Uncanny dehumanization

Freud had his own ideas about how feelings of uncanniness come about, but it will be more productive for us to turn to the work of his contemporary, a German psychiatrist named Ernst Jentsch. In his 1906 paper “On the psychology of the uncanny,” Jentsch set out to determine exactly what it is to experience something as uncanny, what sorts of things elicit this response, and why they do so. He argued that uncanny things produce a sense of disorientation and uncertainty. But he was aware that not all uncertainties generate the uncanny feeling, or produce it to the same degree. There must be a special sort of uncertainty involved.

“Among all the psychical uncertainties that can become a cause for the uncanny feeling to arise, there is one in particular that is able to develop a fairly regular, powerful and very general effect: namely, doubt as to whether an apparently living being really is animate and, conversely, doubt as to whether a lifeless object may not in fact be animate – and more precisely, when this doubt only makes itself felt obscurely in one’s consciousness. The mood lasts until these doubts are resolved and then usually makes way for another kind of feeling.” ⁱ

Skillfully crafted human figures in a wax museum are one of his prime examples. Even though one knows that these are inanimate figures, they are so lifelike that one cannot help regarding them as real people. The outcome is seeing them as both animate and inanimate. But nothing can be wholly animate and wholly inanimate, and it is this that produces the uncanny feeling. To the extent that the uncanny feeling persists, he conjectures, “it is probably a matter of semi-conscious secondary doubts which are repeatedly and automatically aroused anew” or “the lively recollection of the first awkward impression lingering in one’s mind.” ⁱⁱ  

Having got to this point it is helpful to push back at the translation of Unheimlich as “uncanny.” The English word does not quite capture the state of mind that Jentsch is gesturing towards. Something “uncanny” can be simply odd or astonishing – even enjoyably fascinating, as is exemplified by open-mouthed wonder at the “uncanny” feats of a champion athlete. But Jentsch is clearly talking about a disturbing quality of experience, one that sends chills down your spine and makes your blood run cold. To find a good English equivalent, just ask yourself how uncanny things such as wax figures in a dimly-lit room strike you. One word that is likely to come to mind is “creepy.” Uncanny things are creepy things, and the state of mind that they produce is the state of being “creeped out.” Now, imagine that the wax figures in the dimly-lit room begin to move. They turn their heads towards you, open their mouths, and blink their waxy eyes. How would that make you feel? Here is how Jentsch discusses this kind of scenario, using the example of humanoid automata. 

This peculiar effect makes its appearance even more clearly when imitations of the human form not only reach one’s perception, but when on top of everything they appear to be united with certain bodily or mental functions. This is where the impression easily produced by the automatic figures belongs that is so awkward for many people. Once again, those cases must here be discounted in which the objects are very small or very familiar in the course of daily usage. A doll which closes and opens its eyes by itself, or a small automatic toy, will cause no notable sensation of this kind, while on the other hand, for example, the life-size machines that perform complicated tasks, blow trumpets, dance and so forth, very easily give one a feeling of unease.

Sixty-four years after Jentsch’s paper, Japanese roboticist Masahiro Mori published a paper that drew the same conclusion. Although very short and highly speculative, Mori’s paper has been immensely influential, especially in the cognitive sciences. It is titled “Bukimi No Tani,” which was translated into English by the art critic Jasia Reichardt as “The Uncanny Valley." ⁱⁱⁱ Like the German Unheimlich, the Japanese Bukimi can be rendered as “creepy” – so the title of Mori’s paper can equally well be translated as “The valley of creepiness.” The thrust of the article is straightforward. Mori speculated that as robots become more and more human-like, we will feel more and more comfortable with them until technology reaches the point where robots are almost, but not quite, indistinguishable from human beings. At that point, he suggested, there will be a precipitous drop in the feeling of affinity, and the humanoid robot will be experienced as Bukimi. He called this “the uncanny valley.” Mori made the same prediction about prosthetic limbs. An artificial hand that is not quite indistinguishable from a flesh-and-blood will, he supposed, produce feelings of repugnance. And just like Jentsch, Mori argued that adding movement will only augment the disturbing effect. “Since the negative effects of movement are apparent even with a prosthetic hand,” he wrote, “to build a whole robot would magnify the creepiness. This is just one robot. Imagine a craftsman being awakened suddenly in the dead of the night. He searches downstairs for something among a crowd of mannequins in his workshop. If the mannequins started to move, it would be like a horror story.” ^iv

Mori’s use of the expression “horror story” is significant, because there is a quantitative dimension to the uncanny. Uncanny things are on a spectrum extending from the merely creepy (for example, a prosthetic hand) to those that elicit feelings of sheer horror (for example, the moving mannequins). The feeling of horror is not the same as the feeling of fear. As we proceed, I will articulate more clearly what it is that separates horror from fear. For now, though, I want to focus on just one element—one that applies in equal measure to all uncanny things. Things that are uncanny, whether creepy or horrifying have a peculiar allure that distinguishes them from things that are merely repulsive or terrifying. This is nicely illustrated by Plato’s story about Leontius in The Republic.

Leontius, the son of Aglaeon, was on his way up to the town from the Piraeus. As he was walking below the North Wall, on the outside, he saw the public executioner with some dead bodies lying beside him. He wanted to look at the bodies, but at the same time felt disgust and held himself back, but at the same time he was disgusted and turned away. For a time he struggled and covered his eyes. Then, desire got the better of him. He rushed over to where the bodies were, and forced his eyes wide open saying “There you are, curse you. Have a really good look. Isn’t it a lovely sight?” ^v

I do not think that “disgust” is the right word for what drove Leontius’ rubbernecking impulse. Truly disgusting things do not draw the eye towards them, as these corpses did. Instead of being repelled, Leontius behaved like a person watching a horror film who covers their eyes with their hands when things get intense, but then just can’t resist peeking out through the gaps between their fingers. ^vi

For all their insightfulness, neither Jentsch nor Mori got to the bottom of the uncanny. To get closer to the core of the kind of uncertainty or ambiguity that makes things seem uncanny, consider this chillingly evocative passage from Arthur Machen’s novel The House of Souls: 

What would your feelings be, seriously, if your cat or your dog began to talk to you, and to dispute with you in human accents? You would be overwhelmed with horror. I am sure of it. And if the roses in your garden sang a weird song, you would go mad. And suppose the stones in the road began to swell and grow before your eyes, and if the pebble that you noticed at night had shot out stony blossoms in the morning? ^vii

Machen’s examples are at the horror end of the uncanniness spectrum. He was, after all, a writer of horror fiction. Consider the first example – that of a talking pet. Being confronted with a talking dog would certainly be disorienting, but this would not be because you are uncertain whether this entity is a dog or a human being. Rather, it would be because the talking dog has properties that are unique to dogs (its canine appearance) as well as properties that are unique to humans (the ability to speak). You are not wondering whether this creature is a dog or a human. Your reaction is driven by the fact that it seems to be both, but being a dog is incompatible with being a human. It is not uncertainty that elicits the uncanny feeling, it is contradiction. Similar considerations apply to singing roses, and the examples of stones that grow and shoot out blossoms similarly involve an impossible combination of the mineral and the botanical. 

Machen’s passage points to several fundamental features of uncanny things. The first is that they all involve categorical contradiction. Uncanny things seem to transgress the categories that we use to make sense of the world. Whereas Mori’s paper was about the categorical boundary between humans and robots (and human limbs and prosthetic limbs), Jentsch had a more expansive view of the sources of uncanniness, and allowed that other kinds of categorical contradiction can elicit this disturbing feeling. This more general conception of categorical contradiction has received experimental support from a study conducted by Eva and Patrick Weiss showing that cognitive conflict occurring at category boundaries need not always involve the category “human,” but also that ambiguity about whether or not an entity is human produces the most pronounced effect.

The second feature of Machen’s examples is that they all involve things that we think of as natural kinds (although artifacts can figure in them, as is shown by the wax museum and robot examples). And third, they all involve living things (although nonliving things can figure in them, as is shown by the example of the sprouting stones). Now, putting these three elements together, we get the following specification: a thing is uncanny if and only if it is a contradictory living (or once living) thing that violates the boundaries that we take to demarcate biological natural kinds from one another.  

A moment’s reflection reveals why this might be. We tend to essentialize natural biological kinds, and because essences do not come in degrees – a thing cannot have more or less of them – we take essences to demarcate those kinds absolutely. So, if a single entity is categorized as belonging to two different natural kinds, these representations are irreconcilable. A being classified as an insect and as a human being is not felt to be “sort of” an insect and “sort of” a human being – an insect-like human or a humanoid insect – but is represented as completely an insect and completely a human being. Dehumanized people are experienced as uncanny by their dehumanizers, because they violate the human/subhuman boundary. They are conceived as wholly human and as wholly subhuman, but these two representations of the dehumanized person cannot be reconciled with one another. The dehumanizer’s mind is pulled in two directions at once, and it cannot settle on either of the two mutually exclusive alternatives. The dehumanizer’s consciousness oscillates between them, thereby giving rise to the problem of humanity.

Why are these contradictory, metaphysically transgressive representations so disturbing?  To answer that question, we need to draw on the work of a different thinker, who approaches the same conceptual territory from a different direction. 

Mary Douglas was a British anthropologist who wrote an extraordinarily influential book, published in 1966, entitled Purity and Danger. The book is an anthropological study of ritual uncleanliness. Its basic insight is that whatever does not fit into the framework of categories that one’s culture uses to order the world, is felt to be dirty, abominable, and polluting. 

“Dirt is never a unique, isolated event,” Douglas writes. “Where there is dirt there is a system. Dirt is the by-product of a systematic ordering and classification of matter, in so far as ordering involves rejecting inappropriate elements. Dirt is the by-product of the systematic ordering and classification of matter…. In short, our pollution behavior is the reaction which condemns any object or idea likely to confuse or contradict cherished classifications. ^viiii

Every culture has some conception of the natural order: a framework of categories that are used to make the world intelligible. And these concepts of the natural order are used to underwrite the social order. Normally, we think of claims as either descriptive or normative. They either state how things are or how they should be. But the form of thinking that Douglas describes does not conform to this pattern. The natural order is how the world is arranged, and it is how the world should be arranged. But there is also a realm of the unnatural. These are unnatural things—things that exist but are outside the natural order. They are harbingers of pollution and chaos. The idea of the unnatural inevitably accompanies systems that purport to describe the natural order, because there are always things that have no proper place in the framework. These anomalous things are experienced as powerful and dangerous, and must therefore be segregated, marginalized, controlled, or destroyed. 

Douglas’ formulation goes significantly further than those set out by Jentsch and Mori. Unlike theirs, it attends to the social and political underpinnings of classificatory schemes and does not limit the disturbing effect to clashing perceptual (primarily visual) signals. For Douglas, abomination is primarily a matter of how we classify things rather than how they appear to us. This comports very well with my theory of dehumanization. Dehumanized people are regarded as anomalous beings, but this is not because of how they appear. We classify them as human on the basis of their appearance, and as subhuman on the basis of what we have been told. And it is this double consciousness of the dehumanized other that derives the most toxic consequences of the dehumanizing process, because it turns people into monsters.

[i] Ernst Jentsch, “On the psychology of the uncanny,” Angelaki: Journal of the Theoretical Humanities, 2, no. 1 (1997): 12.

[ii] Jentsch, “On the psychology of the uncanny,” 12.

[iii] Masahiro Mori, “The uncanny valley,” trans. Karl F. MacDorman and Nori Kagegi, Robotics and Automation Magazine (June 2012): 98-100. Also see, for example Chin Chang Ho and Karl F. MacDorman, “Revisiting the uncanny valley theory: developing and validating an alternative to the Godspeed indices,” Computers in Human Behavior, 26, no. 6 (2010): 1508–1518. Shensheng Wang, Scott O. Lilienfield, and Philippe Rochat, (2015). “The uncanny valley: existence and explanations,” Review of General Psychology 19, no. 3 (2015): 393-407. Jasia Reichardt, Robots: Fact, Fiction (New York: Studio Books, 1978). 

[iv] Mori, “The uncanny valley,” 100.

[v] Plato, The Republic, ed. G. R. F. Ferrari, trans. Tom Griffith (Cambridge: Cambridge University Press, 2000), 136.

[vi] There are numerous interpretations of what is going on in Plato’s account. See, for example, Rana Saadi Liebert, “Pity and disgust in Plato’s Republic: the case of Leontius,” Classical Philology 108, no. 3 (2013): 179-201.

[vii] Arthur Machen, The House of Souls (New York: Alfred A. Knopf, 1922), 116. 

[viii] Mary Douglas, Purity and Danger, 44-5.