लकारार्थप्रकरणम्

Covers SK 2773–2828 (56 sūtras): all present. Every unit has its own kaumudī (2804's k1 is a lone daṇḍa; its prose is k3). The prakaraṇa is semantic: it does not derive forms, it says which lakāra comes in which sense (and with which co-occurring word). So the "running example" of each group is the example sentence that shows the lakāra choice, and each sūtra line names the condition and the lakāra or lakāras it yields.

The spine is one Aṣṭādhyāyī run in four stretches: 3.2.112–122 (2773–2782), 3.3.4–9 (2783–2788), 3.3.131–172 (2789–2823) and 3.4.1–5 (2824–2828). Sūtras of the run that the SK treats elsewhere are recalled at their Aṣṭādhyāyī place without a new number:

• parokṣe liṭ (3.2.115, SK 2171) inside 2775;
• liṅnimitte lṛṅ kriyātipattau (3.3.139, SK 2229) inside 2796;
• vidhinimantraṇāmantraṇādhīṣṭasaṃpraśnaprārthaneṣu liṅ (3.3.161, SK 2208) and loṭ ca (3.3.162, SK 2194) inside 2816;
• māṅi luṅ (3.3.175, SK 2219) inside 2823.

The kṛt sūtras of the run are left to the kṛt section: 3.3.1–3 and 3.3.10–130, samānakartṛkeṣu tumun (3.3.158, SK 3176), kālasamayavelāsu tumun (3.3.167, SK 3179), 3.3.170–171 (SK 3311–3312) and 3.3.174 (SK 3313). āśiṣi liṅloṭau (3.3.173, SK 2195) and smottare laṅ ca (3.3.176, SK 2220) were given in the bhvādi section and are not recalled here. Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json. All 56 match our source.json, and the skn matches ours for all 56. Note that 2788 (3.3.9) and 2818 (3.3.164) have the same wording, liṅ cordhvamauhūrtike, but are two different sūtras.

• Anuvṛtti from later in the SK.
• bhūte (3.2.84, SK 2995) is the heading over 2773–2782.
• bhaviṣyati gamyādayaḥ (3.3.3, SK 3171) gives bhaviṣyati to 2783–2788 (announced at the end of 2782).
• uṇādayo bahulam (SK 3169) marks the end of the "present" section that 2789 extends.
• samānakartṛkeṣu tumun (SK 3176) gives the same-agent condition to 2815.
• kālasamayavelāsu tumun (SK 3179) gives kāla, samaya, velā to 2821.
• Anuvṛtti from earlier prakaraṇas.
• anadyatane from anadyatane laṅ (3.2.111, SK 2205) at 2773 onwards, and at 2782 by the frog's leap.
• vartamāne laṭ (SK 2151) is the starting point of the section that 2789 extends.
• liṅnimitte lṛṅ kriyātipattau (SK 2229) is carried whole into 2797 and 2798.
• loṭ ca (SK 2194) is drawn in by the ca of 2817.
• śeṣe lṛḍ ayadau is internal (2808 → 2812, BM).

lakārārthaprakaraṇam — outline (p65, SK 2773–2828, all 56 sūtras)

p64 (karmakartṛtiṅprakaraṇam) closed at 2772 kuṣirajoḥ prācāṃ śyan parasmaipadaṃ ca with the print's colophon "iti karmakartṛprakriyā".

• Where the prakaraṇa begins. The print heads it "tiṅante lakārārthaprakriyā 62" (NSP p. 471, scan 0481). Our name and number are lakārārthaprakaraṇam 65. It opens with abhijñāvacane lṛṭ (3.2.112, SK 2773). TB on 2773 is printed at the tail of sk.2771.t1, after the p64 colophon.
• Where it ends. At 2828 with the Māgha verse purīm avaskanda… and Bhaṭṭoji's rejection of two misreadings of it. The print closes with two colophons (NSP p. 478, scan 0488): "iti lakārārthaprakriyā" and "iti … siddhāntakaumudyām uttarārdhe tiṅantaṃ samāptam". Ours lacks both, so the English supplies them in brackets. The whole tiṅanta ends here. p66 (kṛdantakṛtyaprakaraṇam) opens with dhātoḥ (3.1.91, SK 2829).
• Contents in outline.
• 2773–2775: lṛṭ for a past (not of today) after a word of recollection. smarasi kṛṣṇa gokule vatsyāmaḥ. Not with yad (laṅ), but optionally again when the two actions mark each other. parokṣe liṭ is recalled for the first person, and liṭ is given for total denial.
• 2776–2782: the past with particles. laṅ besides liṭ with ha and śaśvat and in a question about recent time; laṭ with sma, in a nanu-reply, and optionally with na or nu; luṅ and laṭ with purā.
• 2783–2788: laṭ for the future with particles and senses. yāvat and purā; kadā and karhi; a form of kim with a wish to obtain; what is sought; what marks loṭ's meaning. liṅ too after a muhūrta.
• 2789–2792: time by extension. Near-present as present; hope as past and as present; lṛṭ with "quickly"; liṅ with a word of hope.
• 2793–2796: where the non-today lakāras (laṅ, luṭ) are barred. Continuity, proximity, a limit (nearer part; farther part optionally), divisions of time other than days and nights. liṅnimitte lṛṅ kriyātipattau is recalled.
• 2797–2801: the lṛṅ headings; censure. lṛṅ in the past too, optionally up to 2809. laṭ with api and jātu; liṅ (and laṭ) with katham; liṅ and lṛṭ with a form of kim.
• 2802–2808: incredibility and intolerance; censure and astonishment with yac ca / yatra. liṅ and lṛṭ; lṛṭ alone with kiṃkila and "is"; liṅ alone with jātu, yad (yadā, yadi), yac ca, yatra; lṛṭ in astonishment otherwise, but not with yadi.
• 2809–2816: liṅ for certainty, a disclosed wish, crediting with ability, cause and result, desire. liṅ and loṭ after desire-verbs (different agents), liṅ alone (same agent), liṅ optionally in the present after desire-verbs. The six vidhi-senses of liṅ (and loṭ) are recalled.
• 2817–2823: the injunctive suffixes. The kṛtyas and loṭ in urging, permission and "time come"; liṅ too after a muhūrta; loṭ with sma; loṭ in respectful urging; liṅ with yad + "time"; the kṛtyas, tṛc and liṅ for a fit agent; liṅ and the kṛtyas for ability. māṅi luṅ is recalled.
• 2824: suffixes of one time used in another, when two dhātu-meanings are related: vasan dadarśa.
• 2825–2828: loṭ for repetition and for aggregation. hi and sva replace it, for every person and number, and a finite verb follows: yāhi yāhīti yāti; saktūn piba dhānāḥ khādety abhyavaharati.

Threads that run through the prakaraṇa:

• The lakāra map the rules work against (from the bhvādi section).
• laṭ: the present.
• Past: laṅ for not of today; liṭ for not of today and out of the speaker's sight; luṅ for the past in general.
• Future: luṭ for not of today; lṛṭ for the future in general.
• liṅ and loṭ: the vidhi-senses.
• lṛṅ: non-accomplishment (kriyātipatti) where there is a cause for liṅ.

Almost every rule here is an apavāda of one or more of these. The commentators name it each time: "laṅo 'pavādaḥ" (2773), "liṭo 'pavādaḥ" (2778), "an apavāda of luṭ and the rest" (2783), "liṅo 'pavādaḥ" (2803), "lṛṭo 'pavādaḥ" (2804), "an apavāda of all the lakāras" (2801, 2806, 2808, 2810, 2814, 2821, 2825).

• Time headings and where they stop.
• bhūte (SK 2995) is the heading for 2773–2782.
• anadyatane runs from 2773. 2780 says it and parokṣa have ceased; 2782 brings anadyatane back by the frog's leap (maṇḍūkapluti).
• bhaviṣyati (SK 3171) holds from 2783 to 2788. At 2790 "in future time" comes from the sense of hope, and at 2796 'bhaviṣyati ity eva'.
• From 2799 on, rules hold in all three times (kālatraye). 'bhaviṣyati' has ceased, so laṭ, liṅ and lṛṭ there stand for past, present and future alike (BM, TB at 2799–2810).
• lṛṅ in non-accomplishment (kriyātipatti). liṅnimitte lṛṅ kriyātipattau (SK 2229) is recalled at 2796: in the future, lṛṅ is obligatory. 2797 bhūte ca makes it a heading for the past as well. 2798 votāpyoḥ (vā ā utāpyoḥ) makes it optional in the past up to, not including, 2809. From 2809 on it is obligatory in the past too.
• This applies only where a cause for liṅ is present. So there is no lṛṅ at 2799 (laṭ is enjoined) or at 2803 (lṛṭ alone). Elsewhere the kaumudī's "lṛṅ prāgvat" (2801, 2802, 2804) means "obligatory in the future, optional in the past" (BM, TB).
• Short anuvṛtti chains, each stopped by a nivṛtti or a yogavibhāga that the kaumudī or BM names:
• garhāyām: 2799 → 2800–2801; ceased at 2802.
• anavakḷptyamarṣayoḥ: 2802 → 2803–2805; ceased at 2806.
• yaccayatrayoḥ: 2805 → 2806–2807. 2805 is split off from 2804 so that only these two words continue.
• praiṣa etc.: 2817 → 2818–2819.
• ūrdhvamauhūrtike: 2818 → 2819. It is stopped at 2820 by the split from 2819.
• kriyāsamabhihāre: 2825, and not into 2826 (2828).
• Splits of a rule (yogavibhāga) made for what follows: 2795 (so that only the temporal limit continues into 2796), 2805 and 2820. The ca that draws something in (anukarṣaṇa): liṭ at 2776; laṭ at 2782, 2788, 2800; "as for the present" at 2790; loṭ and the kṛtyas at 2818; liṅ at 2822; the kṛtyas at 2823; loṭ at 2817 (for prāptakāla).
• Where the commentary lies. Only the kaumudī blocks are translated; the commentators' settled readings come from the unit notes.
• BM:
• none in our corpus for 2793 itself (sk.2793.b1 is BM on 2794–2795; Wikisource has BM on 2793), 2815 and 2816 (Wikisource has 2816);
• sk.2815.b1 is BM on 2821;
• the second part of sk.2826.b1 is BM on 2827.
• TB:
• on 2773: the tail of sk.2771.t1;
• sk.2777.t1 carries 2778, 2780 and 2781;
• sk.2784.t1 carries 2785;
• sk.2787.t1 carries 2788;
• sk.2797.t1 tail carries 2798;
• sk.2806.t1 carries 2799–2805;
• sk.2818.t1 tail carries 2819;
• sk.2822.t1 from 'dhātusaṃbandhe' carries 2824;
• the tail of sk.2828.t1 is TB on SK 2829 (p66).
• No TB anywhere for 2776, 2779 or 2823 (print confirms).

The p62–66 audit (.work/sk-run/p62-66-audit/, list .work/sk-run/p62-66-audit-fixes.md) passed every block's English. MUST-FIX items are being applied by fixers at note level, in parallel; the outline already reads them the corrected way:

• 2776: the domain is liṭ. Ours has लिङ्विषये; the print and the sense have लिड्विषये, and ca adds liṭ (cakāra). A textual note is being added.
• 2785: BM's own conclusion (Wikisource BM, uttara-3/0029, a line our BM lost by eye-skip): following the Bhāṣya on yadvṛttān nityam, every kiṃvṛtta, kadā and the rest included, is taken. The restriction to the case-forms of kim and its ḍatara/ḍatama forms is the Vṛtti's view, which BM reports and rejects. TB's 'kecit' (kadā, kutra not taken, for want of usage) sides with the Vṛtti. This is a BM/TB disagreement; the block English ("a form of kim") is neutral.
• 2828: "'kriyāsamabhihāre' does not carry over into the second sūtra" (reading ananuvṛtteḥ). Ours and NSP read anuvṛtteḥ. The negative is supported by BM's pratīka 'ananuvṛtter iti', TB's 'ity ananuvartanāt', the Wikisource kaumudī, Bhaṭṭoji's undoubled samuccaya examples and a likely haplography at the print's line break 'anu-'.

Note-level MINORs that bear on the outline:

• 2792: keep Bhaṭṭoji's "na tu bhūtavat". TB's "na vartamānavat" too stays in the note.
• 2793: sāmīpya = no intervening item of the same kind. For example, no other full moon in between, though some days may intervene (BM, Wikisource).
• 2788: BM construes ūrdhvamauhūrtike with the loḍarthalakṣaṇa.
• 2800: BM attaches kathaṃ nāma … atyakṣyat to the obligatory future lṛṅ.
• 2809: the rule applies when uta or api is construed.
• 2822: arha = a fit agent.
• 2804, 2806, 2807: these rules also hold in all three times (TB).
• 2826: samuccaya is from ci (cinoti).

Print readings (NSP, scan-confirmed) that the outline follows:

• 2774: abhuñjmahi.
• 2775: uktaviṣaye.
• 2776: liḍviṣaye, and the missing 'cāl liṭ'.
• 2778: yudhiṣṭhiraḥ.
• 2781: ahaṃ nu karomi.
• 2782: anadyatana-grahaṇam.
• 2783: purā yāsyati.
• 2788: muhūrtād upari.
• 2791: lṛḍgrahaṇam (the sense requires it; the scan cannot separate ṅ and ḍ).
• 2793: somenāyaṣṭa; amāvāsyā.
• 2795: āgrahāyaṇyāḥ.
• 2796: lṛṅ in the recalled SK 2229.
• 2800: atyakṣyat.
• 2813: hantīti.
• 2828: yathāyatham, ayāsīt, dādhikam, harāmarāṅganāḥ, lavanādi, and the two colophons.

Still open for Ajai (mūla):

• 2828 'asyānuvṛtteḥ' is ours and the print. BM, TB and Wikisource read ananuvṛtteḥ, and the English follows them in a marked bracket.
• 2776 'liṅviṣaye' is ours; the print has liḍviṣaye. This is an ours ≠ print slip, for the print check.
• 2828 'avaskandesyādi' and 'bhuṅdhvam' are ours and the print. The expected forms are avaskandety ādi (so Wikisource) and bhuṅgdhvam. The English gives the printed bhuṅdhvam with a note.
• The other ours ≠ print slips above are for Ajai's printed copy. source.json is untouched.
• Commentary lemma divergences (not mūla):
• 2781 and 2785: BM glosses kaumudī words absent from NSP but present in Wikisource;
• 2796: TB reads varjyam;
• 2802: BM refers to a kiṃvṛtta example not found here;
• 2808: BM reads mūko 'dhīyīta.

Ordering: strict SK order, which is also Aṣṭādhyāyī order. 2824 forms a group of its own (J), because its outcome (a suffix of one time used in another) is not tied to any particle. The long unit 2828 carries the examples of 2825–2827 as well; group K lays out its forms once, under the sūtra that owns them.

A. lṛṭ for the past not of today after a word of recollection (and liṭ for out of sight) — running example: smarasi kṛṣṇa gokule vatsyāmaḥ (lṛṭ, meaning "we lived") against abhijānāsi kṛṣṇa yad vane 'bhuñjmahi (with yad: laṅ)

• 2773 abhijñāvacane lṛṭ 3.2.112 — when a word that makes recollection known is the upapada, lṛṭ comes in the past not of today. An apavāda of laṅ (SK 2205). bhūte is a heading; anadyatane is carried from SK 2205 (BM, TB).
• In smarasi kṛṣṇa gokule vatsyāmaḥ, smarasi is the recollection-word, and the past residence in Gokula is put in lṛṭ: vatsyāmaḥ = avasāma. The whole clause is the object of smarasi (BM, TB).
• The same holds with budhyase, cetayase and the like, which make recollection known by context.
• 2774 na yadi 3.2.113 — not in connection with yad: the lṛṭ of 2773 does not come, and the general rule laṅ applies (TB). abhijānāsi kṛṣṇa yad vane abhuñjmahi (laṅ; print abhuñjmahi, ours abhuñjamahi).
• 2775 vibhāṣā sākāṅkṣe 3.2.114 — in the same domain (print uktaviṣaye), lṛṭ optionally when the root-meaning is in mutual expectancy, as what marks and what is marked.
• smarasi kṛṣṇa vane vatsyāmas tatra gāś cārayiṣyāmaḥ. The living is the mark and the grazing the marked. In the other option laṅ: avasāma, acārayāma (TB).
• The option applies even with yad, overriding 2774 because it is the later rule (paratvāt). TB: that is why it was not placed before na yadi; BM cites the Bhāṣya 'yadi ca ayadi cāyaṃ vikalpaḥ'.
• Recalled here: parokṣe liṭ (3.2.115, SK 2171): cakāra. In the first person, being out of sight comes about through distraction of mind (sleep, intoxication, absorption): supto 'haṃ kila vilalāpa; bahu jagada purastāt tasya mattā kilāham.
• Vārttika atyantāpahnave liḍ vaktavyaḥ: liṭ in total denial, even of what was not out of sight. kaliṅgeṣv avātsīḥ? — nāhaṃ kaliṅgān jagāma. The denial is total because the going too is denied. Denying only the going gives laṅ: na kaliṅgeṣv agaccham (TB).

B. The past with particles: laṅ beside liṭ; laṭ; luṅ and laṭ — running example: vasantīha purā chātrāḥ (laṭ) / avātsuḥ (luṅ) / avasan (laṅ) / ūṣuḥ (liṭ), "formerly students lived here"

• 2776 haśaśvator laṅ ca 3.2.116 — with ha or śaśvat as upapada, laṅ in the domain of liṭ (past, not of today, out of sight), and by ca liṭ too.
• iti ha akarot / cakāra; śaśvad akarot / cakāra.
• The print reads liḍviṣaye and 'laṅ syāt cāl liṭ'; ours has liṅviṣaye and lacks cāl liṭ (see "open").
• 2777 praśne cāsannakāle 3.2.117 — when the matter asked about is past, not of today and out of sight, but of recent time: laṅ and liṭ.
• agacchat kim / jagāma kim. But when not recent: kaṃsaṃ jaghāna kim (liṭ only).
• "Recent": TB gives within five years; BM gives within the user's range of sight and calls the five years the Vṛtti's. The English stays neutral.
• 2778 laṭ sme 3.2.118 — with sma, laṭ in the domain of liṭ; an apavāda of liṭ. sma makes past time manifest (BM, TB). yajati sma yudhiṣṭhiraḥ, "Yudhiṣṭhira sacrificed".
• 2779 aparokṣe ca 3.2.119 — laṭ with sma also when the past not of today is not out of sight. evaṃ sma pitā bravīti, "thus father said".
• 2780 nanau pṛṣṭaprativacane 3.2.120 — in the reply of one who has been asked, with nanu: laṭ in the past.
• anadyatana and parokṣa have ceased.
• akārṣīḥ kim? — nanu karomi bhoḥ ("why, I did").
• TB: it applies in the reply only, and 'pṛṣṭa' is needed (Kaiyaṭa).
• 2781 nanvor vibhāṣā 3.2.121 — with na or nu (in a reply), laṭ optionally in the past.
• akārṣīḥ kim? — na karomi / na akārṣam; ahaṃ nu karomi / ahaṃ nv akārṣam (print nu; ours tu).
• In the other option luṅ, but laṅ in the past not of today (TB).
• 2782 puri luṅ cāsme 3.2.122 — with purā, in the past not of today (anadyatana back by the frog's leap): luṅ optionally, and by ca laṭ, but not with sma.
• In the other option the lakāra as otherwise obtained: liṭ if out of sight, otherwise laṅ (BM, TB). Hence the four forms of the running example.
• Counter-example: yajati sma purā. Here laṭ comes by 2778, not luṅ.
• Close of 3.2: "[what follows is taught] while bhaviṣyati is carried forward", from SK 3171 (BM).

C. laṭ (and liṅ) for the future with particles and senses — running example: kadā bhuṅkte / karhi bhuṅkte (laṭ) / bhokṣyate (lṛṭ) / bhoktā (luṭ), "when will he eat?"

• 2783 yāvatpurānipātayor laṭ 3.3.4 — with the nipātas yāvat and purā (which convey certainty), laṭ in the future: yāvad bhuṅkte, purā bhuṅkte, "he will surely eat". An apavāda of luṭ and the rest (BM).
• Counter-examples:
• yāvad dāsyate tāvad bhokṣyate: here yāvat is formed with vatup and is not a nipāta;
• karaṇībhūtayā purā yāsyati: here purā is the instrumental of pur (print purā; ours pūrā).
• TB adds the luṭ counter-example yāvad dātā tāvad bhoktā.
• 2784 vibhāṣā kadākarhyoḥ 3.3.5 — with kadā or karhi, laṭ optionally in the future; otherwise lṛṭ and luṭ as obtained.
• karhi bhokṣyate is disputed, since karhi is taught for not of today:
• BM says the lṛṭ example is for kadā alone;
• TB says it is a general future, with not-of-today understood by lakṣaṇā.
• TB: luṭ alternates because these sūtras are carried into anadyatane luṭ (SK 2185).
• 2785 kiṃvṛtte lipsāyām 3.3.6 — with a form of kim (kiṃvṛtta), when the questioner shows a wish to obtain (lipsā): laṭ optionally in the future.
• kaṃ / kataraṃ / katamaṃ bhojayasi / bhojayiṣyasi / bhojayitāsi ("[me, hungry]" understood, BM).
• Counter-example (no lipsā): kaḥ pāṭaliputraṃ gamiṣyati.
• Scope of kiṃvṛtta:
• BM (Wikisource): the Vṛtti takes only the case-forms and the ḍatara and ḍatama forms (so with kadā and the like, laṭ comes by 2784). BM's own conclusion, following the Bhāṣya on yadvṛttān nityam: all kiṃvṛttas, kadā included.
• TB's 'kecit' exclude kadā, kutra and the like.
• 2786 lipsyamānasiddhau ca 3.3.7 — when the accomplishment of heaven and the like through the thing sought (food and the like) is understood: laṭ optionally in the future. This is for the case without a form of kim (BM, TB).
• yo 'nnaṃ dadāti / dāsyati / dātā sa svargaṃ yāti / yāsyati / yātā, construed pairwise.
• 2787 loḍarthalakṣaṇe ca 3.3.8 — after a root whose sense marks loṭ's meaning (urging and the like): laṭ optionally in the future. Otherwise luṭ and lṛṭ.
• kṛṣṇaś ced bhuṅkte tvaṃ gāś cāraya: Kṛṣṇa's eating marks the urging to graze the cows (BM, TB).
• 2788 liṅ cordhvamauhūrtike 3.3.9 — for what lies beyond a muhūrta (ūrdhvamauhūrtika), in the same loḍarthalakṣaṇa sense: liṅ and by ca laṭ, optionally.
• ūrdhvamauhūrtika takes ṭhañ by kālāṭ ṭhañ. Its compound and uttarapada-vṛddhi come by nipātana.
• muhūrtād upari upādhyāyaś ced āgacchet / āgacchati / āgamiṣyati / āgantā, atha tvaṃ chando 'dhīṣva.
• BM construes ūrdhvamauhūrtike with the loḍarthalakṣaṇa.

D. Time by extension (atideśa): near-present as present; hope as past and present; lṛṭ with "quickly"; liṅ with a word of hope — running example: devaś ced avarṣīt / varṣati / varṣiṣyati, dhānyam avāpsma / vapāmaḥ / vapsyāmaḥ ("if the god rains, we shall sow"; luṅ as for the past, laṭ as for the present, lṛṭ)

• 2789 vartamānasāmīpye vartamānavad vā 3.3.131 — in a past or future near the present, all the suffixes taught for the present (from vartamāne laṭ, SK 2151, up to uṇādayo bahulam, SK 3169) come optionally, under the same conditions.
• kadā āgato 'si? — ayam āgacchāmi (laṭ: I have just come) / ayam āgamam (luṅ).
• kadā gamiṣyasi? — eṣa gacchāmi (laṭ: I am about to go) / gamiṣyāmi (lṛṭ).
• sāmīpya is ṣyañ in its own sense. TB: the Bhāṣya rejects this sūtra (superimposition, as for what is near the Gaṅgā).
• 2790 āśaṃsāyāṃ bhūtavac ca 3.3.132 — when hope is expressed (necessarily about the future): the suffixes come optionally as for the past, and by ca as for the present. 'vartamānasāmīpye' is not carried.
• The running example is construed pairwise (BM):
• avarṣīt … avāpsma, luṅ as for the past;
• varṣati … vapāmaḥ, laṭ;
• varṣiṣyati … vapsyāmaḥ, lṛṭ.
• sāmānyātideśe viśeṣānatideśaḥ: only the general past (luṅ) is extended, so no laṅ or liṭ.
• 2791 kṣipravacane lṛṭ 3.3.133 — in hope, with a word meaning "quickly" as upapada: lṛṭ. An apavāda of 2790.
• vṛṣṭiś cet kṣipram / āśu / tvaritaṃ vā yāsyati, śīghraṃ vapsyāmaḥ.
• Why lṛṭ is named (lṛḍgrahaṇam, print) where 'na' would have done: so that lṛṭ comes in luṭ's domain too: śvaḥ śīghraṃ vapsyāmaḥ. BM (Wikisource): with 'na' only the general future's lṛṭ would be negated, not luṭ.
• 2792 āśaṃsāvacane liṅ 3.3.134 — with a word expressing hope as upapada: liṅ in the future, "and not as for the past" (Bhaṭṭoji; TB adds "nor as for the present").
• gurus ced upeyād āśaṃse adhīyīya.
• āśaṃse kṣipram adhīyīya: liṅ wins over 2791's lṛṭ, because this rule is later (BM, TB).

E. Where the non-today lakāras (laṅ, luṭ) are barred: continuity, proximity, limits — running example: yāvajjīvam annam adāt (luṅ, not laṅ) / dāsyati (lṛṭ, not luṭ)

• 2793 nānadyatanavat kriyāprabandhasāmīpyayoḥ 3.3.135 — when continuity of the action or proximity is understood, the suffixes are not as for non-today time: laṅ (past) and luṭ (future) do not come. So luṅ and lṛṭ (TB).
• Continuity: yāvajjīvam annam adād / dāsyati.
• Proximity = no intervening item of the same kind. Days of the dark fortnight may intervene, but no other full moon (BM, Wikisource).
• yeyaṃ paurṇamāsy atikrāntā tasyām agnīn ādhita, somena ayaṣṭa (luṅ; print somenāyaṣṭa).
• yeyam amāvāsyā''gāminī tasyām agnīn ādhāsyate, somena yakṣyate (lṛṭ).
• 2794 bhaviṣyati maryādāvacane 'varasmin 3.3.136 — in the future, when a limit is stated, for the nearer part: not as for non-today time, so no luṭ, lṛṭ only.
• This is for cases with no continuity and no proximity (Bhāṣya, Kaiyaṭa via BM).
• yo 'yam adhvā gantavya ā pāṭaliputrāt tasya yad avaraṃ kauśāmbyās tatra saktūn pāsyāmaḥ.
• TB's counter-example for the farther part: … yat paraṃ kauśāmbyās tatra saktūn pātāsmaḥ (luṭ).
• 2795 kālavibhāge cānahorātrāṇām 3.3.137 — the whole of 2794 is carried. For a division of time the rule holds, provided it is not one of days and nights.
• The split from 2794 is for the sake of 2796 (so that only the temporal limit continues, TB).
• yo 'yaṃ vatsara āgāmī tasya yad avaram āgrahāyaṇyās tatra yuktā adhyeṣyāmahe (lṛṭ).
• Counter-example: … yo 'varaḥ pañcadaśarātras tatra adhyetāsmahe (luṭ comes).
• 2796 parasmin vibhāṣā 3.3.138 — 2794–2795 are carried, without 'avarasmin'. For the farther part the bar is optional (an aprāptavibhāṣā), and 'anahorātrāṇām' still applies (TB).
• … yat param āgrahāyaṇyās tatra adhyeṣyāmahe / adhyetāsmahe.
• Recalled here: liṅnimitte lṛṅ kriyātipattau (3.3.139, SK 2229): where there is a cause for liṅ and the action is not accomplished, lṛṅ, with 'bhaviṣyati' still in force. suvṛṣṭiś ced abhaviṣyat tadā subhikṣam abhaviṣyat.

F. lṛṅ extended to the past; censure (garhā) — running example: kathaṃ dharmaṃ tyajeḥ / tyajasi ("how could you abandon dharma!"; liṅ / laṭ in any time), with kriyātipatti kathaṃ nāma tatrabhavān dharmam atyakṣyat (lṛṅ)

• 2797 bhūte ca 3.3.140 — a heading. The whole of 3.3.139 is carried: where there is a cause for liṅ, lṛṅ in non-accomplishment in the past too. The past here is the past in general (TB).
• 2798 votāpyoḥ 3.3.141 — vā ā utāpyoḥ. Up to, not including, 2809 the heading is "in the past, where there is a cause for liṅ, lṛṅ optionally". 2797 operates from 2809 on.
• Both headings work only where liṅ is enjoined (BM). In the future lṛṅ is always obligatory (TB).
• 2799 garhāyāṃ laḍ apijātvoḥ 3.3.142 — in censure, with api or jātu: laṭ in all three times. It blocks luṅ and the others, being later and otherwise without scope (paratvāt, anavakāśatva).
• api jāyāṃ tyajasi, jātu gaṇikām ādhatse.
• votāpyoḥ does not apply here, since laṭ is enjoined and there is no cause for liṅ.
• 2800 vibhāṣā kathami liṅ ca 3.3.143 — in censure, with katham: liṅ and by ca laṭ, optionally, in the three times.
• kathaṃ dharmaṃ tyajeḥ / tyajasi.
• In the other option, the lakāras proper to each time (TB): tatyaktha, atyajaḥ, atyākṣīḥ, tyaktāsi, tyakṣyasi.
• In kriyātipatti, lṛṅ is obligatory in the future and optional in the past: kathaṃ nāma tatrabhavān dharmam atyakṣyat (print; ours atyakṣat) / atyākṣīt. BM gives this as the example of the obligatory future lṛṅ.
• 2801 kiṃvṛtte liṅlṛṭau 3.3.144 — in censure, with a form of kim: liṅ and lṛṭ.
• 'vibhāṣā' is not carried, so the ordinary lakāras do not come as an alternative (TB).
• kaḥ / kataraḥ / katamo vā hariṃ nindet / nindiṣyati. lṛṅ as before.

G. Incredibility and intolerance; censure and astonishment with yac ca / yatra — running example: na saṃbhāvayāmi na marṣaye vā bhavān hariṃ nindet / nindiṣyati (liṅ / lṛṭ: "that your honour should revile Hari — I do not think it possible / I do not tolerate it")

• 2802 anavakḷptyamarṣayor akiṃvṛtte 'pi 3.3.145 — in incredibility (asaṃbhāvanā) or intolerance (akṣamā), even without a form of kim: liṅ and lṛṭ (carried from 2801), in the three times.
• 'garhāyām' has ceased.
• There is no respective correlation (yathāsaṃkhya), since amarṣa, the shorter word, is not placed first (BM, TB).
• The running example; lṛṅ as before.
• 2803 kiṃkilāstyartheṣu lṛṭ 3.3.146 — with kiṃkila (anger) or a verb meaning "is" (asti, bhavati, vidyate): lṛṭ alone, an apavāda of liṅ. 'anavakḷptyamarṣayoḥ' runs up to, not into, 2806.
• na śraddadhe na marṣaye vā kiṃkila tvaṃ śūdrānnaṃ bhokṣyase.
• asti / bhavati / vidyate vā śūdrīṃ gamiṣyasi.
• No lṛṅ here: liṅ has been excluded, so there is no cause for it.
• 2804 jātuyador liṅ 3.3.147 — with jātu or yad (an indeclinable resembling a case-form): liṅ alone, an apavāda of lṛṭ.
• Vārttika yadāyadyor upasaṃkhyānam: yadā and yadi too.
• jātu / yad / yadā / yadi vā tvādṛśo hariṃ nindet, nāvakalpayāmi na marṣayāmi. lṛṅ as before: jātu bhavān harim anindiṣyat (TB).
• Our k1 is a lone daṇḍa (markup only).
• 2805 yaccayatrayoḥ 3.3.148 — with yac ca or yatra: liṅ (in the same senses).
• yac ca / yatra vā tvam evaṃ kuryāḥ, na śraddadhe na marṣayāmi.
• It is split from 2804 so that only these two continue (BM: into the next sūtra; TB: into the next two).
• 2806 garhāyāṃ ca 3.3.149 — in censure, with yac ca or yatra: liṅ alone. 'anavakḷptyamarṣayoḥ' has ceased.
• TB: it is the apavāda of the suffixes of 2800 and 2801.
• yac ca / yatra vā tvaṃ śūdraṃ yājayeḥ: that is improper. lṛṅ as before: … ayājayiṣyaḥ (TB).
• 2807 citrīkaraṇe ca 3.3.150 — in astonishment, with yac ca or yatra: liṅ. yac ca / yatra vā tvaṃ śūdraṃ yājayeḥ: this is astonishing.
• 2808 śeṣe lṛḍ ayadau 3.3.151 — in astonishment with any other co-occurring word, lṛṭ, an apavāda of all the lakāras. But not with yadi.
• āścaryam andho nāma kṛṣṇaṃ drakṣyati.
• Counter-example: āścaryaṃ yadi so 'dhīyīta, with liṅ by the 2804 vārttika (TB). BM's lemma reads mūko 'dhīyīta.

H. liṅ for certainty, a disclosed wish, crediting with ability, cause and result, and desire — running example: the eating sentences, kāmo me bhuñjīta bhavān (2810) → sambhāvayāmi bhuñjīta / bhokṣyate vā bhavān (2812) → icchāmi bhuñjīta / bhuṅktāṃ vā bhavān (2814) → bhuñjīyeti icchati (2815)

• 2809 utāpyoḥ samarthayor liṅ 3.3.152 — when uta or api is construed in the shared meaning "certainly" (bāḍham): liṅ, in the three times. uta / api hanyād aghaṃ hariḥ.
• Counter-examples, with other meanings: uta daṇḍaḥ patiṣyati (a question); apidhāsyati dvāram (api = covering).
• From here on, lṛṅ in kriyātipatti is obligatory in the past too (the limit of 2798): utāhaniṣyad dasyuṃ rājā (TB).
• 2810 kāmapravedane 'kacciti 3.3.153 — when one's own wish is disclosed, liṅ, but not with kaccit. An apavāda of all the lakāras.
• kāmo me bhuñjīta bhavān.
• Counter-example: kaccij jīvati. With kaccit, laṭ comes by 2789 (TB).
• 2811 sambhāvane 'lam iti cet siddhāprayoge 3.3.154 — liṅ (carried from 2809) in the sense of sambhāvana (crediting someone with an ability), if the sambhāvana is "alam" (he is equal to it), when alam is not used but understood.
• The kaumudī recasts sambhāvane as sambhāvanam and alam as alami. BM and TB: the words are repeated (āvṛtti) and one copy is recast.
• api giriṃ śirasā bhindyāt.
• Counter-example (alam is used): alaṃ kṛṣṇo hastinaṃ haniṣyati.
• BM and TB differ on how sambhāvana is defined and on the force of iti. The English is neutral.
• 2812 vibhāṣā dhātau sambhāvanavacane 'yadi 3.3.155 — when a dhātu meaning sambhāvana is the upapada: liṅ optionally, but not with yad. Without this rule 2811 would make liṅ obligatory.
• sambhāvayāmi bhuñjīta / bhokṣyate vā bhavān. The lṛṭ comes from 2808 (BM).
• Counter-example: sambhāvayāmi yad bhuñjīthās tvam. Here liṅ is obligatory by 2811.
• 2813 hetuhetumator liṅ 3.3.156 — after a root in the sense of a cause or of its result: liṅ optionally (vibhāṣā carried); otherwise lṛṭ.
• kṛṣṇaṃ namec cet sukhaṃ yāyāt; kṛṣṇaṃ naṃsyati cet sukhaṃ yāsyati.
• Vārttika bhaviṣyaty eveṣyate: only for the future. So not in hantīti palāyate (print hantīti; ours hantāti).
• 2814 icchārtheṣu liṅloṭau 3.3.157 — with a verb of desire as upapada, when the agents differ (BM, TB): liṅ and loṭ, an apavāda of all the lakāras.
• icchāmi bhuñjīta / bhuṅktāṃ vā bhavān. Likewise with kāmaye and prārthaye.
• Vārttika kāmapravedana iti vaktavyam: only when the wish is disclosed to another. So not in icchan karoti.
• 2815 liṅ ca 3.3.159 — with a desire-verb of the same agent: liṅ. bhuñjīyeti icchati.
• TB: it restores liṅ (pratiprasava) against samānakartṛkeṣu tumun (3.3.158, SK 3176), and only liṅ, not loṭ.
• 2816 icchārthebhyo vibhāṣā vartamāne 3.3.160 — after desire-verbs themselves, liṅ optionally in the present; otherwise laṭ. icchet / icchati; kāmayeta / kāmayate.
• Recalled here: vidhinimantraṇāmantraṇādhīṣṭasaṃpraśnaprārthaneṣu liṅ (3.3.161, SK 2208). liṅ in its six senses:
• vidhi: yajeta;
• nimantraṇa: iha bhuñjīta bhavān;
• āmantraṇa: iha āsīta;
• adhīṣṭa: putram adhyāpayed bhavān;
• saṃpraśna: kiṃ bho vedam adhīyīya uta tarkam;
• prārthana: bhojanaṃ labheya.
• "evaṃ loṭ" recalls loṭ ca (3.3.162, SK 2194): loṭ in the same senses. BM (Wikisource) says these are recalled because they come next in order.

I. The injunctive suffixes: kṛtyas, loṭ, liṅ, tṛc — running example: bhavatā yaṣṭavyam (kṛtya) / bhavān yajatām (loṭ) → muhūrtād ūrdhvaṃ yajeta / yajatām / yaṣṭavyam (2818) → ūrdhvaṃ muhūrtād yajatāṃ sma (2819)

• 2817 praiṣātisargaprāptakāleṣu kṛtyāś ca 3.3.163 — in urging (praiṣa = vidhi), permission to act at will (atisarga) and "the time has come" (prāptakāla): the kṛtyas, and by ca loṭ, drawn in from SK 2194.
• bhavatā yaṣṭavyam; bhavān yajatām.
• Drawing in loṭ is for prāptakāla only, since in praiṣa and atisarga loṭ already comes by loṭ ca (BM, TB).
• 2818 liṅ cordhvamauhūrtike 3.3.164 — in the same three senses (carried), for what lies beyond a muhūrta: liṅ, and by ca loṭ and the kṛtyas are not blocked.
• muhūrtād ūrdhvaṃ yajeta / yajatām; yaṣṭavyam.
• Here liṅ is new only in prāptakāla (BM, TB).
• Compare 2788, a different sūtra with the same words; there BM lists liṅ, laṭ, luṭ and lṛṭ.
• 2819 sme loṭ 3.3.165 — in the domain of 2818 (praiṣa etc., beyond a muhūrta), with sma: loṭ alone, an apavāda of liṅ and the kṛtyas. ūrdhvaṃ muhūrtād yajatāṃ sma.
• 2820 adhīṣṭe ca 3.3.166 — with sma, in respectful urging (adhīṣṭa): loṭ, an apavāda of the adhīṣṭa-liṅ of SK 2208. tvaṃ sma adhyāpaya (said to a teacher).
• It is split from 2819 to stop 'ūrdhvamauhūrtike' (BM, TB).
• The next sūtra, 3.3.167 (SK 3179), is left to the kṛt section.
• 2821 liṅ yadi 3.3.168 — with yad and kāla, samaya or velā as upapadas (carried from 3.3.167, SK 3179): liṅ. It is an apavāda of all the lakāras and of tumun.
• kālaḥ samayo velā vā yad bhuñjīta bhavān.
• TB also carries praiṣa etc., hence not in yatkālaḥ pacati. BM is silent on this.
• 2822 arhe kṛtyatṛcaś ca 3.3.169 — when a fit agent (arha) is understood: the kṛtyas and tṛc, and by ca liṅ.
• tvaṃ kanyāṃ vaheḥ.
• TB's kṛtya and tṛc forms: tvayā kanyā voḍhavyā / vahanīyā; tvaṃ kanyāṃ voḍhā.
• liṅ is new here. The kṛtyas and tṛc are named so that liṅ does not block them (TB).
• 2823 śaki liṅ ca 3.3.172 — in capability (śak, a kvip-form in the bhāva sense): liṅ, and by ca the kṛtyas. tvaṃ bhāraṃ vaheḥ.
• Recalled here: māṅi luṅ (3.3.175, SK 2219): mā kārṣīḥ.
• Then how can one say mā bhavatu, mā bhaviṣyati? That is the word mā, not māṅ. māṅi luṅ is still needed for the aṭ-less forms by na māṅyoge (SK 2228) (BM).

J. Suffixes of one time used in another, when dhātu-meanings are related — running example: vasan dadarśa ("dwelling, he saw": laṭ, replaced by śatṛ, for the past)

• 2824 dhātusambandhe pratyayāḥ 3.4.1 — when two dhātu-meanings are related, suffixes taught for one time may come in another. They do so after the dhātu of the subordinate action, to conform to the principal action, which the tiṅ-form expresses.
• vasan dadarśa = uṣitvā dadarśa: the seeing has as agent the agent of a past dwelling.
• somayājy asya putro bhavitā: ṇini, taught for the past by karaṇe yajaḥ (SK 2996), is here read as future to match the luṭ bhavitā.
• BM adds gomān āsīt: matup, taught for a present "is", used of a past "was". So "dhātu" means the dhātu-meaning, not two dhātus actually used (BM, TB).

K. loṭ for repetition and for aggregation: hi and sva, then the finite verb that follows — running example: yāhi yāhīti yāti (kriyāsamabhihāra: "he keeps going") and saktūn piba dhānāḥ khādety abhyavaharati (samuccaya: "'drink the gruel, chew the grain' — thus he takes his meal")

• 2825 kriyāsamabhihāre loṭ; loṭo hisvau; vā ca tadhvamoḥ 3.4.2 (ours prints it in three parts; NSP has one sūtra, one number). When repetition or intensity is to be made known, loṭ after the dhātu; in its place hi and sva, an apavāda of the tiṅ.
• With loṭ carried into 'loṭo hisvau' (BM, TB), hi and sva receive the designations parasmaipada and ātmanepada respectively, and tiṅ (hence padahood).
• In the domain of ta and dhvam (2nd plural), hi and sva are optional.
• By force of their prescription they do not take the person and singular designations. So hi comes in every person and number after parasmaipadin roots, in the agent sense; sva comes after ātmanepadin roots, in bhāva, karman and kartṛ.
• The forms are given under 2828.
• 2826 samuccaye 'nyatarasyām 3.4.3 — when an aggregation of several actions is to be made known: the provisions of 2825 come optionally. These are loṭ, hi and sva, and the option for ta and dhvam; 'samabhihāre' is not carried (BM). The examples are under 2828.
• 2827 yathāvidhy anuprayogaḥ pūrvasmin 3.4.4 — in the first loṭ-rule (2825), the dhātu used afterwards (anuprayoga) must be the loṭ's own base. So yāhi yāhīti yāti, not gacchati (TB).
• Immediate contiguity is not required (BM). TB: the Bhāṣya rejects the sūtra.
• 2828 samuccaye sāmānyavacanasya 3.4.5 — in the aggregation rule (2826), the dhātu used afterwards is one of general meaning. laṭ etc. and tip etc. after it make the number, time and person manifest.
• Vārttika kriyāsamabhihāre dve vācye: in repetition, the loṭ-form is doubled.
• Repetition (2825, 2827), doubled:
• yāhi yāhīti yāti, and likewise yātaḥ, yānti, yāsi, …;
• in the ta-domain, the other option: yāta yāteti yūyaṃ yātha;
• luṅ: yāhi yāhīty ayāsīt (print; ours ayāsīti); lṛṭ: yāsyati (BM knows the variant ayāsyat);
• with sva: adhīṣva adhīṣvety adhīte; dhvam-domain, the other option: adhīdhvam adhīdhvam iti yūyam adhīdhve.
• Aggregation (2826, 2828), undoubled:
• saktūn piba dhānāḥ khādety abhyavaharati; annaṃ bhuṅkṣva dādhikam āsvādayasvety abhyavaharate;
• in the ta- and dhvam-domain: pibata khādatety abhyavaharatha; bhuṅdhvam (print; standard bhuṅgdhvam) āsvādayadhvam ity abhyavaharadhve, or with hi and sva in the other option;
• without loṭ: saktūn pibati, dhānāḥ khādati.
• The particular actions are construed as identical with the general action of the following verb.
• The verse purīm avaskanda lunīhi nandanaṃ muṣāṇa ratnāni harāmarāṅganāḥ… (Māgha, per BM) is samuccaya loṭ. Its anuprayoga is asvāsthyaṃ cakre, a past, not-of-today, out-of-sight, single-agent action.
• "He stormed again and again" is wrong, because 'kriyāsamabhihāre' does not carry into 2826 (ananuvṛtteḥ; see "open"), and because the loṭ-forms would then have to be doubled.
• Taking avaskanda etc. as 2nd singular is also wrong, since hi and sva have no person or number (2825).
• Close: "[iti lakārārthaprakriyā]" and "[iti … uttarārdhe tiṅantaṃ samāptam]", supplied from the print.