1. Introduction
As the saying goes, “beauty is in the eye of the beholder”. This means that different people have different views on what is beautiful. It is a typical cognitive penetration which refers to the phenomenon that occurs when perception is subject to higher-level cognitive influences like affection [
1,
2], motivations [
3], theoretical presuppositions [
4], or linguistic representations [
5]. In other words, what a person believes, desires, intends, etc., may alter what one sees, hears, etc.
For the proposal that sensory perception is cognitively penetrable, it is currently a matter of debate as to what triggers such a phenomenon to take place, specifically, under which circumstances it is expected to be observed.
So far, there has been a substantial increase in the arguments describing the mechanisms underlying the occurrence of cognitive penetration. On the one hand, there are some philosophical and psychological theories explaining this phenomenon. For example, Macpherson [
6] proposed a two-stage and indirect mechanism to explain the occurrence of cognitive penetration. In their proposed two-stage mechanism, firstly, penetrating beliefs cause non-perceptual character. Secondly, that character would initiate the perceptual character that exists in humans. On the other hand, empirical explanations for cognitive penetration abound. For example, the study by Radel and Clément-Guillotin [
7] proved that cognition exerted a top-down influence via motivational factors. Lupyan et al. [
8] provided evidence that conceptual information can penetrate early visual processing. They revealed that cognitive penetration derives from category learning rather than simply a bias of perceptual system. Another experiment from Bruner and Postman [
2] associated positive, negative, and neutral symbols with different discs. The result turned out to be an overestimation for positive or negative associations, but not for the neutral one. The authors explained that salient visual input would appear to be larger for human beings. Conclusively, no agreement has been reached regarding the mechanism underlying the occurrence of cognitive penetration.
Meanwhile, documented works against the existence of cognitive penetration cannot be ignored. For example, in the arguments of Firestone and Scholl [
6], they pointed out several pitfalls that previous cognitive penetration demonstrations were caught in. Because most studies committed one or more of these mistakes, they held that none of the evidence for top-down influence on perception or the alleged cognitive penetration was compelling enough. In addition, some early works opposing cognitive penetrability held that the process for visual stimuli was cognitively impermeable, and top-down influence only biased the decision-making process [
9,
10]. The work of Klein et al. [
11] failed to replicate the findings from Bruner and Postman [
2]. They argued that distortions in size perception were merely derived from the method used to ask for size judgments, rather than value cognition. In a recent review on cognitive penetration, the authors stated that those works against the existence of cognitive penetrability emphasized the distinction between cognition and perception. This is the classical modular understanding of perception and cognition, according to which perception is an automatic modular system independent of conscious control [
4].
It can be found that previous debates about the existence and mechanism of cognitive penetration mainly focused on the distinguishing boundary between cognition and perception. Considering the long-standing controversy of the occurrence of cognitive penetration, the current research aimed to provide empirical evidence for cognitive penetration from the perspective that the cognitive processing priority over perceptual processing is critical for cognitive penetration. We made this proposal because we found that inconsistent findings on the occurrence of cognitive penetration were related to processing priority between cognition and perception being differently manipulated by researchers. For example, in a work typically proving the existence of cognitive penetration, researchers presented participants with words close to the threshold of conscious perception. The results suggested that participants who had fasted were more likely than satiated participants to perceive masked food-related words [
7]. In contrast, studies demonstrating that cognition does not influence perception manipulated the two aspects in another way. One central phenomenon offered in support of impenetrability is the Muller-Lyer illusion. That is, even if we know that the two arrows have the same length, we continue to perceive one as being shorter than the other [
12]. Therefore, it can be noted that studies showing cognitive penetration deployed ambiguous stimuli during perception decision, leading to cognitive priority over perception [
1], whereas those studies directly presenting stimuli during perception failed to demonstrate cognitive penetration. From the distinct manners of manipulating the relative processing priority of cognition and perception, it can be inferred that cognitive priority over perception may be a critical factor for triggering the occurrence of cognitive penetration. Thus, this study provided empirical evidence for the critical trigger for cognitive penetration. The work of Radel and Clément-Guillotin [
7], which used ambiguous experimental stimuli to induce cognitive penetration, concluded that humans interpreted unclear stimuli based on their willingness and desires. This finding was consistent with a consequent proposal from Dunning and Balcetis [
5]. Their review held that people categorized ambiguous information from their surroundings in a manner that was in line with their desires.
Considering these debates on the occurrence of cognitive penetration, we assumed that the ambiguous stimuli which led to the top-down influence from cognitive penetration in previous works may derive from the weaker perceptual processing priority compared with cognitive processing priority for ambiguous stimuli. According to previous work about working memory functions, contents that were absent but stored in working memory were worse in perceptual fidelity but better in cognitive parts like values and rewards [
13,
14,
15,
16]. Therefore, stimuli that are absent but stored in working memory may also lead to cognitive penetration in a way similar to that of ambiguous stimuli. Following this logical line, we hypothesized that cognitive penetration was the consequence of stronger cognitive processing over perceptual processing. Therefore, this current research aimed to examine this assumption.
We constructed experimental conditions in the realm of perception judgment based on coin value information. Following the classical paradigm in the work of Bruner and Goodman [
17], who revealed that poor children have a bias to overrate coin size, the current study designed three experiments in the paradigm of perceptual judgment for coins of different values. We chose this paradigm for two reasons. Firstly, the coin size perception task was one of those classical paradigms demonstrating cognitive penetration. Manipulations of a critical independent variable, i.e., the cognitive processing priority over perception, in this paradigm would make our findings more general and robust. Secondly, the coin stimuli were reported to be more related to the cognitive control network than other stimuli such as faces, which are related to emotional pathways as well as the cognitive control network in the brain [
18,
19,
20]. Experiment 1 was designed to disclose the fact that cognitive information from social value cognition would penetrate visual perception. It has been reported that ambiguous stimuli lead to cognitive penetration [
2,
7]. Experiment 1 aimed to test whether absent stimuli during perceptual decisions had the same effect of cognitive priority over perception during decision making. Specifically, participants were asked to judge the sizes of coins which were associated with different values before making a perceptual judgment. Experiment 2 aimed to assess the contribution of perception in triggering cognitive penetration. Considering previous studies highlighting the role of stimulus ambiguity in the top-down influence of perception [
7,
9], we hypothesized that perceptual weakness is a critical factor for cognitive penetration. In Experiment 2, participants went through a reproducing task where they were asked to adjust the size of a circle to reproduce the size of a coin simultaneously presented with the adjusted circle. According to our hypothesis, no significant difference between the adjusted radius for the circle associated with a high or low value should be observed. Experiment 3 was designed to trigger cognitive penetration again by withdrawing the perceptual input when making size judgment. We expected to observe a significant difference between circles associated with different values under the condition of poor perceptual input.
5. Discussion
The reported experiments examined the influence of value on size perception. Three experiments in the paradigm of perceptual judgment for coins of different values were conducted. In Experiment 1, we observed a cognitive penetration phenomenon where participants’ size judgment was subject to value information. Specifically, participants were asked to judge the size of coins which were different in value before making a perceptual judgment, but the coins were actually the same physical size. The result suggested that participants chose the coin associated with high value as the physically larger one, which means that perceptual judgment can be influenced by cognitive processing of stimulus values, even though the stimuli are absent. Experiment 2 was designed to assess the contribution of perception in triggering cognitive penetration. Participants went through a reproducing task where they were asked to adjust the size of a circle to reproduce the size of a coin simultaneously presented with the adjusted circle. We found no significant difference between the adjusted radius for the circle associated with high or low values, suggesting that it is possible that no cognitive penetration occurred during this situation. This suggested that when visual input was present during the size judgment, the perceptual judgment would be immune to valuable information. Experiment 3 was designed to trigger cognitive penetration again by segregating the perceptual input and cognitive input to enlarge the processing difference between perception and cognition. The results suggested that cognitive penetration came back after the perceptual input was not presented. With three experiments, we proved that the processing priority of cognition over perception was a necessary trigger for cognitive penetration.
Our findings contain two aspects. Firstly, the results suggest that value information associated with an object impacted the size judgment of that object. There are several plausible perspectives to explain this phenomenon. From the perspective of cognitive control, the top-down modulation of cognition on perception has been empirically evident for a long time [
1,
7]. According to Bar’s model, ambiguous information from visual input may be interpreted unconsciously as information matched with representations in memory, which reveals how top-down factors can influence early visual processing. From the perspective of prediction error minimization theory [
4], perception and cognition are continuous, which contributes to pattern matching between incoming external stimuli and internal mental representations. Therefore, information from the external surroundings should be extracted in a way that is consistent with predictions. From the perspective of cognitive neuroscience, a recent study, which suggested a neural basis for cognitive penetration, proved that the visual cortex encodes value information associated with objects. This study pointed out that representations within the visual area were alterable with associated values [
21]. In addition, it was reported that a node of the executive attentional network involving the anterior cingulate cortex (ACC) was closely connected to sensory systems based on dynamic bottom-up and top-down interactions [
22]. The attention network was recorded to be frequently activated by conflicts between stimulus attributes [
23]. Another neuroscientific investigation presented participants with arrays of moving dots whose direction of motion was difficult to discern. Participants were required to report the predominant motion direction, before which participants learned that they would receive a reward if the dots predominantly moved in a particular direction. The results suggested that people tended to see the dots moving in the direction associated with reward, and did so by biasing their visual search in favor of the desired perception. Moreover, occipital regions involved in perceptual encoding and prefrontal regions involved in top-down control were modulated by these motivational effects on vision [
24]. Combined with the findings of the current study, it can be inferred that value might influence size perception via the connections between the ACC and sensory system of humans, providing evidence for the proposal that factors of internal state-like values could modulate brain mechanisms through connection pathways. Further research is required for this demonstration.
Secondly, but more importantly, the cognitive processing advantage is a necessary condition for the occurrence of cognitive penetration. In our study, the cognitive processing advantage was achieved via the absence of perceptual stimuli, which resulted in insufficient or impaired perceptual processing, when making perceptual judgments. This finding makes sense when considering the proposal that there is no definite boundary between cognition and perception. It has been proposed that conscious feelings of human beings are combinations of both top-down stored knowledge and bottom-up incoming external information [
4,
25]. Perception and cognition are continually interacting with each other [
26]. Under this premise, the findings from the current study confirmed cognitive penetration in that the absence of visual stimuli contributed to insufficient sensory information to make judgments about the external world. In other words, insufficient perceptual input led the information coming from other systems to be indispensable, thus triggering cognitive penetration. When perceptual input was presented, the cognitive penetration disappeared, as was observed in Experiment 2 of our study. However, it must be mentioned that the current study did not involve finding the boundary condition to trigger cognitive penetration. Further investigation is needed before the boundary condition for perceptual input to trigger cognitive perception is detected. Our current findings also fuel the notion of embodied cognition by revealing that visual size illusion may result from the social value of the stimuli [
27]. From the perspective of embodied cognition, a plausible explanation for our experimental findings is that judgment based on the physical size of perceptual stimuli is greatly influenced by cognitive learning of the social information contained in the stimuli.
Despite the consistency with previous findings, the current study proposed another interpretation from the perspective of cognitive resources. Top-down cognitive processing and bottom-up input are continuously competing for limited cognitive resources. When the external input was withdrawn in Experiment 1 and Experiment 3 in this study, cognitive processing took the advantage of more cognitive resources. Therefore, we proposed from this study that the critical point for the occurrence of cognitive penetration is the processing priority of cognition over perception.
Some limitations of this study are worth noting. Firstly, the insignificant results in Experiment 2 implied that cognitive penetration was likely dismissed in the situation of implementing perceptual input when making size judgment. Nevertheless, the possibility that the experimental paradigm in Experiment 2 was not suitable for testing cognitive penetration cannot be ignored. We ask for future studies to test this experimental paradigm in search of cognitive penetration. Secondly, the findings of this study were obtained from college students, but whether these findings can be generalized to broader populations, especially those whose income level is different from that of college students’, still requires further exploration. Lastly, but most importantly, according to a previous discussion about how beliefs on stimulus colors influence how people perceive the color [
28], there is a distinction between cognitive penetration and perceptual decision when stimuli are absent. Thus, rigorously speaking, Experiment 1 did not prove that cognitive penetration occurs, but that perceptual decision is influenced by cognitive processing regarding valuable coins.