How do you prove that two colors like red and blue are different? I'm not talking about their difference in frequency, I'm talking about its perception. It seems to me that the only possible argument to justify they're different is because you can indeed, differentiate them, but then, with this I'd say this only proves that red and blue make me feel something different in order for me to tell which is which, but not necessarily a difference in my visual perception.

So how could I prove that this difference in visual perception exists?

  • 2
    "red and blue make me feel something different in order for me to tell which is which, but not necessarily a difference in my visual perception" - this does not make sense. Red and blue are perceptually different for you exactly because you can differentiate them. Color blind (or deficient) people may not pass Ishihara tests. – rus9384 Aug 29 '18 at 11:33
  • Relevant: Wilfried Sellars, Science and Metaphysics, esp. Part IV - if there was no (call it "qualitative" or whatever you like - but it surely has to be non-conceptual as it has to be pre-conceptual) difference, you would not have been able to learn the language game of different colours, i.e. to differentiate in response by calling the colours differently. At the same time, it is important to point out that this is not a propositional or conceptual state of the perceiver, as this would mean falling prey to the Myth of the Given (see Empiricism and the Philosophy of Mind) – Philip Klöcking Aug 29 '18 at 13:20
  • 1
    I don't know much about the philosophy of color, a very tricky subject. I'm surprised nobody's mentioned Mary's Room. Long story short Mary knows everything about the physics and physiology of color, but she's been raised in a black and white environment. One day she goes outside. Before, she thought she knew everything about color. But now, she knows what it's like to see red. She knows something she didn't know before. Therefore qualia are meaningful and real. They have to be accounted for any any physicalist theory. en.wikipedia.org/wiki/Knowledge_argument – user4894 Aug 29 '18 at 21:47
  • @user4894, because that's irrelevant? For physicalist brain has different parts. And know conceptually vs. empirically are different things. You can know how bicycle works, how to ride on it, but still riding a bicycle first time is a new experience. – rus9384 Sep 1 '18 at 12:59
  • @rus9384 I summarized the argument. Take it up with Frank Jackson. – user4894 Sep 1 '18 at 15:52

Your definitions of your terms will ultimately decide about the answer, and they are far from clear here.

But indeed, your sentiment can be understood as correct: The difference is trivially there (as we can differentiate) and nevertheless cannot be rigorously proven. Let me explain why ...

(There are other philosophical positions out there, but I consider all that I know of either metaphysically or logically incoherent)

Long (Sellarsian) answer

As you point out, light of different frequency - i.e. physical differences (I omit the technicalities here like that it is rather the absence of a frequency and that it is a combination of the signals of different rods etc. pp.) - may lead to different "feelings", but it is all but clear how they relate to colours.

A philosophical account of your sentiment can be found in the epistemology of Wilfried Sellars. There are at least three aspects of this relevant for your question:

First: Colours do not "exist" (but as perceived qualities)

As this answer correctly points out, Sellars thinks that science (and he probably means 'natural sciences' and not 'cultural sciences' here) determines what is and what is not - and qualia like colours are not part of the "scientific image" referred to in the answer. This understanding of "existence" is closely linked to the metaphysical branch of ontology, i.e. it does not preclude that it is part of our everyday experience and pragmatic context. In other words: Colours do not have independent being and only exist in the sense that they are our way to perceive certain attributes of physical things.

Second: We surely experience different colours and need an explanation for that

Colours are part of our "manifest image", i.e. common sense talk about the world. We experience the grass as green and the sky as blue as a matter of fact. Now, how is this to be explained? Sellars uses "sense impressions" for that (the following references are to his Science and Metaphysics: Variations on Kantian Themes).

Sense impressions have the following properties:

  1. They are (mental?) states of the perceiver.
  2. They are non-conceptual, i.e. they are outcomes of our sensual apparatus (receptivity) without intake from our cognitive faculty to think and speak, but can be described by minimal physical vocabulary (like "red rectangular shape", "white hemispherical surface"). (§§36-37,55)
  3. They "guide" (even minimal) conceptual representations without being an analysable "part" of those conceptual representations. (§39) This is one of his main points: Intuitions proper, i.e. apprehended, conceptual representations are in the space of reason (which is not to be assimilated into the physical or non-conceptual and vice versa) and can only occur (and constitute knowledge) in a person accomplished in language usage, see esp. his Empiricism and the Philosophy of Mind.
  4. Even if they "are" not red, blue, or rectangular, they have attributes systematically analogous to these (conceptual) properties. (§47)

To summarise:

physical objects bring about states of perceivers which have attributes systematically analogous to perceptible colour and shape, without literally having perceptible colour and shape (Science and Metaphysics, §47, p. 19)

This means, in the context of your question, that while there surely are states of the perceiver that correspond to colours without being colourful themselves (like inner images/perceivable representations/sense data), there "are" no colours, neither in the physical reality out there, nor in sense impressions. Colours "are" only insofar conceptual representations brought about by a combination of sensual input and conceptual understanding, sorting, and interpretation of this input are "conceived" as being of a colour, i.e. it is a conceptual attribute (logically) ascribed as a property:

neither the sense impression of a red rectangle nor the conceptual representation of a red rectangle is either red or rectangular (§52, p. 20)

In other words: Even if we perceive and attribute something as "being-red", this does not mean that "it" really "is" red in a material sense, but rather that we learned to (linguistically) respond to a certain attribute of "it" that stands in a certain systematic relation to the real physical object out there by ascribing "redness" to "it".

Third: What does all this mean for your "proof"?

As the short introduction into Sellars' epistemology has (I hope) shown, a difference in (visual) perception can only be understood if we presume a difference in sense impressions, which in turn means a difference in physical circumstances (either of the perceptual apparatus or the environment).

But the relations postulated here (physical object - sense impression - conceptual representation) are merely systematical and analogous relations, and not ones of (necessary, lawful) causation - that is exactly why we can "err" in perception. But in the end, the epistemological theory laid out here and by your description of "difference in feeling" is only that: A theory. (see §§53-55, pp. 21-22)

This theory is there to explain colour perception, i.e. the logic of the language we use talking about perception, its conditions and limits. Theoretical entities in discourses about the "manifest image", i.e. our everyday understanding and experience, are different from the theoretical entities in (natural) sciences - ultimately physical entities - in that they are metaphysical, i.e. in a certain sense not able of proof.

Therefore, only the simple fact that you can distinguish colours in perception "proves" that there is, in some sense, a difference in visual perception. Sellars agrees with your sentiment that there is no further proof and describes the "mature" theory (in reference to what Kant should have said) as follows:

If, per impossibile, Kant had developed the idea of the manifold of sense as characterized by analogical counterparts of the perceptible qualities and relations of physical things and events he could have given an explicit account of the ability of the impressions of receptivity to guide minds, endowed with the conceptual framework he takes us to have, to form the conceptual representations we do of individual physical objects and events in Space and Time. He could thus have argued that when on a certain occasion we come to have an intuitive conceptual representation that this green square adjoins that red square, we do so by virtue of having a complex of non-conceptual representations which, although non-spatial and without colour, have characteristics which are the counterparts of *square, red, green and adjoining, and which make them such as to account for the fact that we have this conceptual representation rather than that of there being a purple pentagon above an orange ellipse. (§78, p.30)


Since Kant, one of the main purposes of philosophy has been to discuss the coherence of metaphysical theories in the sense developed here. His main contribution to philosophy was the insistence that only those metaphysical entities that can consistently (and explicitly) be related to the real physical world are part of metaphysics proper.

That is why Sellars criticised many theories of perception of his time as they thought the transition from brute physical fact to meaningful perception as "given" or - in other words - mixed (contents and language of) the space of physical reality (brute fact) with (contents and language of) the space of reason (knowledge, inference, meaning), producing inconsistencies.

Fun-fact: Even people acquainted with Sellars (even citing him!) re-do the same philosophical mistakes up to contemporary philosophy, see e.g. John McDowell, 'Avoiding the Myth of the Given', in: Having the World in View: Essays on Kant, Hegel, and Sellars, Cambridge, MA: Harvard UP, 2013, pp. 256-72. He basically thinks the Myth avoidable if we only understand that it was (sheer, ineffable) content of receptivity that is part of and unified in our intuitions - by our understanding. He falls back behind Sellars with that sloppy analysis.


By asking for a (presumably scientific) proof for a difference in visual perception that is not part of the talk about visual perception, you would make a category error, asking from something that is bound by, but not identical with, the physical world (language of perception, common sense talk) that it should follow the same laws and logic (here: being able of proof) like it:

In everyday talk and common perception (special conditions aside), we obviously can distinguish between colours and talk about objects having different colours successfully in intersubjective discourse. We can only do so if we perceive these objects differently regarding something that makes us attribute different colours. Therefore, asking for further proof in this context indeed seems odd.

In 'hard' science, talk about "colour" is either nonsense or analogous, i.e. "proving" a difference in colour perception in this context is impossible as there is no such thing as colour perception here, only physical states of neurons, photons with a certain frequency, etc.

Judging across these very different discourses, like "There is no brown table over there, there is only a lot of void with loads of quarks forming something we could call "molecules" exerting forces on each other and absorbing and re-emitting certain frequencies of electromagnetic radiation over there!", is what philosophy calls "category error".

Therefore, no other kind of "proof" is possible as all there can be are analogous relations, not lawful or necessary ones.

  • "Colours do not "exist" - well, my prediction is that colors are just certain patterns in neural structure and then we may say they do exist. Assuming that identical material structures the same it is true, but then we must assume all perceptions exist in material (physical world). If not, then colors exist somewhere else, but still identical structures behave the same. – rus9384 Aug 29 '18 at 15:48
  • @rus9384: Identifying colours with certain patterns in the neural structure is reductionist physicalism and cannot explain the relations prevalent in this discourse (which is why a) dualism has always been so charming and b) there are so many careful arguments, e.g. by Sellars and Dennett, not being so blunt out there). In saying that they "are just... " you make an ontological judgement. This judgement is not justified - and due to category difference, probably not justifiable in principle - as of yet. – Philip Klöcking Aug 29 '18 at 15:54
  • "you make an ontological judgement" - rather, I make a testable theoretical statement and it may appear to be correct or incorrect. The most important thing is that we can know if this statement is incorrect, unlike in case of idealism. – rus9384 Aug 29 '18 at 16:23
  • @rus9384: You overestimate the methodological rigour of science (and the possibility thereof). In the end, you have to draw some relation between "neural states" and "colour". In physics, there is no "colour". Even if there is an experiment that is able to show that certain neural patterns occur (only) if there is a certain colour perception expressed, all that is shown is a relation of correspondence/correlation, not one of identity. Identity and causation are conceptual relations (not material ones), and they should not be drawn between objects of different conceptual frameworks. – Philip Klöcking Aug 29 '18 at 16:45
  • Then this really just reduces to what to count for "colour". We cannot know, of course, if other people really perceive colors as us. We can't even know if solipsism isn't the case. In either way, I define knowledge but its potential usefulness and understanding if others perceive colors as us doesn't make any use therefore there is no knowledge. Anything outside pf potential transportation of information does not make sense. If colors are informative, then they (or their perception) could be transferred through telepathy. – rus9384 Aug 29 '18 at 16:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.