संस्कृते विश्वासटिप्पन्यः

In my opinion, “1.4.2 vipratiṣedhe paraṁ kāryam” means: in the event of DOI, the RHS rule wins. As stated before, I have not used the term ‘rule conflict’ in my interpretation of 1.4.2. This is because I think that Pāṇini does not require us to use such a concept to understand 1.4.2, and consequently, to perform derivations correctly.

However, as shown above, all post-Pāṇinian discussion pertaining to 1.4.2 has focused on conflict. So, I do need to deal with the topic of conflict to contextualize my findings in the contemporary discourse. In other words, I need to show that my interpretation of 1.4.2 correctly resolves examples of DOI conflict, which I will call Type 2a henceforth. For each example, I will first prove that the example involves conflict, then discuss my solution to it, and finally present the traditional solution to it.

Even though the tradition is not very interested in non-conflict, I will also show that 1.4.2 helps deal with examples of DOI non-conflict, which I will call Type 2b henceforth.

Before we start looking at examples, here is a diagram which summarizes this topic:

[[48]]

Type 2 (DOI = vipratiṣedha)

unidirectional blocking (RE¹ 3)
- DOI conflict (Type 2a) [of significant interest to the tradition]
mutual blocking (RE 4)
- DOI conflict (Type 2a) [of significant interest to the tradition]
no blocking (RE 6)
- DOI non-conflict (Type 2b) [not of much interest to the tradition]

+++(In all cases - )+++
My solution: 1.4.2 (RHS wins)

Note the difference between vipratiṣedha, as interpreted by me, and the concept of conflict, which is popularly discussed in modern post-Pāṇinian literature, in the diagram above.

In this section, I have chosen examples from nominal inflection.²

In all derivations performed in this thesis, I present only those steps diagrammatically at which multiple rules are simultaneously applicable. For example, at step a + b, if rules R1 and R2 are applicable to a and b respectively, then I draw the following kind of diagram to illustrate the same:

(R1→ )a + (R2→ )b

[[49]]

However, if only one rule K1 is applicable (to c) at a given step c + d, then I do not draw diagrams of the following kind to represent this:

(K1 →) c + d

Instead, I simply describe this in words, or symbolically, as follows: c + d 🡪 c* + d (K1).

dēvaiḥ

(1) deva + bhis – ‘God’ (masculine), instrumental plural

(7.3.103 →) deva + (7.1.9→) bhis

7.3.103 bahuvacane jhaly et (ataḥ supi): an e replaces the final a of a nominal base when a plural declensional affix starting with jhaL (a non-nasal stop or a fricative) follows.

7.1.9 ato bhisa ais: ais replaces bhis when bhis occurs after an a-final base.

If bhis is replaced with vowel-initial ais by 7.1.9, then 7.3.103, which applies to only those bases which are followed by a jhaL-initial affix, will not be applicable at the following step. Similarly, if the a of deva is replaced with e by 7.3.103, then 7.1.9, which applies to affixes that are preceded by a-final bases, will not be applicable at the following step.

Therefore, 7.1.9 and 7.3.103 block each other. This is a case of mutual blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.1.9 wins, leading to the correct form: deva + ais 🡪 devais (6.1.88 vr̥ddhir eci) 🡪 devair (8.2.66 sasajuṣoḥ ruḥ) 🡪 devaiḥ (8.3.15 kharavasānayor visarjanīyaḥ).

In his comments on 7.1.9, Patañjali tries to solve this conflict by using paratva (the rule that comes later in the serial order of the Aṣṭādhyāyī wins) but that gives the wrong answer: *devebhis. He then asserts that 7.1.9 is nitya and thus wins, giving the correct form: devaiḥ.³ His explanation for calling 7.1.9 nitya is illogical at best, and we will not delve into it.+++(5)+++ Suffice it to say that the nitya tool, which can only solve cases of unidirectional blocking, cannot be applied to the present case of mutual blocking.

Pradīpa and Uddyota, the two popular commentaries on the Mahābhāṣya suggest that the rule 7.1.9 is anavakāśa whereas 7.3.103 is ⁴ sāvakāśa. So, the former wins. The anavakāśa tool is simply a technical way of arguing the following:

(i) 7.1.9 does not apply anywhere else.
(ii) Surely, Pāṇini must have composed 7.1.9 because it applies somewhere.

From (i) and (ii), the tradition concludes that it has to apply here.

For this and many other examples, instead of following a systematic procedure of rule conflict resolution, the tradition adopts a trial-and-error approach to come up with a justification for the application of the rule which leads to the correct form.

hariṇā

(2) hari + āṄ – ‘green’ (masculine), instrumental singular

(6.1.77 →) hari + (7.3.120 ») āṄ⁵

6.1.77 iko yaṇ aci: iK (i, u, r̥, l̥) is replaced with yaṆ (y, v, r, l) when aC (any vowel) follows.

7.3.120 āṅo nāstriyām: nā replaces the affix āṄ, when it occurs after a non-feminine base termed ghi (a base ending in i or u except sakhi).

If the i of hari is replaced with y by 6.1.77, then 7.3.120 which applies only to bases ending in i or u, will not be applicable at the following step. And if āṄ is replaced with consonant-initial nā, then 6.1.77, which could have applied to the i of hari when it is followed by a vowel, will no longer be applicable. Thus, 7.3.120 and 6.1.77 block each other.

This is a case of mutual blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.3.120 wins, leading to the correct form: hariṇā.⁶

To the best of my knowledge, the tradition does not discuss this conflict. But I would guess that it would have used its interpretation of 1.4.2 (the rule that comes later in the serial order of the Aṣṭādhyāyī wins) or niravakāśatva to solve it.

vāriṇā

(3) vāri + āṄ – ‘water’ (neuter), instrumental singular

For reasons I will clarify below, let us look at the three rules which are applicable at this step without resorting to a diagram:

6.1.77 iko yaṇ aci: same as above.

7.1.73 iko’ci vibhaktau (num napuṁsakasya): augment nUM is attached to a neuter base ending in iK (i, u, r̥, l̥) when a vowel-initial declensional affix follows.

7.3.120 āṅo nāstriyām: same as above.

The important question to ask here is, how should we regard the interaction between 6.1.77 iko yaṇ aci and 7.1.73 iko’ci vibhaktau? Is it a case of SOI or that of DOI?

One could argue that it is a case of DOI. Let me explain why. 6.1.77 is applicable to i of vāri. On the other hand, by 1.1.47 mid aco’ntyāt paraḥ (which teaches that an item marked with M is placed after i.e., to the RHS of the last vowel of the morpheme), 7.1.73 is applicable, not to i, but instead to the (currently empty) position that is to the right-hand side of i. Additionally, note that 1.1.47 uses the term para which is also used in the rule governing DOI (cf. 1.4.2 vipratiṣedhe paraṁ kāryam). For these reasons, one could say that the two rules are applicable to two different operands. Here is the diagrammatic representation of the same:

vār ( 6.1.77→) i (7.1.73→) [] + (7.3.120→) āṄ

On the other hand, one could argue that the interaction between 6.1.77 and 7.1.73 is a case of SOI because the whole base vāri itself is the common operand of both 6.1.77 and 7.1.73. Here is the diagrammatic representation of the same:

(6.1.77, 7.1.73 ») [vāri] + (7.3.120 →) āṄ

In fact, we ought to answer other similar questions before moving forward: if there is an SSRI between a rule teaching the attachment of an augment marked with Ṭ (cf. 1.1.46 ādyantau ṭakitau⁷) to a given item and a rule teaching the substitution of the first sound of that item, then should that interaction be treated as an SOI or as a DOI? Similarly, if there is an SSRI between a rule teaching the attachment of an augment marked with K (cf. 1.1.46 ādyantau ṭakitau) to a given item and a rule teaching the substitution of the last sound of that item, then should that interaction be treated as an SOI or as a DOI?+++(5)+++

Without looking at a large number of examples of SSRI involving augments marked with M, Ṭ or K, it would be difficult to decide which of the two positions is correct. In my thesis, I do not focus on augments and thus am not in a position to definitively answer the aforementioned questions. For the sake of this thesis, I have treated examples of the aforementioned kind involving M-marked augments as cases of DOI and those of the aforementioned kind involving Ṭ- or K-marked augments as cases of SOI. I have done this so that the reader may get exposure to both positions – one, that these are cases of SOI and the other, that these are cases of DOI. This will help set the stage for future research on this topic. +++(5 But, from private communication, he prefers SOI approach here.)+++

Coming back to the present example, this is a case of DOI between the three rules.

vār ( 6.1.77→) i (7.1.73→) + (7.3.120→) āṄ

Now let us look at the relationships between these rules. We have already seen in the previous example that 6.1.77 and 7.3.120 block each other.

Let us look at the DOI interaction between 6.1.77 and 7.1.73. If vāri takes the augment nUM by 7.1.73, then we get vārin which does not end in vowel i and thus, 6.1.77 will not be applicable at the following step. If i of vāri is replaced with y by 6.1.77, then we get vāry which does not end in i, thus 7.1.73 will not be applicable at the following step. Thus, 6.1.77 and 7.1.73 block each other. This is a case of mutual blocking and thus of Type 2a (DOI conflict).

Now let us look at the DOI interaction between 7.1.73 and 7.3.120. If vāri takes the augment nUM by 7.1.73, thereby becoming consonant-final vārin, then 7.3.120, which applies only to those affixes that are preceded by ghi bases ending in i or u, will not be applicable at the following step. And if consonant-initial nā replaces Ṭā by 7.3.120, then 7.1.73 which only applies to certain bases followed by vowel-initial affixes will not be applicable at the following step. Thus, 7.3.120 and 7.1.73 block each other. This is a case of mutual blocking, and thus of DOI conflict.

By my interpretation of 1.4.2, we apply the right-most rule 7.3.120⁸ and get the correct form: vāriṇā.⁹

We have already discussed the traditional position on the conflict between 6.1.77 and 7.3.120 in the previous example. I do not think the tradition discusses the conflict between 6.1.77 and 7.1.73. We can assume that it would use its interpretation of 1.4.2 (the rule that comes later in the serial order of the Aṣṭādhyāyī wins) or the apavāda tool to solve this conflict. As for the conflict between 7.1.73 and 7.3.120, the Bālamanoramā commentary on the Siddhāntakaumudī solves it using the traditional interpretation of 1.4.2.

strīṇām

(4) strī + ām – ‘woman’ (feminine), genitive plural

strī + ām

6.4.79 7.1.54

6.4.79 striyāḥ (aci iyaṅ): the final sound of the base strī is replaced with iyAṄ when a vowel initial affix follows.

7.1.54 hrasvanadyāpo nuṭ (āmi): augment nUṬ is introduced to affix ām when it occurs after a nominal base which ends in a short vowel, or is termed nadī (feminine long ī- and ū-final bases), or has taken the feminine affix ṬāP.

If the ī of strī is replaced with iyAṄ by 6.4.79, thereby making it striy, then 7.1.54, which applies to ām when preceded by nadī-final vowels ī and ū, will not be applicable at the following step.

[[54]]

If the augment nUṬ is added to the affix ām by 7.1.54, thereby making it consonant-initial nām, then 6.4.79, which is only applicable to the base strī when it is followed by vowel-initial affixes, will not be applicable at the following step.

This is a case of mutual blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.1.54 wins, leading to the correct form: strīṇām.¹⁰

The Bhaimī commentary on the Laghusiddhāntakaumudī solves the conflict between 6.4.79 and 7.1.54 using the traditional interpretation of 1.4.2 (i.e., the rule that comes later in the serial order of the Aṣṭādhyāyī wins).

vārīṇām

(5) vāri + ām – ‘water’ (neuter), genitive plural

vāri + ām

7.1.73 7.1.54

7.1.73 iko’ci vibhaktau (num napuṁsakasya): augment nUM is attached to a neuter base ending in iK (i, u, r̥, l̥) when a vowel-initial declensional affix follows.

7.1.54 hrasvanadyāpo nuṭ (āmi): same as above.

If the augment nUM is attached to vāri by 7.1.73, thereby making it consonant-final vārin (1.1.47 mid aco’ntyāt paraḥ), then 7.1.54, which only applies to ām when it is preceded by certain vowel-final bases, will not be applicable at the following step.

On the other hand, if the augment nUṬ is attached to the affix ām by 7.1.54, thereby making it consonant-initial nām, then 7.1.73, which is only applicable to certain bases that are followed by vowel-initial affixes, will not be applicable at the following step.

Both 7.1.54 and 7.1.73 block each other. This is a case of mutual blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.1.54 wins, leading to the correct form: vārīṇām¹¹ (6.4.3 nāmi, 8.4.2 aṭkupvāṅnumvyavāye’pi).

The tradition resorts to Kātyāyana’s vārttika ‘numaciratr̥jvadbhāvebhyo nuṭ pūrvavipratiṣiddham’ (vt. 1139 on 7.1.96 striyāṁ ca) to solve this conflict. This vārttika teaches that even though the rule teaching the attachment of the augment nUṬ (7.1.54) comes before the rule teaching the attachment of the augment of nUM (7.1.73) in the serial order of the Aṣṭādhyāyī, the former wins. In this and other pūrvavipratiṣiddha vārttikas, Kātyāyana simply lists those conflicts which cannot be correctly solved using the traditional interpretation of 1.4.2.

kroṣṭr̥̄ṇām

(6) kroṣṭu + ām – ‘jackal’ (masculine), genitive plural

kroṣṭu + ām

7.1.97 7.1.54

7.1.97 vibhāṣā tr̥tīyādiṣv aci (tr̥jvat kroṣṭuḥ): ‘the base kroṣṭu, is treated as if ending in affix tr̥C optionally, when a vowel-initial ending of the tr̥tīyā triplet (instrumental) or any of the following triplets (namely dative, ablative, genitive or locative) follows.40

7.1.54 hrasvanadyāpo nuṭ (āmi): same as above.

If the u of kroṣṭu becomes r̥ by 7.1.97, then 7.1.54, which applies to ām when it is preceded by any of the short vowels, will be applicable to ām at the following step. But if the augment nUṬ is added to ām by 7.1.54, thereby making it (consonant-initial) nām, then 7.1.97, which applies to kroṣṭu only when it is followed by a vowel-initial tritīyādi affix, will not be applicable at the following step.

[[56]]

7.1.54 blocks 7.1.97, but 7.1.97 does not block 7.1.54. This is a case of unidirectional blocking, and thus of DOI conflict.

By my interpretation of 1.4.2, the RHS operation 7.1.54 wins, leading to the correct form: kroṣṭūnām (6.4.3 nāmi).

Since this is a case of unidirectional blocking, the tradition could have used the nitya tool to solve this conflict. However, it does not do so.¹² Instead, Kātyāyana has written the vārttika ‘numaciratr̥jvadbhāvebhyo nuṭ pūrvavipratiṣiddham’ (vt. 1142 on 7.1.96 striyāṁ ca) to solve it. This vārttika teaches that even though the rule teaching the attachment of the augment nUṬ (7.1.54) comes before the rule teaching tr̥jvadbhāva (7.1.97) in the serial order of the Aṣṭādhyāyī, the former wins.

kartr̥

(7) kartr̥ + sU – ‘doer’ (neuter), nominative singular

kartr̥ + sU

7.1.94 7.1.23

7.1.94 r̥duśanaspurudaṁso’nehasāṁ ca (asambuddhau anaṅ sau): the final sound of a base ending in r̥Ṭ or of the bases uśanas, purudaṁsas and anehas is substituted with anAṄ when followed by non-vocative sU.

7.1.23 svamor napuṁsakāt (luk): affixes sU and am occurring after a neuter base are substituted with LUK.

If we apply 7.1.23, then 7.1.94, which applies only when followed by sU, will not be applicable at the following step. If we apply 7.1.94, then 7.1.23, which applies to any neuter base regardless of its final sound, will be applicable at the following step.

This is a case of unidirectional blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.1.23 wins, thereby giving the correct form: kartr̥.

To the best of my knowledge, the tradition does not discuss this conflict. However, I think it would use the nitya tool (the rule that unidirectionally blocks the other wins) to solve it.

tad

(8) tad + sU – ‘that’ (neuter), nominative singular

tad + sU

7.2.102 7.1.23

7.2.102 tyadādīnām aḥ (vibhaktau): the final sound of a base belonging to the group headed by tyad ‘that’ is replaced with a when a declensional affix follows.

7.1.23 svamor napuṁsakāt (luk): same as above.

What kind of interaction occurs between the two rules? The tradition seems to be confused about this. So, let us start by looking at my solution.

This is a case of DOI. By my interpretation of 1.4.2, the RHS rule 7.1.23 wins, giving us the correct answer: tad.¹³

In his commentary[^44] on 7.1.23, Patañjali first tries to use the traditional interpretation of 1.4.2 (the rule that comes later in the serial order of the Aṣṭādhyāyī wins) to determine which of the two rules he must apply. But he gets the wrong answer upon doing so. Then, he tries to use the nitya tool.

If we apply 7.1.23 at this step, 7.2.102 will not be applicable at the following step. On the other hand, if we apply 7.2.102 at this step, 7.1.23 will still be applicable at the following step. Thus, this is a case of unidirectional blocking, and of Type 2a (DOI conflict). Therefore, the nitya tool can be used here.

However, Patañjali then says that 7.1.23 is not nitya with respect to 7.2.102. This is because, after the hypothetical application of 7.2.102, 7.1.23 is not the only rule that will be applicable. 7.1.24 ato ’m45 will also be applicable. Since 7.1.24 is an apavāda of 7.1.23, the former will win. So 7.1.23 will, despite being applicable, fail to apply, following the application of 7.2.102.

For this reason, Patañjali says that 7.1.23 cannot be called nitya with respect to 7.2.102. To deal with this problem, Patañjali suggests some changes in the wording of 7.1.23 svamor napuṁsakāt. We will not dwell on his argument, because it is beyond our scope.

Contrary to Patañjali’s conclusion that 7.1.23 cannot be called nitya, according to paribhāṣā 47 of the Paribhāṣenduśekhara, 7.1.23 is nitya and thus should win. It reads: yasya ca lakṣaṇāntareṇa nimittaṁ vihanyate na tad anityam. Kielhorn translates it as follows: ‘(an operation [here 7.1.23]), the cause of which would, (after the taking effect of another operation [here, 7.2.102] that applies simultaneously), be removed by another (third) rule [here, 7.1.24], is not (on that account regarded as) anitya.’

vāriṇe

(9) vāri + Ṅe – ‘water’ (neuter), dative singular

vār (7.3.111→) i (7.1.73 →)[] + Ṅe

7.3.111 gher ṅiti (guṇaḥ supi): the final vowel of a ghi base (a base ending in i or u, except sakhi) is replaced with guṇa (here, e / o) when followed by a declensional affix marked with Ṅ.

7.1.73 iko’ci vibhaktau (num napuṁsakasya): augment nUM is attached to a neuter base which ends in iK (i, u, r̥, l̥), provided a vowel-initial declensional affix follows.

विश्वास-टिप्पनी - SOI view

vār (7.3.111, 7.1.73 →) i + Ṅe

Operand in case of 7.1.73 - i of a nueter base.
Operand in case of 7.3.111: i.
The former is more “specific”.
Hence it wins and is applied.

This is a case of DOI. If we apply 7.3.111 at this step, then 7.1.73 will not be applicable at the following step. If we apply 7.1.73 at this step, then 7.3.111 will not be applicable at the following step. This is a case of mutual blocking and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, we apply the RHS rule 7.1.73 and get the correct form: vāriṇe.

The tradition uses the vārttika, guṇa-vr̥ddhy-auttva-tr̥jvad-bhāvebhyo num pūrva-vipratiṣiddham (vt. 1046 on 7.1.96 striyāṁ ca) to solve this conflict. This vārttika teaches that even though the rule teaching the attachment of the augment nUM (7.1.73) comes before the rule teaching guṇa (7.3.111) in the serial order of the Aṣṭādhyāyī, the former wins.

[[59]]

Having looked at examples of DOI conflict (Type 2a), now let us look at examples of DOI non conflict (Type 2b).

rājā

(10) rājan + sU – ‘king’ (masculine), nominative singular

rājan + sU

6.4.8 6.1.68

6.4.8 sarvanāmasthāne cāsambuddhau (nopadhāyāḥ dīrghaḥ): the penultimate sound of a base ending in n is replaced with its dīrgha ‘long’ equivalent when a non-vocative sarvanāmasthāna affix (sU, au, Jas, am, auṬ in non-neuter forms or Śi) follows.

6.1.68 halṅyābbhyo dīrghāt sutisyapr̥ktaṁ hal (lopaḥ): there is elision by LOPA of the finite verb affixes ti and si, when they consist of a single sound and follow a form which ends in a consonant, and of the nominative singular case affix sU, when it follows a form which ends in a consonant or the long final vowel of feminine affixes Ṅī or āP.

If 6.4.8 applies at step K, we get rājān, which still ends in a consonant. So 6.1.68 will be applicable at the step K+1. If sU is replaced with LOPA by 6.1.68 at step K, the properties of the affix sU still hold (cf. 1.1.62 pratyayalope pratyayalakṣaṇam), so 6.4.8 will be applicable at step K+1.

We see that 6.4.8 and 6.1.68 do not block each other. This is a case of Type 2b (DOI non conflict).

By my interpretation of 1.4.2, the RHS rule 6.1.68 wins and we get rājan. Now thanks to 1.1.62 pratyayalope pratyayalakṣaṇam, we apply 6.4.8 and get rājān. At this juncture, we apply 8.2.7 nalopaḥ prātipadikāntasya47, which teaches that n is replaced with LOPA at the end of a nominal stem which is termed pada, and get the correct form: rājā.

Note that even if we had applied 6.4.8 (the LHS rule) at the first step, we could have still applied 6.1.68 at the following step. And applying these two rules in this order too would have given us the correct form.

Why then did Pāṇini need to say anything about DOI non-conflict at all? Why did he prescribe that the RHS be applied in such cases (cf. my interpretation of 1.4.2)? We will answer this question while discussing the following examples.

The tradition is not interested in such cases of non-conflict.

tisr̥ṇām

(11) tri + ām – ‘three’ (feminine), genitive plural

(7.2.99 →) tri + (7.1.54→) ām

7.2.99 tricaturoḥ striyāṁ tisr̥catasr̥: tri and catur are replaced with tisr̥ and catasr̥ respectively in the feminine.

7.1.54 hrasvanadyāpo nuṭ (āmi): augment nUṬ is introduced to affix ām48 when it occurs after a nominal base which ends in a short vowel, or is termed nadī (feminine bases ending with ī and ū), or has taken the feminine affix ṬāP.

If we replace tri with tisr̥ at this step, 7.1.54 will still be applicable at the following step because tisr̥ ends in a short vowel. And if we apply 7.1.54 at this step, 7.2.99 will still be applicable at the following step, because its application does not depend on the affix.

Neither of the two rules blocks the other, and so this is a case of Type 2b (DOI non-conflict). By my interpretation of 1.4.2, we apply the RHS rule 7.1.54 and get tri + nām. Thereafter, we apply 7.2.99 tricaturoḥ striyāṁ tisr̥catasr̥ and get the correct form: tisr̥ṇām[^49].

[[61]]

In order to understand why Pāṇini has prescribed that we pick the RHS rule in cases of DOI non-conflict, let us perform this derivation again, this time by picking the LHS rule in case of DOI non-conflict.

(7.2.99→)tri + (7.1.54→) ām

7.2.99 tricaturoḥ striyāṁ tisr̥catasr̥: same as above.

7.1.54 hrasvanadyāpo nuṭ (āmi): same as above.

This is a Type 2b (DOI non-conflict). As stated above, as an experiment, we are going to apply the LHS rule 7.2.99 in this case (of DOI non-conflict). Upon applying 7.2.99, we get tisr̥ + ām. Here, two rules are applicable:

(7.2.100→) tisr̥ + (7.1.54→) ām

7.2.100 aci ra r̥taḥ (vibhaktau tricaturoḥ tisr̥catasr̥): a r replaces r̥ of the bases tisr̥ and catasr̥, when a vowel-initial declensional affix follows.

7.1.54 hrasvanadyāpo nuṭ (āmi): same as above.

If r̥ is replaced with consonant r by 7.2.100, then 7.1.54, which applies to ām when it is preceded by certain vowel-final bases will not be applicable at the following step. And if ām takes augment nUṬ by 7.1.54, thereby becoming consonant-initial nām, then 7.2.100 which applies to r̥ when a vowel-initial affix follows will not be applicable at the following step.

Thus, 7.2.100 and 7.1.54 block each other. This is a case of mutual blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, the RHS operation 7.1.54 wins, leading to the correct form: tisṛṇām.

We have seen that, regardless of whether we pick the LHS or the RHS rule in case of Type 2b (DOI non-conflict) here, we get the same answer: tisr̥ṇām. However, the two derivational paths look different from each other. The first path, in which we pick the RHS rule at the first step (as taught by Pāṇini in [my interpretation of] 1.4.2), is significantly shorter than the second path, in which we pick the LHS rule at the first step. In other derivations too, I have noticed that the derivation looks relatively shorter when we pick the RHS rule in case of type 2b (DOI non-conflict) and relatively longer when we pick the LHS rule.

[[62]]

But is it merely to keep derivations compact that Pāṇini has prescribed the choice of the RHS rule in cases of DOI non-conflict? No. In the next example, we will see that we cannot get the correct answer without picking the RHS rule in case of DOI non-conflict.

How does the tradition perform this derivation? Vārttikas 11 to 1450 on 7.1.96 striyāṁ ca, and Patañjali’s comments on them, deal with this topic in detail and propose various tools like pūrvavipratiṣiddha and apavāda to solve this problem. We will not delve into this topic here.

asmai

(12) idam + Ṅe – ‘this’ (masculine), dative singular

All cases of DOI in this derivation are of Type 2b (DOI non-conflict). I will not prove this at each step.

( 7.2.112 →) id a (7.2.102→) m + Ṅe

7.2.112 an āpy akaḥ (vibhaktau idamaḥ idaḥ): the id of idam is substituted with an, when it does not include a k, and when a declensional affix belonging to āP, i.e., any instrumental, dative, ablative, genitive or locative affix, follows.

7.2.102 tyadādīnām aḥ (vibhaktau): the final sound of a base belonging to the group headed by tyad ‘that’ is replaced with a when a declensional affix follows.

By my interpretation of 1.4.2, the RHS rule 7.2.102 wins, and we get: ida-a + Ṅe. Here, multiple rules are applicable:

(7.2.112.→) id (6.1.97→) [ a-a ] + (7.1.14→) Ṅe

7.2.112 an āpy akaḥ (vibhaktau idamaḥ idaḥ): same as above.

6.1.97 ato guṇe: when a short a, which 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.

[[63]]

7.1.14 sarvanāmnaḥ smai (ṅer yaḥ ataḥ): the affix Ṅe, when occurring after a pronominal base ending in a, is replaced with smai.

By my interpretation of 1.4.2, we apply the right-most rule 7.1.14, and get ida-a + smai. Here, multiple rules are applicable:

( 7.2.112 7.2.113→) id (6.1.97→) [ a-a ] + smai

7.2.113 hali lopaḥ (vibhaktau idamaḥ idaḥ akaḥ): the id of idam is replaced with LOPA, when it does not include a k, and when a consonant-initial declensional affix belonging to āP, i.e., any instrumental, dative, ablative, genitive or locative affix, follows.

6.1.97 ato guṇe: same as above.

By my interpretation of 1.4.2, we apply the RHS rule 6.1.97 and get ida + smai. Here multiple rules are applicable:

(7.2.112 7.2.113→) id a + smai

We see that there is a case of SOI between 7.2.112 and 7.2.113. Because 7.2.113 applies only when the base is followed by a consonant initial affix, it is more specific than, and defeats 7.2.112. Thus, we get the correct form: asmai.

At the very first step of this derivation, where we see the two rules 7.2.112 and 7.2.102 involved in DOI non-conflict, if we had chosen to apply the LHS rule 7.2.112 instead of the RHS rule 7.2.102, we would have got the wrong form at the end of the derivation: *anasmai. The same can be said about the second step too: picking 7.2.112 at the second step instead of 7.1.14 too would have given us the wrong form: *anasmai.

This shows that, even though whether we choose the LHS rule or the RHS rule may not matter in certain cases of DOI non-conflict (see examples 10 and 11 above), in cases of DOI non conflict like this one, choosing the RHS rule alone gives the correct answer.

[[64]]

aham

Finally, let us look at an example which involves cases of both DOI conflict and DOI non conflict.

(13) asmad + sU – ‘I’ (any gender), nominative singular

(7.2.94→) asma (7.2.86→) d + (7.1.28→) sU

7.2.94 tvāhau sau (yuṣmadasmador maparyantasya vibhaktau): the parts of yuṣmad and asmad extending up to ma¹⁴ are replaced with tva and aha respectively when followed by the case affix sU.

7.2.86 yuṣmadasmador anādeśe (vibhaktau āḥ): the final sounds of yuṣmad and asmad are replaced with ā when followed by consonant-initial case affixes which have not undergone any substitution.

7.1.28 ṅeprathamayor am (yuṣmadasmadbhyāṁ vibhaktau): Ṅe, and nominative, accusative affixes are replaced with am when preceded by yuṣmad and asmad.

Let us determine the relationship between 7.2.94 and the two other rules.

If we apply 7.2.94 at this step, 7.2.86 will be applicable at the following step. Similarly, if we apply 7.2.86 at this step, 7.2.94 will be applicable at the following step. There is a Type 2b (DOI non-blocking) relationship between 7.2.94 and 7.2.86.

Similarly, if we apply 7.2.94 at this step, 7.1.28 will be applicable at the following step. If we apply 7.1.28 at this step, 7.2.94 will be applicable at the following step.¹⁵ There is a Type 2b (DOI non-blocking) relationship between 7.2.94 and 7.1.28.

[[65]]

Thus 7.2.94 has a Type 2b (DOI non-conflict) with the other two rules.

Now let us determine the relationship between 7.2.86 and 7.1.28. If we apply 7.2.86 at this step, 7.1.28 will still be applicable at the following step. However, if we apply 7.1.28 at this step, then the affix sU will undergo ādeśa ‘substitution’ with am. 7.2.86 can only apply to asmad when followed by a non-substituted, consonant-initial affix. Thus, 7.2.86 will not be applicable at the following step. This is a case of unidirectional blocking, and thus of Type 2a (DOI conflict).

By my interpretation of 1.4.2, we apply the right-most rule 7.1.28 and get: asmad + am. Here again, two rules are applicable:

(7.2.94→) asma (7.2.90→) d +am

7.2.90 śeṣe lopaḥ: the final sounds of yuṣmad and asmad are replaced with LOPA when followed by case affixes not listed in the preceding rules (7.2.86-89).¹⁶

If we apply 7.2.94 at this step, 7.2.90 will still be applicable at the following step. If we apply 7.2.90 at this step, 7.2.94 will be applicable at the following step. This is a case of no blocking, and thus of Type 2b (DOI non-conflict). By my interpretation of 1.4.2, we apply the RHS rule 7.2.90 and get asma + am. Lastly, we apply 7.2.94 and get aha + am, to which we apply 6.1.97 ato guṇe.¹⁷ This gives the correct form: aham.

As stated in a footnote on 7.2.94, the traditional interpretation of 7.2.94 is different from mine. Thus, its derivational process is different and slightly longer. We will not delve into it here. I will simply say that the tradition would have resolved the DOI conflict in this example using the nitya tool.

[[66]]

This brings us to the end of examples of DOI in this chapter. We will, of course, study more examples of DOI conflict in later chapters. Before we go to the next section, here I want to emphasize that I have discussed blocking and conflict in these derivations only because post Pāṇinian scholarship is interested in these topics. In other words, I have attempted to show that examples of conflict can be solved by my interpretation of 1.4.2.

Note that, if we had simply avoided talking about blocking and conflict, we would have completed these derivations almost effortlessly, by simply picking the right-most rule (cf. my interpretation of 1.4.2) in every case of DOI, irrespective of whether or not the rules in question are involved in any kind of conflict.

RE = ‘Representative Example’. ↩︎
Note that examples from nominal inflection are simpler than those from verbal inflection. One of the many reasons behind this is that, while nominal inflection involves only two items, i.e., a base and affix, verbal inflection generally involves at least three items, i.e., a base followed by two affixes. We will look at examples from verbal inflections as well as primary and secondary derivatives in the following chapters. ↩︎
Mbh III.244.13-21. ↩︎
Mbh III.276.6-22. ↩︎
The instrumental singular affix taught by 4.1.2 sv-au-jas… is Ṭā and not āṄ. The use of āṄ instead of Ṭā “is best understood as reflecting earlier traditions” (Cardona 1997: 51). ↩︎
n > ṇ, by 8.4.2 aṭkupvāṅnumvyavāye’pi. ↩︎
Items marked with Ṭ and items marked with K should be attached to the beginning and end respectively. ↩︎
Note that this is one of the rare cases in which even if we had applied another rule, namely 7.1.73, we would still have got the correct form. ↩︎
n > ṇ, by 8.4.2 aṭkupvāṅnumvyavāye’pi. ↩︎
n > ṇ, by 8.4.2 aṭkupvāṅnumvyavāye’pi. ↩︎
Note that both augments i.e., nUM and nUṬ essentially refer to the same sound n. However, if we applied the rule prescribing nUM, we would get vārin + ām (1.1.47 mid aco’ntyāt paraḥ). In such a situation, we would not be able to elongate the ī of vārin because 6.4.3 nāmi would not apply here. 39 Mbh III.276.6. ↩︎
This is discussed in Pradīpa on vt. 11, 7.1.96. ↩︎
Note that 7.2.102 is not applicable at this point, thanks to 1.1.63 na lumatāṅgasya. 44 Mbh III.248.23-249.2. ↩︎
The tradition translates maparyantasya as ‘up to m’ but I think that Pāṇini means ‘up to ma’. Both interpretations lead to correct answers for all forms of yuṣmad-asmad. My interpretation makes derivations simpler and shorter. ↩︎
Given that sU has been replaced with am, how will 7.2.94 apply at the following step? This is because, by 1.1.56 sthānivadādeśo’nalvidhau, am is treated like sU. How do we know this is not an aL-vidhi? asma and am are not adjacent to each other (d sits in the middle of the two), and so this is not an aL operation. ↩︎
Note that, here, the affix sU has undergone ādeśa ‘substitution’ with am. So, 7.2.86, which can only apply to asmad when followed by a non-substituted and consonant initial affix, and which was applicable in the previous step, is no longer applicable at this step. Instead of that rule, 7.2.90 has become applicable. ↩︎
An a which is not at the end of a pada and the guṇa vowel following it are both replaced with the latter. ↩︎