Abstract
The preceding paper proposed combat participation rate — the share of a mustered force that actually made contact — as a diagnostic tool for identifying battles settled by pre-battle negotiation rather than tactical decision. At Bosworth and Sekigahara, a large fraction of the losing army stood idle, and that idleness was purchased. This paper argues that the metric, taken alone, is close to useless, because it measures an outcome that many unrelated causes produce. Two American Civil War battles fought three weeks apart in the autumn of 1862 make the point with unusual clarity. At Perryville, roughly two-thirds of a Union army stood within a few miles of a general engagement and never entered it — not because any part of it had been bought, but because its commander could not hear the battle. At Antietam, a Union army committed the majority of its strength and still failed, because commitment was sequential rather than concurrent. Neither case involves a single defector. Both produce participation figures comparable to or worse than Sekigahara’s. The paper therefore proposes a decomposition of non-participation into seven causes, a set of companion metrics, and an exclusion procedure that must be run before low participation can be read as evidence of negotiated absence.
1. The Defect in the Instrument
A rate is not an explanation. The prior paper’s argument required that low participation at Bosworth and Sekigahara be read as the visible trace of prior contracts, and the reading was correct for those cases — but it was correct because of features of those cases that the metric itself does not capture. Kikkawa Hiroie’s men were within a mile of the fighting, on a road, in daylight, fed, in formation, and fully aware that a battle was underway. Nothing prevented them from participating except a decision. Strip away any one of those conditions and the same number means something entirely different.
Two cases from 1862 demonstrate how different. They are unusually well suited to the test: same year, same national army culture, same weapons, comparable staff practice, three weeks apart. Whatever varies between them is not technology or period.
2. Perryville: The Army That Could Not Hear Its Own Battle
On 8 October 1862, Don Carlos Buell’s Army of the Ohio, something on the order of fifty-five to sixty thousand men, converged on Perryville, Kentucky, in three columns. Braxton Bragg attacked with roughly sixteen thousand. Bragg attacked because he believed the bulk of Buell’s army was elsewhere — the Union demonstration toward Frankfort, where Bragg had just presided over the installation of a Confederate provisional governor of Kentucky, had persuaded him that Perryville held only a detachment.
He was wrong by a factor of three, and it did not matter, because Buell did not fight the battle either.
Buell’s headquarters lay roughly two and a half miles from the fighting. He had been thrown from his horse the previous day and was in pain and largely immobile. And for reasons of wind direction, terrain, and air layering, he did not hear the engagement. Neither, in part, did Thomas Crittenden’s corps. Buell learned that a general action was underway only in the late afternoon, when the fighting had been going on for hours and Alexander McCook’s corps had been badly handled. Crittenden’s entire corps — some twenty thousand men — never fired a shot. Most of Charles Gilbert’s corps did not either, apart from Philip Sheridan’s and Robert Mitchell’s divisions on the Springfield Pike. Perhaps twenty-two thousand Union troops engaged out of nearly sixty thousand present: a participation rate near or below forty percent, worse than the Western Army’s at Sekigahara.
The acoustic shadow at Perryville is the best-documented instance of a phenomenon that recurs across the war and remains under-theorized as a cause of non-participation rather than as a curiosity. Iuka, fought nineteen days earlier, is the near-perfect companion case: Grant and Edward Ord, positioned north of the town with something like eight thousand men, did not hear William Rosecrans’s engagement to the south and consequently took no part in a battle they had been positioned to join. Two battles, three weeks apart, in which tens of thousands of men were rendered combat-irrelevant by atmospheric conditions.
The battle’s other conditions compound the picture. Kentucky was in severe drought; the fighting began over control of water holes in Doctor’s Creek, and the approach march of both armies had been shaped throughout by the location of standing water rather than by operational preference. Buell’s three corps had come up separate roads and lacked working lateral communication. Gilbert held corps command on an irregular appointment that his own subordinates did not fully credit.
Perryville’s participation rate is therefore genuinely low and genuinely meaningless as evidence of anything the prior framework was built to detect. There was no defection market in Kentucky in October 1862. There was an injured commander, an atmospheric anomaly, a fragmented road net, and a drought.
It is worth adding the mirror-image result. Bragg’s own participation rate at Perryville approached totality — he put in essentially everything he had at Perryville and won the tactical fight, driving McCook’s corps back a mile or more. He then withdrew and abandoned Kentucky. A commander who engages his whole force is a commander with nothing in hand, and the profile is not a mark of decisiveness but of thinness. Meanwhile Bragg’s theater participation rate was dismal: Kirby Smith’s command, drawn off toward Frankfort by the Union feint, was not present at all. Which denominator one selects — men on the field, men in the theater, men in the campaign — changes the number and the verdict.
3. Antietam: High Participation, Low Concurrency
Antietam, 17 September 1862, presents the opposite defect. George McClellan brought roughly seventy-five thousand men against Lee’s force of perhaps thirty-eight thousand at dawn, rising toward forty-five thousand as A. P. Hill’s division came up from Harpers Ferry late in the afternoon after a march of some seventeen miles. The day produced approximately twenty-two thousand seven hundred casualties, the bloodiest single day in American history. Roughly two-thirds to three-quarters of the Union army engaged. By the crude metric, this was a well-committed army.
It was also a defeat in everything but the ledger, and the reason is that the commitment was serial. Hooker’s First Corps attacked at dawn through the Cornfield. Mansfield’s Twelfth followed after Hooker was spent. Sumner’s Second went in later still, and Sedgwick’s division was wrecked in the West Woods by a counterattack that could only have been assembled because no other Union attack was pressing at that moment. Burnside’s Ninth Corps did not force the Rohrbach Bridge and the downstream ford until the early afternoon, and its advance on the Confederate right was broken by A. P. Hill’s arrival — a division that would have been irrelevant against a simultaneous assault but was decisive against a late and isolated one.
Lee held the field with half the numbers because he was permitted to fight four battles in succession rather than one battle at once. Hood, McLaws, Walker, and finally Hill were shifted from sector to sector across a compact interior line. Every one of those transfers presupposed quiet somewhere else.
Layered onto this is the deliberate reserve. Fitz John Porter’s Fifth Corps was held in the center and never committed. Franklin’s Sixth Corps was only lightly engaged; Franklin proposed an attack in the afternoon and was overruled with McClellan’s concurrence. The withholding was doctrine, not failure of nerve alone — a commander is supposed to hold a reserve — but McClellan never spent it, and the reason lies in intelligence rather than tactics. Pinkerton’s estimates had convinced him he faced something on the order of a hundred thousand men. A commander who believes himself outnumbered two to one will hold a reserve against a counterstroke that is not coming. Low participation here is the downstream product of a bad number, not of a bad temperament, and certainly not of anyone’s purchase.
Lee’s own participation problem was worse and occurred before the battle. The Army of Northern Virginia had bled men continuously during the Maryland campaign to straggling, barefoot exhaustion, short rations, and in some measure to a reluctance among men who had enlisted to defend their own states to campaign north of the Potomac. Estimates of the loss vary widely and should be handled with care, but the army that fought on the seventeenth was drastically smaller than the army that had crossed the river. Those men do not appear in any participation calculation, because they were not present to be counted. The metric’s denominator silently absorbs the largest category of non-participation in the entire campaign.
4. A Taxonomy of Non-Participation
Set the four armies of the two battles beside Bosworth and Sekigahara and the same output number turns out to have at least seven distinct inputs.
Contracted. The contingent was bought, either to abstain or to switch. Kikkawa Hiroie, the Stanleys, Kobayakawa. Diagnostic of political disintegration on the losing side.
Uninformed. The contingent did not know a battle was occurring, or did not know it was general. Crittenden at Perryville, Ord at Iuka. Diagnostic of nothing about loyalty; diagnostic of the fragility of pre-telegraphic battlefield awareness and of atmospheric conditions treated as causal rather than anecdotal.
Unreachable. The contingent knew and tried, and distance or road capacity prevented arrival in the window that mattered. A. P. Hill’s division is the inverse case — it arrived just barely in time and thereby saved a flank; Humphreys’s division of Porter’s corps arrived the following day and mattered not at all.
Unfrontable. There was no room. Antietam’s compartmented ground, the single bridge over the creek, the narrow corridors of the Cornfield and the Sunken Road. Frontage imposes a ceiling on how many men can be in contact regardless of intent, and armies routinely bring more men to a field than the field can absorb at once.
Reserved. Withheld on purpose. Porter’s Fifth Corps. Diagnostic of the commander’s estimate of enemy strength and his theory of risk, not of his subordinates’ reliability.
Dissolved. Lost before contact to straggling, sickness, desertion, or exhaustion. The Army of Northern Virginia in Maryland. This category is invisible in any calculation whose denominator is force present on the field, which is precisely why the denominator must be declared.
Misdirected. Sent elsewhere by a faulty appreciation of where the battle would be. Kirby Smith at Frankfort, drawn off by a demonstration; and, from the other direction, Bragg’s decision to attack at all rested on the same misreading.
Only the first of these tells us anything about the political cohesion of an army. The remaining six are, in varying proportions, statements about information, terrain, logistics, doctrine, and weather.
5. Companion Metrics
If participation rate is to survive as a tool, it needs support.
Peak concurrent engagement. The largest share of a force in contact at any single moment. Antietam’s aggregate figure is respectable; its concurrency figure is dreadful, and the concurrency figure is the one that explains the result. This is the single most useful addition, because it captures the failure mode — piecemeal commitment — that aggregate participation actively conceals.
Declared denominator. Civil War returns distinguish aggregate present, present for duty, and present for duty equipped; Confederate returns are considerably less consistent. Premodern musters are worse. Any comparative work must state whether it counts men on the field, men in the theater, or men on the rolls, because the three produce different battles.
Fire delivered against fire received. A regiment that stood four hours under artillery fire without discharging a weapon took casualties and absorbed enemy attention. By a contact-based definition it participated; by a fires-based definition it did not. Reporting casualties-per-unit-present alongside the participation figure separates units that were present and inert from units that were present and expended.
Time-to-commitment. How long after the opening of the engagement each formation entered it. At Antietam this single series would display the entire causal structure of the defeat more clearly than any narrative.
6. An Exclusion Procedure
The consequence for the prior paper is not that its argument fails but that it was underspecified. Before low participation may be read as evidence of negotiated absence, five exclusions must be run:
- Could the contingent perceive the battle? Distance, terrain masking, wind, and air layering are the relevant variables. If not, the case is Perryville, not Sekigahara.
- Could it physically reach the point of decision within the operative window? Road capacity, river crossings, and marching distance.
- Was there frontage available for it? If the line was already full, non-participation carries no information.
- Was it deliberately withheld under a defensible reading of enemy strength? A reserve held against a phantom is an intelligence failure, not a betrayal.
- Was it intact on the morning of the battle? Straggling and sickness must be netted out of the denominator explicitly rather than absorbed silently.
Sekigahara passes all five. Kikkawa’s men could see and hear the battle from Nangū, stood on the road they would have used, faced an unengaged sector with room to deploy, were not held back by any order from Mitsunari, and were intact. That is what makes the contracted reading strong, and it is exactly the demonstration the prior paper assumed rather than performed. Bosworth passes four of five cleanly and the fourth — whether Northumberland was obstructed or abstaining — remains genuinely open, which is the honest result.
7. Two Further Corrections
Participation is not a maximand. The implicit normative reading of a participation figure is that higher is better. Bragg at Perryville disproves it: near-total commitment with no reserve, a tactical success he could not exploit, and a withdrawal that surrendered a state. McClellan’s withholding, whatever its tactical cost, preserved intact the only field army standing between Lee and Washington three weeks after Second Manassas. There is a context-dependent optimum, not a ceiling to be approached.
Non-participation is often the enemy’s achievement. The Frankfort demonstration drew Kirby Smith out of the campaign. Jackson’s operation against Harpers Ferry pulled Confederate divisions away from Sharpsburg and then returned them at the last possible moment. Ieyasu’s letter campaign is the same category of act performed by other means. A low participation figure on one side may be the direct product of skilled work on the other, and reading it purely as an internal defect of the army that displays it mistakes an effect for a symptom.
8. Conclusion
Perryville and Antietam were fought by armies whose soldiers were not for sale, in a war where battlefield defection at the formation level was close to nonexistent, and they produced participation figures at or below those of the two battles the prior paper treated as canonical examples of purchased absence. That result is fatal to the metric as a standalone diagnostic and useful to it as a component of something better.
The corrected proposal is this: participation rate is worth computing everywhere and worth interpreting nowhere on its own. Paired with peak concurrent engagement, a declared denominator, and a run through the exclusion procedure, it becomes a genuine instrument — one that distinguishes an army that was bought from an army that was deafened, an army that was outmaneuvered from an army that simply had nowhere to stand.
The wider point extends beyond the metric. Military history has an extensive apparatus for explaining what troops did and almost none for explaining, in disciplined terms, why troops did nothing. The causes of nothing are numerous, distinguishable, and evidentially tractable — Buell’s silent afternoon, Porter’s unspent corps, Lee’s vanished thousands on the Maryland roads, Kikkawa’s blocked column. Each is a different kind of absence, and until the field has vocabulary that separates them, the single number that lumps them together will keep flattering the historian who reports it.
