Gideon asked:

What conclusions, if any, can one draw from the inverted spectrum thought experiment?

Answer by Craig Skinner

None of note in my view.

The thought is that, for all you and I know, the sensations you have when looking at colours are the inverse of mine. So, looking at grass, you have the sensation I have when looking at red things, and, looking at a ripe tomato, you have the sensation I have looking at grass. Of course we both call grass green and ripe tomatoes red, having been so taught, so that there is no communication problem.

But what’s the point ?

The point is that this is one of the so-called conceivability arguments against physicalism.

Physicalism says that the physical world is all that exists, so that mental states (sensations for example) are just physical brain states, or aspects of these.

A conceivability argument says that we can conceive a brainstate occurring with a mental accompaniment different from usual, or with no mental state at all, so that whatever causes mental states, they are not wholly determined by physical states, hence physicalism is false.

The two best known conceivability arguments are the inverted spectrum argument and the zombie argument. The latter says that we can conceive of an atom-for-atom duplicate of you, with exactly the same brainstates as you, behaving exactly like you, but without any consciousness at all. Hence, whatever causes consciousness, it is not a result of physical brainstates.

I think these are poor arguments. I have two objections.

First, conceivability doesn’t necessarily mean possibility. Our imagination can outrun possibility. Right now, I can conceive my cat jumping up and typing the rest of this answer. But this is metaphysically impossible. There could of course be worlds in which cat-like creatures with superior intelligence do such things, but they would not be cats.

Secondly, advances in our understanding may show that the arguments contain conceptual confusions. Two such conceivability arguments that might have been advanced in the 19th Century illustrate this:

1. We can conceive of a container of gas in which the molecules move faster and faster but the temperature of the gas doesn’t rise. So, whatever temperature is, it’s nothing to do with particle speed, right ? Wrong, temperature just IS mean particle velocity.

2. We can conceive of a world containing tiny, replicating bags of chemicals undergoing complex interactions (let’s call them ‘cells’), but these cells are not alive, just little bags of dead chemicals. So, whatever life is, it’s not explained by complex chemical interactions, right? Again, wrong.

Life just IS complex interaction of dead chemicals in units drawing energy from the outside, maintaining dynamic stability and replicating.

So, I think the spectrum and zombie arguments may likewise fall to advances in cognitive science as we learn how particular brainstates necessarily entail, say, seeing red or being conscious.

Also, I find it completely implausible that healthy members of the same species would see colours differently. But few people hold that we really see colours differently, as opposed to this being merely conceivable. And, of course, colour-blind people with abnormal rods and cones do see colours differently, while birds, with more complex colour vision than primates, see colours we cant imagine and differentiate shades we would judge identical.


Answer by Jürgen Lawrenz

This is an enormously complicated issue on which many people have written. But it is possible to give a simple answer, when broken down into three fairly easy conditions:

1) If everyone has an inverted spectrum, it is undetectable and makes no difference.

2) If a small minority have an inverted spectrum, they will suffer from disablement in respect of social living (maybe with technology it could be remedied with special glasses which invert the inverted spectrum).

3) If an inverted spectrum is very prevalent, communications would break down. This might lead to societies based on the spectra of individuals. Power struggles would ensue between groups who claim to see things the “right” and the “wrong” way. But you understand this is a purely conjectural point. Nothing of the kind is actually known.

Point (1) above refers to the fact that there are no ‘natural’ colours, but only radiant information that is conveyed as colour impressions to our mind. When we associate words with those colours, we are assuming a universal sensitivity of all humans to these forms of radiation and a universal consensus (excepting pathological cases) among all humans as to colours, such that ‘red’, ‘rot’, ‘rosso’ etc. all bear the same meaning in different languages.

There is an evolutionary history behind this, and the rather important fact that our sensory equipment is living tissue. Its importance in the context of the inverted spectrum proposition is that the latter presupposes a form of sensory sensitivity that is derived from computer models. On this score one’s first doubts should arise if the proposition has any bearing on human existence at all.

On the above-mentioned assumption of a universal species characteristic, it is altogether reasonable to assume that all humans (except colour blind or jaundiced people) see the same colours. What we call these subjective impression is quite irrelevant as long as we identify objects in common and call the appearance of blood ‘red’ by any word in any of the languages we employ. If one person refers to blood as ‘red’ and another as ‘green’, there is still no problem, as long as the variance is consistent.

Yet the notion of an inverted spectrum would make some difference in terms of the secondary phenomenon of subjective moods which colours tend elicit. E.g. people vary much in their response to colours, feeling good or bad, aggressive or docile, cosy or uncomfortable, warm or cold depending on the colours in their surroundings. A person who genuinely saw red as green would have some trouble relating psychologically to colours in a way that would be understood by others. Again, however, this is a phenomenon only known to apply to pathological cases (e.g. jaundice).

But the strongest argument in favour of universality is art and design, and especially advertising art. It would be instantly dead if we did not all have the same kind of receptivity to colours.

And in this respect it is interesting that the colouring of many natural features carries information. Some fruits and insects wear colouring that scares off the birds from eating them. Whether these are seen as colours by the birds cannot be ascertained; but the case of ultraviolet vision by bees suggests that the answer is “yes”. This in turn suggests that colour vision is survival equipment. An inverted spectrum that differs from the norm would be very harmful to the survival chances of its owners, and this might be the very simple answer why there seems to be such nearly total agreement on sensory information across all humans and many species of animals.

The last point to deal with (which I mentioned in passing above) relates to the mechanical fallacy. A machine can be tuned to receive particular frequencies of radiation. It can therefore easily be made to output red as green, blue as yellow–which in fact we do routinely with some visual phenomena which contain colour information (e.g. from outer space) that our senses cannot deal with unassisted. But a human sensorium is not an aggregate of antennae. It is a congregation of living cells which receive and pass on the information by which they are physically affected. Mistakes can occur (constantly!); but there are safeguards in place to avoid the inverted colour spectrum possibility. Namely: huge numbers that must agree with each other. This is the way evolving organisms ensure their survival soundness.

On that basis, an inverted spectrum is not a feasible occurrence in any sound genetic pool, since survival fitness is tied to accuracy and stability of sensory information. And now colour blindness is not an inversion, but a genetic defect resulting in deficient production of photopigments, or it can be the result of damage to the retina. Jaundice is also a pathological state without any relation to spectrum inversion. So these two cannot be accepted as reference points for any notions associated with spectrum inversion.

The conclusion to be drawn from the thought experiment of an inversion of the colour spectrum is this: (a) It relies for its cogency on pathological states (i.e. deficiencies). (b) It does not meet with any known facts in the world, and (c) presupposes the applicability of electromechanical machinery as suitable models for its arguments (which in fact are inapplicable to living organisms like humans). Finally (d) it largely ignores the evolutionary trend from which colour vision emerged, i.e. the tuning of biological fibres to radiant information which ends up being inscribed on genes that are shared across a wide variety of life forms for uniform information processing.

For these reasons the thought experiment does not appear to be a coherent proposition. It lacks a sufficient reason for it to be taken seriously. It is a close kin of another game of the same ilk known ‘Mary the blind scientist’. These are nice game to play and eminently suitable for logical exercises – but only as long as the players are aware that they are not dealing with real life, but only with the rules of their game.