It is commonly claimed that natural law jurisprudence wishes, in Hart's words, "to assert mysteriously that there is some fused identity between law as it is and as it ought to be." 1 This claim, which goes back to Bentham and Austin, is, however, simply confused. Natural law jurisprudence (such as that found in Blackstone, whom Bentham and Austin are criticizing) does not traditionally make such a distinction, because such a distinction is too crude. Neither 'law as it is' nor 'law as it ought to be' are precise phrases. In the mouth of a positivist, the sense of 'law' in the phrase 'law as it is' that has to be considered is what in the natural law tradition would be called positive law or human law. But it is clear that natural law jurisprudence does not posit a "fused identity" between positive law as it is and positive law as it ought to be; for one thing, 'positive law as it ought to be' can mean several different things. As an example: ideally, the medievals thought, positive law would outlaw prostitution; this is one sense of 'positive law as it ought to be'. However, they insisted that this was not possible; that if you did actually outlaw prostitution, you would get something worse (the destabilization of society). Despite the fact that it ought not to exist, as a matter of fact prostitution on the medieval view as a lesser evil that just happened to exclude a greater evil by (as it were) channeling violations of the marriage bond into a very narrow channel whose adverse effects could easily be limited by law; if you took away this lesser evil, violations of the marriage bond would not be channeled into this narrow set of limits where law could keep a close eye on it. In one sense, 'law as it ought to be' excludes prostitution; in another sense, 'law as it ought to be' tolerates it with approving of it. In the first sense there is no "fused identity" between positive law as it is and positive law as it ought to be; this is explicitly denied. In the second sense, however, there is no "fused identity," either, because it is never denied that positive law as it is might deviate from positive law as it ought to be.
The trap into which legal positivists continually fall is, from the natural law point of view, thinking that 'law' means exactly the same thing whenever it is used. Natural law jurisprudence denies that; it insists (1) that 'law' means several related but different things; (2) that the difference among these senses is in some cases so great as to result in equivocation; and (3) that these different senses can be ranked in the sense that some things are more properly called 'law' than other things [added later: A fourth needs to be added here, since on some views it is what is really at issue: (4) the primary criterion for this ranking is moral, i.e., moral laws are more properly laws than immoral ones; although most positivist arguments don't really do much against this principle, it is often in the background as the primary disagreement. But here I'm merely considering the "fused identity" claim, which is false because of (3).]. Thus when we are talking about positive law, 'law' may be used for just and unjust laws. If we have an unjust law, natural law jurisprudence denies that there is a "fused identity" between law as it is and law as it ought to be, just as much as any positivist. The difference is that natural law theorists hold that in this sort of case the distinction between law as it is and law as it ought is a sharper distinction than the legal positivist thinks it is. On legal positivism, the word 'law' in 'law as it is' and 'law as it ought to be' always has exactly the same sense; on natural law theory, in all cases where positive law as it is and natural law (one of several possible senses of the phrase 'law as it ought to be') conflict, the word 'law' is used equivocally in the phrases 'law as it is' and 'law as it ought to be' - there is a slightly different sense. Both senses are legitimate senses of the word 'law', but in one case we are using 'law' in a more extended sense than in the other. Another medieval analogy: How many human beings are in a morgue? In one way one can easily count human corpses as human being. It is also clear that in another way we could deny that the human corpses are human beings and restrict 'human being' to the living. The word 'human being' is used equivocally when applied to corpses or to the living; sometimes not recognizing this distinction doesn't make any difference, but sometimes it does. Roughly, one may say that in natural law jurisprudence, unjust laws are the corpses of law. In one sense they are laws, but in another, more fundamental, sense they are not. Sometimes not making this distinction doesn't matter; but sometimes it genuinely does. From the natural law perspective, legal positivism is guilty of failing to recognize this distinction in senses; it tries to act as if there were only one sense of 'law', as if things weren't able to be called 'law' in stricter and looser senses. What bothers the natural law theorist is not that legal positivists make a distinction between law and morality; it is that they make the distinction in an overly simplistic way.
Hart treats the attempt to reject this over-simplistic distinction as "an invitation to revise our conception of what a legal rule is." 2 This, however, is exactly wrong. Natural law jurisprudence precedes positivism in time; it became far more pervasive in society than positivism ever has; its residual traces are still found throughout our legal system; legal positivism grew out of a criticism of this. It was Bentham and Austin who attacked Blackstone, not vice versa; it is the legal positivist who is inviting the revision of our conception of law, a revision that has not completely been effected. The natural law theorist is not offering an invitation to revise our conception of law; the natural law theorist is refusing the legal positivist's invitation to revise our conception of law, on the basis that there are perfectly good reasons not to oversimplify it in the way legal positivists generally seem to do. [One can argue that not all do, i.e., that it is not essential to legal positivism; but this is a different issue. What I am pointing out here is that from the perspective of natural law jurisprudence most positivist arguments against natural law jurisprudence appear to commit them to some such oversimiplification.] 'Law' and 'legal rule' are words that indicate complicated interrelations of different senses, and, from the natural law perspective, the legal positivist is committed to severing these interrelations and saying that one, and only one, of these senses is law (and, what is more, from the natural law perspective the legal positivist chooses the sense least capable of standing alone as their stand-alone sense).
 H. L. A. Hart, "Positivism and the Separation of Law and Morals," in Law and Morality, ed. by David Dyzenhaus and Arthur Ripstein, University of Toronto Press (Toronto: 1996) p. 41.