- About
- Deep Parsing
- Syntactic Complexity
- Cues for Text-To-Speech
- Question Generation
- Sanskrit parsing
- Contact
- Privacy Policy
2.29
shrutvApyenaM v ... [8.3.23] mo'nusvAraHa
kashchit ... [8.4.56, 8.2.39] vA'vasAne
AshcharyavachchEnamanyaHa sh ... [8.3.15, 1.3.2, 8.2.66] kharavasAnayo visarjanIyaHa
kashchidenamAshcharyavadvadati ... [6.1.72] saNhitAyAm
kashchidenamAshcharyavadvadati ... [8.2.39] jhalAm jasho'nte
AshcharyavachchEnamanyaHa ... [8.4.40, 8.2.39] stoHa shchunA shchuHa
tathEva ... [6.1.88, 1.1.1] vRuddhirechi
chAnyaHa ... [6.1.101] akaHa savarNNe dIrghaHa
chEnamanyaHa ... [6.1.88, 1.1.1] vRuddhirechi
chEva ... [6.1.88, 1.1.1] vRuddhirechi
Ashcharyavatpashyati ... [6.1.72] saNhitAyAm
shrutvApyenam ... [6.1.72] saNhitAyAm
shrutvApyenam ... [6.1.101] akaHa savarNNe dIrghaHa
shrutvApyenam ... [6.1.77] iko yaNNachi
anyaHa shRuNNoti ... [8.3.36, 8.3.15, 1.3.2, 8.2.66] vA shari
anyaHa sh ... [8.3.15, 1.3.2, 8.2.66] kharavasAnayo visarjanIyaHa
Underlying: Ashcharyavat pashyati kashchit enam Ashcharyavat vadati tathA eva cha anyas , Ashcharyavat cha enam anyas shRuNNoti shrutvA api enam veda na cha eva kashchit .
COMMENTS:
- In clauses A.1 and A.5, the verbs ('pashyati' and 'veda') are supposed to be in the III-Person-Singular-Present conjugation and should, in theory, agree in Person and Number with the Subject. However, the only term that has the possibility of being the Subject in either clause is 'kashchit' ('anyone'), a term that is defined as an Indeclinable in both [MM2015] and [KK2015]. Given the need for agreement in Person and Number between the Subject and the Verb, we have defined an additional term in the Lexicon for 'kashchit' as a NOM-S-Masculine ('kashchit' is formed from 'kaHa+chit', where 'kaHa' is the NOM-S-Masculine declension of the interrogative 'ka'). This Nominal form of 'kashchit' is shown in clauses A.1 and A.5, agreeing with the Person and Number of the verbs therein (i.e. 'pashyati' and 'veda'). Had it been defined exclusively as an indeclinable (as in [MM2105] and [KAL2015]), we would have been required to provide an explanation for the agreement in Person and Number between an Indeclinable and a Verb. As this would go against the general rules of Sanskrit grammar, such an explanation would be suspect. Hence, there is clearly a difference of opinion between this author and [MM2015] and [KAL2015] on this topic.
- The 'agentive' role of the indeclinable 'kashchit' requires further study and consultation with other grammar experts. However, there is no denying the fact that agreement on Person and Number is required between a Verb and its Subject in Sanskrit. While there may be special conditions governing the situations where these constraints may be waived, the fact remains that the verb 'veda' has two conjugations with the identical form (i.e. III-Singular-VerbPresent and I-Singular-VerbPresent), and the correct conjugation can only be selected using agreement with the Subject (i.e. if it is considered to be absent, then the parser would have to assume an elided Subject for a First-Person verb, and therefore wrongly select the I-Singular-VerbPresent conjugation).
- The parser attaches the first ACC-S 'enam' token as the direct object of clause A.1 (of the verb 'pashyati:to see'), and assumes that clause A.2 does not have an object (i.e. assuming that the verb 'vadati:to speak' is in intransitive use in this clause). In a similar manner, [MM2015] [1] attaches ACC-S 'enam' to clause A.1, but also assumes that clause A.2 has a missing object that is shared with the object of clause A.1 (but with no syntactic cue suggesting such sharing). As this token is on the boundary between these two clauses (both of which may be transitive), a syntactic parser must use a heuristic to decide between the two clauses. Note that there is only one candidate ACC term between these two clauses, hence, its attachment to one clause rules out an object for the other clause. While the parser permits sharing of subjects across 'adjacent' clauses, sharing of objects is not permitted in order to avoid egregious errors (for e.g., where one of the verbs must have an Intransitive form). This is a semantic issue that cannot be resolved by a syntactic parser.
- NOTE: The term 'Ashcharyavat' ('equally with wonder') is an Indeclinable (Adverb), being derived with the 'vatiM' affix (and not the 'matup' affix assumed by [KAL2015] [2]). Unfortunately, there is nothing in its form to suggest that it is not derived from the 'matup' affix, nor is this term defined in the major dictionaries (such as [MW2011] [3]). However, treating this term as a 'matup'-derived Adjective ('in whom there is wonder') results in a failure of agreement on Gender between 'anyaHa' (NOM-S-Masculine) and 'Ashcharyavat' (NOM-S-Neuter) in clauses A.2 and A.3. This prompted a closer look at the alternate 'vatiM' taddhita affix that results in Adverbs (unlike the 'matup' affix that results in Adjectives). [MM2015] [1] assumes the 'vatiM' affix, whereas [KAL2015] incorrectly assumes the 'matup' affix. The 'vatiM' affix is applicable when the semantics suggest an Adverb with the semantics 'like that' or 'equally with'. The reader must deduce from the semantic context of the sentence (rather than from the form of the term) whether the term has a 'matup' affix or a 'vatiM' affix.
A: Ashcharyavat pashyati kashchit enam Ashcharyavat vadati tathA eva cha anyaHa Ashcharyavat cha enam anyaHa shRuNNoti shrutvA api enam veda na cha eva kashchitA.1:
Ashcharyavat:Indeclinable pashyati:III-S:dRush:1:P:VerbPresent kashchit:NOM-S:kashchit:Masc.:Noun* enam:ACC-S:idam:Masc.:Pronoun A.2:
Ashcharyavat:Indeclinable vadati:III-S:vad:1:P:VerbPresent tathA:Indeclinable eva:Indeclinable cha:Indeclinable anyaHa:NOM-S:anya:Masc.:Pronoun A.3:
Ashcharyavat:Indeclinable cha:Indeclinable enam:ACC-S:idam:Masc.:Pronoun anyaHa:NOM-S:anya:Masc.:Pronoun shRuNNoti:III-S:shru:5:P:VerbPresent A.4:
shrutvA:-:shru:1:P:VerbGerund api:Indeclinable enam:ACC-S:idam:Masc.:Pronoun A.5:
veda:III-S:vid:2:P:VerbPresent na:Indeclinable cha:Indeclinable eva:Indeclinable kashchit:NOM-S:kashchit:Masc.:Noun*