Congress has finally passed the much-awaited America Invents Act, which the President promises to sign. Patent reform is finally here.

Excuse me while I yawn.

If this was our nation's big push to make patents more predictable and litigation less expensive, then I'm afraid we've swung and missed, or at least with respect to the area of patent law that needs reform the most—claim construction.

I studied philosophy in college. I am familiar with the scholastics, and I used to know—it's been a while—how many angels can dance on the head of a pin. But I still don't understand claim construction—not enough, that is, to assure clients—and neither do many principled and conscientious judges.

Case in point: On July 8, 2011, two Federal Circuit judges construed the term “body” in a medical device patent to exclude a body made of two-pieces. After all, the patent described “the invention” as having a one-piece body, criticized numerous prior art devices for having two-piece bodies, and every example and drawing in the patent showed the invention with a one-piece body.

But another judge dissented, arguing that the patent “clearly” included devices having two-piece bodies. And he was reading the same patent.

How can this still be happening a decade after the Federal Circuit promised to clean up the rules of claim construction?

I think the problem is inherent in the rules. On the one hand, we're taught that patent claims—the meets and bounds of the property right—should not be limited to the examples disclosed in the written description of the invention. On the other hand, we're taught that claims should never be construed more broadly than the written description of the invention.

Got that?

Me neither, and I've been litigating patents for two decades.

We need to return to first principles. One is that the judiciary's job is not to guess about the truth. I don't mean the “truth” as in who is lying and who is not. I mean the truth as in what the patentee really invented when the patent itself doesn't clearly say. Instead, in American jurisprudence, the judiciary deals with ambiguity by assigning burdens of proof—that is to say, we decide who wins and who loses when the truth can't be known.

A criminal defendant is not acquitted because a jury determines he didn't do it. He's acquitted because the government didn't carry its burden of proof. Why do we assign this burden to the government? Because we live in a liberal democracy, and a liberal democracy makes a policy choice: It is better that many guilty go free than a single innocent man be deprived his liberty. We could easily assign this burden the other way, like the world's tyrannies, which would solve the problem of the guilty walking. But that's not an efficient way to achieve our goals as a liberal democracy.

The same can be said for patents and claim construction. Courts should not be in the business of divining the truth about what an inventor really invented when a patent's claims are susceptible to two or more meanings. Instead, we need a burden of proof, or at least a “burden of clarity.” If a claim is ambiguous, someone should bear the social cost. The only question is, “Who?”

Law and economics say that for any social cost, the law should identify the party who is best able to reduce or eliminate it, and assign that party the burden of doing so. Car manufacturers are better able than drivers to make sure the brakes work, so the law assigns Ford, the manufacturer—not Fred, the consumer—the burden of doing so.

The same should go for claim construction. As between the patentee and the public, the patentee is in a far better position to make sure his invention is clearly described and claimed. If he doesn't, he or his lawyer should bear the social cost. It makes no sense to strap the public or the courts with the burden of divining what the patentee meant when the patentee was free, ex parte, without meaningful time constraints, to do so.

And yet that's our current system. Actually, our current system is far worse. A widely criticized rule of claim construction says that a patentee does not limit the scope of even ambiguous claims unless he does so clearly and unmistakably. And so the social cost of this ambiguity is borne by the public—the party least able to reduce it. That's just wrong. It's inefficient. It's Alice and the looking glass—everything is backwards.

And it's not likely to change anytime soon, even after patent reform.