We like the original versions of abstract artworks more than colour-shifted ones

Take a look at this 1930 painting, "Rhythm, Joy of Life", by French artist Robert Delaunay.

Do you find it colourful?

And do you like it?

Now what if every pixel in a digital version was rotated an equal distance on a "colour wheel" that represents every colour that people can see?

Technically, the number of different colours in the image would be the same — but you'd probably perceive it to be less colourful. And, even if you'd never seen the original before, you'd probably like it less.

That, at least, is the conclusion of a fascinating new paper in the Journal of Experimental Psychology: Human Perception and Performance. The work contributes to our understanding not only of why certain colours are more common in art, but also of how we perceive colour.

Carolin S. Altmann and colleagues at the Experimental Aesthetics Group at Friedrich Schiller University Jena, Germany, first identified 100 colourful abstract artworks by well-known artists. They then rotated each through a 360-degree wheel of perceivable colours, in steps of 60 degrees at a time. This gave them six differently coloured digital versions of each artwork, so 600 images in total.

Twenty young participants with normal vision used six-point scales to rate each version of every artwork for "colourfulness" and, separately (on another day), for how much they liked them.

The team found that all but one of the participants felt the originals to be more colourful than the other versions. Also, about three quarters of the participants preferred the originals over the colour-shifted versions.

"This effect is not only surprisingly consistent across participants, but also across images, and even survived rigorous statistical control of potentially confounding factors," the team reports. In fact, 81% of the original images were both more liked and judged to be more colourful. The team's analysis also revealed that the participants didn't simply like very colourful images per se — rather, "they prefer versions of artworks that they perceive as more colourful".

Each rotation included the same number of perceivable colours. So — assuming that the participants were unfamiliar with the artworks (and this isn't reported on in the paper) — there must be something special about the colour palettes chosen by the artists, the team reasoned.

Their analysis revealed that three quarters of the artworks featured more yellow and orange than other hues. (This was also the case for another 113 artworks that the team had originally identified for possible inclusion in the study). Other research has also found find a bias among artists towards the yellow/orange range. So why might this be?

More yellow in a picture might make it seem closer to a scene lit by natural daylight (even if it's an abstract image), and so preferred, the researchers suggest. But also, the human visual system is less sensitive to blueish-yellowish input than to red, say. (It's been suggested that this made it easier for our ancestors to quickly spot reddish ripe fruit under changing light levels, from dawn to dusk.)

"Potentially, artists thus have to overrepresent yellow colours in their palette to make them just as noticeable to the human observer as other colour hues with higher perceptual sensitivity," the team writes.

"Yellow" is also a narrow "colour category". That is, the number of different hues that we perceive to be examples of yellow is pretty narrow; it's much larger for red, say. So while a rotation may easily have changed a yellow pixel to one in another colour category, a red pixel may still be a type of "red" — which could make the artwork seem to be less colourful and also less complex. Other studies have found that more complex artworks tend to get higher aesthetic ratings, so more apparently complex versions of the same painting might be preferred.

Not all languages use the same colour categories, however. For example, "light blue" and "dark blue" are fundamentally different categories in Russian and Greek — in which a colour cannot simply be called "blue".

It has been suggested that the colour categories that a person is familiar with affects their colour perception. It would be interesting, then, to see if a study using participants who use different colour categories would give different results.

It's also worth noting that the majority of participants in this study were women, and other work has found that men and women don't perceive colours in quite the same way. There are, then, still a lot of questions in relation to this work — but it'll be really interesting to see what the answers turn out to be. 

Further reading

– Liking of Art and the Perception of Color

About the author

Emma Young (@EmmaELYoung) is a staff writer at BPS Research Digest