Cognitive learning is a concept of heated debate in the past. Followers of behaviorism, i.e. behaviorists, stressed that psychology is a scientific study and therefore may only objectively investigate observable phenomena. Cognitive learning, they say, cannot be observed and since researchers must made inferences, objectiveness is compromised.
Nevertheless, evidence supporting the existence of this phenomenon has been found. In addition, these phenomena cannot be explained with neither classical conditioning or operant learning or any other branches of behaviorism. Cognitive learning must be used. How could researchers argue for the existence of the concept? They devised scientific experiments. In one of these, a rat is put into a maze. It is free to wander around and explore the maze without any reinforcements of any kind from the researcher. Afterwards, it is put on one end and food on the other. The rat negotiates its way through the maze for efficiently than one who is unaccustomed to it. Psychologists coined the term cognitive map to describe this phenomenon. The rat has created a mental model of the maze during its exploration. This cannot be explained purely in terms of classical conditioning and operant learning.
There are two other forms of cognitive learning. One is called insight learning. It occurs when the subject suddenly re-arranges its perception of the problem presented and arrives at the answer. In one famous study, a chimpanzee is put into a room with boxes and some bananas hanged from the ceiling. After some failed trials, the chimpanzee decides to put the boxes on top of one another then climb on top of them, deservedly earning its reward. The other form is called observational learning. Animals observe others then learn from them. In some cases, observing animals are indirectly reinforced just by seeing the one in action rewarded somehow. It is favorable because the process of what the best response is for a certain stimulus can be averted. Unfortunately, it is also the factor being the spread of unwanted behavior such as excessive violence.
Personally, I believe that behaviorism and cognitive learning are by no means mutually exclusive. They are simply different forms of learning. They should be united somehow in the future when neural processes, fundamental to any form of learning, are better understood. As a student, I think I use a combination of behavioral and cognitive learning. Of course, the instructor can reinforce students with praises or good marks. But the bulk of learning comes from cognitive processes because we think (a lot in some classes such as AP Calculus) rather than simply doing exactly as told.