Ven. Buddhadatta's Grammar and Prosody Home PageNext Section
Studies in Ven. Buddhadatta's Prosody
1: The Text
The first of the texts to be examined is Vinayavinicchaya, a verse summary of the Vinayapiṭaka, Excluding the Parivārā which is summarised in the Uttaravinicchaya, which is the next text to be analysed in this series.01 together with a short section at the end on Kammaṭṭhāna. There are some 3,200 verses in the work. The figure is derived by taking the number of verses, which equals 3183, and adding the 17 verses which are found in the Sedamocanakathā, after verse 3028.02 Most of the verses consist of 4 lines, but occasionally there are 6, which means that there are altogether 13,026 lines in the work. Of these 12,550 are in Siloka metre (96.3%); and the remainder are in a mixture of Classical Metres.
2: Sarabhatti & Resolution
Sarabhatti vowels are fairly frequent, being found on 32 occasions, and are mainly of the –riy– type, only –arahati at 2567d, 2703d, 2704d being of a different kind. The following table lists the occurrence of sarabhatti vowels, which occur in the following lines: –pāricariyāya 363b; pāricariyāya 365d, 366d; veḷuriya– 750b; –pariyanto 774c; pariyāya– 899a; sūriye Normally sūriye reads as 2 syll, as in the list here, but on one occasion we must scan it as 3 syllables, at 1150c.03 914c, 1145c, 1147a, 1576b; pariyāya– 937c; –pariyāyaṁ 1106c; –pariyāye 1110a, 1111a; nevāhariyamānaṁ 1448a; ācariyānam– 1608a; pariyāpuṇissasi 1723b; sameniriyā– 1889d; –brahmacariyehi 2208a; brahmacariyañ– 2209b; pariyaye 2302d; pariyāpuṇeyya 2310a; pariyesitvā 2340a; pariyesanādismiṁ 2353a; –athācariyaṁ 2471a; ācariyo 2546b; –arahati 2567d, 2703d, 2704d; ācariyupajjhāyaka 2915a; cariyānukūlato 3130b; –pariyantaṁ 3163c.
Resolution of one heavy syllable into 2 light ones is also quite frequent. There are 34 occasions where this occurs. 22 times at the 1st syllable; 11 times at the 4th, and once at the 6th.
They are found in the following verses: The compliment of this phenomena, contraction of 2 light syllables into one heavy syllable is not found in this text.04 pari– 43a; tati– 389c; pari– 531a; chada– 1106c, 1108b; vara– 1246c; ati– 1327b; bada– 1355a; –gahi– 1448b, 1532b; na pa– 1519d, 1527d, 3019c, 3020a, 3021c; puri– 1572d; puna 1579c; –hari– 1742c; –paṭi– 1779b; paṭi– 1780a, 2659b; gahi– 1905c, 2630a; upa– 2001c, 2550c; nadi– 1067b; –kara– 2140b; sara– 2470b; kathi– 2719c; saha– 2736a; labha– 2905c; sama– 3009d; asu– 3132b; nava– 3161b.
3: Analysis of the metres
By far the most frequent of the metres employed is the Siloka, accounting for some 6275 pādayugas. Of these 6051 are pathyā (96.4%), having the prior line structure: ⏓⏓⏓⏓¦⏑−−×.
As an example of the pathyā Siloka, we can quote the Ganthārambhakathā (vss. 1- 5):
vanditvā sirasā seṭṭhaṁ, buddham-appaṭipuggalaṁ.
bhavābhāvakaraṁ dhammaṁ, gaṇañ-ceva niraṅgaṇaṁ. 
bhikkhūnaṁ bhikkhunīnañ-ca, hitatthāya samāhito.
pavakkhāmi samāsena, vinayassa vinicchayaṁ. 
paṭubhāvakaraṁ etaṁ, paramaṁ vinayakkame. 
apāraṁ otarantānaṁ, sāraṁ vinayasāgaraṁ;
bhikkhūnaṁ bhikkhunīnañ-ca, nāvābhūtaṁ manoramaṁ. 
tasmā vinayanūpāyaṁ, vinayassa vinicchayaṁ.
avikkhittena cittena, vadato me nibodhatha. 
Of the variations, mavipulā is the most frequent, though there is a problem in counting the lines, as what prosody Ven. Buddhadatta was allowing for is uncertain. The normal mavipulā has a profile like this: ⏓⏓⏑⏓¦−,−−×. That is, besides the −−−× structure, there is a short syllable at the 3rd syllable, and a word-break after the 5th in the prior line.
100 of the mavipulā lines actually adhere to this, they are found at 13a, 42a, 84a, 87c, 93a, 94c, 107a, 163a, 207c, 208c, 241a, 274a, 350c, 353a, 356a, 411a, 421a, 447a, 488c, 510c, 542a, 558a, 600c, 652a, 653c, 665c, 700c, 722c, 781c, 787c, 08a, 850a, 857c, 910c, 927c, 1072c, 1085a, 1121a, 1294a, 1303a, 1323a, 1380c, 1393c, 1405a, 1416a, 1422c, 1432c, 1433a, 1437c, 1454a, 1462a, 1473a, 1477a, 1489c, 1503a, 1507a, 1512a, 1521a, 1528a, 1531a, 1555a, 1571a, 1673a, 1698a, 1729c, 1777a, 1787a, 1807c, 1877a, 1885c, 1989a, 2059c, 2060c, 2068c, 2185c, 2204a, 2382c, 2561c, 2563a, 2580a, 2592a, 2593c,
2635c, 2652c, 2676a, 2715a, 2744a, 2771a, 2772a, 2833a, 2839a, 2905a, 2949c, 2950ac, 2951a, 2981c, 3051a, 3053a, 3088c, 3161a.
5 mavipulā-type lines have an irregular opening, but have the normal word-break, they are found at: 1831a, 1847a, 1959a, 2015c, 2015c.
15 mavipulā-type lines omit the regular word-break, but have the normal opening: 332a, 608c, 698a, 915a, 950c, 1170a, 1393c, 1401c, 1912a, 2338c, 2434a, 2446c, 2728c, 3051a, 3137c.
There are a further 20 mavipulā-type lines that have neither the normal word-break, nor the normal opening. They account for all the otherwise irregular Siloka lines: 209a, 268c, 621a, 635a, 637a, 849c, 858a, 1059c, 1321c, 1326a, 1339a, 1345a, 1390c, 2544c, 2546c, 2919c, 2928c, 2946a, 2993a, 2994c, 3028. Kha.c. I can see no pattern in these lines that would allow us to establish a new profile for the mavipulā, unless we simply accept the basic profile ⏓⏓⏓⏓¦−−−×, which they do indeed all fit into.
As in the Medieval prosodies the word-break and the opening are not defined, it seems we should accept all these lines as being of the mavipulā-type. In this case there are 140 mavipulā lines in this work, which amounts to 2.2% of the total Siloka lines.
Of the other variations which occur (1.2% of the total) the figures are as follows:
We also find that there are 12 Anuṭṭhubha (or javipulā) lines in the work, which have the same cadence in the prior and posterior lines (⏓⏓⏓⏓¦⏑−⏑×): 300a, 671c, 789c, 1150c, 1811c, 1970c, 2144a, 2506c, 2665c, 3008c, 3089a, 3172a.
Some of the Siloka pāda-s have 9 syllables: 584c; 1356a; 1635c; 1810c; 2631a; 2669c; 2730a; 3016a; & 3105a. Two of these lines (2669c & 2730a) could be corrected if we accepted resolution against the rule, but that would still leave seven which cannot be corrected in this way, so it may be that the other two are simply wrong also.05 There does not seem to be any particular pattern to these hypermetric lines.
avītivatte majjhaṇhe, gharam-aññassa gacchati.
gharūpacārokkamane, paṭhamena hi dukkaṭaṁ. 
muṭṭhikam-adhikaraṇī saṇḍāso vā tulādikaṁ.
kiñci saṅghassa dinnaṁ ce, taṁ sabbaṁ garubhaṇḍakaṁ. 
palaṇḍuko paṇḍuvaṇṇo, bhañjano lohito pi ca.
harito haritavaṇṇo, cāpalo setako pi ca. 
avassutassāvassutā, manussa puggalassa yā.
sarīram-assa tena vā, phuṭṭhā pārājikā siyā. 
As is normal in the Medieval texts compounds are allowed to cross the line in Siloka pādayugas. These are very numerous, and need not be listed, but we can quote the following as an example:
ñātikāya pi aññātisaññissa vimatissa vā.
ekato-upasampannahatthā gaṇhāti dukkaṭaṁ. 
In these cases the normal rule that the last syllable in a line be taken as heavy is ignored.
A curiosity of the Siloka found in Ven. Buddhadatta's prosody is that the normal avoidance of 2 light syllables in 2nd and 3rd positions does not seem to be adhered to. There are certain texts in the Canon which also seem to allow for this variation, see e.g. The Prosody of the Dhammapada, elsewhere on this website.07 There are some 72 lines in this text where light syllables are found in these positions: 42a, 43a, 213c, 276d, 354a, 355a, 387a, 420c, 436c, 456a, 501c, 659b, 753a, 924a, 931c, 1070a, 1104a, 1125d, 1134d, 1181b, 1218a, 1319c, 1324c, 1357c, 1379b, 1428a, 1444c, 1470a, 1481ac, 1499a, 1548ab, 1609c, 1672c, 1724d, 1731c, 1747c, 1780b, 1954a, 2045a, 2093c, 2208ac, 2256a, 2333c, 2334c, 2365b, 2376c, 2437d, 2468d, 2474d, 2491a, 2504b, 2505c, 2515a, 2553a, 2610b, 2660c, 2709a, 2721a, 2842a, 2851c, 2869a, 2912c, 2990c, 2999a, 3001a, 3005c, 3012a, 3014c, 3028. Ka.a, 3028. Kha.c, 3065b, 3117a, 3147a, 3153a, 3160c.
As an example of this we can quote the following pādayuga, where we have light syllables in 2nd & 3rd positions in both prior and posterior lines:
asayanighare tassa, sayanigharasaññino [1548ab]
b: Other Metres
There are 58 other verses in this work that appear in a variety of other metres, which are mainly of the Classical Samavutta type (54 examples). In these examples watch for the complex internal rhyming schemes which is characteristic of Ven. Buddhadatta's Classical style, particularly in those metres that lend themselves to repetition, like Aparavatta & Saṅgata.
One verse is written in Mattāchandas metre - that is a mixed Vetālīya and Opacchandasakā verse:
sakalo asamāsato mayā.
kathito ettha vinicchayo tato. 
There are 2 verses in Aparavatta metre (541, 3124) which is a Classical Addhasamavutta metre derived from Vetālīya:
vividhanayanayutaṁ upenti ye.
vinayanaye paṭutaṁ upenti te. 
There is 1 verse in Pupphitaggā metre (1352) which is a Classical Addhasamavutta metre derived from Opacchanasaka:
kakudhakapitthakakundakaḷīnaṁ. ChSa reads: kakundaṁkaḷīnaṁ. Notice the alliteration on the sound ka.08
kusumam-idaṁ pana yāvajīvikaṁ.
The remainder are of the verses are in Classical Samavutta metres, here arranged in ascending syllabic number:
I have been able to find 4 examples of the rare Pamāṇikā metre in this work (1969, 2078, 2240, 2324); its compliment Samānikā (−⏑−⏑−⏑−⏑ x 4) unfortunately doesn't occur.
adubbale tu cīvare,
ubhosu kaṅkhitāya vā,
avoca dukkaṭaṁ jino. 
The rather special Vijjummālā metre, consisting of 8 heavy syllables is found in 6 places (117, 170, 2062, 2112, 2278, 2283):
dvinnaṁ maggaṁ gacchantīnaṁ,
ekā gantuṁ no sakkoti.
ohīnā ce tassāpatti. 
The 10 syllabic metres are all fairly rare in the literature, in this work there are 2 verses, mainly in Campakamālā metre, with one line in each verse being in a different metre. Campakamālā/Meghavitāna (1320) & Campakamālā/Rummavatī (2214; line c = Rummavatī).
khādaniyaṁ vā bhojaniyaṁ vā, I read short -i- twice in line a to give a Campakamālā line; otherwise the metre would read −⏑−−−−⏑−−−, which is unknown to me.09
kiñci vikāle yo pana bhikkhu.
khādati bhuñjati vā pi ca taṁ,
so jinavuttaṁ dosam-upeti. 
The most frequently used metre apart from the Siloka is the Upajāti, which accounts for 25 of the verses in this work (137, 240, 305, 306, 307, 308, 310, 323, 489, 560, 732, 785, 888, 1223, 1273, 1305, 2254, 2284, 2834, 2843, 2914, 2915, 3077, 3123, 3165).
Upajāti normally consisits of a mixture of Indavajirā (−−⏑−¦−⏑⏑¦−⏑−×) & Upindavajirā (⏑−⏑−¦−⏑⏑¦−⏑−×) metres, but the definition allowed for in the Medieval prosodies, and the one employed in this analysis, also allows for other mixtures, including metres of different syllabic lengths.
The break is very well established and in this work never takes another form. There is only one example here of Upajāti containing a metre other than Indavajirā and Upindavajirā, and that is at the example we quote here:
manussapāṇimhi ca pāṇasaññitā,
sacassa cittaṁ maraṇūpasaṁhitaṁ.
upakkamo tena ca tassa nāso,
pañcettha aṅgāni manussaghāte. 
One of the characteristics of this work is that some of the words in compound cross the break in the middle of the pādayuga, which is a feature of Medieval prosody that is not found in the Tipiṭaka; in these verses the normal word-break occurs in the seam of the compound. This is very common in the Siloka verses here (see the examples above), but extends occasionally to the Upajāti verses also. In this work, the compounds cross the line in the following places: 2914ab & cd; 2915ab.
vattāni bhatte gamikassa jantāghare tathā vaccakuṭippavese. 
There are 4 verses written solely in Indavajirā metre (118, 309, 311, 2612):
mātāpitūnaṁ pana dassanatthaṁ,
daṭṭhuṁ gilānaṁ tad-upaṭṭhakānaṁ,
bhattādi nesaṁ pariyesanatthaṁ. 
At 2011 there is a verse in Dodhaka metre, though we need to make some changes for it to scan properly:
yā pana bhikkhunĭ ussayavādā;
aṭṭakarī mukharī vihareyya.
kenaci yena narenidha saddhiṁ. ChSa reads: yena kenaci here.10
sā garukaṁ kira dosam-upeti. 
There is a rare metre in this class found in this work. It is identified by Prof. Helmer Smith (Saddanīti IV, p. 1167) as a metre called Saṅgatā. The name is recorded in the late Vttodayapariśiṣṭaya, but I have not yet seen it elsewhere.11 There are 3 examples (vss. 38 1963, & 2441). Again notice the rhyme (yamaka) achieved by internal repetition of word-sounds in the line:
imaṁ hitavibhāvanaṁ bhāvanaṁ,
avedi surasambhavaṁ sambhavaṁ.
sa mārabaḷisāsane sāsane,
samo bhavatupālinā pālinā. 
Vaṁsaṭṭhā by itself accounts for only 3 verses (171, 1962, 2328):
varāhabyagghacchataracchakādito, Note that -by- fails to make position in this line.12
upaddavā muccitukāmatāya yo.
tatheva taṁ ṭhānam-atikkameti ce,
na koci doso pana bhaṇḍadeyyakaṁ. 
There are 4 verses written in the Toṭaka metre (37, 324, 2442, 2443):
avagacchati yo pana bhikkhu imaṁ,
amaraṁ ajaraṁ arajaṁ arujaṁ,
adhigacchati santipadaṁ pana so.
The last verse in the work is written in Bhujaṅgappayāta metre, which is quite rare, and is the only example of the metre found in this work:
imaṁ sārabhūtaṁ hitaṁ atthayuttaṁ,
karontena pattaṁ mayā yaṁ tu puññaṁ.
ayaṁ tena loko munindappayātaṁ,
sivaṁ vītasokaṁ puraṁ pāpuṇātu. 
There is only one metre is this class, which the ever-popular Vasantatilakā, although there are only 2 examples here (1530, 3040):
sikkhappahānamaraṇehi ca liṅgabhedā.
dānena tassa ca parassa abhikkhukassa,
sabbaṁ paṭiggahaṇam-eti vināsam-evaṁ. 
Ven. Buddhadatta's Grammar and Prosody Home PageNext Section
last updated: March 2006