Covers SK 2139–2150 (12 sūtras): all present. The spine is Aṣṭādhyāyī 8.1.1–8.1.15 with three omissions.
- Omissions:
- 8.1.2 tasya param āmreḍitam is taught at SK 83 (sandhi). It names the second member of the pair āmreḍita.
- 8.1.3 anudāttaṃ ca is taught at SK 3670 (svara).
- 8.1.6 prasamupodaḥ pādapūraṇe is taught at SK 3599 (Vaidika).
- No reordering and no insertions. 8.1.16 padasya (SK 401) is not taught here, but it is drawn back (apakṛṣṭa) into 2140 (BM, TB).
dviruktaprakaraṇam — outline (p43, SK 2139–2150, all 12 sūtras)
p42 (taddhitādhikāre svārthikaprakaraṇam) closed at 2138 with madrāt parivāpaṇe (5.4.67).
- Where the prakaraṇa begins. NSP heads the section "dviruktaprakriyā 40" (p. 336). It opens with the adhikāra sarvasya dve (8.1.1).
- Where it ends. It ends at 2150 with dvandvam by nipātana and the yogavibhāga "dvandvam elsewhere too". The 1915 print then closes both the section and the first half of the SK: [1915 print: iti dviruktaprakriyā] and [1915 print: iti śrībhaṭṭojidīkṣitaviracitāyāṃ siddhāntakaumudyāṃ pūrvārdhaṃ samāptam]. Our digital mūla lacks both lines (open for Ajai). The second half opens in p44 with vartamāne laṭ (3.2.123, SK 2151).
- Contents in outline.
- 2139–2140: the heading, and doubling for repetition and pervasion.
- 2141–2143: doubling of particular words in fixed senses. pari "except", upari/adhi/adhas "close by", and an addressed word at the head of a sentence.
- 2144–2145: the doubled word treated like a bahuvrīhi (eka; pain).
- 2146–2148: the doubled word treated like a karmadhāraya. "Somewhat X", with 2147's vārttikas on order, alarm, intensity and "each other"; and priya/sukha for "with ease".
- 2149–2150: two fixed forms by nipātana, yathāyatham and dvandvam.
Threads that run through the prakaraṇa:
- What "dve" means. sarvasya dve is read as substitution (sthāne-dvirvacana). "dve" supplies the substitutes, two word-forms (śabdarūpe understood), and "sarvasya" is a sixth case of place. BM (sk.2140.b1) and TB (on 2139, 2140) both take this reading, and TB calls it the principal one (mukhya). The other view is that the whole is pronounced twice (dviḥprayoga, "dve uccāraṇe staḥ"). It is recorded but not followed. The outline says "two word-forms" throughout, 2147 included.
- "sarvasya" makes the whole word, ending and all, the thing replaced: in a form like vṛkṣābhyām the base alone is not doubled (BM, TB).
- The heading works by adhikāra, not anuvṛtti. sarvasya dve is svarita and heads 2140–2150 (TB). The kaumudī itself says "ity adhikṛtya" (2139) and writes "dve staḥ" at 2145 and 2147. BM supplies "dve staḥ" as śeṣa at 2143. At 2144 TB says "dve" anuvartate, which gives the kaumudī's "dviruktaḥ".
- "padasya" is drawn back from 8.1.16 (SK 401; BM names no number, so it is the translator's). This keeps doubling for repetition off a bare root and does not block the antaraṅga yaṅ (TB). So the doubled unit is always a finished pada.
- Two extensions (atideśa) shape what happens after the doubling:
- bahuvrīhivat (2144), carried into 2145 by "ca";
- karmadhārayavat (2146), a heading over 2147–2150. By the Vṛtti (TB) the wording "uttareṣu" is only for explicitness.
- Either extension makes the pair one prātipadika. That gives luk of both sup endings, a sup after the whole, puṃvadbhāva of the first part, and (for karmadhāraya) final udātta by samāsasya (6.1.223, SK 3734).
- The order is prātipadikatva → luk → sup. At 2144 BM construes "bahuvrīhivadbhāvād eva prātipadikatvāt" with both the luk and the new sup (madhyamaṇinyāya).
- karmadhārayavat is not redundant after bahuvrīhivat. It brings puṃvadbhāva by puṃvat karmadhāraya… (6.3.42, SK 746), which reaches even words with k before the final (kārikā, kālakakālikā) (BM, TB on 2147).
- Not every doubling is compound-like. Under 2147, the ānupūrvya and saṃbhrama vārttikas keep the sup (mūle mūle, sarpa sarpa: BM). Under the karmavyatihāra vārttika, compound treatment is bahulam: none for anya/para, always for itara.
Only the kaumudī blocks are translated (SK-TRANSLATE §25). The BM and TB readings cited here are those recorded in the unit notes.
- TB printed under a neighbouring unit: 2139 under sk.2138.t1; 2149 and part of 2150 under sk.2147.t1.
- No TB is printed on 2148 (NSP p. 340 goes from TB on 2147 straight to 2149).
- The NSP footnotes on 2149 ("dve") and 2150 ("mithunīyanti") are missing from editorial_notes. The en.yml notes record them.
Where the p43 audit (.work/sk-run/p43-audit-fixes.md and its two reports) corrects a reading, this outline follows the corrected reading.
- 2147 (must-fix): dve = "two word-forms" (substitution), not "two utterances". This holds in the sūtra and in the vārttikas ānupūrvye dve vācye and karmavyatihāre sarvanāmno dve vācye.
- 2147 (must-fix): in the verse, adḍ replaces su and am (7.1.25, SK 315; BM "adḍādeśaviraha"). It is not something added before them. Also, read daladvaye (not "dalaadvaye").
- 2144 (must-fix): the order is prātipadikatva (by bahuvrīhivadbhāva) → luk of both sups → sup after the whole. The vowel combination eka + eka → ekaika is supplied below and marked as implicit in the kaumudī.
- 2141–2143, 2146, 2148: the anuvṛtti/adhikāra apparatus belongs in the notes, not in the kaumudī's English. "sarvasya dve" works by adhikāra.
- 2145: BM's supplement to gatagataḥ is "priyāṃ vinā kālaḥ" ("[time] is gone, gone, without the beloved"). That is why the form is masculine. The source is not corrupt.
- 2146: "prayojanaṃ sublopapuṃvadbhāvāntodāttatvāni" is a vārttika (Mahābhāṣya 8.1.11 v1, which reads "karmadhārayavattve prayojanam …"; Wikisource vā 4689). NSP prints it run-on, without the sign.
- 2149: the NSP footnote says yathā does not itself function (denote its object as qualified by totality) in vīpsā, so 2140 cannot double it. It does not mean "yathā is not repeated".
- 2150: in dvandvaṃ mantrayate, dvandvam denotes the secret ("in secret"). "Two by two" is TB's etymology.
- 2150: the restored closing lines are mūla from the print, marked "[1915 print: …]". They are not translator's glosses.
Print readings (NSP, scan-confirmed) that the outline follows:
- antodāttatvam, ānupūrvye dve vācye, sarpa sarpa (our "sarva" is a slip), samuccitya and supaḥ (2147);
- prayunakti (2150; ours has "payunakti").
Variants where Wikisource agrees with our text, which the outline keeps:
- iha / NSP ita uttareṣu (2146);
- supi luki / NSP subluki punaḥ (2148);
- āmbhāvo vā (NSP without vā), ādeśam / NSP āmādeśam, kātarākṣī, aparasparaiḥ, bāhulakabalena and anyonyasaṃśrayaṃ tv etad (2147);
- maryādīkṛtya / NSP …prapautreṇāpīty arthaḥ (2150).
Still open for Ajai (mūla):
- the missing closing lines at 2150 (with the p42 closing at 2138);
- payunakti (2150);
- the 2147 slips: aācye, sarva, antodātatvam, supaṃ and samu-itya;
- the editorial "(vārttikam)" tags and the split of sarvanāmno vṛttimātre across k11–k13.
Ordering: SK order throughout, which is AS order here.
A. The heading, and doubling for repetition and pervasion — running example: pacati → pacati pacati; vṛkṣam → vṛkṣaṃ vṛkṣam
- 2139 sarvasya dve 8.1.1 — adhikāra: in place of the whole word (with all its parts), two word-forms come. The kaumudī says "ity adhikṛtya", and BM supplies the rest: "[the injunctions of doubling will be set out]".
- TB: the rule names neither an operation nor a substituend, and it is svarita. So it is a heading, and the heading is its meaning, not its wording. nāmreḍitasyāntyasya tu vā (6.1.99, SK 82) shows this.
- It feeds every doubling below. In pacati pacati, the whole pada pacati is the substituend, and the two word-forms pacati + pacati replace it.
- 2140 nityavīpsayoḥ 8.1.4 — a pada is doubled when repetition (nitya = ābhīkṣṇya, paunaḥpunya) or pervasion (vīpsā, the wish to convey totality) is to be made manifest. The sense is in the base; the doubling only manifests it (dyotaka; BM).
- Repetition belongs to the principal action, so it is found only in tiṅantas and avyaya-kṛdantas (BM, TB):
pacati → pacati pacati ("he cooks again and again"); bhuktvā → bhuktvā bhuktvā. - Pervasion:
vṛkṣam → vṛkṣaṃ vṛkṣaṃ siñcati ("he waters every tree"); grāmo grāmo ramaṇīyaḥ. - The totality is limited by context (the trees of this garden) and appears in the singular, one by one (BM, TB). Where nothing limits it, it is unlimited: jāto jāto nidhanam upaiti (TB).
- "padasya" is drawn back here (BM, TB).
B. Doubling of particular words in fixed senses: pari "except", upari/adhi/adhas "close by", a sentence-opening vocative — running example: pari vaṅgebhyaḥ → pari pari vaṅgebhyaḥ
- 2141 parer varjane 8.1.5 — pari is doubled when it means exclusion.
pari pari vaṅgebhyo vṛṣṭo devaḥ: "the [rain-]god has rained except on the Vaṅgas" (parihṛtya, "leaving them out").- pari is a karmapravacanīya by apaparī varjane (1.4.88, SK 596), and vaṅgebhyaḥ is fifth case by pañcamy apāṅparibhiḥ (2.3.10, SK 598) (BM, TB).
- Vārttika parer varjane vāvacanam: the doubling is optional, so
pari vaṅgebhyaḥ also stands. - TB adds the vārttika's "asamāse": there is no doubling in the compound paritrigartaṃ vṛṣṭo devaḥ.
- 2142 uparyadhyadhasaḥ sāmīpye 8.1.7 — upari, adhi, adhas are doubled when nearness is meant.
- Nearness in space:
uparyupari grāmam ("in the region close above the village"); adho 'dho lokam ("close below the world"). - Nearness in time:
adhyadhi sukham ("sorrow close upon happiness") (BM, TB). - TB: in upari candramāḥ only the relation of higher and lower is meant, so there is no doubling.
- 2143 vākyāder āmantritasyāsūyāsammatikopakutsanabhartsaneṣu 8.1.8 — an addressed form at the head of a sentence is doubled in five senses. BM supplies "dve staḥ" (śeṣa).
- Envy:
sundara sundara vṛthā te saundaryam. - Approval:
deva deva vandyo 'si. - Anger:
durvinīta durvinīta idānīṃ jñāsyasi (BM supplies "[the fruit of your insolence]"). - Censure:
dhānuṣka dhānuṣka vṛthā te dhanuḥ. - Threat:
cora cora ghātayiṣyāmi tvām. - Threat and censure are named apart from anger and envy because a teacher threatens or censures a pupil without either (BM, TB).
- The accent rules for the āmreḍita (svaritam āmreḍite… 8.2.103, SK 3622; āmreḍitaṃ bhartsane 8.2.95, SK 3614) apply. Pluti is optional and not shown (TB).
C. The doubled word treated like a bahuvrīhi: eka, and pain — running example: ekam ekam → ekaikam; gatā gatā → gatagatā
- 2144 ekaṃ bahuvrīhivat 8.1.9 — eka, once doubled ("dve" carried forward, TB; in vīpsā only, per the Bhāṣya, BM), is treated like a bahuvrīhi. This gives sublopa and puṃvadbhāva.
ekaikam akṣaram ("each single syllable"). The chain, in BM's order:
1. ekam ekam: the whole is a prātipadika by bahuvrīhivadbhāva;
2. luk of both sups: eka eka, giving ekaika- (the vowel combination eka + eka → ekaika is implicit in the kaumudī);
3. a sup after the whole: ekaikam.
- Without the extension the pair is no prātipadika ("yatra saṅghāte pūrvo bhāgaḥ padaṃ … samāsasyaiva"), so there could be no luk (BM). The new sup is singular only (TB).
ekaikayā āhutyā ("with each single oblation"): puṃvadbhāva of the first part. The form comes out the same either way, so the difference shows only in the avagraha: eka-ekayā, not ekā-ekayā (BM, TB; the Taittirīyas divide it ekā ekayā).- The prāñcaḥ read the repeated "bahuvrīhi" in na bahuvrīhau (1.1.29, SK 222) as the principal bahuvrīhi only, so an extended bahuvrīhi keeps sarvanāman status. In fact the Bhāṣya rejects that sūtra. Even where it is kept, it bars sarvanāman status only in the alaukika vigraha. So
ekaikasmai dehi: the whole ends in eka, a sarvādi word, and takes smai (BM). - "bahuvrīhi" rather than "samāsa" brings the accent of bahuvrīhau prakṛtyā pūrvapadam (6.2.1, SK 3735) (BM).
- 2145 ābādhe ca 8.1.10 — when pain (ābādha = pīḍā) is to be made manifest, two word-forms come, and by "ca" the pair is also bahuvrīhivat. The kaumudī states both: "dve staḥ bahuvrīhivac ca".
gatagataḥ ("[time] is gone, gone [without the beloved]", BM: priyāṃ vinā kālaḥ) is the cry of one tormented by separation. bahuvrīhivadbhāva gives luk of the sups.gatagatā [priyā]: the first part becomes masculine (puṃvadbhāva) by striyāḥ puṃvat… (6.3.34, SK 831). The second part counts as an uttarapada through the extension (BM, TB).- The extension gives only what a bahuvrīhi shows by rule, not samāsāntas:
dhūrdhūḥ, panthāḥ panthāḥ (TB).
D. The doubled word treated like a karmadhāraya: "somewhat X", order, alarm, intensity, "each other", "with ease" — running example: paṭvī → paṭupaṭvī
- 2146 karmadhārayavad uttareṣu 8.1.11 — heading: in the doublings that follow, the operation (kāryam, BM śeṣa) is as for a karmadhāraya.
- Vārttika (Bhāṣya 8.1.11 v1, quoted without "karmadhārayavattve"; NSP prints it run-on) prayojanaṃ sublopapuṃvadbhāvāntodāttatvāni. Its fruits are sup-deletion, masculine treatment and final udātta. In paṭupaṭvī (2147) all three are at work.
- anudāttaṃ ca (8.1.3, SK 3670), though an adhikāra, is not connected with the final udātta (BM, on the Bhāṣya).
- 2147 prakāre guṇavacanasya 8.1.12 — when similarity (prakāra = sādṛśya, by vyākhyāna; BM, TB) is to be conveyed, two word-forms of a guṇavacana come ("dve staḥ"), treated karmadhārayavat by 2146.
- The sūtra's own example.
paṭvī paṭvī → paṭupaṭvī. The prior part is made masculine by puṃvat karmadhāraya… (6.3.42, SK 746), and the whole takes final udātta by samāsasya (6.1.223, SK 3734). paṭupaṭuḥ = paṭusadṛśaḥ, i.e. īṣatpaṭuḥ, "somewhat skilful".- Both kinds of guṇavacana are taken. A word for a substance with the quality subordinate, and a word for the quality alone:
śuklaśuklaṃ rūpam ("a whitish colour", the quality alone; TB: shown by rūpam);śuklaśuklaḥ paṭaḥ ("a whitish cloth").- Vārttika ānupūrvye dve vācye: two word-forms for succession in order.
mūle mūle sthūlaḥ: each part nearer the root is thicker than the one beyond. - There is no karmadhāraya treatment, so the seventh-case ending stays (BM, from the Bhāṣya's example).
- This is not vīpsā, since a stick has only one real root (TB).
- Vārttika saṃbhrameṇa pravṛttau yatheṣṭam anekadhā prayogo nyāyasiddhaḥ: in alarm a word is used as many times as one wishes, which is established by reasoning. "anekadhā" suspends "dve" (BM).
sarpa sarpa, budhyasva budhyasva; sarpa sarpa sarpa, budhyasva budhyasva budhyasva. The sup stays (BM).- Vārttika kriyāsamabhihāre ca: two word-forms for repetition or intensity of an action.
lunīhi lunīhīty evāyaṃ lunāti. The loṭ comes by kriyāsamabhihāre loṭ… (3.4.2, SK 2825), and the anuprayoga lunāti by yathāvidhy anuprayogaḥ pūrvasmin (3.4.4, SK 2827) (BM).- 2140 already doubles for repetition, so this vārttika is for intensity (bhṛśa). Or else it lets doubling signify repetition jointly with loṭ. That is why a yaṅ form such as pāpacyate is not doubled: yaṅ signifies the sense by itself (BM, TB).
- Vārttika karmavyatihāre sarvanāmno dve vācye, samāsavac ca bahulam: for exchange of action, two word-forms of a pronoun. The doubling is obligatory, and bahulam goes only with the compound treatment (BM, TB).
- bahula means anya and para are not compound-like and itara always is. Other pronouns are not doubled this way at all (BM).
- Vārttika asamāsavadbhāve pūrvapadasthasya supaḥ sur vaktavyaḥ: where there is no compound treatment (anya, para), the prior member's sup, of any case, is replaced by su (BM).
anyam anyam → anyaḥ anyam → anyo'nyaṃ viprā namanti; likewise anyo'nyau, anyo'nyān, anyo'nyena kṛtam, anyo'nyasmai dattam.- Māgha has
anyo'nyeṣāṃ puṣkarair āmṛśantaḥ, so poets do not accept the view that this applies to the first-case singular only (BM, TB). parasparam: s for the visarga because paraspara is kaskādi (kaskādiṣu ca 8.3.48, SK 144).itaretaram, itaretareṇa are compound-like: both sups undergo luk and a fresh sup comes after the whole (BM).- Vārttika strīnapuṃsakayor uttarapadasthāyā vibhakter āmbhāvo vā vaktavyaḥ: in the feminine and neuter, the second member's ending is optionally replaced by ām.
anyonyām / anyonyam; parasparām / parasparam; itaretarām / itaretaram — ime brāhmaṇyau kule vā bhojayataḥ ("these two brahmin women, or two families, feed each other").- kecit: ām comes for the second case only, following the Bhāṣya's examples, so the third case and the rest are as in the masculine.
- anye: ām comes for every vibhakti, because the example is merely indicative (diṅmātra). BM takes this as the settled view.
- The verse (a prācīna śloka, BM) daladvaye ṭābabhāvaḥ klībe cādḍvirahaḥ svamoḥ | samāse sor aluk ceti siddhaṃ bāhulakāt trayam. bāhulaka yields three results:
1. no ṭāp in either half of feminine anyonyam and parasparam;
2. in the neuter, no adḍ substituting for su and am (anyonyam, itaretaram, against adḍ ḍatarādibhyaḥ… 7.1.25, SK 315);
3. aluk of su in the prior segment when the doubled form enters a further compound.
- Why bāhulaka is needed. The explanation first shows that each result is due and has no other source:
- ṭāp: masculine treatment cannot come from the vārttika sarvanāmno vṛttimātre [puṃvadbhāvaḥ], because anya and para are not compound-like. Nor is doubling itself a vṛtti: that would overextend to
yāṃ yāṃ priyaḥ praikṣata kātarākṣī sā sā. - adḍ: it is due by 7.1.25.
- luk of su: it is due in
anyonyasaṃsaktam ahastriyāmam, anyonyāśrayaḥ, parasparākṣisādṛśyam and aparasparaiḥ. - All three are settled "by the force of bāhulaka". The authority is the Bhāṣya's examples on this vārttika and its "anyonyasaṃśrayaṃ tv etad" under striyām (4.1.3, SK 453).
- 2148 akṛcchre priyasukhayor anyatarasyām 8.1.13 — priya and sukha are optionally doubled when they mean ease (akṛcchra = anāyāsa). The karmadhāraya treatment comes from the adhikāra of 2146, not from the vārttika's "samāsavat" (BM).
priyeṇa priyeṇa → priya priya (luk of both sups) → priyapriyeṇa (the same ending after the whole) dadāti. Or simply priyeṇa dadāti.- Likewise
sukhasukhena / sukhena dadāti: "he gives even a very dear thing without effort".
E. Two fixed forms by nipātana: yathāyatham, dvandvam — running example: yathā → yathāyatham; dvi dvi → dvandvam
- 2149 yathāsve yathāyatham 8.1.14 — in the sense of yathāsva, the form yathāyatham is fixed by nipātana.
- yathāsvam is an avyayībhāva in the sense of vīpsā: "whatever is one's self and whatever is one's own" (sva here = self and one's own only, by vyākhyāna; BM).
- Two things are fixed by nipātana: the doubling of yathā, and neuter gender, so the whole yathā-yatha has a gender and a short final (BM).
yathāyathaṃ jñātā = yathāsvabhāvam or yathātmīyam, "one who knows [things] each according to its own nature". jñātā ends in tṛn, so the sixth case is barred by na lokāvyaya… (2.3.69, SK 627) (BM, TB).- Why nipātana? 2140 doubles only a word that itself denotes its object as qualified by totality (otherwise sarva in sarvo ghaṭaḥ would double). yathā does not function that way here (BM; NSP fn.).
- 2150 dvandvaṃ rahasyamaryādāvacanavyutkramaṇayajñapātraprayogābhivyaktiṣu 8.1.15 — dvandvam is fixed by nipātana in five senses.
- The derivation. From
dvau dvau, karmadhārayavattva gives luk of the sups and a sup again (BM). dvi is doubled, am replaces the final of the first member, a replaces the final of the second, and the word is neuter; the "ca" adds singular treatment (TB). So ato 'm gives dvandvam. - Only rahasya is what the word denotes (vācya). The other senses are domains (viṣaya).
- Secret:
dvandvaṃ mantrayate, "he takes counsel in secret" (TB's etymology: "two by two, for so it is secret"). - A limit expressed:
ācaturaṃ hīme paśavo dvandvaṃ mithunīyanti, "for these animals, up to the fourth [generation], [they] desire to mate pair by pair". - The mother comes to son, grandson and great-grandson.
- ācaturam is an avyayībhāva by āṅ maryādābhividhyoḥ (2.1.13, SK 667) with ṭac; BM reads the āṅ as inclusive. mithunīyanti takes kyac (BM, TB).
- Separation:
dvandvaṃ vyutkrāntāḥ, "they stand apart by their connection with two groups". - Use of sacrificial vessels:
dvandvaṃ yajñapātrāṇi prayunakti, "he sets out the vessels in pairs" (NSP prayunakti). - Manifestation:
dvandvaṃ saṅkarṣaṇavāsudevau, "known as a pair through constant association" (sāhacarya). - By yogavibhāga, dvandvam is accepted elsewhere too. BM, TB:
dvandvāni sahate ("he bears the pairs": heat and cold, pleasure and pain, hunger and thirst); dvandvaṃ yuddhaṃ pravartate (vīpsā). - Close: [1915 print: iti dviruktaprakriyā ‖ iti śrībhaṭṭojidīkṣitaviracitāyāṃ siddhāntakaumudyāṃ pūrvārdhaṃ samāptam]. The first half of the SK ends here, and p44 opens the second half with vartamāne laṭ (3.2.123, SK 2151).