How to Read a Historical Film

Peter Burkholder

     In April 2018, the editorial board of The American Historical Review indicated it would reinstate film reviews after a hiatus of many years.¹ On the one hand, this was a reason for celebration: Who better to evaluate dramatic recreations of the past than history experts? On the other hand, the AHR's decision to resurrect a previous feature serves to underscore the fraught relationship professional historians have had with movies: Should academics embrace this medium as an engagement opportunity with the general public? Should they insist that filmmakers adhere to the same standards as professional historians? Should they disregard filmic recreations of the past, believing such artifacts ultimately do more harm than good? In short, even if many in the historical profession see value in cinema, they may not know quite what to do with it.

     This article takes the stance that, not only should historians engage with the works of Hollywood, but they should recognize that cinema can grapple with real historical issues, bringing insights or filling in gaps that may elude academics. In order to appreciate this, one must be aware of the inherent strengths, limitations, and standards of the film medium. In what follows, I offer a primer on "reading" historical films in a way that recognizes the conventions of both cinematic and textual recreations of past events, thus leading to a more sympathetic and productive understanding of the historical film genre. This should prove useful to world historians for research, teaching, or film reviewing purposes. My own background as a medievalist leads me to favor selections from that area of history, though I will invoke titles from other eras as well. Furthermore, my cases draw from mainstream Hollywood productions, since they tend to follow similar modes of storytelling. Those wishing to delve into foreign cinema and its nuances as a function of national origin are directed to Martha Nochimson's World on Film: An Introduction as a useful starting point.²

A Suspicious Medium

     Imagine the following scene. Queen Cleopatra, her beauty not so remarkable, sails upriver on a splendid barge, its sails purple and oars silver. Boys fan her as she reclines under a magnificent canopy, plumes of incense infusing the air. Marc Antony joins her for dinner, desperately trying to impress her with his knowledge and manners, but Cleopatra bests the Roman at every turn. The queen seems to shape-shift on command, shrewdly becoming the woman a soldier like Antony would desire, a quality she effortlessly employs with others, too. Her charisma, to say nothing of her abilities to converse in a multitude of languages, disarm the Roman general to the point that the evening's would-be conqueror quickly becomes the conquered.

     Is the above paragraph an objective account of a historical event? Hardly. One detects not only sympathy toward Cleopatra balanced by an anti-Antony bias, but a carefully crafted, unexpected inversion, whereby the general becomes the dupe of a foreign power, and a female, no less. Antony's emasculation at the hands of the bewitching Egyptian is hard to miss, most likely a product of film industry gender roles or even anti-feminist backlash. And what of the "beauty not so remarkable," the "purple sails," the "silver oars," the "fan boys" and "plumes of incense"? Pure Hollywood schlock. No wonder we can't rely on movies to learn about the past! Best just to stick with the facts and leave out all the drama and invention.

     Except the account here is not based off of Hollywood at all. The description and details of Antony and Cleopatra's meeting around 40 bce are drawn from Plutarch's Parallel Lives, a text written nearly nineteen hundred years ago, but at least a century after the events it describes here.³ Although an ancient account, historians would not read Plutarch without deploying an array of analytical tools, carefully scrutinizing it at every turn, placing it into context, and being fully aware that the author likely seeks more than to relate facts. Indeed, given Plutarch's distance from the event, and given the reputation Antony and Cleopatra would have enjoyed in the author's early Empire, it is perhaps more productive to read the vignette as a first-century morality piece. To echo the conclusion reached above: No wonder we can't rely on written accounts to learn about the past!⁴

     The key takeaway here is that one can profitably "read" filmic history using a heuristic parallel to that employed when deciphering historical texts. The latter has been extensively studied, most notably by Sam Wineburg, who juxtaposes the reading habits and skills of novices and experts. The apprentice group tends to see a text as a text as a text: primary source, journal article, textbook, it is all just words on a page conveying information that is assumed to be true and accurate. The reader's goal under this line of thinking is to memorize as much information as possible from the text, and then to reproduce it faithfully to demonstrate mastery.

     Experts see things very differently. They understand that authors have biases and agendas, and that such variables as proximity to or distance from a described event can have profound implications. They automatically cross-check texts against other accounts, making determinations about reliability and consistency. Finally, they seek to imagine the setting described, empathizing with the author and participants as much as possible, and they are aware of their own limitations to reconstruct the past. Perhaps most importantly, professionals realize that what a text does is likely more important than what it says. Thus, although both novice and expert may read the same words on the page, their takeaways from the same text are vastly dissimilar.⁵

     Can we bridge the chasm separating these two groups? Certainly, practice and experience come into play, but these things alone are insufficient for most learners. Rather, academics, both historians and others, have embarked on "decoding" the techniques and habits of experts, breaking them down into constituent parts that are often unseen and unappreciated by learners. Even when deconstructed, learning these steps is challenging and unnatural for novices. Yet, the process of decoding offers new initiates the chance of reading the past in ways that are more authentic to the methods of professionals.⁶

The Tendencies of Historical Films—And Texts

     How does this apply to "reading" films depicting the past? Similar to the discussion above, cinematic historians understand that their objects of study are not created in a vacuum, that the goals of popular film producers often diverge from those of scholars, and that movies function according to industry, not academic, conventions. Historians are familiar with and highly skilled in interpreting written text, which offers a useful framework for interpreting film. The operative learning principle, then, is for readers to use their prior knowledge and abilities to better understand a medium they might be less familiar with.⁷

     Fortunately, useful decoding of how cinematic history operates was produced some years ago by historian Robert Toplin, who took a sympathetic approach toward the subject. Keeping Toplin's observed tendencies in mind while watching Hollywood history is akin to teaching our own students how to read historical texts more productively: it is more than just the content and facts that matter. To further reinforce the parallels between written and cinematic storytelling, one can compare Toplin's observations against the medieval French epic poem, The Song of Roland. Written probably in the early twelfth century but loosely based on events hundreds of years earlier (notably, the battle of Roncesvalles in 778), Roland is the type of text often used productively in undergraduate courses, its contours, if not actual content, familiar to world historians.⁸

     Toplin found nine recurring structures or tendencies in the historical film genre.⁹ One might object that some of these structures are generic to virtually any film, not just those depicting the past. While perhaps true, Toplin's observed trends serve to underscore that cinema, like textual narrative, unfolds in fairly predictable patterns of storytelling. The nine trends, with illustrative examples and comparisons with the Song of Roland text, are as follows.

     First, historical film "simplifies evidence and removes many details."¹⁰ There are many reasons for this. Time constraints are one important factor, but equally limiting is the fact that the film medium is poor at conveying large amounts of complex historical detail. Thus, character motives, background information, societal values and norms: such things are necessarily truncated or even omitted so as not to overwhelm audiences. The fact that The Song of Roland would have been read or sung for medieval audiences likewise dictates that it not bog down in too much backstory. Thus, the epic devotes little time to describing the reasons for or logic of Charlemagne's campaign in Spain, while understandably omitting what an audience would consider trivial or irrelevant (e.g., logistical issues, weather conditions, bathroom breaks, although such information would prove valuable to historians). Movies likewise necessarily gloss over or remove complicated material. The difficult issue of slavery in the American colonies is hinted at in Mel Gibson's The Patriot, but it mostly (and problematically) becomes just a plot point to showcase British cruelty.¹¹ Christopher Nolan's Dunkirk suggests that men of the British Expeditionary Force were frustrated with the Royal Air Force, but never explains why the RAF was seemingly absent from the beach evacuation.¹²

     Yet another common attribute of films and texts is their tendency to utilize a three-act approach to storytelling.¹³ That is, the story usually unfolds in the form of exposition, complication, and resolution. Movie producers did not invent this approach, as the great majority of epics through history are structured in the same way. The Song of Roland begins by introducing the reader to the main characters and giving a cursory rationale for their being in Spain. Complication arises from an antagonist's conspiracy to destroy the hero's rearguard as it passes through the Pyrenees, resulting in an extended set piece battle. The story finally resolves with the trial and punishment of the evildoers. Historical films typically employ a similar structure. Steven Spielberg's Empire of the Sun begins by showcasing the bucolic life of an English boy born to a privileged diplomatic family in 1930s Shanghai. The Japanese takeover introduces complication, with the youth becoming separated from his parents and eking out an existence under hostile occupation. Resolution comes with Allied victory and familial reunion.¹⁴ In short, the three-act formula has proven its utility across time and space. Storytellers, be they writers or filmmakers, have artistic license to depart from it, but there are serious risks in doing so.

     Whereas professional historians strive for objectivity, film depictions of the past rarely do. Instead, according to Toplin, cinematic history "offers partisan views of the past, clear[ly] identifying heroes and villains."¹⁵ Such depictions continue a theme of storytelling going back thousands of years. In The Song of Roland, the characterizations are unambiguous: the titular figure and his twelve peers (here, the story evokes Jesus and his disciples) are beyond reproach as they sacrifice themselves for God, king, and country. The Muslims whom they face in battle, meanwhile, are not only ugly and devious, but pagans, polytheists, and idol-worshipers, much in violation of basic Muslim tenets, and conveniently ignoring the Abrahamic roots of both Christianity and Islam. This parallels any number of historical film titles. Marcus Nispel's Pathfinder, a throwback to pre-revisionist portrayals of Vikings as unidimensional sadists, juxtaposes peaceful Native Americans with marauding Norse at the point of First Contact. Drawing on every stereotype in the book, there is never any ambiguity about who the heroes and villains are in this work.¹⁶ Cal Hockley, the foil to the protagonist Jack's dreams in James Cameron's Titanic was, in Ann Hornaday's estimation, such a caricature of evil that "he might as well have twirled a waxy mustache à la Snidely Whiplash throughout."¹⁷

     Closely related to the above theme, filmic history depicts "morally uplifting stories" about contests between "Davids and Goliaths."¹⁸ This element is eminently on display in Roland, as the main character and his hopelessly outnumbered rearguard stave off the attacks of a much larger Muslim army. Although Roland and his compatriots are killed, their deaths are not in vain: Charlemagne wheels his army around, destroys the Muslims, and brings Ganelon, Roland's betrayer, to justice. The medieval-set fantasy trilogy The Lord of the Rings, based on J.R.R. Tolkien's novels, is similarly predicated on an ostensibly forlorn struggle for the race of Men and their allies against the monster armies of Sauron. Self-sacrifice is a running theme in the films, ultimately resulting in an unlikely triumph of Good over Evil.¹⁹ Oppressed individuals or groups are often cast as sympathetic David figures, facing down cruel persons or uncaring institutions. Recent films about slavery, such as 12 Years a Slave, and the civil rights movement, such as Selma, are ready examples of this dynamic.²⁰

     The film medium "simplifies plots by featuring only a few representative characters,"²¹ an element on ample display in Song of Roland as well. Although the number of combatants involved at Roncesvalles goes well into the millions according to the epic, only a few are highlighted in any depth: Roland is the always-loyal, self-sacrificing soldier, the archetype for would-be crusaders; Turpin is the fighting archbishop, a stand-in for a militant Roman Catholic Church; Roland's sidekick, Oliver, serves as the titular figure's conscience, alternately begging Roland to seek, then not seek help from King Charlemagne by blowing a signal horn. (Oliver's "conscience role" is explicit in Luc Besson's The Messenger: The Story of Joan of Arc, where Dustin Hoffman's character, identified in the credits as "Conscience," voices Joan's insecurities.²²) World War II "platoon films" repeatedly feature individuals representing a cross-section of American society and viewpoints. Jeanine Basinger's study of this cinema identifies such standard characters as an Italian, a Jew, a (cynical) Brooklynite, a religious sharpshooter (typically from the South), a Midwesterner, and a new initiate, all of whom are led by a reluctant leader.²³

     Successful storytelling, including historical film, speaks to the present.²⁴ The "present" in the case of The Song of Roland is hazy: the story grew out of oral traditions and earlier written accounts, but achieved its epic form probably in the early twelfth century. A towering issue, then, is the crusades, which had been pitched to Frenchmen in 1095. This context is critical to understanding the greatest historical deviation of the Roland story, the changing of the historical antagonists (Christian Basques) to Muslims. Given this alteration, the epic makes sense as a tribute to and recruitment tool for campaigns to the eastern Mediterranean, with promises of riches, glory, and meaningful deaths to those who heeded the call. In this sense, Roland echoes Pope Urban II's original Crusade appeal, where both earthly and spiritual rewards were part of the bargain.²⁵ Taking a very different stance on crusading is Ridley Scott's Kingdom of Heaven, where the dangers of fanaticism, both past and present, abound. The film's protagonist, the historical Balian of Ibelin, is the sole level-headed character drowning in a sea of senseless bloodletting.²⁶ Given the context of U.S. military intervention in the Middle East, Scott's commentary on current events was unmistakable. The fact that American audiences had not yet soured on the Bush administration's war in Iraq meant that Scott's presentist message fell on many deaf ears, helping lead to box office disappointment.²⁷ And while appeals to "freedom" in Gibson's Braveheart may readily resonate with any American viewer, they were much more powerful to Scottish audiences on the eve of a vote for limited autonomy within the United Kingdom.²⁸

     Stories, both written and cinematic, often feature romance, even if love angles are only tangential to the historical plotline.²⁹ Roland features mere allusions to romance, such as the grief experienced by the titular figure's fiancée and the conversion of the Muslim queen Bramimonde to Christianity (a common topos of Crusade-era literature).³⁰ If ever a superfluous screen romance incensed historians, it was Braveheart's affair between the historical figures William Wallace and Princess Isabella, wife of Edward II. The plot point was seen as particularly egregious, given that Isabella was only a child (and not even in England) at the time of Wallace's execution in 1305.³¹ Yet, it is typical to infuse such amorous relations into historical cinema, and in the case of Braveheart, the invented affair provides for the protagonist's beyond-the-grave revenge against the ruthless King Edward I. François Girard's The Red Violin, a fascinating sweep through time and space, cleverly tweaks the romance angle. Although traditional love interests are on display, it is the characters' passion for a magical instrument that drives the plot.³²

     How do filmmakers transmit a sense of the past to audiences? According to Toplin, a feel for history is accomplished largely through "attention to details of an earlier age,"³³ though note that such details need only be authentic, not accurate, to an audience (more on this later). Roland invokes names (e.g., Roland, Charlemagne, Geoffrey of Anjou) that would have resonated with a medieval audience as historical figures, while the characters' exploits unfold in Spanish and borderland locations that would likewise conjure memories of a bygone era. Descriptions of arms, armor, and fighting tactics, though often unrealistic, border on the obsessive. Through props and visuals, filmmakers can immediately situate an audience in time and place by drawing on well-known features of previous films. Helmet crests, capes, short swords, catapults, and disciplined infantry formations unmistakably convey ancient Rome to viewers in the opening scenes of Scott's Gladiator, as those props and features have become staples of ancient cinematic history. And of course, no film set in the Roman era would be complete without a Coliseum, a requirement so strong that the building has been depicted in some movies set prior to its construction in the first century ce.³⁴ Film historian Andrew Elliott calls these props and settings "historicons," which he defines as "a signifying aspect which is iconically linked to a historical period, person or event."³⁵ These become a convenient shorthand for moviemakers, obviating the need to provide extensive backstory for audiences.

     Finally, film producers go beyond character dialogue, making powerful connections with viewers using images and sounds.³⁶ This takes us outside the realm of a poem like The Song of Roland, though the readers or singers of the epic may have invoked telltale movements, props, music, and sound effects to punctuate passages for their audiences. Films are especially potent at captivating viewers through cinematography, visual and aural manipulations, and soundtracks. Consider Spielberg's masterful Saving Private Ryan. Evoking early color newsreels of World War II, Spielberg coated his treatment in sepia tones, a standard technique for films depicting that era. His scene of a mother receiving death notifications of her sons is a veritable Norman Rockwell tableau, transpiring on an idyllic Iowa farm (could it be anywhere but the American Heartland?) and featuring a white picket fence that doubles as grave markers reflected in a window. Containing no dialogue, the passage is overlaid with a swelling orchestral soundtrack that leaves no doubt about the enormity of American losses in the war effort.³⁷ Sounds that are diegetic (i.e., that take place within the film space and characters could be expected to hear) and non-diegetic (i.e., that exist outside of the story space, particularly soundtracks) play essential roles in movies. Soundtracks can be quite generic and are often anachronistic to the events portrayed on screen (for example, a synthesized or modern orchestral arrangement for a pre-modern film setting). Some diegetic sounds—the clanging of a blacksmith, a fife and drum ditty, the roar of piston airplane engines—are immediately recognizable to audiences to specify time and place, thus serving as aural historicons.

     The foregoing comparisons between a medieval epic poem and a selection of recent films underscore Toplin's points about the workings of historical cinema. One should not be too dogmatic about Toplin's scheme, as there are no hard-and-fast "rules" that storytellers must abide by. For example, Robert Altman's Gosford Park, an Edwardian-era period piece, features a dizzying array of characters and complicated storylines which nonetheless work. But it is niche piece, requiring concentration (and perhaps repeated viewings) to understand fully.³⁸ Rather, time and experience have shown that certain techniques and motifs are more likely to succeed than others. Just as important, there should be no expectation that a dramatic retelling of historical events, whether in written or filmic mode, should be the equivalent of a facts-driven textbook brought to the screen: the latter constitutes documentary, which is beyond the purview of this article. Within the context of storytelling, strict historical accuracy becomes just one variable, and is perhaps not even the most interesting.

     All of this reinforces the fact that dramatic films must be "read" in ways that are fair and sympathetic to the medium. Historians deploy a bevy of heuristics, unconsciously and unseen to observers, to help make sense of writings from and about the past. Oftentimes these accounts strain credulity, given issues of authorship, conflicting evidence, and the limits of what is physically possible. Yet, historians do not simply dismiss such accounts as so much junk. Rather, they seek to understand why the story is told the way that it is, and what it says about the author who created it and the societies that accepted or rejected it. The same holds true for historical cinema. More to the point, filmmakers, unbound by the fact-based standards of the historian but beholden to their own industry conventions, have great latitude to experiment with issues of the past that remain enigmatic or forever lost to academics.

Which History Is on Display?

     The film industry is a revolutionary one, allowing audiences to see and experience things that static art, literature, and stage companies could scarcely imagine. That said, film grew out of much older storytelling techniques, as reflected in the earlier comparisons between historical movies and The Song of Roland. Being a primarily visual experience, film owes much of its workings to previous art forms, as the following investigation will show. This section likewise demonstrates how traditional techniques of historical research can be applied to cinematic recreations, but only if conventions of the film medium are borne in mind.

     The Morgan Picture Bible or MPB (also known as the Maciejowski Bible) contains fifty-nine miniature illustrations from the thirteenth-century.³⁹ Like the slightly earlier Song of Roland, holy war is a major theme of the MPB illuminations. Despite being generally considered an "accurate" representation of fighting in that era, historian Richard Abels determined that the MPB's version of medieval combat is highly stylized, exaggerated, and inconsistent with other types of evidence. For example, "war" in the MPB is synonymous with battles, even though sieges were the bedrock of the medieval military landscape. In addition, soldiers' armor, which is correctly depicted in the MPB, appears to be of no protective value in the illustrations: weapons cut through and pierce it with ease. This is at odds both with modern tests of armor's efficacy (in short, it worked pretty well) and with skeletal evidence of wound sites on soldiers' bodies (where damage to the lower legs and forearms, i.e., areas that were unarmored or lightly armored, were common). Abels's systematic tally of wound locations in the MPB miniatures shows they are all above the waist and predominantly in areas of the body that are protected by armor.⁴⁰

     These findings are significant, but they should not necessarily surprise us. For one thing, medieval epics do in words what the MPB does in illustrations. The Song of Roland focuses on a great battle with sieges lurking only in the background, while wound sites and the utility of armor play out in exactly the same way as in the MPB. The sword swings and lance thrusts in the epic not only cut through armor and shields with ease, but cleave bodies (both human and equine) in ways that defy physical limits. These over-the-top representations of warfare are consistent across painting and written mediums of the era, but they are just that: representational, not to be taken literally. Art historian Sara Lipton cautions as much, saying that researchers "have learned to see images not as 'illustration' but as 'representation', and to be wary of assuming that they depict 'reality'."⁴¹

     If filmmakers take their broad storytelling cues from older forms of art and literature, we might expect them to do likewise in their more specific depictions of combat in the Middle Ages. Indeed, this seems to be the case. Like in the MPB, cinematic medieval warfare is heavily tilted toward set piece battles. Even when sieges occur in films, they typically consist of assaults on positions that end up taking the form of melees and individual combat, all the better to highlight individual prowess and keep the action going.

     Wound sites likewise reflect the stylized representations found in the MPB and Song of Roland. In Braveheart's depiction of the battle of Stirling in 1297, students in one of my history courses were able to identify 102 discrete injuries (there is admittedly some uncertainty, due to fast cutting and editing techniques to mimic the chaos of battle). Like the MPB, these wounds are predominantly from the waist up (86%), while injuries to the legs and arms are rare. The Scots (i.e., the protagonists) are mostly unarmored, but the defenses worn by the English seem to do them little good: fifty-nine instances show the defeat of their woven plate armor, while their helmets fail in another eleven cases. Statistical analyses of battle scenes in Antoine Fuqua's King Arthur and Scott's Robin Hood reveal similar patterns.⁴² In a sense, these films' renditions of medieval warfare are the MPB brought to life and put into motion, if only because filmmakers are the heirs to longstanding traditions of combat portrayals. The amount of daylight between a thirteenth-century illuminated manuscript and modern action pictures is vanishingly small.

Accuracy vs. Authenticity

     Academics are understandably concerned with the accuracy of historical cinema. Sometimes this can amount to little more than a turf war, where scholars jealously guard their academic fiefdoms against all intruders, be they filmmakers or anyone else. The case of the sky's depiction on the night of the ship's demise in Cameron's Titanic is a ready example. Astrophysicist Neil deGrasse Tyson objected that the star field Rose saw while lying on driftwood did not align with known celestial positions. One could reasonably raise a So what? defense, but Cameron, always the stickler for accuracy, reacted by going back and digitally editing the scene years after the film's release.⁴³

     Other times, filmmakers can independently go to seemingly ridiculous lengths of verisimilitude. The production designer for the Watergate-era All the President's Men insisted that actual garbage from Washington Post reporters' trash bins be flown in to the Los Angeles film set.⁴⁴ But some obvious, recurring errors have calcified to the point that fixing them is perhaps hopeless: bald eagles on screen, for example, are accompanied by the screech of red-tailed hawks, much to the chagrin of ornithologists.⁴⁵ Historians often raise legitimate objections about factual transgressions, but these disputes can be more complicated than they seem at first glance.

     The case of the interplay between the MPB and screen representations of medieval warfare is but one example. On the one hand, the MPB's and filmmakers' tendencies to favor battles over sieges is at odds with the consensus view of medievalists.⁴⁶ Yet, not all scholars are on the same page here. Such prominent historians as Clifford Rogers, Stephen Morillo, and Kelly DeVries make compelling arguments for battles as more central, due to the important functions of mounted forces and archers.⁴⁷ In other words, to insist that directors create films in some sort of Rankean, wie es eigentlich gewesen fashion, is to deny the existence of historiography.⁴⁸

     If historians fail to agree on many basic aspects of the past, it can be both disingenuous and unfair to expect filmmakers to tell a singular, accurate story about contested events. Granted, movie producers can court problems by making outrageous assertions. Bruckheimer and Fuqua's King Arthur opens with a ridiculous textual "truth claim" about the historicity of its titular character and the contours of his actions.⁴⁹ That such a claim is made in written form lends it even more credence, since text is connected with academic history in ways that film is not.⁵⁰ Filmmakers Joel and Ethan Coen playfully leveraged this connection to give Fargo greater substance, opening with script that their story was based on real events but knowing full well it was fictitious.⁵¹

     In fact, one could argue that many films about medieval warfare essentially "get it right," insofar as they feature depictions that are largely consistent with the evidence contained in epics and artwork from the Middle Ages. The martial protagonists in Braveheart, King Arthur, and Kingdom of Heaven, for instance, have much in common with the stories and illustrations from the period they represent. Like such epic heroes as Roland, the Cid, or Raoul de Cambrai,⁵² the film protagonists are larger than life, killing dozens with seeming ease while being role-model citizens. In that sense, what could be more accurate than to depict principal characters as such on screen?

     One way to make sense of this is film historian Robert Rosenstone's identification of "false invention" vs. "true invention." The former "ignores the discourse of history," whereas the latter consciously "engages" the same.⁵³ But even here, we run into problems: Is William Wallace's affair with Princess Isabella in Braveheart wholly gratuitous, a blatant distortion of the historical record and thus a "false invention"? One could reasonably take such a position. Conversely, one might rationally argue, per Toplin's earlier criteria, that the affair is consistent with the historical film genre's propensity to insert romance, whether factual or not.

     Thus, aiming for strict accuracy can be like trying to hit a moving target. Instead, movie producers look primarily to earlier film representations of their subject matter, using them as primary sources analogous to historians' reliance on contemporary materials. The goal becomes one of authenticity in the eyes of the audience, rather than accuracy in the estimations of subject-matter experts. Because such authenticity may be based on earlier, factually problematic film representations of the past, it assumes a reality (indeed, a place of necessity) in audience expectations. Elliott describes the dynamic as one where "ironically the inaccurate thus becomes the authentic."⁵⁴

     Costuming and props offer compelling examples of the complicated interplay between cinematic authenticity and historical accuracy. Would William Wallace and his Scottish compatriots have dressed in tartan kilts in the late thirteenth century? Most likely not, but the director's failure to dress them as such would sow confusion and consternation among moviegoers, who "know" that medieval Scotsmen wore kilts.⁵⁵ Would the invading Saxons of King Arthur's Britain have relied on crossbows as their missile weapon of choice? Available evidence militates against it, but the crossbow, through both its medieval reputation and repeated appearance in the hands of evildoers on screen, offers a powerful shorthand to establishing characterizations.⁵⁶ What drives the accuracy-minded historian crazy is, to the filmmaker, an expeditious and perhaps dictated factual error, given the conventions of the latter's craft and the expectations of audiences.

     And make no mistake: audience matters. The researcher analyzing a text prudently takes audience into account when trying to make sense of the artifact. Audience buy-in can be dictated by societal norms, as an examination of child soldiers in film demonstrates. World historians can easily imagine how defining a "child" invites all sorts of hazards, depending on time and place. Even in the United States today, a confusing range of ages signals the threshold between youth and adult: movie theaters often regard those thirteen and older as adults for ticketing purposes, but bar viewers under seventeen from R-rated films; a driver's license usually comes at sixteen, though exceptions for operating farm machinery apply; ages for service in the military, to vote, to use alcohol and tobacco, to benefit from lower car insurance rates, and to run for president are not all in agreement. The list goes on.

     Medieval Europe likewise had an array of ages marking adulthood, depending on context, that often do not align with the present. Thus, if a filmmaker chooses to depict child soldiers in a medieval-inspired story, does she opt for historical fidelity (with all the attendant problems signaled above), or does she cast actors of an age that would most readily resonate as "children" to twenty-first-century viewers? And whereas it is an expectation that child soldiers appearing in ostensibly "Western" medieval films partake in meaningful conflicts between Good and Evil, movies showcasing militant youths in fungible African settings commit senseless acts of violence amid failed states. Viewers' extant beliefs about the two historical contexts help dictate these very different portrayals.⁵⁷

The Utility of Cinematic History

     Can filmmakers act as historians, teasing out answers to genuine historical questions? Rosenstone believes they can, writing that the problems both professions face, and how they creatively grapple with them in their particular mediums, suggest historical value in the film producer's craft.⁵⁸ Although this is not the case in all dramatic recreations of the past, there are times when the cinema offers much food for thought.

     Consider for example the ground-level realities of pre-modern combat. Despite copious handbooks and narrative sources on how battles played out, there is still much that remains enigmatic. Military historian John Keegan famously attempted to delineate the "face of battle," combining available evidence, imaginative thought experiments, and the known physical limits of soldiers and their weapon systems. Yet, he remained stymied at various turns. In regard to battlefield communications and troop movements, he concluded that such matters were "unanswerable" in pre-modern contexts. How did bodies of soldiers actually engage one another? Keegan could only offer potential scenarios.⁵⁹

     Philip Sabin's follow-up study of ancient combat owed much to Keegan. Sabin identified various problems that would have afflicted engaged soldiers (e.g., physical and psychological fatigue, communication failure, loss of a leader, unit incoherence, wound rates) as a function of three possible modes of combat. Those modes are the "shoving match," the "melee" (the default for most Hollywood depictions), and the "mutual dread/tentative engagement" model. Although Sabin ultimately settled on the last option as the most likely, he, like Keegan, was limited in his conclusions.⁶⁰

     The opening scene in the premier episode of HBO's Rome dramatic series shows evidence of filmmakers who are not just aware of the evidentiary limitations of understanding combat, but who offer well conceived, if only tentative, solutions to some of the issues that thwarted Keegan and Sabin.⁶¹ After a set-up narration over a montage of maps, artwork, and leading Roman figures, the series begins in medias res at a skirmish during the siege of Alesia in 52 bce. Taking a claustrophobic, soldier's eye view of the Roman lines, the audience sees troops who are dirty and bloodied from prior action, and who grimly await a renewed attack from Gallic forces.

     As the assault comes, the unit commander blows a whistle, signaling his men to assume battle formations. And then, contrary to most film depictions of pre-modern combat as one big melee, the Romans hunker down and let the Gauls come to them, fighting conservatively and methodically with simple sword thrusts from behind their shields. Unit cohesion is maintained by rows further back grasping leather straps attached to those forward. It is a wonderful display of Roman discipline, putting into moving pictures the advice of the later Imperial military writer Vegetius, who warned that assuming a defensive stance and "standing like a wall" were essential to victory.⁶² There is no individual bravado, no wild and dramatic sword slashing here. This is disciplined, impersonal, lethal combat at eye level, where the style of fighting very much fits Sabin's mutual dread/tentative engagement model.

     As fatigue sets in, as it must set in, the Roman lines require rotation, and it is here that the filmmakers take a brilliant step into the realm of reasoned historical speculation. On the commander's second whistle signal, the camera switches to bird's eye view of the Roman formation and shows the lines turn sideways, then march and countermarch as the first line egresses through the second, bringing the latter fresh troops to the front. It is a brief moment, lasting only a few seconds and surely unremarkable to the vast majority of viewers. But in that instant, directors Michael Apted and Mikael Salomon followed "the latest geeky research" advice of their historical consultant, Jonathan Stamp, to present a historically plausible solution to one of the vexing questions about pre-modern battle: how lines of soldiers, in the midst of a deadly fight, could rotate in and out of combat.⁶³

     Historian Lee Brice was impressed with the scene. The filmmakers' portrayal of the line shift, in particular, "provides a visualization of how such a process might have worked." He continued: "While it may not be the definitive reconstruction, the scene is valuable because it provides viewers with a sense of how complicated Roman tactics could be and of the importance of discipline and drill to success."⁶⁴ Although Rome may exist primarily for entertainment purposes, its producers, here and elsewhere in the series, enter into the realm of evidence-based historical speculation often considered the preserve of academics. Their creative solutions offer much for general audiences and historians alike to ponder.

Conclusion

     The American public has a rather grim view toward cinematic history. A national survey conducted by Roy Rosenzweig and David Thelen revealed that, although popular vehicles for interacting with history, movies and television are considered the least trustworthy sources of the past.⁶⁵ Despite this suspicion toward the medium, filmic history strongly influences viewers' understandings of earlier peoples and events. Wineburg's in-depth study of one family's views on the Vietnam war concluded that movies played crucial roles in its knowledge about the conflict. The idea that war is good for the economy? A byproduct of having watched Schindler's List.⁶⁶ Returned American soldiers from Southeast Asia being spat on and called baby killers? Largely a result of repeated viewings of Forrest Gump.⁶⁷

     In each of these cases, the collective memories created by filmmakers can crowd out or serve as the only source of putative historical facts, a phenomenon Wineburg labels "collective occlusion" of the past.⁶⁸ This is even more likely to occur when considering events from long ago and/or in places far away. These are things about which viewers may know nothing factual, so that the "history" portrayed on screen becomes the sole reference point for understanding. Paul Sturtevant's prolonged work with British focus groups concluded that non-academics can be aware of their historical blind spots, but that films nonetheless inject notions (e.g., the Middle Ages were full of mud and filth; medieval masculinity was decidedly aggressive and macho) that become expectations.⁶⁹ Hornaday reaffirms this more generally, writing that, "by virtue of their mass audience," filmic histories "become our consensus version of history itself."⁷⁰

     All of this is to acknowledge that film impacts, if not determines, the ways the general public perceives and understands the past. That is not necessarily a bad thing, insofar as cinematic recreations can question, complement, or even bring new academic interpretations to light. At the same time, general audiences and academics alike must understand that the conventions of dramatic film predispose the medium to alterations and inventions that may not sit well with fact-focused historians. However, these conventions have a long history in storytelling and art forms that pre-date the film industry. Comparing medieval sources such as The Song of Roland or the Morgan Picture Bible against cinematic recreations reveals tropes and tendencies of which historians are readily aware when it comes to primary sources, but which too often go unappreciated in movies and television. Moreover, the conflicts of historiography serve as prima facie evidence of the vagaries of the past: claims that a unified portrayal of anything beyond the trivial is possible amount to little more than hubris. We are all prisoners of our own knowledge and understandings about the past, so that our professional skills serve as both an aid and a detriment to our interpretations of cinematic history. Fortunately, the doors of our interpretive cells swing open if we employ the right keys.

Peter Burkholder is Professor of History at Fairleigh Dickinson University in Madison, NJ. He is on the national advisory board of the Society for History Education and the editorial board of The Teaching Professor, and he serves as a consulting editor for College Teaching. He has produced numerous publications on cinematic history and college-level pedagogy.

¹ "Film Reviews: Introductory Note," American Historical Review 123 no. 2 (2018), 529. In fairness, film treatment had continued intermittently in AHR's monthly companion publication, Perspectives on History.

² Malden, MA: Wiley-Blackwell, 2010.

³ Plutarch's Lives, Bernadotte Perin, trans. (London: William Heinemann, 1920), 9:193–97; available online at <http://penelope.uchicago.edu/Thayer/E/Roman/Texts/Plutarch/Lives/Antony*.html>, 26.1–27.1 (accessed July 30, 2018).

⁴ Readers are forgiven if they conjured up the barge scene from Cleopatra, directed by Joseph Mankiewicz (Twentieth Century Fox, 1963), which obviously took inspiration from Plutarch's account. But the portrayals of Cleopatra and Antony differ in each medium.

⁵ Sam Wineburg, Historical Thinking and Other Unnatural Acts (Philadelphia: Temple University Press, 2001), chs. 3–4. See also Peter Burkholder, "Reading Right Past Each Other: Divergent Faculty and Student Perspectives on Texts," Teaching United States History (January 11, 2018): < http://www.teachingushistory.co/2018/01/reading-right-past-each-other-divergent-faculty-and-student-perspectives-on-texts.html>.

⁶ For decoding, David Pace, The Decoding the Disciplines Paradigm (Bloomington: Indiana University Press, 2017).

⁷ On leveraging prior knowledge for new learning, see Susan Ambrose et al., How Learning Works: 7 Research-Based Principles for Smart Teaching (San Francisco: Jossey-Bass, 2010), 10–39.

⁸ The Song of Roland, Glyn Burgess, trans. (New York: Penguin, 1990).

⁹ Robert Toplin, Reel History: In Defense of Hollywood (Lawrence, KS: University Press of Kansas, 2002), 17–53. A shorter version featuring only six themes is Toplin, "Cinematic History: An Anatomy of the Genre," Cineaste 29 no. 2 (2004), 34–39. Toplin's structure also provides a useful organizational approach to teaching historical cinema; see Peter Burkholder, "From Passive Viewer to Active Learner: Strategies for Teaching Medieval Film," Studies in Medieval and Renaissance Teaching 24 no. 1 (2017), 61–78.

¹⁰ Toplin, Reel History, 17–19.

¹¹ The Patriot, directed by Mel Gibson (Columbia Pictures, 2000).

¹² Dunkirk, directed by Christopher Nolan (Syncopy, 2017).

¹³ Toplin, Reel History, 19–23.

¹⁴ Empire of the Sun, directed by Steven Spielberg (Amblin Entertainment, 1987).

¹⁵ Toplin, Reel History, 23–30.

¹⁶ Pathfinder, directed by Marcus Nispel (Twentieth Century Fox, 2007). On pre- and post-revisionist portrayals of Hollywood Vikings, John Aberth, A Knight at the Movies: Medieval History on Film (New York: Routledge, 2003), 42–61.

¹⁷ Titanic, directed by James Cameron (Twentieth Century Fox, 1997); Ann Hornaday, Talking Pictures: How to Watch Movies (New York: Basic Books, 2017), 17.

¹⁸ Toplin, Reel History, 30–36.

¹⁹ The Lord of the Rings, directed by Peter Jackson (New Line Cinema, 2001–03).

²⁰ 12 Years a Slave, directed by Steve McQueen (Regency Enterprises, 2013); Selma, directed by Ava DuVernay (Pathé, 2014).

²¹ Toplin, Reel History, 36–41.

²² The Messenger: The Story of Joan of Arc, directed by Luc Besson (Gaumont, 1999).

²³ Jeanine Basinger, "Translating War: The Combat Film Genre and Saving Private Ryan," Perspectives on History (October 1998): <https://www.historians.org/publications-and-directories/perspectives-on-history/october-1998/translating-war-the-combat-film-genre-and-saving-private-ryan>; Basinger, The World War II Combat Film: Anatomy of a Genre (Middletown, CT: Wesleyan University Press, 2003).

²⁴ Toplin, Reel History, 41–46.

²⁵ Versions of the pope's speech at Clermont are available at "Urban II (1088–1099): Speech at Council of Clermont (1095), Five Versions of the Speech," Medieval Sourcebook: <http://www.fordham.edu/halsall/source/urban2-5vers.html> (accessed August 5, 2018). For analysis, Michael Lower, "Pope Urban II's Call to Crusade," in Brian Bonhomme, ed., Milestone Documents in World History (Dallas: Schlager Press, 2010), 2:497–512.

²⁶ Kingdom of Heaven, directed by Ridley Scott (Twentieth Century Fox, 2005).

²⁷ The first time a slight majority (52%) of Americans showed disapproval for the U.S. intervention in Iraq was mid-2005, though 50% still approved of President Bush's war on terrorism; Dana Milbank and Claudia Deane, "Poll Finds Dimmer View of Iraq War," The Washington Post (June 8, 2005): <http://www.washingtonpost.com/wp-dyn/content/article/2005/06/07/AR2005060700296.html>. For domestic box office woes, see <http://www.the-numbers.com/movies/2005/KINGH.php> (accessed August 12, 2018).

²⁸ Braveheart, directed by Mel Gibson (Icon Entertainment International, 1995); Colin McArthur, Brigadoon, Braveheart and the Scots (London: I.B. Taurus, 2003), 123–29.

²⁹ Toplin, Reel History, 46–47.

³⁰ On the literary topos, Malcolm Barber, "How the West Saw Islam," History Today 47 no. 5 (1997), 44–50 at 46.

³¹ See the review of Braveheart by Elizabeth Ewan in American Historical Review 100 (1995), 1219–21.

³² The Red Violin, directed by François Girard (Rhombus Media, 1998).

³³ Toplin, Reel History, 47–50.

³⁴ Gladiator, directed by Ridley Scott (DreamWorks, 2000); Martin Winkler, "Gladiator and the Colosseum: Ambiguities of Spectacle," in Winkler, ed., Gladiator: Film and History (Malden, MA: Blackwell, 2005), 87–110.

³⁵ Andrew Elliott, Remaking the Middle Ages: The Methods of Cinema and History in Portraying the Medieval World (Jefferson, NC: McFarland, 2011), 224.

³⁶ Toplin, Reel History, 50–53.

³⁷ Saving Private Ryan, directed by Steven Spielberg (DreamWorks, 1998).

³⁸ Gosford Park, directed by Robert Altman (USA Films, 2001).

³⁹ An online version of this source is at <https://www.themorgan.org/collection/Crusader-Bible> (accessed August 10, 2018).

⁴⁰ Richard Abels, "Cultural Representations of Warfare in the High Middle Ages: The Morgan Picture Bible," in Simon John and Nicholas Morton, eds., Crusading and Warfare in the Middle Ages (Surrey, England: Ashgate, 2014), 13–35. For a much wider survey of wound types and sites, see the collection of essays in Wounds and Wound Repair in Medieval Culture, Larissa Tracy and Kelly DeVries, eds. (Leiden: Brill, 2015).

⁴¹ Sara Lipton, "Images and Objects as Sources for Medieval History," in Joel Rosenthal, ed., Understanding Medieval Primary Sources (New York: Routledge, 2012), 225–42 at 226.

⁴² King Arthur, directed by Antoine Fuqua (Touchstone Pictures, 2004); Robin Hood, directed by Ridley Scott (Universal Pictures, 2010).

⁴³ Simon Brew, "How Neil deGrasse Tyson Got James Cameron to Edit Titanic--15 Years Later," Mental Floss UK (September 16, 2015): <http://mentalfloss.com/article/68595/how-neil-degrasse-tyson-got-james-cameron-edit-titanic-15-years-later>.

⁴⁴ All the President's Men, directed by Alan Pakula (Warner Bros., 1976); Hornaday, Talking Pictures, 77.

⁴⁵ Hornaday, Talking Pictures, 184.

⁴⁶ See for example Jim Bradbury, The Medieval Siege (Woodbridge: Boydell, 1992); Bernard Bachrach, "Medieval Siege Warfare: A Reconnaissance," The Journal of Military History 58 (1994), 119–33; Peter Burkholder, "Popular [Mis]conceptions of Medieval Warfare," History Compass 5 no. 2 (2007), 507–24; Denis Sullivan and Sean McGlynn, "Siege Warfare," in Clifford Rogers, ed., The Oxford Encyclopedia of Medieval Warfare and Military Technology (Oxford: Oxford University Press, 2010), 3:264–70.

⁴⁷ Clifford Rogers, "Carolingian Cavalry in Battle: The Evidence Reconsidered," in John and Morton, eds., Crusading Warfare in the Middle Ages, 1–11; Stephen Morillo, "Battle Seeking: The Contexts and Limits of Vegetian Strategy," Journal of Medieval Military History 1 (2002), 21–41; Kelly DeVries, Infantry Warfare in the Early Fourteenth Century: Discipline, Tactics, and Technology (Woodbridge, U.K.: Boydell, 1996).

⁴⁸ For a more extended analysis of the importance of historiography to understanding the cinematic past, see Elliott, Remaking the Middle Ages, ch. 1. Some films do manage to grapple with historiographical issues. Despite rough reviews at the hands of academics, Mel Gibson's Apocalypto (Icon Entertainment International, 2006) hints at the tension between violence and ecological destruction as contributing factors to the downfall of the Maya.

⁴⁹ The historical consultant's claims to historical veracity are in John Matthews, "A Knightly Endeavor: The Making of Jerry Bruckheimer's King Arthur," Arthuriana 14 no. 3 (2004), 112–15. But Caroline Jewers points out numerous flaws in "Mission Historical, or '[T]here were a hell of a lot of knights': Ethnicity and Alterity in Jerry Bruckheimer's King Arthur," in Lynn Ramey and Tison Pugh, eds., Race, Class, and Gender in "Medieval" Cinema (New York: Palgrave Macmillan, 2007), 91–106.

⁵⁰ Richard Burt, "Getting Schmedieval: Of Manuscript and Film Prologues, Paratexts, and Paradies," Exemplaria 19 no. 2 (2007), 217–42.

⁵¹ Fargo, directed by Joel Coen and Ethan Coen (PolyGram Filmed Entertainment, 1996). See the Coens' tongue-in-cheek explanation in Bill Bradley, "The Coen Brothers Reveal Fargo Is Based on a True Story After All," Huffington Post (March 8, 2016): <https://www.huffingtonpost.com/entry/coen-brothers-fargo-true-story_us_56de2c53e4b0ffe6f8ea78c4>.

⁵² A citation for the Roland epic is in note 8 above. For the Cid, The Song of the Cid, Burton Raffel, trans. (New York: Penguin, 2009); for Raoul, Raoul de Cambrai, Sarah Kay, trans. (Oxford: Oxford University Press, 1992).

⁵³ Robert Rosenstone, Visions of the Past: The Challenge of Film to Our Idea of History (Cambridge, MA: Harvard University Press, 1995), 64–76, with quotes at 72.

⁵⁴ Elliott, Remaking the Middle Ages, 216 (emphasis in original).

⁵⁵ On costuming and props in historical dramas, see Richard LaMotte, "Designing Costumes for the Historical Film," Cineaste 29 no. 2 (2004), 50–54, and Charles Tashiro, "Passing for the Past: Production Design and the Historical Film," Cineaste 29 no. 2 (2004), 40–44.

⁵⁶ Peter Burkholder, "X Marks the Plot: Crossbows in Medieval Film," Studies in Popular Culture 38 no. 1 (2015), 19–40.

⁵⁷ Peter Burkholder and David Rosen, "Child Soldiers in Medieval(esque) Cinema," in Sara Buttsworth and Maartje Abbenhuis, eds., War, Myths, and Fairy Tales (New York and Singapore: Palgrave Macmillan/Springer, 2017), 147–73.

⁵⁸ Robert Rosenstone, History on Film/Film on History (Harlow, England: Pearson Publishing, 2006), ch. 7: "Film maker/Historian."

⁵⁹ John Keegan, The Face of Battle (New York: Viking Press, 1976), with quote at 93.

⁶⁰ Philip Sabin, "The Face of Roman Battle," Journal of Roman History 90 (2000), 1–17.

⁶¹ "The Stolen Eagle," Rome, Season 1, Episode 1, directed by Michael Apted and Mikael Salomon (HD Vision Studios, 2005).

⁶² Vegetius: Epitome of Military Science, N.P. Milner, trans. (Liverpool: Liverpool University Press, 2001), 49–50. Only when one of the characters, Titus Pullo, breaks formation in the HBO scene is the unit put at risk.

⁶³ Sabin, "Face of Roman Battle," 7, discusses our limitations in understanding how this process worked. The "latest geeky research" line is from the DVD commentary to the episode.

⁶⁴ Lee Brice, "The Fog of War: The Army in Rome," in Monica Cyrino, ed., Rome, Season One: History Makes Television (Malden, MA: Blackwell, 2008), 61–77 at 67.

⁶⁵ Roy Rosenzweig and David Thelen, The Presence of the Past: Popular Uses of History in American Life (New York: Columbia University Press, 1998), 20–21.

⁶⁶ Schindler's List, directed by Steven Spielberg (Universal Pictures, 1993).

⁶⁷ Forrest Gump, directed by Robert Zemeckis (Paramount Pictures, 1994).

⁶⁸ Wineburg, Historical Thinking, ch. 10: "Making (Historical) Sense of the New Millennium," with "collective occlusion" at 242.

⁶⁹ Paul Sturtevant, The Middle Ages in Popular Imagination: Memory, Film and Medievalism (London: I.B. Taurus, 2018).

⁷⁰ Hornaday, Talking Pictures, 249.