Command for Turkish Genitive Suffix

Question

In English, the genitive suffix ’s never changes its shape. For example, the genitive suffix ’s does not change its shape in the expressions Alice’s bicycle or Frank’s bicycle. In Turkish, however, the genitive suffix ’{n}{i}n changes its shape, where {n} is a variable depending on the last letter of the word to which the genitive suffix is added, and {i} is another variable depending on the vowel in the last syllable of the word to which the genitive suffix is added.

Here are some examples:

• Ezgi’{n}{i}n -> Ezgi’nin [Ezgi's]

• Sıtkı’{n}{i}n -> Sıtkı’nın [Sıtkı's]

• Utku’{n}{i}n -> Utku’nun [Utku's]

• Hüsnü’{n}{i}n -> Hüsnü’nün [Hüsnü’s]

• Nefel’{n}{i}n -> Nefel’in [Nefel's]

• Anıl’{n}{i}n -> Anıl’ın [Anıl's]

• Ufuk’{n}{i}n -> Ufuk’un [Ufuk's]

• Gönül’{n}{i}n -> Gönül’ün [Gönül's]

Imagine a \trgenitive command which is attached to a TURKISHWORD like so TURKISHWORD\trgenitive, and which prints TURKISKWORD’{n}{i}n, such that:

1. if the last letter of TURKISHWORD is a consonant, then {n} is nothing/empty
2. if the last letter of TURKISHWORD is a vowel, then {n} is n
3. if the last vowel of TURKISHWORD is i or e, then {i} is i
4. if the last vowel of TURKISHWORD is ı, then {i} is ı
5. if the last vowel of TURKISHWORD is u, then {i} is u
6. if the last vowel of TURKISHWORD is ü or ö, then {i} is ü.

Is it possible to define such a \trgenitive command?

Answer 1

You can perhaps implement the proposed postfix syntax with LuaTeX.

With either XeLaTeX or LuaLaTeX, the following prefix syntax works:

\documentclass{article}
\usepackage{xparse}
\usepackage{fontspec}

\ExplSyntaxOn

\NewDocumentCommand{\trgen}{m}
 {
  #1' \kutt_trgen:n { #1 }
 }

\prg_generate_conditional_variant:Nnn \tl_if_in:nn { ne } { T,F,TF }

\tl_new:N \l__kutt_trgen_word_tl

\cs_new_protected:Nn \kutt_trgen:n
 {
  \tl_if_in:neT {aeıioöuü} { \tl_item:nn { #1 } { -1 } } { n }
  \__kutt_trgen_vowel:n { #1 } n
 }

\cs_new_protected:Nn \__kutt_trgen_vowel:n
 {
  \tl_set:Nn \l__kutt_trgen_word_tl { #1 }
  \regex_replace_all:nnN { [^aeıioöuü] } { } \l__kutt_trgen_word_tl
  \str_case_e:nn { \tl_item:Nn \l__kutt_trgen_word_tl { -1 } }
   {
    {i}{i}
    {e}{i}
    {a}{ı}
    {ı}{ı}
    {u}{u}
    {ö}{ü}
    {ü}{ü}
   }
 }

\ExplSyntaxOff

\begin{document}

\trgen{Ezgi}\par
\trgen{Sıtkı}\par
\trgen{Utku}\par
\trgen{Hüsnü}\par
\trgen{Nefel}\par
\trgen{Anıl}\par
\trgen{Ufuk}\par
\trgen{Gönül}\par

\end{document}

The idea is to first check whether the last letter in the argument is a vowel; after the word and the apostrophe, an “n” is printed, otherwise nothing. Then the word is stripped of its consonants and a case switch is applied to the last remaining item (which is a vowel).