Doors, privacy and the public sphere: a conceptual discussion on the spatial structure of early modern Istanbul

Abstract

The aim of this article is to offer a new conceptual framework for the study of various spaces in Istanbul during the seventeenth and the eighteenth centuries. I contend that rather than the sharp distinction manifested with the public–private dichotomy, we need to focus on gradations of privacy. I offer qāpū and bāb as two new concepts borrowed from the period's own repertoire for representing the macro and micro ends of the spatial gradation. Theoretically, I draw on George Simmel's definition of doors as the interfaces of spatial formation.

Introduction

Ottoman society's cultural diversity, and the moments and spaces of encounter that it brings about, make Ottoman public spaces a very attractive research subject. These public spaces include coffeehouses, public baths, barber shops, taverns, bazaars, marketplaces, squares, social complexes and mosques. The difficulty associated with their study in the Ottoman world arises from the unfathomable functional diversity of all these spaces. For example, it is impossible to clearly identify the role of a tavern in the public sphere by distinguishing between its warehouse, retail, social gathering and consumption functions. Moreover, as I reveal in this article, this difficulty is further intensified by the spatial complexity of early modern Istanbul. Thus, drawing a concrete border between public and private spaces from architectural and functional perspectives is a difficult process. The starting point of this article is the conceptual shortcoming regarding the public–private binary opposition for early modern Istanbul. As a solution, I ask whether a conceptual framework can be developed to more accurately represent the spatial structure of early modern Istanbul, and I extend this inquiry towards the specific conceptual repertoire of the period. ¹ I contend that we should focus less on the sharp distinction between so-called public and private spaces, and more on the gradation of privacy from macro to micro scale. I offer qāpū and bāb as two new concepts borrowed from the period's own repertoire for representing the macro and micro ends of the spatial gradation. Theoretically, I draw on George Simmel's definition of doors as the interfaces of spatial formation. Thus, I transform the concepts of qāpū and bāb, both of which are synonyms of door, but differ in their connotations, as the interfaces of the spatial structure of early modern Istanbul.

The early modern period in European history refers to a time span from the end of the fifteenth century, the beginning of the long-term disintegration of the feudal system, and continues until the beginning of the industrial revolution. A similar breadth and uncertainty are also in question for the Ottoman early modern period. In the Ottoman context, scholars use the term to cover the period from the sixteenth to the nineteenth centuries. In this study, I focus on the seventeenth and the eighteenth centuries, specifically on the period until the end of the Tulip Era (1718–30). This period was marked by conflicts between different factions over the sharing of political power, as noted by revisionist historians who opposed the decline paradigm in classical historiography. ² The deep effects of these conflicts led to the founding and shaping of the modern Ottoman state. Besides, since the middle of the eighteenth century, the urban structure of Istanbul began to undergo a major transformation. In this period, as revealed in detail by Shirine Hamadeh, new urban spaces such as promenades, gardens and squares emerged in Istanbul, or rather, these old spaces gained a new quality. ³

Early modern Istanbul essentially corresponded with the walled city, the area labelled as the historical peninsula today. This article focuses on the area within the walls together with the districts and piers located just outside the city gates. The inner wall region has a historically continuous relationship with relatively distant districts such as Eyüp, and Galata in particular – on the opposite side of the Golden Horn – as well as places outside the Golden Horn, such as Tophane, and finally, Üsküdar on the opposite side of the Bosphorus. This relationship continued uninterrupted during the seventeenth and eighteenth centuries, and the city even expanded towards the villages on the Black Sea coast of the Bosphorus. Both the continuity and the transformation of the walled city's connection with Istanbul's immediate surroundings is clearly visible in contemporary maps. For example, in Henry Beauvau's map from 1615 (Figure 1), only Galata and some of the prince islands were depicted outside the city wall, while Jacques Nicolas Bellin represented the city more fully in 1764 (Figure 2), framing the city as covering the entire Golden Horn and most of the Bosphorus. Another remarkable point in Beauvau's map is that it depicts Istanbul in clusters, reflecting the city's early modern morphology; it then consisted of various spatial clusters or units such as neighbourhoods, palaces, monumental buildings and social complexes, as Beauvau had clearly visualized.

Figure 1. Pictorial map ‘Constantinople’. Henry Beauvau, Relation iornaliere du voyage du Levant (Nancy: Iacob Garnich, 1615), 48. Source: gallica.bnf.fr / Bibliothèque nationale de France, https://gallica.bnf.fr/ark:/12148/btv1b8612079t, accessed 5 Sep. 2022.

Figure 2. Jacques Nicolas Bellin, ‘Plan de la ville de Constantinople’. Tomes IV, No: 125 (1764). Source: SALT Research, https://archives.saltresearch.org/handle/123456789/110007, accessed 5 Sep. 2022.

This article uses a variety of primary sources which help us to understand different aspects of Istanbul's spatial configuration during the seventeenth and eighteenth centuries. Of these, the most important are court registers from three different districts of the broader Istanbul area (Istanbul, Galata and Hasköy). ⁴ In addition to these, I also draw on other sources, such as fatwa collections. ⁵ Court registries were the main documentation tools of the Ottoman legal system. Judges (qāḍī) and their clerks scattered across the empire's provinces and districts made registers of different kinds of provisions, including legal cases that they judged, orders from the central administration, inheritance registers and real estate transactions. The last two type of registers have both potential and limitations in terms of understanding the spatial properties of a private house. These inheritance and transaction registers list all the main physical features of a house such as living rooms, kitchen, toilet, courtyard, garden, terrace, etc., allowing the identification of the physical elements of a house. However, the registers are devoid of key information, such as the size or layout of a house, providing neither pictorial nor textual information. It is also important to note that the quality and style of information about residential places in court registers varies between different regions and periods. ⁶ Despite these limitations, the court registers from the seventeenth and the eighteenth centuries from different parts of Istanbul allow at least some insight into the appearance of a place, and an understanding of the vocabulary used for defining it. ⁷

In the first section of my article, I demonstrate the central role of the public–private dichotomy as the main vocabulary of the literature on Ottoman public spaces. In the second, I focus on the door as the interface of spatial configurations, and its potential to overcome the limitations exposed by the public–private dichotomy. In the third, I introduce two types of doors from the Ottoman world, qāpū and bāb, which constitute the main vocabulary of my theoretical discussion. In the fourth, I demonstrate the socio-political and spatial role of qāpūs embodied in the households of the ruling elites. In the fifth section, I deal with bāb, which I consider as the key to the ultimate level of privacy. In the last section, I introduce other vocabulary to further enhance the study of the gradations of privacy in Istanbul. In the conclusion, I reflect on how using the conceptual repertoire of the period allows us to better grasp the concept of privacy, and the spatial configuration.

Public–private dichotomy

Most literature on the history of the Ottoman public sphere focuses on a particular space, the coffeehouse, as a very suitable case-study. ⁸ Considering that coffee is a completely modern beverage, and that the coffeehouse is a modern space, such an orientation is unsurprising. Yet, in his touchstone book, Jürgen Habermas attributes a great deal of importance to the coffeehouse in the context of the public sphere and rational discussion. In fact, to understand the dynamics of the literature on Ottoman public spaces, we need to focus on the fundamental role of Habermas’ theory. Habermas' work was published in German in 1962. ⁹ Although the book had a significant impact on philosophy, sociology and political science circles in the following years, this effect did not fully penetrate historical studies until its English translation in 1989. ¹⁰ The Turkish translation in 1997 triggered a wave of studies dealing with the public sphere in Ottoman society. ¹¹ By this time, Habermas had written a new and comprehensive preface for the 1990 edition, responding to nearly 30 years of accumulated critiques. Therefore, those who encountered the book The Structural Transformation of the Public Sphere after 1990, including the Turkish translation, were provided with various arguments, approaches and responses along with Habermas' main text. This is important in terms of understanding the relationship established with Habermas in Ottoman public sphere literature because this set of texts and arguments, which I describe as the Habermas set, constitutes the cornerstone of Ottoman public sphere studies. Nevertheless, although Habermas' conceptual framework is used predominantly in the literature, this does not imply that researchers established a purely affirmative relationship with it. ¹² On the contrary, very few researchers approach the Ottoman public sphere from a totally Habermasian perspective; instead, an important critical dialogue has emerged, aimed at overcoming weaknesses in his notion of the public sphere. ¹³ However, even studies that were critical continued to generate ideas that were in some way inspired by the body of texts and arguments defined as the Habermas set, above.

The critical conversation with the Habermas set revolves around two fundamental limitations to his theory: the excessive emphasis placed on rational discussion, and the exclusion of some social groups from the public sphere. Regarding Habermas’ ascription of a central role to rational discussion, Serdar Öztürk asserts that conflating public opinion with the notion of a reading public effectively ignores the existence of public opinion in Turkish history before the introduction of the newspapers. ¹⁴ In a similar manner, Eminegül Karababa and Güliz Ger state that Habermas considered the existence of rational individuals as a prerequisite for the formation of the public sphere, but that early modern Ottoman coffeehouse clients did not fit this description. ¹⁵ According to Cengiz Kırlı, who follows the same critical path, public opinion does not exist only when there is a public that participates in rational discussion; on the contrary, every thought expressed in public contributes to the formation of that opinion. ¹⁶ As the second path of criticism suggests, Habermas assigned bourgeois male individuals a central role in his theory while excluding other subjects/social groups. According to Elif Akşit, women have a central role in the formation of publicity, but they are excluded in the process of abstraction of the concept, and ironically, through their exclusion, they have a constitutive contribution. ¹⁷ Öztürk, on the other hand, imagines a more inclusive Ottoman public sphere, objecting to the view that the public sphere belongs only to the educated and wealthy bourgeoisie. ¹⁸ Both of these two critical paths evident in the Ottoman public sphere literature had already been answered by Habermas’ 1990 preface.

Despite establishing a critical relationship with Habermas, the most concrete evidence of researchers’ continued commitment to his conceptual vocabulary is the prevalence of the use of the public–private binary opposition According to Jeff Weintraub, the distinction between public and private is variable and mouldable, rather than holistic. This dichotomy consists not of a single opposition, but of many in complex relationships, and the various uses of the pair of concepts not only point to different phenomena, but also raise different problems and questions drawn from different concerns. Multiple public–private distinctions emerge from different theoretical languages and discursive universes, and all are laden with their own historical assumptions and connotations. ¹⁹ In his article, exceptional in the Ottoman coffeehouse literature, Alan Mikhail argues that no simple spatial dichotomy can accurately reflect the complex reality of the Ottoman city, as grounds for his claim that the coffeehouse is a space where different overlapping functions merge fluidly. For Mikhail, who is critical of Habermas’ boundary between the public and the private, Ottoman coffeehouses clearly demonstrate that this simplistic distinction alone cannot account for the entire Ottoman world. ²⁰

According to Uğur Tanyeli, the conceptual distinction indicated by the public–private split applies to Turkey only in the nineteenth century and beyond, and applying it to previous centuries leads to a misuse of current terminology. ²¹ However, if the public–private distinction was invalid in the Ottoman world before the nineteenth century, this does not necessarily mean that these two concepts were completely intertwined. Rather, what is at issue here is a conceptualization of the physical environment that is ‘more complex than this pair of terms can describe’. ²² ‘If one insists on naming it in those terms’, he argues, ‘it would be necessary to say that public life in the Ottoman Empire was as private as it was public, and public as well as private. It is almost impossible to draw a line between the two.’ ²³ However, this situation cannot be interpreted as meaning that there was a monolithic spatial structure in the Ottoman world. On the contrary, rather than a public–private dichotomy, there was a complex living order and spatial formation with numerous intermediate values of privacy and social life. ²⁴ Expressing a similar criticism regarding residential space, Hatice Özkaya did not believe that this space was free from public influence, in either practical or physical terms. ²⁵ Rather than arguing for a simple distinction of the kind implied by the public–private dichotomy within this structure, she states that ‘it would be more meaningful to think that the public and the private have different levels and forms’. ²⁶ In other words, there was a transition from the public to the private, and this gradation was created by various spatial components.

I should point out that my emphasis on the conceptual shortcomings of the public–private dichotomy does not mean that binary oppositions were completely invalid or meaningless for the early modern Ottoman world. On the contrary, some binary concept pairs such as bīrūn–enderūn (outer palace–inner palace) or khāriciyye–dākhiliyye (exterior–interior) were intrinsic to the Ottoman conceptual repertoire. In this regard, instead of simply accepting or rejecting binary oppositions, we need, as Tülay Artan suggests, to understand the blurred or transitional areas between these spatial oppositions. When ‘identifying ‘amm(e) and hass(a), terms used in the court registers (sicils) of the period, as referring respectively to the public and non-public’, Artan proposes an intermediate area, inspired from the existence of ‘privacies occurring within the public sphere’. ²⁷ The core problem with binary oppositions is not so much the concepts themselves, but rather, the historiographic and methodological issues caused by their anachronistic and political uses. For instance, Leslie Peirce, who has provided one of the best examples of such criticism, argues that it is a myth to describe the harem both as the women's world and private space. The main reason for the myth's persistence is historians’ adherence to Western assumptions when examining non-Western cultures, particularly about gender and politics. Peirce reminds us of the contemporary feminist criticism of the Western-oriented public–private dichotomy, which assumes that the family constitutes the non-political private sphere. Yet, this dichotomy was a construct created within historical processes. ²⁸

Doors and privacy

The effort to overcome the conceptual limitations of the public–private binary opposition to understand and conceptualize the spatial structure of Istanbul in the early modern period led me to consider doors as the interfaces of spatial formation, as argued by Georg Simmel. Simmel's starting point is the relationship between the acts of separating and connecting as human physical and mental activities. According to Simmel, by defining objects/things in nature as ‘separate’, we have already related them, and vice versa, and we then try to relate what we have already in some way defined as separate. However, trying to relate things that are not conceived of separately will inevitably be a vain effort from a practical and logical point of view. ²⁹ Simmel argues that the resistance created by the spatial separation between the two banks of a river has given a different meaning to the construction of bridges, which have come to symbolize the extension of the will of man over space. The bridge not only practically connects two physically separate shores for the benefit of the human body, but also visualizes this connection. ³⁰ In the relationship between separateness and unity, the bridge emphasizes the latter. The unique meaning Simmel attributes to doors emerges at this point, as the door represents both of the two actions.

By virtue of the fact that the door forms, as it were, a linkage between the space of human beings and everything that remains outside it, it transcends the separation between the inner and the outer. Precisely because it can also be opened, its closure provides the feeling of a stronger isolation against everything outside this space than the mere unstructured wall. The latter is mute, but the door speaks. It is absolutely essential for humanity that it set itself a boundary, but with freedom, that is, in such a way that it can also remove this boundary again, that it can place itself outside it. ³¹

An interface that fulfils both Simmel's separating and connecting functions, the door is the most basic component in the grading of privacy in the spatial structure. So much so that each new door creates a new privacy circle, and each new door in this circle makes the privacy even more distinct. This sequence continues until reaching the level of an individual room, the ultimate state of privacy and, finally, a door separates that room from the surrounding spaces.

In her ground-breaking book A Room of One's Own, ³² Virginia Woolf states that until the beginning of the nineteenth century, it was impossible for the great majority of women to have a room of her own. A woman was deprived of even the simplest facilities that a man could enjoy. For writing, a woman could only use the family living room, and thus, for a long while, without the seclusion needed for intense concentration, women authors tended to write prose, rather than poetry. ³³ The demand for a woman to have a room of her own seems to be related to privacy, specifically the privacy of intellectual production. However, Woolf's emphasis on the room becomes meaningful when considered together with the deprivation of women of rights, in contrast to the opportunities afforded to men. When the author tries to enter the famous library of Oxbridge, a fictional combination of Oxford and Cambridge universities, a male official warns her that ‘ladies are only admitted to the library if accompanied by a Fellow of the College or furnished with a letter of introduction’. ³⁴ So, for Woolf, the issue of privacy is also a question of public rights.

There is a striking similarity between Virginia Woolf's emphasis on a room of one's own and issues of privacy in the seventeenth- and the eighteenth-century fatwa collections. In the collection that contains the fatwas of Feyzullah Efendi, who served as sheikh al-Islam in 1688 and between 1695 and 1703, one of the fatwas on husband–wife relations concerns the privacy of a man's two wives. The fatwa considers whether, if a man has more than one room in his house, and each of his wives lives in a separate room with a door with a lock, one of his wives has the right to object to this situation and demand to live in a separate house. ³⁵ Fatwas are formulated in the form of long interrogative sentences, shaped around anonymous characters, and always end with either a positive or negative answer. In this fatwa, the demand of the anonymous wife named Hind for a separate house is answered negatively. We encounter the anonymous Ottoman woman Hind in a similar example in the fatwa collection of Abdullah Efendi, who held the office of sheikh al-Islam between 1718 and 1730. This fatwa regulates the relationship between a man's spouses and concubines in the same dwelling. The fatwa asks whether, if a man brings his concubine, with whom he is having sexual relations, to live in a separate room of the family home, his wife has the right to demand a separate house. Like the previous fatwa, the answer is no. ³⁶

At this point, we inevitably have to ask what unites Virginia Woolf, writing in the first quarter of the twentieth century, and Hind, an anonymous Ottoman woman two centuries earlier, around the theme of ‘a room of her own’? This question can be answered in anachronistic and teleological ways, such that even the simplest demands of the Ottoman woman to overcome polygamy were rejected by the legal system, or that an Ottoman woman could be entitled to a room of her own as early as in the seventeenth century, a privilege that British women did not achieve until the nineteenth century. Apart from importing all the prejudices of orientalist and occidentalist perspectives, these answers would be no help in understanding privacy. The important point emerging from these fatwas is that a separate room with a lockable door was considered sufficient to establish a woman's privacy in the household. Therefore, I argue that the room, and, if possible, a room of one's own, constitutes the nucleus of privacy in the spatial structure, and that this function of a room can only be fulfilled with the existence of a door.

Two types of doors in the Ottoman world: qāpū and bāb

The role and importance of doors is reflected in Ottoman spatial culture, in physical, lingual and textual terms, so much so that there are a variety of words that expressed differently nuanced meanings of door. In Ottoman Turkish, there are four common words for door: Turkish qāpū, Arabic bāb, Persian der and dergāh. In this article, I focus on two of these words, qāpū and bāb, both of which have the same meaning and are heavily used in historical texts, particularly with reference to the state and urban places. ³⁷ Even though it is difficult to draw clear boundaries between the usage of the words qāpū, bāb, der and dergāh, we can determine some criteria that will demonstrate the conceptual differences. The words der and dergāh are mainly used in Sufism and for more abstract expressions, but are also used to a relatively limited extent in expressions related to the state such as der-i āmed (income) and dergāh-ı ‘ālī (sultan's palace). We can draw a clearer distinction between the words qāpū and bāb. Many offices and official roles in the Ottoman state were referred to as bāb, such as bāb-ı ‘ālī (palace of the grand vizier, government), bāb-ı defterī (ministry of finance) and bāb āls‘āde āghāsī (manager of the interior palace). Similarly, numerous official terms include the word qāpū, such as qāpū qūlī (people of the sultan), qāpū çūqadārı (broadcloth servant of the sultan) and qāpū āghāsī (head of the eunuch servants). The main difference between qāpū and bāb can be explained with the gate–door distinction, which has no direct counterpart in either Ottoman or contemporary Turkish, but which I find very meaningful for the conceptual discussion I am conducting here. The palaces, retinues and spheres of influence of the sultan, princes and high bureaucrats are known as qāpū, or qāpū people. Where the word qāpū refers to the palaces of the dynasty, high bureaucrats and elites, it represents an entire spatial complex, thus symbolically representing its gate. There is no such similar use of the word bāb. However, looking at the various archival sources and especially the court registries, we observe that the word bāb is extensively used for describing buildings’ physical properties. In sales or inheritance records that mention a house, mansion or workplace, bāb is often used for the units within these spaces. In these specific cases, bāb bears the meaning of door, an interface that exists in multiple ways, in both the horizontal and vertical diversification of a place.

To avoid misconceptions, I should clarify that the idea of representing the distinction between qāpū and bāb as parallel to the notions of gate and door is not always applicable, and should not be taken as a generalization. For instance, an examination of city gates reveals that qāpū and bāb are used interchangeably in their naming such as Bāb-ı cedīd and Yeni qāpū (the new gate) or Bāb-ı Edirne and Edirne qāpū (Edirne gate).

Qāpū as a socio-political and spatial institution

Qāpū is one of the oldest expressions used to describe the Ottoman state. The term devlet kapısı (gate of the state) is similarly used today and means ‘to be in public service’. ³⁸ Interestingly, today the Turkish government's central online portal used for almost all official procedures is also named e-devlet kapısı. ³⁹ Since the central state mechanism in the Ottoman Empire operated only at the Topkapı Palace, qāpū referred to the sultan's palace as well as to the state. The term qāpū khalqī (door people/household people), as noted above, was used to refer to all those who are in a position of service. ⁴⁰ All officials and soldiers paid from their master's revenues were counted among the people of that household. In this context, what we now call state bureaucracy was nothing but the sultan's qāpū, even until the second half of the sixteenth century. ⁴¹

By the second half of the seventeenth century, the central role of the sultan's qāpū in the state bureaucracy changed significantly. Rather than the palace or the military, the ruling elites were increasingly selected from the vizier and pasha qāpūs, the two main human resources of the Ottoman administrative staff. ⁴² In fact, the growth of the vizier and pasha qāpūs, and the increasing numbers living within or employed by these, had been developing for some time. What then does the increasing importance of qāpūs indicate in the seventeenth century? To answer this question, Rifaat Abou-El-Haj conducted a statistical analysis of the origins of Ottoman ruling elites who came to high-level positions in the 20-year period from 1683 to the 1703 rebellion. By studying ruling elites’ growth and progress through the bureaucracy, both in the central government and in the provinces, Abou-El-Haj was able to classify the origins of administrators by determining categories such as palace, military, beyzade, civilian and vizier–pasha qāpū. He demonstrates that, in the central administration, the largest proportion were from the vizier and pasha qāpūs, followed by those with palace and military origins, and correspondingly, there was a great decrease in administrators with palace and military origins. ⁴³

Vizier and pasha qāpūs initially served the interests of the Ottoman dynasty, like the palace and military institutions, but eventually replaced these and created their own justification for existence. Qāpūs had no uniform position, and could align with the court or military factions according to circumstances. In any case, as previously claimed, struggles for the throne during the seventeenth century took place not only in the palace or between factions in the army. The rise of the qāpūs brought a new force into this power struggle. ⁴⁴ The sultan's power began to be redistributed and shared out more broadly and as a result, the perception of the state consisting of a single imperial qāpū changed, and power shifted to elite qāpūs as competing factions. ⁴⁵ Despite the lingering persistence of the religious roles of the sultan, such as ghazi or hādīm'ül-haremeyn'ül-şerifeyn (servant of the two sacred mosques in Mecca and Medina), the warrior sultan gradually transformed into a more symbolic figure. ⁴⁶ Depending on the growth and differentiation of the state, various bureaucratic units moved out of the palace to new headquarters. ⁴⁷ This transformation of the machinery of state also created a significant differentiation in the texture of the ruling class, with the erosion of its boundaries with the reaya (subjects of the sultan). As the civil servant class grew, the hierarchical mobility within that class also began to increase, and high-ranking officials began to rise to positions such as beylerbeyi (governor-general) and vizierate, which had been largely occupied by the sultan's servants in the fifteenth and the sixteenth centuries. In this process, accompanied by factions around different qāpūs, the ‘masters’ of the civil service system became the new ‘pashas’. ⁴⁸

Undoubtedly, the prototype of the Ottoman elite households was that of the sultan, the Topkapi Palace, which combined the characteristics of the domestic and military households. Various palace officials, such as baltacı or bostancı, combined domestic and military roles. ⁴⁹ Although the sultan's palace was the supreme household not only in Istanbul, but across the empire, it faced rivalry in the seventeenth and the eighteenth centuries from the viziers and the provincial administrators’ qāpūs, most of whom had started their careers in the palace, leading to a dissolution of the power of the sultan's household. The Janissaries, traditionally regarded as an element of the sultan's household, also became actors in this process of competition and alliances. A similar tension was found, albeit on a smaller scale, in the provinces, where local rulers imitated the sultan's household, but found themselves facing competition from the local elite households. ⁵⁰

The social, cultural and political importance of households in the early modern Ottoman world, especially in Istanbul, may be understood in term of Baki Tezcan's ‘Second Ottoman Empire’ theory. According to Tezcan, during the late sixteenth and the seventeenth centuries, a great socio-economic transformation brought to an end the patrimonial empire period (1453–1580), whose ideal form was identified with the period of Suleyman the Magnificent. ⁵¹ Instead, in the late sixteenth and early seventeenth centuries, the vizier households emerged as an alternative political focus. However, during the reign of Murad III (1574–95), new actors supported by the palace entered the political arena and the sultan, disturbed by the vizier households’ extensive network, developed his own rival network. ⁵²

Thus, the transformation of power that dominated the Ottoman Empire in the early modern period was embodied in the qāpūs. Both the qāpū's role in socio-political processes and its spatial function as household make it a key concept for understanding the spatial formation of Istanbul and, I argue, a suitable term for framing the spatial units of the urban topography. Obviously, in its core spatial meaning of household, qāpū only refers to the palaces and mansions of the sultan and the ruling elites. However, given both its socio-political significance and other spatial functions, I believe qāpū contains new conceptual possibilities. Based on this, I propose to use the term qāpū to refer to every single spatial unit of Istanbul's topography that combined blended public and private roles, such as palaces of the high bureaucrats and elites, embassy palaces, social vaqf (foundation) complexes, sufi lodges, military barracks, khāns (inns), neighbourhoods, etc. I will take the last two of these, namely neighbourhoods and khāns as two unambiguous examples of what I propose.

After the Ottoman conquest in 1453, Istanbul was organized in the form of nahiyes (districts), each composed of multiple neighbourhoods. ⁵³ While the nahiyes were more formal and administrative units, neighbourhoods had an administrative role, as well as hosting most activities of daily life. A neighbourhood ‘was an organic unity, a community with its own identity, settled around a mosque, a church or a synagogue’. People living in a neighbourhood were connected in multiple ways, including having common origins, being members of the same religion and undertaking joint responsibility for taxes, security and order. ⁵⁴ The main socialization areas of a neighbourhood were places such as the square, place of worship, hammam, fountain and coffeehouse. There was a direct connection between a particular house and the rest of the neighbourhood. Most residents of Istanbul by then considered their neighbourhood as an extension of their dwellings, and places such as doorsteps, cul-de-sacs and shared courtyards had transitional roles between public and private functions. ⁵⁵ Many neighbourhoods had gates that limited entry, and were, in theory, but not always in practice, locked at night. ⁵⁶ In other words, the analogy I create by defining the neighbourhood as a qāpū was actually embodied by a physical gate separating the neighbourhood from the outside.

Another example that conceptually fits qāpū is the khān, which was an essential part of Ottoman commercial life, as a place of lodging, storage and wholesale. ⁵⁷ As Ahmet Yaşar argues, a khān, particularly an urban khān in eighteenth-century Istanbul, was a ‘microcosm of what was happening in the architectural, social, economic and political milieu of the capital’. ⁵⁸ A khān might comprise rooms, courtyards, small mosque, coffee shop, stables, shops and vaults ⁵⁹ and as such was a site of multiple spatial functions that can be attributed both to private and public lives, and all the transitory moments between. Yet, it was self-sufficient, as well as being connected with the outside.

Imagining early modern Istanbul as composed of multiple qāpūs is the first step of the visualizing of gates and doors as the interfaces of spatial formation. After using qāpū as the first key concept to undermine the rigid concept of public–private opposition, in the next section, I will focus on another kind of door, bāb, to shed light on how privacy is graded within a qāpū.

Bāb, the ultimate unit of privacy

Bāb is only one of the words for door in Ottoman Turkish. However, it is conceptually distinguished from its synonyms because it was commonly used in the court registers to identify various places, especially residential ones, and the subdivisions within them. Just as the qāpū represented the households of the high bureaucrats and the sphere of their power, the bāb refers to the places defined as menzil (house), ev (house) and dükkān (shop), and each of the rooms that made up these places. Istanbul residences, except for those consisting of a single room, were defined using the concept of menzil. Despite numerous different words referring to dwellings, bāb is indisputably the most common concept in the definition of early modern Istanbul residences, and below I give examples from court registers demonstrating its usage.

The main usage area of the word bāb in the court registers was to determine the parts of the house. Bāb is found in almost every provision, from short definitions of ordinary dwellings to comprehensive definitions of large dwellings. In the case of a house consisting of one or two rooms, we often encounter the following types of records:

• ‘A house including one bāb taḥtānī [ground floor] room.’ ⁶⁰

• ‘A house including two bāb rooms, one taḥtānī and one fevqānī [upper floor].’ ⁶¹

• ‘A house including one bāb fevqānī room, one bāb taḥtānī room and a courtyard.’ ⁶²

• ‘A house including a garden, one bāb taḥtānī and one bāb fevqānī room.’ ⁶³

• ‘A house including one bāb room, storeroom and a barn.’ ⁶⁴

• ‘A house including two bāb fevqānī rooms and underneath a wine vault.’ ⁶⁵

Houses that were slightly larger than the examples above, with more than two rooms and with more variety in units were defined with the following expressions:

• ‘A house…including three bāb taḥtānī and one bāb fevqānī rooms, enclosed courtyard and toilet.’ ⁶⁶

• ‘A house including two bāb fevqānī and one bāb taḥtānī rooms, woodshed, water well and toilet.’ ⁶⁷

• ‘A house…including two bāb fevqānī and one bāb taḥtānī rooms, a rooftop terrace, a water well, a faucet and a toilet.’ ⁶⁸

• ‘A house including three bāb taḥtānī rooms, a hammam, a dressing room and underneath a masonry vault, a rooftop terrace, a kitchen, four toilets, four water wells and garden with fruit bearing and non-fruit bearing trees.’ ⁶⁹

Court records also contain information about residences much larger than the examples above, referred to either as a mansion, or denoted by a word with a similar meaning. The complexity of the descriptions of these dwellings, some of which had separate exterior (khāriciyye) and interior (dākhiliyye) sections, depended on the size of the house. ⁷⁰ In court records, the components in the various sections and floors of such large houses are individually listed, and the convention of describing the rooms with the word bāb is maintained. ⁷¹

Examining a wide range of provisions, we can easily observe patterns regarding the definition of houses; the number of rooms is generally specified in terms of bābs. The alternative word bayt (room, pl. büyūt) is used in some provisions, but this creates no change in the meaning. In addition to the number of rooms, there are details of other open, semi-open and closed spaces that make up the house. These spaces varied according to the location, quality and size of the house. The interesting point is that the number of units that do not fulfil the criteria of rooms is rarely stated in terms of bāb, which was a concept that was solely identified with rooms. However, it is important to clarify that while such a pattern is observed in sale and inheritance provisions in the court registers, there are other usages of the term bāb in different kind of provisions with different subjects. In other words, in definition of residences, rooms are associated with the word bāb, but in specific circumstances, bāb may also be used for spaces other than rooms. For instance, bāb refers to the entire space in some short descriptions, such as ‘two bāb house’, ⁷² ‘one bāb property house’ ⁷³ or ‘one bāb fevqānī house including water well, toilet and small garden’. ⁷⁴ However, it is notable that, in these descriptions, no specific rooms are mentioned. In early modern Istanbul, as in many other cities, not everyone could afford detached houses. Many poor people were living in complexes composed of multiple rooms with common facilities, such as toilets. In descriptions such as ‘thirty-six bāb rooms’ ⁷⁵ or ‘thirteen bāb yahūdikhāne’, ⁷⁶ the word bāb refers to the separate rooms of these spaces.

Another usage of bāb in court registers relates to workspaces. In some provisions, shops are referred to in terms of the number of bābs, as in ‘one bāb shop’, ⁷⁷ or ‘three bāb shops’. ⁷⁸ In other examples, the type of shop is also included in the definition, either with or without the ‘shop [dükkān]’ word.

• ‘One bāb grocery shop.’ ⁷⁹

• ‘One bāb tripemaker shop.’ ⁸⁰

• ‘One bāb sherbet and kebab shop.’ ⁸¹

• ‘Two bāb jewelry shop.’ ⁸²

• ‘Two bāb fish shop.’ ⁸³

• ‘Two bāb grocery with masonry vault.’ ⁸⁴

• ‘Forty bāb bazaar shops.’ ⁸⁵

• ‘One bāb bakery oven.’ ⁸⁶

• ‘One bāb windmill.’ ⁸⁷

• ‘One bāb tile kiln.’ ⁸⁸

In the light of all these examples, I argue that bāb was used to denote spaces clearly separated from other spaces that surrounded them, and that could be isolated from these if necessary, regardless of whether they were used for a house, shop or room. When used for the whole house, it expresses the specific independence of the whole place from other places, and when used for the rooms of the house, it highlights the same independence between the house and each of its rooms. As a result, the bāb directly affected the establishment of privacy by protecting a space from the spaces outside it, from the entrance and gaze of outsiders. Bāb had the same separating function in a qāpū. Remembering that qāpū was a complex and large spatial unit with various functions, bāb also provided gradation within a qāpū. However, as I show in the next section, there were other spatial segments and partitions in a qāpū that contributed to the gradation of privacy.

Circles and clusters of privacy

In the last section of my article, to further elaborate on the gradation of privacy, I will deal with two additional features of early modern Istanbul spaces. These additional features and related vocabulary help us to better grasp how multifunctionality was maintained between a qāpū and the bābs within it. Among these, I will focus on the taḥtānī–fevqānī (ground–upper) and khāriciyye–dākhiliyye (exterior–interior) distinctions, the former relating to the multiple floors of a building, and the latter to the exterior and interior sections of large houses.

The words taḥtānī and fevqānī were used very widely in the definitions of residences in the court registers, and the dozens of residences subject to the provisions were defined according to the number of taḥtānī and fevqānī rooms. This is evident in most of the examples examined above in the section discussing the bāb. The word taḥtānī means ‘below, the one below’ and fevqānī means ‘the one on top, the one above’, and when used in relation to each other, they refer to the lower/ground and upper floors of a building. However, the ground and upper floors may have had different functions, which are also reflected in the taḥtānī–fevqānī terminology. The fevqānī was the main living space in a mansion and was raised above ground level. Separating the fevqānī from the ground was the taḥtānī floor, made of wooden pillars and stone walls. Traditionally, the taḥtānī floor was simply an open space over which the house was constructed, but over time it became an increasingly complex concept. ⁸⁹ The simplest residences encountered in court registers are defined as ‘one bāb taḥtānī house’ ⁹⁰ or ‘a house featuring one bāb taḥtānī room and toilet and courtyard’. ⁹¹ These houses consisted of a single room, with basic features such as a toilet. Even without a fevqānī floor, the taḥtānī quality of the existing floor was described, as it gave information about the quality and building technique of the dwelling. As Cem Behar explains, fevqānī referred to houses with more than one floor (usually two floors), whereas taḥtānī houses were average-quality single-storey buildings. Sufli (regular) houses were also single storey, but smaller and simpler. In contrast, big and impressive houses of the high-ranking bureaucrats were defined as mükellef (grand). ⁹²

It is difficult to make a clear distinction between the functions and architectural roles of taḥtānī and fevqānī floors in the context of early modern Istanbul residences, but fortunately some registers provide important details. One house, registered due to an unresolved debt between Abdi Çavuş and his wife Rahime Hatun from Debbâğzâde neighbourhood, is defined as ‘including three fevqānī and one taḥtānī house, a barn and a hayloft’. ⁹³ There was only one taḥtānī room in contrast to three fevqānī rooms, suggesting that the barn and the hayloft were also on the taḥtānī floor. In some registers, it is clearly indicated whether units were located above (fevk) or under the other units. For instance, a house registered as ‘including two bāb fevqānī rooms and a wine vault underneath’ ⁹⁴ leaves no doubt that the taḥtānī floor was reserved for non-residential functions. The same functional distinction is visible in a house ‘including one bāb fevqānī room and underneath a barn and a shared water well’. ⁹⁵ In these examples, if the house consisted of more than one floor, the fevqānī floor served as the main accommodation area, while the taḥtānī floor was reserved for functions such as entrance, barn, vault and storage. Therefore, within the same building, there was a significant difference between the floors in terms of the practices of private life.

Another pair of concepts to consider regarding the gradation of spaces is the khāriciyye and dākhiliyye. Most middle and large-scale houses consisted of two main sections, namely khāriciyye, the men's section that was more connected with the outside world, and dākhiliyye, the core domestic and private section. ⁹⁶ In the court registers, there are many provisions about houses that consisted of both khāriciyye and dākhiliyye sections. One such house is described as ‘a house including many rooms in dākhiliyye and khāriciyye and a water well and two toilets and a garden with fruit bearing and non-fruit bearing trees’. ⁹⁷ This demonstrates that there were rooms in both the dākhiliyye and the khāriciyye sections, even though the distribution of the other facilities between the two sections is not clear. According to a more detailed register, the khāriciyye section of a medium-sized house consisted of two rooms, a hall, barn, courtyard and toilet. In the dākhiliyye of the house, there were two rooms, two cellars, a gateway (dehliz), a kitchen, a well, a small garden (cüneyne) and a toilet. ⁹⁸ The two sections of this house both had two rooms and a toilet respectively, but the main difference was that the barn was in the khāriciyye, and the kitchen and the cellars were in the dākhiliyye. Another provision in the same register book, this time concerning the exchange of two houses, provides even more detailed information. According to this long provision, which includes detailed descriptions of both houses, in the khāriciyye section of the first house, there were two rooms, a balcony (gurfe), a barn, a water well and a courtyard; in the dākhiliyye section, there were four rooms, three halls, an oven, a well and a garden. In the khāriciyye section of the second property there were three rooms, a large central hall, a side hall, a gateway (dehliz), a passage hall (mabeyn dehlizi), two servants' rooms, a barn, a well, a courtyard and a garden; in the dākhiliyye section, there were four rooms, side halls, a kitchen, a fenced courtyard, a pavilion, a courtyard and a garden. ⁹⁹ In both houses, as in the previous example, the barn was located in the khāriciyye. In cases where the kitchen and the oven were listed among the features, both were in the dākhiliyye. Another important point is that the servants' rooms were located in the khāriciyye of the second house. The servants resided in these rooms when not working, i.e., their dwellings were in the khāriciyye section, suggesting that the comparatively domestic characteristics of dākhiliyye in the context of khāriciyye–dākhiliyye distinction may not always have applied, depending on who was using the space. Even though the dākhiliyye was the owners’ main residential and private space, for some other members of this small community, such as servants, their main private space was located in the khāriciyye. The difference between the khāriciyye and dākhiliyye parts is especially evident in visual privacy. When we compare these parts of a house in the Debbâğzâde neighbourhood of Istanbul, we see that the cihannüma, a roof terrace with views on all sides, was in the khāriciyye and the bath, requiring the highest level of privacy, in the dākhiliyye. ¹⁰⁰

While the khāriciyye and dākhiliyye parts of many houses were sufficient and complete units in themselves, the difference emerges in the functional and privacy-related details such as those mentioned above. In some houses, it is relatively difficult to define the distinction between these two parts. For example, the house that Janissary Hasan Çelebi sold to İbrahim Ağa had a barn and an oven in both sections, which is very exceptional indeed. However, in the last instance the khāriciyye differed from the dākhiliyye, as it included a water well with a cupboard and a floor mill. ¹⁰¹ It is uncertain whether the mill was for the residents’ own needs or for generating income. However, in some houses, there were shops with clear economic purposes. ¹⁰² If a house with the khāriciyye–dākhiliyye division included a shop, it was inevitably located in the former section.

Provisions regarding houses that only had either a khāriciyye or a dākhiliyye part seems misleading and self-contradictory. However, such examples are encountered when the khāriciyye–dākhiliyye parts were, for various reasons, turned into independent residences, built separately, or when one of the two parts was endowed for a vaqf (foundation). ¹⁰³ Therefore, houses with only a khāriciyye or dākhiliyye, rather than blurring the conceptual distinction between them, in fact support the integrity and self-sufficient quality of these two parts in themselves. So much so that each could be transformed into an independent residence when necessary. However, although both were self-sufficient, the ultimate distinction remains valid; architecturally and functionally, the khāriciyye was the part of the house that was more connected with the outside and the public space, whereas the dākhiliyye was the more private. ¹⁰⁴ In this regard, it is no coincidence that spatial components such as a shop or a coffee room were in the khāriciyye.

Conclusion

Many places in early modern Istanbul embodied different functions that can be associated both with public and private life, as demonstrated in the previous sections of this article. Various spatial components helped to sustain these functions, such as the taḥtānī and fevqānī floors or the khāriciyye–dākhiliyye sections. This multifunctionality suggests a complexity beyond the binary opposition found in the public–private distinction, and opens the way to other conceptual possibilities. In this article, I introduce qāpū and bāb as the two ends of this conceptual spectrum. Rather than using the concepts of qāpū and bāb with the meaning attributed in the sources, it was necessary, to some extent, to redefine these two concepts, resulting in a combination of current theoretical needs with the conceptual repertoire of the historical sources, and thus, an appropriate method for overcoming the limitations of the public sphere studies as applied to Ottoman history.

In his Second Ottoman Empire theory, Baki Tezcan uses the analogy of a spider web for better articulating the new relations among different power holders. ¹⁰⁵ Tezcan's analogy is also applicable to the spatial configuration and practices of Istanbul, in reciprocal relation with the political and social transformations of the period. In this spider web, the cells represent qāpūs of varying sizes, either close to the centre or at the periphery. Depending on its size or functions, a qāpū could be the sultan's palace(s), the palace of a high-ranking bureaucrat, palaces of an embassy, a sufi lodge, a vaqf complex, a neighbourhood or other kinds of spaces that were self-sufficient, featuring both public and private functions. Whichever of these it was, it occupied a certain space in the urban topography, as one of its building blocks. Neither the distinction between a qāpū and the spaces surrounding it, nor the diversity within a qāpū, can be explained simply in terms of the public–private dichotomy; the spatial gradation within a qāpū can only be explained with the concept of bāb. Each bāb created a new degree of privacy, but public and private functions could continue to co-exist in various proportions until a final level was reached, again identified with a bāb, which was wholly private. Therefore, by advancing Tezcan's spider web analogy, we can imagine new circles of privacy in every cell of the web.

Using qāpū and bāb as key concepts is of course closely related to the synonymous meaning of these words as ‘door’. As Simmel points out, a door both separates and connects spaces functionally and culturally. In addition, separation and connection roles work bidirectionally between the inside and outside of a door. The richness of the connotations of these two kinds of doors have a much better potential to explain the multifunctional and dynamic structure of Istanbul spaces than the public–private dichotomy.

Considering the greater potential of using qāpū and bāb compared to the public–private opposition, these concepts should be used in future studies attempting to re-establish the relationship between daily life and spatial formation. Thus, we can arrive at more realistic explanations for the role of different spaces in processes, such as the circulation of ideas and opinion making. The case of reading groups is another related area. This meeting type, where one person read aloud a book for the benefit of others, was practised in different places such as coffeehouses, mosques, village rooms or houses. ¹⁰⁶ To what extent did the group reading practice give the private house a public role? How did conducting the same practice in different places strengthen the links between these places? What were the differences between the public and private practices in a particular space? I believe the concepts advanced in this article will be helpful for guiding future studies.