अधिकारः — the seventy-three governing domains

A rule that names nothing

Orientation

Read 5.1.19 cold: . Four compounded conditions and one affix, and not one word tells you what kind of thing is getting the affix, or what comes before it, or that this is even about word-formation at all. Printed on its own, it is close to unreadable.

It is not unreadable to Pāṇini's reader, because nobody reads 5.1.19 on its own. By the time you reach it you are already standing inside eight headings that opened earlier and have not yet closed — each contributing one silent word to every sūtra underneath it, this one included. Strip those away and read the sūtra bare, as above: a fragment. Restore them, in order, and it resolves into a full sentence. That restoration is the whole subject of this page, and the last section below walks 5.1.19 through all nine layers, one at a time.

Page 08 already showed you one of these. कारके at 1.4.23 is an अधिकार — every kāraka definition from 1.4.23 to 1.4.55 depends on it without saying so. That page taught what the six roles mean; this one asks a different question about the same fact: how does a sūtra that says nothing about roles end up defining one? — and how far can that stack before a single terse rule is standing on the shoulders of eight others.

What a heading actually does

6.4.1 · 1.1.1

An अधिकार is a sūtra with two properties an ordinary rule does not have: it supplies a word to every sūtra in a stated range, and it says — in the tradition's commentary, if not always in its own eight syllables — exactly where that range ends.

अङ्गस्यOne word, six hundred and thirteen sūtras

6.4.1 — two syllables, genitive case: "of the stem." Nothing else. It opens the largest domain in the grammar, running to 7.4.97, and the Kāśikā says so in words that recur, near enough verbatim, at every heading in this book:

"This is an अधिकार, up to the end of the seventh adhyāya. Whatever we enumerate from here onward is to be understood together with 'of the stem.'" — the formula is यदित ऊर्ध्वम् अनुक्रमिष्यामः … तद् वेदितव्यम्, "whatever we will enumerate hereafter, that is to be so understood," and it is how the tradition itself flags a heading. You will meet it again below, word for word, at 1.4.1 and at 8.2.1.

Two things travel with the word: a slot, and a stop. The slot is which argument of the rule the word fills — अङ्गस्य is a genitive, so it attaches to whatever the rule is talking about and says of the stem. The stop is the sūtra id the domain's own record names as its close — 7.4.97 for this one. A rule with only the first property and not the second is not an अधिकार; it is अनुवृत्ति, and the next section is about exactly that difference.

Not the same thing as अनुवृत्ति

1.4.23 · 1.4.24

अनुवृत्ति — the word a sūtra does not say, carried forward from one nearby into the next by sense alone — is its own page's subject (अनुवृत्तिः, A2). Here is the distinction from this page's side: an अधिकार has a stated range; an inherited word by अनुवृत्ति simply runs until the sense stops fitting, and no sūtra ever announces the point where it stopped.

कारकेOne heading, one word that lapses beside it

1.4.23 opens a domain with a stated close: its own record gives 1.4.55 as the end, thirty-three sūtras away. Every one of them — from the कर्ता at 1.4.54 back to here — is a कारक definition because this word says so, and will go on being one all the way to 1.4.55 whether or not any single sūtra in between needs reminding.

The very next sūtra behaves completely differently. 1.4.24 defines अपादान — "the fixed point something moves away from" — and states no domain of its own. Its word, अपादानम्, simply keeps being understood in the sūtras after it, by sense, for as long as the topic holds:

Highlighted: अपादानम् is still understood at that sūtra, by अनुवृत्ति, though no sūtra says so. Grey: it no longer is — because 1.4.32 changes the subject to संप्रदान, not because any rule retired अपादानम्.

Nobody closes it. कारके's range ends at a sūtra that says so — the domain's own record names 1.4.55. अपादानम्'s run ends at 1.4.32 because 1.4.32 is busy defining something else, and 1.4.32 itself is still inside कारके (its own domain record shows the same two headings 1.4.23 and 1.4.1 still open) even though अपादानम् has already lapsed beside it. Two different mechanisms, stacked on the same sūtra, stopping at two different points for two different reasons.

Unwind a stack, layer by layer

Interactive

Three sūtras, three depths. Step through each and watch a fragment turn into a sentence, one silently-supplied word at a time — outermost heading first, exactly the order a reader meets them coming down the text.

Two headings of a different kind

6.4.22 · 8.2.1

Every domain so far hands down a word — a genitive, a locative, something that slots into the rule underneath it. These two hand down something else: a mode, an instruction about how rules see each other. Both are अधिकार by the same mechanism — a heading sūtra, a stated range — but neither fills a syntactic slot in anything.

आभीयःThe ābhīya zone — rules blind to their neighbours

6.4.22 — "here, up to भ, [rules are] as if not yet applied." For the sūtras from 6.4.22 to 6.4.129, each rule is evaluated as though the others inside the same zone had not yet fired — not invisible to what came before the zone, only to each other, and only inside it.

Worked example — a famous rule, doubly governed. 6.4.24 — for a root not marked अनिट्, elide the penultimate न् before a क्/ङ्-marked affix (नम्→नत, यम्→यत). Read cold it says nothing about roots, stems, or simultaneity. It inherits all three: अङ्गस्य (this is about a stem), and असिद्धवत् (evaluate it as if the other ābhīya rules had not yet applied — which matters, because whether the न् really is penultimate can depend on what those other rules did).

The corpus draws this boundary twice, and they disagree. The zone — where rules are mutually blind — is stated as 6.4.22–6.4.129, sūtras. But the same heading's word, असिद्धवत् itself, is still being cited as far as — sūtras in all. The mutual-blindness rule and the word that names it do not close at the same place. This page shows both numbers rather than quietly picking one.

पूर्वत्रासिद्धम्The tripādī — the last 295 sūtras play by different rules

8.2.1 — "asiddha with respect to what precedes." From here to the very end of the text, 8.4.68, every rule is असिद्ध — invisible to, and unable to see the results of, every rule before this point and every rule after it within these three pādas. You are 3,688 sūtras into a 3,983-sūत्र text when this heading opens, and from here on the physics changes for the last 295.

Why this matters is its own page. Getting order of application right inside the त्रिपादी — which rule can see which — is most of what makes derivations hard, and it is asiddhatva's own subject (असिद्धत्वम्, A4). Here it only needs to be placed on the map: it is the widest of the 73, and unlike the other 72 it never fully closes — its word is a mode that lasts to the last sūtra Pāṇini wrote.

All 3,983, and what governs each one

Interactive

Every sūtra in Pāṇini's order, left to right. The 73 domains run beneath as stacked lanes — nine of them, because nine is the most that are ever open at once. Move over the strip, or tap it on a touch screen, to see exactly where you are: which headings are open, what each one is silently supplying, and how deep you are nested.

—

The measurement, in full

73 domains

Word	Opens	Closes	Sūtras governed	Lane

Where these numbers come from. C.architecture() gives the 73 spans, the 32 pādas, the tripādī and ābhīya boundaries and the canonical order of all 3,983 sūtras. Nesting depth, the longest and shortest domains and the count of ungoverned sūtras are computed on this page by walking that order and counting span coverage at every position — not estimated, and cross-checked against the corpus's own maxDepth figure at build time. The literal word each domain supplies is pulled with C.sutra(ref) from its opening sūत्र, not typed from the descriptive name the span carries. C.sutra_rec(ref) gives the hand-tagged adhikāra stack and anuvṛत्ति edges used in the worked examples above, and the one place on this page it disagrees with the measured stack (5.1.19, in the nesting player) is reported rather than silently reconciled.

Check yourself

अभ्यासः

Six questions. Pick an answer to see which fact decides it.