In a series of posts, I’ll review the current state of the field of the Evolution of Colour Categories. It has been argued that universals in colour naming across cultures can be traced back to constraints from many domains including genetic, perceptual and environmental. I’ll review these arguments and show that if our perception is affected by our language, then many conflicts can be resolved. Furthermore, it undermines the Universalist assumption that universal patterns in colour terms are evidence for innate constraints.
Part 1: Domains of Constraint
Part 2: Universal patterns are not evidence for innate constraints
For the full dissertation and for references, go here.
Our understanding of the world begins with perception. When we perceive an object, such as a piece of fruit, our sensory equipment measures features of the world such as pressure (touch), chemical composition (taste and smell) or the spectral intensity of light (vision). Our brains then process this analogue data into meaningful, categorical structures such as ‘a banana’. Thus we abstract away from the complexity of the world to create concepts. This ability helps us function in the world, and is not unique to humans (Hurford, 2007, p.9-16). However, humans are unique in the ability to attach labels to these concepts using language. We can use language to convey our thoughts to other humans and reconstruct the thoughts of others in our own mind. This study is concerned with two approaches to language. Universalism takes the position that language is subject to innate constraints from perception and the environment and therefore all cultures will have fundamentally similar understandings of the world. Linguistic Relativism, on the other hand, hypothesises that the structures of a language affect the way individuals think about the world. In this case, concepts can affect cognition and perception.
Since Berlin & Kay’s (1969) work on the World Colour Survey, this debate has been researched through the domain of colour perception. This has made the problem more tractable, since it restricts the domain to one sensory modality and the physiology of colour perception is well understood. However, with essentially no variance across populations in terms of the physiology of vision (Mollon, 1999), different cultures nevertheless seem to have strikingly different approaches to colour. For example, the Mangyans of the Philippines seem to have only 2 colour terms – that of ‘fresh’ and ‘dry’ colours (Ball, 2001, p. 15-16). In contrast, speakers of Tzotzil from Mexico appear to have hundreds of grammaticalised forms that describe colour (MacKeigan & Muth, 2006). Neither are colour categories stable across time – the English terms ‘yellow’ and ‘blue’ have the same etymological root, reflecting a time when yellow and blue were seen as different shades of the same colour (Ball, 2001, p.264). Nevertheless, universal patterns in colour term systems (e.g., Berlin & Kay, 1969) have been taken as evidence for universal, innate biases (Bornstein, 1973). This study argues that this inference is not valid. It is true that some colours are more functionally salient than others. For example, leaves are green whereas fruit are red-ish. Therefore, it is functionally useful to have labels that separate green and red. Furthermore, some colours are more perceptually salient than others (Jameson & D’Andrade, 1997). It has also been shown that natural colour term systems are optimised for constraints on the hardware of vision (Regier, Kay and Khetarpal, 2006) and for functionally classifying objects (Griffin, 2006). Universalism would argue that these innate biases have lead to universal patterns in cross- cultural colour term systems.
This section defines the domains of constraint on colour terms in a cultural system, and how they relate. By considering the importance of each domain, conclusions will be drawn about the Universalist/Relativist debate and whether the Cultural implication is supported. Explanations of the cross-cultural variation in colour terms have considered the effects and interrelationships of genetics, culture and the environment. For example, Bornstein (1973)first pointed out that the number of colour terms in a language is likely to be higher if that language is spoken further from the equator (quantified by Ember, 1978). However, the explanations of this pattern have included genetic, cultural and environmental constraints. For example, greater Ultraviolet radiation nearer the equator leads to changes in the eye which favour fewer colour terms (Lindsey & Brown, 2002, see section 5.2.1). This environmental constraint will lead to a genetic adaptation (Bornstein, 1973, Lindsey & Brown, 2002). Onthe other hand, it has been argued that cultures near the equator tend to be less technologically advanced, leading to less of a requirement for exact specification of colour (Magnus, 1880,Regier & Kay, 2004).
Under conventional interpretations, Universalism and Relativism differ with respect to the pressures that they predict will have the greatest effect on colour categorisations. Universalism predicts that linguistic categories are constrained primarily by perception(Nativism) and environment (Empiricism). Relativism predicts that linguistic categories are primarily constrained primarily by language and culture (Culturalism). I argue that if one takes an Embodied approach which emphasises the functional role of cognition, a different prediction is made for Relativism which allows it to explain more of the variation in cross-cultural colour term systems.
In order to do this, the relationships between the constraints on colour terms in a cultural system need to be outlined. First, the entities and processes involved with colour categorisation are identified (entities are in bold and processes are italicised): Genes produce Phenotypes through Ontogeny. This phenotype is immersed in an Environment. The Phenotype passes on its genes through Genetic Inheritance (e.g., sexual reproduction), dependent on Natural Selection pressures from the environment. The phenotype has sensory organs which interact with the environment to activate a perceptual space by a process of Perception. According to Embodied cognition, this space is also used for comprehension, so it is a Conceptual/Sensory Space. Motor control is also included in this space. This conceptual/sensory space converts sensory data into categorical concepts through Cognition. This cognition can affect the behaviour of the phenotype so as to change the Environment through Action. With an ability for Language, labels can be learned for concepts through Learning. These labels are also learned from other people by a process of Cultural Transmission.
Each of the processes above constrains the configuration of colour naming systems:
- Genetic inheritance constrains which genes an entity has.
- Genes constrain the sensory organs that can develop.
- Both the sensory organs and the environment constrain the range of stimuli that are perceived.
- Learning mechanisms and cultural transmission constrain which concepts are labelled.
The Figure below shows this relationship diagrammatically.
Two entities in a genetically and culturally related population are shown. Boxes denote entities and arrows denote processes. From this, we can see that some domains are linked by chains of causality, but others are causally separate. For instance, genes influence the phenotype, and so the perceptual constraints by extension. For example, having genes which produce two or three cone types will influence the physiology of the phenotype and the perceptions it is able to perceive. However, an individual’s genes are not affected by a phenotype’s physiology or perceptions (removing the eyes or turning off a light does not change an individual’s genes).
However, I hypothesise that the effect of categorisation on the conceptual/sensory space complicates the dynamics of the constraints on colour terms. Cross-cultural patterns could occur without innate biases through the following process: If comprehension is based on perception (the Embodied Cognition hypothesis), an adaptive pressure to improve communication will try to minimise perceptual differences between people. That is, by synchronising perception, people’s conceptual understanding of the world becomes closer. This requires a process that can change an individual’s perceptions to fit a shared configuration. If categories can affect perception (the Categorical Perception hypothesis), then, since categories can be shared through language, categories can warp perceptual spaces (e.g., Kuhl, 1991, Iverson & Kuhl, 1995) so as to bring people’s perceptual experiences closer. Systems are likely to converge on colour terms that optimally allow functional discrimination and that involve the least modification of the average perceptual bias. This will cause colour term systems to become optimised for constraints on the hardware of vision and for functionally classifying objects in the environment. Given similar perceptual hardware and survival pressures, this will lead to universal patterns in colour term systems. However, by this line of argument, the innate biases reflected in the colour terms are not the driving adaptive force, but a product of a communicative pressure acting in a cultural system. Therefore, even a Relativist approach to language would expect universal patterns in language. It is concluded that universal patterns do not necessarily support Universalism. Furthermore, an Embodied approach is favoured, since it causes a pressure for colour term systems to become more efficient at describing the real world.
The next few sections take a look at each of the domains of constraint above.
Mollon, J. (1999). Color vision: Opsins and options Proceedings of the National Academy of Sciences, 96 (9), 4743-4745 DOI: 10.1073/pnas.96.9.4743
Bornstein, M. (1973). Color vision and color naming: A psychophysiological hypothesis of cultural difference. Psychological Bulletin, 80 (4), 257-285 DOI: 10.1037/h0034837
Jameson, K., & D’Andrade, R.G. (1997). It’s not really Red, Green, Yellow, Blue: An Inquiry into cognitive color space Color Categories in Thought and Language DOI: 10.1017/CBO9780511519819.014
Regier, T., Kay, P., & Khetarpal, N. (2007). Color naming reflects optimal partitions of color space Proceedings of the National Academy of Sciences, 104 (4), 1436-1441 DOI: 10.1073/pnas.0610341104
Griffin, L. (2004). Optimality of the Basic Colours Categories Journal of Vision, 4 (8), 309-309 DOI: 10.1167/4.8.309
Ember, M. (1978). Size of Color Lexicon: Interaction of Cultural and Biological Factors American Anthropologist, 80 (2), 364-367 DOI: 10.1525/aa.1978.80.2.02a00100
Lindsey, D., & Brown, A. (2002). Color Naming and the Phototoxic Effects of Sunlight on the Eye Psychological Science, 13 (6), 506-512 DOI: 10.1111/1467-9280.00489
Regier T, & Kay P (2004). Color naming and sunlight: commentary on Lindsey and Brown (2002). Psychological science : a journal of the American Psychological Society / APS, 15 (4) PMID: 15043652
Kuhl PK (1991). Human adults and human infants show a “perceptual magnet effect” for the prototypes of speech categories, monkeys do not. Perception & psychophysics, 50 (2), 93-107 PMID: 1945741
Iverson P, & Kuhl PK (1995). Mapping the perceptual magnet effect for speech using signal detection theory and multidimensional scaling. The Journal of the Acoustical Society of America, 97 (1), 553-62 PMID: 7860832