It's question natural enough that it has been asked by many a philosopher. Descartes always seems to be the first who comes to mind, since his whole method was based on doubting everything until he could determine which things were most certain.
In the modern world, you don't have to be in a Cartesian state of doubt to wonder. The standard example is that when you look around, you see many solid objects, but in fact physics tells us that this is mostly empty space. Or if that is too abstract, there are plenty of common illusions of a more mundane sort. Are my eyes telling me lies? It's worth probing the question a bit.
The only illusion, in fact, is the idea that there are direct experiences of things, as opposed to interpretations of them. This illusion is caused by the fact that most of our interpretations of the objects around us are spontaneous. I do not consciously invent the shape and color of the computer screen in front of me, it just happens. To imagine my computer as really a collection of atoms and empty space, or even to imagine it in terms of its constituent electronic parts, requires conscious effort based on training and contemplation.
Yet the difference is in degree, not kind. Whether I am casually observing something or rigorously studying it, my only access to a thing is to imagine it. Indeed, this applies to all of external reality, to such an extent that it is possible to understand the world as not being composed of "things" at all. There are some religions and philosophies which posit a universe which is really all one, and the separateness of the things around us is just an illusion--it is all really one thing.
Imagination is perhaps most naturally associated with fiction, but even in everyday language, we use it more generally, as when someone might say, "I imagine it is difficult to move to a new a country," when that person has never done so. That person means to state a belief, not to invent a world. So it is with all of our beliefs: they are expressions of how we imagine the world.
To clarify, the difference between imagination and merely pretending is that in the latter case we make no attempt to test or change what it is we imagine, but we leave it entirely up to our choice. The kind of imagination which leads to a greater understanding of the world might be called "theorizing" or "hypothesizing." There is a kind of naïve view of the sciences which imagines (indeed!) that a theory is basically reducible to a sentence with causal structure, as in, "If X, then Y." But most scientific theories are really based on what one might call "models," and these are really ways of imagining the world that one can test. For instance, Newton's laws of motion are far from being mere causal statements. They can be best expressed by imagining the world as a three-dimensional Euclidean space measured by spatial coordinates, and objects as points in that space whose coordinates change in time and are governed by differential equations. If you want to know what really makes Einsteinian relativity different from Newtonian mechanics, you need to start by imagining the universe in four dimensions rather than three and go from there. Certainly, what proves that Einstein was more correct than Newton is that certain testable predictions are made, and Newton is shown to be lacking where Einstein is not. But you will not truly have a theory if all you have is a set of prediction statements.
So even our most sophisticated theories about the universe are no different in kind, but only in degree, from the spontaneous images our minds produce when casually observing it. I look at the coffee mug on the table in front of me, and I spontaneously see its colors, a cylindrical shape and a little handle protruding from it. With a little more conscious imagination, I might imagine Cartesian coordinates enveloping it, so that I can more precisely analyze its geometric relation to other objects on the table. Or with a little more imagination, I imagine it as little atoms quickly vibrating against one another, and then even the air around it as atoms moving much more freely. Everything seems to change as I move from one mode to another, yet it is the same reality. There is no reason to say that, after this mental exercise, I have understand my coffee mug "better" than before. And that is because the purpose of my coffee mug is to hold my coffee, which anyone is perfectly capable of seeing even with only a limited scientific imagination, or none at all.
Why, then, should I want to imagine things in different ways? That is a good question, and I don't think there is one single response that applies to all cases. It depends on our intentions, our desires, on our relationship with the world around us. If there were never anything puzzling or difficult or otherwise problematic about the world, I imagine we would be perfectly content with the way we spontaneously envision it--a coffee mug would be a coffee mug, and that's that. And yet there wouldn't even exist such things as coffee mugs, because such an invention comes only in the face of solving a problem--namely that coffee is a hot liquid which I want to drink! Even if something had the size and shape of one, we would not call it a coffee mug nor even have any other sort of name for it were it not for our intention to drink coffee.
An illusion, then, is a way of imagining the world which damages, rather than helps, our relationship to it. This is in relation to a certain context. For instance, if you look at the Müller-Lyer illusion, you will see two lines which appear to have different length, even though in fact they have the same. The reason you are fooled is that your visual senses are calibrated to see such things in a three-dimensional world, and transported into a two-dimensional context you are fooled. I should think you would be rather thankful that your brain is incapable of "unseeing" this illusion, because your ability to navigate the three-dimensional world may very well depend on it!
So the pursuit of truth and the rejection of illusion cannot be devoid of value judgments. How you imagine the world depends greatly on how you desire to live in it. If you seek to master it and control it, you will continually be testing your theories to try to reimagine the world more accurately. If you seek to understand reality as a whole, you will be continually inventing or perfecting robust abstract concepts which capture the essence of everything.
If you seek to appreciate and enjoy the world, perhaps you will not be concerned with theories, but you certainly will be engaged in imagination. You will seek to refine your ability to notice colors, smells, tastes, sounds, and their relationship to one another. This is yet another way of imagining the world, one which often takes the form not only of observation but response, in the form of literature, art, music, and other sorts of performance. All of these are forms of knowledge, as sure as any science. All knowledge is simply a refinement of imagination.
Whether or not I can trust my senses or other faculties is ultimately the same question as whether or not I can live well in this world. Is the world sufficiently inviting to my presence that my attempts to understand it will be fruitful? It is not a question that can be answered a priori; it must be explored as with any other relationship, though this relationship might be decidedly one way. I don't want to go further down this trail of thought, but I mention it only to raise an objection to the notion, common today, that questions of truth are unrelated to questions of goodness.
Now, it seems to me, one thing you can never fully understand is your own mind. Why? Because it is precisely that mind which is doing the imagining, and to imagine itself is a hopelessly recursive task. Take an analogy to some physical object. When you try to understand physical things, you imagine them not only as they are but as they would be in hypothetical situations. So you imagine not only the baseball as a round object in space but also as something being hit by a baseball bat, and you visualize its reaction. You can even quantify such reactions with physical laws (differential equations). Now suppose you try to do the same thing with your own mind. You try to imagine it in such a way that you will be able to predict its behavior. How will you test this, when it is in fact that very mind which decides how to act? Could I have predicted that my mind would formulate this question, and then would seek to respond to it? Can I predict what its response will be? It is like a dog chasing its own tail--the effect is dizzying.
There are, of course, scientific theories of the brain which are rapidly developing. But I do not think any of these can constitute a complete theory of mind, and indeed I do not think there can be any sort of complete theory of mind whatsoever. It is the mind that has theories, the mind that imagines, and the mind that tests itself against the world. The mind cannot test itself against itself.
There is nothing tragic in this. When I read the state of modern day philosophy of mind, I sense an air of sadness that there might be one thing in this universe which we can never understand. Yet this notion is based on the strange separation of our minds from ourselves. To fully understand the mind is futile, because the mind is not in fact external to itself.
Instead of lamenting this futility, perhaps we need to remind ourselves about the point of trying to understand anything at all. Things are not simply there to be understood, but to be lived with. The goal is to increasingly perfect the relationship between our minds--or rather our whole selves--and the world we live in. We cannot step outside the world to understand it is a whole, and why would we want to? If we were removed from this world, there would be no point in understanding it.
It could be objected that there would be great value in understanding another mind, which is not my own, on the assumption that my own mind is much like that other mind. There is a point to this--it is a principle which bears much fruit in fields of study like psychology. But there is still an insurmountable problem, which is that I cannot trascend my own freedom. There can never be a complete theory of my own mind which gives me the ability to predict my mind's own actions, because in fact those actions are choices--my choices. To make this clearer, perhaps one should go read about Newcomb's paradox and then ask, "Am I a one-boxer or a two-boxer?" The problem is that your answer will not simply be a prediction, but rather a conclusion of your own thoughts--a choice.
So I think I have finally articulated an explanation, however faulty or incomplete, of why I think the whole project of comprehensively explaining the mind is flawed to begin with. Of course, words like "choice" and "imagination" are sure to raise controversy, but for now I simply have no other way to see the problem.