लेट् लकारः is showing the same as लट्

[Stormchaser5775]

Right now the the लेट् लकार forms are showing the same as लट्. However, this is not the case. According to Whitney §562 it should look like this:

आति/आत् | आतः | आन्ति
आसि/आः | आथः | आथ
आनि | आव | आम

आते/आतै | ऐते | ऐन्ते/अन्त/
आसे/आसै | ऐथे | आध्वै
ऐ | आवहै | आमहै

«First person endings are the same as लोट् first person as लोट् borrowed these forms from लेट्»

Could this be fixed?

[akprasad]

Sure, I'll add some basic support here.

[akprasad]

This is more complicated than I expected because many of the लेट् rules are optional or apply only in certain ad-hoc conditions:

• https://ashtadhyayi.com/sutraani/3/1/34
• https://ashtadhyayi.com/sutraani/3/4/96

Since लेट् is a marginal/rare form, I'll deprioritize this for now.

[Stormchaser5775]

Ok, but if it's not going to be fixed could it at least be removed; the improper forms can be misleading. Wiktionary lists a lot of the forms without सिप्. The forms without सिप् are somewhat more recognized then with सिप्.

[Stormchaser5775]

Someone has attempted to fix this in the mots recent update but it is still wrong. The current forms for भू look like this:

However a good table should look like:
भवाति, भवात् | भवातः | भवान् (not ान्ति sorry for the mistake in the previous post)
भवासि, भवाः | भवाथः | भवाथ
भवानि | भवाव | भवाम
Then for atmanepadi:
वन्दाते, वन्दातै | वन्दैते | वन्दन्त, वन्दान्तै
वन्दासे, वन्दासै | वन्दैथे | वन्दाध्वै
वन्दै | वन्दावहै | वन्दामहै

[akprasad]

There was no intentional update to these forms. I'll see if I can add basic support, but right now I'm mainly focusing on krdantas.

[akprasad]

I've implemented some basic rules that slightly improve the quality here, but I have done no serious vetting. If you are excited about this feature, please test more examples and update this issues with the errors you notice.

[Stormchaser5775]

Thank you for helping. Yes, I would really like to have a fully functioning leṭ conjugator to use so I will continue to update here. While mostly correct now, I think other leṭ forms for atmanepadi should also be included (the ones with तै instead of ते). Moreover, the उत्तमपुरुषः forms should be the exact same as loṭ's first person forms. The third person plural for parasmaipadi should be like भवान् not भवान्ति.

[akprasad]

The third person plural for parasmaipadi should be like भवान् not भवान्ति.

Which sutra does this? is it https://ashtadhyayi.com/sutraani/3/4/97 ? Does it hold for all prathama-bahu leT forms?

+cc @neeleshb

[akprasad]

Also, thank you for filing issues. In the future I will be short of time and attention for this project, so I need all of the help I can get to improve the project. See https://groups.google.com/g/sanskrit-programmers/c/zoZ4-9eitSY for details and please tell your friends who are interested in Sanskrit programming or vyakarana.

[Stormchaser5775]

Which sutra does this? is it https://ashtadhyayi.com/sutraani/3/4/97 ? Does it hold for all prathama-bahu leT forms?

I think its that one. I derived the usage of only ान् from Whitney's analysis of लेट्.

[neeleshb]

I would like to put on record that forms of लेट्-लकार should not be considered valid unless their actual usage is attested in the Vedas. Consequently, discussions such as भवान् vs भवान्ति are moot unless such a usage is identified. Once a specific usage is pinpointed, the appropriate सूत्र should simply be applied to derive the form. This is the essence of the term "बहुलम्".

This principle has been emphasized repeatedly in various grammatical texts. For instance, let me quote from अष्टाध्यायी सहजबोध on this matter:

हमनें लेट् लकार के सारे प्रत्यय दिये हैं, किन्तु हमें यह अधिकार नहीं है कि लेट्-लकार के इन सारे प्रत्ययों से हम लेट् लकार के सारे रूप बना डालें। पाणिनीय प्रक्रिया हमारे पास है। हम वेद में लेट् लकार का जो भी रूप पाएं, इस पाणिनीय प्रक्रिया से उसे निष्पन्न कर लें।

Translation: "We have provided all the प्रत्यय-s of लेट्-लकार, but we do not have the authority to generate all its forms using these प्रत्यय-s. We have the Paninian procedure, and it must be used to derive only those forms that are found in the Vedas."

[akprasad]

I see four options:

1. Don't support लेट् at all. Delete Lakara::Let and all rules associated with it.
2. Support लेट् by overgenerating. This may be useful and interesting to people like @Stormchaser5775 , but it goes against the tradition given @neeleshb 's point above.
3. Limit लेट् production to only the specific forms attested in Vedic usage and in the commentaries. We can do this with a lookup table, for example.
4. Default to (3) and let the user opt in to (2) with a config parameter.

(3) and (4) seem like reasonable options. If so, the incremental step forward is to continue adding supporting लेट् sutras and confirming that we generate forms attested by usage and the commentaries.