English glues clauses together with the conjunction "that": "I know that he came." Turkish has no such word in the standard construction. Instead it turns the entire embedded clause into a noun phrase — literally "his-coming I-know" — using the -DIK participle for realized events and the -(y)AcAK participle for future ones. Once you see that the "that"-clause is just a big noun, the whole system falls into place.
The core idea: a clause that acts like a noun
Take "I know that he came." Turkish encodes it as a single object noun phrase that istemek-style verbs swallow whole:
geldiğini biliyorum — "I know [his-having-come]."
Inside geldiğini there are four pieces stacked in order:
- gel- — the verb stem ("come")
- -diğ- — the -DIK nominalizer (k softened to ğ before a vowel)
- -i — the possessive, here 3sg, naming the embedded subject ("his")
- -ni — the accusative case (with buffer -n-), because the main verb bilmek wants an accusative object
So the embedded subject becomes a possessor, the embedded verb becomes a possessed nominal, and the whole package takes whatever case the main verb demands. There is no word for "that" anywhere.
Onun çoktan geldiğini biliyorum.
I know that he already came.
Yalan söylediğini hemen anladım.
I immediately realized that you were lying.
Dün seni aradığımı duymadın mı?
Didn't you hear that I called you yesterday?
-DIK for realized, -(y)AcAK for future
The choice of nominalizer carries the tense of the embedded clause:
- -DIK → the event is realized, factual, present, or past (relative to the main verb). "that he came," "that he is coming," "that he was here."
- -(y)AcAK → the event is future or not-yet-realized. "that he will come."
| Subject | -DIK (realized) | -(y)AcAK (future) |
|---|---|---|
| 1sg | geldiğimi | geleceğimi |
| 2sg | geldiğini | geleceğini |
| 3sg | geldiğini | geleceğini |
| 1pl | geldiğimizi | geleceğimizi |
| 2pl | geldiğinizi | geleceğinizi |
| 3pl | geldiklerini | geleceklerini |
Both nominalizers end in k, which softens to ğ before the vowel-initial possessive: -DIK → -diği-, -(y)AcAK → -eceği-. Watch the 2sg/3sg syncretism: geldiğini is both "that you came" and "that he came" — only context (or a spelled-out genitive subject) tells them apart.
Yarın yağmur yağacağını söylediler.
They said that it will rain tomorrow.
Onun bu sınavı geçeceğini hiç sanmıyorum.
I don't think at all that he'll pass this exam.
Bir daha geç kalmayacağına söz verdi.
He promised that he won't be late again.
The embedded subject is genitive — and usually dropped
When you do spell out the embedded subject, it takes the genitive case, agreeing with the possessive on the nominalized verb: onun geldiğini "that he came" (genitive onun ↔ 3sg possessive -i). But because the possessive already encodes the person, the genitive pronoun is normally dropped, exactly as subject pronouns are dropped in main clauses.
Senin haklı olduğunu sonunda kabul etti.
He finally admitted that you were right.
Bu işi benim bitirdiğimi kimse bilmiyor.
Nobody knows that I finished this job.
A full lexical subject (not a pronoun) likewise goes genitive: Ali'nin geldiğini duydum "I heard that Ali came."
Komşunun taşınacağını duydum.
I heard that the neighbor is going to move.
The case is dictated by the main verb
The nominalized clause is an argument of the main verb, so it takes whatever case that verb governs — usually accusative (for bilmek, duymak, düşünmek, anlamak, görmek), but the dative or another case when the verb demands it:
- bilmek / duymak / düşünmek → accusative: geldiğini biliyorum
- emin olmak ("be sure of") → dative: geleceğine eminim "I'm sure that he'll come"
- söz vermek ("promise") → dative: gelmeyeceğine söz verdi
Onun haklı olduğunu düşünüyorum.
I think that he is right.
Bu işi başaracağına eminim.
I'm sure that you'll pull this off.
Geleceğini söylediğini duydum.
I heard that you said you would come.
Why not just use "ki"?
Turkish does have a borrowed conjunction, ki, that can introduce a finite clause resembling English word order: Biliyorum ki o geldi. But this ki-clause is a Persian-derived, marked construction — heavier, more emphatic or literary, and not the default. Reaching for ki on every "that" is the single biggest tell of an English speaker who hasn't internalized nominalization. The unmarked, native way is the -DIK / -(y)AcAK nominal you've built here.
Geç kaldığını fark etmedim bile.
I didn't even notice that you were late.
A note for English speakers
The mental flip required here is large. In English the embedded clause stays a clause: subject in the nominative, verb finite, joined by "that." Turkish demotes the embedded clause to a noun phrase: the subject becomes a possessor (genitive), the verb becomes a possessed nominal, and the conjunction vanishes entirely. The same possessive-on-nominalization machinery you met in the -mA action nominal is at work — -DIK and -(y)AcAK simply add tense to it. If you can say gelmeni istiyorum "I want you to come," you are one step from geldiğini biliyorum "I know that you came": swap the verbal noun for the tensed participle and apply the case the new verb wants.
Common mistakes
❌ Biliyorum ki o geldi.
Not wrong, but marked/heavy as a default — the unmarked native pattern nominalizes instead of using ki.
✅ Onun geldiğini biliyorum.
I know that he came.
❌ geldikini biliyorum
Wrong — the k of -DIK must soften to ğ before the possessive: geldiğini.
✅ Geldiğini biliyorum.
I know that he came.
❌ O geleceğini umuyorum.
Wrong — the embedded subject must be genitive (onun), not nominative (o).
✅ Onun geleceğini umuyorum.
I hope that he'll come.
❌ Yarın geldiğini söylediler.
Wrong tense — a future event needs -(y)AcAK, not -DIK: geleceğini.
✅ Yarın geleceğini söylediler.
They said that he'll come tomorrow.
❌ Geleceğini eminim.
Wrong case — emin olmak governs the dative, not the accusative: geleceğine.
✅ Geleceğine eminim.
I'm sure that he'll come.
Key takeaways
- A "that"-clause becomes a noun phrase: embedded subject → genitive (often dropped), embedded verb → possessed nominal, no word for "that."
- -DIK = realized/factual; -(y)AcAK = future. The k softens to ğ before the possessive: geldiğini, geleceğini, olduğunu.
- The whole clause takes the case the main verb governs — accusative with bilmek / duymak / düşünmek, dative with emin olmak / söz vermek.
- Build it inside-out: stem + nominalizer + possessive + case; negation goes on the stem.
- The ki-clause exists but is marked and literary — don't default to it for every English "that."
Now practice Turkish
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Turkish→Related Topics
- The Object/Factive Participle -DIKB1 — How -DIK plus a possessive suffix relativizes objects and obliques (gördüğüm adam) and nominalizes past/non-future facts in complement clauses.
- The Future Participle -(y)AcAKB2 — How -(y)AcAK builds future-oriented relative clauses and complements with the same possessive-agreement machinery as -DIK.
- The Action Nominal -mAB1 — The -mA verbal noun and how its possessive suffix encodes a subject, enabling different-subject complement clauses like gelmeni istiyorum.
- ki-Clauses: Finite SubordinationB2 — The borrowed conjunction ki as a finite 'that' — Sanıyorum ki haklısın — its result and exclamative uses, and why native nominalization is preferred in neutral prose.