The combination of the connectionist and symbolic approach of artificial intelligence has been an area of interest in developing more reliable and explainable intelligent systems, giving rise to the Neuro-Symbolic (NeSy) paradigm. Addressing these new trends in the development of artificial intelligence, this research project aims to implement a neural network capable of recognizing complex symbols in terms of simpler components, which will be denoted as “sub-algebraic terms”. The operation of these terms can be interpreted as logical operations between symbols, offering a novel approach to symbol recognition and reasoning. Using the Neocognitron architecture developed by Kunihiko Fukushima in combination with current NeSy frameworks and ideas, our study proposes a new view of how the current trends can be combined with the selective architecture of the Neocognitron and how this can clarify the symbolic nature of the NeSy ideas for improvements in classification and reasoning tasks using “sub-algebraic terms”.