1986: Facets of morphology

1986

MADHAV M. DESHPANDE

SOME FACETS OF PĀŅINIAN MORPHOLOGY

1. Basic Morphological Categories

Pāṇini’s - Aştādhyāyi makes a clear terminological distinction between prātipadika ’nominal stem’, ‘dhātu ‘verb root’ and pratyaya ‘affix’. It can be shown that these three are most probably the ultimate three classes of Pāņini’s morphology. The preliminary proof of such a proposition may be presented as follows.

Pāņini defines dhātu ‘verb root’ by referring to his list of these items, viz. P.1.3.1 (bhūvādayo dhātavaḥ) which basically says: ‘bhū etc. (as they are listed in the appendix called Dhātupātha) are given the designation dhātu.1

The tradition of commentators reads into this rule an additional condition that these items must be kriyā-vācinaḥ denoting action. It is not clear whether Pāṇini himself expressly intended such a condition and it is not explicitly stated in this rule. It seems certain that he did not believe that a semantic definition could function very well, and therefore decided to offer a complete listing in this Dhātupātha.

The next class of items is pratyaya ‘affix’. P. 3. 1. 1. (pratyayaḥ) says that items, which are prescribed by rules beginning with P.3.1.1. and upto the end of the fifth chapter of the Astādhyāyi are termed pratyaya. It again seems that he did not offer a semantic definition, but a definition by listing. This is a different kind of listing compared to the Dhātupātha, and yet it is a kind of listing nonetheless.

We can show that the universe of Pāṇini’s morphology was divided into three exclusive domains, two of which were defined in terms of some kind of listing. How can one define the remaining domain of prātipadika ‘nominal stem’? If the universe of morphology is a closed universe, or at least if two of its three domains are closed domains, then the third domain can simply be defined by exclusion. This is what Pāņini does in rule P. 1.2. 45 (arthavad adhātur apratyayaḥ prātipadikam): An item which is meaningful and (yet) is neither a verb root nor an affix is termed prātipadika.’

This is the definition of the primary (=avyutpanna ‘underived’) nominal stems. The derived nominal stems which are combinations of previously defined primary items are defined separately (Ref. P.1.2.46.). For the present purpose, P.1.2.45 is a very crucial rule. It shows that the only primary items of Pāņini’s morphology are dhātu, pratyaya and prātipadika.

If this is true, then what about the categories of vikaraṇa and ägama? The term vikaraṇa as used in the tradition of post-Pāṇinian commentators refers to con jugation-marking items such as -&- in a form like bhavati derived from the sequence bhū+a+ti. The term agama which is translated as ‘augment’ refers to items such as the past-tense marker a- in a form such as a-gacchat ‘he went’. These two terms go back to Kätyāyana and Patanjali in the tradition of Pāṇinian commentators. However, these terms are not found in the Astådhyayi itself. this raises some very interesting questions concerning the status of these items in the original system of Panini. Since the commentatorial tradition beginning with Kätyāyana makes use of these terms, without explicitly defining them, one may consider that these terms were already well established in older non-Pāṇinian systems and that the Pāṇinian commentators simply took over their usage. However, since Pānini does not explicitly define or use these terms, it becomes important for us to try to figure out the exact position of these items in Pāņini’s original system. To put the main question more succinctly: Are vikarana-s and agama-s subsumed in the larger category of pratyaya-s ‘affixes’ in Panini’s grammar? Since they are most obviously neither verb roots nor nominal stems, there is no need to question the status of vikarana-s and agama-s in relation to these categories.

1 For instance, the Mahabhāsya, Vol. I, p. 14, separately discusses the categories of agama, vikara, pratyaya, dhatu and prdfipadika. The term vikarana is also found in the Värttika-s and in the Mahabhāsya.

2. The Status of Vikarana-

From all indications, it seems clear that the category of vikarana is and must be subsumed under the larger category of pratyaya. We may consider the following evidence.

2.1. The major rules prescribing items which are later called vikarana are as follows:

  • P.3.1.68 (kartari sap): ‘When a root is followed by a särvadhätuka affix denoting the agent, then the root is followed by SaP.’
  • P.3.1.69 divādibhyah syan): ‘Under the same conditions as above: SyaN occurs after roots div etc.’
  • P.3.1.73 (svādibhyaḥ śnuh): ‘Snu occurs after roots su etc.
  • P.3.1.77 (tudādibhyah sah): Sa occurs after roots tud etc.
  • P.3.1.78 (rudhädibhyaḥ śnam): SnaM’ occurs in the case of roots rudh etc.?
  • P.3.1.79 (tanādi-krñbhya uh) ’ occurs after roots tan etc. and krñ etc.’
  • P.3.1.81 (kryādibhyaḥ śnā) *Snā occurs after roots kri etc.

All these rules occur in the section governed by P.3.1.1 (pratyayaḥ), and hence the items prescribed by them are eligible for the designation pratyaya.

2.2. Among the items prescribed by the rules given above, a large number begins with the diacritic S. This initial S can be classified as a diacritic only by P.1.3.8 (lasaku ataddhite). With the words ādiņ and pratyaya sya continued from P.1.3.5 and P.1.3.6, respectively, P.1.3.8 means: ?, s and consonants of the K-series are diacritics if they occur at the beginning of non-taddhita affixes. This is the only rule which can make an initial diacritic, and for this the element must be an affic

2.3. In the derivation of forms such as bhavati and tudati, the affixal status of śap in bhū+śaP+tip and tud+śa+tiP is absolutely necessary. P.7.3.84 (sārvadhātukārdhadhātukayoḥ, gunaḥ from 7.3.82) says that i, u, r, and I vowels at the end of the root are replaced by their corresponding guna vowels, if followed by a dhātuka or an ārdhadhātuka affix. For bhū in the sequence bhū+śap to become bho and later bhav, it is important that śap be a sārvadhătuka or an ārdhadhātuka affix. P 34 113 (tiṅ-sit sārvadhātukam) says that all finite verb endings (tiṅ) and affixes marked with the diacritic Ś are sārvadhātuka affixes. P.3.4.114 (ārdhadhātukaṁ seșaḥ) says that the rest of the affixes prescribed after roots are ārdhadhătuka affixes. śaP and śa are both sārvadhātuka affixes. While the affix śap has the diacritic P, the affix śa does not have it. P.1.2.4 (sārvadhātukam a-pit) says that a sārvadhātuka affix which does not have the diacritic P is treated like an affix with a diacritic ṅ. P.1.1.5 (kniti ca) says that the vowel of the base does not undergo gunu and vrddhi substitutions, if it is followed by an affix marked with a diacritic K or ṅ. Therefore, the base tud does not take a guna substitution before śa in the form tudati. For this whole process to work, śap and śa must have the designation pratyaya. Even for the term anga ‘base’ to apply to roots before these items, these items must be pratyayas (Ref: P.1.4.13 (yasmāt pratyaya vidhis tadādi pratyaye’ṅgam).)

2.4. By P.2.4.72 (adiprabhịtibhyaḥ sapaḥ) the vikaraṇa śap is substituted by the zero substitute LUK after roots ad etc. By P.2.4.75 ( juhotrādibhyaḥ śluh), śap is substituted by the zero substitute ŚLU. P.1.1.16 (pratyayasya luk-ślu-lupaḥ) says that the deletion of a pratyaya is termed LUK, SLU and LUP. Thus the fact that P.2.4.72 and P.2.4.75 prescribe substitution of śap by LUK and ŚLU is itself an indication that śap is a pratyaya.

3. The Status of Āgama-s

3.1. The category of āgama-s is somewhat more difficult to define within the Pāṇinian framework. As we will see, there are many loose ends regarding this category. Typically according to the tradition, those items which have the diacritics ṭ, K and M are considered to be āgama.2 Patañjali often talks about the categories of pratyaya, ādeśa, ṭit, kit and mit.3 This would imply that for Patañjali these were exclusive categories.

As augments, the āgama-s do not replace another item, and hence they are certainly not ādeśa-s ‘substitutes’. The question of whether āgama-s can be subsumed under the basic morphological category of pratyaya is a difficult question and needs to be analysed carefully.

3.2. For his part, Patañjali assumes that āgama-s are not pratyaya-s. For instance, P.1.1.69 (and savarṇasya cāpratyayaḥ) says that non-affixal a-N sound and sounds with a diacritic U represent the whole sbare of their homogeneous sounds. How about the augment is prescribed by P.7.2.35 (ārdhadhātukasyed valādeḥ) ? Does it represent the whole class of i-sounds ? If this is is a pratyaya, this question can be automatically answered in the negative. But Patañjali does not treat is to be a pratyaya, and hence it is not automatically prevented from representing the class of its homogeneous sounds. For that purpose, he needs to find a clever jñāpaka ‘indication’. I have discussed this jñāpaka elsewhere. 4 Suffice it to say that this needs further analysis.

3.3. Let us view the evidence in Pāṇini which seems to suggest that at least some of the agama-s have affixal features.

3.3.1. P.3.1.78 (rudhādibhyaḥ śnam) prescribes śnam as the substitute for śap in the case of root rudh etc. We have already seen that the vikaraṇa śap must be categorized as a pratyaya in Pāņini’s system, and hence a substitute for a pratyaya must also be a pratyaya by definition. However, why then does it have the diacritic M? For that we must look at P.1.1.47 (mid aco’ntyāt paraḥ). This rule says that an item with the diacritic M is appended after the last vowel of the item for which it is prescribed. For instance, without the dia critic M, we would have had the sequence rudha+Snat ti. However, with the diacritic M, we now get ru+śnam dh+ti. Items with diacritics ṭ and K respectively are appended to the front and the end of the item for which they are prescribed (Ref. P.1.1.46 (ādy-antau ta-kitau). The Kāśikāvịtti on these rules says that these rules represent an exception to the substitute replacing an original item and an affix occurring after the item for which it is prescribed.5 This would imply that potentially the so-called agama-s are liable to be treated either as substitutes or as affixes, and that these specific rules put some restrictions on their behaviour. But the same Kāśikāvrtti on P.3.1.1. (pratyayaḥ) says that the domain of pratyaya excludes āgama-s.6 At least śnam is a clear example of a pratyaya which has so-to-say an agama diacritic. On P.3.1.78 (rudhādibhyaḥ śnam), the Kāśikāvrtti clearly admits that śnam is a pratyaya and that it still has the diacritic M for its specific placement.7

3.3.2. As we have seen above the absence or presence of the diacritic F is a typical affixal feature and so also the diacritics K and Ň to prevent guna and vrddhi for the base-vowels (Ref: P.1.2.4. (sārvadhătukam a-pit)) P.1.1.5 (kriti ca) and P.3.1.4 (anudāttau sup-pitau). However, in a number of cases, what are traditionally regarded as āgama-s, have these affixal diacritics in Pāṇini’s rules.

486

This does raise a question whether these are pratyaya-s of some sort. Consider the the following examples:

A: P.3.4.92 (ād uttamasya pic ca) prescribe the item as for the first person imperative affixes karavāni. The normal syntax of Sanskrit would map us interpret this rule to mean that it is aT which o becomes pit ‘has the diacritic P’.

B: P.3.4.103 (yāsut parasmaipadeșūdātto ricca says that the optative active affixes get the augment yās Uṭ which is both high-pitch and ñit has a diacritic. This is the normal syntax of the rule. However, given the belief that āgama-s are not pratyaya-s, the Käsikäs vrtti argues that this property of becoming ñit actually applies to the optaṭive ending and not to the augment yāsUṭ, because it is of no use for an agama. barch

C: P.3.4.104 (kid āśiși) says that in the case of āšīrlin, the augment yās Uṭ is treated as a kit has the diacritic K’. Even on this rule, the Kāśikāvrtti feels the same difficulty and aruges that actually it is the final affix which become kit, and not the āgama, because a kit agama would serve no purpose.2

D: P.6.1.187 (ādiḥ sico’nyatarasyām) deals with sic which is traditionally considered to be an agama. But the diacritic G is a typical affixal diacritic and functions according to P.6.1.163 (citaḥ) which makes the whole word antodātta ‘with a high-pitch on the final vowel’. The Kāšikāvrtti treats this as an exceptional case for an agama which is supposed to be unaccented. 8

E: P.6.3.68 (ica ekāco’mpratyayavacca) is an interesting rule, but has a somewhat doubtful reading. This rule is supposed to prescribe, according to the tradition, the agama am where the m is not a diacritic, but at the same time this am is said to behave like a pratyaya. The tradition repeats the word am twice and says that the am augment is to behave like the affix -am. This interpretation evades the question whether it has any inherent affixal character. However, if the rule is read to mean that am in this particular case behaves like a pratyaya, and if am is really an agama, then it would mean that āgama-s generally do not have any affixal character. However, this rule is too debat able to shed any light on the larger question.9

  1. Conclusion

In conclusion, it may be stated that the above discussion shows that the category of vikaraña is clearly subsumed under Pāņini’s basic morphological category of pratyaya. On the other hand, the category of what is known as āgama-s from the post-Pāṇinian literature remains somewhat unclear. Many of these āgama-s seem

488

THE ADYAR LIBRARY BULLETIN

ubsumed directly

to have at least some affixal features, though it is totally clear whether they can be subsumed di under the category of pratyaya. The tradition tren āgama-s as a dependent category. They are viewed being fully dependent upon the item to which they on appended, and are subsumed under that catagory 1 However, they remain a mixed bag of miscellaneorie items whose only common feature is that they are appended to some other item which seems to be more primary in the derivation. The main difficulty is that there is no definition of an āgama in the Astādhyāyī and hence the identity of these items is not immediately obvious without the help of commentatorial interpre tation which sometimes may stray away from Pānini’s original intentions. It is hoped that futher research will shed further light on this topic.

BIBLIOGRAPHY

Deshpande, Madhav M, 1972. ‘Pāṇinian Procedure

of Taparakaraña: A historical investigation’. Zeitschrift für vergleichende Sprachforschung, Band

86, Heft 2. Kāśikāvrtti by Vāmana and Jayāditya, with the com

mentaries Nyāsa by Jinendrabuddhi and Padamañ jari by Haradatta. Six Volumes. Prācya Bhārati Series 2. Tara Publications. Banaras, 1965-67.

1 See Paribhāsendusekhara for a discussion of the Paribhāṣā gadāgamās tadgunibhūtās tad grahenena grhyante, and also the Mahābhāsya on P.1.1.20, Vol. I, p.75: ādeśās tarhime bhavisyanty anā gamakanam sāgamakāh/SOME FACETS OF PĀŅINIAN MORPHOLOGY

489

Mahābhāsya by Patañjali. Edited by Franz Kielhorn.

Revised edition by K.V. Abhyankar. Vol. I, 1962; Vol. II, 1965; Vol. III, 1972. Bhandarkar Oriental

Research Institute, Poona. Paribhāṣendusekhara by Nāgeśabhatta. With the com

mentary Tattvādarśa by Mahamahopadhyaya Vasudeva Shastri Abhyankar. Edited by K. V. Abhyanakar. Revised edn. Bhandarkar Oriental Research Institute, Poona.

  1. See: Mahābhāsya on P.1.3.1., Vol. I. pp. 253ff. The later standard works such as the Käsikāvr̥tti also include this condition in their explanation of this rule. ↩︎

  2. However, it must be kept in mind that there is no automatic way of distinguishing between pratyaya-s with these diacritics and agama-s with these diacrtics. This has to be decided by looking at the context. ↩︎

  3. See the Mahābhāṣya on P.1.1.69, Vol. I, p. 177. ↩︎

  4. Deshpande 1972, pp. 214-5, fn. 20. ↩︎

  5. sthāneyoga-pratyayaparatvasyāyam apavādah, Kāśikāvrtti on P. 1.1.47. ↩︎

  6. pañcamādhyāyaparisamāpter yān ita urdhvam anukramisyāmah pratyayasanjñās te veditavyāḥ, prakrty-upapadopādhi-vikārāgamān varja vitvā, Kāśikāvịtti on P.3.1.1. ↩︎

  7. śnam-pratyayo bhavati , sapo ‘pavadah makāro deśa-vidhya-arthah, Kāśikāvýtti on p.3.1.78. ↩︎

  8. The Kāśikāvrtti says on P.6.1.187: sicaś citkaraņād āgamānudāt tatua hi bādhyate ↩︎

  9. For details, see the Mahābhāsya and other commentaries on P.6.3.68. ↩︎