The necessity to expand traditional researches on intelligence is discussed. The dual model of pragmatics and mechanics of intelligence is failing mainly because of a low ecological validity of traditional measurements of intelligence. Therefore, we are proposing to consider the reflexics of intelligence, a mental process enabling subjective representation of cognitive activity in the experiencing of an individual. Further, the structure of implicit theories of intelligence was studied. Another subject of discussion is the analysis of intelligence through perceived attributes and constructs similar to them (creativity or wisdom). The author presents his own research findings. In the conclusion, several generalizations about the advantages and limits of implicit theories are given.