Chapter 7. Gates, Orders and Assemblies Compared

7.1What the restrictions were

Four records hold four acts on disease, and three of them are read in full in Chapters 3 and 6: Florence in 1348, the Ordinance of Labourers of June 1349 and Ragusa's month apart of 27 July 1377. This chapter puts those three beside a fourth, Philadelphia in the autumn of 1918, and compares them rather than restating them. Each record gives what it holds as the act, the period and the scope; none of the four holds an account of what was weighed in the deliberation before it. What they do hold is enough to set them side by side on four questions: who made the act, what was restricted, when it was made, and on what knowledge.

Boccaccio's text, in the Introduction to the First Day, names the measures in one clause: "the cleansing of the city from many impurities by officials appointed for the purpose, the refusal of entrance to all sick folk, and the adoption of many precautions for the preservation of health;" — Giovanni Boccaccio, The Decameron, the First Day's introduction (translated by J. M. Rigg). The record's own answer sets the limit of what may be read off those measures: it holds that the causal interpretation of the time may be retained only as a symbolic or interpretive frame, since the independent effect of each measure cannot be derived from the mere ordering of preventive measures and the outbreak.

The Ordinance of Labourers survives in English translation in the collection of Adams and Stephens; the text printed by Ernest F. Henderson addresses the king's ordinance to the archbishop of Canterbury and gives its reason in the king's own words, that many labourers, seeing the necessity of masters and the scarcity of servants, "are not willing to serve unless they receive excessive wages" — The Ordinance of Labourers, in the translation printed by Ernest F. Henderson (1892). The record's answer to whom the ordinance reached lists the exceptions for one's own livelihood, land and occupation, the requirement that able-bodied persons under sixty work, and the restriction of wages to customary existing practice.

Ragusa's record gives the rule and the vote in one answer: a stay of one month in Merkan or the old city for persons coming from places of plague, with 33 votes in favour out of 47 present. The record keeps that month distinct from the later forty-day practice and adds that institutional enactment alone does not confirm any reduction in infection, a boundary its answer to the third question repeats. Philadelphia is read in the next section; the table carries its three dates because the record carries them.

All four records carry Skeptron's standing caution that announcement, decision, implementation and outcome are not to be merged into one event. That caution is Skeptron's method and is set out in Chapter 2; it appears in these records as it does in every history case of the series, and it is not a finding about any one of them. What this table adds is the shape of the four acts as restrictions: each fixes who must move, serve or assemble differently, and none of the four records carries a count of what the restriction did.

7.2Philadelphia in 1918

Skeptron's record for this case is built on a study published in 2007, and its Scenario states the sequence the study reconstructed:

The Scenario carries three dates in one line: civilian patient reports on 17 September, the large-scale parade on 28 September, and the introduction of school closures, assembly restrictions and the like on 3 October. None of the three is a document of the hour; all three are entries the 2007 research reconstructs, and the record's dating answer says so, holding that the citation it draws on is a 2007 research announcement. The fifth question of the record asks what post-hoc confusion is prevented by keeping the two actions separate stages, and its answer names the first interval in days:

The same answer's third part states the boundary of the judgment: the pre- and post-difference of one city cannot determine with certainty that a single event was the cause of all deaths, and the record distinguishes the observational research comparison from a causal assertion about a single event. The record names the interval between the first reports and the parade, 11 days. It names a third date but does not state the interval between the parade and the closing of gatherings; the Scenario places the two five days apart in calendar terms the record itself supplies, and that arithmetic is not written into the answer.

Beside the record stands what a reader in those weeks could see. The reports read here from the autumn of 1918 are the United States Public Health Service's weekly summaries on the prevalence of epidemic influenza in the United States, printed from telegraphic reports received by the Surgeon General. The text given for them is uncorrected machine OCR, so this study gives their sense in its own words rather than quoting the transcription. The first issue states that influenza is not a reportable disease in most of the States, that it is therefore impossible to secure complete reports of its prevalence among civilians, and that what follows is a summary of reports received up to the time of going to press. Each issue closes at the press, so a reply received later falls into a later issue. This study reads the weekly report as waiting not for a decision but for a reply.

The four issues run in order from 27 September to 18 October 1918, and the Pennsylvania entry in each shows what the national record knew about Philadelphia week by week. In the first, of 27 September, the Pennsylvania line reports comparatively few cases among the civil population, a few at Hog Island, and some said to have been found in Pittsburgh; Philadelphia is not named at all. In the second, of 4 October, the line reports that in Philadelphia and Pittsburgh the disease is epidemic — the transcription carries a stray character at the head of the entry, and the city's name stands inside a sentence about two cities at once. In the third, of 11 October, the Pennsylvania entry says the disease is especially prevalent in the industrial districts of Philadelphia and vicinity, together with the Lehigh Valley anthracite region and scattered sections of the bituminous coal fields; the same issue records that schools have been closed and public gatherings prohibited at several towns in Washington, and separately records that in New York City schools and theaters have not been closed. The served transcription places that closing line at the end of the New York entry, while the header note to the source text names Maryland's entry of 11 October. In the fourth, of 18 October, the Pennsylvania entry gives 6,081 deaths from influenza and 2,651 deaths from pneumonia registered in the state up to 14 October, and adds that the greater number of these deaths occurred in Philadelphia.

So the national weekly record for Philadelphia in these weeks carries a change of wording, then a word for a state, then a state's deaths. It carries no entry naming a closure of schools or a prohibition of gatherings for Philadelphia, while naming such measures for other places; and its own first issue states that its counts cannot be complete. Chapter 5 sets out why counts made during an outbreak carry such biases, and Chapter 10 applies them to later outbreaks.

The comparison the record draws on was made later still. Two studies published in 2007 — one of the autumn of 1918, one across 1918 and 1919 — set the measures of American cities against the timing of each city's local epidemic, and both work from observational data. Both studies therefore compare cities by the timing of measures; neither isolates a single measure, and neither is a study of one city's before and after, which is the distinction the record's own answer draws.

The later accounts read that autumn afterwards. Alfred W. Crosby, in America's Forgotten Pandemic: The Influenza of 1918 (1989), narrates the autumn wave in American cities, among them Philadelphia, where the epidemic overwhelmed hospitals, doctors, nurses and the means of burial, and asks why so deadly an epidemic left so little mark on American memory. Crosby's history gives the city institutions overrun and a city in its aftermath; it does not supply the date of a closing order, and neither the record nor the weekly summaries carry one for Philadelphia in these weeks.

The record and these accounts therefore sit at three distances from the act. The contemporaneous national record of the act's own weeks is the weekly telegraphic summary, which knows no parade and no closing order for Philadelphia. The three dates a later reader uses come from a study published eighty-nine years afterwards, and the record's answer says as much. The comparison of cities by timing comes later again and, by its authors' own notes, is observational. What the record holds at the end is a date, a sequence and one interval in days: the hour of the act, the hour of the evidence and the hour of the count are three moments, and only the first of them, 28 September, is written in the Scenario as a fact of the city's own day.

7.3Compared on the axes

The table below sets the ledger's cells for these four cases beside this study's own fields, both drawn from "The ledger's numbers". The columns differ in what they ask: the series' cells are readings of setting made by the coders, while this study's own fields ask what the case's own text shows. Unsettled means the two codings of one model run twice gave different values; unknown means neither found the answer in the record; a class means read from the setting rather than from a sentence in the record. The three European cases are read in Chapters 3 and 6 and Philadelphia in 7.2; this section compares them on the cells.

Who decides. The coders agreed on the decider cell in 53 of 89 cases across the ledger, and the count of cases whose record names a decider reads 0 of 89 — that count reads the records' answers rather than their titles; the agreed values stand as 36 unsettled, 30 unknown, 9 institution-or-state read from the setting, 8 team-or-organisation read from the setting, 3 institution-or-state stated in the record's words, 1 multilateral-or-civilisation stated, 1 multilateral-or-civilisation implied and 1 team-or-organisation stated. Three of the four cases here carry a reading rather than a stated office: the Ordinance and Ragusa both read institution-or-state, Florence reads team-or-organisation, and all three rest on the setting — the ordinance's address to the king, Ragusa's Great Council, the officials appointed for the purpose in Florence. Philadelphia's cell is unsettled, meaning the codings divided rather than that the record was silent on authority.

Horizon. The records fix dates and, in Ragusa's case, a period of one month; they do not state how long any act was expected to run or be reviewed. Only Ragusa yields a horizon reading — days to months — because the restriction itself is a stated period; Florence, the Ordinance and Philadelphia are unsettled in the ledger's horizon cell. Across the ledger the coders agreed on horizon in 32 of 89 cases, the lowest of the cells named here; the agreed values are 57 unsettled, 14 unknown, 6 years read from the setting, 4 centuries or more, 4 decades, 3 days to months and 1 day to months stated. The ledger's horizon table records one case at days to months in the strict coding and, in the lenient coding, four at days to months, six at years, four at decades and four at centuries or more, and gives each of those groups no conditions of revisiting. The note on the concept of quarantine in history, by Gensini, Yacoub and Conti (2004), states that the forty days to which Ragusa's month was later extended had no basis in observed incubation and that the number came from custom and symbolism; it describes the later practice, not the council's own reasoning in 1377. No comparison of horizons across these four can be made from the ledger's cells, and this chapter does not make one.

Reversal. The ledger keeps apart the constraint on reversing an act — whether a restriction can be lifted — and the action of reversing it — whether what it stopped can be restored. It reads Florence's measures as easy to reverse, the Ordinance and Ragusa as unknown, and Philadelphia as unsettled; across the eighty-nine the agreed values for the constraint read 51 unsettled, 24 unknown and 14 easy, and for the action 53 unsettled, 24 unknown, 11 easy and 1 costly, with every easy or costly reading implied by the setting rather than stated in a record's own words. Two observations come from the records rather than the cells. The ordinance's own text contains one reversal provision: a covenant made before the ordinance for higher wages shall not bind the covenanting party to pay more than was wont to be paid; the record's reading of unknown does not carry that clause. And none of the four records names what would follow a lifting — no clause on restoration, no date, no authority — so what a restriction stopped is left open in all four cases. Chapter 6 reads the Ordinance's enforcement, where the record of what the restriction actually reached begins. The coders agreed on the constraint cell in 38 of 89 cases and on the action cell in 36.

Who bore the wait. The ledger's waits-and-pays cell reads those-nearest-the-stoppage for the Ordinance and Ragusa, unsettled for Florence and unknown for Philadelphia; across the ledger the coders agreed on that cell in 52 of 89 cases, with agreed values of 38 unknown, 37 unsettled, 8 those nearest the stoppage read from the setting, 5 third parties or the vulnerable read from the setting and 1 those nearest the stoppage stated in the record's own words — the last being none of these two cases. The columns part company for Philadelphia: the ledger reads unknown, this study's own field reads the bearer as named, because the Scenario names civilian patient reports. Boccaccio's text says that citizens avoided one another, that fathers and mothers abandoned their own children, and that the sick were left to the charity of friends, of whom there were few, or to servants who could barely be had. The Ordinance's text names the people the act reaches: every man and woman under sixty, of whatever condition, without land, craft or other means of living. Ragusa's record names arrivals from places of plague — a class, not individuals. Philadelphia's record names civilian patient reports.

Act before proof. The cell reads yes for Florence, Ragusa and Philadelphia, and not-said for the Ordinance: the record does not say whether that act preceded the establishment of the cause it addressed. The ordinance's preamble, which the record does not carry, treats the mortality as already past: a great part of the people, especially the workmen and servants, has now died in that pestilence, and the reason given is the labour market, not the pestilence itself. The act thus follows the deaths and names no cause of the disease; the record's reading of not-said and the text's silence on the cause stand side by side here, and neither is read as a verdict. Later work supplies what the hour did not have: the note on a draft genome of Yersinia pestis, by Bos and colleagues (2011), states that the Black Death was caused by that organism, recovered from teeth of people buried in London, while an earlier study, by Cohn (2002), had questioned the identification from the patterns contemporaries recorded. Neither note is part of these records, and the ordinance itself names a labour market rather than a germ. Across the ledger the coders agreed on this field in 69 of 89 cases: among them 39 are not applicable, 17 read yes on the setting, 8 read yes on the record's words, 4 read not-said and 1 reads no, while in the remaining 20 the codings did not settle the cell.

Counts and gaps. Three of the four carry no count in this study's own field; Ragusa's count is the council's vote, qualified by the scope of the assembly present. Across the ledger the count cell was agreed in 78 of 89 cases, or 88 per cent, with agreed values of 48 no-count, 14 no-count read from the setting, 11 unsettled, 7 qualified by a definition, scope or denominator stated in the record's words, 5 settled stated, 3 no-count stated and 1 provisional or revised stated. Gap named — two moments with different dates written into the case's own text — reads named for Philadelphia and not-named for the three European cases; that cell was agreed in 76 of 89 cases, with 39 named stated, 34 not-named, 13 unsettled and 3 not-named read from the setting. Each of the three European cases carries a date or a period but no interval between an event and its record: 1348 stands in Boccaccio's account of measures and symptoms together, June 1349 is the date of the act itself, and 27 July 1377 is the date of the resolution. Philadelphia's record carries three dates in order and states the first interval in days, which is why it is the only one of the four that could be judged against the study's test; Chapter 12 judges that test and where these cases stand among the eighty-nine. This study's own revision field — what later observation would change how the act is read — reads named for Florence, Ragusa and Philadelphia and not-named for the Ordinance; the codings agreed on that field in 67 of 89 cases, with 15 named stated, 10 not-named and 3 named read from the setting. These four cases stand in the ledger's cells the way the ledger's cases stand in general: on cells the codings read from settings, on cells they could not settle, and, in one case of the four, on a cell where the record's own words carry the value.

7.4What would have changed these acts

Three of the four records carry a fifth question asking what must be observed to maintain or withdraw the comparison with a later matter, and each of those three answers names observations; Philadelphia's fifth question asks instead how many days elapsed between 1918-09-17 and 1918-09-28 and what confusion is prevented by keeping the two actions separate, and its answer names the first interval in days. The Ordinance's answer asks for the delay and the distribution of burden between the agreed payment and the actual payment, and holds that where the payment unit, the scope of labour and the coercive conditions differ, the arrangement is not to be read as the same wage structure. Ragusa's answer asks for records of recovery tests, alternate routes and remaining shortfalls, and withdraws the determination of full recovery if only some functions have resumed or if the alternate route depends on the same cause. Philadelphia's answer sets what the observational comparison cannot settle: one city's before-and-after difference cannot fix a single cause for all deaths.

These are conditions for a later comparison, written into the record form. The ledger counts no case among the eighty-nine as carrying a condition for revisiting the judgment itself (A5, 0 of 89), because the history records have no such field; what a record would need to carry instead is Chapter 11's subject, and the verdict the ledger draws from this count belongs to Chapter 12. Each of the four records does list what is unverified in its own Unknown field, and the ledger counts every one of the eighty-nine cases as listing what is missing, 264 items in all. The closing sentences of each fifth question belong to Skeptron's frame, set out in Chapter 2; they are the studio's method, not findings about these acts.

What the records leave unsaid is as specific. None of the four names who would lift a restriction, on what evidence, or by when; none names a date for review; none records what anyone weighed before deciding. The Ordinance's source text gives a reason for the act, names who must serve and what they may be paid, and provides penalties for departing from those provisions, but no review. Boccaccio's text says the measures were made despite all that human wisdom and forethought could devise to avert the pestilence, and gives no account of a debate about them. The Ragusa record gives a rule, a place and a vote, and no argument. The Philadelphia record gives three dates and no authority; the contemporaneous weekly reports of the autumn name no closure for Philadelphia in the Pennsylvania entries of any of the four issues.

What would have changed each reading, on the record's own terms, differs by case. For Florence, a count that kept the measures apart; the record's answer says the independent effect of each measure cannot be derived from the mere ordering of preventive measures and the outbreak. For the Ordinance, evidence of what was actually paid and what was actually enforced, and any comparison of wage structures would have to keep the payment unit, the labour's scope and the coercive conditions apart, as the record's own answer says; Chapter 6 reads the enforcement record that such evidence begins to be. For Ragusa, records of whether arrivals stayed the month and whether cases followed; the record states that institutional enactment alone does not confirm any reduction in infection. For Philadelphia, the city's own documents of the closing of gatherings — no such document is named in the record, and the weekly reports carry none for Philadelphia in these weeks — and, beyond the record, a denominator that the study's comparison uses but the record does not carry.

None of this is a verdict on any of the four acts. Each is read for what its record holds at the hour it was taken, what its own text says was still open then, and what only later accounts can say.

7.5What the comparison shows and does not

Set side by side on the ledger's cells, three of the four acts were taken before the cause they addressed, or their own effect, was established, and for the fourth the record is silent; the two codings agreed on the field in 69 of 89 cases, reading yes on the setting in 17, yes on the record's words in 8, not-said in 4 and no in 1, with 39 not applicable and 20 unsettled. What the four records keep apart is one separation in four materials, and each holds it differently.

Florence's measures stand in the text before the symptoms. The record dates the case 1348 and adds that serious symptoms appeared around the spring of 1348; it gives no date for the measures, so their order within the year rests on the narrator's sequence rather than on a calendar entry. The figure the Introduction gives for the dead — upwards of a hundred thousand human beings between March and the ensuing July — is the narrator's estimate, not a count, and nothing ties it to the measures.

Ragusa's rule precedes any record of its effect. The record fixes 27 July 1377 · Great Council Original Date, holds the rule and the vote in one answer — 33 votes in favour out of 47 present — and keeps that month apart from the later forty-day practice. What the arrivals did afterwards is undated in the record, which states that institutional enactment alone does not confirm any reduction in infection.

The Ordinance follows a mortality that is a premise rather than a counted figure. What it gives as its reason is the labour market — masters in straits, servants scarce, survivors refusing to serve except at excessive wages — and it addresses no cause of the deaths at all. This study reads the act as a settlement of wages and work written after a mortality, not as a control of disease taken while its cause was unknown; the ledger's cell for that question reads not-said, because the record's own words do not settle it.

Philadelphia's parade stands between reports and restrictions. The Scenario gives civilian patient reports on 17 September, the large-scale parade on 28 September and the closing of gatherings on 3 October 1918; the answer names the first interval in days, 11, and the record's dating answer holds that the citation it draws on is a 2007 research announcement, so the three dates are a later reconstruction rather than the city's own documents. A later reader holds two records of what was known at the hour. The reconstruction says civilian reports were already in hand before the parade. The national weekly summary of 27 September — whose own first issue states that influenza is not a reportable disease in most States and that complete reports of its prevalence among civilians could not be secured — records comparatively few cases among the civil population in Pennsylvania, a few at Hog Island and some said to have been found in Pittsburgh, and no parade at all. The summary of 4 October first names Philadelphia, in the same line as Pittsburgh, as a place where the disease is epidemic. So a later reader can date what the national record knew about the city week by week; what the city itself held on 28 September the record does not carry.

The asymmetry among the four is one the ledger's own fields make plain. Only Philadelphia's record names the gap between the event and its notice in the case's own text, and only that record names a bearer — civilian patient reports — while the codings agreed on the gap field in 76 of 89 cases and on the bearer field in 70 of 89; among the twenty-three cases read closely they read a named gap in 13 and a named bearer in 5. The three European cases name who was reached — the sick refused entry in Florence, arrivals from places of plague in Ragusa, every able-bodied person under sixty without land or craft in the Ordinance — but none names two dated moments in its own text.

The comparison shows what each record cannot carry. Whether any restriction worked is not answered: three of the four carry no count in this study's own field, Ragusa's count is a vote of an assembly present, and the codings agreed on the count field in 78 of 89 cases while reading the values among the four as no-count for three and qualified by scope for one. The weekly reports carry state counts on a basis their own first issue calls incomplete; the Pennsylvania entry of 18 October gives 6,081 deaths from influenza and 2,651 deaths from pneumonia up to 14 October and adds that the greater number occurred in Philadelphia, which is a state total with a city named inside it, not an outcome of any measure taken in that city. What any decider weighed is shown by none of the four records; the Ordinance records that the king held deliberation with the prelates and nobles and other learned men, but gives no account of what was weighed. Who decided is shown by name for two of the four — the Great Council in Ragusa's record, the king by counsel in the Ordinance — while Philadelphia's record names no authority and Florence's names only officials appointed for the purpose. The ledger's field for a decider named in the record reads 0 of 89; across the eighty-nine the coders agreed on the decider cell in 53 of 89 cases, on the horizon cell in 32 of 89, on the reversal cells in 38 of 89 and 36 of 89, on the place cell in 37 of 89, and on this study's own revision-in-case field in 67 of 89. Where a comparison here rests on a cell the two codings divided on, it rests on a reading they did not settle; 7.3 shows which cells those are. None of the four records carries an observable condition for revisiting the judgment itself: the ledger counts 0 of the eighty-nine as carrying one, while 89 of the eighty-nine list what is missing, 264 items in all; 7.4 sets out what each record lists in its own Unknown field.

The comparison does not decide what the shape means. It could be a feature of restrictions in these records, or a feature of the records themselves: these are history records, not decision records, and the cells that would settle the question are the ones the coders read least consistently. It cannot show what any later reader should have preferred; none of the acts is offered here as right or wrong. The hypothesis that most acts were taken before the cause they addressed, or their own effect, was established is judged in Chapter 12, where the strict counts leave it unsettled and the lenient counts read it as supported; likewise for the hypothesis about who bore the wait. This chapter compares four acts in a study of twenty-three cases drawn from eighty-nine, and what that selection represents belongs to those verdicts.

What the comparison does establish is a reading order: a restriction's record can be asked, before anything else, what was known at the hour of the act, who bore the wait while the act was taken or not yet in force, and whether the record itself names the interval between the event and its notice. Three of these four records answer the first question and leave the third unnamed; the fourth names all three, in a form reconstructed eighty-nine years later. Declarations come next: acts whose timing is a dated act about a process already under way, where the interval between the event and the record is the subject of the act itself. Chapter 8 reads the World Health Organization's Constitution, Minamata's first notice, the declarations that close on smallpox and polio, and the characterisation of 11 March 2020.