The root of the uncanny: why are things creepy?

Testing AD 2222 \ please ignore


A dark alley. At the end of it the silhouette of a person, child like, facing you. A chuckle in the distance. Suddenly, one arm jerks up. Then it begins to walk, closing in with sudden movements of each limb, as if pulled by strings…

Now, would that be creepy at all? There is no doubt that we can be ‘‘creeped out’’ by certain sights or situations. Nevertheless, it is still debated why feelings of unease or eeriness exist, be it in in real situations or evoked by movies, stories or computer games. What is it that makes things uncanny? And what do people find so creepy about clowns and ghostly children?

While the nature of creepiness is still a scientific conundrum, the entertainment industry has been exploring the recipes to create creepy situations for decades – and quite successfully so. A common theme in scary films is to include creatures that are somewhat human but have a very unnatural aspect to them: ghostly children, haunted puppets or dolls, all sorts of undead, and of course, demonic clowns. All these seem to work wonders in scaring us on purpose.

The Uncanny Valley

The fact that human-like entities can evoke unease is well known in robotics. This goes back to a paper in 1970 by robotics professor Masahiro Mori where he suggested a general increase in our affinity to an object, such as a robot or puppet, the more human-like it becomes. Only just before being indistinguishable from an actual human being, Mori described a prominent dip in the curve: the uncanny valley. It depicts the low-point in our emotional response, the uncanny feeling upon encountering an entity that blurs the line between a human and something unknown (see the figure below).

Depiction of Mori’s proposed relation between the human likeliness of an object and the human affinity for it. Although not part of Mori’s original figure, a human-like robot such as Hanson Robotic’s Sophia would likely be close to the centre of the uncanny valley.

To picture this just imagine Disney’s Wall-e: a not very human-like robot and therefore far for the uncanny valley. Now compare this to Hanson Robotic’s newest creation of a human like robot, called Sophia. This robot has the right body proportion, skin imitates and even facial expressions. Sophia is much more human-like than Wall-e but is usually considered creepy – even more so when smiling. To many of us, robots like Sophia are close to the centre of Mori’s uncanny valley. But what is it exactly that makes them uncanny and could it help us understand what makes things creepy in general?

The perhaps most prominent mechanistic explanation for the uncanny valley effect is the perceptual mismatch theory, which describes inconsistent features such as robot eyes in a human face as trigger for the uncanny feeling. This theory has been applied in multiple studies that used altered images to reliably create feelings of eeriness in subjects presented with them.

The adaptive value of creepiness

Still, this describes merely a recipe for triggering eeriness. It begs the question of why we can be creeped out in the first place. Wouldn’t it be much nicer if we had evolved without this emotion? Monkeys too, when presented with altered images of conspecifics, express signs of aversion. This does not rule out that the feeling of eeriness is merely a side effect of how our brain processes information.  But it may well be that it has evolved due to some selection pressure: similar to other emotions such as the experience of pain, uncanniness might tell us about the presence of danger.

One idea that builds on this assumption is the pathogen avoidance hypothesis, which states that the uncanny feeling has evolved as protection against a potential carrier of pathogens. It describes changes in behaviour or appearance of an individual compared to the healthy norm as signs of unspecified infection. This hypothesis could well explain why a ‘‘strange’’ human (or uncanny robot) evokes a stronger aversion than members of another species such as a ‘‘strange’’ dog: infected humans would be more contagious to us. However, even in the light of the many zombie films out there one could wonder why horror movies are not focusing entirely on carriers of mysterious infections if this is the root of the uncanny valley effect. Further, the hypothesis only focuses on human-like entities and cannot explain why certain places and locations are deemed as creepy.

The eeriness of the unknown

It is tempting to liken the feeling of eeriness to blunt fear in the face of danger, like heights or a wild tiger. But then there is a clear difference between fear of an approaching tiger and something we deem uncanny: whereas danger coming from the tiger is obvious, it is much less clear in a creepy situation. Just imagine an eerie scene in a movie. The situation is not outright dangerous but rather ambiguous in what will happen next. Scientists McAndrew and Koehnke suggested in 2016 that creepiness is created by an encounter with ambiguous signs of danger. In case of a creepy human they argue that these ambiguous signs would all serve as indicators for unpredictable behaviour and thus for potential danger.

To explore this idea, McAndrew and Koehnke asked more than 1300 adult participants to select from a list of given features those that a creepy person would show. Interestingly, most participants agreed on signs that made the imaginative person deviate from the norm in both appearance and habits – which might then suggest general deviations in behaviour. According to the results a typical creepy person could be oddly dressed, extremely pale, and would laugh inadequately – just to name a few examples. It is perhaps not surprising that some of these are quite applicable to clowns. The authors conclude that all selected ‘‘creepy’’ traits signal for unpredictable behaviour and therefore trigger increased alertness towards unspecified danger, expressed as a feeling of creepiness.

This study has limitations such as self-selection of subjects together with the restricted informative value of questionnaires. But still, the ambiguity hypothesis has gained some ground since, and was extended from creepy encounters to situations in general. Langer and König showed in 2018 that urban locations are rated as more creepy in the dark, when approaching danger should be less predicable. Thus, the ambiguity hypothesis can explain creepiness of both situations and social encounters, which makes it perhaps the most convincing explanation so far.

And maybe it is true that the sight of human-like creatures in films and games triggers alertness in the face of the unknown, be it unpredictable behaviour or potential contagion. It could actually be telling that classic film monsters, which are now widely known, such as Frankenstein’s Monster or Dracula, are scarcely used in contemporary horror films. And is it not the case that many movie monsters loose some of their impression the more we learn about them, maybe even their personal story and motivations? Perhaps eeriness really comes down to ambiguity in the face of potential danger. And once the veil of the unknown is lifted, the former monster becomes more amicable – or outright dangerous, such as the killer in a slasher movie. The feeling of eeriness would suddenly make room for other emotions, be it empathy or fear.

Even though the ambiguity hypothesis might fit well with our personal experience, it is not the definite answer yet. It is connected to the assumption that eeriness evolved as a mechanism to help us avoid potential danger and more studies are needed to show this empirically. What we can be certain about is that monsters will keep coming, given the ever-lasting demand in the horror genre. Where this demand comes from, and why many of us even enjoy the thrill of being somewhat creeped out is yet another question. But then it might be the same one as asking why people enjoy rollercoasters. Only, in contrast to leisure park attractions, horror movies evolve much quicker. Who knows what entities will follow in the footsteps of Pennywise and other major protagonists in contemporary horror? We don’t know. And perhaps that is exactly what will make them scary.

Testing AD 1 1 1 1