4 2 Examples of Application of 1 4 13 and 6 4 1

Now I will discuss some examples through which I will show that my interpretation of 1.4.13 and 6.4.1 alone can help us derive the correct final form. But first, let me offer a clarification.

In many of the examples discussed in this chapter, the derivation should begin with: verbal base + lakāra. At this stage, there are two possibilities:

(i) only one rule, i.e., the rule which teaches the replacement of the lakāra is applicable, and this rule applies.

(ii) multiple rules, including the rule which teaches the replacement of the lakāra are applicable.

( M») verbal base ( N») + ( O») lakāra

The rule O, which teaches the replacement of the lakāra is the right-most. Thus, by my interpretation of 1.4.2, we apply rule O.

Note that, in both cases (i) and (ii), the rule that replaces the lakāra applies at the first step. So, in order to simplify the presentation, in all the examples where the derivation should start with verbal base + lakāra, I simply start it with verbal base + tiṄ (one of the eighteen finite replacements of the lakāras) instead. For instance, in the first derivation discussed in this section, technically the derivation should proceed as follows: edh + LAṬ (3.2.123 vartamāne laṭ) 🡪 edh + jha (3.4.77 lasya, 3.4.78 tiptasjhi…). However, I start with edh + jha, purely for the purpose of avoiding redundancy.

[[112]]

edhante

(1) edh + jha – ‘to grow’, present third-person plural1

As stated in section 4.1 of this chapter, we cannot call edh an aṅga with respect to jha. Consequently, at this step, rules taught in the aṅgādhikāra (6.4 – 7.4), such as 7.1.3 jho’ntaḥ (which teaches that a jh which is the initial sound an affix is replaced with ant) or 7.1.5 ātmanepadeṣv anataḥ (which teaches that a jh, which is the initial sound of an ātmanepada affix preceded by a verbal base that does not end in a, is replaced with at) cannot apply to jh.

Here, only two rules are applicable:

edh ( 3.1.68»)+ jh ( 3.4.79») a

3.1.68 kartari śap: same as above.

3.4.79 ṭita ātmanepadānāṁ ṭer e: the part that begins with the last vowel (ṭi)[^9] of an ātmanepada replacement of a lakāra marked with Ṭ is replaced with e.

By my interpretation of 1.4.2, we apply the RHS rule 3.4.79 and get: edh + jhe.

At this stage too, we cannot call edh an aṅga with respect to jhe. Thus, 7.1.3 and 7.1.5 are not applicable here. Only one rule, namely 3.1.68 is applicable. Upon applying it, we get edh + ŚaP + jhe. At this step, edh and ŚaP cannot undergo any further operations which are not triggered by jhe, so we can simply write edh + ŚaP as edha. edha is an aṅga with respect to jhe. At this stage, of the two aforementioned rules which belong to the aṅgādhikāra, 7.1.3 is applicable but 7.1.5 is not. We apply 7.1.3, and get edha + ante. By 6.1.97 ato guṇe2, we get the correct form: edhante.

113

Error path

To the best of my knowledge, the tradition does not discuss this example. However, let us consider what would have happened if we had not accepted my interpretations of 1.4.13 and 6.4.1 respectively. At the step edh + jha, four rules would become applicable:

edh ( 3.1.68»)+ (7.1.3 7.1.5») jh (3.4.79» ) a

Note that all the DOI relationships here are of Type 2b (DOI non-conflict). As stated before, the tradition is not interested in non-conflict and mostly applies rules in a haphazard order in such cases.

There is an SOI between 7.1.3 jho’ntaḥ and 7.1.5 ātmanepadeṣv anataḥ. 7.1.5 is more specific than and thus wins against 7.1.3 jho’ntaḥ. If the tradition applies 7.1.5, which replaces jh with at first and applies 3.1.68 kartari śap at a later step, that gives *edhate, which is not the correct form.

Conclusion

My interpretations of 1.4.13 yasmāt pratyayavidhis tadādi pratyaye’ṅgam and 6.4.1 aṅgasya respectively ensure that jh replacement, which is taught in the aṅgādhikāra, takes place only after the application of 3.1.68 kartari śap, which is taught outside the aṅgādhikāra. After the application of 3.1.68, 7.1.5 ātmanepadeṣv anataḥ, which is an exception of 7.1.3 jho’ntaḥ, is no longer applicable to jh, and thus 7.1.3 jho’ntaḥ applies to jh. This gives the correct form, edhante.

dadhati

(2) dhā + jhi – ‘to place’, present third-person plural

As stated before, dhā cannot be called an aṅga with respect to jhi. Consequently, rules taught in the aṅgādhikāra (6.4 – 7.4) cannot apply to dhā or jhi. For example, 7.1.3 jho’ntaḥ cannot apply here. The derivation proceeds as follows:

dhā + ŚaP + jhi (3.1.68 kartari śap) 🡪
dhā + ŚLU + jhi (2.4.75 juhotyādibhyaḥ śluḥ3) 🡪
dhādhā + ŚLU + jhi (6.1.10 ślau4) 🡪
dhadhā + ŚLU + jhi (7.4.59 hrasvaḥ5).

At this point, we notice that dhadhā and ŚLU cannot undergo any other operations which are not triggered by jhi. So, we can write dhadhā + ŚLU as dhadhā.

[[114]]

In dhadhā + jhi, dhadhā can be called an aṅga with respect to jhi. Therefore, the following rules from the aṅgādhikāra are applicable here:

dh a dh (6.4.112»)ā + (7.1.3, 7.1.4») jh i

  • 6.4.112 śnābhyastayor ātaḥ: the ā of the affix Śnā or the ā at the end of a reduplicated verbal base is replaced with LOPA when a sārvadhātuka affix marked with K or Ṅ follows.
  • 7.1.3 jho’ntaḥ: a jh which is the initial sound an affix is replaced with ant.
  • 7.1.4 ad abhyastāt: when preceded by a reduplicated base, a jh which is the initial sound an affix is replaced with at.

By my interpretation of 1.4.2, we perform the RHS operation. But there is an SOI between 7.1.3 and 7.1.4, both of which apply to the RHS operand. Because 7.1.4 has been taught for jh when it is preceded by a reduplicated base, it is more specific and wins. Thus, we get: dhadhā + ati. Here 6.4.112 śnābhyastayor ātaḥ applies and we get dhadh + ati. Now that all the possible rules from the sapādasaptādhyāyī have applied, a rule from the tripādī section applies6:

8.4.54 abhyāse car ca: in an abhyāsa (first of two reduplicated syllables), jhaL (a non-nasal stop or a fricative) is replaced with caR (c, ṭ, t, k, p, ś, ṣ, s) or jaŚ (j, b, g, ḍ, d).

Thus, we get dhadhati 🡪 dadhati, which is the correct answer.

Error path

Let us now look at how the tradition tackles this problem. Like in the previous example, in this example too, there are no cases of DOI conflict, and so the tradition chooses to apply rules in a random order. But some sequences of rule application can give the wrong answer. For example: dhā + jhi 🡪 dhā + ŚaP + jhi (3.1.68) 🡪 dhā + ŚaP + anti (7.1.3) 🡪 dhā + ŚLU + anti (2.4.75) 🡪 *dadhanti (6.1.10 ślau etc.). In sum, if jh undergoes replacement before the reduplication of dhā, we get the wrong answer.

115

Traditional fix

To address this issue, the tradition has come up with the following ideas. Consider paribhāṣās 62 and 63 of the Paribhāṣenduśekhara and their translation by Kielhorn:

pūrvaṁ hy apavādā abhiniviśante paścād utsargāḥ (62).

‘apavādas, it is certain, are considered first (in order to find out where they apply); afterwards the general rules (are made to take effect in all cases to which it has thus been ascertained that the apavādas do not apply)’.

prakalpya vāpavādaviṣayaṁ tata utsargo’bhiniviśate (63).7

‘Or (we may say that) first all (forms) which fall under the apavāda are set aside, and that subsequently the general rule is employed (in the formation of the remaining forms).”

Let us see what happens if we follow these paribhāṣās at the first step (dhā + jhi). At this step, 7.1.4 ad abhyastāt, which is the apavāda of 7.1.3 jho’ntaḥ, is not applicable. Since the apavāda is not applicable, we go ahead and apply the utsarga 7.1.3. But this gives us the wrong form *dadhanti. Taking cognizance of this problem, the tradition has come up with the following metarule:

upasaṁjaniṣyamāṇanimitto’py apavāda upasaṁjātanimittam apy utsargaṁ bādhata iti (64).

‘An apavāda supersedes, even though the causes of its (application) are still to present themselves, a general rule the causes (of the application) of which are already present.’

In other words, this paribhāṣā teaches that even though 7.1.3 is applicable to jh from the beginning of the derivation, one must not replace jh until the apavāda 7.1.4 becomes applicable. This gives the correct answer, dadhati.

As stated in the first chapter, the tradition often comes up with a new paribhāṣā to address individual problems like this one. Paribhāṣā 64 is a good case in point.

Conclusion

My method ensures that the replacement of jha, which is taught in the aṅgādhikāra, takes place after the reduplication of dhā, which is taught outside the aṅgādhikāra. Therefore, 7.1.3 jho’ntaḥ does not become applicable until 7.1.4 ad abhyastāt, its exception, also becomes applicable. 7.1.4 wins, thereby giving the correct form dadhati.

[[116]]

My method is able to tackle this issue without relying on paribhāṣās like Pbh 64, which require us to look ahead into the derivation.

Before we move on to discussing other examples, note that Pāṇini teaches most substitutions and other operations pertaining to the eighteen finite affixes from 3.4.77 to 3.4.112. For example, 3.4.87 ser hy apic ca8, 3.4.101 tasthasthamipāṁ tāṁtaṁtāmaḥ9 etc. He teaches the substitution of jhi from 3.4.108 jher jus to 3.4.112 dviṣaś ca. However, the three rules teaching the replacement of jh, i.e., 7.1.3 jho’ntaḥ, 7.1.4 ad abhyastāt, and 7.1.5 ātmanepadeṣv anataḥ are found in the aṅgādhikāra, and not in the section 3.4.77-3.4.112. This strongly suggests that Pāṇini wants us to treat 7.1.3-7.1.5 differently, that is, he wants us to apply them only when jh is part of an affix which is preceded by what I define as an aṅga.

jahitaḥ

(3) hā + tas – ‘to abandon’, present third-person dual

hā is not an aṅga with respect to tas. So here, we cannot apply rules from the aṅgādhikāra, such as 6.4.116 jahāteś ca (see translation below). The derivation proceeds as follows:

hā + tas 🡪
hā + ŚaP + tas (3.1.68 kartari śap) 🡪
hā + ŚLU + tas (2.4.75 juhotyādibhyaḥ śluḥ) 🡪
hāhā + ŚLU + tas (6.1.10 ślau).

Here, two rules are applicable, which are from the aṅgādhikāra, but which are not triggered by tas:

(7.4.62→) h (7.4.59→) ā hā + ŚLU + tas

7.4.62 kuhoś cuḥ: a consonant of the k-series (kU), or a h, that is part of the abhyāsa (first of two reduplicated syllables) is replaced with a consonant of the c-series (cU).

7.4.59 hrasvaḥ: the vowel of the abhyāsa (first of two reduplicated syllables) is replaced with its short counterpart.

By my interpretation of 1.4.2, we apply the RHS rule 7.4.59 and get hahā + ŚLU + tas. To this, we apply 7.4.62 and get jhahā + ŚLU + tas. Now, jhahā and ŚLU cannot undergo any further operations which are not triggered by tas, so we can write jhahā + ŚLU as jhahā. Now,

[[117]]

jhahā is an aṅga with respect to tas. Thus, the following rules from the aṅgādhikāra, which are triggered by tas, become applicable:

jhah (6.4.113, 6.4.116→) ā + tas

6.4.113 ī haly aghoḥ: the final ā of a base which ends in Śnā, or of a reduplicated stem (abhyasta) excluding those termed ghu, is replaced with ī when a sārvadhātuka affix beginning with a consonant and marked with K or Ṅ follows.

6.4.116 jahāteś ca: the final ā of hā ’to abandon’, is optionally replaced with i, when a sārvadhātuka affix beginning with a consonant and marked with K or Ṅ follows.

There is an SOI relationship between 6.4.113 and 6.4.116. Since 6.4.116 has been taught specifically for hā, it wins, as a result of which we get jhahitas. Finally, since all rules from the sapādasaptādhyāyī have been applied, we apply 8.4.54 abhyāse car ca from the tripādī and get jhahitaḥ 🡪 jahitaḥ, which is the correct answer.

Note that 6.4.116 jahāteś ca is an optional rule. If we do not implement18 it, 6.4.113 ī haly aghoḥ applies, giving us jahītaḥ, which is also correct.

Wrong path

To the best of my knowledge, the tradition has not discussed this problem. But, since this derivation does not involve any DOI conflicts, the tradition would have applied rules in any haphazard order. Let us look at one of the possible paths this derivation would have taken if we had not accepted my interpretations of 1.4.13 and 6.4.1 respectively:

hā + tas 🡪
hā + ŚaP + tas (3.1.68 kartari śap) 🡪
hā + ŚLU + tas (2.4.75 juhotyādibhyaḥ śluḥ) 🡪
hi + ŚLU + tas (6.4.116 jahāteś ca) 🡪
*jihitaḥ (6.1.10 ślau etc).

The possibility of getting such a wrong answer is completely eliminated by following my interpretations of 1.4.13 and 6.4.1 respectively. This is because, my method ensures that 6.4.116, which is taught in the aṅgādhikāra and replaces ā of hā with i, applies only after the reduplication of root hā by 6.1.10 ślau, which is taught outside the aṅgādhikāra.

118

aupyata

(4) vap + ta – ‘to sow’, imperfect passive third-person singular

Note that vap is not an aṅga with respect to ta, so rules like 6.4.71 luṅlaṅlr̥ṅṣv aḍ udāttaḥ (see translation below) which are part of the aṅgādhikāra, cannot apply at this step. The following rules are applicable to vap + ta:

(6.1.15») v ap (3.1.67») + ta

6.1.15 vacisvapiyajādīnām ca: roots vac ‘to speak’, svap ‘to sleep’, and those headed by yaj ‘to perform sacrifice’ undergo samprasāraṇa when an affix marked with K follows.19

3.1.67 sārvadhātuke yak: affix yaK occurs after a verbal root when a sārvadhātuka affix which denotes bhāva or karman follows.

By my interpretation of 1.4.2, the RHS rule 3.1.67 applies, and we get: vap + yaK + ta. Thereafter the derivation proceeds as follows:

vap + yaK + ta 🡪
uap + yaK + ta (6.1.15) 🡪
up + yaK + ta (6.1.108 samprasāraṇāc ca20).

Since up and yaK cannot undergo any other operations which are not triggered by ta, we can write up + yaK as upya. In upya + ta, upya is an aṅga with respect to ta. Thus, the following rules from the aṅgādhikāra which are triggered by ta become applicable:

(6.4.71, 6.4.72») upya + ta

6.4.71 luṅlaṅlr̥ṅṣv aḍ udāttaḥ: the udātta ‘high-pitched’ augment aṬ is attached to a verbal base when affixes LUṄ, LAṄ and LR̥Ṅ follow.

6.4.72 āḍ ajādīnām: the udāttaḥ ‘high-pitched’ augment āṬ is attached to a verbal base which begins with a vowel (aC) when affixes LUṄ, LAṄ and LR̥Ṅ follow.

[[119]]

This is a case of SOI. 6.4.72 has been taught specifically for bases which begin with a vowel and thus wins, thereby giving us the correct form ā-upya + ta 🡪 aupyata (6.1.90 āṭaś ca10).

Wrong path

Let us now consider how the tradition deals with this example. Like in the previous examples, here too, we do not find any instances of DOI conflict. Therefore, the tradition applies rules in a random order. If the attachment of the augment had been undertaken before samprasāraṇa, we would have got the incorrect form:

a-vapyata (6.4.71 luṅlaṅlr̥ṅṣv aḍ udāttaḥ) 🡪
a-uapyata (6.1.15 vacisvapiyajādīnām ca) 🡪
*opyata (6.1.108 samprasāraṇāc ca, 6.1.78 eco’yavāyāvaḥ).

In order to overcome this problem, the Kāśikā, on 6.4.72 āḍ ajādīnām, suggests that there is a conflict between augment addition and processes such as replacement of LAṄ and introduction of the vikaraṇa yaK, and by nityatva and antaraṅgatva respectively these two processes defeat the addition of the augment aṬ.11

We may conclude that the tradition comes up with a tailored solution to this problem. In contrast with this, my method eliminates the need to rely on post-Pāṇinian tools and paribhāṣās. My respective interpretations of 1.4.13 and 6.4.1 ensure that the addition of the augment, which is taught in the aṅgādhikāra, takes place only after samprasāraṇa, which is taught outside the aṅgādhikāra. As a result of this, 6.4.71 luṅlaṅlr̥ṅṣv aḍ udāttaḥ does not become applicable until 6.4.72 āḍ ajādīnām, which is its exception, also becomes applicable. 6.4.72 wins, thereby giving the correct form aupyata.

Conclusion

In sum, these four examples prove that my interpretations of 1.4.13 and 6.4.1 respectively are correct. In all four derivations, the tradition applies rules in a haphazard order, as a result of which it often gets the wrong form at the end of the derivation. It is forced to come up with individual solutions for each of these problems.

[[120]]

It is also noteworthy that in cases of the type ‘base + affix (1) + affix (2)’, Pāṇini teaches those processes which contribute towards the construction of the aṅga with respect to affix (2) before 6.4.1, in the Aṣṭādhyāyī’s serial order. For example, he teaches the addition of vikaraṇas in pāda 3.1 and vowel sandhi, reduplication and samprasāraṇa in pāda 6.1.

  1. Unless I explicitly state that the form being derived is passive, it must be assumed that it is active. 9 1.1.64 aco’ntyādi ṭi. ↩︎

  2. When a short a, that is not pada-final (word-final) is followed by a guṇa vowel i.e., a, e, or o, then both a and the following guṇa are replaced with the latter. ↩︎

  3. When the affix ŚaP is preceded by any verbal root belonging to the list headed by hu ‘to perform sacrifice’, it is replaced with ŚLU (cf. 1.1.61 pratyayasya lukślulupaḥ). ↩︎

  4. A verbal base which has not already undergone reduplication undergoes reduplication when it is followed by ŚLU. (Note that, the whole base does not undergo reduplication. Instead, only one syllable does. See 6.1.1 ekāco dve prathamasya and 6.1.2 ajāder dvitīyasya. Unless necessary, I will not repeat this clarification in this chapter). ↩︎

  5. The vowel of the abhyāsa ‘first of two reduplicated syllables’ is replaced with its short counterpart. ↩︎

  6. 8.2.1 pūrvatrāsiddham teaches that from this rule onwards, a following rule is asiddha ‘suspended’ with respect to a preceding rule. So, if 8.4.54 and any rule that precedes it in the Aṣṭādhyāyī’s serial order are simultaneously applicable, then the latter will not acknowledge 8.4.54 and will thus apply at that step. 8.4.54 can apply only after this. I will demonstrate this more elaborately in the following chapter which is devoted to the concepts asiddha and asiddhavat. ↩︎

  7. Paribhāṣās 62 and 63 are found mentioned together on numerous occasions in the Mahābhāṣya (See Bronkhorst 2004: 18, footnote 11 for details). ↩︎

  8. A siP replacement of LOṬ is replaced with hi and is treated as if not marked with P. ↩︎

  9. The tas, thas, tha and miP replacements for any lakāra marked with Ṅ, are replaced with tām, tam, ta and am, respectively. ↩︎

  10. A single vr̥ddhi vowel replaces both āṬ and the vowel following it. ↩︎

  11. Iha aijyata, aupyata, auhyata iti laṅi kr̥te lāvasthāyām aḍāgamād antaraṅgatvāl lādeśaḥ kriyate, tatra kr̥te vikaraṇo nityatvād aḍāgamaṁ bādhate ‘Here [with reference to the derivation of the forms] aijyata, aupyata, auhyata, after the addition of the affix LAṄ, in that state of the lakāra, by antaraṅgatva, the substitution of the lakāra is done [rather than] the addition of the augment aṬ, and thereafter, by nityatva, the [addition of] vikaraṇa defeats [the insertion of] augment aṬ.’ ↩︎