The Paul Ward case is interesting and unusual from a legal point of view. Ward was convicted on the evidence of Charles Bowden, a man who was on his own admission equally involved in the murder of Veronica Guerin. He was therefore an accomplice. Bowden, serving a sentence for other crimes, has been taken into a witness protection programme, with whatever benefits that confers.
The other pillars of the case against Ward - alleged admissions by him while in custody - were rejected by the Special Criminal Court in strong terms. The court doubted if the admissions had been made at all but, even if they were, regarded them as tainted by breaches of Ward's constitutional rights in custody. They were therefore inadmissible. Moreover, the court found that in certain aspects of his evidence Bowden was a self-serving liar.
The court found, however, that Bowden's account of Ward's involvement was accurate and credible and further held that Bowden had no vested interest in trying to shift responsibility for his own actions on to Ward.
Bowden's is serving a six-year sentence on which he could expect one-quarter remission in the ordinary way. He has not been told what will become of him in the witness protection scheme.
Because of the lack of corroboration the case presents the classic dilemmas of accomplice evidence in an unusually pure form. It can be anticipated that any appeal will proceed on the basis that there is no corroboration and that Bowden had a great deal to gain.
Accomplice evidence is found where a person who admits to being implicated in a crime nevertheless makes himself available as a prosecution witness in a case against another person accused of that crime or a series of crimes.
By its very nature, the evidence of an accomplice is unreliable. Despite this, prosecutors have made use of the evidence of accomplices, in varying degrees, as far back as history records.
Contemporary experience shows that prosecutors are particularly prone to use accomplice evidence when confronted with crime by tightly organised bodies of people, paramilitary or otherwise. In such circumstances the authorities may be driven to the distasteful expedient of relying on the evidence of the hardened criminal. Accomplice evidence may be largely true (for example as to his own role) but false in its implication of another person or persons. This aspect - a general appearance of truth which makes any inventions hard to detect - has been found in Northern Ireland to be a great risk of this sort of evidence.
In an attempt to redress the obvious dangers of accomplice evidence, the courts have for centuries treated it with extreme caution.
If there is a jury, it must be warned to treat such evidence with "extreme care and even suspicion" and that it is dangerous to convict in the absence of "some independent testimony which confirms not only some material particular of the accomplice's evidence but which connects the defendant himself to the offence charged" - corroboration, in other words.
The term supergrass (and the court found that Bowden was not one) originated in Northern Ireland in the early and middle 1980s when there was large-scale reliance on accomplice evidence. This was in the context of sustained terrorism.
From small beginnings this system grew to the remarkable extent that in the Steenson case (1986), 27 men were convicted of multiple offences including murder, after a trial lasting 99 sitting days on an indictment with 198 counts, all on the evidence of a single accomplice.
It is almost inevitable that any supergrass will have been thoroughly schooled. The very process whereby he has convinced the police of his accuracy and reliability will itself be a refining and training process.
It is also extremely difficult to cross-examine an accomplice on the substance of his allegations. A defendant who is guilty cannot instruct his lawyers as to inaccuracies without revealing his own guilty knowledge.
Cross-examination must, therefore, be largely directed at the witness's credibility, the process whereby he became a state witness, any inducements given him, his general character and any specific details on which the defence can hope to contradict him.
Northern Ireland experience has shown that, if the use of accomplices, and particularly supergrasses, becomes widespread, the unreliability of any particular witness is compounded.
Harry Kirkpatrick, the supergrass in the Steenson case, was a bridge too far. He implicated a huge number of people; he denied any possibility of error in any part of his very detailed evidence; he swore to demonstrable falsehoods; and he sought to protect himself from cross-examination by feigning failure of memory or uncertainty about any matter outside his core evidence.
His evidence was accepted by the trial judge who concluded that although he had lied about almost everything else, he told the truth about the crimes themselves. This was too much for the Northern Ireland Court of Appeal which quashed the convictions, apart from a few in which there was confession evidence, in a judgment so revealing of the inherent weakness of this sort of evidence that reliance on it effectively ceased.
There will always be a role for accomplice evidence, but it is a small one. If the accomplice is a hardened criminal, or worse still a supergrass, his evidence is always profoundly suspect and a conviction relying on it, in the absence of corroboration, will be rare.
Adrian Hardiman is a Senior Counsel