¹ Smith, John D., The Vīsaḷadevarāsa: a restoration of the text, Cambridge, 1976 Google Scholar (henceforth ‘Smith (1976)’), vii. Since publication I have noticed a number of misprints in this work, and take this opportunity of publishing a list of corrections: (a) p. 10, I. 18, add full stop at end of line; (b) p. 15, l. 7, for -r- read -r-; (c) p. 23, l. 6, close parenthesis after (i.e. jūṭh[ar]au); (d) p. 24, l. 22, for budhavara read budhavāra; (e) p. 47, n. 13, for vanaṣǎṃḍa, cauṣǎṃḍī read vanaṣǎṃḍa, cauṣǎṃḍī; (f) p. 63, n. b, for nisǎṇe read nῐsāṇe; (g) p. 75, n. b, for ususāi read usasāi; (h) p. 159, n. a, for -d- read -d-; for havadaï read havaḍaï; (i) p. 184, n. 4, for na read ne; (j) p. 232, 1. 1 of text, for [fa] read [rā]; (k) p. 236, I. 3 of text, the first word is sauṇa; (l) p. 296, no. 1, 1. 2, for of read or; (m) p. 297, l. 11, for other read others; (n) p. 298, s.v. ahara, for adhara- read adhara-; (o) p. 299, s.v. āsis-, for ASTS- read AŚÍṢ-. I cannot guarantee that this list is complete.

² Smith (1976), 26.

³ Ibid., 23–6.

⁴ Mātāprasād Gupta, Agarcand Nāhṭā, Bīsala Deva Rāsa, Prayāg, 1953; 2nd ed., 1960. We have here used the 1960 edition.

⁵ A mātrā is a metrical instant with the value of one ‘light’ or half a ‘heavy’ syllable.

⁶ In the small scale this ‘partitioning’ of medieval texts according to metricality is a scholarly commonplace: one thinks, for instance, of the frequency with which the lines of a Hindi dohā are expanded to contain relative-correlative constructions, names of purported authors, etc. But in the Vīsaḷadevarāsa we encounter this on an unprecedented scale: expanded lines are the norm, unexpanded lines a relative rarity.

⁷ The normal stanza consists of six lines in the sequence long, long, short, long, short, long, with a rhyme-scheme AABCDC.

⁸ It will be noticed that various modifications in spelling etc. have been carried out for non-metrical reasons: see Smith (1976), 14–15, 21–2, and the notes to the stanzas in question.

⁹ Smith (1976), 22–3.

¹⁰ Ibid., 22.

¹¹ The complete epic text is so long that it would be impossible to sing it all during a single night.

¹² I have myself been working on the Pābūji epic since the winter of 1973–4. For information on the Devnārāyaṇ epic I am deeply indebted to my colleague Joseph C. Miller Jr. of the University of Pennsylvania, who has kindly permitted me to make use of his recordings, transcriptions and background material.

¹³ ‘The singer or the song? A reassessment of Lord's “Oral theory”’, Man, xii, 1977, 141–53Google Scholar (henceforth ‘Smith (1977)’), examples 1–3, p. 145.

¹⁴ By the brothers Javārjī and Rāṇṇ, recorded on tapes 83–6 in the collection of the Jodhpur branch of the Sangeet Natak Akademi. This extract occurs at the point where Pābūjī's niece Kelam and her friends go to swing in the garden.

¹⁵ This festival falls on the third day of the bright fortnight of the lunar month of Śrāvaṇa (July–August); it is celebrated in particular by girls swinging.

¹⁶ Pl.Gl.7 in pt. 2 of Joseph C. Miller Jr.'s unpublished typescript ‘The epic of the twenty-four Bagarāvat brothers and Devnārāyaṇ: A preliminary investigation of poetic form in an oral tradition of Rajasthan, India’. With Miller's consent I have here slightly modified his system of transcription for conformity with my own.

¹⁷ Readers who are puzzled or alarmed by my regular use of the word ‘text’ in dealing with an oral epic are referred to Smith (1977), where it is demonstrated that, pace the ‘Harvard School’, the Pābūji epic has a fixed text which its performers know by heart. It is, a priori, probable that the same applies to the Devnārāyaṇ epic, but this remains to be investigated.

¹⁸ Thus my use of the sign in the word surǎṅgī on p. 350, above (in the nuclear couplet, but not in the sung strophe, since musical rhythm largely takes over from verbal rhythm): the word is pronounced (see p. 354, below) with the scansion ˇ ˇ –, not the w ˇ – – which a ‘classical’ Indologist might expect. (To write surāgī is tempting, but the nasal is fully realized as a consonant, and it is anyway better to retain a single sign meaning ‘apparently metrically long but actually short’.)

¹⁹ For ease of comparison I continue to use the word ‘line’ in the sense ‘passage of text containing a single final cadence’; musically speaking, the line is a recurring melodic unit answering to two lines of text.

²⁰ It is normally very easy to identify the fillers: firstly, the same filler tends to recur redundantly in line after line; secondly, the inclusion of a filler generally forces the performer to sing metrically long syllables to musically short notes; thirdly, occurring fillers tend to be textually ‘unstable’, that is, to be prone to replacement or omission in other performances or at equivalent points in the same performance.

²¹ P. 348, above.

²² Smith (1976), 22.

²³ In ‘The two Sanskrit epics’, a contribution to the forthcoming transactions of the London Seminar on Epic.

²⁴ An equivalent investigation into the epic of Devnārāyaṇ remains to be carried out.

²⁵ As in stanza 7 on p. 354, above (where the -m- is merely scribal).

²⁶ As can be seen on p. 354, below, this form actually replaces bandhāvālā in the spoken arthāv equivalent of this strophe (for the term arthāv see p. 353, below).

²⁷ Smith (1976), 39–40.

²⁸ But metrical syncopation (using the word in its musical sense) is permissible and fairly common—e.g. Vīsaḷadevarāsa 23.5, rūpa deṣi rājā hǎsyau, where the eight mātrās preceding the cadence – – – are grouped in the pattern 2 + 1 + 2 + 1 + 2, which is not divisible into 4-mātrā feet.

²⁹ These songs, together with a number of others, are to be distinguished from the narrative songs which constitute the bulk of the epic: they do not advance the story, and they are both musically and textually different in nature from the narrative songs.

³⁰ Hence my perhaps presumptuous title. I am grateful to my colleague I. M. P. Raeside for checking my comments on this text for factual inaccuracies, etc.

³¹ Catalogue of Gujarati and Rajasthani manuscripts in the India Office Library, by Blumhardt, J. F., revised and enlarged by Master, A., Oxford, 1954, 131–7Google Scholar.

³² The crotchets over the word rāe in 1. 2 of the couplet printed on p. 132 should not be dotted; note also that in the footnote to the same page the bracketed alternatives are inverted.

⁸³ For something approaching a complete list of possibilities see Smith (1976), 12–13.

³⁴ In the epic of Pābūjī it is customary to stop giving any arthāv at or about the point in the story where the avenger Rūpnāth is born; yet performers seem to experience little or no difficulty in producing arthāv for this ‘missing’ section if they are requested to do so—an example is quoted on p. 357, below.

³⁵ These figures are based on the text published by Gupta, not on any single MS; they are therefore likely to be on the low side, since Gupta's editing will have already suppressed some of the less well-attested embellishments.

³⁶ The making of Homeric verse: the collected papers of Milman Parry, ed. Parry, Adam, Oxford, 1971, 272. Emphasis mineGoogle Scholar.

³⁷ A word subject to various spellings: see notes 38 and 41 below.

³⁸ (1) Malhotra, Uṣā, ‘Phaṛ ko Pavāṣo’, Maru-Bhāratī, 1960, 47–57 CrossRefGoogle Scholar; (2) Sahal, Kanhaiyālāl, Mūlyānkan, Jaypur, 1963, 286–330 Google Scholar; (3) ibid., 331–44; (4) Sahal, Kanhaiyālāl, ‘Satiyō kā Pavāṛā’ (Maru-Bhāratī, 1953, 22–39)Google Scholar; (5) Malhotra, Uṣā, ‘Nānaṛiye kā Pavāṛā’ (Maru-Bhārati, 1957, 41–54)CrossRefGoogle Scholar.

³⁹ The originals are deposited in the Birla Central Library, Pilāṇi.

⁴⁰ I have, however, to refer to the attempt made by Kanhaiyālāl Sahal (see p. 26 of no. 4 in the list in n. 38, above) to analyse the metre of the published material, for, in his view, each line contains four mātrās more than I have allowed for the text of the Pābūjī epic. The reason for this is simply that Sahal is seeking to determine the metre of what are clearly passages of arthāv, whilst I have sought to isolate a nuclear text underlying both the arthāv and the sung text. The published ‘episodes’, like all other known arthāv-passages, modify the nuclear text by processes of addition, substitution, and (sometimes) omission, and are thus not themselves—as Sahal admits—strictly metrical.

⁴¹ Svāmī, Narottamdās, ‘Pābūji-rā Payāṛā—(1) Soḍhījī-ro Payāṛo ', Rājasthān-Bhāratī, 1953, 79–86 Google Scholar.

⁴² Cāraṇ, Candradān, Gogājī Cauhān rī Rājasthānī Gāthā, Bīkāner, 1962 Google Scholar.

⁴³ The form quoted is that which underlies the various versions of the couplet Aa + Bb discussed on 146–7 of Smith (1977). For further discussion see my ‘The words and music of Pābūjī's epic’, a contribution to the forthcoming work by E. Hartkamp-Jonxis, J. C. Miller Jr., J. D. Smith and E. van de Wetering: Pābūjī's Paṛ: essays on an Indian cloth-painting and its function in an oral tradition.

⁴⁴ The repetition is a standard feature of Devnārāyaṇ epic performance.

⁴⁵ Words indistinct.

⁴⁶ BSOAS, XLI, 1, 1978, 179 Google Scholar.