An affix that makes
a नया धातुः, not a new form

3.1.32 — सन् (the desiderative), णिच् (the causative), यङ् (the intensive) and the handful of others this page covers are grouped by tradition under one name, सनादि — "सन् and the rest" — and this sūtra says what they share: whatever they attach to, the result is a धातु. Not a form of the old one. A new one.

That single fact is why every other rule on this page reads the way it does. 3.1.68 कर्तरि शप् inserts शप् before a सार्वधातुक ending whenever the अङ्ग in front of it needs one — and once कृ+णिच् counts as a धातु, it is exactly the kind of thing 3.1.68 was written for, with no special clause required. The causative stem takes शप् (or whatever vikaraṇa its own shape calls for), takes every लकार, takes क्त्वा and तुमुन् and शतृ, can itself be turned desiderative — none of that needs its own rule. 3.1.32 already licensed all of it, in four words.

3.1.26 — continuing हेतुमति (from 3.1.25's own list, "of a causer") — णिच् attaches wherever the sentence's sense is हेतु, causation. Once it has attached, 7.2.115 fires: णिच् carries its own इत् letter ण्, which makes it णित्, and before a णित् or ञित् affix the aṅga's vowel gets not गुण but वृद्धि. That is the whole reason कृ becomes कार्- and not a mere "कर्-": the ण् of णिच् is doing that, not any property of कृ itself.

This is the part learners get wrong, because it looks like it should not be grammatical. Read पठ् (व्यक्तायां वाचि, "to recite/read aloud") on its own: शिष्यः ग्रन्थम् पठति, "the pupil reads the book" — शिष्य is कर्ता (प्रथमा), ग्रन्थ is कर्म (द्वितीया). Causativise it and the original agent stops being कर्ता at all.

1.4.52 is the rule doing this: for roots of motion, cognition, eating, or — as here — sound (शब्दकर्मकाणाम्, which पठ् falls under), when a lower causative is embedded under a further णिच्, the original agent (शिष्यः, the one who reads) is reassigned कर्मसंज्ञा. It does not stop being an agent in the real-world sense — it is still शिष्य who does the reading — but the grammar's कर्तृ slot is now occupied by रामः, the one who makes the reading happen, and 1.4.52 gives शिष्य's role to कर्म instead. ग्रन्थ, meanwhile, was already कर्म by the plain 1.4.49 कर्तुरीप्सिततमं कर्म, and stays कर्म. Two participants now both carry कर्मसंज्ञा, and 2.3.2 कर्मणि द्वितीया gives both of them द्वितीया — not because there is one rule for "double accusative", but because two separate, ordinary rules both point at द्वितीया for two different nouns in the same sentence. See page 08's कारकम् for कर्तृ/कर्म/अनभिहिते in general; this is the one construction where a sentence legitimately carries two of the same case for two different reasons.

3.1.7 — सन् attaches when the sense is इच्छा, "wanting", and the one who wants is the same as the one who would do the thing wanted. जि (जये, "to prosper/win") gives जिगीषति, "wants to win"; पा (पाने, "to drink") gives पिपासति, "wants to drink". Both reduplicate with इ, and both stay परस्मैपदी — the base root's own पद carries through unchanged.

3.1.22 attaches यङ् to a root of one vowel starting with a consonant, and reduplicates it heavily: पच् (पाके, "to cook") gives पापच्यते, "cooks again and again / cooks intensely"; भ्रम् (चलने, "to wander") gives बंभ्रम्यते, "wanders restlessly".

Two spellings of बंभ्रम्यते are both real — — an अनुस्वार/म् distinction in how the reduplicated nasal is written, not two different words.

यङ्लुक् — the same affix, dropped

2.4.74 — यङ् itself can be elided (लुक्) after it has done its reduplicating work, leaving only the heavy reduplication behind with ordinary तिङ् endings attached directly. The corpus's own examples for this affix (core/vidyut.json) list four spellings for two roots, which is itself the honest report: . This variability, more than any single form, is why यङ्लुक् is worth being able to recognise on a page rather than expecting to produce reliably.

3.1.8 — a सुबन्त (an inflected noun form), wanting it for oneself, takes क्यच्: पुत्रम् इच्छति आत्मनः → , "wants a son [for himself]". 3.1.9 — the same sense, an alternative affix: , "wants a son". 3.1.10 and 3.1.11 — from a standard of comparison, in the sense of behaving like it: , "behaves like a hawk"; , "behaves violently/intensely".

Traditional accounts of the सनादि also name णिङ् as a further denominative-forming affix. The corpus's own list — C.krt()['sanadi'], seven entries — does not include it: this build of vidyut-prakriya implements काम्यच्, क्यङ्, क्यच्, णिच्, यङ्, यङ्लुक् and सन्, and no eighth. That is a real gap in this build's coverage, not an oversight of this page's.