Almost unnoticed, a fundamental change in penal policy is gathering pace. The main factor in the length of a sentence is, increasingly, not the severity of a crime, but the supposed risk that an offender will do something worse if released.
Risk assessment is at best an inexact science - often, as we shall see later, shockingly so. But its emerging role in the sentencing process is having dramatic consequences: hundreds, soon to be thousands, of petty arsonists, pub brawlers and street muggers are in effect being given life, usually on the basis of highly subjective pre-sentence reports.
The change is certain to cause a further great rise in the prison population, already at record levels, having grown faster under new Labour than under any previous government. It is also arousing deep concern among lawyers, and will top the agenda at a special conference organised by the Criminal Bar Association in Birmingham next month.
The consequences for classical notions of justice are profound. Old lags have a saying: "If you can't do the time, don't do the crime." It needs reformulation: "If you can't do the time, don't have a background that might make a bureaucrat think if ever you're set free, you might be dangerous to the public."
At the heart of this shift is a piece of legislation whose import was barely appreciated when it passed through parliament: the Criminal Justice Act 2003, with its indeterminate public protection (IPP) sentences. According to the act, judges must impose an IPP - life in all but name - on any person convicted of any one of 153 separate violent and sexual offences, if they believe, in the words of the act, that there is "a significant risk to members of the public of serious harm from the commission of further specified offences". In an appeal judgment last year, one of the most senior criminal appeal judges, Lord Justice Judge, made clear how huge a departure this is: "Although punitive in its effect, with far-reaching consequences for the offender on whom it is imposed, it [the IPP] does not represent punishment for past offending. The decision is directed not to the past, but to the future."
It is not as if it had been difficult to lock up the truly dangerous under existing legislation. For decades, judges have been able to give "discretionary" life sentences to those convicted of certain crimes short of murder, such as rape and wounding with intent to kill. After 1998, the "two strikes and you're out" law (now replaced by the IPP) meant that anyone convicted of one of a short list of very serious crimes for a second time got life automatically.
However, the scale on which the IPP is being used dwarfs these older measures. Discretionary and automatic life sentences used to be given about 200 times each year. Since coming into force in April 2005, the IPP has been imposed more than 2,000 times, with a rate of over 100 new IPPs each month. As an official told me, Home Office models predict that by 2011, there will be 12,500 inmates serving IPPs - more than three times as many as those doing life for murder.
Hard cases, as the saying goes, make bad law, and the cases that the government cited to justify the 2003 act were very hard indeed - such as the release of the paedophiles who killed the 14-year-old Jason Swift in 1985. Such debate as there was helped to fuel a "risk panic", in which the media have focused obsessively on crimes committed by previous offenders who should have been under supervision, such as the 2005 murder in Reading of the teenager Mary Ann Leneghan.
It cannot be stressed sufficiently that there is no empirical basis for this panic at all. A study of sex offenders emerging from long-term imprisonment, published in 2002 by a team led by Oxford University's Roger Hood, found (as had earlier, similar projects) that their reconviction rate is reassuringly low. Of the 94 followed for six years after release, only eight were reconvicted for a further sexual offence. Another four were reconvicted and jailed for a non-sexual violent crime. Since the study, multi-agency schemes to monitor such offenders after release and more widespread sex offender treatment in prison may well have reduced this risk.
Even freed lifers commit relatively few crimes. The number of homicides by those previously convicted of homicide and released varies each year between zero and two - roughly 0.3 per cent of murders. In all, about 3 per cent of freed lifers will eventually be convicted of an imprisonable offence.
Nevertheless, public rhetoric is at fever pitch. A recent Observer article claimed that the system for monitoring freed sex offenders in the community is close to collapse. This was accompanied by an editorial headlined: "Control these terrifying predators". Robert Whelan, of the think-tank Civitas, told the Sunday Times that the lesson from the Mary Ann Leneghan case was that all offenders should be kept much longer in prison. Those who disagreed, he said, were "diehard utopians".
Radical as the 2003 act is, it was not the first attempt to protect society from its most dangerous members. From 1857, government criminal statistics began to include figures for "known thieves and depredators". The long search for ways to deal with them began.
By the end of the 19th century, social Darwinism and the work of writers such as Cesare Lombroso had added a veneer of science to the notion of the predatory criminal, and preventive detention for periods far in excess of the ordinary prison sentence was increasingly seen as the solution. Some, such as the writer Bruce Thomson, argued that these inmates should also be castrated: "With cattle, this kind of selection is in fact almost always followed: for hardly anyone is careless enough to allow his worst animals to breed. Why, then, should incorrigible criminals go into prison for short periods only, only to be sent out again in renovated health, to propagate a race so low in physical organisation?"
The term "personality disorder" had yet to be invented, but many argued that dangerous criminals had something like a disease, which made their behaviour incorrigible. As the Westminster Review put it in 1898, "the criminal, while not in the ordinary sense lunatic, is thoroughly irresponsible, hopelessly perverted and mentally and physically incapable of reformation. He is a dangerous animal, and society must be protected against him."
Tyranny of bureaucrats
Against this background, the first public protection sentencing measure took shape: Herbert Gladstone's Prevention of Crime Act 1908. When sentencing a criminal, the court would pay close attention to his record, and so deduce whether there was a high risk of recidivism. If the court thought there was, indeed, "evidence of habituality", it could impose a "dual-track" sentence - first an ordinary penal element, equating to what would previously have been the total sentence, and then "preventive detention", usually for five years. Thus, Gladstone argued, the most dangerous villains would be incapacitated.
The Liberal MP and writer Hilaire Belloc argued in vain that this was "utterly at variance with every political or social principle that western Europe had ever known" for more than 3,000 years. The act, Belloc said, would enshrine the "tyranny of bureaucrats".
One of the sharpest critics of the 1908 act was the home secretary who took office two years later, Winston Churchill. His fear was that it was likely to fall hardest not on the most dangerous, but on prolific petty criminals. "The general test should be - is the nature of the crime such as to indicate that the offender is not merely a nuisance but a serious danger to society?" Churchill wrote in an official circular. Thus he identified a crucial issue of enormous relevance today.
Compared with its 2003 successor, the 1908 act was relatively little used - usually there were fewer than 100 preventive sentences a year. But when it was finally evaluated by a departmental committee in 1932, the findings were devastating. As Churchill had predicted, most of those sentenced were not dangerous at all, but "men of little mental capacity or strength . . . whose frequent convictions testify as much to their clumsiness as their persistence in crime".
Undeterred by this failure, Attlee's Labour government passed another preventive detention act in 1948. In 1963 an inquiry by the home secretary's Advisory Council on Sentencing showed that this was equally unjust. Most of those given the new form of preventive detention - up to 180 people a year - were, in the words of a report from Cambridge University, not predators, but "passive inadequate deviants". In 1967, the 1948 act was repealed.
Evidence is now emerging that the defect spotted by Churchill in the 1908 act is equally manifest in that of 2003. According to official figures, just 28 - 4 per cent - of the first 707 IPPs were imposed on those convicted of crimes against children, and a further 40 on rapists. Forty-four of those sentenced were arsonists, and 149 had been convicted of wounding. But by far the biggest group - 284 prisoners, or roughly 40 per cent of the total - were given IPPs for robbery, almost all of them for street crimes or mugging.
To be sure, some muggers may, on release, go on to commit murder. A minority of arsonists may one day set a fire in which someone is badly hurt. Yet it is clear that the crimes actually committed by many of those getting IPPs are relatively insignificant - and nothing like as serious as those which would previously have merited life. The median tariff set for the first 707 IPP prisoners - that is to say, the penal element of the sentence, the term that would have been imposed before the 2003 act - is just 30 months.
In theory when the tariff expires, IPP prisoners, as with those serving life for murder, become eligible for Parole Board review, and hence possible release. But as the board's chairman, Sir Duncan Nichol, pointed out last December, even if they do get out at their first opportunity, they will probably spend much longer in jail than before. Most IPP inmates will have spent months on remand, and by the time they have been sentenced and settled in a long-term prison, their review date may well be looming.
It takes months, however, for the various prison and other officials necessary to this process to prepare and write their reports. In addition, before freeing someone already deemed dangerous, the Parole Board will need to see that their so-called "dynamic risk factors" have declined - perhaps by graduation from a long course of therapy while in prison. The likelihood is that the time many inmates spend in jail being assessed as future risks will be longer than the time served as punishment.
Nichol commented: "The global impact of IPPs will be that prison overcrowding will increase; places on offending behaviour courses will be scarce; prisoners may spend more time in custody awaiting such courses when they might otherwise have been released earlier; crucial time will be spent writing parole reports by prison and probation staff who have other duties; and the Parole Board will need increased resources to deal with a quadrupling of our indeterminate casework."
A further factor seems likely to make sentences longer still. The 2002 study of long-term sex offenders by Professor Hood and his colleagues suggested that the Parole Board has an innate and understandable bias towards being overcautious: no one wants to be responsible for setting a maniac free. The study looked at what happened to 82 men whom the board decided were "high-risk" and hence refused parole. Four years after their eventual release, just seven had been reconvicted of a further sexual offence, and four for a violent offence - a false positive rate of 92 per cent for a sexual crime, and 87 per cent for either sex or violence.
These men had been serving determinate, fixed-term sentences, so knew the latest they would get out: perhaps, in their cases, such apparent overcaution was acceptable. Now, however, many of the same type of prisoners will be serving IPPs and may never be freed at all, or be freed only after an exceedingly long time.
Sex offenders sentenced today will have been assessed using one or more structured, psychological "tools". These are not, in any sense, infallible: derived from analysis of prob abilistic risk factors shared by groups, they are a very blunt instrument when applied to individuals - when even their staunchest advocates admit that they are no more than about 75 per cent accurate. Hood's team also looked at what happened after release to prisoners deemed to be high-risk using the tool known as Static-99, and found a false positive rate only slightly better than the Parole Board's.
But at least these tools have been evaluated. Perhaps the most disturbing aspect of IPPs is that the method being used by judges in most cases has no such basis at all.
According to the Home Office, most non-sex offenders considered for an IPP - that is, more than 90 per cent - will have been subject only to the standard OASys pre-sentencing assessment, a box-ticking exercise by a probation officer who often will have no direct experience of the offender. Having considered, in a formulaic manner, a variety of factors derived from the paper record, the officer ends the process by making what amounts to a purely subjective judgement: is this offender low, medium, high or very high risk? If either of the latter two, he will probably get an IPP.
I spoke to officials who designed the OASys forms. To date, there has not been any evaluation of their predictive powers, even though they are now being used daily to justify indefinite incarceration. It is, as Belloc put it, "the tyranny of bureaucrats". A new study by the criminologist Diana Fitzgibbon can only amplify such concerns. Based on her investigation of OASys risk analysis in a large city, it finds evidence of "risk inflation", leading to the "non-transformative warehousing" in prison of those thought to be risky. She also discovers widespread gaps in offenders' OASys files - omissions of crucial episodes in their histories that might, had they been considered, have led to their being given a lower risk category.
Perhaps this risk inflation explains some of the cases of IPPs that have already been upheld by the Court of Appeal. For example, there was the 18-year-old with no previous convictions who set fire to two rubbish bins at a seaside resort; and the middle-aged man who set fire to his house when his mother went into a care home - he didn't even do the job very effectively: he poured cooking oil on some cushions and tried to set them alight. Among the numerous robbery cases, there was a man who punched a pizza delivery driver in the chest, and tried unsuccessfully to steal his car; later that night he threatened - though did not hurt - someone else and made off with his mobile phone.
In his story (later a Tom Cruise movie) "The Minority Report", Philip K Dick wrote of a system governed by the notion of "precrime", where people who had yet to do anything wrong were convicted and sentenced because the authorities "knew" they would.
Dick was writing science fiction, but even minus his "precogs" - weird beings with the gift of second sight - the precrime world is already here. It shows every sign of becoming an authoritarian dystopia.
The facts of crime and punishment
Research by Sarah O'Connor
2,000 number of preventive detention orders issued in the UK since April 2005
3% only this proportion of freed lifers is reimprisoned for new offences
17,000 increase in UK prison population since new Labour came to power
8,000 number of extra prison places the Home Office plans to create by 2012
£100,000 cost of creating each new prison place