Evidence of undercounting: Collecting data on mental illness in Germany (c. 1825-1925)

Argument

Collecting data about people with mental disorders living outside of asylums became a heightened concern from the early nineteenth century onwards. In Germany, so-called “insanity counts” targeted the number and sometimes the type the mentally ill who were living unattended and untreated by professional care throughout the country. An eagerly expressed assumption that the “true” extent of the gathered numbers must be much higher than the surveys could reveal came hand in glove with the emerging task of “managing” insanity and its potential dangers in a modern society. The doorstep of the family home became a crucial site in psychiatrists’ and enumerators’ efforts to register the most sensitive of personal data. This article traces the ever more diligent methods that were employed to obtain the desired information, as well as the hidden agenda of the postulate of missing data itself. It also addresses the profound impact that the presumption of having only incomplete data has had on the practice of counting and surveying, as well as on the understanding of the need for professional monitoring of mental illness.

1. The struggle for data

A physician in Silesia is asked to examine the mental state of a baker’s wife in a place two miles distant. The following conversation unfolds: “How long has your wife been mentally unsound?” – “For five years.” – “Why have you not already had her added to the List of the Insane?” – “The local authority did not wish to inflict that upon me.” – “Is anyone else here insane?” – “There is another in this place and two of her siblings’ children are the same, they have not been listed either.” The List of the Insane had been submitted to the local court only shortly beforehand. Hence in one small community four sick residents had not been registered! (Laehr 1865, v) ¹

Heinrich Laehr, one of Prussia’s most prominent and influential psychiatrists, was complaining about how difficult it was to record the non-institutionalized people living with mental illness. The task of this Silesian physician was to ensure that the statistics were properly adjusted to account for people with mental disorders who were living with their relatives rather than in institutions. The dialogue quoted by Laehr vividly conveys the personal effort required of the doctors charged with the enumeration. It sheds light, whether intended or not, on the many hurdles to be overcome when gathering information about the mental condition of every single resident of a village or population and reveals a panoply of practices set in motion to find the “true” number of the mentally ill throughout a country.

Scouting, registering, counting, snooping, denouncing, maintaining silence, keeping secrets, calling the thing by another name: identifying the mentally ill who were cared for at home encountered many forms of witting or unwitting obfuscation and called for various ruses to access information known only within the family or neighborhood. Framing a mental “otherness” outside the asylums meant exploring uncharted territory and pinning down elusive entities.

The prevalence of mental illness within a population and the underlying issue of classification has long been—and still remains—a matter of complex surveys conducted in service of health policy. Sporadic “insanity counts” (Irrenzählungen), targeting the number and sometimes the type of the disease, were already being performed in the late eighteenth century. Heightened concern gave rise to a great abundance of specialized tallies of people with mental disorders. ² In the nineteenth century, the need to count, enclose, treat and, in brief, to govern insanity was felt not only by the emerging profession of psychiatrists, but increasingly by government officials and local administrators. These activities yielded some of the most sensitive information obtainable in the early history of personal data, as paths crossed between expanding psychiatric care, emerging public responsibility and a growing awareness of mental illness as something that must be dealt with in a modern society.

Going out into the field to track down first-hand knowledge everywhere beyond the walls of the asylums was part and parcel of the “avalanche of printed numbers” that has been produced from the mid-1820s onwards (Hacking 1982), but it has received little attention so far. The impetus to count and to intervene were two sides to the same coin. Hacking has already highlighted the widespread belief that “one can improve—control—a deviant subpopulation by enumeration and classification” (Hacking 1990, 3) and he has thus pointed toward the biopolitical dimension of increased medical statistics during the nineteenth century. All these effects, which hold true for medical statistics, could be said to appear in a concentrated form in data collections about mental illness. Yet, the manifold “insanity counts” have until now received little scholarly attention (Ritter 2003; Brückweh 2015).

Other provinces of serialized data collection on mental health conditions, especially asylums, have of course been investigated with great success. One important strand in the quantification of insanity, as Bernd Gausemeier and Theodore Porter have shown, was the search for the causes of disease and the putative role played by heredity. Statistical ambitions emerged in an institutional setting that provided ideal conditions for the observation, interrogation and classification of large numbers of patients (Gausemeier 2015). The asylums, as the main sites of this data-driven science avant la lettre, scrutinized patients and their families for hereditary traits and characteristics. From this perspective, Porter also sheds light on the large amounts of data collected from social surveys and national censuses (Porter 2018).

It is the crossroads of general population counts and medical statistics, however, that is surprising. In addition to the keen alienist interest in collecting data from patients, there is an ever-expanding statistical enterprise that explicitly reached beyond the walls of the institutions—a broad administrative and scientific endeavor that sent emissaries to the very doorstep of individual homes. This not only raises questions about how such personal, intra-familial and sensitive data were elicited, recorded and used, but also invites us to test the precarious validity of information created and gathered at the thresholds of family life.

This paper analyses the methods as well as the perceived limits of obtaining data about the situation outside of professional care. A preliminary review of the surveying cultures conducted by a host of often unnamed statistical experts throughout the German lands yielded rich results. Both the practices of public authorities and psychiatric specialists could be traced in detail, as they set out to record the non-institutionalized mentally ill in Germany. Judged against the history of statistics, it is noteworthy that the search targeted a type of information that was neither easily accessible nor able to be clearly assigned to distinct categories. The fear of incomplete statistical data encouraged techniques for penetrating more deeply into this forbidding terrain. My analysis will address, on the one hand, the many strategic adaptations undertaken in response to these concerns. On the other hand, and closely tied to the push for better numbers, I shall devote specific consideration to the peculiar assumption by experts that they were only identifying a fraction of their target population.

Medical statisticians of the day held that the “true” extent of mental illness defied capture, and this resulted in specific consequences for survey practices. From the mid-1820s until the mid-1920s, as the next four sections will illustrate, a broad spectrum of methods was adopted for carrying out reconnaissance to the furthest corner of every region. The first section covers the first fifty years period, in which the instruments in the service of the alienists’ surveys initially comprised one-off lists, regularly updated registers and sophisticated questionnaires to identify the mentally ill on the very doorstep of their home. These mostly regional surveys were judged to be a flop, occasioning demands that governments should initiate nationwide efforts, as we shall see from the leading Prussian census reforms of the 1860s and 1870s. The second section moves on to the introduction of counting slips (Zählblättchen) and counting cards (Zählkarten), which revolutionized data processing and raised hopes of valid statistics due to the collection of specific personal data in family homes. Hitherto hardly noticed by research, however, was the strategy to recount and “complete” the results achieved by enumeration: The third section describes how archives stuffed with “raw” census material that provided the personal data of almost every citizen gave medical experts an ideal pathway to pursue tightly focused follow-up inquiries. ³ The fourth and final section focuses on the early twentieth century, when the object of psychiatric care was recharted and its investigation intensified, a development that, at the same time, promoted the centralization of personal data. On the one hand, asylum admission policies changed. On the other, the expansion of extramural care paved the way to addressing forms of mental illness and mental deviance not previously in the psychiatrists’ focus. ⁴ The ambition grew to collect and connect ever more data from wide-ranging origins in centralized indices, which in the long run held out the promise of predicting the behavior of the mentally ill.

The search for so-called “lunatics,” “idiots,” “cretins” and “epileptics” produced, as indicated above, more than mere numbers. The belief that not every case of mental illness could be precisely tracked to every house and every corner of the country, I argue, took on a role of its own in the production of evidence. Haunted by the ideal of statistical certainty, the conductors of the counts and surveys proclaimed almost unanimously that the “actual” prevalence of insanity was far greater. Data combined with what I will tentatively call “non-data” based on the experts’ estimates and assumptions played a crucial role in calculating how many mentally unsound people remained untreated and unattended. As we might expect, a desire to capture “all” the mentally ill necessarily produces knowledge gaps. Undoubtedly, however, proclamations of “non-data,” as an expression of incomplete data sets, triggered the development of techniques for gathering data that could stand up as robust.

The article explores the peculiar role of claims, voiced by historical statistical experts, that there must be an unknown and presumably unfathomable totality of insane people living throughout the country. Allegations that family or even community matters were escaping the all-seeing eye of statistical evidence with some regularity were not just repeated over and over again: strategies were devised to counter the problem. The enumerators’ own estimate of their undercount, as Porter observed in another context, was anything but politically neutral (Porter 1995, 33). The presumed “real” ratio of the insane in the entire population was pursued at the margins of medical, socio-political and administrative responsibility. But a surprising outcome of this first comparative overview of the counting practices of mental illness in German lands is that, rather than treating incomplete data as an inevitable by-product of statistical inquiries in fairly impermeable territories, the lines of access themselves were improved upon, to cope with the systematic bias. The doorstep thus became an important site for the awareness that the knowledge produced by psychiatric surveying was limited, as experts embarked on a quest for more effective data gathering techniques. ⁵

2. “Preventative measures”: Lists, tables, questionnaires (c. 1825–1875)

The effort to tally and locate those mentally ill living outside the asylums arose in the context of a broad reform movement in psychiatry. A series of state mental institutions, founded from the early nineteenth century onwards, were designed as Heilanstalten, with curative as opposed to custodial ambitions (Engstrom 2003, 17–18). One of the first major surveys of mentally ill persons in the German states was conducted in 1824 in the Rhine Province of Prussia. A year before the inauguration of the asylum at Siegburg—celebrated as the most progressive of its time—an effort was launched to identify “curable” individuals who would make suitable patients for the facility under construction, and at the same time to collect general information about the mentally ill living in the region. Local officials were required, with the assistance of doctors, to ascertain the nature of each disease and other personal details. Unfortunately, we do not know exactly how the counts were conducted and whether questionnaires were used. What we do know is that difficulties were anticipated even before the survey began, and the people entrusted with it were urged to take particular care. Nevertheless, Maximilian Jacobi, the director of the asylum and the person in charge of the survey, expressed disappointment with its findings. The information provided about the illnesses was too imprecise for statistical use, and he did not find the number of persons listed very plausible. The inadequacy of the results led Jacobi to conclude that families, and sometimes local authorities too, were unacceptably secretive about the existence of mentally ill residents, relatives, and friends.

The claim that people were concealing information about their mental condition and the assumption that families were hiding facts out of a sense of shame or precaution became a commonplace, unquestioned for the entire period studied here. As with so many of these surveys, the data obtained told one story, but the reality appeared to be another. The director of the Siegburg asylum was sure that only about half, and in some areas only a third, of the “lunatics really existing in the province” had been traced (Jacobi 1830, 168–170). He based his assumption about this undercount on the much higher numbers recorded in a British survey performed at the same time. He, like many other heads of such surveys, preferred, as Porter shows, to rely on a seemingly excellent census from somewhere else rather than what was presumed to be a flawed one at home. If lunacy rates were relatively uniform among European populations, they reasoned, it followed that the highest ratios of insane persons found in a population were generally the most accurate (Porter 2012, 591).

The Prussian province of Silesia hoped for better figures from a survey carried out on a single day in 1830. The state-run asylum that had opened a few months earlier in Leubus was already caring for its first patients, but to form a picture of “all the lunatics in Silesia,” Moritz Martini, the institution’s director, designed a questionnaire with thirteen questions. However, just as they had been in the Rhine province, the findings were judged to be hardly serviceable (Martini 1832, 311–12). ⁶ Martini stoically noted that he had no evidence to refute the figures. Nevertheless, he argued, it was striking that some areas had recorded no mentally ill residents at all and that among the higher social strata it was extremely rare to reveal any information of this kind. The “purpose of the tally”—a complaint also expressed elsewhere—had been grasped neither by the local authorities nor by the families concerned (Martini 1832, 317). The instigation of these extra-institutional accountings reflects a psychiatric as well as bureaucratic desire for systematic scrutiny of little-known territories to prepare the ground for preventive and curative measures. Statistical data, of course, are in no way the same as given facts or the real state of affairs. It is all the more urgent, then, to shed some light on the epistemic role of those “non-data” in the face of systematic bias.

In investigative practice, any resistance to disclosing the requested information was duly noted down, as it posed a direct obstacle to the need for medical intervention. Gaining precise figures and containing mental illness within the rapidly proliferating asylum system was intended to lay the foundations for future treatment options. Psychiatry’s claim to heal was crucially based on the principle of interning the patient as soon as possible. This conviction—or, as Eric Engstrom puts it, dogma—was one of the most resilient and commonly heard refrains in nineteenth-century psychiatric discourse. It was founded on the assumption that the sooner people could be diagnosed, removed from the detrimental environment of the family or community, and delivered up to the therapeutic offices of the alienist, the better their chances of being cured and of making a full recovery (Engstrom 2003, 32–33). ⁷ The directors of the asylums dedicated to taking in “curable” patients were keen to gain the trust of families and urged their enumerators to collect accurate data.

What had failed at the regional level, they assumed, might be resolved by Prussia as a whole. It was Martini who, in 1832, called for a state-wide count of the mentally ill. Here again, the data were intended to serve both administration and science (Martini 1832, 405; see also Jacobi 1830, 172; Damerow 1846, 18). A preventive procedure aimed at reducing the number of “lunatics” could not be confined to the asylum itself. Counting them would lay the groundwork for demonstrating the importance of public responsibility. Every province must—so the alienists’ appealed to the authorities—know exactly how many mentally ill persons were living in it. Martini advocated handing over the task of counting to the clergy, especially in rural areas. Their deployment in Scotland in 1815 and in Norway in 1825 had shown that the data obtained were incomparably more precise. Priests, he maintained, predicting resistance among those surveyed, were better able than anyone else to “intrude into family life” and hence ideally suited for the acquisition of sensitive information (Martini 1832, 406; for the Norwegian census: Porter 2018, 82-85).

In the Prussian province of Westphalia, greater faith was placed in medical competence. In 1829, before the extensions to the Marsberg asylum went ahead, director Wilhelm Ruer was entrusted by the authorities with a double brief. First, he was to complete the lists of names drawn up in the past, but regarded as deficient, and to examine the persons concerned. To this end Ruer spent several years travelling all over the province. By observing first-hand in the field and by tapping into local knowledge, he published a statistical study in 1837 that was soon widely acknowledged (Irrenstatistik 1838). Second, Ruer was required in the course of his travels to train local doctors inexperienced in psychiatry. They should be able to distinguish between “curable” and “incurable” mental disorders and to recognize in good time how dangerous a condition was. Those who carried out the insanity counts, it was frequently lamented, had to confront indifference, lack of insight, ignorance and unwillingness among local authorities and families. This made it all the more expedient to train the physicians responsible locally for the referral of the mentally ill and also to convince them of the usefulness of early internment (Ruer 1837). ⁸

Both the authorities and the psychiatrists had a vested interest in registering the existence of the insane. ⁹ Instead of conducting time-consuming, costly and mostly unsatisfactory surveys, ways were increasingly sought to bypass those who were directly concerned and look elsewhere for information. In Baden from the early 1840s, every community had to keep tables of the mentally ill. To enhance the accuracy of these lists and identify “every cretin,” the management of the famous asylum in Illenau in Baden was instructed to strengthen its links with local authorities and clergymen (Roller 1847). Silesia was seen as a good example because of its “sweeping controls” (Damerow 1852, 678): the mentally ill were subject to supervision and monitoring by district medical officers, who were tasked with recording the data in ledgers, updating them when necessary and submitting them annually to the authorities. ¹⁰ In 1863 Baden implemented a thorough system to identify all mentally ill persons receiving non-institutional care. Local councils were required to submit tables with the relevant data in triplicate every three years. This meant that the information gathered in this way was available at all times not only locally, but also to the district medical officer and even to the ministry of home affairs (Roller 1874, 124-125).

Baden was the German state with the highest proportion of institutionalized patients. Welfare, provision, and the desire for accurate statistics on every lunatic in the territory worked hand in glove. Christian Roller, for many years director of the Illenau asylum, made a great effort to win over public confidence. Prejudice about psychiatry and the topos of “family shame” was to be tackled with the aid of educational pamphlets, the daily newspapers and leaflets (ibid., 49). His and other alienist propaganda indicates that domestic thresholds had to be overcome first and families convinced of the merits of professional psychiatric treatment.

Collecting this sensitive information was seen as a “cultural task of the utmost importance” (ibid., 130), but in practice it turned out to be trickier than it sounded. Roller was aware that the success or failure of statistical work depended crucially on the person whose job it was to find out the information. Using doctors for the purpose struck him as far too costly. The clergy should not be entrusted with the task, he claimed, because state and church matters must be kept separate. Mayors would be out of their depth with complicated questions. Although Roller, unlike Martini, warned against too much “intruding into family matters” when gathering data (ibid., 123–124), he still held that it should be a priority for the state to obtain knowledge of all its lunatics.

This keeping of lists was fiercely criticized only a few years later by Julius August Koch, director of the state mental hospital at Zwiefalten in Württemberg. To his mind, naming individuals was the key reason why people withheld information. This approach aroused public distrust and generated far too much opposition. The lists to be kept by local officials, doctors, or even pastors would only benefit from greater accuracy, he argued, if relatives or perhaps neighbors had a duty to register lunatics with the authorities (Koch 1878, 40–41). A notification duty for doctors or authorities had been contemplated on repeated occasions (Nasse 1850, 3-9), only to be either discarded as unacceptable (Jacobi 1841, 10) and unfeasible (Wille 1880, 313–314) or else—usually by insisting that these individuals were potentially dangerous—vehemently championed (Gauster 1877, 21; Roller 1879, 515; see also Roemer 1913/14, 590–592). ¹¹ Koch himself had major reservations about granting the state such unlimited rights and strictly rejected imposing a duty to report. He had no such qualms about another legal requirement enacted by Württemberg in 1836, with a renewed enforcement drive in 1872. It obliged the leader of the parish council to declare whether there were any individuals who should possibly be transferred to an asylum (Koch 1878, 40–41). What was out of the question for a public authority appeared to be legitimate or even necessary for medical reasons. Generating data and triggering psychiatric interventions for preventive and custodial purposes had become mutually supportive ventures.

3. From house to house: Censuses as remedies against under-registration

By the middle of the nineteenth century, the public registration of the mentally ill had entered both national and international agendas. Insanity had come up, as Porter points out, at the First and Second International Statistical Congress in Brussels (1853) and Paris (1855). The question of how to best gather data on the number and the situation of the insane, given the tendency of family members to dissemble, remained crucial (Porter 2018, 158). ¹² Linking the counting of the mentally ill to the general census meant limiting the scope to scant information and quantitative data, but it raised high hopes for accomplishing “complete” enumerations.

Ernst Engel, head of the Prussian Statistical Bureau and co-founder of the International Statistical Congress, initiated a ground-breaking reform of the Prussian census immediately after taking office in 1860. Pastoral practices to detect “real” figures were to be discarded in favor of paper tools and their epistemic virtues. Engel envisaged describing the nation on the basis of the broadest possible data. The process of counting was to be transformed into one that both involved and investigated the individual. Every question on the census form must be answered truthfully and independently. This method of “self-counting,” as Engel put it, not only increased the accuracy of the results, but also enlarged the field of inquiry upon which statistical studies could be conducted (Schneider 2013, 224–225; Hansen 2015, 44–45; von Oertzen 2023, this volume). Looking back on the self-counting exercise he had launched in Saxony almost a decade earlier, Engel admitted it would take some “practice,” but that the method of sending cards for self-inscription to every home had proved decisively beneficial (Erster Jahresbericht 1869, 109–116). ¹³

A year after taking up his post in Berlin, Engel complained that the Prussian tables were silent on two important points: “imbecility and lunacy,” which were described in the language of his day two types of illness, depending on whether the disorder was congenital or had not manifested until later in life (Engel 1861, 157). What the head of the Prussian Statistical Bureau considered indispensable information, about which every province needed to produce exact data, elicited the disapproval of the Central Commission for Statistics (Statistische Centralcommission). Citing the very quality that Engel valued most, the principle of self-counting, the question on mental state was rejected as impracticable and was therefore not included in the forms for the 1864 census (Böckh 1863, 100).

Three years later, new paper technologies revolutionized the compilation of the census. Counting slips, first introduced in the Prussian census in 1867, allowed numbers to be abstracted from the enumeration lists. This paper tool was an intermediate, moveable data carrier designed to facilitate and enhance the counting and sorting of individual data (von Oertzen 2017, 136–137; Porter 2018, 176). The Berlin Medical-Psychological Society (medicinisch-psychologische Gesellschaft) lobbied for the inclusion of the question about mental condition into the census. Just a few months before the forms were printed, the recently founded association managed to persuade the ministries responsible for the census of their case. Here, again, there were concerns in advance that people might be unwilling or unable to answer the questions correctly, thereby blurring the accuracy of the results (Protokoll 1868/69). To help enumerators in making a difficult assessment, they were told to confine themselves to those who were “considered mentally ill in their family or community”. ¹⁴ However, what had seemed an obvious enough request produced unsatisfactory results. Carl Westphal, the chairman of the Medical-Psychological Society, concluded—again by comparing the figures with those of the Scandinavian countries—that barely a third of the mentally ill living in Prussia had been covered (Protokoll 1870, 507).

The Prussian census of 1871 used individual counting cards for the first time. They made the complex intermediate step of transferring the data from enumeration lists to counting slips a thing of the past. The novel form was large enough to display an entire set of characteristics requested for census taking. Each counting card encompassed the complete dataset of one individual on a single page (von Oertzen 2017, 141; see also this volume). Every Prussian head of household was asked whether any members of his household were “idiotic” or “insane”. The resultant numbers rose by about ten percent in comparison to those obtained four years earlier. However, this was not interpreted as evidence of an actual increase of the mentally ill, but rather as a validation of the new paper tools (Die Ergebnisse 1875, 130; Gauster 1877, 5).

The practice of census taking remained a grey area. The very design of the counting cards took into account that people shied away from providing details about family members. The form singled out the question about “deficits affecting learning or employability” such as “blind,” “deaf-and-dumb,” “idiotic” or “insane”: unlike all the other questions on the census, instead of the subject answering in their own words, the item had to be ticked or circled (von Oertzen 2019, 112). But apart from this formal requirement, the instructions about how to fill in the counting card had nothing to say regarding this point. Those responsible for the census were well aware that this was a difficult and sensitive issue and argued that this information should be requested from neighbors instead. ¹⁵ To avoid embarrassing situations on doorsteps and strain on relatives, it would be better—according to another pragmatic suggestion—not to ask the question at all. Instead, the enumerators and the commissions entrusted with data controls should make these entries themselves (Fabricius 1868, 195), or else the local doctors deployed to check the counting cards should simply fill in the data required (Kollmann 1876, 169–170).

The year 1871 marks the beginning of a series of censuses, held at first every four years (1867, 1871, 1875) and then every five, until this pattern was interrupted by war and crisis. The states in the new, united Reich had ten core variables about which they were required to ask, and were free to collect additional information at their own discretion (Gehrmann 2012, 15). The question about physical and mental deficits ranked as one of these “add-ons”. It was up to each state government whether or not to include it in their census. ¹⁶ In Prussia the question on mental condition remained extremely controversial. With every census, debate flared up again between the ministries and commissions involved about whether it was ethical or meaningful. Those in favor, including the Prussian Statistical Bureau under Ernst Engel and the Ministry of Ecclesiastical, Educational and Medical Affairs (Ministerium der geistlichen, Unterrichts- und Medizinalangelegenheiten), argued that these data were in the public interest. The Central Commission for Statistics remained skeptical. Some of its representatives took the view that it was not for the state to ask such questions and that valid data would in any case not be forthcoming. ¹⁷ In spite of these grave reservations, the question about mental deficiencies was included again in Prussia in 1880, 1895 and 1905. However, as we are about to observe, the figures obtained in the census were no longer the only focus, for now the counting cards themselves, the “raw material,” opened entirely new doors in the quest for data.

4. Personal data become mobile

Linking the insanity counts to the reformed census did more than enable a nationwide survey of the mentally ill. The information gathered with the counting cards also indicated the citizen’s full name and location. ¹⁸ Connecting all this data created new opportunities. Soon after the 1871 census, the Prussian Statistical Bureau published a parish lexicon in eleven volumes for each of its provinces. ¹⁹ The total number of all blind, deaf-and-dumb and mentally ill was meticulously listed in tables, demonstrating the ability of the state to record all its inhabitants. Formed as data aggregates, they proved to be powerful categories (Bouk 2017). What the census may have missed shifted subtly into the public eye: the figures broken down for each municipality could be viewed by everyone and had themselves become an instrument of control.

According to the regulations, the names recorded in the census could not be used for statistical analysis. Due to the common assumption that families preferred to conceal the mental condition of their members, Ernst Engel gave explicit assurances that people could enter these sensitive data in the forms without any concern that they would be misused (Engel 1861, 188). ²⁰ However, the counting cards for the blind, the deaf-and-dumb and the “insane and idiotic” not only carried a special marking, but were also archived after use. If any of those deficits had been ticked, a red line was to be immediately added in the bottom left corner. The information so urgently sought could thus be spotted at a glance. The marks served local control and, where appropriate, review (Die Ergebnisse 1875, 130). When the counting cards were centrally processed in the Statistical Bureau in Berlin, these markings determined how those cards were to be handled next. While in general all counting cards were destroyed after the compilation was completed, those distinguished by the red stroke were kept and archived by the Statistical Bureau for the Ministry of Ecclesiastical, Educational and Medical Affairs. ²¹

This archive paved the way for entirely new possibilities of data surveillance. The simple system of the red marks laid the basis for the so-called “private” statistical surveys that would later be permitted. The cards preserved beyond the census were available for physicians to borrow for subsequent inquiries. As the names and locations of the mentally ill were recorded on the census forms, they could be visited again later for further research. Information that is subject to data privacy today was then passed on, either on grounds of preventive examination or to comply with public health policy. These private efforts by medical professionals, usually specialists, were justified in terms of their local knowledge. Armed with these counting cards, they were able to conduct incomparably more accurate analyses within restricted areas. First, they could verify and correct the number of mentally ill established by the census. Second, tracing these individuals with the aid of the counting cards enabled the physicians to form a clearer picture of the local situation. ²²

Soon after the 1871 census, the physician Werner Katz drew on the archive of counting cards for his research. ²³ His study of the blind in Düsseldorf, published in 1874, was a resounding endorsement for the architects of the census. The number of blind residents he had managed to identify was only four per cent higher than the figure ascertained by the census. His study was highlighted as perfect proof that the new paper tools had enhanced accuracy (Guttstadt 1874, 228; Katz 1874). Katz did not tire of pointing out that tracing the blind was much easier than tracing the mentally ill. To optimize capture in the former case, he argued, a list of all the names should be published in the regional newspapers straight after the census. If particular individuals were not displayed, Katz confidently reasoned, their families would provide retroactive notifications. In the case of the mentally ill, such public verification was impossible. Publishing their personal data, even Katz recognized, would “encroach too deeply into family sentiment.” ²⁴

According to Albert Guttstadt, one of Prussia’s foremost medical statisticians and employed in the Berlin Statistical Bureau since 1872, the function of the counting cards for the mentally ill was to be a “guide to trace the individuals concerned” (Guttstadt 1874, 248e). Instead of relying on the private commitment of a few doctors, Guttstadt called for a systematization of follow-up investigations. He outlined a meticulous procedure in order to gain unprecedented data. If the census flagged a mental illness, the counting card was to be duplicated on site. According to his proposal, the original was to be dealt with as usual and sent to Berlin for centralized evaluation. The duplicates, however, were to be kept in the respective districts and ought to help local doctors find the individuals in question. The doctors then had to perform an examination and fill out a questionnaire (see figure 1). With this form in their hands, they would be able both to cross the threshold into private homes and to gather additional information. The questionnaire inquired, in great detail, about personal and medical issues, potential threats to self and others, heredity and criminal activity. Security and predictability were to become a key governmental concern. With this two-tier system, keeping track of the mentally ill could remain linked to the census, but also be placed under expanding medical supervision.

Figure 1. Questionnaire for the follow-up investigations. In: Guttstadt 1874, 248e.

Guttstadt published his proposal for using the names and locations of the mentally ill for follow-up investigations in the journal of the Prussian Statistical Bureau. The editors explicitly noted that his call for a two-stage survey throughout the state had not yet been discussed by the bodies responsible (Guttstadt 1874, 248d). The question whether data on mental condition could be passed on to third parties for more detailed investigations remained unsettled for quite a while.²⁵ During the preparatory phase for the 1880 census, the debate about the usefulness of the “add-on” flared up again. But this time reservations about this item on the counting card no longer related to the difficulty of obtaining precise figures on the doorsteps of the families concerned—a method rejected by medical experts for its inaccuracy anyway. All too promising was the “family signposting”²⁶ itself, the names and communities of residence recorded on the cards that facilitated far more probing inquiries further down the line.

The enhanced search for sensitive data was taken up first in Württemberg and then on a national level. In 1875 the above-mentioned Julius August Koch conducted a two-stage survey that was praised as outstanding in both quantitative and qualitative respects (Porter 2018, 303).²⁷ The decisive barrier to gathering valid data, Koch repeatedly argued, were the families who kept the desired information secret. To break down this wall of silence, he managed to persuade priests to take on the more detailed follow-up interviews. They had traditionally been regarded not only as sufficiently educated, but also as best placed to gain the essential trust of the family. There was a conflict, however, between involving the clergy in such surveys and their duties of confidentiality. Koch succeeded in overcoming this dichotomy by anonymizing the data processing once the persons concerned had been found. This principle protected the pastors from breaching their obligations while permitting this reputedly ideal doorway into family life (Koch 1878, 25-36).

The discreet, trust-based interview setting was crowned with great success. Koch used the publication of his widely reviewed study in 1878 to articulate further demands. A more or less “complete” insanity count, surmised Koch, would need to be performed “on the quiet” (ibid., 43), that is, preferably without the public noticing. Again, alleged concealment by families influenced and legitimized his far-reaching proposal: local experts tasked with reviewing the census should create an index and complete it “under the table” (ibid., 42), or in other words without the knowledge of those concerned. The survey to be conducted in the second stage should build on the anonymous data evaluation that had already proven its worth. Furthermore, this essentially public task of generating insanity statistics in Koch’s mind (see also Wille 1880) had to be entrusted to private hands. This would permit incomparably more accurate findings and free these studies from the squalid reputation of state-run surveys (Koch 1878, 42-45). And yet the attempt to escape public attention reveals once more a tenacious desire to obtain resilient figures. The problem persisted: gathering personal data remained a challenge and the results were still regarded as incomplete.

5. Public welfare and forced registration

In the rhetoric of psychiatry, which was still a fledgling discipline, the mentally ill should benefit from well-organized public welfare instead of being locked away in the stable and abandoned to their sad fate. Metaphors of darkness and animal instinct were countered by the language of patient care and medical healing. The number of asylums was declared early on to be a “barometer for the civilized condition of nations” (Laehr 1852, vii). Public responsibility could be quantified, these advocates argued, but this called for a variety of gauges. One, as we have seen, was the proportion of mentally ill in the overall population. In response to highly diverse findings from numerous counts, Germany’s first reference work on medical statistics in 1865 defined a “correct average ratio” of 1:350 or 1:400 (Oesterlen 1865, 517; Gauster 1877, 5). Another and increasingly significant benchmark was the ratio between those who were cared for by relatives and those who were accommodated in asylums and hospitals (e.g. Hasse 1871; Mendel 1874; Laehr 1876). Here too, international comparison was an important factor. Countries that had placed half or two-thirds of their mentally ill in an institution were seen as progressive in the provision of mental care (Protokoll 1877, 47).

The quest for accurate statistics generated new uncertainties. Despite a flurry of new buildings in the second half of the nineteenth century, the psychiatric institutions were always overflowing. It was clear to everyone that the number of the institutionalized was no valid measure of insanity in the entire population. Most experts certainly did not interpret the increase in the number of asylums as reflecting an absolute increase in mental illness. Those involved in the statistical exercise, in particular, saw the numbers they were unearthing as proof that their surveys were founded on more accurate and efficient methods. Moreover, the experts agreed, the public had a better understanding of the treatability of mental illness these days and were increasingly willing to register them (see e.g., Gauster 1877, 5; Koch 1878, 23).²⁸

The enhancement of public welfare for the insane in the decades around 1900 was twofold. First, admissions to asylums were stepped up. Second, provision was being expanded beyond the asylum walls. In the 1890s there was a significant shift in the funding of inpatient care. The Poor Law of 1871 (Unterstützungs-Wohnsitz-Gesetz) made local communities financially responsible for the care of their poor and sick citizens. In the years after German unification, owing chiefly to the impact of rapid industrialization and domestic migration to urban centers, communities were confronted with staggering financial obligations. In 1891, the Poor Law as it applied to Prussia was amended to alleviate the burdens on local communities. Instead, the provinces were obliged to assume the costs of adequate institutional care (Engstrom 2003, 178; see also Brink 2010, 110-115). The legislation that took effect in 1893 extended provincial responsibility for “curable” and “dangerous” patients to those who were “incurable” and “not dangerous”. These new provisions also included epileptics, most of whom had hitherto received little care (Zinn, Pelman 1893).

Of course, the new remit reinforced the need for counting. In the late 1850s, church pastoral services were already complaining that “the Prussian state knows its barns and stables, its sheep and cows, but not its idiot children” (Disselhoff 1857, 155). Provision for “idiots and epileptics” had long since been consigned to church-run asylums. Although some psychiatrists were pressing in the 1870s for this responsibility to change hands, these demands kept disappearing from the agendas of psychiatric associations (Zinn and Pelman 1893, 66–68). With the new legislation, numerous surveys targeted this group, accompanied by the notorious complaint that the findings did not reflect their “real” numbers (Tigges 1884). Epileptics were regarded as potentially dangerous, prompting calls for them to be institutionalized and their families relieved of the burdens of supervision. The scope of psychiatric care, underpinned by statistics, was to expand further in the following decades.

Counting, registering and scouting for information were an integral part of psychiatric provision. When extramural care was launched in the early twentieth century, its advocates pursued a vision of greater knowledge about mental illness and mental deviance, which would both make society safer and broaden the clientele for research (Ley 2004; Roemer, Kolb, and Faltlhauser 1927). Born under the strain of overfilled asylums and vociferous criticism of custodial care, open care was founded on a comprehensive program of social strategies to cope with mental illness that offered the prospect of lower costs, effective control, as well as early diagnosis and prevention (Schmiedebach and Priebe 2003, 278). This program enabled patients to be discharged earlier and cared for in a familiar environment. Beyond the asylum walls the reformers espied a far greater field for their work. Gustav Kolb, a pioneer of open psychiatric care, put the proportion of “mentally ill and mentally deviant people” (geisteskranke und geistig abwegige Menschen) in the overall population at four per cent (Kolb 1931, 118). Of these, he calculated, only ten per cent were living in asylums.

Psychiatric outside care targeted a hitherto unseen model of medical surveillance. It was based on both voluntary subscription and follow-up welfare work (nachgehende Fürsorge), which meant that the recipients were actively tracked down. Welfare workers entered private homes and offered statistics a new path of access to “the free-living mentally ill.” Hans Roemer, an asylum doctor in Baden who played a major role in establishing open care, proposed a far-reaching system for capturing a broad range of “mental otherness” in 1911. A centralized index (Stammliste) would permit the “insane” and “socially unfit” to be counted and make the findings available for scientific and practical purposes (Roemer 1911/12, 94.).²⁹ The aim was to link data from public and private asylums, courts and prisons, poorhouses and medical statistics. Named counting cards in alphabetical order, with cross-references to kinship and regular updating, would allow biographies to be constructed, and these would help to predict the future behavior of those mentally ill persons who were not confined to asylums. This collection of data, Roemer hoped, would answer the simmering question about rising numbers, assist research on kinship and inheritance patterns and provide insights into forensic matters. The index to be created would be a reliable, if not “complete” tool for surveying non-institutionalized persons who showed signs of mental disorder (Roemer 1911/12).³⁰

Roemer perceived the index’s true value in a closer alliance with open care, which had boosted the acceptability of invading private homes. This would grant access to confidential information which, combined with diligent controls and local knowledge about persons displaying mental symptoms, would unleash a “flow of intelligence” of unprecedented magnitude (Roemer 1913/14, 593). Care providers would be required to feed personal data into this “archive of deviance,” and the biographies collected across regional borders would be copied and made available to local councils. This “growing penetration of socio-psychiatric measures in public life,” Roemer felt, was urgently needed given the threat to society from potentially dangerous individuals with mental disorders. He was aware, he hastened to add, of the tightrope between “considerate care and discreet surveillance” at one end and the “spectre of uniformed policing” at the other (Roemer 1913/14, 590, 592). His warning made it patently obvious how hard it would be in this enterprise to distinguish prevention from control.

Roemer’s proposal for a “Stammliste” was never implemented, but his ideas on meticulously registering mental illness and suspicious social behavior were taken up elsewhere. From 1919 onwards the alienists Gustav Kolb and Valentin Faltlhauser turned the psychiatric outside care in Erlangen into a model institution which made a closely-knit system for recording social, criminal and mental non-conformity very real. Faltlhauser’s vision of the future was to build up a “topography of mental abnormality” (Topographie der geistig Abnormen) for the whole German Reich (Ley 2004, 186-198). Counting is, as Ian Hacking has stated, hungry for categories (1982, 280). Identifying and capturing personal data encouraged greater vigilance and closer attention to the entities that had to be established. The extramural care system and its survey methods not only triggered a new wave of data, but also created a social fact out of mental deviance with all its obscure connotations.

If the earlier attempts to promote open psychiatry had been sporadic, the devastating impact of the First World War made a nationwide count of all infirmities essential. The two-stage procedure tested in the private statistical studies and in Württemberg was adopted by the entire German state for the first time in 1925 and 1926. A nationwide census set out to determine the number of blind, deaf-and-dumb, physically disabled and mentally ill citizens in the republic. The second step was to interview the individuals identified in more detail, depending on the nature of their affliction. Counting cards were used to record the data, which allowed the information obtained to be analyzed in various ways: for the mentally ill, eighteen columns captured their personal and social circumstances and medical data (see Figure 2). This part of the ambitious enumeration was the task of local nurses, who visited the sick and infirm in their homes. Meanwhile, in the asylums, doctors filled in the counting cards for their patients. The local authorities were advised to make duplicates of the cards to create accurate registers of the mentally ill in their communities and to update them regularly.

Figure 2. Counting Card of the census 1925/26.³¹

The count of 1925/26 was by no means uncontested: doctors refused to break their duty of confidentiality, and overstressed nurses kept postponing the deadlines for handing in their questionnaires. The results of this complex count were not published until 1931. Prominent representatives of open psychiatric care were vehemently critical of the findings. They were convinced—due to their operational experience of the extramural territories—that the figures obtained fell “far short of reality” (Kolb and Eitner 1931, 98). Despite all these objections, whether they were appropriate or not, this census differed from all previous enumerations in one crucial methodical respect: in 1925 and 1926 nurses acting on behalf of the state throughout the German Reich had crossed the thresholds of the homes belonging to the families concerned.

6. Conclusion: Evidence of the Unknown

There is a notable common factor in all the surveys of the non-institutionalized mentally ill undertaken in the course of a century that have been analyzed here merely by way of examples. Rarely was an opportunity missed in an unending succession of surveys to point out that only a fraction of the “real” extent of the mentally unsound had been picked up. One might conclude that psychiatrists, enumerators and health administrators were haunted by the idea of not being able to capture the totality of the statistic. The claim of data insufficiency came hand in glove with an assertion by experts that the mentally ill and their families were withholding information about their true condition. This alleged reluctance to pass on information to the enumerators and surveyors legitimated ever more diligent methods to obtain the desired data. References to a mentally ill population that could not be numerically determined, but that was assumed to be out there, had won a firm place in the nineteenth-century canon of psychiatry and census management.

The eagerly expressed view that not all the “free-living” mentally ill had been detected, mostly upheld by well-informed practitioners and therefore hard to contradict, lay in a limbo between knowing and not-knowing. However, the aim is not to determine whether these educated estimates, assumptions or concerns were valid or not. Rather, the acknowledgement of an undercount must in itself be taken into account as an existing practice in the business of gathering hard-to-detect data. On the one hand, the disclosure by those responsible for the surveys that they had not ascertained the “real” figures permitted them to lend transparency and scientific quality to their procedures. On the other hand, “non-data” constantly blurred the results obtained and their messiness was itself crucial evidence of the necessity for further interventions at the margins of growing psychiatric and public responsibility.

Survey practice was influenced in three concrete ways by the assumption that the statistics were incomplete. First, the concern that not every house and every corner of the land could be accessed turned out to be a crucial factor in the interview scenario. Those responsible for the insanity counts realized that the quality of their studies depended heavily on the situation in which the data were requested, and hence on the individuals sent out into the field to collect them. People were being asked to place their trust in the healing promise of psychiatry, and yet they were assumed to be distrustful of the enumerators. Doctors clutching questionnaires, pastors exuding trustworthiness and enumerators tasked with describing the population by capturing facts had to rely on their professional status to overcome the (assumed) coy shame of relatives and expose sensitive information to the public eye. If the details could not be extracted by the direct route, local knowledge must be tapped. The survey and census practice demonstrates how the protagonists were repeatedly operating within a grey zone. Neighbors, doctors and community officials were asked to reveal, complement and correct the information required and make it available in lists and counting cards. The need for untainted and higher figures prompted efforts to bypass the doorsteps of homes. At that point, it appears to be no coincidence that the recommendation to intrude into family life gave way to a warning against intruding too much. Instead of elaborate, costly and dissatisfying surveys, communities had to keep tables of their sick citizens and update them regularly, without the knowledge of those concerned.

Second, the difficulty of gathering valid data was a catalyst for counting, registering and circulating the numbers of mentally ill citizens who were not being treated professionally. Linking the formerly regional surveys to the census, as initiated by the Berlin Medical-Psychological Society, permitted an overview of the entire population. Paper tools made it possible to record, mark, extract and process the precious data of every single citizen. If no anomalies could be identified on the threshold of family homes, medical professionals were able to draw on the archive of counting cards and, usually with the support of local bodies, to revise the census outcomes. For these follow-up-investigations, the individuals identified were visited at home and detailed questionnaires collated an abundance of quantitative as well as qualitative information. Calls to protect personal data were drowned out by the findings of these two-stage surveys, which attracted national as well as international acclaim.

Access to information about non-institutionalized individuals was gradually reinforced in the early decades of the twentieth century by extramural provision. Open psychiatric care and its so called nachgehende Fürsorge (compulsory follow-up welfare supervision) set in motion a flow of sensitive information gathered in the client setting. The registers of sensitive data were founded on a breach of the confidential relationship between the recipients of care and the professionals providing it. Moreover, abundant information was obtained and systematically interlinked without the involvement of the persons concerned, pioneering a formidable permanent information architecture around mental illness and mental deviance. Increasingly networked records documented and consolidated a change in contemporary interpretations of “mental otherness” and its potential dangers.

Third, persistent references to the limits of capture and laments about the inaccuracy and incompleteness of findings sustained a dynamic response and a quest for better numbers on mental illness throughout the country. The “fact” of missing data claimed truth value. Estimates, extrapolations and international comparisons generated figures that attracted greater credibility than the numbers actually gathered. The notion that reliable surveys of the insane were nearly impossible became an integral part of the complex relationship between a modern state bureaucracy, with its urge to control all its citizens, and the scientific need for valid data. The awareness that statistics had a blind side turned out to be the main driver behind methods to facilitate access to sensitive data about the mental condition of every citizen. This went hand in hand not just with an expansion in psychiatric care during the nineteenth and early twentieth century, but with increasingly differentiated categories of mental “otherness” calling for professional support. To underpin the promise of safety and care, both data and “non-data” about mental illness and mental deviance were elevated to a genuine focus of public responsibility.