¹ Abbreviations: (M)RE = (Minor) Rock Edict; PE = Pillar Edict; SepE = Separate Edict; G = Girnār; K = Kālsī; Dh = Dhauli; J = Jaugaḍa; M = Mānsehrā; Sh = Shāhbāzgaṛhi; Y = Yerraguḍi. I follow Hultzsch's division into sections, and adopt his abbreviations for the PE site names.

² Bloch (p. 129) translates this as “de façon qu'ils se repentant et cessent de tuer”, but Edgerton (JAOS, 72, 1952, 117) has pointed out that haṃñeyasu is passive (it is so translated by Hultzsch).

³ At K only -[ney]u is legible, and at Y [haṃ]ne[yu]. At both places haniyeyu might have been expected as a passive, rather than haṃneyu. It is possible that the -n- which editors read is really -y-, since the akṣaras na and ya are sometimes confused (see IIJ, X, 160–1).

⁴ Useful contributions to the study of PE IV have been made by Senart, E., Les inscriptions de Piyadasi, Vol. II, Paris, 1886Google Scholar, translated into English by G. Grierson, Indian Antiquary, XVIII, 1889, 3–9 (I quote from the English version); Bühler, G., Epigraphia Indica, II, 1894, 251–256Google Scholar; H. Lüders, SPAW, 1913, 988–1028 = Philologica Indica, Göttingen, 1940, 274–312 (I quote from Phil. Ind.); Hultzsch, E., The inscriptions of Asoka, Oxford, 1925Google Scholar; Bloch, J., Les inscriptions d'Asoka, Paris, 1950Google Scholar.

⁵ The two groups of PE sites (the NW group consisting of Tōp, Mīr, and All, and the SE group of Ara Nan, and Rām) differ in their treatment of final vowels. I use to indicate that different versions show different lengths of vowels. Here Tōp differs from the rest in reading a- for the expected ā- (<yāvat). For the whole question of vowel length see Janert, K. L., Abstände und Schlussvokalverzeichnungen in Aśoka-Inschriften, Wiesbaden, 1972Google Scholar.

⁶ Alsdorf, L., Aśokas Separatedikte von Dhauli und Jaugaḍa, Mainz, 1962, 16–17Google Scholar.

⁷ SirMonier-Williams, Monier, Sanskrit-English dictionary, Oxford, 1899Google Scholar.

⁸ tāḍanâdyaṅgacchedarūpaṃ tasya kuryān, na māraṇaṃ.

⁹ Barua, B. M., Asoka and his inscriptions, Calcutta, 1946Google Scholar.

¹⁰ Bühler, G., The laws of Manu, Sacred Books of the East XXV, Oxford, 1886, 382Google Scholar. Kullūka explains: tīritam iti, yatra kvacid ṛṇâdānâdivyavahāre yat kāryaṃ dharmatas tīritam.

¹¹ (L) v ite pi ca me āvuti. baṃdhanabadhānaṃ munisānaṃ tīlitadaṃḍānaṃ patavadhānaṃ tiṃni divasāni me yote diṃne. (M) nātikā va kāni nijhapayisaṃti jīvitāye tānaṃ. nāsaṃtaṃ v nijhapayitā (v.1. nijhapayitave) dānaṃ dāhaṃti pālatikaṃ upavāsaṃ v kachaṃti. (N) ichā hi me hevaṃ niludhasi pi kālasi pālataṃ ālādhayevū ti.

¹² Although Bloch (p. 58) and Hultzsch (p. lxxvi) state that -āni is the masculine plural accusative ending, Edgerton has pointed out (loc. cit., n. 2) that -āni (and -īni) are both nominative and accusative. Lüders had already indicated this (pp. 290–1).

¹³ Lüders, 282 n. 2. Bloch (p. 164 n. 6) refers to Wackernagel, Altindische Grammatik, III, 569.

¹⁴ Lüders, 283.

¹⁵ “Middle Indo-Aryan studies VII”, JOI (Baroda), XVIII, 3, 226 n. 3. The fact that all versions read va shows that, if this is a mistake, it was already in the exemplar underlying all the versions. For the suggestion that other readings in the PE's are based upon mistakes in the exemplar underlying them, see my “Notes on Aśoka's Fifth Pillar Edict”, JRAS, 1967, 26–32. Janert seems to have misunderstood the point of this article, and rejects the suggestion that jatūkā aṃbākipilikā is based upon a mis-writing of jatū kādaṃbā kipilikā on the grounds that no facsimile of PE V supports this reading (p. 71).

¹⁶ “Middle Indo-Aryan studies VII”, 226 n. 4.

¹⁷ e.g. vā for ca at K in RE IV(I), RE V(J), RE X(A); voḍā for coḍā at K in RE II(A), following Janert's reading (p. 106).

¹⁸ Davids, T. W. Rhys and Stede, W., Pāli-English dictionary, London, 1925Google Scholar.

¹⁹ Doubtless a Buddhist connotation is intended. Aśoka says of himself: mayā saṃghe upayite (MRE I), “By me the Order was approached,” cf. the common Pāli phrase saṅghaṃ saraṇaṃ gacchāmi, “I go to the Order as a refuge”. There are contexts in Pāli where tāṇa- is a synonym for saraṇa-, e.g. Bhagavantaṃ yeva tāṇaṃ leṇaṃ saraṇaṃ gavesi (D i 95 M i 232).

²⁰ He justified the form nā saṃtaṃ instead of the expected no saṃtaṃ on the grounds that the negative form of the verb “to be” is nathi, not nothi. This is certainly true of the RE's, but in the PE's the negative particle is no, except for one occurrence of na in PE V(B) no eti na ca khādiyati, where considerations of style arising from the repetition of the negative particle doubtless account for no … na. (Hultzsch reads no … no at Ara, but this is a mistake as the photograph clearly shows na.) Lüders (p. 308 n. 2) referred to two other occurrences of nā at K. One in RE I(B) is now read as no by Hultzsch and Janert. The other is in RE IV(C) nā huta-puluve. This, however, is simply a scribal error, since Dh and Y read no, and K reads no when the phrase recurs in RE V(H) and RE VI(B). For the mis-writing of -o as -ā, cf. karāto in RE XII(H) at G against karoto in (G), pāti for hoti in RE XIII(M), hāti for hoti in RE XIII(H), and tatā for tato in RE XIII(C).

²¹ See the Appendix.

²² Another, but less important, grammatical variation may be shown by the reading paṭipadāye at Tōp in PE V(H). This shows a different case ending from paṭipadaṃ at Ara, Nan, and Rām. The Tōp ending may, however, be a scribal emendation influenced by the following dhuvāye. Mīr paṭipadā may be interpreted as supporting either reading: it could show -ā for -aṃ, or arise from the loss of -ye (Janert, p. 76). All is defective here.

²³ The difference between Tōp Mīr tisāyaṃ in PE V(H) and Ara Rām tisyaṃ / Nan tisiyaṃ arises from a miswriting of tisāyaṃ as tisiyaṃ (for the confusion of -ā and -i see IIJ, X, 163–4). There is a tendency in the SE versions to remove svarabhakti vowels, cf. SE avadhya- in PE V(C) where Tōp has avadhiya- (Mīr and All are defective), so that Ara and Nan removed the (faulty) -i- they received. In PE V(J) they all received and wrote the correct tisāye.

²⁴ The writing of ve in conjunction with the preceding word is in accordance with Janert's statement (p. 20) that mono-syllabic words are not written alone. We should, however, have expected the final -a of nijhapayita to be lengthened in the SE versions by the enclitic following it (Janert, 21). Failure to do this must be due to the scribe responsible for the SE exemplar assuming that the ending -itāve in the version before him was a mistake for -itave, the infinitive ending, of which three examples had already occurred in PE IV. For scribes “correcting” their exemplars see “Lexical variation in the Aśokan edicts”, TPS, 1970, 133–4.

²⁵ See IIJ, X, 162 n. 13.

²⁶ Mīr is defective here.

²⁷ Childers, R. C., Dictionary of the Pali language, London, 1875Google Scholar, s.v. niruddho.

²⁸ Thomas, F. W., “Notes on the edicts of Asoka”, JRAS, 1916, 113–123Google Scholar.

²⁹ R. L. Turner, CDIAL, 7286.

³⁰ Kangle, R. P., The Kauṭilīya Arthaśāstra, Part II, Bombay, 1963, 381Google Scholar.

³¹ For the salutary effect of meditating upon death at the end of life, cf. the story of Aśoka's younger brother Tissa told in Mahāvaṃsa V, 154–60. Aśoka allowed him to reign over his kingdom for one week, but informed him that he would be put to death at the end of that time. As a result of this experience Tissa was turned towards faith in the Buddha's doctrine.

³² There are examples in Pāli literature of relatives and friends pleading on behalf of prisoners, e.g. nijjhapetun ti, sakhīhi attano niraparādhabhāvaṃ pakāsetvā pasādetvā (J iv 498); yathā dhana-yasa-siriñāti-balena balavā puriso attano ñātiṃ vā mittaṃ vā raññā garu-daṇḍaṃ dharentaṃ attano bahu-vissatthabhāvena samatthatāya garukaṃ daṇḍaṃ lahukaṃ kāreti (Miln 109).

³³ Introduction, § 23. He explains the ending -aṃ as being neuter: “Mais la forme est généralement celle du neutre” (p. 108 n. 9). There is, however, no reason for assuming that in the Aśokan inscriptions the neuter nominative would differ in form from the masculine nominative. On the analogy of mahaṃte “great” in RE XIV(C) we should expect this to be *saṃte in an eastern dialect.

³⁴ See the Critical Pāli dictionary, s.v. atatha-.

³⁵ The abbreviations adopted for Pāli titles are those of the Critical Pāli dictionary.

³⁶ Hultzsch (p. 97) takes saṃpaṭipajamīne as being in apposition to the subject of ālādhayisatha, and translates “and if (you) act thus, …”. Although this is possible, the other occurrences of “act successfully, unsuccessfully” in the SepE's involve a causative form of the root pad-. I prefer to take the non-causative form in the sense of “be successful”.

³⁷ See Lüders, 298.

³⁸ “Notes on the Bahapur version of Aśoka's Minor Rock Edict”, JRAS, 1971, 43.

³⁹ Epigraphia Indica, XXXI, 1956, Plate facing p. 209.

⁴⁰ Epigraphia Indica, XXXVIII, 1969–70, Plate facing p. 2.

⁴¹ JRAS, 1967, Plates facing pp. 96 and 97.