षड् मार्गाः — six ways through one text

What the reordering buys

रामः

Derive one real word — रामः, "Rāma," nominative singular — and watch where its four sūtras live in Pāṇini's own text versus where a beginner's Kaumudī puts them.

In Pāṇini's own order, deriving रामः means opening the book near its very beginning for two saṃjñā rules, then jumping to its very end for two more — a span of sūtras, nearly the whole text, crossed twice. In the Laghu-siddhānta-kaumudī, all four sit inside its first steps — close enough together that a learner meets every rule a first derivation needs before meeting almost anything else. That is the entire case for reordering, made concrete.

Sūtra	अष्टाध्यायी	सिद्धान्तकौमुदी	प्रक्रियासर्वस्वम्	मध्यसिद्धान्तकौमुदी	लघुसिद्धान्तकौमुदी	सारसिद्धान्तकौमुदी

Position out of the route's own total, earliest first — "—" means that route never states the sūtra at all. Even प्रक्रियासर्वस्वम्, at 92% coverage the fullest of the five Kaumudīs, drops one of these four.

Six walks over one set of stones

C.route()

Every route below visits some subset of the same 3,983 sūtras — never a different sūtra, never a reworded one, only a different order and, for three of the five Kaumudīs, its own topical chaptering laid across that order.

Route	Devanāgarī	Sūtras reached	Coverage	Topical chapters

Two of the five have no chapters at all. The corpus's own chapters field is populated for सिद्धान्तकौमुदी, लघुसिद्धान्तकौमुदी and प्रक्रियासर्वस्वम् — 68, 60 and 16 named groupings respectively. मध्यसिद्धान्तकौमुदी and सारसिद्धान्तकौमुदी carry an order with no chapters at all: they are measured here as a flat resequencing, and this page does not invent topic names for them that the corpus does not have.

What the reordering breaks

अनुवृत्तिः

A sūtra that inherits a word by अनुवृत्ति — the subject of page A2 — depends on its reader having already met that word's source. Pāṇini's own order guarantees this: a source always precedes what depends on it. A Kaumudī's walk makes no such promise. Every sūtra the LSK includes was checked here against every word it inherits: has the LSK, at that exact point in its own walk, already stated the source?

षष्ठी स्थानेयोगा A metarule five later sūtras lean on, and never arrives

1.1.49 — "a genitive [ending, in a rule,] when connected [with what follows], denotes 'in place of.'" This one परिभाषा is what makes every replacement rule in the grammar parse: without it, a rule like "अलोऽन्त्यस्य" (1.1.52, "of the last sound") does not say replace anything — it is 1.1.49 that supplies "in place of."

The LSK never states it. Not late, not in a different chapter — absent from its 1,269, and from मध्यसिद्धान्तकौमुदी and सारसिद्धान्तकौमुदी too. Five sūtras the LSK does include still inherit "षष्ठी" and "स्थाने" from it by अनुवृत्ति:

Position 20 of the LSK's own walk — the strip below is steps 14 to 29, in the order the LSK actually presents them. The one outlined in red is अलोऽन्त्यस्य, silently leaning on a rule two chapters away in Pāṇini's text that this route never states.

Nobody flags the gap. 1.1.52 reads, in the LSK, as a complete sentence — "of the last sound." A reader who has only ever seen the LSK has no way to know it is silently short one word, because the word never had anywhere to come from in this route at all.

Three more, spread across the LSK's whole walk — one relies on a source that never arrives, one on a source that arrives 81 steps after the sūtra that needs it:

LSK step	Sūtra	Needs (अनुवृत्ति)	Source	Source status

The 68% the LSK omits

C.index_rows()

2,714 of the 3,983 sūtras never appear in the LSK at all. Not evenly: what survives depends heavily on what kind of statement a sūtra is, and where in the Aṣṭādhyāyī it sits.

Kind	Total	In the LSK	Kept

परिभाषा (metarules like 1.1.49 above) and अधिकार (governing headings, page A3's subject) are kept at the highest rate of the five kinds — and still lose more than half. विधि — ordinary operative rules, 91% of the whole grammar — is what most of the 68% actually is: 2,552 of the 2,714 dropped sūtras are विधि.

अध्यायः	Total	In the LSK	Kept

Adhyāyas 1 and 7 — saṃjñā-heavy and thick with अङ्ग rules a first derivation cannot skip — survive at roughly double the rate of adhyāyas 2 through 5, where most of the grammar's sheer bulk of विधि lives. कारके (1.4.23) and अङ्गस्य (6.4.1) — page A3's two largest अधिकार domains — are themselves among the sūtras the LSK never states.

One sūtra, six positions

Interactive

Type any reference — 6.4.1, 1.4.23, your own — or pick one below. See where it falls in every route, or that a route skips it entirely.

The opening stretch

Interactive

The first 120 steps of a route, plotted by which adhyāya each step comes from. Pāṇini's own order can only climb — the first 120 sūtras of the अष्टाध्यायी itself never leave adhyāya 1. Watch what a Kaumudī does instead.

compare to Pāṇini's own first 120

Hover or tap the chart

Where these numbers come from

Source

Everything on this page is measured, not summarised. C.route(name) gives each Kaumudī's own order — the ids it visits, in the sequence it visits them — and chapters where it groups them topically. The break count walks every LSK sūtra's C.sutra_rec(ref).anuvritti edges against the LSK's own position map; the 68% table walks C.index_rows() against the LSK's id set. The रामः derivation is pulled unmodified from the site's own prakriyā trace (Grammar/site/data/shabda/prakriya/), the same engine every other page in this series draws its forms from.

The picker above supports any of the 3,983 references for position lookup — the full order array for all six routes ships with this page — but shows sūtra text only for the ~700 sūtras this page happens to cite or open with: the first 120 steps of each route, plus a handful named elsewhere on the page. A reference outside that set still resolves correctly; it links to the full reading view instead of showing text inline, and says so rather than showing nothing.

Check yourself

अभ्यासः

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