Agreement hierarchy overview

Hindi predicates often display gender and number, but the controller is not always the subject. The central generalization is a hierarchy: an eligible unmarked subject normally controls agreement; in a perfective transitive clause with a ने-marked subject, an eligible unmarked direct object may control it; if no eligible controller is available, the predicate uses default masculine singular morphology.

That hierarchy explains forms that seem irregular when translated word by word into English. English finite verbs show little gender agreement, so learners must learn to inspect Hindi case marking before choosing a participle, adjective, or auxiliary.

Start inside one clause

Agreement is local to a clause. First find the finite predicate and the noun phrases that belong to it. Then ask which arguments bear a postposition and which remain unmarked. Do not simply select the nearest noun or the real-world agent.

For the core patterns on this page, use this decision sequence:

StepStructural conditionTypical result
1An eligible unmarked subject is presentpredicate agrees with that subject
2Perfective transitive subject bears ने; direct object is eligible and unmarkedperfective predicate agrees with that object
3No eligible unmarked controller remainsdefault masculine singular

This is a learner's hierarchy, not a claim that every predicative construction has identical syntax. Copular adjectives, passives, dative-experiencer clauses, infinitives, and some complex predicates add construction-specific details. The hierarchy gives the right starting question: what is structurally eligible here?

First choice: an unmarked subject

In ordinary nonperfective clauses, the unmarked subject controls participial gender and number:

सीमा जल्दी उठती है।

sīmā jaldī uṭhtī hai.

Seema gets up early.

लड़के जल्दी उठते हैं।

laṛke jaldī uṭhte haiṃ.

The boys get up early.

The subject determines उठती versus उठते, and number determines है versus हैं. No object or nearby adverb competes for agreement.

Honorificity participates in the same system. Singular आप takes plural/honorific agreement:

आप कल कब आएँगे?

āp kal kab āẽge?

When will you come tomorrow? (respectful; addressed to a man or mixed/unspecified reference)

For a woman addressed respectfully, आप कब आएँगी? is standard. The plural form is grammatical honorific agreement, not a claim that several people are present.

The unmarked-subject page gives the full ordinary paradigm, including coordinated subjects and predicative adjectives.

Second choice: an unmarked object after ने

With a transitive verb in the perfective, an overt subject commonly takes ने. That postposition makes the subject unavailable for the ordinary agreement relation. If the direct object is unmarked and otherwise eligible, perfective morphology agrees with it:

सीमा ने रोटी खाई।

sīmā ne roṭī khāī.

Seema ate the flatbread.

सीमा ने आम खाया।

sīmā ne ām khāyā.

Seema ate the mango.

The agent is feminine सीमा in both sentences. The change from feminine खाई to masculine खाया tracks रोटी versus आम. Agreement is therefore evidence that the object—not the agent—is the controller.

Plural objects produce plural agreement:

सीमा ने चिट्ठियाँ लिखीं।

sīmā ne ciṭṭhiyā̃ likhīṃ.

Seema wrote letters.

The feminine plural perfective लिखीं agrees with unmarked चिट्ठियाँ. Notice the final nasalization on a standalone feminine plural perfective. This differs from a feminine plural attributive participle directly before a noun, whose morphology follows the nominal paradigm of that construction.

💡
In a perfective transitive clause, temporarily bracket off the ने-marked subject. Then inspect the direct object: if it is unmarked, its gender and number are the first agreement candidate.

Third choice: default masculine singular

If ने marks the perfective subject and को marks the direct object, neither controls the ordinary perfective agreement. The predicate takes default masculine singular:

सीमा ने बच्चों को देखा।

sīmā ne baccoṃ ko dekhā.

Seema saw the children.

Masculine singular देखा does not describe Seema or the plural children. It is the default form produced when the construction provides no eligible controller.

लड़कियों ने सीता को देखा।

laṛkiyoṃ ne sītā ko dekhā.

The girls saw Sita.

Both referential noun phrases are feminine, yet देखा is masculine singular. This controlled contrast proves that semantic gender does not override case licensing.

The same logic can apply when the object is a pronoun marked by को or a को-marked specific human. Object marking has discourse and semantic motivations, but once present it also changes agreement availability.

Dative experiencers require construction-level analysis

In clauses of experience, possession, necessity, and sensation, the experiencer often bears को. It does not control ordinary nominative agreement. Another unmarked nominal may control, or the predicate may be default depending on the construction:

मुझे ये किताबें पसंद हैं।

mujhe ye kitābeṃ pasand haiṃ.

I like these books.

The dative experiencer मुझे does not determine agreement. Plural हैं tracks the unmarked plural stimulus ये किताबें; invariant loan adjective पसंद itself does not show gender morphology.

मुझे ठंड लग रही है।

mujhe ṭhaṇḍ lag rahī hai.

I'm feeling cold.

Feminine रही aligns with unmarked feminine ठंड, not with the speaker. Translating मुझे as English subject “I” is semantically useful but grammatically insufficient.

Other dative predicates may lack an overt unmarked controller and surface with default masculine singular. Learn their case frames with the predicates themselves; do not force every English subject into the first step of the hierarchy.

Negation and focus do not reset agreement

Negation changes polarity but not the controller-selection algorithm:

शायद राधा आज नहीं आई।

śāyad rādhā āj nahīṃ āī.

Perhaps Radha didn't come today.

Unmarked feminine subject राधा still controls आई. Similarly, focusing an object with ही does not automatically make it an agreement controller; case and construction remain decisive.

सीमा ने रोटी ही खाई।

sīmā ne roṭī hī khāī.

Seema ate only the flatbread.

Here रोटी is both unmarked and focused, and it controls खाई. If it were रोटी को ही in a context licensing marked-object को, the focus would remain on the object but ordinary perfective object agreement would be blocked. Focus scope and agreement eligibility are separate properties.

Embedded clauses have separate controllers

The closest noun in the sentence may belong to another clause and be irrelevant to agreement:

अर्जुन ने कहा कि सीमा थकी हुई है।

arjun ne kahā ki sīmā thakī huī hai.

Arjun said that Seema was tired.

Matrix कहा is default masculine singular: अर्जुन ने is postposition-marked and the content clause does not function like an unmarked nominal object controlling its gender. Inside the कि-clause, feminine थकी हुई agrees with local unmarked subject सीमा. The feminine form does not reach backward or forward across the clause boundary.

Long-distance-looking patterns exist with selected infinitival and restructuring constructions, but they are tightly constrained; they do not license general agreement with any semantically associated noun. See long-distance agreement limits after mastering the local hierarchy.

💡
Agreement is a diagnostic, not a translation exercise. Clause boundary, aspect, transitivity, and overt postpositions must be established before gender and number are chosen.

Common Mistakes

1. Always agreeing with the agent

❌ सीमा ने गरम रोटी खाया।

Incorrect — ने blocks subject agreement, and unmarked feminine रोटी controls the perfective.

✅ सीमा ने गरम रोटी खाई।

sīmā ne garam roṭī khāī.

Seema ate the hot flatbread.

2. Agreeing with a को-marked object

❌ लड़कियों ने कल सीता को देखी।

Incorrect — को blocks ordinary object agreement, so the predicate is default masculine singular.

✅ लड़कियों ने कल सीता को देखा।

laṛkiyoṃ ne kal sītā ko dekhā.

The girls saw Sita yesterday.

3. Treating an English experiencer subject as Hindi nominative

❌ मैं ये नई किताबें पसंद हूँ।

Incorrect — पसंद uses a dative experiencer here, not an English-style nominative ‘I.’

✅ मुझे ये नई किताबें पसंद हैं।

mujhe ye naī kitābeṃ pasand haiṃ.

I like these new books.

4. Letting a matrix noun control an embedded predicate

❌ अर्जुन ने सबको बताया कि सीमा थका हुआ है।

Incorrect — embedded feminine subject सीमा requires थकी हुई.

✅ अर्जुन ने सबको बताया कि सीमा थकी हुई है।

arjun ne sabko batāyā ki sīmā thakī huī hai.

Arjun told everyone that Seema was tired.

Key Takeaways

  • Find the clause and its case frame before choosing agreement.
  • An eligible unmarked subject normally controls first.
  • In perfective transitives with a ने-marked subject, an eligible unmarked direct object can control.
  • If all potential controllers are blocked by postpositions, default masculine singular is common.
  • Honorificity, dative-experiencer syntax, and clause boundaries belong to the hierarchy; semantic intuition alone does not.

Now practice Hindi

Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.

Start learning Hindi →

Related Topics

  • Agreement with unmarked subjectsA2 — Form Hindi agreement with ordinary unmarked subjects across gender, number, aspect, honorificity, and coordination.
  • Agreement with ergative subjectsA2 — Track Hindi agreement when a perfective transitive subject bears ने, distinguishing syntactic subjecthood from agreement control.
  • Agreement with को-marked objectsA2 — Understand why a Hindi object marked with को cannot control agreement and how the remaining clause structure selects subject agreement or default form.
  • Default masculine singular agreementA2 — Recognize masculine singular as Hindi’s fallback agreement when case marking or clause type leaves the predicate without an eligible controller.
  • Split ergativity and perfective aspectB1 — Understand how Hindi changes subject case and agreement alignment in perfective transitive clauses while other clauses retain ordinary subject agreement.
  • Long-distance agreement limitsA2 — See when Hindi agreement can extend through a tightly integrated infinitival predicate and why finite clauses, case marking, and lexical structure block it.