22. Sarvasamāsaśeṣa-prakaraṇa — What remains to be said about all compounds

After this chapter you will be able to place the compound among the five "complex formations" (vṛtti) of the grammar. You will be able to tell apart the two ways of analysing a compound: the ordinary sentence rājñaḥ puruṣaḥ and the technical string rājan as puruṣa su. You will know what an obligatory compound (nitya-samāsa) is. You will see why "compounds are of four kinds" is only a rough statement, and why the older rule of thumb about which member's meaning leads is rough too. You will know how the karmadhāraya and the dvigu fit inside the tatpuruṣa, and which compounds may have many members. Last comes a verse that sorts every compound into six kinds by what joins what. They run from rājapuruṣaḥ (a noun with a noun) to kṛntavicakṣaṇā (a finite verb with a noun). This chapter has no sūtras. It is one short passage of Bhaṭṭoji's own prose, which closes the compound section.

Before you start

This chapter sums up the whole compound section. Bring with you the following.

• From chapter 8 you know that a form in a kṛt affix (added to a root), a form in a taddhita affix (added to a noun), and a compound are all prātipadika, noun stems that take case-endings (SK 179). You also know the first ekaśeṣa, "one remaining": rāma + rāma + au → rāmau (SK 188).
• From chapter 17 (the avyayībhāva) you know the machinery every compound uses. Words compound only if they are samartha, connected in meaning (SK 647). Saha supā (SK 649) compounds a subanta (a word ending in a case-ending, sup) with a subanta; its first word, saha, read as a rule by itself, also joins a subanta to a finite verb in the Veda (paryabhūṣayat). The case-endings inside a compound drop by luk (SK 650). The member named in the nominative in the rule is the upasarjana and goes first (SK 653–654). Compounds taught before SK 665 are obligatory, except the plain subanta-with-subanta compound of SK 649. You also know unmattagaṅgam (SK 675) and śākaprati "a bit of vegetable" (SK 663).
• From chapter 18 (the tatpuruṣa) you know rājñaḥ puruṣaḥ → rājapuruṣaḥ (SK 702). You know the karmadhāraya, the tatpuruṣa of co-referential members (SK 745), and the numeral-first dvigu (SK 730). You know the upapada compound kumbhakāraḥ (SK 782). You know the prādi compound, whose first member is pra or another word of the list headed by pra (SK 761): atikrānto mālām → atimālaḥ (SK 780). You know that a group dvandva such as dantoṣṭham is neuter (SK 821). You know the list mayūravyaṃsakādi (SK 754) with its attached rules for compounds of finite verbs.
• From chapter 19 (the bahuvrīhi) you know the compound that names a thing outside its members, pītāmbaro hariḥ (SK 830), and the numeral bahuvrīhi dvitrāḥ "two or three", made by SK 843 and listed under SK 844.
• From chapter 20 (the dvandva) you know the "and"-compound in its two shapes, the pair dhavakhadirau and the group saṃjñāparibhāṣam (SK 901). You know that a dvandva of body-parts is singular (SK 906), and you met dantoṣṭham "teeth and lips" under SK 919–920.
• From chapter 21 (the ekaśeṣa) you know how one word stands for several of unlike form: mātā ca pitā ca → pitarau.

The passage

Bhaṭṭoji's passage runs through ten points in a fixed order. It starts wide, with every kind of complex formation, and narrows to the compound. It then tests the two traditional ways of classifying compounds and finds both only roughly true. It ends with a verse that classifies compounds a third way, by the kinds of word joined.

Bhaṭṭoji's text The five vṛttis, the vigraha, and the six kinds of compound

1. The five vṛttis

Bhaṭṭoji begins: there are five vṛttis, "complex formations". They are:

• kṛt — a word made by a kṛt affix on a root, such as kāra "maker" from kṛ with aṇ (the kāra of kumbhakāraḥ, chapter 18);
• taddhita — a word made by a taddhita affix on a noun, such as gārgya from garga with yañ (chapter 21);
• samāsa — a compound, such as rājapuruṣa;
• ekaśeṣa — one word left standing for several, such as pitarau for mātā ca pitā ca;
• sanādyanta-dhātu — a root that ends in the affix san or one of the affixes listed with it; such a form is itself a root (A 3.1.32). The verb chapters teach these.

The compound is thus one of five. What the passage says next about vṛtti and vigraha holds for all five.

2. What a vṛtti is, and what a vigraha is

A vṛtti is "the expression of a meaning that belongs to another" (parārtha-abhidhāna). Bhaṭṭoji gives the definition and does not unpack it here.

A vigraha is "the sentence (vākya) that makes known the meaning of the vṛtti". It is the analysis that states, in separate words, what the vṛtti says in one. rājñaḥ puruṣaḥ "the king's man" is the vigraha of rājapuruṣaḥ.

3. Two kinds of vigraha: laukika and alaukika

The vigraha is of two kinds.

• The laukika vigraha, "of ordinary usage", is correct (sādhu), because it is fully formed (pariniṣṭhita). Every operation has run; the words are finished words you could say.
• The alaukika vigraha, "not of ordinary usage", is incorrect (asādhu) and unfit for use (prayoga). It is the string of stems and endings on which the rules operate.

Bhaṭṭoji's example is the pair for rājapuruṣaḥ:

• laukika: rājñaḥ puruṣaḥ;
• alaukika: rājan as puruṣa su.

In the alaukika string, as is the genitive singular ending (taught with a marker as ṅas) and su is the nominative singular. Nothing has been joined or changed yet. That is why no speaker would ever use it.

Here is the compound built from the alaukika string, step by step, with the rules of chapters 8, 17 and 18:

rājan as puruṣa su — the alaukika vigraha: stem + genitive, stem + nominative

→ ṣaṣṭhī (SK 702): a genitive word compounds with a subanta → a samāsa, a tatpuruṣa

→ rājan as, named in the rule in the nominative (ṣaṣṭhī), is upasarjana and goes first (SK 653–654) → rājan as + puruṣa su

→ a compound is a prātipadika (SK 179)

→ the case-endings inside it drop by luk (SK 650) → rājan + puruṣa

→ the final n of rājan drops, as at the end of a word (SK 236) → rājapuruṣa

→ the new stem takes its own ending, nominative singular su → rājapuruṣaḥ.

The laukika rājñaḥ puruṣaḥ stands beside all this as the plain sentence that the compound means.

4. The nitya-samāsa

An obligatory compound (nitya-samāsa) is one of two things:

• a compound with no vigraha at all (avigraha); or
• a compound whose vigraha is not made with its own words (asvapada-vigraha).

The second kind is the one you have met. Chapter 18 showed atikrānto mālām → atimālaḥ "one who has surpassed a garland" (SK 780). Bhaṭṭoji's commentators on SK 780 add that ati itself means krānta "surpassed". So the analysis in other words is krānto mālām. The member ati does not appear in it; krānta stands in for it. The commentators use such an analysis because the compound is obligatory.

Compare upakṛṣṇam of chapter 17. It is made by SK 652, before the heading of SK 665, so it is obligatory (SK 665). Bhaṭṭoji's commentators note on SK 662 that an obligatory compound takes an analysis not in its own words. Its analysis kṛṣṇasya samīpam says samīpam where the compound has upa.

Bhaṭṭoji gives no example of the first kind here.

5. "Compounds are of four kinds" is a rough statement

People say that compounds are of four kinds: avyayībhāva, tatpuruṣa, bahuvrīhi and dvandva. Bhaṭṭoji calls this a prāyovāda, a statement of what generally holds.

His reason: saha supā (SK 649) prescribes compounds also for words that lie outside the four headings (adhikāra) of those four names. In chapter 17, SK 649 stood before the avyayībhāva heading (SK 651), so a compound made by it alone falls under none of the four names. It is still a samāsa.

Two examples from chapter 17 show this. paryabhūṣayat is made by saha alone: pari joined to the finite verb abhūṣayat. And Bhaṭṭoji's commentators note that for bhūtapūrva, analysed pūrvaṃ bhūtaḥ, the name samāsa is settled by supā, subanta with subanta.

6. "Which member's meaning leads" is a rough statement too

The older grammarians (prāc) had a saying (pravāda) about the four kinds:

The pattern is easy to see in the running examples of chapters 17–20. In upakṛṣṇam "near Kṛṣṇa", nearness, the meaning of upa, leads. In rājapuruṣaḥ the man leads, not the king. pītāmbaraḥ means Hari, who is neither "yellow" nor "garment". dhavakhadirau means both trees.

Bhaṭṭoji says this saying too is meant only as what generally holds (prāyo 'bhiprāyaḥ). The pattern is absent in four sets of forms, one for each kind:

• avyayībhāva: sūpaprati and unmattagaṅgam. Read by SK 663, like śākaprati "a bit of vegetable", sūpaprati is "a bit of soup". There the prior member is sūpa "soup", and the "bit" is the later member prati. unmattagaṅgam names a country (SK 675), which is the meaning of neither member.
• tatpuruṣa: atimāla and the like. By the analysis in SK 780, atimālaḥ is "one who has surpassed a garland". The later member is mālā "garland", and the one meant is not the garland.
• bahuvrīhi: dvitrāḥ and the like. It means "two or three" (dvau vā trayo vā). Here what leads is the "or" joining its members (the commentators on SK 844), not an outside word.
• dvandva: dantoṣṭham and the like. As a group-compound it names "teeth and lips" as one thing: singular by SK 906, neuter by SK 821. It does not name the two members each in its own right.

Bhaṭṭoji only lists these forms. The reasons given here are read from the earlier rules and their commentators.

So the four-way meaning test, like the four-way count, is a guide and not a definition.

7. Karmadhāraya and dvigu are kinds of tatpuruṣa

The karmadhāraya is a particular kind of tatpuruṣa. The dvigu is a particular kind of karmadhāraya.

Both facts match chapter 18. SK 745 gives the name karmadhāraya to a tatpuruṣa whose members are co-referential, as in nīlam utpalam → nīlotpalam. The dvigu (SK 730) is taught inside the co-referential family headed by samānādhikaraṇena (SK 727), with a numeral first: pañcānāṃ gavāṃ samāhāraḥ → pañcagavam. So:

dvigu ⊂ karmadhāraya ⊂ tatpuruṣa.

The karmadhāraya and the dvigu thus fall inside the tatpuruṣa.

8. Which compounds may have many members

Having many members (anekapadatva) belongs only to the dvandva and the bahuvrīhi. The bahuvrīhi rule says aneka, "more than one"; the dvandva rule carries it forward (from A 2.2.24):

• dvandva (SK 901): hotṛpotṛneṣṭodgātāraḥ "the hotṛ, the potṛ, the neṣṭṛ and the udgātṛ";
• bahuvrīhi (SK 830): citrājaradguḥ "whose cow is dappled and old".

For the tatpuruṣa, Bhaṭṭoji says, it has been stated that many members occur only in some cases. He does not repeat the details here.

9. The verse: six kinds of compound

Bhaṭṭoji now adds (kiṃ ca, "moreover") a verse. It sorts compounds by the kind of word that joins with the kind of word:

> Of sup with sup, with tiṅ, with a nāman, and with a dhātu; then of tiṅ with tiṅ, and of tiṅ with a subanta — thus the wise know compounding to be of six kinds.

The four kinds of word are:

• sup — here, a word ending in a case-ending, a subanta;
• tiṅ — a word ending in a personal verb-ending, a finite verb (tiṅanta);
• nāman — a nominal stem, not yet carrying a case-ending;
• dhātu — a verbal root.

Bhaṭṭoji gives an example of each kind. Here is what joins what in each.

(i) sup with sup: rājapuruṣaḥ.

rājan as (a subanta, genitive) + puruṣa su (a subanta, nominative) → SK 702 → rājapuruṣaḥ.

This is the ordinary case, the one worked in point 3.

(ii) sup with tiṅ: paryabhūṣayat.

pari (an indeclinable, a subanta) + abhūṣayat (a finite verb) → saha, read alone (SK 649) → paryabhūṣayat.

The later member is a finite verb. Chapter 17 noted that such compounds occur only in the Veda.

(iii) sup with nāman: kumbhakāraḥ.

kumbha (a subanta, the object) + kāra (a nominal stem made by the kṛt affix aṇ) → upapada compound (SK 782) → kumbhakāraḥ "pot-maker".

In chapter 18 you saw that an upapada compounds with the kṛt-word before any case-ending arises on it. So the later member is a bare stem, a nāman, not a subanta.

(iv) sup with dhātu: kaṭaprūḥ; ajasram.

kaṭa (a subanta) + prū (a root) → kaṭaprū → with the nominative su → kaṭaprūḥ.

The later member is the root itself. Bhaṭṭoji adds ajasram as a second instance of the same kind, without analysis.

(v) tiṅ with tiṅ: pibatakhādatā; khādatamodatā.

pibata "drink!" (a finite verb) + khādata "eat!" (a finite verb) → pibatakhādatā.

khādata + modata (both finite verbs) → khādatamodatā.

Neither member is a subanta. In chapter 18 you met khādatamodatā, with aśnītapibatā and pacatabhṛjjatā. They came from a rule attached to the mayūravyaṃsakādi list, ākhyātam ākhyātena kriyāsātatye. It says: a finite verb compounds with a finite verb when an action goes on continuously (kriyāsātatya). pibatakhādatā is of the same type. Bhaṭṭoji's commentators explain the final ā: the compound names the action, which is feminine, so the feminine affix ṭāp is added.

(vi) tiṅ with sup: kṛntavicakṣaṇā.

kṛnta "cut!" (a finite verb) + vicakṣaṇa "O clever one!" (a subanta, the vocative) → kṛntavicakṣaṇā.

Bhaṭṭoji gives the meaning: the action in which one says kṛnta vicakṣaṇa "cut, O clever one!" is kṛntavicakṣaṇā. Here the first member is the finite verb, and the second is the noun.

10. Why kṛntavicakṣaṇā is allowed

Bhaṭṭoji ends by naming his authority for kṛntavicakṣaṇā. The compound holds because the rule ehīḍādayo 'nyapadārthe ca — "and ehīḍa and the rest, in the sense of another word's meaning" — is read in the list headed by mayūravyaṃsaka (SK 754).

You met this rule in chapter 18. It gives compounds such as ehīḍam, the action in which one says ehi īḍa "come, praise!", and uddharotsṛjā. Each names an action outside its members, just as kṛntavicakṣaṇā names the action in which one says "cut, O clever one!". And SK 754 lays its list down ready-made (nipātana), irregularities included. So a compound whose first member is a finite verb, which saha supā would not give, is still correct.

Terms used in this chapter

• vṛtti — a "complex formation": a kṛt or taddhita derivative, a compound, an ekaśeṣa, or a root in san and the like; defined as "the expression of a meaning that belongs to another" (parārtha-abhidhāna).
• kṛt / taddhita — an affix added to a root / an affix added to a noun, each making a new noun stem (SK 179).
• aṇ — a kṛt affix; on kṛ it gives kāra "maker", as in kumbhakāraḥ (SK 782).
• yañ — a taddhita affix for a family descendant; on garga it gives gārgya (chapter 21).
• samāsa — a compound: words fused into one stem.
• ekaśeṣa — "one remaining": one word left standing for several (SK 188; chapter 21).
• sanādyanta-dhātu — a root ending in the affix san or one of the affixes listed with it; such a form is itself a root (A 3.1.32).
• prātipadika — a noun stem, to which case-endings are added; a compound is one (SK 179).
• vigraha — the sentence that makes known the meaning of a vṛtti; the analysis of a compound.
• vākya — a sentence.
• laukika / alaukika — "of ordinary usage" / "not of ordinary usage": the two kinds of vigraha.
• sādhu / asādhu — correct / incorrect.
• pariniṣṭhita — fully formed: with every operation complete.
• prayoga — actual use in speech.
• sup / subanta — a case-ending / a word ending in one.
• tiṅ / tiṅanta — a personal verb-ending / a finite verb.
• nāman — in the verse, a nominal stem not yet carrying a case-ending.
• dhātu — a verbal root.
• luk — the disappearance of an affix; here of the case-endings inside a compound (SK 650).
• upasarjana — the member named in the nominative in a compound rule; it goes first (SK 653–654).
• nitya-samāsa — an obligatory compound: one with no vigraha, or one whose vigraha is not made with its own words.
• avigraha — having no analysis at all.
• prādi — pra and the other words of the list headed by pra; a prādi compound has one of them first (SK 761, SK 780), as atimālaḥ.
• asvapada-vigraha — an analysis not made with the compound's own words, as krānto mālām for atimālaḥ.
• adhikāra — a heading rule whose wording is read into the rules that follow.
• prāyovāda / prāyo 'bhiprāyaḥ — a statement of what generally holds / meant as holding generally: not a strict definition.
• avyayībhāva, tatpuruṣa, bahuvrīhi, dvandva — the indeclinable, determinative, outside-referent and "and"-compounds of chapters 17–20.
• prāc / pravāda — the older grammarians / a traditional saying.
• pradhāna — predominant: the member whose meaning leads.
• pūrvapada / uttarapada / anyapada / ubhayapada — the prior member / the later member / another word outside the compound / both members.
• samāhāra — the group sense of a dvandva, singular neuter (SK 901).
• karmadhāraya — the tatpuruṣa whose members are co-referential (SK 745).
• samānādhikaraṇa — co-referential: naming one and the same thing.
• dvigu — the co-referential compound with a numeral first (SK 730); a kind of karmadhāraya.
• aneka / anekapadatva — more than one / having many members; found only in the dvandva and bahuvrīhi, and in the tatpuruṣa only in some cases.
• upapada — a word named in the locative in a kṛt rule; it compounds with the kṛt-word before a case-ending arises (SK 781–782).
• kriyāsātatya — continuity of an action; the condition for a finite verb compounding with a finite verb.
• ṭāp — a feminine affix, the sound ā with markers.
• mayūravyaṃsakādi — the list headed by mayūravyaṃsaka, laid down ready-made by SK 754, with rules attached to it.
• nipātana — laying down a form ready-made, irregularities included.
• ehīḍādayo 'nyapadārthe — the rule attached to that list by which ehīḍa and the like compound in the sense of an outside thing, here an action.

Summary