"The test of a first-rate intelligence," F. Scott Fitzgerald once said, "is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function."

Fitzgerald, meet Waverly Yates. He's been functioning, for the past few years, as director of Bonabond, a local self-help group whose principal work is to help ex-convicts become useful, gainfully employed citizens.

And on the notion of criminal sentencing, he also holds the requisite number of opposed ideas: 1) There should be greater flexibility in sentencing -- including a range of alternatives to prison. 2) There should be greater uniformity in sentencing -- with less opportunity for judges to be swayed by their prejudices.

He reached his first conclusion at least in part as the result of a study his organization undertook a couple of years ago. What he found, he said, was an extraordinarily high and direct correlation between joblessness and crimes committed for monetary gain.

For instance, of 30 convicted armed robbers referred to Bonabond during a three-month period, 26 had been unemployed at the time of their arrest.Thirteen of 16 persons convicted of second-degree burglary had been unemployed; 14 of 14 petty thieves; 12 of 13 Mann Act conspirators; 15 of 17 prostitutes.

It may not be unfair to suggest that some of these convicts -- notably the pimps, the prostitutes and the armed robbers -- may well have been "professionals" and, therefore, unemployed only in the technical sense. But Yates's main point is this: The fact that so many of the economic criminals are out of work at the time of their offenses, and the fact that the incidence of these crimes rises and falls with the rise and fall of overall unemployment, suggests that the state of the economy -- and the economic state of the particular offender -- is a factor in criminality.

That being the case, he said, it seems reasonable that the inability of suspects to find gainful work should be taken into account at the time of sentencing.

That is one of Yates's notions. Its opposite, which Yates holds with equal passion, is this: It is unfair that judges, depending on how they feel, where they sit and how the offender impresses them, will hand out wildly variant sentences for essentially the same offense.

The shoplifting son of a well-to-do businessman may get probation, particularly if he is careful to appear remorseful, while a youngster from the ghetto may impress the judge as a hopeless ne'er-do-well and draw a maximum term. A general rule, says Yates, may be that a judge's sentencing is influenced by the degree to which he sees himself (or a member of his class) in the face of the offender.

When the Bonabond director sees such judicial unfairness, he finds himself arguing for a sort of supermarket justice: armed robbery and petty theft, like sirloin and mayonnaise, should have the same price tag for everyone. Judges should have no more discretion than check-out clerks to change the prices, for discretion will lead inevitably to abuse and discrimination.

This notion is very much in line with one of the provisions now under consideration in the bill to revamp the federal criminal code: a provision that would severely restrict the sentencing discretion of judges.

Under the provision, a Federal Sentencing Commission would develop guidelines for federal judges to follow in shopping-list fashion, with no more than a 25 percent difference between minimum and maximum sentences.

Yates would buy it -- until he reflects on his belief that a lot of people are pushed into crime by economic circumstances over which they have no control.

He cannot bring himself to believe that an out-of-work father who steals to forestall his family's eviction should be treated exactly the same as another man who steals simply because the opportunity presents itself. The morality of the two offenses is different, he believes, and that ought to be taken into account at time of sentencing.

For offenders for whom the "driving force" is their economic condition, Yates would urge shorter sentences, probation, halfway houses and job training. He argues that you simply cannot know enough about proper sentencing if you look only at the offense and not at the peculiar circumstances of the offender. It is an argument for judicial discretion.

Then he looks at how judges actually use their discretion and he can think of only one cure: flat time.