3.11

parasparaM bh ... [8.3.23] mo'nusvAraHa
devA ... [8.3.17, 8.3.22, 1.3.2, 8.2.66] bho bhago agho apUrvasya yo'shi
bhAvayantaHa sh ... [8.3.15, 1.3.2, 8.2.66] kharavasAnayo visarjanIyaHa
shreyaHa p ... [8.3.15, 1.3.2, 8.2.66] kharavasAnayo visarjanIyaHa
devAnbhAvayatAnena ... [6.1.72] saNhitAyAm
devAnbhAvayatAnena ... [6.1.101] akaHa savarNNe dIrghaHa
paramavApsyatha ... [6.1.72] saNhitAyAm
bhAvayantaHa shreyaHa ... [8.3.36, 8.3.15, 1.3.2, 8.2.66] vA shari

Underlying: devAn bhAvayatA bhAvayata anena te devAs bhAvayantu vas , parasparam bhAvayantas shreyas param avApsyatha .

COMMENTS: This stanza is difficult to parse correctly because of a term that has definitions both as a Verb conjugation ('bhAvayata:VerbImperativeCausative') as well as a present participle Nominal declension ('INS-S bhAvayatA').

• The sandhi analysis stage chose the incorrect candidate 'bhAvayatA' while splitting the euphonic combination 'devAnbhAvayatAnena'. This was a case of inherent syntactic ambiguity, where there were multiple, valid ways of splitting an euphonic combination. However, the subsequent parsing stage was able to identify the correct term ('bhAvayata:VerbImperativeCausative'), after the term was marked as being unresolved (i.e. marked as 'bhAvayatA/bhAvayata').
• The short pronouns ('te' and 'vaHa') are also difficult to handle correctly, as a syntactic parser cannot definitively select the correct declension from multiple Case/Number possibilities, as this usually requires a semantic understanding of the context.