There are various visual illusions that exhibit different perceptual processes. In scientific psychology, the question that always draws attention is to what extent perceptual processes are driven by the top-down perception theory as opposed to bottom-up information. Eysenck (1998) stated that bottom-up processing is reliant on external stimuli, whereas top-down one is influenced by prior knowledge, expectations, context, and so on. Carlson, Miller, Heth, Donahoe and Martin (2010) stated that the process of perception, in the case of bottom-up processing, starts with the features of the image that the retina captures. This information is then processed in a hierarchical manner by the successively higher levels of the visual system until the top of the system receives the information, and the perception of the object occurs. In contrast, top-down processing entails contextual information extracted from memory, and here the information is passed on from higher to lower neural areas (Carlson et al., 2010). Top-down effects are voluntary, whereas bottom-up is more of physiological in nature (Watanabe, Nanez and Sasaki, 2001). Top-down theory of perception is supported by constructivists like Bruner and Neisser, whereas bottom-up theory is upheld by direct-realist like J.J. Gibson (Shea, 2013). The conflict between these two groups of psychologists arises while interpreting optical illusions, which are images ambiguous in nature, and therefore, can be interpreted in dual ways. This paper will argue that illusions are the product of top-down perceptual rather than bottom-up sensory errors.
The top-down processing theory was proposed by psychologist Richard Gregory who argued that stimulus information gathered from our environment is ambiguous in nature, and hence in order to interpret it correctly, we need higher cognitive information based on our past experiences or knowledge stored in our brain to deduce a perception (Gregory, 1974). He believed that a lot of information (about 90%) our eyes see are lost by the time this information reaches the brain and hence, our brain needs to rely on the prior knowledge to make a perception of reality (Gregory, 1970). Our perceptions are hypotheses influenced by our past and therefore, if we form incorrect ones, then it will lead to errors in perception, such as forming visual illusions like the Necker cube. The optical illusion of Necker cube was first observed by Louis Albert Necker in 1982. The cube is ambiguous, because it can be differently interpreted by different people. At the intersection point, where two lines meet, it is unclear which is at the front, and therefore, it can be perceived in two different ways (Woodson, 1979).
The question that arises is if Gregory's theory of top-down perceptions involves interpreting an image based on a hypothesis, what kinds of hypotheses are they. Hypotheses vary from one person to another, because they are situational and are based on memory and past experiences. Therefore, the same optic illusion can be interpreted in two ways in two different situations. For instance, the picture below showcases a box.
When viewed alone, the picture engages your brain in bottom-up processing, because apparently, it does not involve any optical illusion. The three thin horizontal lines and two thick vertical lines seen in the picture do not give any special context to derive a specific meaning. Hence, it does not involve any top-down processing (Carbon, 2014). But if the picture is positioned in two different situations, the brain engages in making a perception based on hypotheses. For example, when the image is placed in between two letters A and B, the brain perceives the image as letter 'B', whereas when the same image is put between two numbers 11 and 12, the brain hypothesizes it as number '13'. This shows that hypotheses vary in different situations.
On the other hand, Gibson, who proposed the bottom-up perception theory, argued against Gregory's concept of top-down visual illusions on the grounds that Gregory's perceptions of reality are based on artificial images and not real ones found in normal visual environments. Gibson (1966) believed that perceptions are survival mechanisms influenced by evolution, and it does not involve learning. He claims that perceptions are direct and not the result of hypotheses as Gregory proposed. Gibson argued that when we receive information from the visual environment about an object's size, shape, or color, our brain perceives it as what we see (Gibson, 1972). The perception does not involve interpretation because the whole perception is based on sensory stimuli. According to Gibson (1972), perception is a bottom-up process because it follows a pattern of how the information is received. The whole process begins with an analysis of the pattern of light, known as optic array, that reaches the eyes, with all the visual information required for further processing. The information received is passed on to the retina where the process of transduction into electrical impulses starts, and these impulses are then relayed to the brain which triggers further responses along the visual trajectories until they reach the visual cortex for the final processing of information (Gibson, 1972).
Though Gibson's theory provided an explanation for the direct perception of the environment, he, however, could not provide any logical reasoning for the optical illusions. He reasoned that the images used for optical illusions are highly artificial in nature, and the chances for such images to be encountered in real word are very rare (Gibson, 1972). However, it is not true that optical illusions cannot be experienced in natural environmental conditions. For instance, a stick looks straight if you hold it in front of your eyes, but if you put the stick in a glass of water, it looks bent. The stick looking bent is not an artificial image that we do not encounter in real life. The difference between the two perceptions is that in the case of the stick looking straight, the stimulus energy reflecting from the stick to the retina sent information to the brain that is the direct mental representation, which is supported by Gibson’s theory of direct perception (Favela & Chemero, 2016). However, in the case of the stick looking bent in water, the stimulus energy reflecting from the stick to the retina sent information to the brain that results in an illusion (Favela & Chemero, 2016). The stick looks bent, however, it is not really bent. This example of a bent stick suggests that Gibson's theory explains perceptions based on ideal viewing conditions, whereas Gregory's theory involves viewing in conditions that are not ideal.
When Gibson argued that the ambiguous images used for top-down theory of perceptions are unreal, he failed to take a note of a few factors. Firstly, it is true that some pictures with ambiguous features or two-dimensional or three-dimensional drawings used for experimentation purpose are unnatural, but we encounter a lot of ambiguous images in real-life situations that are not the result of unnatural lab experiments. The straight stick looking bent in water is one such example. The illusion of the stick looking bent can be experienced by two individuals at the same time. Secondly, perceptual apparatuses of different animals are different, and therefore, a natural image may produce optical illusion in a natural environmental condition because of how the brain perceives information (Favela & Chemero, 2016). For example, though the illusion of a bent stick in water can appear same to the eyes of two people who have the same perceptual mechanisms, the stick may look straight in water to the eyes of a gannet that has the capacity of accommodating light refraction at the boundary of air and water. Gibson failed to consider the visual capacity of the perceiver while dismissing optical illusions.
Direct realists like Gibson argue that since the top-down processing of perception depends on hypotheses while interpreting information, if the hypotheses are misleading, then it will result in a wrong perception of an image or a situation (Carbon, 2014). For example, making typos during texting or typing is common among people and the readers while reading through a typo use the filling-in phenomena by filling in missing details. In such a case, the brain simply extracts prior information from memory and rearranges or fills in the missing letters within the words. If the prior knowledge registered in the brain is faulty, then there are chances for the brain to make wrong hypotheses and erroneous processing of typos.
What direct realists overlook is that the fundamental principle behind top-down processing theory lies in making a connection between sensory inputs and our semantic networks. The knowledge base or the semantic network determines what kind of hypotheses an individual is likely to make (Costall, 1980). The hypotheses made by one individual could be different from that of another one, depending on their previous experience (Kohn, 2007). Since no two people have similar past experiences, their perceptual system varies. For example, if we read an old book where coffee stains have obscured partial information, some of the letters have become blurred, and the decay processes have made the white paper turn into a yellowish substance, we can read only the fragments of the text and then reconstruct the general meaning of what we read by making hypotheses (Carbon, 2008). In other words, we fill the gaps of missing texts and passages depending on our prior experience. The meaning gathered from the text could vary from one individual to another, as both have different past experiences.
Another good example is the famous man-rat illusion where the sketching is ambiguous in nature and not easily decipherable. So, one can interpret that the image is of a man while another can interpret it as an image of a rat (Carbon, 2014). For most people, the perception of the images switches from man to rat. This is a fascinating example of an illusion that demonstrates the mental capacity of a human being to switch from one meaning to another (Carbon, 2014). It also reveals the intriguing process of our brain that makes the initial perception through the activation of our semantic network. If one has the experience of seeing a picture of a man before, or if one knows what a man looks like or has heard the word "man", then one is more likely to interpret the ambiguous picture as an image of a man. If one's previous experiences were more associated with a rat, or another animal of such kind, then one is more likely to interpret the image as a rat (Carbon, 2014).
Another factor that counteracts Gibson's claim that if the knowledge base is erroneous, then it will lead to perceptual errors is how our perception is based on the expectation we keep from a scene. For example, while entering a classroom, we expect to see a black or whiteboard and desks. If these objects are present in the classroom, we rapidly interpret them. If these elements that we are expecting to see in a classroom are not present in the scene, it leads to perceptual errors. Biederman (1981) showed in his experiment that there was an increased error in identifying fire hydrants when they were placed in kitchens. The same error happened in identifying sofas when they were placed in city streets. It happens because our brain is used to seeing an object in its designated place. The brain will have difficulty in processing information when the object is not seen in its usual context.
In conclusion, it is worth mentioning that there are two groups of perceptual theorists: top-down perceptual theorists and bottom-up theorists. The first ones, known as constructivists, believe that perception is influenced by contexts, prior experience, and expectations. Bottom-up theorists, also called realists, believe that perception is sensation and is the direct result of external stimuli. Constructivists interpret optical illusions on the ground of prior knowledge, contextual information, and expectations, whereas direct realists dismiss the existence of optical illusion, calling it an unnatural phenomenon that does not take place in natural environment condition. However, optical illusions are not only two-dimensional or three-dimensional distorted images. They can be encountered in the natural environment too, as seen in the example of a straight stick looking bent in water. Direct realists also argue that the interpretation of illusions is based on hypotheses and hypotheses influenced by prior knowledge are prone to errors if the knowledge base is misleading. However, the determination of hypotheses to be erroneous in nature is not an easy process, because hypotheses are influenced by prior experience, expectations, and contextual information. All these three variables could be subjective in nature, and therefore, it cannot be stated with conviction that hypotheses could be misleading.
Biederman, I. (1981). On the semantics of a glance at a scene. In M. Kubovy and J. Pomerantz, Eds., Perceptual Organization. Hillsdale, NJ: Erlbaum
Carbon C. C. (2008). Famous faces as icons. The illusion of being an expert in the recognition of famous faces. Perception, 37, 801–806
Carbon, C. (2014). Understanding human perception by human-made illusions. Front Hum Neurosci. 8, 566. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4116780/
Carlson, N. R., Miller, H., Heth, C. D., Donahoe, J. W., & Martin, G. N. (2010). Psychology: the science of behavior (7th (International) ed.). Boston, MA: Allyn & Bacon (Pearson).
Costall, A. (1980). The three faces of Edgar Rubin. Perception. 9, 115.
Eysenck, M. W. (1998). Psychology: an integrated approach. Harlow: Addison Wesley Longman.
Favela, L.H. & Chemero, A. (2016). An ecological account of visual “Illusions”. Florida Philosophical Society, 16 (1), 68-93. Retrieved from https://philosophy.cah.ucf.edu/fpr/files/16_1/Favela_and_Chemero.pdf
Gibson, J. J. (1972). A Theory of Direct Visual Perception. In J. Royce, W. Rozenboom (Eds.). The Psychology of Knowing. New York: Gordon & Breach.
Gibson, J. J.(1966).The Senses Considered as Perceptual Systems. Boston: Houghton Mifflin.
Gregory, R. (1970). The Intelligent Eye. London: Weidenfeld and Nicolson.
Gregory, R. (1974). Concepts and Mechanisms of Perception. London: Duckworth.
Kohn, A. (2007). Visual Adaptation: Physiology, Mechanisms, and Functional Benefits. Journal of neurophysiology, 97, 3155-3164.
Shea, N. (2013). Distinguishing Top-Down From Bottom-Up Effects. Perception and Its Modalities. Retrieved from http://www.nyu.edu/gsas/dept/philo/courses/readings/2013.topdown.final.pdf
Watanabe, T., Nanez, J. E., & Sasaki, Y. (2001). Perceptual learning without perception. Nature, 413, 844-848.
Woodson, P.P. (1979). A Test of Two Theories of the Necker Cube Reversal Illusion. University of Richmond. 1265. Retrieved from <https://pdfs.semanticscholar.org/7573/8c68d6eae784b9dc4192061226132d126e89.pdf>