Deep Nominalization in Formal Turkish

English builds long thoughts out of strings of finite clauses: "The fact that it has not yet been understood why the project failed shows that…" Each "that" opens a new tensed clause with its own verb. Turkish, in its formal and academic register, builds the same thought a completely different way: it turns each clause into a noun, marks it for possessor and case, and then stuffs one inside the next like a set of nested boxes. The result is a single, towering noun phrase that says what English needs four finite clauses to say. Learning to assemble — and above all to parse — these nominalization stacks is one of the genuine dividing lines between an upper-intermediate reader and someone who can actually read a Turkish law, a scientific abstract, or a leading article. This page is about that skill.

The building blocks: -mA and -DIK as clause-nouns

Turkish has no all-purpose word like English "that." Instead it converts a whole clause into a noun using one of three nominalizers, and then case-marks the noun for its job in the larger sentence. Two of them carry most of the load in expository prose.

• -mA (the "action nominal," -ma / -me) names the bare action or fact, tenselessly: gel-me "(the) coming," anla-ma "(the) understanding."
• -DIK (the factive participle, -dık / -dik / -duk / -dük, with k→ğ softening before a vowel) names a realised fact, something that is or was the case: gel-diğ-i "(the fact) that he came / his coming."
• -(y)AcAK (-acak / -ecek) is the same machine pointed at the future/irrealis: gel-eceğ-i "(the fact) that he will come."

The crucial move is that each of these is a noun. A noun in Turkish takes a possessive suffix (marking the clause's subject) and then a case suffix (marking the clause's role). So anlaşılması is not a verb form to translate as "is understood" — it is a noun, "the being-understood-of-it," ready to be possessed, cased, and embedded.

In the second sentence, ertele-n-me-si is one word doing the work of an entire English clause "that the meeting was postponed": ertelen- "be postponed," -me action nominal, -si third-person possessive ("its"). The whole noun then sits as the subject of şaşırttı.

Subject becomes possessor, clause becomes case

Here is the mental flip an English speaker must make. In an English subordinate clause, the embedded subject stays a subject ("that the project failed"). In Turkish, the embedded subject becomes a genitive possessor, and the nominalized verb agrees with it via a possessive suffix. "That the project failed" is literally "the project's its-failing": projenin başarısız oluşu / olması.

In gelmediğini: gel-me-diğ-i-n-i = "come" + negative -me + factive -dik (softened to -diğ-) + 3rd-person possessive -i + buffer -n- + accusative -i. One word: "the-fact-of-your-not-coming," in the accusative because it is the object of fark etmedim. The subject "you" surfaces as the genitive senin. That is the entire grammar of clause embedding in Turkish, and once you internalise the order — stem → nominalizer → possessive → case — you can both build and unpick any single layer.

Stacking: clauses inside clauses inside clauses

Because a nominalized clause is just a noun, it can serve as the possessor of another nominalized clause, which can be the possessor of a third. This recursion is where formal Turkish gets its reputation for density. Each layer harmonises its own vowels, and the k→ğ softening reappears every time a -DIK meets a vowel-initial possessive — so the same shift recurs at each level of the stack.

Watch a two-layer phrase grow into a three-layer one:

• Layer 1: yanlış olması — "its being wrong" (yanlış ol-ma-sı).
• Layer 2: yanlış olduğunun anlaşılması — "the understanding of (the fact) that it is wrong" (yanlış ol-duğ-u-nun anla-şıl-ma-sı).
• Layer 3: yanlış olduğunun anlaşılmasının gerekmesi — "the necessity of (the fact) that it being wrong should be understood."

Now the required example — a triple nominalization fully parsed:

Parse it from the innermost box outward:

1. belgenin sahte olduğu — "(the fact) that the document is forged": belge-nin (genitive subject) + sahte ol-duğ-u (factive -DIK
   • 3rd-poss). This whole noun phrase is the inner box.
2. …olduğunun kanıtlanması — "the proving of [that the document is forged]": the box above goes into the genitive (olduğu-nun) to become the possessor of kanıt-lan-ma-sı (action nominal -mA
   • 3rd-poss, "its being-proved").
3. …kanıtlanmasının imkânsız olduğu — "(the fact) that [the proving of it] is impossible": box 2 goes genitive again (kanıtlanması-nın) and feeds imkânsız ol-duğ-u, another -DIK clause.
4. The outermost -DIK clause is then the bare complement of söyleniyor "is said."

Three genitive possessors (belgenin → olduğunun → kanıtlanmasının) chain into one another; each step the -DIK meets a possessive vowel and softens k→ğ (olduk → olduğu, kanıtlanmak → kanıtlanması). English unrolls this into "it is said to be impossible to prove that the document is forged" — four predicate ideas. Turkish welds them into a single subject noun phrase plus söyleniyor.

Why formal Turkish prefers this to finite clauses

This is not merely an alternative style; it is the default of careful written Turkish. The native finite-subordination strategy with the borrowed word ki (söyleniyor ki…) exists, but in neutral academic and legal prose it sounds conversational or old-fashioned. Nominalization keeps the verb at the end (Turkish is rigidly head-final), keeps the clause inside the matrix rather than dangling after it, and lets the writer pack background propositions into a single subject or object slot. Newspapers, judgments, and scientific abstracts lean on it precisely because it compresses argument structure.

Compare the everyday, finite-flavoured rendering a friend might speak — Baktık ki bulgular eski çalışmalarla uyuşmuyor, ilginç — and you feel the register gulf: the nominalized version is what a journal sentence looks like.

How English speakers get lost — and the fix

The error is almost never grammatical production; it is parsing collapse. Faced with olmadığının anlaşılamamış olmasının, the reader tries to translate left to right and drowns. Two habits fix it. First, read the stack from the inside out: locate the innermost verb stem and its nominalizer, treat everything attached after it (possessive, then case, then the next genitive) as wrapping, and resolve one box before moving outward. Second, let the genitives be your map: every -(n)In tells you "a clause-noun ends here and feeds the next one." Do not try to keep the whole stack in working memory as English clauses; unwrap it physically, box by box.

Common mistakes

Key takeaways

• Formal Turkish embeds clauses by nominalizing them with -mA / -DIK / -(y)AcAK, not with a finite "that."
• An embedded clause is a noun: its subject becomes a genitive possessor, the verb takes a possessive suffix, then a case suffix for its role in the larger sentence (slot order: stem → nominalizer → possessive → case).
• Because each nominalized clause is a noun, clauses stack: one becomes the genitive possessor of the next, three or more deep. The chain of genitive -(n)In suffixes is the skeleton of the stack.
• Each layer re-applies vowel harmony, and -DIK repeatedly softens k→ğ before the possessive vowel (olduk → olduğu).
• The English-speaker pitfall is parsing collapse; the cure is to read the stack inside-out, box by box, using the genitives as your map.