I meant "grammar" in a very loose and general way. I.e. things about the language other than the sounds of the language. So for example I wondered why in sanskrit should ī ,ū or e when dual terminations (and only when dual terminations) remain unchanged before vowels.. In other words why should whether a termination is "dual" or not effect its being changed by following sounds. Or why should internal sandhi be any different from external sandhi.
one has to look at Sandhi as the product of various historical sound changes, some Indo-Iranian (RUKI rule), some pre-Indo-Aryan, some Vedic…
Thus: ī ,ū or e when dual terminations remain unchanged before vowels.
This is historical: ī, ū are from < i+H (laryngeal), u + H > with regular change to ī,ū. The laryngeal disappeared with regular lengthening of the vowel, but speakers “remembered” the gap it left.
Like French Le Havre, not l’Avre. (from Germanic H-)