After this chapter you will be able to read and form the ekaśeṣa, "one remaining": the device by which, when several words are meant together, one of them is kept and the rest drop, the survivor taking the number of the whole. Chapter 8 gave the basic case, words of the same form (vṛkṣaś ca vṛkṣaś ca → vṛkṣau, SK 188). This chapter gives the nine rules by which Pāṇini allows one word to stand for others of different form, and names in each case which word survives: the family-name word beside its junior derivative (gārgyaś ca gārgyāyaṇaś ca → gārgyau), the masculine beside the feminine (haṃsī ca haṃsaś ca → haṃsau), the neuter beside the non-neuter (śuklaś ca śuklā ca śuklaṃ ca → śuklam), the named kin-word in four family pairs (bhrātā ca svasā ca → bhrātarau, mātā ca pitā ca → pitarau), the pronoun beside anything at all (sa ca devadattaś ca → tau), and, for herds, the feminine (gāva imāḥ). You will also learn the one proviso that runs through the whole chapter — the rule works only when the entire difference between the words is the one the rule names — and why, once an ekaśeṣa has been made, no dvandva can follow.
mātāpitarau (SK 921). The ekaśeṣa is the dvandva's rival: both answer to "X and Y", and this chapter closes by saying which of the two wins.gārgyaś ca gārgyāyaṇaś ca → gārgyauThis group has two rules. The first says that when a family-name word (a gotra derivative) is meant together with its junior derivative (a yuvan derivative), the family-name word alone is kept: gārgyaś ca gārgyāyaṇaś ca → gārgyau, "Gārgya and Gārgyāyaṇa (his junior)". The first rule also states the proviso that governs the whole chapter: the two words must differ by nothing but the thing the rule names. The second rule extends the first to a feminine family-name word, and adds that the survivor is then treated as masculine.
Rule. When a vṛddha word is expressed together with a yuvan word, the vṛddha word alone is retained — provided the entire difference in form between the two is made by the gotra and yuvan affixes alone. A vṛddha is a gotra derivative: a word ending in the gotra affix, the affix that names a descendant from the grandson onward (apatyaṃ pautraprabhṛti gotram, A 4.1.162). Vṛddha is the older teachers' word for such a descendant. A yuvan word ends in the yuvan affix of A 4.1.163, which marks a younger member of the line while an elder is alive. "Expressed together" (sahokti) is the situation in which two words are meant jointly, as a sentence would join them with ca. It is not in the sūtra but is supplied by the reader (adhyāhāra). "Is retained" (śiṣyate) is the word śeṣa of SK 188, carried forward by anuvṛtti. Tallakṣaṇa means "having those (the two affixes) as its cause"; viśeṣa is the difference in form; and the word eva "only" is what gives "entire".
Example.
gārgya "descendant of Garga" → garga + the gotra affix yañ (A 4.1.105) → a vṛddha word.
gārgyāyaṇa "junior descendant of Gārgya" → gārgya + the yuvan affix phak (A 4.1.101) → a yuvan word.
gārgyaś ca gārgyāyaṇaś ca "Gārgya and Gārgyāyaṇa" → the two differ only by yañ against yañ + phak → the whole difference is affix-made → the vṛddha word gārgya alone is kept → it takes the number of both → gārgya in the nominative dual → gārgyau.
Applies when / blocked by. Each word of the sūtra bars a case, and Bhaṭṭoji gives all four.
gargagārgyāyaṇau "Garga and Gārgyāyaṇa": garga has no gotra affix, so it is not a vṛddha word. No ekaśeṣa; the two form a dvandva.gargagārgyau "Garga and Gārgya": there is no yuvan word here. No ekaśeṣa.bhāgavittibhāgavittikau. Bhāgavitti is the gotra descendant of Bhagavitta, by the affix iñ (A 4.1.95). Bhāgavittika is a junior descendant of Bhāgavitti among the Sauvīras, spoken of with contempt, by the affix ṭhak (A 4.1.148). The difference is made not by the two affixes alone but also by the contempt and the Sauvīra-hood. No ekaśeṣa.gārgyavātsyāyanau "Gārgya and Vātsyāyana". Vātsyāyana is the yuvan derivative of vātsya, the gotra descendant of Vatsa. Gotra and yuvan affixes are present, but the bases garga and vatsa also differ, so the difference is not entirely affix-made. No ekaśeṣa. (Scholars discuss the exact reading of this example.)Watch out. Bhaṭṭoji's commentators ask why Pāṇini wrote vṛddha and not simply gotra. The answer: outside the section on descendants (the apatya-adhikāra), gotra means the everyday family name, not the technical one. Writing vṛddha keeps the technical sense. One consequence they draw: aupagava, an immediate son of Upagu, and aupagavi, his junior, get no ekaśeṣa, because the son's name is not a gotra derivative in the technical sense.
Rule. When a feminine vṛddha word is expressed together with a yuvan word, the feminine vṛddha word alone is retained, and its meaning is treated as masculine (puṃvat, "like the masculine"). The word vṛddha is carried forward from SK 931 and made feminine. This rule is needed because femininity is an extra difference between the two words. Without it, the proviso of SK 931 would bar the ekaśeṣa.
Example.
gārgī "a woman descendant of Garga" → garga + yañ → then the feminine affix ṅīp (A 4.1.16) → a feminine vṛddha word.
gārgī ca gārgyāyaṇau ca "Gārgī and two Gārgyāyaṇas" → the feminine vṛddha word survives → three persons, so plural → her meaning is now masculine → the gotra affix yañ is dropped by luk (SK 1108, A 2.4.64), which deletes the gotra affix in the plural when the word is not feminine → garga in the nominative plural → gargāḥ.
dākṣī "a woman descendant of Dakṣa" → dakṣa + iñ (A 4.1.95) → then ṅīṣ (A 4.1.65).
dākṣī ca dākṣāyaṇaś ca "Dākṣī and Dākṣāyaṇa" → dākṣī survives → treated as masculine, so the feminine affix ṅīṣ is withdrawn → the masculine stem dākṣi in the dual → dākṣī. The final form looks like the feminine singular but is the masculine dual.
Applies when / blocked by. Luk is the name for the disappearance of an affix, here of yañ. The rule that deletes it, SK 1108, carries in three conditions from A 2.4.62: "in the plural", "when not feminine", and "by that affix alone". It is the treatment as masculine that meets "not feminine"; that is the fruit of puṃvat. Bhaṭṭoji's commentators explain why the example says gārgyāyaṇau (two juniors) and not gārgyāyaṇaḥ (one). With only two persons there would be no plural, hence no luk, and the dual gārgyau looks the same in either gender; puṃvat would show no effect. They also note that puṃvat extends the meaning, not the form. Its fruit is the withdrawal of femininity, seen again in the accusative gargān paśya, whose final n is given by a rule that applies only to the masculine (A 6.1.103). The proviso of SK 931 still binds: bhāgavittī ca bhāgavittikaś ca and gārgī ca vātsyāyanau ca receive no ekaśeṣa.
haṃsī ca haṃsaś ca → haṃsauOne rule. When a masculine and a feminine word are meant together, the masculine alone is kept: haṃsī ca haṃsaś ca → haṃsau, "a goose and a gander". The pair must differ by gender and by nothing else. This is the rule that SK 939, at the end of the chapter, overrides for herds of animals.
Rule. When a masculine word is expressed together with a feminine word, the masculine alone is retained — provided the difference between them is due to gender alone. The proviso tallakṣaṇa eva is carried forward from SK 931 by anuvṛtti.
Example.
haṃsī "goose" → haṃsa + the feminine affix ṅīṣ.
haṃsī ca haṃsaś ca "a goose and a gander" → the two differ only by the feminine affix → the masculine haṃsa alone is kept → dual → haṃsau.
Applies when / blocked by. SK 188, the rule for words of like form, cannot reach this pair, precisely because the gender-made difference makes their forms unlike. That is why a separate rule is needed. Bhaṭṭoji's commentators give two pairs that fall outside it. haṃsaś ca varaṭā ca "a gander and a female goose (varaṭā)": the difference is one of gender only, but the words are of different form, and sarūpāṇām "of like form" is still carried forward from SK 188. No ekaśeṣa. mātṛ "mother" (feminine) and mātṛ "measurer": the pair differs by gender, and also in form, by the lengthening that A 6.4.11 gives to one and not the other (mātārau against mātarau). No ekaśeṣa by this rule.
Watch out. The wording "together with a yuvan" of SK 931 stops here; the proviso, the word śeṣa, and sarūpāṇām continue. Bhaṭṭoji's commentators read the fresh start of the next rule, bhrātṛputrau, as the sign that sarūpāṇām is still in force.
śuklaś ca śuklā ca śuklaṃ ca → śuklamOne rule. When a neuter word is meant together with a masculine or feminine word, the neuter alone is kept, and the survivor may be put in the singular: śuklaś ca śuklā ca śuklaṃ ca → śuklam, "a white (cloth), a white (wrap) and a white (garment)". Bhaṭṭoji shows it on the sentence tad idaṃ śuklam, which hides three ekaśeṣas at once.
Rule. When a neuter word is expressed together with a non-neuter word (masculine or feminine), the neuter alone is retained; and that retained neuter is optionally treated as singular (ekavat, "as one"). The proviso holds again: the difference must be made by neuterness against non-neuterness alone.
Example.
śuklaḥ paṭaḥ, śuklā śāṭī, śuklaṃ vastram "a white cloth, a white wrap, a white garment" → the adjective śukla takes the gender of the noun it qualifies, so it is masculine, feminine and neuter in turn → the nouns are shown only to bring this out; the words to be joined are śuklaḥ, śuklā, śuklam.
śuklaś ca śuklā ca śuklaṃ ca → the three differ by gender alone → the neuter śukla alone is kept → treated as singular → śuklam.
The pronouns of the sentence go the same way: sa ca sā ca tac ca → tat; ayaṃ ca iyaṃ ca idaṃ ca → idam. Hence tad idaṃ śuklam, "this and that (are) white".
Without the singular treatment → the neuter survivor takes the number of the whole → tāni imāni śuklāni.
Applies when / blocked by. Bhaṭṭoji's commentators settle two questions of construction. First, the option-word anyatarasyām goes with ekavat only, by proximity; the ekaśeṣa itself is not optional, only the singular. Bhaṭṭoji's own wording, "and that optionally as singular", follows this. Second, the word asya "of this" confines the singular treatment to this ekaśeṣa alone, so that no one carries it forward into the rules that follow. Why "with a non-neuter"? śuklaṃ ca śuklaṃ ca → śukle: two neuters of like form fall under SK 188, and no singular treatment is available there.
Watch out. The ekaśeṣa never gives a group-singular in the way a samāhāra dvandva does. Bhaṭṭoji's commentators say that ekavat is what shows this: if the section could work in the group sense, the singular would already be available and ekavat would be idle. So odanaś ca odanaṃ ca (a word used in both masculine and neuter) does not yield odanam as a group.
bhrātā ca svasā ca → bhrātarau; mātā ca pitā ca → pitarau / mātāpitarauThree rules, for four pairs of relatives. Brother and sister, son and daughter, father and mother, father-in-law and mother-in-law differ from each other in their whole shape, not by a gender affix. So SK 933 cannot join them, and Pāṇini says outright which word survives. For the first two pairs the ekaśeṣa is obligatory; for the last two it is optional, the dvandva of chapter 20 being the alternative. SK 934 stands between SK 933 and SK 935 in Bhaṭṭoji's order; it is taken here with its fellow kinship rules.
Rule. Bhrātṛ "brother" alone is retained when expressed together with svasṛ "sister", and putra "son" alone when expressed together with duhitṛ "daughter" — the two pairs taken respectively.
Example.
bhrātā ca svasā ca "a brother and a sister" → bhrātṛ alone is kept → dual → bhrātarau.
putraś ca duhitā ca "a son and a daughter" → putra alone is kept → dual → putrau.
Applies when / blocked by. Bhaṭṭoji's commentators say why the rule is needed: in each pair the two words differ in their very form, not by gender alone, so SK 933 would not apply. There is no option-word; the ekaśeṣa is obligatory.
Rule. Pitṛ "father" alone is optionally retained when expressed together with mātṛ "mother". The option is not in the sūtra's own words. Bhaṭṭoji marks it with vā; the option is usually taken as the anyatarasyām of SK 935 carried down.
Example.
mātā ca pitā ca "mother and father" → pitṛ alone is kept → dual → pitarau.
In the alternative → no ekaśeṣa → the dvandva of chapter 20 → mātṛ is placed first → its final ṛ becomes ā by ānaṅ (SK 921) → mātāpitarau.
Applies when / blocked by. Bhaṭṭoji's commentators say that mātṛ stands first in the dvandva because the mother is the more venerable, citing the smṛti verse that a mother exceeds a father tenfold in honour. They also meet an objection: are this rule and the next idle, since pitṛ alone could mean "parents" by extension, and the rules cannot be defended as blocking the dvandva when the dvandva is wanted in the alternative? Their answer: the two rules must be stated in order to bar mātarau alone (and śvaśrvau alone, SK 937) in the joint sense. To reject them as unused would be improper.
Rule. Śvaśura "father-in-law" alone is optionally retained when expressed together with śvaśrū "mother-in-law" — provided the difference is due to that alone.
Example.
śvaśrūś ca śvaśuraś ca "mother-in-law and father-in-law" → śvaśura alone is kept → dual → śvaśurau.
In the alternative → the dvandva → śvaśrū placed first → śvaśrūśvaśurau.
Applies when / blocked by. Bhaṭṭoji's commentators say śvaśrū stands first in the dvandva because the mother-in-law is venerable, citing the smṛti that ranks her with a mother. The defence of this rule against the charge of redundancy is the one given under SK 936.
sa ca devadattaś ca → tauOne rule and three vārttikas. A pronoun of the tyad-list, meant together with any word whatever, always survives: sa ca devadattaś ca → tau, "he and Devadatta". The vārttikas answer the questions the rule leaves open. When two such pronouns meet, which survives? What gender does the surviving pronoun show? And what happens when the pronoun is only the qualifier of a compound?
Rule. The tyad-ādi words — the pronouns read from tyad onward in the sarva-list — are always (nityam) retained when expressed together with any words whatever (sarvaiḥ). "Any words" means both other tyad-ādi words and ordinary nouns.
Example.
sa ca devadattaś ca "he and Devadatta" → tad is a tyad-ādi word → tad alone is kept, devadatta drops → dual → tau. The dvandva taddevadattau does not occur.
Applies when / blocked by. Bhaṭṭoji's commentators explain the word sarvaiḥ: it is stated so that, by proximity to the previous rules, the rule is not narrowed to "only together with other pronouns". The three vārttikas follow.
Vārttika 1: tyadādīnāṃ mithaḥ sahoktau yat paraṃ tac chiṣyate. When tyad-ādi words are expressed together with one another, the one that is later in the list is retained. This rests on the maxim that in a conflict the later word prevails.
sa ca yaś ca "he and who" → yad is read after tad in the list → yad alone is kept → yau. Likewise sā ca yā ca → ye.
The Bhāṣya adds that the earlier one is also sometimes retained: sa ca yaś ca → tau. Bhaṭṭoji's commentators explain that para "later" can also mean "the desired one". They also note that the list is not cut off at dvi here; the Bhāṣya's own example is ahaṃ ca bhavāṃś ca → āvām, "I and you, sir → we two".
Vārttika 2: tyadāditaḥ śeṣe puṃnapuṃsakato liṅgavacanāni. When a tyad-ādi word is retained, it expresses gender according to the masculine or the neuter word that was expressed with it. (The suffix -tas in tyadāditaḥ has the sense of the genitive; liṅgavacana means "gender-expressing", not "gender and number".)
sā ca devadattaś ca "she and Devadatta" → tad is kept → it takes the masculine of devadatta → tau.
tac ca devadattaś ca yajñadattā ca "that (thing), Devadatta and Yajñadattā" → tad is kept → it takes the neuter, the neuter being named in the vārttika → plural → tāni.
When a masculine and a neuter are expressed together, the neuter wins, because in the vārttika's puṃnapuṃsaka the neuter is read later than the masculine: tac ca devadattaś ca → te.
Vārttika 3: advandvatatpuruṣaviśeṣaṇānām iti vaktavyam. It should be stated that vārttika 2 does not hold for a tyad-ādi word that qualifies a dvandva or a tatpuruṣa. Such a pronoun simply follows the gender of the noun it qualifies.
kukkuṭamayūryāv ime "this cock and this peahen" → the pronoun resolves as ayaṃ ca iyaṃ ca → SK 933 would keep the masculine, and vārttika 2 would make it masculine → but the dvandva kukkuṭamayūrī is feminine, taking the gender of its last member (paravalliṅga, A 2.4.26) → its qualifier follows it → feminine ime.
tac ca sā ca → ardhapippalyau te "the half-pepper (neuter) and the half-pepper (feminine) → these two" → tat stands for pippalyardham, a neuter tatpuruṣa; sā for ardhapippalī, a tatpuruṣa by A 2.2.2 and feminine by its last member → SK 935 would keep the neuter → but the tatpuruṣa is feminine, and so is its qualifier → feminine te. Since te is the same in feminine and neuter dual, the telling form is the plural, ardhapippalyas tāḥ.
Watch out. Some copies of Bhaṭṭoji's text add mayūrīkukkuṭāv imau "this peahen and this cock". Bhaṭṭoji's commentators set it aside: that dvandva is masculine by its last member, so imau needs no special rule and the example proves nothing.
gauś ca gauś ca … → gāva imāḥOne rule and one vārttika. For a herd of domestic animals that are not young, it is the feminine that survives: cows and bulls together are gāvaḥ, and the pronoun that goes with them is feminine, gāva imāḥ. This is the chapter's one exception to SK 933. Bhaṭṭoji ends the section with a note on the relation of ekaśeṣa to dvandva that applies to every rule of the chapter.
Rule. When words are meant together (sahavivakṣā) with reference to herds (saṃgha) of domestic (grāmya) animals (paśu) that are not young (ataruṇa), the feminine alone is retained. This is an apavāda, an exception, to pumān striyā (SK 933). The words "together with a non-neuter" are carried forward from SK 935, so the feminine survives beside the masculine.
Example.
gauś ca gauś ca gauś ca … "cows and bulls, a herd" → masculine and feminine go-words meant together → not SK 933's masculine but the feminine is kept → plural → gāvaḥ. The surviving feminine and the masculine would look the same, gāvaḥ either way. So the fruit shows in the pronoun that accompanies the noun: gāva imāḥ, feminine, not gāva ime.
Applies when / blocked by. Each word of the sūtra bars a case, and in each the masculine of SK 933 returns.
rurava ime "these black antelopes": the ruru is not a domestic animal, so the masculine survives and the pronoun is masculine.brāhmaṇāḥ, for brāhmaṇī ca brāhmaṇāś ca "a brāhmaṇa woman and brāhmaṇa men": persons, not animals; the masculine survives. Otherwise it would be brāhmaṇya imāḥ.etau gāvau "this cow and this bull": one female and one male together are not a herd. Bhaṭṭoji's commentators add that saṃgha can only mean many, since an ekaśeṣa is in any case of more than one; otherwise the pronoun would be ete. (Scholars discuss whether etau gāvau itself is regular.)vatsā ime "these calves": young, so the masculine survives. Otherwise vatsā imāḥ. (Scholars discuss whether vatsa is the best example.)Why is the pronoun imāḥ feminine at all, when vārttika 2 of SK 938 would make a surviving pronoun follow the masculine? Bhaṭṭoji's commentators answer that that vārttika does not intervene here, on the strength of the Bhāṣya's own usage gāva imāḥ; others hold that imāḥ is itself a feminine ekaśeṣa, since the pronoun too refers to the herd. Scholars discuss this.
Vārttika: anekaśapheṣv iti vācyam. It should be stated that the rule holds only among animals that are not single-hoofed (anekaśapha). Horses are therefore excluded: aśvā ime "these horses", with the masculine surviving and a masculine pronoun.
Watch out. Bhaṭṭoji closes the chapter with a remark that applies to every ekaśeṣa in it. Once the ekaśeṣa has been made there is no dvandva, because there is no longer a plurality of case-ending words (subanta) to compound. Bhaṭṭoji's commentators put the point this way: the ekaśeṣa is antaraṅga, a rule that acts inside the words before any case-ending is generated, and it blocks the dvandva that would otherwise arise; if it did not, it would have no scope of its own. One might object that the ekaśeṣa should apply after the dvandva, since the Bhāṣya says "the ekaśeṣa in a dvandva"; the answer is that this phrase means "where a dvandva would otherwise obtain". Two consequences follow.
śirasī, śirāṃsi "two heads, heads" (śiras + śiras) → an ekaśeṣa, not a dvandva → so they receive neither the final udātta accent that a compound gets by samāsasya (A 6.1.223), nor the singular treatment that a dvandva of parts of a living body gets (prāṇyaṅga, SK 906, A 2.4.2).panthānau, panthānaḥ "two roads, roads" (pathin + pathin) → an ekaśeṣa, not a dvandva → so they receive no closing affix (samāsānta), the a that A 5.4.74 adds to a compound ending in pathin.One might say the closing affix was never a threat, because the rule that shapes panthānau (A 7.1.86) already presupposes its absence. Bhaṭṭoji's commentators answer that A 7.1.86 has its own scope in apanthānau "two non-roads", where A 5.4.72 makes the closing affix optional, so nothing can be inferred from it.
Thus ends the ekaśeṣa-prakaraṇa.
| SK | What it does | Example |
|---|---|---|
| 931 | Gotra word survives beside its yuvan derivative; only if the whole difference is made by the two affixes | gārgyaś ca gārgyāyaṇaś ca → gārgyau |
| 932 | Feminine gotra word survives beside a yuvan word and is treated as masculine; the gotra affix then drops in the plural | gārgī ca gārgyāyaṇau ca → gargāḥ; dākṣī ca dākṣāyaṇaś ca → dākṣī |
| 933 | Masculine survives beside feminine; only if gender is the whole difference | haṃsī ca haṃsaś ca → haṃsau |
| 935 | Neuter survives beside non-neuter; optionally singular | śuklaś ca śuklā ca śuklaṃ ca → śuklam / śuklāni |
| 934 | bhrātṛ survives beside svasṛ, putra beside duhitṛ; obligatory | bhrātā ca svasā ca → bhrātarau; putraś ca duhitā ca → putrau |
| 936 | pitṛ optionally survives beside mātṛ; else the dvandva | mātā ca pitā ca → pitarau / mātāpitarau |
| 937 | śvaśura optionally survives beside śvaśrū; else the dvandva | śvaśrūś ca śvaśuraś ca → śvaśurau / śvaśrūśvaśurau |
| 938 | A tyad-ādi pronoun always survives beside anything; among pronouns the later in the list; it takes the masculine or neuter of what it stood with (neuter if both are present); not when it qualifies a dvandva or tatpuruṣa | sa ca devadattaś ca → tau; sa ca yaś ca → yau; tac ca devadattaś ca → te; kukkuṭamayūryāv ime |
| 939 | In herds of grown domestic animals, not single-hoofed, the feminine survives; exception to SK 933; after any ekaśeṣa there is no dvandva | gāva imāḥ; aśvā ime; śirasī, panthānau |