Introduction
In Ottoman Turkish, millet is a term that has received considerable scholarly attention for two reasons. First, given the multireligious and multiconfessional composition of the Ottoman Empire, scholars have perceived millets as confessional communities of non-Muslims that the empire used as a unit to categorize non-dominant groups under its rule. From this perspective, they have paid particular attention to its use as a clue to understanding how the empire ruled over diverse populations (Braude Reference Braude1982; Konortas Reference Konortas, Gondicas and Issawi1999; Masters Reference Masters, Winter and Ade2019; Sigalas Reference Sigalas2022; Ursinus Reference Ursinus1989; Reference Ursinus1993). Second, millet acquired the modern meaning of nation at the turn of the 19th century and is still used in this sense in Turkish. Thus, scholars have considered it a key concept in understanding how people in the Ottoman Empire perceived and adopted nationalism (Çalen Reference Çalen and Şimşek2025; Dressler Reference Dressler2022; Karabıçak Reference Karabıçak2019). This article, building on findings in conceptual history, mainly concerns the first area of study. However, it produces insights that will inform research in the second area of study by examining how the term millet was used in Ottoman documents to refer to non-Muslim collectives in relation to their religious authorities and how its use changed over time. Its chief insight is that we may reject the assumption that the use of millet was related to Ottoman governance of non-Muslims and reflected the preference of Muslim ruling elites to control them through a rigidly defined community. Rather, it proposes that we accept the term’s vague contours and asserts that millet’s increased use in Ottoman documents in the first decades of the 19th century derived from non-Muslims’ response to millet’s newfound significance as a term denoting modern nationhood.
In recent years, with the introduction of the conceptual history approach to Ottoman Studies and the understanding that concepts are “tools we use to move the social world, to interact with others, and the means by which we categorize and perceive the world” (Topal and Wigen Reference Topal and Wigen2019, quote on 95), there has been a resurgence of scholarly interest in the word millet. Scholars such as Yusuf Ziya Karabıçak (Reference Karabıçak2019), Nikos Sigalas (Reference Sigalas2022), Markus Dressler (Reference Dressler2022), and Mehmet Kaan Çalen (Reference Çalen and Şimşek2025) have published articles on the term, contributing to a better understanding of the meaning it had in narrative sources in the early Ottoman centuries and the process by which Ottoman actors in the later history came to accept the modern concept of the nation.Footnote 1 Meanwhile, scholars have reflected on the use of millet for non-Muslim collectives, but with rather fragmentary information, as they have not conducted meticulous research on Ottoman documents, which are most likely to provide crucial evidence on Ottoman governance. This situation leaves gaps in scholars’ understanding, as discussed below.
Until the 1980s, scholars generally accepted the millet system theory, according to which the empire categorized non-Muslim subjects under its rule into three confessional communities, called millets, and delegated the autonomy of these millets to their respective religious authorities, which were the Istanbul patriarchs of Orthodox Christians and Armenians and the Chief Rabbi of Jews, without significant changes from the mid-15th century until the collapse of the empire (Carleton Reference Carleton1937; Gibb and Bowen Reference Gibb and Bowen1957, chap. 14). In the 1980s and 1990s, Benjamin Braude and others proposed that such a simplistic model was a projection of the 19th-century situation onto the earlier period, and that the Ottoman arrangements with non-Muslims that 20th-century scholars called the millet system actually emerged in the late 18th and early 19th centuries. The revisionist scholarship also held that the emergence of these arrangements was accompanied by the government’s use of the term millet for non-Muslim collectives in the sense of a state-recognized confessional community encompassing the empire’s entire territory. In their view, it was a tool that the imperial government used to stabilize its rule over multireligious and multiconfessional subjects in its later history (Bardakjian Reference Bardakjian, Braude and Lewis1982; Braude Reference Braude1982; Gara Reference Gara2017; Goffman Reference Goffman1994; Konortas Reference Konortas, Gondicas and Issawi1999; Masters Reference Masters, Winter and Ade2019; Ursinus Reference Ursinus1989).
At present, Ottomanist scholars generally accept that millet was the official term used by the imperial government to refer to non-Muslim communities in the 19th century, and they use it in their publications not only in Turkish but also in English as an untranslatable technical term that reflects the emic perspective of the Ottomans of the time (Cohen Reference Cohen2014; Koçunyan Reference Koçunyan2017; Stamatopoulos Reference Stamatopoulos, Ellis, Hálfdanarson and Isaacs2006; Tzoreff Reference Tzoreff2023).Footnote 2 Based on this understanding, scholars have further discussed that there was an official status of millet in the late Ottoman Empire, that the existence of such a status itself constituted the millet system, and that some minor non-Muslim groups, including Catholics, Protestants, Bulgarians, and Vlachs, received that status while others did not (Arslan Reference Arslan2004; Artinian Reference Artinian1975; Beydilli Reference Beydilli1995; Clements Reference Clements2025; Doyle Reference Doyle, de Obaldía and Monge2022, 86; Evstatiev Reference Evstatiev2019, 90; Katsikas Reference Katsikas2009, 180; Macar Reference Macar2023; Sarıyıldız Reference Sarıyıldız2002; Turhan Reference Turhan2025). Scholars have also considered how millet changed semantically from a confessional community to a nation, how the status of non-Muslims changed from millet as an officially recognized community to become a minority when the empire collapsed, and how the successor states of the Ottoman Empire drew on the millet model in their treatment of non-dominant religious groups (Barkey and Gavrilis Reference Barkey and Gavrilis2016; Evstatiev Reference Evstatiev2019; Katsikas Reference Katsikas2009; Roudometof Reference Roudometof1998; Sezgin Reference Sezgin2010; Sharkey Reference Sharkey2018; Sluglett Reference Sluglett and Robson2016; Stamatopoulos Reference Stamatopoulos, Ellis, Hálfdanarson and Isaacs2006). In all cases, scholars have generally assumed that the term formed the basis for imperial rule over non-Muslims in the 19th century, without questioning that the source of this assumption was mid-20th-century literature based on the accounts of European observers, and without investigating when and how the use of millet for non-Muslim collectives became common in Ottoman documents. Moreover, their conception of millet perpetuates the idea that non-Muslim communities were homogeneous entities strictly defined by the state, which contradicts recent scholarly findings that have revealed the ambiguity of non-Muslim communities by examining how they functioned in the late Ottoman Empire (Ozil Reference Ozil2013).
It is true that in Ottoman documents from the 1830s to the 1870s, millet was frequently used to refer to non-Muslim collectives with predicates such as Rum,Footnote 3 Armenian, Jewish, and Catholic. It is also true that the relationship between the Ottoman government and non-Muslim religious authorities changed in the 18th and early 19th centuries. At the turn of the 18th century, the government strengthened the authority of the Orthodox Christian and Armenian Patriarchates of Istanbul in order to use them in its struggle against Catholic missionaries. In the late 18th century, the government also relied on their ecclesiastical networks as means of managing uprisings in the provinces; then, during the Greek War of Independence, it further promoted a reliance on non-Muslim religious authorities during efforts to monitor non-Muslim subjects in Istanbul (Greene Reference Greene2015; Karabıçak Reference Karabıçak2020; Ueno Reference Ueno2025a; Reference Ueno2025b). However, no scholar has produced historical documentation proving that these changes led to millet’s frequent use in Ottoman documents or that millet had any institutional connotation in the 19th-century Ottoman Empire.
In his recent article, Nikos Sigalas (Reference Sigalas2022, 333) asserted that we should distinguish between the use of millet in the sense of “religion” in higher linguistic registers and “people” in more vernacular registers, arguing that the latter usage went underground from the 17th to the early 18th century due to the ulama’s increasing influence. Sigalas’ proposition, based mainly on narrative sources, offers insight into the particularities of millet’s application; however, it must be modified when we consider the enormous volume of Ottoman state documents written during this period, which were not necessarily produced under the influence of a higher linguistic culture, as they were not the products of madrasa-trained ulama, but of scribes who learned their skills on the job. Moreover, although scholars often assume that the terminology used in Ottoman documents, including imperial edicts, reflects the intentions of the imperial government and state officials, a significant portion of these documents were prepared in response to petitions and reports from outside the central government and included quotations and paraphrases from petitions that, although usually formulaic, left room for vernacular usage.
Sigalas’s article is also suggestive in that it discusses the long span from pre-Ottoman times to the early 19th century to understand the semantic changes of millet, which cannot be understood by examining short periods alone. Following his lead, this article addresses long-term changes from the 16th to the 19th centuries, focusing more on changes in millet’s usage in Ottoman documents for non-Muslim collectives than on its meaning, ultimately leading to different conclusions. It particularly tests the hypothesis from prior research that millet’s use to refer to non-Muslim communities expanded from the late 18th to the early 19th centuries and provides context for this change by carefully examining the preceding and subsequent periods. The ideal source materials for this purpose are Piskopos Mukataası registers.
Within the Ottoman government, the Office of Bishop Taxation Unit (Piskopos Mukataası Kalemi), which was part of the financial department, was responsible for administering affairs related to the state’s arrangements with the Christian high clergy. For bureaucratic purposes, the office created registers that recorded the high clergy’s charters of investiture and related imperial orders (İnalcık Reference İnalcık1982, 440–447). Today, the existence of 37 of these registers has been confirmed, mainly from the early 18th to the early 19th centuries. While these Piskopos Mukataası registers include a small number of documents concerning Syriacs, Nestorians, and Catholics of the Aegean Islands, most concern the Orthodox Christians and Armenians. Contrary to the supposition of millet system theory, the Ottoman Empire did not make similar arrangements with Jewish religious functionaries until the 1830s, which results in a lack of comparable information about them and limits this article’s scope. Nevertheless, among the enormous quantities of surviving Ottoman documents – which cannot all be examined – Piskopos Mukataası registers are the most suitable for achieving its aim: examining the relationship between the use of millet for non-Muslim collectives and the Ottoman treatment of their religious authorities. Unlike other types of documents, the registers frequently contain references to Christian collectives in relation to the rights and privileges of their religious authorities, and the consistent recording of similar documents for over a century enables long-term observation.
As will be discussed in this article, an extensive study of Ottoman documents concerning non-Muslim clergy shows that some uses highlighted in previous studies of millet for non-Muslim collectives in the 18th century were merely sporadic examples. A gradual but substantial increase in the frequency of millet’s use is observable in the first decades of the 19th century. Interestingly, there is a gap in the increasing frequency between documents concerning Rum and Armenians. In addition, there is a difference in frequency depending on the type of document. Benjamin Braude’s argument (Reference Braude1982) is suggestive for understanding these differences. He argued that the Ottoman government’s increased use of millet for Ottoman non-Muslims in the mid-19th century resulted from Europeans using the term, which they had learned from the Ottomans when it was used for European Christians. Although this argument has already been criticized for being based on a misunderstanding of millet’s changing usage, this article finds merit in its focus on actors outside the Ottoman government. Rather than supporting the prevailing state-centric explanation that assumed the increase in millet’s use occurred due to the ruling elites’ desire to categorize non-Muslim subjects uniformly from the top down, this article considers the possibility that non-Muslims influenced the choice of terms in Ottoman documents and provides an explanation for the variations observable in millet’s usage frequency. It argues that non-Muslims were also part of the formation of Ottoman discourse, which is generally considered to have been monopolized by the Muslim ruling elite.
The first two sections of this article carefully consider the use of millet from the 16th to the 18th centuries and demonstrate that the term was occasionally, yet consistently, used in two different senses until the turn of the 19th century. One was religion in the sense of Islam, Christianity, and Judaism, though millet was not generally used to refer to confessions as subcategories of these religions.Footnote 4 In parallel, millet was used to refer to non-Muslim collectives, as it meant people, a premodern nation, or ethnie in Anthony Smith’s (Reference Smith1986) terms, offering an equivalent to taife and kavim in Turkish. The third section discusses the change in the first decades of the 19th century, showing how the frequency of millet’s use for non-Muslim collectives gradually increased. The final section discusses the period after the 1850s, when the Ottoman government began to avoid using millet to refer to non-Muslim collectives, a tendency that strengthened in the 1880s.
Two Senses of Millet
Millet in Turkish is derived from the Arabic word milla, meaning religion, and was used in this sense in the Ottoman Empire. Accordingly, scholars have assumed that the Turkish millet acquired the meaning of not only a religious community but also a non-Muslim confessional community from its original meaning of religion. This assumption needs to be reconsidered, however, in light of Sigalas’ recent research (Reference Sigalas2022), which shows that in the early centuries of the Ottomans, millet meant people and was used to translate laos and genos in Greek and populus in Latin, suggesting that millet did not necessarily carry a religious or confessional connotation when referring to a group of people.
This idea is supported by the examination of several dictionaries produced in the 17th and 18th centuries. In 1641, Hovhannes Ankiwratsi, known as Giovanni Molino, a Turkish-speaking Armenian from Ankara who learned Italian in Rome and worked as an interpreter for the Venetians in Istanbul, published a concise Italian–Turkish dictionary (Święcicka Reference Święcicka2000). Its long index allows us to search Turkish words backward. In his dictionary, millet appears as an equivalent of nation (natione), people (gente), and race (razza), along with taife, which was used in Turkish to refer to various kinds of groups, not only religious and ethnic but also those related to profession and gender (Molino Reference Molino1641, 161, 267, 338). Another equivalent of natione was cins, which is known as a Turkish term for ethno-regional belonging in the 17th century (Kunt Reference Kunt1974). The author had limited knowledge of higher Arabic linguistic culture, so religion did not appear in his dictionary as an equivalent of millet, reinforcing the impression that the two meanings of millet – religion and people – were distinct.
In contrast, the 1665 Italian–Turkish dictionary prepared by a French Capuchin missionary who spent many years in the Ottoman realm listed millet as equivalent of religione but not of natione, gente, or razza (da Parigi, Reference da Parigi1665, 947, 1419, 1749, 1762). Later in the century, Franciscus à Mesgnien Meninski, who worked for the Polish embassy, published a well-known Turkish–Latin dictionary that included explanations in German, Italian, French, and Polish. His explanation of millet can be summarized as religion, people, and nation (Meninski Reference Meninski1680, 4883). This understanding can also be observed in the Turkish–French dictionaries from the late 17th to the 18th centuries.Footnote 5 Taken together, these dictionaries suggest that in the 17th and 18th centuries, millet meant both religion and people in the general sense, as well as a group of people who supposedly shared a common origin and culture but were not necessarily defined by a common religion or confession.
This understanding is confirmed by Evliya Çelebi’s extensive travelogue. It included ethnographic explanations of the people living in the various regions of the empire and beyond, providing a useful corpus on how different identities were explained in mid-17th-century Ottoman Turkish.
When Evliya Çelebi spoke of the identity of people, he distinguished between religious and ethnic identities. Regarding the religious identity of Christians, he differentiated religion and confession in a manner similar to modern understanding by using different terms. When referring to religion, he strongly preferred the word millet in the phrases millet-i Mesihiye and millet-i Nasara, meaning millet of the Messiah and the Nazarenes, respectively, which are quite common expressions in Ottoman Turkish referring to both Christianity and Christians (Evliya Reference Evliya2001a, 122, 170, 315, 319, 320, 322; Reference Evliya2002, 15, 58, 159, 219, 263, 279). Meanwhile, he used mezhep mainly to refer to the confessions of Christians and Jews, including Catholicism (papişte mezhebi, İrim Papa mezhebi), Lutheranism (Luturyan mezhebi), and Karaitism (Karayi mezhebi) (Evliya Reference Evliya2001b, 262; Reference Evliya2000, 86, 125, 222). When writing about the confession of the Russians and the Cossacks, he seemed unable to find a concise word or phrase corresponding to Orthodoxy, and thus used the expressions, “the confession of the Greek people” (Urum kavmi mezhebi) and “the confession of the Greek infidels” (Urum kefereleri mezhebi) (Evliya Reference Evliya2001b, 81, 84; Reference Evliya2000, 337).
Evliya Çelebi perceived groups of people in different regions as kavim, a Turkish term equivalent to ethnie. He combined this term with various ethnonyms, including Urum (Greek), Ermeni (Armenian), Yahudi (Jewish), Kıpti (Coptic), Habeş (Ethiopian), Latin (Ragusan), Eflak (Wallachian), and Macar (Magyar) (Evliya Reference Evliya1999, 74, 130; Reference Evliya2000, 185; Reference Evliya2001b, 81, 220; Reference Evliya2002, 263; Reference Evliya2007, 319, 407, 487).Footnote 6 He also combined the above ethnonyms with kefere (infidel) and taife, but did not directly associate them with millet. Nevertheless, when explaining these collectives, he referred to some as “ancient people” using millet: “the Tatar people is also an ancient millet” (kavm-i Tatar dahi millet-i kadimdir) (Evliya Reference Evliya2001a, 353). In his view, the Serbs, Armenians, and Ragusans were also an “ancient millet” (Evliya Reference Evliya1999, 130; Reference Evliya2001b, 199; Reference Evliya2002, 263). Beyond that, he used millet to refer to “other people,” writing that “there are also two neighborhoods of Roma people (Çingane kavmi) and one neighborhood and one church of Armenian infidels, but Greeks, Franks, and other people have no neighborhood (Rum ve Frenk ve gayri millet mahallesi yokdur)” (Evliya Reference Evliya1999, 225; Reference Evliya2000, 215).Footnote 7
What about the use of millet in Ottoman documents? One might assume that the Ottomans were generally concerned about the confessional identity of non-Muslims and treated them as collectives. However, examined documents suggest that, in the actual work of justice and administration, they did not always find it necessary to use terms referring to confessional or ethnic identities, nor did they use words meaning collectives in the early modern period. Rather, they often used terms referring to non-Muslim individuals, such as kefere (infidel) and zimmi (non-Muslims under the protection of the Muslim polity).
We can find a consistent use of millet in the sense of religion in sharia court records from the 16th to the 18th centuries. In these records, court scribes recorded not only documents issued by the court but also edicts sent from the central government, which will be discussed later. While, especially in the earlier period, some documents contained vernacular expressions, the language of court documents was generally influenced by the higher linguistic culture of the ulama due to their monopoly over the judiciary. This influence is strongly evident in the frequent use of formulaic expressions, in which millet was used to denote identity in the early modern period. Although court records did not necessarily refer to people’s religious, confessional, and ethnic identities, entries about slaves usually identified them through a combination of their origin, religion, and appearance. While court records used asl to refer to a slave’s origin, which usually appeared as their ethnicity, millet (or rather milla, as the Arabic expression was used in Turkish texts) was the word used to express a slave’s religious (but not confessional) identity, whether they were Muslim, Christian, Jew, or “infidel.”Footnote 8 Millet was also used to refer to the religious identity of Christian and Jewish witnesses, in the phrases udul-i millet-i Nasara and udul-i millet-i Yahudi. Footnote 9 Thus, millet was used to denote an individual’s religious identity, not a group to which they belonged.
State documents also used millet in the sense of religion in formulaic expressions; in the titles of dignitaries, mostly foreign ones, we see references to the millet of the Messiah.Footnote 10 Meanwhile, we can find the occasional use of millet for non-Muslims in the sense of a group of people in 17th-century state documents, which contain a wide variety of linguistic expressions, including quotations and paraphrases of petitions, making it difficult to determine what their use of millet meant. The matter we need to consider carefully is whether it was confessional.
Similar to Evliya Çelebi, relatively wordy expressions were used to refer to confessions in state documents, and the words within those were varied. For example, we see “the ritual (ayin) of the Greek (Rum), Serbian, and Wallachian infidels,” “the religion (din) of the Serbs and Wallachians,” and “the religion and ritual of the Franks and Ragusans” (Boškov Reference Boškov1992, 29, 34, 41, 45, 78, 81, 86, 88), suggesting that established, concise expressions were not in common use. Based on these examples, the use of millet in “the monks of the millet of Serbs and Wallachians” (Sırf ve Eflak milletinde olan rahipler) seems to mean religion (Boškov, Reference Boškov1992, 36, 82).
As Vjeran Kursar (Reference Kursar, Baramova, Mitev, Parvev and Racheva2013) pointed out, Ottoman documents related to the Balkans contain various ethnonyms. Some, including Rum, Ermeni (Armenian), and Yahudi (Jew), could be seen as confessional or religious identities,Footnote 11 but many were primarily concerned with ethnic affiliation. It is important to note that ethnonyms from both groups may appear in the same line: “Greek (Rum), Serbian, and Wallachian infidels” and “Albanian, Bulgarian, and Greek (Rum) groups” (Arnavud ve Bulgar ve Rum taifeleri) (Boškov Reference Boškov1992, 29, 78).Footnote 12 Among these ethnonyms was Latin, which sounds confessional but was also used to refer to the Ragusans: “the Ragusan (Latin), Šokci, Albanian, and Magyar groups” (Boškov, Reference Boškov1992, 32, 80; Meninski, Reference Meninski1680, 4131). This indicates that confessions such as Orthodoxy and Catholicism did not necessarily shape a non-Muslim collective; instead, various ethnic groups of the same confession appear in Ottoman documents.
The ethnonyms referred to above were generally associated with kefere and taife. In contrast, their association with millet was usually indirect. For example, documents contain the expressions “Šokci, Bosnian, and other infidel millets” (Şokça ve Boşnak ve sair kefere milleti), “Serbs, Šokcis, Armenians, Jews, and other millets” (Sırf ve Şokça ve Ermeniyan ve Yahudiyandan ve sair milelden), “the monks of Greek, Armenian, and other millets” (Rum ve Ermeni ve sair millet rahipleri), and “Armenians and other millets as well as the group of the Franks, the Georgians, Ethiopians and other millets” (Ermeni ve gayri millet ve taife-i Frenk ve Gürcü ve Habeş ve sair milleti) (Kursar Reference Kursar, Baramova, Mitev, Parvev and Racheva2013, 103; Yekeler et al. Reference Yekeler, Karaca, Temel, Türe and Atik2019, 67; Zahirović Reference Zahirović2020, 111).Footnote 13 Meanwhile, a direct association with ethnonyms can be found in documents where translation from European languages was involved. Millet was used to refer to the Union of Three Nations (Unio Trium Nationum) in the Ottoman vassal state of Transylvania, whose three groups, the Magyars, Székelys, and Saxons (Macar ve Sigel ve Saz), were directly associated with millet (Ursinus Reference Ursinus1993, 63).Footnote 14 An important exception was found in the edicts concerning the relationship between the Armenians and other Christian groups in Jerusalem, which were repeatedly issued with modified wording. It was decreed that Syriacs, Ethiopians, and Copts would use the holy places of Jerusalem as subordinates of the Armenians, and while Armenians and Rum were associated with taife, Syriacs, Ethiopians, and Copts – whom Evliya Çelebi associated with kavim – were consistently combined with millet (Habeş ve Kıpti ve Süryani milletleri) from 1517 to the 18th century (Ercan Reference Ercan1988, 33–42).Footnote 15
An examination of Ottoman documents refutes Sigalas’ argument that the use of millet in the sense of people went underground in the 17th century, throughout which millet in the sense of religion and people was used continuously. In general, millet was used to describe an identity in the sense of religion. Meanwhile, as the examples in this section indicate, millet in the sense of people referred to a group of people defined by origin, language, and culture, rather than by confession.
The 18th Century
What about in the 18th century, when millet began to appear more frequently in Ottoman writings when referring to nations in European countries (Sigalas Reference Sigalas2022, 346–352)?
Compared to the earlier centuries, the number of documents we can consult in the 18th century is enormous, thanks to the development of the Ottoman bureaucracy and the emergence of the Piskopos Mukataası registers. Their emergence also coincided with a changing relationship between the Ottoman government and the Istanbul patriarchates of the Orthodox Christians and the Armenians. As the unsuccessful and protracted war against the Holy League was waged, the Ottoman ruling elites found the proselytizing activities of Catholic missionaries much more problematic in the 1690s than they had previously. To prevent the conversion of Christian subjects to Catholicism, they strengthened the authority of the two patriarchates, expanding their geographical jurisdiction and encouraging them to monitor their flocks (Ueno Reference Ueno2025a, chap. 1). How did this changing treatment of the patriarchates affect the terms used for non-Muslim collectives?
An apparent change in terminology was the decline of the derogatory term kefere (infidel) in Ottoman documents and an increase in the relatively neutral terms zimmi and reaya (subject) in its place. Similar to previous centuries, the strongly preferred term to refer to non-Muslim collectives was taife. Although scholars have highlighted some instances of millet’s use for non-Muslim collectives in the 18th century, these are only occasional uses compared to the frequency of taife. Nevertheless, thanks to the vast number of documents, we can find many examples of the use of millet in direct association with ethnonyms, and the interchangeable use of taife and millet became clear in the 18th century.
Previous studies have assumed that the expressions Rum milleti and Ermeni milleti (or millet-i Rumiyan and millet-i Ermeniyan) began appearing in the mid to late 18th century, with millet in these expressions meaning a confessional community encompassing the entire empire. The idea behind that is that this change in terminology occurred when the Ottoman government entrusted the Orthodox Christian and Armenian patriarchates of Istanbul with the jurisdiction over their entire respective confessional communities; since the people under the jurisdiction of the Orthodox Christian Patriarchate included not only Greeks but also Serbs, Bulgarians, and others, the government chose to use a new term, millet, to denote a confessional community regardless of ethnicity (Doja Reference Doja2022; Konortas Reference Konortas, Gondicas and Issawi1999; Masters Reference Masters, Winter and Ade2019; Sigalas Reference Sigalas2022). However, a general observation of the Piskopos Mukataası registers and other documents does not support these assumptions. We can instead find these expressions from the late 17th century; for instance, an early example of “Rum milleti” can be found in a 1696 edict addressed to the judge of Jerusalem.Footnote 16 Then, their use was consistently found throughout the 18th century, though much less frequently than taife. Footnote 17 Collectively, it is hard to say that the rise of the two patriarchates and the expansion of their jurisdictions was accompanied by the adoption of a new institutional term, suggesting that the government might not have intended to rigidly define the extent of non-Muslim confessional communities from the top down, turning them into unitary entities during this period.
If the state bureaucracy had given a new meaning to millet as an official confessional community encompassing the entire empire, the use of millet in this sense would most likely be found in documents defining the rights and privileges of the clergy, especially their jurisdiction. The Piskopos Mukataası registers record charters of investiture given to the Orthodox Christian and Armenian patriarchs of Istanbul in the 18th century. In the preambles of these documents, millet was sometimes used in the sense of people and in the formulaic titles of the addressees as “the Chosen of the Millet of the Messiah.” However, throughout the 18th century, Ottoman scribes always used taife, not millet, in explaining the jurisdictions of the patriarchates, which was an indispensable part of their charters and for which the use of a technical term, if any, was expected. The charter given to the Orthodox Christian patriarch of Istanbul in 1714 states that the addressee would be recognized as the patriarch by clergymen of different ranks and “the old and the young of other infidels” (sair kefere taifesinin ulusu ve kiçisi). Over time, the term kefere changed to zimmi, and then to Nasara, but the term used to refer to their collective remained taife (Çolak and Bayraktar-Tellan Reference Çolak and Bayraktar-Tellan2019). The charters given to the Armenian patriarchs of Istanbul consistently used “Ermeni(yan) taifesi” throughout the 18th century.Footnote 18
Meanwhile, in the 18th century, Ottoman scribes did not refrain from using millet for smaller non-Muslim groups, such as Syriacs and Maronites. Importantly, although the Ottoman government clearly disapproved of granting the Maronites legitimate status in 1722, and there is no evidence that this policy changed thereafter, they used “Maronite millet” in 1758.Footnote 19 Millet was also used to replace taife and reaya for non-Muslim collectives; for example, Ermeni taifesi and Acem Ermeni reayası were both referred to as “aforesaid millet” in later passages.Footnote 20 The general choice of words suggests that taife had a more formal connotation than millet during this period.
Millet itself is considered to have had a confessional undertone when used for non-Muslim collectives. This understanding, however, hinges on the assumption that the millet system theory is at least somewhat correct. From the observations in this and the previous sections, it is hard to see that millet itself had a confessional connotation in the 17th and 18th centuries. Rather, millet, like taife, could only take on a confessional meaning when used with certain predicates. It seems merely to have been a term used when referring to a group of people, whether ethnic and/or confessional, and therefore could be used for both non-Muslim collectives under Ottoman rule and for European nations in the 18th century.
Shift in the Early 19th Century
After the French Revolution, the concept of the nation acquired a new aura that legitimized the rights of ethnic/national collectives as political entities. As a natural consequence of millet’s use for ethnie and a premodern nation in the 17th and 18th centuries, it was also used to translate a nation in the modern sense immediately after the French Revolution (Karabıçak Reference Karabıçak2019, 325–329). How, then, did the semantic shift regarding a nation in Europe affect the use of millet for Ottoman non-Muslim collectives?
In general, the use of millet for non-Muslim collectives in Ottoman documents increased with the decline of the use of taife in the first decades of the 19th century. To explore whether this trend was related to changes in the state’s treatment of religious authorities, it is fruitful to consult the Piskopos Mukataası registers. Interestingly, strikingly different trends can be observed between the documents related to Armenians and Orthodox Christians.
In the Piskopos Mukataası registers, the use of millet for Armenians gradually increased from around 1805, and became as common as taife around 1815. By the end of the 1830s, millet was almost exclusively used, and taife became an exception.Footnote 21 Unlike the immediate replacement of derogatory terms with neutral ones after the 1856 Reform Edict (al-Qattan Reference al-Qattan and Suleiman1994), the shift from taife to millet was gradual, suggesting that the shift may not have been caused by a political decision by the government.
A similar increase was not observed in the orders related to the Orthodox Christians in the Piskopos Mukataası registers.Footnote 22 Taife continued to be widely used, while millet was used quite occasionally.Footnote 23 Thus, the terminology used for Armenians and Orthodox Christians in the Piskopos Mukataası registers differed in the early 19th century. This is unexpected for two reasons. First, from the early 18th century, the empire increasingly treated the patriarchates of these two groups similarly (Ueno Reference Ueno2025a, chap. 1; Reference Ueno2025b). Second, we can observe the fairly common use of millet for Rum in other types of documents of the same period, such as petitions, grand vizier’s reports (telhis), and sultans’ replies to them (hatt-ı hümayun);Footnote 24 thus, it seems the terminology also differed according to the type of document. How are these differences to be interpreted?
Considering that a large portion of the orders in the Piskopos Mukataası registers were prepared in response to the petitions of the patriarchates and quoted from them, one possible answer is that the scribes of the central government followed their terminology. Although scholars often assume that the terminology used in Ottoman documents reflected the intentions of the imperial government and state officials, numerous petitions and reports in Ottoman Turkish were created outside the central government, and a significant portion of documents composed within the central government were prepared in response to these petitions and reports, rather than through consultation and decision-making within. Therefore, in many cases, Ottoman documents written within the central government quoted or paraphrased language chosen by actors outside the central government. It is likely that although the Armenian Patriarchate actively used millet when composing petitions in Ottoman Turkish, and the term was being used within the government, the Orthodox Christian Patriarchate was reluctant to use it, and this reluctance is reflected in the rare appearance of millet in orders concerning the Orthodox Christians.
This hypothesis, based on archival evidence, contradicts the assumption of previous studies that the use of millet in the 19th century had institutional significance, especially for Orthodox Christians, because it signified a confessional community and legitimized the patriarchate’s jurisdiction over different ethnic groups. If this assumption were correct, the Orthodox Christian Patriarchate, with its ethnic diversity of Greeks, Bulgarians, Serbs, and Arabs, would have preferred to use millet. Meanwhile, the Armenians would have had no reason to use it, since, as an ethnic/national collective, they suffered confessional division due to their large number of Catholic converts (Frazee Reference Frazee1983). However, the Piskopos Mukataası registers show the opposite.
It should be noted that the period during which the term millet gradually became more prevalent overlapped with the Serbian revolts of 1804–1813 and 1815–1817, as well as the Greek revolt of 1821–1829. During these rebellions, millet was used to refer to these two nations (Erdem Reference Erdem, Birtek and Dragonas2005; Karabıçak Reference Karabıçak2019; Reference Karabıçak2020). It is likely that the Orthodox Christian Patriarchate sought a confessional community connotation in taife, a long-established term with that meaning, and avoided using millet. The patriarchate needed to be extremely cautious because using millet would have implied the Greek nation. This could have limited the scope of the community and caused it to be viewed as rebels. Meanwhile, millet was a useful term for the Armenian Patriarchate because it emphasized national unity in the face of confessional divisions within the community.
The use of millet by the Armenian Patriarchate and state officials suggests that the Orthodox Christian Patriarchate was overly cautious. Even in the era when the term was adopted to translate the modern concept of nation, it was not necessarily associated with assertions of political rights (Manoukian Reference Manoukian2022). To reinforce this idea, we will briefly depart from our main examination of Ottoman documents and consider how Armenians explained millet. Armenians were known for being the most widespread users of Turkish among non-Muslims of the Ottoman Empire (Manoukian Reference Manoukian2024). This is reflected in the supplementary Turkish explanations for major terms in the voluminous dictionaries published in 1749 and 1836 that defined Armenian words in Armenian. In these dictionaries, the Armenian term azk was defined as a group of people, with an emphasis on “kinship, genus, or ‘race’.” Because azk was equated to millet in these dictionaries, Sebouh David Aslanian argues that azk acquired a confessional connotation during this period (Reference Aslanian2023, 355–357). However, the assumption that millet had a confessional connotation is unfounded, and the authors of the dictionaries did not mention any such connotations when defining azk in Armenian. Rather, these dictionaries reveal that the Armenians used azk as an equivalent of millet to denote a kinship-based collective.
This view is consistent with the explanation of millet given in 1860 by Garabed Panosian, a Turkophone Armenian who published newspapers in Turkish using Armenian script (Ueno Reference Ueno2016). Faced with the changing meaning of the term nation in Europe, he distinguished between the meanings of nation/millet in Asia and Europe, claiming that Armenians belonged to the former. His definition of nation/millet in Asia coincided with that of ethnie. In his argument, “anyone can change her/his religion, language, state, and customs,” but millet is by no means changeable. Thus “an Armenian can become a Muslim by changing her or his religion, but s/he cannot become an Ottoman or a Turk.Footnote 25” “S/he is not able to get out of being an Armenian” (Miwnadiyi Ērjiyas 1860a, 1860b). Although he accepted that religion was one of the ties that bounded the members of a millet, his explanation left no room for considering that millet itself had the meaning of a confessional community, let alone one with state recognition. The separation of religious and ethnic/national identities and the use of millet for the latter was also observed in the bilingual regulation of Armenians’ agricultural association, which limited its membership to nationally Armenian and religiously Illuminator’s Christian (milletce Hay ve mezhebce Lusaworchagan Khristiyan/azkaw Hay ew Groniw Lusaworchagan Krisdoneay) (Ganonakir Hoghakordzagan Ěngerut‘ean 1863, 5). Meanwhile, the preamble of an 1835 Armeno-Turkish book suggests that Catholic Armenians also expressed their Armenian ethnicity by using millet (Kollaro Reference Kollaro1835, 4).
Let us return to examining Ottoman documents. The different attitudes towards the use of millet can be seen in the titles engraved on the official seals of the Armenian and Orthodox Christian patriarchs in the 19th century. An 18th-century case suggests that patriarchs had a say in choosing the expression of their titles but needed state authorization.Footnote 26 They were required to use their seals when submitting petitions to the Ottoman government. These seals served to prevent others from impersonating the patriarch to request appointment as high clergy and other benefits (Bayraktar Tellan Reference Bayraktar Tellan2011, 136).Footnote 27 The title used by the Armenian patriarchs in Turkish was “the Armenian Patriarch of Islambol” or “the Armenian Patriarch of Istanbul and the Affiliated Regions” until the 1810s.Footnote 28 However, around 1820, it was changed to “the Patriarch of the Armenian Millet.”Footnote 29 The title temporarily reverted to its former form in the seals of Madteos and Hagopos, who assumed the position in 1844 and 1848, respectively, but in the mid-1850s, it changed to “the Patriarch of Armenian Millet in Istanbul and the Affiliated Regions,” a title that remained in use until 1882.Footnote 30 The Orthodox Christians, on the other hand, adopted millet for their patriarchal title in the seal relatively late, in the early 1840s, until which time they used the title “the Orthodox Christian Patriarch of Istanbul (Islambol),” with an exception around 1824, when millet was used only in the handwritten signature.Footnote 31 In the early 1840s, the title in the seal was changed to “the Patriarch of Orthodox Christian Millet in Istanbul and the Protected Domains.”Footnote 32 However, the use of millet in the title was abandoned in the early 1860s – earlier than in the case of the Armenians.Footnote 33 Collectively, these findings support the argument based on the Piskopos Mukataası registers that there was a difference in the willingness of the Armenian and Orthodox Christian patriarchates to use millet.
The expression of jurisdiction in their charters of investiture was also varied. As mentioned above, the charters of the patriarchs explicitly stated over whom they had jurisdiction. In the article on the jurisdiction of the Istanbul patriarchs, the Armenians used the term millet; those over whom they had jurisdiction were called Ermeni taifesi before 1844, Ermeni reayası in 1849, Ermeni milleti from the 1860s to the 1880s, and Ermeni cemaati from the end of the 19th century.Footnote 34 Meanwhile, the Orthodox Christians were referred to as the “group of the Nazarenes” (Nasara taifesi) until 1848, which turned into “my Orthodox Christian imperial subjects” (Rum tebaa-i şahanem) in the 1850s and then remained so.Footnote 35
It should be noted that the Orthodox Christian Patriarchate was not always reluctant in adapting to changes in terminology over time. In fact, the Piskopos Mukataası registers and their successors show a change in one expression used. Until the 1830s, the registers often used the expression “Rum reaya taifesi” to describe Orthodox Christians, but from the mid-1830s, it was gradually replaced with “Rum taifesi.”Footnote 36 Based on the idea of distinguishing between the tax-exempt ruling class and the tax-paying subjects, the word reaya was used to refer to the latter in the early centuries of the Ottoman Empire. In the 18th century, however, reaya was also used to refer to non-Muslims in general (Fotić Reference Fotić2017; Tezcan Reference Tezcan and Woodhead2011, 165–167). Özgür Türesay (Reference Türesay, Lignereux, Messaoudi, Peters-Custot and Wilgaux2021), who analyzed Ottoman terminology from the early 19th century, has shown that the term reaya gradually shifted to tebaa in the 1830s to refer to Ottoman subjects in general. It seems the government’s choice of terminology led to the avoidance of reaya by the Orthodox Christian Patriarchate as well, which occurred at exactly the same time. In contrast to this change, the rarity of the use of millet in the edicts concerning the Orthodox Christian patriarchs in the first half of the 19th century can be understood as a sign that the government was not willing to push for it, nor was it wanted by the Orthodox Christians in the patriarchate.
Nevertheless, if we consider Ottoman documents from various ministries, councils, and departments, the use of millet for non-Muslim collectives clearly dominates from the 1830s to the 1870s, while the use of taife declines. This cannot be explained by Armenian preference alone. One possible factor is the development in the late-1820s of terminology differentiating Greek rebels in the Morea from Orthodox Christian Ottoman subjects, referring to the former as Yunan and the latter as Rum (Karabıçak Reference Karabıçak2020, 445–447). Another reason is the avoidance of reaya. As noted above, Türesay (Reference Türesay, Lignereux, Messaoudi, Peters-Custot and Wilgaux2021) proposes that the use of reaya declined in the 1830s and that tebaa was used instead to refer to subjects of the empire in general. However, tebaa was not the only word whose usage increased as reaya’s decreased. Previously, reaya was sometimes used to refer to all tax-paying subjects and sometimes only to non-Muslims – tebaa took the place of reaya in the former meaning. Meanwhile, though, a change in terminology occurred in the 1820s. Until the 1820s, Ottoman documents sometimes used the phrase “people of Islam and reaya” (ehl-i İslam ve reaya) to refer to the entirety of the empire’s subjects. In the 1820s, however, this phrase was partially, if not completely, replaced by “people of Islam and three millets” (ehl-i İslam ve milel-i selase) (Ueno Reference Ueno2025b, 293).
Along with the gradual increase in the use of millet for non-Muslim groups in the early 19th century, the term “three millets” also came into use. Initially, it seems to have been limited to the local context of major cities, including Istanbul, Izmir, and Edirne.Footnote 37 In 1794, a scribe took the time to paraphrase it as “Orthodox Christian, Armenian, and Jewish communities” (milel-i selase ki Rum ve Ermeni ve Yahud taifesi), suggesting that the phrase had not gained currency.Footnote 38 The use of “people of Islam and three millets” became common in the 1820s, when the Ottoman ruling elites – strongly motivated by the Greek War of Independence to control the population of Istanbul through the introduction of an internal passport system and the census – found it necessary to impart the entirety of Istanbul’s diverse population, and by extension, that of the empire. Numerous documents related to the internal passport were produced and sent throughout the empire (Ueno Reference Ueno2025b, 293). This change in terminology, combined with the tendency to avoid reaya, seems to have led to the spread of the phrase “three millets” and the broader use of millet to refer to non-Muslim collectives.Footnote 39
New Communities and New Terminology
How should we understand the independence of the Catholic community in 1830? Based on the assumption that there was an institutional status of confessional community called millet, scholars have mentioned that the Catholics achieved independence as a millet and saw it as a precursor to the creation of other millets in the last century of the empire (Artinian Reference Artinian1975; Beydilli Reference Beydilli1995). However, the Ottoman documents written about the Catholics do not refer to their independence as a millet.
Among the Ottoman documents from the 1820s and 1830s addressing the treatment of Catholics, a document recording negotiations between Ottoman and French representatives holds particular significance because of its decisive role in their independence. The document also provides valuable insight into millet usage. The two sides were negotiating the treatment of Armenian Catholic converts, who were numerous in Istanbul, and agreed to separate them from the jurisdiction of the Armenian Patriarchate. The French side then demanded that the Ottoman government appoint a patriarch to head the Catholics and “place them in the position of a glorious and free millet” (bunları azim ve serbest millet hükmüne koyacak). Without assuming that millet itself meant a state-recognized confessional community, this seems to be a demand for the liberation of Catholics from the control of the Armenian Patriarchate and ongoing persecution in the form of expulsion from Istanbul.Footnote 40 The Ottoman representatives rejected the demand, claiming that two patriarchs should not be appointed for a single millet, even though they had already agreed to the independence of the Armenian Catholics (Beydilli Reference Beydilli1995, 226–231). This may have been a pretext for refusing to appoint a patriarch, but suggests that the Ottoman representatives did not necessarily believe that this independence would create another millet; even after the separation of the communities, both the Armenians under the superintendence of the Armenian Patriarchate and those under the Catholic clergy would remain a single millet, that is, the Armenian nation. Although millet was used for the Catholics along with taife after independence was gained in 1830, this may simply have been because the Armenians, with whom the Catholics had a fierce rivalry, used the term millet so often that the Catholics, claiming to be their equals, were willing to use the same term, and the state bureaucracy simply used the term that the Catholics used. The use of millet for the Catholics is not surprising, considering that the French term nation was polysemic and that its abstract form was also used for a religious collective in France during this period (Noiriel Reference Noiriel2001, 152).
The Ottoman government first appointed a bishop and, from 1834, a patriarch to oversee not just Armenian Catholics but all Catholic subjects in the empire, except for Latin subjects. This overarching jurisdiction was clearly stated in the charters of appointment given to the Catholic patriarchs in the mid-19th century. Ottoman scribes wrote that the patriarch would be recognized both by those of the “aforementioned millet,” presumably referring to the Armenians, and by “all other Catholic Christians of the Chaldean, Syriac, Melkite, and Maronite millets” (sair bi’l-cümle Nasara-yı Katolikandan Keldani ve Süryani ve Melkit ve Maruni milletleri).Footnote 41 Thus, they distinguished confession and ethnicity, using millet for the latter; the Catholic community, defined by confession, subsumed plural millets or ethnicities, even though the community as a whole was sometimes referred to as a millet. Such flexible usage strongly suggests that the use of millet for non-Muslim collectives had no institutional connotations in the mid-19th century.
While the use of millet for non-Muslim collectives gradually gained currency in Ottoman documents due to the term’s popularity in the sense of a nation among non-Muslims, the birth of the Catholic community and its preference for using the term for itself kept its meaning ambiguous. What about the independence of other non-Muslim groups?
In 1850, the Ottoman government decided to separate the Protestant converts from the Armenian community. In the edict officially announcing this independence, the reason given by the government was that the Protestants were a religiously separate cemaat (hasbel-mezhep ayrıca bir cemaat olduklarından).Footnote 42 Later, in the 1870 edict officially announcing the establishment of the Bulgarian Exarchate, the government did not include any word to denote a community or group.Footnote 43 Then, in 1905, when discussing the independence of the Vlachs, the Ottoman cabinet clearly stated that “in terms of civil law, the Vlach people are considered a cemaat, similar to the Greek, Bulgarian, and Serbian people” (Ulah cinsi hukuk-ı medeniyece Rum ve Bulgar ve Sırp ahali gibi bir cemaat add olunarak).Footnote 44 While scholars today refer to the independence of the Protestants and Vlachs and the establishment of the Bulgarian Exarchate as forming millets, it is clear that the Ottoman government did not see these actions as such. Rather, it is noteworthy that the word cemaat appeared, which had rarely been used to refer to non-Muslim collectives as a whole and was instead mainly used to refer to small local congregations until the mid-19th century.
From the 1850s, cemaat was added to the terms used to refer to non-Muslim collectives, along with millet and taife. In the Reform Edict of 1856, which marked a milestone in the Ottoman treatment of non-Muslims by declaring equality between Muslims and non-Muslims, cemaat seems to have been the term the ruling elites preferred to use for non-Muslim communities (Takvim-i Vekayi‘ 1856). However, while they were forcefully removing the use of discriminatory terms against non-Muslims from official documents during the same period (al-Qattan Reference al-Qattan and Suleiman1994), they did not push hard for the use of cemaat; it can be noted that the Reform Edict used millet twice for non-Muslims. One was with “the patriarch or the notables of millet,” which was translated in the French version prepared by the Ottoman government as communauté. On the other occasion, the edict proclaimed that all subjects were eligible to be employed as state officials, and millet was used in “without distinction of millet,” which was translated as “sans distinction de nationalité.” Interestingly, the edict also twice used milletçe, an adjective form of millet; once in the phrase “the administration of milletçe affairs of Christian and other non-Muslim subject communities” (Hıristiyan ve sair tebaa-i gayrimüslime cemaatlerinin milletçe olan maslahatlarının idaresi). In the contemporary French and Armenian translations, milletçe in this phrase was translated as lay or secular (temporalle, ashkharagan). The other occasion of its use was with “milletçe schools,” which was translated into “écoles publiques” in French and simply “schools” in Armenian.Footnote 45
In the constitution of 1876, the government officially promoted the idea of the Ottoman nation by declaring that all people of Ottoman nationality, regardless of religion, were to be called “Ottoman” (Osmanlı). The constitution called non-Muslim communities cemaat (Kanun-ı Esasi Reference EsasiH1324, 7, 8). After that, the government began to push hard to use cemaat and avoid millet for non-Muslim collectives in state documents. This change in terminology was mentioned by a contemporary foreign observer, F. van den Steen de Jehay, who was the author of a fairly detailed analysis of non-Muslims in the Ottoman Empire. In the main text, he used the term millet, which he said means a nation, to describe various non-Muslim groups, but he clearly stated in a footnote that in recent years, the use of millet in official documents was replaced by cemaat (van den Steen de Jehay Reference van den Steen de Jehay1906, 83). Cemaat, not millet, was also the word used in the 1889/90 textbook of Ottoman administrative law used in the Ottoman Law School, which included a chapter on the treatment of non-Muslim communities and religious authorities (İbrahim Hakkı Reference HakkıH1307, 288–315; Fujinami Reference Fujinami2025, 101–107). As mentioned earlier, the use of millet disappeared from the seals of the Armenian Patriarchs in 1882, as well as from the explanations of jurisdiction of the Armenian and Catholic patriarchal appointment charters, and was replaced by cemaat. Footnote 46
Scholars have assumed that millet was the term that the late Ottoman Empire used to categorize people under its rule, and some have further claimed that the application of this categorization itself is the millet system. However, that assumption is not consistent with the laws governing the classification of people. The terms used in the first census law of 1882 to classify people are mezhep and cemaat, meaning confession and community, respectively. In the official language of the empire, the population was divided by mezhep, and a group of people belonging to the same mezhep was called cemaat (Düstur H1299, 15–24). The 1900 amendment to the census law replaced mezhep with millet, while non-Muslim collectives continued to be called cemaat (Düstur 1941, 433–451). Contrary to scholars’ assumption, millet did not refer to confession, but to the religions of Islam, Christianity, and Judaism, as explicitly stated in the reamended census law of 1902 (Düstur 1941, 865).Footnote 47
It is important to note that the term milliyet, an abstract form of millet commonly used to denote nationality, was originally mentioned in the 1882 census law as an entry on identity papers (Düstur H1299, 20).Footnote 48 This was changed to the “cemaat that he or she belongs to” in the 1902 amendment, but what did the original term milliyet mean? Answering this question is difficult because the actual identity papers did not have the milliyet entry. Identity papers prepared between 1882 and about 1904 used the term mezhep as a standard, with no reference to milliyet, although the term was used in the law. The mezhep entries included Muslim, Orthodox Christian (Rum), Jewish, and Armenian.Footnote 49 Beginning around 1904, the name of the entry gradually changed from mezhep to millet, and what was listed as millet corresponded to a religious faith, such as Muslim, Christian (Hıristiyan), or Jewish.Footnote 50 At the beginning of the 20th century, the Ottoman government decided to change the criteria for the census and identity papers. These changes indicate that the Ottoman government was still experimenting with different ways of classifying people at the beginning of the 20th century. As this reveals, the management of such a religiously and ethnically diverse population remained a challenge for the Ottoman Empire until its last decades.
Conclusion
In his widely cited article, Benjamin Braude (Reference Braude1982) criticized the “foundation myths of the millet system” and paved the way for revisionist scholarship. Meanwhile, his demystifying article led to the creation of another myth; if the so-called millet system was a projection of the 19th-century situation onto the preceding period, it reflected the 19th-century reality to a certain extent, and therefore the term millet had the meaning of a state-recognized confessional community in that period. Although this assumption had no basis, scholars accepted it and opted to call the non-Muslim communities of the 19th century millets for decades. It was a convenient term that allowed scholars to briefly explain (or not to explain) the supposedly unique approach of an exceptional empire at the expense of portraying the Ottoman treatment of non-Muslims as exotic, something out of step with global trends. However, as discussed in this article, the use of millet for non-Muslim collectives did not reflect the intentions of the empire’s ruling elites.
In the study of the Ottoman Empire, state documents in the Ottoman archives are the most important sources of information, but at the same time, they have forged a state-centric view of Ottoman history. What makes these documents appear more state-centric is the gaze of scholars, who have searched for the state behind them. Combined with the lingering influence of the millet system theory, this perspective led scholars to assume that the use of millet in Ottoman documents reflected the intentions of the imperial ruling elites, who wanted to stabilize their rule over a religiously diverse population by using this institutional term that rigidly defined the scope of non-Muslim communities. However, as discussed in this article, millet was an ambiguous term with no institutional meaning. Furthermore, its changing use should be considered instead, as it reveals how Christian subjects responded to the global shift toward a new aura of nationhood. For the Armenian clergy, who suffered from confessional strife within their ethnic/national community, a term denoting their unity was beneficial, and so it was quickly adopted on their part. Meanwhile, the Orthodox clergy were wary of using the term, which could have emphasized ethnic divisions within the confessional community under their jurisdiction. In this period, the Ottoman government did not actively forbid the use of millet or give it any special meaning, but accepted its use by non-Muslims as a general term and incorporated it into state documents. Thus, behind the Ottoman documents reviewed lies the use of language by a culturally diverse population, suggesting a need to revise the state-centered discussion of the arrangements that the Ottoman state had with non-Muslim religious authorities regarding their rights and privileges.
Acknowledgments
I would like to thank Yusuf Ziya Karabıçak for his insightful feedback on an earlier draft of this article. I would also like to thank Yaşar Tolga Cora for providing valuable information about materials.
Financial support
This work was supported by JSPS KAKENHI Grant Numbers JP20KK0266 and JP23K00870.
Disclosure
I acknowledge using DeepL Write (https://www.deepl.com/ja/write) and DeepL Translate (https://www.deepl.com/ja/translator) to improve the manuscript’s word choice, grammar, clarity, and readability. After using these tools, I carefully reviewed and edited the text and take full responsibility for the final content.
Open access
