First, modern Turkish has a large amount of dialectical variation across its primary area of distribution from the Balkans to Iraq. Each local dialect absorbs words from languages in the surrounding area. For example, Turkish speakers in Serbia have incorporated a number of Serbian words, and many of these have spread to other Turkish dialects. The standard dialect of Turkish is the Istanbul dialect, so that is what we will focus on now.

Standard Turkish has a shocking 86% of words of native origin with the remaining 14% coming from languages such as Arabic, Persian, and French. For comparison, about 26% of words in English are of Germanic origin (from the same language family as English). Hungarian, a language which has undergone active changes to replace foreign words with native words, has only about 50% native vocabulary. This is particularly interesting because Turkish has not been historically isolated from other languages.

On the other hand, Ottoman Turkish (spoken from the 15th to the 20th centuries), the predecessor of modern Turkish, had a huge number of loan words from Arabic and Persian (up to 88% in some texts). We will look into how Turkish redeveloped a native vocabulary shortly, but can we go back even further? What about the ancestor of Ottoman Turkish, commonly called Old Anatolian Turkish (spoken around the 11th to 15th centuries)? It's hard to directly say how much of Old Anatolian Turkish was made up of loanwords due to a scarcity of evidence. However, Ottoman Turkish loanwords from Persian and Arabic were heavily altered from their original forms even in early writings, suggesting that these words were incorporated before Turkic people migrated to Anatolia around the 11th century. Thus, our best guess is that Turkish was heavily influenced by other languages going back at least 1000 years. Going even further back to over 1000 years ago is nearly impossible (in terms of analyzing vocabulary) due to a lack of written evidence. Ancestors of modern Turkish from that time are only known through linguistic reconstruction; that is, inferring the contents of an older language based on relationships between existing languages. The process of reconstruction is very limited and cannot hope to rebuild most of the lexicon of an old language. Thus, we have no idea how "pure", from the standpoint of vocabulary, Turkish languages were 1000 years ago. And the same goes for 2000 years ago.

Instead of analyzing vocabulary, we can investigate other features of ancient Turkic languages. Unfortunately, we don't know a direct ancestor of Old Anatolian Turkish. What we can infer is that there were a large number of Turkish languages (the ancestors of modern Turkish, Azerbaijani, Turkmen and other languages) all mixing together within the Oghuz Turkic tribes (and it gets even messier the further back we go). When there is no clear ancestor, it's not clear what it means to have a pure linguistic lineage. In addition, historical linguists have found evidence that Turkish languages have undergone many changes in pronunciation, grammar, and vocabulary from their hypothesized common ancestor over 2000 years ago. In other words, we don't really know what "Turkish" was 2000 years ago, but it almost certainly was very different from modern Turkish.

But where does the idea of the purity of modern Turkish come from? And how does it suddenly have an 88% native lexicon? In 1932, the Turkish Language Association (TDK) was established as an institution with the mission of studying the Turkish language and its roots. The TDK has published many papers and books on Turkish languages. While many of their materials are academically useful, some of them are quite dubious. In the 1930s, they promoted the pseudoscientific Sun Language Theory, which hypothesized that a proto-Turkic language was the ancestor of all languages spoken by European, Middle Eastern, and Asian civilizations; this hypothesis is rejected by linguists. In addition to their publications, the TDK has replaced numerous loanwords in Turkish with native words. This has been done through a mix of deriving new words from Turkish roots, borrowing words from Old Turkic (another old Turkish language), and inventing new words with Turkish-sounding roots. The TDK continues to do this in the present day, regularly replacing loanwords for new technology with newly invented native words.

So the modern Turkish lexical "purity" is artificially constructed. The replacement of loanwords may make sense from an aesthetic perspective, since it might offer a more uniform "sound" and spelling to the language. However, it is not particularly "historical" even if words are occasionally taken from Old Turkic. The reality is that modern Turkish has changes in pronunciation and grammar that differentiate it from its ancestors, and there is no way of knowing what most ancient Turkish words actually were. Even if one could perfectly reconstruct an ancient language, it would be fundamentally different because it would be deployed in a new cultural context with new technologies, social structures, and ideas.

Languages are constantly evolving. Every feature of a language from pronunciation to vocabulary to grammar evolves slowly over time, even now. Each language is a synthesis of cultural and historical influences that reflects the needs and context of the people who speak it. In this way, no language is truly pure because no language is static or independent of its social context. Ancient languages only seem pure because we have limited knowledge of their predecessors. Languages have existed long before writing, and it's hard to identify loanwords and changes in grammar when there is no written record.

Modern linguistics rejects many notions of purity, standardness, and hierarchy of languages as pseudoscience, pointing to the long history of how such concepts are typically imposed with directly political agendas. All languages evolve naturally, and every existing "standard" language used to be an "impure", "incorrect", "barbaric", "uneducated", or "vulgar" dialect of another language. This is important to keep in mind when you hear someone talk about the purity of a language.