These Strawberries Are Not Red
Why illusions reveal how good our perception is
Carl Jennings
Like most people, you probably see a group of red strawberries in the image above. However, you might be surprised to learn that there is not a single red pixel in the entire image. If you look at fig.2 below, (a detail of the above image) you will see that the strawberries are in fact grey!
How can this be? Are your eyes playing tricks?
Not really.
For many, visual illusions are confirmation that our senses are unreliable and prone to error, that our ability to be fooled by such illusions is proof positive of a fundamental ‘glitch in the system’. The reality, however, can be very different. Many visual illusions work because they actually reveal a sophisticated and creative intelligence inherent in our perceptual systems — one that demonstrates just how well-adapted we are to the physical world. To understand this discrepancy, it helps to keep in mind that our perceptual systems evolved long before the advent of ‘pictures’. They were developed to help us navigate the physical, three-dimensional world of people, places, and predators! — not the illusory space of flat, two-dimensional images. Like a magic trick, two-dimensional images can pull off a conceptual sleight of hand, by taking hundreds of thousands of years of evolutionary development and applying it to a relatively recent phenomenon that it was never meant for — 2-D images of the real world. Because the eye and brain are designed for real-world interaction, their complex processes, when applied to illusory two-dimensional representations, can produce some startling results, not because they are faulty but because they are highly sophisticated. This is precisely why we perceive what appears to be red strawberries in the image — when, in fact, they are grey.
So why do we perceive them as red?
Short answer: if we saw this image in the real world, in three dimensions, the strawberries would have to be red, and therefore, our brain makes us see red.
Long answer: everything that we view in the world, we view in context — in a particular environment and under a specific set of circumstances. We rarely see things in isolation. This is important because a given context will create a set of assumptions and expectations about what we are looking at. In the strawberry image above, our brain is seeing the strawberries in an environment that involves a cyan (greenish-blue) color that we can assume to be either a filter that we are looking through, or equally, a cyan light source (illuminant) which appears to be bathing the whole scene. Our perceptual system takes in the entire scene and compares all the different elements to each other. It then creates a best-guess scenario about what exactly we are looking at, and it is this guess, or inference, that we actually see — not what is objectively ‘entering’ our eyes. Our eyes aren’t being fooled, they are in fact seeing/detecting grey strawberries, but because we are perceiving them in a cyan-colored environment, our brain interjects and infers that the strawberries have to be red objects in order to appear grey in that context. How? There are two ways. We are either seeing them through a colored filter or we are seeing them lit with a colored light -but it doesn’t matter because the mechanism is essentially the same. If we assume it is a filter, colored filters remove those wavelengths that are spectrally opposite to the color of the filter. For something to appear achromatic or colorless (grey), it must be what’s called the additive opposite color of the filter, and since red is the additive opposite of cyan, our perceptual system concludes that what we are looking at must, in the real world, be red. In other words, only red objects can appear grey through a cyan filter. Likewise, if we assume it is a colored light source (illuminant) — our brains have evolved to automatically, and naturally, subtract the color bias of the light source when we look at things. This evolutionary adaptation, known as color constancy or chromatic adaption, explains why objects appear to retain their original color regardless of changing lighting conditions, even though the actual color of the object might have changed. Without such an evolutionary adaption our paleolithic ancestors might not be able to recognize ripe fruit and fresh meat under different lighting conditions. According to the neuroscientist Bevil Conway at the National Eye Institute, commenting on a well-known version of the same illusion, “Your brain says, ‘the light source that I’m viewing these strawberries under has some blue component to it, so I’m going to subtract that automatically from every pixel.’ And when you take grey pixels and subtract out this blue bias, you end up with red.” In other words, whether we are looking through a filter or looking at a scene lit with a colored light, our brain would conclude that, through a process of unconscious inference, the strawberries would have to be red in order to register as grey to the eye under such conditions.
This example points to an even more important and fundamental aspect of perception and the brain: namely, that our perception is less about the light hitting the eye and more about what such light means. This can be difficult to grasp because we often misunderstand the true purpose of perception. For many of us, what we see, hear, smell, touch, and taste are understood to be more or less accurate features of the world we inhabit, and that the function of our perceptual systems is to faithfully represent this world to our conscious awareness. But this is only partly true. The real purpose and function, is to help us survive and navigate our environment, and this does not always correspond with conveying an accurate representation of the external world. This might seem like a subtle difference, but it’s not. Perceptual stimuli — the light that enters our eyes, the soundwaves that enter our ears, etc. — are inherently ambiguous. The brain doesn’t hear sounds, see light, taste tastes, feel touch, or smell smells. It only has access to streams of chemical and electrical signals that are indirectly related to the outside world. From these signals, it constructs a prediction model of how the world most likely is. These models, which are built up over our lifetime and which are constantly being updated with new perceptual stimuli as we move through and navigate the physical world, provide the foundation for a series of best-case scenario guesses and inferences that we continuously make, moment by moment. According to cognitive neuroscientist Anil Seth, perception is “a process of informed guesswork in which the brain combines these sensory signals with its prior expectations or beliefs about the way the world is to form its best guess of what caused those signals.” This idea, an integral part of the theory of predictive processing, which has emerged as the leading theory of perception and cognition in recent decades, maintains that what we perceive is always structured and interpreted within a wider context of understanding. Our brains are constantly figuring out what works and what doesn’t through a process of prediction, inference and model-making. In the case of the strawberries however, some might object that it is simply our past experience with ‘red’ strawberries, that allows us to see them as red. But, by ‘past experience’ the theory of predictive processing means something much more foundational; namely our experience with changing lighting conditions as we move through the world. It’s not that we know strawberries are red, which makes them appear red but rather a more basic visual inference that our eye and brain make based on our experience with changing lighting and viewing conditions. This can be demonstrated by simply reversing the color of the filter/illuminant from cyan to magenta (fig 3). The same rules apply, and now the strawberries appear green, when in fact, there is no green in the image! It is not our memory of strawberries but rather our experience of navigating the world and the different ways in which we interpret and predict what we see, that causes this illusion.
Though these illusions are specific to color, the basic concept applies to all aspects of our visual perception: size, shape, value, movement, etc.
Take the following image (fig. 4). Like the strawberry illusion, our brain’s job is to understand the world so that we can act appropriately, and sometimes this means overriding what we actually see. In the example below, the two red lines appear to be two different sizes, when in fact they are the same length.
Why is it so hard to see this? Again, because our perception is adapted to the real world, and in the real world, the line on the right would be larger; it has to be, as it not only runs the entire height of the wall, but we know that things appear smaller the further away they are. Even though the length of the lines entering our eye are the same, we would know that they are not, and only when we flatten the image into a two-dimensional representation would this discrepancy become apparent. Illusions therefore are capable of showing us just how well adapted we are to living in the real world, and that it is only in artificial situations, like 2-D images, do they can appear to ‘trick’ us.
In short, we create meaning, not just at the conceptual or psychological level but at every step of the way, even at the most basic level of image processing. We don’t just ‘take in’ and see what is before us; we edit, manipulate, interpret, infer, and give sense to visual information. Visual illusions reveal this meaning-making ability by exposing the underlying mechanisms of our perceptual systems. Our visual perception, even at the most basic and unconscious level of image recognition, is both creative and intelligent.
