Two Basic Principles of Turkish Syntax and Other General Properties (As Compared to English)
Updated: Sep 8
The two basic principles of Turkish syntax are:
Some general properties of the Turkish language, as opposed to the English language, are provided in the table below:
Head-Initial (Complements) / Head-Medial (Modifiers)
Canonical Word Order
Subject–Object–VERB (syntactically flexible, pragmatically fixed)
Subject–VERB–Object (syntactically fixed, pragmatically depending on intonation)
Asyndetic Parataxis preferred
Syndetic Parataxis preferred
Nonfinite (Nominalization-Based Verbals and Postpositionals) & Finite
Finite & Nonfinite (Verbals)
Style of Describing
Left-Branching (+ Mid-Branching)
Right-Branching (+ Left- and Mid-Branching)
Subject vs. Topic
Determined by position and case marking
Determined by article: a(n), the, zero article
Determined by case marking
Determined by position within a sentence
Below, I'll explain in more detail some aspects of these principles and properties.
I. The Principle of Preceding Qualification
The most prominent characteristic of the Turkish language is that, syntactically, it follows the so-called rectum-before-regens, or dependent-before-head, principle of phrasal and clausal structuring. In Turkish syntaxis, any qualifier precedes the qualified element, the secondary element is placed before the principal constituent, and all the words that complete the sense of (depend on) another word are placed before it, their head. This makes Turkish a head-final language.
Another term for the [dependent+head] relationship is phrase, or rather a dependent phrase, as opposed to a coordinated phrase, which is based on an independent relationship. A phrase connected through an independent relationship (meaning that it does not have any heads or dependent elements) consists of equal, parallel items connected as a series or a compound construction.
The head plays a central role in defining the phrase: The grammatical category of the phrase depends on that of the head. If the most important part of the phrase, its head, is an adjective, the phrase is an adjective phrase; if the most important part of the phrase is a noun, the phrase is a noun phrase, etc. Clearly delineating the phrases helps make the structure of the sentence clearer and less ambiguous.
There are two kinds of dependents:
1. Complements (obligatory)
2. Modifiers (optional), including determiners (articles, demonstratives, possessives, quantifiers)
While complements complete the meaning of their heads (and, as such, usually cannot be omitted), modifiers optionally modify their heads and, therefore, can be dispensed with.
In both Turkish and English, the common phrases with dependent complements include:
Verb phrases with direct and indirect objects, complement that-clauses, subject and object predicatives
Noun phrases with nominal complement that-clauses
Adjective phrases with adjectival complements (to-infinitival clauses)
Pre-/post-positional phrases with prepositional/postpositional complements
English: Head-Initial (with Complements) & Head-Medial (with Modifiers)
In English, complements always come after their heads, which makes English head-initial (with complements). English modifiers, however, may come before or after their heads, which makes English head-medial (with modifiers).
As a head-initial language (for complements) and a head-medial language (for modifiers), English words can have both pre- and post-dependent modifiers and post-dependent complements. The table below shows all kinds of phrases, with the heads highlighted in bold and their dependents identified in the brackets [...]:
HEAD OF PHRASE
if you can, call
call when you can
say that you are OK
[that-clause as an object]
(the) people present
(the) reason [that] he left
(the) then director
(the) people that we met
glad to know
tall as me
guilty of a crime
glad about this
glad that you could came
Turkish: Strictly Head-Final
In the head-final Turkish, in its basic syntactic units—phrases—the dependent element (or a complement) must come before its heads.
For example, compare the Turkish and English phrases with complements, with their heads shown in bold:
TR: öğrenciydik [noun (subject complement) + verb (copula)] ⟹ a verb phrase
EN: (we) were students [verb (copula) + noun (subject complement)] ⟹ a verb phrase
TR: çikolata yedi [obj.+ verb] ⟹ a verb phrase
EN: (she) ate chocolate [verb + obj.] ⟹ a verb phrase
TR: mutlu etti [adj. + verb] ⟹ a verb phrase
EN: made happy [verb + adj.] ⟹ a verb phrase
TR: işinden memnun [postposit. phrase + adj.] ⟹ an adjective phrase
EN: happy about the work [adj. + preposit. phrase] ⟹ an adjective phrase
TR: okulda [noun + locative case suffix] ⟹ a postpositional phrase
EN: at school [prep. + noun] ⟹ a prepositional phrase
TR: memnuniyet ile [noun + postposit.] ⟹ a postpositional phrase
EN: with pleasure [preposit. + noun] ⟹ a prepositional phrase
TR: sağdan ikinci [postposit. + numeral] ⟹ a postpositional phrase
EN: (the) second on the right [preposit. + preposit. phrase] ⟹ a prepositional phrase
As for modifiers, Turkish modifiers, again, come before their heads, whereas English modifiers come before and after the words they depend on:
TR: yavaşça konuş [adverb + verb] ⟹ a verb phrase
EN: ate chocolate [verb + adverb] ⟹ a verb phrase
TR: güzel okudu [adverb + verb] ⟹ a verb phrase
EN: read beautifully [verb + adverb] ⟹ a verb phrase
TR: bugün bitirdik [noun+ verb] ⟹ a verb phrase
EN: (we) finished today [verb + noun] ⟹ a verb phrase
TR: ilk olarak söyledi [converb + verb] ⟹ a verb phrase
EN: said first [verb + numeral (adverbial)] ⟹ a verb phrase
TR: belki görür [verb + verb] ⟹ a verb phrase
EN: maybe see [modal adverb + verb] ⟹ a verb phrase
TR: masadaki bardak [relativized noun + noun] ⟹ a noun phrase
EN: (the) glass on the table [noun + preposit. phrase] ⟹ a noun phrase
TR: naylon torba [qualifying noun + noun] ⟹ a noun phrase
EN: nylon bag [qualifying noun + noun] ⟹ a noun phrase
TR: yeşil kitap [adj. + noun] ⟹ a noun phrase
EN: green book [adj. + noun] ⟹ a noun phrase
TR: hep gülen yüz [participle + noun] ⟹ a noun phrase
EN: (the) face that always smiles [noun + relative clause] ⟹ a noun phrase
TR: annemin gelmesi [genitive noun + noun] ⟹ a noun phrase
EN: mother's arrival [genitive noun + noun] ⟹ a noun phrase
TR: senden daha yavaş [adjective + comparative clause] ⟹ an adjective phrase
EN: slower than you [adjective + comparative clause] ⟹ an adjective phrase
TR: beklediğimiz gibi meşgul [adjective + postpost. clause] ⟹ an adjective phrase
EN: busy as we have expected [adjective + comparison clause] ⟹ an adjective phrase
TR: çok yavaşça [adverb + adverb e] ⟹ an adverb phrase
EN: very slowly [adverb + adverb] ⟹ an adverb phrase
TR: çamur içinde [noun + postposition] ⟹ a postpositional phrase
EN: in the mud [preposit. phrase] ⟹ a prepositional phrase
TR: her ikisi [determiner + numeral] ⟹ a numeral phrase
EN: both of them [determiner + preposit. phrase] ⟹ a numeral phrase
While Turkish is a head-final language with both complements and modifiers, English is head-initial with complements but head-medial with modifiers.
Turkish Independent Phrases
In Turkish, like in English, the basic syntactic unit is a phrase, which can be presented as having dependent or independent relationships.
In a phrase with independent relationship, the elements are syntactically equal, or parallel to each other, which makes them a pair or a series of parallel items, or serial items. For example, a series of items, including subjects, objects, complements, modifiers, or verbals (ben ve sen; araba ve tren ile; sevimli, akıllı; etc.) are equal and parallel.
As in English, Turkish series can be syndetic or coordinated (linked with a coordinating conjunction), asyndetic or juxtaposed (without any conjunctions linking them), or polysyndetic (linked with multiple conjunctions).
Turkish Dependent Phrases
In a phrase with dependent relationship, the central element is the head, while the other one is its dependent. The head and its dependent must be in agreement with each other, with the head containing the key information about its dependent.
In addition to the obligatory complements, optional dependent elements, e.g., modifiers and adverbials, can also exhibit the head vs. dependent relationship.
TR: onunki masa [relativized noun phrase + noun] ⟹ a noun phrase
EN: (the) table of his [noun + preposit. phrase] ⟹ a noun phrase
TR: okulda kal [postposit. phrase + verb] ⟹ a verb phrase
EN: stay at the school [verb + preposit. phrase] ⟹ a verb phrase
Turkish Phrases = Phrases Within Phrases
As we can see, in both Turkish and English, phrases are recursive, meaning that phrases can be made of other phrases, which, too, stand either as heads vs. dependents or as independently linked items:
TR: [[dün ve bugün] [benimle [koşan adam]]] ⟹ a noun phrase
EN: [[[(the) man who ran] with me] [yesterday and today]] ⟹ a noun phrase
dün ve bugün
dependent _ head
dün ve bugün koşan
dün ve bugün benimle koşan
dün ve bugün benimle koşan adam
If we add another phrase, pek konuşkandır, another verb phrase that contains an adjective phrase (pek konuşkan) and the copular -dır, we get a complex sentence (birleşik cümle) with a verbal subordinate clause (girişik tümce) as its subject:
[Dün ve bugün benimle koşan adam] [pek konuşkandır].
The man who ran with me yesterday and today is very talkative.
Since phrases constitute the clausal elements that can function as subjects, predicates, complements (objects), adverbials, etc., in sentences, a basic sentence formula can be expressed as a sum of two phrases:
Simple Sentence = Subject + Predicate = Subject [Noun Phrase] + Predicate [Verb Phrase]
Subject + Predicate
Subject [Noun Phrase] + Predicate [Verb Phrase]
Turkish Subject–Verb Agreement
In any Turkish and English sentence, the subject and predicate exhibit agreement: if the subject changes, the predicate changes accordingly, which means that the subject modifies the predicate.
Ayşe kitabı okuyor.
Ayşe is reading the book.
Ayşe’yle Ali kitabı okuyorlar.
Ayşe and Ali are reading the book.
Oğlum her aksam okur.
My son reads every evening.
As illustrated by these examples, the subject and the predicate form a phrase, in which the subject is the modifier, and the predicate is the modified, which makes the predicate the head and the subject the dependent of the predicate (verb) phrase.
Both Turkish and English predicates agree with the respective subjects: in number and person in Turkish and in number and partially in person in English.
In Turkish, this agreement is expressed through suffixes affixed to the verb stem, while in English, the agreement is maintained through the use of suffixes as well as auxiliary verbs.
Turkish Subject vs. Predicate = Dependent vs. Head
This subject vs. predicate relationship aligns with the dependent vs. head relationship, in which the dependent modifies the head. This means that we can reduce the sentence formula down to this:
Sentence = Dependent [Noun Phrase (Subject)] + Head [Verb Phrase (Predicate)] = Verb Phrase [Subject + Predicate]
Verb Phrase [Subject + Predicate]
This is how any basic sentence, whether Turkish or English, can be interpreted:
The head of a sentence is usually taken to be the tensed verb, and every other word is either dependent of the sentence head (the subject, complements, adverbials), or connects to it through a path of other dependencies.
Although both subject and object are dependent on the main verb, they are not dependent in the same way. Ranked higher than any other complements, the subject has a special, sentence-forming relationship with the predicate. Linguists refer to such difference as subject-object asymmetry, exhibited by most languages.
II. The Principle of Verb-Final Word Order (SOV)
As I showed before, in a simple English and Turkish sentence, the verb functions as the head with two dependents: the subject and the object. The ordering of these core constituents, however, in a sentence depends on the head directionality of the language.
Head-Directionality & Word Order
Most languages can be classified in terms of the head directionality, which aligns with the arrangement of the basic (canonical) elements of the sentence, its core.
In Turkish, which is classified as a head-final language, any simple sentence has the two constituent verb phrases (subject+verb + object+verb), sharing the same head, merge into one verb phrase (subject + object) + verb. Since pragmatically, the canonical subject is also the topic that is positioned initially, the canonical Turkish sentence is ordered as subject–object–verb (SOV). This also classifies Turkish as a verb-final language.
TR: Head-Final, Verb-Final
Ayşe kitabı okuyor.
Ayşe ⟵ okuyor
Dependent ⟵ Head
kitabı ⟵ okuyor
Dependent ⟵ Head
Ayşe ⟵ kitabı ⟵ okuyor
Subject ⟵ Object ⟵ Verb
Topic ⟵ Focus ⟵ Verb
English, on the other hand, is a head-initial language. So, the two constituent verb phrases (verb+subject + verb+object), sharing the same head, merge into one verb phrase verb+(subject + object).
However, because the subject syntactically modifies the verb (and English modifiers precede their heads), the verb is moved to the post-subject position while remaining positioned before the object (its complement). Thus, the verb is positioned medially, which is why English is classified as a verb-medial language, reflected in its canonical ordering, subject–verb–object (SVO).
EN: Head-Initial (Medial), Verb-Medial
Ayşe is reading the book.
Ayşe ⟵ is reading
Dependent ⟵ Head
is reading ⟶ the book
Head ⟶ Dependent
Ayşe ⟵ is reading ⟶ the book
Subject ⟵ Verb ⟶ Object
Topic ⟵ Verb ⟶ Focus
Although commonly classified as a head-initial language (based on the head-complement relationships in English), English is sometimes reclassified as a head-medial language to additionally account for the modifier-head relationships.
Dependency Links (DLs) & Dependency Distance (DD)
Whatever language we speak, it would be safe to assume that, in our communication, we are driven by the same goal: we want to get our point across, and we want to do it fast. In our utterances, we strive both to reduce our planning effort and to effectively convey our meaning. Linguistically, the point of any utterance, in its most concise (and fastest) form, is expressed through its core: the subject, the verb, and the object if the verb requires it to make sense.
To express the point (which requires the utterance of the core of the sentence) effectively and efficiently, languages have developed their own strategies, of which the most important is the strategy for minimizing the dependency links and reducing the dependency distance, which facilitates comprehension.
As a head-initial language (for complement dependencies) and a head-medial language (for modifier dependencies), English words can have both pre-dependent constituents (with modifiers) and post-dependent constituents (with complements).
This makes English rather efficient in terms of reducing the distance between heads and their dependents, or dependencies. Turkish, as it happens, has devised other strategies to accomplish that, including reducing the number of dependencies (or dependency links) by reducing the count of words in a sentence (e.g., through agglutinating the function words, dropping pronominal subjects and objects).
Dependencies = Dependency Links (DLs)
Understanding a sentence as a combination of phrases highlights the structural dependencies within the sentence, which is why some linguists use the term dependency instead of the term phrase.
The number of dependencies, or dependency links (DLs), measures the number of phrasal connections within a clause/sentence. It directly depends on the count of words in the clause/sentence.
Psycholinguistic measures of memory load also show that the load on our memory increases with the length of a dependency (measured in terms of the number of intervening words).
The head directionality tells us on which side of the head its dependent is positioned. As phrases/dependencies are recursive, the same word can be both head and dependent, depending on the phrase. If each word had only one dependent, we would probably have perfectly structured sentences.
However, things get complicated when words have more than one dependent constituent. As most languages arrange sentences linearly, they have to find ways to arrange words with several dependents in a linear fashion.
Dependencies = Dependency Distance (DD)
The dependency distance (DD) measures the distance between the head and the dependent: from 0 (for the immediately adjacent ones, with no words interrupting) to 1 (with 1 word interrupting) ... to X (with X words interrupting).
For example, in the previous sentence Ayşe is reading the book, the dependency distance between the subject (Ayşe) and the verb (is reading) is 0 (zero). In its Turkish equivalent, Ayşe kitabı okuyor, the dependency distance between the subject (Ayşe) and the verb (okuyor) is 1 (one) because of the intervening object (kitabı) between them.
It is generally accepted that the head-medial ordering, which allows putting some dependents before the head and others after it, is optimal for reducing the dependency distance.
For example, the English sentence below has 7 dependency links and a 5 word-dependency distance:
____ __ __
￬ ￪ ￬ ￪
She regards her pen as her only weapon.
￪___￬ |_________￪ ￪ | ￪__ __ | ￪
￬___________ | ￬ ________ _|
On the other hand, the Turkish equivalent has 5 dependency links and a 7-word-dependency distance:
___ _____ _____
￬ ￪ ￬ ￪ ￬ |
0, kalemini tek silahı olarak görür.
￪ ￪______________________ |
So, while Turkish has less dependency links to go through, English benefits from the reduced dependency distance.
👉 In fact, these stats represent the general Turkish vs. English picture: English sentences move forward in a more linear fashion, with most units measuring a minimum dependency distance. An English sentence resembles a chain of dependencies hooking on each other, which can expand, or branch out, without much forward planning. English sentences are, thus, often loose and right-branching.
👉 Any Turkish sentence has the longest dependency distance encoded into its core: With fewer but longer words and the restrictions imposed by its canonical verb-final word order, Turkish sentences resemble Russian nesting dolls, with its recursive embedding-within-embedding clause-linking. This makes Turkish sentences periodic and left-branching.
English As a Subject-Prominent Language
Ideal for human memory is to have a dependent positioned immediately next to its head. Therefore, logically, if a word has two dependents, the way to guarantee that each of them is placed immediately next to its head is to put one on each side.
And this is exactly what a head-medial language such as English does.
For example, the sentence My husband called immediately, where both My husband and immediately depend on called and are placed right next to it, produces the least possible load on memory. This also explains why English prefers such arrangement to My husband immediately called, where the adverb of manner, immediately, intervenes between the subject and verb.
In fact, the subject-verb link is so “sacred” in English that almost nothing is allowed to intervene between them. This is why English does not favor “heavy” (long) subjects.
English strongly favors light (minimally worded) subjects (the best of which are personal pronouns, since they do not take any dependents) to minimize any intervention between the subject and the predicate.
Moreover, in a narrative where the subject is often a recurrent topic, it is replaced with a respective pronoun. Pronouns are not only convenient placeholders for repetitive referents, they are also inherently much “lighter” than other nouns because they tend not to have any dependents.
English: Lighter Subjects = Heavier Predicates
“Lighter” subjects inevitably mean “heavier” predicates. This is why the preference in an English sentence is to place longer and heavier structures towards the end of the clause—in line with the end-weight principle of the English language.
Every English sentence must have a subject. The English subject almost always is positioned sentence-initially, expressing the recurrent referent of the narrative. The subject commonly (but by no means invariably) identifies a topic for the clause, i.e., what the clause is primarily about, and the predicate makes some sort of comment about that topic.
The longer part of the sentence also contains more new information, which aligns with the end-focus principle (and more broadly the universal information-flow principle) that postulates that any new or important information tends to be given at the end of the sentence.
English syntax has a strong tendency to package information in a sentence so that the sentence-initial subject, which cannot be omitted, represents old or given information. The subject is, therefore, typically the topic of the sentence. This makes English a subject-prominent language.
For example, in English sentences with prepositional verbs (e.g., to give away), direct objects should be placed between the verb and the preposition. However, a long direct object triggers the end-weight principle, which prompts a delay of the direct object to the position after the preposition.
For example, in the example below, if we measure the total dependency distance of the original sentence and the revised variant, we can see why the end-weight principle makes sense:
He gave [all his valuable possessions] away.
The dependency distance is 10 words.
He gave away [all his valuable possessions].
The dependency distance is 7 words.
The end-weight principle can also explain why, in English, there is a strong stylistic preference for longer predicates. For example, English speakers tend to avoid predicates consisting of a single intransitive verb (a verb that does not take a direct object, e.g., to swim, to look, etc.). Instead, they prefer replacing it with a general transitive verb (e.g., to have, to take, to give, to do) while filling the object position with an abstract noun phrase derived from the intransitive verb. The new verb adds little information but helps to give more weight to the object.
Instead of She’s swimming, use She’s having a swim.
Instead of He’s bathing, use He's taking a bath.
Instead of The driver shouted, use The driver gave a shout.
Instead of She works very little, use She does very little work.
Instead of I was asked to look, use I was asked to take/have a look.
Sometimes, English writers replace even transitive verbs (that take direct objects) by an indirect object construction with the general verbs to give, to pay, etc.
Instead of He kicked the door, the writer may write He gave the door a kick.
Instead of They visited her, the writer may write They paid her a visit.
English Fixed Word Order
In both Turkish and English, the subject and object tend to be nouns or noun phrases, i.e., words that look very much alike, which makes them potentially confusing. Moreover, without being connected to each other, the subject and object are both connected to the main verb, the latter easily distinguishable from a noun.
By butting in between the subject and object, the English verb, in its finite, tense-signaling form, serves to mark the boundary between them. Thus, in canonical English sentences, any noun to the left from the verb cannot be an object. Likewise, the right side of the verb is mainly preserved for non-subjects. To keep the verb pinned between the subject and object, English has developed a fixed word order.
Turkish as a Null-Subject, Pro-Drop, and Topic-Prominent Language
While English personal pronouns are extremely useful as subject/topic-placeholders, Turkish pronominal subjects and objects are often omitted (or dropped) thanks to being encoded (or incorporated) into the predicate’s structure. Moreover, Turkish personal pronouns are relatively useless as identifiers: for example, the astonishingly laconic pronoun o can be translated into English as she, he, or it, and as the demonstrative pronouns that or those.
Instead, Turkish personal pronouns are used mainly for contrasting or focusing functions.
Once stated in a narrative, the common subject can be referred to without being explicitly stated over and over again. Such subject thus functions as the syntactic pivot Turkish, around which sentences revolve. This means that if two clauses share a subject, its subsequent occurrences may be omitted: the second clause is understood to have the same subject as the first one.
Turkish, which often either pushes its subjects away from the sentence-initial position or drops its pronominal subjects altogether while keeping its topics, is thus classified as a null-subject, pronoun-dropping (pro-drop), and topic-prominent language.
However, sentences with dropped subjects typically have topics, which may be sentence-initial objects, other complements, or adverbials. Studies have shown that about one thirds of Turkish sentences do not have subjects positioned initially.
Turkish: Word Count-Reducing Strategies
Turkish pronominal subjects (and objects) dropping aligns with the general need to reduce the word count in a sentence to balance off the strictly head-final structure of the Turkish language. In addition, the agglutinative nature of the Turkish language, which allows it to “glue” most of its function words to lexical words, also contributes to the significant reduction of its word load. This way, by reducing the word count, Turkish ensures that sentences have fewer dependency links.
As a pro-drop language, Turkish omits (drops) its pronominal subjects and objects if they can be inferred from previous sentences (pronominal subject and object dropping).
Turkish also employs a mechanism for reducing the longest dependency distance within sentences—namely, between the subject and the verb—by bringing the subject next to the verb in a process called backgrounding. This is part of a larger mechanism of creating irregular sentences, devrik cümleler, in Turkish.
Turkish: Backgrounding and Presupposability
In Turkish, pronominal and other subjects frequently appear in the postverbal (backgrounded) position.
Traditionally, it has been explained by the fact that some constituents in a sentence can be presupposed (inferred from the context or assumed to be shared by the hearer). As the least-informative piece of information, the speaker delays it by moving it to the background (the area after the verb), but then, because of the potential ambiguity, chooses to recover or repair the implicit information by stating it explicitly in the background area.
For example (the backgrounded constituents shown in red):
“Padişahımızın en usta nakkaşları onlar,” dedim. “Kıymayın onlara.”
“These men are the most expert miniaturists serving Our Sultan,” I said. “Make certain no harm befalls them.”
Orhan Pamuk, Benim Adım Kırmızı
Backgrounding of the subject and the oblique object (dative).
Dertli bir türküydü bu.
This was a sad song.
Yaşar Kemal, İnce Memed-1
Backgrounding of the subject (demonstrative pronoun).
Backgrounded constituents can also be multiple:
Iraz: “Demedim mi ben size?” dedi.
Iraz asked, “Didn't I tell you?”
Yaşar Kemal, İnce Memed-1
Backgrounding of the subject and oblique object (dative).
Given that non-contrastive pronouns functioning as subjects or objects can be omitted, why aren’t these dropped then? Why mention something that is presupposed at all? For example, why did Yaşar Kemal keep bu in the example before, since it can be inferred from the predicate’s suffix?
There is no clear explanation for such backgrounding. Gerjan van Schaaik, in his The Oxford Turkish Grammar, gives a hardly “worthwhile” explanation (underlining is mine):
In many a discourse certain matters are so well-known (given) that making reference to them is hardly worthwhile. In many cases such matters are merely mentioned for safety’s sake and then only by way of an afterthought.
Another explanation, provided by Aslı Göksel and Celia Kerslake in their foundational Turkish Comprehensive Grammar and Emine Erguvanlı in her seminal Function of Word Order, acknowledges the role played by backgrounding in shifting stresses within sentences.
Turkish: Backgrounding as a Stress-Shifting Strategy
By placing certain elements after the verb, other elements can be moved to the front or closer to the verb to render them tropicalized or focused, respectively. For example, in the sentence below, the backgrounding of the pronominal object (onu) allows for the subject (ben) to be focused (shown in CAPITAL LETTERS):
Kırk yıl öncesinin Kazvinli ustalarından ders almış gibi geniş, rahat, mutlu, yuvarlacık çizgiler çeker, parlak, katıksız renklerini cesaretle sürer ve resminin gizli nizamında hep afili bir yuvarlak vardır, ama BEN yetiştirdim onu.
He makes wide, easy, blithe curves, as if he’d taken lessons from the masters of Kazvin forty years ago; he confidently applies his bright, pure colors, and there’s always a gentle circularity hidden in the arrangement of his paintings; but I’m the one who trained him.
Orhan Pamuk, Benim Adım Kırmızı
This does not explain, however, why backgrounding can take place even if no shifting in the stress positions occurs. One such example is provided by Emine Erguvanlı, even though she uses her example to illustrate an obscure sociolinguistic factor that motivates Turkish speakers to background constituents that are new information.
Her example also illustrates how Turkish speakers background not only pronominal but also non-pronominal words:
Annemin çok selamı var size. Aksam misafir geliyor bize yemeğe. Onun için o yemek yapmakla meşgul.
My mother sends her greetings to you. Guests are coming to dinner tonight; therefore, she’s occupied with preparing dinner.
Of the two backgrounded constituents, bize (to us) and yemeğe (to dinner), the former is clearly given information, “since the speaker believes this to be in the consciousness of the hearer, and it is not emphatic, so it naturally occurs in the post-predicate position.”
Further, Erguvanlı states that, since yemeğe (to dinner) is new information, it would have normally been focused; yet the speaker chooses to de-emphasize it by backgrounding. The reason, she explains, is that the speaker, following a certain sociolinguistic norm—namely, politeness (an important factor in Turkish society)—wishes to avoid boasting about an event in their house.
There are several issues with this explanation:
For one, technically speaking, the canonical version of the sentence, which is all new information, save for the potentially presupposed bize (to us), would have had the word misafir (guests) focused, not yemeğe (to dinner), because the word misafir is a categorical (non-specific) subject and must be placed immediately before the verb, as the focused constituent:
Akşam MİSAFİR geliyor bize yemeğe.
(Akşam) bize yemeğe MİSAFİR geliyor.
This means that, whether backgrounded or not, the focus falls on the same constituent. The only difference is that the potential topic bize, which is presupposed info, is backgrounded (or detopicalized), as well as the new information constituent yemeğe.
Second, Erguvanlı fails to notice that the devrik version of the sentence is easier to process thanks to the reduced load on memory measured by the reduced dependency distance (DD): While the DD in the devrik sentence is 2 words, the canonical sentence would have had the DD as long as 6 words!
This is why the speaker chooses to change the word order: not so much to background the two words, since at least one of them is new, as to place the head (verb) exactly in the middle between its dependents, which makes the sentence easier to process, for both the speaker and the listener:
Devrik version (DD = 2 words):
akşam 🠚 misafir 🠚 geliyor 🠘 bize 🠘 yemeğe
Non-devrik version (DD = 6 words):
akşam 🠚 bize 🠚 yemeğe 🠚 misafir 🠚 geliyor
Syntactically and pragmatically, such backgrounding is the most optimal way because:
1. Syntactically, the head is placed at the shortest distance to all its dependents.
2. Pragmatically, in the sentence with most of the constituents denoting new information, which would normally require a higher intonational effort, the process of backgrounding allows the speaker to move one of the new information constituents to the postverbal, de-stressed area, thus reducing the number of stressing positions within the sentence, which means a lesser intonational effort on the speaker's part.
Turkish: Backgrounding as a DD-Reducing Technique
Like any other speakers in the world, Turkish speakers seek to optimize the utterance planning and delivering effort. To accomplish that, they strive to express themselves in the most economical yet effective way—namely, by reducing the dependency distance within each utterance.
How do they do that? Studies have repeatedly shown that, in Turkish spoken and written language, the most frequently backgrounded constituent is the subject, or rather topic, whether it’s pronominal or lexical, grammatical (nominative-case-marked) subject or logical (genitive-case-marked) subject, whether it’s a topicalized direct or oblique object or adverbial, or whether it’s old, presupposed, or new information. By placing the topic/subject after the verb, i.e., next to it, both Turkish speakers and writers, effectively cut the longest dependency distance in a Turkish sentence.
The significant (and largely understudied and underestimated) portion of sentences in Turkish literature are of the irregular structure, the majority of which have their would-be-topics backgrounded to bring them as close to the verb as possible:
Sanki eşyanın, atılmış hayat parçalarının yaptığı bir romandı bu.
(Bu sanki eşyanın, atılmış hayat parçalarının yaptığı bir romandı.)
… composing a novel of material objects and discarded life fragments.
Ahmet Hamdi Tanpınar, Huzur
Bir odaya kapanıp, bir masaya oturup, bir köşeye çekilip kağıtla kalemle kendini ifade eden insanın yaptığı şeyin, yani edebiyatın anlamı demek bu.
(Bu, bir odaya kapanıp, bir masaya oturup, bir köşeye çekilip kağıtla kalemle kendini ifade eden insanın yaptığı şeyin, yani edebiyatın anlamı demek.)
It is what a person creates when he shuts himself up in a room, sits down at a table, and retires to a corner to express his thoughts—that is, the meaning of literature.
Orhan Pamuk, Babamın Bavulu
Tenhaydı otobüsün içi.
(Otobüsün içi tenhaydı.)
The bus was secluded inside.
Murathan Mungan, Aynalı Kırk Oda
The devrik cümle continues to be (mis)understood and (under)appreciated as either a quirky element or an aberration in the Turkish language.
Bugün imrenerek, neredeyse ağzımızın suyu akarak izleriz biz o güzel günleri!
(Biz, o güzel günleri bugün imrenerek, neredeyse ağzımızın suyu akarak izleriz!)
Today, we watch those beautiful days with envy, almost watering our mouths!
Ezra Erhat, Sevgi Yönetimi
Ağzından tek söz çıkmıyordu(,) kadının.
(Kadının, ağzından tek söz çıkmıyordu.)
Not a word came out of her mouth.
Murathan Mungan, Aynalı Kırk Oda
In Turkish, the most frequently backgrounded constituent is topic (which can be a subject, a topicalized complement (object), adverbial, a verbal or adverbial clause, etc.).
Bir isteksizlik sardı içimi. Acıya alışıyor muyum, diye sordum kendime.
(İçimi bir isteksizlik sardı. Kendime, acıya alışıyor muyum, diye sordum.)
I was overcome with reluctance. I asked myself if I were growing accustomed to the pain.
Orhan Pamuk, Benim Adım Kırmızı
Durum nedir bugün?
(Bugün durum nedir?)
What is the situation today?
Ezra Erhat, Sevgi Yönetimi
İstediği anda da topuklarına dayandığı gibi fırladı yerinden.
(Yerinden, istediği anda da topuklarına dayandığı gibi fırladı.)
When he wanted to, after resting on his heels, he jumped up.
Uzaylıların, öcünü alacağı günü beklemekten başka bir şey kalmamıştı ona.
(Ona, uzaylıların, öcünü alacağı günü beklemekten başka bir şey kalmamıştı.)
All she had to do was wait for the day when the aliens would take their revenge.
Murathan Mungan, Aynalı Kırk Oda
Pek çok kitap okudum o yaz.
(O yaz, pek çok kitap okudum.)
I read a lot that summer.
Orhan Pamuk, Kırmızı Saçlı Kadın
Turkish: Backgrounding as a Focusing Strategy
Backgrounding is also an important technique for shifting the stress positions within a sentence. In most cases, the backgrounded constituent is a verb complement, which, unless backgrounded, has to be placed immediately before the verb (in the focused position).
Such backgrounded complements are often oblique objects of intransitive verbs that together constitute compound verbs, such as içi kaplamak. Such compound verbs are typically two- or three-place verbs, meaning that they can have two or three dependents (the subject; direct, indirect, or oblique objects; adverbials, etc.). Speakers and writers often background one of such dependents to bring it closer to its head and thus to reduce the intonational effort.
For example, in the sentence below, Orhan Pamuk backgrounds the oblique object gözümün önüne, which allows for the focusing of the other dependent, the subject insan:
Yazı deyince önce romanlar, şiirler, edebiyat geleneği değil, bir odaya kapanıp, masaya oturup, tek başına kendi içine dönen ve bu sayede kelimelerle bir yeni alem kuran İNSAN gelir gözümün önüne.
(Yazı deyince önce romanlar, şiirler, edebiyat geleneği değil, bir odaya kapanıp, masaya oturup, tek başına kendi içine dönen ve bu sayede kelimelerle bir yeni alem kuran insan gözümün ÖNÜNE gelir.)
When I speak of writing, what comes first to my mind is not a novel, a poem, or literary tradition, it is a person who shuts himself up in a room, sits down at a table, and alone, turns inward; amid its shadows, he builds a new world with words.
O. Pamuk, Babamın Bavulu
In another sentence, Pamuk backgrounds the object içimi and thus focuses the subject pişmanlık:
Ondan sonraki sessizlikte yavaş yavaş PİŞMANLIK kapladı içimi.
(Ondan sonraki sessizlikte pişmanlık yavaş yavaş İÇİMİ kapladı.)
In the silence that followed, I was gradually overcome by regret.
O. Pamuk, Benim Adım Kırmızı
Turkish Free-ish Word Order
In Turkish, even simple sentences start with strings of noun phrases, comprised of the subject, objects, nominal adverbials, or nominalized verbals (which themselves may be modified by nouns, nominalized verbals) with the verb that connects them positioned in the very end of the sentence.
Although Turkish has developed an effective case-marking system (that signals the relations between the elements in the sentence through inflection suffixes of nouns) to keep such strings of nouns apart, the boundaries between noun phrases, including their modifiers, are often not immediately clear, not least because some markings coincide.
In Turkish, the most frequently backgrounded constituent is topic, which can be a subject, a complement (object), an adverbial, or a verbal adverbial clause.
Regarded as a syntactically flexible word-order language, Turkish is, nonetheless, pragmatically fixed. Unless the sentence has a contrastive (emphatic) stressing, the topic is invariably sentence-initial, the focus is immediately preverbal, while the post-verbal position is assigned for defocused, de-stressed, or backgrounded constituents.
The process of backgrounding, which involves moving constituents to the postverbal position is not only useful but essential for clause-linking in Turkish. The main purpose of backgrounding is to bring the core constituents closer to each other, specifically, closer to the verb. This is also a great way to ensure that the subject and object, which are easily confusable in Turkish, are clearly demarcated.
Although presupposability may be a factor in backgrounding (as generally believed), considering that even new information can be backgrounded—and often it is far from clear whether the backgrounded information is given, old, or new—what actually matters for efficient sentence processing is the reduction of the distances between the dependent constituents of the sentence. This is why the most frequently backgrounded constituents are would-be-topics (i.e., subjects or any other topicalized constituents).
The irregular (devrik) sentence continues to be unfavorably contrasted with the regular (kurallı) sentence and viewed as “a strategy restricted to spoken and informal written Turkish” (A. Göksel and C. Kerslake). In reality, Turkish clausal structural variations are much richer than is allowed by the current Turkish language school.
Despite the freedom allowed by the rich morphology of the Turkish language, the preceding qualification principle established by Ottoman scholars is still prescriptively enforced, without much challenge or recognition of all the existing clause-linking mechanisms.
This means that, no matter the underlying preceding qualification principle and the verb-final word ordering principle, taken to be the incontrovertible rule of Turkish syntax, Turkish speakers and writers, just like English ones, strive to bring as close as possible to each other the core constituents of the sentence—ostensibly in violation of the established principles.