Constraint in information theory is the degree of statistical dependence between or among variables.
Garner[1] provides a thorough discussion of various forms of constraint (internal constraint, external constraint, total constraint) with application to pattern recognition and psychology.
