UNIT 1 – WHY STUDY SOUNDS?

PHONETICS AND PHONOLODY

• The domain of both phonetics and phonology is speech.
• Phonology: It deals with how speech is organized.
• Phonetic: It is concerned with the units of speech.
• Phonetics tries to arrive to a physical description of sound.
• Both phonetic and phonological competents of the grammar has to be referred to when
describing
speech as there are two levels of representation is speech. –phonological and phonetic
representation-
• In the phonological competent of the grammar the sounds are stringed together where
phonological rules apply. Therefore, phonological representation is more abstract.
• The output of the phonological competent is fed into the phonetic competent.
• The output of the phonetic competent is the realization of speech (the pronunciation of a
sentence).
• Ex. A suffix may have more than one phonetic realization.
• Take the plural suffix in English which is represented as “-s”. It may be realized as “s” (cats),
as “z”
(dogs), “əz” (watches).
• The plural suffix in English has three different pronunciations.
• Thus, the plural suffix is represented /s/ in the phonological competent, but it is realized as
[s] or
[z] or [əz] phonetically.
• The underlying (phonological) representation of the plural is /s/ while the phonetic
(surface) realization is [s], [z] or [əz].
• The Turkish plural suffix has two different surface realization, -ler(as in ev-ler) and –lar(as in
dolap-lar).
• We know that the suffix – çü and – cı are the same suffix although these two forms are
different
phonetically (the same is true for the suffixes – ler/-lar and –den/-dan), and that these
differences
in the form are due to in the roots in terms of the sounds each root contains.
• The underlying representation is notated between slashes (//).
• The surface realization is notated between square brackets ([]).
• The differences in the surface realization are systematic regularities in that they are
predictable.
They are predictable because they are rule governed.
• As the root-final sound of chair is a voiced consonant, the plural suffix is realized as [z].
Because
this is a grammar rule.
• Similarly, a native speaker of Turkish knows that the plural suffix of, say, ‘masa’ would be
realized
as ‘-lar’ as in ‘masalar’, but the same suffix would be realized as ‘-ler’ in ‘sebzeler’. The choice of
the appropriate realization is determined by the last vowel in the root, therefore it is
predictable.
• Some properties are idiosyncratic in that they are not predictable. The fact that the initial
sound
of ‘sap’, for example, is an idiosyncratic property of that word as there is no way of predicting
that
particular word should have ‘s’ as the initial sound. The same word could have been ‘rap’, ‘tap’,
‘map’, etc. and a native speaker would have no way of knowing which sound would occur in the
initial position for that particular word.
• Such sounds which cause a change in the meaning are called phonemes. If we take ‘sap’ as an
example again, changing the initial sound to ‘k’ would result in a different word ‘kap’. ‘Kap’ and
‘sap’ are two separate words with different meanings. This difference is not predictable,
therefore
an idiosyncratic property of these words.
• Pairs such as ‘kap’ and ‘sap’ are called minimal pairs.
• ‘Kel’ and ‘kal’ are also the minimal pairs showing that /e/ and /a/ are phonemes. In each of
these
pairs, a change in one sound, the other sounds being the same, changes the meaning of the word.
Thus those sounds are phonemes as they change the meaning.
• In Turkish, /k/ has two different surface realizations or two variations, [k] and [c]. ([k] kalın
/k/
ex. kal, kıl etc. / [c] ise ince /k/dir. Ex. kel, kil etc.)
• Such variations which do not change the meaning are called allophones. Allophone is then a
variation of a sound which does not cause a change in the meaning of a word. Allophones are
generally predictable.
2
• In Turkish, the occurrence of [k] and [c] is determined by the vowel. The vowels in the first
column
are called ‘front vowels’ thus the formation of /k/ is affected and is fronted. The vowels in the
second column are called ‘back vowels’ and with the effect of the vowels /k/ is formed further
back.
SPEECH CHAIN
• The speech process is comprised of a speaker and a listener between which an acoustic signal is
transmitted. The speech process is a very complex system. To produce speech sounds, neural
commands are sent to muscles and as a result, muscles contract. In speech, precise and rapid
coordination of more than eighty muscles are used. Air, the source of sound, is pushed out of the
lungs with the contraction of muscles into the vocal apparatus. The air that passes through the
vocal tract gains different qualities as a result of different shapes of the vocal tract. The
sounds are
transmitted through air which is an acoustic signal. That signal then reaches the listener’s ear
where the nerves in the ear are put into motion. Those nerve pulses are then translated into a
linguistic message, which then ends up in the brain.
• As there is a speaker and a listener and an acoustic signal produced by the speaker and
received
by the listener, sounds are described in three different ways. The three different descriptions
of a
sound are,
• Speaker-based description
• Acoustic description
• Listener-based description
• Speaker-based description of sounds describes sounds in terms of the speaker. This in itself
can be
described in terms of the muscles involved in speech production, the airflow through the vocal
tract, and articulation (the movement of organs in the vocal tract).
• Acoustic description is considered with the sound waves that are transmitted in the air. Within
acoustic description the acoustic properties of speech sounds such as the waveforms, duration.
• Listener-based description involves the reception of sound waves in the air, the neurons
responsible for receiving these waves, then translation of the nerve pulses into linguistic
message.
SOUND vs. LETTER
• IPA (International Phonetic Alphabet) is the tradition within linguistics. The reasons IPA,
rather
than letters, is used to represent sounds are two fold.
• One, letters are language specific. Each language may use different letters to represent the
same
sound. For example, the initial sounds of the Turkish word “sair” and the English word “sheep”
are
similar. But the two languages represent these similar sounds by using different symbols or
letters;
“s” in Turkish, “sh” in English.
• When IPA is used, a symbol is used to represent the same sound in all languages. If the same
letter represents different sounds in different words, then different IPA symbols representing
those
sounds are used.
• The second reason why IPA is used to represent sounds is that we want to be able to represent
all
the sounds that occur in the languages of the world. As letters are language specific, letters of
a
specific language cannot represent sounds that do not occur in that language.
UNIT 2 - TURKISH CONSONANTS
CONSONANT DESCRIPTION
Consonants are described in terms of
• Voicing (presence or absence of vocal cord vibration)
• Place of articulation (where in the vocal tract sounds are produced)
• Manner of articulation (how sounds are produced in terms of airflow)
1. Voicing:
• Voicing is defined as the presence or absence of vocal cord vibration. Vocal cords are two
muscular folds within the larynx behind the Adam’s apple. You can manipulate your vocal
cords by opening and closing them. The vocal cords are open when breathing. You can also
close your vocal cords. For example, you say ‘a-a’ when you are surprised. Stop after you
make the first ‘a’. Try breathing. You should not be able to because you have closed your
vocal cords.
• As the air passes between the vocal cords, vocal cords may be apart or close together. If
the vocal cords are apart, the air passes through the vocal cords freely. If the vocal cords
are close together, then there is a narrow passage between them which causes a build up of
air pressure.
• This air pressure in turn causes vocal cords to vibrate.
• If the vocal cords vibrate in the production of sounds, then those sounds are voiced.
• If the vocal cords do not vibrate, then those sounds are voiceless.
• For example, /s/ is voiceless as there is no vocal cord vibration; /z/ is voiced as there is
vocal cord vibration.
• The consonants in “fıstıkçı sahap” are all voiceless. The others are voiced.
2. Places of Articulation
• Consonants are described in terms of where in the vocal tract they are produced:
• Bilabial: two lips together. ‘b’, ‘p’ and ‘m’ (bellek, pınar, masa)
• IPA symbols for these sounds are /b/, /p/, /m/.
• Labiodentals: lower lip and upper front teeth. ‘f’ and ‘v’ (fare, vali)
• IPA symbols for these sounds are /f/, /v/.
• Alveolar: tongue tip or blade and the alveolar ridge. ‘d’, ‘t’, ‘s’, ‘z’, ‘n’, ‘l’ and ‘r’
(sezenlerde dertli)
• IPA symbols for these sounds are /d/, /t/, /s/, /z/, /n/, /l/, /r/
• Palato-alveolar: tongue blade and back of the alveolar ridge. ‘s’, ‘j’, ‘’4 and ‘c’
• IPA symbols for these sounds are / ȓ /, /Ȣ/, / ȷ /, /ȴ/
• Palatal: front of the tongue and the hard plate. Only ‘y’ (yıl, ye)
• IPA symbols for this sound is /Ј/.
• Velar: back of the tongue and the velum (or the soft palate). ‘k’ and ‘g’
• IPA symbol for these sounds are /k/ and /g/.
• Glottal: articulation involving the glottis. Only ‘h’ (han, hamam)
• IPA symbol for this sound is /h/.
3. Manners of Articulation
• How sounds are produced is referred to as manner of articulation.
• Stop: The air is stopped somewhere in the oral tract so that the airstream cannot escape
from the mouth. The velum is raised so that air cannot escape from the nose either.
• Stop sounds of Turkish are /p/, /b/, /t/, /d/, /k/, /g/.
2
• Nasal: The velum is lowered so that the air escapes through the nose. Notice that when
you produce /b/ you cannot hear a sound until you open your lips. That is because you stop
the airstream by closing your lips so that air cannot escape from your mouth. When you
produce [m] on the other hand, although your lips are together not letting air go out your
mouth, you can hear the sound [m].
• Nasal sounds of Turkish are /m/ and /n/.
• Fricative: In the production of fricatives two articulators are close together so that a
narrow passage is formed between the two articulators. As the air goes through that narrow
passage, turbulence is produced causing friction.
• Fricative sounds of Turkish are /f/, /s/, /z/, /ȓ/, /Ȣ/, /h/.
(s), (j)
• Approximant: To produce an approximant, two articulators are close together but not so
close together to form a narrow passage. As the passage is not narrowed to an extend that
a turbulent airflow is produced, the airstream flows more freely without causing friction.
• Approximants in Turkish are /v/, /Ј/ and /l/.
(y)
• These approximants are of two types depending on their airflow.
a. Central approximant: air flows over the center of the tongue.
/Ј/ and /v/ are central approximants.
(y)
b. Lateral approximant: with a central obstruction, air flows over the sides of the tongue.
Only /l/ is a lateral approximant.
• Flap/tap: The Turkish ‘r’ (/Ȏ/) is a flap whereby the tip of the tongue makes a single tap
against the alveolar ridge. Notice what you do with your tongue when you say the word
‘ara’. The tongue rapidly taps against the alveolar ridge. (genellikle kelimenin ortasındaki
‘r’de “tap” g・ ・)
• Flap also occurs in American English in the medial position of words such as latter and
ladder, causing both words to be pronounced the same way.
• Affricate: The production of affricates involves a stop and a fricative. The affricate starts
with a stop which is immediately followed by a fricative. Say the words 軋y and cam.
• The affricates of Turkish are /ȷ/ and /dȢ/.
(5) (c)
• In the description of Turkish sounds, we have not discussed ‘ğ’, what is called ‘soft g’ in the
literature.
• ‘ğ’ is a letter that occurs in the Turkish orthography (the writing system). But there is no
distinct corresponding sound. Many agree that ‘ğ’ has the function of lengthening the vowel
preceding it.
• The difference between the members of the pairs such as ‘ada-ağda’, ‘tura-tuğra’, ‘yamayağma’,
etc., is the length of the vowels preceding ‘ğ’. Thus, ‘ğ’ is not treated as a distinct
consonant here.
• English has consonants that do not occur in Turkish such as /θ/, /./, /w/, /ȉ/, /ŋ/, /v/.
• /θ/ and /./ are interdental fricatives produced with the tongue tip between the teeth as in
the words ‘thumb’ and ‘those’ respectively.
• /w/ is a labiovelar central approximant produced at the velum with the lips rounded.
• /ȉ/, the initial sounds of words, ‘right’, ‘real’, ‘real’, ‘rock’ is an alveolar central approximant
‘r’.
• /ŋ/ is a velar nasal produced by a complete obstruction in the oral tract by raising the back
of the tongue to the velum. The words ‘sing’, ‘rang’, ‘doing’ have [ŋ] as the final sound.
• The letter ‘v’ represents two different sounds in Turkish and English. Although the place of
articulation is the same in the two languages, ‘v’, is a central approximant in Turkish but a
fricative in English.
3
ALLOPHONES OF TURKISH CONSONANTS
[k] with (a, ı , o, u) _ voiceless velar stop (kalın ‘k’)
k
[c] with (e, i, 5, 5) _ voiceless palatal stop (ince ‘k’
________________________________________________________________________
__________
[g] with (a, ı , o, u) _ voiced velar stop
g
[†] with (e, i, 5, 5) _ voiced palatal stop
________________________________________________________________________
__________
[n] elsewhere _ voiced alveolar nasal
n
[ŋ] before velar consonant _ voiced velar nasal ( renk, denge…)
________________________________________________________________________
__________
[Ȏ] inter (ara, sıra…) _ voiced alveolar flap
r [ř] initial (resim, r・a…) _ voiced, fricated alveolar, flap
[ř] final (bir, her…) _ voiceless, fricated alveolar, flap
˚
__________________________________________________________
[v] after voiceless consonants (e.g., cetvel, tesvik)
v [∅] before labial consonants (e.g. , kovmak, smek)
 (‘v’ ı輒mıyor gibi)
[ț] elsewhere
________________________________________________________________________
__________
[l] after e, i, 6, 6 and initial (bele, sel, kel, lamba, lokum…) ince ‘l’
l
[Ǽ] after a, ı, o, u _ voiced (kal, kıl, kola, kul…)
TURKISH CONSONANTS
Bilabials Labiodentals Alveolar Palatoalveolar
Palatals Velars Glottal
-V +V -V +V -V +V -V +V -V +V -V +V -V +V
Stop p b t d k g
Nasal m n
Fricative f s z s j *h
Approxi central *v y
mant
lateral l
*r
Flap/tap
çc
Affricate
ENGLISH CONSONANTS
(*) Consonants which have differences between Turkish and English consonants.
Bilabials Labiodentals *dentals Alveolar Palatoalveolar
Palatals Velars Glottal
-V +V -V +V -V +V -V +V -V +V -V +V -V +V -V +V
Stop p b t d k g
Nasal m n
Fricative f *v ȅ ğ s z s j
Approx glides *w y
imant
liquids l
*r
*h
Affricate çc
UNIT 3 – TURKISH VOWELS
VOWEL PRODUCTION
• Vowel production is a continuum whereby there are no clear boundaries of tongue position for
different vowels. Because vowel production is a continuum, it is possible to produce different
vowels between any two vowels.
• Produce the vowels presented with the letters “i” and then move to “ı”. Notice that the
movement
from one vowel to another changes the auditory quality of vowels.
• The change in the quality of the vowels is a result of the movement of the tongue and the lips.
• All vowels are voiced.
VOWEL DESCRIPTION
• Vowel description is based on auditory quality of the vowels because the position of the tongue
is
not clear in the production of vowels.
• Traditionally, vowels are described in terms of
• Height
• Backness
• Lip position
• Vowels are described in terms of auditory quality rather than the position of the tongue. The
terms
‘high-low, front-back’ are just labels referring to auditory quality rather than tongue position.
• These terms are used to describe how vowels sound relative to other vowels.
• The reference point is the vowel ‘schwa’ [ə]; also called a neutral vowel because it is produced
with
the tongue at rest position. Vowels are said to be higher or more back than [ə].
HEIGHT
• For the pronunciation of high vowels the tongue is raised towards the roof of the mouth.
• When the tongue is lowered, to a lower position than that of [ə], low vowels are produced.
• If the tongue height is similar to that of [ə], then the vowel is a mid vowel.
• The initial sounds of the following words contain high vowels ‘isim’, ‘ırmak’, ‘・ ・,
’ ‘umut’. The
IPA symbols for these vowels are /i/, /ǿ/, /y/, /u/
(i), (ı), (7), (u).
• The following words have mid vowels in word-initial position, ‘ekmek’, ‘・’, ‘orman’. When
compared to [ə], the height of the tongue for the mid vowels is the same as that of [ə], which
is
also a mid vowel.
• The IPA symbols for the mid vowels are /Ǭ/, /oe/, /o/
(e), (7), (o).
• Turkish has only one low vowel – ‘a’ as in the words ‘abla’ and ‘aslan’.
• The IPA symbol for ‘a’ is /a/.
BACKNESS
• Backness refers to the position of the highest position of tongue body in the vocal tract.
• Height refers to how high or low the tongue is in reference to [ə].
• Backness has three degrees: Front, central, back.
• Front vowels are /i/, /y/ and /Ǭ/--
(i), (7), (e)
• There is only one central vowel which is [a]
• Back vowels are /ǿ/, /u/ and /o/--
(ı), (u), (o)
• Kitap “”7 ye bir grup bulamamıs ama sanırım (/oe/) “”7 de front vowel. ☺
2
ROUNDING
• In the production of some vowels, the lips are protruded or rounded to form what is called
rounded vowels.
• For some vowels the lips are not protruded, rather they are spread. Such vowels are called
unrounded vowels.
TENSE-TAX
• Turkish distinguishes vowels in terms of height, backness, and rounding. English, for example,
has
tense-lax distinction.
• Tense-lax is a cover term to distinguish pairs of vowels such as /i/ in the word ‘eat’ and /ı/ in
the
word ‘if’ in English.
• Vowel length – tense vowels are longer
• Tongue position – tense vowels have more extreme tongue position
• Lip position – tense vowels have more extreme lip position
• All back vowels are rounded, all rounded vowels are back. Inversely, all non-back vowels are
unrounded, all unrounded vowels are non-back. (YA KĐTAP YANLIS YADA BEN ANLAMADIM!!
DIPHTHONGS
• Diphthong is a vowel in which there is a change in quality during a single syllable.
• Thus, it is a vowel which starts as one vowel but ends as a different vowel within the same
syllable.
• There are 3 diphthongs in American English:
• /ai/ my, lie, die, rye
/au/ cow, how, now
/oi/ boy, toy, royal
• Turkish does not have diphthongs.
• Although the Turkish word ‘ay’ and the English word ‘eye’ are similar in their pronunciation, the
Turkish word does not contain a diphthong while the English word does. The Turkish word is
transcribed as /aj/, a vowel and a consonant, the English word is transcribed as /ai/ with a
diphthong. The word ‘ay’ when added the suffix –ı becomes ‘a-yı’, the sound ‘y’ goes to the next
syllable.
ALLOPHONES OF TURKISH VOWELS
[i] before “ğ” _ tense (uzun) _(iğne)
/i/
[ı] elsewhere _ tax (kısa) _ (isim)
[u] before “ğ” _ tense _ (tuğra)
/u/
[Ț] elsewhere _ tax _ (tura)
[e] before “ğ” _ tense _ (eğer)
/ǫ/ [訃 speaker-dependent_ (gen8)
[ǫ] elsewhere _ tax_ (gen8)
[a] speaker-dependent_ (para)
/a/ [ƽ] in fist syllable _ (masa, halı)
[Ȝ] elsewhere _ (masa, sıra)
UNIT 4 - PHONOLOGICAL PROCESSES IN TURKISH
PHONOLOGY
• All languages have rules about how speech is organized. These rules are language specific in that
each
language has its own phonological rules and a phonological rule may occur in one language while not in
other languages.
• Knowing a language means knowing which sound sequences are permissible and which sound sequences
are not permissible.
• For example while the following nonsense words (words that do not have meaning in Turkish) ‘cunur’,
‘kambat’ and ‘sortunk’ are acceptable in Turkish; ‘skinez’ and arktrat’ are not acceptable according to
Turkish phonology.
• “Cunur” could be a Turkish word because all the sounds in it occur in Turkish.
• Similarly, because the sounds and the order of the sounds comply with the rules of Turkish,
“kambat”
could be a Turkish word also.
• “Sortunk” doesn’t seem like a Turkish word at the first glance because of the word-final “-nk”. But
that
is permissible combination word-finally as it occurs in the words “renk, denk…”.
• “Skinez” and “arktrat” can’t be Turkish words because both have non-permissible combination.
• “Skinez” has two consonants together without an intervening vowel (consonant cluster) in word-initial
position which Turkish doesn’t allow. Turkish has borrowed words such as “spor”. It is pronounced as
two syllables with a vowel in between in two consonant “sı-por”, thus not breaking the rules of Turkish.
• “Arktrat” also has a consonant cluster in syllable-initial position. In the first case, there are three
consonant together syllables initially. In the second case, there are consonant clusters both in
syllablefinal
and syllable-initial position.
• Syllable-final ‘-rk’ sequence is permissible as it occurs in words such as ‘T・k’, ‘fark’, ‘kork’.
• But, syllable-initial consonant clusters are not allowed in Turkish.
FINAL DEVICING
• Study the words ‘sahab’, ‘alad’, ‘sorag’ and ‘akac’. These words are not acceptable in Turkish.
• The reason why these words cannot be acceptable in Turkish is that Turkish does not allow voiced
stops
(b, d, g) and affricate (ȴ) to occur in word-final position.
• This is a rule called Final Devoicing Rule which states that voiced sounds become voiceless in word or
syllable final position.
• In Turkish, final devicing rule is stipulated as voiceless stops and affricatives becoming voiced when
a
suffix beginning with a vowel is attached.
• “p, 9, t, k” _ “b, c, d, g” (e.g. ağaç_ağac-ı, kitap_kitab-ı)
• However, this rule doesn’t account for words in which voiceless stops don’t become voiced when
vowel
initial suffix is attached. (e.g. sap_sap-ı, T・k_ T・k-9)
• In phonology a distinction between underlying representation (UR) and surface realization (SR) is
made
to explain such phenomena.
• Underlying representation means that a word may have a different form than its actual realization
(i.e., how the word is said). Surface realization, on the other hand, refers to how it is actually said.
• UR_ -i_(SR) singular plural
/kȜb/ [kȜbǿ] [ kȜp ] [kȜpǼȜř]
˚

/sȜp/ [sȜpǿ] [sȜp] [sȜpǼȜř]

˚

• The second column in which the suffix –i is attached illustrate the underlying voicing of the final
stops.
ASSIMILATION
• Assimilation is a phonological process whereby one sound becomes like the neighboring sound.
_ Place of articulation
• Notice that in “Đstanbul” the alveolar nasal /n/ preceding the bilabial stop /b/ is pronounced as a
bilabial nasal /m/.
• While producing /n/ we begin to for /b/ for which lips come together, resulting in /m/.
• Place of articulation means that the place of articulation of one sound becomes like the
neighboring sound.
• Also, when /n/ is followed by a velar stop /k/ or /g/, /n/ is realized as /ŋ/. (e.g. Ankara)
/ŋ/
2
• Place of articulation assimilation in Turkish is regressive (from right to left; going backward) in
that the consonant following the nasal determines the place of articulation.
• Ex: Kanbur _ /kȜnbuȎ/ _ [kȜmbȚř]
°

Sandık _ /sȜndǿk/ _ [sȜndǿk] (değisiklik yok)

Kangal _ /kȜngȜǼ/ _ [kȜŋȜǼ]
_ Voicing Assimilation
• Voicing articulation is a process whereby the voicing of a consonant becomes similar to what of
the neighboring consonant.
• Ex: ev_ ev-de, yurt_ yurt-ta
• When the suffix _de is attached to a root two different forms with respect to initial consonant
‘_de’ and ‘_te’ surface.
• The word in the first example above end with a voiced sound /v/. The initial consonant of the
suffix _de therefore is also voiced.
(voiced + voiced)
• The word in the second example above end with a voiceless sound /t/. the initial consonant of
the suffix _te is also voiceless.
(voiceless + voiceless)
• There are a number of stop-initial or affricative-initial suffixes in Turkish which exhibit voicing
assimilation. (e.g. fırın-cı, 輟rap-ı…10)
• But not all stop and affricate-initial suffixes assimilate in voicing. (ex, “-ken, -sa..”)
Ex: af-ken (voiced + voiced), av-ken (voiceless + voiced)
• Suffixes with initial consonants other than stops and affricates don’t assimilate in voicing (ex, “-
lar”)
• Voicing assimilation is progressive. It is the root-final sound that determines the voicing of the
alternating suffix-initial consonant.
VOWEL HARMONY
• Vowel harmony can be considered a type of assimilation since neighboring vowels.
• Ex: is-ler, kuzu-lar
ev-ler, muz-lar
• The root-vowels in the first group of are / Ǭ, i, oe, y/ - (e, i, 10, 10). The property these vowels
share that
are all front vowels. Thus the suffix –ler, which also contains a front vowel, is attached. (front +
front)
• The vowels in the second group, /a, ǿ o, u/ - (a, ı, o, u) are all non-front vowels. Thus the plural
suffix,
has the form –lar, with a non-front vowel. (non-front + non-front)
• In Turkish, the rules is when a word contains more than one syllable, it is the vowel of the last
syllable
which determines the backness of the suffix vowel. (e.g. kalem-ler, selam-lar)
• There are two types of harmony.
1. If the vowel of the suffix is not a high vowel, then it agrees in backness. This is called backness
harmony.
2. If the vowel of the suffix is high vowel, then is agrees in both backness and rounding. They are
called rounding harmony.
_ (ı, i, u, 10) değilse backness; (ı, i, u, 10) ise rounding harmony olur.