On the Persistence of a War in the Census

The southern war has lately been credited with eighteen thousand four hundred casualties by the Royal Census Office, seventy-five thousand eight hundred by the Dwarven Bureau of Households at Keld, and three hundred and forty-four thousand by the Elven Registry of Avarren. Each office examined the same northern territories; each possessed substantially the same returns; each printed its answer beneath a heading sufficiently similar to deceive a minister. I was asked which figure deserved adoption. Having examined the calculations, I recommended the withdrawal of all three headings. This advice was received with the reserve commonly accorded to an inexpensive remedy which would require several important men to admit that they had misunderstood their own tables.

The difficulty begins with the word loss. A person killed during the fighting is a loss. A person who would probably remain alive today, had the fighting spared him, represents a surviving deficiency in his generation. A child never conceived because a prospective parent died contributes to a different deficiency, whose existence must be inferred. These quantities require separate accounts. The offices had measured different combinations of them, across the interval between the peace of 1758 and the census of 1868, and then allowed a printer to confer upon their disagreement the appearance of a historical controversy. Their arithmetic was mostly competent; their explanation of what they had counted was deplorable.

The Royal Census Office obtained its eighteen thousand four hundred by comparing the present population with a projection of the population expected without the war. Its calculation allowed for subsequent migration, though with a confidence which the frontier registers scarcely merited. In the Human population, recovery had produced fifty-seven thousand four hundred persons above the projected number. Against this excess stood deficiencies of fifty-one thousand two hundred Dwarves and twenty-four thousand six hundred Elves. Subtraction yielded the published figure. Thus an additional Human railway porter at Northbridge could cancel, upon the official sheet, an Elven woman killed before his grandfather was born. When I put this consequence to the superintendent, he replied that persons were the unit of population. I agreed, and asked whether he would accept payment of a missing locomotive in the same number of teaspoons, articles being the unit of inventory.

His calculation nevertheless answered a legitimate question. A government arranging grain imports or waterworks may wish to know how many people inhabit its territory compared with an earlier expectation. For that purpose the descendants of newcomers require bread and drains quite as urgently as anyone else. The error entered when the office described the difference as the casualties of a war. Over the intervening century, Human births had occupied abandoned streets, migration had restored deserted holdings, and industrial employment had drawn families into districts whose mills had once stood roofless. Meanwhile, almost every Human who would have survived the war was now dead from some subsequent cause. The original wound had become difficult to locate in the living population. Nothing in this process altered the number buried.

I initially suspected that immigration explained nearly the whole Human recovery. The surviving municipal totals appeared to support me, and I expressed this judgement with an assurance which made its correction particularly agreeable to my colleague Mrs Ferriby. She had followed households through the rural registers instead of remaining, as I had, among the excellent chairs of the capital. Her linked returns showed substantial recovery among resident families before the great railway migration began. Earlier marriage in districts with vacant holdings contributed; so did improved survival among children after the reconstruction of the water supply. The precise division remains uncertain. My assertion that the recovery was chiefly imported was wrong, and Mrs Ferriby has earned the right to mention this whenever I become impatient with her footnotes.

At Keld, the clerks followed generations. They asked how many members of the populations exposed to the war should still have been living in 1868, then compared that expectation with the survivors actually recorded. Their seventy-five thousand eight hundred consisted almost entirely of the Dwarven and Elven deficiencies already mentioned. Human survivors from the exposed generations were too few to sustain a useful estimate. The Bureau therefore preserved a visible absence which the Royal Office had partly cancelled against later Human abundance. Its schedules were admirably constructed, although the director’s claim that they represented the dead required correction. They represented people whose deaths still reduced the expected number of living members of those generations.

A Dwarven birth ledger explains the distinction with very little assistance from philosophy. Among those who died during the war were infants, young adults and people approaching the end of their ordinary span. By 1868, many of the last group would have died in any event. Their wartime deaths remained historical facts, while their contribution to the present deficiency had disappeared. Conversely, many who had been young should still have been alive. The surviving generation consequently retained a broad depression in its numbers. Later births could increase the population beneath that depression without supplying the absent people within it. The Keld director understood this perfectly when discussing households. I found it necessary to insist that he retain the understanding when composing his preface.

The Elven absence was clearer and, in one respect, more difficult to measure. A century removed relatively few expected survivors from the younger generations, so the interruption persisted across successive registers. Yet Elven households displaced during the invasion could remain abroad for decades without regarding their departure as permanent. An empty dwelling offered poor evidence of a death. At Avarren I refused to admit forty-seven names from one district until enquiries had been sent to the western ports. Eleven answers arrived from the supposed dead themselves. One correspondent supplied a lengthy objection to my terminology and a shorter confirmation of her continued existence. Both were entered in the record, although only the latter improved the estimate.

Avarren’s larger figure came from carrying the reconstruction backwards. Its clerks sought the deaths attributable to the war when those deaths occurred, including people who would no longer be alive today. They combined burial accounts and military records with household disappearances, then estimated how many deaths would ordinarily have occurred during the same years. I rejected their first allowance for unregistered rural deaths because it treated the destruction of a register as proof that everyone entered in it had perished. After revision, the central estimate for the three northern kingdoms was approximately two hundred and forty-two thousand Human deaths, seventy-one thousand Dwarven deaths and thirty-one thousand Elven deaths above ordinary mortality: three hundred and forty-four thousand altogether. The surviving evidence supports a range of roughly three hundred and twenty-six to three hundred and sixty-three thousand. It does not support the last three zeroes as a claim to exact enumeration.

These are deaths attributable to the war, including the excess mortality of hunger and epidemic disease. The returns cannot always distinguish a fever contracted in a crowded refuge from one that would have occurred at home. They can show that mortality rose sharply where stores were destroyed and families driven together, and that comparable districts beyond the armies’ reach suffered less. Nor have I included hypothetical descendants among the dead. Such descendants matter greatly to later population estimates, but assigning them a number requires assumptions about unions and reproductive decisions which become increasingly fragile with every generation. An Elven woman who postponed a birth for eighty years has confounded an annual forecast; she has supplied no evidence that the child will never be born.

The pre-war population of the northern kingdoms was approximately four million eight hundred thousand. Our central estimate therefore represents a little over seven deaths above ordinary mortality for every hundred inhabitants during the conflict. That proportion understates the local devastation. Losses clustered along the campaigning roads and in the grain districts from which successive armies requisitioned their food. In the Lower Grelth, reconstructed excess mortality approaches one person in five. I walked through villages whose present prosperity had been offered to me as evidence against these estimates. Their substantial new houses established that someone had built houses afterwards. The village accounts, including payments for the removal of bodies from wells, addressed the earlier question rather more directly.

The same arithmetic helps explain why the fighting ended when it did. Sudermark had committed approximately one hundred and forty-six thousand troops to the northern campaigns, counting successive drafts without counting transfers twice. At the withdrawal, its surviving muster accounts assigned about fifty-seven thousand to death, thirty-eight thousand to captivity or continuing incapacity, and twenty-four thousand to indispensable garrisons and communications duties. This left twenty-seven thousand available for field operations. Some assignments are doubtful, especially where a missing soldier was eventually declared dead, but no plausible correction restores the army with which the thanedom had begun. These southern military losses are additional to the northern population losses discussed above; the southern civilian record remains too incomplete for a comparable estimate.

The commander who assured his council that fresh levies could repair the deficiency had also requested exemptions for mine pumps, harvest labour and the railway works serving his depots. I have read both documents. Taken together, they constitute an unusually effective criticism of their author. Men transferred from those occupations could enlarge a muster while reducing the food, fuel and transport available to the men already serving. Replacement equipment also arrived more slowly as skilled labour diminished. Dwarven apprentices recruited into the infantry could be given muskets within a month; the workshops from which they had been removed could not manufacture experienced foremen on that schedule. Exhaustion appeared in the accounts as a succession of worsening exchanges, each nominal addition to the army imposing a greater cost upon its supply.

The northern kingdoms had little strength left to exploit the retreat. Their combined establishments still listed one hundred and twelve thousand soldiers, but only thirty-nine thousand were reported fit and present. Of these, approximately eighteen thousand were required for depot protection, escorts and the security of the recovering districts. Twenty-one thousand remained for sustained operations. I therefore dispute the familiar assertion that the northern commanders declined pursuit through an excess of generosity. Their correspondence shows concern for food and draught teams, and their returns explain it. Sudermark had been driven back with its military capacity exhausted; its opponents had secured an opportunity to survive and rebuild. The subsequent unification of Calder gave that recovery a political form which the casualty columns alone could never explain.

I have recommended that future publications print historical excess deaths separately from deficiencies among surviving generations and from estimates of descendants foregone. The Human gap has largely closed in the present population; the Dwarven and Elven gaps remain open because many of the absent people should still be here. Neither circumstance measures the value of the lives lost, though both help explain the different persistence of the war in households and public memory. At Keld, a registrar showed me the place where his mother’s name should have appeared in the current household return. He required no assistance from me in understanding the entry. I confined my intervention to the printed instructions, which had contrived to make his perfectly intelligible absence appear statistically obscure.