In Anderton v. Ryan1 and Haughton v. Smith2 the House of Lords reintroduced the idea that there will, on occasions, be no liability for ‘attempting the impossible’. It came to that view even though the Criminal Attempts Act fairly clearly indicates that the opposite result is appropriate in cases of ‘missing circumstances-impossibility’ such as these were.3 In truth, as Chapter 4 must imply, the Act too misses the point; impossibility is irrelevant to an understanding of attempts. This is not merely to agree with some that impossibility should not affect our view of how blameworthy certain defendants are. It is rather to say that the ‘impossibility’ referred to in the context of criminal attempts is just not about attempts at all. In contrast, possibility is relevant to an understanding of attempts, but its relevance is universal. No normative or categorical distinctions among attempts can be made on the basis of possibility or impossibility.4
To understand the irrelevance of impossibility is to avoid a number of traps. The first is the inexplicably recurrent confusion between what truly is attempted (theft of a violin, say) and what actually occurs in an attempt (the case is empty, say) which leads to the fallacious conclusion that the defendant attempts the impossible; the courts have been no less immune to its allure than scholars. The second appears in analyses that propose categorical distinctions between possible and impossible attempts. That this is a trap ought to be clear from the inconsistencies that arise; anyone seeking to produce a defensible list of ‘impossible attempts’ drawing from the literature will be utterly frustrated. There is no communality, not even significant overlap, to observe. The third diversion is more attractive and involves assessment of what the phrase ‘attempting the impossible’ might mean. It appears compelling for if we can understand that much we may have a clearer sense of what such a category might consist in. Unsurprisingly, even this turns out to be a red herring. There is nothing that the phrase refers to.
To avoid the traps is to show that impossibility is a concept that does not belong to criminal attempts any more than it belongs to attempts more generally. It then must follow that putatively criminal attempts, which are treated as different from ‘central cases’ of attempts in virtue of impossibility, either (a) are not different at all or, (b) are different for reasons other than impossibility. If some attempts belong to category (b) then, crucially, it must be considered that the equal blame that ordinarily is taken to survive ‘impossibility’ does not survive these other reasons. The chapter concludes by examining whether and why some such extraordinary attempts ought not to attract criminal liability.
6.1 Ex post analysis in criminal attempts: practice and scholarship
As indicated in Chapter 4, the surprisingly frequent failure to observe the distinction between what was done and what was attempted is usually a symptom of perspectives that do not belong to attempting; these are represented most damagingly in objective standpoints and in ex post identifications. From these perspectives we may wish to conclude that where the defendant attempts on Sunday to rob a bank that physically holds no liquid assets on a Sunday, he attempts the impossible! If, instead, we adopt, the only available, subjective perspective identified in Chapters 3 and 4, the folly in this conclusion becomes clear.
Is the same error evidenced in legal practice? Consider the reasoning of Lord Hailsham in Haughton v. Smith:
…I do not think that it is possible to convert a completed act of handling, which is not itself criminal because it was not the handling of stolen goods, into a criminal act by the simple device of alleging that it was an attempt to handle stolen goods on the ground that at the time of handling the accused falsely believed them to be stolen.5
In Haughton v. Smith the substantive crime failed because the corned beef that the defendants believed to be stolen was in police custody at the relevant time; the defendants did not therefore handle stolen goods. Accordingly, there is naturally a valid description of the action that goes: ‘the defendants handled the goods that they wished to handle and these were not stolen goods’. However, if it is criminal culpability that interests us this is not a description we would give. The more natural accounts are: ‘the defendants handled goods believing them to be stolen’ and ‘the defendants attempted to handle stolen goods’. The first description need not entail the second for, as Chapter 4 indicated, only some beliefs will feature in our actions as a dimension of what we set out to do, and agents’ recognition alone is authoritative in this regard. But, here, it can hardly be doubted that the defendants set out to handle stolen goods; every aspect of their handling points to the same. If their general goal was simply to make money, they nonetheless had to set intermediate ends such that at some point they will have considered, ‘What kind of goods will we sell?’ A centrally relevant answer to this question is not; ‘goods that are in blue containers’ or ‘large goods’ or some such like, but ‘stolen goods’ or ‘goods that are likely to yield a high return; those that are not legitimate’. An end so construed is not undone by any more general orientating goals that the defendants have. So the decision is flawed in so far as it suggests that characteristic-based beliefs of this kind cannot also constitute what we set out to do; they can.
Of course this is not a dispute that gains anything – indeed it loses much – by being considered (as it here was) in the context of impossibility for it does not go to impossibility at all. The issues involved are more clearly addressed and resolved when understood (correctly) to be about the nature of attempting; about what it means to set ends of action; and about the extent to which these ends accommodate or overlap with beliefs or the objects of our other human capacities. The (false) observation that the defendants cannot attempt a crime where no crime is possible – the goods were not stolen as they believed them to be – adds nothing illuminating to this debate and obscures much. That was how the House of Lords conceptualised the matter nonetheless. In doing so, it drew a false distinction between this kind of attempt and others, missing the fact that no reason for the failure of an attempt, in this regard, is relevantly different from any other. Universally, it is not possible for the crimes that correspond to mere attempts somehow actually to be; the defendant could not have stolen the contents of the pocket [because there were none] (reason for failure). The defendant could not have handled stolen goods [for the goods were not stolen] (reason for failure). The defendant could not have hit his target [for he was prevented from doing so by the police] (reason for failure). In suggesting that the defendants in Smith attempted the impossible, The House of Lords conflates ‘what turns out to be the case’ (the goods were not stolen) with what is attempted (handling stolen goods), giving an objective account of the attempt which, as such, is categorically erroneous.
What about the possibility that there is a relevant distinction between acts that always diverge from what the defendant sets out to do and acts that only diverge from what the defendant sets out to do near their completion? Can it be said that if a defendant could never at any stage possibly be handling stolen goods (since they were always in charge of the police) that she cannot be attempting to do so, whereas the defendant who fails just at the last moment is attempting to steal from the pocket that turns out to be empty? Certainly Lord Hailsham appears, impliedly, to indicate that acts diverging, at the outset, from what is ‘intended’ or believed, cannot be attempts for they can never be crimes at all.6 Where in setting out permanently to deprive the owner, the defendant puts his hand in an empty pocket, the failure of the substantive actus reus of theft is caused, it may seem, near the end of its completion; only at this late stage is it found that there is no property to appropriate. The outcome sought is possible up until this point. It seems, by contrast, that in the case of handling goods that are wrongly believed to be stolen, the reason for (substantive) failure exists all along the actus reus, which could hardly constitute an actus reus, since at no time is there anything to which it can be proximate. There is never any possibility of success. This is a circular argument. It is only if we determine in the first place that the acts performed do indeed constitute an actus reus in the pickpocket example that we will find the existence of an attempt rather than a non-criminal act. But we may equally decide that since there is nothing in the pocket, there is never any possibility of picking it and, therefore, there is no possibility of actus reus. Likewise, by this reasoning, since there is no prospect of killing the target by aiming a firearm in the wrong direction, it may be said that there is no actus reus and thus no attempt. It is just not possible for a crime to succeed once it has not succeeded and the point during the attempt at which failure becomes real or manifest alters nothing in this regard.7
Various kinds of ‘impossible attempts’
Those decisions that followed Smith and Ryan reaffirmed the position in the Criminal Attempts Act that the defendant is to be ‘judged’ on the facts as he believed them to be.8 So, it seems that the defendant's belief that the goods are stolen, say, will suffice to show the presence of a criminal attempt in the relevant circumstances. But the Act itself is infamously opaque. Section 1 (3) provides that ‘a person may be guilty of attempting to commit an offence to which this section applies even though the facts are such that the commission of the offence is impossible’. Of course if it is accepted that in any mere attempt the actual commission of an offence is impossible, it must follow that the provision fails to illuminate anything. Most damagingly, the reference to ‘impossibility’ encourages the persistent, futile and apparently frustrating efforts to categorise attempts under the headings of possibility and impossibility; to identify a class of attempts that the Act might properly refer to.
Moreover, the Act unfortunately preserves the idea that facts matter in understanding what we set out to do; a defendant can ‘intend’ to commit an offence in virtue of the presence of relevant facts (the goods believed to be stolen are stolen) and not ‘intend’ to commit an offence where the same facts are absent (the goods believed to be stolen are not stolen).9 Of course there is a desire to retain culpability in the latter category and in order to do so the Act stipulates a requirement to impose an intention where, apparently, there is none. This approach to the facts is in keeping with Duff's in so far as it indicates that context, as much as content, can fix what we ‘intend’. The Act implies that if, in Ryan, the recorder was stolen, there would have been an ‘intention’ to handle stolen goods, that accompanied the defendant's belief in the same. Since the goods are not stolen and just believed to be stolen, there is no such ‘intention’ although we are entitled to impose one in virtue of the facts as the defendant believed them to be.10 In this way, the Act allows, what is, to Duff, the incorrect outcome. Whatever the proper contours of these provisions there can be no doubt that the Act (like Duff) errs in implying both that there is a distinctive category of ‘impossible attempts’ and that what I set out to do can be altered according to whether or not certain facts persist; no such transformation occurs.
The Criminal Attempts Act has the same effect as an outright denial of ‘impossible attempts’ (i.e., it very clearly preserves the category but prevents it from affecting liability). The category is equally well preserved among scholars of criminal law and in those cases that properly interpret the legislation. At the same time there are some vitally important insights that emerge following the erroneous application of the law in Smith and in Ryan.
In Shivpuri, the House of Lords upheld the conviction of a man who received, from India, imported harmless vegetable matter believing it to be heroin or cannabis. Lord Bridge considered that liability was warranted for if the facts had been as the defendant believed them to be he would have been guilty of the full offence. More importantly, Lord Bridge observed that any attempt to commit an offence that involves ‘an act that is more than merely preparatory to the commission of the offence but which for any reason fails so that in the event no offence is committed, must ex hypothesi from the point of view of the criminal law be objectively innocent. What turns what would otherwise, from the point of view of the criminal law, be an innocent act into a crime is the intent of the actor to commit an offence’. The principle here elucidated is convincing. Law's perspective is an external one; indeed, any perspective other than the attempter's is external to the attempt. In implying as much, Lord Bridge comes close to admitting the universal role of possibility in attempting; it is our external perspective that may lead us to identify ‘impossible attempts’ and to view the ‘objective innocence’ that the category mistakenly ‘reveals’. Now, once it is recognised that ‘impossibility’ and ‘objective innocence’ are actually universal in objectively observed attempts, it is a short if still complex journey to the conclusion that, in fact, possibility belongs to their anatomy.
If Lord Bridge advanced the underlying principles a great deal, the same cannot be said for all contributors to the debate. White suggests that unsuccessful attempts have either possible or impossible goals and that some attempts commonly thought to be ‘attempting the impossible’ actually have possible goals. In this category he includes an attempt to poison another using an innocuous substance, and lifting a rock that is too heavy. These goals are deemed possible. In the ‘impossible goals’ category he includes stealing one's own umbrella, picking an empty pocket and killing a dead man.11 The goals here set are deemed impossible. However, in the latter category the goals are misrepresented. Respectively our agents’ goals are, of course, to steal someone else's umbrella, to pick a pocket with valuable contents and to kill a living being. Furthermore, the distinction between the two categories seems not to be sustainable. Why equate the ‘goals’ with the reason for failure in the case of stealing one's own umbrella and not in poisoning with an empty substance? It is no more possible actually to poison someone with an innocuous substance than it is to steal one's own umbrella.12
George Fletcher distinguishes between apt and inapt attempts and suggests that there should be no liability for the latter category. Mistakenly shooting at stumps (rather than the victim) and ‘poisoning’ with sugar (rather than poison) are said to be inapt attempts in that ‘objectively’ they are unrelated to the ‘intent’ (to kill) whereas there is said to be nothing inapt about the empty pocket cases. Fletcher argues that apt attempts differ from inapt attempts in their propensity to generate harm. However, it is difficult to see why harm is generated by attempting to steal from a pocket that turns out to be empty and not by attempting to poison with a substance that turns out to be harmless.13 Of these cases it might be said (incorrectly) that they evidence no propensity to generate harm, for no harm occurs, or (correctly) a propensity to generate harm since it is to the defendant possible to bring about the harm sought. Neither is it clear why the poisoner is inapt and not the pickpocket. One might expect a pickpocket to be confident that the pocket actually contains something valuable as one might expect a poisoner to make sure that the substance actually contains a poison. Of course there may be good reasons to distinguish between offenders who are apt and those who are truly inapt but impossibility never provides such a reason.14
Gideon Yaffe identifies two sorts of impossibility. In the case of ‘circumstantial impossibility’ he evidences a number of commitments. Take the receipt of property that the defendant believes to be stolen but is not stolen. For Yaffe, it appears that the following propositions are true: the property could have been stolen; the attempt to handle stolen goods is, at the same time, impossible; guilt, generally, is present in such cases, although ‘missing circumstances’ can, if rarely, indicate that the defendant lacked the relevant ‘intention’ to commit the crime so that liability is not warranted, despite commission.
‘Inherent impossibility’, by contrast, refers to attempts that are not just impossible but ‘wildly impossible’, like killing by voodoo. These kinds of attempts are classified alongside but should be contrasted with attempts that involve ‘monumentally stupid plans’. A case in point concerns a man and his wife who make an insurance claim for a stereo that they had already sold, thereby attempting the offence of theft by deception.15 The actual offence could not be committed because the claim for the stereo took them above the maximum amount claimable.16 Both the ‘voodoo’ defendant and the ‘theft by deception’ defendants commit, what are by Yaffe's account, offences. At the same time, the former is not to be deemed guilty of an offence whereas the latter is. Whatever the merits of Yaffe's views on liability for impossible attempts and given his contention that, at least, some of these attempts could occur (which implies that some and not others are actually possible), it is clear that he certainly retains the category of ‘impossible attempts’ as a sensible one.
The disparity evidenced in these views and others is remarkable. For some ‘empty pocket’ cases are impossible attempts, for others they are not. The view persists that ‘poisoning’ with an innocuous substance is an impossible attempt, whilst those that deny the same nonetheless identify alternative instances of impossible attempts. Nor will it be clear the implications of impossible attempts; liability may be deemed to survive impossibility or not. All this confusion, and more, represents a failure to explore a principled analysis of the area, the potential for which ought to have been evident in Hart's work and in Lord Bridge's judgement.17 Moreover, Chapter 4 makes plain that these resurrections of impossibility represent a fundamental misunderstanding of the anatomy of attempting.
6.2 Blameworthiness for extraordinary attempts
Courts and commentators are reluctant to conclude that all ‘impossible attempts’ are culpable in the ordinary way; their hesitation reflects the view that some kinds of impossibility (perhaps voodoo say) are particularly troubling when it comes to matters of blame.18 The philosophical problem is that those who hesitate in this way offer no convincing criteria to distinguish between culpable ‘impossible attempts’ and non-culpable ‘impossible attempts’. It follows that either the hesitation is unwarranted or that reasons other than impossibility underpin it. If these reasons are legitimate then it seems likely that the law on attempts captures a (very narrow) subcategory of attempts that ought not to fall within the criminal definition. The claim that: ‘of course these are attempts, according to the legislation, but there would/should never be prosecutions’ is an unsatisfactory response to cases like voodoo or in respect of the jumper who wishes to cause genocide.19 It is equally an inadequate approach to the defendant who attempts to steal an umbrella, not knowing that it is his own, or to the accused who attempts to steal a violin from a case that turns out to be empty. Indeed the view that there should be no liability is certainly not at all convincing in the case of an attempted (consensual) rape where the victim happens to be above the age of consent, although the defendant believes that she is not.20 An explanation to account for the uniqueness of attempters in these cases or a basis for distinguishing among them is needed and that explanation cannot be founded on impossibility.
The previous chapter showed what a principled basis for culpability in criminal attempts, one founded on subjectivity, might look like. A defendant is culpable when she conceives of what she is doing in her attempt as actually bringing about the prohibited end. Moreover, it will be shown in the following chapter that when this holds true we may very well be justified in blaming the defendant as though the prohibited result did occur. Our sense of justification reflects the fact that such attempters, in bringing about the end set, demonstrate the same active commitment to harm that is realised in a substantive crime.21 The debate on moral luck can assist in bringing out the truth in this approach to blame; where a flying bird deflects the trajectory of the defendant's bullet, such that as a matter of luck he misses his target, he ought not to be rewarded in more lenient sentencing. Prima facie, if the subjectivists are right, the argument would also hold that a defendant should not be rewarded in virtue of the fact that the heroin turns out to be sugar, say. Arguably, he has just been morally lucky. However, the point to be determined here is whether ‘impossibility’ as such does something to affect liability or blame and what is of concern, therefore, is not that ‘impossibility’ may exist as a matter of luck,22 but whether in itself it is morally significant.23 An analysis of the anatomy of attempts disposes of this issue for it indicates that we attempt the possible. Impossibility does not distinguish the blameworthy from the not so or the less so for impossibility is simply alien to the concept of attempting.
It was mooted at the start of this chapter that if we cannot sensibly be said to attempt the impossible then criminal attempts (wrongly) so described are either indistinguishable from criminal attempts generally or differ for reasons not connected to impossibility. Moreover, if there are such differences, we may expect these to be relevant to the way in which we blame offenders. The suggestion is at least a plausible one that the relevant differences are found in the defendants who perform attempts rather than ‘in’ the attempts they perform. In the case of those extraordinary agents who attempt to kill by voodoo, say, we may find, by adopting Duff's objective perspective, that what they undeniably set out to do cannot reasonably be believed possible. A defendant who believes he can kill by shooting the victim in the heart and fails by his poor aim is, at least on the face of it, different altogether from a defendant who believes he can kill by voodoo. What are those who commit such extraordinary attempts culpable for and how can they be distinguished, if at all, from those who commit attempts in a rational or conceivably rational manner?
Harmfulness as a rationale for distinguishing extraordinary attempters
Usually a defendant who goes so far as to try to achieve a criminal end is harmful in respect of that end.24 Yet it seems that harmfulness is precisely what is missing in a narrow category of extraordinary attempts. The defendant who attempts to cause genocide by jumping up and down may be harmful, but he is unlikely to be harmful (indeed he seems impotent) in respect of causing the criminal end sought and it is that end which is relevant for criminal culpability. Similarly, the defendant who attempts to kill by voodoo may lack the harmfulness that is normally present in an attempter. Both defendants are engaged in ‘complete attempts’ but the former, at least, is harmless in respect of the end set. There is no immediate potential for the harm to occur nor, usually, is there any obvious suggestion that these sorts of defendants will try again by more effective means. In contrast, the defendant who attempts to shoot his victim not knowing that the target is just out of range is likely to try again and succeed. The potential for harm exists in the immediately morally real sense that the defendant (or victim) may move within range and in the wider morally real sense that the defendant, by these actions, demonstrates that he poses a danger to the community. He lacked neither commitment nor the ultimate capacity to know how to bring about the harm desired. Here we have a central case example of an attempt and the defendant is just ‘lucky’ to avoid the prohibited outcome.
If harmfulness distinguishes some attempters from others it is important to be clear about how to identify that harmfulness. In substantive crimes actual harm is apparent in the facts. We know that the defendant is harmful because she has in fact brought about the harm prohibited. In attempts we cannot infer harmfulness from actual criminal harm (for the harm attempted does not materialise). Equally, we should not assume that merely trying to cause criminal harm makes one harmful in respect of that harm. So, we cannot assume necessarily that the elements constituting a criminal attempt are sufficient to account for harmfulness. When I try to play chess, by playing snakes and ladders, I demonstrate, by my actions, an inability not only to play chess but to know how to go about playing chess. There is no potential ‘danger’ of me succeeding in my objective. I am harmless despite the fully committed nature of my trying. When, by contrast, I try to play chess but fail because a piece is missing, I have demonstrated an understanding of what it means to play chess, I understand the implications of a missing piece and my trying, therefore, is evidence of some, at least basic, ability in respect of the end. Usually in criminal attempts too the defendant's harmfulness is unaltered by the reason for failure (where she just misses her target, say) and it is sufficiently demonstrated in her trying itself. Equally, it appears likely that in some extraordinary attempts the reasons for failure may provide evidence of/disclose the defendant's inability to be harmful in a criminally relevant way in the first place. Where I attempt my end by voodoo, for example, there is something about the nature of my attempt that is relevant to an understanding of my harmfulness and I may very well be in an analogous position to the person who attempts to play chess by playing snakes and ladders.
Harmfulness deriving from what could have happened?
In an attempt, where the harm does not materialise, it is tempting to consider that the defendant is harmful if she: ‘could have caused harm’. In other words, it is tempting to consider whether the actions that occurred (rather than the actor) were themselves harmful. But this is just to raise once more the tricky concept of ‘possibility’. ‘What could have occurred’ offers no basis for distinguishing the harmfulness of various attempts because there is just no reason to permit the counterfactual in our analysis of some ‘particular’ attempts and not others. As Chapter 4 noted, we cannot say, for example: ‘if the victim was within range, as the defendant believed, the crime could have occurred’ unless we are prepared to say the same whether the victim is 1 cm out of range or 300 miles out of range or on the moon.25 If we still suspect that there is a lack of potential for harm in some of these attempts and not others the distinction cannot lie in what ‘could have been’.
Harmfulness deriving from proximity of the attempt to the end sought?
A further source of harmfulness may seem to lie in the proximity of the specific attempt to the end sought. Where the defendant shoots just out of range he gets close to achieving his purpose. He needs only to move physically to be within range. In contrast, there is nothing the defendant feasibly can do to make killing the man on the moon more proximate. Similarly the jumper cannot, by jumping more forcefully on the earth's surface, bring about mass genocide of the human race. However, this relation does not tell us about those who shoot well out of range but at a distance one may reasonably think feasible and it does not dispose of the complication that someone who thinks he can kill, whilst very obviously far removed from his target, will certainly know how to kill within range. Moreover, the distinction is just very poor, rendered linguistically plausible in examples apparently involving matters of degree only through the term ‘close’. The fallacy becomes clear when it is considered that where the case is empty, a person who attempts to appropriate a violin from it is just as ‘far’ from achieving his end as an agent who attempts to appropriate an elephant from that case or an agent who attempts to shoot the moon. People who shoot at others, however far away they are, seem by any sensible account to be harmful in respect of shooting. In a similar vein, the defendant who imports sugar rather than heroin may in the past have successfully and often imported heroin and he may succeed in the future too.26 He has been harmful and by any ordinary understanding he is harmful. Furthermore, there is some link between the act he actually did (importing sugar believing it to be heroin) and his harmfulness so that in describing the defendant as harmful we are speaking of a harm that is relevant by virtue of its connection to the attempt. Yet there is nothing that this defendant can do to make sugar be anything other than sugar. There is no proximity between the facts and what the defendant set out to do.
Harmfulness deriving from attempters
Normally, a defendant who tries sufficiently to achieve a criminal purpose, and who grasps the basic operations of cause and effect, is harmful. The potential for harm inheres in the individual. Such an individual not only has tried, in their attempt, to bring about harm they also have some, at least minimal, idea of how effectively to do it. The person who sets out actually to do something that they know cannot be done – the absurd actor whose attempt at the impossible is self-contradictory – does not understand the basic relations of cause and effect and does not in fact perform an attempt at all. Indeed he thinks something like: ‘I can bring something about without bringing it about.’ Attempts involving voodoo, for example, are not self-contradictory in this way but come close to the same level of absurdity. Here, the attempter understands cause and effect but is incapable of connecting the reasons for the failure of his attempt to that failure. More so, these same reasons at least intuitively seem somehow implicated in the attempter's setting of his end in the first place. This defendant is insufficiently harmful because (a) even with full knowledge, he cannot understand that such an attempt will lead to failure, and (b) he performs his attempt because of those factors that, unbeknown to him, lead inevitably to failure. An attempter of this sort is ineffective and may as well lack the ability to understand cause and effect. Two questions may be asked to determine whether an attempter lacks the harmfulness to be liable for an attempt:
(a) Had the defendant been aware of the facts that caused the substantive crime to fail would the defendant have understood or agreed that the crime failed because of those facts?
(b) Was the defendant's decision to act a product of his commitment to the facts as he believed them to be?
If the answer to (a) is ‘no’ and (b) is ‘yes’ then the defendant is insufficiently harmful to be held liable for the attempt. For most attempts, including those sometimes thought to fall into the impossibility category, (a) just cannot be satisfied. A defendant who is caught in his attempt will realise that he failed because he was apprehended. A defendant who is told: ‘this gun has a range of 15 feet and you were 16 feet away’ will know why the attempt failed. Similarly, defendants knowing that the goods are not stolen after all, that the umbrella is their own, that the aim was misdirected and the murder victim, actually a corpse, will realise that these facts caused their substantive crimes to fail. All these defendants display not a lack of capacity for harmfulness in respect of the ends sought. They demonstrated rather a failure effectively to execute their plans, or they were merely lucky to avoid committing a substantive offence and for this they ought not to be rewarded. Where there is some semblance of rationality involved in the defendant's intentional action, he is harmful in respect of the prohibited end. Not only is he bringing about X, he also has some idea of what really it might mean to bring X about. The facts that lead to failure do nothing to affect the criminally relevant harmfulness that inheres in such individuals.
In contrast the defendant who is told: ‘you attempted to kill by jumping up and down on the earth's surface’ will most likely not understand the connection between failure and these facts. Even if he has some understanding of the physical world he may not agree that its operations of cause and effect were the reasons for failure; perhaps he may consider that the unicorn who normally gives him special powers failed, on this occasion, to do so. In the voodoo example, too, the defendant will not appreciate or agree that the attempt is bound to fail when told of the facts that lead to failure. Now this defendant may well know that society at large does not believe that voodoo can cause death, but he is not interested in how other people treat cause and effect; he has his own concept of how ends are brought about.
In (b) the issue is to determine whether the defendant is harmful notwithstanding the inability to appreciate why failure ensued. If, for example, the decision to act is unconnected to the facts that produce failure then it may be regarded as a matter of luck only that the failed method, however bizarre, was selected and that failure therefore ensued. (The defendant who does not understand the relevance of being out of range may however be just as willing to shoot within range.) The feature of certain attempts that (b) is designed to accommodate is the connection between ends that cannot be brought about in a certain way and the decision to attempt those ends in that particular way. For example, it seems to be true of voodoo that the setting out to kill is not something that is done independently from the means selected to achieve the end. It is rather an element of the commitment to those means. This is a complex connection, but it at least seems likely that such attempters do not decide to kill and only then choose voodoo as the most appropriate method. Rather, the decision to kill is bound up in the commitment to voodoo in the first place.27 The defendant's commitment to his own ability is almost the motivator that, ironically, makes the decision to act a feasible one to the attempter. For the jumper too it is significant from the outset that he decides to kill in this way. There is something that connects his wanting to kill to his perceived ability to cause death by disrupting the earth's atmospheric stability. Here we can say that the defendant's commitment to certain facts accounts for his embarking on this attempt. These defendants are not merely utterly incapable; they are, by any account, wholly extraordinary; a reason for distinguishing them from other attempters can be found in who they are.
Hart hints at the likelihood that ‘impossible attempts’ are just attempts in the ordinary sense:
…there is a sense of ‘impossible’ in which it is true to say of many clear cases of attempt that it was impossible in the circumstances for it to have succeeded. A man shoots to kill but the bullet is stopped by the cigarette case in the victim's breast pocket; or a burglar attempts to open a safe with a jemmy that snaps in the door. It is impossible ‘factually’ or ‘physically’ (or perhaps more perspicuously ‘causally’) that in precisely those circumstances the shot should have killed the victim or that the burglar should have opened the safe. But when empty pocket cases are specially singled out as cases of ‘factual’ or ‘physical’ impossibility, it is left quite unclear how this differs from the impossibility of success in these ordinary cases of attempt.28
It is to be regretted that Hart did not go further explicitly to expose the fallacy of ‘attempting the impossible’, but it is surprising nonetheless that the categories he raised to expose their lack of specificity are resurrected in the literature. ‘Impossible attempts’ are forensically dissected as though there is a need to explain away the apparent paradox between ‘attempting the impossible’ and the ‘surviving culpability’ that we are willing to attribute to such attempts. In truth, there is no paradox to be explained away for the category is a mythical one. There is, however, a real category of extraordinary offenders and it is their extraordinariness that stands in need of explanation.
6.3 Legal impossibility
There can be no liability for that category of attempts that has come to be defined as ‘legally impossible’.29 These attempts are said to be indicated where a person sets out to do what they believe to be a crime, which is not in fact a crime. So, Bill may attempt adultery believing that committing adultery is a crime.30 Can Bill attempt adultery? Yes. Can Bill attempt adultery as a crime? Yes. Is Bill therefore attempting a crime? Yes. Is Bill liable for an offence? No. The answers to the first three questions are ‘yes’ for the reason that the chapters on possibility and subjectivity make clear. My attempt is set by me and it is possibility to me that matters.
Why, then, given the ‘yes’ answer to these questions is there no liability falling on Bill? Why is the answer to the final question, ‘no’? The absence of liability is attributable solely to the fact that it is not an offence to attempt a crime where the end set is not in fact a crime. Sure, I can attempt to commit the offence of wearing purple trousers just in the same way as a defendant can attempt to kill by voodoo or by shooting at his victim. The distinction between the two has nothing to do with kinds of impossibility, or with impossibility at all; it is simply that to attempt to commit the crime of wearing purple trousers is no crime whereas to attempt to kill is a crime. Essentially, the phrase ‘legal impossibility’, in so far as it is attributed to the nature of the attempt, is even more unhelpful than the phrase ‘factual impossibility’.31 Consider how things would change if I actually succeeded in wearing purple trousers or where a person does commit adultery. In neither case would there be an offence, regardless of what the agents believed (had attempted). This should not, of course, lead us to conclude ‘it is not possible to commit the offence so it is not possible to attempt it so there should be no liability’. It is entirely possible for agents to attempt crimes of adultery and purple-trouser wearing. The relevant point is just that the attempt and the acts alike are no crimes.
