Crown Court for Northern Ireland Decisions

Neutral Citation No:  [2024] NICC 10       Judgment: approved by the court for handing down (subject to editorial corrections)*

Ref:                OHA12470                        ICOS No:        14/126080   Delivered:     21/03/2024

 

 

IN THE CROWN COURT IN NORTHERN IRELAND

SITTING AT LAGANSIDE COURTHOUSE

___________

 

THE KING

 

v

 

HENRY FITZSIMMONS

and

COLIN DUFFY

___________

 

Mr C Murphy KC with Mr S Magee KC & Mr D Russell KC (instructed by the PPS) for the Crown
Ms E McDermott KC with Mr Jonpaul Shields (instructed by Breen Rankin Lenzi, Solicitors) for the defendant Fitzsimmons

Mr M Mulholland KC with Mr J O’Keefe KC (instructed by Phoenix Law, Solicitors) for the defendant Duffy

___________

 

O’HARA J

 

Introduction

 

[1]       On the evening of 5 December 2013 at approximately 19:00 a convoy of three police vehicles was driving along the Crumlin Road, Belfast, in the direction of Twaddell Avenue.  A camp had been established there as part of a loyalist protest against a decision on marching taken some months earlier by the Parades Commission.  As the convoy moved northwards out of the city it came under attack.  The vehicles did not stop when shots were fired until they reached Woodvale Road where the vehicles were inspected. 

 

[2]       That inspection revealed a possible strike mark on the first vehicle, a liveried armoured Land Rover.  The second vehicle was a Shogun towing a communications trailer which showed damage consistent with a gun attack.  The clearest evidence of the attack was found on a Volkswagen Transporter (the third vehicle), which was also towing a communications trailer.  Bullet fragments were found in the Transporter showing that bullets had struck and penetrated it. 

 

[3]       At around the same time police in another vehicle in Ardoyne saw a car on fire at Butler Place.  After the Fire Service and police reinforcements were called to the scene a burnt AKM assault rifle was found in the front passenger seat of the vehicle.  The following morning a second weapon was found nearby in a flower bed.  It was an AK47 rifle in working condition and had been made in Romania though its cartridges had been produced in Yugoslavia in 1982.

 

[4]       Close to this part of Ardoyne stands a high brick wall running alongside the Crumlin Road.  The police found a trestle scaffold against the wall.  Fourteen spent shell casings were recovered from the grass area around the scaffold.  Forensic analysis showed that they had come from the Yugoslavian cartridges of the AK47 found in the flower bed.  A person standing on the trestle scaffold would have a view over the high wall onto the Crumlin Road, the spot along which the police convoy had been travelling.

 

[5]       The burnt out AKM assault rifle was forensically examined.  It was found that 27 cartridges in its magazine had exploded due to the heat of the fire in the car.  They too were cartridges produced in Yugoslavia in 1982.  Further analysis led to the conclusion that an attempt had been made to fire this weapon.  However, the first cartridge loaded into the rifle was defective and misfired and had not been cleared before the rifle was left in the vehicle which was then set alight. 

 

[6]       At 13:47 on 6 December a silver Lexus car entered Forest Glade, a street in Lurgan, Co Armagh.  The defendant Duffy lives at 30 Forest Glade.  The model of the Lexus was IS 200D SE.  Its arrival was caught by a camera at the junction of Forest Glade and Antrim Road in Lurgan.  The full registration of the car could not be made out, but enhancement of images disclosed a partial registration as *FZ 4*3*. 

 

[7]       At approximately 13:55 three men were recorded by the camera walking up Forest Glade, crossing the road and then disappearing briefly out of view.  Three men were then captured seconds later on another camera which was deployed at a laneway into Lurgan Park from the Antrim Road.  (Lurgan Park is a large public park owned by the local council and was also referred to in evidence as Demesne House/Park.)

 

[8]       It is the prosecution case that the Security Services had secreted audio devices at various points in Lurgan Park.  From 13:58 until 15:00 approximately on 6 December the audio devices recorded a conversation which the prosecution contend was an incriminating conversation between the two defendants and another.  The other individual was alleged to be Alex McCrory, but in his case, I gave a direction of no case to answer on 31 May 2023.

 

[9]       At 15:00 the camera shows three men, allegedly and by inference the same three men, walking down Forest Glade and turning off in the direction which would take them towards the defendant Duffy’s house. 

 

[10]     At 15:56 the Lexus car (or a Lexus car) which had entered Forest Glade at 13:47 was captured on camera coming out of Forest Glade and driving off.

 

[11]     The defendants face a number of serious criminal charges.  Fitzsimmons is charged with attempted murder on 5 December 2013 in connection with the gun attack on the police on the Crumlin Road, Belfast.  He is also charged with possession of firearms and ammunition from the same incident, the firearms being the rifles found in the burnt out car and the flower bed.  Both defendants are charged with preparing for terrorist acts, contrary to section 5 of the Terrorism Act 2005 and with directing a terrorist organisation, contrary to section 56 of the Terrorism Act 2000.  Each of them is also charged with membership of a proscribed organisation, contrary to section 11 of the 2000 Act.       

 

[12]     Critical to these charges is what was recorded on the audio devices.  On the prosecution case the recordings show that Fitzsimmons was intimately involved in the Crumlin Road gun attack and that he knew of it, approved it and, in effect, signed off on it before it took place.  If the recordings are authentic and reliable, they can or should be interpreted, says the prosecution, as revealing that Fitzsimmons had a level of knowledge about the attack which is more than enough to implicate him in it.

 

[13]     So far as the other charges are concerned the prosecution case is that the nature and detail of the conversation in Lurgan can only be interpreted as meaning that the two defendants are members of an illegal organisation (the IRA), that they are deeply involved in it to the extent of directing its activities and learning lessons from past attacks to plan future crimes. 

 

Circumstantial evidence

 

[14]     The prosecution has presented its case as one which is founded on circumstantial evidence.  As is stated in Blackstone’s Criminal Practice 2024 at F1.22:

 

“Circumstantial evidence is evidence of relevant facts, ie facts from which the existence or non-existence of facts in issue may be inferred.  It does not necessarily follow that the weight to attached to circumstantial evidence will be less than to be attached to direct evidence.  For example, the tribunal of fact is likely to attach more weight to a variety of individual items of circumstantial evidence, all of which lead to the same conclusion, than to direct evidence to the contrary coming from a witness lacking in credibility.”

 

[15]     There is nothing unusual about the Crown presenting a case depending on circumstantial evidence and the whole concept has been repeatedly considered and analysed by courts over many decades.  From these authorities the following lessons can be drawn:

 

• Circumstantial evidence works by cumulatively eliminating other possibilities.

 

• It might be likened to a rope comprised of several cords.  One strand of the cord might be insufficient to sustain the weight but three strands together may be of quite sufficient strength.

 

• There may be a combination of circumstances, not one of which would raise a reasonable conviction or more than a mere suspicion, but the whole taken together may create a strong conclusion of guilt. 

 

• While circumstantial evidence may sometimes be conclusive it must always be narrowly examined.  This must be done so that before drawing any inference of the accused’s guilt, the court is sure that there are no other co‑existing circumstances which would weaken or destroy the inference. 

 

• In this context the court must guard against distorting the facts, or the significance of the facts, to fit a certain proposition. 

 

• It must be remembered that any fact which is proved to be inconsistent with the conclusion advanced by the prosecution may be more important than all of the other facts put together.

 

[16]     In R v Courtney [2007] NICA 6 the Court of Appeal said at para.31:

 

“In a case depending on circumstantial evidence, it is essential that the evidence be dealt with as a whole because it is the overall strength or weakness of the complete case rather than the frailties or potency of individual elements by which it must be judged. A globalised approach is required not only to test the overall strength of the case but also to obtain an appropriate insight into the interdependence of the various elements of the prosecution case.”

 

[17]     It must also be borne in mind, to use the words of Treacy LJ in R v McLaughlin [2020] NICA 58 at para [23]:

 

“Fragile threads do not make a strong rope.”

 

[18]     It is within these parameters that I must analyse each piece of evidence relied on by the prosecution to prove its case.  I must be satisfied that each piece of evidence is truly relevant.  If it is not relevant, it cannot be used to form any part of the rope and if there is a piece of evidence which is inconsistent with the evidence advanced by the prosecution, the rope may be fatally weakened.

 

[19]     Ultimately, the question for me will be whether I am satisfied beyond a reasonable doubt that the case has been proved against each defendant on each charge.  The guilt or innocence of each defendant is to be considered separately even though evidence of association between them may be a relevant matter.  And each charge has to be considered in turn and on its own specific elements. 

 

Elements of case

 

[20]     I will now turn to deal one by one with the strands of evidence which the prosecution rely on to prove the case against each defendant.  As will appear below there is inevitably an overlap between the strands but it is necessary to identify what they are and how far they advance the prosecution case, if they advance it at all.

 

(i)        Phone evidence

 

[21]     A mobile phone with the number ending in … 2339 is attributed to Fitzsimmons.  That attribution is not challenged on his behalf although it is pointed out that when police arrested him on 15 December 2013 the phone was found in McCrory’s Audi car which was parked outside Fitzsimmons’ house.

 

[22]     The undisputed prosecution evidence proves that on 6 December at 13:15 a call was made using 2339 while the phone was in the Whiterock area of Belfast.  There were then no other calls made using that phone until 16:07 when the phone was within the range of the Turmoyra/Lurgan masts.  Further cell site analysis proves that the phone was used after 16:07 as the individual using it drove or was driven along the M1 towards Belfast from the Lurgan area.  It was not, however, in use during the time when the Lexus car alleged by the prosecution to belong to Fitzsimmons was seen at 13:47 entering Forest Glade in Lurgan and 15:56 when the Lexus is said to have left Forest Glade. 

 

[23]     On behalf of Fitzsimmons this phone evidence is not challenged.  There is, however, an issue about the identification of Fitzsimmons’ Lexus and whether his Lexus was, in fact, the one seen entering and leaving Forest Glade on the afternoon of 6 December.  For Fitzsimmons it was submitted that if the Lexus in Forest Glade cannot be proved to belong to Fitzsimmons, the court should not be satisfied of anything beyond the fact that his Lexus was, indeed, in the Lurgan area on the afternoon in question, coincidentally during the times which the prosecution say are relevant.  On this approach the analysis of the phone records does not advance the prosecution case.

 

[24]     It was further submitted for Fitzsimmons that even if his Lexus and his phone were in Forest Glade, that is not enough to place him in the laneway as one of the men who were discussing terrorist matters.  In this context it is emphasised that there is no evidence (save for the disputed audio) that he himself was in Lurgan or that if he was, he ever got out of the car.  Specifically, there is no identification evidence from any witness for Fitzsimmons to answer.

 

(ii)       Lexus PFZ 4638

 

[25]     It is not disputed that on 6 December 2013 Fitzsimmons was the owner of a silver coloured Lexus PFZ 4638.  The movements of that vehicle on 6 December were confirmed by analysis of Automatic Number Plate Recognition (“ANPR”).  This showed that the Lexus was seen on five occasions between 13:30 and 16:40. At approximately 13:30 it was seen going west on the M1 at Sprucefield.  Then at 13:42 it was seen at Lurgan/Lough Road going south east towards Lurgan. 

 

[26]     At 16:04 it was again seen at Lurgan/Lough Road, this time going north west towards the M1.  At 16:17 it was seen at Sprucefield going east on the M1 towards Belfast.  Finally, at 16:40 it was seen at New Barnsley in West Belfast.

 

[27]     There is no dispute that this proves that Fitzsimmons’ car was seen being driven towards and into the Lurgan area before the three men were seen in the park and away from Lurgan towards and into Belfast after the men had left the park.  It is also accepted that the movements of the car tie in with the cell site analysis for Fitzsimmons’ phone. 

 

[28]     What is disputed is whether this Lexus was the model seen driving into Forest Glade at 13:47 and driving out of Forest Glade at about 15:56.  Although these timings are consistent with the ANPR evidence and the cell site analysis, it was contended for Fitzsimmons that the evidence of an expert witness called by the prosecution proves that the car in Forest Glade was a different model of Lexus or at least cannot be presumed to have been the same model.  The witness in question was a Mr Lavery, a sales manager at Lexus in Belfast.  He was asked whether the Lexus caught on video evidence entering and leaving Forest Glade was a Lexus IS 220D SE, the make of Fitzsimmons’ vehicle.  For this purpose Mr Lavery was shown an album of photographs which are not of the highest quality.  In fact, it was put to him in cross-examination that the photographs were quite blurred.

 

[29]     Mr Lavery identified the car in the photographs as being an IS 220D model.  He was able to say that it was an IS from the style of the vehicle and a 220 because it only had one exhaust at the rear.  He then went on to say it was an SEI model and to explain that the differences between an SE and an SEI would be “slight differences in the specification in the interior, more so than exterior.”

 

[30]     In cross-examination by Mr Shields, he was asked about the difference between an SE and SEI.  He said the SE was an older model and that the SEI was a “face-lifted model” with “slightly smoother tail lights.”  He then said that “from what I can see, the one in Forest Glade was an SEI.”  Ultimately, in answer to a question from me, he said “it could be either model”, meaning either an SE or an SEI.  He expressed no doubt that it was an IS or that it was a 220.  The only ambivalence was whether it was an SE or SEI, the latter being “slightly different” and more so in the interior than exterior.

 

[31]     What is not in dispute is that the available images show that insofar as the registration number can be made out in the imperfect photographs, that number is “*FZ 4*3*.”  That coincides with the registration of Fitzsimmons’ Lexus to the striking extent that it has four of the same letters and digits and in the same sequence.

 

[32]     For Fitzsimmons it was submitted that Mr Lavery’s evidence proves that the Lexus in Forest Glade was an SEI model rather than an SE.  It was further submitted that at the very least Mr Lavery’s evidence left doubt about the model which means that critically there is no proof that Fitzsimmons’ car was ever near Forest Glade.  All that can be proved, it was submitted, by ANPR and cell site analysis is that Fitzsimmons’ car and phone were in the general Lurgan area.

 

[33]     I do not accept that submission.  I do not interpret Mr Lavery’s evidence as being as categorical as has been submitted for Fitzsimmons.  I accept that he left open the question as to the precise model, suggesting it could be one of two, but when one adds the very striking coincidence of the elements of the registration number it seems to me that there could be no doubt about the vehicle which was driven into Forest Glade.  On the face of it, it would be a huge coincidence if there was another Lexus in the Lurgan area within the same timescales as Fitzsimmons’ Lexus, looking so similar to his and with four of the correct letters and digits in the same sequence as in his registration plate.  To that I add the fact that the movements of the car to and from Lurgan coincide with the use of Fitzsimmons’ phone, referred to above.  From this I am satisfied that Fitzsimmons’ car and phone moved from Belfast to Lurgan as established by ANPR and cell site analysis and that they were in Forest Glade between 13:47 and 15:50. 

 

(iii)     Use by Fitzimmons of the Lexus

 

[34]     It is submitted on behalf of Fitzsimmons that he is not identified by any witness as being the driver of his own Lexus on 6 December.  That is correct.  However, it is relevant to note that in the Statement of Agreed Facts No.3, it was confirmed that he was seen driving it on 2 November 2013, 7 November 2013 and 16 November 2013.  It was registered to him from 23 September 2013 to 9 December 2013.  From this it can be inferred that he owned the vehicle and drove it regularly until shortly after 6 December. 

 

(iv)      Association between Fitzsimmons and Duffy

 

[35]     One of the strands of evidence relied on by the prosecution is that there is an established connection between Fitzsimmons and Duffy.  The point advanced here is that if the prosecution can prove that they are known to each other, it arguably increases the likelihood of them being two of the three men seen in the park in Lurgan on 6 December.

 

[36]     The association between them over a period of time is established in the Statement of Agreed Facts No.2, in which the following is accepted by them:

 

• On 6 May 2011, at approximately 22:00 Duffy and Fitzsimmons were present at the same location.

 

• On 13 September 2013, an Audi car was seen on the driveway of Duffy’s home in Lurgan.  Two men got into it.  When the vehicle was stopped a short time later the passenger in the vehicle was Fitzsimmons.

 

• On 22 October 2013, Duffy and Fitzsimmons were seen with two other males getting into a Mercedes parked in William Street, Lurgan. 

 

[37]     While the 2011 association is not close in time to 6 December 2013, the second and third connections on 13 September and 22 October are.

 

[38]     In addition to this, there is evidence before me that on 22 September 2013 there was video footage from Lurgan Park, essentially the same location as the footage from 6 December 2013.  In respect of that footage, five police officers asserted that they recognised Duffy and further identifications were also made in relation to Fitzsimmons.  The extent to which those identifications prove anything and can be relied on was challenged as a result of failings to follow the precise protective procedures laid down for witnesses who are shown footage to identify individuals who are regarded as suspects.  This will be referred to again below in relation to the asserted recognition of Duffy in the lane on 6 December.  At this stage, however, I note that:

 

• The fact of association between Duffy and Fitzsimmons is not in dispute, nor is the fact that that association was proximate in time to 6 December.

 

• The fact that they are known to some police officers means that such evidence is by way of recognition evidence, not identification evidence.

 

• There is no recognition or identification evidence of Fitzsimmons relating to 6 December.

 

• Only two police witnesses assert that they recognised Duffy from footage of 6 December.

 

(v)       Clothing

 

[39]     A further strand of circumstantial evidence advanced by the prosecution is the contention that clothing which is visible from video footage of the three men in the laneway on 6 December can be seen to be entirely consistent with clothing found in Duffy’s home when he was arrested, and clothing worn by Fitzsimmons on the evening of 5 December when he was in an off-licence in Belfast.

 

[40]     I can deal with this strand of evidence in brief terms.  The evidence is presented by a police officer who could not be presented as an expert witness on clothing.  Rather, he pointed to what he suggested were similarities between different items of clothing, the proposition being that I should conclude from these comparisons that the clothing was so similar as to be persuasive incriminating evidence.

 

[41]     Some of the items of clothing were produced in court as exhibits.  In my judgement they were garments which were not in any way distinctive, much less unique.  The quality of the available video footage from the laneway is less than perfect.  To illustrate that point, as already stated above, despite the fact that Fitzsimmons is clearly known to the police, no officer was called to identify him as being in Lurgan Park on 6 December and only two officers were called as being able to identify or recognise Duffy. The limited quality of the video footage makes it even more difficult to attach weight to generic clothing allegedly worn by either of the defendants.

 

[42]     In these circumstances, I do not accept that the evidence as to the clothing advances the prosecution case at all.  It does not undermine any of the other strands of evidence, but it is simply not sufficiently particular for the prosecution to be able to rely on it to advance the case.     

 

(vi)      Audio Evidence

 

[43]     I turn now to what has been the most controversial part of the evidence in this intensely contested trial.  In this context I refer to my ruling of 30 September 2022 on the voir dire and to my ruling on 31 March 2023 on applications by the then three defendants for a direction of no case to answer.  This judgment should be read in conjunction with those two earlier rulings.

 

[44]     The prosecution case has always been that listening devices were secreted along the lane in Lurgan Park at some point prior to 6 December 2013.  Those devices then recorded the conversation between three men who entered the park just before 14:00 and left approximately one hour later.  It is contended that the devices were removed from the Park on 8 December and downloaded with the result being the recordings which were produced in evidence.  For the defendants, especially Duffy, there was a wholesale attack on the audio evidence and its authenticity.  That attack failed in large part because of the evidence of a Mr Koenig from the United States - see paras [37]-[43] of my voir dire ruling.  I accepted the authenticity and reliability of the audio evidence because of his evidence that it is effectively impossible to edit or manipulate recordings of a conversation picked up by 14 separate devices.  At paras [42] and [43] I ruled as follows:

 

“[42] What Mr Koenig went on to say about the Santa Ponsa audio and the Lurgan audio is of real significance in the context of the issue about the authenticity of the Lurgan audio.  He said that one cannot take words out of one or more sentences in order to form a new but unspoken sentence - that just won’t ever sound right.  He also accepted that if, as in Lurgan, there is more than one recording device one would have to interfere with each of those which had captured a snippet or excerpt of the conversation.  Furthermore, he accepted that any attempt to manipulate or move words around is even more difficult to achieve when 14 devices have background noises or sounds which vary e.g. traffic noise.

 

[43] In a further concession, quite properly made by Mr Koenig, he said that it is virtually impossible to add into a recording or recordings words which had not actually been spoken at all by the people who were being recorded.  On his evidence interference with the start or the end of a conversation is possible to achieve by omitting the first or last parts but interference with the substance of a conversation, particularly one which is recorded across a number of devices, is close to impossible.”

 

[45]     In large part the attempt to exclude the audio evidence failed because of Mr Koenig’s evidence.  I held it to be authentic evidence at the admissibility stage and I have no reason to form any different view now, notwithstanding the closing submission on behalf of Duffy to revert to that ruling and to exclude the audio or at least not rely on it.

 

[46]     As the voir dire ruling continued, however, I ruled against the prosecution on the separate question of the admissibility of expert evidence as to voice attribution and transcripts of what was said by the men in the park.  When the experts for the prosecution were engaged, they were provided not just with the audio but also with a transcript prepared by the police setting out what the police could make out as having been said and attributions as to which of the three accused said each part.  This issue is dealt with at paras [64]-[85] of the voir dire ruling which concluded with me admitting the transcripts of the Lurgan audio but with the attributions being removed.  The attribution by the police of words spoken was always going to be inadmissible because of the ruling of the NICA in R v O’Doherty [2002] NI 263.  In my ruling I held that the attributions by the experts had been fatally tainted by reason of the fact that the experts had been briefed with the police attributions. 

 

[47]     On that basis, and with no further evidence having been called after the voir dire ruling, I next heard applications for a direction of no case to answer.  At that point issues emerged which I summarised and dealt with as follows on 31 March 2023:

 

“[5] Although I admitted the Lurgan audio in evidence, I was not satisfied on the issue of the transcripts of the recording.  I ruled that the attributions of speech to each of the three defendants were fundamentally flawed and at para [84] I admitted in evidence the transcripts of the Lurgan audio but without the attributions.

 

[6] The precise terms of that ruling became problematic during the course of the applications of no case to answer because there are a number of different transcripts, some of which are poisoned, at least to some degree, by their origins.

 

[7] Accordingly, for the purposes of this ruling, I have listened intently a number of times to the Lurgan audio itself.  By doing so, I have sought to avoid the issue that different transcripts include references to “Harry”, “Collie/Colin” and “Alec/Alex” whereas others do not.  Those references are inconsistent, but the important point is that I have relied almost exclusively on the audio alone.  The only exception to this approach is that I referred to the transcript prepared blind by Dr Philip Harrison, a course suggested to me by the defendants.  In fact it was accepted by all sides during the oral submissions that overwhelmingly the best course for me to take is to listen to the audio.”

 

[48]     In light of closing submissions now received on behalf of Duffy I emphasise that the course I took in relation to the audio was accepted by Mr Mulholland during the course of exchanges with him.  His only caveat was that I should be alert to the risk that I had been primed by the evidence to date about what was in the transcripts with the result that I might wrongly conclude that I could hear words or names or exchanges which are, in fact, unclear.

 

[49]     I bore that warning in mind when considering the direction application.  Having done so, I ruled that McCrory had no case to answer at para [21] in the following terms:

 

“[21] This prosecution has been brought significantly, though not exclusively, on the basis of the Lurgan audio.  Having listened to the audio recordings again, I have concluded that I just cannot make a confident finding that there is any reference to Mr McCrory or to his name in that audio.  That difficulty for the prosecution is insuperable and is added to by the fact that there is no identification evidence nor is there evidence of the sort referred to above in relation to Mr Fitzsimmons upon which a judge could find that he was in Lurgan on the afternoon of 6 December 2013.”

 

[50]     I did not form the same view in relation to the remaining two defendants.

 

[51]     It has been necessary to explain that history because for Duffy it has been asserted in the closing submission from para [95] that, in effect, I made a fundamental error because “the court stepped into the shoes of an ad hoc expert at the invitation of the prosecution.”  This proposition is developed in the following paragraphs including para [105] which states:

 

“The voice attribution process that the court conducted has made an attribution of a voice on poor quality audio to an accused after expert evidence had been excluded …”

 

At para [106] the submission continues:

 

“This exercise goes beyond finding a name on the audio, rather (i) it attributes this name to someone on the audio and (ii) it makes a finding that this person was the accused.  This is voice identification contrary to O’Doherty [2002] NI 263 …”

 

[52]     With respect to counsel that submission is wrong and misconceived.  I listened then, and have listened again, to the audio for two purposes:

 

• Can I confidently make out either the name Collie/Colin or the name Harry being referred to at any point?

 

• If I can do so, is the person named Collie/Colin or Harry named or referred to in context as a participant in the conversation?  It is not enough that a name is mentioned - see para [9] of my ruling on the direction application.

 

[53]     After the direction ruling on 31 March 2023, Mr Mulholland made an application for Duffy on 25 May 2023 that I should recuse myself.  The basis of that application was the real possibility that I might be subconsciously biased arising from the fact that I had simply heard so much prejudicial evidence which I had then ruled was not to be admitted as part of the trial.  I rejected that application.

 

[54]     The next stage of the case was that it was confirmed that neither defendant would give evidence.  However, for Fitzsimmons Ms McDermott confirmed that expert evidence was being obtained from a Dr Frederika Holmes who specialises in forensic speech and voice analysis.  Her report dated 14 June 2023 was engaged with for the prosecution by Dr Richard Rhodes, an expert in the same field.  His report dated 29 August 2023 prompted a reply from Dr Holmes on 21 October 2023.  After various discussions between the experts and between counsel for the prosecution and Fitzsimmons, I was presented on 24 November 2023 with a note signed by Mr Murphy and Ms McDermott setting out an agreed position that the reports of Drs Holmes and Rhodes together with a joint statement agreed by them which could be considered by the court and admitted in evidence.  The note signed by counsel included the following:

 

“All the recent expert reports from Dr Holmes and Dr Rhodes may be considered by the court and are admitted by agreement. 

 

Dr Holmes accepts in respect of extracts 1 and 2 as referred to at page 6 of her joint report with Dr Rhodes that the name “Harry” is a plausible interpretation.  In this regard Dr Holmes states that “plausible” means that this interpretation is consistent with the limited phonetic detail at this section of the recording, but it is impossible to exclude other possible interpretations. 

 

                        Dr Rhodes’ position is as set out in the joint report.

 

Ultimately, it is a matter for the court to determine what it hears, and it can derive whatever assistance it considers appropriate from the various expert reports.”

 

[55]     When this development was announced in court on 24 November 2023, Mr Mulholland for Duffy indicated that he neither consented nor objected to this approach but that the reports and agreed documents were not relevant to Duffy whose case he then closed without calling any evidence.

 

[56]     Since the reports of Dr Holmes and Dr Rhodes are the only expert evidence before me, apart from the transcript originally prepared blind by Dr Harrison, they demand some analysis. 

 

[57]     Dr Holmes was provided with three recordings - DL4, DL9 and DL13.  Her instructions in May 2023 were to provide a transcript for the first 15 minutes and the section from 40 minutes to the end of the approximately hour long recordings.  In the introduction to her first report she stated as follows:

 

“In keeping with best practice and to minimise the risks of cognitive bias, I was provided with very limited background information regarding this material.  I was told that the recording had been made via a number of microphones placed in an outdoor location and that a compilation had been prepared from those sections with the best signal strength across each of the recording devices.  I was further told that the technical aspects of the recording process and the preparation of the compilation were not in dispute and had been accepted by all parties.”

 

[58]     Later on 9 June she was asked to focus particularly on four limited sections which lasted for a total of 67 seconds.  Then on 13 June she was asked to consider specifically whether a specific name, Harry, could be found in those 67 seconds. 

 

[59]     In her report Dr Holmes provided an evaluation of the material which had been made available to her.  She said at section 4.2 that there appeared to be at least three male speakers and that the technical quality of the recording varied between fairly poor and extremely poor.  She also contended that the recorded signal was further degraded by “the presence of a high level of masking noise.”  In addition, she confirmed that the attribution of words was not a principal focus of her instructions.  She then went on to prepare a transcript of the two sections of the recording which is under scrutiny.  As a general comment it is fair to say that there are much greater gaps in what she has been able to transcribe than was the case with others who have done this work before her. 

 

[60]     On the specific issue of whether she could identify the name Harry she said as follows at section 6.3.3:

 

“Neither of my first two draft versions of the transcript included this name as a spontaneous decoding and further detailed differential analysis of the questioned sections including signal processing and slowing down of the signal showed no evidence of an utterance which was sufficiently clear as to make this name an acceptably plausible interpretation of the recorded evidence.”

That section of the report ends with the following:

 

“In the present context it is my opinion that while much of the speech content of the recording is sufficiently poor that the presence of specific lexical items or names cannot be excluded, the phonetic and acoustic evidence does not provide support for the inclusion of this specific name in the transcript.”

 

[61]     As indicated above, Dr Rhodes was engaged to respond to this.  Apart from reviewing Dr Holmes’ report he was also instructed, he says, to review her two transcripts of the first 15 minutes and the last approximately 15 minutes and “to compare this with the content of transcript JPF1.”  He then produced a transcript comparison document which presented the content of the transcript prepared by Dr Holmes from the two sections she had focused on and the transcript JPF1 in a single table.  This was done to assist me in considering the alternative interpretations provided in the two transcripts. 

 

[62]     Dr Rhodes advanced a number of conclusions which are as follows:

 

• There are widespread and significant differences between Dr Holmes’ transcript and the JPF1 transcript.

 

• In his view while some sections of the audio were poor and difficult to make out there were also clearer sections of the audio where most of the words were clear.  On this basis he thought that Dr Holmes was being too harsh by far in saying that the quality varied from fairly poor to extremely poor.

 

• He also commented that much of the clear transcribable speech was not, in fact, transcribed by Dr Holmes.

 

• Critically, he reported that he could hear the name Harry clearly at two points and probably at two further points.  He identified specifically where in the recording each of the four points was. 

 

[63]     In her response of 21 October, Dr Holmes emphasised the following at section 3.2:

 

“I was instructed to prepare an initial transcript without access to any background information regarding the material, in order to minimise the risk of cognitive bias.  This is in keeping with best practice for forensic transcription of complex audio and meant that I had no information regarding the subject matter or context of the conversation, nor any information about the number and identity of speakers who could be heard on the recording.  In these circumstances, the need to avoid introducing error is paramount and the resulting transcript will err on the side of caution on the interpretation of sections of lower clarity.”

 

[64]     She then stated the following in relation to Dr Rhodes:

 

“3.3 I understand that Dr Rhodes had been instructed in previous iterations of transcript work in this material and that some of his work had been carried out with prior access to a police transcript which included attribution of utterances between named individual speakers.”

 

[65]     That understanding on the part of Dr Holmes is entirely correct.  Dr Rhodes was not introduced in 2023 to the case as a new expert.  He had previous involvement with JP French Associates, whose work I commented on in the voir dire ruling in 2022.  To be fair to Dr Rhodes he openly acknowledged that fact at section 5.16 and 5.17 of his August report where he stated as follows:

 

“5.16 One difference in the procedure is that Dr Holmes is the only analyst who worked on her transcript whereas JPF1 was produced by a team of three analysts (Prof French, Dr Harrison, Dr Rhodes).

 

5.17 The approach of transcribing through multiple drafts by multiple experts is used because what people hear, and what they can therefore transcribe, in a recording is inherently subjective; the purpose of using multiple transcribers is (a) to reduce the degree of subjectivity by cross-checking interpretations with other experts, (b) to locate and reduce errors in content and formatting, and (c) to improve the amount of speech that can be transcribed.”

 

[66]     In my judgement, in light of my previous rulings, someone other than Dr Rhodes should have been engaged to respond to Dr Holmes.  And that someone should have been a person independent of JP French Associates or someone there without previous involvement with this case.  An equally important point is that if Dr Holmes’ transcript was to be compared to anything, it should not have been the work of JP French Associates which I have already rejected in the direction ruling.  That transcript cannot be relied upon because it may well have been, and most probably was, influenced by the transcript provided by the police and discussed with the experts as their work got under way.

 

[67]     For this reason I can only conclude that I cannot rely on Dr Rhodes’ report.  I have reservations about the very limited extent to which Dr Holmes contends that she could make out what was said on the audio but I cannot simply disregard her opinion, based on an untainted analysis, that the name Harry cannot be clearly and unambiguously heard on the recording. 

 

[68]     At the stage at which I refused a direction to Duffy and Fitzsimmons I distinguished them from McCrory because I could not confidently make out a reference to his name.  At this stage of the trial, the final stage, I must be quite satisfied that I can hear their first names clearly.

 

[69]     I still think that is a possible conclusion to draw from a careful listening to the audio.  However, I must bear in mind the rationale already quoted above from Dr Rhodes at section 5.17 of his report for not relying on what I myself hear or think that I can hear.  In effect, as a prosecution expert, he warned me against relying on Dr Holmes because she had worked alone on the transcript.  The same warning must apply to me with even greater force.  I am not working from a transcript but perhaps even more uncertainly I am trying, or the defence might argue straining, to see if I can make out the names Harry and Collie/Colin. 

 

[70]     In this context, I must also acknowledge the fact that in the table of facts presented to me on 24 November 2023 by agreement between Mr Murphy and Ms McDermott, it is confirmed that Dr Harrison of JPF Associates who started his work by ignoring the police document and preparing the transcript blind did not hear the name Harry at all.  The same applies to Dr Rhodes’ first transcript.  Even on his second transcript he mentioned Harry only once, not four times as he does now, and even then he put that name in brackets, thereby indicating a lower level of confidence in how clearly it could be made out. In other words, he regarded it as a plausible interpretation but not to the exclusion of other possibilities.

 

[71]     I turn back now to Duffy.  Mr Mulholland, having said on 24 November that the reports of Dr Holmes and Dr Rhodes were not relevant to Duffy, then enthusiastically adopted Dr Holmes’ report in his submission.  Furthermore, he highlighted the inconsistencies in the findings of the name Collie or Colin reported by the various experts at different parts in the trial process and submitted that they could not form the basis of a safe conviction. 

 

[72]     There is other evidence against Duffy including the identification by way of recognition of him by two police officers in the lane.  It was argued against that identification evidence that they had failed to follow guidelines specifically introduced to safeguard against poor quality identification/recognition.  Notwithstanding those failings, that evidence adds to the case against Duffy but is not enough to convict without the audio connection.

 

[73]     There is also evidence about Duffy’s activities on holiday in Santa Ponsa in August 2013, only a few months before the events which are critical to this trial.  Again, that is evidence which adds to the case, but not enough to convict.

 

[74]    As against both defendants I was invited to draw an adverse inference by reason of their failure to give evidence in their own defence. Had I been able to conclude that either of their names was audible on the audio, that is a course I would most probably have taken. However given the centrality of the audio to this trial, my inability to conclude with the required confidence that I can hear either name means that the question is not a live one.

 

Conclusion

 

[74]     At the start of this long trial, and throughout the period during which it has lasted, the prosecution has accepted that the audio evidence is at the heart of its case.  It has done so because unless it can prove who was walking along the lane in Lurgan Park, nobody can be connected to any of the charges.  When I excluded the evidence of attribution at the voir dire stage the prosecution case was weakened.  This is what led to the acquittal of McCrory.  However, the weight of the evidence is such that it would be unsafe for me to rely on my own subjective impression of the audio evidence in the face of the clear warnings from the experts for both prosecution and defence about relying on what I think I can make out from the audio which the experts themselves have reached different views on at different times. Those warnings apply with added force in a case in which I may be regarded as having been primed by reason of the fact that I heard incriminating evidence which I then ruled inadmissible.

 

[75]     In light of this conclusion, I cannot say that I am satisfied to the requisite standard that the audio evidence proves that either Fitzsimmons or Duffy was in the lane in Lurgan Park on the afternoon of 6 December 2013.  The reasonable and legitimate suspicions raised about their presence and their intent from other evidence are not enough to form the basis of a conviction of either of them. 

 

[76]     In these circumstances, I find both defendants not guilty on all charges.