NOTES

 

Voluntary withdrawal in the context of attempt - a defence?

 

 

1 Introduction

Once a crime has been committed, full repentance and restoration do not have any bearing on liability (Simester and Sullivan Criminal Law: Theory and Doctrine (2001) 305), but may be taken into account in mitigation of sentence. (For a discussion of the concept of remorse in sentencing, in respect of which repentance and restoration may be strong indicators, see Terblanche A Guide to Sentencing in South Africa 3ed (2016) 229-230). On the other hand, there is no question of criminal liability ensuing for an attempt at a crime if there is a withdrawal from the envisaged crime while still in the stage of preparation, and before, in South African law, reaching the watershed moment of the "commencement of the consummation" (S v Kudangirana 1976 (3) SA 563 (RA) 565-566; Snyman Criminal Law 6ed (2014) 284; on the commencement of the consummation test, see Hoctor "The (Surprising) Roots of the Test for Criminal Liability for Interrupted Attempt in South African Law" 2015 SACJ 363). However, what occurs between the moment when the attempt begins, and the moment when the crime has been completed, where there has been a withdrawal from the criminal purpose, is more contested terrain. The disagreement does not apparently arise in the South African case law, where the few judgments that refer to this question have consistently held that where the accused withdraws after the commencement of the consummation of the crime, there will be attempt liability and, at best, the accused may rely on the abandonment as a mitigating factor in sentencing (see Rabie "Die Verweer van Vrywillige Terugtrede by Poging: 'n Tweede Mening" 1981 SACC 56, 61; the view that abandonment should only be a mitigating factor, rather than a substantive defence, is supported by Lee "Cancelling Crime" 1997 Connecticut Law Review 117 152 and Yaffe Attempts (2010) 291). However, as is discussed, prominent South African academic commentators, along with comparative sources in both the civil-law and common-law jurisdictions, demur from such an "unyielding analysis" (the phrase is that of Simester and Sullivan Criminal Law 305), and would regard such withdrawal as giving rise to a defence to criminal liability. Which approach ought to be applied in South African law? The difficulties inherent in making this choice are evident in the words of Yaffe (Attempts 309):

"Change of mind, as common as it is, is puzzling from a moral point of view. It is hard to know how to react to the person who acts in a way of which we disapprove, but changes his mind later. As the agent of a bad act, he seems worthy of censure; as the agent who prevented the bad act's occurrence, he seems worthy of praise. The puzzle arises from the fact that he is, indeed, both of these agents."

The question may be posed as to how to categorise a defence of voluntary withdrawal? It is neither a justification ground nor a ground excluding fault, but rather a ground excluding punishment (Rabie 1981 SACC 58). The uniqueness of the defence is demonstrated in that the accused has already met all the requirements for liability, and thus it is not an intending criminal, but an actual criminal who is being considered (Rabie 1981 SACC 58 (author's own emphasis). This is at least true of the common-law approach (also adopted by South African law), where a two-stage approach is applied to the trial, relating first to establishing criminal liability and followed, if guilt is so established, by an inquiry into sentence. At the outset, it may be stated that the view that is taken in the discussion that follows is that there is no good reason to treat voluntary abandonment as a special defence. (For further evaluation, see Rabie 1981 SACC 56). As Yaffe (Attempts 293) has stated, to grant a defence on the basis of abandonment is to mistake the absence of a reason to issue a particular sanction rather than a lower one for a sufficient reason to issue no sanction at all.

In the discussion that follows, the current case law is examined, whereafter the alternative approach contended for by some academic writers (and used in other jurisdictions) is discussed; the arguments for and against a renunciation defence are set out, before these aspects are drawn together in a final concluding analysis.

 

2 Case law

The first case dealing with the issue of abandonment - that of Q v Töpken and Skelly ((1880-1884) 1 Buch AC 471), which related to a successful appeal against convictions of attempted robbery and attempted murder -does not provide much clear authority or guidance as to the question of what the legal consequences of voluntary withdrawal from an attempt might be. De Wet (Strafreg 4ed (1985) 170) comments that the court did not pay serious attention to the matter. The court comments (474) that

"[t]he evidence is quite consistent with the view that, however criminal their intentions may have been, they repented before the time arrived for carrying their intentions into practical effect; and if this view be correct, the charge of an attempt to commit a crime falls to the ground."

The facts were that the accused intended robbery, but got the day wrong as to when the postal coach passed by the place where they were waiting, and so their waiting was in vain. They then went home. While the cited statement is compatible with a defence based on voluntary abandonment of the attempt, it may equally be a simple acknowledgement that the accused were still in the stage of preparation, and had yet to commit an attempt. The relatively undeveloped nature of attempt liability in South African law at this point is indicative that the latter option is more plausible. Rabie (1981 SACC 58) is in accord with this view.

The case of Andrew (1916 TPD 20 24) provided clearer, albeit obiter, guidance when it stated, in the context of a factual scenario that pertained to an interrupted attempt, that an attempt may be committed even though the completion of the offence was interrupted by the voluntary act of the accused. In casu, despite his attempt to pour the brandy out of the glasses that he was carrying when he saw the policeman, the accused was convicted of attempting to supply liquor contrary to the liquor laws. Rabie (1981 SACC 56) notes that this case should preferably be regarded as an interrupted attempt. A rather similar scenario applied in Hlatwayo (1933 TPD 441), where a young domestic worker wanted to poison her employers, and to this end threw caustic soda into their porridge. When this changed the colour of the porridge, she threw it away. She was nevertheless convicted of attempted murder, with the court holding that voluntary withdrawal has no bearing on liability.

The case of B (1958 (1) SA 199 (A)) related to apartheid legislation prohibiting inter-racial intercourse. When the accused was unable to perform sexually owing to alcohol and stress, he abandoned the attempt, but it was held that he was nonetheless guilty of contravening the statutory provision. The court confirmed the correctness of the approach taken in Hlatwayo with regard to voluntary withdrawal (supra 203A-B). The approach of the appeal court in B was cited with approval in the (erstwhile) Rhodesian case of R v Khalpey (1960 (2) SA 182 (SR) 186-187), which confirmed that an accused may still be guilty of an attempt even where he voluntarily desists from carrying out his original intention.

The question of withdrawal arose once again in S v Agmat (1965 (2) SA 874 (C)), where a pickpocket had unclasped the complainant's bag with a view to stealing her purse. The appellant argued that any theft was still in the stage of preparation, which would allow for the possibility that he had abandoned the attempt (supra 875D-E). However, the court decided that the appellant was indeed in the stage of execution and, that being so, it was quite immaterial whether the appellant went to look into a shop window because he now had changed his mind as to the carrying out of the crime, or whether he had merely gone to stand there to see if the complainant became aware that her bag had been opened so that, if she did not become aware, he would then follow her and complete the crime (supra 875E-F).

In S v Du Plessis (1981 (3) SA 382 (A)), which dealt primarily with an appeal against convictions of contravening the Official Secrets Act 16 of 1956, it was confirmed that voluntary withdrawal after commencement of consummation is not a defence, and that if the change of mind occurred before the commencement of the consummation, then the person concerned cannot be found guilty of an attempt, but "if it occurred after the commencement, then there is an attempt and it does not avail the person concerned to say that he changed his mind and desisted from his purpose" (supra 400D-E). The Appellate Division cited the cases of Hlatwayo (supra 444-445) and B (supra 203B) with approval. However, this statement was obiter as no definitive intention to commit the offence could be identified, and consequently no attempt was committed.

Whether the cases of Hlatwayo and B are indeed sound authority for the proposition that voluntary withdrawal does not constitute a defence to attempt, once the stage of preparation has been passed, has been doubted by some writers. Kemp, Walker, Palmer, Baqwa, Gevers, Leslie and Steynberg (Criminal Law in South Africa 3ed (2018) 302) simply follow the authority in Hlatwayo in excluding the possibility of a defence where there has been a change of mind. Snyman argues that the withdrawal in Hlatwayo was not voluntary, as the accused was "caught out by other people (other domestic workers in the house who saw what she had done) and for that reason decided not to proceed with her plan" (Criminal Law 284). Rabie (1981 SACC 56) agrees that this was a case of interrupted attempt. Similarly, Snyman (Criminal Law 4ed (2002) 289) comments regarding the case of B that it was not a genuine case of voluntary withdrawal, as external factors that came to the accused's attention at a late stage - thus founding an interrupted attempt rather than a withdrawal - induced him not to continue with his acts. De Wet (Strafreg 170) agrees that this was not a pure case of withdrawal, as it was not that the accused abandoned his intent as a result of better insights, but because the flesh was too weak, though the spirit was still willing. Rabie (1981 SACC 56) states that the facts of this case are more properly classified as a voluntary withdrawal. Burchell (Principles of Criminal Law 5ed (2016) 555) is of the view, given the less-than-determinative authority on the point, that the relevance of a truly voluntary withdrawal from an attempt to commit a crime - as opposed to an involuntary withdrawal as a result of being detected for instance - still falls to be determined in South Africa.

 

3 Is there a need for a voluntary withdrawal defence, and how should it be structured?

It is clear that a possible voluntary abandonment defence can only have application once the course of events, along with the accused's intent, satisfies the test for attempt liability. However, protagonists and those systems supportive of the defence do not describe it uniformly. One of the leading supporters of the test in South Africa, Snyman, does not dispute that if the withdrawal takes place after the first harm has already been done, the attempt should give rise to conviction and punishment. Hence, he argues, if in the course of committing assault, X "withdraws" after having struck the first blow, or if in the course of committing arson, she "withdraws" after the first flames have already damaged the building, the "withdrawal" is too late to afford X a defence (Snyman Criminal Law 284). However, Snyman contends that if X withdraws before having inflicted any harm or damage, even if her conduct up to that stage can be construed as having already passed the point where the "consummation has commenced", then there should not be liability for attempt (Snyman Criminal Law 284). Snyman relies on comparative jurisdictions that favour this defence, including those US states that follow the Model Penal Code and continental legal systems such as the Netherlands and Germany (for discussion of these provisions, see heading 4 below).

Burchell (Principles of Criminal Law 556) notes that the complexity associated with the issue of voluntary withdrawal relates to the balance that has to be drawn between antagonistic interests: on the one hand, the view taken by Snyman is that while harm has not yet been done to society, the law should encourage desistance from the intended purpose; on the other hand, there is the concern that the provision of a "fall-back" defence would encourage those who wish to embark on a criminal endeavour to do so in the knowledge that there will be an opportunity to avoid liability on the basis of withdrawing from the purpose. The point of departure for Burchell, like Snyman, is that there are "strong reasons" to allow for a voluntary withdrawal defence (Principles of Criminal Law 557), with two qualifications - namely, that there would nevertheless be attempt liability, despite withdrawal: (i) where the accused's intent to offend is still clear at the moment of renunciation, but he withdraws because he has decided that carrying out the purpose to offend at that particular opportunity was no longer suitable; and (ii) where the attempt has already been completed. Burchell further narrows the ambit of his conception of the voluntary withdrawal defence by pointing out that even where the withdrawal takes place in the context of an uncompleted attempt, there may yet be liability where the court simply does not believe in the authenticity of the withdrawal, or where the accused's acts already constitute an attempt at the impossible (Principles of Criminal Law 557-558).

Another proponent of the voluntary withdrawal defence is Labuschagne ("Vrywillige Terugtrede uit 'n Misdaadpoging en Menslike Gedragsbeheermeganismes: Opmerkinge oor die Persoonlikheidsregtelike Begrensing van die Strafreg" 1995 Stell LR 186), who associates himself with views expressed by both Snyman and Burchell, although his own specific point of departure is that the accused who withdraws has not yet broken through the psychological barrier to crime ("interne gedragsbeheermeganismes"); that criminal law should not apply in this private space; and that liability should consequently not follow.

The arguments raised by these writers in favour of the voluntary withdrawal defence are set out below. At this juncture, a further question may be posed: is it appropriate to base a voluntary withdrawal defence on a subjective basis? Rabie points out that if the issue is the demonstration of a remorseful state of mind, then it ought not to matter whether the accused had actually committed the crime originally intended, or had only committed the crime of attempt (1981 SACC 59). Proponents of the defence would certainly want to recognise the difference between these scenarios, given that there is no debate that remorse after completing a crime does not affect the attribution of criminal liability; Rabie suggests that such a defence may better be founded on the objective criterion that an accused who withdraws after an attempt would cause less harm to the community than an accused who is remorseful about committing a crime, even if he takes steps to ameliorate the harm caused after the commission of the crime (1981 SACC 59). Whether this is indeed an appropriate basis for such a distinction remains a moot question.

 

4 Some comparative observations

English common law has not recognised the possibility of a voluntary withdrawal defence (Ormerod and Laird Smith, Hogan, and Ormerod's Criminal Law 15ed (2018) 434, citing the case of Taylor (1859) 1 F&F 511; Horder Ashworth's Principles of Criminal Law 9ed (2019) 524; Haughton v Smith [1973] 3 All ER 1109 1115; but see Simester and Sullivan Criminal Law 305, who are more equivocal). The question was considered by the English Law Commission prior to the drafting of the 1981 Criminal Attempts Act, but no change was proposed to the common-law approach. Following the approach adopted in the United States Model Penal Code (MPC), which adopted an almost exclusively subjective approach to attempt liability (Dressler Understanding Criminal Law 5ed (2009) 412-413), a substantial number (roughly half of US states - see the list in Lee 1997 Connecticut Law Review 120n13) do provide for a voluntary withdrawal defence in the context of attempt (although exact requirements differ from state to state). The model provision may be found at section 5.01(4) of the MPC, which states:

"Renunciation of criminal purpose: When the actor's conduct would otherwise constitute an attempt ... it is an affirmative defense that he abandoned his effort to commit the crime or otherwise prevented its commission, under circumstances manifesting a complete and voluntary renunciation of his criminal purpose ... Within the meaning of this Article, renunciation of criminal purpose is not voluntary if it is motivated, in whole or in part, by circumstances, not present or apparent at the inception of the actor's course of conduct, that increase the probability of detection or apprehension or that make more difficult the accomplishment of the criminal purpose. Renunciation is not complete if it is motivated by a decision to postpone the criminal conduct until a more advantageous time or to transfer the criminal effort to another but similar objective or victim."

The nature of the voluntary abandonment defence contained in s 5.01(4) of the MPC bears closer scrutiny, however. First, its availability is limited by an "affirmative defence", which means that X bears the burden of proof to establish the defence on clear and convincing evidence. Simester and Sullivan (Criminal Law 306) also postulate the possibility of incorporating the requirement that the accused is required to prove the defence on a balance of probabilities, given the "highly exceptional" nature of the defence. However, as Burchell (Principles of Criminal Law 555) correctly (it is submitted) points out, given the antipathy towards placing a reverse onus on the accused in South African law, such a solution is unlikely to find favour in this jurisdiction. Secondly, not only is there a procedural limitation on the defence in terms of the MPC model, but the substantive qualifications on what constitutes "voluntary" renunciation in terms of this defence narrow its availability even further. In short, wherever the actor is motivated by circumstances that increase the probability of the actor being arrested for his actions, or that increase the difficulty of achieving the envisaged purpose, then the renunciation is not regarded as "voluntary". Hence, renunciation on the basis that the accused is choosing to transfer the criminal purpose to another target or until a more appropriate time does not avail the accused. At the least, it can be said that the American defence is limited to certain situations and/or motivations only, and so it is clear that it is by no means a complete defence.

As Snyman notes (Criminal Law 285), voluntary withdrawal before the completion of the crime is treated as a defence in the Continental legal systems. The comparison with continental or civil legal systems should always be borne in mind; unlike the South African (and common law) two-stage trial, in these systems, the criminal trial comprises a single-phase hearing that examines both liability and punishment (Du Plessis "Hans Welzel's Final-Conduct Doctrine: An Importation From Germany We Could Well Do Without" 1984 SALJ 301 317). Hence, in continental systems, attempt liability is only seen as a stage of development in the consummation of the offence (Keiler and Roef (eds) Comparative Concepts of Criminal Law 2ed (2016) 228). While space constraints preclude a detailed examination of these systems, for present purposes it bears noting that there are significant differences in application between certain systems. Thus, the German system applies an essentially subjective approach to the defence, in terms of which the actor can rely on the defence if, at the moment of withdrawal, he has "in his view not done everything necessary to achieve his goal, but believes that by simply continuing to act he could reach it" (Keiler and Roef Comparative Concepts of Criminal Law 230, original emphasis). The Dutch system is much more objective in nature (for discussion on the Dutch law relating to attempt, see Jörg, Kelk and Klip Strafrecht Met Mate 14ed (2019) 151ff). Hence, for example, withdrawal from an attempt at murder, having inflicted severe injuries, would be regarded as a completed attempt in Dutch law, renunciation of which would not negate attempt liability. In Germany, such facts would merely give rise to an incomplete attempt, resiling from which would provide a voluntary withdrawal defence (Jörg et al Strafrecht Met Mate 151ff). Nevertheless, though applying different modes of classification of attempt, both systems allow a defence where the withdrawal is deemed "voluntary", and both seem to apply the following formula as a standard: "a withdrawal will be considered voluntary if the offender thinks 'I do not wish to carry on even if I could' while it will be involuntary if he thinks 'I cannot carry on even if I wanted to'" (Keiler and Roef Comparative Concepts of Criminal Law 232). As with the Model Penal Code definition of the defence, it is therefore clear that in Dutch and German law, an "involuntary" withdrawal on the basis of police intervention or fear of detection would not constitute a defence (Keiler and Roef Comparative Concepts of Criminal Law 232).

 

5 Arguments for and against the voluntary withdrawal defence

What is the rationale for allowing a defence based on voluntary withdrawal? It has been argued that the rationale for punishing attempt is to be found in the utilitarian theories of punishment (see e.g., Williams Textbook of Criminal Law 2ed (1983) 404) such as deterrence, prevention and reformation, which are justified by "the advantage [punishment] brings to the social order" (Burchell Principles of Criminal Law 3ed (2005) 73). If somebody therefore voluntarily resiles from his criminal scheme it means that he has already been deterred from committing the crime (Horder Ashworth's Principles of Criminal Law 525) and its commission has already been prevented. No danger to society remains, if ever there was, and consequently no punishment is required (this argument is raised by Snyman Criminal Law 285). Simester and Sullivan (Criminal Law 305) explain that this approach is consistent with the principal rationales of attempt: "to allow timely interventions by law enforcement personnel prior to realisation of any harm and to allow lawful restraint of the socially dangerous". They argue that a genuine withdrawal prior to the harm being realised indicates that such intervention and restraint were not required (Criminal Law 305). As for the reformative theory, it is contended that there is nobody to be reformed because the accused has already reformed himself (Snyman Criminal Law 285). These are the arguments in favour of the defence based on sentencing theory.

Such arguments may be countered by the fact that once the requirements for an attempt have been met, the accused has consequently committed a crime, and the only issue to be determined is not guilt, but punishment. This accords with the "temporal logic of the law" (Horder Ashworth's Principles of Criminal Law 524), as explained by Hoeber ("The Abandonment Defense to Criminal Attempt and Other Problems of Temporal Individuation" 1986 California Law Review 377 380) in the following terms:

"Criminal prohibitions have temporal dimensions in the sense that, given the definition of any crime, we may refer to later occurrences as 'post-crime' conduct or events. In relation to the crime of theft, for example, the return of stolen property is a post-crime event. In relation to the crime of attempt, abandonment is a post-crime event. Criminally prohibited states of affairs are thus ... temporally individuated".

It is interesting that Hoeber's words are cited with approval by Labuschagne (1995 Stell LR 186), despite his support for a voluntary withdrawal defence. Nevertheless, Labuschagne goes on (198) to state that it amounts to circular reasoning to argue that since an attempt is an offence there must be liability, since, if the defence were accepted, there would not be any liability. Labuschagne's contention does not detract from the central truth that (as he acknowledges) liability for attempt is indeed established once the requirements are met. The point is that any voluntary withdrawal defence is a singular one, applied post-crime, expunging the liability already incurred through the application of the rules of attempt liability. Whether such a peculiar defence should exist is the point of contention, not whether liability for attempt arises.

Moreover, arguably the most fundamental theory of punishment, the theory of retribution, is entirely consistent with the approach that excludes a voluntary withdrawal defence (Rabie 1981 SACC 60-61; significantly (and somewhat contradictorily) Snyman Criminal Law 11-15 contends for the preeminence of retributive theory despite his support for a voluntary withdrawal defence; while Labuschagne 1995 Stell LR 196 is of the view that retribution is not consonant with modern punishment theory, which would allow for the voluntary withdrawal defence). As Lee (1997 Connecticut Law Review 142) notes, the concept of retribution provides a basis for "some mitigation of punishment where abandonment is concerned".

This may be contrasted with the idea underlying the theories of punishment that are proffered in support of the withdrawal defence, where the inference of non-dangerousness on the part of the actor, and his consequent unsuitability to be punished, are inferred from the abandonment. Fletcher, who favours the voluntary withdrawal defence, states that this claim "raises the basic question whether the criminal law should be grounded in case-by-case assessments of personal dangerousness" (Rethinking Criminal Law (1978) 187). As Fletcher further inquires (Rethinking Criminal Law 187), would it not then be sound to inquire whether the actor was generally dangerous (author's own emphasis), even though on a particular occasion he abandoned the offence? He concludes that this is an unsatisfactory argument in the context of the attribution of criminal liability. It is submitted that this approach is to be preferred to that of Labuschagne (1995 Stell LR 197), who finds the argument founded on non-dangerousness compelling. Lee notes (1997 Connecticut Law Review 145-151) that unless the courts are convinced of a "genuine moral conversion", they are generally not receptive to the argument that non-dangerousness may be inferred from abandonment. In any event, Lee reasons (1997 Connecticut Law Review 152), while attempters who renounce their attempted crime are almost certainly less dangerous than attempters who persist, attempters who renounce are most probably more dangerous than members of the public at large.

A further argument adduced in favour of allowing voluntary withdrawal to function as a defence is that it accords with one of the basic reasons for distinguishing between acts of preparation and acts of consummation: a person ought not to be punished as long as there is still a possibility that she may change her mind for the better (Snyman Criminal Law 285). Snyman argues that "this is possibly exactly what Watermeyer CJ had in mind" in R v Schoombie (1945 AD 541 547-548) when he spoke of "the last series of acts which would constitute a continuous operation, unbroken by intervals of time which might give an opportunity for reconsideration" as acts of consummation (Snyman's emphasis; Burchell Principles of Criminal Law 556 cites this argument with apparent approval). However, what Watermeyer CJ was seeking to describe was the stage of attempt liability (the commencement of the consummation) that had progressed beyond the stage of preparation, and which could thus found criminal liability for attempt. Once this stage has been reached, liability for attempt has been established. The voluntary withdrawal defence only arises for consideration after this point, and so this dictum is not applicable to the current discussion. (For some comments on this dictum in the general context of incomplete attempt, see heading 6 below.)

Nevertheless, the question remains whether this argument in favour of a voluntary withdrawal defence is indeed valid, allowing maximum scope to the potential offender to change his mind so as not to transgress the law. This idea dovetails with the idea that the law ought to encourage prospective wrongdoers not to transgress, and that it cannot do this by punishing people who decide to abandon their criminal plans (Snyman Criminal Law 285). The argument avers that the prospective criminal should know that he will be rewarded if he voluntarily abandons his criminal project (Snyman Criminal Law 285). Burchell (Principles of Criminal Law 556) contends that this will be more likely to serve to nullify the materialising of future harm, "since the accused in an attempt situation is often the sole author of his or her own fate", with his or her liability not customarily linked to that of others. However, whether the existence of a withdrawal defence really does motivate the accused not to complete her intended crime is by no means clear (Ormerod and Laird (Smith, Hogan, and Ormerod's Criminal Law 434) state that this "seems unlikely"; Fletcher (Rethinking Criminal Law 186) states that this claim "needs far more proof than any of its proponents have offered"; Labuschagne (1995 Stell LR 195) states that this approach is now regarded as naïve, although he contends that it has symbolic significance). If the accused has weighed up the risk of being caught, and nevertheless proceeded with her intended crime, why would her decision now change when the same criteria apply later in the course of conduct (see the discussion in Lee 1997 Connecticut Law Review 142-144)? If the reason for not proceeding is that the accused sees, rather than the error of her ways, the enhanced likelihood of being caught, then, as is the case in the Model Penal Code version of the defence as well as the Dutch and German law (discussed above at heading 4), it is rather doubtful, on policy grounds, that the accused should not be punished. After all, the accused's withdrawal is not voluntary in such circumstances, but rather dictated by self-preservation. Although withdrawal from a common purpose is crucially different from withdrawal from an attempt (as is argued below), it is nevertheless noteworthy that similar policy concerns would apply - the withdrawal must be voluntary.

Lee raises a further consequentialist argument against allowing the abandonment (or as he refers to it, renunciation) defence (1997 Connecticut Law Review 145). He notes, using the analogy of the money-back guarantee, which encourages a purchase on the basis that the decision to purchase can always later be reversed, that such a defence may reduce the deterrent effect of the criminal law and "embolden otherwise ambivalent actors" (1997 Connecticut Law Review 145):

"The availability of a renunciation defense cannot entice people to take steps toward crime if they have no interest in committing crimes in the first place. But those who harbor some interest in committing crimes may decide that the renunciation defense effectively buys them more time to think."

A further argument in favour of the voluntary withdrawal defence is that the fact of withdrawal proves that the accused did not in fact have the intention at all material times to complete her act; in other words, that the accused's intention was not so strong as to "motivate" her to complete the crime (Fletcher Rethinking Criminal Law 187ff; Horder Ashworth's Principles of Criminal Law 524). After all, the argument contends, for a conviction of attempt to commit a crime, the State must prove that the accused had the intention to commit the completed crime, and not merely an intention to attempt to commit the crime (Snyman Criminal Law 285). It may be countered that this argument does not fully accord with the nature of attempt liability - it is not simply focused on the intent of the actor, but on whether the actor has expressed that intent in terms of identifiable objective criteria (see Lee 1997 Connecticut Law Review 140). Simply put, where the accused has reached the stage of commencement of the consummation of the intended crime, she is already blameworthy. Renunciation does not avail her. Neither does remorse (as Labuschagne (1995 Stell LR 199) notes, voluntary withdrawal does not of itself indicate remorse). Neither does previous good character. All these factors will no doubt play a role in mitigating sentence, but liability has already been established (see generally Yaffe Attempts ch 11). Burchell (Principles of Criminal Law 557) suggests that a subsequent change of mind can indeed override an earlier convergence of actus reus and mens rea, given that intention "involves a continuing state of mind" (original emphasis). However, this logic would inevitably require that any change of mind, at any stage of the course of criminal conduct, would negate liability. This is not seriously defensible. Once liability has been established, the point of no return has been reached.

Proponents of a voluntary withdrawal defence also struggle to answer satisfactorily a crucial question: how do you know why someone abandons their criminal purpose? (Simester and Sullivan Criminal Law 305). The Australian case of R v Page ([1933] Argus LR 374 (Victoria)) is instructive. The accused, who had climbed a ladder, put the point of a crowbar under a window frame. Then he climbed down, and claimed that thoughts of his mother brought him back to his good self. Or perhaps it was that he had noticed the police watching him?

The final salvo in favour of a defence of voluntary withdrawal is that it is inconsistent not to recognise renunciation as a defence to a charge of attempt given the existence of the defence in voluntary withdrawal from a common purpose to commit a crime (on this defence, see Burchell Principles of Criminal Law 501 -504; Snyman Criminal Law 263-264; Kemp et al Criminal Law in South Africa 281 -283) or from a conspiracy (on conspiracy, see Burchell Principles of Criminal Law 540-545, and specifically on withdrawal 544-545; Snyman Criminal Law 286-289; Kemp et al Criminal Law in South Africa 306-307). Both Burchell (Principles of Criminal Law 556) and Snyman (Criminal Law 286) highlight this aspect. No such inconsistency exists however. Withdrawal from a common purpose occurs before the crime has been completed. In respect of attempt liability, the rationale for such liability is to intervene in the criminal course of conduct at the earliest moment when there is sufficient evidence of a criminal purpose. The person who successfully withdraws from a common purpose could well be convicted for attempt, and so it should be if she has commenced the execution of the crime, but she could not be held liable for the completed crime unless the common purpose continues throughout the commission of the crime. And so it should be. As for withdrawal from a conspiracy, this can only successfully occur before the requirements for conspiracy liability have been met (by way of comparison, the US Model Penal Code at section 5.03(6) and section 5.02(3) respectively allows for renunciation as a defence relating to conspiracy and solicitation, provided the actor prevents the crime from being committed - see Lee 1997 Connecticut Law Review 119-120). A person who withdraws from a conspiracy before liability can ensue is in exactly the same position as one who withdraws from an attempt while still in the stage of preparation. No liability can apply, and there is no question of inconsistency of approach on this score.

 

6 Concluding remarks

In conclusion, it may be noted that, as Horder observes, those systems that have a voluntary withdrawal defence typically do not find it problematic, and members of the public in these jurisdictions seem to regard it as fair (Ashworth's Principles of Criminal Law 525). Moreover, the defence is rarely raised and considered in the courts, apparently affirming Fletcher's comment that the defence "is likely to remain an elegant artifact of criminal codes, viewed with admiration, yet rarely employed in the day-to-day affairs of the courts" (Rethinking Criminal Law 197).

Whether the defence really is or should be "viewed with admiration" is however uncertain, it is submitted. Having briefly adverted to the arguments generally raised for and against the defence in the previous discussion, the last section of this piece seeks to justify a final contention in favour of not allowing the defence in South African law, by examining selected theoretical and practical concerns.

In principle, if the voluntary withdrawal defence in the context of intent is compelled by the idea of recognising remorse, and responding with mercy (see Labuschagne 1995 Stell LR 199), then the same considerations should underpin an exculpatory defence in respect of completed crimes. However, as stated earlier, in relation to completed crimes, remorse is merely a mitigating factor. While withdrawal can reduce culpability (Simester and Sullivan Criminal Law 305) this begs the question why remorse in the context of attempt should be treated differently (i.e. as a substantive defence), as opposed to a factor that only has a bearing on sentencing. The question is further complicated by the existence of statutory offences that are formulated in terms not only of the completed act, but also an attempt to commit the completed act (see e.g., the offences in s 10 of the Sexual Offences Act 23 of 1957; s 47(2) of the Land and Agricultural Development Bank Act 15 of 2002; and s 215(2)(/) of the Companies Act 71 of 2008). The voluntary withdrawal defence has no application to these prohibitions. But if one were to allow a defence of withdrawal in the context of attempt, what would be the basis for distinction between the offence of attempt per se and these provisions? As Rabie points out, it would be much easier, and more consistent in principle, to apply the same approach to attempt generally, and for withdrawal to be considered, at most, in mitigation of sentence (1981 SACC 60).

As mentioned earlier, it might be argued that a distinction could be drawn between an actor who has already caused the harm, in the context of a completed offence, and one who has renounced his planned criminal course of action, and so has only inflicted limited harm, if any harm at all. It could be contended (as does Labuschagne (1995 Stell LR 197)) that the actor who withdraws from the attempt is no longer dangerous to the community.

However, this reasoning does not exactly accord with the approach adopted in the US Model Penal Code, and the Dutch and German law, which excludes the defence from those who withdraw "involuntarily", such as where the withdrawal is motivated by reasons pertaining to the probability of apprehension, or by a choice to postpone the criminal conduct to a more advantageous time. Surely this exclusion from the ambit of the defence recognises that those who withdraw on these grounds remain dangerous to the community, and like those who have already committed an offence, are blameworthy?

Another practical argument relates to the problems of law enforcement if such a defence existed, as accused persons caught before the final act could very well claim that they were in the process of withdrawing from their course of action and had no intention of going beyond the stage they had reached (Ormerod and Laird Smith, Hogan, and Ormerod's Criminal Law 434; Horder Ashworth's Principles of Criminal Law 524). Even Burchell (Principles of Criminal Law 556) concedes that the defence may be subject to abuse if applied in every case - but this then begs the question when it should apply.

It may be noted that the question whether such a defence ought to exist in South African law is not as obviously justifiable from the comparative sources as Snyman has suggested. In particular, as mentioned above, while the defence is favoured both in German law and in those US states following the Model Penal Code formula, the general approach to attempt liability in these jurisdictions is primarily subjective in nature, with a directed focus at the state of mind of the accused. The accepted approach in the South African law of attempt (with the exception of attempt at the impossible) is, however, objective in its orientation, as per the "commencement of the consummation" test applied to incomplete attempts (see, generally, Hoctor 2015 SACJ 363).

In fact, this point bears further amplification in that given that the objective "commencement of the consummation" test applies to both instances of interrupted attempt and voluntary withdrawal attempt, there is no good reason to deal with these categories of attempt differently. Moreover, the practical difficulties of distinguishing between these categories of attempt (see discussion at heading 2 above) militate against any efforts to do so. Thus, irrespective of whether the accused has experienced a moral epiphany, once the accused has been assessed to have gone beyond the preparatory stage of the attempt, "it makes no difference ... whether a plan is not carried out because of some outside intervention [interrupted attempt] or because the doer changes his mind [voluntary withdrawal]" (S v B supra 203G). Schreiner JA, on behalf of a unanimous Appellate Division court, clarified that the phrase in the Schoombie judgment describing the consummation as including "all the last series of acts which would constitute a continuous operation, unbroken by intervals of time which might give an opportunity for reconsideration" (Schoombie supra 547-548, cited in B supra 203C) should not be understood to mean that the mere chance for reassessment, even if "ample" and "leisurely" in nature (B supra 203D) is any obstacle in itself to the "commencement of the consummation" being attained:

"[T]he factor of intervals of time obtains such importance as it has rather from the notion of a normal break between preparatory acts and acts that begin the consummation than from the notion of the opportunity for reconsideration that time affords" (B supra 203E-F).

This approach was followed in R v Katz (1959 (3) SA 408 (C)), where the court explained that there should be no uncertainty "as to intent or desire to proceed to completion" as opposed to "mere uncertainty as to whether the machinery or means being adopted would in fact lead to completion" (424A). In this case, the court held that any opportunity for reconsideration afforded by a time interval "loses practically all its force through the fact that voluntary turning back by the accused was 'in the natural course of events' extremely unlikely" (424G). The inevitable conclusion of the court was that the accused had proceeded to the stage where they were guilty of an attempt. Where the accused has reached the point of no return then whether the subsequent interruption proceeds from an external source, or from within, should be irrelevant for the purposes of liability.

The submission is therefore that there is no good reason to treat voluntary withdrawal as a separate defence (Rabie 1981 SACC 61). Instead, withdrawal should at most amount to a mitigating factor in the inquiry into sentence (see Yaffe Attempts 291; Horder Ashworth's Principles of Criminal Law 524; for a contrary view see Brink ("First Acts, Last Acts and Abandonment" 2013 19 Legal Theory 114 120), who regards mitigation as unsatisfactory as it is "typically discretionary and partial"). Yaffe ("Trying to Defend Attempts: Replies to Bratman, Brink, Alexander, and Moore" 2013 19 Legal Theory 178 188) compares two hypothetical accused who commit exactly the same acts, and inquires why the mere renunciation of the attempt should mean that the conduct of one of these would not result in any criminal liability for attempt, as opposed to mitigation. It may be concluded that voluntary withdrawal from the course of criminal conduct is not "sufficiently fundamental to warrant a complete defence to criminal liability" (Horder Ashworth's Principles of Criminal Law 525); and allowing the defence amounts to logic giving way to policy (Ormerod and Laird Smith, Hogan, and Ormerod's Criminal Law 434). Whatever the accused's moral character before or after the attempted crime may be, criminal law concerns itself with one act at a time - this is the temporal logic of the law (Lee 1997 Connecticut Law Review 141). While there certainly may be reduced culpability for a genuine change of heart, once the accused's firm intention to commit the crime, as revealed in his conduct, amounts to an attempt, there should be no basis for a substantive defence of voluntary withdrawal in the context of attempt liability.

Shannon Hoctor

University of KwaZulu-Natal, Pietermaritzburg