The amount of information reaching our sensory organs every second would be overwhelming if it were not for our ability to categorize it. Colour perception is a good example of this phenomenon. When we pick strawberries, we can easily discriminate between unripe fruit and fruit of the many different shades of red that indicate ripeness. Writing in Nature, Caves et al.1 report that zebra finches (Taeniopygia guttata) can also perceive a continuum of colours as belonging to distinct categories, a phenomenon that affects birds’ ability to distinguish similar colours.
Although we can easily discriminate between the different shades of ripe strawberries, we tend to generalize and treat these shades as being equivalent. When comparing colours, if the differences between them are on the same scale of separation, our ability to perceive differences between colours from two separate categories, say ‘red’ and ‘orange’, is enhanced compared with our ability to perceive differences in colours that are both within one of these categories2,3. This enhanced ability to distinguish between colours if the colours are in separate categories is called categorical colour perception.
The preconditions necessary for the ability to perceive colours in distinct categories had already been demonstrated in birds. Humans and our close relatives have evolved to have three types of colour-sensing cone cell in the eye, and birds have evolved to have four types4,5. Birds have impressive colour-discrimination abilities4, including the capacity to perceive the ultraviolet range of the spectrum. A remarkable earlier study5 provided clear evidence that birds can generalize among certain colours, and thus divide the continuum of the colours that they perceive into discrete categories. But it was not known whether this ability affects how birds perceive similar colours and whether it helps them to spot key colour differences. Caves and colleagues investigated whether birds’ ability to categorize colours affects their colour-discrimination abilities, and thus whether these animals have categorical colour perception.
The authors created an ingenious experimental set-up. Female zebra finches were presented with a device in which food was hidden beneath coloured discs. Food was present beneath bicoloured discs and absent below discs composed of a single colour. This training scheme allowed the authors to test how well the birds recognized colour differences by their ability to identify bicoloured discs when searching for food.
The authors studied a range of colours from orange to red, evenly dividing this part of the spectrum into eight shades of colour. Caves and colleagues made great efforts, using physiological models of bird colour vision, to make all of the steps between the shades equivalently sized on the birds’ colour scale. These colours are worthy of attention because the zebra finch beak is red or orange. Beak colour depends on the amount of astaxanthin pigment deposited, which reflects the health of an individual’s immune system4, hence these colours might provide information about an individual’s fitness. Females seem to be able to discriminate not only between males that have red or orange beaks, but also between males that have beaks of differing red shades6. However, whether female preference for males depends on male beak shade is debated6.
Caves and colleagues first tested the finches using neighbouring pairs of shades from their eight-step colour scale and observed that birds distinguished between two of the shades better than between any other pair of neighbouring hues. This suggests that a putative border is present between the red and orange shades. The authors then investigated whether the birds were better at discriminating between pairs of colours of a similar level of shade separation that cross the proposed category boundary, compared with their ability to discriminate between colour pairs from one side of the category boundary (Fig. 1). Zebra finches passed this key test, demonstrating their capacity for categorical colour perception.
This result is fascinating and thought-provoking for many reasons. Birds are the only animals, besides primates2,3, in which categorical colour perception has now been demonstrated. More work should be done to investigate whether other aspects of colour, such as intensity and spectral purity, influence categorical perception in birds. It would also be interesting to determine whether zebra finches’ ability to group colours into ‘red’ and ‘orange’ has relevance for mate choice. However, this could be difficult to test because mate selection might depend on a range of male characteristics, such as the rate of male courtship displays7, rather than only beak colour.
The work also has implications for our understanding of human colour perception. There is an ongoing debate about whether language — including colour terms such as red, blue, green and yellow — influences colour perception. One school of thought holds that colour categories have a cultural and linguistic basis2. The hallmark of categorical perception — faster and more-accurate discrimination of colours in different colour categories — is seen only if a subject’s language has names for the specific colour categories being compared2.
The other school of thought contends that colour perception has a biological basis that is not dependent on cultural and linguistic influences. Evidence to support this viewpoint includes the observation that terms for specific colours cluster around the same hues across different languages3, and the fact that infants can discriminate between red, green, blue, yellow and purple before they have learnt the words for these colours3. Caves and colleagues’ finding that birds have the capacity for categorical colour perception adds more evidence to support the biological basis of this phenomenon.
Why might categorization be important, and how does it fit into the broader context of signal perception? The term ‘categorical perception’ was coined to describe the human ability to distinguish sounds in discrete units, called phonemes, that help to discriminate one word from another (such as the sounds ‘d’, ‘t’, ‘b’ and ‘p’ in the English words bad, bat, pad and pat8). Perception of phoneme-like elements also occurs in other animals, including birds9. Categorical perception could be described as a top-down mechanism to focus on key sensory cues by separating such signals from the enormous volume of irrelevant information. Another way to achieve this separation is a bottom-up approach termed ‘matched filter’, a concept which proposes that many animals’ sensory organs are designed as filters that perceive only the range of information that is relevant to the organism10. These two approaches could together enable animals to handle the vast amount of sensory input that is needed to inform their choices and behaviours.
The level of contribution of these processes, and how they evolved in different animal clades, are topics worthy of further study. Caves and colleagues’ work on zebra finches might be the start of a wider survey of categorical perception of colour in other animals.
Nature 560, 311-312 (2018)
Caves, E. M. et al. Nature 560, 365–367 (2018).
Roberson, D., Davies, I. & Davidoff, J. J. Exp. Psychol. Gen. 129, 369–398 (2000).
Skelton, A. E., Catchpole, G., Abbott, J. T., Bosten, J. M. & Franklin, A. Proc. Natl Acad. Sci. USA 114, 5545–5550 (2017).
Olsson, P., Lind, O. & Kelber, A. J. Exp. Biol. 218, 184–193 (2015)
Jones, C. D., Osorio, D. & Baddeley, R. J. Proc. R. Soc. B 268, 2077–2084 (2001).
Blount, J. D., Metcalfe, N. B., Birkhead, T. R. & Surai, P. F. Science 300, 125–127 (2003).
Collins, S. A. & Ten Cate, C. Anim. Behav. 52, 105–112 (1996).
Liberman, A. M., Harris, K. S., Hoffman, H. S. & Griffith, B. C. J. Exp. Psychol. 54, 358–368 (1957).
Nelson, D. A. & Marler, P. Science 244, 976–978 (1989).
Wehner, R. J. Comp. Physiol. A 161, 511–531 (1987).