MESSENGERS, umpires, and judges share a similar experience: Nobody wants to learn the worst, so everybody hates the bearer of bad tidings. In my line of work, the announcement of misfortune most often comes during (or immediately after) a criminal case, when the judge, obedient to the Constitution, has to exclude vital evidence, obtained with unfortunate disregard for the Bill of Rights, and thus allow a blatant miscreant to walk free.
If a shortstop performs gymnastic marvels to cut off a sure single, but guns a wide throw to first, the umpire has to call the runner safe. And if a detective shrewdly tracks the suspect to his lair, but neglects to obtain a search warrant before ransacking the bedroom and finding the stolen jewels under the mattress, the judge has to keep the evidence from the jury.
Sometimes the miscue proceeds from ordinary, indeed extraordinary, stupidity or neglect. Rarely - at least in my experience - is the constitutional misbehavior a product of deliberate prosecutorial chicanery. Most often the problem arises from an unfortunate disinclination to recall the difference between a suspect and a convict; between popular certainty and legal proof; between indictment and guilt.
Once law enforcement people obtain information that persuades them that a given ''individual'' performed a particular act, they understandably assume that a jury will agree. Similarly, many citizens, learning that the police have arrested X for the murder of Y, conclude without more evidence that X is indeed guilty.
I do not suggest that either the prosecutors or the public regard guilt as an inescapable conclusion. Everyone understands that a trial - a fair trial - is a necessary element of justice. But people do not always remember that a trial, because it involves a testing, offers the possibility of alternative results.
Of course, in this country, it does not matter how conclusive the police or the public believes the evidence to be. It does not matter how strong the case is in fact: Every defendant is innocent until proved guilty.
No judge could argue with that: All of us regularly take it as the text of our jury sermons/charges. Whether the public truly accepts the principle is less certain.
A recent survey sponsored by the Hearst Corporation produced startling results. Of people who had actually served as jurors, 49.9 percent agreed that ''it is up to the person who is accused of the crime to prove his innocence.'' And 31 percent of college graduates said the same thing.
The simple truth is that the verdict in a criminal trial has nothing to say about a defendant's innocence. The jury decides only one thing: Has the government proved the defendant's guilt beyond a reasonable doubt? Indeed, the verdict the jury returns when the government has not done so refers, not to innocence or even to lack of guilt. A verdict of ''not guilty'' signifies only that the government's evidence - strong, weak, or middling - did not satisfy the jury to a moral certainty that the defendant did what the prosecution claims he did.
When we forget that simple truth, we begin to talk about the defendant's burden of proof. That concept leads in turn to a belief that accusation alone established at least a presumption of guilt. It also leads to the disquieting suggestion that any not-guilty verdict demonstrates, if not a failure of justice , then at least a miscreant blessed with luck and a good lawyer.
We have national difficulty distinguishing between the certainty that a crime has been committed and the uncertainty that its perpetrator was the particular person standing trial.
In Scotland, that land of good sense, they have a custom we might profitably import. There, the jury can choose from among three verdicts: the two we use, and ''not proven.'' With that option, a jury can clearly announce the difference between a serious accusation and legal proof, between a fault in the system and a deficit in the evidence.