My hypothesis of reality

I've started working on my hypothesis on how people grasp reality. It is a part of my life-time goals of understanding the human mind, and creating an artificial being.
It goes as follows:
We do not live inside the reality. We live in a model of reality that exists in our mind and is built by our senses and perceptions. When something is "caught" in one of the senses the mind compares the input to the model and analyzes the input by it. What we hear, see, feel and think are actually an analysis of this input from our surroundings by this model. The model is not static and changes with time, especially when we learn or comprehend something new. Fundamental elements of the model (e.g. water runs from high places to lower ones) changes only rarely normally, and requires a source that is considered reliable enough by the model itself in-order to change. The less fundamental the element, the easier it is to change it, annex it, or remove it from the model and vise-versa. The model is frequently integral in the aspect of it rarely including contradictions and/or self contradictions and these mostly happen while the model is undergoing any sort of a big change (e.g. religious conversion). There is a constant effort to minimize and eliminate this contradictions as soon as possible as the mental health of the person at this situation is usually fast declining and the symptoms are often deep confusion, depression and frustration. The more this contradictions are fundamentals and exists more time they increase the odds of seeking external aid and guidance. This guidance often causes even much more extreme changes and results in a much more stable model (or seldom a mental breakdown) that settles the contradictions. After the implantation of such a change the new improved model is much more fixed in the elements caused the initial contradiction, presumably to prevent repetition of this events. In order for new elements to be added to the model, they are compared to the existing model for analysis. If there is a need to change or decline existing elements to accept the new ones a full scale analysis of possible outcomes of accepting and rejecting the change is being carried out, and the decision will be carried to support elements that fits better in the general model. All in all, new elements are usually accepted only when they settle with the existing general model and mostly explain it better and in a more extensive way.

So this is it so far...
It still needs much more work, and will be constantly improve. Tell me what you think about it!