Notes

n.1Published from Cambridge University Library Ms. Add. 1680.8.3 and marked at the time as “unidentified” in Hidas 2021, p. 42.

n.2Lee 2004, p. 79: Sarva­karmāvaraṇaviśodhanī haritā tri­śūka­vajrāṅkasitaraktakamaladharā; Toh 3141, folios 127.b–128.a: las dang sgrib pa thams cad rnam par sel ba ljang gu ste rdo rje rtse gsum pas mtshan pa’i padma dkar dmar ’dzin pa’o ||. The Tibetan translation las dang is surprising, as we would expect las kyi.

n.3Denkarma, folio 303.a; no. 415 in Herrmann-Pfandt, pp. 237–38; and no. 382 in Kawagoe 2005, p. 21.

n.4Pelliot tibétain 49.3; Pelliot tibétain 49 itself is a kind of proto-Compilation of Dhāraṇīs. The identity of the text was first determined by Lalou 1939, pp. 16–17.

n.5The translation of this dhāraṇī is: “Obeisance to the Three Jewels! Oṃ, kaṅkani kaṅkani radiant, radiant, destroyer, destroyer, trembler, trembler, remove, remove my entire succession of karma, svāhā!” The reading in the Dunhuang manuscript (henceforth marked as Z) is corrupt at the end: na mo rad na tra ya ya | om ka ka ni | ka ka ni | ro tsa ni | ro tsa ni | tro tha ni | tro tha ni | tra sa ni | tra sa ni | pra ti hA ni | pra ti hA ni | sa rwa pA ra mA | pA ra mA | pA ri ni mA swA hA. Also, note that the Sanskrit differs slightly.

n.6Z is marred by a serious eyeskip here: the text jumps two sentences, reading, “If one recites it constantly […] bad omens,” etc., with the bracketed ellipsis marking the omitted text.

n.7Normally “the three times” (dus gsum) refers to “the past, the present, and the future,” but in the present context, it must mean “at the three junctures” (i.e., dawn, noon, and dusk). Cf. Candrakīrti’s Pradīpoddyotana (Chakravarti 1984, p. 189 and p. 215 respectively): triṣkālaṃ sandhyātraye and triṣkālaṃ trisandhyam.

n.8Z reads “or” (sam) instead of “and” (nas). The canonical reading is not impossible if we understand the booklet to be very small.

n.9Z reads “at the time of death” (’chi ba’i tshe) instead of “of a moribund” (’chi ba’i).

n.10That is to say, the dhāraṇī is customized by replacing the me (“my”) with the name of the beneficiary in the genitive.

n.11If the empowered substance was water.

n.12The Stok Palace manuscript reads “worshipped” (mchod) instead of “cremated” (bsreg), which is also the reading of Z.

n.13A funerary caitya (mchod rten); Z has a fascinating variant, “tomb” (mchad pa). It is not impossible that the text was customized to fit local practices around Dunhuang.

n.14See n.­10.

n.15Z reads “shadow” (bsgrib pa) for “practitioner” (sgrub pa po); if the reading is not a corruption, it must mean the shadow cast by the aforementioned caitya . In that case, understand the following elements to refer to the caitya , not the ritualist.

n.16The string “manifest his own appearance” is not found in Z. It is possible that it is a contamination from a similarly phrased passage below, in the penultimate paragraph.

n.17Lit. “will say, ‘Well done!’ ” ( legs so), which usually corresponds to sādhu in Sanskrit. Z repeats the words (legs so legs so).

n.18These are clearly small votive caityas .

n.19The difference between the two, at least according to Bhavabhaṭṭa, a commentator of the Catuṣpīṭha­tantra, is that banners (dhvaja) have an emblem or design (cihna) on them. Ad 2.3.11: patāketi cihnarahitā matsyapakṣyādicihnāṅko dhvajaḥ (Szántó 2012, vol. 1, p. 288 and vol. 2, pp. 91–92); the Tibetan is slightly different, but the import is the same, see Toh 1607, folio 180.b: ba dan zhes bya ba ni nya dang ngang pa la sogs pa’i ri mo med pa’o || rgyal mtshan zhes bya ba ni ri mo dang bcas pa’o.

n.20Z does not transmit “etc.”

n.21Up to here, Z reads merely, “Or, having fashioned a caitya , having worshiped it…”; it is quite likely that some text was lost here. Having said that, it is odd that the canonical reading has “having performed worship in the same way” and then repeats the articles of worship. Therefore, it appears that we have some contamination.

n.22Instead of “so-and-so” (che ge mo zhig), Z transmits “this person” (myi ’di).

n.23Instead of “indeed, by this [merit]” (’di kho nas), Z transmits “from this transmigration” (’di ’i ’khor nas), which may well be the original reading.

n.24See n.­17.

n.25This last paragraph is not transmitted in Z. Some parts of it are, however, attested in the Sanskrit fragment. The sentence with “written on a wall” is attested in the fragment as antaśaḥ kuḍyalikhitām api. It is also noteworthy that antaśaḥ here means “ultimately” and not “at the time of death” as the Tibetan would suggest.

n.26This passage proves that the initial obeisance to Akṣobhya (also see the expanded obeisance in the Sanskrit fragment) is part of the text and not the translators’ obeisance, despite the fact that it is also not transmitted in Z. No other tathāgata was mentioned before. The invitation refers to Akṣobhya’s pure land, Abhirati. Alternatively, emend de nyid to nyid and understand it as a reference to a generic tathāgata with an emphasis (i.e., “the tathāgata himself”).