In his ongoing debate with Clark Glymour and other scientific realists, Bas van Fraassen (e.g. 1983 pp. 165–168, 1985 p. 247 pp. 280–281 pp. 294–295) has often appealed to a widely accepted doctrine that strength and security are conflicting virtues which must be traded off one against the other. In limiting his commitment to only the empirical adequacy of a theory van Fraassen claims to be simply more cautious than his realist opponents. Certainly, as he delights in pointing out, a theory T cannot be more probable than its empirical consequences E. For Probability is monotone with respect to entailment, so P(T) ≤ P(E) if T entails E. If security is measured by any function which, like probability, is monotone with entailment, then it would seem that the kind of trade-off between strength and security van Fraassen appeals to is unavoidable.¹