Abstract
In her article ‘The soul factor: Deception in intimations of life in computer-generated characters’ (2009), Kathryn S Egan argues that digital characters in cinema have no soul and are incapable of being empathized with. For evidence, Egan assesses the character of Gollum from Peter Jackson’s adaptations of The Lord of the Rings trilogy (2001–2003, NZ, USA). This article finds, however, that there are many who have empathized with Gollum, in direct contradiction to Egan’s findings. The author argues that Egan produces academic scholarship rather than phenomenological research, and that there is an opening for empathizing with and relating to digital characters such as Gollum. He posits that further research is required in the area of mediated characters and relationships, and that more precise language is required around terms like ‘empathy’ to accurately communicate the nuance in relationships between real people and digital or artificial characters. In addition, he suggests a term, ‘directed empathy’, to evoke the kind of empathetic relationship that exists between digital characters and viewers.
In her article ‘The soul factor: Deception in intimations of life in computer-generated characters’ (2009), Kathryn S Egan argues that digital characters in cinema have no soul, and thus cannot be the source of any form of meaningful connection between a viewer and that character (p. 106). She treads relatively familiar and even understandable ground in making this argument, moving from Aristotle’s conception of the soul to Edith Stein’s work on empathy to a computer’s inability to calculate the spontaneous actions and reactions that characterize human consciousness and action. Her analysis centers around the character Gollum in Peter Jackson’s The Lord of the Rings: The Two Towers, ultimately concluding that this character’s attempts at vying for empathetic connection are deceptive and hopeless.
My argument in this article is that Egan’s work is a fine piece of scholarship, but is emphatically not phenomenology, meaning that it eloquently uses phenomenological scholarship and logic to make a case while actually ignoring the entrenched, lived experience of people – including Egan herself. Looking once again at the character of Gollum and the personal experiences of my wife and other viewers, I argue that, in spite of the rigid logic that Egan brings to bear on the issue, there is a phenomenological precedent for the consideration of computer-generated characters as subjects which engender empathetic engagement and reactions from real, human viewers. The implication of such a consideration for the broader field of phenomenology and media studies is the requirement for greater precision in the language we use to consider characters and artifacts which we may be too quick to write off as entirely artificial. To this end I would suggest the concept of ‘directed empathy’ for consideration as descriptive of the ways that digital characters appeal and relate to human viewers.
Egan, soul, and empathy
One evening in early 2016, my wife and I sat on the couch, watching The Lord of the Rings: The Two Towers. I had seen the film many times, while this was her first exposure to Peter Jackson’s depiction of Middle Earth. On the screen before us, the slinking character Gollum deftly crawled, head-first, down a sheer cliff face. His large eyes glowing in the dim light, his ragged, stringy hair hanging down over his face, Gollum muttered to himself. He readied his skeletal body to pounce on the two Hobbits – Frodo and Sam – lying beneath him, preparing to kill them to obtain his ‘Precious’, the One Ring that had enslaved him and his will. As Gollum’s hand neared Frodo’s head, the Hobbits jumped into action, having only been pretending to sleep to trap the creature they knew to be following them. In the struggle that ensued, Gollum viciously fought back against his would-be captors, biting, shoving, and clawing in an attempt to simultaneously escape and obtain his precious trinket. Beside me on the couch, my wife writhed uncomfortably watching the scuffle, cringing and gasping audibly throughout. The scene ends in a breathless face-off, Gollum choking Sam, Frodo pointing a sword at Gollum’s throat and bargaining for the release of his companion. Knowing that he has been beaten, Gollum lets out a pained cry. Gollum’s screams echo as the scene switches to another encounter with Frodo, Sam, and Gollum. The two Hobbits lead Gollum by a rope tied around his neck, which the gangly creature cries is burning him. Gollum begs for the rope to be removed and negotiating ensues. Towards the end of this second scene, with Gollum’s pale face large in the frame, my wife reached over and paused the film. Turning to me and gesturing to the image onscreen she said, ‘Now I know that that isn’t a real actor, but – is it?’
Particularly with the character of Gollum, sentiments such as these are not entirely uncommon. One reviewer even went so far as to say that ‘The only protagonist with any human range is Gollum, who is computer-animated’, naming him ‘the only character worth watching in Lord of the Rings’ (Sands and Monahan, 2002). Even today, Gollum’s appearance and performance in the concluding two chapters of Jackson’s trilogy are considered to be watershed moments in the creation of convincing, realistic digital characters (Blair, 2007; Letteri, 2013; Prince, 2012). Indeed, many in Hollywood believed Andy Serkis’s performance as the forlorn creature to be worthy of an Academy Award (Fleming, 2002; Holloway, 2011), though questions lingered about just who was most responsible for the character’s performance onscreen. Was it Serkis, whose voice, motions, and presence on set established the baseline performance for the character, or the litany of animators and technicians whose work was also responsible for the character’s appearance and many of the subtleties of his behavior in the finished film?
Putting aside issues of performance attribution, the presence of visually and emotionally convincing digital characters raises significant questions regarding the role that digital elements as a whole have on the cinematic experience. What makes digital elements of cinema, particularly digital characters, emotionally and empathetically compelling to viewers? Why is it that some digital characters fail in attempting to be compelling? Does the digital nature and origin of an element or character alter the experience of the film they help comprise?
In Gollum, we see a wholly digital character with a notably convincing visual appearance of reality, and with a human actor at the core of his performance, usefully exemplifying many of the issues at stake in these discussions – no wonder that Egan chose him as her object of analysis in examining the idea of digital characters having heart and soul. Egan specifically focuses on the fact that Gollum’s creators used these terms (‘heart’ and ‘soul’) to describe him, with such descriptors being evidence of the ‘realness’ of Gollum and other digital characters, though she contests the conclusion that these qualities exist or contribute to the formation of a relationship between these characters and a film’s viewer: ‘I argue’, she says, ‘that the claim for “soul” in these characters is ironic: soul is what they lack, and therefore there is no hope to be found in them’ (Egan, 2009: 106).
Egan begins her exploration into the ‘soul factor’ of computer-generated characters with Aristotle, noting his belief that ‘the soul is the principle of animal life’ (p. 107), and further identifying independent movement as a significant indicator of the presence of a soul. From here, her discussion ventures into the realm of empathy, referencing Edith Stein’s (1989) research and even her definition of the phenomena: ‘Empathy . . . is the givenness of foreign objects and their experiences to a psycho-physical “I” that is body and soul together’ (p. 107). From here, Egan elaborates on the process of computer animation, crediting both animators and computers for facilitating the process of engendering animated characters with life through the appearance of volitional movement, which she says is ‘evidence of soul’ (p. 109). Quickly, however, the notion of animated characters having soul is quashed by the logic of limitations inherent in animation technology: The goal is a perfect reproduction of reality. But the computer is limited because it processes information in bits of atomistic information that cannot arrive at anything that approximates our understanding of the world. What the computer cannot calculate is soul. (p. 110)
The standard that Egan sets here is the ability for a character to engender empathy, or to ‘convey that the brain is driving the action’ (p. 110). She acknowledges that some classically animated characters have been successful at accomplishing this feat (though she does not tell us which), but this seems not to bother Egan’s argument because these characters are intentionally not photorealistic, or in any way making an implied claim to being present alongside real, human actors in a film.
Beyond empathy, Egan explores the notion of authenticity in digital characters through the concept of anomaly. She specifically details how the movements of an animated character’s eyes are essential in imbuing qualities characteristic of life. This is not merely attentiveness to a character’s gaze, eye-line, or consistency in blinking, but rather specific glances or seemingly unconscious behaviors that provoke a sense of consciousness behind the gaze of the animated character. Details such as these are essential for a character’s believability to an audience.
It is worth noting that Egan’s probing of empathy and authenticity in seeking answers to the conundrum of Gollum and other digital characters is not unique. Rather, such avenues of exploration are foundational to the practice and phenomenological inquiries of fiction at large, particularly within the cinematic medium. Questions of empathy are prominent in all sorts of scholarship relating to film and animation (Misselhorn, 2009), in part due to the cinematic medium’s inherent history ‘as a medium capable of exciting all the bodily senses’ (Stephens, 2012: 529). Anomalous behavior as an indicator of authenticity is particularly noted as a factor in the exploration of digital effects and characters because of the way human beings interpret and enact sincerity. Animators working on The Life of Pi (Ang Lee, 2012, USA, Taiwan, UK, Canada) strove to introduce elements of spontaneity into the animation of the tiger Richard Parker so as to avoid anthropomorphizing his behavior, and to circumvent the appearance of a rigid, obviously staged performance (Giralt, 2017: 13).
What these examples serve to show is that the topics Egan uses to evaluate Gollum’s appearance and behavior for the presence of a soul are not foreign concepts to either animators or film scholars. She largely admits these qualities to Gollum but remains unconvinced that empathy towards him is possible for ontological reasons. She emphasizes that Gollum is: inauthentic, because no matter how like a living person the computer-generated image may be, he is not really a ‘person’ but rather a composite of idealizations and typifications of Others . . . He fails to appear as a phenomenal being constantly in flux, graspable only in an essential intuition . . . Therefore, the Gollum is a deceitful sign, a fiction constructed by the animator. (p. 113)
Egan concludes: I look at Gollum’s face looking out from his world seemingly into mine (but never right at me), and know that, having no body of his own, and, therefore, no possible psycho-physical relationship with me, there can be no empathy. (p. 114)
Yet, the question that my wife asked in response to her emotional experience persists. Gollum’s introductory scene had been fraught with tension, and the violence had been enough to elicit genuine emotional and physical responses from her as she concentrated on the heroes, Frodo and Sam, and the danger that they were in. The following dialog-heavy scene allowed time for the shock to settle and Gollum’s appearance to sink in. His movements and his emaciated, sickly-seeming body seemed entirely foreign to what a real, human actor would be capable of doing or looking like, but even with his face filling the frame in high definition, Gollum looked real, looked like he could be a real person. More importantly, Gollum felt real, and this is the critical area of investigation. Digital characters are no longer a novelty, but the modeling, animation and performance of this character in the (then) 14-year-old film was in fact more convincing to her emotionally and aesthetically than the multitude of others that have become so commonplace since Gollum. How, then, are we supposed to make sense of this?
Bringing phenomenology to bear
Egan’s approach to the question of empathy in digital characters is highly rational and engages the phenomenological literature one would expect to inform such an investigation. But, for all its emphasis on the phenomenal, her article not only minimally applies phenomenological methodology itself, but it also comes to conclusions that are in stark opposition to the phenomenal experience Egan herself admits.
Following Husserl’s principles of concentrating on things and experiences themselves, Max van Manen (2016: 48) urges that the aim of phenomenology is not for arguing a particular understanding of the meaning of an experience, but rather for showing how meaning is revealed in that experience. More than simply assembling sources and citations to make an argument, phenomenological research ought to privilege the description of experience as it applies to a particular phenomenological question. Van Manen even titles one section of his book ‘The Research is the Writing’ to emphasize this point (p. 389).
In contrast, Egan’s analysis clearly privileges the logic of her research even over the evidence of her own experiences. Here, Egan briefly verges on practicing phenomenological description as she discusses her engagement with Gollum as she watches The Two Towers: In the case of Gollum, I surround myself with his situation in the context of the film. I see him hissing and moving erratically – the mood emerges and I experience perceiving Gollum in anguish. His anguish is not my own, but I am pulled into it and experience it in Gollum’s place, and then I am to feel his anguish though it is separate from me. Gollum is represented to feel that anguish primordially and I am led by it to experience it non-primordially . . . I look on Gollum’s struggle with compassion. (Egan, 2009: 112–113, emphasis added)
In her own way, and in her own words, Egan describes having an emotional encounter and connection with Gollum – just as my wife did. But she ultimately concludes that even given the experience she describes, ‘empathy eludes me’ (p. 113) in regard to Gollum. This closing statement leads directly into the passage quoted further above, where Egan applies the logic of Stein’s examination of empathy to Gollum’s presence in the film, which is where Egan’s account becomes problematic. In privileging the explorations of others above her own experience, Egan characterizes her empathy with Gollum as ‘self-deception’, noting that she has ‘falsely inferred, on the basis of my own orientation, that he has a soul’ (pp. 113–114).
Such a perspective fails to privilege the primacy of perception in fictional experience above the reflection that eventually follows that experience, ignoring that the starting point of what one perceives and experiences is essential to the doing of phenomenology. The pre-eminent phenomenologist of perception, Maurice Merleau-Ponty, noted that if we discuss phenomena as illusion or illusory than we have already presupposed categories of real/illusion rather than describing what is evidently true to our perceptions at a given moment or in a given experience (Merleau-Ponty, 2013: lxxx). Don Ihde (2012: 18) likewise emphasizes the importance of recognizing what is given to our perceptual faculties in a given experience: Thus, while it may or may not be the case that the dark shadowy figure I perceive in the hallway really is a person rather than a hatrack [sic], at the moment of the shadow-presentation, I cannot doubt that the presentation is given as I see it, within the limits of the moment of perception.
In addition, Egan’s application of Stein’s conception of empathy and authenticity may be right in concluding that Gollum and other digital characters are soulless, but this may not actually matter given Stein’s own definition of empathy. The portion of Stein that Egan quotes emphasizes the givenness of foreign objects to a particular ‘I’ that is ‘body and soul together’ (Egan, 2009: 107), but this only comments on the one feeling empathy, not the source of those empathetic feelings, i.e. the one who has the capacity for empathy is defined as requiring a body and soul intertwined, but this says nothing specific about the foreign objects for which the receiver feels empathy. Indeed, Peter Shum’s (2012) analysis of Stein’s work merely emphasizes that what one feels empathy for is inherently ‘something which is alien, and not something which belongs to one’s sphere of ownness’ (p. 178). Digital characters such as Gollum certainly fit the bill here, and the experiences of empathy that viewers have with these characters – my wife’s visceral reaction to the onscreen action and the Oscar-buzz that Serkis’s performance as Gollum garnered here, acting as evidence of audience identification or emotional connection with that character – cannot be discounted. This empathy may not be empathy in the same fashion that empathy with another living, soul-possessing human may be, but it is arguably still a form of empathy nonetheless. Such a notion of a ‘form of empathy’ is an idea that I will return to below.
What the problem of analyzing Gollum in relation to empathy ultimately points to is a lack of precision in the language that we sometimes use to describe our experiences of mediated Otherness, thus providing a clear area of continued investigation and expansion for scholars. Many phenomenologists – including Merleau-Ponty and Ihde – tend to describe the emotions and experiences elicited by mediated phenomena as ‘quasi’ emotions or experiences. Technology, in Ihde’s (1990) description, is a ‘quasi-“other”’ that bears a ‘quasi-autonomy’ from users (pp. 98, 100).
These descriptors are not inherently poor by any means, and may, in fact, be found to most comprehensively express and describe these experiences. However, Van Manen (2016) convincingly argues that the practice of phenomenological description ought to emphasize accuracy in terminology so as to approach the innate awe and wonder of phenomena in relation to experience (pp. 43, 52). This involves being detailed in one’s description, but also using evocative and even poetic language to get at the heart of an experience (p. 221). With this belief in mind, I would argue that the emotions evoked via mediated experiences ought not to be characterized in terms such as ‘quasi’ or any of its synonyms. This strikes me as imprecise and ultimately unfair to the legitimate emotions that are provoked in the spectator.
Towards ‘directed empathy’
I am aware that up to this point in this article I may be criticized for the same fault that I have leveled at Egan: here I have presented a scholarly assemblage of argument instead of doing phenomenology. But my intention thus far has been merely to illustrate areas where significant work in phenomenologies of digital characters and technologies is still to be done. Though there are fundamental consistencies in experience over time, the blanket application of old ideas to relatively new phenomena is not, I hope I have shown, always the best way to approach things. In my reading, Egan’s own characterization of her experience with Gollum as self-deception is self-deception, just not of the type that she argues it is. It is my belief that Egan, like my wife, and like countless other viewers, empathized with and through Gollum more than they may care to admit. Just what form of empathy this was, or how it was made or experienced, remains to be more fully explored.
To this end, I would like to engage in a kind of phenomenological thematization to suggest a concept and a system for describing the empathy that emerges between human viewers and digital characters in contemporary cinema. This is not pure descriptive phenomenology, but rather an aggregate of my own reflective experience along with some selected support from online fan and critic reviews of The Two Towers that evidence my claims for Gollum’s empathetic relations with viewers. In this I have attempted to engage not only the phenomenological scholarship surrounding issues of empathy, but those in whom empathy with the digital character was felt.
The concept that I am elaborating on here is named ‘directed empathy’, and is intended to imply that the human characters in a film direct the viewer’s empathetic experience with the film’s digital characters, using the status of the human characters as human to legitimize the emotions a viewer may feel about the digitally animated character. There are four elements that contribute to this conceptualization: (1) genuine, convincing human performances; (2) sufficient screen time for dramatic development as a character; (3) a moderately realist, multisensory aesthetic; and (4) photographically relative visual verisimilitude. In adequately explaining and supporting the idea of directed empathy, the metaphor of a four-legged stool is apt, with each of these four elements comprising one leg of the stool: a four-legged stool is strongest when all four legs are functioning perfectly; when one is weak or taken away, the stool may still stand, just not as surely as before; and to take away two or more legs significantly decreases the support that the stool can offer. Similarly, taking away or weakening one of the factors of directed empathy may not significantly alter a viewer’s experience with particular characters but, if any of the other factors are diminished, the viewer’s ability to see a character as legitimate, or to empathize with them, is likely to disappear. The first three of these four elements may be applicable to a viewer’s relationship with any character in a film, whether digital or played by a real, human actor. The fourth, however, is distinctive in its application to digital characters. Important to note is that these four elements are not four explicitly required elements or steps that are necessary to enable a character to be legitimized or an audience to feel empathy with them. Instead, audiences vacillate (to use the language of phenomenologist Vivian Sobchack, 2004) between their reliance upon each of the four factors in regulating or enabling their engagement with the character at any given time.
Genuine, convincing human performances
Gollum is not a fully-formed, engaging character right off the bat. As introduced in The Fellowship of the Ring, Gollum merely fulfills the trope of a lurking, scheming villain. In his introductory moments in The Two Towers, he may be seen as exemplifying this same trope. Crawling down to attack the ostensibly sleeping Hobbits does not immediately position Gollum as a significant character. However, even in his approach towards the Hobbits, Gollum’s presence elicits an emotional response from viewers, one of fear or anticipation. This is not directed towards Gollum, however, but rather towards the Hobbits, a response that carries throughout the scene. Gollum’s presence in the situation is what initially provokes the scene’s emotion, but the excitement we feel is directed towards the Hobbit characters as we fear for their safety and well-being. There is also something significant in the performances of Frodo and Sam in this struggle: never once in the scene does the viewer doubt that the Hobbits are under attack. From their exasperated grunts to the pained expressions of surprise and fury on their faces, the fear produced initially by Gollum’s attack is perpetuated and emphasized by the actions and reactions of the Hobbits. Their convincing pain and confusion amidst the bedlam of the tussle justify the tension audiences feel at Gollum’s approach – viewers are right to take the threat of Gollum seriously because, as the Hobbits themselves make clear, he is a legitimate threat.
Frodo and Sam’s clear engagement with Gollum at both physical and emotional levels primes viewers for the scene immediately following, where the harrowing emotion felt in the physical struggle is replaced with nuanced emotions that invite Gollum to be considered as more than merely a trope-fulfilling device of the plot. For one thing, in this scene, Gollum himself displays a new set of emotions – he is crying out in pain at the elvish rope around his neck, not hissing menacingly or physically attacking the heroes. Frodo and Sam’s reactions are significant here as they discuss the problems raised by Gollum’s shrieks, and debate over what to do about the problem. A bargaining discussion breaks out between Frodo and Gollum, which is interrupted by Sam, bringing about a tense discussion between the Hobbits. Frodo’s exasperation at Sam is noteworthy here because it is an emotion felt on behalf of Gollum by Frodo. Although Frodo is himself surely frustrated with Sam for reasons emanating from himself, the pity that then plays across his face for Gollum is also certainly a catalyst for his feelings towards Sam. Effectively what you see here is that Frodo feels empathy with Gollum.
Frodo’s visibly recognizable emotions in the scene form the basis for a kind of transfer of the emotional experience in the film, or a rupture in expectations surrounding the availability and tangibility of emotion to be found in a digital character. His recognition of Gollum’s consciousness directs the viewer’s recognition of Gollum’s consciousness, and paves the way for the character’s independence and emotional vitality within the narrative. Because Frodo sees Gollum as a legitimate Other, viewers are invited to do so as well. And, importantly, they have. Fan reviews emphatically mention the ways that Gollum ‘doesn’t feel disconnected from us [viewers]’ (Loving_Silence, nd), noting that ‘there are areas where you feel sorry for the poor guy, or just want to kill him’, declaring Gollum ‘the most diverse character I’ve ever seen’ (JohnLennon1985, nd). That such comments are always closely tied to the relationship between Gollum, Frodo, and Sam speaks volumes.
The implication of recognizing the beginnings of an empathetic relationship with a digital character in the performance of a human character is that such performances must be convincing performances. Important to note is that, in the moment of Frodo’s empathy with Gollum, the viewer does not necessarily feel empathy with Frodo. Empathetic engagement with a human character is not necessary for the enabling of empathy with a digital one. Instead, the human performance must be emotionally engaging, realistic, and convincing within the boundaries and confines of the narrative in which it is contained. Though technology can certainly be at fault for poor-looking digital characters, even in the contemporary era where creating realistic images is far easier than it has ever been, it may be argued that some considerations of inferior CGI are equally a result of poor acting, or unconvincing performances. To act ‘with’ characters who will eventually be digitally placed into a film is no simple process, technically or artistically for the other human actors involved (see The Genius of Andy Serkis, 2011; Pellerin, 2003). Such a process is no doubt stressful and difficult for other actors used to having the presence of other people to direct, inform, and inspire their own performances. Asking the same level of commitment when faced with in many cases a distinct lack of presence from what will eventually be a digital character is bound to be something that not every person succeeds in doing, a factor which can contribute to the failure of a digital character to achieve compelling connections with audiences.
In the case of Gollum, after Frodo’s initial act of bringing him into full recognition as a consciousness in the film, other characters also serve to support this development. Sam is the most prominent, obviously, in that the trio spend a significant amount of time together onscreen, but the character of Faramir also later relates significantly with Gollum. The involvement of these other human actors interestingly may bring about a doubled application of some of phenomenologist Maurice Merleau-Ponty’s notions regarding movement as revelatory of the Other. Merleau-Ponty (2013) argues that an individual’s movement or motion through physical space is inherently also an act of involvement in perceiving the world. Seeing others move, then, allows me to ‘interpret their behavior through analogy with my own behavior and my own inner experience, which teaches me the sense and the intention of the perceived gestures’ (p. 364). This principle is on display in how the perceiving viewer examines both the movement of human characters on screen and their examination of the movement of digital characters. In watching Frodo, Sam, and Faramir act, their innate humanness carries with it the kind of intentional, consciousness-inhabited movement that is taken for granted as present in other human beings. We watch them onscreen, displaying their consciousness before us, even as they occupy these fictional roles. And further, their intentional movement is reflecting the consciousness of the digital character with whom they are (apparently) dynamically interacting, who is then seen to explicitly have his own consciousness, as reflected in intentional movement.
Sufficient screen time for dramatic development as a character
The Lord of the Rings films are long, complex stories that involve a large number of characters in many different locations. Yet, Gollum, even as a digital creation, is not lost in the shuffle of storylines and plot points. Instead, he is incorporated into the story in an essential fashion, complete with his own independent character arc that also contributes to the overarching development and themes of the larger story. This may be illustrated through the significance of instances in which Gollum’s performance elicits an emotional or empathetic response that is entirely of his own making. This primarily occurs in a scene where he debates with himself, and temporarily banishes the disruptive and evil Gollum personality. For one thing, Gollum is alone onscreen for the duration of this scene and others like it. The story being told in these instances is Gollum’s, particularly in this first scene. Later scenes involve Gollum discussing other characters and his relationship to them, but this is the first major moment in which Gollum is considered on his own, as himself, by himself.
And this moment succeeds in being dramatic, in drawing viewers in to the drama of Gollum’s split personality, inviting them to root for him as he attempts to liberate himself from such a devious, pervasive captor. As one fan reviewer noted, ‘there was so much emotion in this character that I couldn’t help believe he was real’ (docmonster, nd) and moments like this one are paramount in provoking that kind of experience When he succeeds in throwing off the bonds of his own darker half, we are excited for him, in addition to being happy for what this means for Frodo and Sam in terms of the dangers that the split personality of Gollum previously presented for them.
This is not to say that Gollum only develops in this scene, or other scenes where he is by himself. Throughout the second and third parts of the trilogy, Gollum is an essential part of the trio alongside Frodo and Sam, with the three characters dramatically pushing and pulling at one another in compelling ways to inspire growth and development. But what this moment of solely focusing on Gollum does is emphasize that Gollum as a character has value beyond his ability to interact dynamically with the two Hobbits. It legitimizes his presence in the story for the sake of his own character, not just the larger plot, and also illustrates that, by this point in the story, we already care enough about the character on his own to engage dramatically with him at the level of empathy in his moment of triumph.
Related to this is some consideration of the actual shots in which empathetic emotions are stirred in audiences. The element that stands out in consideration of many of Gollum’s scenes is the sheer amount of time that we spend with him onscreen. What this suggests is that effective digital characters must have sufficiently long takes and sequences in which to demonstrate their emotion. As a rule, montage is highly effective at juxtaposing imagery for informational and emotional effect but, in the case of a digital character, montage may ultimately defeat the purpose of imbuing such characters with a vitality of character with which audiences can empathize with. Instead, time is necessary for the audience to fully apprehend and appreciate the character’s visible emotions. Perhaps the term ‘realist aesthetic’ pushes the concept too far – this is not to say that films must be comprised of long takes or fewer cuts, but merely that some longer takes and instances must be provided for the digital character to perform, rather than merely appear in various shots emoting as is necessary for the single shot’s impact within the montage. This principle, it must be noted, holds true even in relation to real human actors – an actor’s physical presence in the midst of a harried montage does little to emphasize or gild their performance, but instead turns them into objects merely pictured to elicit emotion.
A moderately realist, multisensory aesthetic
Related to the visual component of long takes, there are other sensory elements that may contribute to the grounding of a digital character in the world of a live-action film alongside human actors and real, physical elements. These may also be considered under the banner of a moderately realist aesthetic, with the word ‘multisensory’ as an addendum. After film’s visual elements, sound is perhaps the most obvious area to explore, and Don Ihde’s (2007) work on the phenomenology of sound contributes to our understanding of how this omnipresent phenomenon establishes the overall sense of reality in a film all around, and with characters specifically.
Digital imagery is significant in that it opens cinema up to a host of questions regarding ontology and indexicality. And though similar questions may be brought to bear on audio files and sound recording, most actors cast as digital characters end up having their actual voice used in that character’s portrayal. Gollum is Andy Serkis’s voice, which is also heard in his roles as Caesar (Planet of the Apes trilogy, 2011–2017, USA, UK, Canada, NZ), Captain Haddock (The Adventures of Tintin, 2011, USA, NZ, UK), and Supreme Leader Snoke (both entries in Disney’s Star Wars revival 2015/2017, USA). Ihde (2007: 14) notes that sound is both omnipresent and has a distinct deference for invisible phenomena, which may serve to ground an unreal, digital character in the real world. If the actor for a digital character may be said to not be there at all, if the character is arguably ‘invisible’ in terms of its relationship to reality, then perhaps the sense that provides the most grounding in reality is that which is inherently linked to that invisibility. Detractors may argue about the vagaries of digital sound recording, remixing, and alteration, and they may be right to raise such objections. But, as Ihde notes, sound has often been disregarded or diminished in phenomenology, and many others have acknowledged its seeming lack of import in the realm of film studies (Beck and Grajeda, 2008; Belton, 1985; Schiffer, 2012; Sterne, 2008). Perhaps it is time to give sound its due, and perhaps this is one area where the use of sound may be acknowledged or even praised. To my knowledge, there has yet to be a digital character whose sound is entirely divorced from some human origin. Some robotic characters may arguably fit that bill, but in terms of digital main characters in film, it may be that a human (or animal) voice is the only connection to reality that the character actually has.
In addition to the visual and aural senses, Vivian Sobchack’s (2004) notion of the ‘cinesthetic subject’ includes elements of taste and touch as prominent areas of discussion. Sobchack notes that ‘we do not experience any movie only through our eyes. We see and comprehend and feel films with our entire bodily being, informed by the full history and carnal knowledge of our acculturated sensorium’ (p. 63). Sobchack is specifically attending to the senses of touch and taste that are easily evoked in the cinema and yet generally left behind in consideration of the visual and aural aspects of film. The examples that she uses rather descriptively are unspecific and yet universal – the experience of seeing food onscreen and tasting it on your tongue, or of seeing a texture on a piece of clothing and simply knowing in your own body how that clothing would feel. The thrust of Sobchack’s argument is that, as viewers seated in the theater watching a film, our intentionality is directed towards the images before us onscreen. All of our senses are engaged – we watch the film sensibly – but which senses we rely on at different times vacillate in response to what is happening in the film. Sometimes visual elements lead our identification with the cinematic, cinesthetic subject, other times perhaps we lean more heavily upon our sense of touch or taste to drive our engagement with the cinematic world.
With specific regard to Gollum, the element of taste is fairly minimized. At two different points in The Two Towers and The Return of the King, Gollum eats raw meat – taking a brief bite of a rabbit in the former and nastily crunching a fish in close-up in The Return of the King. The rabbit moment is rather quickly deflected by Sam coming in and taking the rabbit from Gollum to prepare it properly, but the fish is prominently played up for all its multisensory, cringe-inducing possibilities. The visual of the bite is close-up; the sounds of the crunch and the sopping wetness of the bite send shivers down the spine; and though there is not, typically, a taste sensation when Gollum takes the bite, there can be a kind of oral retreat that happens in the viewer’s mouth, as if you were being made to feel the slimy texture in your throat and taste the fish’s pungent flavor. Sobchack’s emphasis on the embodied experience of viewing accounts for this phenomenon, and the disgust it evokes in viewers.
Additionally, there are elements of touch and tactility that are quite prominent in Gollum’s actions, nicely supported by his overall behavior. Throughout the films, Gollum is seen as a capable climber, often scrambling up sheer rock faces and jumping off of gut-wrenching precipices. For all his slinking, creeping qualities, Gollum is often lithe and impressively deliberate in his movements. Other times, he scrabbles and flops with reckless abandon, slamming his body against the ground and the surrounding rocks with seemingly no regard for his own physical well-being. Watching these moments bring alternating feelings to the viewer’s body. When Gollum crashes into what looks like razor-sharp rocks, our bodies cry out as if they too are hurting; and when Gollum elastically and elaborately jumps and moves around we almost feel like we are getting into his rhythm, like our bodies could do those same things if we were in his position. Though we remain seated we feel like we too are dodging, jumping, rolling.
Sobchack (2004: 79) notes that these responses are ‘not consciously reflective’, meaning that we do not think about our bodily responses to a film, because to do so would be to jar ourselves out of the direct cinematic experience. Instead, we respond bodily ‘as only one side of an irreducible and dynamic relational structure of reversibility and reciprocity that has as its other side the figural objects of bodily provocation on the screen’ (p. 79, original emphasis). In our direct experiences of the film in all its multisensory facets, our bodies are often bound up with the narrative such that our bodily experience is not yet differentiated into the neat categories of reality and illusion. Sobchack here finds reasoning for why some people become sick in some movies or cover their eyes to protect themselves from some sights (p. 79). While our reactions to Gollum are unlikely to be so extreme, we nonetheless feel and respond to his body with our own in subtle, yet important ways.
A related conceptual tool that aligns with the notion of a multisensory aesthetic is Lilly Husbands’ conception of a fantastical empathy that provokes feeling into or along with even abstract imagery. Citing the presence of the ‘gap’ between onscreen stimulations and the body’s sensual reciprocation, Husbands (2018: 80) argues that ‘Fantastical empathy attempts to bridge the gap because, phenomenologically, we are never free from our compulsions to assimilate new stimulation by associating it with what our minds and bodies already know.’ Intentionally evoked here is Sobchack’s cinesthetic subject, who brings his or her prior lived experience to bear in redirecting sensual responses to familiar figural objects, even if, like Gollum, they are not totally known or understood.
Photographically relative visual verisimilitude
The elements of directed empathy that I have discussed so far may be argued to apply to any character, not necessarily exclusively digital ones. Here we come to that element which does explicitly distinguish the digital character – the element of visual realism. What I wish to suggest is that some measure of visual verisimilitude is necessary in order for a digital character to appear realistic in a live-action setting, beside a real, human actor. But this verisimilitude does not need to be the kind of hyper-realism that characterizes many contemporary digital characters and effects.
Julie Turnock (2012) helpfully elaborates on the idea of ‘photographic realism’ in film as being distinct from ‘realism’. Turnock’s perspective emerges from the special effects photography/cinematography that Industrial Light & Magic (ILM) established in the 1970s, which sought to integrate animated or modeled elements seamlessly into an already photographed reality. This meant that aesthetic and design choices were made in the creation of the effects to match with the cinematography and other photographed elements of the film in which the effects were intended to appear. Just as elements on set were altered from reality through their being photographed, so too were the effects changed to give that same feeling of the altered, photographed reality that the film’s director and cinematographer were aiming to achieve. Thus, the effects were not intended to have a pure basis in reality, but rather they were intended to fit into the photographed, already artificial reality that the film was seeking to create.
Turnock argues that this aesthetic has continued into the contemporary era as cinematography and other shooting choices in some modern films are designed specifically with the effects in mind, so that the capabilities of digital effects to mimic photographic reality end up driving the cinematography of the film in which they are intended to appear (p. 165). This leads to the conclusion that effects do not necessarily have to maintain a basis in reality to be accepted but must maintain a consistency with the filmic world in which they appear. The aesthetic of the overall film drives the aesthetic needs of the effects, such that the verisimilitude of the effects depends rather largely upon audience acceptance of the film’s broader aesthetic and how those effects convincingly fit into that aesthetic. This could, in fact, mean that contemporary films may be progressively becoming less and less ‘realistic’ (by historical standards) in order to compensate for deficiencies in the capabilities or appearances of their digital effects elements.
Lilly Husbands (2018: 80) notes that ‘the tendency in feature-length narrative animation towards photorealism can be seen as an attempt to narrow [the gap between visual stimulus and our body’s sensual responses] and to smooth over the sensual dissonance created by computer-animated imagery.’ However, this issue of lifelike realism is something I find remarkably distracting in even recent releases featuring prominent digital characters. Supreme Leader Snoke in Star Wars: The Last Jedi serves as the most prominent example in recent memory. The character model of Snoke (also played by Andy Serkis) is detailed to the point of distraction; in one close-up where Snoke and Daisy Ridley’s character Rey stare menacingly into one another’s eyes, I found myself glancing back and forth between the two characters’ eyebrows. Snoke’s were hideously bushy with nearly every hair on his brow clearly distinguishable and identifiable; Rey’s, in contrast, were a simple black line with a distinct shape to the overall brow, but with nowhere near the amount of detail. It was strange to me how it seemed that the human characters had effectively lower fidelity in their visual appearance than did the digital character. I find myself agreeing more with Daniel Dockery (2018), who argues that what filmmakers lack now is not the power of technology that enables the creation of realistic images, but rather the restraint that the most effective use of that technology requires.
What perhaps matters most is that filmmakers effectively communicate the physical traits or characteristics of a character or effect that audiences most expect. Helpful here is Stephen Prince’s (1996) notion of correspondences as a procedural explanation for how we assess the reality of appearance and behavior in digital characters. In examining the creature special effects in Jurassic Park, Prince notes that the images of dinosaurs are ‘perceptually realistic’ in that, even though we know they are not (in fact cannot) be real, they still appear to be in part through a complex process of correspondences in which viewers assess whether the images onscreen hold up to the things which we know to be true about the world. Prince writes that ‘these correspondences which anchor the computer-generated image in apparent three-dimensional space, routinely include such variables as surface texture, color, light, shadow, reflectance, motion speed and direction’ (p. 33). In this manner, a dinosaur may be evaluated based upon what we know about large animals in the real world. Based on the appearance of the dinosaurs’ skin texture and their sluggish movements, they may be argued to be consciously evoking the movements and appearance of creatures like elephants or rhinos, and so we might assess the validity of their movements and appearance based on what we know about these real animals. Thus, a certain level of realistic detail is required in the appearance of a digital character, as is some semblance of realistic, multisensory expectation-satisfying movement and behavior, but what Prince once again emphasizes is the importance of perceptual realism over the exacting hyper-realism that is emphasized in innumerable behind-the-scenes featurettes about CG characters in the wake of Gollum (see, for instance, Pellerin, 2006, 2013).
What all of this suggests is that visual verisimilitude is important to some extent, and that there is a minimum threshold of verisimilitude required to create a convincing digital character. But the need for continuity with the other visual elements may create a maximum threshold of realism and detail that may prove to disenchant viewers, even if it is executed in a way that would otherwise suggest circumventing the ‘uncanny valley’. A film may establish its own definition of what appears to be visually ‘normal’ in that story’s world, and so our assessment of what passes for realistic (or realistic enough) may be made against the film’s visual grounding. Additionally, as has been emphasized above, the multisensory component of a film means that visual verisimilitude may also be mitigated and negotiated by other sensory factors that contribute to the overall perceived realism of a digital character.
Conclusion
The system of ‘directed empathy’ that I have outlined here is an initial stab at articulating the experience of watching Gollum in The Lord of the Rings films and feeling something with and through his character, a kind of empathy that simply was not accorded to his character in the literature of phenomenology or film studies. I recognize that the system may be found lacking, that there may be deficiencies that need shoring up, or elements that may need to be reconfigured in their entirety. My hope is that this initial exploration sparks conversation and future research on the matter, and refinements in the language used to describe and discuss relationships between viewers and various types of characters onscreen.
One interesting consequence of the notion of ‘directed empathy’ is that it may point to the idea that the experiential differences or consequences that a digital character brings to the film seem mainly to be one of heightened attention to the other, real elements within the film. As I recently watched The Lord of the Rings, I noticed that my gaze was often drawn less to the specific actions of Gollum and more to the things with which he was interacting, as well as the people he was looking at and who were looking at him. In this way, the digital character may be said to alter our experience of the film they are appearing in by actually heightening our sensory engagement with the film. Whether consciously or unconsciously, we assess these characters based on their interaction with many different facets of reality, calling upon our multisensory experiences with reality to validate what it is we are experiencing with the film and its characters.
When my wife first paused The Two Towers to ask me about Gollum, I believe that she was experiencing some of the same sorts of tension that many film theorists and thinkers have undergone in their consideration of digital characters. She knew at one level that what she was seeing could not be possible in reality, and yet the evidence of her experience of the film was such that she believed in what was there before her, thus ascribing to Gollum all the emotional connection that the film to that point had suggested he was due. As many others have noted, this is one significant aspect of the cinema’s incredible magic: its ability to carry us away to fantastic realms and cause us to engage and feel something with the most unlikely and, perhaps, unimaginable of creatures. This, I believe, truly is as close to magic as we typically experience, and while some may feel that the magic ought to be more logically thought out and explained, my final argument is that the best and most appropriate response is simply to acknowledge the experience for what it is, and, maybe, to revel in it just a little, too.
Footnotes
Funding
The author received no financial support for the research, authorship, and publication of this article and there is no conflict of interest.