Here is a link to a review by Scordato on Tamanaha's book on legal realism and formalism (which I actually read early in 2011 and enjoyed very much).
I liked T's book, but as hard as it was on the 20th century legal realists, I don't think he was hard enough. Their version of legal history seems not just wrong, but bordering on an ideologically inspired, deliberate falsification of the past. On second thought, strike the "bordering on". But T still deserves a lot of credit for starting the work of undoing the damage that the legal realists' false characterization of the legal formalists, if that is even the right term for classical jurisprudence, has done. But I do have a problem with T's account which still seems to be saying something like, the formalists were a lot more reasonable, that is realist, than they get credit for. He does not seem to have in mind a version of formalism that a reasonable person might subscribe to, but assumes some form of realism is what all reasonable jurists must believe deep down. So his book has something of the flavor of a liberal telling his friends that his conservative friend is really more reasonable, i.e. liberal, that he seems. Not a defense of conservatism.
Scordato makes the good point that formalism in the form of, I'm just here to enforce the law, and interpret it when necessary, not to create law or policy, is the official line of everyone who is nominated to the bench and every legislator who opines on this. How could this be, if we are all reasonable realists now? Good question, and the answer is, because we're not. The problem is actually worse than this. Not just judges appearing before the Senate, but the vast majority of sitting judges and serving legislators actually believe in common sense formalism, and a good thing too. It is only in law schools, I would claim, that some form of realism is nearly universal. I think this gives academics a very distorted view of the supposed consensus on realism.
I personally have always found realism an incoherent philosophy of law as I'm sure some jurisprudes do. S is absolutely right that realism does not explain the hierarchical structure of the judicial system, which sure seems to presuppose that there are legal rules that errors can be made about and subsequently corrected by better reasoning or interpretation. But it goes far beyond that. The whole idea of reconciling legal authorities that are in tension with one another, a process from which Western jurisprudence might be said to have grown, presupposes that logical coherence and consistency, those features of rules, play an essential role in any processes that can be thought of as legal. The whole idea of a legal rule that controls what officials can do is inseparable from the idea of a legal system, which is a system of rules, be they crisp or squishy. Legal realism of the sort we were exposed to at Yale back in the 80s, which was left over from the 1930s I suppose, seemed little more than disparaging the flimsiest of straw men, a mechanical jurisprudence that never existed -- Tanahama shows that -- but also mischaracterizes badly what a legal system that takes rules, and lots of 'em, seriously, would actually look like.
Of course every legal system, if it is a legal system, is to a degree mechanical and if it's not just a mathematical exercise, to a degree not. A twenty year old pickup, that wiggles and squeaks a great deal, is still a mechanical system. That it does a great deal it can do in part because a lot of it is not rigid, hardly makes it less mechanical. Part of how it works may not be explicable in terms of standard mechanics. You might need chaos theory to predict exactly when it will start or which way it will hop on a bump. But if you want to understand the thing, you are going to need to appreciate its mechanics at a minimum.
I'm sure I'm not the first to notice that legal realism seems to have been mostly an attack on legal rule-following by those who found those rules inconvenient for their political projects (Progress! Science!), like the man who discovers that marriage interferes with his self-realization and so is actually morally non-binding shortly before he hopes to commit adultery, as prescribed by some embarassing book about personal relationships circa 1974 and inspired by the availability of a hot receptionist. Now that's realism. One wonders why such special pleading should be taken seriously at all, but it has to be, if only because it has done so much damage to the project of the rule of law.