ManticMoo.COM All Articles Jeff's Articles Scholarly

The Ambiguous Existence of Colors
Review by Jeffrey P. Bigham

Bertrand Russell formulated two main arguments that dealt with the question of whether or not objects in the external world are actually colored. In these he addressed his concern that there may be no basis for the assumption that they do even though it seems intuitively obvious.  In this paper these arguments will be explained and scrutinized in an attempt to arrive at the true nature of colors and their existence as an inherent quality of objects.

Russell's first argument relies much on the premise that objects can appear to be made up of different colors, even if they are completely uniform, depending on the observer's prospective and other variables.  Russell used the example of a brown table to illustrate his point.  He said that while "I believe the table is 'really' of the same colour all over, the parts that reflect the light look much brighter than the other parts, and some parts look white because of reflected light (8)."  This means that while the table is regarded as uniformly brown, in response to looking at it from different points of view the color may change.

As a result of this, two people viewing the table from different places will not perceive it to be the same color.  Furthermore, even from the same viewing point the table will appear dissimilar under various types of light, to a person that is color-blind, to a person wearing colored glasses, and when the light is otherwise altered between an observer and the table.  For these reasons Russell says that there is then no color that can be considered to be the one and only color of the table.  Instead, the color is completely dependent on the conditions under which it is viewed and the qualities of the particular person who is viewing it.

Russell further contends that while we may refer to a certain color as being the color of the table, this is only "the sort of colour which it will seem to have to a normal spectator from an ordinary point of view under usual conditions of light."  The table has many different colors dependent on the circumstances under which it is viewed, and none of these colors can be considered more real than can any of the others that are present.  Russell argues that because of this, the table cannot be said to have any particular color.

As unintuitive as Russell's first argument may be, it does appear to be valid as long as its underlying basis is correct.  One premise that Russell gave only minimal attention to is the precise definition of what exactly it means for an object to have a particular color.  He seems to take the view that if colors exist then they would be something inherent in an object and virtually unchanging regardless of viewpoint if the object is uniform.  While at first this seems correct, a different, more specific definition of what exactly colors are may not fail under Russell's argument.

Objects that do not produce their own light in some sense have no color of their own.  The perceived color "is not something which is inherent in the table, but something dependent upon the table and the spectator, and the way the light falls on the table (9)."  But what if this property is the true color of the table?  In such situations, it seems that the color of the light of this object would be better defined as the way in which incoming light is altered and reflected rather than by any one specific color of the rainbow.  Instead of saying that an object's color is "red", the color could be described as the property that when light is shined at the object the red wavelengths tend to be reflected while the others tend to be absorbed.  Further constraints on the properties of the object’s color could be added to take into account other qualities of the object that affect the perceived color, such as how the light reflected from the object changes as the incident angle of the incoming light rays and the angle to the observer changes.

Such a definition could nullify much that Russell used in his argument to reason that colors do not exist.  For example, it would explain why the table might appear to have bright splotches at certain locations.  Using this new definition of colors, these bright splotches would be an expected and necessary part of the color itself.  Because standing in different locations changes the angle between the light source, the object, and the observer, it would also not be a surprise to discover that the perceived color changes according to viewpoint.

However, this new definition does not necessarily contradict the conclusion of the argument that Russell presented because it could be challenged on the basis that it has only been shown that objects contain this alteration and reflection property and not that they contain what are normally referred to as colors.  However, the properties presented in this new definition basically are what is accounting for the color of the object and, therefore, could be considered the "real" color of the object.  Regardless, Russell's conclusion becomes more ambiguous when this definition is taken into account, and becomes even more so when objects that produce light on their own are considered.

All objects that are above absolute zero in temperature release some wavelength of radiation dependent largely on their temperature and molecular makeup.  Of course the wavelengths which are included in visible light are usually only produced in this manner by objects that are fairly hot, like glowing coals or the Sun, so most objects appear to not have any color of their own.  If, however, the definition of "color" is extended to include radiation of all frequencies - not just those included in visible light - then every object has a color at any given moment because every object in the universe has a temperature above absolute zero.  If these objects are viewed without the colors (as defined as alterations and reflections in the above paragraph) of the intervening matter changing the light before it is perceived, then every object would have a certain, predictable color at any one point in time.

Using this definition, objects appear to have inherent colors as long as the objects themselves exist.  Of course this definition of color is far from certain as it may again be said to describe something other than what is normally referred to as the color of an object.  Furthermore, while the color may be relatively certain at any given moment, this color will change if the temperature of the object changes and, if the object's color is variable, any one color cannot be said to be the true color of the object.  These differences are somewhat unlike the ones observed by changing the viewpoint in Russell's original argument, however, in that they are predictable and are the same for everyone with "normal" vision looking at the object at any given moment.

This examination of what exactly it means for an object to have a certain color does not by any means prove that Russell's argument is based on an invalid premise, but it does offer support for that possibility.  While most people think it is easy to know what a color is like after they have seen an example of it, it is much more difficult to describe what it means for an object outside of their own minds to possess the corresponding quality.  Until this is possible, Russell's first argument is based on a premise that is only likely to be true but not certain.

As part of his first argument against the existence of colors, Russell concludes that the perceived color of an object is different when the person looking at it is color-blind or wearing colored glasses, among other things.  Based on that conclusion, he says that since no one color can be said to be the color of the object, then the object cannot be said to have color.  This conclusion raises the question of whether all of the properties of something have to be known or even be knowable to determine other properties of that thing.  Particularly, does the precise color of the object need to be known in order to know that the color exists? This does not seem to be the case in regards to other relations of this type.  For example, if a person is known to be playing either basketball or soccer, it would seem that the particular sport that he or she was playing need not be known in order to determine that the person is playing one of the two.  Similarly, if a color is seen when looking at an object and that color varies all across the spectrum, it would seem that the particular color of the object need not be known in order to determine that it has a color of some sort.  If a color is perceived from an object, it may be impossible to determine precisely what color is the true color of the object, but it seems perfectly reasonable to say that a color exists.

Additionally, just because a color-blind person and a person wearing colored spectacles receive a different sense datum when viewing an object does not mean that the color of the object itself has changed. The human eye is only able to detect three wavelengths of light and the brain combines and interprets these to come up with the colors that are perceived. A color-blind person is usually missing one or more of these detectors, which means that he or she can detect only one or two unique wavelengths and, therefore, can combine them into far fewer distinct colors. Furthermore, a small number of people are able to detect four wavelengths of light, which allows them to distinguish many more colors than people with "normal" vision can. While such a person may not be able to see what those with normal vision consider to be "red", for example, they still see the same color whenever they are presented with something red.

Regardless of how many colors can be interpreted, the color that is perceived is still unchanging for the person viewing it, and would appear to the person to represent the same color that the rest of us see. Of course, problems arise when a color-blind person tries to distinguish two colors that look the same for him or her but different for those with normal vision. Concluding that because they perceive a different representation of this color then no particular color exists is like determining that colors do not exist for those that see because those that are blind do not see them.  Both conclusions seem unintuitive and incorrect.

As Russell stated, colored spectacles do alter the color perceived to be emanating from an object, but this does not mean that the color of the object itself changes. Another view of this occurrence is that when looking at objects through colored glasses a person is able to look simultaneously both at the color of the glasses and at the color of object (the coloring of the spectacles happens to allow some incident colors to pass through and others not). The blending of these colors results in the change, but since it occurs after the light from the object has left the object the true color of the object can hardly be said to have changed.  Russell deals with this alteration of light rays by intervening matter in his second argument.

He begins this argument by formulating a possible technique for arriving at the true color of an object.  Because an object will generally have similar colors when looked upon from various points of view, he hypothesizes that the true color of the object might be the average color over this range.  This method would allow the determination of the true color of any object, and, if the method were found to be correct, it would provide additional evidence to the position that colors exist.  While Russell says that it might not be possible to completely disprove this method, he does see promise in the attempt to disprove its basis.

The colors that humans perceive are based solely on the light waves that reach the eye, and therefore are "modified by the medium intervening between us and the object, as well as by the manner in which the light is reflected from the object in the direction of the eye (35)."  First, Russell contends that because air affects the perceived color unless it is perfectly clear and because any strong reflection will alter them completely, colors are not just a property of the object from which the ray comes.  Because colors from the objects cannot be viewed alone, it is impossible to determine whether or not they exist independently.

Secondly, Russell says that as long as the necessary light waves reach the eye, colors will be perceived regardless if the object has a color or not.  This statement connects Russell's color argument to the more general theory that human beings may just be "brains in a vat" who are being fed in what Russell calls sense data.  Because it is impossible to determine whether these sense data are actually coming from some real object or not "it is quite gratuitous to suppose that physical objects have colours, and therefore there is no justification for making such a supposition (35)."  Certainly, they seem to have them, but they would appear exactly the same if the required signals were otherwise sent to the brain.

The colors that are perceived in daily life may be a product of more than just the object from which they originate, but if conditions are set up carefully a better understanding of the true color of an object that would hold up more diligently under Russell's scrutiny may be found.  To satisfy his first objection dealing with how intervening air or other matter might affect the perceived color, the object could be placed in a vacuum.  To deal with any reflection that might occur a standard light source could be used when determining colors or, as described in the analysis of Russell's first argument, the object could be heated to a standard temperature.  These steps should help insure that the true color of the object is seen.

Unfortunately, it is often unwise to assume that a particular case will extend to the general case as has been done here.  While it seems that the particular case in which all intervening matter has been removed and a standard light source applied would provide the true nature of any light rays coming from an object, this is not necessarily correct.  The standard light source might help establish what is determined to be the color of the object, but, because the object itself cannot know such a standard, it cannot be used to determine the object's true color.  Furthermore, the situation is not even as simple as this because there is no clear way to address the fact that the light rays will still be altered by the observer's eye before becoming incident on the retina and being transformed into sense data.

Even if matter will always be in between an object and the observer, that still does not necessarily mean that it is impossible to determine if colors exist.  Russell based his conclusion on the fact that any colors that are seen in this situation will be a property not only of the object but of the intervening matter as well.  This seems to suggest that if objects do not have colors themselves then these colors would not exist without the intervening matter.  But this seems to suggest that either the intervening matter – which could also be considered an object – would be required to have a color or somehow the combination of the two results in a color.  The first option seems incorrect because there is little reason for the intervening matter to have a color while the original object does not, and, while the second option is possible, it does not seem probable.

Objects usually have a similar color regardless of the medium through which they are viewed unless that medium itself has a color, as with the colored glasses mentioned in Russell's first argument.  In addition, objects that generally have different colors will still have those different colors even if they are viewed through the exact same intervening matter.  Because the intervening matter seems to add only trivially to the color of an object, it does not seem likely that they are responsible for it.  These facts help cast doubt on Russell's conclusion, for, while the color that is perceived might be a property of both the object and the intervening matter, it is likely that this matter only alters the original color from the object.  If it only alters the color from the original object, then the original object would seem to necessarily have a color from the beginning.

The problem, as Russell mentioned, is that there is no way to test whether an object has a color all by itself because there is no way to completely remove all of the intervening matter.  It is possible to approach this situation in the manner mentioned earlier, but, as shown before, the desired conditions can never be completely obtained, so no definite conclusion can be drawn.  Therefore, while this argument adds support to the basis that Russell argued against, it cannot by any means can prove that objects truly have colors.

In his arguments, Russell presented some interesting ideas that were, for the most part, valid, but some of the assumptions that he made were quite questionable and others really didn't hold up when scrutinized closely.  While the analysis presented in this paper was not able to show that colors necessarily exist, it did raise many ambiguous details of Russell’s arguments that he did not address adequately. Colors may not exist, but Russell's attempt at proving this was not completely successful.

Related Links

  • On God: Why Being a Good Person is Sufficient
  • ManticMoo.COM All Articles Jeff's Articles Scholarly