Men Against Fire

I read S.L.A. Marshall’s 1947 Men Against Fire after seeing it cited over and over again, sometimes even in conversation. The common takeaway is that most infantrymen in the Second World War didn’t really do much in terms of killing or even shooting at the enemy. Indeed Marshall does claim that, with the qualification that “didn’t shoot” isn’t the same as “worthless”, but that’s not the only point the book makes — more broadly, it’s an attempt at a systematic if ultimately somewhat anecdotal study of infantry combat, something subject even in WWII to far less quantification and contemporary study than combat in the air or at sea.

I had a devil of a time finding this because of some TV show.

After years in print, and particularly after Marshall’s death, analysis of Marshall’s notes and methods strongly suggests that Men Against Fire and other works were far less scientific and systematic than he presented them. The famous “ratio of fire” claim, that 90% of infantrymen in combat never fired their weapons, was made up. More broadly, perhaps Marshall generally had a habit of imposing his own narratives on the information he did collect. So why bother? First, because of the book’s influence. Rather than resorting to second hand narratives (“You know in WWII most soldiers never even shot at the enemy when given the opportunity” “nuh-uh, I heard that was all bullshit”) it’s best to actually read the sources, at least if one cares to form some semblance of expertise on the subject in question. And the cost for doing so in this case is rather low; the book is quite short (I read the entire thing on a slow day in the field). For what it’s worth, I’ve also read the 1988 article debunking the “1-in-10 ratio of fire” claim and it’s easily summed up: “There is no evidence whatsoever he gathered any data to support this claim”.

The other reason is that regardless of whether Marshall pulled numbers out of his ass, I think he’s correct about some non-intuitive or non-obvious aspects of the modern battlefield, the expected performance of groups of men in combat, and the role of leaders.

With Men Against Fire, Marshall describes contemporary infantry combat with a focus on its less intuitive aspects, arguing both that military decision makers and the general public have several fundamentally wrong ideas about what goes on and that the armed forces can improve the performance of their soldiers and leaders by making changes to their training regimen. In some cases, he describes specific case study-type examples; in others, he resorts to abstract descriptions. I noticed that his abstract descriptions tended to dominate his explanations for soldiers not firing his weapons and more generally for his generalized description of the modern infantry battlefield. This was probably the first red flag regarding the quality of Marshall’s evidence. Nevertheless, I do believe he was on to something; charitably, his inability to quantify this may have led to a bit of embellishment on his part. Since this is both the most well-known and most controversial aspect of Men Against Fire, I’ll start with it.

Non-Participating Combatants and the “Ratio of Fire”

In Marshall’s description, the modern battlefield is generally not densely populated, and certainly does not appear to be. The accuracy and lethality of modern weapons means troops must fight in open order and seek cover and concealment at all times. Enemy entrenchments, troop formations, and individual soldiers and positions simply cannot be seen by the advancing infantryman. The “battlefield”, then, consists not of an actual field on which move or wait the enemy soldiers, formations, and equipment which the recruit expects from his training. The visibility of _friendly_ troops and equipment in fact decreases steadily as the soldier moves towards the fighting. The front or battlefield or combat zone is in fact eerily absent of any visible life or movement, and usually quiet. When ordered to advance into this place, the soldier sees only the terrain. When under enemy fire, he can rarely identify its origin. He seeks cover, and looks for targets. When he doesn’t see anything to shoot at…he doesn’t shoot at anything. Even if the enemy is defeated and falls back, this fact is not immediately apparent and doesn’t appear at all related to anything the soldier has done (or hasn’t done). Other times, the attack fails and the soldier’s own unit is forced to fall back, and he makes his way back to more secure cover, again probably without really knowing what’s going on and without having *done* much of anything (this is different than a panic or collapse, which Marshall addresses separately). For a few individuals, this isn’t the case; I’ll return to this subject later.

Here we see the real basis for Marshall’s “ratio of fire” claim — most soldiers didn’t shoot, because they didn’t have anything to shoot at. I immediately noted — as Marshall did — that machine guns typically *did* fire whether they saw the enemy or not, because they were directly ordered to shoot at terrain features as “known, likely, or suspected” enemy positions. The same went even more so for indirect fire (artillery) for which every individual round was ordered to be fired onto a specific location. In most cases these weapons were fired at directly observed enemy only as targets of opportunity and otherwise used their effects for suppression — and note that in spite of this these weapons inflicted most casualties, a fact widely known then and now.

Aftermath of a failed Japanese assault on Guadalcanal

This explanation is a tremendous contradiction to David Grossman’s ideas about the psychological effects of “crew served” weapons (“On Killing”). Marshall does say that some soldiers didn’t fire even at clearly exposed enemies, or reported an unwillingness to kill when interviewed by psychiatrists. This might be true for some soldiers some of the time, but Marshall’s lack of quantification prevents the reader from guessing how much this weighed on the willingness of soldiers to shoot vs their lack of targets. I find the former accounting of this phenomenon far more convincing than Marshall’s tentative (and later, Grossman’s bold) assertion of some universal fear of or aversion to killing. Soldiers in the First World War universally feared to stick their heads above the trench parapet in daylight because of how quickly this earned one a bullet through the skull. And my observation of traumatized soldiers now suggests that victims rather than killers become traumatized as long as those killed weren’t obviously innocent (such as the archetypal and unfortunately quite real “station wagon full of children barrelling through a checkpoint”). Marshall himself undermines the “killing aversion” argument further by claiming that operators of one-man heavy weapons like flamethrowers or rocket launchers didn’t display much aversion to firing. Confusion coupled with fear of death through exposure to enemy fire explains much more about “frozen” or non-participating soldiers than being afraid of shooting at or killing the enemy. Likewise, the effectiveness of “crew served” weapons having more to do with a combination of increased firepower and more effective control systems than “psychological distance” or “diffusion of responsibility” matches my own observation.

Marshall suggests that training must be more realistic in order to correct this deficiency. My understanding is that infantry training was heavily marksmanship-focused at the outbreak of the First World War and, while this focus had lessened somewhat in the 1930s and 1940s, still constituted a major part of instruction. (I think that modern infantry training focuses too little on marksmanship, but that’s another story.) However, soldiers practiced firing at circular, high-contrast targets with high or unlimited exposure time. When they didn’t encounter these sorts of targets in combat, they didn’t fire. Marshall suggests that rifle targets be made to look roughly like the enemy, with limited exposure time. Likewise — while acknowledging that a complete replication of the combat experience was impossible — Marshall suggests more realistic field exercises would give inexperienced soldiers a better idea of what to expect and lessen the shock of encountering combat for the first time. Modern armies have indeed implemented these measures and did see increases in the qualitative effectiveness of their soldiers, even if Marshall exaggerates the unwillingness of the 1940s soldier to engage the enemy.

There is one technological factor that Marshall doesn’t account for, however: the introduction of the assault rifle into general issue. I would be very interested in whether American infantry equipped with the autoloading M1 rifle were more likely to fire and/or advance than soldiers armed with the bolt-action M1903; Marshall doesn’t even attempt to investigate this. Very limited anecdotal evidence from the Japanese in WWII and later the Chinese in the Korean War suggests they were somewhat taken aback by the firepower of small (company down to squad) American units, although it’s difficult to disaggregate the volume and effect of small arms from that of artillery and airstrikes in these usually vague comments. Much better evidence suggests that a generation later ARVN troops armed with M1s and M14s felt outgunned by AKM-wielding enemies. Not only is rate of fire higher, but the soldier carries more ammunition.

Additional factors that might make the assault rifle an enabler of action involve its automatic fire capability. A rifleman can, in a pinch, act more like a machine gun than a “rifle” by switching to automatic fire, though note that while this superficially improves his “firepower” it doesn’t improve his control of the weapon at all either in the colloquial or the cybernetic sense. It is also my opinion that the soldier views, or can be trained to view, the automatic fire capability of his weapon as a sort of bayonet, something that instills in him the belief that he can get up close to the enemy and prevail and so encourages him to advance.

The ammunition load unquestionably makes a difference; WW2 soldiers appear to have carried a minimum of ten pouches of two 5-round stripper clips for bolt action rifles (100 rounds plus loaded magazine for 105 rounds) or ten pouches of one 8-round M1 clip (88 rounds, probably with another few clips to make it about 100 likewise). An American rifleman today carries at least six extra 30-round magazines, giving him 210 rounds with the magazine in his rifle. Ten 20-round magazines seems to have been normal in Vietnam. Of course soldiers in both eras carried more ammo out of personal preference or in expectation of long patrols and intense engagements. Even light irregular fighters now usually have about three 30-round magazines on them. So, the assault rifle and its intermediate cartridge synergizes well with training soldiers to fire in the general direction of the enemy or at suspected and likely positions — he has at least twice as much ammo at his disposal as his forebears.

Not to say the problem was purely technical. Clausewitz commented that troops had to have a certain mindset distinct from the regular line soldiers of his time in order to be effective at fighting in highly dispersed formations, and it is common knowledge in military history that certain social principles substantially increased the use of dispersed skirmishers among line units by the armies of Revolutionary France. On the other hand, another thread of military thought runs through the intensely supervised close-order firing line. The ghost of close-order combat haunted civilized armies for a long time after and sometimes continues to do so today. So, the problem of troops training in close order (note Marshall’s emphasis on the visibility of both friend and foe) and then faltering when dispersed is a real tendency, and something must be done to counteract it. Realistic training including focus on individual and small-team tactics and skills is Marshall’s proposed answer, which the American ground forces adopted.

The Combat Unit

Not everyone in a unit is a non-participant, and most non-participants are far from being cowards. To summarize Marshall, out of 100 men, 80 are just there to occupy frontage and maybe take the occasional opportunity shot, or operate crew-served weapons under comparatively close supervision. Ten soldiers possess substantial initiative and will aggressively contest with the enemy through movement and fire without much prompting; perhaps one of these is a hero capable of habitual feats of incredible and successful daring. On the other hand, ten other soldiers will actively attempt to shirk even the most basic combat duties and consciously resist efforts to move them forward; perhaps one of these is an out-and-out craven. Marshall emphasizes that the middle group are not cowards, and that they serve a real purpose in denying terrain to the enemy who, remember, is in the same situation. They also can be motivated by example somewhat to imitate the go-getters and in so doing amplify the efforts of those more audacious soldiers. Of course, the shirkers could end up winning over the mass as well.

A contemporary (WW2) American infantry squad. 

More aggressive soldiers are of course gambling with their lives and no matter how calculated their risks are usually putting more on the line than the cautious majority. Their high relative casualty rate offers an intriguing explanation, or at least factor, for the oft-remarked phenomenon of “over-seasoned” units — after progressing through their first trial by fire, units increase in proficiency up to a certain point before being “burned out” and noticeably less willing to take even clearly tactically advantageous risks.

Whatever its causes, and whatever the quality of Marshall’s evidence, there exist fundamental reasons to expect something like this occurs. In other contexts — and here we also return to the “ratio of fire” issue — a minority fraction of effort or personnel having a disproportionate effect on some final outcome is normal, a.k.a. the Pareto principle. This survey includes data indicating that a wildly disproportionate share of air-to-air kills were achieved by a small fraction of fighter pilots. I saw the same thing maneuvering tanks at the National Training Center (where simulated kills are tracked by a central system). There’s no reason to suspect infantry combat to be any different.

Marshall says training aides like these gave new recruits a false sense of certainty about enemy dispositions.

Command in Battle

Marshall believes that identifying these soldiers before the trial of combat to be impossible. It is here, he says, that company officers often run into trouble. The platoon leader or company commander naturally expects himself to be the motive force of his unit. Nevertheless, he finds that in combat he is outshone by some junior soldier — a seasoned senior NCO, a daring young sergeant, a battle-savant private — who not only accomplishes more individually but, under intense pressure, suddenly starts making snap decisions — and good ones, correct ones — that the men follow eagerly while the company officer perhaps finds himself personally “pinned down”, “out of the loop”, relaying his observations up to higher headquarters, or at best simply parroting the “orders” of his platoon champion.

After the action, these officers may feel like failures and cowards, and worry that they have lost the respect of their men (a worry that threatens to become a self-fulfilling prophecy whatever the facts). Marshall believes this to be the result of unrealistic expectations and naivete: in his view, the officers have done their duty if they get their men into combat properly disposed, and consolidate & reorganize in the aftermath. As for the heat of battle: since no one can tell who the real heroes are before battle, and the platoon leader is no more or less likely than anyone else to be one of them, he shouldn’t worry himself overmuch.

WW2 platoon headquarters.  Modern platoons have radio operators, not messengers, and there are no “guides”.

I don’t know quite what to think of this. I will say that it doesn’t appear to have been an issue in the most recent wars. One explanation there might be something common in professional armies, the simple and obvious gulf in basic soldiering experience between a junior officer and his NCOs who, unlike in the mass-conscripted armies Marshall describes, have simply been around longer and of whom more might be naturally expected by a new officer in relation to himself. But from other reading, this wasn’t a common feeling probably even in Korea and certainly not in Vietnam (even before American morale fell apart there). A junior officer has enough tasks to keep himself busy — keeping track of and reporting his unit’s disposition being an easy fallback, but also adjacent unit communication and fire support coordination — that even if he actually hasn’t taken appropriate initiative he can likely feel like he accomplished something. “Freezing” or total paralysis is a completely different problem. Perhaps WW2 company officers simply had less to do, or a combination of training and distance from the closed-order paradigm (of the officer four steps ahead on the advance and on the line during engagement) succeeded in changing expectations.

Overall Men Against Fire is worthwhile, if requiring a healthy dose of salt and critical reading. Some of Marshall’s criticisms or descriptions may be nothing more than a now-archaic army working out its kinks. Others may be fabricated, or at least supported by fabrications. Still, the book requires little investment in time and does offer some insights into the experience of open-order infantry combat.