Turkish has two different suffixes for relative clauses. The difference is due to the subjectivity in relative clause. In other words; if it defines a subjective, it is constructed with the suffix -En; if not, it is constructed with -DIK/AcAK. (-DIK and -AcAk make time difference in the structure and they cooccur in other structures such as nominalizer.)

  (1) Kitap okuyan kızı gördüm. 
      book  read-part. girl-acc. see-past-1stp.
      "I saw the girl who read a book."

  (2) Kızın okuduğu kitabı gördüm.
      girl-gen. read-part. book-poss. see-past-1stp.
      "I saw the book which the girl is reading."

  (3) Kızın okuyacağı kitabı gördüm.
      girl-gent. read-part. book-poss. see-past-1stp.
      "I saw the book which the girl will read."

In these sentences, the change in the suffixes create ungrammaticality. However, there are also structures allow the both structure even though it should be nonsubjective suffix.

  (4) Bacasından dumanın çıktığı evi gördüm.
      chimney-poss.-abl. smog-gen. came out-part. house-acc. see-past-1stp.
      "I saw the house from whose chimney the smog came out."

  (5) Bacasından duman çıkan evi gördüm.
      chimney-gen.-abl. smog come out-part. house-acc. see-past-1stp.
      "I saw the house from whose chimney the smog came out."

It does not actually define the subject of relative clause but it can form by both suffixes. Furthermore, when I asked it to the native speakers, most of them pointed that the structure with subjective suffix sounds more natural. Do you have any suggestion or reference why or how it can be possible?

  • I haven't studied much Turkish, but on the basis of your examples, the obvious difference it that çıkmek is intransitive. In the case of okumak, the different suffixes distinguish whether the subject or object is relativised; but with çıkmek there is no object to be distinguished. Without more examples, I've no idea whether this is really the difference,
    – Colin Fine
    Nov 2, 2013 at 23:37
  • I am not a linguist, but your examples here are difficult to sort out because there are other issues. For example 5 is almost nonsensical, but could be rendered with "çıkaran" instead. Also the primary difference between the -en and -dik- suffixes is not actually subjectivity. I think that is what is throwing you off.
    – Caleb
    Nov 6, 2013 at 19:36
  • @Caleb, 5 means "there is a house and the smog came out from its chimney." When you change it with "çıkaran", it sounds as if "the chimney makes the smog came out." and it is totally different. Nov 6, 2013 at 21:44

4 Answers 4


There are two things going on here.

As it is correctly stated in the question, the primary use case of the -(y)En suffix is to modify the head noun when it's the subject of the embedded clause. Let's call it the use case A.

    Çocuğu uyandırma.
    Don't wake up the child.

    Çocuk odada uyuyor.
    The child is sleeping in the room.

    Evde uyuyan çocuğu uyandırma.
    Don't wake up the child who's sleeping in the room.

The first thing that is going on is that -(y)En is also used when the head noun is the the modifier of a possessive construction functioning as the subject of the embedded clause. Since it's an extension of the use case A, let's call it the use case A'.

    Çocuğun odası yeni boyandı.
    The child's room was recently painted.

    Odası yeni boyanan çocuğu uyandırma.
    Don't wake up the child whose room was recently painted.

The second thing that is going on is that sometimes -(y)En can also be used when the subject of the embedded clause has low definiteness and animacy and the head noun is in dative -(y)E, locative -DE or ablative -DEN with a directional or locative function (not when they function as indirect objects). In rare cases, it can also be used when the head noun is the direct object in accusative case -(y)İ. Let's label all these use cases the use case B.

    Odaya güneş giriyor -> Güneş giren oda
    Çekmecede para duruyor -> Para duran çekmece
    Caddeden tramvay geçiyor -> Tramvay geçen cadde
    Kızı arı soktu -> Arı sokan kız

The example in the original question is what we can label as the use case B'. It happens when the head noun modifies an argument that would be eligible for the use case B in a possessive construction.


In Turkish, some subject + verb groups tend to form a close association, a tendency to stay together and not to allow other words to appear in between. It appears to be motivated by the low animacy of the subject. It is mildly similar to split ergativity. For example:

Adamı köpek ısırdı.
man+definite-accusative dog bite+past-tense

Here, the subject + verb group "köpek ısır-" shows a strong tendency to remain together at the expense of natural SOV word order. Because the word "adam" (man) has higher animacy than the word "köpek" (dog), the topic of the sentence shifts from the subject to the direct object. The subject + verb group "duman çık-" shows a similar tendency. Sometimes these rigid groups may accept a subjective "-en" participle instead of the expected objective.

The actual rules are way more complicated (and possibly not entirely analyzed) but there is a chapter devoted to the subject in this book.

  • As a native speaker of Turkish, "Köpek adamı ısırdı." is totally natural to me. Furthermore, I prefer two sentences in different contexts. The referentiality and defineteness change in these sentences. 1st one is nondefinite and nonreferential while the second one is definite and referential for "köpek". Nov 6, 2013 at 21:50
  • @SerpilKarabüklü Yes, it's perfectly natural. My point was that inanimate indefinite subjects tend to form a rigid group with the verb and it motivates the usage of "-en" for secondary arguments if those have the main focus. "Bacasından dumanın çıktığı ev" breaks the indefiniteness of the original subject "duman" and makes it definite. The only neutral choice is "bacasından duman çıkan ev".
    – cyco130
    Nov 10, 2013 at 15:49
  • In the said book, the authors base their explanation on definiteness but give a couple of examples where animacy does matter: "İçine çocuk düşen havuz ..." is ok but "içine çocuk atlayan havuz ..." is not. Because the original subject "çocuk" has more animacy in the second example.
    – cyco130
    Nov 10, 2013 at 16:04
  • @SerpilKarabüklü Please see my second answer because I've noticed that there is more going on in your example.
    – cyco130
    Nov 11, 2013 at 16:28

Relative clauses introduced by some sort of agent specifically two different ones hark to two different traditions that were dialectical.

  • Welcome to Linguistics SE! Currently I can't really see how your contribution answers the question. Do you think you could expand it?
    – robert
    May 11, 2014 at 20:47

While the suffixes may be translated with a relative clause, they are participles.

I think the difference is not only the subjectivity but also the concurrency.

My translations:

Present participle:

(1) I saw the girl, who was reading a book

  • At the moment i saw her, she was reading

(5) I saw the house, from whose chimney the smog was coming out.

Past participle:

(2) I saw the book, which the girl had read (some time before).

(6) I saw the house from whose chimney the smog came out (some time before)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.