Turkish by Robert Underhill

1

Table of Contents

1 History and Classification ................................................................................................2

2 Phonology and Orthography ............................................................................................4

2.1 Vowels .....................................................................................................................4

2.2 Consonants ...............................................................................................................5

2.3 Stress ........................................................................................................................7

3 Morphophonemics ...........................................................................................................8

3.1 Vowel Harmony .......................................................................................................8

3.2 Other Morphophonemic Processes ............................................................................9

4 Morphology................................................................................................................... 10

5 Syntax ........................................................................................................................... 12

6 Sources .......................................................................................................................... 16

2

TURKISH

ROBERT UNDERHILL San Diego State University

1 History and Classification

Turkish is a member of the Turkic family of languages, which extends over a vast area

in southern and eastern Siberia and adjacent portions of Iran, Afghanistan, and China. The

more widely spoken Central Asian Turkic Languages include Karakalpak, Kazak, Kirghiz,

Uygur, and Uzbek. To the east of these, there is another group of Turkic languages north

of the Altai Mountains: this group includes Yakut in eastern Siberia. To the west, Tatar is

spoken in the Volga area and in the Urals, and there is a group of related languages north

of the Caucasus. Chuvash, descended from the language of the Huns, is also spoken in the

Volga region.

Turkic, in turn, belongs to the Altaic family of languages, which also includes Mongol

and Manchu-Tunguz languages of north-eastern Siberia. Some scholars have suggested

that Korean, and perhaps even Japanese, may be related to Altaic. Further, there are

typological and lexical similarities between Altaic and Uralic languages, which include

Finnish, Estonian, Hungarian, and a number of Siberian languages, notably Samoyed.

These similarities may be evidence for a Ural-Altaic language family, but they may also

be explained by prolonged cultural contact between Altaic and Uralic tribes in Siberia.

The subgroup of the Turkic family that particularly concerns us here is the

southwestern or Oğuz group. These languages are: (1) Türkmen, spoken east of the

Caspian Sea, in the Türkmen SSR and adjacent portions of north-eastern Iran; (2)

Azerbaijani, or Azeri, west of the Caspian Sea in the Azerbaijan SSR and north-western

Iran; (3) Ghashghai, spoken by a number of nomadic tribes in the Zagros Mountains of

southwestern Iran; (4) Gagauz, in two small areas on the coast of the Black Sea, in the

Moldavian SSR and Bulgaria; (5) Turkish, the standard language of the Republic of

Turkey. Turkish is also spoken in small areas throughout the Balkans, notably in Greece,

3

Bulgaria, and Macedonia; and on Cyprus. There is a Turkish-speaking population in

northern Iraq, in the area of Kirkuk; and smaller groups, including Turkish-speaking

Armenians, throughout the Middle East, particularly in Syria and Lebanon.

While Turkish speakers arrived in the area that is now Turkey in the 11th

century, the

earliest written materials in Anatolian Turkish date from the 13th

century. Turkish scholars

divide the history of Turkish into three periods. Old Anatolian Turkish (Eski Anadolu

Türkçesi) includes the 13th

century through 15th centuries. Ottoman Turkish (Osmanlıca)

includes the period of the height and decline of the Ottoman Empire. The most important

characteristic of Ottoman which distinguishes it from Modern Turkish is the very heavy

influence of Arabic and Persian, a consequence of Arabic and Persian influence on

Turkish literature and culture during that period. Ottoman Turkish was written in Arabic

script, used a higher proportion of Arabic and Persian words, particularly in literary or

learned writing, and borrowed certain syntactic rules from Persian. One of these rules was

the Persian rule for the noun phrase, by which the adjective or possessor follows the head

noun: Bâb-ı âli „great gate‟, bâb „door‟, âli „big‟; in Turkish syntax, adjectives and

possessors precede the head noun. Another was the use of Persian ki „that‟ and its

derivatives as subordinating conjunctions preceding embedded sentences; Turkish syntax

uses participial and nominal verb forms rather than subordinating conjunctions.

The transition from Ottoman to Modern Turkish (Yeni Türkçe) is given by the political

events connected with the fall of the Ottoman Empire in the Civil War of 1919-22, and by

the Language Reform movement of the late 1920‟s and 30‟s. The language reform

movement must be understood in the political and social context of the Kemalist

revolution, a nationalist and secularist movement aimed at the modernization and

westernization of Turkey. The reduction of Arabic and Persian influence on the language

thus coincided with the broader political goals of the reduction of Oriental and Islamic

influence on Turkish culture in favor of native or Western influence.

The modern language reform movement is considered to date from 1928, when the

Arabic script was replaced by a Latin orthography. While the Latin alphabet was claimed

to be easier to learn and use, it is not clear that the change in orthography could be

justified on linguistic grounds alone, without accompanying political motivation. Older

people who know both scripts continue to use the Arabic script in private correspondence.

The Arabic writing system was phonological in the modern theoretical sense, in that it

4

tended to write each morpheme in a constant form regardless of its phonetic variations,

while the Latin writing system is closer to an autonomous phonemic level, in that it

records much predictable variation.1

During the decade following the orthographic reform, and continuing until the present

time, the Turkish Language Society (Türk Dil Kurumu) has supervised a steady program

aimed at the reduction of the number of Arabic and Persian loanwords. Turkish

replacements were taken from non-standard dialects or other Turkic languages,

constructed with Turkish derivational suffixes (some of which were revived for this

purpose), or simply invented: okul „school‟ is a mixture of French école and Turkish oku-

„read, study‟. Sometimes loanwords were replaced with simpler loanwords: kütüphane

„library‟, made of the Arabic plural kütüp „books‟ plus Persian hane „house of‟, has been

replaced (in some quarters) with kitaplık, made of the Arabic singular kitap „book‟ plus a

Turkish abstract noun suffix –lIk. The Arabic and Persian component of the vocabulary

has been by no means eliminated; a count of the basic vocabulary in one recent textbook

(Underhill 1976) gives 35% Arabic or Persian against 62% Turkic and 3% European; the

percentage would be higher in a more complete lexicon or in literary of learned text. It is

significant that there has been little attempt to reduce the number of European loanwords.

Some Greek and Italian loans are very old, while more recently many French and English

loans have accompanied the modernization and Westernization of Turkey.

In syntax, the Persian noun phrase rule has been dropped in favor of the Turkish rule,

which works as well. However, the Persian rule for subordinate clauses after ki has been

retained; this rule has some utility, since it permits Extraposition. As a result of these

grammatical and lexical changes, even nineteenth century Ottoman texts look very

different from modern Turkish, and are unintelligible to most modern speakers. (Language

reform: Heyd 1954, Gallagher 1972.)2

2 Phonology and Orthography

2.1 Vowels

Turkish has an eight-vowel system, made up of all possible combinations of the

distinctive features front/back, high/low, and rounded/unrounded. The resulting system is

often displayed as a cube:

5

ї u

i ü üü

o

e ö

The short vowels are normally pronounced lax. /ö ü/ are front rounded vowels, and /ї/ is a

high back unrounded vowel, written ı (an undotted i) in the Turkish orthography.3

There are also long vowels, which come from two sources. Arabic and Persian loans

have introduced the long vowels /a: e: i: u: /, thus sakin [sa:kin] „quiet‟ vs. sakın [sakın]

„beware‟, memur [me:mur] „official‟ vs. meme [meme] „breast‟. Long vowels have also

arisen in native words through the loss of ğ. The orthographic symbol ğ stands for an

abstract phonological element which formerly was a voiced velar fricative [ɣ]; it is

preserved as such in many non-standard dialects of Turkish and in other Oğuz languages.

In standard Turkish, however, it has no consonantal articulation. In intervocalic position,

it disappears, creating a bisyllabic two-vowel sequence: ağaç [aač] „tree‟, eğer [eer] „if‟.

In syllable-final position, it is lost with lengthening of the preceding vowel: dağ [da:]

„mountain‟, iğne [i:ne] „needle‟. Standard Turkish thus has eight long vowels beside the

eight short ones, for a total of sixteen vowel phonemes.4

2.2 Consonants

The consonants of Turkish are the following:

p t č k k

b d ǰ g g

f s š h

v z ž

m n r l l

y

ü

a

e

6

The front velars /k g / and the palatal /l/ are distinctive only in loans. In words of

Turkic origin, they occur with front vowels, while /k g l/ occur with back vowels: kara

[kara] „black‟, kere [kere] „time‟; ak [ak] „white‟, ek [ek] „affix‟ . This is often called

„consonant harmony‟ in the literature on Turkish . /k g l/ have become separate phonemes

through the influence of loanwords. Turkish, like many other Turkic languages, uses the

contrast between /k/ and /k/ to preserve the Arabic contrast between /k/ and /q/, and uses

[l] for Arabic /l/ which is normally palatal. As a result, /k g l / have come to appear with

back vowels, producing pairs like kar /kar/ „snow‟, kâr /kar/ „profit‟; bal /bal/ „honey‟,

hal /hal/ „condition‟. /l/, furthermore, is spreading at the expense of /l/ in both initial and

final position.

The sonorants /m n r l l / form a structural class, particularly with regard to the

important rule for the formation of syllable-final clusters, which consist most typically of

sonorant plus voiceless stop ( /č/ being considered a stop). /y/ is not a member of this

class.

On an abstract phonological level, one might want to recognize two more segments.

[ğ], which we have already seen, can be analyzed as a separate segment, or as the

postvocalic reflex of /g/, or as derived from /k/ under various conditions. Glottal stop,

which is the reflex of both glottal stop and ain in Arabic loans, plays a limited role in

syllabification and morphophonemics.

The orthography uses the following special symbols for consonants:

Phoneme Letter

č ç

ǰ c

š ş

ž j

/k g l / before back vowels are sometimes indicated by a circumflex accent on the

vowel: gâvur [gavur] „infidel‟, lâle [lale] „tulip‟.

7

2.3 Stress

Stress in Turkish, which is actually pitch accent rather than dynamic stress, is usually

on the last syllable of the word, so that stress appears to move rightward as suffixes are

added:

evlér „houses‟

evlerím „my houses‟

evlerimdé „in my houses‟

However, there are cases where stress appears non-finally, and these fall into certain

regular classes. (1) Certain suffixes are unstressable, and stress falls on the syllable

preceding the leftmost unstressable suffix. Examples of unstressable suffixes are the

negative –mE, and all suffixes belonging to auxiliary constituent:

gítme „don‟t go!‟

gitmé ‘going’

várdı „it existed‟

vardí „it arrived‟

(2) Certain lexical or syntactic classes call for initial stress; these include place

names, vocatives, and some types of adverbs:

bebék „baby‟

Bébek suburb of Istanbul

babá „father‟

Bába „Father!‟

(3) Many lexical items have inherent stress on a non-final syllable; most of these are

European loans, although a few native Turkic words are included. In such words, the

stress stays in the same place even when suffixes are added: lokánta „restaurant‟,

lokántalar ˙(pl.); mása „table‟, másalar ˙(pl.). (Stress: Lees 1961, Swift 1962, Foster

1969, Zimmer 1970b, Lightner 1972, Sezer 1983.)

8

3 Morphophonemics

3.1 Vowel Harmony

The best-known morphophonemic process in Turkish is vowel harmony, which is a

process by which the vowels in all syllables of a word except the first assimilate to the

preceding vowel with respect certain phonetic features. This situation may be illustrated

with the following forms:

Nominative Objective Dative

„hand‟ el eli ele

„dog‟ it iti ite

„eye‟ göz gözü göze

„ashes‟ hül külü küle

„horse‟ at atı ata

„girl‟ kız kızı kıza

„arm‟ kol kolu kola

„slave‟ kul kulu kula

Most specialists will agree that vowel harmony in Turkish is a left-to-right process

operating sequentially from syllable to syllable. The rules are:

1. A non-initial vowel assimilates to the preceding vowel in frontness.

2. A non-initial high vowel assimilates to the preceding vowel in rounding.

3. A non-initial low vowel must be unrounded; that is, o and ö do not occur except in

first syllable of words.

Thus while any of the eight vowels may occur in the first syllable of a word, the vowel

of the following syllable is restricted to a choice of two. The features front/back and

rounded/unrounded are entirely predictable, and only high/low vowels remain distinctive.

In citing morphemes, we may use the cover symbol E for a low vowel, appearing on the

surface as a or e, and I for a high vowel, appearing as i, ı, u, or ü. The objective suffix is

thus –I and the dative suffix is –E.

Some of the evidence for talking vowel harmony to be a left-to-right process also

poses interesting problems for the formalization of the vowel harmony rules. This

9

evidence is: (1) high vowels are rounded only if there is a rounded vowel in the

immediately preceding syllable; if an unrounded low vowel intervenes assimilation does

not carry over: kol „arm‟, kolum „my arm‟; kollar „arms‟, kollarım „my arms‟

(*kollarum). (2) Non-harmonic suffixes, which are exceptions to harmonic conditioning

from the vowel on their left, condition the vowel on their right normally. Thus the

progressive –Iyor contains an invariant /o/ which controls the vowel harmony on the

following suffix: geliyorum „I am coming‟, bakıyorum „I am looking‟. (3) Non-harmonic

stems, which are not subject to vowel harmony internally, take suffixes conditioned by

the last vowel of the stem: elma „apple‟, elmalar „apples‟; kardeş „sibling‟, kardeşler

„siblings‟.

Recently the autosegmental theory has explored promising ways of treating the

features [Back] and [Round] on the separate level, spreading them across a word by

association rules. This also offers new approaches to the much-discussed issue of whether

vowel and consonant harmony can be handled by the same rules. (Autosegmental:

Clements 1980, Clements& Sezer 1982.)

3.2 Other Morphophonemic Processes

There are many other morphophonemic processes in Turkish. Some of these, like

vowel harmony, are assimilatory in nature. A suffix-initial stop must agree in voicing

with the preceding segment: sandalye „chair‟, sandalyede „in the chair‟; sepet „basket‟,

sepette „in the basket‟. Suffixes also alternate according to whether they follow a vowel

or a consonant: ev „house‟, evim „my house‟, evi „his house‟; deve „camel‟, devem „my

camel‟, devesi „his camel‟.5

Other rules can be explained in terms of constraint on surface syllable structure. These

rules reduce syllable-final geminate consonants by deleting one; break up syllable-final

consonant clusters, except those consisting of sonorant plus stop, 6

by inserting an

epenthetic vowel; devoice syllable-final voiced stops;7

and shorten long vowels in close

syllables. A velar consonant, which is either /ğ/, /g/, or /k/ depending on one‟s

phonological theory, deletes in intervocalic position. (K-deletion: Zimmer 1975,

Zimmer&Abott 1978, Sezer 1981.) the following examples will illustrate some of these

processes.

In each case the underlying stem may be found by deleting the final –I from the

objective form:

10

Nominative Objective

„right‟ hak hakkı

„city‟ şehir şehri

„reason‟ sebep sebebi

„color‟ renk rengi

„time‟ zaman zamanı (/zama:nı/)

„butterfly‟ kelebek kelebeği

The result of these processes is that Turkish suffixes tend to have a highly protean nature.

As an extreme example, a nominal and participial suffix which we might transcribe as –

DIk has 16 different forms: -dik/dık/duk/dük/tik/tık/tuk/tük/diğ/dığ/duğ/

düğ/tiğ/tığ/tuğ/tüğ.

4 Morphology

Turkish is often cited as a canonical example of an agglutinating language, meaning a

language in which the grammatical elements are joined together in such a way that

segmentation is relatively easy.

ev „house‟

evler „houses‟

evlerim „my houses‟

evlerimde „in my houses‟

evlerimdeki „that which is in my houses‟

evlerimdekiler „those which are in my houses‟

Except for two prefixal processes involving reduplication, Turkish is exclusively

suffixing. The inflectional suffixes may be divided into two groups, a noun paradigm and

a verb paradigm. The elements of the noun paradigm, in order, are:

1. Noun Stem.

2. Plural –lEr.

3. Possessive.

11

4. Case: objective-(y)I, genitive –(n)In, dative –(y)E, locative –DE, ablative –DEn.

5. Relative –ki. This suffix is added only to genitive or locative suffixes. The result is

a noun stem and one may start over

All the elements of the noun paradigm except (1), the noun stem, are optional.

However, if one includes the nominative among the cases, with a zero ending, then

element (4), case, is obligatory. A new comitative/instrumental case is in the process of

formation by the suffixation of a postposition meaning “with” to noun stems: Hasan ile

‘with Hasan‟ Hasanla.

The verb paradigm is more complex:

1. Verb Stem.

2. Derivation: reflexive –In, reciprocal –Iş, causative –Dır, passive –Il. The reflexive

and reciprocal are obviously mutually exclusive; otherwise, almost any

combination of these suffixes is possible, but they must appear in the indicated

order. The causative may be repeated.

3. Impossible –(y)E and –mE. The negative may occur alone, or in combination with

the impossible: gelmiyorum „I am not coming‟, gelemiyorum „I cannot come‟.

4. Tense. This element is obligatory for a finite verb. For a non-finite verb in a

subordinate clause, a nominal, participial, or adverbial suffix replaces the tense.

There are eight tenses, not all of which are tenses in the strict sense of the term:

aorist –Ir, progressive –Iyor, definite past –DI, narrative past –mIş, future

–(y)EcEk, optative –(y)E, necessitative –mElI, condtional –sE. Of particular

interest in the tense system is the contrast between the definite past and the

narrative past. The definite past is used to describe events which the speaker has

personally witnessed, while the narrative past is used for actions about which the

speaker knows through report or inference. As a result, the narrative past is the one

normally used for narration, although the definite past is used for historical events

which are well established. (Tense/aspect: F. Yavaş 1980, 1982a-b.)

5. Auxiliary. Four suffixes may be added both to verbal and non-verbal predicates.

Their syntactic status is discussed fuehrer under „Syntax‟. These include: past –

(y)DI, dubitative –(y)mIş, conditional –(y)sE, adverbial –(y)ken. The dubitative

auxiliary –(y)mIş is distinguished phonologically, syntactically, and semantically

from the narrative past tense –mIş. It is used when the speaker wishes to express

doubt, uncertainty, or surprise about the truth value of the statement: sen

tembelsin „you are lazy‟, sen tembelmişsin „they say you are lazy‟; bu oda meşgul

12

„this room is occupied‟, bu oda meşgulmüş „this room seems to be occupied‟.

(Dubitative: Slobin&Aksu 1982.)

6. Person. This is also obligatory for a finite verb form. The minimal elements of a

finite verb, then, are (1) stem, (4) tense, (6) person.

Just as Turkic languages can create new noun cases by the suffixation of postposition

onto the noun, they can create new verb tenses by the reduction of periphrastic

constructions. For example, the “abilitative” suffix –(y)Ebil, as in gelebilirim „I can

come‟, is derived historically from an adverbial form of the main verb with a suffix

–(y)E, plus the auxiliary bil- „know‟. Several similar constructions are in the process

of formation.

5 Syntax

There are two types of predicate in Turkish: verbal, consisting minimally of a verb,

and non-verbal, consisting of something other than a verb, normally a noun phrase or

adjective.

(Ben) yorgun –um.

I tired -1SG

„I am tired‟

(Ben) gel -iyor -um.

I come PROG 1SG

„I am coming.‟

If the subject is a pronoun, it is normally omitted: Yorgunum „I am tired‟, Geliyorum ‘I

am coming.‟ (That is, Turkish is a Pro-Drop language.) Subject pronouns may be retained

for emphatic or contrastive purposes: Ben geliyorum, sen gelmiyorsun „I am coming, you

are not coming‟. The predicate is followed by a personal ending, here 1sg. –(y)Im.

The personal ending may be preceded by one of three suffixes: past –(y)DI, dubitative

–(y)mIş, conditional –(ysE. These may be added to predicates of all types: if the predicate

is a verb, they follow the verbal tense suffix:

13

(Ben) yorgun -du -m.

I tired PAST 1SG

„I was tired‟

(Ben) gel -iyor -du -m.

I come PROG PAST 1SG

„I was coming‟

We have adopted the term “auxiliary” to describe those suffixes, including personal

endings, which are added to predicates of all types. They are to be distinguished from

suffixes such as tense which are added to verbs only.

The unmarked word order is SOV (subject-object-verb):

(Ben) mektub -u gönder -iyor -um.

I letter OBJ send PROG 1SG

„I am sending the letter.‟

Ahmet mektub -u gönder -iyor.

Ahmet letter OBJ send PROG

„Ahmet is sending the letter.‟

However, word order is in fact pragmatically controlled and thus appears highly fluid to

English speakers. There are two basic principles. The position immediately to the left of

the verb is the focus position:

Şimdi mektub -u gönder -iyor -um.

Now letter OBJ send PROG 1SG

„Now, I am sending the letter.‟

Mektub -u şimdi gönder -iyor -um.

Letter OBJ now send PROG 1SG

„I am sending the letter now.‟ (or right now)

14

Mektub -u Ahmet gönder -iyor.

Letter OBJ Ahmet send PROG

„Ahmet is sending the letter.‟

Certain types of constituents obligatorily appear in focus position: these include indefinite

object (which do not have the objective suffix), wh-words, and when other principles do

not obstruct, indefinite subjects. (Indefinite movement: Underhill 1972.)

Ahmet mektub yaz -ıyor.

Ahmet letter write PROG

„Ahmet is writing letter(s).‟

Mektub -u kim yaz -dı?

Letter OBJ who write PAST

„Who wrote the letter?‟

San -a (bir) mektup gel -di

You DAT a letter come PAST

„A letter came for you.‟

Secondly, constituents may “leak” to the right of the verb (thus violating verb-final

order); post-verbal position is used for backgrounded, presupposed, or afterthought

information.

En sonunda iş bul -du Ahmet.

-at last- job find PAST Ahmet

„At last Ahmet found a job.‟

(Where Ahmet is already “given” information in the context);

Şimdi gönder -di -m mektub -u.

Now send PAST 1SG letter OBJ

„I just now sent (it), the letter.‟

15

This order is particularly favored for imperatives:

Bas -ma kedi -ye (kediye basma).

Step NEG cat DAT

„Don‟t step on the cat!‟

Ver -me çocuğ -a kibrit -ler -i.

Give NEG child DAT match PL OBJ

„Don‟t give the child the matches!‟

(Pragmatic control of the word order: Erguvanlı 1979a.)

The relative fluidity of the word order compared with regularity and transparency of

the morphology produces an interesting result in children‟s acquisition of Turkish. While

English-speaking children key very early on word order strategies, Turkish children key

very early on morphology and essentially ignore word order. In tests, basic SOV

sentences were scrambled into all possible permutations and Turkish children almost

unerringly interpreted them correctly. (Acquisition of Turkish: Slobin 1982, Aksu-

Koç&Slobin 1985, Slobin&Bever 1982.)

Turkish has the canonical properties of an SOV or left branching language, as

suggested by Greenberg. For example, it uses postpositions rather than prepositions.

Nominal modifiers, including genitives, precede the noun. A special property of the

Turkish noun phrase is that a possessed noun must have a suffix agreeing with the

possessor:

Hasan -ın kedi -si.

Hasan GEN cat 3SG POSS

„Hasan‟s cat‟

(Biz -im) kedi -miz.

Our GEN cat 1SG POSS

„Our cat‟

Relative clauses also precede the noun they modify. Relative clauses are made by

replacing the verbal tense suffix with a participle suffix, and deleting the repetition of the

head noun in the lower sentence. There are two types of participles, and the choice

between them is highly complex. The subject of the lower sentence, if it is not deleted,

16

becomes genitive and the corresponding possessive is added to the participle. (Relative

Clauses: Underhill 1972, Hankamer&Knecht 1976.)

Çocuk mektub -u oku -yor.

Child letter GEN read PROG

„The child is reading the letter.‟

Mektub -u oku -yan çocuk

Letter GEN read PART child

„The child who is reading the letter.‟

Çocuğ -un oku duğ -u mektup

Child GEN read PART 3SG POSS letter

„The letter which the child is reading.‟

Embedded nominal complement constructions are made in a similar manner. The tense

suffix of the verb is replaced with one of a number of nominal suffixes. The subject of the

complement, if it is not deleted, is marked genitive, and the corresponding possessive

suffix is placed on the nominalized verb. The nominalized verb then gets whatever case

suffix is appropriate for its function in the higher sentence.

(Biz) Hasan -ın mektub -u gönder -diğ -in -i bil -iyor -uz.

We Hasan GEN letter OBJ send NOM 3SG OBJ know PROG 1PL

„We know that Hasan sent the letter.‟

(Biz) Hasan -ın mektub -u gönder -me -sin -i iste -mi -yor -uz.

We Hasan GEN letter OBJ send INF 3SG POSS OBJ want NEG PROG 1PL

„We don‟t want Hasan to send the letter.‟

6 Sources

The most useful comprehensive grammars are Kononov 1956, Swift 1963, Lewis

1967, and Underhill 1976. Kononov and Lewis are traditional grammars but excellent

17

data sources, Swift is a structuralist, and Underhill is a pedagogical grammar but

generative in orientation.

In phonology, Lees 1961, a comprehensive phonological description, is outdated but

still useful; is should be used in conjunction with Zimmer‟s review (1965). Foster 1969 is

a good overview in classical phonological terms, and Lightner 1972 goes over many of

the main problems in phonology although it should be used with great caution. Vowel

harmony is the most popular topic in Turkish linguistics and no attempt is made here to

list the many references dealing it. Other popular phonological topics include stress and

the deletion of velars, both cited earlier in the text.

Good overviews of Turkish syntax from a classical or standard theory viewpoint may

be founded in Hankamer 1971, Lees 1972, Aissen 1974a-b, Knecht 1976, and Kornfilt

1976. A Government and Binding approach is taken by George&Kornfilt 1981. Turkish

has played a major role in the development of Relational Grammar: Perlmutter 1978,

Özkaragöz 1980b, Gibson&Özkaragöz 1981. Some interesting papers from a

functionalist viewpoint are Kuno 1971, 1980a-b, and Kornfilt, Kuno&Sezer 1980. A

variety of other syntactic approaches, including lexicalist, interpretive, Montague and so

on may be found in the bibliography (next chapter)

As footnote 2 points out, only the most basic or most important sources have been

cited here for each topic. Anyone wishing to pursue any of these topics in greater depth

should search through the bibliography for additional references.

NOTES

1) As an example, the predicative suffix which can have the phonetic forms –dir/dır/dur/dür/tir/tır/tur/tür was normally written

in Ottoman with the single form د ر.

2) References are to the article “Bibliography of Modern Linguistic Work on Turkish” in this volume. For most topics, only the

most important references are given; those wanting to pursue any particular area will usually find additional references by

searching through the bibliography. Some general sources are discussed at the end of this article.

3) This requires that the distinction between dotted and undotted i must be made for capitals also: I vs İ

4) The dialects in which ğ is pronounced [ɣ] have only the four long vowels in loans, for a total of twelve. In another group of

dialects, ğ behaves as we have described it only in the environment of back vowels, while it becomes /y/ in the environment

of front vowels

5) In citing morphemes, we write a consonant in parentheses if the consonant is present when the suffix is post-vocalic, absent

when it is post-consonantal. Thus the correct form of the objective is –(y)I.cf. evi ‘house (obj)‟, deveyi „camel (obj)‟. We

also write D for suffix-initial consonant that surfaces as t after unvoiced consonant , d elsewhere.

6) Or a few consisting of unvoiced fricative plus stop.

7) Syllable-final voiced stops are permitted morpheme-internally in loans: tedbir „precaution‟ (Arabic): radyo