AbstractThe interpretation of factors constitutes a major problem in factor analysis. Until now the underlying semantic aspects of the question have been studied only occasionally. However, there are some indications that the mechanical interpretation of the results may constitute in the immediate future a real need in the realm of very large problems. On the other hand, the semantic study of multivariate statistical methods is the only way to warrant the validity of the mechanization procedures and the means to identify their limits.