Skip to content

Kazuo Ishiguro and the Uncanny Cascade

We live in a transitional period, when the possibility of being duped by incomprehensible intelligences—and thereby duping ourselves—has grown exponentially.

· 16 min read
Kazuo Ishiguro and the Uncanny Cascade
Author Kazuo Ishiguro at the PEN Literary Cafe, interviewed by PEN Deputy President Lindsay Mackie. (Image: Robert Sharp/English PEN; Flickr)

In a crucial scene from Kazuo Ishiguro’s first novel, A Pale View of Hills, a woman named Sachiko attempts to account for her daughter’s odd behavior by telling the story of a drowning the young girl witnessed as a toddler during the Second World War. After the Allied firebombing campaign that reduced almost all of Tokyo to rubble and left most survivors inhabiting derelict buildings and tunnels, Mariko took off through the wreckage:

Mariko ran down an alleyway, and I followed after her. There was a canal at the end and this woman was kneeling there, up to her elbows in water. A young woman, very thin. I knew something was wrong as soon as I saw her. You see … she turned around and smiled at Mariko. I knew something was wrong and Mariko must have done too because she stopped running. At first I thought the woman was blind, she had that kind of look, her eyes didn’t seem to actually see anything. Well, she brought her arms out of the canal and showed us what she’d been holding under the water. It was a baby.

Whatever emotions this scene could foreseeably provoke in a reader, one overwhelming aesthetic effect is the uncanniness of the woman’s smile. One of the details Sachiko recalls is that Mariko stopped short before becoming aware that she was witnessing the killing of an infant. Something in the woman’s smile, in her unseeing eyes, simultaneously concealed and revealed a profound threat—in this case, the gravest threat imaginable to a five-year-old girl: death by maternity. For what is that smile disguising but moral vacancy, a murderous derangement that wishes to disguise itself as placid, even welcoming? Or is it the mother’s superficial self-denial of her own atrocious deed, as though the most natural thing in the world is to kill our own children? Perhaps, like Niobe, she is petrified by affliction, her capacity for expression surpassed by her grief. Whatever the case, her smile does paradoxically signify one thing to the child: danger.

It may seem strange to refer to the woman’s smile as uncanny. We are probably most familiar with the technical meaning of the term, thanks to Masahiro Mori’s concept of the uncanny valley, where the more a robot appears human without completely deceiving us, the greater the chance it will be creepy. As Mori put it in his 1970 essay, “I have noticed that, in climbing towards the goal of making robots appear human, our affinity for them increases until we come to a valley, which I call the uncanny valley.” According to Mori, we find automatons charming or cute so long as we can tell instantaneously that they are artificial (consider R2D2 in Star Wars). The moment an automaton passes a certain threshold of human likeness, however, we are repulsed by a nervous sensation—as though its exterior form conceals a deeper threat (consider the Diego-San model manufactured by Hanson Robotics).

But like the face of Diego-San, the crazed mother’s smile is a simulation—a symbolic matrix concealing the fact that she has relinquished her moral sovereignty as a mother. It betrays the inhumanity that has taken control of her through the carnage of war. Just as the machine cloaked in flesh says “I am human” when it is not, the psychotic smile of the infanticidal mother simulates humanity when the woman has become less than human. Both are indeed simulations, and as Jean Baudrillard once put it in his book Simulacra and Simulation: “To simulate is to feign to have what one doesn’t have.”

The concept of the uncanny originated in German philosophy and psychology around the turn of the 18th century. The German word for it is unheimlich, which is actually closer to the English term “unhomelike.” This more literal translation of the German coinage is helpful to consider, in part because no one would call something “unhomelike” unless it was in some fashion reminiscent of the deeply familiar (no one from Cleveland, Ohio, for instance, would visit the Sahara and think to say that the view is unhomelike). In German, the term encompasses a contradiction—a simultaneous attraction and repulsion—implying that the comfort of a certain recognizable pattern also induces fear or doubt.

In 1906, the German psychologist Ernst Jentsch wrote that the unheimlich feeling arises when a person “doubts whether an apparently animate being is really alive; or, conversely, whether a lifeless object might not in fact be animate.” Summarizing Jentsch’s argument in 1919, Sigmund Freud further noted that Jentsch “refers in this connection to the impression made by wax-work figures, ingeniously constructed dolls and automata. To these [Jentsch] adds the uncanny effect of epileptic fits, and of manifestations of insanity, because these excite in the spectator the impression of automatic, mechanical processes at work behind the ordinary appearance of mental activity.”

In addition to the uncanniness of automata and psychotic smiles, there can be an uncanniness to more ordinary human experience, as is evident from the milder and less narrow use of the word we hear in casual conversation—for instance when a strong resemblance between two individuals is mentioned. If someone says, “She bears an uncanny resemblance to my sister,” the uncanniness is obviously greater the more the lookalikes are physically indiscernible. The uncanny element here is not the resemblance, strictly speaking, but the knowledge in the observer—often momentarily delayed—that despite the resemblance, the familiar sister and the unfamiliar stranger are distinct volitional agents.

Although we probably connect the word “uncanny” to resemblance, repetition, or coincidence more or less instinctively, without full knowledge of the word’s technical significance, it is more likely that the word will be used—or, what is more important, the uncanny effect is more likely to occur—the more the resemblance would allow one of the individuals to replace the other, steal her identity, or manipulate people on the basis of their perceived similarity. The same may be said of a statement like, “This building feels uncannily familiar,” almost as though a mysterious architect designed it to lure us into a sense of vulnerable comfort.

Conversationally, uncanniness generally has to do with an imitation, doubling, or repetition that houses a possible threat, as though someone or something has deceitful potential but has yet to exploit it. The uncanny is the sudden feeling that a projected familiarity is a type of camouflage outfitting possible danger. This is why anyone who has experienced déjà vu could rightfully call the unsolicited “repetition” of events uncanny, since the momentary disorientation concerns a familiar unfamiliarity that inspires our curiosity concerning what lies beneath the veil of ordinary perception.

There is a delay effect followed by a bewilderment whenever déjà vu occurs—the intuition that we are either not in control of our own intentions, or that the empirical world, like Descartes’s evil demon, is misleading us. One of the cleverest elements of The Matrix plays on this bewilderment: famously, déjà vu signals a “glitch in the matrix,” a moment when the intelligence that enslaves humanity inadvertently ruffles the simulation-curtain to give us a glimpse into the calculating menace controlling what is presented as familiar. In The Matrix, as in everyday experience, déjà vu calls into question, even if only momentarily, the assumptions of mental life.

Masahiro Mori concludes his own essay on the uncanny valley with some questions and a hypothesis concerning our capacity to spot and name the uncanny. “Why were we equipped with this eerie sensation?” he wonders. “Is it essential for human beings? … I have no doubt it is an integral part of our instinct for self-preservation.” The reason is that our “sense of eeriness is probably a form of instinct that prevents us from proximal, rather than distal, sources of danger. Proximal sources of danger are corpses, members of different species, and other entities we can closely approach. Distal sources of danger include windstorms and floods.” In Mori’s picture, the uncanny object makes us recoil as though we have suddenly intuited that what we thought was a broken branch at our feet is, in fact, a venomous snake. What appeared to be non-agential has allowed us to get close, only to reveal that it has malicious intent; or, in the case of a robot, the apparently self-directed human inadvertently reveals that it is operated by an algorithm or some remote operator whose intentions are obscure.

There are at least two layers of uncanniness to the drowning scene in A Pale View of Hills: the affective layer attributable to the young Mariko as she approached the woman and noticed her psychotic smile, and the vicarious layer attributable to the reader engaged with the novel. Unlike Mariko, the reader experiences no proximal danger but only the reverberation of uncanny imagery—the uncanny rendered artfully. Thrilled and repulsed, we are left distantly wondering what the psychotic smile is concealing. War-torn insanity? True evil? The momentary self-deceit that surely takes place during a mercy killing? A willed blindness to the suppressed intensity of maternal instinct? Inexpressible horror at her own action? What forces are acting through this afflicted woman, and why? And how, so to speak, does such an ordinarily unfamiliar impulse emerge and transform a mother into an agent of death? Has this mother become the inadvertent puppet of some massive modern machine designed by the invisible hand of global warfare? Or—and this may be the uncanniest question of all—is this murder an act of catastrophic love, where familiar emotion is inverted and expressed in the most unfamiliar of actions?

Like so many of Ishiguro’s novels, A Pale View of Hills is packed with situations of intergenerational negligence, forsaken responsibilities, and outsourced morality. The novel’s thematic amplitude is generated, in part, because it raises more questions than it can answer about whether moral disavowal is the cause or the effect of our devastated social and physical landscapes—or whether cause and effect exist in an endless feedback loop. For the remainder of her childhood, Mariko will be so haunted by the uncanny smile of the infanticidal woman that, in effect, she will become a barometer of uncanniness, suspicious of all grinning maternal figures. In the years following the war, once Sachiko and Mariko have moved from Tokyo to Nagasaki, Mariko still reports seeing the smiling mother on a regular basis, even though everyone knows that she sliced her own throat shortly after drowning her child. On top of this, while the book’s narrator and protagonist, Etsuko, does not explicitly admit to it in her selective memory of events, we are given to understand that Mariko detects a certain resemblance between the homicidal mother and Etsuko, who happens to be pregnant during the period of their acquaintance.

Soon enough we learn that Mariko’s barometer was accurate in signaling a likeness between Etsuko and the homicidal mother. The daughter Etsuko is carrying in utero will eventually hang herself for reasons that are not entirely clear (given Etsuko’s self-deceptive narrative reticence). Despite narrative obscurities, however, we gather that Etsuko failed to nurture her daughter or protect her interests. If it is true that Etsuko did eventually facilitate her daughter’s suicide, then Mariko’s initial fear in looking into her face may be seen as a déjà vu-like premonition of Etsuko’s own maternal vacancy. Mariko detects some sort of infanticidal coefficient in Etsuko and therefore rightfully fears looking at her as one fears staring at a ghost. While indeed traumatized and perhaps over-cautious, Mariko’s heightened sensitivity is an accurate gauge of the distrustful, just as the uncanny sense in all of us may, as Masahiro Mori suggested, serve as an evolutionary instinct, the purpose of which is to shield us from proximal danger. The gravest threats often loom beneath the recognizable, friendly, and familiar.

The uncanniness of the drowning scene is crucial to the entire novel because it crystallizes so many questions about what is natural and what is artificial when we act in times of desperation. As the novel unfolds with its slow reveals, the reader begins to suspect that Etsuko’s entire social element in postwar Nagasaki falls along the uncanny spectrum. The ladies of Nagasaki fear seeming imperfect, offensive, unhappy, and insufficiently maternal, and one way of looking at the drowning mother is as a woman who, in the most horrendous moment of her life, presents the world’s most unsuccessful facade. She is merely an extreme version of Etsuko and her friend Sachiko (who drowns Mariko’s kittens in front of her in a pathological repetition of the canal incident, and who prioritizes a dalliance with an American seducer over Mariko’s wellbeing). Etsuko and Sachiko may not literally kill their own children, but they do endanger and traumatize them repeatedly.

In fact, almost all of Ishiguro’s novels depict characters who are, themselves, at least minimally uncanny. His butlers act like robots, his robots act like butlers. His clones act like artists, his artists act like clones. Almost everyone wears an emotional mask in circumstances when soulful self-exposure is the needful thing. All perform their roles according to narrow expectations, all exonerate their prior moral failures by reference to the indelibility of these roles, and all are afraid of violating the taboos that negatively structure their insular habitats. Their outward actions seldom correspond to their inward actions—for their lives are lived as performances calculated to gain the admiration of an imaginary consensus.

Hence, all have committed themselves to some form of rule-based or algorithmic servitude that masquerades as complete human happiness or perfection. These damaged, robotic figures are frightened of moral independence, love, and responsible freedom (which entails the possibilities of guilt, shame, offense, judgment, and vulnerability). Thus, they outsource their tasks and decisions in a variety of ways, and they become the shadow puppets of an unseen hand we might call “history.” Their internal bargaining tactics—with which they sophistically relinquish moral sovereignty—only intensify the social structures that have generated their dilemmas.

Ishiguro, whom we might call our laureate of the uncanny, has repeatedly explored the consequences of outsourced agency, disavowal of responsibility, and empathy-malfunction, maintaining his essential theme of the human trending towards the nonhuman under the effects of peer pressure, warfare, tyranny, difficulty, intellectual ambiguity, technology, and eroticism. At the same time, he has adapted the explicit conflicts of his novels to the sweeping social and technological changes that have occurred since the Second World War.

Readers who have followed Ishiguro’s development as a novelist from A Pale View of Hills through his most recent book, Klara and The Sun, will note how his focus has drifted geographically from Japan to England, to an unidentifiable European country, and finally to America—and from the tribal/imperialistic past to the technocratic near-future. With this general shift, his writing, although still deeply recognizable thematically and stylistically, has tended more towards science-fiction gossamer—where genre tropes like cloning, genetic engineering, and robotics tint the atmosphere of a novel rather than being the constituents of a fully fledged alternate or unknown world. His more recent novels construct unheimlich—and even highly likely—variations of the world we currently inhabit. What is essential is that his humans are the truly uncanny ones.

It is not that Ishiguro reveals uncanny technological advances, although he does, but that he reveals humans leaning on these advances for the avoidance of life’s most difficult tasks. Although usually relatively powerless figures, their protocols of appeasement contribute to what I call the uncanny cascade. While Mori’s concept of the uncanny valley is useful for many reasons, it remains anchored to a bygone era—an age in which human sensitivity to alternate forms of engineered intelligence was sharper. Now those forms of intelligence hide behind keystrokes, search engines, databases, the digital humanities, our “archive fever,” privacy terms and conditions—the fine digital print—NDAs, online marketplaces, and social media apps.

When Mori authored his paper in 1970, he was a robotics professor with an explicit purpose. As he put it, “We hope to design and build robots and prosthetic hands that will not fall into the uncanny valley.” Yet what is more urgent than overcoming bad robot aesthetics is an Ishiguro-like attention to the ways in which the silicon-military-industrial-pharmaceutical-surveillance-censorship-complex renders humans more uncanny than ever was possible before. We outsource our aesthetic decisions, crowdsource our ethics, and have a deep fear of violating the established order (committing offenses), even though that order is a mere illusion advanced by the belief in a panopticon.

Technological proliferation has brought about an uncanny cascade conditioned by (but not limited to) unforeseen forms of AI, neural networks, deep fakes, chatbots that split the difference between search engines and task engines, online avatars that bear little relation to the individuals controlling them, metaverses that allow us to outsource experience itself, and the proliferation of surveillance mechanisms and databases inhibiting public behavior without altering private intent, we are witnessing a transitional period in which this new outpouring of the cascade has become more evident generally while hiding its particular “valleys” in plain sight. As I understand things, all of human history is an uncanny cascade, but only during transitional epochs such as ours is this truly noteworthy. If César Aira is right that literature was invented to explain cave-paintings, then from the start of advanced forms of communication, we have the representation of a representation. As time unfolds from our civilizational origins, we continually repeat this replication process, writing about writing about writing about an absent original.

From the invention of myth, storytelling, and art—when representation, simulation, dissimulation, replication, masking, disinformation, doubling, puppetry, and abstraction first became human trades—until now, each epoch’s uncanny total product has derived from and multiplied the uncanny factors of all the prior stages combined: the concept of fate, weaponry and warfare, an all-seeing monotheistic God, trompe l’oeil, the printing press, the camera, motion pictures, the bomb, computers, the Internet, the cloud, social media, smartphones, machine learning—each development begets a new uncanny epoch at a faster rate. Each development threatens the contemporary grasp of sovereignty, and thus each must beget an equivalent adaptation of moral tactics. At crucial points of innovation, thinkers tend to discuss or represent uncanniness more openly.

To take an older example (since we have already seen the responses of Jentsch, Freud, Mori, Ishiguro, and Baudrillard): Don Quixote is a humorously uncanny figure who has outsourced his decision-making to an obsolete code he calls chivalry. In part, the novel in which he appears was a reaction to the authority crisis instigated by the printing press. Like Don Quixote’s author, Cervantes, we are merely in a transitional period, when the possibility of being duped by incomprehensible intelligences—and thereby duping ourselves—has grown exponentially.

This presents the incredibly difficult task of staking out a smaller moral ground with greater moral fortitude. We are so immersed in the uncanny that there is little sense in denying that its attraction factor overwhelms its repulsion factor. As with bureaucracy, which is a network of disavowal that renders its human constituents unaccountable, yielding to the uncanny cascade reinforces its current.

Ishiguro has been tracking this. For his entire career, he has been issuing novelistic warnings to the effect that the human being loves technology for its emergent ability to make us more uncanny. His novels express a deep and instinctive knowledge that new opportunities for moral disavowal lie where uncanniness is felt. We can tether the “artificial” and the “human” forms of uncanniness by saying that the uncanny emerges when we witness beings, mechanical, algorithmic, or real, performing tasks without moral sovereignty. And whether we like it or not, Ishiguro’s recent trends as an author suggest that the uncanny cascade is a major aesthetic issue of our time, and thereby a major ethical stumbling block: how to create levees to prevent the uncanny cascade from sweeping us into its current?

It is entirely apt that the uncanniest scene from his latest work, Klara and the Suna scene in which a mother and a robot perform a simulation together—occurs at a waterfall. Published in 2021, the novel takes place in a near-future when gene editing and artificial intelligence have become far more viable than they are today (although, as we slowly learn, the viability of these technologies is anything but comprehensive). Klara, the narrator, is not a human female but a “Girl AF,” a solar-powered machine-learning robot with feminine traits. Because the gene-editing techniques designed to enhance the intellectual capacities of human children often leave them fatally ill, one function of AFs (Artificial Friends) is to care for human children who are too sick to function normally. The scene in question once again involves a negligent mother and child mortality.

Klara has become an Artificial Friend to a genetically edited and subsequently fatally ill child named Josie. As Klara and the mother drive alone to the scenic falls that had been the favorite picnic spot of Josie’s now-deceased older sibling, Sal, they hold a conversation about whether Klara is capable of feeling emotion. “The more I observe, the more feelings become available to me,” Klara says, to which the mother responds: “In that case, maybe you shouldn’t be so keen to observe,” making it clear that she finds the idea of an emotionless sentience enviable.

When they reach the falls, Klara consoles the mother for the loss of her first daughter, Sal, prompting the mother to say, “Okay Klara. Since Josie isn’t here, I want you to be Josie. Just for a little while, since we’re up here.” She reminds Klara of the day the family purchased Klara from a retail shop, when, in a strange audition, the mother requested that Klara imitate her daughter’s idiosyncratic limp. Now, she asks Klara to simulate Josie’s seated posture, to move and gesture like Josie, and finally to speak and act as though she were trying to inhabit Josie’s psyche.

“Now say something. Let me hear you speak.”
“I’m sorry, I’m not sure…”
“No, that’s Klara. I want Josie.”
“Hi, Mom. Josie here.”
“Good. More. Come on.”
“Hi, Mom. Nothing to worry about, right? I got here and I’m fine.”

As the simulation progresses, the reader begins to suspect something that was only faintly apparent earlier on—that the mother has brought Klara into their broken home for a more extraordinary purpose than the explicit functions for which AFs were designed. She intends to have a replica Josie in case the sick child finally dies—Klara is a form of insurance against the mortality of a second child that the mother has inadvertently sickened. Well beyond merely testing Klara’s observational abilities and capacity for feeling, she is priming her to be a robotic continuance of human life. The mother’s request comes to an abrupt end when Klara goes too far—she imitates Josie too well and makes everything too familiarly unfamiliar, so that it becomes unclear what or whose agency is directing her simulation.

The situation the mother has put herself in can be likened to the self-infliction of a double-bind. Consider that the mother must hope for Klara’s imitation to be both a success and a failure. If the imitation is a success, then Josie will “live” on after her death—and the mother can keep Josie/Klara and continue loving this pre-programmed double (who is in herself already doubled, being both Klara and Josie, all the while pretending to be only Josie but betraying signs of Klara). On the other hand, if the imitation is indeed a success, then the very notion of soulfulness, human individuality, spirit, personality, character—or whatever one’s preferred terminology—will be extinguished. Whether or not the mother is aware of this on the surface of her mind, the effects of this double-bind are already present prior to and regardless of Josie’s possible premature death. Enthralled by technological advances, the mother has dehumanized her daughter and herself. Her own desire is no longer actionable because its object has been eliminated.

The doubling here is almost dizzying, since it seems the mother is fated repeatedly to kill her own daughters. First she initiates Sal’s death, then experiences it, then Josie’s death becomes likely. Next she threatens the idea of Josie as an unrepeatable soul in the universe. Finally, she must intermittently kill the idea of “Klara as Josie” whenever the thought of a reanimated Josie becomes too uncanny, as it does in the chilling moment beside the waterfall. But really, simply by entertaining the prospect of replacing Josie with a robot, the mother has already killed her second daughter to a certain degree. The hope that she can be replicated constitutes an act of bad faith against the supposition that her beloved daughter is a besouled, bounded, and singular being with free will, imagination, unpredictability, and the capacity for love. As Jean Baudrillard put it, “when the double appears, when it becomes visible, it signifies imminent death.” Ironically, this is doubly so when the double is not merely encountered but actually summoned or created in an effort to eliminate death. Recall the fate of Dorian Gray, who outsourced his aging process to a painting of himself.

The mother has generated a situation in which she is already mourning the living, but she refuses to admit that she is the killer, which is why she doubles down on the technologies of death. Her repulsed reaction to Klara’s simulation makes it clear that, from now on, both Josie and Klara will be uncanny figures in her life. But for readers, the mother is the uncanniest of all. She lives in bondage to technological forces whose seductive strategies she refuses to contemplate, creating uncanny valleys she is unwilling to acknowledge, and perpetuating the uncanny cascade by outsourcing life’s difficulties in ways that only dampen her diminishing humanity.

Latest Podcast

Join the newsletter to receive the latest updates in your inbox.


On Instagram @quillette