Book 3 - Pada-kāṇḍa (1): Jāti-samuddeśa (On the Universal)

3.1.1 unformalized

द्विधा कैश्चित्पदं भिन्नं चतुर्धा पञ्चधापि वा ।
अपौद्धृत्यैव वाक्येभ्यः प्रकृतिप्रत्ययादिवत् ॥ १ ॥

dvidhā kaiścitpadaṃ bhinnaṃ caturdhā pañcadhāpi vā |
apauddhṛtyaiva vākyebhyaḥ prakṛtipratyayādivat || 1 ||

The word is divided by some in two ways, in four, or even in five — having been abstracted from sentences, just as stem, suffix and the rest are.

The Padakanda opens by conceding, in its first breath, everything the Vakyakanda argued. The word — the object this entire kanda will anatomize across fourteen samuddesas and twelve hundred verses — is apoddhrtya eva vakyebhyah: *abstracted from sentences*, and the comparison seals it: prakrti-pratyayadivat, *just as stem and suffix are* — which is 2.10 verbatim, where the extraction of words from sentences was said to have exactly the standing of the extraction of stems from words. So the kanda's subject is announced as a construct. The divisions canvassed are the tradition's: dvidha, twofold (noun and verb); caturdha, fourfold (Yaska's nama, akhyata, upasarga, nipata — the Nirukta's opening classification); pancadha, fivefold (adding the karmapravacaniya of 2.199-204). Kaiscit, 'by some': even the count is disputed, as 2.341 noted when it priced the reduction that would abolish the fourfold scheme. The verse is a masterstroke of intellectual honesty at a book's threshold: here is my subject, and it does not exist except as my analysis makes it. What follows is not thereby idle — the scaffold's statics are the science (2.75, 2.88) — but the reader is told at the outset in what spirit to take twelve hundred verses about word-meanings.

3.1.2 unformalized

पदार्थानामपोद्धारे जातिर्वा द्रव्यमेव वा ।
पदार्थौ सर्वशब्दानां नित्यावेवोपवर्णितौ ॥ २ ॥

padārthānāmapoddhāre jātirvā dravyameva vā |
padārthau sarvaśabdānāṃ nityāvevopavarṇitau || 2 ||

In the abstraction of word-meanings: either the universal, or else substance. The two word-meanings of all words are described as eternal indeed.

The kanda's program, stated with the economy of a table of contents. Padarthanam apoddhare — in the *abstraction* of word-meanings (3.1.1's apoddhrtya sustained: what follows concerns extracts) — the candidates are two: jatir va dravyam eva va, either the *universal* or *substance*. Every word, on analysis, means a type or a thing; there is no third. And both are nityau eva upavarnitau — described as *eternal indeed*: whichever a word means, what it means does not come and go (1.23's nitya sambandha requiring an eternal relatum, and 3.1.13's dravya turning out to be eternal in a sense the Vaisesikas would not recognize). The two options name the kanda's first two samuddesas — Jati and Dravya — and the fact that they are canvassed as *alternatives* rather than combined is the tradition's long dispute (Vajapyayana for the universal, Vyadi for the substance, as the Mahabhasya reports). Bhartrhari will not choose in the way either party wanted: 3.1.6 is about to relocate the whole question by asking what a word means *first*, before it means anything about the world at all.
contested: nityau eva ('eternal indeed') is emphatic and its scope is disputed — whether both candidates are eternal, or whether the verse reports that each party claims eternality for its own candidate; we take the first, with 3.1.13's dravya-doctrine as warrant

3.1.3 unformalized

केषाञ्चित्साहचर्येण जातिः शक्त्युपलक्षणम् ।
खदिरादिष्वशक्तेषु शक्तः प्रतिनिधीयते ॥ ३ ॥

keṣāñcitsāhacaryeṇa jātiḥ śaktyupalakṣaṇam |
khadirādiṣvaśakteṣu śaktaḥ pratinidhīyate || 3 ||

For some, by association, the universal is the indicator of capacity; among khadira-woods and the like that are incapable, a capable one is substituted.

The first argument for the universal, and it is drawn — characteristically — from ritual practice rather than from metaphysics. The injunction prescribes a khadira post for the sacrifice. Why khadira? Not for khadira-ness as such: jatih sakty-upalaksanam — the universal is the *indicator of capacity*, sahacaryena, *by association*: khadira-wood is durable, and the universal is named because it reliably accompanies the durability the rite needs. The proof is the substitution-rule: khadiradisu asaktesu — among khadiras that are *incapable* (rotten, unsound) — saktah pratinidhiyate: a *capable* one is *substituted*. If khadira-ness itself were what the injunction wanted, a rotten khadira would satisfy it and a sound non-khadira would not; ritual practice says the reverse. So the universal named is a *mark* for a capacity meant. This is 2.64-66's pratinidhi-doctrine (substitution proving that the general survives its specification) redeployed: here it proves that what a universal-word conveys may be a capacity the universal merely flags.

3.1.4 unformalized

अस्वातन्त्र्यफलो बन्धिः प्रमाणा�ीव शिष्यते ।
अतो जात्यभिधानेऽपि शक्तिहीनं न गृह्यते ॥ ४ ॥

asvātantryaphalo bandhiḥ pramāṇādīva śiṣyate |
ato jātyabhidhāne'pi śaktihīnaṃ na gṛhyate || 4 ||

The binding, whose fruit is non-independence, is taught as (governed) by measure and the rest; therefore, even where the universal is expressed, what lacks capacity is not taken.

The argument of 3.1.3 given its grammatical anchor. The rite's verb is bandhi — *binding* (the victim bound to the post), and its fruit is asvatantrya-phalah: *non-independence* — the binding's whole purpose is that the bound thing not get free. Such a binding is sisyate pramanadi iva — taught as governed by *measure* and the rest: the post's dimensions are prescribed, and dimensions are prescribed because they serve the holding. Hence: atah jaty-abhidhane 'pi sakti-hinam na grhyate — therefore, *even where the universal is expressed*, what *lacks capacity* is *not taken*. The reasoning is tight and it is the samuddesa's opening move: the injunction's own apparatus (the measures, the substitution-rule) shows that its universal-word is in service of a capacity, so universals as word-meanings are, at least here, instrumental rather than terminal. What a word means and what a word is *for* have come apart again (2.113's abhidheya versus prayojana, now inside a single noun).

3.1.5 unformalized

संश्लेषमात्रं बध्नातिर्यदि स्यात्तु विवक्षितः ।
शक्त्याश्रये ततो लिङ्गं प्रमाणाद्यनुशासनम् ॥ ५ ॥

saṃśleṣamātraṃ badhnātiryadi syāttu vivakṣitaḥ |
śaktyāśraye tato liṅgaṃ pramāṇādyanuśāsanam || 5 ||

But if 'binds' were intended as mere contact, then the teaching of measure and the rest would be a mark of a substratum of capacity.

The alternative construal, and it reaches the same destination. Suppose samslesa-matram badhnatih — 'binds' means *mere contact*, nothing about holding fast: the verb merely joins victim to post. Then the prescriptions of measure would not be explained by the binding's own sense — and yet they are prescribed. So: sakty-asraye tato lingam pramanady-anusasanam — the teaching of measure and the rest would then be a *mark* (linga) of a *substratum of capacity*: the dimensions would be evidence that something capacity-bearing is wanted, even though the verb does not say so. Either way, on both construals of the verb, capacity is what the injunction is after and the universal-word is its indicator. Bhartrhari's technique here is worth naming: he does not argue that his opponent's reading is wrong; he shows that both readings deliver his conclusion. The same move closed 1.28 (eternal or produced, no beginning either way) and 2.433 (all-natured or self-less, the word is sovereign either way).

3.1.6 unformalized

स्वा जातिः प्रथमं शब्दैः सर्वैरेवाभिधीयते ।
ततोऽर्थजातिरूपेषु तदध्यारोपकल्पना ॥ ६ ॥

svā jātiḥ prathamaṃ śabdaiḥ sarvairevābhidhīyate |
tato'rthajātirūpeṣu tadadhyāropakalpanā || 6 ||

Its own universal is first expressed by all words indeed; thereafter comes the construction of superimposing that upon the forms of the object-universals.

The samuddesa's central doctrine, and one of the most consequential verses in the work. Sva jatih prathamam sabdaih sarvair eva abhidhiyate: *its own universal* is expressed *first* by *all* words — every word, before it means anything about the world, means *itself*: the word-universal, sabdatva, the type of which its utterances are tokens (1.71's recurrence-datum: the same word across sentences). Only thereafter: tatah artha-jati-rupesu tad-adhyaropa-kalpana — the *construction of superimposing* that upon the forms of the *object*-universals. The word's own type is laid over the object's type, and reference is that superimposition. Every word is thus doubly universal, and the second universal is a projection of the first. The verse gathers the whole work: it is 1.50's double luminosity (the word shows itself and its meaning) given priority-order; it is 2.127-128's abhijalpa (word and meaning fused in cognition) given its genesis; it is 1.59-68's svam rupam (grammar's technical practice of citing word-forms) shown to be not a special usage but the *primary* one, of which ordinary reference is derivative. And adhyaropa — superimposition — is the term the Vedantic tradition would make its own: what we take to be the world's joints are the word's own type projected outward. The Padakanda's fourteen categories are about to be examined, and the first thing said about all of them is that they are projections of sabdatva.
contested: prathamam ('first') can be temporal (the word-universal grasped before the object-universal, in each act of hearing) or logical (prior in the order of explanation, without any felt priority in experience). The kanda's method favours the logical reading, but the doctrine is stronger on the temporal one, and Bhartrhari does not disambiguate

3.1.7-8 unformalized

यथा रक्ते गुणे तत्त्वं कषाये व्यपदिश्यते ।
संयोगिसन्निकर्षाच्च वस्त्रादिष्वपि गृह्यते ॥ ७ ॥
तथा शब्दार्थसंबन्धाच्छब्दे जातिरवस्थिता ।
व्यपदेशेऽर्थजातीनां जातिकार्याय कल्पते ॥ ८ ॥

yathā rakte guṇe tattvaṃ kaṣāye vyapadiśyate |
saṃyogisannikarṣācca vastrādiṣvapi gṛhyate || 7 ||
tathā śabdārthasaṃbandhācchabde jātiravasthitā |
vyapadeśe'rthajātīnāṃ jātikāryāya kalpate || 8 ||

(7) As the reality in the red quality is designated in the dye, and, from the proximity of what is conjoined with it, is grasped also in cloth and the like — (8) so, from the word-meaning connection, the universal standing in the word serves, in the designating of the object-universals, for the effect of the universal.

The superimposition of 3.1.6 given its simile, and the simile is a dyer's. (7): redness — rakte gune tattvam — is really *in* the dye, kasaye vyapadisyate, designated there: the dye is what is red, properly speaking. And samyogi-sannikarsat — from the *proximity of what is conjoined* with it — vastradisu api grhyate: it is *grasped also* in cloth and the like: the cloth is called red, and the redness it is called by is the dye's, transferred by contact. Nobody thinks the cloth manufactured its own redness. (8) applies it: sabdartha-sambandhat sabde jatih avasthita — from the word-meaning connection, the universal *standing in the word* — vyapadese artha-jatinam jati-karyaya kalpate: serves, in the *designating* of object-universals, *for the effect of the universal*. The object-universal is the cloth's redness: real as designated, borrowed by contact, the dye's own. This is the sharpest formulation the samuddesa gives of 3.1.6: what we call the universal *in things* is the word's universal, worn by the thing through the contact that reference is. And the choice of a dyeing-simile rather than a mirroring one is deliberate — dye does not merely appear on cloth; it *takes*, and the cloth is thereafter red.

3.1.9 unformalized

जातिशब्दैकशेषे सा जातीनां जातिरिष्यते ।
शब्दजातय इत्यत्र तज्जातिश्शब्दजातिषु ॥ ९ ॥

jātiśabdaikaśeṣe sā jātīnāṃ jātiriṣyate |
śabdajātaya ityatra tajjātiśśabdajātiṣu || 9 ||

In the single-remainder of universal-words, that is held to be the universal of universals; in 'word-universals', that universal is among the word-universals.

The regress opened, and opened deliberately. If every word expresses its own universal (3.1.6), then the word 'jati' expresses *its* own universal — and when universal-words are gathered by ekasesa (2.108's single-remainder: one form standing for many), sa jatinam jatir isyate: that is held to be the *universal of universals*. And the same again one level up: sabda-jatayah iti atra — in the expression 'word-universals' — taj-jatih sabda-jatisu: that universal is *among the word-universals*. Universals of universals of universals, without evident end. Bhartrhari states the regress without alarm because his doctrine is not embarrassed by it: on the sarva-sakti account (2.433, 2.472) the word can do anything, including take itself as object indefinitely, and each level is a carving of the same plenitude rather than a new entity. It is the *realist* about universals for whom the regress is a bill — each level a further real thing, inhering in the last — and the tradition's third-man problems are precisely there. The verse hands the bill over.

3.1.10 unformalized

या शब्दजातिः शब्देषु शब्देभ्यो भिन्नलक्षणा ।
जातिस्सा शब्दजातित्वमप्यतिक्रम्य वर्तते ॥ १० ॥

yā śabdajātiḥ śabdeṣu śabdebhyo bhinnalakṣaṇā |
jātissā śabdajātitvamapyatikramya vartate || 10 ||

The word-universal in words, whose mark is distinct from the words — that universal operates transcending even word-universal-ness.

The regress's terminus, and it is not a top level but a dissolution. Ya sabda-jatih sabdesu sabdebhyo bhinna-laksana: the word-universal *in* words, whose *mark is distinct from* the words — the type is not the tokens, and its criterion is other than theirs (2.104's capacity worn by the invariant stream). And then: jatih sa sabda-jatitvam api atikramya vartate — that universal *operates transcending even word-universal-ness*: it exceeds the very category it was invoked to fill. The formulation is careful and its direction is unmistakable. What is reached at the end of the ascent is not the highest universal in a hierarchy but something that has left the hierarchy — atikramya, having *stepped beyond*. This is the sabda-tattva of 1.1, met from the other end: the Brahmakanda began with the word-essence and descended; the Padakanda begins with word-meanings and ascends to the same, and finds that it is not a padartha at all. The samuddesa on universals thus reaches, in its tenth verse, the point where universals stop being the right category — which is what one should expect from an author for whom every category is a posting.

3.1.11 unformalized

अर्थजात्यभिधानेऽपि सर्वे जात्यभिधायिनः ।
व्यापारलक्षणा यस्मात्पदार्थाः समवस्थिताः ॥ ११ ॥

arthajātyabhidhāne'pi sarve jātyabhidhāyinaḥ |
vyāpāralakṣaṇā yasmātpadārthāḥ samavasthitāḥ || 11 ||

Even in the expressing of object-universals, all words are universal-expressers — since word-meanings stand characterized by operation.

The doctrine of 3.1.6 stated as a thesis about all words whatever. Artha-jaty-abhidhane 'pi — *even* where a word is expressing an *object*-universal (the ordinary case: 'cow' meaning cowhood) — sarve jaty-abhidhayinah: *all* are universal-expressers. There is no class of words that escapes into pure particularity, because the expressing of the object-universal is itself the superimposition of the word's own (3.1.6-8). And the ground offered: vyapara-laksanah yasmat padarthah samavasthitah — since word-meanings stand *characterized by operation*: what a padartha *is* is what it *does* — vyapara, the functioning, the office (2.104's capacity, 2.83's individuation by operation). And offices are repeatable by nature: to be characterized by an operation is to be the sort of thing that recurs, which is to be a universal. So universality is not a feature some meanings have; it is what having a meaning *is*, once meanings are understood functionally. The argument is elegant and it is the Padakanda's charter for treating every category as it will.

3.1.12 unformalized

जातौ पदार्थे जातिर्वा विशेषो वापि जातिवत् ।
शब्दैरपेक्ष्यते यस्मादतस्ते जातिवाचिनः ॥ १२ ॥

jātau padārthe jātirvā viśeṣo vāpi jātivat |
śabdairapekṣyate yasmādataste jātivācinaḥ || 12 ||

When the universal is the word-meaning, either a universal, or even a particular as though a universal, is what is required by words; therefore they are universal-expressing.

The thesis defended against its obvious counterexample. If words express universals (3.1.11), what of words that plainly pick out particulars — proper names, demonstratives, 'this cow'? The reply: jatau padarthe — when the universal is (taken as) the word-meaning — sabdaih apeksyate jatih visesah va api jativat: what words *require* is either a universal, or *even a particular as though a universal* — jativat, in the manner of a universal. The particular, to be meant, must be meant *as* something: as this-cow, as Devadatta — that is, under a characterization that is repeatable in principle even if instantiated once (the office-structure of 3.1.11). So particulars enter language only in universal dress. Atas te jati-vacinah: therefore words are universal-expressing, without exception. The doctrine anticipates a long argument in Indian philosophy about whether the particular as such is sayable, and Bhartrhari's answer is the one his whole system requires: it is not — not because particulars are unreal, but because saying is a universal-making operation (3.1.6's superimposition), and what will not wear the word's type cannot be said.

3.1.13 unformalized

द्रव्यधर्मा पदार्थे तु द्रव्ये सर्वोऽर्थ उच्यते ।
द्रव्यधर्माश्रयाद् द्रव्यमतः सर्वोऽर्थ इष्यते ॥ १३ ॥

dravyadharmā padārthe tu dravye sarvo'rtha ucyate |
dravyadharmāśrayād dravyamataḥ sarvo'rtha iṣyate || 13 ||

But when substance is the word-meaning, every meaning is called substance; from resting on the properties of substance, everything is held to be substance.

The rival option (3.1.2's dravyam eva va) shown to be equally total. Take the other horn: dravya-dharma padarthe — when substance is the word-meaning — dravye sarvo 'rtha ucyate: *every* meaning is called substance. And the reason: dravya-dharmasrayat — from *resting on* the properties of substance: whatever a word means is treated as bearing properties, standing in relations, being qualified — which is to be a substance, whatever else it may be. Even universals and actions, once spoken of, take adjectives and cases and behave as things. Atah sarvo 'rtha isyate: therefore *everything* is held to be substance. Read together with 3.1.11-12, the pair is the samuddesa's most striking result: on the jati-option, *everything* words mean is a universal; on the dravya-option, *everything* is a substance. Both options are unrestricted, which means neither is a discovery about the world — each is a way of taking whatever is said (1.74's two darsanas; 2.179's licensed double description). The great dispute between Vajapyayana and Vyadi is not a dispute about the facts. It is a choice of ledger.
contested: the verse's first pada (dravya-dharma padarthe tu) is awkward and our construal supplies the option-structure from 3.1.2; some read dravya-dharma as a compound qualifying padartha ('the word-meaning having substance-properties'), which yields the same doctrine by a rougher route

3.1.14 unformalized

अनुवृत्तिधर्मो वा जातिस्स्यातसर्वजातिषु ।
व्यावृत्तिधर्मसामान्यं विशेषे जातिरिष्यते ॥ १४ ॥

anuvṛttidharmo vā jātissyātasarvajātiṣu |
vyāvṛttidharmasāmānyaṃ viśeṣe jātiriṣyate || 14 ||

Or the universal would be the property of recurrence in all universals; the general character of the property of exclusion is held to be the universal in the particular.

The two definitions of the universal, and the verse is doing more than cataloguing. Anuvrtti-dharma va jatih syat sarva-jatisu: the universal would be the property of *recurrence* — anuvrtti, the running-along, the same-again across instances — in all universals: cowhood is what recurs in cows. That is the realist's definition, and it is the tradition's default. Then: vyavrtti-dharma-samanyam visese jatir isyate — the *general character of the property of exclusion* is held to be the universal *in the particular*: vyavrtti is the turning-away, the excluding, the not-being-other — and on this second definition what makes a cow a cow is its exclusion from non-cows. Bhartrhari records both without deciding, and a reader who knows what came next in Indian philosophy will feel the ground move: vyavrtti is *apoha*, and the Buddhist logicians would build their entire semantics on it — the word 'cow' meaning nothing but exclusion-from-non-cows, universals dissolved into negations. That the option is already here, stated neutrally as one of two available definitions, is among the most valuable pieces of information the Vakyapadiya carries about the state of the question in its own time.
contested: whether vyavrtti-dharma-samanya here anticipates the developed apoha-doctrine or merely names a definitional alternative is a genuine question of intellectual history; the karika is neutral, and reading Dignaga backwards into it would be an error we flag rather than commit

3.1.15 unformalized

संयोगिधर्मभेदेन देशे च परिकल्पिते ।
तेषु देशेषु सामान्यमाकाशस्यापि विद्यते ॥ १५ ॥

saṃyogidharmabhedena deśe ca parikalpite |
teṣu deśeṣu sāmānyamākāśasyāpi vidyate || 15 ||

And when place has been constructed by difference of the properties of the things conjoined with it, in those places there exists a general even of ether.

A hard case for the recurrence-definition, and it is chosen with precision. Akasa — ether, space — is by every school's account *one*, partless, all-pervading: it has no instances, so nothing recurs in it, and by 3.1.14's first definition it can have no universal. And yet we speak of 'places' — desa — and places are akasa: 'here' and 'there' are ether, differentiated. How? Samyogi-dharma-bhedena dese parikalpite: place is *constructed* by difference of the properties of the things *conjoined* with it — the jar's place is where the jar is; the difference is the jar's, lent to the space. And then: tesu desesu samanyam akasasya api vidyate — in those places, a *general* even *of ether* exists: once places are so constructed, ether has instances after all, and place-hood recurs among them. The verse is a small demonstration of the samuddesa's method: universals appear wherever differentiation is constructed, even in what is by definition undifferentiable — which suggests that universals track our constructions rather than the world's joints, and that 3.1.6's superimposition-doctrine is the honest account of what has happened.

3.1.16 unformalized

अदेशानां घटादीनां देशास्संबन्धिनो यथा ।
आकाशस्याप्यदेशस्य देशास्संबन्धिनस्तथा ॥ १६ ॥

adeśānāṃ ghaṭādīnāṃ deśāssaṃbandhino yathā |
ākāśasyāpyadeśasya deśāssaṃbandhinastathā || 16 ||

As places, being relata, belong to pots and the like which are not places — so places, being relata, belong even to ether, which is not a place.

The construction of 3.1.15 shown to run in both directions, and the symmetry is the argument. A pot is adesa — *not a place*; it is a thing, not a location. And yet desah sambandhinah: *places* belong to it as *relata* — the pot has a place, is somewhere, occupies. Where did that come from? From relation, sambandha: being conjoined with the space it sits in. Akasasya api adesasya — and ether *likewise* is not a place (it is undifferentiated, 3.1.15) — desah sambandhinah tatha: places belong to it too, and by the same route: from what is conjoined with it. So neither the pot nor the ether is intrinsically a place, and both have places, and the places are wholly a matter of the relation between them. Placehood belongs to neither relatum and appears between them — which is exactly the status 2.46 gave relation itself (inferable from its effect, its form never seen, utterly beyond thinghood). The verse is the Sambandha-samuddesa reaching back into the Jati-samuddesa: what looks like a property of things turns out to be a posting of the relation, and the relation is not a thing.

3.1.17-18 unformalized

भिन्नवस्त्वाश्रया बुद्धिः संयोगिष्वनुवर्तते ।
समवायिषु भेदस्य ग्रहणं विनिवर्तते ॥ १७ ॥
अतः संयोगिदेशानां गौणत्वं परिकल्प्यते ।
अविवेकात्प्रदेशेभ्यो मुख्यत्वं समवायिनाम् ॥ १८ ॥

bhinnavastvāśrayā buddhiḥ saṃyogiṣvanuvartate |
samavāyiṣu bhedasya grahaṇaṃ vinivartate || 17 ||
ataḥ saṃyogideśānāṃ gauṇatvaṃ parikalpyate |
avivekātpradeśebhyo mukhyatvaṃ samavāyinām || 18 ||

(17) A cognition resting on distinct things runs along among things conjoined; among things inherent, the grasping of difference withdraws. (18) Therefore the places of the conjoined are constructed as secondary; from the non-discrimination of the sites, primariness belongs to the inherent.

Two ways of being related, and the criterion is what the cognition does. (17): among samyogi — things *conjoined*, in mere contact (the pot and its space) — bhinna-vastv-asraya buddhih anuvartate: a cognition resting on *distinct things* runs along: one keeps seeing two, the contact never fusing them. But among samavayi — things *inherent*, bound by samavaya (the Vaisesika's intimate relation: quality in substance, universal in particular) — bhedasya grahanam vinivartate: the grasping of difference *withdraws*: one does not see the redness *and* the cloth; one sees red cloth. (18) draws the consequence for place: samyogi-desanam gaunatvam parikalpyate — the places of the *conjoined* are constructed as *secondary*; and mukhyatvam samavayinam — *primariness* belongs to the *inherent*, avivekat pradesebhyah, from the *non-discrimination* of the sites. Where difference is not grasped, place is primary; where it is, place is borrowed. Note what has just happened: the primary/secondary distinction — which 2.253 reduced to established currency — is here made to track a *phenomenological* fact about whether a cognition discriminates. The two treatments are compatible: currency is sedimented discrimination.

3.1.19 unformalized

अनुप्रवृत्तिरूपां यां प्रख्यातामाकृतिं विदुः ।
केचिद्व्यावृत्तिरूपां तु द्रव्यत्वेन प्रचक्षते ॥ १९ ॥

anupravṛttirūpāṃ yāṃ prakhyātāmākṛtiṃ viduḥ |
kecidvyāvṛttirūpāṃ tu dravyatvena pracakṣate || 19 ||

That well-known configuration which they know as having the form of recurrence — some, however, declare it, in the form of exclusion, to be substance-hood.

The two definitions of 3.1.14 given their partisans, and the second is startling. Akrti — the *configuration*, the shape-universal, a term the tradition has used for the universal since Panini's own akrtyadhikarana debates — is known prakhyatam anupravrtti-rupam: as having the form of *recurrence* (3.1.14's anuvrtti: the same-again across instances). That is the standard view. Kecid vyavrtti-rupam tu dravyatvena pracaksate: but *some* declare it, in the form of *exclusion*, to be *substance-hood*. Two moves at once. The universal is redefined as exclusion (3.1.14's vyavrtti, the apoha-shaped option) — and then identified with dravyatva, substance-hood: what makes a thing a substance is its exclusion from what it is not. So the two candidates of 3.1.2, jati and dravya, collapse into each other under the exclusion-definition: universals *are* substances, both being nothing but not-being-other. The verse is a compressed report of a position that would have large consequences, and Bhartrhari gives it without comment — which, given 3.1.13's finding that both options are unrestricted, is consistent: if everything is a universal and everything is a substance, their identification costs nothing.

3.1.20-21 unformalized

भिन्ना इति परोपाधिरभिन्ना इति वा पुनः ।
भावत्मसु प्रपञ्चोऽयं संसृष्टेष्वेव जायते ॥ २० ॥
नैकत्वं नापि नानात्वं न सत्त्वं न च नास्तिता ।
आत्मतत्त्वेषु भावानामसंसृष्टेषु विद्यते ॥ २१ ॥

bhinnā iti paropādhirabhinnā iti vā punaḥ |
bhāvatmasu prapañco'yaṃ saṃsṛṣṭeṣveva jāyate || 20 ||
naikatvaṃ nāpi nānātvaṃ na sattvaṃ na ca nāstitā |
ātmatattveṣu bhāvānāmasaṃsṛṣṭeṣu vidyate || 21 ||

(20) 'They are different' — or again 'they are not different' — this is an adjunct imposed by another; this elaboration arises only among the selves of things when they are commingled. (21) Neither oneness, nor manifoldness, nor existence, nor non-existence, is found in the self-realities of things when they are uncommingled.

Two of the most severe verses in the work, and they dissolve the samuddesa's own subject. (20): bhinnah iti — 'they are different'; abhinnah iti va punah — 'or again, they are not different': *both* are para-upadhih, an *adjunct imposed by another* — neither difference nor identity belongs to things; both are laid on from outside. And the location: ayam prapancah samsrstesu eva bhavatmasu jayate — this *elaboration* arises *only* among the selves of things *when they are commingled*: difference and identity appear where things have been brought into relation — by us, in cognition, in speech (3.1.16's places, appearing between relata) — and nowhere else. Then (21) the full list of what goes: na ekatvam na api nanatvam na sattvam na ca nastita — *neither oneness, nor manifoldness, nor existence, nor non-existence* — atma-tattvesu bhavanam asamsrstesu vidyate: is found in the *self-realities* of things *uncommingled*. Unrelated, a thing is not one, not many, not existent, not non-existent. The four negations are the catuskoti's own four corners, and their appearance here — in a grammarian's treatise on universals, applied to the uncommingled thing-in-itself — is not accidental. Bhartrhari's point is his own, though: what the four corners fail to apply to is what has not been *related*, and relating is what cognition and speech do. The categories are artifacts of commingling; strip the commingling and no category applies, including existence.
contested: whether 21's asamsrstesu atma-tattvesu names a real unrelated substratum (things as they are apart from our relating) or is a reductio (there is no such condition to describe) is the deepest question the passage raises, and the karika does not settle it. The Madhyamaka resonance is unmistakable and its direction of influence is unrecoverable

3.1.22-24 unformalized

शर्वशक्त्यात्मभूतत्वमेकस्यैवेति निर्णयः ।
भावानामात्मभेदस्य कल्पना स्यादनर्थिका ॥ २२ ॥
तस्माद् द्रव्यादयः सर्वाः शक्तयो भिन्नलक्षणाः ।
संसृष्टाः पुरुषार्थस्य साधिका न तु केवलाः ॥ २३ ॥
यथैव चेन्द्रियादीनामात्मभूता समग्रता ।
तथा संबन्धिसंबन्धसंसर्गेऽपि प्रतीयते ॥ २४ ॥

śarvaśaktyātmabhūtatvamekasyaiveti nirṇayaḥ |
bhāvānāmātmabhedasya kalpanā syādanarthikā || 22 ||
tasmād dravyādayaḥ sarvāḥ śaktayo bhinnalakṣaṇāḥ |
saṃsṛṣṭāḥ puruṣārthasya sādhikā na tu kevalāḥ || 23 ||
yathaiva cendriyādīnāmātmabhūtā samagratā |
tathā saṃbandhisaṃbandhasaṃsarge'pi pratīyate || 24 ||

(22) The conclusion is that the being-all-powers is of one alone; the construction of a difference of selves among things would be purposeless. (23) Therefore all the powers — substance and the rest — of distinct characters, being commingled, are what accomplish the human end, and not when taken alone. (24) And just as the wholeness of the sense-faculties and the rest is their very self, so it is understood also in the commingling of relatum and relation.

The positive doctrine that 3.1.20-21 cleared ground for, and it is the Brahmakanda's metaphysics arriving in the Padakanda. (22): sarva-sakty-atma-bhutatvam ekasya eva iti nirnayah — the *conclusion* is that being-all-powers belongs to *one alone* (1.2's Brahman resting on its differentiated powers; 2.433's sarva-saktitva); bhavanam atma-bhedasya kalpana syad anarthika: the construction of a *difference of selves* among things would be *purposeless* — not false merely, but idle, doing no work. (23): tasmat — therefore — dravyadayah sarvah saktayah bhinna-laksanah: all of them, *substance and the rest* — which is to say the very categories this kanda's fourteen samuddesas will examine — are *powers*, of distinct characters; and samsrstah purusarthasya sadhikah na tu kevalah: *commingled*, they accomplish the human end, and *not when taken alone*. The categories are powers of the one, functional only in combination. (24) gives the model: as the samagrata, the *wholeness*, of the sense-faculties is atma-bhuta — their very *self*, not something added to them (2.419's eyes, which are nothing outside the body) — so with sambandhi-sambandha-samsarga, the commingling of relatum and relation. The passage is the Padakanda's charter: everything it is about to anatomize is a power of one thing, real in commingling, nothing alone.

3.1.25-26 unformalized

न तदुत्पद्यते किञ्चिद्यस्य जातिर्न विद्यते ।
आत्माभिव्यक्तये जातिः कारणानां प्रयोजिका ॥ २५ ॥
कारणेषु पदं कृत्वा नित्यानित्येषु जातयः ।
क्वचित्कार्येष्वभिव्यक्तिमुपयान्ति पुनः पुनः ॥ २६ ॥

na tadutpadyate kiñcidyasya jātirna vidyate |
ātmābhivyaktaye jātiḥ kāraṇānāṃ prayojikā || 25 ||
kāraṇeṣu padaṃ kṛtvā nityānityeṣu jātayaḥ |
kvacitkāryeṣvabhivyaktimupayānti punaḥ punaḥ || 26 ||

(25) Nothing whatever comes to be of which there is no universal; the universal is the occasioner of the causes, for the manifestation of its own self. (26) Universals, having taken their station in causes both eternal and non-eternal, come somewhere to manifestation in effects, again and again.

A causal role for the universal that inverts the usual picture. (25): na tad utpadyate kimcid yasya jatir na vidyate — *nothing* comes to be of which there is *no universal*: nothing is produced except as a something, an instance of a type; bare particulars are not born. And then the inversion: atma-abhivyaktaye jatih karananam prayojika — the universal is the *occasioner of the causes*, for the *manifestation of its own self*. Not: the causes produce the thing, which then has a universal. Rather: the universal deploys the causes so as to manifest itself — the type is the agent, the causes its instruments, the individual its appearing. (26) extends it: jatayah karanesu padam krtva nityanityesu — universals, *having taken their station* in causes eternal and non-eternal — punah punah abhivyaktim upayanti: come to manifestation *again and again*. The eternal type, stationed in the causal order, surfacing repeatedly as instances. This is the manifestation-doctrine of Kanda I (1.44-49: the sphota manifested by sounds it is not made of) transposed to ontology entire: individuals are to their universals what utterances are to the word — occasions of appearing, not constituents.
contested: prayojika ('occasioner') gives the universal genuine causal agency, which is a strong reading; a weaker construal ('is the occasion for the causes') keeps the causes primary. The abhivyakti-vocabulary of 26 and the parallel with Kanda I's manifestation-doctrine support the strong reading and we adopt it

3.1.27 unformalized

निर्वर्त्त्यमानं यत्कर्म जातिस्तत्रापि साधनम् ।
स्वाश्रयस्याभिनिष्पत्तौ सा क्रियाणां प्रयोजिका ॥ २७ ॥

nirvarttyamānaṃ yatkarma jātistatrāpi sādhanam |
svāśrayasyābhiniṣpattau sā kriyāṇāṃ prayojikā || 27 ||

The action that is being brought about — there too the universal is the means; in the completion of its own substratum, it is the occasioner of the actions.

The doctrine of 3.1.25-26 extended from things to actions. Nirvartyamanam yat karma — the action that is *being brought about*, in process of accomplishment — jatih tatra api sadhanam: there too the *universal* is the *means*. And: svasrayasya abhinispattau sa kriyanam prayojika — in the *completion of its own substratum*, it is the *occasioner of the actions*. The cooking-universal deploys the cookings; what makes this stirring part of a cooking is the type it is realizing, and the type is what drives the process to its completion. Note how far this is from the Vaisesika's inert universals sitting in their instances: Bhartrhari's jati is dynamic, purposive, the engine of its own instantiation (3.1.25's atma-abhivyaktaye, 'for the manifestation of its own self'). And note the term sadhanam — *means*: the seventh samuddesa is titled Sadhana-samuddesa, and it will treat the karakas. Universals as means is thus not a metaphor here but a placement: the jati belongs among the factors that bring actions about.

3.1.28 unformalized

विधौ वा प्रतिषेधे वा ब्राह्मणत्वादि साधनम् ।
व्यक्त्याश्रिताश्रिता जातेः संख्याजातिर्विशेषिका ॥ २८ ॥

vidhau vā pratiṣedhe vā brāhmaṇatvādi sādhanam |
vyaktyāśritāśritā jāteḥ saṃkhyājātirviśeṣikā || 28 ||

In injunction or in prohibition, brahminhood and the like is the means; the universal of number, resting on what rests on the individual, is what differentiates the universal.

The doctrine in ritual application, and then a fine piece of stratification. Vidhau va pratisedhe va — in *injunction* or in *prohibition* — brahmanatvadi sadhanam: brahminhood and the like is the *means*: what qualifies a man for the rite, or bars him from it, is the universal he instantiates, not his individuality (the adhikara-doctrine of 2.79, now grounded). Then the second line's stratification: samkhya-jatih visesika jateh — the *universal of number* is what *differentiates* the universal — vyakty-asrita-asrita, *resting on what rests on the individual*: number is a universal that inheres not in the individual directly but in what inheres in it. So there is a hierarchy: individual, then the universal in it, then number as a universal of *that*. Universals of universals (3.1.9's regress) here doing real grammatical work: the plural of a universal-word is a number-universal riding a quality-universal riding a substance. The verse shows the machinery the grammar needs and the ontology it costs — which is the samuddesa's constant theme.

3.1.29 unformalized

यथा जलादिभिर्व्यक्तं मुखमेवाभिधीयते ।
तथा द्रव्यैरभिव्यक्ता जातिरेवाभिधीयते ॥ २९ ॥

yathā jalādibhirvyaktaṃ mukhamevābhidhīyate |
tathā dravyairabhivyaktā jātirevābhidhīyate || 29 ||

As the face itself, manifested by water and the like, is what is expressed — so the universal itself, manifested by substances, is what is expressed.

The reflection returns, and this time to a constructive purpose. Yatha jaladibhih vyaktam mukham eva abhidhiyate: as the *face itself*, manifested by water and the like, is what is *expressed* — pointing at a reflection, one says 'that is my face', and one is right: the reflection is not a second face one names by courtesy; the face is what is named, manifested there (1.49's borrowed properties; 2.294's reflection failing the effect-test, but naming succeeding). Tatha dravyaih abhivyakta jatih eva abhidhiyate: so the *universal itself*, manifested by *substances*, is what is expressed. Point at a cow and say 'cow': what is named is cowhood, manifested in this beast as the face is in the water. The simile does exactly the work the jati-vadin needs — it explains how a word can name a universal while one is looking at a particular — and it is chosen for a further reason: the water manifests the face without the face being *in* the water, which is precisely how 3.1.25-26 want universals to stand to their instances. Manifestation, not inherence.

3.1.30-31 unformalized

यथेन्द्रियगतो भेद इन्द्रियग्रहणादृते ।
इन्द्रियार्थेष्वदृष्टोऽपि ज्ञानभेदाय कल्पते ॥ ३० ॥
तथात्मरूपग्रहणात् केषांचिद् व्यक्तयो विना ।
सामान्यज्ञानभेदानामुपयान्ति निमित्तताम् ॥ ३१ ॥

yathendriyagato bheda indriyagrahaṇādṛte |
indriyārtheṣvadṛṣṭo'pi jñānabhedāya kalpate || 30 ||
tathātmarūpagrahaṇāt keṣāṃcid vyaktayo vinā |
sāmānyajñānabhedānāmupayānti nimittatām || 31 ||

(30) As a difference residing in the sense-faculty, though unseen in the objects of the senses, serves — apart from any grasping of the faculty — for a difference of cognition; (31) so, for some, individuals, without any grasping of their own form, come to be the occasioning ground of the differences in the cognition of the general.

The samuddesa's subtlest argument, and it turns on a genuine observation about perception. (30): the sense-faculty has its own differences — the eye's constitution differs from the ear's, and within a faculty from moment to moment — and these are indriya-gato bhedah, residing in the *faculty*, adrsto indriya-arthesu, *unseen in the objects*: nothing in the pot corresponds to my eye's condition. Yet indriya-grahanad rte — *apart from any grasping of the faculty* (one never perceives one's own eye in perceiving) — the difference jnana-bhedaya kalpate: serves for a *difference of cognition*: my seeing differs though neither its object nor its organ is given to me. (31) applies it: tatha atma-rupa-grahanat kesamcid vyaktayo vina — so, for some, *individuals*, *without any grasping of their own form* — samanya-jnana-bhedanam upayanti nimittatam: come to be the *occasioning ground* of the *differences in the cognition of the general*. The individuals are never themselves grasped (3.1.12: particulars enter only in universal dress), and they nonetheless make my cowhood-cognition here differ from my cowhood-cognition there. Ungrasped, they condition; like the eye, they work by not appearing. It is an elegant rescue of the individual's reality from a doctrine that denies it can ever be meant.
contested: 31's atma-rupa-grahanat vina is syntactically awkward and our construal ('without grasping of their own form') follows the parallel with 30's indriya-grahanad rte; an alternative reads it as 'from the grasping of the self-form', reversing the sense — but that loses the analogy the pair is built on

3.1.32-33 unformalized

सत्यासत्यौ तु यो भावौ प्रतिभावं व्यवस्थितौ ।
सत्यं यत्तत्र सा जातिरसत्या व्यक्तयः स्मृताः ॥ ३२ ॥
सम्बन्धिभेदात्सत्तैव भिद्यमाना गवादिषु ।
जातिरित्युच्यते तस्यां सर्वे शब्दा व्यवस्थिताः ॥ ३३ ॥

satyāsatyau tu yo bhāvau pratibhāvaṃ vyavasthitau |
satyaṃ yattatra sā jātirasatyā vyaktayaḥ smṛtāḥ || 32 ||
sambandhibhedātsattaiva bhidyamānā gavādiṣu |
jātirityucyate tasyāṃ sarve śabdā vyavasthitāḥ || 33 ||

(32) But of the two entities, the real and the unreal, which stand in each thing — what is real there is the universal; the individuals are held to be unreal. (33) Existence alone, being differentiated by the difference of its relata among cows and the like, is called the universal; in it all words stand established.

The samuddesa's culmination, and it is as radical as anything in the work. (32): every thing has two entities standing in it — satya and asatya, the *real* and the *unreal*: satyam yat tatra sa jatih — what is *real* there is the *universal*; asatya vyaktayah smrtah — the *individuals* are held to be *unreal*. The realist's world is inverted at a stroke: types are what there is, tokens are appearance. Then (33) the final identification, which dissolves even the types: satta eva sambandhi-bhedat bhidyamana gavadisu jatir ity ucyate — *Existence alone*, being *differentiated* by the difference of its *relata* among cows and the like, is *called* the universal. There are not many universals: there is *satta*, Being, one; and what we call cowhood is Being as differentiated by its relata (3.1.20's difference imposed from outside; 3.1.16's places appearing between). The plural universals are Existence carved. And the last line brings the whole kanda home: tasyam sarve sabdah vyavasthitah — in *that*, *all words* stand established. Every word, whatever it seems to name, is established in Being — which is 1.1's sabda-tattva reached by the Padakanda's own road, and which explains 3.1.6's superimposition at last: the word's own universal is laid over the object's because both are the one Being, differently carved. The Jati-samuddesa began by asking what universals are and ends by saying there is only one, and it is not really a universal.
contested: 32's satya/asatya we render 'real/unreal'; the pair can be softened to 'true/false' or 'abiding/transient', and how strongly one takes the individuals' unreality is the classical crux of the samuddesa. 33's satta-doctrine (the maha-samanya, Being as the sole universal) is the Bhartrharian signature and is not in doubt

3.1.34 unformalized

तां प्रातिपदिकार्थं च धात्वर्थं च प्रचक्षते ।
सा नित्या सा महानात्मा तामाहुस्त्वतलादयः ॥ ३४ ॥

tāṃ prātipadikārthaṃ ca dhātvarthaṃ ca pracakṣate |
sā nityā sā mahānātmā tāmāhustvatalādayaḥ || 34 ||

And they declare that to be the meaning of the nominal stem and the meaning of the root; it is eternal, it is the great Self — the affixes 'tva', 'tal' and the rest express it.

Satta — Being, the sole universal reached at 3.1.33 — now claims the whole of grammar. Tam pratipadika-artham ca dhatv-artham ca pracaksate: they declare *that* to be the meaning of the *nominal stem* and the meaning of the *root* — both, the two great classes of meaning-bearing element the science recognizes: what nouns are built on and what verbs are built on alike mean Being. Then: sa nitya, it is *eternal*; sa mahan atma — it is the *great Self*, and the phrase is deliberate, importing the upanisadic register into a discussion of stems and roots. And the grammatical clincher: tam ahuh tvatal-adayah — the affixes *tva* and *tal* and the rest *express* it: these are the abstract-noun suffixes (gotva, 'cowness'; gota, the same), and what they do, on this account, is not manufacture an abstraction but *say Being* — 'cowness' means Being-as-carved-among-cows (3.1.33). The verse is the samuddesa's boldest identification: the abstract suffix is the grammar's own name for the mahan atma, and every stem and every root has been meaning it all along.

3.1.35 unformalized

प्राप्तक्रमा विशेषेषु क्रिया सैवाभिधीयते ।
क्रमरूपस्य संहारे तत्सत्त्वमिति कथ्यते ॥ ३५ ॥

prāptakramā viśeṣeṣu kriyā saivābhidhīyate |
kramarūpasya saṃhāre tatsattvamiti kathyate || 35 ||

That very one, having attained sequence among the particulars, is called action; in the withdrawal of the form of sequence, that same is spoken of as substance.

The unification, and it is one of the great verses of the work. Sa eva — *that very one*, satta, Being — prapta-krama visesesu kriya abhidhiyate: *having attained sequence* among the particulars, is called *action*. And: krama-rupasya samhare tat sattvam iti kathyate — in the *withdrawal* of the form of sequence, that same is spoken of as *substance*. Action and substance — the two great categories, verb-meaning and noun-meaning, the whole dichotomy the grammar is built on (2.195's asattva-bhuto bhavah versus sattva) — are *one thing* under two aspects: Being with sequence, Being with sequence withdrawn. There are not two kinds of meaning; there is Being, and whether we meet it as verb or noun depends on whether krama is in or out. And krama, the whole work has argued, is the mind's contribution (2.25's sequencing power; 2.27's sequenceless capacity generating ordered appearance). So the noun/verb distinction — the deepest structural fact about language — is a difference in how we take one thing, not a difference in what there is. This is 3.1.2's alternatives (jati or dravya) dissolved: both are satta, sequenced or not.

3.1.36 unformalized

सैव भावविकारेषु षडवस्थाः प्रपद्यते ।
क्रमेण शक्तिभिः स्वाभिरेवं प्रत्ययभासते ॥ ३६ ॥

saiva bhāvavikāreṣu ṣaḍavasthāḥ prapadyate |
krameṇa śaktibhiḥ svābhirevaṃ pratyayabhāsate || 36 ||

That same one attains the six states among the modifications of being; by its own powers, in sequence, it thus appears in cognition.

The doctrine of 3.1.35 given its classical furniture. Sa eva bhava-vikaresu sad-avasthah prapadyate: that same one — Being — attains the *six states* among the *modifications of being*: the sad-bhava-vikara of Yaska's Nirukta, which 1.3 already installed as the wombs of differentiation: is born, exists, changes, grows, declines, perishes. Kramena saktibhih svabhih: *in sequence*, by its *own powers* — the powers are Being's, the sequence is the mode of their exercise (3.1.35's prapta-krama). Evam pratyaya-bhasate: thus it *appears in cognition*. Note the verb: appears — the six states are how Being shows, not six things that happen to it. And note the closing of a circle: 1.3 introduced the six vikaras at the Brahmakanda's opening, resting on Brahman's kala-sakti; here, twelve hundred verses later, the same six are Being's own powers exercised in sequence. The Padakanda's ontology and the Brahmakanda's cosmology have turned out to be one doctrine approached from opposite ends.

3.1.37 unformalized

आत्मभूतः क्रमोऽप्यस्या यत्रेदं कालदर्शनम् ।
पौर्वापर्यादिरूपेण प्रविभक्तमिव स्थितम् ॥ ३७ ॥

ātmabhūtaḥ kramo'pyasyā yatredaṃ kāladarśanam |
paurvāparyādirūpeṇa pravibhaktamiva sthitam || 37 ||

Even the sequence belonging to it is its very self, wherein this seeing of time, in the form of before-and-after and the rest, stands as though divided.

The precision that keeps 3.1.35-36 from incoherence. If Being is sequenceless and action is Being *with* sequence, whence the sequence? Atma-bhutah kramo 'pi asyah: *even the sequence* belonging to it is *its very self* — not an addition, not an alien imposition, but Being's own (3.1.24's samagrata, the faculties' wholeness being their self). Yatra idam kala-darsanam paurvaparyadi-rupena pravibhaktam iva sthitam: wherein *this seeing of time*, in the form of before-and-after and the rest, stands *as though divided*. The iva does its usual office. Time's divisions are how the seeing takes it; the sequence that is Being's own self is undivided, and paurvaparya — before-and-after — is a division laid on. So sequence is real (it is Being's self) and its articulation into befores and afters is not (that is darsana, seeing). The verse threads a needle the whole work has been approaching: 2.24 argued that time admits no intrinsic divisions and 2.25 located sequence in the cognition; here the two findings are reconciled by distinguishing sequence-as-such (real, Being's own) from ordered sequence (constructed). The Kala-samuddesa, the ninth, will take this up at length.

3.1.38 unformalized

तिरोभावाभ्युपगमे भावानां सैव नास्तिता ।
लब्धक्रमे तिरोभावे नश्यतीति प्रतीयते ॥ ३८ ॥

tirobhāvābhyupagame bhāvānāṃ saiva nāstitā |
labdhakrame tirobhāve naśyatīti pratīyate || 38 ||

On the admission of the disappearance of things, that same one is their non-existence; in a disappearance that has attained sequence, it is understood as 'it perishes'.

Being made to account for non-being — the hardest thing the doctrine must do. Tirobhava-abhyupagame bhavanam — on the admission of the *disappearance* of things (granting that things do vanish) — sa eva nastita: *that same one* — Being — *is* their *non-existence*. The formulation is deliberately paradoxical and it is exact: what we call a thing's absence is Being, carved so as to exclude it (3.1.33: Being differentiated by its relata). Absence is not a further entity opposed to Being; it is Being under a determination. And the verb-form: labdha-krame tirobhave nasyati iti pratiyate — in a disappearance that has *attained sequence*, it is understood as 'it *perishes*': the same non-existence, sequenced (3.1.35's rule), presents as an action, the perishing. The verse thus delivers what 2.33 promised — that the capacities for existence *and* non-existence stand established in the one — and it disposes of the negation-problem that occupied 2.240-245: there is no cancelled deposit to account for, because the negative sentence's content is Being, carved to exclude, and nothing was ever added to be taken away.

3.1.39 unformalized

पूर्वस्मात्प्रच्युता धर्मादप्राप्ता चोत्तरं पदम् ।
तदन्तराले भेदानामाश्रयाज्जन्म कथ्यते ॥ ३९ ॥

pūrvasmātpracyutā dharmādaprāptā cottaraṃ padam |
tadantarāle bhedānāmāśrayājjanma kathyate || 39 ||

Fallen away from the prior property, and not having attained the subsequent station — in that interval, from the resort of the differences, birth is spoken of.

Birth analysed as an interval, and the analysis is subtle. Purvasmat pracyuta dharmat — *fallen away* from the prior property; aprapta ca uttaram padam — and *not having attained* the subsequent station: the thing is between, no longer what it was and not yet what it will be. Tad-antarale bhedanam asrayat janma kathyate: in *that interval*, from the *resort of the differences*, *birth* is spoken of. Birth is not an event that happens at an instant; it is the name for a gap — the interval where the differences take their stand, where a thing is neither of its two determinations. And this is precisely 3.1.20-21's condition: uncommingled, a thing is neither one nor many, neither existent nor non-existent — and here that condition is located not in some inaccessible thing-in-itself but in the ordinary interval of becoming. Everything that comes to be passes through the state to which no category applies. The verse gives the catuskoti a home in the biology of change.

3.1.40 unformalized

आश्रयः स्वात्ममात्रा वा भावा वा व्यतिरेकिणः ।
स्वशक्तयो वा सत्ताया भेददर्शनहेतवः ॥ ४० ॥

āśrayaḥ svātmamātrā vā bhāvā vā vyatirekiṇaḥ |
svaśaktayo vā sattāyā bhedadarśanahetavaḥ || 40 ||

The resort is either mere portions of its own self, or things that are distinct from it, or its own powers — these being the causes of the seeing of difference in Existence.

What the 'resort of the differences' (3.1.39) might be, canvassed in three options. Asrayah sva-atma-matra va — the resort is *mere portions of its own self*: Being's own aspects, nothing external; bhava va vyatirekinah — or *things distinct* from it: real others, standing outside; sva-saktayo va — or its *own powers* (the sakti-doctrine of 1.2 and 3.1.22-23). And all three are sattayah bheda-darsana-hetavah — *the causes of the seeing of difference in Existence*: whichever they are, what they do is cause the *seeing* of difference. Note the term: darsana, seeing (3.1.37's kala-darsanam). Bhartrhari canvasses three metaphysical accounts of what differentiates Being and does not choose — because on his method the choice is a ledger-choice (1.74, 2.179), and what all three deliver is the same: difference is *seen*, and its causes are what they are. The samuddesa's characteristic tolerance at the level of ontology, rigour at the level of phenomenology.
contested: the three options are compressed and their individuation is our reading; some take sva-atma-matra and sva-saktayah as one option variously phrased, yielding a two-fold canvass (own aspects versus real others), which is the more usual dichotomy

3.1.41 unformalized

पृथिव्यादिष्वभिव्यक्तौ न संस्थानमपेक्षते ।
अनुच्छिन्नाश्रयाज्जातिरनित्येऽप्याश्रये स्थिता ॥ ४१ ॥

pṛthivyādiṣvabhivyaktau na saṃsthānamapekṣate |
anucchinnāśrayājjātiranitye'pyāśraye sthitā || 41 ||

In its manifestation among earth and the rest, it does not depend on configuration; the universal, its substratum being uninterrupted, stands even in a non-eternal substratum.

A defence of universals against the transience of their instances. Prthivy-adisu abhivyaktau na samsthanam apeksate: in its manifestation among earth and the rest, it does not depend on *configuration* — the universal is not the shape; a lump of earth and a pot are both earth, and earthness cares nothing for their forms (against any akrti-doctrine that identifies the universal with configuration, 3.1.19's prakhyata akrti). And the second line: anucchinnasrayat jatih anitye 'pi asraye sthita — the universal, its *substratum being uninterrupted*, stands even in a *non-eternal* substratum. The pots come and go; the *series* of pots does not; and the universal rides the unbroken succession rather than any member of it. This is 1.28's vyavastha-nityata exactly — beginninglessness as regular arrangement rather than as any item's permanence, the analogy there being living beings, each produced and the series without a first term. The eternal universal needs no eternal instance, only an unbroken supply.

3.1.42 unformalized

अनुच्छेद्याश्रयामेके सर्वा जातिं प्रचक्षते ।
न यौगपद्यं प्रलये सर्वस्येति व्यवस्थिताः ॥ ४२ ॥

anucchedyāśrayāmeke sarvā jātiṃ pracakṣate |
na yaugapadyaṃ pralaye sarvasyeti vyavasthitāḥ || 42 ||

Some declare every universal to have an uninterruptible substratum; they are settled in the view that there is no simultaneity of everything at the dissolution.

The doctrine of 3.1.41 pressed to a cosmological commitment. Anucchedyasrayam eke sarva jatim pracaksate: some declare *every* universal to have an *uninterruptible* substratum — not merely unbroken in fact but incapable of interruption. Why so strong? Because of the pralaya, the world-dissolution: if at some moment *all* substrata perished together, every universal would lose its footing at once and could not return (3.1.41's succession requires that the chain never fully break). So these theorists are settled: na yaugapadyam pralaye sarvasya iti vyavasthitah — there is *no simultaneity* of everything at the dissolution: the destruction is staggered, something always remaining, the chain never quite parting. This is the seed-doctrine of 1.132 (when all traditions perish, a seed remains) transposed to ontology, and it shows how much metaphysical weight the universals' eternality is made to carry: rather than let the chain break, the school revises its cosmology.

3.1.43 unformalized

प्रकृतौ प्रविलीनेषु भेदेष्वेकत्वदर्शिनाम् ।
द्रव्यसत्त्लं प्रपद्यन्ते स्वाश्रया एव जातयः ॥ ४३ ॥

prakṛtau pravilīneṣu bhedeṣvekatvadarśinām |
dravyasattlaṃ prapadyante svāśrayā eva jātayaḥ || 43 ||

For those who see oneness, when the differences have dissolved into the primal nature, the universals — their own substrata — attain the state of substance.

The alternative cosmology, for those who can afford it. Prakrtau pravilinesu bhedesu — when the *differences* have *dissolved* into the *primal nature* (prakrti, the Samkhya's unmanifest matrix) — ekatva-darsinam: for those who *see oneness* (the abheda-purvaka party of 2.57, Bhartrhari's own). What happens to universals then? Dravya-sattvam prapadyante svasraya eva jatayah — the universals, *being their own substrata*, attain the state of *substance*. Where all differentiated instances have dissolved, the universals do not lose their footing: they *become* their own footing — cease to need instances and stand as substance. And by 3.1.35 this is exactly right: substance is Being with sequence withdrawn, and at the dissolution sequence has withdrawn. So the pralaya is not a catastrophe for universals but their return to what they always were: satta, undifferentiated, before its relata carve it (3.1.33). The verse resolves 3.1.42's anxiety by not sharing it — those who see oneness need no unbroken chain, because the chain was never what held the universal up.

3.1.44 unformalized

ब्राह्मणात्वादयो भावाः सर्वप्राणिष्ववस्थिताः ।
अभिव्यक्ताः स्वकार्याणां साधका इत्यापि स्मृतिः ॥ ४४ ॥

brāhmaṇātvādayo bhāvāḥ sarvaprāṇiṣvavasthitāḥ |
abhivyaktāḥ svakāryāṇāṃ sādhakā ityāpi smṛtiḥ || 44 ||

Brahminhood and such entities stand present in all living beings; manifested, they are the accomplishers of their own effects — such too is the received teaching.

A doctrine of remarkable reach, reported as smrti. Brahmanatvadayah bhavah sarva-pranisu avasthitah: brahminhood and the like *stand present in all living beings* — not in brahmins only: every universal is everywhere, latent, in everything that lives. What distinguishes a brahmin is not that he has brahminhood and others lack it but that in him it is abhivyaktah, *manifested*. And manifested, svakaryanam sadhakah — they are the *accomplishers of their own effects* (3.1.25's prayojika: the universal deploying the causes for its own manifestation). Iti api smrtih: *such too is the received teaching* — the report-marker, and Bhartrhari does not press it further. The doctrine follows strictly from 3.1.33 (there is only satta, carved by relata) and 3.1.25-27 (universals as self-manifesting powers): if all universals are one Being differentiated, then all of them are latent wherever Being is, which is everywhere. Its social implications are not drawn here, and it would be anachronistic to draw them for the text; but the metaphysics is unambiguous, and it is the metaphysics that the verse states.

3.1.45 unformalized

चित्रादिष्वप्यभिव्यक्तिर्जातीनां कैश्चीदिष्यते ।
प्राण्याश्रितास्तु ताः प्राप्तौ निमित्तं पुण्यपापयोः ॥ ४५ ॥

citrādiṣvapyabhivyaktirjātīnāṃ kaiścīdiṣyate |
prāṇyāśritāstu tāḥ prāptau nimittaṃ puṇyapāpayoḥ || 45 ||

By some, a manifestation of universals is accepted even in paintings and the like; but those that rest in living beings are the occasion, in their attainment, of merit and demerit.

A boundary drawn on 3.1.44's universality. Kaiscid citradisu api abhivyaktih jatinam isyate: by some, a manifestation of universals is accepted *even in paintings* — the painted brahmin has brahminhood manifested in him, the painted cow cowhood: manifestation does not require flesh (and 2.290 already granted the painting its likeness while denying it efficacy). But the qualification that matters: prany-asritas tu tah praptau nimittam punya-papayoh — those that rest in *living beings* are, in their attainment, the occasion of *merit and demerit*. Only the living manifestation carries dharmic weight: honouring a painted brahmin earns nothing, killing a painted cow costs nothing. So the boundary is not ontological but ethical — the painting has the universal and not the consequences — and it is drawn by exactly the criterion 2.288-290 used to sort the real from the unreal: asadharanam karyam, the effect that only the genuine article produces. Here the effect is merit, and merit is what the painting cannot do.

3.1.46 unformalized

ज्ञानं त्वस्मद्विशिष्टानां तासु सर्वेन्द्रियं विदुः ।
अभ्यासान्मणिरूप्यादिविशेषेष्विव तद्विदाम् ॥ ४६ ॥

jñānaṃ tvasmadviśiṣṭānāṃ tāsu sarvendriyaṃ viduḥ |
abhyāsānmaṇirūpyādiviśeṣeṣviva tadvidām || 46 ||

But the knowledge of those, in men distinguished from us, they know to be through every sense-faculty — as, from practice, among the particulars of gems and silver and the like, for those who know them.

How universals are known, and the answer returns to the work's oldest example. Asmad-visistanam — of men *distinguished from us*, those beyond our grade (the rsis of 1.37-38, whose knowledge of past and future is not different from perception) — jnanam tasu sarvendriyam viduh: their knowledge of universals is *through every sense-faculty*: not inferred, not reasoned to, but perceived, and by any sense whatever. And the analogy that makes it credible: abhyasat mani-rupyadi-visesesu iva tad-vidam — as, *from practice*, among the particulars of *gems and silver*, for *those who know them*. The jeweller of 1.35 returns, and his epistemic situation is now given its metaphysical import: what he perceives, unable to say how, is the universal — and what the seer perceives is the same, at every grade. Abhyasa, practice, is what closes the gap (2.117's pratibha-hetuh abhyasat; 2.152's six-fold flash). The doctrine of universals thus rests, in the end, not on argument but on connoisseurship — the trained eye that sees types where the untrained sees only this stone.

3.1.47 unformalized

जात्युत्पलादिगन्धादौ भेदतत्त्वं यदाश्रितम् ।
तद्भावप्रत्ययैर्लोकेऽनित्यत्वान्नाभिधीयते ॥ ४७ ॥

jātyutpalādigandhādau bhedatattvaṃ yadāśritam |
tadbhāvapratyayairloke'nityatvānnābhidhīyate || 47 ||

The reality of difference which resides in the scent of jasmine, lotus and the like — in the world it is not expressed by cognitions of that being, because of its non-eternality.

A limit on what universals language reaches. There is a genuine bheda-tattvam — a *reality of difference* — residing in the scent of jasmine as against lotus: the noses of the world discriminate them infallibly, so the difference is real. And yet: tad-bhava-pratyayaih loke na abhidhiyate — in the world it is *not expressed* by cognitions of that being: there is no word for it, no abstract noun, no 'jasmine-scent-ness' that ordinary usage employs. Why? Anityatvat — because of its *non-eternality*: the scent is too fleeting to have earned a universal-word. The verse concedes something important: real differences may go unnamed, and language's inventory of universals is not a map of the world's distinctions but of the ones usage has settled on (2.253's prasiddhi; 2.296's loka-prasiddhi deciding what is real). What we have words for is what has stayed put long enough to be worth a word.

3.1.48 unformalized

अस्वशब्दाभिधानास्तु नरसिंहादिजातयः ।
सरूपावयवेवान्या तासु श्रुतिरवस्थिता ॥ ४८ ॥

asvaśabdābhidhānāstu narasiṃhādijātayaḥ |
sarūpāvayavevānyā tāsu śrutiravasthitā || 48 ||

But the universals such as man-lion and the like are not expressed by their own words; in them stands a hearing that is, as it were, other, resting on like-formed parts.

The man-lion returns (2.90's specimen) and now as a problem for universals. Asva-sabda-abhidhanah — *not expressed by their own words*: narasimha-adi-jatayah, the universals of man-lion and the like, have no word of their own; 'narasimha' is a compound, borrowing from 'nara' and 'simha'. So a real kind (the tradition grants the man-lion its reality) with no proper name, denoted only by pieces that name other kinds. And what stands in such a word: sarupavayaveva anya tasu srutir avasthita — a *hearing* that is, as it were, *other*, resting on *like-formed parts*. The compound's members sound like the words for man and lion and are not those words (2.177's rule: different meaning, different base; 2.169's kupa-supa-yupa). So the man-lion's universal is reached by a word whose parts merely resemble words. The verse is a fine specimen of the samuddesa's method: a case where the smooth doctrine (words express universals) meets a kind that has to be named obliquely, and the obliquity is described rather than smoothed.

3.1.49 unformalized

जात्यवस्थापरिच्छेदे सङ्ख्या सङ्ख्यात्वमेव वा ।
विप्रकर्षेऽपि संसर्गादुपकाराय कल्पते ॥ ४९ ॥

jātyavasthāparicchede saṅkhyā saṅkhyātvameva vā |
viprakarṣe'pi saṃsargādupakārāya kalpate || 49 ||

In the delimiting of the universal's state, number, or number-hood itself, serves for assistance through commingling, even where there is remoteness.

Number's role in fixing universals. Jaty-avastha-paricchede — in the *delimiting of the universal's state* (which cowhood: the one, the many, the herd) — samkhya samkhyatvam eva va — *number*, or *number-hood* itself — upakaraya kalpate: serves for *assistance*; and the striking qualification: viprakarse 'pi samsargat — *even where there is remoteness*, through commingling. Number assists the universal even at a distance — the singular or plural of a word conditioning which state of the universal is meant, though number is not part of the universal and stands well away from it (3.1.28's stratification: number as a universal riding on what rests on the individual). The verse is preparing the Sankhya-samuddesa (the eleventh) and, more immediately, the discussion of vivaksa that follows: if number assists from a distance, then whether it assists at all in a given utterance is a matter of what the speaker means it to do.

3.1.50 unformalized

लक्षणा शब्दसंस्कारे व्यापारः कार्यसिद्धये ।
सङ्ख्याकर्मादिशक्तीनां श्रुतिसाम्येऽपि दृश्यते ॥ ५० ॥

lakṣaṇā śabdasaṃskāre vyāpāraḥ kāryasiddhaye |
saṅkhyākarmādiśaktīnāṃ śrutisāmye'pi dṛśyate || 50 ||

Indication is the operation, in the refining of the word, for the accomplishing of the effect; the capacities of number, action and the rest are seen even where the hearing is the same.

The grammar's own use of the foregoing. Laksana sabda-samskare vyaparah karya-siddhaye: *indication* is the *operation*, in the *refining of the word* (sabda-samskara: 1.131's term, the grammarian's whole business), *for the accomplishing of the effect*: the science indicates in order to get its operations right, and indication is a doing, not a describing (2.436: word-meanings stand by way of characterization, not in reality). And: samkhya-karmadi-saktinam sruti-samye 'pi drsyate — the capacities of *number*, *action* and the rest are *seen* even where the *hearing is the same*: identical forms carrying different capacities, sorted by nothing audible (2.104's capacity-assumption in the invariant stream). So the grammar's indications are postings that track capacities the sound does not carry — which is why the science is necessary and why its categories are not discoveries.

3.1.51 unformalized

न विना सङ्ख्यया कश्चित्सत्त्वभूतोऽर्थ उच्यते ।
अतः सर्वस्य निर्देशे सङ्ख्या स्यादविवक्षिता ॥ ५१ ॥

na vinā saṅkhyayā kaścitsattvabhūto'rtha ucyate |
ataḥ sarvasya nirdeśe saṅkhyā syādavivakṣitā || 51 ||

Nothing that is a substance-entity is spoken of at all without number; therefore, in the designation of everything, number would be unintended.

A dilemma about number, sharply put. Na vina samkhyaya kascit sattva-bhuto 'rtha ucyate: *nothing* that is a substance-entity is spoken of *without number* — Sanskrit has no numberless noun; to say 'cow' is to say one cow, and there is no way to mention a thing without counting it. But then: atah sarvasya nirdese samkhya syad avivaksita — therefore, in the designation of *everything*, number would be *unintended*: if number is unavoidable, it cannot be meant — what one cannot help saying, one does not thereby intend (1.136's canon of sato 'vivaksa, the non-intention of the actual; 2.69's unintended qualities of the rice). The reasoning is exact and it produces a genuine puzzle: the grammar's number-apparatus is obligatory and therefore, on the tradition's own principle, semantically idle. What is forced is not meant.

3.1.52 unformalized

एकत्वं वा बहुत्वं वा केषाञ्चिदविवक्षितम् ।
तद्धि जात्यभिधानाय द्वित्वं तु स्याद्विवक्षितम् ॥ ५२ ॥

ekatvaṃ vā bahutvaṃ vā keṣāñcidavivakṣitam |
taddhi jātyabhidhānāya dvitvaṃ tu syādvivakṣitam || 52 ||

For some, either oneness or manifoldness is unintended, for that serves for the expressing of the universal; but duality would be intended.

The puzzle of 3.1.51 resolved by a discrimination, and it is a fine one. Ekatvam va bahutvam va kesamcid avivaksitam: for some, either *oneness* or *manifoldness* is *unintended* — the singular and the plural are the forms one uses when saying nothing about number, tad hi jaty-abhidhanaya, for *that* (either of them) serves for the *expressing of the universal*: 'the cow is an animal' uses a singular and means cowhood; 'cows are animals' uses a plural and means the same. Neither number is doing number-work. Dvitvam tu syad vivaksitam: but *duality* would be *intended* — the dual is never idle. Sanskrit's dual is used only when exactly two are meant; there is no generic dual. So the obligatory numbers (singular, plural) are unmeant and the optional one (dual) is meant, and the criterion is precisely 3.1.51's: what one could have avoided saying, one intended. The verse is a small triumph of the vivaksa-doctrine (1.155, 2.400): meaning is what was aimed, and grammar's forced categories are outside the aim.

3.1.53 unformalized

यद्येतौ व्याधितौ स्यातां देयं स्यादिदमौषधम् ।
इत्येवं लक्षणेऽर्थस्य द्वित्वं स्यादविवक्षितम् ॥ ५३ ॥

yadyetau vyādhitau syātāṃ deyaṃ syādidamauṣadham |
ityevaṃ lakṣaṇe'rthasya dvitvaṃ syādavivakṣitam || 53 ||

'If these two should fall ill, this medicine is to be given' — in such an indicating of the meaning, duality would be unintended.

The rule of 3.1.52 (the dual is always meant) immediately qualified, and the qualification is exact. 'If these two fall ill, give this medicine': the sentence's purpose is laksana, *indicating* a condition — it tells you when to administer, not how many patients matter. Would one withhold the medicine from a third? Obviously not; and would one wait until both were ill? Not that either. So dvitvam syad avivaksitam — *duality would be unintended*, even in the dual, whose whole distinction was that it is never idle. What overrides the rule is purpose: where the sentence is indicating rather than specifying, the number is carried along and not aimed (2.301's smoke: what an operation cannot help producing, it does not intend). The verse guards 3.1.52 from being taken as a formal rule about the dual. It is not. It is an application of vivaksa, and vivaksa answers to what the sentence is *for* (2.113), which the dual's morphology cannot settle.

3.1.54 unformalized

एकादिशब्दवाच्यायाः कर्मण्यङ्गत्वमिष्यते ।
सङ्ख्यायाः खनति द्वाभ्यामिति रूपाद्धि साश्रिता ॥ ५४ ॥

ekādiśabdavācyāyāḥ karmaṇyaṅgatvamiṣyate |
saṅkhyāyāḥ khanati dvābhyāmiti rūpāddhi sāśritā || 54 ||

For number expressible by 'one' and the rest, auxiliary-status in the action is accepted; for in 'he digs with two', it is resorted to from the form.

The contrary case, where number *is* doing work. Ekadi-sabda-vacyayah samkhyayah karmani angatvam isyate: for number expressible by 'one' and the rest, *auxiliary-status* in the action is accepted — angatva, the Mimamsa relation of a subsidiary to its principal (2.81-82): the number is a genuine contributing factor to the act. The evidence: khanati dvabhyam — 'he digs *with two*' (two spades, two hands) — rupad dhi sa asrita: from the *form* it is resorted to. Digging with two is a different action from digging with one; the number enters the act's own description. So numbers sort into two classes by the same criterion as everything else in this samuddesa: those the action needs (auxiliary, intended) and those it merely carries (unintended, 3.1.51). And the sorting is not readable off the case-ending — 'with two' and 'these two' are both duals, and one is meant and one is not.

3.1.55 unformalized

यजेत पशुनेत्यत्र संस्कारस्यापि सम्भवे ।
यथा जातिस्तथैकत्वं साधनत्वेन गम्यते ॥ ५५ ॥

yajeta paśunetyatra saṃskārasyāpi sambhave |
yathā jātistathaikatvaṃ sādhanatvena gamyate || 55 ||

In 'let him sacrifice with an animal', even where refinement is possible — as the universal is, so oneness is understood as a means.

The test-case that will occupy the rest of the movement. Yajeta pasuna — 'let him sacrifice with an animal': the injunction names a beast, in the singular, and the question is whether the singular is meant. Samskarasya api sambhave — *even where refinement is possible*: even granting that the animal must be ritually prepared, which might seem to fix attention on the individual. And the answer offered: yatha jatih tatha ekatvam sadhanatvena gamyate — *as the universal is*, so *oneness* is understood *as a means*: both the animal-universal and the oneness are contributing factors, sadhana, to the rite. If cowhood is a means (3.1.27's jati as sadhana), then so is the number. The parity-argument is clean and it sets the trap the following verses spring: if number is a means here, the sacrificer must use exactly one animal and no more — and the exegetical tradition has views about that.

3.1.56 unformalized

लिङ्गात्तु स्याद् द्वितीयादेस्तदेकत्वं विवक्षितम् ।
एकार्थविषयत्वे च तल्लिङ्गं जातिसङ्ख्ययोः ॥ ५६ ॥

liṅgāttu syād dvitīyādestadekatvaṃ vivakṣitam |
ekārthaviṣayatve ca talliṅgaṃ jātisaṅkhyayoḥ || 56 ||

But from the mark of 'a second' and the rest, that oneness would be intended; and where the object is single, that mark belongs to both universal and number.

Evidence for 3.1.55's reading, from elsewhere in the ritual code. Lingat tu syad dvitiyadeh — from the *mark* of 'a second' and the rest: other injunctions speak of a second animal, a further one — which presupposes that the first injunction meant *one*; tad ekatvam vivaksitam, that oneness would be *intended*. The reasoning is the linga-inference of 2.73 (a mark elsewhere disclosing a meaning here; 2.86's difference inferred from a mark seen in another sentence). And the second line notes the mark's ambiguity: eka-artha-visayatve ca tal lingam jati-samkhyayoh — where the object is *single*, that mark belongs to *both* universal and number: 'a second animal' equally evidences that the first injunction meant one animal and that it meant animal-kind. The mark does not discriminate. Which is why the movement must go on: the evidence is real and it underdetermines the question.

3.1.57 unformalized

अन्यत्राविहितस्यैव स विधिः प्रथमं पशोः ।
क्रियायामङ्गभावश्च तत्त्वेतस्माद्विवक्षितम् ॥ ५७ ॥

anyatrāvihitasyaiva sa vidhiḥ prathamaṃ paśoḥ |
kriyāyāmaṅgabhāvaśca tattvetasmādvivakṣitam || 57 ||

That injunction is first, of an animal not enjoined elsewhere; and there is auxiliary-status in the action — therefore, from this, it is intended.

The argument for intention completed, and its two premises are the exegete's standard tools. Anyatra avihitasya eva sa vidhih prathamam pasoh: that injunction is *first*, of an animal *not enjoined elsewhere* — this is the rite's own animal, introduced here, not carried in from another context (an apurva prescription, 3.1.69's term: what is not otherwise obtained must be enjoined). And kriyayam anga-bhavah ca — and there is *auxiliary-status* in the action (3.1.54's criterion). Tat tu etasmat vivaksitam: therefore, from this, it is *intended*. The reasoning: a number that is freshly prescribed, in a rite that has no other source for it, and which contributes to the act, must be meant — nothing else explains its being said. This is the strongest case for number-as-intended, and the verses following will set against it a class of cases where every one of these conditions fails.

3.1.58-59 unformalized

ग्रहास्त्वन्यत्र विहिता भिन्नसङ्ख्याः पृथक् पृथक् ।
प्राजापत्या नवेत्येवमादिभेदसमन्विताः ॥ ५८ ॥
अङ्गत्वेन प्रतीतानां सम्मार्गे त्वङ्गिनां पुनः ।
निर्देशं प्रति या सङ्ख्या सा कथं स्याद्विवक्षिता ॥ ५९ ॥

grahāstvanyatra vihitā bhinnasaṅkhyāḥ pṛthak pṛthak |
prājāpatyā navetyevamādibhedasamanvitāḥ || 58 ||
aṅgatvena pratītānāṃ sammārge tvaṅgināṃ punaḥ |
nirdeśaṃ prati yā saṅkhyā sā kathaṃ syādvivakṣitā || 59 ||

(58) But the graha-vessels are enjoined elsewhere, of differing numbers, severally — 'nine Prajapatya' and so on, accompanied by such differences. (59) But in the cleansing of the principals, which were understood as auxiliaries — the number that is there with respect to the designation, how would that be intended?

The counter-case, and it is precisely constructed. The grahas — the soma-cups — are anyatra vihitah, *enjoined elsewhere*: their numbers come from other injunctions, prthak prthak, severally, each rite specifying its own ('nine Prajapatya', and so on: bhinna-samkhyah, of differing numbers). So when a *cleansing* rule mentions them, their number is not being prescribed; it is being *referred to*. (59): sammarge angi nam punah — in the *cleansing* of the *principals* (which, in the earlier context, were understood angatvena, as auxiliaries — the role-inversion of 2.306, and 3.1.69 will name the principle) — nirdesam prati ya samkhya sa katham syad vivaksita: the number that is there *with respect to the designation*, how would *that* be intended? It could not: a number one merely mentions in order to identify what to clean is not a number one is enjoining. Every condition of 3.1.57 fails — the number is obtained elsewhere, is not fresh, contributes nothing to the cleansing — and the conclusion inverts.

3.1.60 unformalized

नान्यत्र विधिरस्तीति संस्कारो नापि चाङ्गिता ।
हेतुः सङ्ख्याविवक्षायां यत्नात्सा हि विवक्षिता ॥ ६० ॥

nānyatra vidhirastīti saṃskāro nāpi cāṅgitā |
hetuḥ saṅkhyāvivakṣāyāṃ yatnātsā hi vivakṣitā || 60 ||

'There is no injunction elsewhere' — thus neither refinement nor principal-status is the cause in the intending of number; for it is intended from effort.

The criterion isolated, and the isolation is the movement's real work. Two candidate explanations of why number is meant are ruled out. Not samskara — *refinement*: that the item must be ritually prepared does not make its number meant (3.1.55 raised this and it fails). Not angita — *principal-status*: that the item is the rite's principal does not make its number meant either. What then? Na anyatra vidhir asti iti — *because there is no injunction elsewhere*: number is meant exactly where it is not obtainable from another source, so that saying it here is the only way to have it. And the formulation that names the principle: yatnat sa hi vivaksita — for it is intended *from effort*: what the text goes to the trouble of prescribing, it means. Yatna — effort, trouble taken — is the criterion, and it is 3.1.51-52's in another guise: what is forced is not meant, what costs something to say is.

3.1.61 unformalized

सम्मार्जने विशेषश्च न ग्रहे क्वचिदाश्रितः ।
विहितास्ते च संस्कार्याः सर्वेषामाश्रयस्ततः ॥ ६१ ॥

sammārjane viśeṣaśca na grahe kvacidāśritaḥ |
vihitāste ca saṃskāryāḥ sarveṣāmāśrayastataḥ || 61 ||

And in the cleansing, no particular is anywhere resorted to in the graha; and those that are enjoined are what is to be refined — therefore the substratum of all.

The counter-case sealed. Sammarjane visesah ca na grahe kvacid asritah: in the *cleansing*, *no particular* is anywhere resorted to in the graha — the rule does not care which cup, or how many; it says clean them. Vihitah te ca samskaryah — and *those that are enjoined* (elsewhere, with their numbers, 3.1.58) are *what is to be refined*: the cleansing rule takes its objects readymade from the injunctions that prescribed them. Sarvesam asrayah tatah — *therefore* it is the *substratum of all*: the cleansing applies across the board, to whatever the other injunctions supplied, in whatever number they supplied it. The verse completes the demonstration that a number can be uttered, heard, and grammatically present while being semantically inert — not because number is generally idle (3.1.54 showed it working) but because *this* rule's purpose does not reach it.

3.1.62 unformalized

प्रत्याश्रयं समाप्तायां जातावेकेन चेत्क्रिया ।
पशुना न प्रकल्पेत तत्स्यादेव प्रकल्पनम् ॥ ६२ ॥

pratyāśrayaṃ samāptāyāṃ jātāvekena cetkriyā |
paśunā na prakalpeta tatsyādeva prakalpanam || 62 ||

If, the universal being completed substratum by substratum, the action would not be constructed with one animal — then that construction would indeed occur.

The distribution-question of the Vakyakanda (2.371-455) reappears, now as a ritual worry. Praty-asrayam samaptayam jatau — the *universal* being *completed substratum by substratum* (2.18's whole-at-each principle: animal-hood complete in each animal) — ekena cet kriya na prakalpeta pasuna: *if* the action would *not* be constructed with *one* animal — then tat syad eva prakalpanam: *that* construction would indeed occur — i.e. more animals would have to be brought. The conditional is the exegete's live question: does the rite need one beast or many? And the verse's structure makes the answer depend on the *action's* requirements, not the noun's number — 2.371's samarthya-criterion exactly. If sacrificing can be done with one, one suffices, whatever the singular does or does not mean.

3.1.63-64 unformalized

एकेन चेत्प्रसिद्धायां क्रियायां यदि सम्भवात् ।
पश्वन्तरमुपादेयमुपादानमनर्थकम् ॥ ६३ ॥
यथैवाहितगर्भायां गर्भाधानमनर्थकम् ।
तथैकेन प्रसिद्धायां पश्वन्तरमनर्थकम् ॥ ६४ ॥

ekena cetprasiddhāyāṃ kriyāyāṃ yadi sambhavāt |
paśvantaramupādeyamupādānamanarthakam || 63 ||
yathaivāhitagarbhāyāṃ garbhādhānamanarthakam |
tathaikena prasiddhāyāṃ paśvantaramanarthakam || 64 ||

(63) If, the action being already accomplished by one, another animal were to be taken up merely because it is possible — the taking-up would be purposeless. (64) Just as, in one whose embryo is already placed, the placing of an embryo is purposeless — so, the action being accomplished by one, another animal is purposeless.

The answer to 3.1.62, and its simile is memorable. (63): if the rite is prasiddha, *accomplished*, by one animal, then bringing a second sambhavat, merely *because it is possible*, is anarthakam — *purposeless*: possibility is not a reason; superfluity is not piety. (64) drives it home: yatha eva ahita-garbhayam garbhadhanam anarthakam — just as, in a woman *whose embryo is already placed*, the *placing of an embryo* is purposeless: the garbhadhana rite, performed on the already-pregnant, accomplishes nothing — the thing it exists to bring about has been brought about. Tatha ekena prasiddhayam pasvantaram anarthakam: so, the action accomplished by one, another animal is purposeless. The principle is that of ritual economy — a rite's requirements are satisfied or not, and beyond satisfaction there is no more merit to be had — and the simile is chosen for a reason: the pregnancy-case is one where the redundancy is undeniable to everyone. Nobody thinks a second placing helps.

3.1.65 unformalized

तावतार्थस्य सिद्धत्वादेकत्वस्याव्यतिक्रमम् ।
केचिदिच्छन्ति न त्वत्र सङ्ख्याङ्गत्वेन गृह्यते ॥ ६५ ॥

tāvatārthasya siddhatvādekatvasyāvyatikramam |
kecidicchanti na tvatra saṅkhyāṅgatvena gṛhyate || 65 ||

Because the purpose is accomplished by that much, some hold to the non-transgression of oneness; but here number is not grasped as an auxiliary.

The position that emerges from 3.1.62-64, and its careful limitation. Tavata arthasya siddhatvat — because the purpose is *accomplished by that much* (by one animal) — kecid icchanti ekatvasya avyatikramam: some hold to the *non-transgression* of oneness: one animal, and no more, is what the rite gets. So far this looks like 3.1.55's conclusion (oneness as means). But: na tu atra samkhya angatvena grhyate — *but here number is not grasped as an auxiliary*. The distinction is fine and it is the movement's yield. One animal is used — not because the singular *enjoined* oneness, but because one suffices and surplus is purposeless (3.1.63-64). The practical outcome is identical; the semantics is opposite. Number is not doing the work; economy is. This is exactly the discrimination the samuddesa has been building toward: a form can be reliably correlated with an outcome and contribute nothing to the meaning that produces it.

3.1.66 unformalized

द्वितीयादिषु यल्लिङ्गमुक्तं न्यायानुवादि तत् ।
न सङ्ख्या साधनत्वेन जातिवत्तेन गम्यते ॥ ६६ ॥

dvitīyādiṣu yalliṅgamuktaṃ nyāyānuvādi tat |
na saṅkhyā sādhanatvena jātivattena gamyate || 66 ||

The mark stated in the case of 'a second' and the rest is a restatement of the principle; number is not thereby understood as a means, as the universal is.

The evidence of 3.1.56 disarmed. Dvitiyadisu yal lingam uktam nyaya-anuvadi tat: the *mark* stated in the case of 'a second' and the rest is nyaya-anuvadi — a *restatement of the principle*, not an independent witness. When another text speaks of 'a second animal', it is not disclosing that the first injunction enjoined oneness; it is registering what the economy-principle (3.1.63-64) already established — that one was enough and this is an addition. Anuvada is 2.115's term for idle restatement, and its application here is deft: the mark that seemed to prove number-intention turns out to be repeating what was known on other grounds. Hence: na samkhya sadhanatvena jativat tena gamyate — number is *not* thereby understood as a *means*, as the *universal* is. The parity of 3.1.55 is broken: cowhood is a genuine sadhana of the rite, oneness is not, and the mark that seemed to equate them says nothing.

3.1.67 unformalized

अन्वयव्यतिरेकाभ्यां सङ्ख्याभ्युपगमे सति ।
युक्तं यत्साधनत्वं स्यान्न त्वन्यार्थोपलक्षणम् ॥ ६७ ॥

anvayavyatirekābhyāṃ saṅkhyābhyupagame sati |
yuktaṃ yatsādhanatvaṃ syānna tvanyārthopalakṣaṇam || 67 ||

When number is admitted by way of agreement and difference, it is right that it should be a means — but not that it should be an indicator of another meaning.

The methodological point beneath the casuistry. Anvaya-vyatirekabhyam samkhya-abhyupagame sati — when number is admitted by *agreement and difference* (the distributional method of 2.12 and 2.166: vary the form, watch the meaning) — yuktam yat sadhanatvam syat: it is *right* that it should be a *means*: if number covaries with outcome, credit it as a contributing factor. Na tu anya-artha-upalaksanam: *but not* that it should be an *indicator of another meaning* — one may not, on distributional evidence, promote number into a *sign* of something else (a mark of capacity, of principality, of whatever the exegete needs). The restriction matters because 3.1.3-5 did exactly that with the universal (jatih sakty-upalaksanam: the universal as indicator of capacity), and the verse is refusing the same licence to number. Covariance establishes contribution; it does not establish signification (2.167's sap: a unit certified by distribution and meaning nothing).

3.1.68 unformalized

साधनत्वे पदार्थस्य सामर्थ्यं न प्रदीयते ।
सङ्ख्या व्यापारधर्मोऽतस्तेन लिङ्गेन गम्यते ॥ ६८ ॥

sādhanatve padārthasya sāmarthyaṃ na pradīyate |
saṅkhyā vyāpāradharmo'tastena liṅgena gamyate || 68 ||

In being a means, the word-meaning's capacity is not given; number is a property of the operation — therefore it is understood by that mark.

Number's real status, at last. Sadhanatve padarthasya samarthyam na pradiyate: in being a means, the *word-meaning's* capacity is *not given* — being a factor in an action does not confer semantic power on the word that carries it (3.1.67's restriction, generalized). And the positive placement: samkhya vyapara-dharmah — number is a *property of the operation*: it belongs to the doing, not to the thing done or to the word that names it (2.83's individuation by operation; 3.1.11's padarthas characterized by vyapara). Atah tena lingena gamyate: therefore it is understood *by that mark* — number is inferred from how the action goes, not read off the noun. The verse resolves the movement's question in a way neither party expected: number is neither meant by the word (3.1.51's forced category) nor idle (3.1.54's digging with two) — it is a feature of operations, detected by their marks, and the case-ending is a symptom rather than a source.

3.1.69-70 unformalized

अपूर्वस्य विधेयत्वात्प्राधान्यमवसीयते ।
विहितस्य परार्थत्वात्वाच्छेषभावः प्रतीयते ॥ ६९ ॥
संमार्गस्य विधेयत्वादन्यत्र विहिते ग्रहे ।
विधिवाक्ये श्रुता सङ्ख्या लक्षणाया न बाध्यते ॥ ७० ॥

apūrvasya vidheyatvātprādhānyamavasīyate |
vihitasya parārthatvātvāccheṣabhāvaḥ pratīyate || 69 ||
saṃmārgasya vidheyatvādanyatra vihite grahe |
vidhivākye śrutā saṅkhyā lakṣaṇāyā na bādhyate || 70 ||

(69) From the enjoinability of the unprecedented, principality is determined; from the being-for-another of what is enjoined, subordinate-status is understood. (70) From the enjoinability of the cleansing, the graha being enjoined elsewhere — the number heard in the injunctive sentence is not annulled from its indicating.

The movement closes with the two principles that governed it, stated as maxims. (69): apurvasya vidheyatvat pradhanyam avasiyate — from the *enjoinability of the unprecedented*, *principality* is determined: what the text must prescribe, because it is available nowhere else (apurva: the exegete's technical term for the not-otherwise-obtained), is the rule's principal business. Vihitasya para-arthatvat sesa-bhavah pratiyate: from the *being-for-another* of what is *already enjoined*, *subordinate*-status is understood — what the rule merely mentions, having got it from elsewhere, is in service of something else (2.113's parartha, 2.82's sesa). Together these sort every element of every injunction into what it enjoins and what it presupposes. (70) applies them to the case: the *cleansing* is what this rule enjoins; the *graha* with its number is presupposed, enjoined elsewhere (3.1.58); therefore vidhi-vakye sruta samkhya laksanaya na badhyate — the number *heard* in the injunctive sentence is *not annulled* from its *indicating*: it goes on doing its referential work, identifying which cups, without being enjoined and without being contradicted. Heard, useful, and unmeant — which is precisely the status this long movement was constructed to establish.

3.1.71 unformalized

विधिवाक्यान्तरे सङख्या पशोर्नास्ति विरोधिनी ।
तस्मात्सगुण एवासौ सदेकत्वेन गम्यते ॥ ७१ ॥

vidhivākyāntare saṅakhyā paśornāsti virodhinī |
tasmātsaguṇa evāsau sadekatvena gamyate || 71 ||

In another injunctive sentence there is no number of the animal that conflicts; therefore that very one, together with its quality, is understood as existing with oneness.

The case of 3.1.55 resolved by the absence of a competitor. Vidhi-vakyantare samkhya pasoh na asti virodhini: in *another* injunctive sentence there is *no* number of the animal that *conflicts* — nothing elsewhere prescribes a different count, so nothing contests what this rule's singular presents. Tasmat sa-guna eva asau sad-ekatvena gamyate: therefore *that very one*, sa-guna, *together with its quality*, is understood as *existing with oneness*. The reasoning is negative and it is characteristic of the exegetical method: number stands here not because it was argued into place but because nothing dislodges it — the badha-machinery (2.66's annulment, 2.73's hierarchy of evidences) having found no conflict to adjudicate. Where nothing competes, what is heard holds. This is the settlement of the long movement on the sacrificial animal: one beast, and the oneness is neither enjoined (3.1.65-66) nor annulled — it simply survives, unopposed.

3.1.72 unformalized

निर्ज्ञातद्रव्यसंबन्धे यः कर्मण्युपदिश्यते ।
गुणस्तेनार्थिता तस्य द्रव्येणेव प्रतीयते ॥ ७२ ॥

nirjñātadravyasaṃbandhe yaḥ karmaṇyupadiśyate |
guṇastenārthitā tasya dravyeṇeva pratīyate || 72 ||

A quality that is taught in the action, the substance-connection being already known — by that, its need for the substance is understood as though for an instrument.

The relation of quality to substance in an injunction, and it inverts the expected direction. Nirjnata-dravya-sambandhe — the *substance-connection* being *already known* (the rite's substance settled by an earlier rule) — yah karmani upadisyate gunah: a *quality* that is taught in the action: 'with a red cow', where the cow is given and the redness is added. Tena arthita tasya dravyena iva pratiyate: by *that* (quality), its *need* for the substance is understood *as though* for an instrument. The quality needs the substance the way an action needs its means — the redness cannot present itself except in a red thing (1.62's paratantra guna: the subordinate has no standing alone). So the taught quality does not qualify a substance already independently there; it *requires* one, and the requirement is what the injunction registers. The iva keeps the ontology honest: the quality does not literally use the substance as a tool, but its dependence has the same structure.

3.1.73-74 unformalized

कश्चिदेव गुणो द्रव्ये यथा सामर्थ्यलक्षणः ।
आधारोऽपि गुणस्यैवं प्राप्तः सामर्थ्यलक्षणः ॥ ७३ ॥
तयोस्तु पृथगर्थित्वे संबंधो यः प्रतीयते ।
न तस्मिन्नुपघातोऽस्ति कल्प्यमन्मन्न चाश्रुतम् ॥ ७४ ॥

kaścideva guṇo dravye yathā sāmarthyalakṣaṇaḥ |
ādhāro'pi guṇasyaivaṃ prāptaḥ sāmarthyalakṣaṇaḥ || 73 ||
tayostu pṛthagarthitve saṃbaṃdho yaḥ pratīyate |
na tasminnupaghāto'sti kalpyamanmanna cāśrutam || 74 ||

(73) Just as some quality in a substance is characterized by capacity, so too a substratum for the quality, obtained, is characterized by capacity. (74) But when the two are needed separately, the connection that is understood — in that there is no injury, nor is anything unheard to be constructed.

The dependence of 3.1.72 shown to be mutual. (73): kascid eva guno dravye yatha samarthya-laksanah — as *some quality* in a substance is characterized by *capacity* (3.1.3's jatih sakty-upalaksanam: the property flagged because it flags a capacity) — adharo 'pi gunasya evam praptah samarthya-laksanah: so too the *substratum* for the quality, obtained, is characterized by capacity. The redness needs a cow, and the cow is wanted for what it can do. Neither is prior; each is there for the capacity the other completes. (74): tayoh tu prthag-arthitve — but when the two are *needed separately*, each on its own account — sambandho yah pratiyate, the *connection* that is understood: na tasmin upaghato 'sti — in that there is *no injury*: nothing is strained, no rule violated; kalpyam anyat na ca asrutam: nor is *anything unheard* to be *constructed* — no supplement must be imported. The maxim is the exegete's economy (2.84's guru/laghu: prefer the lighter assumption): a construal that needs no additions and breaks nothing is the one to take.

3.1.75 unformalized

क्रियया योऽभिसंबन्धः स श्रुतिप्रापितस्तयोः ।
आश्रयाश्रयिणार्वाक्यान्नियमस्त्ववतिष्ठते ॥ ७५ ॥

kriyayā yo'bhisaṃbandhaḥ sa śrutiprāpitastayoḥ |
āśrayāśrayiṇārvākyānniyamastvavatiṣṭhate || 75 ||

The connection with the action is what is brought by the explicit text for those two; but the restriction of substratum and what rests on it stands from the sentence.

The two channels distinguished, and the distinction is the hierarchy of 2.73 at work. Kriyaya yo 'bhisambandhah sa sruti-prapitah tayoh: the connection *with the action* is *brought by the explicit text* for both quality and substance — sruti, what is actually said, delivers that they figure in the rite. But: asraya-asrayinoh vakyat niyamah tu avatisthate — the *restriction* of substratum and what rests on it stands *from the sentence*: which is the bearer and which the borne is settled not by any word but by the construal, vakya — the third and weakest of the evidences (sruti > linga > vakya). So the rite's participation is *said*; their internal architecture is *construed*. The stratification matters for what follows: a connection resting on vakya can be overridden by one resting on sruti, and 3.1.76 immediately draws the consequence.

3.1.76 unformalized

तत्र द्रव्यगुणाभावे प्रत्येकं स्याद्विकल्पनम् ।
श्रुतिप्राप्तो हि संबन्धो बलवान्वाक्यलक्षणात् ॥ ७६ ॥

tatra dravyaguṇābhāve pratyekaṃ syādvikalpanam |
śrutiprāpto hi saṃbandho balavānvākyalakṣaṇāt || 76 ||

There, in the absence of substance and quality, there would be an option for each severally; for a connection obtained by the explicit text is stronger than one characterized by construal.

The consequence, and it is a rule of practical adjudication. Tatra dravya-guna-abhave pratyekam syad vikalpanam: in the *absence* of substance and quality — where the prescribed cow or the prescribed redness is unavailable — there would be *option* for each severally: one may take the substance without the quality, or the quality's bearer without this substance; the rite does not simply fail. And the ground: sruti-prapto hi sambandho balavan vakya-laksanat — a connection obtained by the *explicit text* is *stronger* than one characterized by *construal*. Since the text said that quality and substance figure (3.1.75's sruti-channel) but only construal linked them to each other (the vakya-channel), the link is the weak member and gives way first. So when the red cow cannot be had, one does not abandon the rite; one relaxes the construed bond and keeps what was said. The verse is fine exegetical engineering, and its principle — stronger evidence survives, weaker yields — is the same badha-machinery that 2.73 catalogued.

3.1.77 unformalized

यदा तु जातिः शक्तिर्वा क्रियां प्रत्युपदिश्यते ।
सामर्थ्यात्संनिधीयेते तत्र द्रव्यगुणौ तदा ॥ ७७ ॥

yadā tu jātiḥ śaktirvā kriyāṃ pratyupadiśyate |
sāmarthyātsaṃnidhīyete tatra dravyaguṇau tadā || 77 ||

But when a universal, or a capacity, is taught with respect to the action — then substance and quality are brought near, from capacity.

The reverse configuration. Where the injunction names not a substance and a quality but a jati or a sakti — 'sacrifice with cattle', 'use something that holds' — then dravya-gunau samarthyat sannidhiyete: substance and quality are *brought near*, from *capacity*: they are not said, and they arrive, because a universal cannot be sacrificed and a capacity cannot be handled — something must bear them (3.1.72's arthita, the need running the other way). Sannidhi is the term of 2.47 (fixity shining forth through mere proximity) and 2.51 (particulars manifested in co-presence): the unsaid participants are not inferred but *drawn in*, by what the said ones cannot do without. So the injunction's silence about the beast is not a gap: naming cattle-hood summons cattle, because capacity does the summoning. The pair 3.1.76-77 thus completes the picture: what is said may be relaxed under pressure, and what is unsaid arrives under necessity — and neither transaction runs through the words.

3.1.78 unformalized

जातीनां च गुणानां च तुल्येऽङ्गत्वे क्रियां प्रति ।
गुणाः प्रतिनिधीयन्ते छागादीनां न जातयः ॥ ७८ ॥

jātīnāṃ ca guṇānāṃ ca tulye'ṅgatve kriyāṃ prati |
guṇāḥ pratinidhīyante chāgādīnāṃ na jātayaḥ || 78 ||

Though universals and qualities are equally auxiliary with respect to the action, qualities are substituted; the universals of goats and the like are not.

An asymmetry that the parity-doctrine did not predict, and it is decisive. Jatinam ca gunanam ca tulye 'ngatve kriyam prati: universals and qualities are *equally* auxiliary — both contribute to the rite, both were called sadhana (3.1.27, 3.1.55). And yet: gunah pratinidhiyante chagadinam na jatayah — *qualities* are *substituted*; the *universals* of goats and the like are *not*. When the prescribed red goat cannot be had, one may use a black one — the quality yields. When no goat can be had, one may not use a sheep — the universal does not yield. The substitution-rule (3.1.3's pratinidhi, the samuddesa's opening evidence) discriminates sharply between two things it had treated as equals. So angatva, auxiliary-status, is not one status: within it, some auxiliaries are essential and some are negotiable, and ritual practice knows the difference infallibly while no doctrine so far advanced predicts it. The next verse explains the asymmetry.

3.1.79 unformalized

व्यक्तिशक्तेः समासन्ना जातयो न तथा गुणाः ।
साक्षाद्द्रव्यं क्रियायोगि गुणस्तस्माद्विकल्पते ॥ ७९ ॥

vyaktiśakteḥ samāsannā jātayo na tathā guṇāḥ |
sākṣāddravyaṃ kriyāyogi guṇastasmādvikalpate || 79 ||

Universals are very near to the capacity of the individual; qualities are not so. The substance is directly conjoined with the action; therefore the quality is optional.

The asymmetry of 3.1.78 explained, and the explanation is a metaphysics of proximity. Vyakti-sakteh samasanna jatayah — *universals* are *very near* to the *capacity of the individual*: goathood is what makes this goat able to do what goats do in a rite; remove it and the capacity goes. Na tatha gunah: *not so* the qualities — the redness is not what the goat's ritual capacity rests on; it can go and the capacity stays. And: saksat dravyam kriya-yogi — the *substance* is *directly* conjoined with the action: it is what is handled, bound, offered. Gunas tasmat vikalpate: *therefore* the quality is *optional*. The account is exact and it rescues the substitution-rule from arbitrariness: what may be substituted is what stands at a remove from the capacity the action needs, and what may not is what the capacity is. This is 3.1.3-5's doctrine (the universal as indicator of capacity) doing real work at last — and it explains why the tradition never let a sheep stand for a goat while cheerfully accepting a discoloured one.

3.1.80 unformalized

साम्येनान्यतराभावे विकल्पः कैश्चिदिष्यते ।
अतद्गुणोऽतश्छागः स्यान्मेषो वा तद्गुणो भवेत् ॥ ८० ॥

sāmyenānyatarābhāve vikalpaḥ kaiścidiṣyate |
atadguṇo'taśchāgaḥ syānmeṣo vā tadguṇo bhavet || 80 ||

By some, option is accepted where, both being equal, one or the other is absent; therefore the goat would be one lacking that quality, or the sheep would be one possessing it.

The hard case, where the asymmetry of 3.1.78-79 gives out. Samyena anyatara-abhave vikalpah kaiscid isyate: by *some*, *option* is accepted where — the two being *equal* — one or the other is *absent*. And the dilemma spelled out: atad-guno 'tas chagah syat — *therefore* the goat would be one *lacking* that quality (a goat, wrong colour); mesah va tad-guno bhavet — *or* the sheep would be one *possessing* it (right colour, wrong beast). Which? The verse reports that some allow a free option — take either. Bhartrhari does not adjudicate, and the case is genuinely hard: 3.1.79's rule says keep the universal and drop the quality, but that rule was derived from cases where the quality was clearly remote from the capacity, and here the colour may be doing ritual work of its own. The exegetical tradition's willingness to accept vikalpa — free option — as an outcome is itself notable: 2.85 recorded that option is the science's last resort, and here is a case that reaches it.

3.1.81-82 unformalized

जातेराश्रितसंख्यायाः प्रवृत्तिरुपलभ्यते ।
संख्याविशेषमुत्सृज्य क्वचित्सैव प्रवर्तते ॥ ८१ ॥
पराङ्गभूतं सामान्यं युज्यते द्रव्यसंख्यया ।
स्वार्थं प्रवर्तमानं तु न संख्यामवलम्बते ॥ ८२ ॥

jāterāśritasaṃkhyāyāḥ pravṛttirupalabhyate |
saṃkhyāviśeṣamutsṛjya kvacitsaiva pravartate || 81 ||
parāṅgabhūtaṃ sāmānyaṃ yujyate dravyasaṃkhyayā |
svārthaṃ pravartamānaṃ tu na saṃkhyāmavalambate || 82 ||

(81) The operation of a universal whose number is dependent is observed; somewhere that same one operates, having relinquished any particular number. (82) A general that has become auxiliary to another is joined with the substance's number; but when it operates for its own sake, it does not lean upon number.

Number's dependence on the universal's role, and the rule is elegant. (82) states it: para-anga-bhutam samanyam yujyate dravya-samkhyaya — a general that has become *auxiliary to another* is *joined with the substance's number*: where cowhood is serving a cow ('bring the two cows'), it takes the number of what it serves. Sva-artham pravartamanam tu na samkhyam avalambate: but when it operates *for its own sake* — where cowhood is itself the topic ('cowhood is a universal') — it *does not lean upon* number: the abstract noun takes a singular that means nothing (3.1.52's unintended ekatva). So number attaches to universals only parasitically, through the substances they qualify, and drops away the moment the universal is spoken of in its own right. (81) records the observation the rule explains: sometimes the universal operates samkhya-visesam utsrjya, *having relinquished* any particular number. The verse-pair gives the grammar's number-apparatus its exact semantic scope: it reaches things, and reaches types only through things.

3.1.83 unformalized

यजेत पशुनेत्यत्र यज्यर्थायां पशुश्रुतौ ।
कृतार्थैकेन पशुना प्रधानं भवति क्रिया ॥ ८३ ॥

yajeta paśunetyatra yajyarthāyāṃ paśuśrutau |
kṛtārthaikena paśunā pradhānaṃ bhavati kriyā || 83 ||

In 'let him sacrifice with an animal' — the animal being heard for the sake of the sacrificing — the action is the principal, one animal having served the purpose.

The test-case's final settlement. Yajeta pasuna iti atra — in 'let him sacrifice with an animal' — yajy-arthayam pasu-srutau: the *animal* being *heard for the sake of the sacrificing*: the noun is there in service of the verb (2.48's karaka-architecture: nouns orbit the action). Krtartha ekena pasuna pradhanam bhavati kriya: *one animal* having *served the purpose*, the *action* is the *principal*. Every element of the long movement is gathered here. The action is principal (3.1.69's rule: what the text must enjoin is its business); the animal is its means, heard for its sake; one suffices because purpose is what governs (3.1.63-64's economy); and the oneness is neither enjoined nor annulled but simply what serving the purpose required (3.1.65's discrimination). Krtartha — 'having served the purpose' — is the operative word and it is the whole doctrine: the rite wanted a sacrificing, the animal was for that, and when the purpose is met the question of number does not arise.

3.1.84-85 unformalized

यावतां संभवो यस्य स कुर्यात्तावतां यदि ।
आलम्भनं गुणैस्तेन प्रधानं स्यात्प्रयोजितम् ॥ ८४ ॥
संमृज्यमानतन्त्रे तु ग्रहे यत्र क्रियाश्रुतिः ।
संख्याविशेषग्रहणं नैव तत्राद्रियामहे ॥ ८५ ॥

yāvatāṃ saṃbhavo yasya sa kuryāttāvatāṃ yadi |
ālambhanaṃ guṇaistena pradhānaṃ syātprayojitam || 84 ||
saṃmṛjyamānatantre tu grahe yatra kriyāśrutiḥ |
saṃkhyāviśeṣagrahaṇaṃ naiva tatrādriyāmahe || 85 ||

(84) If one should perform the slaying of as many as are possible for him — then by those qualities the principal would be occasioned. (85) But in the graha, where the ritual apparatus is being cleansed and there is a hearing of the action — we do not at all insist there on the grasping of a particular number.

The two poles of the movement set side by side. (84) states the maximalist temptation: yavatam sambhavo yasya sa kuryat tavatam yadi alambhanam — *if* one should slay *as many as are possible* for him — gunaih tena pradhanam syat prayojitam: then by *those qualities* the *principal* would be *occasioned*: more offerings, more merit, the principal served by abundance. The verse states it as a conditional and 3.1.63-64 has already refused it: beyond satisfaction there is no more to be had, and surplus is anarthaka. (85) gives the settled contrary: sammrjyamana-tantre grahe yatra kriya-srutih — in the *graha*, where the apparatus is being *cleansed* and the action is *heard* — samkhya-visesa-grahanam na eva tatra adriyamahe: we do *not at all insist* there on the grasping of a *particular number*. The first-person plural is worth noticing — adriyamahe, 'we do not insist': the author speaking as a member of the tradition, in its voice, settling a case. Between 84's refused maximalism and 85's relaxed cleansing, the movement's whole range is on display: number matters exactly where purpose makes it matter, and nowhere else.

3.1.86 unformalized

शिष्यमाणपरे वाक्ये यदेकग्रहणं कृतम् ।
शेषे विशिष्टसंख्येऽपि व्यक्तं तल्लिङ्गदर्शनम् ॥ ८६ ॥

śiṣyamāṇapare vākye yadekagrahaṇaṃ kṛtam |
śeṣe viśiṣṭasaṃkhye'pi vyaktaṃ talliṅgadarśanam || 86 ||

In a sentence whose purport concerns what is to be taught, the mention of 'one' that has been made — in the remainder, even where the number is distinct, that mark's exhibition is evident.

A refinement on how the number-word functions in a teaching-sentence. Sisyamana-pare vakye — in a sentence whose *purport* concerns *what is to be taught* (the sentence's business is instruction, not specification) — yad eka-grahanam krtam: the mention of *'one'* that has been made — sese visista-samkhye 'pi vyaktam tal-linga-darsanam: in the *remainder*, even where the number is *distinct*, that *mark's exhibition* is *evident*. Where the sentence teaches, its 'one' is a specimen: the number is exhibited as a mark, an illustration of the pattern, and applies to the remainder whatever their actual counts. This is the pedagogical function of forms that the whole work has insisted on (2.13's divisions for the unlearned; 2.176's instruction by the general for brevity's sake) — and it identifies a third status for number, beside intended and unintended: *exhibited*, offered as a sample rather than a stipulation.

3.1.87 unformalized

समासप्रत्ययविधौ यथा निपतिता श्रुतिः ।
गुणानां परतन्त्राणां न्यायेनैवोपपद्यते ॥ ८७ ॥

samāsapratyayavidhau yathā nipatitā śrutiḥ |
guṇānāṃ paratantrāṇāṃ nyāyenaivopapadyate || 87 ||

As the explicit text has fallen, in the rule for compounds and affixes — of qualities that are dependent, it is established by the principle itself.

The doctrine confirmed from the grammar's side. Samasa-pratyaya-vidhau yatha nipatita srutih: as the *explicit text* has *fallen* — nipatita, dropped into place, the wording as it happens to stand — in the rule for *compounds and affixes*. Gunanam paratantranam nyayena eva upapadyate: *of qualities that are dependent*, it is established *by the principle itself*. Panini's rules for compounds and affixes mention qualities, and the mentions carry numbers and genders that nobody supposes are prescribed: the wording fell that way, and the *principle* — that dependents take their determinations from what they depend on (3.1.82's rule) — supplies what the wording does not mean. So the grammar's own text exhibits exactly the phenomenon the ritual science has been sorting: forms present, heard, and semantically inert, their apparent content supplied by structure rather than by the words. The verse's nipatita is dry and exact: the text fell there; do not read fate into where it landed.

3.1.88 unformalized

गुणेऽपि नाङ्गीक्रियते प्रधानान्तरसिद्धये ।
संख्या कर्ता तथा कर्मण्यविशिष्टः प्रतीयते ॥ ८८ ॥

guṇe'pi nāṅgīkriyate pradhānāntarasiddhaye |
saṃkhyā kartā tathā karmaṇyaviśiṣṭaḥ pratīyate || 88 ||

Even in the quality, number is not admitted as auxiliary for the accomplishing of another principal; the agent, likewise, is understood in the object as undifferentiated.

The number-doctrine stated in its final restriction. Gune 'pi na angikriyate pradhanantara-siddhaye samkhya: *even in the quality*, number is *not admitted as auxiliary* for the accomplishing of *another* principal — a quality's number does not enter the service of a rite whose business is elsewhere; it does not get promoted into a factor merely because it is present (3.1.67's restriction: covariance establishes contribution, not signification). And the parallel: karta tatha karmani avisistah pratiyate — the *agent*, likewise, is understood in the *object* as *undifferentiated*: where the sentence's business is the object, the agent comes along unspecified, its determinations unmeant. Two categories, one rule: what the sentence is not about arrives without its features meant. This is 1.136's canon of sato 'vivaksa — the non-intention of what is actually there — applied across the karaka-system, and it prepares the Sadhana-samuddesa, the seventh, where the karakas get their full treatment.
contested: the second line's karta tatha karmani avisistah is compressed and our construal (the agent unspecified where the object is at issue) follows the parallel with the first line; some read it as a separate observation about agent-object identity in certain constructions

3.1.89 unformalized

यस्यान्यस्य प्रसक्तस्य नियमार्था पुनः श्रुतिः ।
निवृत्तौ चरितार्थत्वात्संख्या तत्राविवक्षिता ॥ ८८ ॥

yasyānyasya prasaktasya niyamārthā punaḥ śrutiḥ |
nivṛttau caritārthatvātsaṃkhyā tatrāvivakṣitā || 88 ||

That other thing which, being applicable, has a further explicit text for the sake of restriction — since it has served its purpose in the exclusion, number there is unintended.

A last case, and it turns on what a restricting text has already done. Yasya anyasya prasaktasya niyama-artha punah srutih: that other thing which, being *applicable* — prasakta, already in play, obtainable — has a *further explicit text* for the sake of *restriction* (the sruti here doing niyama-work: narrowing what was antecedently possible, as 'rice' narrows 'substance' at 2.66). Nivrttau caritarthatvat samkhya tatra avivaksita: since it has *served its purpose in the exclusion*, number there is *unintended*. The restricting text's whole office was to exclude the alternatives; having done that, it is caritartha — spent, its errand discharged (2.54's caritartha: the words that vanish into their meaning; 2.270's carita-asti-kriya). Whatever else it carries — its number among them — is residue. The verse completes the movement's inventory of ways a number can fail to be meant: forced (3.1.51), merely referential (3.1.59), exhibited as a mark (3.1.86), fallen with the wording (3.1.87), riding a dependent (3.1.88), and now spent along with its sentence's discharged errand.
contested: the source numbers this verse 88 in its final line while labelling the page 3.1.89; we follow the page-label, and note the numbering fault

3.1.90 unformalized

सरूपसमुदायात्तु विभक्तिर्या विधीयते ।
एकस्तत्रार्थवान्सिद्धः समुदायस्य वाचकः ॥ ९० ॥

sarūpasamudāyāttu vibhaktiryā vidhīyate |
ekastatrārthavānsiddhaḥ samudāyasya vācakaḥ || 90 ||

But the case-ending which is prescribed after an aggregate of like-formed things — there, one alone, being meaningful, is established as the expressor of the aggregate.

The movement closes by returning to ekasesa, and the return is pointed. Sarupa-samudayat tu vibhaktir ya vidhiyate: the *case-ending* prescribed after an *aggregate of like-formed* things — the single-remainder construction (2.108-110: 'ramau' for two Ramas, one form serving many). Ekas tatra arthavan siddhah samudayasya vacakah: *one alone*, being *meaningful*, is *established* as the *expressor of the aggregate*. One word, meaningful, expressing many — and the number that does the expressing is carried by an ending on a form that stands for all of them. So the samuddesa's long interrogation of number ends where the Vakyakanda's boundary-machinery ended: with the device in which a single form bears plural office (2.109: uttering them severally would be *incorrect*), and in which, therefore, number is neither distributed among the words nor absent from the meaning. It is a capacity of the one form, carved by the construction — which is what the samuddesa has been arguing about number all along, and what 3.1.6 said about every word's meaning in its sixth verse.

3.1.91 unformalized

प्रत्ययस्य प्रधानस्य समासस्यापि वा विधौ ।
सिद्धः संख्याविवक्षायां सर्वथानुग्रहो गुणे ॥ ९१ ॥

pratyayasya pradhānasya samāsasyāpi vā vidhau |
siddhaḥ saṃkhyāvivakṣāyāṃ sarvathānugraho guṇe || 91 ||

In the rule for an affix, or for a principal, or even for a compound — when number is intended, assistance in the quality is established in every way.

A closing rule for the number-movement, gathering its cases. Pratyayasya pradhanasya samasasya api va vidhau — in the rule for an *affix*, for a *principal*, or even for a *compound*: the three sites where the grammar's number-apparatus operates. Siddhah samkhya-vivaksayam sarvatha anugraho gune: when number is *intended*, *assistance in the quality* is established *in every way*. Where number is genuinely meant (3.1.60's yatnat: what costs effort to say), it does its work wherever it stands — affix, principal, or compound — and the quality it attaches to is thereby served. The rule's interest is its scope: having spent forty verses establishing how rarely number is intended, the samuddesa closes the topic by noting that where it *is* intended, nothing blocks it. Sarvatha, 'in every way': the exceptions were all about vivaksa, never about capacity.

3.1.92 unformalized

अभेदरूपं सादृश्यामात्मभूताश्च शक्तयः ।
जातिपर्यायवाचित्वमेषामप्युपवर्ण्यते ॥ ९२ ॥

abhedarūpaṃ sādṛśyāmātmabhūtāśca śaktayaḥ |
jātiparyāyavācitvameṣāmapyupavarṇyate || 92 ||

The form of non-difference, similarity, and the powers that are one's very self — the expressing of these too, as synonyms of the universal, is described.

A widening of the universal's vocabulary, and the widening is philosophically loaded. Three things are named as jati-paryaya-vacitva — expressed as *synonyms of the universal*: abheda-rupam, the *form of non-difference*; sadrsya, *similarity*; and atma-bhutah saktayah, the *powers that are one's very self* (3.1.22-24's doctrine: the categories as powers of the one). So 'universal' names indifferently: what is not-different, what resembles, and what a thing can do. The three are not obviously the same, and their identification is the samuddesa's compressed argument: non-difference is what recurrence amounts to (3.1.14's anuvrtti), similarity is how recurrence is met in experience (2.90-94's respect-wise resemblance without shared parts), and capacity is what any of it is *for* (3.1.11's padarthas characterized by operation). Three descriptions, one phenomenon — and the verse's paryaya, 'synonym', is the term 2.150 used to deny that corrupt and correct forms are interchangeable. Here the interchange is granted.

3.1.93-94 unformalized

दण्डोपादित्सया दण्डं यद्यपि प्रतिपद्यते ।
न तस्मादेव सामर्थ्यात्स दण्डीति प्रतीयते ॥ ९३ ॥
नेच्छानिमित्तादिच्छावानिति ज्ञानं प्रवर्तते ।
तस्मात्सत्यपि सामर्थ्ये बुद्धिरर्थान्तराश्रया ॥ ९४ ॥

daṇḍopāditsayā daṇḍaṃ yadyapi pratipadyate |
na tasmādeva sāmarthyātsa daṇḍīti pratīyate || 93 ||
necchānimittādicchāvāniti jñānaṃ pravartate |
tasmātsatyapi sāmarthye buddhirarthāntarāśrayā || 94 ||

(93) Although one obtains a staff through the desire to take up a staff — not from that capacity alone is he understood as 'staff-bearer'. (94) The cognition 'he has desire' does not proceed from the occasioning ground of desire; therefore, even where the capacity exists, the cognition rests on another object.

A sharp argument that capacity does not by itself deliver a description. (93): a man's iccha, his *desire to take up a staff*, is what gets him the staff — the capacity is real and it operates. And yet: na tasmad eva samarthyat sa dandi iti pratiyate — *not from that capacity alone* is he understood as *'staff-bearer'*: he is called dandi because he *holds* one, not because he wanted one. Capacity does not name. (94) generalizes: na iccha-nimittad iccha-van iti jnanam pravartate — the cognition 'he *has* desire' does not proceed from the *occasioning ground* of desire: one does not know a man is desirous by inspecting his desire's cause. Tasmat sati api samarthye buddhih arthantarasraya: therefore, *even where the capacity exists*, the cognition *rests on another object*. The verse pair is doing quiet damage to 3.1.92's identification of the universal with capacity: if capacity does not deliver descriptions, then the universal — which does — is not simply capacity. The samuddesa is auditing its own equations.

3.1.95 unformalized

स्वभावोऽव्यपदेश्यो वा सामर्थ्यं वावतिष्ठते ।
सर्वस्यान्ते यतस्तस्माव्द्यवहारो न कल्पते ॥ ९५ ॥

svabhāvo'vyapadeśyo vā sāmarthyaṃ vāvatiṣṭhate |
sarvasyānte yatastasmāvdyavahāro na kalpate || 95 ||

Either an indescribable own-nature, or capacity, stands at the end of everything — wherefore, from that, practice is not constructed.

The terminus of analysis, and its consequence. Svabhavo 'vyapadesyo va samarthyam va avatisthate sarvasya ante: either an *indescribable own-nature*, or *capacity*, stands *at the end of everything* — push any analysis far enough and one reaches either a nature that cannot be told (3.1.20-21's uncommingled thing, to which no category applies; 2.416's what-has-own-form-cannot-be-described) or a bare capacity. Yatah tasmat vyavaharo na kalpate: *wherefore*, *from that*, *practice is not constructed*. Neither terminus supports transaction: the indescribable cannot be said, and capacity does not name (3.1.93-94). So whatever is at the bottom, vyavahara does not rest on it — practice runs on the constructed middle (2.297's sabdah loka-nibandhanah: words are grounded in the world, the shared audited domain, not in the ultimate). The verse is the samuddesa's most economical statement of the work's persistent finding: the real is not what we speak from, and speech's ground is not the ground.

3.1.96 unformalized

यदा भेदान्परित्यज्य बुद्धयैक इव गृह्यते ।
व्यक्त्यात्मैव तदा तत्र बुद्धिरेका प्रवर्तते ॥ ९६ ॥

yadā bhedānparityajya buddhayaika iva gṛhyate |
vyaktyātmaiva tadā tatra buddhirekā pravartate || 96 ||

When, having abandoned the differences, it is grasped by the intellect as though one — then it is the individual's very self, and a single cognition operates there.

The first of two accounts of how a plurality comes to be taken as one. Yada bhedan parityajya buddhya eka iva grhyate: when, *having abandoned the differences*, it is grasped by the *intellect* as though *one* — the mind drops the distinctions and takes what remains. Tada vyakty-atma eva tatra buddhih eka pravartate: then it is the *individual's very self*, and a *single cognition* operates there. Note what this account says: what is grasped is not a universal over the individuals but the individual's own self, the differences merely set aside. So oneness here is subtractive — an achievement of ignoring, not a perception of a type (against 3.1.29's face-in-water, where the universal itself was manifested). The iva marks it: grasped *as though* one. The verse offers the deflationary account of unity, and 3.1.97 will offer its companion.

3.1.97 unformalized

भेदरूपैरनुस्यूतं यदैकमिव मन्यते ।
समूहावग्रहा बुद्धिर्बहुभ्यो जायते तदा ॥ ९७ ॥

bhedarūpairanusyūtaṃ yadaikamiva manyate |
samūhāvagrahā buddhirbahubhyo jāyate tadā || 97 ||

When one thinks of what is interwoven with the forms of difference as though it were one — then a cognition grasping the collection arises from the many.

The second account, and it inverts the first. Bheda-rupaih anusyutam — what is *interwoven with the forms of difference* (anusyuta: the weaver's word, 1.123's anuviddham iva sabdena — cognition shot through with word) — yada ekam iva manyate: when one thinks of it as though *one*. Then: samuha-avagraha buddhih bahubhyah jayate tada — a cognition *grasping the collection* arises *from the many*. Here the unity is not achieved by abandoning differences (3.1.96) but by taking the differentiated manifold *as* a collection: the many are retained and gathered. So there are two routes to oneness — subtractive (ignore the differences) and collective (gather them) — and they produce different cognitions: one of the individual's self, one of the samuha, the group. The pair is the samuddesa's contribution to a question the whole work has circled: what is it to take many as one? And its answer is that there is no single such operation.

3.1.98 unformalized

सकृत्प्रवृत्तावेकत्वमावृत्तौ सदृशात्मताम् ।
भिन्नात्मिकानां व्यक्तीनां भेदापोहात्प्रपद्यते ॥ ९८ ॥

sakṛtpravṛttāvekatvamāvṛttau sadṛśātmatām |
bhinnātmikānāṃ vyaktīnāṃ bhedāpohātprapadyate || 98 ||

In a single operation, oneness; in repetition, the state of being of like self — the individuals, whose selves are distinct, attain this from the exclusion of difference.

The two routes of 3.1.96-97 given their conditions, and the last word is the loaded one. Sakrt-pravrttau ekatvam: in a *single* operation — one glance, one act — *oneness*; avrttau sadrsatmatam — in *repetition*, the state of being *of like self*: met again, the thing shows as similar rather than as identical. So oneness is what a single act delivers and similarity is what repeated acts deliver — a fine observation about how types are built out of encounters. And the mechanism: bhinnatmikanam vyaktinam bheda-apohat prapadyate — the individuals, *whose selves are distinct*, attain this *from the exclusion of difference*. Apoha — *exclusion*: the word that the Buddhist logicians would make the foundation of their entire semantics, and which 3.1.14 already canvassed as one of the two available definitions of the universal (vyavrtti). Here it names the operation by which distinct individuals come to present as one or as alike: not by sharing something but by having their differences excluded. The samuddesa on universals has arrived at the doctrine that would be used to abolish universals, and states it as its own mechanism.
contested: bheda-apohat is the crux: whether Bhartrhari uses apoha here as a settled technical term (and if so, whose) or in its plain sense of 'removal' cannot be determined from the karika. The relation between this verse and the developed apoha-doctrine of Dignaga is a genuine and unresolved question of intellectual history

3.1.99 unformalized

अनुप्रवृत्तेति यथाभिन्ना बुद्धिः प्रतीयते ।
अर्थो व्यावृत्तरूपोऽपि तथा तत्त्वेन गृह्यते ॥ ९९ ॥

anupravṛtteti yathābhinnā buddhiḥ pratīyate |
artho vyāvṛttarūpo'pi tathā tattvena gṛhyate || 99 ||

As a cognition, by way of recurrence, is understood as undivided — so the object too, though its form is exclusion, is grasped as the reality.

The parallel that closes the argument of 3.1.96-98. Anupravrtta iti yatha abhinna buddhih pratiyate: as a *cognition*, by way of *recurrence* (anupravrtti, 3.1.19's mark of the universal), is understood as *undivided* — the repeated cognition presents as one continuing thing, not as a series. Artho vyavrtta-rupo 'pi tatha tattvena grhyate: so the *object* too, *though its form is exclusion* (vyavrtta: 3.1.14's second definition, 3.1.98's apoha), is grasped *as the reality*. Both sides of the transaction are the same structure: what is really a recurrence presents as a unity; what is really an exclusion presents as a positive thing. Tattvena — 'as the reality': we take the excluded-from-others *to be* something. The verse states the illusion-mechanism with complete evenhandedness between the two definitions: whichever the universal is, it is not what it presents as, and the presenting is what makes speech possible (2.238's asatya vartman: standing on the untrue path to seek the true).

3.1.100-101 unformalized

सरूपाणां च सर्वेषां न भेदोपनिपातिनः ।
विद्यन्ते वाचकाः शब्दा नापि भेदोऽवधार्यते ॥ १०० ॥
ज्ञानशब्दार्थविषया विशेषा ये व्यवस्थिताः ।
तेषां दुरवधारत्वाज्ज्ञानाद्येकत्वदर्शनम् ॥ १०१ ॥

sarūpāṇāṃ ca sarveṣāṃ na bhedopanipātinaḥ |
vidyante vācakāḥ śabdā nāpi bhedo'vadhāryate || 100 ||
jñānaśabdārthaviṣayā viśeṣā ye vyavasthitāḥ |
teṣāṃ duravadhāratvājjñānādyekatvadarśanam || 101 ||

(100) And of all like-formed things there exist no expressing words that fall upon their difference; nor is the difference ascertained. (101) The particulars that stand as the objects of cognition, word and meaning — because those are hard to ascertain, there is the seeing of oneness in cognition and the rest.

Why we do not see the differences we do not name. (100): sarupanam ca sarvesam na bheda-upanipatinah vidyante vacakah sabdah — of all *like-formed* things there exist no *expressing words* that *fall upon their difference*: language has no vocabulary for the distinctions among things it groups; na api bhedo 'vadharyate — *nor is the difference ascertained*: and consequently we do not make them out. The order of that sentence is the doctrine: no words, therefore no ascertainment (3.1.47's unnamed scent-difference, real and unexpressed). (101) applies it reflexively: the visesah — particulars — that are the objects of jnana, sabda and artha — of *cognition*, *word*, and *meaning* — are duravadharatvat, *hard to ascertain*; hence jnanady-ekatva-darsanam: the *seeing of oneness* in cognition and the rest. We take our cognitions to be one kind of thing, our words to be one kind, because their particulars escape us. So the very categories in which this treatise is written are unities-by-default, produced by the same failure that produces every other universal.

3.1.102 unformalized

ज्ञानेष्वपि यथार्थेषु तथा सर्वेषु जातयः ।
संसर्गदर्शने सन्ति ताश्चार्थस्य प्रसाधिकाः ॥ १०२ ॥

jñāneṣvapi yathārtheṣu tathā sarveṣu jātayaḥ |
saṃsargadarśane santi tāścārthasya prasādhikāḥ || 102 ||

As among objects, so among all cognitions too there are universals; in the seeing of commingling they exist, and they are what accomplish the meaning.

The universals extended to the cognitive side. Jnanesu api yatha arthesu tatha sarvesu jatayah: as among *objects*, so among *all cognitions* too there are universals — cognition-types, kinds of knowing, as real (or as constructed) as kinds of thing. And their conditions: samsarga-darsane santi — in the *seeing of commingling* they *exist* (3.1.20's prapanca arising only among things commingled: difference and identity appear where relation has been made); tah ca arthasya prasadhikah — and they are what *accomplish the meaning*. So the cognition-universals are what make meaning possible: it is because knowings fall into kinds that a word can wake the same knowing twice (2.152's transmitted pratibha; 1.71's recurrence-datum on the hearer's side). The verse quietly completes 3.1.6's doctrine — the word's own universal superimposed on the object's — by supplying the third term: the cognition's universal, which is what the superimposition happens *in*.

3.1.103 unformalized

ज्ञेयस्थमेव सामान्यं ज्ञानानामुपकारकम् ।
न जातु ज्ञेयवज्ज्ञानं पररूपेण रूप्यते ॥ १०३ ॥

jñeyasthameva sāmānyaṃ jñānānāmupakārakam |
na jātu jñeyavajjñānaṃ pararūpeṇa rūpyate || 103 ||

The general residing in the knowable alone is what assists the cognitions; never is cognition, like the knowable, given form by an alien form.

A restriction on 3.1.102, and it is the samuddesa's turn to epistemology proper. Jneya-stham eva samanyam jnananam upakarakam: the general residing *in the knowable alone* is what *assists* the cognitions — cognition-types are conditioned by object-types; that this knowing is a cow-knowing is the cow's doing. But the converse is denied: na jatu jneyavat jnanam para-rupena rupyate — *never* is cognition, *like the knowable*, given form by an *alien* form. The knowable takes its form from another (the pot's form is the pot's); cognition does not — its form is its own, self-luminous, not conferred (1.50's cognition displaying its own form and the object's; the svaprakasa-doctrine the Prabhakaras would defend). So the assistance runs one way only: objects shape knowings, and knowings are not shaped from outside as objects are. The asymmetry is the ground of the next verse's simile and of the samuddesa's closing argument.

3.1.104 unformalized

यथा ज्योतिः प्रकाशेन नान्येनाभिप्रकाश्यते ।
ज्ञानाकारस्तथान्येन न ज्ञानेनोपगृह्यते ॥ १०४ ॥

yathā jyotiḥ prakāśena nānyenābhiprakāśyate |
jñānākārastathānyena na jñānenopagṛhyate || 104 ||

As a light is not illuminated by another illumination — so the form of cognition is not grasped by another cognition.

The simile that has served the work at every load-bearing joint (1.46, 1.55, 2.298, 2.397, 2.472), here in its epistemological office. Yatha jyotih prakasena na anyena abhiprakasyate: as a *light* is not illuminated *by another* illumination — one does not need a second lamp to see the first; light shows itself in showing. Jnana-akaras tatha anyena na jnanena upagrhyate: so the *form of cognition* is not *grasped by another* cognition. The doctrine is self-luminosity, svaprakasatva: awareness is not known by a further awareness (on pain of regress: what would know *that* one?), but in knowing its object it displays itself. This is 1.50's thesis (in a cognition, its own form and the object's are both seen) now defended by the regress-argument, and it is the anti-Nyaya position — for the Naiyayika, a cognition is known by a subsequent introspective cognition (anuvyavasaya), and the regress is stopped by fiat. Bhartrhari's lamp stops it by structure.

3.1.105 unformalized

घटज्ञानमिति ज्ञानं घटज्ञानविलक्षणम् ।
घट इत्यापि यज्ज्ञानं विषयोपनिपाति तत् ॥ १०५ ॥

ghaṭajñānamiti jñānaṃ ghaṭajñānavilakṣaṇam |
ghaṭa ityāpi yajjñānaṃ viṣayopanipāti tat || 105 ||

The cognition 'pot-cognition' is distinct from the pot-cognition; and the cognition that is 'pot' — that one falls upon its object.

The self-luminosity doctrine given its decisive analysis. Ghata-jnanam iti jnanam ghata-jnana-vilaksanam: the cognition *'pot-cognition'* — the reflective act by which I think about my seeing — is *distinct from* the pot-cognition itself. They are two, and the second is not the first inspecting itself; it is a new act with a new object (namely, the first). And: ghata iti api yaj jnanam visaya-upanipati tat — the cognition that is *'pot'* — that one *falls upon its object*: the original seeing is directed outward, at the pot, not inward at itself. So introspection does not deliver the original cognition; it replaces it with another about it. The analysis is exact and it is why 3.1.104's regress bites: if knowing a cognition requires a further cognition, and that one is distinct again, no cognition is ever known — unless each shows itself in showing its object, which is the lamp. The verse dismantles the introspective model with a distinction anyone can verify: thinking about your seeing is not seeing.

3.1.106 unformalized

यतो विषयरूपेण ज्ञानरूपं न गृह्यते ।
अर्थरूपविविक्तं च स्वरूपं नावधार्यते ॥ १०६ ॥

yato viṣayarūpeṇa jñānarūpaṃ na gṛhyate |
artharūpaviviktaṃ ca svarūpaṃ nāvadhāryate || 106 ||

Since the form of cognition is not grasped in the manner of an object — and its own form, separated from the object's form, is not ascertained.

The Jati-samuddeśa's last verse, and it ends on a limit rather than a thesis. Two negations. Yato visaya-rupena jnana-rupam na grhyate: since the *form of cognition* is *not grasped in the manner of an object* — cognition never stands before us as things do (3.1.105's analysis: try to catch it and you have another cognition). And: artha-rupa-viviktam ca svarupam na avadharyate — and its *own form*, *separated from* the object's form, is *not ascertained*: neither can we strip the object away and inspect awareness bare. So cognition is inaccessible from both directions: not as an object, and not in isolation from objects. What we have is always knowing-of-something, self-luminous and never separately given (1.124's transcendental argument: strip the word-form from awareness and the light would not shine). The samuddesa that began by asking what universals are ends by finding that the very cognition in which universals appear cannot be brought into view — which is 3.1.95's finding again (an indescribable nature at the end of everything, from which practice is not constructed), now reached on the subject's side. The Jati-samuddeśa closes as it opened: with an honest account of what its own instrument cannot reach.
contested: the verse is grammatically incomplete as transmitted (two subordinate clauses with no main clause), and readers supply the conclusion variously — 'therefore cognition is self-luminous', or 'therefore its universal is not ascertained'. We translate the fragment as it stands and leave the completion open