funderburkjim
commented
6 years ago endings for m_F and n_F models
| Case | S | D | P |
|---|---|---|---|
| Nominative | IH/FH | irO/rO | iraH/raH |
| Accusative | iram/Fm | irO/rO | iraH/Fn |
| Instrumental | irA/rA | IrByAm/FByAm | IrBiH/FBiH |
| Dative | ire/re | IrByAm/FByAm | IrByaH/FByaH |
| Ablative | iraH/ruH | IrByAm/FByAm | IrByaH/FByaH |
| Genitive | iraH/ruH | iroH/roH | irAm/rAm |
| Locative | iri/ri | iroH/roH | Irzu/Fzu |
| Vocative | IH/FH | irO/rO | iraH/raH |
base for m_F model
Obtain base by dropping the final 'F'
m_F declension algorithm
Completely analogous to the other vowel-ending declension algorithms:
- Detach prefix from final pada
- Form base from final pada, as above
- Join base to the endings, then apply nR sandhi as needed.
- Reattach the base to the results of preceding step.
gasyoun
There are only 3 nominals in lexnorm-all2 that end with one of the vowels 'F','x','X' (ṝ, ḷ, ḹ ); and they are the 'words' F,x,X !
Kale does not mention ḹ, and MW says "entirely artificial and only appearing in the works of some grammarians and lexicographers". Thus we remove 'X' from lexnorm-all2 and make no attempt to decline.
Of ṝ, ḷ, Kale says: "There are no words ending in ṝ or ḷ,; therefore, the imitative sounds of the roots kṝ and tṝ and gamḷ and śakḷ are taken to show what the forms of such words will be, if there be need to use such words".
Kale shows declension of the masculine of kṝ and gamḷ, and the endings and declension algorithm described below are based on these.
The algorithms also allow for feminine declensions, and assume they would be same as for masculine; this is by analogy with the equivalence of masculine and feminine declensions for nominals ending in e, E, o, O.
Since Kale mentions nothing regarding a neuter declension of kṝ and gamḷ, we make no guess here.