संस्कृतव्याकरणम् · 13 of 24 Home
अष्टाध्यायी १.३ · २.३ · ३.१ · ३.४ · प्रयोगप्रकरणम्

One event,
told तीन ways

Every finite Sanskrit sentence makes one choice that English never has to make explicit: which of the event's participants does the verb ending speak for? Not "active or passive" — that is the translation, not the grammar. कर्तरि · कर्मणि · भावे are three answers to one question, and page 08's single governing word — अनभिहिते — decides every case in the sentence once that one answer is chosen.

हन्
+
कर्तरि
→
हन्ति
agent expressed
धातुः — unchanged throughout प्रयोगः — which participant the ending speaks for विभक्तिः — moves as a consequence

Not a European "voice"

recap · 2.3.1

Page 08 (कारकम्) showed that a case ending appears only when the verb has not already claimed that role — . This page asks a question that sentence had not yet asked: which role does the verb claim? There are exactly three answers, and nothing else about the event changes when you pick a different one.

प्रयोगःThree claims a finite verb can make

A तिङन्त ending is always third person, always agrees with something in number — the only question is what. Pāṇini's grammar recognises three possible answers, and every other fact on this page is a consequence of which one a speaker picks:

कर्तरिthe ending speaks for the agent कर्मणिthe ending speaks for the object भावेthe ending speaks for neither — only the action itself
Textbook "active/passive" hides the third option. English only distinguishes two of these — "kills" and "is killed" — because English has no way to mark an action with no subject and no object at all. Sanskrit does: भावे. That absence is not a gap in the translation; it is a real grammatical category English simply lacks a word for.

राम, मृग, and हन् — three ways to say it

worked examples

One event — Rāma kills a deer — three sentences. Watch what stays fixed (the root, the meaning) and what moves (every case ending, and which noun the verb agrees with).

कर्तरिThe ending speaks for the agent

agrees with राम, so राम's role is already expressed — it takes प्रथमा, bare naming, by . The deer's role is unexpressed, so gives it द्वितीया.

कर्मणिThe ending speaks for the object

now agrees with मृग — the deer's role is what the ending expresses, so मृग takes the bare प्रथमा instead. राम's role is now unexpressed, and — the very sūtra that gives an instrument its case — gives राम तृतीया. This is why the "passive agent" and the instrument are never two different case-uses; they are one rule.

भावेThe ending speaks for neitherBRIEF's own examples

भावे has no room for a second participant on stage at all — it is what a grammar reaches for precisely when there is only one, so every real example of it is built on a root used with nothing to promote:

gives राम/तद् तृतीया — unexpressed, exactly as in कर्मणि above. There is no object anywhere in either sentence: हस् is intransitive and has none to offer, and गम् is used here with none in view. The ending (-type) agrees with nothing — not person, not number in the ordinary sense — it simply names that the action occurred.

The two case-endings on राम are the same sūtra. covers both कर्मणि's unexpressed agent and भावे's unexpressed agent, because from the ending's point of view they are the identical situation: a participant the verb has said nothing about.

One stem-sign marks both

3.1.67 · 1.3.13

हन्ति and हन्यते do not differ by accident. Watch where the two derivations split — and where they do not split at all.

यक्The affix that builds both कर्मणि and भावे stems

3.1.67 — "in the सार्वधातुक [sense], यक्" — is the single rule behind every कर्मणि or भावे तिङन्त in the language. It never distinguishes the two; nothing in its wording could, since यक् only ever sees the धातु, never whether a sentence around it happens to mention a कर्म.

Before यक् can even apply, though, one earlier choice has already been made: which set of personal endings the verb will use at all. 1.3.13 — "in भाव or कर्म [meaning]" — is what selects आत्मनेपद for both, against 's parasmaipada default for कर्तरि. That is the actual first fork in the derivation, one step before यक् is even reached.

Both derivations above are the engine's own output, read step by step. The कर्मणि and भावे chip is not shown twice with two different histories — there is only one history, because deriveTinantas() called with prayoga:"Karmani" and with prayoga:"Bhave" returns the same steps, in the same order, every time. The engine does not know which one a sentence is until a reader decides whether a कर्म is on stage — which is exactly this page's point, made by the derivation graph rather than by prose about it.

Watch the sentence rebuild itself

Interactive

Same event — राम, मृग, हन् — one dial. Nothing about what happened changes. Only which participant the verb chooses to speak for does, and every case ending in the sentence follows from that one choice alone.

प्रयोगः

The same logic, in the क्त participle

3.4.67 – 3.4.72

प्रयोग is a property of तिङन्त endings, and क्त — the "done" participle — is not one. Yet it carries the identical choice in a different dress, decided not by a parameter you pick but by whether the root is transitive at all.

निष्ठाक्त's default meaning, and the one exception

3.2.102 names the affix — क्त and क्तवतु together are called निष्ठा, by . The general rule for कृत् affixes is कर्तरि meaning, ; but निष्ठा is one of the affixes 3.4.69–70 pulls out of that default and reassigns: / — "of those [तिङ्], कर्मणि and भावे, [and] अकर्मक roots [only भावे]; of those same two, कृत्य, क्त and खल्" alike. So a transitive root's क्त means कर्मणि by default — the thing acted on, not the one who acted:

कृ, करणे — "the thing done" हन्, हिंसायाम् — "the one killed"

But two more sūtras immediately narrow that back. 3.4.71 and 3.4.72 hand क्त back to कर्तरि meaning for a named list of roots — motion verbs (गत्यर्थ) and, in the same breath, every अकर्मक (intransitive) root outright. There is nothing for a transitive-default को कर्मणि to attach to on a root with no object, so the rule simply says so:

गम्, गतौ — गत्यर्थ, अकर्मक — "the one who went"
Same शब्द-आकृति, opposite grammatical role. गतः looks exactly like कृतः — a root, plus , plus a प्रातिपदिक ending. Nothing in its shape marks which meaning it carries; only the transitivity of the root underneath decides, by rule, which of 3.4.70 or 3.4.72 applies. This is the same अनभिहिते logic as the तिङन्त dial above, wearing a participle's clothes instead of a verb ending's.

हन्, all three प्रयोग, every पुरुष and वचन

Reference · 27 cells

Every cell below came off the engine independently — 27 calls to deriveTinantas(), one per पुरुष-वचन-प्रयोग combination. Compare the कर्मणि and भावे columns cell by cell: they do not merely resemble each other, they are identical, letter for letter, in every one of the nine rows.

पुरुषः · वचनम्कर्तरिकर्मणिभावे

Check yourself

अभ्यासः

Six questions. Every one is answered by a sūtra, not by feel.