Volume 7, Issue 1/2/3, Octubre 1992
Toma A and Toma B
Una aproximación a las bases neurales deI pensamiento lógico
In this article we describe the logical performances displayed by a context-dependent associative memory model. This model requires the existence of a network able to construct the Kroneker product of two vectors, and then to send the composed vector to a correlation distributed memory. This system of nets is capable to sustain all the operations of the classical propositional calculus. This fact implies the existence of vector logics where the logical functions are displayed by matrix operators constructed using the properties of the Kronecker product. When the basic binary matrix operators act on fuzzy inputs, a probabilistic many-valued logic emerges. The present approach implies the potential existence of alternative vector logics. We describe a vector Shefferian “Iogic”, and we comment the potentialities of multidimensional vector logics.