On the main metarule of the Sanskrit grammar of Pāṇini

Ulanskii Evgenii Aleksandrovich ORCID: 0000-0002-9710-8979 PhD in Physics and Mathematics Postgraduate student of the Department of Indian Philology of the Institute of Asian and African Studies of Lomonosov Moscow State University 119991, Russia, g. Moscow, ul. Leninskie Gory, 1 ulanskiy@mail.ru Other publications by this author

The grammar of Sanskrit Adhy?y? "Octateuch", compiled by the ancient Indian linguist Panini (PIni, ~ V century BC) in the genre of sutra (s?tra), is a sequence of short and succinct statements, each of which is traditionally also called a sutra. When formulating the sutras, Panini used numerous abbreviations, symbols and other information compression techniques. In this regard, in addition to the sutras that define and describe the rules of Sanskrit grammar, paribhashasutras (paribhs?tra) — sutras of an interpretive nature were included in the "Octateuch". They reveal the principles of composing the grammar of Panini, describe its metalanguage and provide guidance on reading and interpreting all other sutras ^[1].

The concept of interpretative sutras was developed in the Shrauta Sutras ^[2-5], from where it was presumably borrowed by Panini. The most complete list of Panini Paribhashasutras is compiled by Abhyankar ^{[6, p. 4-5]}. We have translated and annotated the Abhyankar-listed and some other Paribhashasutras ^[7].

Due to the extreme conciseness of the Panini sutras, even with the presence of paribhashasutras, there were places in the "Octateuch" that allowed ambiguous interpretation, in connection with which commentators — Katyayana (K?ty?yana, ? III century BC), Patanjali (Patanjali, II century BC) and their followers — formulated additional meta-guidelines that allowed correctly interpret the Panini sutras and resolve conflicts between the grammatical operations prescribed by them ^[1]. The Katyayana commentaries are sutra-like utterances called varttikas, with the help of which the essence of a detailed commentary-vritti - is usually stated, but in the case of Katyayana, only varttikas have reached us. References to the varttika indicate the number of the book, chapter and sutra of the "Octateuch" to which it refers. It is customary to refer to the Panini sutras themselves with the same three numbers. The Great Commentary compiled by Patanjali (Mah?bhYam) interprets in great detail both the sutras of the "Octateuch" and the varttikas of Katyayana.

The meta—directions formulated by commentators were called "paribhasa" (paribh) - this term, not used by Panini, originated in two varttikas of Katyayana (1.1.69 vartt. 4 and 1.3.11 vartt. 3) and Patanjali was actively used. Many varttik Katyayana are metapravils or references to metapravils, from which it can be concluded that they were already in use among grammarians of that time. Patanjali not only comments on these varttikas, but also cites hundreds of new rules of an interpretative nature, many of which were later included by Sanskrit grammarians in the consolidated lists of meta-rules necessary for the interpretation of the "Octateuch". Such codes have been compiled for many centuries both by adherents of the grammatical tradition of Panini, and in other grammar schools (Shakatayana, Chandra, Katantra, Jainendra, etc.), who developed their own sutras, processing the work of Panini. In many cases, these consolidated lists of metaprograms were accompanied by comments not only interpreting the metaprograms, but also illustrating their application with concrete examples. The most authoritative commented set of metapravels is the "Lunar Diadem of Interpretation" (ParibhEndu?ekhara), compiled by Nagojibhatta (N?goj?bhaa) in the XVIII century AD, although the first such works appeared at least at the beginning of our era ^{[6, pp. 5-11]}.

Metapravila 37-69 from Nagojibhatta's list is the central section of his entire work. They speak of the superiority of one Panini prescription over another in cases where there are no visible obstacles to the application of both ^{[8, p. 140]}. Such meta-guides are sometimes called bal?balas?tra "the sutra about power and its absence" ^{[6, p. 19]}. The culmination of this section is the meta-rule 50 asiddam bahira?gam antara?ge "(The rule) of the bahiranga appears as not applied during the application of the (rule) of the antaranga" ^{[8, p. 142]}, known as the bahiranga-paribhasha (bahira?gaparibh). The concepts of antaranga and bahiranga are interdependent and situational and will be defined by us a little later.

The purpose of this article is to give a detailed explanation of the bahiranga—paribhasha in the edition of Nagojibhatta, as well as to highlight the history of the mention of this meta-rule and its interpretation in Sanskrit grammatical works.

Grammarians justify the necessity of each meta—rule in different ways - by logical reasoning, the general law of the world order, simply by the authority of Katyayana or Patanjali, who voiced this meta-rule, as well as by using a sign ^{[8, pp. 136-139]}. A sign (jpaka) is considered an element that seems, at first glance, redundant in the "Octateuch", which is unacceptable, given Panini's desire for maximum brevity of formulations. However, if we assume that Panini took into account a certain meta-rule, then this element not only ceases to be redundant, but also becomes absolutely necessary. The meta-guide, which Panini was supposed to keep in mind, is declared "achievable with the help of a sign" (jpakasiddh?) ^[9-13]. Kielhorn gives the distribution of metapravils from the Nagojibhatta list according to the method of justifying their necessity ^{[12, pp. vii-xvii]}. After some preparations, we will describe what is considered to be a sign of bahiranga-paribhashi.

Sutra 6.4.131 vaso? sa?pras?ra?am prescribes the replacement of the type "samprasarana" (sa?pras?ra?a) in the suffix vas, which serves to form the basis of the perfect participle. Replacement of "sampradaya" for those times when it is not explicitly specified, is determined by the Sutra 1.1.45 i"k" ya"?"a? sa?pras?ra?am "near replacement "of sampradaya" means the substitution of vowels i u this ? (as well as ? ? ? ?) is polyglossic y v r l (respectively)". The brackets "" here and further highlight grammatical marks — one of the metalanguage techniques of Panini, which allows, among other things, to abbreviate groups of phonemes, suffixes, etc. In this case, a group of four vowels is briefly designated i"k", and of four semivowels — ya"?". In order to properly understand Sutra 6.4.131, it is also necessary to take into account the heading sutra 6.4.129 bhasya. The action of this sutra continues until the end of the 4th chapter of the 6th book of the "Octateuch" and implies that any operation prescribed by the sutras of this section is carried out only for an element that can be assigned the technical status of bha. This status is defined in sutras 1.4.18-19 and is assigned to any element to which the endings of the nominal declension, suffixes that serve to form the basis of the feminine gender, or suffixes of secondary word formation starting with a vowel or with y are attached (with the exception of the endings of the nominative case or the endings of the singular and dual numbers of the accusative case). Thus, sutra 6.4.131 reports that the substitution of u instead of v in the suffix vas occurs when suffixes are attached to it, allowing technically to name it bha. This sets the context for the next sutra as well.

Sutra 6.4.132 v?ha? ?"?h" prescribes for the basis of v?h the substitution of "?h" (as before, in cases where this basis can be assigned the technical status of bha). According to one of the standard procedures for the "Octateuch", we must supplement sutra 6.4.132 with the word sa?pras?ra?am from sutra 6.4.131. This allows us to establish that the "?h" in the base of v?h is substituted instead of v. The basis of v?h is formed from the verb root vah when it appears at the end of compound words. This is the root of the verb, which has the meaning "to carry, to carry; to winnow; to endure, to endure", and it can only be part of complex words constructed similarly to the Russian "water-who" or "dry-wei". In such complex words, the verb root, which appears in the last position, expresses the participle "carrying" or "blowing" in meaning. In the grammar of Panini, the verb root cannot in its pure form enter a compound word, and one of the suffixes specially designed for such cases must be attached to it. In the case of the root vah, it is the suffix "?"v i prescribed by sutra 3.2.64. Here "?" is a grammatical litter indicating, according to sutra 7.2.116, the lengthening of the vowel in the root vah. The suffix itself, without a grammatical mark, has the form v i, but only conditionally, because according to sutra 6.1.67, this v i is replaced by zero. As a result, at the end of compound words, the verb root vah appears in the form of vah.

The function of the grammatical mark "?h" in the "?h" is demonstrated by an example. When constructing the basis of the compound word praha-v?h "draft bull" (lit. "lucky at the head") the verb root vah and the suffix "?"v i are attached to the base of praha "front, main". To form the plural form of the accusative case, the ending a s is attached to the base of prahav?h. This ending begins with a vowel, which makes it possible to assign the basis of prahav?h the technical status of bha (see above). Then, according to the already considered sutra 6.4.132, the "?h" is substituted for v in v?h:

praha-v?h+as ?praha-?‹?h›?h+as.

As already mentioned, the replacement made is called "samprasarana", and for it it is necessary to take into account sutra 6.1.108, according to which the vowel that was directly behind the semivowel that was replaced by "samprasarana" is removed (in our case, the vowel a that followed v is removed):

praha-?‹?h›?h+as ?praha-?‹?h›h+as.

At this point it becomes essential that instead of v, the vowel was substituted, not by itself, but with the grammatical mark "?h". The fact is that Sutra 6.1.89 prescribes the substitution of the diphthong a u instead of a combination of the vowel a and the vowel y, equipped with the grammatical mark "?h":

praha-?‹?h›h+as ?prahauh+as.

Whereas without the grammatical mark "?h", instead of a combination of a vowel and a vowel, o would be substituted according to sutra 6.1.87, which would lead to an incorrect construction. Finally, as a result of applying the rules set by sutras 8.2.66 and 8.3.15 to the final s, the final result praha u h a? is obtained.

Katyayana considered the element "?h" redundant in the text of sutra 6.4.132 v?ha? ?"?h", believing that one v?ha? would be enough. Consider his varttik to sutra 6.4.132.

v?ha? ?"?h" vacan a-anarthakyam sa?pras?ra?ena k?tatv?t (vartt. 1) "The utterance "v?ha? ?"?h""is useless, since (the necessary construction) is accomplished by replacing "samprasaran"."

gu?a? pratyaya-lak?a? atv?t (vartt. 2) "Substitution of "gun" (is carried out) according to the suffix rule."

e"c"-graha v?ddhi? (vartt. 3) "(Is) substitution of "vriddhi" prescribed by the wording "e"c""".

According to the very first sutra 1.1.1 of the "Octateuch", when substituting "vriddhi", the replacement element can be a or one of the two diphthongs a i and a u. According to Sutra 1.1.2, when substituting "guna", the replacement element can be a, e or o. The suffix rule mentioned by Katyana is set by sutra 7.3.86, prescribing the substitution of "guna" instead of the penultimate short vowel of the verb root in the case of a suffix attached to it. The prescription with the wording "e"c"" is given by sutra 6.1.88 V?ddhi? e"c"i, according to which, instead of a combination of a vowel a and a vowel e or o, "vriddhi" is substituted, that is, in this case, respectively, a i or a u.

According to Katyayana, one word v?ha? would be sufficient for the formulation of sutra 6.4.132, which would prescribe the replacement of "samprasaran" for the basis of v?h in cases where it can be assigned the technical status of bha. By default (see the paragraph about sutra 6.4.131 above), this would mean substituting u instead of v in v?h:

praha-v?h+as ?praha-u?h+as.

Then, according to the already mentioned sutra 6.1.108, the vowel a following v would be deleted:

praha-u?h+as ?praha-uh+as.

At this stage, the vowel u would be in the penultimate position in the verb root, and the suffix "?"v i attached to this root, although it was replaced by a zero, according to sutra 1.1.62 continues to have the same effect as if it were still present. Therefore, all the conditions for the prescription of sutra 7.3.86 are met, namely the substitution of "guna" instead of u. According to Sutra 1.1.50, o is substituted instead of u from the substitutive elements a, e and o when substituting "guna":

praha-uh+as ?praha-oh+as.

After that , according to sutra 6.1.88 , instead of a combination of the vowel a and the vowel o , the diphthong au would be substituted:

praha-oh+as ?prahauh+as.

It can be seen that the result would have been the same even after removing the "?h" from the wording of sutra 6.4.132. The question arises, what made Panini lengthen the wording of this sutra? The answer is given by Patanjali in the following lines, commenting on the above series of varttik Katyayana.

tarhi siddhe evam sati yat v?ha? ?"?h"am sti tat jpayati ?c?rya? bhavati e paribh asiddham bahira?ga-lak?a?am antara?ga-lak?a?e iti "Because reaching thus in this case (proper construction) teacher (still) requires "v?ha? ?"?h""he wants to indicate the presence of such metaplasia "Rule bahiranga appears as unapplied during the application of the rules antaranga"".

Since Patanjali did not give additional explanations, it can be concluded that he considered the construction of Katyayana impossible due to the fact that one of his steps did not take into account the effect of bahiranga-paribhashi. After reviewing Nagojibhatta's definition of the concepts of antaranga and bahiranga, we will see exactly at which step this happened.

A deep analysis of Nagojibhatta's commentary on the bahiranga paribhasha was carried out by Bronkhorst in his dissertation ^[14]. As Cardona rightly notes, for an objective assessment of Bronkhorst's interpretation, it is necessary to make an extensive study of all the examples given by Nagojibhatta, as well as many passages of the "Great Commentary" ^[15]. However, as Kulikov writes in his review, this work should in any case be of interest to Sanskritologists and specialists in the history of linguistics ^[16]. Indeed, despite the existence of a translation of Nagojibhatta's work, including his commentary on the bahiranga paribhasha ^[12], Bronkhorst's work is extremely useful, since in it he scrupulously collected all the fragments of Nagojibhatta's extensive commentary, allowing to formulate the definition of the concepts of antaranga and bahiranga in its entirety, and also checked the correctness of the received definition on all considered by Nagojibhatta examples, sometimes involving other grammatical works of Nagojibhatta and even his commentators for analysis. We will rely on the analysis made by Bronkhorst, setting out the definitions of Nagojibhatta.

The Sanskrit adjectives antara?ga and bahira?ga are possessive compound words (bahuvrihi), in which the preposition antar "inside" and the postposition bahi s "outside, outside" are combined with the nominal basis a?ga. The noun a?ga Nagojibhatta interprets as "cause". The concepts of antara?ga and bahira?ga themselves refer to grammatical rules or operations prescribed by these rules, and "reason" in this case means a condition provoking the application of a rule or the implementation of an operation. In the Panini system, the names of the elements that set the context of the grammatical operation are mentioned in the sutras in the form of the adjunctive, genitive, local and occasionally the creative case. Nagojibhatta clarifies that in order to assign the antarang or bahirang status to a rule, its reason must be expressed in the Panini sutras in the local or other case. Bronkhorst, after a thorough study of all the examples considered by Nagojibhatta, excluded cases of specifying the cause by the forms of the adjunctive and genitive case and concluded that the only possibility remains the local case. As a possible explanation of Nagojibhatta's words "in the local or other case", Bronkhorst suggests Nagojibhatta's foresight in case of some examples of the use of bahiranga-paribhashi that did not occur to him ^{[14, p. 21]}. We can assume that Nagojibhatta could also imply the creative case, with which the context of performing grammatical operations is sometimes set (for example, in sutras 8.4.40-41). It is important to mention one clarification of Nagojibhatta regarding the element, the mention of whose name in the form of the local (or creative) case becomes the reason for performing a grammatical operation. This element should be considered exclusively formally, as a sequence of sounds, without reference to the meaning that it can carry. In most of the Panini sutras, this principle is observed by agreement set by sutra 1.1.68, but there are exceptions when it is the meaning carried by the element that can provoke the performance of grammatical operations. Nagojibhatta excludes the reasons of this kind from the definition of the noun a?ga, and hence from the definition of the concepts antaranga and bahiranga.

We are ready to define the concepts of antaranga and bahiranga. These concepts are interdependent and in order for any rule to receive the status of an antarang, it is necessary to have a second rule that would simultaneously receive the status of a bahirang, and vice versa. A rule gets the status of an antarang if the set of reasons provoking its execution is strictly contained in the set of reasons provoking the execution of some other rule, which at the same time receives the status of a bahirang. Thus, antara?ga can be literally translated as "who has causes inside", and bahira?ga as "who has causes outside". The strict content of one set in another means that there are more elements in the second set than in the first, and that among the elements of the second set there are all the elements of the first set. This definition also fits the case when there are no elements at all in the first set, that is, it is empty, whereas in the second there is at least one element. The concept of cause here should be understood only in the sense described above, that is, as an element considered formally — in the form of a sequence of sounds, and the name of which is used in the sutra in the form of the local or creative case. Nagojibhatta complements the definition by assigning the status of antarang and bahirang, respectively, to two rules, the first of which refers only to one word (or the basis of the word in the composition), and the second to two or more.

It is time to explain in the interpretation of Nagojibhatta the above formulation of the bahiranga-paribhashi "(The rule) of the bahiranga appears as not applied during the application of the (rule) of the antaranga." If two rules are assigned the status of antarang and bahirang and each of them can be applied at the moment with some grammatical construction, then preference should be given to the antarang rule. If the bahirang rule has already been applied, and then the antarang rule is applied, then at the time of its application it should be assumed that the bahirang rule has not yet been applied ^{[14, p. 54]}.

The last point should be explained by the example of the construction of the plural form of the accusative case from the base prahav?h considered above. According to Patanjali, because of the bahiranga-paribhasha, the construction proposed by Katyayana was impossible. Let's recall the first steps of this construction:

praha-v?h+as?praha-u?h+as?praha-uh+as?praha-oh+as.

The first substitution (u instead of v) in this chain was made according to sutra 6.4.132, the wording of which Katyayana proposed to reduce to a single word v?ha?, removing from it the element "?h" that seemed to him redundant. The reason for the implementation of this replacement is the ending of as, thanks to which the basis of v?h acquires the status of bha, which makes the replacement possible. The reason for the implementation of the last replacement (o instead of u) according to Sutra 7.3.86 is the null suffix, which replaced the suffix "?"v i at the stage of building the basis of prahav?h. The reason in the form of a null suffix does not fall under the criterion of Nagojibhatta, according to which the reason should be an element in the form of a sequence of sounds. Therefore, formally, the rule set by Sutra 7.3.86 has no reason, and the rule set by sutra 6.4.132 has one reason. This allows you to assign the first of these rules the status of antarang, and the second — bahirang. According to the bahiranga-paribhashi condition, since the substitution of u instead of v, which has the status of bahiranga, has already been carried out, then at the time of the implementation of the substitution of o instead of u, which has the status of antaranga, it must be assumed that there was no substitution of u instead of v. But then the substitution of o instead of u itself becomes impossible due to the absence of the substituted element.

As can be seen, the elimination of the "?h" from the wording of sutra 6.4.132 does not allow us to present a correct construction of the form praha u h a?, if we take into account the bahiranga-paribhasa. Whereas without the need to resort to this meta-rule, the construction of Katyayana would be satisfactory. In this regard, Sanskrit grammarians consider the presence of the "?h" in the wording of sutra 6.4.132 to be a sign of bahiranga-paribhashi (in the understanding of the sign described above).

In the example considered, the set of reasons for the antarang rule was empty and naturally contained in the set of reasons for the bahirang rule. In other examples suggested by Nagojibhatta, the same pattern is observed. An example when both of these sets are non-empty is found in one of the commentaries to the work of Nagojibhatta, in which the construction of the form of the 3rd person singular aorist atihi pat is given from the causative verb sth?payat i "to put", derived from the verb root sth? "to stand". In this construction, two rules are applied that have the status of antarang and bahirang, and the first of them has one reason, which is simultaneously one of the two reasons for the second ^{[14, p. 56]}.

The predecessors of Nagojibhatta included in their lists another meta-rule relating to a pair of rules of antarang and bahirang: antara?ga-bahira?gayo ? antara?ga m bal?ya? "From the rules of antarang and bahirang more powerful than antarang". This meta-rule gives priority to the antarang rule in the case when each of the antarang and bahirang rules can be applied at a given moment with some grammatical construction. As we have seen, Nagojibhatta considers this to be an integral part of the bahiranga-paribhasha, so he does not include this meta-rule in his list. At the same time, it expands the definitions of the concepts of antarang and bahirang for the case of simultaneous applicability of the rules (this extension does not apply to the case when the antarang rule is applied after the application of the bahirang rule and does not take into account the effect of its application). In this expanded sense, two rules receive the status of antarang and bahirang if, when constructing, the reason for the first of them precedes the reason for the second, or if the first rule refers to a logically more important stage of construction than the second ^{[14, p. 54]}. That is, either the reason for the antarang rule is located in the constructed word form to the left of the reason for the bahirang rule, or the anatarang rule is applied at an earlier stage of word form construction in the following sequence: changes in the root of the word, word-formation suffixation, inflectional suffixation, prefix, word composition. Such an extension of the definition of antaranga and bahiranga Nagojibhatta justifies the general law of the world order: waking up in the morning, a person first performs his own affairs, then obligations to friends, after that — to relatives ^{[12, p. 261]}. It is significant that in the grammar of Panini prefixes are considered independent, and the prefix is a kind of word composition. Nevertheless, in the logical sequence of the synthesis of word forms, it is preferred over the word structure in the usual sense. This is confirmed by the mention of the prefix a"?" in the wording of sutra 6.1.95 om-a"?"o ? ca. This sutra sets an exception to the already mentioned general phonetic rule (sutra 6.1.88), according to which, instead of a combination of a vowel a or a and a vowel o, the diphthong a u is substituted. If the vowel o is part of the om mantra or was formed (according to sutra 6.1.87) from a combination of the prefix a"?" with a vowel u or a, then in combination with a vowel a or a and such a vowel o, this vowel o is preserved, and the vowel a or a is elided. If Panini had not given priority to the prefix before the word composition, then the mention of the prefix a"?" in the wording of Sutra 6.1.95 would have been redundant. Let's show this by the example of constructing the basis of the compound word adyo?h a "brought today" from the adverb adya "today", the prefix a"?", meaning "to, at", and the basis of the participle a?ha "brought", formed from the already mentioned verb root vah. According to sutra 6.1.101, instead of a combination of the vowel a ending in adya and the vowel a prefix a"?", the vowel a would be substituted:

adya+?‹?›+??ha?ady?+??ha.

The grammatical litter "?" has been omitted here, since grammatical litters refer to metalanguage reality. With specific constructions, they are discarded according to sutra 1.3.9. Further, according to sutra 6.1.87, instead of a combination of the vowel a and the vowel a, the vowel o would be substituted:

ady?+??ha? adyo?ha,

which would lead to the desired result. However, the prefix is an antarang operation in relation to the word composition operation, and it must be carried out first:

adya+?‹?›+??ha?adya+o?ha.

Then, to obtain the desired result adyo?ha mention of the prefix ?"?" in the formulation of the Sutra-exclusion 6.1.95 already absolutely necessary, and predecessors Negogiate saw it as a sign of metaplasia antara?ga-bahira?gayo? antara?gam bal?ya?. 

Prior to Nagojibhatta, the definition of antaranga and bahiranga was not formulated strictly. Katyayana, despite the assumption of redundancy of the "?h" in the wording of sutra 6.4.132, was familiar with the bahiranga paribhasha.  In his varttikas, he 23 times cited the argument na v a bahira?galak?a?atv?t "however, this is incorrect due to the rule on the status of the bahirang", but without further explanation ^{[17, footnote 246]}. Patanjali, commenting on these references of Katyayana, as well as in his own arguments, gave the full formulation of the bahiranga-paribhasha. He also gave many examples of the application of this meta-rule, but did not offer its exact interpretation and did not explicitly formulate the definition of the concepts of antaranga and bahiranga.

The first definition we find in the commentators of the Katantra-paribhashasutras. Durgasimha (? IX-X centuries AD) defines antaranga and bahiranga as relying on a smaller and larger number of causes, respectively. Most likely, this definition does not imply at all that the reasons for the antarang rule are contained among the reasons for the bahirang rule, as can be seen from the examples of constructions carried out by various grammarians before Nagojibhatta. He himself criticizes such a definition: "No laws of the language for the words antara?ga and bahira?ga can be obtained the meaning of "relying on less and more"" ^{[12, p. 228]}. It is all the more surprising that Vaidyanatha Payagunda, the commentator of Nagojibhatta, who is considered his direct disciple, returns to this definition ^{[14, pp. 143-146]}. Another commentator of the Katantra-paribhashasutras, Bhavamishra (? X century A.D.), adds another condition: the rule relating to the basis has the status of antaranga, and the rule relating to the affix — bahiranga. Both commentaries are available in the collection of Abhyankara ^[6].

The closest to the interpretation of Nagojibhatta came Abhyanandin (?X-XI centuries AD), who gives the following conditions in the commentary to the grammar of Jainendra. The reasons for the antarang rule are among the reasons for the bahirang rule. The reasons for the antarang rule are less than the reasons for the bahirang rule. A rule relating to one word has the status of an antarang, and to two or more — a bahirang. However, it should be noted that Abhyanandin did not specify what exactly can be considered as the reason for the rule. Abhyanandin's commentary can be found in the same collection ^[6].

In the famous treatise "Siddhantakaumudi" ^[18], in which the famous grammarian Bhattoji Dikshita (XVI-XVII centuries AD) thematically rearranged and annotated the Panini sutras, the bahiranga-paribhasha is contained at the end of the chapter on metapravils. It is also quoted five times in the Kashikavritti (VII century AD) ^[19] — the first commentary that has come down to us, in which every Panini sutra is commented on. But in both of these works, only the formulation of the bahiranga-paribhasha is given.

The following conclusions can be drawn. Although the historical prerequisites for the interpretation of Nagojibhatta bahiranga-paribhashi existed in the commentaries of Durgasimha, Bhavamishra and, especially, Abhyanandin, the formulation of this meta-rule, elaborated, verified and tested on many examples, became a revolutionary event among Sanskrit grammarians. To the greatest extent, this was facilitated by the fact that Nagojibhatta for the first time strictly defined the concept of cause (a?ga), on which the definition of the concepts of antaranga and bahiranga significantly depends. It is not surprising that the "Moon Diadem of Interpretation" has become a canonical work in the field of interpretation of Panini grammar.