O'Brien, R (on the application of) v Director of Public Prosecution

This is the judgment of the Court to which both members of the court have contributed. On 12 October 1987, a Cardiff newsagent Phillip Saunders was murdered. The Claimant (Mr O’Brien) and two other men, Darren Hall and Ellis Sherwood, were charged and subsequently convicted of the murder, in July 1988. Mr Hall implicated his two co-defendants in the robbery and murder of Mr Saunders. All three were sentenced to life imprisonment.

On 25 January 2000, the Court of Appeal (Criminal Division) heard an appeal by all three men against their convictions. The appeal came before the court by way of a reference by the Criminal Cases Review Commission, Case No. 98/6926/27/28/S1. The central plank of the appeal was that the “confessions” of Hall and his implication of his co-defendants were unreliable, because of his psychological state and mental makeup. After receiving fresh evidence bearing on that point, the Court of Appeal quashed the convictions.

An ancillary point in the appeal related to evidence given at the trial by Detective Inspector Lewis. It was his evidence that he had overheard an incriminating conversation between Mr O’Brien and Mr Sherwood in the cells of Canton Street Police Station. The Court of Appeal reached no conclusion about the veracity of DI Lewis’s evidence, but did conclude, for reasons we shall explain below, that his evidence would be opened to fuller challenge at any retrial. Following the appeal, Mr O’Brien was released. He had spent more than eleven years in custody.

After his release, Mr O’Brien sued the South Wales Police for malicious prosecution and misfeasance in public office. There was a broad range of allegations in the civil claim, but one important plank was the allegation that DI Lewis had fabricated the evidence of the conversation in the police station and perjured himself when giving evidence about it to the jury. Very shortly before the due trial date, the South Wales Police settled the claim for a substantial sum of money: this court was told there was a payment of £300,000 for non pecuniary losses. The settlement was in October 2006. It was reached shortly after service of some of the expert linguistics evidence placed before this court.

Following the settlement Mr O’Brien sought to have DI Lewis, who by then had retired from the police with the rank of Detective Chief Inspector, prosecuted for perverting the course of justice, perjury and/or misfeasance in public office. A fresh police investigation took place the results of which were submitted to the Crown Prosecution Service. On 16 November 2012, the Specialist Prosecutor of the Special Crime and Counter Terrorism Division of the CPS concluded that there was insufficient evidence to provide a realistic prospect of conviction against Stuart Lewis for any offence. This decision was communicated to Mr O’Brien’s solicitors in a letter of 19 November 2012.

Mr O’Brien seeks judicial review of the decision not to prosecute. Through his counsel, Ms Heather Williams QC, Mr O’Brien accepts that, in order to succeed in such a claim, he must show that the decision was one at which no reasonable prosecutor could have arrived. Ms Williams formulates her challenge under three heads. Firstly she submits that the prosecutor misdirected herself and/or conducted a seriously flawed analysis in relation to the evidence from the linguistic experts. Secondly, she submits that the prosecutor failed to conduct a balanced and even-handed assessment of the evidence, in particular failing critically to examine the accounts provided by Mr Lewis and his police colleagues: and thirdly she suggests that the prosecutor misunderstood and/or failed to take into account highly relevant evidence on a number of specific issues. Taken together, Mr O’Brien contends that the flaws in the prosecutor’s reasoning render the decision unsustainable.

Phillip Saunders was attacked and robbed on 12 October 1987. He died from his injuries on 17 October. The three men who were tried were arrested and interviewed in late October that year, but all three were released on 2 November. As the Court of Appeal in 2000 recited, all three were released despite the fact that Hall had already made admissions of being involved in the attack on Saunders, which admissions the police clearly doubted at that stage. Following further enquiries, the three were re-arrested on 10 November 1987.

At the time of his arrest Mr O’Brien was 20. He was a painter and decorator with one young child and another on the way. He had no previous convictions, although it was his own case that he had met the other two men on the evening of the attack in order to take and drive away motor cars. He had vehemently denied involvement in the murder during interviews before his release on 2 November. The final interview before his release involved a confrontation between Mr Hall and Mr O’Brien, the central passages of which read as follows:

Sgt Fenton (indicated Darren Hall) “I am Sgt Fenton. Do you know this person?”

O’Brien: “Course I do. What you been saying.”

Hall: “It’s no good Michael, you might as well come clean. I told them I was at the entrance and you and Ellis went in the car park.”

O’Brien: “You bastard you’ll fucking get it.”

Hall: “I told them we talked about rolling Saunders.”

O’Brien: “Yeah yeah, I know that but I didn’t.”

Fenton: “Darren [Hall] Isn’t it right that when you were keeping watch you saw Saunders arrive and a few minutes later and you heard noises?”

Hall: “Yes Bang Bang and they come out and told me to run.”

O’Brien: “You bastard you’re a load of shit.”

Hall: “It’s up to you Michael I’ve told them the truth.”

Mr O’Brien was arrested at 8.30 in the morning of 10 November, arriving at the police station just before 9.00am. He was held without legal access through into the early afternoon, when he saw a solicitor. He was interviewed during the afternoon, following which at 18.30 he indicated he was not going to answer any further questions which he had previously been asked. At 20.30 hours he was in the cells at the police station.

DI Lewis was the Detective Inspector in charge of the investigation. He had joined the police force in 1966 and became a detective in 1973. He was promoted to Detective Inspector in 1983. Over the course of his career he had a considerable number of commendations from judges, magistrates and superior officers. As we set out below, one significant case, details of which emerged after these events, would have given some material for hostile cross-examination as to DI Lewis’s record. A number of complaints were made against him between 1978 and 1997. None has been found to be proved or substantiated.

Mr O’Brien’s custody record shows that he was checked in his cell at 8.30pm. Mr Sherwood’s custody record shows that he was returned to the cells at 8.40, after an interview and a meal. DI Lewis’s evidence was that he went to the cell area and overheard a conversation between them. Mr Lewis made an entry in relation to this conversation in Mr O’Brien’s custody record. It is timed 20.43 and reads:

“Conversation between O’Brien and Sherwood overheard by DI Lewis. Note made. Signed S Lewis DI”

It is noteworthy that this entry follows an entry of 20.50 which records Mr O’Brien being removed from the cell area to be finger printed.

DI Lewis made a Criminal Justice Act statement on 4 December 1987. His account was that he had placed himself outside the cell complex at the police station at 20.43, and heard the conversation between the two men. He said that he made a note on the back of an expenses form which he had with him. His statement gives the conversation as follows:

“O’Brien: They are going to charge me and you.

Sherwood: No they are not, all they have is Hall, he is grassing us.

O’Brien: I can’t hold out for much longer, I may have to tell them the truth.

Sherwood: Don’t do that, we’ll be fucked.

O’Brien: I can’t hold out much longer, I might have to tell them what happened.

Sherwood: You’re talking about life, being on remand means nothing.

O’Brien: I can’t hold out much longer, I’m scared. I’ll have to tell them what happened.

Sherwood: Just keep your mouth shut.

O’Brien: Why don’t you tell them what happened.

Sherwood: I can’t can I? If Hall hadn’t opened his mouth we wouldn’t be here.

Sherwood: I think there is someone listening. I’ll catch up with you later, OK?

O’Brien: Yeah, OK.”

The custody record shows Mr Sherwood being removed from his cell at 21.00 for “process” returning to his cell at 21.20. The next entry in Mr Sherwood’s custody record is the entry by DI Lewis relating to the conversation, again timed at 20.43, out of sequence. Mr O’Brien was returned to his cell at 21.50.

Although the matter appears not to have surfaced in the original trial, two officers subsequently gave evidence that DI Lewis showed them the notes of the overheard conversation shortly after it had taken place and before the Defendants were confronted in interview with what DI Lewis suggested they had said.

The custody record then shows Sherwood being removed from the cell by CID Officers at 22.05. In the course of the ensuing interview, the evidence of DI Lewis was that he read from his manuscript notes the conversation he said had taken place between the two men. Since the detailed responses featured in the eventual decision not to prosecute, we reproduce them here. The record of interview interleaves the quotation from the notes, with questions from the officer and responses from the prisoner. In the following quotation, passages from the cell conversation being quoted to the prisoner are in italics:

“Lewis: around 8.43pm this evening the following conversation was heard between you and Michael O’Brien in the cells here by me.

O’Brien: They are going to charge me and you.

Sherwood: No they are not, all, they’ve got is Hall he’s grassing us.

Sherwood: Wrong we did not say that.

O’Brien: I can’t hold out for much longer I may have to tell them the truth.

Sherwood: Go on down and find out the truth, you keep on at him …..long enough he’ll tell you, he’ll end up in Whitchurch, he’ll have a mental breakdown.

Sherwood: Don’t do that we’ll be fucked.

O’Brien: I can’t hold out for much longer I might have to tell them what happened.

Sherwood: There was no such thing like that mentioned I don’t know where you got it.

Sherwood: You’re talking about life. Being on remand means nothing.

Sherwood: I didn’t say that that’s a load of bollocks.

O’Brien: I can’t hold out for much longer I’m scared I’ll have to tell them what happened.

Sherwood: Just keep your mouth shut.

O’Brien: Why don’t you tell them what happened?

Sherwood: I can’t can I? If Hall hadn’t opened his mouth we wouldn’t be here, I think there is someone listening, I’ll catch up with you later.

O’Brien: Yeah OK.

Conversation ended.

Lewis: Did any of this conversation take place at 8.43pm tonight?

Sherwood: About the only thing that took place down there was “someone’s listening I’ll catch you later”

…….”

The interview with Mr Sherwood ended at 22.40. At 22.45, Mr O’Brien was removed from his cell by detectives and at 22.48 was charged with the murder. Sherwood was charged immediately thereafter at 22.50.

At 23.05, after being charged, Mr O’Brien was interviewed about the cell conversation. His solicitor was present. Adopting the same differing type faces, this short interview went as follows:

“Lewis: At 8.45pm tonight whilst you and Sherwood were in the cells, the following conversation took place between you.

O’Brien: They’re going to charge me and you.

Sherwood: No they’re not, all they’ve got is Hall, he’s grassing us.

O’Brien: I can’t hold out for much longer. I may have to tell them the truth.

O’Brien: No way – that’s bare faced lies.

Sherwood: Don’t do that we’re be fucked.

O’Brien: I can’t hold out for much longer.

O’Brien: No way, you’re lying there.

CONTINUED O’Brien: I might have to tell them what happened.

Sherwood. You’re talking about life. Being on remand means nothing.

O’Brien: I can’t hold out for much longer. I’m scared. I’ll have to tell them what happened.

O’Brien: I’m scared I might not see my kids again.

Sherwood: Just keep your mouth shut.

O’Brien: Why don’t you tell them what happened.

Sherwood: I can’t can I? If Hall hadn’t opened his mouth we wouldn’t be here. I think there is someone listening I’ll catch up with you later OK.

O’Brien: Yeah OK. I’ll tell you what Paul, most of that is lies in there, when I said I was scared, I was scared of losing my kids and my wife.

Lewis: Did you not repeat I can’t hold out much longer?

O’Brien: No not in that sense no. When I said I couldn’t handle it much longer, I mean the questioning and the hassle I’ve had.”

Mr O’Brien signed the record of interview as being a correct record. As far as the court is aware, there has never been a challenge to the accuracy of the record of the interview itself.

At the murder trial in the summer of 1988, Mr O’Brien was represented by very experienced senior counsel, Mr John Diehl QC. Mr Lewis gave evidence in support of the overheard conversation. His Criminal Justice Act statement incorporated word for word the conversation as it had been put to the suspects in interview.

No transcript of DI Lewis’s evidence exists. The best available record of the evidence is to be found in the notes of the CPS representative attending counsel at court. According to this record, Mr Lewis told the court that Mr Sherwood and Mr O’Brien were in two separate cells and that this was the first opportunity they would have had to communicate that day. He went there with the intention of listening in, because he wondered if they would say anything incriminating. He had no other officer with him. The note of the cross-examination by Counsel for Mr Sherwood goes on as follows:

“I recorded what I heard on these two pieces of paper. Mr Sherwood and his solicitor must have heard it. As I approached cell block I arrived in the middle of the conversation. I had a file to rest on – standing up. He wasn’t just talking about the fact that his wife was having a baby. He didn’t say he wanted to be bailed ASAP to be with her. The custody record was kept for each……….2043 the conversation was heard by ……Lewis. Note made. It was a retrospective entry. By the time I had discussed what I had heard with others, other officers had been to the cells and made entries. It was more important to discuss the content. This entry must have been made before they were charged. ……..”

The note of cross-examination by leading counsel for Mr O’Brien deals with other matters before turning to the cell conversation in the following terms:

“The conversation on 10 November which I overhead – (papers shown to J Diehl) the date on top of the note is 10 Nov. The zero was a dot. I was positioned at the gate to the cells. It was a verbatim note. There were substantial pauses before each replied. Custody Record – 2043 – note made – ref to conversations. I can’t say when I entered it on CR [Custody Record] – between 2050 and 2150 – anytime in that hour. The time is the time conv. took place – not the time of the entry. I spoke to DS Rogers possibly DC Bishop. He did say “I might have to tell them the truth” I felt it proper to put the conversation to him and his solicitor after I charged him – I had already charged the other two. I asked him about the conversation at 8.05 it was my decision to charge him.”

A number of relevant matters emerge from the judge’s summing up in the trial. Firstly, both Mr Sherwood and Mr O’Brien acknowledged in evidence that they had lied repeatedly in the course of the investigation. Further, the judge recited to the jury the record of the overheard conversation and the responses of each of the Defendants as significant evidence in the trial. He reminded the jury that in evidence, Mr Sherwood had described a whole series of witnesses, eight in number and including DI Lewis as “all liars”. There was no doubt at the trial as to the suggestion being made by the Defence. In dealing with Mr O’Brien’s case, the Judge noted that in cross-examination Mr O’Brien denied making admissions which had been recorded in the record of interview and signed by him as being correct. In relation to the overheard conversation, Mr O’Brien also gave evidence that DI Lewis’s account consisted of “barefaced lies”.

There was no exploration in the course of the trial as to whether such a record could have been written in the time available, as to whether it could have been written either verbatim or accurately, and no expert evidence was brought to bear on this aspect of the case at all.

Pausing there, it seems to us that the statement of DI Lewis, reciting the conversation in the way it does, carried the implication that his record of the conversation was verbatim. He appears to have confirmed that under cross-examination from Mr O’Brien’s counsel. It is agreed that at the trial he had with him the expenses claim form. From the note of evidence, it is clear that this was more than a single sheet. It appears also to be unchallenged that, from these manuscript notes, DI Lewis was able to reconstruct the conversation in identical terms, during the two separate interviews of the Defendants.

No copy of his notes survives. Despite the strong denials by Mr Sherwood and Mr O’Brien that they had said anything incriminating, and despite the challenge to some of the reported conversation, the notes were never made an exhibit, nor were they retained after the trial. Therefore no copy of these notes would be available for any prosecution today.

As we have indicated, when the Court of Appeal considered this case in 2000 following referral from the CCRC, the central issue was a consideration of the confession of Mr Hall and his implication of his co-defendants, in the light of developments emerging after the convictions. We need not concern ourselves with that aspect of the case.

However, the Court of Appeal did give some consideration to the overheard cell conversation. The court observed that at the time of the trial there was “little scope for cross-examination to undermine [Lewis’s] evidence that he had overheard an incriminating cell conversation” other than the timing of the entry in the custody record, which was readily explained by saying that 20.43 was the time of the conversation, not the time of the entry. The court went on to note that events in the course of the trial of Griffiths & Others at Cardiff in September 1983, a case known as the Welsh Bomb Trial, had subsequently given rise to real concern, due to inconsistencies between manuscript notes and typed copies of those notes. Although the principal concern in relation to this discrepancy focused on the West Midlands Serious Crimes Squad, the court did observe:

“it is difficult to see how additional words could be inserted into a typed version of notes which were apparently made by DS Lewis, without his having been aware of that happening. In any event we accept the submission made by counsel for the Appellants that DI Lewis would now be liable to be cross-examined about his part in the Welsh Bomb Trial and about how it could have occurred that additional words appeared in the typed copy of notes made by him.”

The Court of Appeal were explicit that they did not conclude that DI Lewis’s evidence about the conversation must be false. They did conclude that cross-examination of DI Lewis in the light of the information by then available would be “much more effective than the cross-examination he faced at the Appellant’s trial”. They also observed that the fact that the note of the conversation was missing did not overly concern them, given that the conversation was put to Mr Sherwood and to Mr O’Brien so soon after it took place, in the presence of their solicitors:

“Who no doubt had the opportunity to examine the claim form on which the record of the conversation was written. That record was produced to Defence Counsel at the Appellant’s trial and examined by them. The document was not made an exhibit. It is in our view precisely the kind of document that is easily mislaid.”

So matters stood after the appeal.

A file was prepared and submitted to the Crown Prosecution Service in 2000, following the Appeal, to consider whether any criminal offences had been committed. The evidence was reviewed and advice sought from Senior Treasury Counsel David Perry QC. The result of that review was that counsel concluded there was insufficient evidence to provide a realistic prospect of conviction in relation to any officer for any offence. This consideration included the evidence then available as to the overheard cell conversation.

For the purpose of the civil proceedings, Mr O’Brien made a witness statement of some 80 pages in length, dated March 2006. In any trial of DI Lewis, Mr O’Brien would be an essential witness and would be liable to be cross-examined on the basis of this witness statement. The account of events in Mr O’Brien’s witness statement acknowledges that some conversation took place between himself and Mr Sherwood at the relevant time. Indeed, Mr O’Brien conceded that he did use some of the language claimed by DI Lewis, such as “I can’t handle it much longer”. Although he denied that Mr Sherwood said anything like “you’re talking about life”, he conceded that Mr Sherwood did say something like “being on remand means nothing”. The picture painted by this statement is that there was such a conversation at the relevant time, and that some of the language claimed by DI Lewis was used, but not the passages capable of supporting an inference of guilt. Thus if Mr O’Brien was to give evidence in a criminal trial consistent with this witness statement, it would be to the effect that DI Lewis elaborated on a conversation which he had genuinely overheard; and that he did so in notes which were made shortly after the conversation had taken place, shown to officers quickly thereafter, used in the two interviews within about 2 hours, in sight of the Defendants and their solicitors, and were subsequently available for the inspection of counsel at the trial and if so desired, to be the basis of cross-examination.

DI Lewis also made a very full witness statement in the civil case, dated 20 April 2006. The date of this statement is of some significance, in that it was signed before Mr Lewis was aware of any expert forensic or linguistic evidence bearing on the timing (and thus veracity) of the recorded conversation. The statement begins by saying that DI Lewis went to the cells complex intending speak to the custody officer but, whilst there, he could hear the two men talking in the cells and went to listen. The statement goes on as follows:

“113.

Michael O’Brien and Ellis Sherwood were in separate cells which were not adjacent to each other, but rather there was a gap in between. They were talking to each other by speaking between the opening in the cell doors. They were speaking louder than a normal conversational level in order to ensure that each other could hear the other, but they were not shouting. Due to the passage in time I cannot now recall the manner in which the conversation appeared to have started. However I do recall that there were pauses in between what was being said and the conversation was certainly not one in which things were said in a continuous flow.

114.

The record I made was an accurate one. I am clear in my recollection that I did not expand on the text of what was said although I may have added the time and date of the conversation to the note. I would have used personal shorthand when making the record. For example, I would probably have put “O” or “M” for O’Brien and “S” or “E” for Sherwood, or something similar, I would have also shortened other words when making the record but I cannot now say which words these might have been or how they were abbreviated. I can confirm that I did hear the conversation that I recorded. I am satisfied that I recorded the conversation as accurately as I could in the circumstances but I would have to accept that it is possible that the text might not be word for word accurate. Had there been any inaccuracy it would have been due to an honest mistake. I do not believe that any inaccuracies are likely to be significant.

115.

I made the following note of the conversation on a green expense form:

[The conversation is reproduced]

116.

I never considered this cell conversation to be a full confession by either Mr O’Brien or Mr Sherwood and I was surprised that they denied parts of it later in interview.”

In this account of events, DI Lewis claims that the record was “accurate”, but does not use the term “verbatim”. It is also evident that his account was and would be that the notes he made contained abbreviations, not merely for the names, but for other words in the record. Unsurprisingly perhaps, in the absence of the manuscript notes, DI Lewis did not claim to remember the number and degree of abbreviations.

For the purposes of the civil claim, evidence was obtained from two forensic scientists. Dr Hardcastle reported for Mr O’Brien in February 2002 and Dr Day for the South Wales Police in June 2003. Their conclusions were subsequently summarised in a joint memorandum of July 2003. They agreed that:

“1.

In normal conversation, words are spoken at a much faster rate than they can be written down in longhand. In scenario of a police officer attempting to record an overhead conversation he must resort to some form of abbreviation. Whole words may be omitted, words may be abbreviated and the writing may be slurred so as to keep pace with the speech however, with such a short conversation as that in question, it would be possible to record key parts or phrases and then to insert from memory afterwards those parts that were omitted with reasonable if not perfect accuracy.

2.

In the absence of the original handwritten note …….it is not known how many characters were written or how legible they were. In this context we know that the writing of DI Lewis, the note taken, within the custody record entry …..is blurred to the point of illegibility.

3.

The available relevant research studies concerning the maximum speeds of writing ……relate to legible handwriting and to periods of writing generally in access of 30 minutes. It is expected higher speeds would be attainable over very short spaces of time.

4.

The full written version of the recorded conversation contains 551 characters when the full names SHERWOOD and O’BRIEN are included, or 467 characters if they are omitted. A suggested abbreviated version of the text …..contains 400 characters when the initials OB and SH are included or 376 characters if they are omitted. Any further abbreviation would obviously reduce the character count accordingly.

5.

If the characters written were slurred to a degree comparable with that seen in the custody record entry written by DI Lewis, then the character counts ….would be adjusted to 331, 280, 240 and 226 characters respectively.

6.

The absence of the original handwritten notes coupled with the difficulties in estimating the achievable speeds of writing with short passages of text prevent an accurate determination of the minimum time required to write the notes …..at a very fast speed this could be as little as 1.5 minutes or possibly even less.

…….

8.

In view of the many uncertainties involved it is not possible to say whether the record of conversation given in the typed statement of DI Lewis is accurate or not.”

Linguistic experts were instructed, Professor Hoey on behalf of Mr O’Brien and Professor Coulthard on behalf of the South Wales Police. Professor Hoey’s report of May 2006 focused on the pattern of conversation. His principal conclusion was based on the proposition that conversation is made up of “adjacency pairs”, where one utterance requires and expects an appropriate response in the reply, and he concluded that the recorded conversation was not in an expected pattern. His conclusion, to a high level of confidence, was that the transcription did not represent a verbatim record.

Professor Coulthard reported in July 2006, with sight of the report from Professor Hoey. He agreed, following Professor Hoey’s reasoning, that the transcription does not represent a verbatim record. He added that “there are no linguistic grounds to deny that the reported conversation is in the main ‘one that may have taken place’”. It appears to have been Professor Coulthard who first focussed on the timing of the conversation, and whether DI Lewis could possibly have captured in substance all of what was said.

It will be recalled that the forensic scientists concluded that the conversation required at least a minute and a half to transcribe, making all favourable assumptions to DI Lewis. The experts refined that conclusion to give a minimum time of 85 seconds. Professor Coulthard read aloud the recorded conversation in some 35 seconds, leaving 55 seconds of “non speaking time to be accounted for”. He went on to consider how pauses might be distributed, to account for the time needed to transcribe the conversation.

He concluded that:

“A much more likely explanation than speaker pausing for the discrepancy between the amount of time that must have elapsed during the writing down and the amount of speech recorded, is that DI Lewis would have simply been unable to record all of what was said. In other words less important items would have been omitted – this is exactly what happened when I conducted a real-time transcription test in order to prepare a report for the appeal of …….. in 2002.”

Professor Coulthard went on to restate that the record could not be verbatim, but to observe that DI Lewis had not claimed it was verbatim:

“What should concern us more is whether the record is as accurate as it could be.”

These two experts then considered their conclusions in discussion. They agreed that the pattern of language recorded demonstrates either that Mr O’Brien was “a malfunctioning conversationalist” or that there were problems with the transcription of the conversation. Their joint note went on as follows:

“3.4

Pauses

Professor Hoey fully accepts Professor Coulthard’s observations on pausing. We note that the most likely explanation of the discrepancy between the amount of time that must have elapsed during the writing down and the amount of speech recorded is that DI Lewis was unable to record all of what was said. We also note that this explanation does not fully tally with DI Lewis’s own report of the circumstances of transcription.”

The experts went on to agree that the pattern of language raised questions about the reliability of the transcription as evidence.

That joint note was prepared in August 2006. In October 2006 Professor Coulthard prepared a second report. This report sets out the outcome of some fresh experimentation by Professor Coulthard. It is not necessary to recapitulate the experiments here, but they consisted of asking academic linguists to write down contemporaneous long hand notes of a conversation lasting 162 seconds in which 383 words were spoken. The results were abbreviated and very imperfect. Professor Coulthard cites the experiment as demonstrating that DI Lewis’s record is not:

“anything like a transcription of the conversation and so cannot easily carry over concepts derived from transcribed texts.”

Professor Coulthard observed that:

“Despite what DI Lewis himself might believe, his record could never be a verbatim account, in the sense in which I understand verbatim, but at best an honest attempt to record as accurately as possible some of what was said.”

It was shortly after the further report from Professor Coulthard that the civil claim was settled.

Following the settlement, after a considerable period a fresh police investigation was opened into the question of possible offending by former DCI Lewis. Officers confirmed they had seen DI Lewis with the notes of the overheard conversation before they were put to the suspects in interview. Mr O’Brien was video interviewed on 3 February 2012 and reiterated his previous accounts. He accepted his memory was no longer as good as it had been, given the passage of time, but maintained his complaint of fabrication. Mr Sherwood was video interviewed on 27 February 2012. He too maintained that the incriminating comments noted by DI Lewis were fabricated. Mr Sherwood had suffered a stroke with significant consequences: he had difficulty in word retrieval and expressing what he wanted to say.

The Crown Prosecution Service caseworker Mark Salter, who made the notes of evidence in the case, was interviewed. He was asked whether he could recall if DI Lewis had described the record of the speech as verbatim. His reply was:

“It is more likely than not that Stuart Lewis would have used the word verbatim as he goes on to clarify that there were substantial pauses before each reply…..this is only an assumption upon my note taking style and not a clear recollection of the evidence given and I therefore cannot say with any certainty this is an accurate record of his reply.”

In the course of the investigation, in response to a request from Mr O’Brien’s solicitors, police asked Professor Hoey to carry out a further transcription experiment. In fact he conducted two versions of the experiment. In each case the text as recorded by DI Lewis was read to groups of students, who were asked to transcribe what they heard as accurately as possible. In one version of the experiment there were no pauses. In the second there were 5 to 10 second pauses after three of the exchanges. In the experiment without pauses, none of the transcripts was complete and only one was coherent. Professor Hoey did note that the students did not pick up the “acceptance markers in adjacency pairs”. It is remarkable therefore, that the very feature which had caused Professor Hoey to be sceptical of this record in the first place was reproduced in his experiment.

In the second experiment, where pauses were added, Professor Hoey found that 3 of 8 transcripts were coherent and caught the gist of the conversation, although replicating less than half of the clauses which had been read out. The most successful student recorded 50% of the dialogue.

The repetitions of similar phrasing in DI Lewis’s text were not reproduced although the best student participant did include parts of each.

Professor Hoey’s conclusions were as follows:

“Accurate transcription in real time was not possible, even with significant pauses. There were coherent transcripts in three cases in the second experiment and: “the data supported the view that a coherent transcript was possible. Therefore if it was accepted that there were significant pauses between the exchanges, then it follows that this transcript may be a coherent selection of what DI Lewis heard.”

In further comment, Professor Hoey considered whether DI Lewis’s more than 20 years experience as a police officer, with a detailed knowledge of the case, might enhance his capacity to record accurately. Professor Hoey did not consider that the ability to record naturally occurring speech could be enhanced by training: he knew of no research which suggested this. He did consider that an experienced police officer would attend to relevant material which might affect his choice of what to record, although doubting whether the language recorded in this transcript was particularly relevant. Professor Hoey went on to conclude:

“It is therefore possible that the officer could have performed better than any informant in my sample ……….I do not believe however that the conversation the officer heard could have been recorded in its entirety with near complete accuracy unless there were frequent long pauses between utterances, and I feel I can state this to the required criminal standard.”

When Professor Coulthard was asked to consider these experiments, he had a number of reservations about them. Although the experiment was well administered and analysed, no account was taken of background knowledge, no member of either group was primed as to the relevance of the material, and although it was known that DI Lewis claimed to have used a form of personal shorthand, no one had been included in Professor Hoey’s experiments who had such skills. Professor Coulthard went on to analyse the effects of priming, considering that DI Lewis might have been attuned to hearing, mishearing or misinterpreting incriminating material. He concluded that although it was not possible to record the conversation in the terms which DI Lewis originally claimed:

“It does not follow that DI Lewis is lying. Ordinary people do have an enhanced confidence in their ability to accurately remember, record and later report what was said: as a consequence the word “verbatim” is used more loosely than by professional linguists.”

Professor Coulthard considered that there were no linguistic grounds to deny that much of the conversation may have taken place. He stated:

“There is no way in which a linguist can conclude the transcript is fabricated……it is possible to conclude, to the required criminal standard, that the transcript cannot represent a complete or even an accurate record of the wording of the conversation……it cannot be a verbatim record …..but at best an honest attempt to record significant parts of what was said.”

DI Lewis was interviewed under caution on 22 March 2012. As the prosecutor’s Review Note subsequently recorded, he made a number of observations of significance. He conceded that the note taking took place in difficult circumstances and said:

“I would strive at all times to obtain an accurate and verbatim account but you cannot ……..rule out the possibility words or text may be missing. I would not say it was a verbatim note……I will say it was an honest attempt by me to record what they were saying and certainly the whole sense of what they were saying was recorded.”

DI Lewis made the point that, when taking notes in this way, there was often not enough time to note down the question and the answer, and often what is recorded incorporates both together. Confronted by the gist of the expert evidence he responded by saying “the cells note is probably not word for word accurate.”

DI Lewis was tasked on what meaning he gave to the word “verbatim”. He confirmed that he believed it to mean “word for word”, but went on to say that he might use the term verbatim interchangeably with the word “contemporaneous”.

DI Lewis continued to deny any question of fabrication of evidence.

The decision-maker on behalf of the Defendant was Alison Storey, specialist prosecutor with the Special Crime and Counter Terrorism Division. She completed her review on 16 November 2012. The decision was communicated by letter of 19 November. The Review is a sizeable document of 54 pages with annexes. Ms Storey incorporates a full review of the material in her Note.

Ms Storey noted that the Court of Appeal had not concluded that the evidence of DI Lewis was false. She noted that the file submitted in 2000 had led to the conclusion there was no realistic prospect of conviction. She analysed the background of Stuart Lewis and the successive accounts he had given of events. She had available both the CPS notes and the summing up from the trial.

She reviewed the successive statements by Mr O’Brien, noting that he had said:

“I don’t recall any pauses whatsoever. Not that I recall anyway.”

Mr O’Brien claimed that the handwritten notes of the conversation were never produced in court whilst he was present, in conflict with the court record.

Ms Storey looked closely at the transcripts of Mr O’Brien’s interviews in 2012, given in the latest police investigation. She quoted in her report quite extensive passages wherein Mr O’Brien made repeated utterances of the same phrase. The significance of this in her thinking was the comment by the linguistic experts on repetition of similar phrases in the cell conversation.

She noted a conflict between Mr O’Brien and Mr Sherwood. Mr O’Brien’s account was to the effect that some of the recorded conversation had taken place, although expanded and embellished by DI Lewis to render the meaning incriminating. Whereas Mr Sherwood:

“Denied that the conversation as reported took place. He claimed that the only thing said was “someone’s listening, I’ll catch you later.””

Ms Storey laid emphasis on Mr Sherwood’s stroke and his incapacity to give evidence easily.

After noting the other material gathered in the investigation, and considering the documentary evidence from the time of the arrest including the custody record, she analysed in some detail the successive accounts of DI Lewis. She laid emphasis on the fact that Lewis accepted that the text might not be word for word accurate, before the expert evidence was available to him, quoting Lewis’s Civil Evidence Act statement from 2006.

Ms Storey went on to consider the expert evidence which she reviewed extensively, including the experiments conducted by Professor Hoey after the conclusion of the civil proceedings and the response to that material from Professor Coulthard.

Ms Storey reviewed the law of perjury, perverting the course of justice and misconduct in public office. She then moved to the Evidential Stage of the Full Code Test for crown prosecutors. This requires a prosecutor to be satisfied that there is sufficient evidence to provide a realistic prospect of conviction, considering the likely defence case and its effect. The judgement as to a realistic prospect of conviction:

“is an objective test based solely upon the prosecutor’s assessment of the evidence and any information ……about the defence that might be put forward by the suspect. The prosecutor must consider whether the evidence can be used in court, whether it is reliable and will be so regarded by a reasonable jury properly directed.”

Ms Storey began by considering no one had been convicted of the murder of Mr Saunders and that:

“If charges are laid against Lewis, all of the circumstances of the case are liable to be adduced in evidence. There is a very real likelihood that any trial of Lewis would involve a defence which amounted to a retrial of the original murder…….Lewis would be entitled to rebut any suggestion that the conversation was fabricated to bolster a weak case by implicating O’Brien and Sherwood. Since this is essentially a word against word case then the credibility of O’Brien and Sherwood is crucial this evidence would have the ability to severely undermine their credibility.”

She laid emphasis on the potential inconsistencies in Mr O’Brien’s accounts and in particular on the problem (which Mr O’Brien accepts) that some of the things in the record were said which Mr Sherwood rejected: she concluded this represents a “large inconsistency between the evidence of the two”. Ms Storey noted that the evidence of police officers that DI Lewis produced the note to them very soon after it had been taken:

“reduced further the time available to Lewis to compile the note, if it was fabricated after the actual conversation been noted, before making a note on the custody record.”

The prosecutor considered the evidence from others present on the night of the incident but none provided any support to the allegation.

Analysing the sequence of events, the prosecutor considered that the only explanation for a note of a part fabricated conversation would be that DI Lewis left the area and rewrote the note, using genuine material noted at the time but incorporating additional fabricated comments. She commented that there was little time available for this. She further commented that had the matter been a fabrication it was surprising that the content of the transcript was not more incriminating than it was. She noted that the motive for fabrication was not clear, given that the other evidence already present in the case constituted very strong evidence against Mr O’Brien and Mr Sherwood:

“It was not a weak case in need of bolstering.”

The decision to charge both prisoners had already been made, according to the relevant custody record, since “processing” began at 20.10.

Ms Storey noted a further piece of evidence which she considered severely undermined the prospects of successful prosecution. A Detective Sergeant Stephen was contacted by the prison authorities whilst Mr O’Brien and Mr Sherwood were in custody, recording an altercation between Mr O’Brien and Mr Sherwood. In the course of this Mr O’Brien was heard to say to Mr Sherwood “it’s all your fault I’m going to have to say something soon”. The prosecutor observed that “this has the same tenor as the remarks noted by DI Lewis and provides powerful support for his account”.

The prosecutor also noted that the credibility of Mr O’Brien and Mr Sherwood would be liable to attack, both as to their criminal activities and because they admitted lying extensively in the course of the murder investigation.

At the end of her review of the evidence, apart from that of the experts, Ms Storey concluded that she could see no evidence to support the allegations of Mr O’Brien and Mr Sherwood. Further, their allegations were not fully consistent with each other. The prosecutor could see some evidence to support the account of DI Lewis, which would tend to undermine any prosecution. In that context she turned to consider the expert evidence.

Before we turn to her further analysis, we note that Ms Williams QC accepts the proposition that, without the expert evidence, the judgement that prosecution did not satisfy the evidential test could not be challenged. Ms Williams suggests that Ms Storey overstated the difficulties somewhat, a proposition with which we agree in one respect, as appears later in this judgment. However, any difference of emphasis between the position of Ms Williams and the analysis of the prosecutor on the non-expert evidence is in truth academic, given the concession that has been made.

The prosecutor set out a fairly full analysis of the expert evidence. In a passage criticised by Ms Williams, she recited part of the conclusions of Professor Coulthard as containing the suggestion that:

“he would have been surprised if a trained detective with 20 years experience, primed for this case, had not performed better [than the participants in Professor Hoey’s experiments].”

The criticism is just. Professor Coulthard’s view appears to have been that a trained detective might have done better than the best candidate in the experiment.

Based on her earlier analysis, Ms Storey emphasised that, from 2006 onwards, DI Lewis had conceded that his note might not be word for word accurate, and that in his recent interview under caution he had pointed out that the note was made under difficult circumstances. Once it is conceded the note was not verbatim, the prosecutor considered that discrepancies of time might be resolved. She concluded:

“I am compelled to accept that the note is not a verbatim note of the conversation although it could represent a coherent transcription of the conversation and includes significantly more than 50% of the content of the actual conversation.”

Ms Storey concluded there was no linguistic evidence which provided support for fabrication and to rely on the observations of Professor Coulthard that transcribers may have a high degree of confidence in the accuracy of their notes, even when they were inaccurate in fact. As the prosecutor phrased it:

“Lewis’s confidence in the accuracy of his having recorded the full conversation is a common misplaced confidence in transcribers and does not denote dishonesty or fabrication.”

Before considering the issue of bad character evidence, Ms Storey set out her reasoning as follows:

“I have therefore concluded that there is no real support from any quarter for the allegations made by O’Brien and Sherwood. I further conclude that the evidence that they provide is inconsistent and that it is undermined by the note provided by Prison Officer Stephen……..considering the evidence as a whole, including the expert evidence, I consider that the evidence as it stands goes nowhere near providing a realistic prospect of conviction for any offences against Stuart Lewis and I take the view that a jury would most certainly not be able to be sure of guilt.”

The remainder of the Review concerns the potential for the admission of bad character evidence affecting DI Lewis, and is not an aspect which has been relied on in the course of this application. For the avoidance of doubt, we consider the conclusions of counsel, recited by Ms Storey, to the effect that it would be difficult to achieve the admission of bad character evidence, were entirely reasonable.

For constitutional reasons, the courts are very slow to interfere with a prosecutorial decision. Prosecutorial independence is of great importance. The parties are agreed that the proper approach to this application is set out by the Divisional Court in R –v- Director of Public Prosecutions ex parte C [1995] 1 CAR 136. After reviewing previous authority, Kennedy LJ concluded:

“From all of those decisions it seems to me that in the context of the present case this court can be persuaded to act and only if it is demonstrated to us that the Director of Public Prosecutions acting through the Crown Prosecution Service arrived at the decision not to prosecute:

(1)

because of some unlawful policy …….

(2)

because the Director of Public Prosecutions failed to act in accordance with her own settled policy as set out in the Code, or

(3)

because the decision was perverse. It was a decision at which no reasonable prosecutor could have arrived.”

In the instant case, there is no submission of unlawful policy or a failure to act in accordance with the Code. The submission is that the decision was unreasonable or perverse. In the future, as was set out in L -v- DPP [2013] EWHL 1752 (Admin), the court will take into account the CPS victim’s Right to Review Scheme

Because of the context, the test in this case is best formulated as follows: was it unreasonable of Ms Storey to conclude that there was no realistic prospect that a reasonable jury, properly directed, would be able to be sure of guilt?

In the course of oral submissions, Miss Williams QC did add one proposed elaboration of that test. She submitted that if there were public law errors in the approach of the prosecutor, then the decision not to prosecute should be quashed and the matter returned to the Director for reconsideration, even if the court’s own conclusion was there was a low prospect of successful prosecution.

Although for reasons we express below this does not arise in this case, we are doubtful of that approach. Judicial review is a discretionary remedy and there are powerful reasons for a highly sparing exercise of the court’s power in relation to the decisions of an independent prosecutor, particularly in the light of the CPS victim’s Right to Review Scheme, applicable to all decisions after 5 June 2013. Where the court considers there is no real likelihood of successful prosecution, it will only rarely be appropriate to quash the prosecutor’s decision, even where there may be one or more detailed criticism of the reasoning which might amount to a “public law error”.

Mr O’Brien’s submissions have already been indicated in outline. Ms Williams submits the prosecutor overstated the difficulties in relation to the case before considering the expert evidence. Specifically, she criticises reliance on the perceived discrepancy between the accounts of Mr O’Brien and Mr Sherwood on what was said in the cells. Mr Sherwood’s rejection of all of the content of the note should not be carried too far. A reasonable jury would make allowance for the situation of two young men, in the case of her client a young man with no previous convictions, in the cells facing a murder charge. The powerful spontaneous reactions of both Mr O’Brien and Mr Sherwood when denying the suggestions put to them would convince a reasonable jury. The prosecutor too readily fastened on small inconsistencies. The prosecutor did not re-evaluate the credibility of the protagonists on the basis of the expert evidence.

Mr Edis submits to the contrary that the prosecutor’s assessment of the non-expert evidence was entirely realistic. Although Mr O’Brien was a young man with no convictions, his activities that evening were a pursuit of crime in the company of Mr Sherwood, and he was forced to admit he had told a sequence of lies. The lack of evidential support for Mr O’Brien’s account meant that, setting aside the expert evidence, the prospects of successful prosecution were very poor.

In assessing the case apart from the expert evidence, we do consider that Ms Storey importantly overstated one aspect of the case. The anticipation that any trial of DI Lewis would be converted into a repeated murder trial was a clear error. It should have been understood from the beginning that no judge would permit significant satellite litigation of that kind to take place. What would be likely on any trial of DI Lewis would be that the defence could legitimately explore all of the lies told by Mr O’Brien and Mr Sherwood in the course of the investigation of the murder. The relevant issue would be the credibility of those witnesses against DI Lewis. That evidence is clearly separate from the question of their guilt or innocence of murder.

That issue aside, we regard the judgements of Ms Storey about the non-expert part of the potential case against DI Lewis as perfectly reasonable.

We turn to the expert case. The central criticism by Ms Williams is that, properly understood, the expert evidence is capable of making a jury sure. The joint views of the two forensic experts, plus their linguistic colleagues, make it clear that a jury could properly be sure that the note was a fabrication. The central proposition is that the minimum time taken to write the note (85 seconds) is far too long for the time taken to conduct the conversation (35 seconds). The expert evidence militates against significant pauses, hence there is no explanation as to how the time was consumed. This points to fabrication.

Ms Williams also emphasises what she describes as DI Lewis’s shifting account as to whether the note was verbatim or not: she says DI Lewis’s position in the interview under caution was significantly different from any account he gave before he learned of the expert evidence. The prosecutor misunderstood the importance of a meaning of “priming” and overstated Professor Coulthard’s view as to whether a trained police officer could record more than 50% of conversation.

Mr Edis argued that the prosecutor’s understanding of Professor Coulthard’s evidence was perfectly reasonable. In addition he argued that at any trial the defence would mount an attack on the admissibility of the linguistics experts’ evidence on the issue of pauses. The defence would be “taking scissors” to the evidence. Why was the existence of pauses in the course of a conversation a matter for expert evidence? In any event, the expert evidence was over-determined. A jury would reasonably consider that DI Lewis might be writing during pauses, and after the conclusion of the conversation. They would be considering both whether the note could be verbatim and whether it was accurate. Neither of the experts was prepared to say the note was fabricated, or that their expert evidence amounted to the basis for such a conclusion. The prosecutor was right to emphasise that Professor Coulthard’s conclusion in his second report restated that of his first:

“There are no serious linguistic objections that can be raised to the majority of the utterances in the note.”

Mr Edis also submits that the prosecutor was right to place emphasis on the belief of the participants in Professor Hoey’s experiments, that they had done well in recording a truthful account of the conversation in the experiment. Mr Edis’s oral submission was that the prosecutor’s conclusions were not only reasonable and defensible, but correct.

We have expressed our view as to the case apart from the expert evidence. We consider there is one valid detailed criticism of the prosecutor’s consideration of the expert case. Paragraph 6.59 does overstate Professor Coulthard’s view. He thought it was possible that a trained and experienced detective might perform better than the best of the participants in the experiment, but it was an overstatement to suggest he would have been surprised if that had not happened. However, that is a detail. We consider it very likely that a jury would be unsure whether DI Lewis had claimed from the beginning his note was word for word accurate. Although the term “verbatim” was used by the judge in the summing up, no doubt accurately based on DI Lewis’s evidence, in the light of the subsequent series of statements and accounts from him, particularly in reference to the confusion between “verbatim” and “contemporaneous”, it is unlikely that a reasonable jury could be sure he meant “word for word” accurate. The evidence that recorders of speech are over confident of their own accuracy would sound with the jury.

We also consider that it is highly doubtful that the existence of pauses or otherwise in a conversation is properly the subject of expert evidence. On what experimental or other proper basis can an expert say whether or not two young men, no doubt in a highly anxious state facing a murder charge and sitting in the cells, would or would not pause for some seconds in the course of such a conversation? On the evidence of DI Lewis, they concluded their conversation because they thought someone was listening. Even setting aside all the other difficulties identified with the expert evidence, it seems to us that is fatal to Mr O’Brien’s position. Evidence against the existence of pauses is doubtfully admissible and weak if admitted.

For these reasons, we conclude that the decision of the prosecutor in this case was not only reasonable but correct. The application for judicial review is dismissed.