Đại Hữu pedestal fragment (C. 171)

Finot & Goloubew observe that “Le texte subsistant ne doit guère représenter que le quart de l’inscription totale” and that “Il paraît être rédigé en çlokas” (1925, p. 472). The second claim is not confirmed by the more precise metrical analysis presented here. Our reconstruction of the lost text is based on the assumption that the text would have been more or less symmetrically distributed over the two sides of a pedestal, the conduit for libation fluids marking the transition from the first part (A) to the second (B). Part A of the text would have begun on either near the middle of the vertical face straight across the pedestal from the conduit, or in the far left corner viewed from the conduit; part B would have ended on back at the same point on the vertical face across the conduit, or in the far right corner. One line of text would, in each part, have extended over about 65 akṣaras. ]

To attempt to reconstruct the stanza structure of the lost parts, we must consider that even if the entire text is not in anuṣṭubh, it certainly includes a sequence of stanzas in that metre. Even if the extant fragments of text cannot be allocated to the first or second hemistich of an anuṣṭubh stanza, we can be certain that each lacuna must cover a number of syllables equivalent to A, the number of syllables missing from anuṣṭubh hemistichs interrupted by the lacuna, plus B, a multiple of 16 (corresponding to 0 or more lost anuṣṭubh hemistichs). Thus the length of the lacuna at the beginning of A3 must be 11+16n akṣaras, that is, two syllables (to complete the hemistich beginning in A2 before the lacuna), plus nine syllables (to complete the hemistich ending in A3 after the lacuna), plus zero or more times 16 syllables for any fully lost hemistichs in between. Similarly, the lacuna at the end of B1 must comprise 17 +16m akṣaras, that is, nine to complete the hemistich beginning in B1, plus eight to complete the one ending in B2, plus any number of fully lost hemistichs. The contents of B3 seem to be prose, which must have started at some point in B2, so the extent of the lacunae in B2 and B3 cannot be estimated accurately by such a calculation. As for that in A2, the extant text of that line does not fit the requirements of the anuṣṭubh metre. However, yasya prajāsu sutarāṁ kr̥ (after a space that indicates a break) is quite certainly the beginning of a vasantatilakā line, while cañcalā buddhiḥ before that break may well be the end of an āryā stanza. We know of no other metrical template that might fit this text, so the above identification is fairly certain unless the text is prose. Assuming it is not, the lacuna at the end of A2 must consist of 47+16p characters to complete a full vasantatilakā stanza, plus any number of anuṣṭubh hemistichs. Finally, the lacuna in A1 is again uncertain. If the end of an āryā is preserved here, then this lacuna must have comprised at least 19 morae for the partially preserved second hemistich, plus 30 morae for the fully lost first hemistich. The number of necessarily lost syllables can be estimated, with a wide margin of uncertainty, to be slightly under forty (assuming that three akṣaras make up four morae on average). The lacuna in A1 may have been longer than this, as the stanza may have been preceded by some opening prose or, less likely, by another stanza of unknown metre.

Assuming that lines on the two parts were roughly the same length, we must look for possible values of lacuna length in each line that are reasonably close to each other. Some variation is of course expected due to uneven character size. We are quite confident that the lacuna in A2 must cover at least 47 characters, with the effect that considerably shorter lacuna options may be eliminated in the other lines as well. It is thus most likely that the size of the lacuna is around 40 to 45 characters in line A1 (the 37 required plus a short prose opening); exactly 47 characters in A2 (the minimum required); 43 in A3 (the 11 required plus 32 for two lost anuṣṭubh hemistichs); 49 in B1 (the 17 required plus 32 for two lost anuṣṭubh hemistichs); around 45 to 50 in B2 (the 12 required, plus 32 for two lost anuṣṭubh hemistichs, plus a few characters of prose), and an unknown number of characters at the end of B3. Assuming this last line was filled to the end with text, as many as 60 characters may well have been lost here, since the characters of this line are slightly smaller than those of the other lines.

The above estimate is to some extent corroborated by the presence of a symbol in A3. It is a logical assumption that this symbol would occur between stanzas rather than inside a stanza, after the first hemistich. By our estimate, this symbol falls between stanzas 4 and 5. Reconstructions involving a shorter lacuna are ruled out by the presence of a vasantatilakā stanza in A2. An alternative scenario with three, rather than two, anuṣṭubh hemistichs lost in lines A3, B1 and B2, would result in the symbol falling in the middle of a stanza and can thus be ruled out. As a final alternative, it does seem possible that four anuṣṭubh hemistichs have been lost in each of these three lines. At over 90 akṣaras per original line, this is rather long, whereas our preferred reconstruction, with about 60 akṣaras per line, of which 15 or 16 are extant in the lines with regular-sized characters, corresponds exactly to Finot and Golubew’s estimate cited above. Should this latter scenario, nevertheless, be the actual case, an extra stanza of about 32 characters (an anuṣṭubh, possibly an āryā) may well have existed before our stanza I (or there may have been a longer opening in prose), and further extra anuṣṭubh stanzas existed between our stanzas III and IV, between our V and VI, and after our VIII. As a final twist, it is possible that the fragment preserved in line B3 is not prose but part of an āryā stanza, the first hemistich of which ends with śrīmāñ. The presence of a space before, rather than after, this word speaks against this assumption, but the prosody does permit it. In this case, the putative āryā stanza lacks 23 morae (corresponding to 15 characters with a wide margin) at the beginning, that would have been in the lacuna at the end of B2. This in turn means that our stanza VIII did not exist; instead, we would then have to posit about 15 to 20 characters of prose after stanza VII, followed by this āryā (now stanza VIII).