Date: 5th May 2021

Speaker of the day: Adrian Palacios

Article: Thompson, Palacios, Varela, “Ways of coloring: Comparative color vision as a case study for cognitive science” (1992)

Abstract: In the lecture Adrian Palacios presents paper that he wrote together with Francisco Varela and Evan Thompson on the topic of color vision. In presentation and paper, he argues that when studying color vision, it is important to take into the account both objectivistic and subjectivistic account, called comparative research. Essentially, it is an argument how enaction comes to shape in area of color vision.

Keywords: color vision, comparative vision, computation, objectivism, ontology, subjectivism, enaction, codetermination



The main idea of the paper is to show how enaction comes to shape color vision. Palacios defines color vision with a very tiny window in the electromagnetic radiation which – when coupled with our perceptual apparatus – allows us to experience colors. With a short analysis of a newer paper titled ‘’Missing Colors: The Enactivist Approach to Perception’’ at the very beginning, Palacios shows us the difference between computational objectivism and neurophysiological subjectivism. The important thing here is to take into account the comparative argument which states that ‘’although animals differ in their neural apparatuses and environment, they share the ability of color perception. And although they share this ability, individual color experiences may differ regarding their dimensionality and sensitivity.”

Paper itself is organized in three parts which we follow: 1. Explanation in visual science and the ontology of color. 2. The comparative argument. 3. Toward the enactive view of color vision.

Color space: Paper argues that when understanding color vision, one must define a color space in which one maps one’s observations – this is called phenomenal structure of color. When studying color vision, we need to explain how this phenomenal structure is generated. For that we need a mapping system which will be able to transform the space of phenomenal color into a new color space, with which the reseacher will be able to address also the psychophysical and/or neurophysiological processes relevant to color vision. This way one creates a new database, which measures how the resource system transforms the signal that comes from the outside world into something that has physiological meaning for the brain. This approach allows us to start binding two previously separate areas of research together.

But color constancy in humans is not as stable as one might imagine; accordingly the naive computational approach did not take this into account . And that is why it is important to have another point of view on how colors are being processed in biological and experiential domains. Simply put, humans did not perceive along the lines of researcher’s computative models This is the point at which the comparative argument comes into play: even if different animals have similar apparatuses, when it comes to color vision they live in a very diverse environments. These variations imply different phenomenal color spaces. And that is why we need both experientialist and ecological approaches when researching color vision.

The color space in other animals: Palacios, Varela and Thompson created an experiment in which they mixed colors and trained pigeons to select objects which were similar with respect to the wavelength. Then they mapped this space (second paragraph) and realized that colors that can be distinguished by pigeons differ from those belonging to the human point of view. Visual pigment essentially creates something new and points toobjects that are important for animals yet more or less invisible to humans. Newer studies on hummingbirds confirmed that bird retina indeed contains 4 pigments (humans have only 3) by looking at which flowers they visited in an unrestricted environment and then mapping their spectral vision. Researchers showed that some flowers are closer to non-spectral colors that humans simply cannot see, which proved that wild hummingbirds are able to discriminate non-spectral colors, which seem to have behavioral meaning. At this point we can confidently say that subjectivism and objectivism alone by themselves cannot offer proper explanation nor solution. It is very important to mix both, keep in mind animal behavior in nature and to consider coevolution of for example flowers and hummingbirds.

Toward an enactive view of color vision: this is how organisms construct their environment and vice-versa. Animal-environment codetermination. Organisms determine, organisms alter the world, organisms transduce physical signals. The organism-environment relationship defines the traits selected in evolution.

Enactive approach: The approach of enactive theory needs to treat the environment not simply as the ecological setting for animal activity, but as something determined by that very activity. Today it seems even more evident that the relationship between the retina, the brain and the environment, is not finished , but rather an ongoing circular process. Brain, eye, fish, or human are all singular and have genes, proteins, neural networks that change over time. Everything needs to be considered.


1) Sebastjan Vörös (and partly Fabijan Purg): How has the scientific community responded to enactive approach? And in relation to that? When one reads the paper and the replies that it received, one notices how difficult it is for people from different disciplines to talk about this seemingly common phenomenon?

Adrian Palacios: The impact of this paper is the most obvious in computational science. Robotics are for example using the enactive approach (using vision for behavior). When it comes to neurophysiology: it is very difficult to work in enactive program, because you need all 3 parts: behavior, physiology, and ecology. If you do not have all 3 it does not make any sense.

Evan Thompson: Scientific response: for example, color-vision, enactivism, visual ecology, sensory motor and gibsonian ecological psychology have massively developed. Those things are much more interactive today. In philosophy most philosophers fall back into objectivist or subjectivist position. Not many try our proposed way.

2) Sebastjan Vörös: How do we approach the study of color? On the level of phenomena or underlying mechanisms? How could Goethe approach to color connect with this?

Adrian Palacios: You must select what you are experimenting with. The way to ask the question what color is, you need to go out. For me that is a problem. You are not always able to take care of both sides of the experiment. In this newest paper on hummingbirds’, birds were not restrained, and they could prove non-spectral colors – they see colors that humans do not see. But how do they experience it? I do not know. It is difficult. We should perhaps ask people who work on the neurophenomenology of color.

Evan Thompson: So already in the early emergence of color science we have proto neurophenomenology. And Varela has those sciences in mind. Here we have a lesson that we need to learn from Whitehead: not to fabricate nature. Hue or red is an event in nature and we must understand it as such. The task is to re-do the science without fabrication and that is what Goethe tried articulating in their time.

3) Matt Segall: Goethe says that theory is only useful insofar as it makes us believe in the nexuses of appearances. Whitehead says that red is the element of the process of nature. Science becomes the study of the relationship between phenomena. The question is then, to what extent is the enactivist project in-line with Goethe and/or Whitehead?

Adrian Palacios: What allows us to have the possibility to create colors? This is something energetically defined (which is very narrow). It is the physical condition that should be considered.

Evan Thompson: Goethe is someone who, in his dissatisfaction with Newton’s theory, insists on concrete visual experience within phenomenological domain as our touch stone to understanding what color is.

I never use the term enactivism: we always spoke of the enactive approach that can encompass very different theories and philosophies. As a philosopher I think that when we are talking about science, we should combine Whitehead and Husserl. Engaged and constantly opening through science and technology but not to use this to substitute nature.

4) Sebastjan Vörös (and partly Eva Ilič): Goethe was a great source of inspiration for Goldstein. He and other psychologists coming from the school of gestalt psychology developed a notion of color that is multi-dimensional. They observed how colors have something that Merleau-Ponty called ‘atmosphere of color’. Natalie Depraz joins: What I find interesting in Merleau-Ponty is kind of inverted perspective: color comes to me.

Adrian Palacios: I agree, this is a nice point of view. You have photons, but your brain cannot perceive photons directly. Plasticity of neural networks plays a major role.

5) Sebastjan Vörös: idea of color taking over you, evoking a certainly new bodily attitude. Would you say that this is related to the notion of codetermination?

Adrian Palacios: codetermination is much easier when you are trying to explain it – you can just use some data. The harder thing to explain is the experience of color. This is even more obvious when it comes to us humans, because we can use this experience (for example in painting) and work with it. I do not see any equivalent of this in non-primates. For me it makes sense what you mentioned about using color for your own progress, to grow. It is something you can constantly use. We are in another world than animals. We are more cognitive driven, living outside nature.

Natalie Depraz: Co-creation is more dynamic. Codetermination is a limited expression.

6) Sebastjan Vörös: Is the relationship more causal or internal (constitutive)? If there is only causal interrelation – you are still dealing with two separate entities. But there is also a stronger way of reading this (like Whitehead, Varela, Merleau-Ponty): the organism and its world are internally interrelated in a sense of mutually constituting. They are a part of a broader whole with 2 aspects. This has an impact on how we go about doing science.

Natalie Depraz: you could really reformulate it in Husserl words. Co-relation, co-constitution. The same way we talked last time about the fold.

Adrian Palacios: there are different time scales. One is millions of years where you can see coevolution, changes of genomes etc. The other aspect is how your nervous system is changing in your life with aging, pathological affection… so you end up with a brain that has less plasticity. Your experience of color vision is then going to be different. You are losing capacity.

7) Sebastjan Vörös: time is in this way like color. Notion of temporality – taking a processual approach instead of static is very interesting indeed.

Evan Thompson: causal question that you raised – it would be helpful to go back to Varela’s star perspective. Because the distinction between causal and constitutive is context dependent, it should not be set up in opposition. You can think of branching tree structure as a temporal ordered causal process. But when you think of them all as interfolded then they are constitutive of the network and form unity.