Statistics : Deren Ertas
|
Ottoman State Archives, DH.MB.HPS, 148/1 (1130 H.)
Statistics : Deren Ertas 1. Statistics denotes the science of collecting, organizing, and analyzing data.1 It is essentially a method that holds the normative position that quantifying and identifying patterns and norms constitutes a way of knowing. Statistics is etymologically rooted in the Latin word status (condition) and shares a root with the term State, a convergence that reflects more than a linguistic coincidence. From the administration of colonies to contemporary algorithmic governance, statistics has served as a technology of abstraction, rendering populations legible, comparable, governable. For social scientists, statistics are mass data that can be used to reconstruct the lives of ordinary people in the past or present, as well as to understand aggregate conditions and public opinion for policymaking. However, we seldom examine how statistics themselves are collected. In this essay, I focus on the implications of thinking about statistics as a political concept. My contention is that being a technology of government does not immediately make statistics a political concept. Rather, statistics becomes political in moments of data collection where different logics of accounting encounter each other. In other words, statistics is politicized by communities or individuals who refuse to be counted, misreport data, or reject imposed classifications. I suggest that these moments expose the conceptual limits of statistical reason: they reveal that statistics is not simply a mode of knowledge, but a site of political struggle over recognition, legibility, and rule. Furthermore, I argue that the way statistical data is compiled should be critically examined by social scientists, historians, and others who use it beyond the frame of methodology and its numerical constraints. The initial articulations of statistics as a way of knowing and describing the world date back to the seventeenth century. In the l660s Lower Saxony, politics professor Hermann Conring delivered a series of lectures on the finances, populations, and agriculture of European states—a tradition that culminated in the publication of Microscopium Statisticum: quo status imperii Romano-Germanici cum primis extraordinarius, ad vivum repraesentatur in 1672.2 Some decades later, in 1749, Gottfried Achenwall used the term Statistik to denote a descriptive field concerned with the structural features and the relative power of European states.3 These uses of statistics were more about comparing the relative size of workforces, agricultural capacity, and military power of European states vis-à-vis one another. A parallel field, called political arithmetic, developed as a quantitative approach to governance, taxation, population, and trade, which more closely resembles what we associate with statistics today. Modern statistics, I would suggest, is a marriage between early modern Statistik and political arithmetic: it is both a logic of governance and a manner of reducing complex data to abstractions that can be compared across nations. Michel Foucault’s work on governmentality provides one of the most influential accounts of how statistics came to function not simply as a descriptive science but as a tool of governance. In the lectures at the Collège de France in the second half of the 1970s, he traced a shift in the logic of political power beginning in the eighteenth century.4 He explained that sovereignty, which had traditionally operated through territory and the application of law, was supplemented—and in some cases supplanted—by disciplinary and biopolitical power.5 The object of biopolitical power, what it aimed to govern, was population, which it governed not through edict but through the management of life processes: fertility, mortality, disease, and so on.6 Censuses, vital records, and surveys proliferated, as did statistical government.7 Statistics was essential to the growth of biopolitics because it enabled power to view the population as an aggregate and revealed that human behavior, when viewed as a whole, exhibited patterns that could be quantified and projected. These patterns became the basis for “regulatory mechanisms” that sought to stabilize and normalize population-level processes such as life and death, health and disease, work and idleness, and so on.8 As Foucault put it, “It is thanks to the perception of the specific problems of the population, and thanks to the isolation of the level of reality that we call the economy, that it was possible to think, reflect, and calculate the problem of government outside the juridical framework of sovereignty.”9 Statistics enabled a form of governance that could operate without direct coercion, through knowledge and normativity. It produced not only new data but new ways of seeing: a world of means, averages, rates, and tendencies that could be visualized in graphs and charts. Foucault’s explanation of biopolitics is useful for understanding statistics as a technology of governance over a population; however, it only tells one side of the story. Politics, as I understand it following Jacques Rancière, emerges when there is a conflict over the basic structures of power, its norms, and its deployment. To that end, I argue that statistics becomes intelligible as a political concept only when its accounting logic is contested by people who reject its basic premise. This means focusing not on what statistics reveal, but on what they presuppose about why and how people should be counted, what data is necessary, and what it means for the data to cohere. Statistical practices do not simply reflect populations; they constitute them as legible objects, shaping how individuals and groups are seen, understood, and governed. In this view, enumeration appears to be more than a mode of observation: it is, as Louis Althusser has suggested, a mechanism of interpellation. In Althusserian terms, statistics is a mediating technology that hails individuals into subject positions within the different ideological apparatuses of the state.10 To the extent that it is a method of interpellation, statistics is politically consequential. What happens when those who are interpellated refuse to respond in ways that are legible to the state? The history of how statistics are produced contains many moments of evasion, contest over categories, and concealment of facts. It is precisely in these moments that statistics becomes more than a technology of government; it becomes politicized and thereby graspable as a political concept. In developing statistics as a political concept, I mainly draw on Rancière’s theory of politics, which locates the political not in institutional negotiation, but in the interruption of an established order by those who have no part within it.11 Politics is the moment when a subject without a part asserts themselves within a system that does not recognize them—a disruption in what he calls the distribution of the sensible. In this sense, politics does not begin with inclusion but with disagreement: a refusal to be seen, heard, or counted in the way the order prescribes. When people resist being counted—when they falsify information, reject imposed categories, or disappear from statistical visibility—they do not merely frustrate governance. They enact dissensus: a disruption in what Rancière calls the distribution of the sensible, where the apparent neutrality of legibility is fractured by what it cannot account for. But while Rancière sharpens our understanding of how disruption works, he tells us less about how dominance is maintained. To illuminate this dimension, I turn to Stuart Hall, who elaborates a direct critique of Foucault on this point. According to Hall, Foucault’s emphasis on regimes of truth and power ultimately displaces ideology as the object of critique, thereby weakening our grasp of political struggle.12 Resistance, in Foucault’s account, seems to emerge from nowhere: it is ubiquitous and therefore guaranteed, but it lacks a model of social structure that helps us analyze resistance in terms of actual relations of force. As Hall argues, this collapses power into a flat field of effects, without any way of conceptualizing the balance of forces that structure social domination.13 What Foucault calls a regime of truth, Hall insists, must be situated within a field of ideological struggle: there are dominant and subordinated regimes, and their relative power is not merely discursive but historically and materially conditioned.14 This insight helps us understand statistics not just as a mechanism of normalization, but as a site where dominant modes of legibility obscure and suppress competing forms of knowledge. The statistical categories that structure state vision are not neutral abstractions or attempts at representation. They reassert their dominance against “subordinated regimes of truth,” re-narrating refusal, evasion, and concealment as a temporary setback in the path toward statistical universality.15 Yet, the archive is rife with traces of contestation over statistical accounting. In this sense, silence, manipulation, and categorical disidentification are not just a disruption of form. They coincide with a struggle over the truth, or at the very least, a struggle over which truths can and should be known, by whom, and to what end. To observe the political dimensions of statistics requires a methodological shift in both how we understand the history of statistics and how we regard statistical data in research. Rather than treating it as a transparent reflection of the population, or simply as the action of a dominant ideology on the subjects of a polity, this approach reads enumeration as a contested practice: an encounter between the state’s desire for legibility and the complex social realities it seeks to regulate. As Adi Ophir has argued, concepts become political when something is at stake in their use, when their meaning is contested, unstable, or under pressure.16 Statistics, in this sense, is political and requires conceptualization not simply because it refers to something political, but because it is the object of a political struggle: its function, boundaries, and authority are challenged. 2. My approach to delineating the political stakes of statistics is by historicizing the encounter between state statistical projects and the subjects they sought to enumerate.17 Based on cases of statistical flight in the late Ottoman Empire (roughly from 1831 to the 1910s) and the early Turkish Republic (1920s), I examine how imperial subjects and, later, citizens of the nascent nation-state engaged with the statistical practices of their governments. Doing so, I consider different kinds of sources. On the one hand, I engage texts that shaped the conceptualization of statistics as a state science—census instructions, statistical treatises, reports from ministries and advisory councils, etc.—that offer insight into the epistemological assumptions that undergird statistical rationality. They describe what is worth knowing, how populations are imagined, and how numbers are to stand in for lives. On the other hand, I read these sources alongside the archival traces of those who resisted or reconfigured the logic of enumeration: through avoidance, falsification, selective self-reporting, or outright refusal. These moments do not always declare themselves by advancing alternative truth claims in a self-conscious manner, but can be inferred from silences, inconsistencies, and the reports or anxious commentary of state officials.18 What makes this historical context especially illuminating is not the ambitions that undergird the introduction of statistical methods into government, but the degree to which those efforts encountered friction at every scale. From the imperial capital to rural villages, actors resisted, reshaped, or simply ignored the statistical gaze in ways that cannot be reduced to a lack of capacity or technical failure on the part of the state. After the Turkish Republic was founded, and statistics became a matter of nation-making, the temporal lag between the founding of the state and the citizens being convinced that disclosing private information to state officials was part of a common generated new kinds of conflicts. Together, these cases expose statistics-making as the outcome of political contestation rather than merely a top-down process. In the late Ottoman Empire, statistics emerged as both a science of governance and a marker of modern statehood. By the mid-nineteenth century, the early modern practices of recordkeeping and land surveying, whose primary purpose was to facilitate tax collection and the organization of the military, no longer met the needs of governance as understood by imperial officials. Consequently, new technologies of governance were employed to produce statistics on population, agricultural and industrial production, and trade.19 The first census, conducted in 1831 under Mahmud II (r. 1808–1839), was designed primarily to count the empire’s Muslim and non-Muslim male population for purposes of conscription and taxation.20 A sense of failure pervaded this census: officials frequently reported incomplete data; provincial notables resisted its procedures; in some areas, villagers fled to remote areas to avoid being drafted into the military or facing new taxes.21 In 1844, the government sought to count the nomadic populations, who frequently thwarted census officers, as part of a wider effort to settle them. Yet another census was attempted between 1866 and 1873, during which officials went from house to house to record the number of individuals, along with their ages, marital statuses, professions, and real estate holdings. At the end of this census, Ottoman subjects were issued identity cards that served as their proof of citizenship. Amid mounting debt, territorial losses, and demographic upheaval, a dedicated statistics office was established in 1875 within the Ministry of Commerce.22 Its regulations on census methods and registration systems reflected both bureaucratic ambition and political insecurity.23 The planned census of the late 1870s failed to materialize due to numerous crises, including the 1877–78 Russo-Ottoman War, which redrew borders and brought a massive influx of Muslim refugees. These upheavals only heightened the drive for legibility. By the time of the 1881–82 census, statistics came to be described as a civilizational aspiration that was “imperative, urgent, and essential” for the rule of law and the preservation of property and security.24 The methods of census collection also changed during this time: instead of sending agents to the provinces, the government tried a bottom-up approach, requiring all subjects to register with local district offices and receive identity cards, formally embedding enumeration at the center of imperial governance.25 One of the clearest expressions of the Ottoman state’s statistical ambition appears in the work of Abdurrahman Şeref Efendi, a bureaucrat and historian who authored Statistics and Geography of Prosperity (İstatistik ve Coğrafya-yı Umraniye) in 1896. Like many other intellectuals of his time, he had considerable faith in the possibility of rational government based on quantitative data and saw this as a major step toward civilization, in which the Ottomans had an important role to play by elevating their long-standing recordkeeping practices to a scientific basis.26 He noted that statistics was no longer confined to the numerical representation of land and people, but rather had become a tool for comparing, analyzing causation, and generating norms, which could be employed in reforming governance. The act of enumeration promised not just administrative control, but insight into the well-being of the governed. For Abdurrahman Şeref, “knowing the level of well-being of the population” was “the highest monument of orderliness for a government.”27 Yet even in his own writings, the gap between statistical ideal and social reality was evident. He lamented that heads of households often refused to report the number of women in their homes, that village officials adjusted population counts for tactical reasons, and that individuals concealed the truth out of fear of taxation or conscription.28 Elsewhere, population figures were padded to attract schools, hospitals, police stations, and other such public works. In short, legibility depended on cooperation, and cooperation was conditional. At the turn of the twentieth century, these frictions condensed around the status of the categories themselves. Due to growing nationalist movements and competing imperial influences, census tables in the Balkan provinces had become arenas of political performance. In 1903, a group of Bulgarian nationalists circulated a letter to the village of Negorçe, instructing residents to declare themselves “Bulgar” rather than “Rum” to census officials. The appeal was not to local identity or communal memory but to the political future: “Tell them, ‘We are not Rum. We are Bulgars,’ because you speak Bulgarian. And don’t be afraid of anyone. Don’t have yourselves registered Rum, because then no good will come to your children, your goods, and your animals.”29 Future-oriented and instrumental, Bulgarian nationalists’ threat shows that categories such as “Rum” or “Bulgar” were not inert descriptors. Census data could—and did—serve as evidence for territorial claims, arguments for autonomy, or grounds for intervention. For nationalist organizers, villagers within a certain territory appearing in the correct category was a form of futureproofing: a consistent statistical identity served as an avenue for political recognition. In other contexts, people chose to actively disidentify with national categories. In 1906, the mayor of Çaricani refused to sign the census registers after learning that, unlike previous counts, the 1906 forms would record the “national and racial identities” of local residents rather than simply listing their religious affiliations.31 His objection was based on his discomfort with the shift in the classificatory principle away from a theological register of belonging toward a national taxonomy. For him, this was not clarification—as it was for the imperial state—but a violation of communal self-understanding. These episodes demonstrate that Ottoman subjects contested the stability of categories, rather than simply opposing headcounts. Local elites and villagers misreported, reworded, or refused to provide classifications to secure future advantages or defend established ways of identification. What administrators labeled as error or fraud was, in practice, strategic resignification. The census thus became a site where the state’s bid to stabilize identity for comparative governance collided with actors who repurposed or rejected those identities for political ends.32 These struggles foreshadowed Mehmed Cavid’s diagnosis in 1909: the Ottoman Empire had vast statistical ambitions, but its fractured apparatus could not organize numbers or categories in ways that signaled statistical modernity. In the first year after the Young Turk Revolution, he published a book titled Statistics (İhsa’iyât) (1909), in which he argued that the statistical efforts of the last decades were either imitations of the West or data intended to misdirect public opinion.33 While the previous Ottoman administrations had pursued centralization and extensive recordkeeping, they produced a statistical apparatus that was fragmented and lacked dependable integration with the center. In this book, he detailed the existing Ottoman framework—an official statistical administration housed in the Ministry of Public Works, tasked with collecting annual tables from each ministry. Yet even as he described the structure, he admitted that it existed “in name only.”34 There was no coordination among ministries, no binding connection between central and peripheral agencies. The result, he wrote, was a state with vast statistical ambitions and little statistical coherence. There was no coordination among ministries, no binding connection between central and peripheral agencies. The result, he wrote, was a state with vast statistical ambitions and little statistical coherence.35 The scope of what the empire sought to measure—births and deaths, marriage and divorce, migration, forestry, mining, tram companies, orphanages, hospitals, telegraphs, alcohol regulation, and more—reveals that statistics had become crucial to the Ottoman vision of modern governance. But Mehmed Cavid’s account also exposed the fragility of that vision. Numbers were difficult to collect; categories were unstable; ministries worked at cross-purposes. The infrastructure of enumeration—both material and conceptual—remained piecemeal, disjointed, and perpetually incomplete. What the state aspired to as a rational, scientific project appeared instead as a patchwork, at once overextended and incoherent. The CUP transformed Mehmed Cavid’s diagnosis into a mandate to make statistics the engine of social transformation. They embraced a positivist belief in objective, scientific truth, treating statistics as a technology for remaking the social body.36 On an ideological substrate, this was realized on women’s bodies: across different intellectual traditions, forming a family and raising children was declared a woman’s “essential duty.”37 Reformist newspapers prescribed conception and pregnancy regimens while dismissing feminism as a distraction from demographic increase. Maternity was cast as a civic obligation. That cross-ideological consensus preselected the very metrics the state would privilege (births, infant mortality, maternal health) and interpellated women as statistical subjects whose reproductive capacity anchored national arithmetic. Enumeration was no longer just an index of order; it was an instrument of transformation. In practice, this meant using counts, registries, and tabulations to plan resettlement of waves of refugees, set conscription and requisition targets, monitor and reshape the ethno-confessional composition of provinces, and underwrite projects of social reordering.38 The most violent end of the CUP’s statistical project was the Armenian Genocide, which was organized and carried out via information coded in statistical registers.39 The promise of statistical coherence quickly mutated into a program of demographic engineering, exposing the underbelly of numerical governance. The Republic of Turkey (est. 1923) inherited both the conviction and the incoherence of the late Ottoman state and turned populationism into raison d’état. Reformers like Besim Ömer Akalın cast demographic growth as a national imperative, arguing that “the foundation of the state rests on two pillars: land and people.”40 The Interior Ministry echoed that modern rule begins with knowing the population.41 With that end fixed, the government contracted the Belgian statistician Camille Jacquart to design and oversee the 1927 general census. With that end fixed, the government contracted the Belgian statistician Camille Jacquart to design and oversee the 1927 general census.42 Jacquart brought with him the tools and assumptions of European statistical modernity: uniform categories, comparable schedules, the promise of a fully visible people. In Foucauldian terms, the 1927 operation was a bid to perfect the biopolitical gaze: to make life measurable, accountable, and therefore governable. Under Jacquart’s coordination, the General Directorate of Statistics issued a detailed instruction manual for census officers, specifying the attributes a governable subject must possess, including name, gender, birthdate, marital status, profession, native language, literacy, religion, and disability. The manual directed agents to elicit answers and, particularly for literacy, double-check the accuracy of the responses by putting citizens to the test with the census form itself.43 As Jacquart’s reports and interviews make clear, the 1927 census was an experience of the clash between bureaucratic procedure and local moral economies. There was infrastructural underdevelopment: houses lacked numbers, streets lacked names, many citizens—especially in rural districts—did not use surnames or did not report them, there was no centralized registry of births and deaths, and he estimated near-universal illiteracy outside towns and cities.44 Some 50,000 census agents were trained and deployed to conduct census operations across the country, during which they encountered communities and households that neither recognized the purpose of the questions nor accepted that registration was a civic good. Families customarily hid their children to avoid taxation or conscription, delayed registering daughters, or misrepresented their birthdates to manipulate marriageable age. In an interview with the Belgian newspaper Vigntième Siècle, Jacquart related a villager’s rule: “so long as a child doesn’t walk by himself and talk… he is not a man, and there is no use in recording him.”45 What appeared to Jacquart as backwardness or irrationality can also be read as an epistemological misalignment. In Rancière’s terms, these quiet tactics—non-appearance, misreporting, categorical drift—are forms of dissensus that break interpellation and interrupt the passage from rule to recognition. While the villager articulated a “subordinate regime of truth,” he was ultimately relegated to the position of a challenge to be overcome on the path to building Turkey as a statistical state. In fact, the 1927 census faced many problems during the count and in its aftermath, collectively showing that the collection of statistical data required cooperation and faith in enumeration as a public good. While state leaders performed coherence and aptitude for the population—Prime Minister İsmet İnönü praised the census as proof that the new administration functioned “like clockwork”—the cracks began to show within a few days.46 An editorial published in the National Newspaper (Milli Gazete) lamented the fact that in “some corners of the country,” families had intentionally concealed their household size.47 A report organized following the 1931 election of local representatives explains that, in various districts, the omission of itinerant laborers (many of them working in railroad construction), tribes, and refugees, as well as errors by census officers, led to a false population count.48 Beyond the usual neglect by the populace to register births and deaths, some district governors simply refused to follow the regulations. In Bursa, for instance, the officer in charge was disciplined for falsifying census documents.49 Material contingencies also undercut the archive: in Bozüyük, the registry room burned, forcing a full redo in 1928, which yielded different results.50 In short, counting demanded more than infrastructure; it demanded participation and shared beliefs about what counting meant. If refusals make visible the limits of statistical reason, Jacquart’s reflections two years after the initial census show how those limits were narrated away. In a 1929 essay in Life (Hayat) Magazine addressing concerns about Turkey’s statistical underdevelopment, Jacquart clearly stated that the ideological horizon of statistical modernity was full legibility.51 He narrated the present shortcomings in Turkey as a temporary stage on the path to transparent knowledge of the population. He did not treat refusals, omissions, and contradictory returns as evidence of the project’s limits. Instead, he re-narrated them as “not yet,” a transitional condition of national development. This move is central to what I have argued throughout: statistics, as an element of positivist government, did not simply fail when populations refused to appear as intended. It reasserted itself by insisting that enumeration is inevitable and that incoherence can only ever be temporary. This can be read, following Hall, as the ideological effect of statistics: subordinated truths—silence, evasions, and categorical re-significations visible in the archive—are not recognized as truths.52 Rather, they are absorbed into a narrative of eventual legibility. If refusal makes the limits of statistical reason visible, Jacquart’s narrative forecloses that recognition, insisting instead that counting will always catch up. Taken together, the Ottoman and Turkish cases reveal not a weak state failing to impose order, but the structural tension of statistical reason itself. Every attempt at legibility—from nineteenth-century censuses to CUP demographic engineering to the 1927 general count—confronted refusals, evasions, and categorical manipulations that unsettled the coherence of statistics promised to deliver. These, I am suggesting, were not aberrations: They were the very conditions of enumeration. What administrators labeled as errors, absences, or delays were, in fact, traces of dissensus, the points at which subjects declined to appear in the forms demanded by the state. And when these refusals surfaced, as Jacquart’s narrative showed, they were reabsorbed as small setbacks on the path to full enumeration. Taken together, these cases illustrate that statistics do not simply describe social reality but rather constitute it through continual struggle, and that they become political precisely where their universality is fractured and must be reasserted. 3. If the Ottoman and Turkish cases above demonstrate anything, it is that the statistical trace is not innocent of the struggles that produced it. But a number, once recorded, does not dictate the use to which it is put. The same figures a state compiled to govern could be turned by its critics into evidence against it. This is why statistics has proven ideologically flexible: those who opposed the very order that statistics helped govern nonetheless found its numbers indispensable, turning the state’s own instruments against it. Karl Marx is the clearest case. While remaining critical of statistical governance—he dismissed Arthur Young, author of Political Arithmetic (1774), as an “unspeakable statistical prattler”—he nonetheless studied the reports compiled by factory inspectors closely, treating their returns on the working day and on child labor as evidence of the material conditions of surplus and absolute surplus production.53 V. I. Lenin, in turn, leaned on statistical tables—“the summarized results of irrefutable bourgeois statistics”—in his work on imperialism, using them to chart the concentration of capital and the partition of the world market.54 The same tables that states assembled for the purpose of governing became, in their hands, the instruments for anatomizing and indicting imperialism and capitalism. If statistics cannot be cornered by any single ideology, this is not because it stands above politics but because each use of statistical figures reopens the struggle sedimented in their creation. This has a direct consequence for the social scientist. To draw on statistical data is not to retrieve a neutral fact but to take up an artifact whose production was itself contested; and the manner of that taking up—whether the number is read as a transparent window or as the trace of a conflict—is itself a political choice. As a social historian, I am acutely aware of how statistical data can recover the experiences of those who left few written traces: factory workers, peasants, women, children. Census tables, wage reports, and mortality rates furnish the raw material for histories of ordinary lives, and often they are the only material there is. The temptation, precisely because these sources are so indispensable, is to treat them as windows, thereby letting aggregation stand in for the lives it counts, to accept the state’s wager that numbers can substitute for the people they enumerate, and that statistics yield unmediated access to the truth. What the Ottoman and Turkish cases show is that this wager was contested at the very moment of collection, and that its traces survive in the data inconsistencies, newspaper clippings, and anecdotes read against the grain. Statistical traces are themselves the sediments of conflict: they register not only demographic patterns but refusals, evasions, and manipulations. Households withholding women, daughters’ ages adjusted, local officials falsifying returns, villagers repurposing categories—these are not simply reactions to the enumerative impulse of the state but the assertion of deep-rooted logics of accounting that ran counter to it. They are not just noise in the signal; they are the signal, the record of a negotiation between the counter’s demand for legibility and the subjects’ strategies of concealment or manipulation. For the social scientist, this means that statistical traces are not neutral reflections of reality but crystallizations of political struggle. What appears as inconsistency, silence, or error is the mark of that negotiation. Ultimately, to treat statistics as a political concept is to resist reading numbers and series as transparent facts, and instead to interpret them as sites where power was exercised, contested, and re-narrated. Seen in this light, the Ottoman and Turkish archives are not anomalous but exemplary. They reveal the universal tension within statistical reason: its aspiration to render populations fully legible, its dependence on cooperation, and its tendency to narrate away refusals as irrational or not yet complete. Statistics does not simply fail when subjects refuse to appear as designed. Rather, it reasserts its inevitability. This is what Hall called the ideological effect: the conversion of contradiction into narrative, the foreclosure of alternative truths by naturalizing enumeration as the only horizon of governance.55 Hall’s insight helps us see why statistics feel inescapable: its fractures are continually overwritten as incompletion, as delay, as something to be fixed rather than as a structural impossibility. But this is the crucial point. Statistics does not overwrite its fractures because it has closed them; it overwrites them precisely because it cannot close them. The performance of totality is necessary because the totality is never achieved. Legibility must endlessly be reasserted because it is endlessly interrupted, and the very insistence that enumeration will one day be complete is the surest sign that it never is. We now inhabit what might be called a condition of statistical universality, in which there is seemingly no outside to enumeration. From predictive policing to credit scores, algorithmic feeds to biometric screening, subjects are rendered and re-rendered as data points, and every inconsistency is folded back into the system as an error to be corrected or a gap to be filled. Here the apparent paradox—that statistics is at once inescapable and never complete—comes into focus. These are not two opposed claims, but one structure seen from two sides. Statistics is inescapable as a horizon: there is no position outside enumeration from which one might simply opt out, no ground untouched by the demand to be counted. Yet it is never complete as an achievement: the totality it performs is a claim it makes about itself, not a state it has reached. The Ottoman villager who refused to register his children and the contemporary user who evades tracking share a structural position. Their refusals do not place them outside the system; they mark the point, within it, where its claim to totality breaks against what it cannot absorb. In this sense, refusal is not the exception to statistical universality but the permanent condition that universality must continually work to deny. If this is the condition we inhabit, then the obligation it imposes falls not only on those who govern by numbers but on those who study by them. The social sciences have grown extraordinarily sophisticated in the use of statistical data—in reconciling figures, correcting for their gaps, and making them yield arguments—while remaining relatively incurious about how those figures came to exist at all: who compiled them and under what compulsion, who was counted and who slipped the count, which refusals and evasions were smoothed into the totals now treated as evidence. To treat statistics as a political concept is to insist that these two questions cannot be separated. A number is not a transparent fact but the trace of a struggle over who should be counted, how, and to what end; to use it well is to read that struggle in it. Published on August 4, 2026 * Deren Ertas * 1. This paper is the result of many years of work and the input of many valuable individuals. I wrote a version of this paper as a graduate student in Beshara Doumani’s population politics seminar at Brown University in Fall 2020. I wish to thank the organizers of the Political Concepts Graduate Student Conference at the New School for Social Research in March 2021 for inviting me to participate in this special edition. Thanks to Tsiona Lida and Salman Mirza for inviting me to workshop this paper at the Center for History and Economics at Harvard University (2021), as well as to everyone who attended. I am grateful to Samet Budak, Jochen Schmon, and Clover Reshad, as well as the two anonymous reviewers, for their thoughtful edits on this piece. Special thanks to Can Eskinazi. All errors, I am afraid to say, are my own.↩ 2. In English, this title translates to “Statistical Microscope: An Analysis of the State, in which the State of the Germanic Roman Empire is vividly represented, above all extraordinary.” Wallace Ferguson, “Microscopium Statisticum and the Etymology of Statistics,” Significance 19:1 (2022): 6; Ian Hacking, The Taming of Chance (Cambridge: Cambridge University Press, 1990), 23–24.↩ 3. Karl Pearson, The History of Statistics in the 17th and 18th Centuries, ed. E.S. Pearson (New York: Macmillan, 1978), 2–3; Walter F. Willcox et al., “Statistics,” in Encyclopedia of the Social Sciences, ed. Edwin R. A. Seligman and Alvin Saunders Johnson (1937), 356.↩ 4. Michel Foucault, Security, Territory, Population: Lectures at the Collège de France, 1977–1978, trans. Graham Burchell, ed. Michel Senellart (London: Palgrave Macmillan, 2007).↩ 5. Michel Foucault, Society Must Be Defended: Lectures at the Collège de France, 1975–76, trans. David Macey, ed. Mauro Bertani and Alessandro Fontana (London: Picador, 2003), 250.↩ 6. Michel Foucault, Society Must Be Defended: Lectures at the Collège de France, 1975-76, ed. Mauro Bertani and Alessandro Fontana, trans. David Macey (Picador, 2003), 247.↩ 7. Willcox et al., under “Statistics,” 358.↩ 8. Foucault, Society Must Be Defended, 246.↩ 9. Foucault, Security, Territory, Population, 104.↩ 10. Louis Althusser, “Ideology and Ideological State Apparatuses (Notes Towards an Investigation),” in Lenin and Philosophy and Other Essays, trans. Ben Brewster (New York: NYU Press, 2001), 85–126.↩ 11. Jacques Rancière, “Ten Theses on Politics,” in Dissensus: On Politics and Aesthetics (London: Continuum, 2010), 35. He writes: “Political conflict does not involve an opposition between groups with different interests. It forms an opposition between logics that count the parties and parts of the community in different ways.”↩ 12. Stuart Hall and Lawrence Grossberg, “On Postmodernism and Articulation: An Interview with Stuart Hall,” Journal of Communication Inquiry 10, no. 2 (1986): 48–49.↩ 13. Stuart Hall and Lawrence Grossberg, “On Postmodernism and Articulation: An Interview with Stuart Hall,” 49.↩ 14. Stuart Hall and Lawrence Grossberg, “On Postmodernism and Articulation: An Interview with Stuart Hall,” 48. He writes, “If Foucault is to prevent the regime of truth from collapsing into a synonym for the dominant ideology, he has to recognize that there are different regimes of truth in the social formation.”↩ 15. Stuart Hall and Lawrence Grossberg, “On Postmodernism and Articulation: An Interview with Stuart Hall,” 49.↩ 16. Adi Ophir, “Concept,” Political Concepts: A Critical Lexicon, 2012, available at: https://www.politicalconcepts.org/concept-adi-ophir/.↩ 17. Ian Hacking, “How Should We Do the History of Statistics?,” in The Foucault Effect: Studies in Governmentality, ed. Graham Burchell et al. (Chicago: The University of Chicago Press, 1991), 184. In this piece, Hacking writes, “I believe that the organization of our concepts, and the philosophical difficulties that arise from them, sometimes have to do with their historical origins.” Yet, he says, “exposing the historical ground of a problem” does not make it go away. His method is “a use of the past for understanding some of the incoherence in present ideas” by “producing an hypothesis about the relationship between concepts in their historical sites” (ibid.).↩ 18. Ranajit Guha, “The Prose of Counter-Insurgency,” in Subaltern Studies II: Writings on South Asian History and Society, ed. Ranajit Guha (Oxford: Oxford University Press, 1983).↩ 19. Zafer Toprak, “Osmanlı Devleti’nde Sayısallaşma Ya Da Çağdaş İstatistiğin Doğuşu,” in Osmanlı Devletinde Bilgi ve İstatistik (Başbakanlık Devlet İstatistik Enstitüsü Matbaası, 2000).↩ 20. Standford J. Shaw, “The Ottoman Census System and Population, 1831–1914,” International Journal of Middle East Studies 9:3 (1978): 325.↩ 21. Beshara Doumani, “The Political Economy of Population Counts in Ottoman Palestine: Nablus, circa 1850,” International Journal of Middle East Studies 26:1 (1994): 1–17; Nilay Özok-Gündoğan, “Counting the Population and the Wealth in an ‘Unruly’ Land: Census Making as a Social Process in Ottoman Kurdistan, 1830–50,” Journal of Social History 53, no. 3 (2020): 763–91.↩ 22. Toprak, “Osmanlı Devleti’nde Sayısallaşma Ya Da Çağdaş İstatistiğin Doğuşu,” 97.↩ 23. Musa Şaşmaz, “The Ottoman Censuses and the Registration Systems in the Nineteenth and Early Twentieth Centuries,” Ankara Üniversitesi Osmanlı Tarihi Araştırma ve Uygulama Merkezi Dergisi 6 (1995): 292.↩ 24. Report of Şuray-ı Devlet, Tanzimat Dairesi, No. 438, 21 Cemaziyulevvel 1248 (21 April 1881), BA, Irade, Şuray-ı Devlet, 3148, cited in Kemal H. Karpat, “Ottoman Population Records and the Census of 1881/82–1893,” International Journal of Middle East Studies 9 (1978): 237–74.↩ 25. Abdurrahman Şeref, İstatistik ve Coğrafya-yı Umraniye (İstanbul: Karabet Matbaası, 1896), 30–31.↩ 26. Theodore M. Porter, Trust in Numbers: The Pursuit of Objectivity in Science and Public Life (Princeton, NJ: Princeton University Press, 1995).↩ 27. Abdurrahman Şeref, İstatistik ve Coğrafya-yı Umraniye, 36.↩ 28. Abdurrahman Şeref, İstatistik ve Coğrafya-Yı Umraniye, 23–24.↩ 29. İpek Yosmaoğlu, “Counting Bodies, Shaping Souls: The 1903 Census and National Identity in Ottoman Macedonia,” International Journal of Middle East Studies 38 (2006): 56.↩ 30. İpek Yosmaoğlu, Blood Ties: Religion, Violence, and the Politics of Nationhood in Ottoman Macedonia, 1878-1908 (Ithaca: Cornell University Press, 2014), 148–49.↩ 31. BA, TFR.I.SFR, 15 Zilkaade 1323 (11 January 1905). This folder contains a series of communiques concerning the mayor of Çaricani’s imprisonment due to his refusal to sign the census registers.↩ 32. Fatma Müge Göçek and M. Şükrü Hanioğlu, “Western Knowledge, Imperial Control, and the Use of Statistics in the Ottoman Empire,” in Cultural Horizons: A Festschrift in Honor of Talat S. Halman, ed. Jayne L. Warner (Syracuse / Istanbul: Syracuse University Press / Yapı Kredi Yayınları, 2001), 115.↩ 33. Mehmed Cavid, cited in Toprak, “Osmanlı Devleti’nde Sayısallaşma Ya Da Çağdaş İstatistiğin Doğuşu,” 99.↩ 34. Mehmed Cavid, İhsaʾiyât (İstanbul: Maṭbaʿa-ı ʿĀmire, 1909), 87.↩ 35. For more on this, see M. Şükrü Hanioğlu, The Young Turks in Opposition (Oxford: Oxford University Press, 2009).↩ 36. Erik Jan Zürcher, Turkey: A Modern History, 4th ed. (London: I.B. Tauris, 2017), 131.↩ 37. Aynızade Hasan Tahsīn, “Nüfūs Mesʾelesi,” Beyānü’l-Hakk, 1326/1910–11.↩ 38. Erik-Jan Zürcher, “The Late Ottoman Empire as Laboratory of Demographic Engineering,” Il mestiere di storico: Rivista della società italiana per lo studio della storia contemporanea 1:1 (2009), 1–12; Fuat Dündar, Modern Türkiye’nin Şifresi: Ittihat ve Terakki’nin Etnisite Mühendisliği: (1913–1918) (İletişim Yayıncılık, 2015); Vladimir Hamed-Troyansky, Empire of Refugees: North Caucasian Muslims and the Late Ottoman State (Stanford: Stanford University Press, 2024).↩ 39. Fuat Dündar, Crime of Numbers: The Role of Statistics in the Armenian Question (1878–1918) (London: Routledge, 2018).↩ 40. Ömer Besim Akalın, Nüfus Meselesi ve Küçük Çocuklarda Vefeyat (Istanbul: Yeni Matbaa, 1923), 7.↩ 41. Abdülmuttalib, Medenî Devletlerde ve Türkiye’de Tahrir-i Nüfus İcrāsı (Ankara: Vilayet Matbaası, 1925).↩ 42. TCA, Kararlar Daire Başkanlıǧı (1920–1928), 14/40/16; decision regarding Camille Jacquart’s invitation to Turkey, June 10, 1925.↩ 43. Tahrir Memurlarına Mahsûs Mufassal Ta‘lîmâtname (Ankara Matbaası, 1927), 5.↩ 44. “İstatistiki Terakki Ettirmek İçün,” Hayat Dergisi 2:28 (July 9, 1927): 34–35.↩ 45. Turkish Republican Archives (TCA), “Muamelat Genel Müdürlüǧü,” 25/145/22, August 27, 1927; letter from the Foreign Office to the Prime Minister’s Office with Camille Jacquart’s article published in French newspaper, XX Siecle.”↩ 46. See “5 Teşrinisani 1927 Cumartesi, Güvenoyu İsteği Nedeniyle Yaptığı Konuşma,” in Başbakanlarımız ve Genel Kurul Konuşmaları, Cilt 2 (Cumhuriyet Hükümetleri Dönemi), ed. İrfan Neziroǧlu and Tuncer Yılmaz (Ankara: TBMM Basımevi, 2014), 96–97.↩ 47. “Nüfus Meselesi,” Milli Gazete 8 (1927): 2.↩ 48. TCA, “Muamelat Genel Müdürlüğü,” 24/135/15, May 19, 1931; report on the difference between male populations in the 1931 election of representatives and the 1927 census.↩ 49. TCA, “Muamelat Genel Müdürlüğü,” 32/184/7, February 12, 1927; petition to reduce the punishment of Mehmed Efendi, the Chairman of the Bursa Census Office, for falsifying documents.”↩ 50. TCA, “Muamelat Genel Müdürlüğü,” 26/146/6, May 1, 1928; report on the Renewed Census from the Bozyüzük District to the Prime Minister’s Office.↩ 51. Camille Jacquart, “Türkiye’de İstatistik Teşkilatı: İktisad Doktoru Vedat Nedim Bey Efendi’nin Suallerine Cevap,” Hayat Mecmuası 5:114, 195.↩ 52. Stuart Hall and David Morley, “Culture, the Media, and the ‘Ideological Effect,’” in Essential Essays, by Stuart Hall and David Morley, Stuart Hall, Selected Writings (Durham: Duke University Press, 2018), 323.↩ 53. Hacking, “How Should We Do the History of Statistics?,” 184; Karl Marx, Capital: A Critique of Political Economy (Vol. 1), trans. Ben Fowkes (New York: Penguin, 1981), 399–400; 386, ff. 89.↩ 54. V. I. Lenin, Imperialism: The Highest Stage of Capitalism (Bombay: People’s Publishing House, 1944), 36–45. My thanks to Sesil Artuç for pointing me to this reference.↩ 55. Stuart Hall, “Culture, the Media, and the ‘Ideological Effect,’” in Essential Essays, vol. 1: Foundations of Cultural Studies, ed. David Morley (Durham: Duke University Press, 2018), 323.↩ |
