Gemination of the numeral ti

Dear Ole,

Thanks.
> <However, the doubling doesn't always take place, e.g., atirekatirattam,
> dirattatiratta, ekuunati.msa. Why is this so?>
>
> My own explanation would be that the compilers of the Paatimokkha tried to
> avoid excessive gemination which invariably would entail prosodical length
> and thus would have rythmical implications>
>  
Only one of the three compounds (dirattatiratta) mentioned is found in
the Patimokkha. There are many other occurrences of compounds starting
with atirekati-, ekuunati-, etc, all over the Canon, commentaries, etc.
Maybe I misunderstand ''prosodical length'', but the Patimokkha is not
prosody.
The non-gemination apparently only takes place when ti is found in
compounds with more than 2 members. Maybe there is a rule about not
letting this gemination take place in compounds consisting of more than
two members. The gemination would make the compounds more difficult to
pronounce.
This being written, I noticed that in the compound dvattikkhattu.m
gemination occurs twice—both in ti and in khattu.m.

There is also the reading dvi- instead of dva- Not only here but also in
other compounds like dvattipatta. I wonder which one is more authentic.
Why are there various compound forms of the numeral 2, i.e., dvi, di and
dvá? The di- form appears to be the most prakritic.

> The reading baat.msa(.m) appears to be exclusively Burmese. The references I
> have checked read as one would expect batti.msa.
>  
Batti.msa is also found on CSCD, but mostly in the Sinhalese texts.

Kind regards,
                           Bh. Nyanatusita