Snake Bite - Pt 7
Wednesday, December 7, 2005, 03:59 PM - Counterfactuals and Contingency

The "passive" Gen. Joseph E. Johnston


Although I wrote my last post in this series back in August, through the miracle of hyperlinks I can resume the story without much back-tracking. Basically I'm trying to see what light counterfactual theory can shed on one of the bread-and-butter issues of doing operational military history--namely the assessment of battlefield decision-making. I'm doing this by way of a review of Union and Confederate decision-making during the initial stages of the 1864 Atlanta campaign.

So far I've dealt with the Union side of the hill. Now I want to look at things from the Confederate perspective. The key actor here is Gen. Joseph E. Johnston, commander of the Army of Tennessee.

Johnston had taken charge of the army in December 1863 after its defeat at Chattanooga a month earlier. The beaten and demoralized force was then at Dalton, Georgia, about twenty-five miles southeast of Chattanooga. Johnston's guidance from Confederate president Jefferson Davis called upon him to reorganize the army and then undertake an offensive aimed at recovering middle Tennessee. Johnston thought--and most analysts
agree--that Davis's expectations were unrealistic: the Army of Tennessee was outnumbered more than two-to-one by the enemy forces opposing it. Even so, they had the political effect of obliging Johnston to hold his position at Dalton.

That was unfortunate, because in purely military terms Dalton was not the best place to make a stand. True, it was shielded from direct attack by Rocky Face Ridge immediately to the west. But the ridge terminated just three miles to the north, allowing the Federals easy access to a wide valley where they could deploy their superior numbers and force a decisive battle on terms favorable to themselves. Alternatively, the Federals could use the ridge to shield their movements while they maneuvered around Johnston's left flank.

In his 1874 memoirs, Johnston said that would have preferred to make his initial stand at Calhoun, a small town eighteen miles south of Dalton, and that he remained at Dalton only because of "the earnestness with which the President and Secretary of War, in their letters of instructions, wrote of early assumption of offensive operations and apprehension of the bad effect of a retrograde movement upon the spirit of the Southern people."

Consequently, when the Atlanta campaign began, Johnston had three-and-a-half things to worry about, and I've tried to sketch them on this map. In order of seriousness, they can be listed thus:



Worry No. 1 was the chance that Sherman might seek an immediate fight to the finish on the plain north of Dalton.

Worry No. 1.5, so to speak, was the chance that Sherman might seek to pierce the Rocky Face Ridge gap and lunge directly at Dalton. Sherman was unlikely to choose this option--the excellent defensive ground offered Johnston his best chance to inflict a crippling defeat on the Federals--but it might become attractive to Sherman if Johnston did not hold the ridge in sufficient strength.

Worry No. 2 was the chance that Sherman might seize Rome, Georgia (which is what Sherman initially planned to do). With Rome in Union hands, Johnston could not easily receive reinforcements from Alabama and Mississippi, and from Rome an advance of just thirteen miles would allow the Federals to cut Johnston's supply line at Kingston. The Confederates at Dalton would be over twice that distance from Kingston.

Worry No. 3 was the chance that Sherman might slip through the mountains at some point between Rome and Rocky Face Ridge. This of course is what Sherman wound up doing when he sent Maj. Gen. James B. McPherson through Snake Creek Gap.

Johnston has been criticized, legitimately, for leaving the gap unguarded and for persistently interpreting intelligence reports of McPherson's early marches to mean that McPherson was en route to Rome. Nevertheless, except for Worry No. 1.5, Worry No. 3 was the least of Johnston's concerns--the least dangerous thing the enemy could do. And as matters turned out, Johnston evaded the trap Sherman planned to spring on him. He recognized the threat in time, shifted his army to meet it, fought a solid defensive battle at Resaca on May 13-14, and then withdrew safely farther south.

Even so, Johnston's critics consider Snake Creek Gap an irretrievable disaster for the Confederates. A counterfactual analysis of the event asks us to look for the main historical variable used to explain the disaster and then to come up with a "minimum rewrite" that modifies or does away with that variable. In this case, the key variable turns out to be Johnston's character.

Here's Richard M. McMurry on Snake Creek Gap: "In retrospect, we can see that, in all likelihood, [McPherson's] seizure of the gap determined the outcome of the campaign. After May 8 Johnston could not long remain in his Dalton works unless he detached troops to regain Snake Creek Gap. Given the passive way he was determined to conduct his operations, he could not undo the damage." (McMurry, Atlanta 1864 [2000], 63-64; emphasis supplied)

James M. McPherson in Battle Cry of Freedom (1988), 744:
Unlike Lee, whom necessity compelled to adopt a defensive strategy, Johnston by temperament preferred the defensive. He seemed to share with . . . [George B.] McClellan . .. a reluctance to commit troops to all-out combat. . . . This unwillingess to fight until everything was just right may have been rooted in Johnston’s character. A wartime story made the rounds about an antebellum visit by Johnston to a plantation for duck hunting. Though he had a reputation as a crack shot, he never pulled the trigger. "The bird flew too high or too low—the dogs were too far or too near—things never did suit exactly. He was . . . afraid to miss and risk his fine reputation." (Emphasis supplied)

Stephen Davis in Atlanta Will Fall: Sherman, Joe Johnston, and the Yankee Heavy Battalions (2001), ix:
Throughout the campaign Sherman possessed such numerical strength, such sharp strategic thinking, such confidence, logistical mastery, and aggressive determination, that the combination of these advantages against Johnston’s passivity combined to determine the outcome of the campaign. (Emphasis supplied)

Read in this way, it all sounds very plausible. But the virtue of counterfactual analysis is that it puts pressure on causational claims and forces them to work for a living. McMurry, McPherson, and Davis (among others) are in essence claiming that each person has a stable personality, character, or temperament--I'll use character for convenience's sake--and that this thing is such a reliable predictor of a person's decision-making that you needn't look further. Speaking as someone with bipolar disorder, and who therefore has an intimate grasp of how biochemistry can influence moods and how moods can influence decision-making, I wonder about this. Do I not have a stable personality, character, or temperament? I think that most people who know me would affirm that despite my illness I do in fact retain a distinctive character. At the same time, plainly I do make different choices depending on mood. (A different day, a different mood, and I might not be writing this, for instance.)

Beyond question, some historical figures of major consequence have had disorders comparable to my own. A recent book on Abraham Lincoln, for instance, amplifies on the well-known fact that Lincoln throughout his life experienced extensive bouts of depression. But when dealing with war, an historian doesn't need evidence of bioaffective disorder to give heed to the possibility that the moods of a key actor may influence decision-making. Physical and mental stress, illness, fear, and anxiety are fundamental components of the inner world of war. War, indeed, can be seen as an environment is which human beings are constantly and deliberately acted upon to produce extremes of mood. Troops on the attack are whipped up into states of manic excitement; the attacks, in turn, have as their aim the creation of such anxiety in the enemy that the enemy will break, resulting ideally in a state of mass depression and demoralization. Good commanders do their best to shield themselves from the stresses of war, but it is naive to suppose that they always succeed. And because of this, I am intensely skeptical of the "Johnston's innate passivity" explanation of Snake Creek Gap. The basic premise on which the explanation rests is flawed.

In this instance, counterfactual analysis has the effect of causing me not to posit a minimum rewrite of history, but rather to reject existing interpretations as fundamentally misguided.

Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7
4 comments 4 comments ( 94 views )   |  permalink

Snake Bite - Pt 6
Tuesday, August 30, 2005, 05:17 AM - Counterfactuals and Contingency

Situation, May 5-May 9. I've indicated roughly the positions of the Union main body (in black) and Confederate main body (in gray). Solid arrows show McPherson's approach march to Snake Creek Gap. Ignore lines extending south from Resaca; they depict a later stage of operations.

The Atlanta campaign begins on May 5. On that day McPherson receives the following orders from Sherman, reproduced here in their entirety:

SIR: The enemy still lies about Dalton, and from all appearances is on the defensive, guarding approaches mostly from the north and west. He occupies in some force the range of hills known as the Tunnel Hill. By tomorrow night our forces will be about as follows: Schofield at Red Clay; Thomas at Ringgold--his left, Catoosa Springs, center at Ringgold, and right (Hooker) near Wood's Station; and you at Gordon's Mill. If you are all ready I propose on Saturday morning [May 6] to move against the enemy--Thomas directly on Tunnel Hill; Schofield to Varnell's and the gap between it and Catoosa Springs, feeling toward Thomas; Hooker will move through Nickajack Gap on Trickum and threaten the road which runs from Buzzard Roost to Snake [Creek] Gap. As these are in progress I want you to move, via Rock Spring, Tavern Road, to the head of Middle Chickamauga; then to Villanow; then to Snake [Creek] Gap, secure it and from it make a bold attack on the enemy's flank or his railroad at any point between Tilton and Resaca. I am in hopes that Garrard's cavalry will be at Villanow as soon as you, for, you know, I have sent General Corse to meet him at Shellmound and conduct him across the mountain to La Fayette and to you. But, in any event, his movement will cover your right rear and enable you to leave all incumbrances either at Ship's Gap or VillanoW, as you deem best. I hope the enemy will fight at Dalton, in which case he can have no force there that can interfere with you. But, should his policy be to fall back along his railroad, you will hit him in flank. Do not fail in that event to make the most of the opportunity by the most vigorous attack possible, as it may save us what we have most reason to apprehend--a slow pursuit, in which he gains strength as we lose it. In either event you may be sure the forces north of you will prevent his turning on you alone. In the event of hearing the sound of heavy battle about Dalton, the greater necessity for your rapid movement on the railroad. It once broken to an extent that would take them days to repair, you can withdraw to Snake [Creek] Gap and come to us or await the development according to your judgment or information you may received. I want to put this plan in operation, beginning with Saturday morning if possible. The sooner the better for us. (OR 38/4:39-40)


(Parenthetically, one might compare the above to the modern "five paragraph field order".)

The assets McPherson had for the operation consisted of five infantry divisions with artillery (about 24,500 men). A single cavalry division attached to his army was still far to the west, near the Georgia-Alabama state line, and took no part in the coming operation. On May 8 Sherman detached a brigade from a second cavalry division to assist McPherson, but only its commamder, Brig. Gen. Judson Kilpatrick, showed up on May 9. The rest of the brigade arrived next day.

At 2 p.m. on May 8, McPherson informed Sherman that his troops were entering Snake Creek Gap.

Dawn on May 9: His advance guard, the 66th Illinois, passes out the southern end of the gap and begins reconnoitering toward Resaca.

Mid-morning: Two of McPherson's divisions (Thomas W. Sweeny, James C. Veatch, under command of Grenville M. Dodge) emerge from the gap; they also advance toward the east, but are are now moving cross-country and the underbrush in this sparsely-settled region slows them down.

Late morning: McPherson's remaining three divisions (under John A. Logan) emerge, less two brigades left to provide security at the northern end of Snake Creek Gap.

12:30 p.m.: McPherson himself has reached a crossroads about 5 miles east of the gap, and 2 miles from Resaca. From here he writes Sherman, "I propose to cut the railroad, if possible, and then fall back and take a strong position near the gorge on this [side] of the mountain and await your orders.” (OR 38/4:105)

2 p.m. - Dodge's two divisions are now at the crossroads in force. Ahead there is a bald hill held by an estimated 1,400 Confederates. McPherson instructs Dodge to sweep the hill with one division but to place the other facing north to block any potential Confederate advance from the direction of Dalton.

Dodge does so. From the crest of the Bald Hill he can see Resaca, the objective point, less than a mile away. But it is fortified and contains what seems to be a considerable garrison.

4 p.m. - Logan's three divisions are now at the crossroads. McPherson goes to the crest of the hill.

Essentially the choice facing him is whether or not to attack using Logan's divisions. At this time of year, sunset is at about 7 p.m. Sufficient daylight to conduct operations last until about 8 p.m. Night attacks in the Civil War are rarely undertaken, though night marches sometimes are.

Soon after this, McPherson gives the order for his entire force--all five divisions--to retire to the mouth of Snake Creek Gap.

At 10:30 p.m. McPherson updates Sherman on his situation and explains that he did not cut the railroad at Resaca:

The enemy have a strong position at Resaca naturally, and, as far as we could see, have it pretty well fortified. They displayed considerable force, and opened on us with artillery. After skirmishing till nearly dark, or getting to it, I decided to withdraw the command and take up a position
for the night between Sugar Valley and the entrance to the gap for the following reasons: First. Between this point and Resaca there are a half dozen good roads leading north toward Dalton down which a column of the enemy could march, making our advanced position a very exposed one. Second. General Dodge's men are all out of provisions, and some regiments have had nothing to-day. His wagon train is between here and Villanow, and possibly some of them are coming through the gap now, but they could not have reached him near Resaca; besides, I did not with to block up the road with a train. It is very narrow, and the country on either side is heavily wooded. I had no cavalry except Phillips' mounted men [a very small force] to feel out on the flanks. If I could have had a division of good cavalry I could have broken the railroad at some point.


Nor does this exhaust the list of McPherson's misgivings. To give just one more example, he has not heard all day from Joseph Hooker's Twentieth Corps, which is supposed to be covering Dug Gap to the north. In the absence of clear word from Hooker, he can't be sure the gap is covered and that a large force of Confederates cannot sweep through Dug Gap, seize the north end of Snake Creek Gap, and cut off McPherson's column.

The question at this point, from a historian's standpoint, is what to make of this. The usual tack is to judge McPherson's decision to pull back into Snake Creek Gap--or to speculate upon what would have occurred if Thomas's original plan had been followed. In the months and years immediately following the Civil War this was inevitable. But what purpose it serves nearly a century and a half later is less apparent to me. It seems like just another instance of what John Keegan calls "the accusatory approach."

If one were to adopt an "inquisitory approach" instead, what alternative questions might be framed?

Lastly, from the standpoint of counterfactuals and contingency, have we reached a level at which chance developments operate so overwhelmingly that extended historical analysis is no longer appropriate. That one may narrate what occurred on the Union side at Snake Creek Gap, but to assign interpretive weight to any particular factor misleads more than it illuminates.


Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7
12 comments 12 comments ( 175 views )   |  permalink

Snake Bite - Pt 5
Friday, August 26, 2005, 06:32 AM - Counterfactuals and Contingency

Maj. Gen. James B. McPherson, USA; Gen. Joseph E. Johnston, CSA


We come to the part I don't really like, when it comes to doing military history. That's making judgments about people under the pressures of actual combat. There's a reason we have phrases like "armchair general." We all know at some level that there is something tasteless, even mildly offensive about second-guessing field commanders. It's not just that we know more than they do--we have extensive if not complete information about an operation, while they made life-and-death decisions on the basis of missing or mistaken intelligence and conflicting reports. It's not just that we know exactly how things turn out, and can therefore backtrack to reach optimum solutions. It's also that we cannot, in the nature of things, readily think ourselves into the mental and emotional states of commanders--nor their physical condition, which often exerts at least some influence over the choices they make.

Nevertheless, if we military historians are not simply to chronicle military operations, we have to exert an evaluative function. But what sort of evaluations are we to make, and by what criteria?

Years ago John Keegan noted that "most British military historians . . . implicitly out someone in the dock—a general or an army, charge him or it with a crime—defeat if a friend, victory if an enemy—and marshal the evidence to show his or her responsibility.” (Face of Battle, 74). He added that American military historians seem to do the same. He thought we ought to quit it--to set aside the "accusatory" approach for what he termed the "inquisitorial" approach," True, his specific examples of the inquisitorial approach seem to turn away from the business of evaluating commanders. And true, in many of his books since then he has resorted periodically to the accusatory approach. Nevertheless, it's good advice.

What I want to do here, then, is to try to understand the decisions taken with regard to Snake Creek Gap, not to criticize them. That has been done ad nauseum by literally dozens of other commentators and historians ever since May 9, 1864, the day that Maj. Gen. James B. McPherson marched his army through the gap and then back in again.

Why does Snake Creek Gap matter? Two reasons. First, it succeeded in compelling the Confederate army to retreat, quite early in the Atlanta campaign, from a position that they reasonably expected to hold for much longer than they did. That set in motion a series of subsequent retreats that brought Sherman half the distance to Atlanta in the opening weeks of the campaign. If he had been held farther away for a longer period of time, he might not have captured Atlanta in time to affect the 1864 presidential election. If the Atlanta campaign matters at all, you have to look at Snake Creek Gap.

But second, it did not succeed to the degree that it could have. Many have argued that bolder action on the part of the Federal commander at Snake Creek Gap would have created conditions in which the Confederate army could have been "broken up" if not actually destroyed at the campaign's outset. In that event one cannot say what would have transpired next, but it is hard to see a seeming military stalemate developing in 1864. Indeed, it is not unreasonable to imagine the Union winning the war about six months sooner than it did. Maybe more.

The method I'm going to use will be to evaluate the choices made by the two key actors in the Snake Creek Gap affair: Union Maj. Gen. James B. McPherson, commander of the Army of the Tennessee, which comprised the right wing of Sherman's "army of invasion;" and Confederate Gen. Joseph E. Johnston, commander of the Army of Tennessee. (Not that confusing similarity of names: the former army was named for the Tennessee river, the latter for the state of Tennessee.) Like much of Civil War military history, this will be yet another version of a hoary old tale. But by a judicious use of counterfactual theory, by placing the tale in the larger context of the history of warfare, and above all by showing a little basic compassion for men operating under tough circumstances, I hope I can shed at least a little bit of light.

Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7

Did Atlanta Matter?
Wednesday, August 24, 2005, 07:47 PM - Counterfactuals and Contingency


Only the most devoted readers of this blog will recall that I'm under contract to write a book in the Oxford University Press Pivotal Moments in American History series, edited by David Hackett Fischer and James M. McPherson. McPherson published the inaugural volume in the series--Antietam: Crossroads of Freedom--in 2002. Fischer's contribution to the series, Washington's Crossing, came out last year. Six other books in the series have also appeared, with at least one more in press.

My "pivotal moment" deals with 1864. It was a presidential election year, so the most obvious pivot is the question of Lincoln's reelection. Usually this is framed in terms of whether the Democratic nominee, George B. McClellan, might have defeated him in November, and whether that, in turn, might have led to a compromise peace or even Confederate independence. But given that Lincoln faced no fewer than three challenges from within his own party, one must also think about the chances of his being replaced--by Salmon P. Chase, John C. Frémont, or some other candidate. Believe it or not, the name of Benjamin F. Butler was bruited about more than once. And Lincoln quietly but carefully sniffed out Grant for possible presidential aspirations before appointing him general in chief.

Eighteen sixty-four was also the year in which white Americans, North and South, began to come seriously to grips with a change in the racial status quo. African Americans had a significant, albeit informal, influence over this shift, most obviously because Blacks were becoming increasingly important as a reservoir of military manpower.

In the realm of counterfactuals and contingency, Lincoln's reelection is widely thought to have hinged on the perception of Union military success.

“If the election had been held in August 1864 rather than November, Lincoln would have lost. . . . This did not happen, but only because of events on the battlefield—principally Sherman’s capture of Atlanta, and Sheridan’s spectacular victories over Jubal Early in the Shenandoah Valley. These turned northern opinion from deepest // despair in the summer to confident determination by November.” (James M. McPherson, “American Victory, American Defeat,” in Gabor S. Boritt (ed.), The Collapse of the Confederacy (New York: Oxford University Press, 1992), 39-40.

"There was nothing inevitable about northern victory in the Civil War. Nor was Sherman’s capture of Atlanta any more inevitable than, say, McClellan’s capture of Richmond in June 1862 had been. . . .” (Ibid., 41)

Albert Castel agrees, and greatly amplifies this thesis in

Albert Castel, “The Atlanta Campaign and the Election of 1864: How the South Almost Won By Not Losing,” in Castel, Winning and Losing in the Civil War (Columbia: University of South Carolina Press, 1996), 15-32.

However, William W. Freehling--though in agreement that Atlanta “was the Confederacy’s last best hope to escape strangulation,” thinks that it was nonetheless a forlorn hope, that the point of no return had been reached in 1863. And he argues that certain structural factors--e.g., superior Northern military and industrial strength and internal stresses within the Confederacy--make Confederate victory unlikely in any event. As for Lincoln's reelection being dependent on a timely military triumph, and Union victory being dependent on Lincoln's reelection:

“[F]or military historians to be declared right that Sherman’s victory alone could have saved Lincoln’s victory, or that Lincoln’s victory alone could have saved Union victory, political historians must be proved dead wrong about antebellum politics in general and the Democratic Party in particular.” [William W. Freehling, “The Divided South, the Causes of Confederate Defeat, and the Reintegration of Narrative History,” The Reintegration of American History: Slavery and the Civil War (New York: Oxford University Press, 1994), 226-227] The former because 90 percent of 19th century American voters remained loyal to party, the latter because Peace Democrats were a minority within that party.

See also Freehling's The South vs. the South: How Anti-Confederate Southerners Shaped the Course of the Civil War (New York: Oxford University Press, 2001), esp. pp. 177-199.

William C. Davis shares Freehling's skepticism in “The Turning Point That Wasn’t: The Confederates and the Election of 1864,” The Cause Lost: Myths and Realities of the Confederacy (Lawrence: University Press of Kansas, 1996), 127-147. Davis does concede, grudgingly, that Confederate military success in 1864 could have unseated Lincoln, but ups the ante by implying it would have taken more than Atlanta.

“In the end the only Confederate acts that could have—not necessarily would have—affected the outcome were those in the one theater in which the war was being decided from the outset: the battlefield. If Jubal Early had captured Washington and held it for some appreciable time. If Sterling Price had wrested Missouri from the Union and been able to hold it. If the forts at Mobile had been able to repulse Farragut and his fleet. If Lee had been able to take some action against Grant, however small, to embarrass him in the trenches at Petersburg. And most important of all, if Joseph E. Johnston or John Bell Hood had been able to turn Sherman decisively, not just away from Atlanta, but back on his base at Chattanooga. If all these ‘ifs’ had come to pass, they would have constituted a series of body blows to Union morale and Lincoln’s prestige, at the rate of one every few weeks during the last four months of the election campaign. Then quite possibly, even probably, sagging Northern spirits would have translated into Democratic votes.” (137)

This comes fairly close to the famous Saturday Night Life sketch that asked, "What if Eleanor Roosevelt could fly?" Even if the Confederates ran the tables, it would have resulted only in a McClellan victory, and Davis argues that McClellan would have continued the war and would have inherited a military position in which he could hardly have failed to win it.

Larry J. Daniel concurs with Davis and systematically critiques Albert Castel's essay in “The South Almost Won By Not Losing: A Rebuttal,” North and South Magazine vol. 1, no. 3 (February 1998):44-48, 50-51. (BTW, I'm grateful to Eric Wittenberg for the loan of this article, which I was finding hard to locate.)

Most recently we have:

Richard M. McMurry, “The Atlanta Campaign and the Election of 1864,” appendix four of Atlanta 1864: Last Chance for the Confederacy (Lincoln: University of Nebraska Press, 2000), 204-208. He writes:

“[F]orays into counterfactual history can be instructive. They often help us get a better understanding of the past by forcing us to examine roads not taken and the reasons why they were not. Such exercises, however, lose validity as they become more and more complex. They must keep within the bounds of the possible. It helps if we limit them to possibilities that were probable.

“To apply such counterfactual speculations to the Atlanta campaign and the 1864 election, we have to work our way successively through a maze of at least a dozen counterfactual scenarios.”

Which he does on pp. 206-207, and concludes:

“In arguing that Lincoln had to have military success (or perceived success) in 1864 to win reelection, Castel was correct. I believe, however, that success came late on May 8 at Snake Creek Gap, not at Atlanta on September 2. Given the passive way [Confederate Gen. Joseph E.] Johnston was determined to conduct his campaign, loss of that meant that the Rebels could not—or at least really would not attempt to—halt Sherman’s advance into Georgia.” (207)

It's that last contention that justifies the focus on Snake Creek Gap in the Snake Bite series of posts. My purpose, however, is neither to introduce new revelations about this operation nor to argue that America's future necessarily hinged on what occurred here. It's to better understand the role of counterfactuals and contingency in historical interpretation--and figure out how to explain this to the readers of my OUP book.

Snake Bite - Pt 4
Tuesday, August 23, 2005, 06:54 PM - Counterfactuals and Contingency

Capt. William E. Merrill

The blog has received a spike in hits today, which led me to poke around to see from whence they came. Aside from the usual--I get a number of visits each day from people searching for images of "yin and yang", for instance--I see quite a few from members of Civil War West, a moderated Yahoo discussion group.

The instigator, so to speak, was a member who posted this morning: "Everybody’s favorite topic is being covered in Mark Grimsley’s blog" and gave the link. This generated thunderous silence, so after a while the member wrote: "Hmmm…I really thought this would have stimulated some discussion on this group." Which did provoke a smattering of response, mostly along these lines:

What new information does Grimsley present that you feel would have stimulated discussion? Other than some nice visuals, like the satellite image, his articles seemed like the same-old same-old to me.


And indeed, if you search the group's archives for "Snake Creek Gap," you do find a number of hits--although frustratingly, the Yahoo search engine goes back only a few days at a time before calling it quits and forcing you to resend the query.

One recent respondent noted that I need to flesh out the counterfactual. He suggests that Grant might have appointed Maj. Gen. James B. McPherson over Thomas. Which is something I discuss in a comment to Part 1, but that's an easy thing to overlook.

What I have not seen on "Civil War West," nor on what I could find in searches of other Civil War groups, is a discussion that focuses on staff work rather than commanders. I do not claim credit for any great originality here. But it does seem to me that, at least in terms of knowledge of the existence of Snake Creek Gap among the high command, staff work is a key factor.

For example, Federal cavalry during the Atlanta campaign carried this map, lithographed on muslin so that it could be stowed in saddlebags:



You can click on it for a larger image, but it's a big file and if you're using dial-up, you might go make a sandwich while it loads. The key detail is here anyway:



It's still not that easy to make out, because I was limited by the resolution of my source image, but Snake Creek Gap is clearly visible, accurately located and named.

The guy responsible for the map was Captain William E. Merrill. He is mentioned just twice in the Union reports of the Atlanta campaign (OR 38:1, p. 128; OR 38:3, p. 64). He is not found at all in the index to Albert Castel's Lincoln Prize-winning Decision in the West, the standard study of the campaign and generally considered one of the best Civil War campaign studies, period.

Merrill's low rank and virtual invisibility tells a lot about the importance--or rather lack of it--attached to good staff work during the Civil War. He graduated at the top of his West Point class in 1859 but served throughout the war as an engineer, not a line officer. He supervised creation of the maps that helped Maj. Gen. William S. Rosecrans win the campaign of maneuver which resulted in the bloodless capture of Chattanooga, Tennessee, in September 1863. When Rosecrans got the boot after Chickamauga, Merrill went to work for George H. Thomas. Earl B. McElfresh, a student of Civil War mapping, notes that the pair "were a perfect match because Thomas believed wholeheartedly in mapping and engineering. . . . The Army of the Cumberland set the standard for mapping of all Sherman's forces." (McElfresh, Maps and Mapmakers of the Civil War (New York: Henry N. Abrams, 1999), 244). Merrill ended the war as--drum roll!-- a brevet colonel.

Even so, the Union army plainly had staff work capable of gleaning intelligence about the topography of north Georgia, collating it into a reasonably accurate map, and above all mass-producing and distributing that map widely. By contrast, I know very little about the map found among the papers of that Confederate brigade commander (see previous entry), but the fact that nobody noticed its existence for over a century suggests to me that it was not mass produced or distributed. Indeed, one would need to know much more about just how widely the map was circulated. If by some chance the Confederate brigade commander had it made for himself, it's entirely possible it never made it up the chain of command.

On few subjects connected with the Civil War is it possible to say that nothing has been done. But there are quite a few that are significantly understudied, and staff work is one of them. Offhand I can think of one book on staff and headquarters operations in the Army of Northern Virginia and a handful on military intelligence, the medical services, etc. That's about it. One thing's for sure, however. Compared to their European parts, Civil War army commanders did way too much work that properly ought to have been delegated. This is a structural shortcoming, not a matter of "negligent" generals. And the fact that Union staff work was incrementally better than Confederate staff work may have been a significant factor in shaping Union victory. It certainly bears consideration with regard to Snake Creek Gap.

Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7
21 comments 21 comments ( 139 views )   |  permalink

Snake Bite - Pt 3
Tuesday, August 23, 2005, 06:51 AM - Counterfactuals and Contingency
Below is an image and a link to an image showing Snake Creek Gap, the route Thomas wanted to use to get at Johnston's army. Thomas's scouts had discovered the gap after some minor skirmishes in February. They got all the way through the gap to Resaca and discovered that it was unguarded by rebel troops. That seems astounding, but the Federals kept an eye on the gap and it continued to look unoccupied.


Satellite photo of Snake Creek Gap; image created using Google Earth. (Click to enlarge)

Link to TopoZone Map centered on Snake Creek Gap

Why was this? We used to believe that the Confederates may not have been aware of the gap, but in recent years a map has surfaced that was carried by a Confederate brigade commander during the campaign and which does show the gap. Stephen Davis, in his study of the Atlanta campaign, assumes that Johnston must therefore have known of the gap and so was "negligent." That is fairly typical in Civil War military history--and indeed, all military history. It's what I call "what fools they were" history. If a serious mistake is made (and in war they happen all the time), you just say the officer responsible was negligent or an idiot or sulking or some other bad adjective, and you don't have to perform any further analysis.

Personally I think Johnston wanted very much to hold his position at Dalton for as long as possible, blocking every known gap was indispensable to that end, and thus if Snake Creek Gap remained unguarded it would be worthwhile to look elsewhere for an explanation. My guess is that Confederate staff work was inadequate. Civil War staffs tended to be much smaller than their European counterparts, staff officers had no specific training for the job, and without a really good system a lot of details were bound to be lost. I'd want to check out that hypothesis before I started talking about somebody's negligence.

Incidentally, in no area does military history needs to be blogged or flogged or dogged out of the Stone Age--even on its own traditionalist terms--than in the realm of adequate maps. Operational history is fundamentally concerned with time-space relationships, and the more fluid the action, the more you need maps that convey the action and do not resemble a maze of squiggles. When we were restricted to paper, a dearth of maps could be excused (or not: Douglas Southall Freeman had hundreds of maps in his books, simple but serviceable). Nowadays you can create as many maps as you want and post them online. You can create Flash presentations that animate those maps. You can do all sorts of things that would make it a lot easier for readers (or viewers) to follow the action.

And there's certainly lots of material on the web to help, as the images above attest.

Here's a cropped (and slightly modified) version of a map used at West Point. It may serve as a useful supplement, though it shows the Snake Creek Gap maneuver as actually executed. But switch "McPherson" and "Thomas" and you'll have the picture of Thomas's proposal.

Sherman, to repeat, nixed Thomas's proposal. But when he discovered that McPherson's army would be short 10,000 men during the coming campaign--thanks to the unfortunate Red River campaign in north Louisiana, to which the troops had been "loaned"--he scrapped his own original plan and ultimately decided to send McPherson through the Snake River Gap. But there were two important differences. McPherson had only about 40 percent of Thomas's numerical strength, and Sherman's orders instructed McPherson simply to cut the railroad at Resaca and then withdraw back to the southern end of the gap.



Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7

Snake Bite - Pt 2
Tuesday, August 23, 2005, 05:36 AM - Counterfactuals and Contingency
On March 10, while Grant was in Washington to receive his promotion to lieutenant general and command of all Union armies, Sherman penned an adoring and somewhat anxious letter to his newly elevated friend:

Now as the future. Do not stay in Washington. Halleck is better qualified than you are to stand the buffets of intrigue and policy. Come out West; take to yourself the whole Mississippi Valley; let us make it dead-sure, and I tell you the Atlantic slope and pacific shores will follow its destiny assure as the limbs of a tree live or die with the main trunk! . . . For God's sake and your country's sake, come out of Washington! . . . Here lies the seat of the coming empirel and from the West, when our task is done, we will make short work of Charleston and Richmond, and the impoverished coast of the Atlantic.(Sherman to Grant, March 10, 1864, Simpson and Berlin (eds.), Sherman's Civil War: Selected Correspondence of William T. Sherman, 1860-1865, pp. 603-604.

Sherman obviously believed that Georgia, not Virginia, should be the main focus of the Union spring offensive. He wrote his wife Ellen:
Grant in command--Thomas the Center [John] Schofield the Left & Sherman the right, if we can't whip [Confederate general] Joe Johnston we will know the reason why. (WTS to Ellen, March 10, 1864, ibid., 605)

This sounds as if Sherman expected that Grant would wear two hats: as general in chief and commander of the Military Division of the Mississippi. McClellan in 1862 had done something like that--combining the post of general in chief with personal command of the Army of the Potomac. Grant, of course, rejected this idea. Historians usually highlight two reasons: first, he believed that the Army of the Potomac was unduly influenced by politics and that only his personal presence and authority could prevent this; second, he believed Lee's army in Virginia was the most important military objective. Would the disappearance of Sherman have caused him to reverse himself and, ironically, do as Sherman had urged? Not, I think, as long as Grant maintained his opinion that Lee's army was the first great object and that the potential for Washington to interfere in this matter had to be neutralized.

That leaves Thomas operating independently in the Mississippi valley. As he did with Sherman, however, Grant would have asked to know Thomas's plans for the coming offensive. We don't have to guess at this plan. We know its outlines because Thomas pitched it to Sherman in March 1864.

Henceforth the discussion requires some knowledge of the armies involved in the Atlanta Campaign and the way the campaign unfolded. Fortunately we have available the nicely done The Atlanta Campaign: An Album, which combines text, maps, and images (particularly present-day photographs) in a nice package. If you get confused or want further detail, try using that.

The Confederate Army of Tennessee under Gen. Joseph E. Johnston had taken up a strong defensive position on Rocky Face Ridge just west of Dalton, Georgia. Nobody wanted to attack this position directly. (Sherman, for one, called it "the terrible door of death.") It was obvious to everyone--Federals and Confederates--that the position would have to be "enveloped"--which means struck on its more vulnerable side; or "turned," which means a movement aimed at threatening the enemy's line of communications so severely that the enemy must abandon an otherwise good position. (Just as an aside: military historians who bemoan the obscure language of postmodern and new cultural historians ought to think about how our lingo sounds to outside ears.)

Here is Sherman's original plan. It's a turning movement, not an envelopment:



Essentially, Sherman would use most of his force--the Army of the Cumberland (Thomas) and the Army of the Ohio (Schofield) to "fix"--or pin--Johnston in place while the Army of the Tennessee (James B. McPherson) marched from its camps near Huntsville, Alabama, and seized Rome, Georgia. You'll see that a railroad runs east from Rome to Kingston, where it meets the Western & Atlantic Railroad. The W&A was critical to Johnston. It was the main conduit for supplies and reinforcements and his best way to keep in touch with Richmond and the rest of the South was by way of the telegraph lines that paralleled the tracks. If McPherson could get to Rome, he would have an easy shot at Kingston just thirteen miles to the east. Johnston would have to retreat, and Sherman would have gained.

Here is the plan Thomas vainly urged upon Sherman:



Once again, some forces pin Johnston in place--only this time it's the Army of the Tennessee and the Army of the Ohio that receive the job. The Army of the Cumberland will march via Snake Creek Gap to Resaca, also on the Western & Atlantic Railroad but much closer to Johnston's army, which means Johnston has a reduced chance of retreating unscathed. Thomas, in short, imagines an envelopment followed by a decisive battle. But wait--wouldn't the Confederates have foreseen and blocked such a dangerous move?

Part 1 - Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7

Snake Bite - Pt 1
Monday, August 22, 2005, 07:22 PM - Counterfactuals and Contingency

William T. Sherman; George H. Thomas

When Grant was appointed general in chief in March 1864, his first act was to name a successor to his own previous job as commander of the Military Division of the Mississippi, a zone of responsibility stretching from the Appalachian Mountains to the Mississippi River. (See map.) That successor, of course, was Sherman, the subordinate Grant trusted most. But if we erased Grant and/or Sherman, what difference would that make? Were Grant and Sherman indispensable to Northern victory in 1865, as Civil War historians often imply, or did the North have a pool of senior generals who could have achieved equivalent results? Just how important was the military leadership of specific individuals compared with other variables, so-called "second-order" variables that would tend to push events back down the paths they took historically: superior Union military strength and manufacturing capacity, internal stresses within the Confederacy, and so on?

In keeping with the "minimum rewrite" rule of counterfactuals, let's remove only Sherman from the war. We'll assume that he suffered crippling grief over the death of his son Willy, though he could just as easily have contracted a fatal case of pneumonia, typhoid fever, or any of a dozen illnesses. (He did in fact have significant respiratory problems in January 1864 and did not fully recover until April). His most likely replacement would have been Maj. Gen. George H. Thomas. Grant, to be sure, is well known to have had quiet misgivings about Thomas. But few others did. By March 1864 Thomas had acquired the reputation of never having lost a battle. It was his Army of the Cumberland that had played the principal role in winning the battle of Chattanooga and it was Thomas's firm determination that had earlier saved that army (then under command of Maj. Gen. William S. Rosecrans) from complete disaster at the battle of Chickamauga. Thomas was also senior to anyone else Grant would have wanted even remotely for the job. (He was senior to Sherman, for that matter.) And although historian Richard McMurry is correct to point out that Thomas, a native Virginian, had few political patrons in Washington, he did have the full support of Tennessee's military governor, Andrew Johnson, who told Lincoln about this time that Thomas would be "one of the great generals of the war, if not the greatest." (OR 32:3, p. 105) Preferring Sherman to Thomas was one thing. Utterly snubbing Thomas would have been quite another.

We could stop our counterfactual exercise right here by noting that Thomas, historically, continued to be one of the war's most successful generals. Most military historians rate him at or near the same level as Grant, Sherman, and Philip H. Sheridan. Some consider him the best of the lot. But Thomas and Sherman were not interchangeable parts. Can we make any reasoned guesses about what Thomas would have done in Sherman's place, and do these guesses suggest any interesting questions?

But first we have to reflect on the unhappy position in which Grant would have found himself with his most trusted lieutenant incapacitated or extinct.

Grant would almost certainly have given Thomas the same mission that historically he gave to Sherman:

You I propose to move against Johnston's army, to break it up, and to get into the interior of the enemy's country as far as you can, inflicting all the damage you can against their war resources. (OR 32:3, p. 246)

But whether he would have followed those lines with these is another matter:

I do not propose to lay down for you a plan of campaign, but simply to lay down the work it is desirable to have done, and leave you free to execute it in your own way.

The loss of Sherman would have given Grant one of three basic choices. He could indeed have given Thomas the same operational discretion as Sherman. He could have given Thomas more detailed instructions, as he often did with subordinates (unlike his predecessor Maj. Gen. Henry W. Halleck, who believed on principle that field commanders were in the best position to judge what ought to be done). Or he could have scrapped his plan to travel with the Army of the Potomac and returned west to oversee Thomas the way he did historically with the Potomac army's commander, Maj. Gen. George G. Meade. Simply considering which option Grant would most likely have pursued obliges us to revisit Grant's command style, his set of priorities, and the importance to him of his partnership with Sherman.

Part 1- Part 2 - Part 3 - Part 4 - Part 5 - Part 6 - Part 7
7 comments 7 comments ( 113 views )   |  permalink

Erasing Grant and/or Sherman
Saturday, August 20, 2005, 07:41 AM - Counterfactuals and Contingency

Ulysses S. Grant; William T. Sherman.

On March 22, 1864, Lt. Gen. Ulysses S. Grant sat for his portrait in the Washington studio of Mathew Brady. Standing to one side was none other than Secretary of War Edwin M. Stanton, who had personally brought Grant. Stanton understood the value of giving the country a good look at its new top commander.

Seeking to take advantage of the mid-afternoon sunlight, Brady sent an assistant to the roof to uncover a skylight. The assistant took a bad step; the skylight shattered, raining shards of glass all around the Union hero. "It was a miracle that some of the pieces didn’t strike him," Brady said. "And if it had, it would have been the end of Grant; for that glass was two inches thick.” Stanton told Brady not to mention Grant's brush with death to anyone. He feared it would spark rumors of an assassination attempt.

Brady told the story to play up Grant's imperturbality--"the most remarkable display of nerve I ever witnessed.” Novelist Shelby Foote and historian Brooks D. Simpson used it in their work for the same purpose, and rightly so. But the incident also shows how easily a key historical actor can be removed from the scene. So let's assume that the "miracle" didn't occur, and that Grant died under a hail of glass plate just eleven days after assuming command of all Union armies. What can this "rewrite" of history tell us about the final year of the Civil War?

This kind of history is called counterfactual history, and historians and social scientists are finding it a useful thought experiment, because it enables them to think in fresh ways about existing historical explanations. The counterfactual scenario can involve anything--economic historian Robert Fogel once wrote a classic essay on American industrialization by asking what would have happened if mid-19th century America had continued to use rivers and canals, not railroads. It can be plausible, like Grant's mishap in Brady's studio, or utterly implausible, like his sudden abduction by aliens. Either approach, properly handled, can help us interrogate existing interpretations of the past.

"Many explanations in History and the social sciences," writes Geoffrey Hawthorn, "turn not on causal connections between states of affairs that are beyond human control, but on the relevant agents' own practical reasonings." A range of strategic choices were available to Union armies in 1864 and because Grant held the top post, he made the choices. But if choice has any meaning, it implies a range of roads not taken as well as those that were. Would another Union general have chosen the same roads as Grant? Would he have had the same capacity as Grant to direct the armies in the face of the many obstacles and frustrations that beset the Union war effort in 1864? How many roads led to victory? In short, how indispensable was Grant?

One can go further. We know that Maj. Gen. William T. Sherman was Grant's key subordinate, and most Civil War historians lay stress on the importance of their command partnership. We also know--or think we do--that Sherman's capture of Atlanta in September 1864 was crucial to Lincoln's reelection two months later. What if Sherman, too, left the scene in early 1864?

This is more of a stretch, because Sherman suffered no brushes with death during the winter of 1864. But he did lose his nine-year old son Willy to typhoid fever in October 1863, and his grief was profound. Historian Michael Fellman, the closest student of Sherman's personality, writes:

[Sherman] mourned the death of his blessed son deeply and at extraordinary length. Indeed, throughout the year to follow, during the campaign against Atlanta and on down to the sea, his long letters to Ellen and hers to him would be filled with grieving for Willy. Far more than Sherman’s complex and deadly military situation, far more than his bouts of self-inflation, which never ceased, and his eager seizing of victories, Willy would remain the obsessive subject of his correspondence with his wife. (Citizen Sherman: A Life of William Tecumseh Sherman, 200)

Sherman weathered the melancholy. But we know that he suffered from depression several times during his life, that these depressions appear to have been associated with a mood disorder, and that one of these--in the autumn of 1861--put him on inactive status for several months. During that period Sherman is known to have had suicidal thoughts. If we postulate that his grief at Willy's loss triggered an episode of clinical depression that resulted in his removal from command or even his suicide, we have removed a second key historical actor.


The Burnet House hotel, ca. 1910

Or we can remove Grant and Sherman another way, by assassinating them in the Burnet House hotel in Cincinnati on March 20, 1864, as the two of them held the famous meeting in which they discussed strategy for the coming spring offensive. The Civil War is full of such seemingly unlikely incidents actually occurring. In October 1862 one aggrieved Union general shot another to death in the Galt House hotel at Louisville. In November 1864, Southern sympathizers set fire to several buildings in a bid to burn New York City. And in April 1865 the John Wilkes Booth conspirators succeeded in killing Lincoln and seriously wounding his secretary of state.

What I find interesting about this exercise is that it forces me to think about the depth of Union military talent at that stage of the war. Without Grant and Sherman, how many generals capable of winning the war remained?

Beyond Beer and Pretzels - Pt 2
Friday, August 19, 2005, 06:39 AM - Counterfactuals and Contingency
I've finished reading Ransom's The Confederate States of America: What Might Have Been. On the whole I think it's quite responsibly done--far more rigorous and thoughtful than what a group of historians, relaxing after a day at some conference, might come up with over beer and pretzels--which is often the stereotype of counterfactual history. As I read I jotted a few brief notes, which I transcribe herewith:

Ransom: “The fact that they are significant historical events is only one reason that wars are popular subjects for counterfactual speculations. Warfare also invites the consideration of alternative outcomes because it presents the historian with a relatively short window of analysis and a limited number of counterfactual outcomes that can be considered.” (6) Which is why military historians especially should become adept at the potential and pitfalls of the counterfactual approach.

“What is extremely difficult to construct is a meaningful narrative of counterfactual history beyond the first imaginary turning points. How do we keep the counterfactual story from exploding into a myriad of different tales, all going in different directions? One answer is the use of what I have called the historical mold to shape the initial burst of possibilities to fit the world around them. That is what economic historians do best; they provide theoretical paths along which our counterfactual story must travel.” (13)

Ransom thinks that the Union victories of 1862 in the west and 1863 at Gettysburg, Vicksburg, and Chattanooga were decisive. “By this time the Confederates had no more answers: Their cause was indeed lost. They put up a stout resistance for another fifteen months, but the eventual outcome of the war was no longer in doubt.” (109)

One might well ask, to whom was it no longer in doubt. Certainly not the Confederate government--and certainly not to many in the North. And anyway, Ransom is thinking here only of the "consequent" of Confederate independence. Even conceding that by December 1863 there was no doubt that the Confederate cause would ultimately fail, it still matters how long it took to fail. An earlier-than-historical Union victory in the spring of 1864 would produce a different postwar environment than the actual victory in spring 1865, while a later-than-usual Union victory in, say, the summer of 1866 would produce a still different postwar environment. The timing matters, if nothing else, because of its implications for emancipation. Generally speaking, the longer the war continued, the more Northerners became willing to consider more advanced measures in the realm of emancipation and the genesis of free labor. Suppose, for instance, the war lasted long enough for the measures contemplated under Special Field Order No. 15 to become the mainstream Union military policy toward freedpeople?

Ransom, ch. 3 (pp. 121-160), gives an extended counterfactual of the Civil War that departs from the historical timeline at Fort Donelson, which falls to Union forces in February 1862 but only after its 13,000 defenders break out to freedom. It is unnecessary and would spoil the fun to rehearse his counterfactual unfolding of the conflict. But basically he follows the “minimal rewrite” rule as faithfully as possible, with numerous second order variables pushing events back toward the historical path, though not enough to create a Union victory. “If the story seems strangely familiar," Ransom says, "that is because I have made a conscious effort to place the counterfactual changes outlined earlier into the historical mold of the war as it actually took place. Thus the time period for the duration of the war is roughly the same: spring 1861 to spring 1865. Most of the battles occur in roughly the same places at roughly the same times as [they did historically]. Up until the last year of the of the counterfactual war many of them have the same outcomes. The result, as we shall see, is a war that looks familiar in many ways but has a very different ending.” (122)

It takes Ransom 160 pages to get to the postwar period. He then devotes just 85 pages to the development of a counterfactual Confederacy, deftly using his professional training as an economic historian to do so. For all practical purposes he halts on the eve of the First World War. At that point, Ransom says, "we have pushed our counterfactual recipe as far as it can go; the historical mold has become so speculative that it can no longer handle our counterfactual pudding. No matter what mix of ingredients we come up, the result will be too heavily weighted with imagination." (240)

Additional details re the book are here. (Hat tip to Dimitri Rotov at Civil War Bookshelf).

Part 1 - Part 2


Back Next