-
Notifications
You must be signed in to change notification settings - Fork 2
Morphological PostProcessing
Turkish is an agglutinative language, where complex words are formed by adding multiple suffixes. Furthermore, letters in the root word or suffix can change depending on preceding or succeeding letters. These rules are called Vowel and Consonant Harmony. However there are always some exceptions violating these rules (especially borrowed words from other languages). In order to accommodate these exceptions, some characters are changed or injected in root or suffixes.
Vowel Harmony:
^ : make previous vowel back, i.e. aıou->eiöü e.g. kol-A -> kola, rol^-A -> role
H : high vowels, using harmony rules aı->ı, ei->i, uo->u, üö->ü e.g. ev-Hm -> evim, yol-Hm -> yolum, gül-Hm -> gülüm
A : low,unrounded vowels, using harmony rules aıuo->a, eiüö->e e.g. ev-lAr -> evler yol-lAr -> yollar
Consonants:
N : used as last char of same suffixes (e.g. 3d sing possesive), realized to n only if it is followed by a vowel e.g. tereyağıN-ı->tereyağını, tereyağıN-ım ->tereyağım
@ : drops the next vowel if it takes a vowel-starting suffix e.g. bur@un-u -> burnu, bur@un-da -> burunda
+ : duplicates previous letter if it takes a vowel-starting suffix e.g. hak+-ı -> hakkı, hak+-ta -> hakta
? : softens of previous consonant e.g. kitap?-ı -> kitabı, hap-ı -> hapı
Note that ^, @, +, ?, N symbols can be used in word root