Sandhi Statistics

Presented below are the results of an automated sandhi analysis of Chapter 4 of the Srimad Bhagavad Gita. All the sandhi sutras applicable to Chapter 4, as well as one example of each sutra, are listed on the next page.

As discussed in the Sandhi Statistics of Chapter 1, a grammatical analysis of a Sanskrit sentence or stanza must analyze every word in a sentence, as Paninian sandhi sutras transform (or choose not to transform) underlying terms (declensions, conjugations, and indeclinables) into 'single words' as well as 'combination words'. Hence, we use the broader term 'sandhi analysis' instead of 'sandhi splitting'.

The Columns labelled 'False Negative' and 'False Positive' are explained in the Sandhi Statistics of Chapter 1.

As will be noted from the above table, the number of errors in the 'sandhi analysis' is reasonably small (False Negatives < 1% and False Positives < 1%). Most of the errors are due to inherent ambiguity, and some of these errors can be rectified by a syntactic parser that is run as the next step in the grammatical analysis.

The following is a summary of the False Negatives and False Positives discussed in the table above.

A. Row (A) above shows that 107 'combination words' were split into 245 underlying terms. Of these 245 underlying terms, 2 were incorrect (matching False Negatives and False Positives).

In Stanza 4.26, the combination word 'shabdAdInviSHayAnanya' was correctly split into 'shabAdIn+vishayAn+anya', but the sandhi analyzer incorrectly assumed that the final term 'anya' was derived from 'anyaHa' instead of from 'anye'. In other words, the sandhi analysis should have assumed the application of the 6.1.78-echoByavAyAvaHa sutra (with the 8.3.19-lopaHa shAkalyasya elision sutra) instead of assuming the 8.3.17-bho bhago agho apUrvasya yo'shi sutra (with the 8.3.19-lopaHa shAkalyasya elision sutra). This incorrect decision is due to inherent syntactic ambiguity. A syntactic parser that runs as the next step in grammatical analysis may be able to rectify this error.

In Stanza 4.28, the combination word 'yogayajnyAstathApare' was split into 'yogayajnyAHa+tathA+pare' instead of 'yogayajnyAHa+tathA+apare'. Note that the difference is only in the final term 'pare' and 'apare', both of which are valid underlying terms and valid derivations. If the sandhi analyzer had instead used the 6.1.101-6.1.101 akaHa savarNNe dIrghaHa sutra, it would have obtained the correct underlying term 'apare'. This incorrect decision is due to the fact that there were multiple valid ways to split the combination word, and the sandhi analyzer made the wrong choice in this particular case. This cannot be resolved other than through an expert semantic analysis (even a detailed syntactic analysis would not be able to decide between the two valid terms presented).

B. Row (B) shows 2 False Positives.

In Stanza 4.3, 'sa' was wrongly assumed to have been derived from 'se' (a rare conjugation and an even more uncommon declension) instead of from 'sas'. In other words, the sandhi analysis should have assumed the application of the 8.3.17-bho bhago agho apUrvasya yo'shi sutra (with the 8.3.19-lopaHa shAkalyasya elision sutra) instead of assuming the 6.1.78-echo'yavAyAvaHa sutra (with the 8.3.19-lopaHa shAkalyasya elision sutra). This incorrect decision is due to inherent syntactic ambiguity. A syntactic parser that runs as the next step in grammatical analysis may be able to rectify this error.

In Stanza 4.32, the sandhi analyzer incorrectly assumed that 'bahuvidhA' was derived from the underlying term 'bahuvidhAs' by the application of the 8.3.17-bho bhago agho apUrvasya yo'shi sutra (with the 8.3.19-lopaHa shAkalyasya elision sutra). However, in this case, the underlying term was an unchanged 'bahuvidhA' instead. This incorrect decision is due to inherent syntactic ambiguity. A syntactic parser that runs as the next step in grammatical analysis may be able to rectify this error.

C. Row (C) shows a single False Negative, which matches the False Positive discussed in point B above i.e. Stanza 4.32.

A closer look at the errors in this Chapter (and in previous Chapters) will indicate that most of the errors are due to the assumption of an elision sutra (either the 8.3.22-hali sarveSHAm elision sutra, or the 8.3.19-lopaHa shAkalyasya elision sutra) in the derivation (from 'printed form' to underlying term). The application of these two elision sutras results in ambiguous 'printed forms' that cannot be easily traced back to their underlying terms without a complete grammatical analysis.