Constructive Nature of Perception

75 Constructive Nature of Perception

David Kyle Johnson

I would like all people to know that I do believe that this is the Virgin Mary, Mother of God.

Diana Duyser, about a “face” in her grilled cheese sandwich

Many of the things that we believe are generated by our senses interacting with the outside world. If this interaction always generated an accurate representation of the way the world is, that wouldn’t be a problem. What we have discovered, however, is that our senses can lead us astray – and do so far more often than we realize or would even suspect. Why? Because our perception is constructive. Consequently, to have the best chance of having true beliefs about the way the world is, we have to guard against the many ways our perception can lead us astray. Otherwise we can fallaciously draw erroneous conclusions about everything from ghosts in houses and UFOs in the sky to Virgin Marys in grilled cheese sandwiches and satanic messages in popular music.

What does it mean for perception to be constructive? Take vision, for example. Our eyes do not just take a picture and then present the world to us as it is. They receive information about the world, send that information to our brain, and then our brain interprets that information in certain ways based on certain criteria. It decides what is relevant, disregards what is not, and then determines how that information should be presented to us.

In many cases, that presentation is inaccurate. As neuroscientist Steven Novella (2016) put it:

[Our] brains construct [our] image of reality […] we are not passive recorders of reality. What you experience as your stream of consciousness is an incredibly highly filtered, and selected and altered and reconstructed narrative of what’s going on. […. It’s] very useful, but largely a fiction [because] in general your brain errs massively on the side of continuity and internal consistency […] not accuracy. It utterly sacrifices accuracy in order for that narrative to be seamless and consistent. […] It will fill in the gaps – we call that confabulation […] it makes assumptions, it alters how you perceive things in order to […] make it all fit.

Consider solidity and color. A solid object is not a continuous block of matter at the microscopic level (see Matthews 2009), but perceiving it as solid is not therefore an illusion: solidity concerns how it resists penetration and deformation. Color perception depends on reflected light, illumination, and the visual system (see Marder 2015). These examples show that perception is not a literal copy of microscopic structure, rather than that the objects have no physical properties.

The “dressgate” photograph illustrates the role of interpretation. Viewers disagreed over blue and black versus white and gold partly because they interpreted illumination differently. Treating the color one sees in the image as a direct, unmediated fact about the physical dress misses that contribution. The debate does not show that there are no objective facts about the dress’s reflectance or lighting (see Johnson 2015). The same image can support different perceptual experiences.

Perception combines sensory signals with contextual cues and prior expectations, including aspects of our “worldview”. We can see extreme (and interesting) cases of how this can lead us astray in brain‐damaged patients. In “The Sound of One Hand Clapping” – one of my favorite chapters in V.S. Ramachandran’s book Phantoms in the Brain (1999) – Ramachandran tells the story of patients who have had strokes in their right hemisphere but consistently deny the fact that their stroke paralyzed the entire left side of their body. The denial is so strong that when asked to do tasks that require both hands, like tying their shoes, they will try to do them – they’ll fiddle at their shoelaces with their right hand (left hand dangling at their side) – and then (incorrectly) insist they completed the task successfully “with both hands.” When asked to clap, they will raise only their right hand to the middle of their body, but insist that their left hand is raised too. Indeed, they will even see it moving!

Ramachandran suggests that the stroke has impaired the patients’ ability to update an established representation of their body when information conflicts with it. On this account, the prior representation of a functioning left side persists. This is a proposed explanation of these patients’ reports, not evidence that beliefs alone generally determine what people perceive.

The boundary between “normal” perception and perceptual error is not a simple distinction between injured and uninjured brains. Oliver Sacks, in his book Hallucinations (2012), goes into detail about how and why even normal people can suffer from hallucinations. You can probably even think of a few examples from your own personal life. I’ve seen my dog running in the kitchen when he’s not (he’s actually asleep upstairs). While I have been in the shower, I’ve heard the phone ring when it’s not. I have felt my phone vibrating in my pocket when it’s not; it’s not even in my pocket! This is because of the role of expectation in perceptual construction. According to Schick and Vaughn (2014, 101), by simply altering what people expect, studies have shown that you can make people see lights flash, feel electric shocks or warmth, or even smell certain odors all without any of those things actually being present.

When investigating an allegedly haunted house, learning a detailed account of the supposed ghost beforehand can shape expectations and later interpretations. Recording observations before supplying those cues makes it easier to assess that influence. The point is to reduce suggestion, rather than assume that expectation guarantees an experience or explains every report (see Grant 2005).

Another assumption that informs how we interpret the information that our brain receives is that the size, color, and shape of objects is constant. This allows us to perceive an object the same way even if how we are seeing it is changing – for example, even if the surrounding lighting or our distance and angle to the object changes. We see snow as white regardless of whether it is lit by the moon or the sun. We perceive a bus as large even if it is far away.

Obviously this kind of perceptual constancy (as it’s called) can help us understand the world accurately – because usually objects aren’t changing size, color, or shape. But it can also lead us astray. There are of course countless visual illusions that demonstrate this fact; one of my favorites is the checkerboard shadow illusion in which you will see two squares as completely different colors, when in fact they are the same (due to the way light and shadow affects our interpretation of visual information). But you can also demonstrate this experimentally. In one study, donkeys and trees were each cut out of the same green paper and illuminated with the same red light, yet the subjects still saw the donkey as grey and the tree as green.

Experience helps calibrate judgments of size and distance. Familiar objects and reference points make those judgments easier; a small nearby object and a large distant one can be difficult to distinguish against a blank sky. The useful range depends on the available cues, rather than a fixed distance that applies to everyone.

This is why my friend, Tony, and I were once nearly convinced that a plane was going to slam right into a 7‐foot delta kite we were flying in North Dakota one day. The kite was 1,000 feet up, and until the plane passed right over the kite (the plane was much higher), we had no frame of reference to judge the size or altitude of the plane. It’s for similar reasons that we all see the moon as large on the horizon but small in the sky. It’s not because the lower atmosphere acts as a magnifying lens; that’s a myth. One explanation emphasizes perceived distance and size relations, including comparisons with the Ponzo illusion (see Plait 2010); this should not be treated as a settled single cause. Similar kinds of perceptual mistakes explain why two airline pilots and one national guard pilot were all convinced that they saw a whole squadron of UFOs – so near that the national guard pilot radioed a near collision with the UFOs – even though it was later confirmed that what they were actually seeing was a meteor and its fragments, a full 125 miles away (see Schick and Vaughn 2014, 110).

Another reason people think they see UFOs (and another way our perception leads us astray) is the autokinetic effect. If you stare at a stationary bright object against a solid background for a while, it can start to appear as if it is moving. The account discusses possible explanations (see Rucci and Poletti 2009). The effect can contribute to interpreting a stationary light, such as Venus, as a moving object.

This brings me to the last way our perception leads us astray that I’ll mention: pareidolia. One thing that our brain likes to do with the information it receives is impose patterns on it. So when we, for example, see a vague amorphous stimulus, we will often impose a pattern onto it and see something that is not actually there: a human face on Mars, Jesus on a tortilla, the Virgin Mary on a grilled cheese, and so on. It’s often a face because of how much of our brain’s visual centers are dedicated to facial recognition, but it doesn’t have to be. I’ve also seen Godzilla in the clouds, a dog on a stained wall, and even the full body of an ascending Jesus in a dog’s butt (see Campbell 2013).

But vision is not the only sense susceptible to this effect. Our hearing is too. This is why you can sometimes hear “secret messages” in speech or music when you play it backwards. “Another one’s bites the dust” sounds like “It’s fun to smoke marijuana” if you play it backwards, and there are supposedly secret messages about the Beatles replacing Paul McCartney with a lookalike in Beatles songs. It’s called “back masking,” and when I was growing up, many a youth pastor argued that I shouldn’t listen to secular music, supporting their arguments by playing popular songs backwards, fallaciously finding nefarious hidden messages in what they heard, and then claiming they were put there by Satan.

Being told which words to listen for can strongly affect what you hear in an ambiguous recording. When I play videos that have Beatles songs played backwards for my students and hide the subtitles, they can’t hear anything. But when I play the same thing with subtitles, they can’t hear anything else. Subtitles accompanying supposed “voices” on programs such as Ghost Hunters can likewise influence interpretation. To test an ambiguous recording, ask listeners what they hear before suggesting words, then compare their independent responses.

A convincing perceptual experience does not by itself establish its cause. Independent observations, controlled comparisons, and tests that limit suggestion can help us evaluate competing explanations. These methods make perceptual errors easier to identify without assuming that perception is generally useless.