[2024]_IESC_31 — [2024]_IESC_31 — Ireland courts

AN CHÚIRT UACHTARACH

THE SUPREME COURT

S:AP:IE 2023:000070

[2024] IESC 31

Between:

THE PEOPLE (AT THE SUIT OF THE DIRECTOR OF PUBLIC PROSECUTIONS)

Respondent

v.

WILLIAM TWOMEY

Appellant

Charleton J

O’Malley J

Woulfe J

Hogan J

Murray J

Judgment of Ms Justice Iseult O’Malley delivered the 16th day of July 2024 Introduction

1. The appellant was convicted in the Special Criminal Court of the offences of robbery,

assault occasioning harm and making a demand with menaces. He was acquitted on a

charge of false imprisonment. The prosecution did not allege that the appellant took

physical part in the offences but prosecuted him on the basis that he had been part of a

joint enterprise. The defence as put forward in the trial was primarily to the effect that

the appellant was involved only at the very beginning of the plan that culminated in the

commission, by others, of the offences in question and that he was not part of a joint

enterprise because he never intended or envisaged the events that ultimately took place.

Coupled with this, it was contended that the appellant had in any event abandoned or

withdrawn from the enterprise at a certain point before the offences were committed.

The trial court found that the appellant was part of a joint enterprise. He had been

involved for longer than he had admitted, he must have been aware of the nature of

what might happen and the events that did take place were, with the exception of one

charge before the court, within the scope of the agreement. The court further found that

the actions relied upon by the appellant as establishing abandonment or withdrawal

were insufficient for that purpose.

2. The appeal concerns the criteria applicable to the defence of abandonment or

withdrawal (either word may be used in this judgment), and the central question is

whether the appellant took sufficient steps to be entitled to avail of it. Background facts

3. All of the charges in the case stem from an incident that took place in a remote rural

part of County Louth on the 2nd December 2017. The victim was a dealer in second-

hand construction equipment named Edward McAndrew, who was from Co Mayo and

traded in both the Republic, Northern Ireland and rest of the United Kingdom. He was

lured to the location by a false pretence that he could inspect certain items of heavy

machinery that he was interested in buying at a price of €85,000. When he got there he

was severely assaulted by three masked men, two of whom were armed with metal bars,

who demanded money from him. One of the men said that they were from the

Continuity IRA. One feature of the case is that the men appeared to have expected that

Mr McAndrew would have brought money with him to pay for the machinery he

thought he was going to inspect. That expectation would not be shared by anyone with

knowledge of the trade. Mr McAndrew was imprisoned for some lengthy period of time

in the boot of a car. He was robbed of, inter alia, three phones and two laptops, and was

threatened with further assaults unless he agreed to pay the sum of €50,000 in the

coming days. He needed treatment for multiple head lacerations.

4. As noted above, it was not alleged by the prosecution that the appellant took part in the

assault or even that he was present at the scene. The essence of the case against him

was that he devised and began the implementation of the scheme that ultimately brought

Mr McAndrew to the place where the crimes took place, that he recruited the man in

charge of the group that carried out the crimes and that he received part of the benefit

of the crimes in the form of data that he wanted from Mr McAndrew’s phones. The evidence concerning his complicity is somewhat complicated, but the following

summary should suffice.

5. The appellant, a financial adviser who appears to have had financial and personal

difficulties since the financial crash of 2008, had known Mr McAndrew for about ten

years and had had personal as well as business dealings with him. They appear to have

had a falling out in 2017, some time prior to the events in question. It may be noted

here that the trial court did not find either man to be a reliable witness in relation to

either this aspect or other matters dealt with in their evidence, and indeed expressly

sought out corroboration in respect of Mr McAndrew’s account of the offences

committed against him.

6. According to the prosecution, the trouble between the two men in 2017 arose out of an

incident involving the misappropriation of a bank card and the taking of money that

belonged to an associate of Mr McAndrew in Northern Ireland. This matter had led to

the appellant being investigated by the PSNI, and he was eventually charged in August

2017 with fraud in that jurisdiction. There was evidence of an angry phone call between

the two around February 2017, recorded by Mr McAndrew, in which the appellant

uttered what might be regarded as threats to get revenge. There also appears to have

been an incident in which certain files, belonging to the appellant and containing

information about him and his family, were taken by Mr McAndrew from the

appellant’s home in an apparently forceful attempt to get the missing money returned.

7. It is noticeable that each of the two men appeared to downplay the extent of the falling

out between them when giving their own evidence – Mr McAndrew denied that he had ever had a “problem” with the appellant and said that he did not know anyone else who

had had a problem with him. However, he agreed that they had had “a bit of an old

dispute” in early 2017 and had some “words”. The dispute over the money may have

resolved at some stage – the two apparently did some normal business in August 2017.

When Mr McAndrew’s mother required hospital treatment in late December 2017, soon

after the assault on Mr McAndrew, the appellant came to the hospital and was able to

offer some assistance to Mr McAndrew. Mr McAndrew said that he had not thought

that the appellant had anything to do with the assault on him. He did not implicate him

in his evidence.

8. The Northern Ireland charges were ultimately dropped in May 2018. However, it was

apparent from evidence in the Special Criminal Court trial that while the charges were

pending the appellant had blamed Mr McAndrew for his situation. He had drafted

witness statements for various people intended to be called in his defence if the trial

had proceeded, with the apparent purpose of implicating Mr McAndrew in some form

of dishonesty in relation to the money that was the subject-matter of the charges. This

drafting process continued into 2018.

9. The appellant’s evidence was that the fraud charges were not his reason for engaging

in the scheme at the heart of this case. Rather, he said that he was motivated by the

belief that Mr McAndrew owed money to a vulnerable friend of his. (It is not necessary

to describe the circumstances in which this debt was said to have arisen, which may

have involved some business assistance having been given by the friend to Mr

McAndrew, but in truth there seems to have been little genuine basis for any significant

financial claim.) He recruited a Mr Tony Finglas to assist in obtaining repayment of this alleged debt. The appellant was at that time living in a flat in the town of Warrenpoint,

in Co Down, and Mr Finglas lived in a separate flat in the same house. He agreed to

help in return for half of whatever money he could recover.

10. The evidence in relation to Mr Finglas, given by the appellant and also by the estate

agent who had introduced him to the appellant, was that he had previously been

employed for the purpose of debt collection from “difficult” tenants. He was a man of

intimidating appearance and manner, with a number of criminal convictions, who, they

said, turned out to be more violent than they had expected. The appellant said in

evidence that in recruiting Mr Finglas he had thought that he would “annoy” or

“pressurise” Mr McAndrew into paying the debt. (Mr Finglas pleaded guilty to the

charge of demanding money with menaces, before the appellant’s trial came on for

hearing. It should perhaps be said here that in the course of the appellant’s trial, the

members of the Special Criminal Court recalled the case and stated that it was accepted

that Mr Finglas had not anticipated how violent his companions would be and had

prevented them from inflicting even worse violence on Mr McAndrew.)

11. The appellant said that his original plan was to get Mr McAndrew interested in a

potential purchase of some machinery. This would, he said, demonstrate to Mr Finglas

that Mr McAndrew had substantial amounts of money and could afford to pay the

claimed debt, so that Mr Finglas would not be deterred by assertions of inability to pay.

In order to bring about a meeting, he communicated with Mr McAndrew by email on

the 6th September 2017, sending three messages on that date that purported to offer him

various heavy machines for sale. For this purpose, he used a dormant email account that

he had previously set up using the name Alan Mooney and calling himself AlanLiverpoolFC, and he found suitable photographs and descriptions of machinery on

the internet. Those emails, which the appellant accepts were sent by him, were followed

by both further emails from the AlanLiverpoolFC account and by telephone

communications to Mr McAndrew by a man calling himself Barry. The “Barry” phone

calls were found by the trial court to have been made by Mr Finglas, and it was through

those phone calls that the arrangement for the inspection of the machinery was made.

On the 3rd December 2017, the day after the assault, a message from the same email

address threatened Mr McAndrew with further harm if he did not cooperate. A fire was

set at his house (although it is not clear how much damage was done) and he received

a threatening phone call from “Barry”. His daughter also received a call of that nature

– part of the significance of that lies in the fact that the caller must have obtained her

number from the contacts list on Mr McAndrew’s phone.

12. The appellant said in evidence that he had not sent any emails after the 6th September

but had passed control of the email account to Mr Finglas. He did, however, accept that

he had supplied further pictures of machinery that were used in emails on the 21st

October. Having regard to the evidence relating to examination of the appellant’s

laptop, the trial court was satisfied that he had in fact sent the emails on that date and

that the joint enterprise was active and progressing at that time. The court did, however,

find that there was a reasonable doubt as to whether he had sent the threatening email

of the 3rd December, because it could on the evidence have been sent by Mr Finglas.

13. The trial court did not accept that the appellant’s purpose was limited to demonstrating

that Mr McAndrew had funds available, finding that this explanation did not make

sense. The court considered that, apart from getting some money from Mr McAndrew, the evidence demonstrated that the appellant had the additional, stronger motive of

wanting to obtain information about Mr McAndrew and his business and that he had

shown a “prolonged” and “intense” interest in doing so. It took the view that his

objective in the scheme was to find material that might be of assistance to him in the

then-pending fraud trial.

14. Having noted the appellant’s acceptance that Mr Finglas would be likely to progress

the matter “in his own way”, the court was satisfied that a demand for money was

contemplated and that menaces and a degree of physical force might be resorted to. The

court noted that the appellant knew from his experience in such matters that Mr

McAndrew would not bring large sums of money with him when inspecting machinery,

so it would be necessary to make a “significant impression” on him if money was to be

obtained from him subsequently. Thus, an assault and the causing of harm was held to

have been included in the initial understanding of the enterprise, although the court was

prepared to accept that what eventuated went beyond the level of violence contemplated

by the appellant.

15. The appellant was arrested on the 28th March 2018. On being searched he was found to

have in his possession a USB stick that contained a quantity of material relating to Mr

McAndrew’s business and personal affairs. It was established by investigators that this

material had been downloaded from one of the phones stolen from Mr McAndrew on

the 2nd December 2017 and recovered from Mr Finglas on his arrest. A search of the

appellant’s flat turned up, inter alia, a handwritten list of contact details which had also

clearly been taken from one of the phones. It appears from Mr McAndrew’s evidence

that the phones were not password-protected. 16. The appellant’s evidence was that this material had been given to him on the 14th March

2018 by the friend on whose behalf he had tried to recover the debt. He said that it had

not occurred to him that there was a connection with the events of the 2nd December.

The trial court found his account wholly implausible. The technical examination of the

stolen phone from which the material was taken demonstrated that it had been wiped

on the 9th December 2018, and was used by Mr Finglas thereafter. Accordingly, the

court found that the appellant must have been given access to the material during the

week after the robbery. Further, there was reasonably frequent contact between that

phone and the appellant’s phone between the 14th January and the date of his arrest,

suggesting that the appellant and Mr Finglas were back on some sort of terms.

17. At trial, the appellant accepted in evidence that he had devised the original plan that led

to the incident on the 2nd December 2017. (The trial court also noted that he was the

only person involved who had sufficient knowledge, from his professional experience

in the area of the trade in construction equipment, to carry out the plan in a way that

would satisfy Mr McAndrew that a genuine deal was on offer.) However, he said, he

became concerned at the erratic and violent behaviour of Mr Finglas. In early

November 2017 he told Mr Finglas to “cease” on the plan to recover the debt,

whereupon Mr Finglas had assaulted and injured him to an extent requiring medical

treatment. The appellant also described other incidents in which Mr Finglas had

demanded the use of his car – on one occasion he had broken his door in to get his car

keys. 18. The appellant said that after that incident he was concerned about Mr Finglas and tried

to stay away from him. He contacted a detective garda based in Shannon to tell him that

he was concerned that there might be aggressive action towards Mr McAndrew.

D/Garda Hayes testified in the trial and confirmed that there had been such a

conversation between them at some stage in 2017. He could not remember the date, but

it appears clear that the appellant was living in Warrenpoint at the time – on the

evidence, this meant that the call was some date after August 2017 and the trial court

accepted that it was before the 2nd December. D/Garda Hayes said that, as the appellant

told him that the attack might take place in Northern Ireland, he had advised him to

contact the Police Service of Northern Ireland. The appellant said that he rang a general

number there, but the response from the PSNI was that since Mr McAndrew lived in

the South it was a matter for the gardaí. He did not contact Mr McAndrew to warn him,

because he feared what Mr McAndrew might say and to whom he might say it. He took

no further action to prevent the crime.

The verdict of the Special Criminal Court

19. The trial court accepted that the evidence indicated that the false imprisonment of Mr

McAndrew in the boot of the car was something done by other participants in the assault

and that it was not part of a tacit agreement between the appellant and Mr Finglas or

within the scope of their common design. It therefore acquitted the appellant on this

charge. 20. The court then considered the question whether the prosecution had rebutted the

defence of withdrawal from the common design covering the rest of the charges. On

this aspect, the court said:

“This is a matter to be decided on the facts of each case. What is required to

effect withdrawal and avoid criminal responsibility will generally be positive

acts proportionate to the extent of previous involvement with and contributions

of the accused to the enterprise in question. The constitutionally prescribed

objective of social order requires more than silent withdrawal and must involve

reasonable steps to reverse the criminal enterprise previously embarked upon.

The steps required may also differ depending on the spontaneity of the

participation. In our view, an effective withdrawal requires at least the

cancellation of the effects of previous participation and, if possible, prevention

of the offence contemplated by the participant.

What is necessary to withdraw will also depend on the stage of the plan at which

the withdrawal is said to have occurred. It may involve countermanding

permission or going further, if the matter is more advanced, and reporting the

matter to the authorities. Furthermore, there must be timely and unequivocal

communication of notice of withdrawal from the common purpose to others who

wish to continue with the matter. In this case Mr Twomey's participation was

not spontaneous. On the contrary, he had recruited Mr Finglas to the plan

devised by him and provided the directions, information, material and means to

Mr Finglas to set the plot in motion by setting up the ruse to entice Mr

McAndrew to a meeting at a remote location. By the 21st of October Mr Twomey

had continued to involve himself after the initial contact with Mr McAndrew by facilitating and participating in the follow-up communication by way of the

second set of emails and new photographs which were specifically directed to

keeping the joint enterprise alive and in progress. The fact that Mr Twomey had

subsequent qualms about the conduct of Mr Finglas would not be sufficient in

itself to extricate Mr Twomey.”

21. On the evidence relating to the contact with the gardaí and the PSNI, the court noted

that the appellant had not given D/Garda Hayes any detail as to his own participation

in the matter and had not identified the precise source of possible danger. He had not

taken “the first and most obvious step” to effect a full withdrawal, which would have

been to warn Mr McAndrew directly. He could at least have told him to ignore certain

emails or phone calls. Given what had occurred when he told Mr Finglas to desist, he

should have realised the extent of the risk to Mr McAndrew.

“In short, his actions were not proportionate to the task of undoing his previous

actions and of mitigating the risk that his actions had created for Mr McAndrew.

Instead, he left Mr McAndrew ultimately exposed to a continuing risk and Mr

McAndrew was, at the end of the day, obliged to pay the price of Mr Twomey's

desire to protect himself from the consequences of his actions. What he did was

insufficient to counteract, reverse and nullify the effect of his earlier activities.

It seems to us that the impression that his evidence creates is that he hoped that

his efforts would have been enough to forestall violence to Mr McAndrew,

without the inconvenience of any searching inquiry into any previous

wrongdoing on his own part. As it turned out, as a half-measure it was not in

fact effective for any of these purposes. Therefore, we find that the common

design was not effectively terminated prior to the 2nd of December 2017 and therefore remained in place at the time of the commission of the assault and the

other crimes that occurred in relation to Mr McAndrew.”

22. Having made that finding, the court also expressed the view that the appellant’s actions

in copying the material from the stolen phones in the week after the 2nd December, and

his retention of it up to the date of his arrest, were “inconsistent” with his claim of

having withdrawn before that date.

The Court of Appeal

23. The Court of Appeal delivered judgment on the 25th April 2023 (Edwards J. – see [2023]

IECA 101).

24. It is noted in the judgment that both parties relied upon the section in Charleton &

McDermott’s Criminal Law and Evidence (2nd ed., 2020, Bloomsbury Professional)

dealing with the defence of abandonment. The relevant pages are 385 – 390. In

particular, the Court of Appeal cited the following passages:

[8.83][…] To be legally effective, that withdrawal must be (1) clear and

unequivocal, (2) timely and (3) communicated either to the other parties or to

the police. This may be translated into a requirement to clearly disavow the

proposed wrong and to take reasonable steps to reverse the criminal enterprise

embarked upon. Thus it is required that the accused ‘took, in a manner

proportional to his or her participation in the commission of the planned

offence, reasonable steps in the circumstances either to neutralise or otherwise cancel out the effects of his or her participation or to prevent the commission of

the offence’”…

“[8.88][…] the first question is whether the accused was part of a common

purpose to commit the offence in issue. If that person was, but withdraws, the

second question arises as to whether the accused in fact withdrew from it and

as to whether as a matter of law the steps taken, if accepted by the jury, are

sufficient in law to amount to a negation of the offence in which he or she was

originally complicit.

[8.89] The defence of withdrawal is dependent on the facts of each case: what

was done, was it timely, was it at least potentially undermining of the scheme?

That, in turn, hinges on how far things have gone and what the accused needs

to undo:

‘What will suffice in terms of withdrawal from a joint enterprise or from

a situation in which a defendant has counselled and procured or aided

and abetted a crime will vary markedly from case to case. It will involve

an assessment of what was reasonable and practical in the

circumstances. The more the defendant has done by way of planning or

providing information or items to enable completion of the crime, the

more is likely to be required of him by way of withdrawal or

countermand, if he is to avoid criminal responsibility. In some cases,

particularly where the participation or aiding and abetting is

spontaneous, withdrawal by leaving the scene, especially when coupled with advice or other indication to those who remain of the abandonment,

or with the effluxion of time, might be sufficient.’”

25. The internal quotation in that last paragraph is from R v. Sully [2019] SASCFC 9, a case

in which the Supreme Court of South Australia considered a range of leading authorities

from a number of jurisdictions. The quotation comes from paragraph 75, which

concludes with the observation that, although it might seem incongruous to introduce

questions of causation into an aiding and abetting situation, it was clear that withdrawal

could be demonstrated where the secondary party’s encouragement had become “spent”

even if there was no communication. However, in paragraph 76 the proposition that it

was enough that the aiding and abetting had ceased was rejected as too simplistic.

26. The next paragraph from Charleton & McDermott cited in the judgment of the Court

of Appeal reads:

“[8.90] Withdrawal may be verbal in the early stages of a criminal enterprise;

it may involve countermanding permission to use a vehicle or withdrawing from

confederacy in a plan to commit a crime. Where more advanced steps have been

taken, such as supplying the poison in Gauthier [ [2013] 2 SCR 403],

withdrawal may only be valid where it counteracts the help given. This could be

fulfilled by reporting the proposed crime to the authorities. At the more extreme

stages of a criminal enterprise a withdrawal may only be effective where the

accused takes positive action to thwart the plan by, for example, attempting to

protect a potential victim from a murderous attack. If the accused is to withdraw,

his or her duty is to reverse whatever has been the encouragement or assistance. This, in turn, depends on the degree of assistance or encouragement so far

offered. In Eldredge v United States [(1932) 62 F 2d 449] the court pithily stated

that a declared intention to withdraw from a conspiracy to dynamite a building

is not enough if the fuse has been set: the person wishing to withdraw must step

on the fuse.”

27. Applying these principles, the Court of Appeal concluded:

“44. We reject the suggestion that the trial court erred, as was submitted on

behalf of the appellant, by supposedly focussing on whether the appellant

should have admitted his own criminal acts, rather than on whether the elements

of abandonment were present. It is clear to us that the trial court’s entire focus

was on the sufficiency of the acts of the appellant which were said to provide

evidence of withdrawal from the joint enterprise. There was a clear finding of

insufficiency of the steps taken to reverse his complicity, one with which we

agree, having regard to the extent of the preparations that had already taken

place towards the intended criminal conduct, and the extent of the appellant’s

involvement.

45. It was a question of the legal effectiveness of the steps taken by the appellant

supposedly in furtherance of a claimed resolve to withdraw from the common

design. As Charleton, McDermott et al point out at para. [8.83] of the previously

cited work, to be legally effective the withdrawal must be (1) clear and

unequivocal, (2) timely, and (3) communicated either to the other parties or to

the police. The trial court was not satisfied that the appellant’s steps in furtherance of the purported withdrawal satisfied these requirements. Although

they were a non-jury court they were required to consider, as though they were

a jury, the evidence in support of the alleged common purpose/joint enterprise,

and also the evidence being relied upon in support of the claim of effective

withdrawal by the appellant, and then to determine (i) whether, on the totality

of the evidence, they were satisfied beyond a reasonable doubt that a common

purpose/joint enterprise, involving the appellant, had subsisted and was

continuing at the time of the offences, and (ii) that the appellant had not

withdrawn from the common purpose/joint enterprise, by clearly resiling from

what was contemplated such that he was no longer a party thereto, and doing

enough to undo or countermand the plan. The Special Criminal Court judges

approached the matter entirely correctly in our view, and their conclusions,

which were cogent and reasoned in detail, appear to us to have [been] open to

them on the evidence. They are unassailable in the circumstances.”

Submissions in the appeal

28. The appellant refers to a number of leading judgments from the courts of England and

Wales, Australia, Canada and the United States for the proposition that certain elements

of the test for a defence of withdrawal are clear. There must be an intention to withdraw,

which must be communicated clearly and in a timely manner. It is submitted that the

appellant meets each of these two limbs, in that he told Mr Finglas to “cease” in the

month before the assault of Mr McAndrew. 29. What is seen as less clear is the next limb – whether it suffices for the accused to simply

“demonstrate” withdrawal, or whether they need to “countermand” the effect of their

participation. It is acknowledged that there may also be debate about what constitutes

“countermanding”, with a question as to whether the steps taken by the accused must

be actually effective or simply capable of being effective. It is submitted that the

appellant took reasonable and proportionate steps to nullify or countermand his

participation. He had communicated clearly to Mr Finglas and had contacted both the

gardaí and the PSNI. All of this was done before a date and location for the meeting

had been arranged between Mr Finglas and Mr McAndrew, so it would not have been

possible for him to give precise information on those aspects.

30. The appellant further submits that the proportionality of the steps taken must be

assessed in the light of the crime that he anticipated might occur. The trial court found

that the level of violence used had gone beyond what he might have intended or

expected. He says that he had been in fear as to what might happen if he contacted Mr

McAndrew directly, and it would be unreasonable to expect him to put himself at risk

of harm. He could not do anything further to prevent Mr Finglas from going ahead,

given that he had already been assaulted by him. It is submitted that the trial court set

the test too high, requiring that all reasonable steps be taken for the purpose of

preventing the crime. That, the appellant contends, is what is needed if withdrawal is at

a late stage. Where it is timely, he says that the requirement is pitched at a lower level

– to take steps proportionate to the initial steps taken by him in the criminal enterprise.

31. Finally, the appellant argues that the trial court should have considered alternative

explanations for the fact that he was in possession of the material downloaded from the stolen phones. He submits that being presented with the fruits of a crime is not

necessarily inconsistent with having withdrawn from it.

32. The Director of Public Prosecutions acknowledges the public policy interest furthered

by the existence of the defence of withdrawal. She agrees that the criteria cannot be so

onerous as to require the accused to have prevented the crime from being committed,

since otherwise the defence would not need to be raised. The necessity for proportionate

steps to be taken is emphasised, with the question being what is reasonable in the

circumstances of the case.

33. It is submitted that the contact with the two police forces by the appellant was not only

ineffectual but was known by him to have been ineffectual. He could not have thought

that the information he had provided was capable of being effective, or likely to

countermand his participation, or that he had potentially undermined the criminal

enterprise. It was not simply that his efforts failed but that he knew he had given

insufficient information. He had not specified either the source of the danger or how

Mr McAndrew might avoid it. He had also given no warning to Mr McAndrew, whether

anonymously or otherwise. It is submitted that he was required to care at least as much

about the potential harm to the victim as for his own welfare.

34. On the final issue raised by the appellant, in respect of the accessing of information

from the phones, the respondent points out that the trial court did not rely on this matter

in reaching its conclusion on the question of withdrawal.

The defence of withdrawal/abandonment 35. About three hundred years ago Hale stated (in Pleas of the Crown vol. I, p. 618) that a

person would not be liable as an accessory if he had given a "timely countermand", but

that repentance, without an actual countermand before the fact committed, was not

enough to avoid liability. The continued existence of the defence has never been in

doubt. However, writing in 1981, Lanham observed that, while the law had recognised

for centuries that a person “who would otherwise be an accomplice” might escape

liability for the substantive crime by withdrawal before the crime was committed, there

was still a “remarkable area of doubt” about the precise limits of the defence (David

Lanham, ‘Accomplices and Withdrawal’ (1981) 97 LQR 575). Twenty years later,

Smith contended that English law had failed to provide anything approaching a

complete explanation of the conditions under which the defence may be available

(K.J.M. Smith, ‘Withdrawal in complicity: a restatement of principles’ (2001) Oct,

Crim. L.R. 769).

36. Looking at the authorities chronologically, with a view to discerning the criteria

applied, one may start with the example given by Foster:

“A, B and C ride out together with intention to rob on the highway. C taketh the

opportunity to quit the company, turneth into another road, and never joineth A

and B afterwards. They upon the same day commit a robbery. C will not be

considered an accomplice in this fact. Possibly he repenteth of the engagement,

at least he did not pursue it; nor was there at the time the fact was committed

any engagement or reasonable expectation of mutual defence and support so

far as to affect him.” 37. In so far as it is possible to speculate about this brief example, it is clear that C had

carried out no overt act beyond the initial agreement and encouragement. It also seems

that the manner in which C left A and B must have made it clear to them that they could

expect no help from him. It is noteworthy that C’s motive in abandoning the agreed

course of action was seen as irrelevant.

38. In the more modern era, the man accused of murder in R. v Croft [1944] 1 KB 295;

(1944) 29 Cr. App. R. 169 claimed in evidence that he had entered into a suicide pact

with the deceased and had provided a gun for that purpose. He said that after she shot

herself once in the chest, and was still alive, he had gone to get help. He left the revolver

with her, and she then shot herself in the head. The Court of Criminal Appeal declined

to quash his conviction for murder, on the basis that the defence had not shown an

“express and actual countermand or revocation of the advising, counselling, procuring

or abetting” which he had earlier given to the deceased.

“He never said anything to her which could have removed from her mind the

effect of the counsel which he had given her before. He does not say that he said

anything to her; and he left her wounded, possibly fatally wounded, with the

revolver loaded in all chambers except the one which had been fired. In our

opinion that conduct cannot be held to be such a countermanding or

determination of an agreement to commit suicide as would entitle him to an

acquittal.”

39. The liability of the accused seems to have been expressly grounded, therefore, in the

failure to revoke the previous encouragement. However, there also seems to have been an implicit finding that the appellant should have taken away the loaded gun. That

would have been an action that might have prevented the death, if the first shot was not

fatal. It could therefore be seen as a requirement to take reasonable steps to prevent the

crime, although not expressly described as such by the Court.

40. The judgment of Sloane J.A., of the British Columbia Court of Appeal, in R. v.

Whitehouse [1941] 1 D.L.R. 683 has been of continuing influence. In the appeal in

question the appellant had been convicted of murdering a merchant by hitting him with

a heavy pipe. He had struck the blow and was therefore the principal offender. The issue

in the appeal was the possible impact on the jury of the evidence of two witnesses for

the prosecution. Each of them said in evidence that they had, up to a certain point, aided

and abetted the appellant in the common design of violent robbery, but had fled the

scene just before or at the time he was alleged to have struck the fatal blow. The question

therefore was whether they should have been treated as accomplices, such that their

evidence required corroboration.

41. The trial judge directed the jury that if, in running away, the witnesses had abandoned

the common intent then they would not be accomplices because they would not have

been “united” in the commission of the crime when it took place. This analysis seems

to have been based on a view that, despite their previous assistance, the witnesses would

not have been accomplices because they did not have mens rea at the time of the murder.

This view would, I think, have been wrong in any event – the mens rea for an aider and

abettor must exist at the time of doing the acts that constitute the aiding and abetting,

since they must have been done with the intent of furthering the criminal purpose, but

it does not necessarily have to persist at the time of the substantive offence. 42. However, in the appeal the Court saw the matter in terms of the conditions applicable

to the defence of abandonment. Sloane J.A. saw the judge’s instruction as meaning that

there were only two elements to abandonment: (a) a change of mental intention and (b)

quitting the scene before the crime was finally consummated. In his view, this was

insufficient to absolve from all the consequences of the crime those who had already

participated in it by overt acts. Emphasis was placed on the fact that the crime was

alleged to have been completed by a person who was unaware of his companions’

change of heart. The following passage has been cited many times since:

“After a crime has been committed and before a prior abandonment of the

common enterprise may be found by a jury there must be, in my view, in the

absence of exceptional circumstances, something more than a mere mental

change of intention and physical change of place by those associates who wish

to dissociate themselves from the consequences attendant upon their willing

assistance up to the moment of the actual commission of that crime. I would not

attempt to define too closely what must be done in criminal matters involving

participation in a common unlawful purpose to break the chain of causation

and responsibility. That must depend upon the circumstances of each case but

it seems to me that one essential element ought to be established in a case of

this kind: where practicable and reasonable there must be timely

communication of the intention to abandon the common purpose from those who

wish to dissociate themselves from the contemplated crime to those who desire

to continue in it. What is “timely communication” must be determined by the

facts of each case but where practicable and reasonable it ought to be such communication, verbal or otherwise, that will serve unequivocal notice upon

the other party to the common unlawful cause that if he proceeds upon it he does

so without the further aid and assistance of those who withdraw. The unlawful

purpose of him who continues alone is then his own and not one in common

with those who are no longer parties to it nor liable to its full and final

consequences.”

43. Thus, if one has already committed overt acts in furtherance of the crime then changing

one’s mind and running away are not enough – there must be “timely communication”

of the intention to abandon. If “practicable and reasonable”, the communication must

serve unequivocal notice upon the principal offender that the accessory will provide no

further aid or assistance. It is not clear what standard is to be applied to the conduct of

the accused if such communication is not practicable and reasonable in the

circumstances of the case.

44. Sloane J.A.’s analysis was endorsed and applied by the Court of Appeal of England and

Wales in R. v. Becerra (1975) 62 Cr. App. R. 212. In that case, Becerra and his co-

accused Cooper were burgling a house, and had already assaulted the elderly owner,

when they were surprised by a tenant. Becerra shouted “There’s a bloke coming, let’s

go” and climbed out a window. Cooper killed the man with a knife that had earlier been

given to him by Becerra. One ground of defence was that by leaving in the manner he

did, Becerra had withdrawn from the common design by the time of the killing.

45. The Court of Appeal adopted the entirety of the passage cited above from the judgment

of Sloane J.A., and also the decision in Croft. In the circumstances of the case, since Becerra had been found to have given Cooper the knife for the purpose of avoiding by

violent means the hazard of identification, withdrawal would have required a

“countermand” in some manner “vastly different and vastly more effective” than

merely to say “Come on, let’s go” and leave. What Becerra had done was not capable

in law of amounting to withdrawal. In those circumstances, the Court found it

unnecessary to decide whether the trial judge had been correct in saying that he would

have had to take all reasonable steps to prevent the commission of the crime which he

had agreed the others should commit (which the trial judge thought would have meant

trying to take the knife from Cooper, or warning the tenant, or interposing his own body

between them).

46. It seems, however, that the Court in Becerra intended to convey more by the word

“countermand” than simply a verbal communication to the effect that the accessory was

no longer willing to participate. Something more “effective” was needed.

47. In R v Rook [1993] EWCA Crim 3 the Court of Appeal reached a similar conclusion.

The appellant had not in any way communicated to his accomplices his intention not to

turn up and assist in a murder, so could not avail of the defence. Again, the Court found

it unnecessary to decide whether, if there had been communication, it would also have

been necessary for him to have “neutralised” his earlier assistance. It did not, however,

accept the contention on the part of the prosecution that the accused must have broken

the chain of causation by warning the victim or going to the police. The Court noted the

quotation by Glanville Williams (in Criminal Law: the General Part) from the judgment

of the majority in Eldredge v United States – “A declared intent to withdraw from a

conspiracy to dynamite a building is not enough. If the fuse has been set he must step

on the fuse” – and suggested that this might go too far. It might be enough to try to step on the fuse, but this was a question of policy as much as law and did not arise on the

facts of the case.

48. White v Ridley [1978] HCA 38; (1978) 140 CLR 342 is a decision of the High Court of

Australia in a case of importation of drugs. The appellant consigned a box of cannabis

from Singapore to Australia, then flew ahead with the relevant documentation in order

to be in Australia to collect it. On arrival he was questioned about the documentation.

Realising that suspicion had been aroused he attempted to countermand the

consignment and leave the box in Singapore, but his message was not acted upon in

time and it arrived. He was charged with importation. The case came before the

appellate courts because the trial court was unable to determine whether or not the

airline would in fact have been able to take the necessary steps to get the box off the

plane in time if it had acted upon receipt of the appellant’s instruction.

49. Although it did not concern a joint enterprise as such, the case is frequently cited in the

context of discussions of the defence of withdrawal from a joint enterprise because

Gibbs J. saw the principles as being analogous, with the question being the effect of a

retraction of an instruction to an agent (who was, in this case, innocent). Having cited

Croft, Whitehouse and Becerra, he noted Halsbury’s statement (in the 4th edition) that

the countermand or withdrawal must have been effective if the accessory was to escape

liability, but said that in his view what was required was that it be “potentially”

effective. If the crime was not committed, then the accused’s liability as an accessory

would not arise. It would, he thought, be irrational to hold the accused liable if the

countermand was completely ineffective but not liable if it had “some” effect. Gibbs J.

also endorsed the view of Glanville Williams and Smith & Hogan that the accused remained liable, notwithstanding a communicated withdrawal, unless he took steps to

avert the danger which he had helped to create. It seemed to him to be entirely

reasonable that a person who counselled, procured or conspired should accompany his

countermand or withdrawal with “such action as he can reasonably take to undo the

effect of his previous encouragement or participation”.

“The countermand must have been manifested by words or conduct sufficiently

clear to bring it home to the mind of the agent that the accused no longer desires

the agent to do what he was previously asked to do; a vague, ambiguous or

perfunctory countermand would not be enough. And the accused must have

done or said whatever was reasonably possible to counteract the effect of his

earlier request.”

50. On the facts of the case, Gibbs J considered that the appellant had not done all that was

reasonably possible to counteract his initial request because he had not informed the

airline of the contents of the box.

51. It may need to be noted that this was not a majority view of the case. Stephens J., with

whom Aikin J. agreed, saw the case as one where the appellant had deprived himself of

the ability to change the course of events that he had set in motion. No new factor had

intervened, such as might mean that the appellant was not the importer. Stephens J. did

accept that there could be cases where a confederate decided to ignore a countermand,

in which case the confederate’s independent resolve to continue could be seen as

displacing the original instruction as a cause. Jacobs J., dissenting, did not see the

question of causation as relevant, but asked whether the mens rea for importation subsisted at the time when it occurred. The mens rea would cease if the accused had an

honest and reasonable belief that he had stopped importation from occurring. Murphy

J. also dissented, on the ground that there had been no express finding in the court below

that the appellant had not done all he reasonably could.

52. Perhaps the clearest example cited to this Court of how the defence can succeed is the

case of R. v. Whitefield (1984) 79 Cr. App. R. 36. The appellant had told two other

people that the owner of the flat next to his was currently away. They had agreed to

burgle it together on a particular night and discussed how the proceeds would be split.

The plan was to use the appellant’s balcony to gain access. However, the appellant

changed his mind before the agreed night and told the accomplices that he would not

take part. The other two went ahead anyway, and gained access without using his flat.

He knew they were going to do it, and heard them breaking in, but took no step to warn

the occupier or the police. In those circumstances, it can be said that the appellant’s

complicity was clearly established when the initial plan was made – he had provided

vital information, offered tangible assistance and was going to share in the proceeds.

The question was whether his withdrawal was effective. The trial judge ruled that his

actions could not constitute the defence of withdrawal because he had done nothing to

prevent the crime.

53. The Court of Appeal held that the defence should have been left to the jury. The

judgment cites Whitehouse, as approved in Becerra and R v Grundy [1977] Crim LR

543, as establishing that where a person’s participation is confined to advice or

encouragement, he must at least communicate his change of mind to the other, and the

communication must be such as "will serve unequivocal notice upon the other party to the common unlawful cause that if he proceeds upon it he does so without the aid and

assistance of those who withdraw." The Court considered that there was clear evidence,

if the jury accepted it, that the appellant had indeed served unequivocal notice to that

effect. It seems, therefore, that the Court was prepared to accept that the “unequivocal

notice” was sufficient to establish the defence. It may be noted that there is nothing in

the judgment to the effect that the accused had to cancel the effect of having informed

the others about the occupier’s absence, still less that he had to attempt in any way to

prevent the crime.

54. Miller v Miller [2011] HCA 9, another Australian case, was concerned with the question

whether a young woman could recover damages for injuries received when the car in

which she was a passenger crashed. She had stolen the car in order to get home, and

had assented to its being driven by a relative, but was alarmed by his driving and asked

to be let out. He refused to stop. The case was defended on the basis that she was

engaged in a joint illegal enterprise and that therefore the driver did not owe her a duty

of care.

55. The High Court of Australia held that the plaintiff had withdrawn from the joint

enterprise when she asked to be let out. The judgment is, of course, largely concerned

with the role of illegality in the context of civil proceedings and gives lengthy and

nuanced consideration to the authorities and policy considerations in that field. The

discussion of withdrawal from a criminal enterprise must be seen in that context, and it

must also be noted that the Court was dealing with a legislative formulation in the

Criminal Code providing for offences committed on foot of a “common intention to

prosecute an unlawful purpose”, where the offence actually committed was of such a nature that its commission was a probable consequence. The relevant part provided that

a person would not commit an offence if they

(a) withdrew from the prosecution of the unlawful purpose;

(b) by words or conduct, communicated the withdrawal to each

other person with whom the common intention to prosecute

the unlawful purpose was formed; and

(c) having so withdrawn, took all reasonable steps to prevent the commission of

the offence.

56. The majority of the Court found that the plaintiff had withdrawn from the joint

enterprise before the crash and had communicated her withdrawal by her requests to be

let out. Further, it held that in the circumstances there had been no reasonable steps that

she could have taken to prevent the continued unlawful use of the car.

57. In R. v O’Flaherty [2004] EWCA Crim 526; [2004] 2 Cr. App. R. 20, the Court of

Appeal of England and Wales was dealing with a fight between two groups of young

men, in the course of which one participant received fatal injuries. Much of the

difficulty for the prosecution stemmed from the fact that while CCTV captured various

incidents, the fight itself moved around between different streets with different groups

forming and assaulting each other. The victim had sustained both stab wounds and a

fractured skull but it was not possible to say who had caused them. The Court allowed

the appeals of two men who had earlier been involved in fighting the victim but had not

entered the street where he was killed. It was held that there was no evidence that they

were part of a joint enterprise with those who pursued him there. The third appeal was dismissed, because that appellant had followed the victim into the second street,

carrying a weapon.

58. Much of the discussion in the judgment about the principles applicable to joint

enterprise, and in particular about the relevance of foresight, may now have been

overtaken by the decision of the UK Supreme Court in R. v. Jogee [2016] UKSC 8;

[2017] AC 387, and hence may need to be treated with caution. However, for present

purposes, there is some relevant consideration of withdrawal. The Court rejected a

submission by the prosecution to the effect that once a person had engaged in a joint

enterprise (in that case an agreement to engage in unlawful violence) it was irrelevant

that at some stage they decided to take no further part unless the subsequent fatality

was a “separate” event. The Court considered that such a preclusion of the defence of

withdrawal could not be correct as a matter of either principle or policy.

59. At paragraph 60 the Court stated that it was necessary for the person to do “enough” to

demonstrate that they were withdrawing from the enterprise. This was ultimately a

question of fact and degree for the jury. Account would be taken inter alia of the nature

of the assistance and encouragement already given and how imminent the infliction of

the fatal injury or injuries was, as well as the nature of the action said to constitute

withdrawal.

60. Noting that in some cases it had been said that reasonable steps must have been taken

to prevent the crime, the Court (citing Whitehouse, Whitefield and Becerra) considered

that it was “clear” that this was not necessary. Furthermore, in a case of spontaneous

violence there need not necessarily be a clear communication of withdrawal. It was for

the prosecution to prove that the accused remained part of the joint enterprise. 61. It may be noted here that the distinction between pre-planned and spontaneous violence

had previously been drawn by the Court of Appeal in R. v. Mitchell and King (CA, 24

July 1998) but is not to be found in other authorities. The decision in O’Flaherty may

also be open to criticism on the ground that it does not fully distinguish between a denial

of mens rea as to the ultimate offence and the establishment of the defence of

withdrawal.

62. In R. v Sully, referred to above, the Court of Appeal of South Australia reviewed all of

the above authorities in a case where the appellant had been illegally racing his car

against another driver on a city street. He pulled out of the race, turning off the route

just seconds before the other driver was involved in a fatal collision. In his appeal

against conviction for various offences including causing death by careless driving, he

argued that he had ceased to aid and abet the actions of the other driver before the

accident. Having conducted a considered review, the judgment of Vanstone J. (the other

members of the Court agreeing on this aspect) came to the following conclusions:

What will suffice in terms of withdrawal from a joint enterprise or from a

situation which a defendant has counselled and procured or aided and abetted

a crime will vary markedly from case to case. It will involve an assessment of

what was reasonable and practical in the circumstances. The more the

defendant has done by way of planning or providing information or items to

enable completion of the crime, the more is likely to be required of him by way

of withdrawal or countermand, if he is to avoid criminal responsibility. In some

cases, particularly where the participation or aiding and abetting is spontaneous, withdrawal by leaving the scene, especially when coupled with

advice or other indication to those who remain of the abandonment, or with the

effluxion of time, might be sufficient. However, with respect to those who have

expressed a contrary view, I do not agree that there is any distinction in point

of principle between cases where there is a pre-existing agreement to commit

the crime in cases of spontaneous participation, such as by aiding and abetting.

It is a matter of fact and degree. Therefore I do not consider that Mitchell &

King and O’Flaherty should be followed in this Court. Also, while there might

seem to be a degree of incongruence in introducing questions of causation to

an aiding and abetting situation, it seems clear that withdrawal could be

demonstrated where the secondary party’s encouragement has been “spent”,

even where there was no communication.

Returning to the argument in the instant case, I would reject Mr Algie’s

proposition, said to be based on O’Flaherty, to the effect that a person’s

criminal responsibility necessarily ends at the point where he ceases to aid and

abet. This over simplifies the position.”

63. In the Canadian case of R. v Gauthier the appellant had been convicted of murdering

her children. The prosecution case was that, in pursuance of a murder-suicide pact with

her, her husband poisoned them with medication that she had obtained. She did nothing

to stop them drinking it. The alternative defences mounted in the trial were absence of

mens rea due to a dissociated state, and withdrawal from the agreement. The evidence

grounding the latter defence was her testimony that she told her husband that “we

couldn’t do that” and that she had believed that he had accepted it. 64. Under s. 21 of the Canadian Criminal Code, a distinction is drawn between two

categories of persons who are parties to an offence. In the first, they are parties because

they actually commit the offence, or they do or omit to do anything for the purpose of

aiding any person to commit it, or they abet any person in committing it. In the second

category, two or more persons form an intention in common with others to carry out an

unlawful purpose and to assist each other therein, and any one of them commits a crime

that each of them knew or ought to have known would be a probable consequence of

carrying out the unlawful purpose.

65. Giving the majority judgment Wagner J. said that it was well-established that the scope

of the defence varied from case to case, depending on the circumstances. He considered

that a person who came within the second category of s. 21 was party to the offence

because he or she had promised to devote physical and intellectual resources to the

achievement of the common unlawful purpose. The person’s liability for an incidental

offence therefore stemmed from their decision to participate and to contribute

resources. In this context, timely and unequivocal notice of withdrawal was all that was

needed for the defence to succeed. However, the situation in respect of aiders and

abettors was different because they had done more than merely promise support in

carrying out the unlawful purpose in the future. They had performed concrete acts.

Their liability stemmed from those acts and was proportionate to them. In that context,

notice of withdrawal was not sufficient “to break the chain of causation and

responsibility” (the words of Sloane J.A. in Whitehouse, quoted above). The accused

must go further. Wagner J. endorsed the view of Smith (commenting in [1999] Crim.

L. R. 497 on the decision in R v Mitchell and King) that the withdrawal had to be

accompanied by actions that were likely to cancel out the effects of their participation. The question in each case, therefore, would be whether the accused took, in a manner

proportional to his or her participation in the commission of the planned offence,

reasonable steps in the circumstances either to neutralize or otherwise cancel out the

effects of his or her participation or to prevent the commission of the offence.

66. In the case before the Court, Wagner J accepted that a jury might have found that she

had communicated her withdrawal. However, she would have to have taken further

steps, such as hiding the medication, or getting verbal confirmation from her husband

that he agreed, or calling the authorities.

67. It must be noted that Wagner J’s analysis was based at least in part on the fact that the

Criminal Code distinguished between the two kinds of participation – on the one hand,

aiding and abetting, and on the other, liability for an offence committed by an

accomplice in carrying out a mutual unlawful purpose.

68. In a strong dissent, Fish J. said that the Canadian Courts had never, in the 70 years since

Whitehouse, held that the defence of withdrawal required more than a change of

intention and timely and unequivocal communication of that change. That test had been

upheld by the Court in more recent times, but the majority were now proposing a new

test. He did not see the distinction drawn between the two statutory provisions as

carrying the weight ascribed to it, because, as he saw it, both could involve active

participation on the part of the accused. Whitehouse itself had involved persons who

had aided, and the test expressly contemplated that factor – the communication must

convey that there would be no further aid or assistance. Moreover, to require the

accused to have neutralised their prior assistance was to disregard an important

rationale for the existence of the defence, namely, that the communication would give the principal offender an opportunity to abandon the criminal objective. The

requirement that the communication be timely was linked to that objective. Fish J. noted

that in O’Flaherty the Court of Appeal had explicitly rejected the proposition that

effective withdrawal required reasonable steps to prevent the crime. Evidence of such

steps might strengthen the defence, but absence of such evidence was not fatal.

69. Finally, the parties have cited the decision of the New Zealand Supreme Court in Ahsin

v R [2014] NZSC 153. New Zealand legislation provided for a distinction, similar to

that in Canada, between those who assisted a crime (by words or conduct) and those

who were party to a common unlawful purpose where the offence was known to be a

probable consequence. “Probable consequence” is interpreted as meaning a real or

substantial risk of “something that might well happen”.

70. In prosecuting the two female appellants for murder the Crown relied upon both limbs.

It adduced some evidence that was said to show assistance and encouragement in

relation to the murder (mainly, driving the members of the group), and some that was

said to show participation in the unlawful purpose of assaulting members of a rival gang

(confrontations with various people earlier in the day), in which a killing with

murderous intent was known to be a probable consequence. There was evidence that

female voices had been heard shouting at the principal offender after his initial assault

on the victim, telling him that was enough, to get back in the car because the police

would be coming. The defence said that this was evidence of withdrawal, the

prosecution said that it showed that the women were keeping a lookout.

71. The trial judge treated the defence of withdrawal of support as relevant to the question

whether an accused had remained party to a common unlawful purpose, directing the jury that there would have to be a clear and unequivocal communication. It was also

necessary, if a person had been giving assistance, that they would have stopped such

assistance in an obvious way. However, the judge saw withdrawal as irrelevant to that

part of the case based on assistance, because the assistance already given (driving up to

the victim) could not be undone. The Supreme Court allowed the appeals.

72. A majority of the Court agreed that liability was complete when the acts of assistance

were done. In a joint judgment McGrath, Tipping and Glazebrook JJ observed that the

discussion in previous cases had not fully resolved the question whether there was no

liability in a case of withdrawal because an element of party offending had not been

proven, or because a “true defence” had been established. They saw liability for

assistance as complete when the act of assistance (or encouragement, counselling etc)

took place, provided the principal offender went on to commit the crime. Similarly,

liability under the “common unlawful purpose” category was complete at the time when

the accused joined or formed the common purpose, and was contingent only on the

commission of the crime. Therefore, in both cases withdrawal was a true defence. That

meant that there was an evidential burden on the defence to show some evidence of

withdrawal – thereafter, the burden was on the prosecution to disprove it beyond

reasonable doubt.

73. The majority saw the main policy rationale for the defence as being the potential

beneficial effect in preventing the commission of the crime, because it might dissuade

or frustrate the principal offender. Withdrawal might also demonstrate a lack of

entrenched criminal purpose or further dangerousness.

74. Turning to the question of what was required beyond a clear communication of

withdrawal, it was noted that different formulations had been reached in different jurisdictions. It was clear that the judgment was “an intensely contextual one”. The

judgment refers here to O’Flaherty, Whitehouse and Gauthier. It was held that, given

these authorities and the nature of liability under the legislation, New Zealand law

required firstly, clear communication of withdrawal and, secondly, the taking of

reasonable and sufficient steps to undo the effect of previous participation. It was noted,

however, that a clear communication could sometimes both demonstrate withdrawal

and be a step to the prevention of the crime, since it might dissuade the principal from

continuing alone. Likewise, a clear and communicated countermand, revoking earlier

instruction, advice or encouragement would often clearly convey that the party was

withdrawing from further participation and, at the same time, be a step towards undoing

the effect of a prior command or support.

75. It was said that consideration of the “reasonable and sufficient steps” requirement

involved careful factual inquiry. In particular, consideration had to be given to the

nature of the assistance and encouragement already given and the timing of the

attempted withdrawal – “what is done must be proportionate to the impact of the

assistance earlier given”.

76. The three judges concluded, in relation to the nature and extent of the party’s prior

involvement, that:

(a) Communicating discouragement or dissuasion to the principal offender,

orally or in writing, with sufficient clarity and communication might be enough

actually to undo, or be capable of undoing, the influence of previous party

conduct which consisted solely of words.

(b) If the party’s previous participation consisted of actions of assistance, further

reasonable steps to undo that previous assistance or to otherwise prevent the crime would be required for there to be a valid withdrawal. Such steps might

include, for example, retrieval of a weapon provided, warning the victim or

contacting the police.

77. The sufficiency of what was done must be judged by reference to its potential to be

effective in the circumstances of the case. Early actions of withdrawal were more likely

to be effective than late, and leaving it too late might mean that effective withdrawal

would be extremely difficult if not impossible. If the action taken did not have the

potential to either undo earlier involvement or prevent the crime, it would be

insufficient. It was suggested that juries should be asked to consider whether it was

reasonably possible that:

(a) the defendant demonstrated clearly, by words or actions, to the principal

offender that he or she was withdrawing from the offending before the offence

was committed?

(b) the defendant took steps to undo the effect of his or her previous involvement

or to prevent the crime?

(c) the steps taken by the defendant for those purposes amounted to everything

that was reasonable and proportionate, having regard to the nature and extent of

the defendant’s previous involvement?

(d) the steps taken by the defendant were timely, in the sense that the defendant

acted at a time when it was reasonably possible that he or she may be able either

to undo the effect of his or her prior involvement or to prevent the crime?

78. Interestingly, Young J. reached a different conclusion as to the nature of withdrawal.

He did not see it as a true defence, but as something that went to negate components of liability. His analysis rests to a large extent on the interpretation of the relevant statute,

and its predecessors, which is not relevant to the issues before this Court. However, it

is noteworthy that he did not see Whitehouse or Becerra as accepting that withdrawal

was a true or authentic defence. He considered that the Court in Whitefield had

overlooked Whitefield’s liability as an aider and had instead treated him as if he had

only counselled the commission of the burglary. The counselling could be withdrawn

by notice to the principal that if the crime went ahead, it would be without his assistance.

Of the judgments in White v Ridley, he thought that only that of Gibbs J would support

a “true defence” analysis. By contrast, Young J did see Gauthier as accepting that

analysis, but he considered that that was because it was thought to have been implicitly

accepted in the earlier case law.

Conclusions

79. I referred earlier to the articles by Lanham and Smith (both of which are cited in the

judgments in Ahsin). Part of Smith’s thesis was that the case law showed a tendency to

confuse two separate scenarios – one where the act of withdrawal relied upon in fact

constituted a denial of actus reus and/or mens rea, and one where it served as a “true”

or “affirmative” defence. It seems to me that the distinction between the two situations

is one of crucial importance.

80. The first consideration here is that the defence is relevant only in cases of joint

enterprise where the accused person invoking it is not a principal offender – that is,

where the accused did not commit the substantive offence but is liable to be convicted

on the basis of secondary liability. This must be so, given that the principal offender

can never claim to have withdrawn before the crime was committed. 81. In a case of joint enterprise (which, in this jurisdiction, is based on an agreement, tacit

or express, between the participants) it is possible for a person who did not carry out

the actual crime, and perhaps was not even present when it was carried out, to be

charged, convicted and punished in the same way as a principal offender. That would

arise if, with the required intent, they did something that amounted to aiding, abetting,

counselling or procuring the commission of the offence.

82. For example, the accused may have devised the plan for the crime, or recruited the

persons who did carry it out, or provided assistance such as weapons or special

information for it, or in some way encouraged the principal offender(s) to proceed with

its commission. In each such case, the accused person becomes potentially liable once

they have intentionally taken the relevant actions. That liability crystallises when the

offence is carried out. It is not a defence for the accused to say afterwards that they had

changed their minds before the substantive crime was committed and in fact hoped that

it would not be committed.

83. A person who is alleged to be a secondary party in a case of joint enterprise may defend

the case in a number of different ways. They may, for example, argue that the evidence

is insufficient to prove that they did, as a matter of fact, do the action that the

prosecution claims they did. They may say that the evidence does not prove that they

knew of and agreed to the commission of the crime. They may make the case that they

had agreed only to a less serious crime, or to something that was not a crime at all, and

that what the principal offender(s) actually did went materially beyond or outside the

scope of that agreement. If any of these contentions succeeds, it will mean that the

prosecution has not proved that the accused was, in truth, part of a joint enterprise or common design that culminated in the crime in question. An acquittal must follow as a

matter of right in those circumstances, without the imposition of further policy-based

requirements that the accused must have done anything more, whether by way of

communicating his intentions or acting to prevent the crime. Of course, evidence of

such conduct on the part of the accused will strengthen the case for finding that the

prosecution has not proved its case.

84. If the foregoing is correct, then the defence of withdrawal becomes relevant only if the

prosecution case establishes that the accused is otherwise liable to be convicted – that

is, where the participatory acts of the accused are proved beyond reasonable doubt to

have reached the point in the joint enterprise where his or her guilt is established, and

the offence committed was within the scope of that enterprise. In Lanham’s opening

words, the accused who relies on the defence is “a person who would otherwise be an

accomplice”. Accordingly, it is only at the point where guilt has been established that

the public policy interest may provide a defence of withdrawal, excusing the accused

from conviction and punishment for the substantive offence, to an accused who meets

the relevant criteria. I would, therefore, see it as a “true” defence, but one of a special

excusatory character.

85. In determining what exactly those criteria are it is of course essential to bear in mind

that the defence only arises where the crime has actually been committed. It is

necessary, therefore, to be cautious about the use of language in order not to give the

impression that the defence cannot succeed unless the accused prevented the crime from

taking place. 86. It is also necessary to note at this point that the central feature of the defence, accepted

in all of the authorities, is that accused persons who avail of it must have communicated

their withdrawal to the other participants. The focus is, therefore, on the accused’s

relationship with those participants and on actions taken that might have the effect of

dissuading them from pursuing the agreed scheme but will, in any event, make it clear

to them that the accused will not participate further. In that context, the defence would

not appear to be relevant if, for example, a person went to the police and informed them

of the plan in a timely and adequately detailed manner but made no communication of

withdrawal to the accomplices. Such a scenario might well raise other public policy

issues, that might result in a completely different form of outcome. It could for example

result in an offer of immunity and/or protected witness status, or a reduction in charges

brought against the accused. However, it is not what is before the Court in this appeal.

87. While the distinction – between, on the one hand, a failure of the prosecution to prove

its case in respect of joint enterprise, and, on the other hand, the invocation of the

defence by an accomplice who is otherwise liable to be convicted – is clear in theory,

it may of course be that in any given case the evidence of words and actions is equivocal

and open to competing interpretations. Accordingly, it may be open to an accused to

argue, as this appellant did, both that their actions demonstrated that they were not full

participants in the crime, and therefore not guilty, and alternatively that the actions

showed that they had withdrawn effectively from any agreement to which they may

have been a party before it was carried out. Nonetheless, it is important to keep the

distinction in mind. 88. In this particular case, the defence no longer maintains that the appellant was not part

of the joint enterprise. All necessary elements of guilt have been established, as

recorded in the verdict of the Special Criminal Court, and the only issue for the Court

is whether there was sufficient evidence to establish the defence of withdrawal.

89. It will have been seen that there is a unanimous view across the common law world that

the defence requires evidence of a clear, unequivocal and timely communication to the

principal offender, whether by way of words or other conduct. What is less clear is

whether there is a further requirement that the accused should have taken other steps,

and if so, what the nature of such steps must be. Such a requirement has been held to

apply in Australia, New Zealand and Canada, although the extent to which the

development of the concept may have been influenced by the existence of specific

legislation dealing with liability in joint enterprise is not for me to say. Certainly, Fish

J. in Gauthier was confident that it had not previously been part of Canadian law. The

Court of Appeal of England and Wales did not require any steps to have been taken by

the accused in Whitefield. It might seem to have adopted such a requirement to some

extent in Becerra, and to have been open to it (depending on whether it arose on the

facts) in Rook, but explicitly rejected it in O’Flaherty.

90. I agree with the view expressed in Rook that the question is as at least in part a question

of policy. However, since the defence is a policy-based common law development in

the first place, and in the absence of any relevant Irish authority, that is not a bar to the

Court considering it in an appropriate case. In the instant appeal it is relevant because

the appellant certainly took some steps, which were found by the trial court and the

Court of Appeal to have been inadequate. 91. It will have been seen that various policy considerations have been put forward as the

rationale for the defence of withdrawal. Of these, my view is that the most forceful from

the point of view of the public interest is that it gives an incentive to those who become

involved in planning a criminal enterprise to change their minds and to do something

that might reduce the likelihood of the crime being committed. Other considerations,

such as the demonstration of a lesser level of dangerousness or culpability on the part

of an individual who withdraws, seem to me to be more relevant to mitigation of

sentence than to excusing that individual for criminal acts.

92. If this is correct, and the primary consideration is the desirability of reducing the

likelihood of crime, then it seems rational to ask whether what the accused did could

have had that effect in the circumstances of the case.

93. I agree with the analysis of the Supreme Court of New Zealand in Ahsin, in observing

that in some cases the communication itself could be capable of having that effect but

in others more will have been necessary. This is a fact-dependent matter. I will not use

words such as “neutralising” or “cancelling” in relation to the effects of the accused’s

previous contributions to the enterprise, since in some cases neutralisation or

cancellation would be simply impossible. For example, encouragement can be revoked,

weapons may in some cases be recoverable, but information cannot be taken back after

it has been imparted. However, the impossibility may be due to the fact that the accused

failed to take steps until it was too late to prevent the crime. I would prefer to express

the requirements in these terms – the actions taken by the accused must be objectively

proportionate to the effects of the earlier actions of the accused. The more they have

done to further the criminal enterprise, the more they must try to prevent it from coming

to pass. 94. I would adopt the suggestion in Ahsin as to the questions to be addressed by the finders

of fact, which I will repeat here:

Is it reasonably possible that

(a) the defendant demonstrated clearly, by words or actions, to the principal

offender that he or she was withdrawing from the offending before the

offence was committed?

(b) the defendant took steps to undo the effect of his or her previous

involvement or to prevent the crime?

(c) the steps taken by the defendant for those purposes amounted to everything

that was reasonable and proportionate, having regard to the nature and

extent of the defendant’s previous involvement?

(d) the steps taken by the defendant were timely, in the sense that the defendant

acted at a time when it was reasonably possible that he or she may be able

either to undo the effect of his or her prior involvement or to prevent the

crime?

95. The burden on the accused is an evidential one – to adduce evidence, or point to

evidence already in the case, that is sufficient for the issue to be put before the finders

of fact. The question for them is whether it reasonably possible that the accused had

withdrawn. 96. The first requirement – that of communication with the co-conspirators – is essential to

a defence of withdrawal for three reasons: because it establishes the fact that the

accused person has dissociated from the proposed crime, because of its potential to

dissuade all or some of the co-conspirators from proceeding and, therefore, because of

its potential to prevent a crime from being attempted or committed.

97. In some cases, depending on what, exactly, the accused has done up to that point, such

communication may be sufficient for the establishment of the defence because, as a

matter of fact, it undoes the effect of their earlier involvement. In such a case, it will

not be necessary to go any further than the first proposed question.

98. It is in those cases where the accused has been more deeply involved, and has gone

beyond words of encouragement or the giving of peripheral information, that further

steps beyond the communication of dissociation are required. Since these further steps

are for the purposes of undoing the effect of earlier involvement and preventing the

crime, they must be proportionate to that involvement. Proportionality, in this context,

imports the concept of reasonableness.

99. In applying this approach to the appellant’s case, it must be accepted that he did

communicate to Mr Finglas that the plan should cease. Mr Finglas’s reaction would

seem to indicate that he understood the communication clearly, because he reacted

violently to the suggestion that the opportunity should be passed up. However, that

picture becomes somewhat blurred by the fact that the appellant ultimately obtained

what the trial court found he had originally intended to get from the crime – data about

Mr McAndrew. Complaint is made on behalf of the appellant about the fact that the

trial court saw this as inconsistent with withdrawal, because, it is said, there could be an innocent explanation. That seems to me to overlook the fact that the appellant

purported to give an innocent explanation in his evidence. That explanation was

expressly disbelieved by the court. While he did not bear a burden of proof on the issue,

the court was entitled to draw an inference from lies.

100. In any event, the appellant knew that his communication to Mr Finglas had not

had the effect of dissuading the latter, and that he would proceed with the plan that the

appellant had devised and implemented. It was therefore necessary for him to take

further steps.

101. It must also be accepted that the appellant took some steps to prevent the crime

by contacting the two police forces and that those steps were timely. However, his

action did not, to my mind, amount to all that was objectively reasonable and

proportionate having regard to his earlier involvement. It must be remembered that the

plan was the appellant’s idea, that he recruited Mr Finglas, that he sent the emails up

until at least the 21st October and that he sourced the appropriate photographs for them.

Mr Finglas would not have been able to think up a way to get Mr McAndrew to the

location where he was assaulted and threatened without the appellant’s input.

102. The steps that the appellant was required to take, therefore, had to be

commensurate with that central role in the enterprise. The key point, it seems to me,

was that it needed to be conveyed (whether through the gardaí or the PSNI or directly

to Mr McAndrew personally) that the danger was associated with messages from “Alan

Mooney” and/or “Barry”, and that Mr McAndrew should not agree to meet with anyone

contacting him in relation to those messages. It must, in my view, have been obvious

to the appellant that one or both of the police forces would have to be given more detailed information if they were to take preventive action. As it was, D/Garda Hayes

was left with the vague impression that there was a danger to Mr McAndrew and that

something might happen to him in Northern Ireland. The Court has only the appellant’s

account of what he told the PSNI, and that account does not go any further than that he

told them there was a risk to a person who resided in the Republic.

103. I would hold, therefore, that such efforts as he made were simply inadequate for

the purposes of either the undoing of the effect of the actions he had taken earlier or the

prevention of the crime.

104. In the circumstances I would dismiss the appeal.