The brain is amazing. Seemingly complicated, but amazingly simple in its function: input, recognition; test, accept, reject; and finally... combine and create.
Above is an fMRI scan of the axon paths through the human brain. Axons are neurone connectors: to reiterate earlier explanations, every neurone has many receptors (dendrites), but only one output (axon) to carry the neurone charge to another neurone... if it fires. Whether it fires or not (and to what degree) is the result of the sum of input it receives from other neurones.
The network of neurones behind a single memory is a complicated thing... not only are there the neurones forming the memory itself, but its connection to 'inhibitor' and 'accelerator' neurones that will affect their relation to other neurones, namely in the later 'conclusion' part of the thinking process. To make things simpler, I've eliminated all those secondary (yet capital) connections in my illustrations, to show just the process result.
The above is the simplest of memories: empirical memory. First off, your brain will decide for you whether an event is even worth remembering, and if it is, it will be linked with the emotion that made it seem 'important', a link that will be traded off to the neurone used to permanently store the memory (if the brain deems it important enough). These emotion connections can be trained by further 'matching' input from empirical... or 'trusted' sources.
These 'test types' are capital to the human brain. Even before we became rational creatures, we were mimicking ones: whereas 'lower' creatures had to rely on empirical testing for memory 'validation', our ancestors learned to copy the behaviour of 'similars' who had already tried and tested the concerned techniques and circumstances. Unless trained, the brain tends to ignore (and even reject) any information not coming from anyone in the 'similar', or 'trusted' category.
Yet we only 'recognise' or 'match' things already in our memory. If we've had experience with a red ball, and the brain deems it important enough, it will be one of the things our (supposed) subconscious 'watches out for' in our environment. If it is detected, the 'recognition' trigger will be accompanied by the memory's associated emotion (playing with the ball, for example, as opposed to being hit on the head with it) will be triggered too. If at first the emotional 'importance' of each memory may seem quite stark and distinct, this will become 'muted' with repetition ('lesser importance') and other associated/more important events.
Now enter critical thinking, the 'extra level' that makes us human. It's that region of constantly 'slow firing' neurone region of the brain that, when activated, will 'test' existing memories against each other to create a third 'possibility' that can become a 'memory' of its own: if a) a rock is heavy and b) a muddy slope is slippery, therefore c) a heavy rock on a muddy slope will slide. Untested, the new idea will generate but a vague 'warning, watch out for this' emotion attachment, but if someone having that thought later sees that rock slide (and avoids it to survive), that thought becomes an empirically confirmed fact that can be related to others (with that 'warning' emotion) until they can empirically test it themselves.
Yet consider the experiment: if a) All Terriers are dogs, and b) all dogs are animals, then c) all Terriers are animals. This is an exercise concerning only the mind (and communication with other minds) and categorisation - it cannot be tested empirically, and can only be 'confirmed' by the emotional reaction (understanding, approval or not) of the person receiving the idea. Yet it uses the same survival-tool technique as the first example.
I mentioned earlier that emotions can be trained. In the empirical world, this is a clear-cut affair (with, say, an initial fear reaction to fire dulled by an education in its uses, and experience with them), but when it comes to ideas, it all depends on where the ideas come from. Without critical thinking, the origin of an idea is just as, if not more, important than the idea itself. If a trusted source says that circumstance a A will result in A, and you repeat their lesson and get a positive response, then the same tells you that circumstance A will result in B, you will attribute that same emotional 'reward' to both lessons, even if they are contradictory.
Yet to the critical thinker mulling over circumstance A, they may weigh it against other circumstance/criteria/memories to conclude that, in fact, circumstance B because of A doesn't make any sense. The very act of this consideration removes the 'confirmation' emotion dependance on the teacher, and if the result of the reflection is tested empirically to a conclusive result, it may dampen, negate or even convert to 'warning' all former emotional connections to the source of the information. But the concrete conclusion of this exercise is that the emotional reward becomes a personal one.
It seems that once one experiences this personal conclusion/reward for the first time, it seems to 'validate' for the brain the utility of the critical thinking technique as a whole. I guess this is the 'switch' I was trying to locate in my earlier posts on this subject.