The problems of perception are often overlooked, perhaps because the other problems of the mind, for example consciousness and intentionality seem more pressing. Usually we just assume perception works, and leave it at that. We are all realists, we suppose (the position that an external world exists independently of our perceptions), so what more is there to think about? Here is one question then to motivate some interest in perceptions: are we direct realists or indirect realists?
Here I will define perception as the channel by which information enters the conscious mind. In the indirect theory the objects of our perception (where the information is most immediately coming from) are sense data. Sense data is generated by an unconscious process, depending on both the external world and on unconscious states (this is how hallucinations are explained under this theory). On the other hand the direct realist argues that the objects of our perception are the real objects, and sense data is not part of this picture. Below is a diagram to illustrate this difference.
From looking at the diagram it might seem that direct realism makes more sense. After all it is the simpler solution, and what evidence could we possibly have for the existence of sense data anyways? However an indirect realist would argue that we should reject direct realism because there are serious explanatory problems with it, which only indirect realism can overcome.
Specifically the argument from illusion is often considered a good reason to abandon direct realism. The argument from illusion runs as follows: the direct realist argues that the objects of our perceptions are the real contents of the world, however it is a well known fact that our perceptions can be inaccurate or completely illusory, for example a straight stick may appear bent when placed in water, or we may suffer from hallucinations that have no basis in reality whatsoever. Since the object of these perceptions is clearly not part of the objectively real world direct realism contradicts our experience and should be discarded.
Of course other people have attempted to answer this objection before me, but here is my take on it. First let us deal with “physical” illusions, such as a stick appearing bent when in water or the color of an object appearing to be different under different lighting conditions. I would argue that a denial of these phenomena as real is to misuse the word real. The properties that are being perceived here are not simply the object but the object and properties of its surroundings. For example when I perceive a stick that looks bent in water I am not perceiving an illusion but a real fact about the optic properties of water combined with real information about the stick. Likewise when an object looks different colored under different lighting conditions I am perceiving the real fact that the light reflected from an object depends both on the surface conditions of the object and the light striking it.
The serious problem for direct realism is the case of hallucinations, because when we have a hallucination we can’t say that somehow the hallucination had a basis in the objectively real properties of the world. But does this make the hallucination unreal? I would argue that clearly there is some real component to the hallucination, as it has had an effect on you mental states, perhaps even causing you to say something like “that dancing midget sure is green”. Clearly then the hallucination is a real result of some activity in the unconscious. It seems foolish though to deny the existence of the unconscious mind. Thus your perception of the hallucination was about something real after all, a real activity of your unconscious mind. This line of argument can be also used to defend direct realism against similar attacks, for example when you see a color surrounded by a darker color it seems lighter than that same color surrounded by a light color. Once again the difference you perceive is a real property of how you process vision.
In brief then my defense of direct realism runs like this: the objects of perception are always real, however some of them are real features of the mind (which you may not be otherwise conscious of). Some may object to this, claiming that I simply have re-defined what real is. In a way this is true, and if you wish to argue that the only things which are real are those that can be observed by multiple people (objectively real) then I would have to say that the objects of perception do not all fit this definition of real. However, such a definition also denies that the experience of being conscious is real, and given this I would argue that it is a poor definition. Thus I would argue that the definition of what is real proposed here, namely something that can have a casual effect on the world, is preferable, and given this definition direct realism seems to be an acceptable way of describing perception.