ON DEFINITENESS IN TURKIC LANGUAGES AND ON THE NATURE OF THE TURKIC ‘ACCUSATIVE’ (WITH A VIEW TO TURKISH)

 

IVAN G. ILIEV^*

 

 

Summary

 

In the article an attempt is made to prove that the morpheme in Turkic languages considered till now to be an accusative ending is actually a post-positioned definite article.

 

Key Words: Turkish, Turkic languages, accusative, definite article.

 

Introduction

 

In linguistics, and in other sciences, too, it often happens that a wrongfully established opinion is for decades repeated by generations of scholars, no attention being paid to its groundlessness. With this article, I’ll try and correct one just such instance connected with the definition of the so-called ‘accusative case’ in Turkic languages (mostly in Turkish), which practically isn’t a case form but expresses definiteness in the position of direct object – that is it plays the role of a post-positioned definite article. The correct formulation of the issue is relevant not only to Turkology but also to Balkan and Slavic studies, because it presupposes a fresh look at phenomena encountered in some Balkan and Slavic languages, which most probably have to do with the state in contemporary and old Turkic languages.

 

Current definition of ‘-i forms’^[1] in Turkish.

 

It is considered^[2] that: ‘The first comprehensive modern treatment of Turkish grammar was Jean Deny’s monumental ‘Grammaire de la langue turque, dialecte osmanli’, published in Paris, in 1921. This was a significant first step towards the creation of a terminology that would accurately reflect the features of the language without trying to assimilate them to Indo-European preconceptions’. However, regarding the so-called “accusative case” in Turkish in practice there is copying of just an Indo-European model.

Being for centuries part of the Ottoman Empire and territory on which even after the Liberation in 1878 and to this day significant masses of a Turkish-speaking population continue to live, which to a different degree is used by an education in the mother tongue, too, with the exception of a period of 30 or 40 years at the end of the 20^th c., it can be assumed that the Turkish grammatical tradition in the territory of Bulgaria doesn’t hold a second place to those in many other countries (not counting Turkey). That’s why I’ll start with a quotation from the Turkish language grammar most popular in Bulgaria that has undergone several editions – the one by G. Galabov, using the third edition variant of 1957^[3] issued for the fourth time anonymously in the 1980’s after the so-called Revival Process of 1984-1985.

In the grammar in question^[4], regarding the use of the so-called ‘accusative case’ it is written: ‘The noun is used in the accusative case in the sentence when it serves as a definite direct object (bold face here and below mine – Iv. Il.), as for instance: Yarışı kazandınız ’You won the competition.’ Or^[5]: ‘in Turkish the direct object stands in the nominative or accusative case … When the direct object is in the nominative, it generally signifies the object without defining it precisely; in that case it’s called i n d e f i n i t e ; в (Ahmet roman okuyor ‘Ahmed is reading a novel’) ... When the direct object is in an accusative case form, it expresses an exactly definite object we’ve been discussing or are interested in etc. That direct object is called d e f i n i t e (Ahmet romanı okudu ‘Ahmed has read the novel’’) ... The definite direct object in Turkish expressed through the accusative case form is rendered in Bulgarian with a noun used with the article, and the indefinite direct object with a form without the article.’

The attentive unbiased linguist would immediately ask himself why with the so-called ‘accusative case’ in Turkish only the definite direct object is marked, and not the indefinite one, and what is that ‘nominative case’ with which a form, although indefinite, of the same direct object is designated. Obviously here another grammatical category is concerned (definiteness and not accusativeness) and another grammatical category (the definite article and not an accusative ending). ^[6]

The same question I. Hazimov, too, obviously asked himself as early as 1929^[7] and he almost hit the truth but didn’t dare to spell it out: ‘the accusative case replaces the definite article, as there is not one in Turkish’. Yordanova’s comment^[8] about that, without any arguments being supplied, is as follows: ‘Indeed in Turkish there is no definite article but it’s not it that the accusative case replaces although the accusative serves to express definiteness’.

The same treatment is repeated still further in the Turkish grammars published in Bulgaria. Here is what is written in one of the latest of them, issued in Bulgarian^[9] on the same matter: ‘The names in accusative case are objects on which the action immediately falls that is expressed with the predicate and in the sentence, they serve as definite direct objects: Kızım, üzümü yıka ‘Daughter, rinse the grapes.’

Let’s take a look also at a Turkish grammar published in Russia^[10]. The difference in it is that instead of the term, inexact in this case, of ‘nominative case’, another is used (‘indefinite case’) and an exclusively important specification is made about the grammatical essence of proper names and the personal and demonstrative pronouns: ‘The ‘indefinite case’ designates the subject, corresponding to the Russian nominative case (Kitap kalındır ’the book is thick’) … (and to), the direct object, corresponding to the Russian accusative case. The indefinite case appears to be a case of the direct object in these instances: a) The object appears indefinite (logically and grammatically), mentioned for the first time or inseparable from its likes. Proper nouns, personal and demonstrative pronouns, being in themselves definite elements, when objects, are always formed as the accusative case (O kitap okur ‘he a book/books (indef.) is reading’). The accusative case is used with a definite object (kitabı okuyor ‘he is reading the book (a definite one)’) or with a grammatically definite object, i. e. an object defined with a noun, an adjective, a pronoun or affixes of belonging (babanım kitabını okuyorum ‘I am reading my father’s book’)’. 

Let’s see the formulation of the issue in the Turkish grammatical literature as well (grammars and dictionaries of linguistic terms). N. Koç^[11] in the same vein talks about belirtili nesne (determined direct object) – ‘belirtme durumu eki almış nesne’ (direct object that receives a suffix of definite case (! – the exclamation and bolding mine – Iv. Il.) or definite form): Kitaplarınızı, defterlerinizi kapatınız ‘close your books (and) notebooks’ and for belirtisiz nesne (indetermined direct object) – ‘ad durumu eki almamış yalın durumdaki nesne’ (form of the name not having received a suffix or a direct object in the nominative case): Eve giderken ekmek al ‘on the way home fetch some bread’.

А. Eker^[12] presents things almost as in the dictionary of linguistic terms cited, but belirtili nesne is translated into English just as a direct object, belirtisiz nesne is rendered as indetermined direct object – that is, he assumes that the direct object’s normal state is for it to be definite. 

There is a small difference with Т. Demir^[13], in whose view in Turkish besides belirtisiz nesne or indefinite direct object (Annesi ona bir kuzu getirmişti ‘his mother has brought him a lamb’) and belirtili nesne or definite direct object (Ağaçları kestiler ‘they cut the trees down’) there also exists kökteş nesne or cognate (tautological) direct object (Bir bakış baktın ‘you take a look’ etc.).

The German, English, American, Polish grammars I won’t dwell on in detail. Yordanova^[14] has done that exhaustively and it can be seen from her survey that there one can’t find anything very different from what has been shown here regarding the definition of the so-called ‘accusative case’. A more serious attention is due however to the co-authored grammar of Göksel and Kerslake^[15], in which to the question of expressing definiteness and indefiniteness a special attention is paid and again the words that are naturally definite are indicated: ‘In Turkish the minimal requirement for a noun phrase to be interpreted as definite is:

(i) the absence of an indefinite determiner (15. 6. 1) (bir, birkaç, etc.)

(ii) accusative case marking where the noun phrase is functioning as direct object (Garson temiz tabak-lar-ı masa-ya koydu  ‘The waiter put the clean plates down on the table’).’

Similar to Kononov above, here again attention is paid to words whose natural quality is to be definite:^[17]

The following classes of noun phrases are inherently definite: the proper names of people, places and institutions (Osman-ı dün gördük ‘We saw Osman yesterday’); most pronouns – specifically the simple personal pronouns (Murat sen-i seviyor ‘Murat loves you’); the demonstrative pronouns, etc.; a noun phrase that includes one of the definite determiners (Bana bu oda-yı ver-di-ler ‘They have given me this room’)’.

Definite determiners are dealt with on page 180-181 (bu, şu, o, her, etc.). On page 181, it is said again that ‘Noun phrases with definite determiners obligatorily (my bolding – Iv. Il.) have accusative case marking when functioning as direct objects.’

 

About other instances of definiteness in Turkish

 

Definiteness and indefiniteness in Turkish have been the particular subject of several research papers.^[17] In the first one of them I will dwell on^[18] , in the part entitled ‘Turkish marking of indefiniteness with singular nouns, plural nouns and mass nouns’, the expression of definiteness with the preposition is already discussed and it is pointed out that ‘the zero case of the subject in Turkish is called nominative, casus generalis, or absolute case and usually shows definiteness when it is not expressly marked as indefinite: Adam geldi ‘the man came’ и  ‘the plural subject usually neutralizes the opposition definite vs. indefinite: Adamlar geldiler ‘men came/the men came’.

Regarding indefiniteness with the direct object it is indicated that ‘the direct object in the absolute case without modifiers marks indefiniteness: Mehmet mektup yazdı ‘Mehmet wrote a letter/letters’, but: Kitap okumayı severim ‘I like reading a book/books’. And further: ‘the accusative case is tightly bound to the category of definiteness as it usually indicates the definite object: Kitabı aldım ‘I bought the book’.

The new information here is the statement shown on p. 198 that: ‘an indefinite object can also take the accusative case markers if it expresses specific/referential indefiniteness:

Her gün bir gazete okuyorum ‘every day I read a newspaper’ as against:

Her gün bir gazeteyi okuyorum ‘every day I read a (particular) newspaper’. Into Bulgarian however the second sentence is translated as containing a definite numeral – Аз чета единия вестник – literally ‘I read the one newspaper’ (below I’ll discuss such cases in detail).

The same research also examines the expression of definiteness in other syntactic positions – predicate, first member of izafet, etc. On p. 200 the author dwells on the expression of definiteness with the genitive and says that ‘the genitive serves to mark the definite status of the first element of an izafet group: Çocuğun kitabı ilginç ‘the child’s book is interesting’. On the next page^[19] the quoted author examines the function of ‘the third-person possessive suffix as ‘definite article’ (the quotes around ‘definite article’ are by him) and stresses: ‘According to the great turcologist K. Grønbech, the third-person possessive suffix of the Turkic languages originally had the function of a veritable definite article … This viewpoint has often been criticized by later scholars (Johanson, etc.)’.

Even before Rocchi, Aygen paid attention to definiteness-indefiniteness in Turkish – first in the chapter regarding objects in her dissertation^[20] and then in a separate article.^[21] What she says is not different from Rocchi’s statements but she is more detailed in describing the use of adverbs, adjectives and the word-order in sentences that contain direct objects: Ben hızlı kitap okurum ‘I read a book fast’ or Ben belki kitab-ı okurum / Ben kitab-ı belki okurum ‘Maybe I read the book’ but *Ben kitap hızlı okurum. Also: Ben bir kitab-ı okudum ‘I read one of the books’ or Üç çocuk yeni bir araba almış ‘Three children bought a new car’.

What Aygen and Rocchi haven’t noticed however is the fact that the possessive morpheme in question (in different variants according to vocal harmony: -ı/-i/-u) formaly coincides with the accusative ending! For the clarification of the matter what also matters is the existence in Turkish of the so-called ‘Janus constructions’,^[22] with which ‘the third-person possessive suffix serves as definiteness marker of Turkish nouns in the case of third-person linking of the sentence between two related or closely connected people, who are both defined by the suffix in question: Hastası doktorunu arıyor ‘the patient is seeking the doctor’ (literally His patient his doctor is seeking)’.

The basic reason for not defining the Turkish ‘accusative’ as a definite article must have been the fact that it is expressed by different means in the position of subject and in in the position of direct object (that is no precedent however because in many other languages it is the same – German, Greek etc.).  

It’s well-known that in English the definite article the is not placed before a possessive pronoun, which means in that language the latter is also assumed to be definite. But in Bulgarian there are cases (not impossible under influence from Turkish, see below) in which the postpositive definite article also expresses possession: Ženata doyde ‘My wife came’ – literally ‘The wife came’^[23]. The sentence Ženata doyde ‘My wife came’ (literally ‘The wife came’) in a certain context expresses the same as Žena mi doyde ‘My wife came’. That is why it isn’t impossible, in a sentence as Hastası doktorunu arıyor ‘the patient is seeking the doctor’ for hastas-ı to be perceived both as ‘his patient’ and as ‘the patient’. In the same way, Osmanın karısı/kadını ’Osman’s wife/woman’ (in front of a personal name in English the definite article is not placed either) is translated into Bulgarian with an article жената на Осман (literally ’the wife of Osman’), and in colloquial Bulgarian and in some Bulgarian dialects the Turkish word order is used^[24]: na Оsman ženata – literally ’*the wife of Osman’ or na Оsman žena mu – literally ’*Osman’s his wife’.

Now I’ll return to cases of the type Her gün bir gazeteyi okuyorum ‘every day I read a (particular) newspaper’, in which before the definite direct object the numeral bir ‘one’ stands. In another research,^[25] another identical answer is given: Bugün bir avukat-ı görüyorum ‘I am seeing a (particular) lawyer today’ (Bulg. Viždam ediniya advokat dnes – literally ’the one lawyer’).

Bliss^[26], in the chapter ‘Specificity in Turkish’, also gives examples with the three kinds of direct object indicated so far:  

1. Indefinite: Roman bulmak istiyorum ‘I want to find a novel/novels’ (expressed with a ‘bare noun’) – the semantic reason for the direct object,  expressed with roman, to be able to be translated into Bulgarian or into English both with a singular and a plural form will also be discussed below;

2. Definite: Romanı bulmak istiyorum ‘I want to find the novel’ (with an -ı ending – my note – Iv. Il.);

3. Specific indefinite: Bir romanı bulmak istoyorum ‘I want to find a (certain) novel’.

An important specification regarding the grammar essence of defined names in Turkish is indicated in Nakipoğ^[27]: ‘the accusative marked definite noun is related to focus (more exactly, the topic-focus structure – my note – Iv. Il.) structure: Arabada üç kutu vardı. Emre bir kutu-yu taşıdı ‘There were three boxes in the car. Emre carried a box (one of the boxes)’. Because the example Emre bir kutu-yu taşıdı ‘Emre carried one of the boxes’ also belongs to the mentioned third type of direct object that is translated into Bulgarian defined (едната кутия ‘one of the boxes’– literally ‘the one box’), I’ll now comment on this matter in particular.

None of the authors cited, however, dwell on similar examples in a wider context (for instance: Bir kitab-ı aldim, diğerin-i bıraktım) that would show what happens to the remaining boxes, books, novels, lawyers etc. and that would make clear which of the three possible cases they belong to (indefinite, definite, ‘specific indefinite’!).  

Besides the fact that, in Bulgarian, combinations of the type bir gazeteyi – Bulg. ediniya vestnik (‘a particular newspaper’ – literally ‘the one newspaper’),  bir avukat-ı – Bulg. ediniya advokat –  (‘a particular lawyer’ –  literally ‘the one lawyer’),  bir romanı – Bulg. ediniya roman (‘a particular novel’ – literally ‘the one novel’), bir kutu-yu – Bulg. ednata kutiya  (‘a particular box’ – literally ‘the one box’) are definite (when the noun is modified by an adjective in Bulgarian the latter assumes the article instead of the former), another fact got me wondering in these constructions. When I asked several informants, Bulgarian Turks, to translate for me into Bulgarian the sentence Bir kitabı aldım, most of them replied it was wrong! Just one gave me the answer I was expecting – Vzeh ednata kniga (literally – ‘I took the one book’), and when I asked one more informant, of those who said the indicated sentence wasn’t correct if it couldn’t be translated into Bulgarian as Vzeh ednata kniga, the reply was that Vzeh ednata kniga should be rendered in Turkish as Kitabın birini aldım. That shows that a sentences of the type Bir kitabı aldım aren’t characteristic of the Turkish dialects, at least in the Eastern Rhodopes in Bulgaria, where my informants were from.

In the sources cited so far, another case isn’t mentioned, of expressing definiteness in Turkish – after a numeral, concretely in toponymy. While for instance the phrase yedi kız gets translated as ‘seven girls’, the toponym Yedi kızlar (camisi) – the name of a mosque in the village of Chorbadzhiysko, Kirkovo Municipality, District of Kardzhali in Bulgaria – is translated as ‘the seven girls’.

It seems that not only the ‘accusative case’ in Turkish isn’t an accusative case, but the plural ending (which is identical with nouns and verbs,^[28] at least in its primary meaning, did not have as its basic function the expression of plurality. The ‘plural’ ending -lar/-ler is not necessary to express plurality after numerals: yedi kız ‘seven girls’. What’s more, in many cases it is not necessary for number concord to exist between the subject and predicate, for example: Kızlar çıktı ‘The girls went out^[29] – literally ‘*The girl-ren went out’ (as in ‘The brethr-en/childr-en/ox-en went out’). Above it was indicated that with the subject ‘the plural subject usually neutralizes the opposition definite vs. indefinite: Adamlar geldiler ‘men came/the men came’, and with the direct object (emphasized by two separate authors) the same applies to the ‘bare noun’: Mehmet mektup yazdı ‘Mehmet wrote a letter/letters’ or Roman bulmak istiyorum ‘I want to find a novel/novels’. Those last examples also cast the shadow of doubt over the plural essence of the initial meaning of the morpheme -lar/-ler in Turkish, and in Turkic languages in general.

 

Expressing definiteness-indefiniteness in the direct object in other Turkic languages

 

As everything started from the position of direct object (although definiteness with the subject and uncoordinated modifier was mentioned, too, and definiteness with the indirect object is another matter!), I’ll discuss the similar cases in other Turkic languages, too, without exhaustively delineating the situation in all of them. I’ll basically look at Chuvash (a direct successor to or in the worst case the most kin language to Bulgar (or Proto-Bulgarian) which gets fused with Slavic when the Slavic-Bulgarian nation is formed) and Tatar (a direct successor to the Cuman language which has also influenced the formation of the Slavic Bulgarian language until its present state – the Cumans during the Second Bulgarian state in the 12^th -14^th c. play the same role as the Bulgars (Proto-Bulgarians) during the First Bulgarian state 7^th -10^th  c).

 

А. In Tatar

 

I’ll breach chronology and start with Tatar, because in Chuvash the situation is slightly modified. In Tatar the situation with the definite and indefinite direct objects is the same as in Turkish: ‘The accusative case suffix marks the object of a verb when it is definite – kitap-пı ‘the book’^[30]

‘Yesterday that boy gave the book to the (little) girl’

O   bala   tynevin       kitap-пı    qız-tƒıq-qa     ber-di

that   boy yesterday     the book   to the girl       gave

An indefinite object of a verb does not have an overt accusative case marker, so the form of the noun is identical to the nominative:

‘The child reads a book’

Bala       kitap       oqu-j

the child   a book     reads

From the shown Tatar examples, it is also seen that definiteness with the subject (bala ‘child’) and also with the indirect object (qız-tƒıq-qa ‘to the little girl’), which syntactic position is not drawn into attention in this article, is expressed just like in Turkish – with a zero morpheme.

 

B. In Chuvash

 

In Chuvash, as different from Turkish or Tatar, for accusative and dative one and the same form is used. In Degtyarev’s view^[31]: ‘The dative-accusative case can express … a definite (emphasized) object of the action:

Кĕнеке-не ил ‘take this/the book’  - as against kĕnеке il ‘take a/whichever book’.

The same can be seen with Krueger^[32], too: ‘The objective relational morpheme has the function of indicating the person of objects to or for which something is done, as well as indicating the actual object of action, in the event that this is a definite and specific object … When the object of action is general and unspecified, the zero allomorph is employed:

Epĕ śakă kĕneke-ne iletĕp ‘I buy this book’ – but: Epĕ kĕneke iltem ‘I bought a book’.

 

C. In other Turkic languages

 

I won’t enter into details about all the Turkic languages, but for instance, in Bashkir, it is the same^[33] – the accusative … functions as a definite direct object:

‘Thou art reading the book’

Hin  kitap-tï   uqïyhïŋ

you the book  read

In Kazakh, too,^[34] ‘the object of the sentence or the indefinite subject is in the nominative’, and the accusative marks the definite object of a verb:

‘I have not seen this/the house’

Bul      üj-di      körgen  žokyp

this  the house  I have not seen

In Turkmen, the accusative case functions to mark definite grammatical objects, while indefinite objects are formally indistinguishable from nominals in the nominative case^[35]:

‘Students brought by hand all of the bricks for the guardhouse’

Garawulhananyň … kerpiçleri-ni okuwçylar elleri daşadylar

guardhouse    the bricks   students   by hand brought

 

The use of cases with postpositions in Turkish (as another argument for the correctness of my assertion)

 

There exists another fact, besides the use of an ‘accusative’ ending to mark just particular direct objects, which shows that in Turkish at least it isn’t about any accusative ending but the definite article. It’s the lack of a prepositional government with the accusative (at least in the language mentioned). As the numerous times quoted Galabov shows^[36], in Turkish postpositions can govern different cases: nominative (böcekler gibi ‘like insects’), genitive (benim gibi ‘like me’, sizin kadar ‘as much as you’), dative (Kütüphaneye doğru gitti ‘she or he left for the library’), ablative (benden başka ‘besides me’). No use exists, however, of prepositions with the accusative case. And that is indicative of the real character of the latter in Turkish and the rest of the Turkic languages!

 

Conclusion

 

The re-thinking of the essence of the Turkic ‘accusative’ changes a number of obsolete views on Balkan languages and at least two things become clear – Turkic languages as exemplified by Bulgar (Proto-Bulgarian), Cuman, Ottoman Turkish can’t help having influenced the formation of the postposited definite article in the former.