What is a minimum-phase network?

THE terms “minimum-phase” and “nonminimum-phase” as applied to network functions were probably first introduced and defined by Bode.1 Since their introduction these terms have become almost household words in circuit theory circles. There seems to be unanimity of opinion as to the definitions of minimum and nonminimum-phase networks, but there is less than unanimity in understanding of the properties of these two classes of networks. What is it that is minimum about a minimum-phase network? Is it the algebraic value or the absolute value of the phase angle that is a minimum? Is it the total change in phase angle as frequency varies from zero to infinity that is a minimum? Should the angle considered be angle of lead or angle of lag?