Chapter 1. The State of the Question: Four Literatures on Acting Before the Evidence

1.1Why begin with the literature

This study reads twenty-three of Skeptron's history cases, from an Egyptian surgical papyrus of about 1600 BC to the World Health Organization's characterisation of COVID-19 as a pandemic on 11 March 2020. The question the study asks of those cases — what was established at the hour a public body acted on disease, and whether the record lets a later reader keep apart the time of the act, the time of the evidence and the time of the count — has to be answered in a field that is already full. Four large literatures have been asking, for decades, what an epidemic is, what a state should do about one, what a number of the dead proves, and how anyone decides when the evidence is thin. Each has its own arguments, its own disagreements and its own way of reading the past. The cases in this study are read through those ideas rather than beside them.

This chapter therefore maps the four literatures before any case is opened. Each work is reported here at the level of what its own reading note gives it, by the fact it settles: what it takes an epidemic, a response, a count or a decision to be, and what it says those things do in the world. Where two works hold opposed positions, both are named with the author who holds each. Where a tradition has no vocabulary for something this study codes — the gap between an act and its record, the years a certification requires — that absence is stated rather than filled.

The mapping is not a verdict on any of these works. What it does is set out where each argument sits, so that the chapters that read the cases can say plainly two things: where Skeptron's records bear on the debate, and where they cannot speak to it at all. A record of a single act cannot settle whether anticontagionism was a doctrine of commerce; a table of burials from 1665 cannot show what a count does for the state in 2026. Where one of the four literatures fits a case badly, the chapter that reads the case will say so.

1.2The first literature: the histories of epidemics and of public response

The first literature treats an epidemic as something a society does something with, and treats the response as following from how the society reads the disease. Each work below differs on what that reading is and on what follows from it.

Charles E. Rosenberg's The Cholera Years: The United States in 1832, 1849, and 1866 (1962) compares responses to three cholera epidemics in the United States. In 1832 cholera was widely read as a moral and religious judgment on the intemperate and the poor; by 1866 it was read as a sanitary problem that public authorities could prevent, and New York's new Metropolitan Board of Health acted on that reading. Rosenberg's point, in the terms of his reading note, is that the change in reading came before the germ theory was established. An epidemic, on this account, is not yet a cause in the laboratory but a thing a city decides what it is; a public response is the set of measures that follow from that decision. In Explaining Epidemics and Other Studies in the History of Medicine (1992), the essay "What Is an Epidemic?" carries the idea further and states its form: an epidemic has a dramaturgic shape, unfolding as in acts — a slow, often reluctant recognition called progressive revelation, the framing of an explanation that manages randomness, the negotiation of a collective public response, often with a ritual side, and a subsidence followed by retrospection. The same volume distinguishes two explanatory styles: disease read as configuration, an imbalance of the whole person and the environment, and disease read as contamination, a specific agent passing between people. This study reads Rosenberg's progressive revelation as naming the recognition problem that the ledger codes as an act taken before the cause it addressed was established; the comparison is made in Chapter 13, where the two are set beside one another.

Paul Slack's The Impact of Plague in Tudor and Stuart England (1985) is a social history of plague in England from the sixteenth century to 1666. It examines the plague orders issued from the 1570s, which shut up infected households, appointed searchers and watchmen and kept the bills of mortality, lists of burials. These measures, in Slack's reading, fell hardest on the poor, provoked resistance, and divided magistrates, clergy and physicians over whether isolation was just. Slack also argues that public measures, especially quarantine against plague from abroad, may have contributed to plague's disappearance from England after 1666. An epidemic, here, is a thing a governing community handles with rules and officers; a response is the rule-book itself, and its cost is read where it falls. The bearers of that cost are named by Slack's note as the poor; the axis on who waits is treated in Chapter 6, which reads Slack in full beside the Ordinance of Labourers.

Carlo M. Cipolla's Fighting the Plague in Seventeenth-Century Italy (1981) studies the health boards of the Italian states in the plague of 1630–31. These were permanent magistracies with powers to impose quarantine, close borders, issue health passes, run pest houses and burn goods. Their measures rested on a mixture of miasmatic (disease carried by foul air) and contagionist ideas, were costly to trade and to the poor, and brought the boards into conflict with merchants and the clergy; how much the boards achieved against the disease is uncertain, in the terms of the note. An epidemic, for Cipolla, is a problem of administration under mixed doctrine; a response is an institution with a budget, a jurisdiction and enemies. The note's uncertainty about what the boards achieved is worth keeping: Cipolla writes an administrative history in which the effect of the measures stays open, which is the same position this study's records often leave a later reader in.

Ann G. Carmichael's Plague and the Poor in Renaissance Florence (1986) uses Florentine records to argue that in the century after the Black Death plague came to be seen as a disease of the poor, and that the city's plague controls, from the fifteenth century under health officials, increasingly targeted the poor as the source of danger. An epidemic, on this account, is a social category as much as a disease; a response is the identification of a group with the danger. Where Slack records a contested justice, Carmichael records, on this study's reading, a settled attribution. Chapter 6 reads both, because the two together give a record's most delicate question — who is named as bearing the cost — a history that predates the act.

Frank M. Snowden's Epidemics and Society: From the Black Death to the Present (2019) is the widest of the five. It is a survey of epidemics from plague to Ebola, and its argument is that epidemics are not random but follow from the ways societies live, trade, build and govern; their study shows a society's structure and its relation to the environment. The book treats the plague regulations of the Italian states as the first public health measures and the origin of quarantine, the cholera of the nineteenth century and the sanitary movement, and, in its last chapters, emerging diseases, with the West African Ebola epidemic of 2013–16 and the weakness of the health systems it struck. An epidemic, for Snowden, is a mirror held up to a society; a response is the society's own structure acting, for better or worse. This study shares Snowden's starting point that the way a society governs is visible in how it meets disease, and differs in the unit of analysis: Snowden reads whole societies across centuries; this study reads dated records one at a time.

1.3The second literature: contagion, quarantine and the state

The second literature asks a narrower and older question: what holds a state to one disease-control strategy rather than another, and where did the rules of separation come from? Its positions are brief enough to state in full here and are taken up again in Chapters 3 and 6.

Erwin H. Ackerknecht's Anticontagionism between 1821 and 1867 (1948) argues that the anticontagionism that dominated medical opinion in the mid-nineteenth century was tied to liberalism and to commercial interests opposed to quarantine, which restrained trade and individual liberty, while contagionism and quarantine went with bureaucratic, absolutist states. The scientific case, Ackerknecht's note adds, was not settled when these positions were taken. On this reading, a doctrine about disease is also a claim about government, and the choice between them is made before the science is.

Peter Baldwin's Contagion and the State in Europe, 1830–1930 (1999) answers by comparing how Britain, France, Germany and Sweden met cholera, smallpox and syphilis. Against the view that liberal states preferred sanitation and authoritarian ones quarantine, Baldwin argues that a state's strategy followed above all its geographical and epidemiological position, its trade and its administrative capacity. Britain's system of notification, isolation and disinfection, in this account, replaced quarantine at the ports, a shift the note attributes to position and capacity rather than to doctrine. This study reads the disagreement between Ackerknecht and Baldwin as a disagreement about what a decision on disease is evidence of: of the regime's character, or of its position and capacity.

Mark Harrison's Contagion: How Commerce Has Spread Disease (2012) widens the frame again. It is a history of the tension between trade and disease control from the fourteenth century: quarantine and embargoes protected health but restrained commerce and were sometimes used as trade weapons, and the international sanitary conferences from 1851 tried to agree on measures that would protect health while sparing trade, out of which international health regulation grew. Valeska Huber's The Unification of the Globe by Disease? The International Sanitary Conferences on Cholera, 1851–1894 (2006) reads the same conferences as a slow and political process: delegates disagreed for decades over the nature of cholera and the use of quarantine, agreement came late, and the conferences served imperial and commercial interests, with the route from India through Suez at their centre, as much as the control of disease. Both works place an act on disease inside a bargain between health and commerce that the act's own text may never mention.

Michel Foucault's Discipline and Punish: The Birth of the Prison (1977) contributes one image rather than an argument about epidemiology: in the chapter "Panopticism" a late-seventeenth-century town is partitioned, each household shut in, inspected daily by syndics and intendants, and every death and illness registered and passed up the hierarchy, a model of segmentation and constant surveillance contrasted with the earlier exclusion of lepers and treated as an image of disciplinary power. Chapter 6 sets Foucault's plague town beside the acts this study reads, and says where it fits and where it fails them.

On the origin of the rules of separation, two works stand together. Gian Franco Gensini, Magdi H. Yacoub and Andrea A. Conti's The concept of quarantine in history: from plague to SARS (2004) gives a short history of quarantine: Ragusa's thirty-day separation, the trentino, of 1377, later extended to forty days, the quarantena, Venice's lazaretto, a quarantine station, and the use of quarantine down to SARS in 2003. The forty days, in the note's words, had no basis in observed incubation; the number came from custom and symbolism. Zlata Blažina Tomić and Vesna Blažina's Expelling the Plague: The Health Office and the Implementation of Quarantine in Dubrovnik, 1377–1533 (2015) follows the same decision down through its enforcement and revision: the Great Council of 1377 required people coming from places of plague to spend a month on islands or at Cavtat before entering the city, a health office with its own officials grew out of the measure, and the commune balanced protection against its trade. Ragusa's month of separation is read in full in Chapter 3, which asks what a rule whose number came from custom and symbolism rather than from observed incubation asks of the reader who must judge it six centuries later.

1.4The third literature: numbers and their authority

The third literature does not take disease as its subject. It studies what a count is, how counts come to be believed and what they do to the institutions that keep them. All five works are read in full in Chapter 5; what follows is the position each holds.

Ian Hacking's The Taming of Chance (1990) traces how, in the nineteenth century, an avalanche of printed numbers — official counts of births, deaths, diseases, crimes and suicides, especially from about 1820 to 1840 — led to the idea of statistical laws governing society, and how the categories used for counting shaped what was counted. Determinism eroded as chance came to be seen as lawful. A count, on this account, is not only a tally; it carries a category, and the category decides what will be seen next time.

Theodore M. Porter's Trust in Numbers: The Pursuit of Objectivity in Science and Public Life (1995) argues that quantification is a technology of distance: rules for counting and calculating let communication reach beyond a community of trust. Mechanical objectivity, following rules rather than judgment, is pursued most by weak or contested experts and by bureaucracies under outside pressure, as a substitute for personal trust, rather than by secure elites. A count, for Porter, is a claim that survives a journey to an audience that cannot check the people behind it.

Alain Desrosières's The Politics of Large Numbers: A History of Statistical Reasoning (1998) is a history of statistics as bound up with the making of states. Statistical objects, in this account, exist through conventions of equivalence: to count, one must first decide which cases are of the same kind, and those classifications are at once technical and political; once established, they hold, and become things that can be acted upon. A count, for Desrosières, brings a class of cases into existence as one thing that an administration can address.

James C. Scott's Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed (1998) describes how states make society legible by simplifying it — standard names, maps, measures, registers — and argues that high-modernist schemes which impose such simplifications and override local, practical knowledge, mētis, have failed, sometimes disastrously. A count, for Scott, is one of the simplifications, and this study reads Scott's simplifications as making a count's success a matter of what it shows an administration.

Andrea A. Rusnock's Vital Accounts: Quantifying Health and Population in Eighteenth-Century England and France (2002) brings the argument back to medicine. It shows how physicians and others in eighteenth-century England and France turned to numbers, the bills of mortality and new tallies, to settle medical and political questions, notably whether smallpox inoculation was worth its risk, by comparing the deaths from natural and inoculated smallpox; and that the trust such accounts won depended on how the numbers were gathered and who vouched for them. A count, for Rusnock, is a settlement of a live dispute, and its authority is a matter of vouching.

The public-domain text that goes with this part of the literature is John Graunt's Natural and Political Observations Made upon the Bills of Mortality (1662), first published in London. The record this study holds for Graunt is fixed at 1665, the fourth edition, printed at Oxford; the passages read here come from the first edition, and Chapter 5 reads the record. Graunt read the weekly bills that London's parish clerks compiled from reports of the searchers, whom his note describes as mostly ancient matrons who viewed the dead and reported the cause. His own words give the act's shape: "The Searchers hereupon (who are antient Matrons, ſworn to their Office) repair to the place, where the dead Corps lies, and by view of the ſame, and by other enquiries, they examine by what Diſeaſe, or Caſualty the Corps died." Three moments sit in that sentence — a death, a view at the body, and a report that reaches a bill later — and Graunt adds that the count built from them does not reach the number it is said to record: "That the knowledg even of the numbers, which die of the Plague, is not ſufficiently deduced from the meer Report of the Searchers, which onely the Bills afford;" His note lists the reasons he gives for that insufficiency: the searchers' judgments, the burials that were never registered, and the confusion of diseases under one name. This study reads the passage as the third literature's subject existing before any of the five works: a count whose date is weekly, whose cause is a judgment made at a view, and whose reliability its own maker disputes. Nothing here settles how good the bills were. What the text does is date the act and leave the cause to judgment, which is what a later reader must know before reading the count at all.

What the study's own records carry is set out in The ledger's numbers. On this study's count field — does the case's own text give a count of people, cases, deaths, votes or items, and how does it qualify the count — the two codings agreed in 78 of the 89 cases, or 88 per cent. The agreed values stand as the table gives them, with no-count 48 as the largest single value, and qualified-by-definition-scope-or-denominator 7, settled 5 and provisional-or-revised 1 among the cases that give one at all. The registered hypothesis that such counts are qualified rather than settled is unclear on a strict reading (supports 8, against 5, unclear 25, not applicable 51) and supported on a lenient one (supports 8, against 5, unclear 11, not applicable 65). This study reads that split as the same problem these works name from the other side: whether a count reads as settled depends on what the record carries — a definition, a scope, a denominator — and on what a reader is willing to take as given. A later reader can tell, in most cases, what date a record puts on its count; what a record most often does not carry is the unit, scope or population against which a later figure could be compared as the same measurement.

This study reads the five positions as five different answers to one question — what does a count do besides counting? On this reading, Hacking's count generalises, through its categories, about society as a whole; Porter's travels, carrying a claim where trust does not reach; Desrosières's constitutes, making a class of cases exist as one thing that can be acted upon; Scott's simplifies, rendering a society into something an administration can read; Rusnock's settles, ending an argument about risk by producing numbers whose vouching is on the record. This study's question — whether the record keeps the time of the count apart from the time of the act — is not answered in the terms of any one of the five notes; each addresses what a count does with what has already been gathered. Chapter 5 sets the five positions beside the records.

1.5The fourth literature: deciding before the evidence

The fourth literature is the one most directly about acting while the evidence is thin. Its subject is the decision itself: how it is formed, what the situation looks like to the people forming it, and what a later reader can know about the hour it was taken.

Richard E. Neustadt and Harvey V. Fineberg's The Swine Flu Affair: Decision-Making on a Slippery Disease (1978) is a study, written for the Secretary of Health, Education and Welfare, of the 1976 United States decision to immunise the whole population against a swine influenza found at Fort Dix, which no pandemic followed. The study traces overconfidence built on thin evidence, the momentum of a programme once announced, the merging of separate decisions — to make the vaccine, to inoculate everyone — and the absence of planned points at which the decision would be looked at again in the light of new evidence. Its recommendation is to ask what would change the decision and to build such points of reconsideration into the programme. This study reads that recommendation as the nearest thing in this literature to the axis this study calls the condition of a later comparison; what the records leave of it, and what they do not, is Chapter 11's question, and Chapter 12 reports what the codings found.

Richard E. Neustadt and Ernest R. May's Thinking in Time: The Uses of History for Decision Makers (1986) proposes ways for decision makers to use history: to separate what is known, what is unclear and what is presumed about a situation; to ask what the story of the issue is and when it began; to test analogies by listing likenesses and differences; and to place people and organisations in their own histories. The 1976 swine influenza programme is one of their cases. The method is written for people inside a decision; this study applies the same separation from the outside, to records, which is a different position from the one the method assumes.

Reinhart Koselleck's Futures Past: On the Semantics of Historical Time (1985), first published in German in 1979, argues that historical time is shaped by the relation between the space of experience, the past present in memory, and the horizon of expectation, the future anticipated now; and that in the modern age expectation drew away from experience. The two times this study keeps apart — the hour of the act and the hour of the evidence — are not Koselleck's terms, but his distinction makes the separation thinkable: an actor at the hour of an act works with an experience of the past and an expectation of the future, and the record later read is what that experience looked like to the people who wrote it down.

Barry A. Turner's Man-Made Disasters (1978) argues that disasters incubate: warning signs accumulate, unnoticed, misread or discounted, over a long period before the event, because of rigid beliefs, the difficulty of handling information in organisations, and the minimising of danger; after the event, an inquiry reconstructs what should have been seen. Diane Vaughan's The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA (1996) shows how, at NASA, signals of danger came step by step to be accepted as normal within the culture and its rules — the normalisation of deviance — so that following the rules led to the Challenger disaster; structural secrecy kept information from those who decided. Both works are written about events whose later inquiry has already reconstructed the warning signs; this study reads records that mostly stop before such an inquiry, which is why the two works fit some of its cases and not others.

Ulrich Beck's Risk Society: Towards a New Modernity (1992), published in German in 1986, argues that modern societies produce risks — pollution, radiation, toxic substances — that cannot be perceived by the senses and become visible only through scientific knowledge; who defines what counts as a risk, and on what evidence, is therefore a matter of power. Sheila Jasanoff's The Fifth Branch: Science Advisers as Policymakers (1990) studies scientific advisory committees in United States regulation: the science used for regulation differs from research science, since it must produce answers under time pressure and uncertainty, and advisers and agencies negotiate the boundary between science and policy, with a decision's legitimacy depending on that negotiation. Beck names the politics of definition; Jasanoff names the room in which the definition is argued out. Neither work takes disease as its case study, and both belong here because the acts this study reads — to notify, to declare, to certify — are acts of definition in a room.

The dispute inside this literature runs between Cass R. Sunstein's Laws of Fear: Beyond the Precautionary Principle (2005) and the editors of Late Lessons from Early Warnings: The Precautionary Principle 1896–2000 (2001), an issue report of the European Environment Agency, by Poul Harremoës and others. Sunstein argues that the strong precautionary principle is incoherent, since precaution against one risk creates others and risks lie on all sides of a decision; he recommends weighing costs and benefits with margins of safety, and explains public fear through availability and probability neglect. The Late Lessons editors give case studies of hazards — among them asbestos, benzene, PCBs and BSE — in which early warnings were available long before action was taken, and draw twelve late lessons, among them to acknowledge ignorance as well as uncertainty, to keep up long-term monitoring of early warnings, and to look for blind spots in scientific knowledge. This study reads the two positions as differing on what an early warning is: for one, a signal that must be weighed against other signals; for the other, a record of how long a signal sat unused. Chapters 3, 11 and 13 return to this dispute against the acts of this study, and say what the records show of the weighing and what they do not.

1.6What counts as evidence that an act worked

Cutting across the four literatures is a question none of them alone answers: by what a later reader counts as proof that an act on disease did anything. Four works supply the standing answers, and Chapters 4, 5 and 7 take them up in full.

Harry M. Marks's The Progress of Experiment: Science and Therapeutic Reform in the United States, 1900–1990 (1997) argues that the controlled clinical trial grew from therapeutic reformers' distrust of the individual physician's judgment and of commercial claims: reformers sought to discipline therapeutic knowledge by collective, controlled experiment, and the randomised trial's authority was built over decades. The controlled comparison, on this account, is not a natural fact about medicine but an institution that had to be built before it could be trusted.

Marc Lipsitch, Christl A. Donnelly, Christophe Fraser and others' Potential Biases in Estimating Absolute and Relative Case-Fatality Risks during Outbreaks (2015) sets out why estimates of the share of cases who die, made during an outbreak, are biased: deaths lag behind cases, so early ratios of deaths to cases understate it; mild cases are missed, so ratios overstate it; and case definitions, testing and reporting differ between places and over time. Comparisons across places and periods, the note concludes, need these biases taken into account. Lipsitch's paper is the closest thing in this study's library to a general warning about counts taken at the hour of an act; Chapter 10 applies it to the Ebola and MERS records.

Richard J. Hatchett, Carter E. Mecher and Marc Lipsitch's Public health interventions and epidemic intensity during the 1918 influenza pandemic (2007) studied seventeen United States cities in the autumn of 1918. Cities that introduced nonpharmaceutical interventions — closing schools, churches and theatres, banning gatherings — early in their local epidemic had lower peak death rates than cities that acted later; Philadelphia, where the first civilian cases were reported in mid-September and interventions came only in early October, is set beside St Louis, which acted within days. The study is observational: it compares cities by timing, and does not isolate the effect of any one measure. Howard Markel, Harvey B. Lipman, J. Alexander Navarro, Alexandra Sloan, Joseph R. Michalsen, Alexandra Minna Stern and Martin S. Cetron's Nonpharmaceutical interventions implemented by United States cities during the 1918-1919 influenza pandemic (2007) is the companion study, across forty-three United States cities in 1918–19: cities that put school closure and bans on public gatherings in place together, early, and for longer had lower excess death rates, and cities that lifted measures early often saw mortality rise again; the authors argue for layered, early and sustained interventions from historical, observational data.

This study reads these four works as supplying a ladder of evidence about an act on disease: at the bottom, a single case or a single act; above it, a controlled comparison; above that, a comparison of places by timing, which can show association without showing which measure did what. This study does not climb that ladder. It codes only what a record lets a later reader judge about the hour of the act, and it makes no claim that any restriction, remedy or declaration succeeded or failed, because most of the records in hand do not carry the outcomes in a form that would allow such a claim. Chapter 4 reads the trial and the single case, Chapter 5 the outbreak's own counts, Chapter 7 the comparison of cities.

1.7The gap

Each of the four literatures has a horizon of its own. The histories of epidemics read a society across a century: a reading, a set of measures and the cost they bore. The literature on contagion and the state reads a doctrine or a strategy across a continent and a decade. The literature on numbers reads a counting practice across an administration and an age. The literature on deciding reads a decision, and usually reads it from the far side of the outcome. All four are answerable to something larger than one record: a society, an institution, a generalisation, a programme. None of them, in the terms of its own note, reads, case after case on the same axes, the records a later archive keeps of acts on disease, for what they let a reader know of the time of the act, the time of the evidence and the time of the count.

That narrower place is this study's. Its unit is a dated record of a single act on disease: a resolution of a council, an order from a king, a notice from a hospital, a table of burials, a trial of twelve men, a report to an academy, a certification after years of surveillance. Its question about each such record is not whether the act was right, nor what the disease was, nor whether the measures helped. It is whether the record keeps three things apart that a later reader would want apart: the hour the act was taken, the hour the evidence for it was established, and the hour the count that would judge it was made. To answer that, the study uses Skeptron's history cases, written in Korean and read here in an English translation made once for this study, beside the public-domain texts their actors wrote where those survive; and it codes those cases together with the rest of Skeptron's medicine, disaster and industrial histories, eighty-nine cases in all, on the same axes. The method that makes that comparison possible is Chapter 2's subject.

What a record dates and what it leaves undated is this section's ground. Every record in hand fixes a date for the act itself — a resolution, an order, a notice, a table, a report — and leaves undated the process the act names: the spread of a disease, the total of the dead, the hour at which a cause was settled, the years over which a certification ran. The World Health Organization's characterisation of 11 March 2020 is one such record, and its date is the date of the characterisation. The ledger, set out in Chapter 2, holds one row per case and one cell per axis; its two codings are one model run twice, reported as two codings. The ledger's numbers show how much of the three hours a later reader is actually given. Across all eighty-nine cases a deadline appears in the record in 88, and something missing is listed in every case, 264 items in all; no case states a condition for revisiting its own judgment. The study's own field that asks the whole question at once — a public act, a named gap between the event and its record, who bore the wait named, what would change the reading named — is filled in 1 of 36 cases, or 3 per cent, and among the twenty-three cases this study reads closely, in 1. These are counts of what the records carry, not of what the acts did. The series' claim that few decisions write down in advance what would reverse them, and that those that do write it in words no one could observe, is counted as supported in all eighty-nine cases. This section reports that count and draws no conclusion from it; the hypotheses and their verdicts belong to Chapter 12.

The source texts are read where their cases are read, and they answer a different part of the question. Boccaccio's Introduction to the First Day and Thucydides's account of the plague go with the acts of 1348 and 430 BC in Chapter 3; Snow, Lind, Jenner and Pasteur with the trials and single cases of Chapter 4; the bills of mortality and the report of 1858 with the counts of Chapter 5; the WHO Constitution with Chapter 8; the Dongui Bogam's preface with Chapter 9. What a source text holds that a record's date alone cannot is the actor's own account of the hour: Snow's second edition writes, "In consequence of what I said, the handle of the pump was removed on the following day." The record for that case fixes the date 1854-09-08; this study reads the sentence above as fixing what happened after what, which a date cannot do. The texts are unevenly available, and the study says so: no source text goes with the cases whose originals are not in the public domain, whose terms forbid reproducing them, or that could not be reached. Where a text is absent, a later reader has only what the record carries.

Skeptron's records hold one case on COVID-19, the World Health Organization's characterisation of 11 March 2020; lockdowns, vaccines and national responses are not in them, and this study does not use them as evidence. That statement stands wherever the subject is used, and it is worth saying here because the fourth literature, and the debate over precaution in particular, is often read as having been settled by events after 2020; this study's records cannot speak to the rest.

What this study does not add has to be said plainly. It adds no new history of any epidemic: where it reports what a disease was or how it spread, it reports the reading note of a work already in the library above. It adds no medical knowledge, and gives no judgment about how anyone should be diagnosed or treated; the sick and the dead appear only as the records and the source texts name them, in the numbers those texts give. It adds no claim about how often acts on disease succeed or fail: most of the records in hand do not carry the outcome of the act in a form that would support such a claim, and the one comparison that comes close — the observational comparison of American cities in 1918 — is observational. And it adds no verdict on any actor's judgment at the hour it was taken: each act is shown with what was known then, what it cost and who bore it, what evidence came later, and what the record does or does not say about what would change the reading. The findings, and the hypotheses the ledger's counts support or go against, are left to the chapters that carry them, beginning with the method.