For a great many ideas Turkish has two current words, an older one borrowed from Arabic or Persian and a newer one built from Turkic roots during the language reform, and choosing between them is not a matter of correctness. Both are understood by everybody. Which one a person reaches for says something about their age, their education and sometimes their politics, and a learner picking at random is making a statement they did not intend.
That is unusual. Most languages have a formal and an informal register; Turkish has two vocabularies for the same register, produced deliberately within living memory.
Six pieces of vocabulary that are social rather than lexical
Turkish has a set of fixed phrases for situations English simply leaves silent, and knowing them changes how a conversation goes more than another hundred nouns would.
- abi and abla, older brother and older sister, used to address any slightly older stranger. Getting this right is the difference between polite and cold.
- kolay gelsin, may it come easy, said to anybody who is working, including a shopkeeper or a driver. There is no English equivalent and its absence is noticed.
- eyvallah, thanks, or fair enough, or goodbye, depending on tone. Informal and extremely common.
- inşallah, if God wills, used by religious and secular speakers alike to mean hopefully.
- yabancı means foreigner, and also stranger, and it is not an insult.
- Afiyet olsun, said before and after eating, roughly enjoy your meal, and it is expected rather than optional.
Four layers, and one of them was manufactured
The Turkic core is the oldest and it carries the everyday words: the body, the family, the numbers, the basic verbs. On top of it sit centuries of Arabic and Persian borrowing, which supplied the vocabulary of religion, administration, poetry and abstraction, in the way Latin supplied English. A French layer arrived in the nineteenth and twentieth centuries with technology, institutions and fashion, and it is still arriving in the form of English.
The fourth layer is the one with a date. From the 1930s a national language institution set out to replace the Arabic and Persian vocabulary with words built from Turkic roots, some revived from older texts and some coined outright. A great many of those coinages took, and now sit beside the words they were meant to replace rather than having removed them.
Which is why a Turkish dictionary will often give you two words and no guidance, and why the practical question about a new word is not only what it means but which layer it belongs to and whether it is still the ordinary choice.
Learning from older material and sounding like it
Here is the stall, and it catches readers rather than talkers. Older novels, poetry, and a great deal of the free material online use the pre-reform vocabulary heavily, because that is when they were written. A learner who builds vocabulary from them acquires words that are perfectly good Turkish and that a younger speaker will find dated or literary.
The mirror error also exists: material aimed at learners sometimes uses the newest coinages exclusively, which reads as slightly official. Neither is wrong and both are marked, and nothing in a dictionary will tell you which way. The exit is to check a word's currency rather than only its meaning, and to notice when two words for one idea turn up, because that pair is almost always the reform in miniature. Making that check a habit at the moment a word is met is the same storage decision described in making cards that teach rather than cards that pile up.
Same idea, several words
Where each choice sits, from a grandparent to a policy document
This rail is not about formality. It is about which vocabulary a speaker grew up with and which one they choose to use.
Placed by who is most likely to reach for the word today. The placings are this page's judgement rather than a measurement.
- The Arabic or Persian wordUnderstood by everybody, and dated or literary to a younger speaker
- The same word in a fixed phraseWhere the old vocabulary survives whatever the reform intended
- Whichever one is simply currentFor many pairs the question is settled and nobody thinks about it
- The French or English borrowingModern, technical, and outside the argument entirely
- The reform coinageBuilt from Turkic roots, and in official use whether or not it took
The middle mark is the important one and the figure cannot show how wide it is. For a large share of these pairs one word won outright and the other is genuinely archaic, and only a dictionary that dates its entries will tell you which pairs those are.
Two dictionaries, and one of them tells you where the word came from
Nisanyan's etymological dictionary is free online and it does the one job an ordinary dictionary does not: it gives the origin of a word and the date it is first attested in Turkish. It replaced guessing whether a word is Turkic, Arabic, Persian or a modern coinage, which is exactly the fact that decides how it will land. Reading two entries side by side for a pair of synonyms usually settles which one to use in about ten seconds.
Subtitles from films of two different decades are the other half, and the exercise is a comparison rather than a study. Watch a scene from something made in the 1970s and a scene from something recent, both with Turkish subtitles, and the vocabulary shift is visible on the screen rather than described to you. It replaced being told that the language changed, which is true and does nothing, and it is free wherever you already watch.
The mistake: assuming the shared words are safe
A learner who knows Persian or Arabic arrives with a large number of words already in place, and most of them work. The trap is that a shared word may have moved: it can be more formal in Turkish than in the source language, or narrower, or replaced in ordinary speech by a reform coinage while surviving in a fixed phrase.
So the head start is real and it is a reading head start rather than a speaking one, in the same way a Spanish speaker's Portuguese is. The words will be recognised and the register will be wrong, which is the error nobody corrects because the sentence is understood. Checking currency rather than meaning is the specific habit that converts the discount into something usable, and it costs one lookup per word.
What the words attach to
Every one of these words takes the same queue of endings regardless of which layer it came from, which is a chain of endings that turns one root into a clause, and the loanwords that fail to harmonise are almost all from the borrowed layers, in why an ending changes shape to match the word before it. The derivational endings are also what the reform used to build its coinages, so knowing four of them makes the new vocabulary transparent. Where each word sits in a sentence is decided by the focus rules in no gender, no articles, and the verb waiting at the end. And hearing which of a pair somebody chose, at speed, is speech that runs a whole chain of suffixes together.