Four out of five U.S. appellate court decisions are literally unspeakable. What this means is that the vast majority of opinions by federal appeals courts are deemed "unpublished" and cannot, for the most part, be cited in later cases.  They are the law for one case only, which may mean they hardly deserve to be called law at all.  

When a judicial opinion is "unpublished," it doesn't mean it's unavailable or even hard to find. Nearly all federal appeals court decisions are posted on databases maintained by West Publishing or Lexis-Nexis, both private companies. But only certain decisions are published in the Federal Reporter, also published by West, but available in hard copy in law libraries. Where and how cases are reported is important not just because not everyone has access to computer databases, which can be quite expensive, but because most of the circuit courts have rules against citing opinions that are not in the actual books. 

The Second Circuit Court of Appeals, which sits in New York, for instance, says of its unpublished decisions that they "may not be cited as precedential authority to this or any other court."  The unpublished decisions may, however, be cited in subsequent stages of the same case: the decision is good for one case only, but does not govern generally. 

Unpublished decisions are not new, but the problem is getting worse. The Administrative Office of the U.S. Courts first started counting them in 1989, and by that year 64.7% of decisions were unpublished. By 2002, the percent unpublished had grown to 80.5%. Since that time, the number of federal appeals determined on the merits has risen, but the tide of appeals has ebbed since 1994, but the percentage of opinions that are unpublished has continued to climb. 

The percentage of unpublished opinions varies dramatically by circuit. The First Circuit, which sits in Boston, publishes 55% of its decisions, while the Fourth Circuit, in Richmond, publishes a mere 8%. But with less than one in five opinions being published, the judges can pick and choose which cases will be relevant in the future and which will be one-offs for the parties only.   

What's wrong with a decision being unpublished—that is, uncitable—anyway? Judges are not Roman emperors giving the thumb up or thumb down. They don't just to decide who wins and who loses— their job is to declare law.  That law is precedent, and it must be applied in subsequent cases.  Professor Joseph Story, who would later serve on the Supreme Court, put it in 1833: "The case is not alone considered as decided and settled; but the principles of the decision are held, as precedents and authority, to bind future cases of the same nature. This is the constant practice under our whole system of jurisprudence." 

The "principles of the decision" are not just to guide the people or future litigants; they are an important check on the judges themselves.  Indeed, in the federal system where judges have life tenure, they are the only check (apart from the few cases heard by the Supreme Court).  Judges must explain themselves, and be bound by that explanation.   

If some decisions—indeed most decisions-- are deemed non-precedents, there is nothing to prevent a judge from deciding according to favor, prejudice or whim. What he says in one unpublished case won't haunt him in the next. For these reasons, Judge Richard S. Arnold, in an opinion for the Eighth Circuit, declared its rule that unpublished opinions are non-precedential to be not just wrong but unconstitutional, though Arnold's opinion itself was later vacated.  

Judges who defend the current system say that being an appeals court judge is tremendously hard work already, and that if judges had to draft full-blown opinions in every case, it would be impossible. Alex Kozinski, a highly respected judge on the Ninth Circuit, compares the writing of a full opinion to writing a law review article every two-and-a-half weeks. Unpublished opinions can be knocked off in a matter of hours, he says.  Judges must work hard to insure that the law of a circuit is clear and consistent. Thus they work hard on the cases they have decided will count as precedent, and much less on the cases that won't count to anyone but the parties. Unpublished opinions don't explain much, so why bother publishing them?  

The solution to the problem (apart from authorizing more judgeships) may be in Judge Kozinski's description of it. Law review-type decisions are not as necessary as brief statements of the case and the rule of decision. But the real problem with unpublished opinions is not that they have fewer words, or cite fewer cases.  It is that there is no check to insure they follow the law established by Congress, the Supreme Court or the circuit court. The judge also says no one has proven that unpublished opinions do not follow the law.  Such proof is, of course, unobtainable: Providing it would mean re-deciding the unpublished cases.  

But the issue isn't whether unpublished opinions are usually or even often wrong. The issue is that they are not scrutinized by other judges, or even other lawyers. The Supreme Court is also far less likely to review unpublished decisions.  

The solution is not to publish more, but to remove the prohibitions against citation. If the briefer decisions say less, they will naturally be cited less.  But if every case counts, there will be reason to make every case right. Lawyers have long been school on Chief Justice John Marshall's 1803 dictum that "It is emphatically the province and duty of the judicial department to say what the law is." Two centuries later, judges say what the result is, and leave it at that.