Now let us examine the distribution of examples of conflict across various kinds of derivations (e.g., subanta, kr̥danta etc.). Since Pāṇini uses the general-exception framework throughout the Aṣṭādhyāyī, we find cases of SOI conflict in all kinds of derivations. And while we might find more examples of SOI conflict in some kinds of derivations than others, we do not come across any unique or peculiar patterns that merit discussion here.
So, I will focus on the distribution of DOI conflicts in Pāṇinian derivations in this section. Let us inquire why, on the whole, DOI conflicts, and especially certain kinds of DOI conflicts (e.g., mutual blocking), are found more frequently in certain kinds of derivations than others. I request the reader to bear in mind that I will be making some broad generalizations here in order to paint an overarching picture. Therefore, my statements will not be entirely accurate.
Since we are talking about DOI-conflict here, I will not touch upon those instances of DOI which do not involve conflict.
To start with, let us consider subanta derivations.
‘nominal base + declensional affix’
A B
We will focus on cases where the application of A to the base is triggered by the first sound of the affix, and the application of B to the affix is triggered by the last sound of the base. If the first sound of the affix changes, A is not applicable to the base anymore and if the last sound of the base changes, B is not applicable to the affix anymore. Therefore, when two such rules are simultaneously applicable in subanta derivations, A to the base and B to the affix, both rules block each other, leading to a situation of DOI conflict. See examples 1-5 and 9 of section 2.7, chapter 2.
In other cases, we find that the application of B to the affix is triggered simply by the grammatical gender, word category (e.g., pronoun) etc. of the base. In such a case, even if the
165
base undergoes phonological change, B will still be applicable at the following step. On the contrary, we observe that the application of A to the base is triggered by the first sound or the mere presence of an affix. So, if the affix is deleted, for example, by LUK, or if its first sound changes, then A will no longer be applicable at the following step. These are cases of unidirectional blocking. See examples 6-8 and 13 of section 2.7, chapter 2.
Thus, we see both kinds of examples of DOI conflict, namely those of mutual blocking and those of unidirectional blocking, in subanta derivations. Note that I have overlooked, for the sake of simplicity, examples of DOI conflict where both rules apply to two different parts of the base or to two different parts of the affix respectively.
Let us contrast this with tiṅanta derivations. One of the early steps of these derivations looks like this:
‘verbal base + vikaraṇa + finite ending’
C D E
Vikaraṇas on the whole do not undergo many changes. Even when they do, the application of D (which may teach replacement with LUK or other substitutes, augmentation, etc.) is not triggered by the last sound of the verbal root. So even if the verbal root undergoes some changes, D will still be applicable to the vikaraṇa at the following step. On the other hand, the application of C (which may entail guṇa, samprasāraṇa, augmentation, lengthening of the penultimate vowel, deletion of nasal etc.) is dependent on the existence of the vikaraṇa, its being marked with K or Ṅ, etc. So, if the vikaraṇa undergoes certain changes, such as replacement with LUK or attachment of certain augments like iṬ which annul the effect of K/Ṅ (cf. 1.2.18 na ktvā seṭ), C will not be applicable to the base at the following step. These are cases of unidirectional blocking.
Most rules (E) which are applicable to finite endings at this stage, are triggered by the type of lakāra that the ending has replaced, whether that lakāra is marked with Ṭ or Ṅ, the number and person of the ending, whether the ending is parasmaipada or ātmanepada etc. They do not block and are not blocked by other rules (for example, see rules 3.4.77 – 3.4.112 of the Aṣṭādhyāyī). So, we will not focus on them here.
166
Once the aṅga is ready, we get:
aṅga + finite ending
F G
The application of F (such as guṇa, vr̥ddhi, samprasāraṇa etc.) is triggered by the existence of the affix, the first sound of the affix, whether or not it is marked with K / Ṅ etc. Thus, if the affix undergoes certain changes, F is not applicable at the following step. But G is not triggered by the last sound of the aṅga. Thus, even if the aṅga undergoes certain changes, G is still applicable at the following step. This is a case of unidirectional blocking.
Let us now look at kr̥danta derivations.
verbal base + kr̥t affix
H I
The application of H (such as guṇa, samprasāraṇa etc.) is triggered by / depends on the first sound of the affix, whether it has taken the augment iṬ, whether it is marked with K / Ṅ etc. Thus, if the affix undergoes certain changes, H is not applicable at the following step. Let us call H the dependent rule. On the other hand, I is triggered by the first sound of the affix itself (e.g., 7.2.35 ārdhadhātukasyeḍ valādeḥ) and other factors. Essentially, the application of I is not dependent on the final sound of the base. So even if the base changes, I is still applicable at the following step. Let us call I the independent rule. This is a case of unidirectional blocking, where the independent rule blocks the dependent rule.
Before we proceed further, notice that, in almost all cases of unidirectional blocking in DOI discussed in the thesis, it is the RHS rule which unidirectionally blocks the LHS rule, and not vice-versa.62 This is because, it is the RHS rule which is independent and it is the LHS rule which is dependent. In other words, in almost all cases of unidirectional blocking, the applicability of the RHS rule does not depend on whether the penultimate or last sound of the
167
base changes, but the applicability of the LHS rule does depend on whether the affix is marked with K / Ṅ, whether it starts with a vowel, whether it has taken the augment iṬ etc.
Coming back to the larger theme of this section, we see that almost all cases of DOI conflict in both tiṅanta and kr̥danta derivations involve only unidirectional blocking. This can be observed in the examples discussed in section 4.3. We rarely come across examples of DOI conflict that involve mutual blocking. One such exception is example 21 of section 4.3 of this chapter, which does involve mutual blocking.
To sum up my observations, we find examples of both mutual and unidirectional blocking in subanta derivations, but of unidirectional blocking alone in tiṅanta and kr̥danta derivations.
As seen in this thesis, we find relatively fewer examples of DOI conflict in taddhitānta and samāsa derivations than we do in subanta, tiṅanta and kr̥danta sections. How can we explain this phenomenon?
Let us first answer this question in the context of samāsa derivations. The samāsa template is ‘[(base1 suP1) (base2 suP2)] + suP3’. suP1 and suP2 are replaced with LUK by 2.4.71 supo dhātuprātipadikayoḥ. Thus, we are left with ‘base1 base2 + suP3’. Given that the only remaining affix, i.e., suP3 is also a suP affix, there is almost no scope for any other conflicts to arise apart from those that can potentially arise in subanta derivations. The only exceptions to this are those cases wherein the uttarapada can potentially trigger changes in the pūrvapada (see examples 1 and 8 of section 3.2, chapter 3). This explains why we find very few examples of DOI conflict which are exclusive to samāsa derivations, i.e., which are not already found in subanta derivations.
In taddhitānta derivations too, we find very few examples DOI conflict. Even these examples are quite similar to each other (see examples 3-7 of section 3.2, chapter 3) and arise because of the nominal inflection of taddhitānta forms. Why is this the case? The majority of taddhita rules actually teach addition of taddhita affixes, and not any substitutions or modifications. The taddhita template is ‘(nominal base + suP) + taddhita affix’. suP is replaced with LUK by 2.4.71 supo dhātuprātipadikayoḥ. Thus, we are left with ‘nominal base + taddhita affix’.
nominal base + taddhita affix
J L
168
Taddhita affixes undergo very few, generic changes by rules (L) like by 7.1.2 āyaneyīnīyiyaḥ phaḍhakhacchaghāṁ pratyayādīnām, which are independent of the final sound of the nominal base. So, any change in the base by rule J cannot block these operations (L) on taddhita affixes.
The nominal bases preceding taddhita affixes can also undergo certain general changes by rules (J) such as 7.2.117 taddhiteṣv acām ādeḥ, 7.2.118 kiti ca etc. which do not depend on the first sound of the taddhita affix for their application, and thus are not blocked by L in case of DOI. And even those operations (J) such as 6.4.146 or guṇaḥ and 6.4.148 yasyeti ca, which are triggered by the first sound of the following taddhita affix, are seldom blocked, simply because the following taddhita affixes themselves undergo very few changes. So, barring replacement with LUK (see examples 3-7 of section 3.2, chapter 3), most changes in the taddhita affix cannot block these operations (J) on the nominal base. Since there is little scope for DOI blocking between J and L, we come across very few examples of DOI conflict in taddhita derivations.
169