A new paper adds fuel to a long-standing debate.
14 April 2026
By Emma Young
This long-debated question has been investigated in a number of studies, with some suggesting that prior knowledge, emotions, and desires can influence visual perceptions — however, as the authors of a recent paper in the Journal of Experimental Psychology: General write, the idea is still hotly contested.
Michael A. Cohen at MIT and colleagues decided to focus on probing one particular claim: that a person's prior knowledge of the colour of an object can influence the colour that they see. Their work suggests that this can indeed happen — at least, in certain circumstances.
The researchers first identified sixteen colourful objects and images that they were sure would be very familiar to the 15 student participants that they recruited for the main experiment. These included the Facebook logo, the American flag, the Incredible Hulk and a road Stop sign.
Each participant first went into a dimly lit room, where they were shown glossy grey-scale posters of each of the 16 stimuli. Each time, they were asked if the image seemed to be in colour or black and white — and if they did see a colour, they were asked what that colour was, and how vivid it was. The participants then repeated this procedure, but in bright light.
Then, back in dim light, they were shown the posters again, but this time, the images had been mostly covered, so that only a small segment was visible. An experimenter did tell them what the full image was, however. Again, they were asked about what colours, if any, they saw.
When the team analysed the data, they found that, when the participants viewed the full grey-scale images in bright light, they did not see any colours. Neither did they report seeing any colours when they were shown only small, ambiguous portions of the images in dim light — despite being told what the images were.
However, in dim light, when viewing the full grey-scale images, the participants consistently reported seeing the corresponding colours. So, when viewing a grey-scale Facebook logo, for example, in bright light they reported seeing no colour, though in dim light, it had appeared to be blue. "Overall, we believe that these findings demonstrate a robust and subjectively appreciable memory colour effect under ambiguous viewing conditions (i.e. dim light)," the researchers write.
It was notable that even when the participants were told what the full image was, when they saw only a small part of it, they didn't see colours. "This result suggests that although top-down knowledge can induce a memory color effect, it cannot do so in the abstract alone; there must be sufficient visual input that such knowledge can act upon," the team notes.
It's worth noting that the sample size in this study, of just fifteen, is small. However, their findings do fit with the popular 'predictive processing' model of perception. According to this model, the brain uses both raw sensory data and expectations, based on prior experience, to generate sensory perceptions. However, the theory goes, if one is lacking — if the sensory data is of poor quality, or, alternatively, if the brain can't identify prior experiences to draw on — it will rely more heavily on the other source of information to generate perceptions. This is exactly what seems to have happened in this study.
Understanding when and how the brain 'fills in' colour has potential practical implications for improving safety and communication in the dark, for example, and designing signage for low-light settings, the team thinks. But, by providing clear evidence that existing knowledge can influence colour perception, this study also contributes to the ongoing debate about when, and in what circumstances, 'top-down' knowledge can have an impact on what we see.
SOURCE:
No comments:
Post a Comment