching Sanskrit see Samtani 1971, p. 322.

n.75ṛjukā, drang pa (rather than dran pa).

n.76The Tibetan suggests doṣa rather than dveṣa (Samtani 1971, p. 322). It also suggests samyagdarśikā.

n.77Here the Sanskrit has spharaṇa (“pervasive”) (Samtani 1971, p. 322; see also n. 9, reporting the alternative orthography sphuraṇa in the manuscripts that he abbreviates as N2 and N3).

n.78For the matching Sanskrit, see Samtani 1971, p. 322.

n.79“Even” here translates samā; there is a wordplay between samā and samādhi .

n.80The Tibetan suggests *yā samyaktvena samā. The Sanskrit has yā samyaktvena samādhiḥ.

n.81Here the Sanskrit adds ayam ucyate āryāṣṭāṅgo mārgaḥ (Samtani 1971, p. 322): “This is called the noble path with eight parts .” The Tibetan brtan pa’i gnes par gyur pa suggests that sthiratvaṁ niyāmam (Samtani 1971, p. 322) should probably be emended to sthiratva­niyāma.

n.82Here the Sanskrit has iha bhikṣavo bhikṣuḥ smṛta āśvasan smṛta āśvāsāmīti yathābhūtaṁ prajānāti (Samtani 1971, p. 43). The Sanskrit adds iha and specifies that “someone” is a monk (bhikṣuḥ); we think the Sanskrit version here is a better reading, matching standard phrasings also found in Pāli parallels (idha bhikkhave bhikkhu […]).

n.83Following the commentary (abhipramodayan ceti […], Samtani 1971, p. 237) in reading abhipramodayan ca rather than abhipramodayan me (Samtani 1971, p. 44). we read the syntax differently from Samtani (2002, p. 44), as we believe cittam should be taken as the object, rather than as the agent, of abhipramodayan; Ferrari (1944, p. 605) also understands cittam as the agent (“Si rallegra la mia mente […]”), but the wording of her text is indeed different (abhimodati me cittam, Ferrari 1944, p. 576).

n.84Samtani (1971, p. 45) adds bhikṣavaś between square brackets; we understand this as indicating that bhikṣavaś (“monks”) was also omitted in the available Sanskrit manuscripts.

n.85Samtani (1971, p. 45) reports that manuscripts F and ANe here add bhikṣavaḥ.

n.86Samtani, also taking into account Pāli parallels, writes bhagavatā (Samtani 1971, p. 45; Samtani 1971, p. 248, n. 5). However, he reports the manuscript reading as bhagavato (Samtani 1971, p. 45, n. 7); furthermore, the manuscripts of the Nibandhana commentary that Samtani abbreviates as G and N are also reported as reading bhagavato (Samtani 1971, p. 248, n. 5). The Tibetan would suggest bhagavato, we think, rather than bhagavatā.

n.87Here the Sanskrit reads pipāsā­prativinayaḥ (Samtani 1971, p. 46). Samtani suggests that the Tibetan would rather correspond to tṛṣṇā­praṇāśanaḥ (Samtani 1971, p. 46, n. 2). However, the Tibetan sred pa is one of the possible translations of pipāsā; and rab tu sel ba translates a few different terms. It is quite possible that the Tibetan was meant to translate pipāsā­prativinayaḥ.

n.88Here the Sanskrit has dharmopacchedaḥ śūnyatopalambhaḥ (Samtani 1971, p. 46): “the cutting off of the dharmas, the obtainment of emptiness.”

n.89Here the Sanskrit has ṛjudṛṣṭi­pratipannaḥ (Samtani 1971, p. 46): “practicing with upright view.” The Tibetan would most likely correspond to ṛjupratipannaḥ, matching Pāli parallels (ujuppaṭipanno). The Nibandhana commentary too reads ṛjupratipannaḥ (Samtani 1971, p. 251).

n.90The Nibandhana explains that this expression refers to the “Dharma of teaching” (deśanādharma), which is said to “accord with the Dharma” of realization (adhigama), since it elucidates it. See Samtani 1971, pp. 251–52.

n.91prajñā­sampannaḥ (shes rab phun sum tshogs pa) does not appear in the Sanskrit edition of the main text, which has a different order for the first few elements in this list. It does appear, however, in the commentary (Samtani 1971, p. 254).

n.92susamārabdhāni (Samtani 1971, p. 47): “well undertaken” is additionally found in the Sanskrit. Samtani has a note (Samtani 1971, p. 47, n. 11) for vijña­praśastāni, saying that it is missing in the Tibetan, as well as in ANe; we suspect that the note was supposed to be for su­samārabdhāni and was somehow printed in the wrong place. The commentary does not represent the last list of qualifiers in the same vibhakti (the nominal endings indicating syntactical roles) as the root text, but we wonder whether there is some corruption in this part of the commentary, since in two different but nearby sentences we find aparāmṛṣṭair iti and then aparāmṛṣṭāṇīti, where both should be quotes from the root text.

n.93Sanskrit adds bhikṣavaḥ (“monks”) (Samtani 1971, p. 47), but note 14 reports that ANe and F do not have it and thus match the Tibetan.

n.94Samtani (1971, p. 48, nn. 4–5) reports that ANe and F have yathābhūtaṁ prajānāti | idaṁ tathāgatasya prathamaṁ tathāgata­balam, which corresponds to the Tibetan yang dag pa ji lta ba bzhin and ’di ni de bzhin gshegs pa’i ltobs dang po’o; his edition, on the other hand, lacks these portions.

n.95parapudgalānām (Samtani 1971, p. 48); “other persons” is added in the Sanskrit.

n.96Here the Degé block print has the correct reading sems can gzhan, while the Degé as reported by Samtani has sems gzhan, an obvious mistake. It seems that Samtani had been relying on the Zhol version.

n.97punar aparam (Samtani 1971, p. 49); “furthermore” is added in the Sanskrit, but Samtani reports that F and ANe do not have it (Samtani 1971, p. 49, n. 3).

n.98punar aparam (Samtani 1971, p. 49); “furthermore” is added here too in the Sanskrit, but Samtani reports that F and ANe do not have it (Samtani 1971, p. 49, n. 3).

n.99tathāgato 'rhan (Samtani 1971, p. 49); “the tathāgata, the arhat” is added in the Sanskrit, but Samtani reports that F and ANe do not have tathāgata .

n.100The Sanskrit syntax is ambiguous as to what samyak may be qualifying, but the Tibetan reads samyak as qualifying nadati.

n.101The Tibetan zhes bya bar rgyas par sbyar ba matches iti vistaraḥ (Samtani 1971, p. 50). This could have been meant as a scribal abbreviation, meaning that the previous section is understood as repeated. We think Samtani understood it in this way, for he translates by repeating it (Samtani 2002, pp. 196–97). However, as the Tibetan abbreviates rather than repeats, we have conformed to this convention.

n.102Here the Sanskrit has punar mayā (Samtani 1971, p. 50): “moreover, by me.” Samtani (1971, p. 50, n. 11), however, reports ANe and F as having anena, which we believe matches the Tibetan ’dis.

n.103Here too the Tibetan has zhes bya bar rgyas par sbyar ba, which is more likely a translation of iti vistaraḥ, rather than of iti pūrvavat (Samtani 1971, p. 50).

n.104The Sanskrit syntax is ambiguous as to what samyak may be qualifying, but the Tibetan reads samyak as qualifying nadati.

n.105Here the Sanskrit reads artha­pratisaṁvit katamā yaduta paramārthe yad avaivartyajñānam | dharma­pratisaṁvit katamā anāsraveṣu dharmeṣu yad avaivartyajñānam | nirukti­pratisaṁvit katamā abhivyāhāre yad avaivartyajñānam | prabhāna­pratisaṁvit katamā yuktam uktam abhilāpitāyāṁ samādhivaśe saṁprakhyāneṣu yad avaivartyajñānam (Samtani 1971, p. 52): “What is the special knowledge of meaning? It is irremovable awareness of the highest meaning. What is the special knowledge of dharmas? It is irremovable awareness of dharmas without fluxes. What is the special knowledge of explanations? It is irremovable awareness of what is uttered. What is the special knowledge of brilliancy? It is irremovable awareness that what has been said is fit and free when something is expressed within the mastery of samādhi.” This explanation is not found in the Tibetan; Samtani (1971, p. 52, nn. 2–6) reports that F and ANe also omit this part.

n.106yad uta (Samtani 1971, p. 53); “it is thus” is additionally found in the Sanskrit as edited by Samtani, but Samtani (1971, p. 53, n. 1) reports that ANe and F accord with the Tibetan, omitting it.

n.107Samtani’s edition does not have this item, but he reports (Samtani 1971, p. 53, n. 5) that it is found in ANe and F (and in the Tibetan). The only way to get to eighteen items is to count this together with, possibly, the previous one.

n.108yad uta (Samtani 1971, p. 54); “it is thus” is additionally found in the Sanskrit as edited by Samtani, but Samtani (Samtani 1971, p. 54, n. 1) reports that ANe and F agree with the Tibetan in omitting it.

n.109Although the Sanskrit edition lacks “he has the Tathāgata’s mark of a great person,” Samtani reports that ANe and F have it, thus agreeing with the Tibetan (tathāgatasyedaṁ mahā­puru­ṣalakṣaṇam, Samtani 1971, p. 54, n. 2).

n.110As pointed out by Samtani (1971, p. 54, n. 5), in this instance the Tibetan not only translates eṇeya with ri dvags but also offers a transliteration.

n.111Here the Sanskrit does not have anything explicitly matching pus mo (“knee,” often representing jānu). Samtani notices that there is a textual problem and supports his preferred reading with the Nibandhana commentary and with the relevant Mahāvyutpatti entry (Samtani 1971, p. 54, n. 7). It is to be noted that the relevant Mahāvyutpatti entry also has pus mo in the Tibetan but nothing explicitly matching it in the Sanskrit; we thus think that the Tibetan was in fact meant to translate the Sanskrit without jānu, but was meant to make it more explicit.

n.112Here, Samtani (1971, p. 56, n. 7) suggests that the Tibetan g.yog ’khor may be translating parijana rather than parivāra, and he supports this suggestion with the relevant Mahāvyutpatti entry. However, g.yog ’khor is an attested rendering of parivāra (see Negi vol. 13, p. 6095).

n.113Here and in the following sentences the Sanskrit continues to have mahā­puruṣasya mahā­puruṣa­lakṣaṇam, but the Tibetan omits it (see also Samtani 1971, p. 57, nn. 7–8). “Fearlessness” (abhaya) does not appear in the Tibetan. It does however appear in the don rnam par gdon mi za ba’i ’grel pa commentary.

n.114Again, the Sanskrit does not explicitly represent pus mo (but see the relevant note above).

n.115Although the Tibetan does not have a precise correspondent for utsukatayā, we feel that “eager to perform tasks” should still work for kimkaraṇīyatā on its own, as reflected in the Tibetan rendering bya ba ci yod ces bya ba.

n.116Sanskrit adds sarva­sattvāśvāsa­prayogatayā (Samtani 1971, p. 58): “due to having offered consolation to all sentient beings.”

n.117Here the Sanskrit reads samāvirala­dantatā (Samtani 1971, p. 58): “his teeth are even and have no interstices.” Sanskrit adds sama (“even” teeth), which corresponds to how the item had been previously listed; however, Samtani (1971, p. 58, n. 7) reports that F and ANe do not have sama, thus matching the Tibetan.

n.118Here the Sanskrit reads, pūrve ’pramāṇa­puṇya­skandho[pasevitā]tmatatayā (Samtani 1971, p. 59). Samtani reports that “Ms. has faulty and faint reading. So also F and ANe” (Samtani 1971, p. 59, n. 4; F and ANe are abbreviations used by Samtani).

n.119Here the Sanskrit reads snigdha­vacana­satya­pālanatayā (Samtani: 1971, p. 60): “because he guarded affectionate speech and truth.” Samtani proposes that the Tibetan should correspond to sattvebhyaḥ mṛdu­vacanālapanatayā (Samtani 1971, p. 60, n. 2).

n.120Here and in the following item, in the Tibetan de bzhin gshegs pa’i (tathāgatasya) is omitted.

n.121Here the Sanskrit reads dharma­saṅgīti­citta­karmaṇyatayā (Samtani 1971, p. 60): “he had a mind that was workable for chanting the Dharma.” Samtani proposes that the Tibetan chos yang dag par sdud pa could correspond to dharma­saṁgraha­karmaṇyatā (Samtani 1971, p. 60, n. 6). However, the Mahāvyutpatti has chos yang dag par sdud pa as the translation for the entry dharma­saṁgītiḥ (see Negi vol. 3, p. 1293).

n.122Samtani suggests that ’thun par/mthun par should be translating anukūla rather than pradakṣiṇa (Samtani 1971, p. 61, n. 2). However, ’thun par/mthun par is attested as a translation of pradakṣiṇa (see Negi vol. 5, pp. 2116–17; see also De Jong 1975, p. 117). We have used “courteous dexterity” so as to reflect, at least in part, the etymological rationale for this explanation (where the pradakṣiṇa = “turning to the right” is caused by pradakṣiṇa = “courteous dexterity”). This rationale is unfortunately lost in Tibetan translation.

n.123Here the Sanskrit reads sarvaprāṇa (Samtani 1971, p. 61): “all life forces.” Tibetan would correspond to sattvaprāṇa (“the life force of sentient beings”), as pointed out by Samtani. We believe that the Tibetan reading is better, and sattvānāṁ prāṇa- is indeed attested elsewhere. See sattvānāṁ prāṇa­rakṣāya (Hevajra­tantra 2.4.90 in Tripāṭhī and Negi 2001, p. 193) and sattvānāṁ prāṇahāriṇi (Mañjuśrī­mūla­kalpa 12.18; 12.14 in Śāstrī 1920, p. 119).

n.124sku here has no matching term in the Sanskrit.

n.125This mark is not in Samtani’s edition but is reported by Samtani as found in Ms, F, and ANe (Samtani 1971, p. 62, n. 3) with the addition of one reason in the manuscript that Samtani abbreviates as Ms: “and helped broken sentient beings to cross beyond their own troubles” (bhinna­sattva­santāraṇatayā). It is also found in the Nibandhana commentary (Samtani 1971, pp. 305–6).

n.126The Sanskrit has tulyapāṇirekhāś ca (Samtani 1971, p. 63); this is omitted in the Tibetan, but without it we get only seventy-nine marks. Samtani reports that F and ANe also omit this (Samtani 1971, p. 63, n. 4).

n.127The Sanskrit here has vṛttāṅgulayaś ca (Samtani 1971, p. 63), “round fingers,” while the Tibetan has sen mo rnams zlum pa, “round nails.” Samtani (1971, p. 63, n. 5) speculates that the Tibetan may contain a mistake in the transmission wherein sor mo was accidentally substituted with sen mo in this item. This is very plausible. It is worth noting, though, that Samtani (1971, p. 63, n. 5) also reports that F and ANe have vṛttāṅgulinakhāś ca, “round fingernails.” (Also, the Tibetan version consulted by Samtani, as he reports it, contains one further mistake for sor mo, while the Degé Parphud (par phud) printing is fine.) Stok has indeed sor mo, thus supporting our preference and, first of all, Samtani’s insightful proposal. The relevant entry in the Mahāvyutpatti has vṛttāṅguliḥ/sor mo rnams zlum pa.

n.128“Hidden” translates gūḍha/mi mngon pa.

n.129Here the Sanskrit reads pradakṣiṇāvarta­gāminaś ca (the reading of manuscript F; see Samtani 1971, p. 64, n. 1); pradakṣiṇa­gāminaś ca (Samtani 1971, pp. 63–64). Whether we read āvarta or not, we could have the same English translation. We believe that the Tibetan g.yas phyogs su ldog cing gshegs pa translates the reading of F; the relevant Mahāvyutpatti entry (Mahāvyutpatti no. 283) has this expression as the translation of pradakṣiṇāvarta­gāmī (see Negi vol. 13, p. 6064).

n.130Samtani (1971, p. 64, n. 2) reports the Tibetan as having instead mi g.yo bar gshegs pa, but the Degé Parphud edition has it as mdzes par gshegs pa.

n.131Here the Sanskrit has avakra­gātrāś ca (“bodies that are not crooked”), but we wonder whether something might have gone wrong in the transmission: the Tibetan mi g.yo bar gshegs pa (that Samtani thought was in place of the previous item) would correspond to avakra­gāminaś ca (see Negi vol. 10, p. 4380, reporting the relevant Mahāvyutpatti entry), which seems likely to us. However, we find attestations for both avakragāmitā (Abhi­samayālaṁkāra­śāstra­vṛtti, Amano 1975, p. 286; Abhi­samayālaṁkārālokā, Wogihara 1932–73, p. 921, lines 4–5; Dharmasaṁgraha, Müller and Wenzel 1995, p. 19) and avakragātratā (Sāratamā, Jaini 1979, p. 182).

n.132The equivalent, gātratā, is not included in the Sanskrit, but it is found in parallel passages from other texts: pṛthu­cāru­maṇḍala­gātratā (Abhi­samayālaṁkāra­śāstra­vṛtti, Amano 1975, p. 286; Abhi­samayālaṁkārālokā, Wogihara 1932–73, p. 921, line 9) cāru­pṛthu­maṇḍala­gātratā (Sāratamā, Jaini 1979, p. 182; note that this, as well as the phrase in the Vṛtti and in the Ālokā, is commenting on the same expression occurring in the root text of the Abhi­samayālaṁkāra). Importantly, sku kho lag yangs shing bzang ba is attested as the translation of pṛthu­cāru­maṇḍala­gātraḥ (Negi vol. 1, p. 175, referring to Mahāvyutpatti no. 293).

n.133Samtani (1971, p. 64, n. 6) suggests that the Tibetan should rather correspond to spaṣṭa­kukṣayaś ca; however, dku skabs phyin pa is attested as a translation of mṛṣṭakukṣiḥ (Negi vol. 1, p. 105, reporting the relevant Mahāvyutpatti no. 302); thus the Tibetan, we think, matches the Sanskrit well.

n.134This item and the next appear in reversed order in Samtani’s edition; however, the “thin tongue” is added by him on the basis of the Tibetan ljags srab pa, rendered as tanujihvāś ca as per Ferrari’s conjecture (Samtani 1971, p. 64, n. 12; this conjecture is supported by the relevant Mahāvyutpatti entry, no. 317; see Negi vol. 4, p. 1460). This item should probably have gone first anyhow, considering the Tibetan.

n.135Tibetan cing snyen la ’jam pa does not repeat “voice,” and Samtani points out that F has mañjusvarāḥ. We would propose *madhura­cāru­mañju­svarāś ca as the more likely Sanskrit from which the Tibetan was translated. The relevant entry in the Mahāvyutpatti (no. 320) indeed has madhura­cāru­mañju­svaraḥ (although the wording is slightly different: gsung snyan cing mnyen la ’jam pa; see also Negi vol. 16, p. 7363).

n.136Sanskrit does not have anything that could match dkar ba (“white”). We are not sure whether the Tibetan considers these two as one item or two. If these are counted as two, we get eighty marks.

n.137spyan dkar gnag ’byes shing ud pa la’i ’dab ma yangs ba lta bur ’dug pa dang. The Sanskrit reads nīlotpala­dalanayanāś ca (Samtani 1971, p. 65), “eyes like the petals of a blue lotus”; Samtani (1971, p. 65, n. 7) points out that Ferrari adds sitāsita- (“white and nonwhite”) before -nīla- (“blue”). The conjecture resembles the wording of the relevant Mahāvyutpatti entry (no. 331), which is somewhat different: spyan dkar nag ’byes shing padma’i ’dab ma rangs pa lta bu/sitāsita­kamala­daraśa­kalana­yanaḥ (see Negi vol. 8, p. 3397). The wording sitāsita- is also found in the Dharma­saṁgraha (sitāsita­kamala­dalana­yanatā, Müller and Wenzel 1995, p. 20), in the Abhi­samayālaṁkāra­śāstra­vṛtti (sitāsita­kamala­dalana­yanatā, Amano 1975, p. 288), Haribhadra’s Abhi­samayālaṁkārālokā (sitāsita­kamala­dalana­yanatā, Wogihara 1932–73, p. 922, line 9), and also, most likely, in Ratnākaraśānti’s Sāratamā (We think that the printed text sitasita­kamala­dalana­yanatā [Jaini 1979, p. 184] should be emended to sitāsita­kamala­dalana­yanatā).

n.138“Firm” is not in the Sanskrit; we were unable to find parallels for this passage.

n.139“Black” is not explicit in the Sanskrit (bhramara­sadṛśa­keśāś ca, Samtani 1971, p. 66). Samtani (1971, p. 66, n. 2) suggests that the gnag in the Tibetan could correspond to an additional asita in the Sanskrit, but parallels suggest that it is just a slightly explanatory translation, and it corresponds to this very Sanskrit. See for example Negi vol. 9, p. 3978, where the relevant Mahāvyutpatti entry (no. 342) is given.

n.140asaṁhata­keśāś ca (Samtani 1971, p. 66); Samtani (1971, p. 66, n. 4) suggests that the Tibetan might be translating asamṛdita­keśāś ca, corresponding to the reading of F. However, we think that the reading that he reports for ANe, asaṁsudita­keśāś ca, is likely to preserve traces of the right reading; it could be emended to asaṁluḍita­keśāś ca. Parallels suggest that the Tibetan could more likely match asaṁluḍita­keśāḥ (asaṁluḍita­keśatā, Abhi­samayālaṁkāra­śāstra­vṛtti, Amano 1975, p. 290; Abhi­samayālaṁkārālokā, Wogihara 1932–73, p. 922, line 19; Sāratamā, Jaini 1979, p. 284).

n.141The Sanskrit colophon as per Samtani’s edition reads ye dharmā hetu­prabhavā hetus teṣāṁ tathā­gato hy avadat | teṣāṁ ca yo nirodha evaṁvādī mahā­śramaṇaḥ || likhitam idaṁ samvat 319 caitra śukla 9 (Samtani 1971, p. 68): “ ‘For dharmas that come about from causes, the Tathāgata spoke of their causes, and also their cessation: the great śramaṇa speaks in this way.’ This was written in the Saṁvat 319, in the month of Caitra, on the ninth of the white half.” The Sanskrit colophon as per Ferrari’s edition reads śubhaṁ || samvat 1971 dharma­rājena likhitvā divya­deva­śarmaṇāya dattam idam pustakam iti | tat­pustakāt pratilikhitam (Ferrari 1944, p. 587): “Good! In the year 1971, this book was given by Dharma­rāja, after having written it, to Divya­deva­śarmaṇa. It has been copied from that book” (1971 Nepali Samvat = 1915 ᴄᴇ). Stok includes a Tibetan transliteration of the ye dharmā verse, followed by dge’o/ bkra shis par shog. The Comparative Edition (dpe bsdur ma) Kangyur records that the Yongle, Lithang, Kangxi, Narthang, and Choné versions are all missing a colophon and that Stok is also missing a colophon.