CV Text of Publications Links Home

John Hasnas Associate Professor McDonough School of Business Georgetown University

Once More onto the Breach: The Inherent Liberalism of the Criminal Law and Liability for Attempting the Impossible

By John Hasnas ⁽¹⁾

I. Introduction
 

Jean-Claude, a practitioner of voodoo who fled to the United States from Haiti to escape the oppression of the Duvalier regime, passes a man on the street that he recognizes as a former member of the Tonton Macoutes, Duvalier's secret police. Consumed with hatred and the desire for revenge, he goes home, fashions a doll in his enemy's image, and drives a needle through its heart. He does so with the intention of producing his enemy's death, which he firmly believes will result from his action. His wife, another believer in voodoo, is horrified by his action and turns him into the police, who arrest him for attempted murder. Jean-Claude is tried and convicted.

Laura and Frank are rabid anti-environmentalists who believe there should be no restrictions on people's ability to utilize natural resources. Intending to engage in an act of civil disobedience, they go deer hunting on October 31. Although hunting season runs from October 1 until October 31, Laura mistakenly believes the season runs from September 15 to October 15, while Frank, who has just returned from Japan, mistakenly believes the date to be November 1. While hunting, they encounter a forest ranger to whom they proudly confess their defiance of the law. The ranger promptly arrests Frank, who is tried and convicted for attempting to hunt out of season, but does not interfere with Laura, who has committed no crime.

A former dot com millionaire, who has been reduced to liquidating his property to stave off bankruptcy, parks his SUV along a crowded street and offers to sell various electronic gadgets to passing pedestrians. Jenny is one such pedestrian to whom he offers his 700 MHz laptop computer for $400. Jenny concludes that the computer must be stolen property, but buys it nonetheless. Thrilled with her purchase, she shows the laptop to a man coming from the opposite direction and says, "Hey, there's a guy selling 'hot' electronic gear down the street for unbelievable prices. I got this laptop for only $400." Unfortunately for her, the man is an undercover police officer who places her under arrest. Jenny is tried and convicted of attempting to receive stolen property.

Clarisa, the long suffering wife of a philandering husband is in the habit of preparing him a cup of coffee with one spoonful of sugar each morning. However, when he comes home at 3:00 am with lipstick smeared on his shirt collar, she decides that this is the final straw. The next morning, she stirs what she believes to be a spoonful of the arsenic she had purchased for this eventuality into his coffee. When her husband gets up from the breakfast table, kisses her on the cheek, and leaves for work, she realizes that she mistakenly added sugar to his coffee, just as she does every morning. Taking this as a sign from god to repent, she immediately goes to the police station to confess. Clarisa is arrested for attempted murder and is tried and convicted. ⁽²⁾

With the exception of Laura, the protagonist in each of the above dramas is being subjected to criminal punishment for attempting crimes that could not possibly be produced by their actions. Should they be? There is little question that in most American jurisdictions the defendants would have no grounds on which to appeal their convictions. Thirty-seven states have explicitly eliminated impossibility as a defense to a charge of attempt ⁽³⁾ and the federal circuits that have not done likewise have so limited the range of application of the defense as to render it virtually a dead letter. ⁽⁴⁾ As a result, one's susceptibility to punishment for attempting the impossible is today a rather uncontroversial matter of settled law.

This was not always so. Thirty-four years ago Graham Hughes wrote that,

[t]he relevance on a charge of criminal attempt of the impossibility of the accused's attaining his objective has for some time been a subject of sharp dispute among jurists of the criminal law, . . . That teachers of criminal law and writers in the field should devote time and energy to this question is perfectly proper for it is an important question in a number of ways. It raises very basic interrogatories concerning the aims and purposes of the criminal law; it compels us to focus attention on concepts such as "intention" and "purpose," an analysis of which is indispensable to criminal law scholarship; and it provides an excellent opportunity for reflecting on the pervasive and difficult distinction between mistake of fact and mistake of law. For these reasons the problem is a splendid set-piece which exhibits in a short space some of the most difficult issues of criminal law analysis. ⁽⁵⁾
 

How did the "splendid set piece . . . of criminal law analysis" of a generation ago come to be a matter of routine today?

The answer apparently lies in the confusion and frustration courts experienced when trying to apply the common law rule governing impossible attempts. The rule itself was clear. Legal impossibility constituted a defense to a charge of attempt; factual impossibility did not. ⁽⁶⁾ An attempt was legally impossible "where the intended acts, even if completed, would not amount to a crime." ⁽⁷⁾ An attempt was factually impossible "when extraneous circumstances unknown to the actor or beyond his control prevent consummation of the intended crime." ⁽⁸⁾ The confusion arises when courts try to apply the rule to cases because the categories are not mutually exclusive.

Consider, for example, the case of Jenny, the laptop purchaser, described above. She completed all of her intended acts which did not amount to a crime. Therefore, her attempt was legally impossible. ⁽⁹⁾ On the other hand, the fact that the laptop was not stolen property was an extraneous circumstance unknown to the her that prevented the consummation of the intended crime. Therefore, her attempt was factually impossible. ⁽¹⁰⁾ Consider also Clarisa, the long suffering wife. She too completed all of her intended acts which amounted to sweetening her husband's coffee as she does every morning; surely no crime. Therefore, her attempt was legally impossible. However, she failed to murder her husband only because she mistook sugar for arsenic, an extraneous circumstance that prevented the consummation of the crime. Therefore, her attempt was factually impossible. As these examples suggest, whether an attempt is legally or factually impossible appears to depends more on the way the court chooses to characterize the defendant's actions than on the defendant's actions themselves.

This situation produced a howl of frustration from the judges who had to apply the rule. Characterizing legal and factual impossibility as "logically indistinguishable" ⁽¹¹⁾ and the distinction between them as "a matter of semantics," ⁽¹²⁾ courts found the common law rule to be a "morass of confusion," ⁽¹³⁾ and "a source of utter frustration." ⁽¹⁴⁾ One court summed up the situation by stating,

[T]he application of the defense of impossibility is so fraught with intricacies and artificial distinctions that the defense has little value as an analytical method for reaching substantial justice. . . . We think the effort to compartmentalize factual patterns into these categories of factual or legal impossibility is but an illusory test leading to contradictory, and sometimes absurd, results. ⁽¹⁵⁾
 

In the judicial equivalent of throwing up one's hands, courts both rejected the impossibility defense whenever possible ⁽¹⁶⁾ and urged legislatures to supplant the common law doctrine with a "progressive and more modern view." ⁽¹⁷⁾ The majority of state legislatures obliged, ⁽¹⁸⁾ usually adopting an attempt statute patterned on §5.01 of the Model Penal Code that eliminates the impossibility defense. ⁽¹⁹⁾ Thus, courts now rarely if ever have to address the "splendid set piece . . . of criminal law analysis" of a generation ago.

In this article, I will suggest that this is not necessarily a good thing. With a little help from our fictional friends, Jean-Claude, Laura, Frank, Clarisa, and Jenny, I hope to show that there was both significant wisdom in and good normative grounding for the common law impossibility defense. In Part II, I will argue that the distinction between legal and factual impossibility is not nearly as impenetrable as contemporary commentary suggests, and that, in fact, the essence of the distinction can be encapsulated in a single sentence. In Part III, I will argue that the confusion regarding the distinction arose from the difficulty the early courts had in giving a clear explanation for the basis of their decisions. I will claim that the courts were saying one thing while doing another, and that by paying attention to what the courts said rather than what they did, subsequent courts and commentators stumbled into their "morass of confusion." ⁽²⁰⁾ In Part IV, I will examine the normative argument offered for the rejection of the impossibility defense by the authors of the Model Penal Code and other theorists who adopt what is known as a subjectivist approach to the crime of attempt. In doing so, I will review the main objections to the Model Penal Code's subjectivist approach and catalog the responses the authors of the Code can make to each. In Part V, I will argue that there is a principled distinction between moral and criminal responsibility, that the normative arguments supporting the Model Penal Code's subjectivist approach are based on an improper conflation of the two, and that a correct understanding of the nature of criminal responsibility leaves the Code's position without normative support. In Part VI, I will argue that there is an inherent liberal bias built into the Anglo-American criminal law that supplies a principled basis for rejecting the Model Penal Code's definition of attempt. In Part VII, I will suggest a proposed definition for attempt that encompasses and extends the common law defense of impossibility on a firm theoretical basis. Finally, in Part VIII, I will conclude.

In the Introduction, I suggested that the courts lost patience with the common law doctrine of impossibility because the categories of legal and factual impossibility are not mutually exclusive. Indeed, given the definitions assigned to these terms by the common law decisions, they are not. But this implies only that the courts did a poor job of labeling the distinction they were making, not that no viable distinction exists. On the contrary, with some adjustment to our terminology, it is quite easy to identify the class of cases in which impossibility functioned as a defense at common law from those in which it did not.

A perhaps apocryphal story about football coach Vince Lombardi holds that following a bad game by the Green Bay Packers, he began a team meeting by saying, "Let's get back to fundamentals. This is a football." I believe an analogous approach would helpful to elucidate the substantive features of the impossibility defense to attempt. This means beginning with the observation that, "Attempt is a crime." As such, it consists in a guilty act, an actus reus, undertaken with a guilty state of mind, a mens rea. The actus reus encompasses more than merely the bodily movement of the defendant, but refers to an action that produces prohibited consequences and occurs under legally specified circumstances. Thus, the actus reus of a crime consists in three constituent elements: 1) a physical action, 2) its consequences, and 3) its attendant circumstances. ⁽²¹⁾ The mens rea describes the state of mind the actor must have with regard to each element of the actus reus.

Professor J. C. Smith has drawn a distinction between two types of attendant circumstances that is useful for present purposes. Smith distinguishes "pure" from "consequential" circumstances. Pure circumstances are "those the existence of which is not essential to the occurrence of the consequences but is relevant to the legal effect of the consequences. The relevance of these circumstances lies simply in that their existence is required by the definition of the crime as a condition of liability." ⁽²²⁾ An example of a pure circumstance would be the requirement for the crime of burglary that the breaking and entering take place at night. This circumstance has no bearing on whether the consequences of the burglary, the breaking and entering, occur, but because it is explicitly required by the definition of the crime, there can be no liability for burglary without it. Consequential circumstances are "those the existence of which is essential to the occurrence of the consequences" but whose "existence is not necessarily required by the definition of the crime, . . ." ⁽²³⁾ Examples of consequential circumstances would be: on a charge of murder, whether the gun to be used is loaded and "on a charge of larceny, whether there was anything in the pocket into which D put his hand with intent to steal." ⁽²⁴⁾

Because attempt is a crime, it can be fully described by specifying its constituent actus reus and mens rea. What then is the actus reus of attempt? Although it is too early in our analysis to answer this question with precision, ⁽²⁵⁾ we can say at least that the actus reus of attempt consists in trying and failing to commit another, completed crime. The mens rea of attempt is a bit easier to specify. Attempt is a specific intent crime, one whose mens rea requires two instances of intention. To be liable for a specific intent crime, one must 1) intentionally produce the actus reus of that crime while 2) entertaining a further intention to do something or produce certain consequences in the future. ⁽²⁶⁾ This further intention is the specific intent. In the case of attempt, the required specific intent is the intent to commit the completed crime. ⁽²⁷⁾ Thus, the mens rea of attempt consists in intentionally taking the actions that constitute the actus reus of attempt, i.e., the actions of trying and failing to commit a completed crime, with the further intention to thereby commit that crime. Note that this implies that to be culpable for an attempt, one must act with the specific intention to produce a state of affairs prohibited by law.

This review of criminal law fundamentals should be sufficient for us to unravel the common law doctrine of impossibility. Consider first Laura, our environmental rebel, who was trying to hunt out of season. Laura's mistake about the law prevented her from realizing her objective. Laura was not only not convicted of attempting to hunt out of season, she was not even arrested. Why not?

The simple answer is that Laura does not possess the mens rea of attempt. It is fair to say that Laura tried and failed to hunt out of season, and hence that her actions constitute the actus reus of an attempt. It is also fair to say that because she took these actions intentionally, she has one of the two instances of intent requisite for an attempt conviction. What is missing, however, is the specific intent. To be culpable, Laura must also intend to commit a completed crime, i.e., to produce consequences that are prohibited by law. Despite the fact that Laura intends to produce consequences that she mistakenly believes to be prohibited by law, she does not intend to produce consequences that are prohibited by law. Therefore, she does not possess the specific intent required by the mens rea of attempt. ⁽²⁸⁾

Laura has attempted to commit an imaginary crime. ⁽²⁹⁾ One attempts an imaginary crime when one makes a mistake of law such that he or she "either believes that he [or she] violates a criminal prohibition which, in reality, does not exist, or . . . wrongly expands the scope of an existing criminal statute to his disadvantage." ⁽³⁰⁾ One who attempts to engage in illegal gambling in Nevada or violate a motorcycle helmet law in a state that has not enacted such a statute would be attempting to commit an imaginary crime. Laura's mistake caused her to "wrongly expand the scope of an existing criminal statute to [her] disadvantage." ⁽³¹⁾ Our analysis suggests that there should be no culpability in cases of imaginary crimes because the defendant does not possess the mens rea of attempt. Thus, as common sense would suggest, attempting to commit an imaginary crime is not attempting to commit a crime. ⁽³²⁾
 

Consider next the case of Clarisa, our aggrieved wife. Clarisa has made no mistake of law. What she tries to do, murder her husband, is certainly prohibited by law. However, due to a mistake about the condition of the world, i.e., the location of the arsenic, the means she employs can not possibly achieve her end. Clarisa has made a mistake of fact that prevents her from producing the consequences she desires which would constitute a murder. Employing Professor Smith's terminology, we can say that Clarisa is laboring under a mistake of fact with regard to a consequential circumstance.

By stirring sugar into her husband's coffee, Clarisa tired and failed to kill him. It is therefore fair to say that her actions constitute the actus reus of attempted murder. Further, Clarisa clearly possesses the mens rea of attempt. She intentionally performed the actions that constitute the actus reus of the attempt, stirring sugar into her husband's coffee, and she did so with the further specific intention to thereby cause his death. Hence, Clarisa appears to be a good candidate for conviction for attempted murder.

Due to a mistake about a consequential circumstance, Clarisa has attempted to commit a crime by means that make it impossible for her to succeed. Other cases that share the characteristics of Clarisa's failed attempt are attempting to pick an empty pocket, ⁽³³⁾ or to commit murder using an unloaded gun. ⁽³⁴⁾ Our analysis suggests that in cases of this type in which the impossibility results from a factual mistake about a consequential circumstance, conviction for attempt is appropriate.

But now consider the case of Jenny, our pedestrian laptop purchaser. Like Clarisa, Jenny is not mistaken about the law. What she is trying to do, receive stolen property, is legally prohibited. Also like Clarisa, Jenny has made a mistake about the condition of the world, i.e., whether the laptop she purchased had been stolen. Unlike Clarisa, however, Jenny's mistake of fact did not prevent her from producing the consequences she intended to produce, i.e., obtaining a laptop at a low price. It did, however, prevent her from thereby producing the actus reus of the crime of receiving stolen property. This is because Jenny was laboring under a mistake of fact with regard to a pure circumstance, that is, a circumstance "the existence of which is not essential to the occurrence of the consequences but is relevant to the legal effect of the consequences." ⁽³⁵⁾

It appears that Jenny had the mens rea required for an attempt. She intentionally took the actions by which she received the laptop and she did so with the further specific intention to thereby receive stolen property. But what can we say about whether her actions constitute the actus reus of an attempt to receive stolen property? On the one hand, Jenny successfully accomplished her objective of obtaining a laptop computer at a very low price. She does not appear to have failed at anything at all, and had she known that the laptop was not in fact stolen property, she would have been even more delighted with her purchase. Given this, it seems odd to characterize Jenny's activities as trying and failing to commit the crime of receiving stolen property. Thus, one could argue that her activities do not constitute the actus reus of attempt. On the other hand, because Jenny believed that the laptop was stolen property, it is not unreasonable to say that she was trying to receive stolen property, and, because the laptop was not stolen property, that she failed to accomplish this. Thus, one could also argue that her actions do constitute the actus reus of an attempt.

It appears that whether Jenny has produced the actus reus of an attempt to receive stolen property depends on whether the object she seeks to obtain is characterized as a low priced laptop or as putative stolen property. Fortunately, it is not necessary to resolve this question at the moment. For now, we can simply say that a preliminary analysis in terms of the fundamental concepts of criminal law leaves us uncertain as to whether Jenny should be liable to conviction for attempting to receive stolen property.

Due to a mistake of fact about a pure circumstance, Jenny has obtained an objective that she believed to be criminal without thereby committing a crime. Other examples of this type of impossible attempt are attempting to steal one's own umbrella ⁽³⁶⁾ or to distribute a controlled substance by distributing an uncontrolled substance. ⁽³⁷⁾ At present, our analysis leaves us uncertain as to whether there should be a conviction in cases such as these in which the impossibility results from a mistake of fact about a pure circumstance.

We are now in a position to give a clear account of the common law impossibility doctrine. At common law, a factually impossible attempt was one in which the impossibility arose from a mistake of fact with regard to a consequential circumstance, the type of mistake Clarisa made. Factual impossibility was no defense at common law, a result which squares with our analysis of Clarisa's situation above. A legally impossible attempt was either an imaginary crime or one in which the impossibility arose from a mistake of fact with regard to a pure circumstance, the type of mistake Jenny made. Legal impossibility was a common law defense, a result which squares with our analysis of imaginary crimes such as Laura's, and resolves our uncertainty about Jenny's situation in favor of finding no liability.

The common law decisions overwhelmingly support this analysis. With the exception of a few nineteenth century English cases, ⁽³⁸⁾ almost all cases in which an attempt was impossible because of the defendant's mistake of fact with regard to a consequential circumstance have been classified as factually impossible and the conviction upheld. Thus, attempts to pick empty pockets, ⁽³⁹⁾ perform abortions on women who were not pregnant, ⁽⁴⁰⁾ commit rape when impotent, ⁽⁴¹⁾ and kill someone already dead, ⁽⁴²⁾ or with an inadequate dose of poison ⁽⁴³⁾ or an unloaded gun, ⁽⁴⁴⁾ or by shooting into an empty bed ⁽⁴⁵⁾ or in the wrong direction ⁽⁴⁶⁾ have all been held to give rise to attempt liability. ⁽⁴⁷⁾ In addition, almost all cases in which attempt convictions have been reversed on the grounds of legal impossibility are either cases of imaginary crimes or those in which the defendant made a mistake of fact with regard to a pure circumstance. Thus, convictions have been reversed for attempting to commit forgery by altering a non-material portion of a check, ⁽⁴⁸⁾ attempting to receive stolen property by receiving non-stolen property, ⁽⁴⁹⁾ attempting to suborn perjury by soliciting false testimony on an immaterial matter, ⁽⁵⁰⁾ attempting rape by having forcible intercourse with one's wife, ⁽⁵¹⁾ attempting to distribute a controlled substance by distributing an uncontrolled substance, ⁽⁵²⁾ and attempting to smuggle letters out of prison without the warden's knowledge when the warden was aware of the smuggling. ⁽⁵³⁾ Far from involving the "highly abstract issues that are notorious for their degree of difficulty" ⁽⁵⁴⁾ that were decried by courts and commentators, the common law distinction between factual and legal impossibility is really quite simple. Because cases of true imaginary crimes almost never arise, courts can accurately and easily categorize attempts as legally or factually impossible by focusing on whether the defendant was mistaken about the existence of a pure or consequential circumstance. Discounting imaginary crimes, the common law doctrine of impossibility can be summed up in a single sentence: impossibility is a defense to a charge of attempt if it results from the defendant's mistake concerning a pure circumstance.

In Part II, I argued that impossibility constituted a defense to a charge of attempt at common law if it arose from the defendant's mistake concerning the existence of a pure circumstance. If things really are as simple as that, what accounts for all the confusion over the issue? After all, not only the courts regarded the doctrine of impossibility as a "morass of confusion." ⁽⁵⁵⁾ Virtually all academic commentators who address the subject begin with a reference to the conceptual difficulty of the distinctions involved. Thus, the impossibility doctrine is declared to be "quicksand" ⁽⁵⁶⁾ or "an intellectual quagmire" ⁽⁵⁷⁾ that "has resulted in extreme confusion" ⁽⁵⁸⁾ and constitutes "the most intractable problem of all." ⁽⁵⁹⁾ How could what is essentially a simple distinction give rise to so much intellectual tumult?

The answer may be as simple as that courts and commentators paid too much attention to what judges said and not enough to what they did. To explain what I mean, let me begin with a brief digression on the behavioral significance of making a mistake about a consequential as opposed to a pure circumstance. Recall that a consequential circumstance is one "the existence of which is essential to the occurrence of the consequences" ⁽⁶⁰⁾ of the completed crime. Thus, defendants who are mistaken about the existence of a consequential circumstance are prevented from achieving their objectives. Were such defendants to become aware of their mistake, they would change their behavior to make it possible for them to attain their ends. For example, when one attempts to kill with an unloaded gun, he or she is mistaken with regard to the consequential circumstance as to whether the gun is loaded. The mistake makes it impossible for the defendant to achieve his or her objective of killing the victim. If the defendant became aware that the gun was not loaded, he or she would surely load it.

A pure circumstance, on the other hand, is one "the existence of which is not essential to the occurrence of the consequences" ⁽⁶¹⁾ of the completed crime. Defendants who are mistaken about the existence of a pure circumstance are therefore not prevented from producing the consequences they intend to produce and achieving their ends. Were such defendants to become aware of their mistake, they would not change their behavior in any way. For example, when one attempts to burglarize a home believing it to be 5:00 am when it is actually 7:00 am, he or she is mistaken with regard to the pure circumstance of whether it is night. The mistake does not interfere in any way with the defendant achieving his or her objective of breaking into the home. If made aware of the true time, the defendant would simply proceed with the breaking and entering, although he or she might be relieved to know that there could be no liability for burglary for doing so.

Now consider what the common law judges did when faced with a case involving an impossible attempt. Keep in mind that originally such judges did not see themselves as drawing a distinction between legal and factual impossibility. Their task was simply to determine whether the defendant before them should be liable to punishment for an attempt. Because in virtually every case the defendant had the necessary mens rea, ⁽⁶²⁾ this meant deciding whether the defendant's actions constituted the actus reus of an attempt. But because the law provided no specific guidance on how to deal with impossible attempts, ⁽⁶³⁾ judges had to make this determination on the basis ordinary English usage and common sense notions of what constituted an attempt. ⁽⁶⁴⁾ In ordinary usage, an attempt is a try or an effort to achieve an objective. ⁽⁶⁵⁾ In the context of the criminal law in which attempt could be charged only when the defendant had failed to commit a completed crime, ordinary usage and common sense suggested that an attempt was an unsuccessful try or a failed effort to achieve a criminal objective. ⁽⁶⁶⁾ Thus, judges typically, and not unexpectedly, found attempt liability where defendants had tried and failed to obtain a criminal objective and found no liability where the defendants' actions did not fit this model.

What results did this decision process produce? Because mistakes about consequential circumstances prevented defendants from attaining their goals, defendants who made such mistakes always fit the model of those who had tried and failed to achieve a criminal objective. As a result, common law judges almost always found that the actions of such defendants constituted the actus reus of an attempt and upheld their convictions. ⁽⁶⁷⁾ On the other hand, mistakes about pure circumstances did not prevent defendants from attaining their goals. The actions of such defendants are not conveniently described as unsuccessful tries or failed efforts as we saw in our discussion of Jenny's actions in Part II. ⁽⁶⁸⁾ As a result, common law judges were unlikely to view the actions of these defendants as the actus reus of an attempt and would overturn their convictions. ⁽⁶⁹⁾ Thus, by responding in a perfectly natural way to the need to decide the cases before them, the common law courts produced a division between culpable and non-culpable impossible attempts on the basis of whether the defendant's mistake concerned a consequential or pure circumstance. ⁽⁷⁰⁾ Failed attempts to pick empty pockets would support liability; successful attempts to receive property erroneously believed to be stolen would not.

Why then did commentators fail to recognize this distinction? I believe it is because the language the courts used to justify their decisions suggested that a different distinction was being made. In the cases in which convictions for impossible attempts were upheld, the early opinions fairly explicitly justified the convictions on the basis of the ordinary meaning of an attempt. These opinions invariably characterized an attempt as a failed effort. The fact that the effort could not possibly succeed in no way changed its character as a failed effort, and therefore could not shield the defendant from liability for attempt. ⁽⁷¹⁾ In making this point, however, the courts regularly characterized the cause of the effort's failure in terms of unknown facts, ⁽⁷²⁾ unexpected obstacles, ⁽⁷³⁾ extrinsic causes, ⁽⁷⁴⁾ and similar language that suggested a factual error or physical impediment.

The courts had more difficulty explaining their decisions in the cases in which convictions for impossible attempts were overturned. Because in these cases the defendants' errors about pure circumstances did not prevent them from achieving their objectives, their actions did not fit the model of a failed effort and did not square with the ordinary meaning of 'attempt.' However, it would not do for courts to simply declare that a defendant's conviction must be reversed because his or her actions did not look like an attempt. ⁽⁷⁵⁾ But without the benefit of the terminology we are retroactively applying, courts found it difficult to specify the class of cases for which they considered attempt liability inappropriate. As a result, they groped for a description of the distinction they were after with locutions such as: "[i]f all which an accused person intends to do would, if done, constitute no crime, it cannot be a crime to attempt to do with the same purpose a part of the thing intended" ⁽⁷⁶⁾ and "[i]f the thing defendant attempted to do would not and could not, under the statute, have been a crime if accomplished, how can it be said that he attempted to commit the denounced crime, however reprehensible may have been his intent from the standpoint of morals?" ⁽⁷⁷⁾

Unfortunately, these descriptions are highly ambiguous and depend on how one interprets phrases such as 'what an accused person intends to do.' Is 'what an accused person intends to do' merely to produce the consequences he or she seeks or is it to produce these consequences under the attendant circumstances as he or she believes them to be? Should the accused person's intent ignore or incorporate his or her mistaken belief about the pure circumstances? Was Jenny's intention to receive a 700 MHz laptop or was it to receive a stolen laptop?

If the former, then the courts' language would refer to a class of cases that includes both mistakes about pure circumstances and imaginary crimes. When the desired consequences are viewed objectively, a defendant can be said to 'intend' to produce consequences that he or she believes to be criminal but which are not both when a necessary pure circumstance is missing or when he or she mistakenly believes the consequences to be legally prohibited. If Jenny's intention is to receive a 700 MHz laptop, then she can believe this to be a criminal activity either because she believes the laptop to be stolen property or because she believes purchasing such a machine for less than $500 violates a non-existent price regulation.

On the other hand, if 'what an accused person intends to do' incorporates the defendant's mistaken belief, then the courts' language can refer only to imaginary crimes. When the desired consequences are viewed from the defendant's perspective, the defendant intends to produce criminal consequences. But the only way one can intend to produce criminal consequences that are not in fact criminal is to be mistaken about what is legally prohibited. If Jenny's intention is to receive a stolen 700 MHz laptop, then the only way this could not be criminal would be if the jurisdiction did not have a statute prohibiting the receipt of stolen property.

It should be apparent that the courts' language in these cases was sure to create confusion. We know the courts were trying to identify the class of cases in which defendants made a mistake of fact about a pure circumstance because virtually all of the cases in which the relevant language was employed have this characteristic. ⁽⁷⁸⁾ This implies that the courts were viewing 'what an accused person intends to do' from an objective standpoint. But nothing in the courts' language suggests that this rather than the defendant's perspective is the proper interpretative viewpoint. And even if the language is interpreted as the courts apparently intended, it is still problematic in that it refers not only to cases of mistake of fact about a pure circumstance, but also to imaginary crimes involving mistakes of law. In fact, phrases such as "[i]f all which an accused person intends to do would, if done, constitute no crime" and "[i]f the thing defendant attempted to do would not and could not . . . have been a crime if accomplished" seem more suggestive of mistakes of law than mistakes of fact.

The courts' difficulty articulating the basis for their decisions in cases of impossible attempts produced the following situation for those seeking to explain when impossibility could serve as a defense. In virtually all the reported cases of impossible attempts, the defendant had made a mistake of fact. The factor that actually determined whether the defendant's conviction would be upheld was what kind of circumstance the mistake of fact was about; consequential or pure. But judicial opinions in the cases that upheld attempt convictions misleadingly emphasized the factual nature of the defendant's mistake rather than that the mistake caused the failure of the defendant's efforts. And judicial opinions in the cases that overturned attempt convictions used language that could be interpreted to either include or exclude the class of cases the courts' were seeking to identify and, in addition, misleadingly suggested that the defendant had made a mistake of law. Thus, legal analysts were faced with a situation in which the opinions they were analyzing directed their attention away from the distinction that was doing the work and toward a specious distinction between mistake of fact or physical impediment and mistake of law.

Although specious, the distinction between factual mistake and legal mistake was seductive. This is due, in part, to the well-known role the distinction plays with regard to completed crimes. An elementary tenet of criminal law is that mistake of fact can negate mens rea and thus criminal culpability, but mistake or ignorance of the law cannot. ⁽⁷⁹⁾ There is an appealing intellectual elegance to the observation that in the case of the inchoate crime of attempt, it is mistake of law that defeats liability and mistake of fact that does not. Furthermore, for the most part, this observation is correct. Mistake of law, which cannot exculpate in the case of completed crimes, also cannot inculpate in the case of attempt. The belief that gambling is legal will not relieve one of liability for illegal gambling in New York, and the belief that gambling is illegal will not subject one to liability for attempted illegal gambling in Nevada. In addition, because factual mistakes usually concern a consequential circumstance, in most cases, a mistake of fact, which can exculpate in the case of completed crimes, also can inculpate in the case of attempt. The belief that a gun was not loaded can relieve one of liability for murder, and the belief that a gun was loaded can subject one to liability for attempted murder. The desire to assimilate attempt doctrine symmetrically into the established doctrine of mistake of fact and mistake of law created a powerful temptation to overlook the qualifying phrase 'in most cases.' ⁽⁸⁰⁾

In the end, this temptation proved too powerful for many commentators to resist. The very practice of referring to the distinction between culpable and non-culpable impossible attempts as the distinction between factual and legal impossibility reinforced the impression that factual errors were invariably, rather than usually, inculpatory and that exculpatory errors necessarily concerned matters of law. ⁽⁸¹⁾ As a result, many commentators found cases of factual mistakes about pure consequences anomalous. Cases such as Jenny's that contained exculpatory mistakes of fact disrupted the elegant symmetry of the mistake of fact/mistake of law distinction. The commentators who were bothered by this tended to view these as cases in which the courts were confused or had made a mistake. Thus, Francis Sayre, one of the earliest commentators, played upon the ambiguity of the phrases such as 'what an accused person intends to do' to argue that courts that overturned convictions in cases of mistake as to pure circumstances had failed to appreciate that one of the defendant's objectives had been criminal. ⁽⁸²⁾ Once this is recognized, he contended, it would be apparent that impossibility resulting from mistake of fact cannot exculpate. ⁽⁸³⁾ Robert Skilton, another early commentator, exploited the same ambiguity to contend that the courts' mischaracterization of the defendant's intent in cases like Jenny's caused them to incorrectly overturn the convictions. ⁽⁸⁴⁾ Similarly, Jerome Hall regarded the idea of legal impossibility as merely "an awkward expression of the principle of legality" ⁽⁸⁵⁾ indicating that he identified it exclusively with attempts to commit imaginary crimes and regarded the interpretation that included mistakes about pure consequences as "untenable." ⁽⁸⁶⁾

The later commentators who felt the allure of the mistake of law/mistake of fact doctrine similarly had their attention drawn in an inappropriate direction. In the decided cases, the key distinction was between mistakes of fact about consequential circumstances and mistakes of fact about pure circumstances. The question of whether the defendant had made a mistake of law and attempted an imaginary crime was, at best, of tangential significance since almost no cases involved such mistakes. The mistake of fact/mistake of law distinction placed the focus on the difference between imaginary crimes and all mistakes of fact, and relegated the distinction between factual mistakes as to pure versus consequential circumstances to a subsidiary, and apparently superfluous, role. This was eventually reflected in the terminology the commentators employed. Imaginary crimes came to be referred to as cases of "pure legal impossibility," ⁽⁸⁷⁾ while all cases of mistake of fact were referred to as either "pure factual impossibility" ⁽⁸⁸⁾ or simply "factual impossibility." ⁽⁸⁹⁾ Cases of factual mistakes about pure circumstances were given the anomalous appellations of "mixed legal and factual impossibility," ⁽⁹⁰⁾ "mixed fact/law impossibility," ⁽⁹¹⁾ or even "legal (?) impossibility." ⁽⁹²⁾ As this terminology suggests, those employing it almost always regarded the courts' failure to uphold convictions in cases of factual mistake about pure circumstances as erroneous. For example, Professor Simons believes that the very specification of these cases as a separate class is based on an erroneous belief that they involve a mistake of law, ⁽⁹³⁾ and that "[t]he category of traditional 'legal (?) impossibility' is therefore conceptually confused and essentially indistinguishable from the category of factual impossibility." ⁽⁹⁴⁾

Of course, not all commentators were seduced into to viewing the distinction between culpable and non-culpable impossible attempts purely in terms of the mistake of law/mistake of fact distinction. Among those who were not, however, the language the courts had employed was sufficiently misleading so that few saw the distinction in terms of the failed effort/successful effort model. ⁽⁹⁵⁾ Thus, commentators sought the crucial distinguishing element in factors as diverse as whether the defendant's actions were overt or innocent, ⁽⁹⁶⁾ were reasonable or unreasonable, ⁽⁹⁷⁾ invaded a legally protected interest, ⁽⁹⁸⁾ were sufficient to cause public alarm, ⁽⁹⁹⁾ threatened harm, ⁽¹⁰⁰⁾ or were apt. ⁽¹⁰¹⁾ The result of all this was that when the courts looked to the academic commentary for guidance on impossible attempts, they met with a frustrating lack of consensus, ⁽¹⁰²⁾ which was then cited as a reason for abandoning the effort to make a distinction at all.

In sum, it seems likely that impossible attempts came to be regarded as "a morass of confusion" ⁽¹⁰³⁾ because although the early courts distinguished culpable from non-culpable impossible attempts in an entirely natural and common sense way, they did a poor job of describing the distinction they made. By describing the distinction between factual mistakes about pure versus consequential circumstances with language that suggested a distinction between mistake of fact and mistake of law, many commentators were led to view the distinction in terms of the conventional mistake of fact/mistake of law doctrine for completed crimes. These commentators adopted terminology that reinforced the focus on the fact/law distinction: first by labeling the distinction as factual versus legal impossibility, and later, by refining this into the distinctions among pure legal impossibility, mixed fact/law impossibility, and factual impossibility. Because this was not the distinction the courts actually employed, the decided cases did not fit conveniently into the academic categories. Furthermore, even among the commentators who were not seduced by the intellectual elegance of the mistake of fact/mistake of law analogy, the courts' language was imprecise enough to allow for various interpretations of the distinction. As a result, the efforts made by later courts to apply the academic distinctions were maddingly difficult, which, in turn, led the courts to abandon the effort or ask the legislatures to do away with the impossibility defense. Thus, by paying more attention to what the courts said then what they did courts and commentators sounded the death knell of the impossibility defense.

In Part III, I suggested that the confusion that served as the impetus for the rejection of the impossibility defense to a charge of attempt resulted from the courts' difficulty describing what was really a rather simple distinction. Had the courts and commentators been able to describe the distinction between culpable and non-culpable impossible attempts as the difference between a factual mistake as to a consequential or pure circumstance, they may have had no trouble applying it. This suggests that the difficulty of deciphering the common law distinction was not an adequate reason for its rejection.

Of course, the fact that the common law impossibility defense was intelligible does not mean that it was normatively justified. To say that it should not have been discarded as unworkable is not to say that it should not have been discarded. It may well be that under the common law distinction properly understood, Jenny, who made a factual mistake about a pure circumstance, would not be subject to punishment for attempt, while Clarisa, who made a factual mistake about a consequential circumstance, would be. But how can this differential treatment be justified? Aren't Clarisa and Jenny equally blameworthy? Didn't they both intend to violate the law and haven't they each demonstrated their willingness to act on this intention? Both would have violated the law if the facts were as they believed them to be. Why should Clarisa be liable for an attempt, but not Jenny?

Although there is an intelligible difference between Clarisa and Jenny's cases, many commentators would argue that there is no morally relevant difference. Clarisa and Jenny are equally blameworthy, and therefore should be treated the same way by the law. In fact, it is clear that at least some of what caused the early academic commentators to see the distinction between culpable and non-culpable impossible attempts in terms of the mistake of law/mistake of fact distinction was not merely the belief that this is the distinction the courts were making, but that it was the one the courts should be making. When two parties have manifested their willingness to violate a valid provision of criminal law, there seems to be no reason to relieve one of liability for attempt because of the type of mistake of fact he or she has made.

Commentators who subscribe to this conclusion offer a simple, straightforward argument in its support. The purpose of the criminal law is to punish those who manifest their dangerousness or moral depravity through their actions. Those who try to violate the criminal law, but fail to do so due to a mistaken belief about the nature of the world have manifested their dangerousness and/or moral depravity, regardless of whether their mistake was about a consequential or pure circumstance. Therefore, all who make such mistakes should be liable to criminal punishment.

This argument obviously entails the rejection of the common law impossibility defense. A defendant whose attempt to commit a crime is rendered impossible because of his or her mistake of fact about a pure circumstance has nevertheless manifested his or her dangerousness or depravity. Therefore, the fact that the attempt could not result in a completed crime should not bar liability. It remains true that when, due to a mistake of law, the defendant attempts an imaginary crime, there can be no conviction for attempt. But this is because the mens rea of attempt is not present, not because the attempt cannot be completed. ⁽¹⁰⁴⁾

The argument thus provides a principled resolution to the initial uncertainty we felt about Jenny's culpability: ⁽¹⁰⁵⁾ she should be convicted. It also resolves the courts' confusion about whether to characterize the consequences the defendant intends to produce objectively or in accordance with his or her erroneous belief. ⁽¹⁰⁶⁾ Because the defendant's dangerousness or moral depravity is manifested by what the defendant thinks he or she is doing, the relevant consequences are those the defendant thinks he or she is producing. Therefore, the defendant's actions should be described according to the circumstances as he or she believes them to be.

Because of this last feature, commentators who adopt this line of argument are usually known as subjectivists. ⁽¹⁰⁷⁾ These commentators answer the question of which actions can constitute the actus reus of attempt very broadly, allowing any or almost any action the defendant subjectively believes to be in furtherance of his or her criminal intent to support a conviction. ⁽¹⁰⁸⁾ Not all commentators agree that this broad answer is the correct one. Those who would answer the question more narrowly by requiring the defendant's actions to possess particular, objectively identifiable characteristics to constitute an actus reus of attempt are usually referred to as objectivists. ⁽¹⁰⁹⁾

The authors of the Model Penal Code explicitly adopted the subjectivist approach to attempt in drafting the Code. Section 5.01 defines attempt as follows:

(1) Definition of attempt. A person is guilty of an attempt to commit a crime if, acting with the kind of culpability otherwise required for commission of the crime, he:

(a) purposely engages in conduct which would constitute the crime if the attendant circumstances were as he believes them to be; or

(b) when causing a particular result is an element of the crime, does or omits to do anything with the purpose of causing or with the belief that it will cause such result, without further conduct on his part; or

(c) purposely does or omits to do anything which, under the circumstances as he believes them to be, is a substantial step in a course of conduct planned to culminate in his commission of the crime. ⁽¹¹⁰⁾
 

As indicated by the emphasized phrases, the Code explicitly incorporates the defendant's subjective characterization of his or her actions into its description of the acts that will support a conviction for attempt. Thus, under the Code any action at all can serve as the actus reus of an attempt as long as the defendant believes it will further his or her criminal intention.

The definition reflects the Code's acceptance of the fundamental proposition that the criminal sanction should be imposed on those who manifest their dangerous or depraved propensities.

Conduct designed to cause or culminate in the commission of a crime obviously yields an indication that the actor is disposed towards such activity, not alone on this occasion but on others. There is a need, therefore, subject again to proper safeguards, for a legal basis upon which the special danger that such individuals present may be assessed and dealt with. They must be made amenable to the corrective process that the law provides. ⁽¹¹¹⁾
 

This causes the authors of the Code to explicitly reject an objectivist approach to attempt.

The literature and decisions dealing with the definition of a criminal attempt reflect the ambivalence as to how far the governing criterion should be found in the dangerousness of the actor's conduct, measured by objective standards, and how far in the dangerousness of the actor, as a person manifesting a firm disposition to commit a crime. . . . [T]he proper focus of attention is the actor's disposition, and the draft is framed with this in mind. ⁽¹¹²⁾
 

Accordingly, the Code would convict all those who, like Jenny, make a mistake of fact about a pure circumstance.

The basic rationale of these decisions is that, judging the actor's conduct in light of the actual facts, what he intended to do did not amount to a crime. This approach, however, is unsound in that it seeks to evaluate a mental attitude--"intent" or "purpose"--not by looking to the actor's mental frame of reference, but to a situation wholly at variance with the actor's beliefs. In so doing, the courts exonerate defendants in situations where attempt liability most certainly should be imposed. In all of these cases the actor's criminal purpose has been clearly demonstrated; he went as far as he could in implementing that purpose; and as a result, the actor's "dangerousness" is plainly manifested. ⁽¹¹³⁾
 

Thus, the Code is explicitly designed "to extend the criminality of attempts by sweeping aside the defense of impossibility (including the distinction between so-called factual and legal impossibility) . . ." ⁽¹¹⁴⁾
 

Given the judicial confusion and frustration concerning the common law defense of impossibility described in Parts I and III, the Model Penal Code's simple, logically coherent argument for its rejection had sufficient appeal to carry the day. As noted in Part I, impossibility is now almost never a valid defense to a charge of attempt in the United States. With most states adopting some version of the Model Penal Code's definition of attempt, the subjectivist approach has become the orthodox position on attempt. ⁽¹¹⁸⁾ But should it be? Is the Model Penal Code's subjectivist argument sound?

The minority of commentators who favor a narrower construction of the actus reus of attempt do not believe that it is. These objectivist theorists have raised several objections to the Model Penal Code's position, the most significant of which are that: 1) its logic mandates the punishment of imaginary crimes, 2) its standard for liability is impracticable, 3) it subjects the public to improper preventative detention, 4) it permits punishment for thoughts alone, 5) it violates the principle of legality, 6) it improperly requires the punishment of harmless, irrational attempts, and 7) it would allow convictions to be based on unreliable forms of evidence. On close examination, however, none of these objections seem to deal a fatal blow to the Model Penal Code's subjectivist approach. The first five appear to misconstrue the Code's position, while the last two, which do accurately address the Code, fail to provide a principled normative rationale for its rejection.

Neither the Code nor any other subjectivist theorist advocates punishment for attempting imaginary crimes. However, objectivist critics claim that the logic underlying the Code's definition suggests that they should. ⁽¹¹⁹⁾ One who attempts to commit an imaginary crime is one who, laboring under a mistake of law, performs an action he or she believes to be a crime, but which, in fact, is not. The only thing that distinguishes such a case from any other impossible attempt is that in imaginary crimes the defendant has made a mistake of law while in all other cases the defendant has made a mistake of fact. But whether one's mistake is of law or fact seems wholly immaterial to the question of whether one has manifested his or her dangerousness or depravity. In imaginary crimes, no less than in ordinary impossible attempts, the defendant has shown himself or herself willing to violate the law. In both cases, the defendant has intentionally taken actions which if things were as he or she believed them to be would constitute a crime. Such defendants appear equally dangerous or depraved and equally likely, if left unpunished, to try to violate the law in the future. As an illustration, consider Laura and Frank our anti-environmental activists. As one commentator put it,

We fail to see how any rational system of criminal law could justify convicting one and acquitting the other on so fragile and unpersuasive a distinction that one was suffering under a mistake of fact, and the other under a mistake of law. Certainly if the ultimate test is the dangerousness of the actor (i.e., readiness to violate the law), as [the subjectivist] would have it, no distinction is warranted--[Laura] has indicated [her]self to be no less 'dangerous' than [Frank]. ⁽¹²⁰⁾
 

Thus, the objectivist would argue that it is difficult to see how an acquittal in the case of imaginary crimes could be justified under the Code's standard.

The problem with this objection is that it misconstrues the Code and the subjectivist position generally. It is certainly true that whether a defendant's mistake is about a matter of fact or law is irrelevant to how dangerous or depraved he or she is. This is beside the point, however, because the subjectivists are not advocating that all dangerous or depraved people should be subject to criminal punishment. They are proposing a standard for the actus reus of attempt, not for attempt liability in toto. Subjectivists argue that given a criminal mens rea, any action that demonstrates the defendant's dangerousness or depravity can serve as an actus reus, not that all dangerous or depraved people be punished for attempt.

The authors of the Code can freely admit that one who attempts an imaginary crime is every bit as dangerous or depraved as one who engages in a factually impossible attempt since each has equally demonstrated his readiness to break the law. Laura is indeed as dangerous as Frank. However, under the Code's approach, this implies only that there is no difference between Laura and Frank with regard to the actus reus of attempt. It does not imply that Laura and Frank are equally culpable. The subjectivist can consistently maintain that Laura is not liable for an attempt because she does not possess the requisite mens rea. As previously noted, ⁽¹²¹⁾ one who attempts an imaginary crime does not have the specific intent required by the offense. Because such a defendant would not be guilty of an attempt regardless of whether his or her action constitutes an actus reus, the logic of the subjectivist approach does not require convictions in cases of imaginary crimes. Laura and Frank may be equally dangerous or depraved, but they have not equally met the mens rea requirement for an attempt conviction.

Another objection brought against the Code's definition of attempt is that it is fundamentally unworkable. Critics who advance this objection point out that the Code's definition of attempt is based on the assumption that the purpose of the law of attempt is the protection of the public from dangerous or depraved individuals. Realizing this end requires a standard of liability that will convict those with socially dangerous or depraved characters, but acquit those who pose no danger to society. The problem, according to the critics, is that recent psychological and criminological developments have shown that such a standard is unattainable because there is no practicable way of differentiating dangerous from and non-dangerous individuals. This is evidenced by recent efforts to introduce effective rehabilitation into the criminal punishment system which demonstrated the futility of trying to determine the danger to society presented by an individual.

The rehabilitative ideal, of which the concept of dangerousness is a cornerstone, has recently undergone a rather painful process of demystification. . . . The most comprehensive study so far on the effectiveness of rehabilitation programs has led its authors to the devastating judgment that "nothing works." . . . [T]he very foundations of the belief in our ability to identify and cure socially dangerous individuals have now been shattered. ⁽¹²²⁾
 

Because the subjectivist approach to attempt requires precisely this judgment, it must be rejected as unworkable.

As a consequence of the decay of the rehabilitative ideal, a reorientation of the basic assumptions of attempt law seems necessary. The fairly modern idea that a psychiatrist should determine the attemptor's guilt by diagnosing the amount of subconscious internal control which helped to foil the attempt and thus indicated lesser dangerousness, appears as rather grotesque today. Nevertheless, it is but a logical endpoint of the "rational course" proposed by Professor Glanville Williams, which is to "catch intending offenders as soon as possible, and set about curing them of their evil tendencies."

That course seems much less "rational" today as we know more about our ignorance. ⁽¹²³⁾
 

Much like the last objection, the charge of impracticability appears to be an attack on a straw man. There may indeed be no practicable method for identifying who possesses a dangerous or depraved character in the sense of being likely to violate the law in the future. However, the subjectivist definition of attempt requires no such determination. It does not require the prediction of future criminal acts. It advocates the punishment of dangerous or depraved people only in the sense that those who have already taken actions that demonstrate their willingness to break the law should be punished. The subjectivist position asserts that, given the mens rea of attempt, one who has taken any past action that demonstrates his or her willingness to violate the law should be convicted of attempt. Our inability to predict who will violate the law in the future clearly has no bearing on the adequacy of this claim.

C. Preventative Detention

Some objectivist theorists charge that the Model Penal Code's definition of attempt converts the law of attempt into a mechanism for improper preventative detention. Because the chief concern of the Code is to protect the public from dangerous or depraved individuals, its definition of attempt allows the punishment of those who, like Jean-Claude, Frank, Clarisa, and Jenny, have engaged in entirely harmless conduct on the ground that they may do harm in the future. As Glanville Williams, a pre-eminent subjectivist theorist, explains with regard to cases like Clarisa's and Jenny's.

Getting away from dialectics, it is said in favor of the narrower construction of the law that one who attempts to murder with sugar, thinking it to be arsenic, ought not to be held guilty of an attempt because "there is no danger to the public." The short answer to this is that there is danger to the public in leaving uncorrected a man who is bent on murder. ⁽¹²⁴⁾
 

Th[e] attempted receiver [of unstolen property] should be liable to punishment for the same reason as other attempters are: that he has shown himself prone to crime and may well do it again if he does not receive an effective warning. ⁽¹²⁵⁾
 

Objectivist critics claim that this runs afoul of the fundamental tenet of criminal jurisprudence that holds that one can be punished only for what one does, not for what one may do in the future. The essential purpose of the actus reus requirement is to bar preventative detention by limiting the state to proceeding against those who have already engaged in conduct that has been expressly prohibited, rather than those who merely plan to do so in the future. The critics accuse the Code of defeating this purpose by interpreting the actus reus of attempt to allow for the punishment of those who have, in fact, done no harm merely because they may in the future.

This is the price of the choice made long ago in favour of guaranteeing that the innocent be protected, that the heavy weight of the criminal sanction be not imposed too broadly and that the process work moderately rather than oppressively. The [subjectivist approach,] however, seems unconcerned about all this. It would make of the law of attempt a full-scale method of preventative detention. ⁽¹²⁶⁾
 

As was the case with the previous two objections, this objection misconstrues the subjectivist position on attempt, in this case by confusing the concept of harm with the concept of action. Preventative detention consists in the incarceration of an individual who has not yet taken wrongful or harmful action to prevent such action in the future. The Code's definition of attempt does not allow for the detention of one who has not yet acted, it allows for the detention of one who has acted in furtherance of a criminal intention even though that action produces no harm. The subjectivist position is indeed justified on the basis of the need to prevent those who have demonstrated their dangerousness or depravity from causing harm in the future. But this does not imply that the defendant is being incarcerated for actions he or she has not yet and only might take. The defendant is being incarcerated for actions he or she has already taken because those actions demonstrate that he or she is a dangerous or depraved person. This is not preventative detention, but perfectly appropriate punishment for past actions.

D. Punishment for Thoughts Alone

A related objection frequently brought against the Code's subjectivist position is that it permits punishment for thoughts alone. Objectivist critics claim that the Code's approach does not merely place the actus reus requirement in a subsidiary role as the subjectivists admit, ⁽¹²⁷⁾ but renders it entirely devoid of content, and thus is equivalent to punishing defendants for their thoughts alone. ⁽¹²⁸⁾ Although the subjectivists claim to be retaining the actus reus element, they regard this element as fulfilled by any action whatsoever as long as the accused believes it will result in a crime. But, the critics assert, being liable to conviction for doing anything with the belief that it will result in a crime is functionally indistinguishable from being liable to conviction for merely resolving to commit a crime.

[The subjectivist] is inviting us to say that attempted murder can be doing anything while thinking (mistakenly) that you are going to cause [the death of a human being.] This is a dangerous invitation which should be rejected, since it provides no criterion whatsoever for characterizing an act as an attempt other than the mistaken view under which it is being done, and is thus, in spite of [the subjectivist's] denials, tantamount to punishment for intention alone. ⁽¹²⁹⁾
 

This objection can also be illustrated with Clarisa's case. Because Clarisa's act of stirring sugar into her husband's coffee is completely innocent, the only factor that will distinguish her case from non-criminal activity is Clarisa's belief that the sugar is arsenic. If, however, the only factor that distinguishes the culpable from the non-culpable is Clarisa's belief, then the actus reus requirement apparently affords her no substantive protection from conviction. She is being punished exclusively for what is in her mind.

In this case, the authors of the Code can reasonably respond that this objection is wrong on its face. The Code's definition of attempt punishes not mere intention, but only proscribed intentions that have been acted upon.

Superficially it may seem as though . . . the use of the law of attempt would result in punishing a man for mens rea alone, in defiance of the principle that the criminal law requires an act or omission. But it is hoped that this objection has already been sufficiently answered: the accused has gone far beyond mental preparation for a crime; he has adopted a course of conduct which, on the facts as he believes them to be, constitute not merely an attempt but the consummated crime. ⁽¹³⁰⁾
 

The subjectivists can fairly claim that there is a substantive, practical difference between punishing one who intends to commit a crime but takes no action and punishing one who intends to commit a crime and acts upon his intention, albeit mistakenly. Thus, the Code's approach to attempt is not functionally equivalent to punishing for thoughts alone. The requirement of action obviously provides more of a limitation on liability than its absence.

The only sense in which the Code's approach is indistinguishable from punishment for intention alone is to outward appearance; there may be no apparent difference between punishing one who intends to commit a crime but takes no action in furtherance of this intent and punishing one who intends to commit a crime, acts upon his intent, but due to a mistake, acts in a perfectly innocent way. However, punishing an innocent-appearing action taken with a criminal intent is far from a novel event. Driving one's car across a state line is a perfectly innocent action, but if done with the intent to avoid prosecution, it is a crime. ⁽¹³¹⁾ This does not imply that the prohibition on flight from prosecution constitutes punishment for thoughts alone.

The subjectivists appear to be able to distinguish their position from punishment for thoughts alone on a principled basis. Because the criminal law is designed to protect the public from dangerous people, and because taking action in furtherance of a criminal intent is sufficient to demonstrate a defendant's dangerousness or depravity, it is proper to punish such a defendant even if his or her actions appear innocent. However, because one who merely forms a criminal intent but does nothing to effectuate it has demonstrated neither dangerousness nor depravity, he or she should not be punished. Thus, there should be no punishment for thoughts alone.

An objection that is often brought against the Code's approach is that its definition of attempt violates the principle of legality. The principle of legality holds that "there must be no crime or punishment except in accordance with fixed, predetermined law." ⁽¹³²⁾

Legality requires that the forbidden conduct be defined in advance. This is in part so that the citizen will receive advance guidance as to what conduct is forbidden. But it is well recognized that the requirement of advance definition also serves to control the discretion, and thereby the bias, of those officers of the criminal process who make decisions affecting the defendant. ⁽¹³³⁾
 

Objectivists claim that because the subjectivist definition of attempt is essentially "doing anything with the intention of committing a crime," under it, the law of attempt is neither fixed nor predetermined. This definition not only fails to provide the public with advance warning of what conduct is prohibited, it leaves prosecutorial and judicial officials with virtually unlimited discretion to determine what conduct contravenes the law. ⁽¹³⁴⁾ To realize the end of subjecting all those who demonstrate their dangerousness or depravity to the criminal sanction, the Code's definition of attempt must be extremely flexible.

But, of course, this flexibility is bought at the cost of legality and notice. The simple truth is that when any version of the subjective theory holds sway, a potential offender can have no confidence as to how much conduct toward the commission of a crime will be sufficient to constitute an attempt--unless there happens to be a case directly on point. ⁽¹³⁵⁾
 

As with the last objection, the authors of the Code can make a reasonable case that this objection is simply wrong. The Code's definition of attempt entails no lack of advanced warning of what conduct is subject to punishment. The substantive crimes the defendant is trying to commit are all clearly defined and the law of attempt informs the public that doing anything in the effort to commit them is prohibited. This may be a broad definition of the range of attempt liability, but it is perfectly clear and definite. The public is not left in doubt as to which actions intended to result in a crime constitute attempts and which do not; all do. This is perfectly adequate advanced warning of what conduct is subject to punishment. It certainly does not invest prosecutorial agents and judges with discretion to redefine the nature of the offense or to do the type of "criminal equity" that is prohibited by the principle of legality. ⁽¹³⁶⁾ Though broad, the Code's definition of attempt is entirely fixed. ⁽¹³⁷⁾

A common objection raised against the Model Penal Code's approach to attempt implicates our hypothetical friend Jean-Claude, who was convicted of attempted murder for driving a needle through the heart of a voodoo image of his enemy. Under the Code's subjectivist approach, it would appear that Jean-Claude was rightly convicted. He certainly had the mens rea required for attempted murder, and if the circumstances had been as he believed them to be, he would not only have taken a substantial step toward producing his enemy's death, he would have done everything required to bring it about. Thus, under the Code, his actions constitute the actus reus of attempted murder. They certainly demonstrate that he has the dangerous or depraved disposition to commit murder. Therefore, Jean-Claude's culpability appears clear cut.

Objectivist critics of the Code contend that this result is simply wrong. They argue that Jean-Claude's actions were patently harmless, and that any standard of liability that imposes criminal sanctions on someone like Jean-Claude is clearly too broad. George Fletcher, a leading objectivist theorist, points out that Jean-Claude's situation represents:

one case in which virtually everyone agrees that there should be no liability . . . [i.e.,] the case of nominal efforts to inflict harm by superstitious means, say by black magic or witchcraft. The consensus of Western legal systems is that there should be no liability, regardless of the wickedness of intent, for sticking pins in a doll or chanting incantations to banish one's enemy to the nether world. ⁽¹³⁸⁾
 

Thus, objectivist theorists offer Jean-Claude's conviction as a counter-example to the Code's subjectivist approach to attempts. ⁽¹³⁹⁾

This objection is likely to be persuasive only to those who already share the intuitions of those making it. Subjectivist theorists can with some justice claim that the objectivists' assertion constitutes mere disagreement rather than a principled objection. Because there are almost no reported cases of irrational attempts, the consensus Fletcher refers to is merely agreement among commentators who apparently do not accept the subjectivist approach. It is true that Jean-Claude's actions are harmless. However, this is relevant only if one has already discounted the subjectivist approach to attempt for which the salient feature is not the harmfulness of Jean-Claude's actions per se, but whether they demonstrate his dangerousness. One who has made every effort to kill by irrationally ineffective means may well employ more efficacious means the next time. Had Jean-Claude not been arrested, he may well have tried again with a machete.

Viewed in isolation, punishing Jean-Claude for attempted murder may seem to be a harsh result. But the authors of the Code would argue that such cases should not be viewed in isolation. Jean-Claude's attempt may have been irrational in that it employed means that a reasonable person would know could not accomplish his objective, but it demonstrated his willingness to take another's life and hence his dangerousness. Since, according to the Code, the purpose of punishing attempts is to subject those who pose a danger to their fellow citizens to criminal sanction, Jean-Claude is a proper candidate for punishment.

Cases can be imagined in which it might well be accurate to say that the nature of the means selected, say black magic, substantially negates dangerousness of character. On the other hand, there are many cases as well where one who tries to commit a crime by what he later learns to be inadequate methods will recognize the futility of his action and seek more efficacious means. There are, in other words, many instances of dangerous character revealed by "impossible" attempts, and to develop a theory around highly exceptional situations ignores the propriety of convictions in these. ⁽¹⁴⁰⁾
 

Thus, the supporters of the subjectivist approach can fairly contend that a shared intuition among non-subjectivist theorists that people like Jean-Claude should not be subject to punishment does not constitute an adequate reason for rejecting the Code's definition of attempt.

Finally, objectivist critics charge the Code's subjectivist approach to attempt of endangering the innocent by allowing convictions to be based on unreliable forms of evidence. These critics point out that to gain a conviction in criminal cases, the prosecution must establish both that the defendant's conduct conformed to an antecedently defined standard and that he or she possessed the requisite state of mind when engaging in such conduct. This means that in addition to evidence of the defendant's state of mind, the prosecution must produce specific evidence demonstrating that the defendant's conduct constituted the actus reus of the relevant offense. However, under the Code's definition, the defendant's conduct need not meet any particular set of criteria. Any action that the defendant believes will result in a crime will suffice when undertaken with the appropriate mens rea. This means that under the Code, a conviction can be obtained based solely upon evidence of what was in the defendant's mind. The evidence that is usually required to establish what the defendant has done is replaced by evidence of what the defendant believed he or she has done.

The problem, according to objectivist theorists, is that evidence of an defendant's state of mind in the absence of any objectively verifiable supporting facts is notoriously unreliable. Arnold Enker points out that in cases like Jenny's, for example, there is:

a significant statistical correlation between possession of recently stolen goods and knowledge of the fact that they have been stolen. If we can say that in a given percentage of the cases the possessor knows the goods are stolen, then possession of stolen goods is probative of knowledge that the goods are stolen at least in the sense that proof of possession of such goods makes it more likely that the defendant knew the goods were stolen than if there is no proof of possession. . . . [However,] whatever the statistical relation between possession and knowledge may be, the percentage of people possessing unstolen goods who believe the goods are stolen is clearly much lower. ⁽¹⁴¹⁾
 

As a result, one can be convicted of an attempt on highly suspect evidence.

The point is . . . that by eliminating these objective elements we create newly defined crimes in which we replace the statutorily defined fixed reference points for judging the defendant's mens rea with an open-ended sufficiency-of-the-evidence test which may include the less reliable forms of evidence such as questionable admissions, the testimony of informers and accomplices, and proof of prior convictions. ⁽¹⁴²⁾
 

Consider Clarisa's case again. She stirred sugar into her husband's coffee just as she did every morning. It was only because she went to the police to confess that she was arrested. But, as Graham Hughes points out, "apart from [her] confession, there is nothing at all that approaches the threshold of criminal behavior." ⁽¹⁴³⁾ In such cases, "the danger of the [subjectivist] view is that under it cases appear as indictable attempts in situations where proof of the intention of the accused is the only circumstance that could make us begin to think of what has been done as a criminal attempt." ⁽¹⁴⁴⁾

This is a potentially serious objection to the Code's subjectivist position. If the subjectivist theory of attempt really does create an unacceptable risk of convicting the innocent, that would seem to be a good reason for its rejection. However, the supporters of the Code's definition respond to this charge by denying that the subjectivist position creates any greater danger of convicting the innocent than that of many substantive crimes that permit conviction on the basis of conduct that is innocent on its face.

A considerable number of instances can be named in which the offender's act, viewed by itself, is harmless, but is nevertheless punishable if a particular accompanying state of mind on his part can be proved. . . . It is generally unobjectionable, and often highly desirable, to enter into a room [breaking and entering or entering], to make a telephone call [disorderly conduct], to stand upon the sidewalk [loitering with intent to disturb school], to pay money to a party in a lawsuit [communicating with jurors and witnesses], to travel [flight to avoid prosecution or giving testimony] or to transport a women across state borders [White Slave traffic]. Yet all these activities are punishable if done with the "proper" criminal intent. The law of criminal omissions is another concededly atypical example--if there exists a legal duty to act, the offender may be peacefully asleep and yet thereby incur criminal liability. ⁽¹⁴⁵⁾
 

If innocent acts can serve as the actus reus of substantive crimes without posing an unacceptable risk that the innocent will be convicted, subjectivist theorists can reasonably ask why that risk is unacceptable when the crime is attempt?

The force of the objectivist criticism comes from the implication that the subjectivist approach to attempts presents a greater risk of convicting the innocent than does the theory underlying the substantive crimes. The subjectivist response appears to show that this is not the case. Every crime poses some risk that innocent people will be convicted. How much is unacceptable? Subjectivist theorists can argue the proper amount of this risk is the minimum necessary to accomplish the goal of the criminal law, which, according to them, is to protect the public from dangerous or depraved individuals. This is precisely the amount of risk their approach to attempts presents. Unless the objectivists can provide a principled reason why this level of risk is unacceptable, it does not appear that the present objection can effectively undermine the subjectivist position.

This review of the complains objectivist theorists typically bring against the Code's definition of attempt shows that most are ineffective. Several, such as the claims that the logic of the Code's definition requires the punishment of imaginary crimes, that the Code's definition permits preventative detention, and that the Code's standard of attempt liability is impracticable, turn out to be attacks on a straw man. Each of these criticize the proposition that the law of attempt should be defined to permit the punishment of anyone with a dangerous or depraved character. But this is much broader than the Code's position, which is that given the mens rea of attempt, any act in furtherance of that mens rea can serve as the actus reus of an attempt. This more modest proposition does not fall victim to these objections. In addition, the claims that the Code's definition would allow punishment for thoughts alone and violates the principle of legality appear to be inaccurate exaggerations.

The remaining objections, that the Code would punish irrational attempts and allow convictions based on unreliable forms of evidence, amount to the charge that the Code's definition of attempt is too broad. These objections accurately point out that the Code would punish those who employ irrational means to accomplish their criminal ends and those whose actions are outwardly indistinguishable from innocent activity. However, they do not present a compelling case that punishing irrational or innocent-appearing attempts is normatively unacceptable. The authors of the Code have provided a principled argument for the conclusion that it is not. If the criminal law is designed to protect the public from dangerous or depraved individuals and punishing irrational and innocent-appearing attempts will help realize this end, then punishing these attempts is normatively justified. Jean-Claude may have acted in a harmlessly ineffectual manner and Frank's, Clarisa's, and Jenny's conduct may appear outwardly innocent, but each is just as morally blameworthy as the pickpocket who reaches into an empty pocket or the jealous lover who fires a pistol at his rival but misses. Why should they escape punishment when the pickpocket and lover do not? Indeed, Glanville Williams poses precisely this question as a challenge to all those who advocate an objectivist approach to attempt: "The burden lies on those who would restrict the law of attempt to show (1) that there is a reason of policy for so restricting it and (2) that a line between punishable and non-punishable impossible attempts can be satisfactorily drawn." ⁽¹⁴⁶⁾ Can this challenge be met?

I believe that Professor Williams' challenge can be met. In subsequent parts of this article, I will both identify the "reason of policy" that justifies narrowing the Model Penal Code's definition of attempt and show that there is a practicable method for identifying the cases in which impossibility should serve as a defense to a charge of attempt. But to do so, I must begin by highlighting a fundamental flaw in the normative arguments offered in support of the Code's definition: the improper conflation of moral and criminal responsibility.

In the preceding section, we saw that the basic argument for Model Penal Code's definition of attempt rested on the premise that the criminal law is designed to punish those who manifest their dangerousness or moral depravity through their actions. ⁽¹⁴⁷⁾ This is merely a reflection of what the subjectivists regard as the underlying reason for having a criminal justice system in the first place, which is to provide for an orderly society and secure the persons and property of law-abiding citizens public against invasion by their ill-motivated (or negligent) fellows. Punishing those who act in furtherance of a criminal intention even if their efforts are extremely inapt and can not possibly succeed advances this purpose by preventing these individuals from trying again by more effective means. Therefore, the subjectivists argue that the Model Penal Code's definition allowing any action in furtherance of a criminal intention to serve as the actus reus of an attempt advances the fundamental purpose of the criminal law.

The subjectivists' basic argument is typically supplemented with two other normative arguments: that the Code's definition is derived from an analysis of the underlying purposes of punishment and that it is required by the fundamental principle of justice that requires treating like cases alike. The former argument recognizes that there is much controversy over what constitutes the proper purpose of criminal punishment. Debate rages over the relative importance of deterrence, retribution, and rehabilitation and how to reconcile conflicts among these ends. This is immaterial to the subjectivist argument, however, which contends that all of the putative purposes of punishment support the Model Penal Code's standard for attempt liability. For example, if the proper purpose of punishment is deterrence, punishing those who intend to violate the law and fail is fully justified regardless of the reason for the failure. Doing so certainly provides specific deterrence by preventing the particular defendant from trying again. It also provides general deterrence by discouraging those who "are not completely confident that they will succeed in their criminal objective, but will be prepared to run the risk of punishment if they can be assured that they have to pay nothing for attempts which fall; whereas if unsuccessful attempts were also punished the price might appear to them to be too high" ⁽¹⁴⁸⁾ as well as those who "might with good or bad reason believe that if they succeed in committing some crime they will escape, but if they fail they may be caught." ⁽¹⁴⁹⁾ Those who demonstrate their willingness to violate the law by acting in furtherance of a criminal intent are precisely the individuals who need deterring. The particular reason why such attempts fail seems wholly irrelevant to the deterrent value of punishing the attempt. Further, if the purpose of punishment is either retribution or rehabilitation, the case is even clearer. For the retributivist, punishment must be assigned in proportion to the moral deserts of the defendant. But one who has done all he or she can to violate the law but failed is just as morally culpable as one who succeeds, and thus just as deserving of punishment. ⁽¹⁵⁰⁾ Once again, the reason why the attempt failed is irrelevant to the defendant's moral blameworthiness, and thus to his or her liability to punishment. Similarly, anyone who demonstrates his or her willingness to violate the law is in need of rehabilitation regardless of whether his or her efforts were successful. Jean-Claude, Frank, Clarisa, and Jenny are all just as morally culpable and just as in need of deterrence or rehabilitation as anyone else who attempts to commit a crime, and hence just as deserving of punishment. Thus, the subjectivists contend that regardless of which theory of punishment is correct, the Model Penal Code's definition of attempt must be. ⁽¹⁵¹⁾

The subjectivists also argue for the Model Penal Code's definition of attempt indirectly by contending that all objectivist definitions violate the basic principle of justice that requires that like case be treated alike. This is because no matter where the objectivist theorists draw the line between culpable and non-culpable attempts, parties who are identical in all morally relevant respects will end up on opposite sides of it. This point was made by the British Law Commission which argued:

[O]ne may contrast the case of the intending murderer who doses his victim with too weak a solution of poison with that of another who administers an entirely innocent liquid. Both believe that what they are giving is lethal. The only possible explanation which can be given for holding the former guilty of attempted murder and the latter not guilty is that the actions of the former were in some sense more dangerous than those of the latter. The explanation itself poses the question of how weak the mixture has to be before it becomes innocent.

Results as capricious as these do not seem acceptable in the criminal law. ⁽¹⁵²⁾
 

Further, the objectivist approach to attempt by definition distinguishes culpable from non-culpable attempts on the basis of something other than the moral characteristics of the defendant. This seems to guarantee that "it would have an arbitrary operation" ⁽¹⁵³⁾ often "acquitting persons who may even be more dangerous to the public than those who are convicted." ⁽¹⁵⁴⁾ As Professor Michael Cohen put it:

What is being submitted is that the "objectivists" are being seduced by the relative harmlessness of the extreme examples. They would throw away a device that would enable conviction in those analytically indistinguishable cases where, on any view, D is a social danger. The price for not convicting the relatively harmless would be that the criminal law would be impotent to deal with the wicked and the dangerous. ⁽¹⁵⁵⁾
 

It must be admitted that these arguments for the Model Penal Code's definition of attempt are impeccable under the assumption that there is no difference in principle between moral and criminal responsibility. One is morally responsible and hence subject to moral censure when one acts immorally, i.e., when one intentionally or at least negligently violates a moral tenet or fails to fulfill a moral obligation. ⁽¹⁵⁶⁾ The subjectivist arguments proceed as though the class of actions for which one may be held criminally responsible and hence subject to criminal punishment is simply a subset of the class of actions for which one may be held morally responsible; the subset that has been deemed worthy of enforcement by state sanction. Under this view, criminal responsibility may be properly imposed when one has engaged in morally culpable action of the type that the criminal law is designed to suppress. If, indeed, the purpose of the criminal law is to secure the persons and property of citizens against harm from their fellows, as the subjectivists contend, this would mean that criminal responsibility may be imposed whenever a defendant acts in a morally culpable way that demonstrates that the defendant poses the danger of harm to his or her fellow citizens. This conception of criminal responsibility leads directly to the Model Penal Code's subjectivist position on attempt because, given action on the part of the defendant that demonstrates he or she is a danger to others, the determining factor as to whether the defendant should be subjected to criminal punishment must be his or her moral culpability.

The problem with this line of reasoning is that the assumption upon which it is based is incorrect. There is a principled difference between moral and criminal responsibility. Moral responsibility indicates that one is deserving of punishment. Criminal responsibility authorizes some human beings to punish others. Criminal responsibility inherently involves an element of human agency that moral responsibility does not.

In making a determination of moral responsibility, we are concerned only with the actions of one party, the agent whose conduct is being evaluated. The only relevant issue is whether the agent has acted in a morally unacceptable way; whether he or she has violated a moral tenet or failed to fulfill a moral obligation. ⁽¹⁵⁷⁾ Determining that the agent has acted in a morally blameworthy manner does not in itself authorize anyone else to take action against him or her. The inquiry is an abstract one involving no practical enforcement issues.

The case is different when we make a determination of criminal responsibility. Such a determination requires not only a finding that the defendant has acted in a culpable manner that manifests his or her dangerousness or depravity, but also that it is proper for agents of the government to impose a criminal sanction on him or her. Here, we are necessarily concerned with the actions of two parties, the defendant and the government enforcement agents. Unless these agents are both omniscient and incorruptible, the class of cases in which the defendant has culpably manifested his or her dangerousness or depravity can not be coextensive with the class of cases in which the imposition of the criminal sanction is justified.There will always be some cases in which the effort to impose punishment on a class of defendants who morally deserve it would subject the public to an unacceptable risk of harm from the errors or venality of the human beings charged with enforcing the law. If the criminal justice system were guaranteed to be administered with godlike perfection, then realms of criminal responsibility and moral responsibility would indeed be co-extensive. But this is not the world we live in, and thus determinations of criminal responsibility must always consider practical matters of administration that determinations of moral responsibility ignore.

In asserting that criminal responsibility is different in kind from moral responsibility, I am not making an appeal to any Constitutional or extra-legal values. The distinction is not based on the Constitutional restraints on police and prosecutorial practices derived from the Bill of Rights or any philosophical argument about the value of a free society. It is derived from the internal logic of the criminal law itself. The purpose of the criminal law is to provide an orderly society in which citizens are secure in their the persons and property; to protect citizens against harm from the other members of society. Obviously, to accomplish this end, it must protect citizens against threats to their persons and property from other members of society acting in their individual capacities; it must provide protection against "criminals." However, it would be pointless to do so in a way that left the citizens exposed to threats to their persons and property from the individuals acting as governmental enforcement agents. To serve its purpose, the criminal law must be structured to provide citizens with the optimal amount of protection to their persons and property against invasion from all other members of society, whether acting individually or officially.

The need for protection against the human beings who administer the criminal justice system explains why criminal responsibility is inherently different from moral responsibility. Moral responsibility is an all or nothing affair. One has or has not acted in a mortally culpable way. If one has, he or she is liable to moral censure. Criminal responsibility, on the other hand, always involves a balancing of competing interests. Because the criminal law is administered by human beings who are as error-prone and susceptible to temptation as anyone else, every gain in protection against criminal activity that comes from more effective enforcement measures produces a loss in protection against ill-considered or improper official action. Conversely, every enhancement in protection against the state hampers governmental agents' ability to provide protection against individual criminals. Theorists can and do argue about where the line should be drawn to realize the optimal level of protection, but the line must be drawn somewhere. Thus, judgments of criminal responsibility necessarily involve a weighing of competing interests that moral responsibility does not.

The recognition of the fundamental difference between criminal and moral responsibility, i.e., that judgments of criminal responsibility necessarily involve consideration of the risk of harm presented by enforcement agents in a way that judgments of moral responsibility do not, reveals the defects in the subjectivist arguments for the Model Penal Code's definition of attempt. By treating criminal responsibility as though it was isomorphic with moral responsibility, the subjectivist arguments depersonalize the state. They yield judgments of criminal responsibility that would be correct only if the government enforcement mechanism functioned perfectly with no risk of enforcement error or abuse; that is, as though it was not administered by human beings. The subjectivists' basic argument for the Model Penal Code's definition of attempt rests on the fundamental premise that those who manifest their dangerousness or depravity through their actions should be subject to criminal punishment. This is supposed to derive directly from the essential purpose of criminal law, which subjectivists characterize the protection of the persons and property of the citizenry against harm from their ill-motivated fellows, i.e., from "criminals." But this characterization of the purpose of the criminal law completely ignores the danger to citizens' persons and property presented by state enforcement agents. By analyzing the issue as though it were one of moral responsibility, the subjectivists implicitly assume a perfectly functioning enforcement mechanism not burdened by human errors or weakness of will. Approaching the matter as one of criminal rather than moral responsibility suggests that the correct characterization of the purpose of criminal law is the protection of the persons and property of citizens against harm from all human agents, whether acting in their individual capacity as criminals or in their official capacity as state agents of enforcement. But if this is so, it is not true that all individuals who manifest their dangerousness or depravity through their actions should be subject to criminal punishment. Only as many as can be punished without creating too great a risk of harm from enforcement agents should be held criminally responsible. This means that it is possible for one or more of Jean-Claude, Frank, Clarisa, or Jenny, all of whom are morally culpable, to not be criminally responsible for attempt because they represent a class of defendant whose punishment would create too great a risk of enforcement error or abuse.

The depersonalization of the state that results from treating criminal responsibility as though it was moral responsibility also causes the subjectivists to mistake conditions that are merely necessary for criminal responsibility for those that are sufficient for it. It is certainly true that punishment is proper only when one has acted in a blameworthy way; that moral culpability is necessary for punishment. It is also true that criminal punishment is proper only when one has acted in a blameworthy way that is at variance with the ends of the criminal law; that both moral culpability and an action demonstrating one's dangerousness or depravity are necessary for criminal punishment. But this implies only that criminal punishment is improper in the absence of either of these elements, not, as the subjectivist argument assumes, that it is justified when they are present. The latter would be the case only if punishing all morally culpable agents whose actions demonstrate their dangerousness or depravity posed no risk of harm from enforcement error or abuse; in other words, if the mechanism of punishment were not administered by human beings. Because this mechanism is administered by human beings, however, there is an additional necessary condition for criminal responsibility; that punishment can be imposed without creating an unacceptable level of risk of harm from government agents.

The existence of this third condition for criminal responsibility converts the subjectivist argument that the Code's definition of attempt follows from the analysis of the purposes of punishment into a non sequitur. The fact that society's interest in deterrence, retribution, and rehabilitation would all be served by punishing those who take any action in furtherance of a criminal intention proves only that punishing such individuals is not unjustified as a pointless infliction of harm upon them. It does not establish that all such individuals should be punished unless that can be done without creating an unacceptable level of risk of harm from government agents. It is only by viewing the assignment of criminal responsibility as though it were the assignment of moral responsibility, i.e., only by viewing the state as a depersonalized, perfectly functioning enforcement machine, that the subjectivists can argue that all those whose punishment serves the underlying purposes of punishment should in fact be punished.

The existence of this third necessary condition for the imposition of criminal responsibility also undermines the subjectivist argument based on the fundamental normative requirement that like cases be treated alike. It is true that unless all those who act in furtherance of a criminal intention are subject to punishment for attempt, parties that are indistinguishable with regard to their moral blameworthiness will end up on opposite sides of the liability line. An objectivist definition of attempt would allow some parties to escape punishment who are equally or, as Glanville Williams points out, ⁽¹⁵⁸⁾ even more morally blameworthy than those who would be punished. Clarisa or Jenny might get off scot-free while one who attempted murder with an harmlessly small dose of poison or who attempted to receive stolen property that was actually stolen would be incarcerated. But this does not imply that an objectivist definition of attempt "would have an arbitrary operation." ⁽¹⁵⁹⁾ There is no violation of the duty to treat like cases alike because cases such as Clarsa's or Jenny's are like those of their incarcerated counterparts only with respect to their moral responsibility, not their criminal responsibility. It is criminal responsibility, however, that renders one liable to criminal punishment. The principled difference between the cases of those who escape punishment and those who do not is that the former can not be subjected to criminal punishment without an unacceptable risk of enforcement error or abuse while the latter can. Moral responsibility looks only to actions of those to be punished. If such parties are equally blameworthy, a rule which exempts some from moral censure would indeed run afoul of the requirement to treat like cases alike. Criminal responsibility, however, inherently requires the evaluation of the actions of not only the persons to be punished, but also those of the agents imposing the punishment. Cases which are indistinguishable with regard to the actions of the recipient of punishment may be completely different with regard to the actions of those meting it out. For purposes of criminal responsibility, such cases are not alike and do not require similar disposition.

It should be noted that if my argument for the distinction between criminal responsibility and moral responsibility is correct, it implies only that the arguments offered in support of the Model Penal Code's definition of attempt are flawed, not that the Code's definition is not in fact correct. Proving that the Code's definition accurately apportions moral responsibility can not prove that it accurately apportions criminal responsibility, which is the present concern. However, whether the Code's definition does accurately apportion criminal responsibility remains, at this point, an open question. All criminal punishment presents some risk of enforcement error or abuse. The crucial issue is how much is too much. The proper assignment of criminal responsibility requires the optimization of the balance between protection against harm from criminals and protection against harm from enforcement agents. To determine whether the Code's definition is acceptable, we must determine whether it falls within or without the acceptable range of risk of enforcement error and abuse.

But how is this range to be determined? An examination of the internal logic of the criminal law tells us only that the threat of harm from criminal activity must be balanced against the threat of harm from enforcement error and abuse. It tells us nothing about where this balance should be struck. Making this determination requires an appeal to the normative values that underlie the criminal law.

I have argued that a correct understanding of criminal responsibility requires consideration of both the threat presented by criminal activity and that presented by enforcement error and abuse. Criminal responsibility is correctly assigned when the balance between these conflicting values is optimized. But striking this balance requires a normative standard by which to judge the relative importance of the values in conflict. Fortunately, such a standard is close at hand, for the Anglo-American criminal law has a standard of values embedded within it; something we might call the inherent liberal bias of the criminal law.

As a political theory, liberalism gives primacy of place to individual liberty and autonomy. It is usually presented as a deontological theory that declares that the duty to preserve the autonomy of the individual members of society (usually embodied in a set of individual rights) must be respected even if significant gains in the aggregate welfare of society could be achieved by violating it. ⁽¹⁶⁰⁾ Thus, liberalism places a higher normative value on liberty than on improvements in material welfare or physical security, and when such values come into conflict, tips the balance toward the preservation of liberty. ⁽¹⁶¹⁾

In the context of the criminal law, liberalism requires that state enforcement power be curtailed sufficiently to ensure respect for the autonomy of the citizens. Criminal law may be necessary to provide for the security of citizens' persons and property, but it does so by depriving those who violate the law of their liberty. The normative priority of liberty over security requires that criminal responsibility be assigned so that increases in security against criminals are not purchased with decreases in the liberty of law-abiding citizens. Accordingly, liberalism regards the protection of citizens against state enforcement error and abuse as relatively more important than protection against criminal activity. This argument based on the philosophical commitment to the priority of liberty is reinforced with the practical observation that the state generally constitutes a greater danger to citizens than individual criminals. Even in a relatively lawless society, citizens can act to protect themselves against theft and injury. Installing burglar alarms, purchasing a gun, forming a neighborhood watch group, and hiring private security services are all steps citizens can take to protect themselves against the criminal activity of their fellows. However, there is little citizens can do to protect themselves against abuse by state officials. Self-help against official action is itself illegal. The only redress available against the state is an appeal to the state itself. ⁽¹⁶²⁾ This makes the danger of a state that has slipped its bonds significantly greater than that posed by even the most malicious of criminals.

There is, of course, nothing sacrosanct about liberalism, which is under attack from various academic and ideological quarters. It is, however, the value system that is embedded in the Anglo-American criminal law. The idea that a state that does not respect the rights of its citizens is a greater danger than individual criminals is the source of many of the structural features built into the criminal law. The presumption of innocence, the prosecutorial burden to prove every element of a crime beyond a reasonable doubt, and the requirement of a unanimous jury verdict for conviction all exemplify the inherent liberal bias of the criminal law. Even the Constitutional protections of the bill of rights such as the right to trial by jury, to be represented by counsel, to confront one's accuser and be informed of the charges brought against one, and to be free from compulsion to testify against onself and from double jeopardy were originally derived from the common law of crime. ⁽¹⁶³⁾ The oft-quoted maxim that it is better that nine guilty people go free than that one innocent person be wrongly convicted reflects the liberal standard of values that is an imminent feature of Anglo-American criminal law. ⁽¹⁶⁴⁾ Thus, basing judgments of criminal responsibility on a liberal normative standard that privileges citizens' interest in protection against the state enforcement error and abuse over their interest in protection against criminals ensures that criminal responsibility is assigned in a manner consistent with the law's internal value structure; surely a strong consideration in its favor.

A full consideration of the relative merits of a liberal value structure as opposed to more right or left wing alternatives is obviously beyond the scope of the present work. I suggest, however, that in the absence of a reason to reject the fundamental liberal bias of the criminal law as presently constituted, this inherent bias can supply the "reason of policy" for narrowing the scope of attempt liability demanded by the subjectivist challenge.

The argument for this conclusion rests on the claim that achieving the proper protection against excessive governmental interference with citizens' liberty required by the inherent liberalism of the criminal law mandates an objective condition for attempt liability. The requirement that the prosecution establish an objectively verifiable element before convicting a citizen of attempt acts as a check on prosecutorial abuse. Imagine, for example, an crusading anti-drug prosecutor willing to stop at nothing to shut down the drug trade. The prosecutor decides to pressure a suspected drug courier into testifying against his kingpin boss by charging him with the attempt to smuggle a controlled substance into the country when he flies into the United States with talcum powder in his luggage. Under the Model Penal Code, the only thing the prosecutor need prove is that the courier believed the talcum powder to be heroin, something it would be difficult for one with a record of drug smuggling to convincingly deny. Or imagine a prosecutor trying to extort testimony from a suspected member of organized crime by sending a police agent to offer him a valuable product at a below market price and then charging him with the attempt to receive stolen property. Would such a defendant be willing to face a jury if, as is the case under the Model Penal Code definition, all the prosecutor has to prove is that the defendant believed the product to be stolen property? The requirement that the prosecutor establish an objective element of the actus reus acts to prevent this sort of prosecutorial abuse.

Professor Arnold Enker has provided an eloquent argument for the claim that the inherent liberalism of the criminal law requires an objective condition for attempt. Enker begins by appealing to H. L. A. Hart's argument that a major function of the mens rea requirement is to "maximize the individual's power to predict the likelihood that the sanctions of the criminal law will be applied to him." ⁽¹⁶⁵⁾ Enker agrees with Hart that if the purpose of the criminal law is to protect citizens from harm, the mens rea requirement is necessary to protect citizens from the risk of harm from the state.

It is the function of the criminal law to promote the security and well being of the members of society by securing for them a high measure of protection from harmful acts. But since society achieves such protection by inflicting harm on those who would commit such acts, it must take care not to offset this gain in security by unduly increasing the risks that persons will be subjected to official harm unpredictably. Acts can occur accidentally, but the state of mind that accompanies one's act is entirely within the individual's control. Thus, by recognizing mens rea as an indispensable element of crimes, we substantially increase the individual's power to control his freedom from punishment. ⁽¹⁶⁶⁾
 

Enker simply extends this argument to the actus reus. Both actus reus and mens rea are requirements for criminal liability. As such, both serve to protect citizens from oppression by the state. The rationale that underlies the mens rea requirement applies equally to the actus reus requirement.

But it would be shortsighted to think that only the mens rea element serves this function. Mens rea is within one's control but, as already seen, it is not subject to direct proof. More importantly, perhaps, it is not subject to direct refutation either. It is the subject of inference and speculation. The act requirement with its relative fixedness, its greater visibility and difficulty of fabrication, serves to provide additional security and predictability by limiting the scope of the criminal law to those who have engaged in conduct that is itself objectively forbidden and objectively verifiable. Security from officially imposed harm comes not only from the knowledge that one's thoughts are pure but that one's acts are similarly pure. So long as a citizen does not engage in forbidden conduct, he has little need to worry about possible erroneous official conclusions about his guilty mind. ⁽¹⁶⁷⁾
 

Thus, a major purpose of the actus reus is to set "an objective limit to those situations and persons that can become the objects of official assertions of control." ⁽¹⁶⁸⁾ Enker argues that the Model Penal Code's definition of attempt that allows any action at all to serve as the actus reus of attempt when it is accompanied by the required mens rea can not serve this purpose. Only a definition that requires the prosecution to prove an antecedently defined objective feature of a defendant's conduct can.

In asserting that the liberal bias of the criminal law provides the normative justification for requiring an objective condition for attempt liability, I have no desire to overstate the case. Despite the above-quoted maxim that it is better that nine guilty people go free than that one innocent person be wrongfully convicted, I make no claim that the balance between protection against enforcement error and abuse and protection against criminal activity demanded by the inherent liberalism of the criminal law is subject to precise quantification. The appeal to liberalism that I am making may supply nothing more than a general rule of prudence that when a balance is required between the need to protect citizens against enforcement error and abuse and the need to protect them against criminals, it is best to err on the side of protection against enforcement error. If so, it may be insufficient to resolve very close cases in which the competing interests are evenly balanced. Liability for impossible attempts is not such a case, however. In the context of attempting the impossible, even a minimal rule of prudence is sufficient to ground the requirement of an objective condition for liability.

Adding an objective condition to the actus reus of attempt adds somewhat to the danger citizens will experience from their fellows. Assume for the moment that the objective condition mirrors the common law rule that exempted legally impossible attempts from liability. ⁽¹⁶⁹⁾ There will be people like Jenny who will have done everything that they believe to be necessary to break the law without either succeeding or satisfying the objective condition for attempt liability. Such people will not be liable to criminal sanction and will remain free to break the law in the future as the authors of the Model Penal Code contend. ⁽¹⁷⁰⁾ The added danger to society will be rather slight, however. Most of those who attempt to commit crimes and fail will still be liable to the criminal sanction. Only in the somewhat unusual cases in which a defendant has made a mistake about a pure circumstance would he or she constitute an ongoing danger.

On the other hand, the lack of an objective condition for liability poses a significant risk of enforcement error or abuse. Almost all of us jump at the chance to purchase an item we want at a bargain price, often from non-retail sources. How secure can those of us with criminal records or other associations or characteristics that reduce our credibility with a jury be if a prosecutor can obtain a conviction for receiving stolen property merely by convincing a jury that we believed the property to be stolen? A subjectivist definition of attempt would allow one to be convicted of the attempt to possess a controlled substance when one possessed an uncontrolled substance he or she mistakenly believed to be controlled. ⁽¹⁷¹⁾ Most prosecutors are entirely honest. But the incentives within the criminal justice system reward those who obtain convictions and "get the bad guys off the streets." One of the best ways to do this is to get participants in criminal enterprises to become informers or turn state's evidence. A law that allows a prosecutor to pressure drug suspects to do so by charging those who have baby powder or sugar in their homes with the attempt to possess a controlled substance and challenging them to convince a jury that they did not believe it to be heroin provides a temptation to abuse that may be difficult to resist. ⁽¹⁷²⁾ Furthermore, even in the absence of abuse, a subjectivist definition of attempt creates strong incentives for informers and those facing charges themselves to accuse others of possessing what they believed to be controlled substances, whether they do or not, in return for monetary payments or leniency in sentencing. To the extent that police and prosecutors rely on such evidence, the risk of enforcement error is greatly increased.

These considerations suggest that in the context of attempt, removing the requirement of an objective condition for the actus reus significantly reduces the burden of proof on the prosecution in order to punish a relatively small additional number of morally culpable actors. It thus creates an incremental but wide-ranging increase in the risk of enforcement error and abuse for very little added protection against criminal activity. Even if the inherent liberalism of the criminal law provides no more than a prudential rule to err on the side of protection against enforcement error and abuse, that would be sufficient to justify requiring the prosecution to establish an objective condition to obta