Your well-considered observations on this puzzling problem are very much appreciated. They make perfect sense to me, and few doubts would remain if we were dealing with a text written in normal Sanskrit. The mūla Kālacakra-tantra, however, is not written in normal Sanskrit. Nor is it written in Buddhist Hybrid Sanskrit, the language described by Franklin Edgerton that is seen in old verses preserved in Buddhist texts. It is not even written in “just bad Sanskrit,” as David Snellgrove characterized the language of the Buddhist Hevajra-tantra in 1959, often repeated by others since. The peculiarities of its language were briefly described by Puṇḍarīka at the end of the third introductory section of his Vimalaprabhā commentary on the laghu Kālacakra-tantra, pp. 29-30 in the 1986 printed edition (link given earlier). These peculiarities were discussed in detail by John Newman in the article that Matthew referred to, “Buddhist Sanskrit in the Kālacakra Tantra” (attached).
One of the items in Puṇḍarīka’s summary of its linguistic peculiarities specifically pertains to meter: kvacit vṛtte yati-bhaṅgaḥ. While the yati or metrical pause is not the particular problem that we are dealing with, it does serve to show that we should not expect the meter to be regular. Indeed, this is what we find. In the small sample of just twenty-one and one-half verses, besides the pāda containing aja that is one syllable short, there are four pādas that are one syllable long: 11a. tantre 'smin ṛṣikulādīnām; 13c. sarvanivaraṇaviṣkambhī; 15b. trayodaśānye krameṇa te; 22a. laghutantre mañjuvajraś ca. In this text, then, we cannot take for granted that the verses were originally written correctly in regular meter, as we can in normal Sanskrit. Snellgrove found the same thing in his edition of The Hevajra Tantra (vol. 2, p. ix): “More than a hundred lines are quite irregular, and although they clearly represent ślokas of a kind, it is impossible to see how many of them can ever have been anything but irregular.”
If we posit that samudravijayo 'jaḥ in 17b is an error for samudravijayo 'jayaḥ that occurred in the copy of the mūla Kālacakra-tantra used by Puṇḍarīka, we must assume that Puṇḍarīka did not catch the error, but retained it in his quotation from that text and then adopted the erroneous name aja in his commentary on laghu Kālacakra-tantra 1.27. Puṇḍarīka is supposed to be the son of Mañjuśrī Yaśas, who prepared the laghu or condensed Kālacakra-tantra from the now mostly lost mūla or root Kālacakra-tantra. So he should be “in the know” about the linguistic peculiarities that he summarized. Indeed, Puṇḍarīka several times at other places in his Vimalaprabhā commentary points out these very linguistic peculiarities; for example, saying that here in such and such a word the locative case is used for the ablative (several examples are given by John Newman in his article). Yet he did not correct aja to ajaya in the list of the kings of Śambhala, where he himself appears as the second kalkī king, but on the contrary he adopted aja in his commentary at 1.27. While we can explain the perpetuation of an error in the Vimalaprabhā by an unwillingness of copyists to emend the text, it is harder to explain an error by its author himself. It is easier to accept such a seeming error as one of the linguistic peculiarities that he was fully aware of.
As for the meaning of the name, while ajaya, “unconquerable,” clearly makes a better king’s name, there are a few other names in this list that, like aja, “unborn,” seem more fitting for some philosophical or cosmic principle. These are, in the list of kalkī kings: no. 15. ananta, “infinite”; and no. 13. viśvarūpa, “he whose body is the all, i.e., the universe”; and in the list of dharmarāja kings: no. 6. viśvamūrti, “he whose form is the all, i.e., the universe.” We have no indication from Puṇḍarīka what he understood the name to mean. The earliest of the Tibetan translations is the one that apparently took aja as abja, chu skyes, “water-born.” I have no confidence that the next Tibetan translation, rgyal dka', “unconquerable,” was any more correct.
Best regards,
David Reigle
Colorado, U.S.A.
Dear David,My general experience is that copyists in the Sanskrit/Indic tradition are very conservative. Almost always, they do not make emendations and, if they do, they usually note them in the margins. Secondly, your mss do not seem to fall into versions or recensions. In effect, you have only one ms reading aja. Also, you have indicated that the source of the (seemingly) multiple attestations is unlikely to have been distant from the autograph or a copy that was made available for further copying. There is thus room to believe that an oversight made in the autograph or in a released copy of the autograph has been faithfully retained. Therefore, unless another independent occurrence of aja in the sense of ajaya is found, the emendation that occurred to Madhav and you should be accepted. Just as we do not accept in a critically constituted text grammatically wrong forms obviously incongruent with the rest of an author’s/scribe’s style or standard, we should not accept a word for which no external corroboration is available, especially if the meaning of that word (‘unborn’ in this case) is unlikely to be confused or associated with a meaning several of us have viewed as contextually likely and the expresser of that meaning gives us a metrically sound text.I hope I make sense even if I fail to create conviction.