These fragments are all in Prakrit gāthās, in whole or in part, that bear on the grammatical characterization of Prakrit. The first group containts fragments attributed to Harivṛddha. The second contains fragments with no attribution. The third group contains testimonia. I can make no claims to completeness: the Jain commentarial literature is vast, and I rely largely on the findings of A. N. Upadhye (1931–1932) and Hiralal Jain (1945).
FRAGMENTS ATTRIBUTED TO HARIVṚDDHA
These fragments are collected from the following materials:
• Ratnaśrīṭīkā (RāŚrīṬī) of Ratnaśrījñāna on Daṇḍin’s Kāvyādarśa (see Mirror of Literature in the bibliography). Written in 931. This appendix reflects most of the suggestions of Bhayani 1973. Some of Ratnaśrījñāna’s quotations are preserved by Saṅgharakkhita in his Mahāsāmi-ṭīkā on the Subodhālaṃkāra (ed. Padmanabh Jaini [Oxford, 2000]).
• Ṭippaṇī (KāAṬi) of Namisādhu on Rudraṭa’s Kāvyālaṃkāra (see Ornament of Literature in the bibliography). Written in 1069. Other readings are given by Kulkarni 1988 = PVSWP.
1. Ratnaśrīṭīkā on 1.33 (p. 23).
कथं तद्भवं तस्मात् संस्कृतात् †वर्णान्यत्वेन† उत्पत्तिर्यस्य तत्तद्भवं शब्दभवमित्यर्थः । तच्च महिन्द-सिन्धव-बहिरादिकं यथोक्तं हरिवृद्धेन ।
mahiṃda-, siṃdhava-, bahira-, etc. [are śabdabhava words.]
2. Ratnaśrīṭīkā on 1.33 (p. 23).
तत्समम् तेन संस्कृतेन समं तत्समम्, प्राकृतशब्दमपीत्यर्थः । तच्च हरि-हर-कमलादिकं यथोक्तं तत्रैव ।
hari-, hara-, kamalā-, etc. [are śabdasama words.]
3. Ratnaśrīṭīkā on 1.33 (p. 23).
देशी प्राकृतं महाराष्ट्रप्रसिद्धम् । तदुक्तम्—
मरहट्ठ-देस-संकेअएहि सद्देहि भण्णए देसी इति ।
मरहट्ठ] Bhayani; मरहट्ठा RāŚrīṬī
संकेअएहि] Bhayani; संकेतएहि RāŚrīṬī
Deśī is expressed through words that are conventionally recognized in the region of Mahārāṣṭra.
4. Ratnaśrīṭīkā on 1.33 (p. 23). The deśī words in this passage have been restored by Bhayani on the basis of Hemacandra’s Deśīnāmamālā.
तच्च बोक्कण-कंकेल्लिचिरिड्डिहिल्ल-सित्थादिकं यथोक्तं तत्रैव ।
बोक्कण] Bhayani; वोक्कण RāŚrīṬī
चिरिड्डिहिल्ल] Bhayani; चिरिहिरहिर RāŚrīṬī
सित्था] Bhayani; सिच्छा RāŚrīṬī
bokkaṇa- (“crow”), kaṃkelli- (“Aśoka tree”), ciriḍḍihilla- (“curds”), sitthā- (“bowstring”), etc. [are deśī words.]
5. Ratnaśrīṭīkā on 1.34 (p. 24). Although not explicitly attributed to Harivṛddha, the context makes the attribution very probable.
महाराष्ट्राः कुन्तल-मुलकाश्मक-विदर्भ-†महियाचरत्रवैश्या†दि-प्रभेदाः7 आश्रयः अधिष्ठानं यस्यां तां महाराष्ट्राश्रयां भाषां वाचं प्रकृष्टं प्राकृतेषु शोभनतमं प्राकृतं विदुः उपदिशन्ति बहुज्ञाः । तदुक्तम्—
†ऊह इअ विविह-भासा परिरि हिअअं य† मोत्तूणं ।
मरह[ट्ठभासिअं चे]अ अत्थि गहिअं कइल्लेहि ॥
कुन्तल ... प्रभेदाः] conj.; कुन्तलामुरलासकविदर्भमहियाचरत्रवैश्यादिप्रभेदाः RāŚrīṬī
मरह[ट्ठभासिअं चे]अ] conj.; मरहअठिअया अ RāŚrīṬī
...it is the language of Mahārāṣṭra that poets have accepted.
6. Ratnaśrīṭīkā on 1.33 (p. 24).
ननु सामान्यभाषापि प्राकृतप्रकारोऽस्ति । यदुक्तं हरिवृद्धेन—
अण्णेहि अ एएहि अ सरिसं चिअ होइ सामण्णे इति ।
चिअ] चअ RāŚrīṬī
That which these and the others have in common is in the category of “Common” (sāmānya).
7. Ratnaśrīṭīkā on 1.34 (p. 24). Bhayani restored musumūria on the basis of Siddhahemacandra 8.4.106, which teaches this root as a substitute for bhañj-.
व्यवह्रियते [शि]क्षणं व्यवहारः [तत्र] प्रवर्तते एभिरिति सा च मुसुमूरिअ-मूअच्छिअच्छिकेत्यादिका यथोक्तं हरिवृद्धेन ।
मुसुमूरिअ] Bhayani, मुसुमुरिअ RāŚrīṬī
broken...
8. Ratnaśrīṭīkā on 1.34 (pp. 24–25). Although not explicitly attributed to Harivṛddha, the context makes the attribution very likely.
तदुक्तम्—
सद्दभवा सद्दसमा देसि त्ति अ तिण्णि पाअअण्णेहि ।
सामण्ण-पाअअ-सहिअ †आया अअय इतराणि आउ† ॥
पाअअण्णेहि] Bhayani; आ अ अण्णेहि RāŚrīṬī
पाअअ-सहिअ] Bhayani; ययाअ इसहि RāŚrīṬī
“Derived,” “Identical,” and “Regional” are the three [recognized] by those who know Prakrit;
With the addition of “Common Prakrit”...
9. Kāvyālaṃkāraṭippaṇī 2.19 (p. 17) = PVSWP p. 2.
तथा ह्यष्टौ हरिणोक्ताः यथा—
महुरं फरुसं कोमलमोजस्सिं निट्ठुरं च ललियं च ।
गंभीरं सामण्णं च अट्ठ भणिईओ नायव्वा ॥
फरुसं] PVSWP; परुसं KāAṬi
अट्ठ] PVSWP; अद्ध KāAṬi
भणिईओ] conj.; भणितिउ PVSWP, भणिती उ KāAṬi
नायव्वा] PVSWP; नायच्चा KāAṬi
The sweet, the harsh, the soft, the powerful, the severe, the playful, the profound, and the general: these are the eight bhaṇitis.
UNATTRIBUTED FRAGMENTS
These fragments are collected from the following sources:
• The Nāṭyaśāstra (NāŚā) ascribed to Bharata (see Treatise on Theater in the bibliography). Dates very approximately to between the second and fourth centuries CE. It contains a concise grammar of Prakrit, partially composed in Prakrit, at the beginning of the seventeenth chapter. Nitti-Dolci 1972 [1938] and Alsdorf 1975 [1941] made corrections to the reading of the first edition of the Baroda text, which have not been taken into account in subsequent editions. My apparatus only refers to the readings of the second edition; that edition can be consulted for variants in the manuscripts of the Nāṭyaśāstra (of which there are an enormous amount).
• The Gāthālakṣaṇa (GāLa) of Nanditāḍhya (see Definition of the Gāthā in the bibliography). Date unknown; a quotation of a verse from Rājaśekhara, if it is not an interpolation, would put him after the tenth century.
• The Śvetāmbara commentarial literature, especially that of Jinadāsa (seventh century), Haribhadra (ca. eighth century) and Malayagiri (twelfth century) on the Nandisūtra, Anuyogadvārasūtra, Daśavaikālikasūtra, Āvaśyakasūtra, and Sūryaprajñapti. Fragments of Prakrit grammars in these texts were first noted by Upadhye 1931–1932.
• The Digambara commentarial literature, especially the Dhavalā of Vīrasena on the Ṣaṭkhaṇḍāgama of Puṣpadanta and Bhūtabali (completed in 816), and the Jayadhavalā (JaDha) of Vīrasena and Jinasena on the Kaṣāyaprābhṛta of Guṇabhadra (completed in 823). Most of the citations from these sources were noted by Jain 1945.
• Prakrit grammars, namely, the Prākṛtalakṣaṇa (PrāLa) ascribed to Caṇḍa (see Definition of Prakrit in the bibliography) and the Prākṛtasaṃjīvinī (PrāSaṃ) of Vasantarāja on Vararuci’s Prākṛtaprakāśa (see Light on Prakrit in the bibliography). Vasantarāja probably lived in the eleventh century (see chapter 7). The Prākṛtalakṣaṇa is more of a text tradition than a single text, and different manuscripts have different rules, examples, glosses, and so on.
1. Cited by Haribhadra in his Vṛtti to the Nandisūtra 74 (p. 57 l. 12); also in his commentary on the Daśavaikālikasūtra (only the second pāda) and Malayagiri’s commentary on the Nandisūtra (only the second pāda), the Āvaśyakasūtra (see Jain 1945 and Upadhye 1931–1932), and the Sūryaprajñapti (see Weber 1868: 273). Nitti-Dolci 1972 [1938]: §841 notes a different version of the same verse cited in the commentary to Prākṛtalakṣaṇa 2.13 (दुवयणे बहुवयणं चउत्थीविभत्तीए छट्ठी भण्णए । जह हत्था तह पाया वंदामि देवाहिदेवाणं ॥).
बहुवयणेण दुवयणं छट्ठिविभत्तीइ भण्णइ चउत्थी ।
जह हत्था तह पाया नमोत्थु देवाहिदेवाणं ॥
The plural replaces the dual, and the sixth case replaces the fourth case.
For example, “hands” and “feet,” and “reverence to the Jinas.”
2. Cited by Haribhadra in his Vṛtti to the Nandisūtra 51 (p. 28 l. 19).
ए होइ अयारंते पयम्मि बीयाऎ बहुसु पुल्लिंगे ।
तइयाइसु छट्ठी-सत्तमीण एक्कम्मि महिलत्थे ॥
E occurs at the end of a word whose stem ends in a in the masculine accusative plural and in the instrumental, genitive, and locative of the feminine singular.
3. Nāṭyaśāstra 17.6 = Gāthālakṣaṇa 4. Nitti-Dolci (1972 [1938]: §839) notes the close similarity to Prākṛtalakṣaṇa 2.10 (एदोद्रलोपा विसर्जनीयस्य).
एओआरपराइ अ अंकारपरं च पाअए णत्थि ।
वसआरमज्झिमाइ अ कचवग्ग-तवग्ग-णिहणाइं ॥
पराइ अ conj.; पराणिअ NāŚā, पराइं GāLa
अंकारपरं च GāLa (and Alsdorf); अं आरपरं अ NāŚā, अंआरपरं अ Nitti-Dolci
The sounds after e and o (i.e., ai and au),
as well as the sounds after anusvāra (i.e., visarga), do not exist in Prakrit.
Likewise the sounds between v and s (i.e., ś and ṣ)
and the final sounds in the velar, palatal and dental groups (i.e., ṅ, ñ and n).
4. Nāṭyaśāstra 17.7. Also cited in the Dhavalā (pādas ab) and the Jayadhavalā (pādas cd); see Jain (1945).
वच्चंति कगतदयवा लोवं अत्थं च से वहंति सरा ।
खघथधभा उण हत्तं उवेंति अत्थं च मुंचंता ॥
वच्चंति ... वहंति सरा] NāŚā; Dha reads वच्चंति कगतदयवा लोवं अत्थसरा; Jain emends to तिट्ठंति अट्ठ सरा
लोवं] JaDha; लोपं NāŚā
च] conj; अ NāŚā
The sounds k, g, t, d, y and v are lost, and the vowel that follows them bears their meaning.
The sounds kh, gh, th, dh, and bh become h and leave their meaning (?).
5. Nāṭyaśāstra 17.8.
उप्परहुत्तरआरो हेटा हुत्तो अ पाअए णत्थि ।
मोत्तूण भद्र-वोद्रह-रुद्र-ह्रद-चन्द्र-जाईसु ॥
हुत्तो] conj.; हुत्तौ NāŚā
रुद्र] Alsdorf; पद्र NāŚā and Nitti-Dolci
Whether it comes first or last, r as part of a consonant cluster does not exist in Prakrit.
Exceptions include words of the type bhadra-, vodraha-, rudra-, hrada-, and candra-.
6. Nāṭyaśāstra 17.9.
खघथधभाण हआरो मुह-मेह-कहा-वहू-पहूएसु ।
कगतदयवाण णिच्चं वीयम्मि ठिओ सरो होइ ॥
h replaces kh, gh, th, dh, and bh in words like muha- (mukha-),
meha- (megha-), kahā- (kathā-), vahū- (vadhū-) and pahū- (prabhu-).
The following vowel always stands in for the sounds k, g, t, d, y, and v after they disappear.
7. Malayagiri’s commentary to the Nandisūtra (the second half of a gāthā). Cited in Upadhye 1931–1932.
मतुवत्थम्मि मुणिज्जह आलं इल्लं मणं तह य ॥
Know that -ālaṃ, -illaṃ, and -maṇaṃ are possessive suffixes.
8. Vasantarāja, Prākṛtasaṃjīvinī on 4.34. I have restored the verse heavily; it is evidently a gāthā, but the latter half of the first line is very corrupt. Although this verse does not pertain directly to Prakrit grammar, it bears on the regional characterization of Prakrit.
मरहट्ठदेसभासाऎ [संकहिइ जो पसिद्धसोहग्गं] ।
सो तावेण ण सावइ कविअणचिरसाइअं भणिअं ॥
भासाऎ] conj.; भासाअ PrāSaṃ
संकहिइ जो पसिद्धसोहग्ग] conj.; संकइ जो राहिअ स्संदेहिइ जा गो हिअं पसिद्धं PrāSaṃ. I take संदेहिइ to be an explanatory gloss on संकहिइ. The rest of the pāda is jumbled and unmetrical.
तावेण] conj.; दावेण PrāSaṃ
सावइ; conj. metri causa; सावेइ PrāSaṃ
He who doubts the well-known beauty of the regional language of Mahārāṣṭra—does he not thereby curse the words that have been savored for so long by so many poets?
9. Prākṛtalakṣaṇa (manuscript C), commentary to 2.14; see Nitti-Dolci 1972 [1938]: §842. The verse describes the “root sounds” (mūlavaṇṇa-), that is, the phonological inventory of Sanskrit.
तेत्तीस विंजणाइं च सत्तवीसइ सरा तहा भणिया ।
चत्तारि य जोगवहा चउसट्ठी मूलवण्णा ॥
च सत्तवीसइ] conj. metri causa; सत्तवीस PrāLa
Thirty-three consonants, twenty-seven vowels,
and four combining sounds makes sixty-four root sounds.
10. Dhavalā 9: 95 (only the last half); Jayadhavalā (see Jain 1945).
कीरइ पयाण काण वि आई-मज्झंत-वण्ण-सर-लोव ।
Some words undergo an elision of an initial, medial or final consonant or vowel.
11. The first few words are cited widely: by Jinadāsa (Anuyogadvārasūtra-cūrṇi, p. 128), by Haribhadra (Anuyogadvāra-vivṛti, p. 187), by Vīrasena (Dhavalā, vol. 8, p. 90; vol. 9, p. 95; vol. 10, p. 2; vol. 13, pp. 243 and 337). The complete verse is cited only in the Jayadhavalā (see Jain 1945). Since it allows for the substitution of any vowel by any other vowel, it must have been very useful for exegetical purposes.
एए छच्च समाणा दोण्णि अ संझक्खरा सरा अट्ठ ।
अण्णोण्णस्स विरोहा उवेंति सव्वे समाएसं ॥
The eight vowels—these six simple vowels and two compound vowels—
come in place of each other without any restraint (so Jain).
12. Jayadhavalā (see Jain 1945).
दीसंति दोण्णि वण्णा संजुत्ता अह व तिण्णि चत्तारि ।
ताणं दुव्वल-लोवं काऊण कमो पजुत्तव्वो ॥
When two, letters are joined, or three, or four,
elide the weakest of them, and continue the process.
13. Jayadhavalā (see Jain 1945). This transforms voiceless into voiced sounds, which is relatively rare except in Jain texts and in (in the limited context of t to d) in Prakrit used on the stage. As the verse currently stands it is an upagīti/gātha (both halves have just one light syllable in their sixth gaṇa).
वग्गे वग्गे आई अविट्ठया दोण्णि जे वण्णा ।
ते णेयय णिय वग्गे तइअत्तणयं उवणमंति ॥
In every class the two letters that stand at the beginning
are variously changed to the third letter of that class.
TESTIMONIA
1. Vṛttajātisamuccaya 2.8–9. Note that the commentator Gopāla notes that “according to some people Vṛddhakavi is Harivṛddha” (vṛddhakavir harivṛddha iti kecit).
भुअआहिवसालाहणवुड्ढकइनिरूविअं इमं दइए ।
णिहणणिरूपविअधुवअम्मि वत्थुए गीइआ णत्थि ॥
In the opinion of Bhujagādhipa, Sātavāhana, and Vṛddhakavi,
when a strophic vastuka features a dhruvakā in its definition, there is no need for a gītikā.
भुअआहिवसालाहणवुड्ढकइनिरूविआण दुवईण ।
णामाइं जाइं साहेमि तुज्झ ताइं विअ कमेण ॥
I will tell you in sequence all the names for the dvipadas
defined by Bhujagādhipa, Sātavāhana, and Vṛddhakavi.
2. Sarasvatīkaṇṭhābharaṇa 1.99 (ex. 133), p. 93 = Śṛṅgāraprakāśa 9.266, p. 507.
अम्हारिसा वि कइणो हलिवुड्ढहालपमुहा वि ।
मण्डुक्क-मक्कडा वि हु होंति हरी सप्पसिंहा वि ॥
People like me are poets
Just as much as Harivṛddha and Hāla.
Don’t we call frogs and monkeys hari,
besides snakes and lions?
3. Karpūramañjarī pp. 9–10 (ed. Ghosh). The vidūṣaka complains about the servant girl Vicakṣaṇā.
विदूषकः । [सक्रोधम्] ता उज्जुअं ज्जेव किं ण भण्णइ अम्हाणं चेडिआ हरिउड्ढ-णन्दिउड्ढ-पोट्टिस-हाल-प्पहदीणं पि पुरदो सुकइ त्ति ।
हरिउड्ढ ... प्पहदीणं] Konow lists many variants on these names, but the most significant is हरि-बम्हसिद्धि-ओड्डीस-पालित्तअ-चंपअराअ-मल्लसेहराणं, read by witnesses STU.
VIDŪṢAKA: [Angrily.] Well, why don’t you come right out and say it? That this servant girl of ours is a better poet than even Harivṛddha, Nandivṛddha, Poṭṭisa, and Hāla?