We have already looked at several examples of SOI while discussing examples of DOI conflict. Here I will present a few more examples. As I have done earlier, I will spell out and examine the conditions in which the rules apply and then determine which of the two rules is more specific.
(1) cal + tiP – ‘to walk’, simple future third-person singular
cal + tiP
3.1.33 3.1.68
3.1.33 syatāsī lr̥luṭoḥ: affixes sya and tāsI respectively occur after verbal roots when LR̥ and LUṬ follow.
3.1.68 kartari śap: affix ŚaP occurs after a verbal root when a sārvadhātuka affix which denotes kartr̥ ‘agent’ follows.
3.1.33
(when LR̥ and LUṬ follow)
3.1.68
(when LR̥ and LUṬ follow)
(when other sārvadhātuka affixes follow)
The conditions highlighted in bold are exactly the same. This is a case of SOI-M. Thus, we compare the two rules themselves. 3.1.33 has been taught specifically for LR̥ and LUṬ. So, it is more specific and thus wins. We get cal + sya + ti 🡪 caliṣyati (7.2.35 ārdhadhātukasyeḍ valādeḥ), which is the correct form.
160(2) vad + miP – ‘to speak’, imperative first-person singular
vad + miP
3.1.68 3.4.89 3.4.92 3.1.68 kartari śap: same as above.
3.4.89 mer niḥ: a miP substitute of LOṬ is replaced with ni.
3.4.92 āḍ uttamasya pic ca: a first-person substitute of LOṬ receives the initial augment āṬ which is treated as marked with P.
3.1.68 is not in conflict with 3.4.89 or 3.4.92. By my interpretation of 1.4.2 we should perform the RHS operation. But which one of the two RHS rules, namely 3.4.89 and 3.4.92, should we apply? Let us examine the SOI between 3.4.89 and 3.4.92.
3.4.89
miP (replacement of LOṬ)
3.4.92
miP (replacement of LOṬ)
other first-person affixes (replacements of LOṬ)
The conditions highlighted in bold are exactly the same. This is a case of SOI-M. Thus, we compare the two rules themselves. 3.4.89 has been taught specifically for miP. So, it is more specific and thus wins. Thus, we get vad + ni. Thereafter, the derivation proceeds as follows: vad + ni 🡪 vad + āni (3.4.92 āḍ uttamasya pic ca) 🡪 vad + ŚaP + āni (3.1.68 kartari śap) 🡪
vadāni (6.1.97 ato guṇe), which is the correct form.
161
(3) tṝ + tiP – ‘to cross’, present third-person singular
tṝ + tiP 🡪 tṝ + ŚaP + tiP (3.1.68 kartari śap)
tṝ + ŚaP + tiP
7.3.84 7.1.100
7.3.84 sārvadhātukārdhadhātukayoḥ: guṇa replaces the final sound iK (i, u, r̥, l̥) of a verbal base when a sārvadhātuka or ārdhadhātuka affix follows.
7.1.100 ṝta id dhātoḥ: ṝ which occurs at the end of a verbal base is replaced with i.
Note that, we have to take into account rules like 1.1.5 kṅiti ca [which prohibits guṇa and vr̥ddhi of the iK (i, u, r̥, l̥) of a verbal base when the following affix is marked with K, G or Ṅ] when determining the exact conditions in which the aforementioned rules are applicable. Because of 1.1.5, 7.3.84 is applicable only when the affix is not marked with K, G or Ṅ.
7.3.84
ṝ + affix (sārvadhātuka or ārdhadhātuka) (not marked with K, G, Ṅ) other iK sounds + affix (sārvadhātuka or ārdhadhātuka) (not marked with K, G, Ṅ) 7.1.100
ṝ + affix (sārvadhātuka or ārdhadhātuka)
The conditions highlighted in bold are not the same. Thus, this is a case of SOI-L. 7.3.84 is more specific because it is applicable only if the affix is not marked with K, G or Ṅ whereas 7.1.100 is applicable regardless of whether the affix is marked with K, G or Ṅ. Thus, 7.3.84 wins, giving the correct form tarati (cf. 1.1.51 ur aṇ raparaḥ).
Let us now consider Cardona’s (1970: 57-58) method of deriving this form. He uses a principle that he calls ‘limited blocking’ to deal with this aforementioned SOI. He explains it as follows: “though a rule R2 as a whole does not state an apavāda of an R1, as a whole, it can do so for some operands or environments common to both”. Further, he says: “(Consider) rules: 7.3.84 sārvadhātukārdhadhātukayoḥ and 7.1.100 ṝta id dhātoḥ. By the latter, the ṝ of an aṅga which is a verb root is replaced with i. The rules are not related as utsarga and apavāda in their entirety: the operands of 7.3.84 are i, u, r̥ while that of 7.1.100 is ṝ. Nor are the contexts identical. Although 7.1.100 operates when the root is followed by any affix introduced after it
162
and sārvadhātuka and ārdhadhātuka affixes, the contexts for 7.3.84, include all post-radical affixes, the context of 7.3.84 is restricted by 1.1.5 (kṅiti ca). In the case of the single shared operand (ṝ), then, 7.1.100 will counter 7.3.84 (sic)61, since all the contexts of the former are included in those of the latter. Thus, given the root stṝ followed by the affix ana, one obtains the desired form staraṇa- ‘spreading’ without recourse to paratva.”
Kiparsky (1991: 350-351) criticizes this solution, saying that using such arguments, one could have arrived at exactly the opposite conclusion. He says, “So (Cardona’s statement) is compatible with two different procedures yielding opposite results:
“If the environments of R2 are properly included in the environments of R1, and the operands of R1, are properly included in the operands of R2, then
a. R2 blocks R1, (for the environments of R2 are properly included in the environments of R1, in the shared operand domain).
b. R1 blocks R2 (for the operands of R1 are properly included in the operands of R2 in the shared environment domain).
“In case (a) of Cardona 1970 (p. 57) the two rules are: 7.3.84 sārvadhātukārdhadhātukayoḥ (guṇaḥ) and 7.1.100 ṝta id dhātoḥ. So, Cardona applies procedure a: “in the case of the single shared operand (ṝ) then, 7.1.100 will counter 7.3.84 [sic – this is evidently a slip and he must have meant to say ‘7.3.84 will counter 7.1.100’], since all the contexts of the former are included in those of the later”. If the facts were the other way round (i.e., if the outcome was *stiraṇa), he would have said “in the case of the single shared context (non-kit suffixes), 7.1.100 will counter 7.3.84, since all the operands of the former are included in those of the latter (procedure b)”.
I think that Cardona’s limited blocking principle is similar to my method of dealing with SOI. However, Kiparsky correctly points out that the explanation offered by Cardona is ambiguous. On the other hand, my solution overcomes such ambiguity by following the clearly defined procedure which I have developed and used to tackle all examples of SOI in this thesis.
This brings us to the end of our survey of SOI and DOI examples from derivations of finite verbs and primary derivatives.
163