Open funderburkjim opened 5 years ago
The declension algorithm described below uses two bases; the first comes from the MW citation form, and the second comes from the feminine form. The feminine form is simply related to what Antoine calls the 'weak' stem and comes from the 3rd personal plural of the perfect of the underlying verb. Since there are so few citation instances to work with (only about 30, not counting compounds of vidvas), the algorithm was determined 'by hand' using clues from MW, Antoine, Deshpande and Kale.
Here is the list of citation forms and feminine forms and underlying root; this is taken as given in the current algorithm. (The root is not required, but is helpful to know when researching the dictionaries.)
vidvas viduṣī vid
cakṛvas cakruṣī kṛ
papivas papuṣī pā
proṣivas proṣuṣī vas
bibhīvas bibhyuṣī bhī
rarivas raruṣī rā
saścivas saścuṣī sac
īyivas īyuṣī i
udeyivas udeyuṣī i
upeyivas upeyuṣī i
pareyivas pareyuṣī i
khidvas khiduṣī khid
cikitvas cikituṣī cit
jakṣivas jakṣuṣī ghas
jaganvas jagmuṣī gam
jagmivas jagmuṣī gam
jaghanvas jaghnuṣī han
jaghnivas jaghnuṣī han
jajñivas jajñuṣī jan
jūjuvas jūjuvuṣī jū
śuśruvas śuśruvuṣī śru
tasthivas tasthuṣī sthā
pīpivas pipyuṣī pyai
śiśrivas śiśryuṣī śri
vavṛtvas vavṛtuṣī vṛt
sāsahvas sāsahuṣī sah
suṣupvas suṣupuṣī svap
ninīvas ninyuṣī nī -- not in mw, mentioned in Antoine.
I hope our Sanskrit experts (@drdhaval2785 and/or @SergeA ) will take a critical look at the feminine forms and point out any errors.
Of course, the feminine declension is the normal one for feminines ending in long vowel 'I', like nadI.
As mentioned, there are two bases required in the present algorithm for masculine and neuter declensions. I'll just call them base1 and base2.
This base is formed by dropping the final 'vas' from the citation form. For example, from vidvas we get vid
Note that all the feminine stems ending in uzI (uṣī). The second base is formed by dropping this final suffix from the feminine stem. For example, from cakruṣī we get second base of cakr.
In the algorithm, some of the endings will be combined with base1 and some endings with base2. In this case, all the endings which are to be combined with base2 are those which begin with uz. These are identified in italics in the table. The others are to be combined with base1.
Case | S | D | P |
---|---|---|---|
Nominative | vAn | vAMsO | vAMsaH |
Accusative | vAMsam | vAMsO | uzaH |
Instrumental | uzA | vadByAm | vadBiH |
Dative | uze | vadByAm | vadByaH |
Ablative | uzaH | vadByAm | vadByaH |
Genitive | uzaH | uzoH | uzAm |
Locative | uzi | uzoH | vatsu |
Vocative | van | vAMsO | vAMsaH |
As with the endings for the masculine, the base2 endings are italicized in the table.
Case | S | D | P |
---|---|---|---|
Nominative | vat | uzI | vAMsi |
Accusative | vat | uzI | vAMsi |
Instrumental | uzA | vadByAm | vadBiH |
Dative | uze | vadByAm | vadByaH |
Ablative | uzaH | vadByAm | vadByaH |
Genitive | uzaH | uzoH | uzAm |
Locative | uzi | uzoH | vatsu |
Vocative | vat | uzI | vAMsi |
Except for the fact that some endings are joined with base1 and some with base2, the declension algorithm is similar to the declension algorithms for vowel-ending nominals.
Note on sandhi: There are no sandhi adjustments required in this algorithm, including nR sandhi. In standard presentations of the algorithm for these 'vas' participles, there are three bases (strong, middle, and weak), and a general table of endings. As with our algorithm, a given cell in the declension table is formed by joining the general ending to the base (s,m,or) appropriate to the cell, and applying general sandhi rules as part of this joining. In effect, the algorithm used here has already taken into account these sandhis by the choice of endings.
Case | S | D | P |
---|---|---|---|
Nominative | vid + vAn = vidvAn | vid + vAMsO = vidvAMsO | vid + vAMsaH = vidvAMsaH |
Accusative | vid + vAMsam = vidvAMsam | vid + vAMsO = vidvAMsO | vid + uzaH = viduzaH |
Instrumental | vid + uzA = viduzA | vid + vadByAm = vidvadByAm | vid + vadBiH = vidvadBiH |
Dative | vid + uze = viduze | vid + vadByAm = vidvadByAm | vid + vadByaH = vidvadByaH |
Ablative | vid + uzaH = viduzaH | vid + vadByAm = vidvadByAm | vid + vadByaH = vidvadByaH |
Genitive | vid + uzaH = viduzaH | vid + uzoH = viduzoH | vid + uzAm = viduzAm |
Locative | vid + uzi = viduzi | vid + uzoH = viduzoH | vid + vatsu = vidvatsu |
Vocative | vid + van = vidvan | vid + vAMsO = vidvAMsO | vid + vAMsaH = vidvAMsaH |
Case | S | D | P |
---|---|---|---|
Nominative | cakf + vat = cakfvat | cakr + uzI = cakruzI | cakf + vAMsi = cakfvAMsi |
Accusative | cakf + vat = cakfvat | cakr + uzI = cakruzI | cakf + vAMsi = cakfvAMsi |
Instrumental | cakr + uzA = cakruzA | cakf + vadByAm = cakfvadByAm | cakf + vadBiH = cakfvadBiH |
Dative | cakr + uze = cakruze | cakf + vadByAm = cakfvadByAm | cakf + vadByaH = cakfvadByaH |
Ablative | cakr + uzaH = cakruzaH | cakf + vadByAm = cakfvadByAm | cakf + vadByaH = cakfvadByaH |
Genitive | cakr + uzaH = cakruzaH | cakr + uzoH = cakruzoH | cakr + uzAm = cakruzAm |
Locative | cakr + uzi = cakruzi | cakr + uzoH = cakruzoH | cakf + vatsu = cakfvatsu |
Vocative | cakf + vat = cakfvat | cakr + uzI = cakruzI | cakf + vAMsi = cakfvAMsi |
There are about 60 entries ending in 'vas' in lexnorm-all2, which have been identified as being perfect participles active parasmaipada (kvasu). As I understand it, there are many more such participles, indeed one for any root or prefixed root. But we are only concerned here with those which have a separate entry in MW.
Note on spelling: In PWG, these are spelled with ending 'vaMs'. For example vidvas in MW corresponds to vidvaMs in PWG.
The models for declensions of these participles are named m_vas, n_vas, and f_vas_I.