Click here for the Turkish translation 
Yazının Türkçe çevirisi için buraya tıklayınız

Recently, Erdoğan again raised one of the favourite controversies of Islamist ideologues in Turkey anew: the Kemalists’ switch to the Latin script in the early years of the Republic. This decades-old debate can by no means be reduced to a merely technical question and “resolved” thereby. But as both sides of the debate assume and claim that the facts are on their side, it is important to show that criticisms of the Ottoman script and the current Latin script for Turkish are both partially correct. In addition to debunking some myths about the Latin and Arabic scripts as used for Turkish, I will make use of comparison to other cases of 20th century orthographic reforms, and situate these discussions in the broader context of political-cultural-ideological-educational hegemony. 


The Arabic and Latin Scripts of Turkish

Erdoğan, like Islamist ideologues going back decades in Turkey, has repeated the claim that the switch to the Latin script “[made us] illiterate overnight”. While most literate Turkish speakers in the Ottoman Empire did use the Arabic script, the majority of the Turkish-speaking population was functionally illiterate at the dawn of the republic. The Kemalists, of course, credit the switch to the Latin script with the subsequent rise in literacy rates.

Since the embrace of Islam by most Turkic peoples (those peoples living across Eurasia speaking Turkic varieties) centuries prior, the literate classes who spoke Turkic had largely used the Arabic script to write. This was not only for reasons of religious identity, but because their language of education was first and foremost heavily Arabic-influenced Persian. Even when the literate found occasion to write in their own language, it was often a Turkic heavily influenced by that same Arabic-influenced Persian. The result was a written language that, when read aloud, was often incomprehensible to the average Turk.

However, we must separate out the choice of words and grammatical influence from Arabic and Persian from the question of the script. This much broader debate actually has little bearing on the discussion of the Arabic and Latin scripts. Literary Persian, after all, had long been marked by the same Arabic influence. But nationalist Persian speakers in Afghanistan and Iran will eschew the use of Arabic words while maintaining the Arabic script, just as the official language of Tajikistan contains many Arabic words still in spite of being written in the Cyrillic script. The script and the proliferation of foreign words in the language are two questions independent of one another.

Likewise, in Turkey, many “linguistic conservatives” (not only Islamists, as Foti Benlisoy will be the first to tell you) still use many Arabic and Persian phrases while writing exclusively in the Latin script. Could the Arabic script have been preserved for the official language of Turkey, and would the average school pupil have suffered as a result?

First, to provide some historical context: the Kemalists did not invent this question to spite Islamists, or vice versa. In the Ottoman Empire, difficulties in the use of the Arabic script as it was used for Turkish were noted as early as the mid-19th century (Lewis 1999:28). Scholars of this period often advocated changes to the Arabic script, rather than a switch to the Latin script, and modifications of this kind may be observed in the famous Ottoman era dictionary Kamus-ı Türki by Şemseddin Sami.

In his highly accessible history of the language reform “The Turkish Language Reform: A Catastrophic Success” (translated to Turkish by Mehmet Fatih Uslu), Geoffrey Lewis, the late renowned scholar of Turkish, claims that the Iranian Azerbaijani scholar Axundov proposed a switch to the Latin script in the 19th century, as well as having been a popular discussion among Ottoman officers during World War I (Lewis 1999:29). Lewis even claims that “free-thinking” writers associated with Kılıçzade Hakkı attempted to make the case to the religious officials of the Ottoman state on “Islamic” grounds (Lewis 1999:30).

As mentioned, Lewis’s history of the language reform (which mostly does not focus on the script, save for the relevant chapter from which all citations in this article are taken) is highly accessible to readers without linguistic training. Its Turkish translation is also readily available in Turkey. The interested reader may consult it for extensive examples of the flaws of the Arabic script as it existed at the time, as well as the dramatic saga of Latinising Turkish. It would, however, aid the reader to briefly detail what the problem was, rather than take 19th and early 20th century sources at their word that the Arabic script as it was used for Turkish was so deeply flawed:

The Arabic script is of course perfectly suited for representing the Arabic language, for which it was designed, but the claim has always been that it is very ill-suited to writing the native Turkic vocabulary. This is no small matter, as in spite of a heavy influence of foreign languages on Turkish in the Ottoman era in particular, the core of everyday Anatolian Turkish vocabulary is as Turkic as Uyghur, Tatar, or Kazakh. Chief among the issues is that many roots are distinguished by whether a front or back vowel is used, indicated by dots or their lack over ı/i and u/ü in Turkey, where no such distinction exists in either Arabic or Persian in which the Ottoman literati were educated.

Some conventions existed from the earliest years of writing Turkish in the Arabic script to distinguish these contrasts. For example, since Arabic distinguishes some consonant sounds as being closer to the throat (or with additional guttural articulation), these consonants may be used to indicate the articulation of the adjacent vowel. Consider the Arabic /k/ vs. /q/ (which are also the “back” and “front” /k/ sounds in eastern Turkic varieties). While the sounds of “o” and “ö” were both written with the same Arabic letter wāw (و) in Ottoman script, “kor” and “kör” could none the less be written distinctly by differentiating their “k” sounds: کور with a “k” for kör and قور with a “q” for “kor”.

Unfortunately for the literate Turk, if a given sound didn’t have a “back” equivalent in Arabic, there was no way to write the distinction. For example, compare “bol” and “böl”, both of which would be written identically in the Arabic script (بول). Even more troublesome, and frequently mentioned, was the inability to write the distinction between the verb “to become” “ol” and the verb “to die” “öl”.

A switch to the Latin script immediately solved this problem, borrowing the convention of twin dots over ö and ü for the front rounded vowels, and inventing a dotless ı for the back equivalent to i.

But the astute reader will note that these dot distinctions did not exist in the original Latin alphabet, but rather had to be developed for the script to represent the sounds themselves. Therefore, the Latin script was not, by some essential nature without reference to historical development, better suited for Turkic than the Arabic script was.

Indeed, during the same period, linguistic modernisation in other majority-Muslim societies often resulted in changes to the Arabic script: Persian, which like Turkish, has some consonant sounds not found in Arabic (such as the extremely common sound /p/), began consistently marking this and other sounds with new characters which had been more sporadic for reasons of tradition in the pre-modern era. Kurdish, as written in Iraq and Iran, ultimately reformed its own alphabet to mark the vowels consistently and alphabetically as is done in the Latin script. And as previously mentioned, even the Ottoman era Turkish dictionary Kamus-ı Türki adds some marks over wāw to distinguish Turkish’s many round vowel sounds.

And at any rate, flaws could also be identified with the Latin script as it exists today. Most noteworthy is the treatment of front and back consonants which do not match with vowels in loan words. In the early years of the republic, a letter “â” was used to indicate a palatal consonant somewhere (!) near it, as in “kâr” (palatal k) hal (palaltal l), but also to indicate a long vowel (such as in “âşık”, meaning “paramour”). Indeed, this confusion changed the pronunciation of Ali’s sword “Zülfükar” (with a long “a” and normal “back” k) to its present standard pronunciation with a palatal “k” (in contrast to the pronunciation of rural Alawites, who produce an Arabic /q/ sound, quite the opposite of the modern standard form). Problems of this nature are not just found in the ambiguity of the letters which exist, but in the absence of their corollaries: why is there no equivalent letter to indicate palatalisation in the vicinity of “o”, such as in the French loanword “rol” (compare the “back” l in “yol”). Lewis, interestingly, fails to mention the inability to represent palatalisation around vowels other than “a” (Lewis 1999:33, 36).

But the palatalisation of “l” (unlike the back and front forms of /k/) was likewise never indicated in the Arabic script, and indeed, is only indicated in the Cyrillic script in other Turkic varieties today, although, as we see from the Arabic script used for Kurdish or the Latin script used for Spanish, it is possible to indicate this contrast in both of the scripts in question, if one is so inclined.

In summary, on a technical level, as seen above, it was not merely an act of ideological engineering which caused the Kemalists to switch to the Latin script. In fact, groups that otherwise opposed the Kemalist project for religious reasons accepted the switch. On the other hand, it is not that the only objection to the Latin script was one of self-satisfied Ottoman nostalgia (one might propose reforms to the Arabic script). History could have gone more than one direction here, as we have seen above, and I will pass now from the technical discussion of Turkish into a more general discussion about justifying changes to scripts. 


20th Century Script Changes and Reforms: the East is Well-Read

Beyond the debate around suitability in the abstract, the standard Kemalist account of history tells us that the switch to Latin script was responsible for the pronounced increase in literacy rates in Turkish, an expected outcome of a “modernisation” which distanced itself from Ottoman backwardness tied to “the East”.

It cannot be denied that the Ottoman state was indeed backwards in a socio-economic sense relative to the triumph of western imperialist powers armed with all the advantages of industrialised capitalism. Of course, the same could be said for almost every single society to Turkey’s east, societies with diverse scripts and religious associations and so on and so forth. Can we say that the suitability of the Arabic script or other “oriental” scripts played a significant role in rising literacy rates?

Of course, as has been emphasised, each language has their own particular features which might render a script (or a version of a script) better or worse for that language. So it is a fact that without an indication of their difference in pronunciation, a script which renders “die” and “become” the same way falls short. Thus, a defender of the Kemalist switch to the Latin script might state that we cannot compare the retention of the Arabic script for Kurdish or Persian or the languages of Pakistan to the same idea for Turkish. After all, didn’t the Soviet Union also replace the Arabic script for all Turkic varieties which used it, first with the Latin script, then with the Cyrillic? Clearly, reforms or no reforms, Kemalists or no Kemalists, Turkic phonology almost “forces” us to abandon the Arabic script in a quest for modern standards of literacy.

In other words, for us to make the argument that it was (almost) completely arbitrary whether (a modified form of) the Latin script were adopted or (a modified form of) the Arabic script were retained for Turkish, we would actually need to show that not only would it be possible to modify a given script to make it usable, but that the Kemalists are relying on a false assumption by stating that literacy rates are overall affected by the logical ease of acquiring a given script.

Thankfully, even further east than Iran or Pakistan, we have examples of other societies which rapidly modernised during the same period with an even more extremely different choice of scripts than Turkey was faced with. I am referring to the region generally known as “East Asia”, but more specifically those countries which made use of (or still make use of) Chinese characters to write.

Chinese characters, as is widely known, are very complex and number in the thousands, with an individual character for more or less each word. Furthermore, their internal composition is based on meaning as much as sound, the sounds themselves being represented at best obliquely: one cannot “read aloud” or “sound out” Chinese characters as one can even with the English alphabet. One simply must know the sound associated with the word indicated by the character which is committed to memory.

Surely, if a relative increase in clarity between alphabetic scripts could cause an increase in a societal literacy rates, we might expect a gap in literacy rates favouring any alphabetic script over the Chinese script? Perhaps we might compare Vietnam, which since switching to the Latin script in the early 20th century, has steadily achieved mass literacy, particularly thanks to the efforts of the communist movement to educate the peasantry; and on the other hand, we should compare China itself, whose own communist party maintained the complex Chinese character system, in spite of the known difficulty in learning it?

Today, Vietnam and China both report extremely high literacy rates, but we must note that Vietnam (which switched to the Latin script) has a lower reported literacy rate than China (which did not). Clearly, lacking a clear alphabetic script did not itself prevent mass literacy among the Chinese, nor, in the final instance, did it even hold them back relative to the Vietnamese. This is the case today, but is this enough to rebut the Kemalist defence of the switch to Latin script?

In fact, in my experience, even the most die-hard Kemalists would concede that eventually any usable script could have been taught to the Turkish population, but rather the Latin characters facilitated the more rapid growth in literacy among an overwhelmingly illiterate population. This, the Kemalists claim, was only possible through the Latin script, and thus the Latin script was crucial for developing the country overall. It is my contention that the role of the Latin script as such was secondary: that as Turkey developed socio-economically, the state was able to implement successively more effective educational initiatives with the aim of raising the population’s literacy, which could have been done in the Arabic script with reforms thereto (as occurred with Kurdish in Iraq and Iran), or for that matter in Cyrillic. Can these two hypotheses be tested against other societies?

Fortunately for us, Turkey’s low literacy rate in the early 20th century was not an exception on a global scale: while the Ottoman Empire was underdeveloped relative to its European imperialist rivals, the same could hardly be said when we compare to most societies across the globe. In other words, at the time of the adoption of the Latin characters, a journey south and east would’ve shown most Turks comparably underdeveloped societies. The East Asia region, today one of the most developed outside of western Europe and North America, also began the 20th century “backwards”. So, what can we learn by analogy? How did different scripts fare over the course of the 20th century, when industrial capitalism’s new education model opened up pathways to literacy for more of humanity than ever before?

Comparing again the case of Chinese, whose complex character system was reported to have been mastered by about 20% of the entire population by 1949 (Peterson 1994:108), which, we may note, was comparable to the rate of literacy among Turkish women in 1950 (Taeuber 1958:101). Given that Turkish men had almost double the literacy rate at the time, it would appear that the Chinese characters were holding back Chinese literacy. However: this year was chosen for comparison because it was the beginning of the Chinese Communist Party’s attempts to develop their population for a new socialist life complete with a new socialist education. 1950 was decades into the use of Latin characters and modern public education among Turkish citizens.

So Turkish literacy was still lagging behind the United States, which had reported higher literacy rates for decades in spite of a version of the Latin script which is widely agreed to be less easy to learn than the Turkish one (Negroponte 1995:145). At the very same time, China was about to embark on its own script reform. Ostensibly with the goal of raising literacy rates, the Chinese Communist Party “reformed” Chinese characters in order to make them simpler, resulting in a split between Traditional and Simplified character sets which persists in the Chinese-writing world to this day.

The result of these reforms, according to some sources, was next to nothing (Seeberg 1990). Even if one is to disregard foreign scholarship and exiled Chinese academics, the census data from 1982 (after the end of the Cultural Revolution and the writing of the new PRC constitution of 1980) likewise “revealed that illiteracy and low educational attainment remained pervasive in China” (Peterson 1994:96). In other words, China’s current extremely high literacy rate has been achieved not only with one of the most complex scripts in use for any human language on the globe, it has done so shooting ahead of the extremely user-friendly Latin script for Turkish only since the 1980s, after decades of self-reported stagnation by the Chinese state. Thus, as a historical trend, we see a rise in literacy in both Turkey and China was based not on the introduction of an easier-to-use script, but on development and public policy directed towards a cohesive modern education system.

Oddly, in spite of these facts, the myth also persists in the PRC that literacy was raised by the introduction of the simplified version of Chinese characters, which took place before the Cultural Revolution and significantly before the changes to educational policy in the 1980s. Because the communists could not liberate Taiwan from the KMT (Guomindang) nor Hong Kong and Macau from the British and Portuguese respectively, these territories continued to use traditional Chinese characters. And yet, literacy rates in these territories was consistently higher than in the People’s Republic, and they remain comparably or more literate even now.

It would seem that these statistics demonstrate that the choice to simplify or change scripts does not have any significant effect on literacy rates. Rather, as a country develops socio-economically, its population may be afforded more comprehensive education in their own language, and as a result, literacy will rise. Actual ambiguities or problems with individual conventions or words can apparently equally (or at least almost equally) be remedied within any given script, provided the ruling classes and their educational apparatus have the will to do so. 


Conclusion: the reality of ideology and script in Turkey

As has been made clear, the claims of a purely technical motive for switching from the Arabic script to the Latin script are suspect at best. Even to the extent that the Arabic script was insufficient to represent Turkish words, modifications to this script would have been no less possible than the modifications to the Latin script that were necessary to use that script for the Turkish language. Herein lies the weakness in the Kemalists’ position against the Islamists: the Islamists know full well that the real motive in switching to the Latin script had to do with the Kemalist identification with the more socio-economically developed west and embarrassment with their “Oriental” culture.

Furthermore, while (for socio-economic reasons, and not its choice of the Latin script) the new secular education system was plainly superior from the perspective of spreading literacy among the popular classes, it did in fact leave several groups “illiterate”: the new language laws of the Turkish nation-state prevented a similar spread of literacy in the Kurdish language. The real effects of being a linguistic minority forced to be educated in another language as your “native” language are well documented in many countries, and Turkey is no exception (see, for example, Kırdar 2009).

The same can of course be said for the smaller Arab minority in Turkey, whose language, culture, and script were blamed for Turkish backwardness. The Arabs in Turkey today tend to be illiterate in the Arabic script even if they can speak Arabic, and are consequently cut off from the literature of what is a major world language in its own right.

But the Islamist narrative in Turkey is weakened by its tendency to avoid precisely such questions of minority rights. In fact, the Islamist narrative, drawing on Necip Fazıl Kısakürek, tends to resort to an equally mythological and supremacist position in its “culture war” on the Kemalists. When Islamists claim that Turks must write in the Arabic script because this was the script in which Turks “began to learn to think” “after accepting Islam” (!), they are making the inversion of the culturally essentialist argument the Kemalists made whenever they valorised “the West”: Turks should choose their script based on whether they reach for “the fruits of the Arab and Persian trees” or, on the other hand, if they seek to emulate “the westerner” and appropriate “Greek and Latin sources”.

The Kemalist assumption that only the Latin script and “western” culture could bring modernity relies on an unreconstructed Eurocentric conception of how a society can make cultural and social progress. The Islamists, for their part, responded with a rough mix of theological nationalism and reactionary status-quoism. In theory, both approaches are very reductive towards what a culture in the 20th and 21st century can be, and how it can relate to other cultures. But in practice, the Kemalists understood that this culture war was fought on material as much as ideological grounds, and produced programmes to change Turkish society accordingly.

The Islamists, for their part, have mostly relied on the socio-economic advances produced by Kemalists and other non-Islamists in Turkish and foreign society, but stamped their feet in outrage that “Turks” are “losing their own culture”. The question of how to write modern Turkish is a case in point: the Islamists never produced a cohesive programme for how to modernise the Arabic script to be incorporated into a modern education system, they merely stood on the sidelines and accused the Kemalists of infidelity to Islam. With such a weak approach, it is not difficult to see how the Kemalists won this particular battle.



Kırdar, M. G. (2009). Explaining ethnic disparities in school enrollment in Turkey. Economic Development and Cultural Change, 57(2), 297-333.

Kısakürek, N. F. (2014). İdeolocya örgüsü. Büyük Doğu Yayınları Basım Yayın Prodüksiyon Ltd. Şti.

Lewis, G. (1999). The Turkish Language Reform: A Catastrophic Success. OUP Oxford.

National Center for Education Statistics: “Literacy from 1870 to 1979”,, accessed December 15, 2019

Negroponte, N. (1996). Being digital (London: Hodder & Stoughton).

Peterson, G. (1994). State literacy ideologies and the transformation of rural China. The Australian Journal of Chinese Affairs, (32), 95-120.

Rye, J.B. (2007). “SPELLING REFORM And the Real Reason It’s Impossible”,, accessed December 15, 2019

Seeberg, V. (1990) “Literacy in China-The Effect of the National Development Context and Policy on Literacy Levels. 1949-79”.

World Population Review: “Literacy Rate by Country”,, accessed December 14, 2019.