In his recent article “Argumentum ad ignorantiam”, Elmir Mirzoev describes what he sees as the ongoing influence of the Cyrillic alphabet on Azeri orthography, despite the switch to Latin 17 years ago. Elmir bey describes the situation as absurd and claims that it will adversely affect Azeri grammar and even lead to “total illiteracy” in Azerbaijan. In this article I will try to show that a more careful analysis of the situation reveals that what Elmir bey identifies as Cyrillic influence is, in fact, something altogether different and that Azeri orthography is not necessarily in need of any major reforms.

A False Consensus

At the core of Elmir bey’s criticism of the supposed Cyrillic legacy in Azeri is the idea that all Latin-based writing systems are united by certain spelling conventions, which the Azeri writing system, however, systematically violates. One of those conventions is said to be that there are certain common words and abbreviations that are written the same way in all languages with Latin-based scripts except, apparently, Azeri. But checking just a few of the examples given by Elmir bey, this assumption about the unity of Latin-based scripts proves false. For example:

Sex:     seks (Dutch), sekss (Latvian), seks (Polish), seks (Tagalog), etc.

Taxi:    taksi (Finnish), taksi (Indonesian), taksî (Kurdish), taksi (Turkish), etc.

Jazz:   xhezi (Albanian), džäss (Estonian), dzsessz (Hungarian), djass (Icelandic), etc.

The idea that the spellings of commonly borrowed words like sex, taxi, or jazz are somehow universal is simply false. It’s true that the spellings which Elmir bey cites as universal are used in English or French, for example, which are widely studied as second languages, but there are dozens of other languages that transcribe these words phonetically according to their own orthographic rules. In employing the spellings seks, taksi, and caz, Azeri is not exceptional.

Another case of Cyrillic influence cited by Elmir bey is in the spelling of the names of famous people and historical figures. It is true that the spelling of Sigmund Freud remains unchanged in most other Latin-based scripts, as do J. S. Bach, Stefan Zweig, etc., and in these cases Azeri is in the minority. However, there are many exceptions to this tendency, and not only in Azeri, but in all Latin-based writing systems. For example, a name originating from Arabic, Chinese, or any other language with a non-Latin-based script, will be adapted differently in different Latin-based orthographies. Here are just a few examples:

From non-Latin-based scripts:

Çingiz xan:     Čingischán (Czech), Genghis Khan (English), Dschingis Khan (German), etc.

Çaykovski:     Tchaïkovski (French), Tsjajkovskij (Norwegian), Chaikovski (Spanish), etc.

Xamenei:        Chāmeneʾi (German), Jamenei (Spanish), Hamaneyi (Turkish), etc.

But Mongolian, Russian, Farsi, and other non-Latin based writing systems are not the only sources of such orthographic variations. For instance, the names of many historical figures and mythical heroes are written in different languages according to centuries-old traditions which are unique to each language. In such cases, the names may even originate from a language with a Latin-based script:

Historical names from Latin-based scripts:

III Vilhelm:      William III (English), Guillaume III (French), Wilhelm III. (German), etc.

Siseron:          Zizeron (Basque), Cicerone (Italian), Cícero (Portuguese), etc.

XIV Lüdovik: Lodewyk XIV (Afrikaans), Louis XIV (French), Ludwig XIV. (German), etc.

In compiling examples of the alleged absurdities of Azeri orthography, Elmir bey, unaware of the abovementioned patterns, made two mistakes. It is asserted that the name of the legendary Greek poet, known “throughout the world” (bütün dünyada) as Homeros, is written in Azeri as Homer under the influence of the Cyrillic alphabet. The name of the hero Hercules is also said to have been distorted in Azeri as Herakl. The problem is that 1) these names both come from Greek, a language that is not written the Latin alphabet, and 2) they are both ancient and widely famous, and there are many different local traditions for adapting their names. Therefore, just as expected, there is no standard spelling for these names across Latin-based scripts, as is easily demonstrated:

Homer:           Homer (English), Hóiméar (Irish), Omero (Italian), etc.

Herakl:            Hèrcules (Catalan), Ercole (Italian), Herkul (Slovene), etc.

The worldwide consensus on spelling turns out to be a fiction. Both widely borrowed words (e.g. jazz) and many proper names (e.g. Cicero) are often written differently in different Latin-based scripts. Every written language has its own conventions in this respect. It is clear that this phenomenon has little to do with the Cyrillic alphabet, since it affects all languages with Latin-based scripts, from English to Indonesian. And it also seems unlikely that the same phenomenon will lead to “total illiteracy” (total savadsızlıq) in Azerbaijan if it has not yet done so in the rest of the world.

Three Approaches

In a footnote, Elmir bey writes:

.the spelling of some city names – take New York in English, Nowy Jork in Polish, Нью-Йорк in Russian, written in other languages (including Turkish) as in the original, our spelling of this city’s name as Nyu York (once again taken from Cyrillic) looks somewhat illogical.

This passage is significant because it provides us with examples of the three approaches by which Latin-based orthographies typically deal with the spelling of foreign borrowings. Those three approaches can be briefly characterized in the following ways:

  1. Transplantation        New York (English)    →         New York (Turkish)

In Turkish, the spelling of many foreign names is simply transplanted wholesale, even when they include foreign letters which do not otherwise occur in the Turkish language, such as the ‘w’ in ‘New York.’

  1. Translation                New York (English)    →         Nowy Jork (Polish)

In this example from Polish, the foreign city name has been translated to preserve it’s meaning in the original language, i.e. English ‘new’ (yeni) is rendered as Polish ‘nowy’ (yeni).[1]

  1. Transcription            New York (English)    →         Nyu-York (Azeri)

In Azeri, the foreign name has been phonetically transcribed to approximate the pronunciation in the language of origin, and therefore the English word ‘new’ is rendered phonetically according to Azeri orthography as ‘nyu.’

As these examples demonstrate, each strategy preserves a different aspect of the foreign word: the first preserves the spelling, the second preserves the meaning, and the third preserves the phonetics. All three approaches are widely used, and in fact a single language often employs each of them in different contexts. Take English, for example:


As can be seen in this table, English employs all three strategies in writing foreign borrowings. It is also clear that there are no strict rules as to when each strategy should be applied: Spanish Cabo de Hornos is translated as ‘Cape Horn,’ but Tierra del Fuego (‘land of fire’ in Spanish) remains untranslated in English. English-speakers must know something about the Danish language in order to correctly say hygge (pronounced hyuqa), written in English as in Danish, but Hungarian gulyás is transcribed more or less phonetically as goulash. An examination of many other languages with Latin-based orthographies reveals the same diversity of approaches in rendering foreign borrowings.

All three of these approaches are equally logical, and they all have advantages and disadvantages. But by criticizing exclusively the strategy of phonetic transcription, Elmir bey ignores its advantages as well as the disadvantages of other approaches.

Let’s compare, for instance, phonetic transcription with transplantation. The advantage of phonetic transcription is that it establishes a standard pronunciation of foreign borrowings, while a disadvantage of transplantation is that it offers no guidance to pronunciation for the reader. In English, where transplantation is widely used, there is no standard pronunciation of many foreign names. English-speakers pronounce names like Goethe or Zweig in different ways (if at all!) depending on a number of factors, including their level of familiarity with the names, their level of familiarity with the language of origin, by analogy with familiar words with similar spellings, etc. This means that in English texts, foreign names are often accompanied by attempts at phonetic transcription (for which there is no standard method), and in speech English-speakers have to cope with wide varieties of pronunciation of the same name. To take the Polish city Rzeszów from the chart above as an example, an English-speaker with no knowledge of Polish spelling conventions will simply have to guess at the pronunciation, while an Azeri-speaker has no need to consult a Polish textbook, because the name will have been phonetically transcribed for them in any Azeri text as Jeşuv. To put it another way, phonetic transcription facilitates the translation of text into speech, while transplantation doesn’t.

While I think Elmir bey is wrong to claim that the Azeri tradition of phonetic transcription is harmful, it would also be wrong to claim that transplantation, despite the difficulties cited above, is necessarily harmful if kept within certain limits. With the examples above I simply aim to show that the tendency of Azeri to rely on phonetic transcription in rendering foreign borrowings 1) is not a marginal phenomenon specific to Azeri or the Cyrillic alphabet and 2) it has the advantage of establishing and promoting norms of pronunciation.

Phonetic Transcription Predates Cyrillic

The examples above demonstrate that there is no inherent connection between the orthographic strategy of phonetic transcription and the Cyrillic alphabet. But in the case of Azeri, to assert such a connection is simply incorrect chronologically. The Azeri tradition of Latin-to-Latin phonetic transcription actually predates the introduction of the Cyrillic alphabet.

In 1922, Nariman Narimanov ordered the formation of a commission to create a Latin-based writing system for the Azeri language, and the linguist and revolutionary Samad agha Aghamalioghlu was chosen to lead it. In September of that year, the newspaper Yeni yol (then it was written Jeni jol) was created and began publishing in parallel Arabic and Latin columns.[3]

Starting in 1922 and until the creation of Azeri Cyrillic in the 1940’s, phonetic transcription was already being used systematically in Azeri to render all foreign words and names. If you leaf through Yeni yol or any other books or newspapers published in the Latin script at that time, you will find such spellings as Şeqspir, Moljer, and Ƶyl-Vern, Fransa and Bordo.[4]

One of the creators and most active proponents of Azeri’s original Latin alphabet was the linguist Bakir Chobanzadeh. In the 1920’s, Chobanzadeh wrote extensively about problems of orthography and advocated phonetic transcription as the primary guiding principle. He specifically rejected the notion of transplanting the spelling of foreign words from the languages of origin because of the difficulties that creates:

In words borrowed from European languages, attention must be paid not to the spelling but to the pronunciation, because, due to the history of of the orthography of those languages, especially French and English (sometimes German and Russian), today there is a large gap between the spelling of words in those languages and their pronunciations. For instance, in an English dictionary, if the pronunciation is not shown next to each word it becomes difficult for foreigners, and maybe for the English themselves, to use their dictionaries easily.[5]

For the most part, Chobanzadeh’s proposals were adopted and eventually codified in orthographic dictionaries, school textbooks, etc., and phonetic transcription was applied systematically in the original Azeri Latin script of the 1920’s and 1930’s.

But Chobanzadeh was not the first person to promote the principle of phonetic transcription in Azerbaijan. The concept has had widespread support among the intelligentsia from the beginning of Azeri language reform movement, even before Soviet Latinization. One of the problems of Azeri’s former Perso-Arabic writing system, as pointed out by early language reformers, was its inability to render foreign words and names phonetically. Take, for example, this lament from a 1909 issue of Molla Nasreddin:

Without a doubt, thanks to the clear and interesting articles of our Kocharlis, the sweet translations of our Uzeyirs, and the information that our other writers have given about Gogol [غوغول], our readers are acquainted with this writer, and maybe well-acquainted; but I want to say that, nevertheless, despite all this reading and writing, our readers who don’t know Russian still don’t know what this writer’s name is.[6]

The Arabic spelling of Gogol’s name, غوغول, is not a guide to pronunciation at all, or at least it’s an extremely imperfect one. It might be transliterated into modern Azeri as ğuğul, but in fact it could be transliterated in numerous other ways because of the inherent ambiguity of the Perso-Arabic writing system. Although Mirza Jalil’s complaint is directed at Perso-Arabic orthography, the situation is in fact no different when transplanting spellings such as Goethe, Rzeszów, or hygge from their languages of origin into other Latin-based scripts, such as Azeri, English, etc. These spellings, outside the context of their native orthographic systems, no longer reflect the sounds of the words, and each reader is forced to guess at the pronunciations when translating text into speech.

In any case, it is clear that phonetic transcription had support when Azeri was still written with the Perso-Arabic alphabet, and it was first applied systematically in the original Azeri Latin script in the early 1920’s. The Cyrillic script and its orthography were first designed two decades later in the 1940’s.

Cause or Effect?

Reflecting on the connection between cultures and alphabets, Elmir bey writes:

For example, if some language uses the Arabic script, it will influence deeply that language’s lexicon, grammar, and in some cases even phonetics, because then vowels won’t be used, naturally a number of grammar rules and many terms will enter that language from Arabic, etc.  Or, say, if some language uses characters, it would be a reflection of a totally different worldview.

In this passage, Elmir bey asserts that a language’s writing system is intimately connected to its phonetics, grammar, lexicon, and even to a particular worldview. Therefore, he says, when a language adopts a foreign script, it will be altered drastically under the influence of the language for which the borrowed script was originally created. And, to take Azeri as an example, we see that the language has been most deeply influenced by Arabic, Persian, Russian, Turkish and English, and this corresponds with the alphabets that Azeri has employed: Perso-Arabic, Cyrillic, and Latin.

But by merely identifying a correspondence, we have not yet determined a cause-and-effect relationship. In other words, was the Azeri language so deeply influenced by Persian and Arabic because it was written with the Perso-Arabic alphabet, or was the alphabet adopted because Azerbaijan fell into the sphere of Perso-Arabic cultural, political, and linguistic dominance? The same question can be posed for Cyrillic and Latin.

In world history, I am unaware of any change of alphabet that was not accompanied by a broader cultural shift. There are, however, countless examples of cultural and linguistic shifts that were not accompanied by a change of alphabet. For example, Elmir bey notes that all attempts to Latinize the Japanese language failed, but that is not the end of the story. In fact, despite the failure of the Romaji movement, Japanese has absorbed massive linguistic and cultural influence from the West, primarily the United States and the English language. The Westernization of Japan began in the 19th century and was accelerated by the American occupation after WWII. To take just the lexicon as an example, it has been estimated that, in Japanese, English words account for 53% of the vocabulary related to management, 75% in marketing, 80% in trade, and 99% in computer technologies, not to mention the massive impact of English in other fields, in popular culture and slang, etc. It turns out that Japan’s incredibly complex writing system, without the slightest relationship to the Latin script, is no barrier to linguistic and cultural influences from the West.[7]

In the case of Azeri, it is clear that each change of alphabet was preceded by wide-ranging cultural and linguistic shifts. To take the lexicon as an example again, Russian borrowings started to enter the Azeri language en masse not after the adoption of Cyrillic, but after the signing of the Treaty of Gulistan. Whole words and phrases were adopted from Russian starting in the 19th century when Azeri was still written in the Perso-Arabic script: okrujnoy sud, uşkol, milyon, poçt, nəçərnik, student, fotoqraf, etc. These borrowings are reflections of much larger changes that occurred at that time, including new systems of administration and education, new technologies, growing contacts with Russian and European culture, expanding trade, etc. The alphabet proved to be neither a conduit nor a barrier to these developments.

All this is to say that, although alphabets and other writings systems certainly play important symbolic roles in national cultures, it is important not to overestimate their influence on language itself or imagine that a writing system can somehow predetermine a national community’s cultural orientation. To do so would be to confuse a cause with an effect. As Azerbaijan enters a new period of orthographic reforms, the public discussion should be based on a clear understanding of what a writing system is and how it works. Writing is primarily a tool, and any particular script should be judged on how well it does its job of clearly and accurately representing a specific language. All other questions are secondary.


[1] This is also an example of transcription: ‘York’ has been phonetically transcribed into Polish as ‘Jork.’

[2] All of the foreign borrowings used here as examples come from languages with Latin-based scripts.

[3] Audrey L. Altstadt, The Politics of Culture in Soviet Azerbaijan, 1920-40, Routledge, 2016, Kindle Edition, Loc 2513-2526.

[4] These examples are taken from M. F. Axundov, Əsərləri, Bakь: Azərnəşr, 1938; Ƶyl-Vern, Sirrli ada, Bakь: AzLQGI MQ, 1939; and from the first issue of the newspaper Jeni jol from September 21, 1922, reprinted in Bilâl N. Şimşir, Türk Yazı Devrimi, Ankara: Türk Tarih Kurumu Basımevi, 1992.

[5] Bəkir Çobanzadə. Seçilmiş əsərləri, V cild. Bakı: Şərq-Qərb, 2007, s. 233.

[6] “Molla Nəsrəddin”, 5 April, 1909, № 14, from Məmmədquluzadə, Əsərləri, v. II, p. 353, 731

[7] L. J. Loveday, Language Contact in Japan. A Socio-Linguistic History, Oxford, 1996, p. 101-103; В. М. Алпатов, Япония: язык и культура, Москва: Языки славянских культур, 2008, с. 93.

Argumentum ad Ignorantiam: Use of Latin Script in Azerbaijani