Revised and Visualized: “Killers of the Flower Moon” and “The Devil Commands”
It happens all the time: I read two books that have nothing in common, but due to their proximity on my reading list, I notice surprising resonances between them, as if the happenstance of my literary diet was predetermined by some omnipotent TA. This autumn, the unexpected pairing was between the novels Killers of the Flower Moon and The Edge of Running Water, or rather, the films adapted from them. The novels are not much alike: Killers is a nonfiction history of the Osage Indian Murders, Water is the arguably-less-substantial tale of a mad scientist and his machine for talking to ghosts. Both books center around a mystery and its untangling by a heroic protagonist, but the more interesting similarity is that the screenwriters behind each film adaptation chose to forego all mystery, sideline the hero, and tell the whole story from the villain’s point of view. The reasoning behind these revisions, and the narrative choices made to effectuate them, provoke some interesting observations about the function of perspective in novels and feature films.
The choice to radically revise each narrative was made for reasons as distinct as the stories themselves. William Sloane’s The Edge of Running Water is a spooky science fiction novel, published in 1939 and then adapted two years later as a Boris Karloff vehicle with a new title: The Devil Commands. In both versions, the narrative intrigue surrounds a creepy invention which uses radio waves to communicate with the dead, and the murderous activity the inventor and his malicious servant get up to in order to protect their machine. While Sloane made some attempt to elevate his subject matter with psychological realism and polished prose, what artistry the film has is incidental to its purpose as a cheapie schedule-filler for Universal Pictures. 1940s audiences showed sufficient interest in Karloff as a mad scientist to justify the purchase and revision of Sloane’s book, in which the scientist character is more of an ominous presence than a primary character. The actual protagonist and narrator of Water is the scientist’s protégé, psychologist Richard Sayles, whose name I had to Google, as Sayles is not the most memorable or fascinating character in literary history. But then, he doesn’t have to be, since he is not so much a character as he is a conduit for a type of reader identification more particular to literature than cinema.
As narrator, Sayles is psychologically normal, intelligent but not brilliant, principled but not self-righteous, devoid of quirks to the point of being virtually non-descript, and as such he is relatable to just about any young, male sci-fi reader. Despite being the hero of the story, Sayles mostly plays the role of passive witness. After seeking out Julian Blair, his old friend (and present mad scientist) in a remote town in Maine, Sayles spends much of the book gradually piecing together the creepy stuff Blair has been up to in his recent seclusion. Blair has created a machine which he hopes will connect him with the spirit of his deceased wife, but which is actually more likely to create an apocalyptic tear in spacetime; Sayles’ initial skepticism and eventual horror at Blair’s invention are no more or less than the standard reactions of the stock literary everyman. This is Sayles’ purpose, he is straightforwardly an avatar by which the reader can experience the story as if it was happening to them.
Sloane wasn’t doing anything new here; presenting a deranged character and a spooky situation through the first-person perspective of an ordinary, nearly uncharacterized observer is a staple of the genre. H.P. Lovecraft, possibly an influence on Sloane*, structured many of his stories this way. The device operates according to a simple logic: filling the narration with personal pronouns and first-person descriptions automatically gives the reader a deeper stake in the story, and characterizing the narrator exclusively with relatable traits like intelligence and benign morality serves to connect any reader more closely with the narrator’s voice. Then, if the reader-narrator identification is successful, when the scary stuff happens, the narrator’s terror will be felt more powerfully by the reader. The device is so simply and subtly effective that it can be pushed to perverse extremes without breaking the reality of the story. In Lovecraft’s story Herbert West – Reanimator, the nondescript protagonist/narrator is the lifelong assistant to the titular Frankenstein-like scientist. The nameless narrator assists the increasingly-malevolent West at every step of his iniquitous experiments, from grave robbing to murder, and yet for the entirety of the story the narrator maintains the moral horror of West without which reader identification would be difficult to maintain. Witness the balancing act Lovecraft executes in the following paragraph:
West had soon learned that absolute freshness was the prime requisite for useful specimens, and had accordingly resorted to frightful and unnatural experiments in body-snatching. In college, […] my attitude toward him had been largely one of fascinated admiration; but as his boldness in methods grew, I began to develop a gnawing fear. I did not like the way he looked at healthy living bodies; and then there came a nightmarish session in the cellar laboratory when I learned that a certain specimen had been a living body when he secured it. That was the first time he had ever been able to revive the quality of rational thought in a corpse; and his success, obtained at such a loathsome cost, had completely hardened him.
This paragraph occurs only three quarters into the story, but despite his horror and revulsion at West’s behaviour, the narrator continues to assist the villain in even more despicable antics. Though Lovecraft may have found some discreet pleasure in the irony here, the contradiction is never pronounced in the story, and the narrator maintains the moral high ground even when West finally suffers the consequences for acts of which the narrator is equally guilty. This paradox does not hamper the success of the story at all because the narrator barely exists in any realistic way, he is transparently a literary device by which the horror can be executed with maximum impact on the reader.
Although Richard Sayles has a bit more specificity and coherence than the Reanimator protagonist, he serves much the same purpose in Sloane’s novel, so it’s no wonder that in The Devil Commands Sayles is barely a supporting character. While reader identification of the kind I’ve described above is possible in literature, the device does not translate to film at all. The reasons for this are obviously founded in the difference between reading and watching: in literature, one hundred percent of the story is delivered via the narrator’s voice; all the writer needs to do is make that voice first person and they’re halfway to creating a convincing facsimile of the reader’s inner monologue, regardless of the personality of the narrator. This is why a book like Jim Thompson’s The Killer Inside Me is so disturbing and claustrophobic: for as long as you’re reading it, you’re effectively replacing your own thoughts with the thoughts of a twisted psychopath. There is no analogue for this effect in filmmaking; if there’s any doubt, consider that a genuine first-person film has only been attempted a handful of times, with results that are questionable at best. In cinema, a purposefully bland protagonist like Richard Sayles thrills with the excitement of an instruction manual, regardless of his function as a surrogate for the audience. When adapting stories that have been structured this way, the screenwriters must decide how to repurpose the protagonist in a way that makes dramatic sense on screen. In Stuart Gordon’s 1985 Re-Animator, the narrator is given a life of his own, with conflicts that provide coherence to his actions. In The Devil Commands, Sayles’ function as a plot device is not hidden, and he’s given only the minimum amount of screen time necessary to keep the story going.
Julian Blair is the protagonist – or anti-hero – of The Devil Commands, and because the character was discussed more than depicted in The Edge of Running Water, screenwriters Robert Hardy Andrews and Milton Gunzburg needed to employ much invention to restructure the story around Blair. In the novel, whether or not Blair is mad, whether his machine works or not, what exactly the machine does, and whether he and his medium-cum-assistant Mrs. Walters murdered their housekeeper in order to protect the machine are all questions kept mysterious until the very end; these questions are the source of the novel’s suspense, but withholding answers to these questions would have been impossible in a story in which Blair is the main character. As a result, the writers decided, rather intelligently, to leave out mystery altogether, and refashion the narrative as a crime story in which the suspense is generated from watching morally ambivalent characters commit evil acts and then try to get away with them.
Martin Scorsese and Eric Roth made the identical decision when they scripted this fall’s Killers of the Flower Moon. The prestigious epic was adapted from journalist David Grann’s 2015 nonfiction history of the Osage Indian Murders of early 20th century Oklahoma, in which the oil-rich Osage were systematically murdered and robbed of their fortunes by the local white population. Despite Grann’s book being an historical, nonfiction work, he structures the narrative very much as a mystery, with the identities of the killers revealed on about the same schedule as you’d expect from any detective story. Grann even employs some intentional diversion, initially setting up the mastermind behind the killings, cattleman William Hale, as a friendly, benign figure. The bulk of the novel details the painstaking investigation conducted by FBI agent Tom White, a Texan who, by Grann’s account, could be the uncomplicated hero of an old western. In Scorsese’s film, however, White is hardly a primary character; he doesn’t even appear until about two hours into the film’s 3.5 hour running time. The central figures of Scorsese’s version are Hale (Robert De Niro) and more importantly Hale’s pliable nephew and accomplice Ernest Burkhart (Leonardo DiCaprio).
Being a real man who was by all accounts a noble and humane person, and Grann’s story being history and not fiction, Tom White cannot be categorized with Richard Sayles as a protagonist who’s more of a plot device than a character. That said, there are certain similarities to the role each character plays in his story, these similarities stemming from crime story tropes which are so ubiquitous that they create expectations even when the story is nonfiction. Given White’s stolid goodness, and the relative success with which he executed his investigation, I imagine Grann felt a certain atavistic excitement when he looked into White: here was a real-life hero detective who actually caught the bad guys and did not turn out to be corrupt or incompetent all along. What makes Grann’s work superlative, however, is the way he indulges that type of genre pleasure only to make the cold light of reality all the more bracing when it’s unveiled. For the last third of the book, Grann unfolds, for the first time, all of the insight granted by research and retrospect, and he proffers much evidence that the Osage killings were far more pervasive than just those committed by Hale and Burkhart; that the success of White’s investigation was a shallow victory, more of a cosmetic win that ultimately masked the widespread guilt and corruption amongst the white Oklahomans who made such victims of their Osage neighbours. Grann is quite brilliant for the way he gives his readers the formula they want and expect – in which a maverick detective outwits and apprehends the villains – only to expose the limitations of this formula when applied to real history. Therefore, by Grann’s narrative strategy, Tom White is both an historical figure and a literary device.
A close cinematic rendering of this narrative structure would have been intriguing. There are a few ways one can imagine it being pulled off: the apparent victory of the investigation could have been depicted in full, then followed by flashbacks showing everything White missed; or, all of the unprosecuted crimes could have been cut-to simultaneously with the investigation, so that at no point would an illusion of justice have been possible. Interesting possibilities, but perhaps too dry for a cinematic drama, and certainly drier than the film Scorsese has made. Though Scorsese, Roth and producer/collaborator DiCaprio famously struggled over how to tell the story from White’s point of view, the approach was dropped altogether and they opted to center their version around Burkhart and Mollie (Lily Gladstone), the Osage heiress whom Burkhart married in order to possess her fortune. Burkhart, who in the book registers as a simple thug, is transformed via Scorsese’s storytelling and DiCaprio’s performance into a man with the worst-ever case of cognitive dissonance; a complex figure who genuinely loves his wife even as he poisons her and murders her sisters at the beckoning of his satanic uncle. It’s almost surprising that this idea eluded Scorsese and his collaborators for so long, since the director’s lifelong subject has been the tormented humanity of evil men. Scorsese’s narrative focus on the psychological dynamics between Burkhart and Mollie comes at somewhat of a cost, however, as the widespread culpability of the white Oklahomans is not imparted as clearly in Scorsese’s telling as it is in Grann’s. Nevertheless, Scorsese resists presenting lawman White (Jesse Plemons) as anything like a hero or, even worse, a white saviour; White’s investigation comes across as the nothing more or less than the inevitable result of the entitled sloppiness of Hale and Burkart’s crimes, and there’s no implication that the investigation provided anything resembling justice to the Osage.
As in The Devil Commands, the choice to recenter the narrative of Killers necessitated many practical changes from the source material. It’s interesting to observe how in some ways the writers of each film took a similar approach (though I’m not about to suggest Scorsese or Roth were inspired by the little-seen Karloff picture). The first half of both films dramatize backstory that in the books was summarized in paragraphs rather than pages, and the writers employ considerable creativity imagining scenes about which readers of the books could only speculate.
In The Devil Commands, some of the additions improve on the source, providing new dramatic justification and upping the stakes of the story for cinematic suspense. In The Edge of Running Water, Julian Blair is goaded on by opportunistic medium Mrs. Walters, a sort of Lady Macbeth figure who sees Blair’s machine as a path to fame and fortune. Walters’ problematic sway over the otherwise rational Blair is meant as a sign of his weakening mind, but Sloane doesn’t offer much explanation for how Blair initially fell under Walters’ spell. For The Devil Commands, the screenwriters reinvent Mrs. Walters (Anne Revere) as a genetic anomaly whose body conducts electricity without harming her, making her the perfect engine for Blair’s machine, and giving Blair a logical motivation to keep her around. Then, the progression of the initially scientific partnership into something more sinister is depicted with impressive economy for the 65-minute movie. In order to work, Blair’s machine requires multiple human beings to generate the waves necessary to reach the dead; when Blair and Walters secretly use the machine on unlucky janitor Karl (Ralph Penney), the electricity almost kills him. Now quasi-criminals, Blair and Walters escape to Maine, with the brain-damaged Karl in tow as their zombie-like muscle. By the time Sayles (Richard Fiske) arrives in Maine, Blair and his motley crew have been grave-robbing corpses to power Blair’s machine. Writers Andrews and Gunzburg keep raising the stakes until the very end, when after Walters has been killed by too much electricity, Blair opts to use his daughter Anne (Amanda Duff) to power the machine instead. Blair’s willingness to kill his daughter in order to speak to his dead wife is the perfect dramatic irony to cap his total descent into madness; it also gives Sayles the opportunity to play his role as the perfunctory hero and rescue Anne just in time for the happy ending. Literally all of these suspense-inducing elements were invented for the film: the character of Karl, the need for humans to run Blair’s machine, the grave-robbing, Blair’s victimization of his daughter, even Blair having a daughter at all (in the novel, Anne is Blair’s sister-in-law). That Sloane’s plot was much simpler shows the power of mystery to keep a reader engaged when there is not a whole lot, dramatically, going on. Perhaps none of Andrews and Gunzburg’s changes display genius-level brilliance, but they all bespeak the writers’ talents for practical revision.
In writing the screenplay for Killers of the Flower Moon, Scorsese and Roth didn’t make too many changes to the plot (perhaps we should say, to history); their additions rely heavily on empathetic imagination. For at least the first hour, the film is more or less a romance between Burkhart and Mollie, and the actors and filmmakers do an incredible job imagining what kind of relationship might have existed between these two very real, very complicated people. The first act of the film is so quietly beautiful, so elegiac in its depiction of turn-of-the-20th-century Americana, that Scorsese almost tricks you into believing this will be a romantic tale of unconditional love and racial harmony. This way, when Burkhart takes his dark turn and the crimes covered by the book start to ramp up, the effect is far more painful for the viewer than it ever could be otherwise. The emotional impact of the film makes a powerful argument for the importance and purpose of performed drama; as valuable as Grann’s book is as a work of history, its cinematic realization here provides this important story with an empathetic power that would be hard to capture in any other medium.
If readers identify with the first-person narrators of fictions like The Edge of Running Water, do audiences identify in any comparable way with the protagonists of films and television shows? If they did, then I’d think we’d have more cinematic protagonists of the bland everyman[/woman] variety, and more leading actors with something less than world-historical good looks and charisma. Certainly, people talk as if they identify with screen characters; consider the oft-heard clichés: “The characters on [Seinfeld/Friends] are just like my friends and I,” or, “The Office is just like where I work.” However, if you’ll permit some amateur psychologizing, I think in these cases “identification” is used to describe what’s actually a form of aspiration. If pressed, probably no one would claim that their friends or co-workers are actually as consistently hysterical**, quirky and distinctive as the characters from those shows; screen fictions like these are carefully designed to impart a reflection of real life that’s distorted to be funnier, more interesting, more eventful, in order to trigger an aspiration in the viewer that’s just realistic enough to be mistaken for identification. I’d describe the relationship of a reader with a narrator like Sayles in Water as comparable to that of a video game player and his avatar; however, I think viewers relate to screen characters more like the way we’ve all related to that person we know who’s a little smarter, a little funnier, a little more charismatic than we are: with a mix of captivation, jealousy and aspiration.
At least, that’s how we react to the screen characters we like; what of anti-heroes like Ernest Burkhart and Julian Blair? Something interesting happens when such a character is made the lead of a narrative work, something that’s perhaps easier to cop to with a pulpy creation like Blair than a real-life villain like Burkhart. The phenomenon I’m describing is probably best summarized by Jerry Seinfeld’s stand-up routine about how when we watch the Discovery Channel and the subject is the antelope, we root for the antelope to flee the lion; but when next week the subject is the lion, we root for it to kill the antelope. If we don’t exactly root for the likes of DiCaprio’s Burkhart in Killers of the Flower Moon, we may look on in an avid, involved-to-an-uncomfortable-degree way to see just how long he can pull off his awful crimes. If that’s so, and I reckon it is for many viewers, is it because Burkhart is played by mega-charmer DiCaprio? Partly, I think; but perhaps classical narrative is so deeply embedded in the human mind as a means to organize information that just having a main character of any kind will trigger some level of involvement in the viewer, or reader, or listener. Some may argue that given this idea, centering a character like Burkhart could engender unwarranted sympathies for a racist killer; alternatively, doing so could comprise the most effective lesson for audiences that we’re all more capable of horrible acts than we’d like to believe – a far more effective lesson than framing Burkhart as a villain to be caught by a noble hero. Whatever you think about Killers in particular, it remains as true as ever that stories are very powerful things, and perhaps the most important thing about a story is who’s at the center of it.
* - I have no evidence for this, but Lovecraft’s short story From Beyond is so similar to The Edge of Running Water that I’m inclined to think Sloane was aware of and inspired by Lovecraft.
** - When I use the word ‘hysterical’ here I’m describing popular taste more than my own, though I would certainly apply that word to Seinfeld.