Declension of nominals ending in 'j'

[funderburkjim]

This note documents current status of declensions of nominals ending in 'j'.
For lack of a better term, I'll call these the 'Cologne' declensions.

As with declensions of other nominals, the algorithms generate a declension based on two pieces of information:

• a model. This model is used by the software to pick out which part of
  the algorithm to apply.
  • For nominals ending in 'j', the algorithm is one of m_1_j, f_1_j, n_1-j.
    m_1_j is intended to be read masculine as noun with 1 stem ending in j.
• a headword:
  • in SLP1 spelling
  • with hyphenation for compounds, such as 'eka-rAj'

In a local installation, experimentation with the results of declensions is facilitated by a command line program decline_one.py.

For example,

python decline_one.py m_1_j eka-rAj
Declension of m_1_j eka-rAj
Case 1:  ekarAw ekarAjO ekarAjaH
Case 2:  ekarAjam ekarAjO ekarAjaH
Case 3:  ekarAjA ekarAqByAm ekarAqBiH
Case 4:  ekarAje ekarAqByAm ekarAqByaH
Case 5:  ekarAjaH ekarAqByAm ekarAqByaH
Case 6:  ekarAjaH ekarAjoH ekarAjAm
Case 7:  ekarAji ekarAjoH ekarAwsu
Case 8:  ekarAw ekarAjO ekarAjaH

[funderburkjim]

The feminine declensions

are same as masculine declensions:

 python decline_one.py f_1_j eka-rAj
Declension of f_1_j eka-rAj
Case 1:  ekarAw ekarAjO ekarAjaH
Case 2:  ekarAjam ekarAjO ekarAjaH
Case 3:  ekarAjA ekarAqByAm ekarAqBiH
Case 4:  ekarAje ekarAqByAm ekarAqByaH
Case 5:  ekarAjaH ekarAqByAm ekarAqByaH
Case 6:  ekarAjaH ekarAjoH ekarAjAm
Case 7:  ekarAji ekarAjoH ekarAwsu
Case 8:  ekarAw ekarAjO ekarAjaH

The neuter declensions

are same as masculine, EXCEPT for cases 1,2,8:

python decline_one.py n_1_j eka-rAj
Declension of n_1_j eka-rAj
Case 1:  ekarAw ekarAjI ekarAYji
Case 2:  ekarAw ekarAjI ekarAYji
Case 3:  ekarAjA ekarAqByAm ekarAqBiH
Case 4:  ekarAje ekarAqByAm ekarAqByaH
Case 5:  ekarAjaH ekarAqByAm ekarAqByaH
Case 6:  ekarAjaH ekarAjoH ekarAjAm
Case 7:  ekarAji ekarAjoH ekarAwsu
Case 8:  ekarAw ekarAjI ekarAYji

[funderburkjim]

The tricky part

There are rules for what Goldman's text calls Consonants in Absolute Final Position (Goldman, section 3.25).

Palatals cannot appear in absolute final position; the palatal 'j' is replaced, in final position, by either k or w (IAST ṭ).

The tricky part is deciding, for a particular nominal, which of the two replacements to use. There seem to be no definite rules.

However, once you decide whether to use k or w (ṭ), the rest of the declension is determined. 'eka-rAj' shows the w case.

An example where final 'j' is replaced by 'k' is:

 python decline_one.py m_1_j Buj
Declension of m_1_j Buj
Case 1:  Buk BujO BujaH
Case 2:  Bujam BujO BujaH
Case 3:  BujA BugByAm BugBiH
Case 4:  Buje BugByAm BugByaH
Case 5:  BujaH BugByAm BugByaH
Case 6:  BujaH BujoH BujAm
Case 7:  Buji BujoH Bukzu
Case 8:  Buk BujO BujaH

[funderburkjim]

The choice of 'k' or 'w' is confirmed for some words by Huet declension
and, occasionally, by a comment in MW.

Cologne declensions using 'w' confirmed by Huet

• rAj and compounds: aDi- eka-(MW) giri- vi- viSvA- vEdya- sam- (MW)sva- svA- naB-
• vrAj compounds: pari- (MW) pra-
• ij Urj tyaj yaj yAj (MW)

Cologne declensions using 'k' confirmed by Huet

• Also by MW: uSij ftvij baRij vaRij Bizaj a-sfj
• Buj and compounds: kaRa- kziti- gfhabali- tri- dvArabali- BU- mahI- SrAdDa- Svasana- sasya- sTUla- huta-
• BAj and compounds: janma- jAti- daRqa- doza- puRya- vi- naB-
• ruj and compounds: nI-
• yuj compounds: saM- sa-
• ij Urj tyaj yaj (these also take 'k', in current Cologne-Huet declensions)

[funderburkjim]

yuj, not at end of compound

According to mw, when uncompounded, the strong cases have a nasal e.g. nom. yuṅ, yuñjau, yuñjas, but aśva-yuk

The Cologne declension:

Declension of m_1_j yuj
Case 1:  yuN yuYjO yuYjaH
Case 2:  yuYjam yuYjO yujaH
Case 3:  yujA yugByAm yugBiH
Case 4:  yuje yugByAm yugByaH
Case 5:  yujaH yugByAm yugByaH
Case 6:  yujaH yujoH yujAm
Case 7:  yuji yujoH yukzu
Case 8:  yuN yuYjO yuYjaH

Huet shows some additional forms in cases 1,2,8

python decline_one_huet.py m_1_j yuj
Declension of m_1_j yuj
Case 1:  yuk/yuN yujO/yuYjO yujaH/yuYjaH
Case 2:  yujam/yuYjam yujO/yuYjO yujaH
Case 3:  yujA yugByAm yugBiH
Case 4:  yuje yugByAm yugByaH
Case 5:  yujaH yugByAm yugByaH
Case 6:  yujaH yujoH yujAm
Case 7:  yuji yujoH yukzu
Case 8:  yuk/yuN yujO/yuYjO yujaH/yuYjaH

[gasyoun]

For lack of a better term, I'll call these the 'Cologne' declensions.

So be it.

However, once you decide whether to use k or w (ṭ), the rest of the declension is determined.

It's easy. k is for all (J1, H1 how Zaliznjak calls them), other than roots (list of roots-exclusions, called J2 and H2): √yaj совершать жертвоприношение, √vraj продвигаться, √rāj править, √bhrāj сиять, √maj вытирать, чистить, √sarj отпускать, √bharjj (bhrajj) жарить; √vah везти, √sah преодолевать, √lih (rih) лизать, √mih мочиться, √guh (gūh) прятать, √ruh подниматься, √tarh давить, √darh (dṛṁh) укреплять, √varh (barh) рвать, √baṁh [802] усиливать. У корня √muh теряться, смущаться конечное h может вести себя и как h1 и как h2. Во всех прочих случаях j и h традиционной записи с морфонологической точки зрения суть j1 и h1.

[funderburkjim]

Wonderful that you mention Professor Zaliznjak. Others may enjoy this inspiring link.

[funderburkjim]

Paraphrase of Zaliznjak comment for -j roots

Most 'j'-ending nominals take 'k' EXCEPT for nominals based on a list of roots.

That list of j-ending roots whose nominals take ṭ are:

• yaj совершать жертвоприношение = 'make a sacrifice'.
  • MW confirms related 'yAj' nom. yāṭ
  • MW doesn't mention which nominative to use for 'yaj' or for two compounds diviyaj and devayaj.
• vraj
  • MW pari-vrAj nom. ṭ
• rAj MW agrees
• BrAj MW agrees ṭ for compounds
• maj
  • Based on sense 'вытирать, чистить' = wipe clean, this root corresponds
    to MW 'mfj'
  • MW compound 'Darma-mfj' disagrees , mentioning 'nom. mfk'
• sarj (MW spelling sfj)
  • Goldman viSva-sfj confirms ṭ.
  • MW viSva-sfj entry ambiguous: nom. -sṛk or incorrectly -sṛṭ
• Brajj Confirm MW: nom. bhraṭ

[funderburkjim]

Gerard pointed out Whitney Grammar Section 219.

Whitney also delineates two groups, the 'yuj' group and the 'mfj' group.

I think his 'yuj' group is what I call the 'k' group, and his 'mfj' group is the ṭ group. The distinctions of these two groups appear to include declensions, but
also include other forms (exactly which forms not obvious to me).

Whitney 'yuj'-class ('k')

• Baj
• saj
• tyaj (not V.) ?
• raj color
• svaj
• majj
• nij
• tij
• vij
• Buj
• yuj
• ruj
• vfj
• aYj
• BaYj
• SiYj
• Urj
• sraj
• Bizaj
• asfj (Note plain 'sfj' in mfj group below)
• tfzRaj
• vaRij
• ftvij

Whitney 'mfj'-class (ṭ)

• yaj
• Brajj
• vraj
• rAj
• BrAj
• mfj
• sfj -- But 'MS. has ViSva-sfk

[funderburkjim]

Question re Brajj

I am uncertain of the handling of the double-j.
The current method uses

• 'w' in nom. singular
• 'w' in locative plural
• 'q' before 'B'
• 'jj' before vowel endings

Here is neuter declension:

Declension of n_1_j Brajj
Case 1:  Braw BrajjI BraYji
Case 2:  Braw BrajjI BraYji
Case 3:  BrajjA BraqByAm BraqBiH
Case 4:  Brajje BraqByAm BraqByaH
Case 5:  BrajjaH BraqByAm BraqByaH
Case 6:  BrajjaH BrajjoH BrajjAm
Case 7:  Brajji BrajjoH Brawsu
Case 8:  Braw BrajjI BraYji

Can anyone confirm or point out errors?

[funderburkjim]

Question re mfj

Only two compounds ending in mfj from MW.

• Darma-mfj MW says 'mfk' v.l. for -Dfk (Darma-Dfk is a proper name)
• pari-mfj MW doesn't mention nominative

According to Zaliznjak and Whitney above, ṭ should be used.
But I am 'k' in both cases (Darma-mfk, pari-mfk).
Should I change either or both to ṭ ?

[gasyoun]

Should I change either or both to ṭ ?

Let @drdhaval2785 comment. Zaliznjak's source is Whitney with a few finetunings. But it's not enough for our computation tasks in samasas.

[drdhaval2785]

I think the rule which you want to have at hand is https://ashtadhyayi.com/sutraani/8/2/36
vraSc -> vfw
Brasj / Brajj -> Bfw
sfj -> sfw
mfj -> mfw
yaj -> yaw
rAj -> rAw
BrAj -> BrAw

And any verb ending with C and S get converted to ending with 'w'.