The Cybernetic Victorians

Abstract

Who would believe that novel characters are real people? It seems like a rhetorical question. In critical practice, much as narrative scholars deplore such credulous readers, we constantly invoke and imagine them. Like the literary present and the Romantic genius, the reader to whom characters seem real is a seductive hermeneutic tool and a difficult one to abandon. It is at the heart of several popular approaches to novel studies, ranging from new historicism, through ideology critique, to the ethical turn. Depending on the argument, a critic might trot out several varieties of these literary naïfs. Sometimes, they are slightly younger versions of ourselves, to whom great novels offer masses of social information they have not yet experienced elsewhere. These adolescent selves help us explain the novel's philosophical and ethical import. Or else they are much younger versions of ourselves, from a time when our imaginary friends and the real world blended with each other. Such childlike readers undergird theorizations of novels as charming fictional worlds or as hypnotic ideological devices. Still other credulous readers we invoke come from an even earlier stage of human innocence. Indigenes to the novel's prehistory, they bow and weep before Pamela in living proof of the powers of realism.Megan Ward's Seeming Human opens with a bold question: what if we tried to sidestep this rhetoric of extreme readerly naivete? What would characters seem like if we approached their imitation of life with an emphasis on imitation? For Ward, this is not merely a formal question: a question of, say, how reported speech interacts with free indirect discourse or how many perspectives a novel gives us on its fictional beings. Instead, it involves both form and reader response. Ward places us in the uncanny valley, where characters seem both clearly scripted and oddly interesting; she hunts down moments where some reptilian part of the reader's brain might start to anticipate their animation despite protests from the frontal lobes. As critics, we like to swoon over authors who get us considerably past this point of barely suspended disbelief. Ward insists on the equal interest of the twilight zone at which the barest minimum of the illusion of aliveness is continually reapproached and retested.Although she is a Victorianist, Ward finds inspiration for this thought experiment in a branch of philosophy not usually associated with the Victorian period: artificial intelligence (AI). “Rather than reading fictional character through the study of humans,” she explains, “I turn to machines meant to seem human for a theory of character verisimilitude (2). She acknowledges her method to be a partial anachronism—and her purpose is emphatically not historical: I do not suggest a continuous historical development from Victorian realism to AI. Rather, I argue that AI offers us theories of verisimilitude that allow us to see realist character's fictionality anew. In this way of thinking, character and AI are both part of what Jay Clayton has identified as the “long history of artificial beings, which stretches from the early nineteenth century to today's headlines.” Artificial beings both replicate human subjectivity and create it anew, representing and embodying a complex set of interwoven experiences that define what it means to seem—rather than be—human. (3)Later in the introduction she describes this as a “technological formalist method, wherein technology offers us a formal and theoretical mode of reading character rather than historical context” (4).Ward sees this method, more broadly, as an instance of what she describes as scholarly work done in “the historical middle.” Wary of historicism and presentism alike, Ward insists there is value in making different periods' thought systems meet each other even when causal relationships between them are not obvious or completely provable. The appeal of this kind of thinking hinges on the productiveness of the analogies it stages. The acts of historical translation and rapprochement it undertakes try to persuade us that even if two periods and methods have not actually interacted much with each other, they somehow ought to—and we ought to be present for their imagined conversation. Ward defends this approach well, while articulating its risks with much self-awareness. By the end of the book, she convinces me of its usefulness to an extent that exceeds the explicit conclusions she draws from her unusual pairing.A windup doll graces the cover of Ward's book, appropriately. The argument she makes hinges on making novel characters seem doll-like. The characters on whom she focuses are akin to the subtype of Forster's flat characters who vibrate with an oddly seductive intensity. They are not “flat,” in the sense of being stereotypical or unengaging; yet they are also not “deep,” in the sense of surprising us with interior voices and insights. Made from repetitive gestures distributed among large character networks, they tap into our habitual responses to other people's equally habitual social behaviors. Like the AI assistants installed on our phones, they thereby activate the automatisms of our socializing patterns rather than any wells of empathy or recognition.In four chapters, Ward tests out four versions of this aesthetic effect. The first chapter explores the model of “the feedback loop” in the novels of Charlotte Yonge, Elizabeth Gaskell, and Margaret Oliphant (12). Chapter 2 considers “the information system” as an organizing principle in sensation and mystery fiction, including novels by Charles Dickens and Wilkie Collins (13). The third chapter takes up Alan Turing's idea of “the imitation game” to illuminate character construction in Anthony Trollope. The fourth chapter discusses works by Thomas Hardy and Henry James within the context of “the perceptron and the physical symbol system” (16). Finally, a short epilogue applies these theories to more recent fiction by David Lodge.There is much to admire about these chapters, not least the flexibility with which they approach their sources. All too often, theoretical readings subordinate literary examples to abstractions. Ward never does that; it is clear that her interest in AI stems from a deep sensitivity to Victorian fiction. Nor do her readings simply suggest that the novels she studies are meta-narratives, more interested in baring the device than in constructing a credible fictional world. Instead, she argues, these more mechanical modes of character construction are concerned with the mechanization of Victorian life—and, equally so, with the blurred boundaries this mechanization creates between the human and the automaton. Each chapter explores a slightly different angle of these concerns. Some sections focus on representations of human beings as surprisingly robotic; others stress how processes that initially seem automatic acquire a surprising aura of human presence.In the first chapter, Ward shows how Yonge's, Gaskell's, and Oliphant's character construction deploys something like “the cybernetic loop” “as the fundamental rhythm of domestic realism” (21). All three novelists use repetitive household routines to create a sense of everydayness. They also do so to make a proto-feminist point about the stultifying effects of domestic labor. “[T]hese characters,” Ward argues, “develop recursively in order to eradicate, rather than develop, interiority. Lack of interiority here is not a failure of aesthetics; The Daisy Chain shows character formation based on everyday routine that neither attempts nor wants to seem natural” (21). In these characters' repetitiveness, Ward implies, Victorian readers might recognize their own. They speak to these readers' selves at their most bored and quotidian, representing the bare modicum of awareness daily routines leave them room for.Chapter 2 reconsiders Forster's famous distinction between flat and round characters through the prism of character networks. Most theories of character development, Ward argues, tend to assume that a character's perceived depth is tied more to their represented interiority than to their represented interactions with others. Ward overturns this assumption by drawing attention to what she describes as novels' “information systems.” Wilkie Collins's No Name, for instance, “asks how to write a fictional character whose unpredictability comes from her relations to others rather than in the revelation of her ‘true nature’” (55). A character can be made to seem surprising, despite her apparent repetitiveness, because of the shifting balance of new or important information she carries within a larger character system. This balance can shift not only as a result of changes to this character's represented interior states but simply because she enters new social groups and contexts.Chapter 3, devoted to Trollope, reads his fiction through the prism of Turing's famous “Imitation Game.” Turing's thought experiment considers the conditions under which a machine might be able to successfully imitate human conversation. Success, for Turing, is defined as the capacity to deceive a human being into perceiving this machine as a fellow human consciousness. Many of Trollope's novels, Ward argues, create illusions of psychological depth out of similar patterns of deception. Characters are made to seem more fascinating to each other, and implicitly also to the reader, when they scan each other's dialogues for signs of pretense or falsehood. For instance, in The Eustace Diamonds, “the value of Lizzie's and Lucy's interiorities—and their fictional exposition—can only happen through others' perceptions of the truth-value of their vows of fidelity” (77).The fourth chapter takes on two seemingly quite different kinds of characters, those of Thomas Hardy and of Henry James. In both cases, Ward argues, the illusion of consciousness is produced in these novels through a somewhat mechanized focus on the senses. Relating this mode of representation to the way scientists model artificial neural networks, Ward shows how immersive such a depersonalized focus on sensory data gathering can be. Hardy and James accomplish this effect in rather different ways. Hardy, she says, “shifts the emphasis in characterization from seeing to being seen, from being a focalizing subject to being a focalized object. This series of overlapping perceptions constructs an alternative version of interiority” (106). By contrast, “James's narrative techniques have become definitive of depth psychology not necessarily because they portray a more realistic version of the human, but because they portray a naturalized version of the machine” (107). The difference, here, is between imputing this mechanical mode of perception to an implied, depersonalized narrator (as in Hardy) and attributing it to the characters themselves.I followed these analyses with admiration and a sense of recognition. Cutting through the pieties of empathic reading, Ward lets us appreciate the degree to which novels resemble elaborate puppet shows. “The novel,” as she puts it toward the end of Seeming Human, “ultimately demonstrates the limitations of character and machine as simulations of consciousness” (124).All the same, I am not quite satisfied with such a statement as the main conclusion of Ward's monograph. As a coda to a sequence of brilliant and quite moving readings, it rings false in its modesty. It also does not fully capture the mood of Ward's analytic mode, which as I suggest above has more to do with the uncanny than with a dialectic of disenchantment. Part of the marvel of Seeming Human is that the narrative tricks Ward exposes often still work after she has exposed them. Even if the characters on whom she focuses might not inspire our philosophical attention or passionate identification—and they would probably not have done so even without her exposition—they do, still, bear that modicum of lifelikeness that Mel Chen describes as “animacy.” We seem to fall for them (or at least I do) in the same way we fall for visual illusions, by virtue of some automatized gestalt instinct.In a somewhat bolder expression of her book's ambition, Ward suggests that seeing characters as “machinic” helps us relate to the technologies around us: “Fictional characters are machinic as well as human—they capture and anticipate the varied ways in which we design technologies in our own image. As such, they offer a way through and into a more complex relationship with technology, one built not only on resistance but also on the recognition of likeness” (17). I agree—but again, such a conclusion undersells a large set of theoretical issues Ward's book opens. Seeming Human does not merely help us understand how we relate to Siri (although it also does this). It raises much larger questions about the spectrum of standards that undergird our notions of what makes a novel plausible enough to think or feel with—and about the transposability of these standards onto the ways we interact with human beings in real life.I would be curious to know where Ward would place her “human-seeming” characters in the typology of characters with which I began. For all its emphasis on AI, Seeming Human still depends inextricably on reader-response and an assumption of readerly credulousness—even though this credulousness is of a far less extreme, melodramatic or lofty kind than what we are used to seeing in literary criticism. The models of character construction she unravels are not incompatible with those around which critics like Martha Nussbaum construct intricate philosophical theories. Indeed, to think about cognitive automatisms on the same plane on which we consider how our reading practices are inflected by feelings, values, and social conventions is a fascinating proposition. It offers a middle ground between feeling and philosophy as explanatory frameworks of reader engagement. It also complements maximalist discussions of character plausibility by turning our attention to the neglected minimum at which the question of plausibility becomes pertinent at all. Were Ward moved to address these larger questions, the answers might add to the fascination of this already accomplished, thoughtful book. Whatever angle her future work actually takes, I will look forward to learning from it, as I did from this volume.