"The imposition of sentence is probably the most critical point in our system, of administering criminal justice," observed Marvin Frankel, a distinguished jurist, in 1973.

It may, literally, mean the difference between life and death, freedom or confinement, short-or long-term imprisonment.

The power of the sentencing judge, in many jurisdictions, is a judge may decided to sentence one robber to probation and another, different in no relevant respect, to 20 years in prison. Nor can these sentences generally be reviewed by a higher court.

Despite the enormous power of the sentencing judge, the process of imposing sentence is essentially lawless. There are few guidelines and virtually no accountability.

Both observes of, and participants in, the American criminal justice system are almost unanimous in viewing the process of imposing sentences as a dismal failure by any standard.

Yet the imposition of sentence is "crucial" because for many defendants it may be the only point in the criminal justice system - other than bail determination - where a judicial decision is made. Despite popular fascination with the drama of the courtroom trial, the vast majority of criminal cases are disposed of without any trial. The defendant agrees to plead guilty to a given crime, in exchange for some concession by the prosecutor - a reduced charge or a promise to recommend a reduced sentence.

In some jurisdictions, judges participate overtly in this bargaining. In most jurisdictions, however, judges remain aloof from the negotiation. They retain the power - at least in theory - to accept or reject the prosecutor's recommendation and to impose any sentence within the statutory range.

The unfairness and uncertainly of this sentencing system has been amply documented.

In one recent study, 50 federal judge were given 20 indentical files, drawn from actual cases, and asked to indicate the sentence they would impose on each defendant. In a case of possession of barbiturates with intent to distribute, one judge gave the defendant five years in prison, while another put him on probation. One judge sentenced a defendant convicted of securities fraud to two years imprisonment, while another fined him $2,500.

This study, commissioned by a group of judges, concluded that there were "glaring disparities" in sentencing. Similarly, a recent study of sentences imposed during a two-year period in Montgomery County, Ohio, disclosed that certain judges imprison defendants four times as often as other judges for the same offense.

Disparities of this kind cannot be explained by differences among criminals. They are - as one judge recently observed - a function "of the wide spectrum of character, bias, neurosis and daily vagary encountered among occupants of the trial bench."

There is also evidence that some of the disparity is a function of prejudice: social, economic and cultural. An exhaustive study of state and federal sentences for larceny and assault disclosed that blacks have a one-and-a-half times greater chance of being imprisoned than whites with similar records. Other studies have shown that defendants apearing in low status dress are significantly more likely to receive prison sentences than comparable defendants wearing higher status clothing.

Two centuries ago, Blackstone, the great English legal commentator, observed that the sentences handed down by judges are not "their" sentences, but the sentences of the "law." Today, it is the judge - as an individual - who decides who shall be imprisoned: and it is the judge and the members of the parole board, not he "law" as an abstraction, who decide how long an imprisoned defendant shall serve.

Recently, there has been mounting criticism - from the left and right alike - of a sentencing system that makes so much depend on the idiosyncrasies of individual sentencing judges.

Liberal critics believe the sentencing system discriminates against poor and minority criminals and in favor of white-collar and privileged criminals. Conservative critics argue that current sentencing practices result in the early release of dangerous, violent people.

The specific focus of much of this critism has been the so-called indeterminate sentence - a mechanism by which the amount of time a convicted criminal will actually serve is decided by the "parole board" or "adult authority" while the prisoner is serving his sentence. Both the legislature and the sentencing judge still play important roles; the legislatue sets the outer limits of the permissible punishment for the type of crime, while the judge decides on the desirable range for the crime and criminal. But these limits are often broad, and the parole agency thus becomes responsible for deciding what realy counts: When the defendant will be released.

The indeterminate sentence is merely one manifestation of the existing disparity in sentencing.The underlying cause is the unchanneled discretion exercised by all the sentencing decision-makers-judges, prosecutors, parole boards, and adult authorities.

In an effort to impose some uniformity of sentencing, a number of legislature - including Congress - are now considering significant reforms. Some of these reforms, however, address only a small part of the problem.

For example, mandatory minimum sentencing for certain offenses deals only with discretion at the low end of the sentencing spectrum. It "requires" judges to impose a certain minimum sentences (perhaps a year) upon eveyone convicted of a specific offense (for example, illegal possession of a handgun, as in Massachusetts).

Flat-time sentencing retains "judical" discretion by allowing the judge to select the "appropriate" sentence from a wide range of alternative; but it eliminates" parole board" discretion by requiring the inmate to serve his entire term (minus "good time").

The approach that seems to be attracting the most attention is a compromise solution called "presumptive sentencing." Under that approach, or its many variants, the legislature decides not only on the minimum and maximum sentences for a given crime, as it does today, but also on the "presumptive" sentence for a "typical" first offender convicted of a "typical" instance of this crime.

The legislature might thus decide that the typical burglar - an unmarried, unemployed, uneducated male in his early 20s who broke into an inhabited house late at night without a weapon and took several hundred dollars worth of valuables - should generally serve one year. One year would thus become the presumptive sentence for this crime.

In the absence of legislatively specified aggravating or mitigating circumstances, the sentencing judge would be expected to impose that sentence on all first offenders convicted of that crime. If the judge departed from the presumptive sentence, he would have to detail in writing the reasons for his decision. All sentences departing form the presumptive one by more than a specified percentage - for example, 25 per cent - would be automatically appealable. The sentence would be reversed unless the appellate court concluded that the judge's reasons had overcome the presumption in favor of uniformity.

Under this approach, the parole board would retain only limited power under unusual circumstances to release the inmate before the expiration of a statutorily fixed percentage of his sentence (for example, 75 per cent).

In the end, neither this nor any other proposed solution to the dilemma of sentencing will be a panacea. The elusive quest for the fitting punishment has occupied the collective wisdom of mankind since the beginning or recorded history.

The pendulum appears now to have swung in the direction of greater certainly and uniformity in sentencing. Undoubtedly some reform will be forthcoming, and we will see not the demise of individualization in sentencing, but its waning influence. Perhaps a decade from now a reactionwill again set in and the pendulum will swing back in the direction of increased flexibility.