When it comes to how we think and how we learn, there is really only one paradigm that applies. It is the combination of The Structure of Scientific Revolution by Kuhn, and the concept of a neural network, best described through our attempts to create AI, artificial intelligence, or learning computers.
I will not get into each in detail here. Consider this an intro type page to a set of chapters.
Kuhn's book is a seriously tedious read. The actual lesson it intends to impart is very basic. However, it is (or at least was) an attack on the scientific community, so he went through great lengths to give example after example after example in order to make his case. The problem, which his own theory expresses, was that people would challenge many of the individual findings, as if he had called Newton an imbecile for being wrong (Newton not being entirely correct was the issue), instead of seeing what he had expressed on the broad scale.
Basically, any accepted scientific theory will have a devout following. They will disregard anomalous data. Eventually, the amount of anomalous data becomes great, such that a new theory is created which abandons or incorporates the previous theory. Over a short period of time, it becomes accepted. I'd say look no further than "Climate Change" for an example you have most likely lived through. Now widely accepted, early theories on the topic were disregarded out of hand, even laughed at.
WARNING - Anyone that sites Kuhn (including myself) and The Structure of Scientific Revolution has not made their case simply by doing so. Truly, it only stands for the proposition that anomalous data will be disregarded. It does not mean an new theory which incorporates the data is correct, only that the theory MAY be correct, the next step in the evolution of understanding within a certain topic. So, when someone argues Ancient Aliens, it is their actual argument that must be the basis of their theory, not simply citing Kuhn and pointing to structures that can not be built even with today's technology.
A computer neural network would take me too long to explain. Basically, multiple connections at several different levels link an input to an output. As the computer is "taught" to recognize data for desired outputs (a bomb/mine vs. rock program fits best, sonar feedback being interpreted as metal or rock), the various connections are "weighted," some given more strength than others, until the machine can tell if an object is a rock or metal without further adjustments.
This is exactly like the brain, with synapses taking the place of connections, only on a much grander than imaginable scale. Each synaptic firing strengthens the synapse, and just as a neural network connection is "weighted" for an output, the synapses become "weighted" through the trial and error of the human.
When combined, virtually every aspect of thought can be expressed, which is why I consider their combination the ultimate paradigm of thought.
Consider depth perception. The brain first only has inputs it does not comprehend. Moving pictures the baby perceives as reality. Over time, the arms waved in front of the eyes, hitting objects, or the recognition of size to proximity of a mother or her breast, creates a new theory.
More apt would be object permanence, given the child is easier to comprehend at that time. Before becoming aware of object permanence (before the "revolution" of the child's theory of the world), an item disappears when out of view. Visual stimuli, data, of an item suddenly not apparent is believed gone. It is not even questioned. Just like Kuhn's paradigm, however, the anomalous data is eventually confronted by the brain and a new theory of the world evolves, object permanence.
A conceptual take would be the original application of Sabermetrics in baseball compared to the historical reliance on scouting and the disregard for certain statistics. Watch Moneyball for the gist.
The ultimate realization is that everything you know and believe fits into this category, not merely cognitive psychological issues and "theories" about specific topics. Everything, every argument you make, every opinion you have, to some extent or another will disregard "anomalous data" that does not fit within your own ideas (yes, presuming any "anomalous data" has actually been presented, because you may be entirely correct given the limited amount you know, or the scope of the concept is so small, like 2 + 2 = 4, to have further data) until you consciously confront the anomalous data, perhaps being taught something new or being convinced of something by someone else. Then, you develop a new way to look at topic, develop a new opinion, create a new theory.
Basically, though I know I have not flushed it out well, two certainties exist.
First, you can always gather more data and attempt to identify things which you have, until that point, disregarded. There may be data you perceive as anomalous and disregard (which makes it hard to identify, doesn't it?). This can be accomplished by always trying to identify your own presumptions and questioning them with regards to any other information presented. Of note, the very nature of synaptic connections will reinforce the manner one questions their own thoughts, making the adage "question everything," taught to me by Laura Nader, a path of wisdom.
Second, everything you think you know is wrong, at least partially, at some level, so long as anomalous data exists. ["I'm looking at you, physicists that claim dark matter means they completely understand the universe through math!"]