Impossibility Attempts: A Speculative Thesis

Peter Westen*
Courts and commentators have struggled for years to identify rules
to explain and justify certain widely-shared intuitions about impossibility
attempts, and they have proposed rules variously based upon (1) what
mistakes actors make, (2) what intentions actors possess, and (3) what
conduct actors perform. None of the proposals fully succeeds, however,
and none is able to explain the widely-shared intuition, which underlies
Sandy Kadishs inventive hypothetical regarding Mr. Law and Mr.
Fact, that some attempts based upon mistakes of law are just as
blameworthy as attempts based upon mistakes of fact. I propose an
alternative rule that, I believe, not only explains where and why people
possess widely-shared intuitions regarding impossibility attempts
(including regarding Mr. Law and Mr. Fact), but also explains where
and why people have conflicting intuitions. I argue that widely-shared
intuitions of blameworthiness and non-blameworthiness regarding
impossibility attempts are a function, respectively, of whether informed
citizens of the jurisdiction that enacted the statutory offense that the
defendant allegedly attempted to commit widely believe or disbelieve that
he would have been a threat to interests that the statute seeks to
protecta determination, in turn, that is a function of whether they
widely believe or disbelieve that he would have committed the offense
under counterfactual circumstances that they fear could have obtained.

Perhaps no aspect of the criminal law is more confusing and confused

than the common law of impossible attempts. 1

I. INTRODUCTION

I have had the pleasure, for twenty-five years, of teaching from Sandy Kadish
and Steve Schulhofers celebrated casebook, and I have learned something new
from it every year. Considering that impossibility in criminal attempts has been
one of Sandys favorite subjects, I shall use this occasion to ruminate about the
nature of impossibility and its implications for criminal responsibility generally.
Sandy and Steves casebook illustrates the puzzle that haunts impossibility
attempts, as well as an irony regarding the attention that scholars devote to it. The

*
Frank G. Millard Professor of Law, University of Michigan. I am deeply grateful to Larry
Alexander and Doug Husak for their comments.
1
JOSHUA DRESSLER, UNDERSTANDING CRIMINAL LAW 432 (4th ed. 2006).

523
524 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

puzzle, of course, is to identify a principle that explains and justifies the widely-
held intuition that, although certain actors who commit impossibility attempts
should be convicted, certain others deserve a defense of impossibility. 2 The
irony is that scholars devote attention to a few problematic cases that are so lacking
in practical importance 3 that scholars are obliged to invent fanciful hypotheticals to
discuss them. 4
Sandy illustrates the tenacity of the puzzle by trying twice to solve it; first
through the vehicle of The Case of Lady Eldons French Lace, 5 and, then,
through a Comment in a Hypothetical Law Review 6 in which he questions his
original solution and advances a superseding solution. Sandy also illustrates the
irony by devoting sixteen dense pages of his casebook to an issue that he discusses
by reference to an imagined Lady Eldon, who attempts but fails to smuggle lace in
the mistaken belief that the law declares lace of that kind to be contraband, 7 and to
an inventive hypothetical regarding Mr. Fact and Mr. Law, who both attempt
but fail to hunt out of season. 8
I shall explore a thesis that, I believe, resolves the puzzle and, in doing so,
vindicates scholars like Sandy who regard it as a window into criminal
responsibility generally. 9 Specifically, I shall argue that the resolution of

2
Although I disagree with Antony Duff about how to resolve impossibility attempts, see
infra note 29, Duffs magisterial CRIMINAL ATTEMPTS (1996) is the most scholarly and penetrating
philosophical and legal analysis that we possess to date regarding criminal attempts.
3
See Neil P. Cohen, Teaching Criminal Law: Curing the Disconnect, 48 ST. LOUIS U. L.J.
1195, 119798 (2004) (footnote omitted):
I have taught the first-year Criminal Law course for many years . . . [and addressed] the
traditional subjects, including such esoterica as impossible attempts . . . . A few years
ago, I was afforded the chance to leave teaching for a short period and become an
assistant district attorney prosecuting state cases in a medium-sized county. . . . During
my relatively brief tenure as a prosecutor, I came to realize that many important issues
routinely faced by lawyers in criminal law are simply ignored or given short shrift in the
basic Criminal Law course . . . . For example, I did not encounter any impossible attempt
cases, but I did see a significant number of assault-related incidents.
4
See, e.g., Kyle S. Brodie, The Obviously Impossible Attempt: A Proposed Revision to the
Model Penal Code, 15 N. ILL. U. L. REV. 237, 249 (1996) (hypothesizing an actor who gives a target
a dose of two aspirin, mistakenly believing that two aspirin will kill); Larry Alexander, Inculpatory
and Exculpatory Mistakes and the Fact/Law Distinction: An Essay in Memory of Myke Balyes, 12
LAW & PHIL. 33, 51 (1993) (hypothesizing a Mr. A who engages in adultery mistakenly believing
adultery is an offense).
5
SANFORD KADISH & STEPHEN SCHULHOFER, CRIMINAL LAW AND ITS PROCESSES 59498 (7th
ed. 2001).
6
Id. at 598601.
7
Id. at 597.
8
Id. at 599.
9
See, e.g., Graham Hughes, On Further Footnote on Attempting the Impossible, 42 N.Y.U.
L. REV. 1005, 1005 (1967):
[I]mpossibility . . . 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
2008] IMPOSSIBILITY ATTEMPTS 525

impossibility attempts presupposes a stealth requirement of criminal

responsibility that has hitherto gone largely undetecteda requirement of criminal
responsibility that the vast majority of criminal cases so readily satisfy that we
scarcely notice its existence, but that rarefied impossibility attempts force us to
confront.
Nevertheless, while the thesis arguably illuminates impossibility, it also has
the consequence of raising troubling questions regarding criminal attempts
generally. The thesis implies that when we punish persons for attempt and for
crimes of ulterior intent, we are fundamentally punishing them notor, at least,
not onlyfor what they have done, but for what we believe they would have done
under counterfactual circumstances that we fear could have obtained.
I shall proceed by (1) defining the scope and terms of the inquiry, (2)
canvassing existing proposals for solving the puzzle, (3) proposing a thesis to
explain and justify widely-shared intuitions about impossibility attempts, and (4)
discussing the distinctive challenge of punishing attempts based upon mistakes of
law. I will conclude by arguing that the thesis helps explain why people disagree
about impossibility attempts, and why juries rather than judges should be the final
arbiters of the stealth requirement of criminal responsibility that underlies
impossibility attempts.

II. THE SCOPE AND TERMS OF THE INQUIRY

My specific aim regarding impossibility is common to many scholars who

explore issues of criminal responsibility in general: (1) to explain, predict, and
justify whatever settled and widely-held intuitions of justice people may possess
regarding certain subjects, and (2) to account for any sharply conflicting intuitions
of justice they possess in the same field.
Interestinglyat least if my students are representativepeople seem to
possess conflicting intuitions about impossibility cases as often as they possess
shared intuitions. Thus, while my students broadly agree that an actor who
intentionally shoots to kill a person whom he knows is no threat, but misses, is
guilty of attempted homicide, they disagree about whether a person who buys
sharply-discounted property from a street vendor in the mistaken belief that it is
stolen is guilty of attempted receipt of stolen property. Considering these
disagreements, I shall first try to reduce their scope by distinguishing
disagreements that derive, not from conflicting intuitions regarding impossibility
itself, but from conflicting intuitions about attempts generally. Then, having
defined the class of attempts I shall be addressing, I shall define the subset that
consists of impossibility attempts.

time and energy to this question is perfectly proper, for . . . [it] is a splendid set-piece
which exhibits in a short space some of the most difficult issues of criminal law analysis.
526 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

A. Criminal Attempts Generally

Over the years I have polled students about how they would resolve certain
real and hypothetical impossibility cases. Based on the polls, I believe that some
of their differences of opinion are due to disagreements not about what is unique to
impossibility attempts, but about how broadly the underlying crime of attempt
itself ought to be defined. Specifically, I believe that many of their disagreements
are about (1) whether actors with certain mens rea ought to be punished for attempt
at all, (2) whether attempts to commit minor offenses ought to be punished at all,
and (3) how severe penalties for attempts ought to be.

1. The Mens Rea of Attempt

Everyone agrees that with respect to a charge of attempting to commit offense

X, it ought to suffice that an actor possess purpose regarding the elements of
offense X. Students tend to disagree, however, about whether purpose is
necessary. Some take the view that it is, indeed, necessarythat regardless of the
mental states that are otherwise required for offense X, no one ought to be
punished for any attempt to commit offense X, including an impossibility attempt,
unless the person acted with purpose regarding the material elements of offense X.
Others believe it also suffices that he possess mental states of knowledge or belief
regarding circumstance and result elements of offense X, at least with respect to
serious crimes, and regardless of whether it is an impossibility attempt. Still others
take the view that it suffices that actors possess the mens rea that offense X
requires regarding circumstance elements, even if the mens reas consists of mere
recklessness or negligence, and regardless of whether it is an impossibility attempt.
As a consequence, when these students disagree about whether actors with the
latter states of mind ought to be punished for impossibility attempts, their
disagreement is not about impossibility as such, but about something that is
extraneous to impossibility, namely, about whether persons with such mens rea
ought to be punished at all for attempt.

2. Attempts to Commit Minor Offenses

Everyone agrees that with respect to the classes of completed crimes for
which attempt liability obtains it ought to suffice that offense X is a serious
offense. Students tend to disagree, however, regarding whether it is necessary that
offense X be a serious offense. Thus, some take the view that it is, indeed,
necessarythat no one ought to be punished for any attempt to commit offense X,
including an impossibility attempt, unless offense X is itself a serious offense. In
contrast, others take the view that liability ought to exist for attempts to commit
minor offenses, such as receiving stolen property, regardless of whether they are
impossibility attempts. As a consequence, when these students disagree about
whether actors ought to be punished for impossibility attempts to commit minor
2008] IMPOSSIBILITY ATTEMPTS 527

offenses, their disagreement is not about impossibility as such, but about

something that is extraneous to impossibility, namely, about whether the crime of
attempt ought to extend at all to such offenses.

3. The Severity of Penalties for Attempt

Everyone agrees that with respect to penalties for attempt, incarceration can
be appropriate for attempts to commit the most heinous offenses like murder.
Students tend to be uncertain, however, about the penalties for attempt; and they
further disagree about whether any incarcerationand, if so, how much
incarcerationis appropriate for attempts to commit less serious offenses. As a
consequence, when students disagree about whether actors ought to be punished
for impossibility attempts to commit less-than-serious offenses, they may be
expressing uncertainty and disagreement, not about impossibility as such, but
about something that is extraneous to impossibility, namely, about the
appropriateness of certain real or imagined terms of incarceration for such
attempts.
In order to exclude these extraneous considerations, I shall confine the scope
of my inquiry. Rather than seek a principle that explains and justifies peoples
intuitions of just punishment across the range of what counts as a criminal attempt
in Anglo-American law, I shall confine myself to criminal attempts and
punishments of certain kinds. By an attempt, I mean an act or omission by an
actor who, while possessing purpose or belief regarding circumstance and result
elements of a crime, purposely does or omits to do anything which, under the
conditions that he believes exist at the time, is a substantial step in a course of
conduct planned to culminate in a crime. By punishment for attempt, I mean the
public act of officially declaring an actor to be guilty of the attempt and making
it part of his criminal record, regardless of any additional term of incarceration.
These stipulations significantly reduce the areas in which my students
disagree about impossibility attempts. Nevertheless, as we shall see, areas of
disagreement remain to which I shall return in the Conclusion.

B. Impossibility Attempts

Impossibility is one of two overlapping ways in which an actor who

intentionally undertakes to commit an offense (and who does not voluntarily
abandon the undertaking) can nevertheless fail to complete it, thereby leaving
himself guilty of attempt. The two overlapping ways are (1) through interruption,
and (2) by means of impossibility. 10
10
Larry Alexander argues that all attempts are impossibility attempts without distinction.
See Alexander, supra note 4, at 45. As a result, Alexander is implicitly obliged to take the position
that an actor who tries but fails to kill a target by burning down the targets home is guilty of an
impossibility attempt when, after he sets the fire and departs the scene, firemen arrive and narrowly
rescue the victim.
528 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

Criminal undertakings fail because of interruption when events operate to

frustrate actors from, in any sense, carrying out all that they intend to do. Thus, an
actor fails because of interruption when he is unexpectedly arrested at the entry of
a bank before he can make a threatening demand, when his pistol accidentally slips
from his hand before he can pull the trigger, when he suffers a debilitating heart
attack just before grabbing an intended kidnap victim, and when his intended rape
victim successfully fights him off. The hallmark of interruption attempts is that
because the actor is interrupted before he can do all that he intends, it is always
possible, even if it is unlikely, that he would voluntarily repent before effectuating
the offense.
Criminal undertakings fail because of impossibility when actors make
mistakes of a certain kind regarding their ability to commit the offenses they
intend. The mistakes, in turn, are about conditions that actors believe exist at the
time they act. Actors may mistakenly believe that circumstance elements of an
offense exist, 11 e.g., that sexual intercourse with a woman is without her consent.
They may mistakenly believe that means or bases exist for fulfilling conduct or
result elements of an offense, 12 e.g., that a gun is loaded or that an intended
homicide victim is alive. They may mistakenly believe that criminal prohibitions
exist, e.g., that adultery is an offense. In any event, a criminal undertaking fails
because of impossibility when, though the actor would be committing a crime if
he did everything he intends under the conditions that he believes exist at the time,
what he actually doesor what he would do if he fully acted on his intentis not
the offense he intends to commit, because the conditions are not what he believes
them to be. 13
Impossibility attempts, in turn, fall into two categories depending upon
whether the attempts are interrupted. 14 An actor commits an uninterrupted
impossibility attempt when, in committing an impossibility attempt, he, in some
sense, does everything he intends. Thus, an actor commits an uninterrupted
impossibility attempt when he purposefully shoots to kill a person whom he
believes is alive, only to discover his intended victim is already dead; when he

11
J.C. Smith refers to these conditions as pure circumstances because, when present, they
always have the effect of satisfying a circumstance element of an offense. See J.C. Smith, Two
Problems In Criminal Attempts, 70 HARV. L. REV. 422, 424 (1975).
12
J.C. Smith refers to these as consequential circumstances because, rather than being a
required circumstance element of an offense, they can have the consequence of enabling an actor to
commit a conduct element or achieve a result element. See id. at 425.
13
For a different definition of impossibility, see George Fletcher, Constructing a Theory of
Impossible Attempts, 5 CRIM. JUST. ETHICS 53, 5759 (1986), classifying the act of shooting and
missing as not being an instance of impossibility.
14
By the same token, interruption attempts also fall into two categories, depending upon
whether they are also impossibility attempts. An actor commits an interruption attempt that is not
also an impossibility attempt when, though he would be committing a crime if he does everything he
intends under facts and law as he correctly believes them to be, events operate to prevent him from,
in any sense, doing all he intends to do.
2008] IMPOSSIBILITY ATTEMPTS 529

purposefully shoots but misses a person whom he believes is in his line of fire,
only to discover that his intended victim was not in the line of fire; or when he
purposefully absconds with property that he believes he is taking without the
owners consent, only to discover that the owner consented. An actor commits an
interrupted impossibility attempt when, in the course of his committing an
impossibility attempt, events operate to frustrate him from, in any sense, carrying
out all he intends to do. Thus, an actor commits an interrupted impossibility
attempt when the police arrest him just as he is about to pull the trigger of a gun he
believes to be loaded but is actually empty; or when, for the purpose of stealing, he
breaks into a locked safe that he believes may contain money but that turns out to
contain no money at all.

III. EXISTING PROPOSALS TO SOLVE THE IMPOSSIBILITY PUZZLE

The three leading tests for solving the impossibility puzzle focus respectively
on (A) what an actor intends to do, (B) what kind of mistake he makes, and (C)
what kind of act he performs. Unfortunately, all three tests fail to explain some
widely-shared intuitions. In addition, the first two tests tend to be hijacked by
bogus tests that masquerade under similar names, while the third is a rule of law
that seeks to do what is better done by a rule of evidence.

A. What an Actor Intends to Do

Some courts and commentators seek to solve the impossibility puzzle by

asking of an actor, What did he intend to do? As we shall see, this test is capable
of being applied in a manner that is instrumental and moderately useful.
Unfortunately, the intent test can also be hijacked by a bogus test that renders it
either random or conclusory.
One of Sandys featured cases, People v. Jaffe, 15 illustrates the bogus manner
in which the intent test can be hijacked. The Jaffe court purported to resolve the
impossibility puzzle by (1) identifying what act the actor at issue intended to
perform, and (2) determining whether the act was criminal or non-criminal. 16 An
actor is guilty of attempt, the Jaffe court said, if the act he intended to perform is a
crime. 17 Otherwise the actor is innocent. Thus, Jaffe said, an actor who takes
possession of property in a sting operation, mistakenly thinking that it is being
tendered without the owners consent, is not guilty of attempted larceny because
the act he intends to perform (i.e., the act he actually effectuates of taking
possession of particular property and intentionally carrying it away) is not a crime
given that the property is being tendered to him with the owners consent. 18 In
15
78 N.E. 169 (N.Y. 1906).
16
Id. at 170.
17
Id.
18
Id.
530 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

contrast, Jaffe said, an actor who reaches into a pocket for the purpose of picking it
only to discover the pocket is empty, is guilty of attempted pick pocketing because
the act he intends to perform (i.e., the act of picking a pocket of valuables) is a
crime. 19
As Sandy himself points out, 20 the flaw in the Jaffe test is that, although it
purports to inquire into the one act that an actor in an impossibility case intends to
perform, it defines intent in such a novel way that an actor invariably intends to
perform two acts:

Act #1: the act that a person in an impossibility case actually performs
and that is never a crime, e.g., the act of taking and carrying away
property with the consent of the owner; and

Act #2: the act that a person in an impossibility case mistakenly believes
he is performing and that is always a crime, e.g., the act of picking a
pocket.

Because Jaffe seeks a single intended act where two intended acts exist, an
actors liability for attempt ultimately depends not upon which act he really
intendsgiven that he really intends bothbut rather upon which act a court
happens to highlight after the fact.
The defect in Jaffes approach, therefore, is not that it dictates unjust results,
but that it dictates no results at all. Jaffe is a non-test that provides no instrumental
guidanceno set of instructionsfor a court that wishes to know whether to
acquit or convict. It can be applied, willy-nilly, either to acquit or convict a
defendant in an impossibility case, depending upon whether the court emphasizes
act #1 or act #2. Thus, by emphasizing act #2, a court can convict an actor like Mr.
Jaffe by asserting that he intended to do what he mistakenly thought he was doing,
i.e., to take and carry away property without the owners consent. By emphasizing
act #1, a court can acquit a failed pickpocket by asserting that he intended to
perform the act that he actually performed, i.e., to put his hand in a pocket that was
empty. This means that Jaffes approach is either conclusory or random,
depending upon whether a court has already decided whether a given actor should
be acquitted or convicted. For a court that has consciously or unconsciously
decided to acquit or convict, Jaffe is a set of conclusory labels that the court
invokes after the fact to rationalize a decision it has previously reached on unstated
grounds. For a court that has not already decided, Jaffe randomly triggers either
acquittal or conviction, depending upon which of the two acts happen to captivate
the court.
Now this is not to say that an actors intent is always a non-test. George
Fletcher propounds a version of the intent test that is instrumental rather than

19
Id. at 16970.
20
KADISH & SCHULHOFER, supra note 5, at 59496.
2008] IMPOSSIBILITY ATTEMPTS 531

conclusory. Fletcher recognizes that in impossibility cases, an actor can be

regarded as intending both act #1 and act #2 (rather than either act #1 or act #2). 21
Whether the actor possesses a defense of impossibility, Fletcher argues, depends
upon which is the actors motivating intent. 22 The measure of motivating intent,
in turn, is which of the two actsact #1 or act #2he would perform if he were
disabused of his mistake in time to take corrective action. As Fletcher puts it:

[P]eople attempt to achieve only those ends that affect their motivation in
acting. Their mistaken beliefs constitute part of the attempted act only so
far as being disabused of their mistakes would alter their course of
conduct. This is an argument . . . that is rooted in the ordinary
understanding of what it is, in daily life, to try to achieve a particular
goal. 23

Accordingly, an actor is not guilty of attempt who is so motivated that, if he

were disabused of his mistake, he would continue unfazed to perform a harmless
act of the kind he actually performed. In contrast, an actor is guilty of an attempt
who is so motivated that, if he were disabused of his mistake, he would change
course and try to perform the criminal act he thought he was performing.
Fletchers test has the advantage of being able to explain many widely-shared
intuitions. Thus, the test explains why an actor is guilty who reaches into another
persons pocket in the mistaken belief that it contains valuables, and why an
adulterer who engages in adultery in the mistaken belief that adultery is a crime is
not guilty. The former is guilty because, if he were disabused of his mistake in
time to take corrective action, he would find and pick a pocket that contained
valuables. The latter is not guilty because, were he disabused of his mistake, he
would perform the same innocuous act he actually performed, i.e., adultery, while
this time relieved that he was not committing a crime.
Unfortunately, Fletchers test counterintuitively acquits actors whom most
observers would convict. Thus, consider a person who, wishing a romantic rival
dead, shoots to kill in the mistaken belief that his rival is asleep, only to learn that
his rival had already died of heart disease a few hours earlier. If my students are a
measure, most people would convict such an actor. Yet Fletcher would acquit him
because if the actor were apprised of his mistake in time to correct his course of
action, the actor would do something that is even more innocuous than he

21
GEORGE P. FLETCHER, RETHINKING CRIMINAL LAW 179 (1978).
22
Fletcher argued in Rethinking Criminal Law that, although the motivational test is the sole
test of liability in impossibility cases, it is confined to a certain subset of impossibility cases, namely,
attempts to commit crimes other than crimes, such as murder, that involve the central harms that the
criminal law seeks to prevent. Id. at 15566, 184. For the latter crimes, Fletcher advocated a
manifest-criminality test. See id. at 14655, 184. More recently, however, Fletcher has argued that
the motivational test is a necessary part of a two-part test that together applies to all impossibility
cases, the other part being the manifest-criminality test. See Fletcher, supra note 13, at 64.
23
FLETCHER, supra note 21, at 18182.
532 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

originally did: he would quietly leave the scene in good conscience, knowing that
his rival was about to die from natural causes. 24

B. What Kind of Mistake an Actor Makes

Another leading test, including Model Penal Code [MPC] section 5.01, is
based upon whether an actor makes a mistake of fact or a mistake of law. The
clearest version of the law/fact testwhich Sandy himself initially expounds in
The Case of Lady Eldons Lace 25 explicitly invokes the terms law and
fact, by providing a defense to actors who make mistakes of law and not to those
who make mistakes of fact. 26
A less clear, but nevertheless equivalent, version of the law/fact test is the
Model Penal Code test. 27 MPC section 5.01(1) contains two elements that together
functionally duplicate the law/fact test: (1) the first element bases an actors
liability upon what he believes the circumstances 28 to be, even if he is
mistaken in his belief; (2) the second element conditions an actors liability for
attempt to commit offense X upon his possessing the mens rea that offense X itself

24
For another normative critique of Fletchers motivational view of intent, see Kenneth
Simons, Mistake and Impossibility, Law and Fact, and Culpability: A Speculative Essay, 81 J. CRIM.
L. & CRIMINOLOGY 447, 47071 (1990).
Larry Alexander criticizes Fletchers test on the ground that it is premised on counterfactual
determinations that, in fact, are indeterminate. See Alexander, supra note 4, at 6364. Alexander
illustrates this critique by pointing out that Fletcher and Myke Bayles come to different conclusions
about whether the defendant in Jaffe, 78 N.E. 169 (N.Y. 1906), would have altered his conduct if he
had been disabused of his mistake in time to take corrective action. I am skeptical of Alexanders
critique. The counterfactual determination that Fletchers test dictates is clear and simple, for it
requires that one reconceptualize the past by changing only one thing about it. It requires that one
conceptualize what an actor would do if he were made aware of his mistake in time to take correction
action. Fletcher disagrees with Bayles because in applying his test, Fletcher makes the mistake of
also changing something else. Unlike Bayles, who reconceptualizes the past by imagining what Mr.
Jaffe would do if he were apprised of the sting before acting, Fletcher also changes the property
owners reason for consentingchanging it from a desire to entrap the defendant and his accomplice
in a sting, to a desire to confer a veritable gift on the defendant and his accomplice.
25
KADISH & SCHULHOFER, supra note 5, at 59798.
26
See ANDREW ASHWORTH, PRINCIPLES OF CRIMINAL LAW 406 (3d ed. 1999) (discussing
modern English doctrine).
27
But see MODEL PENAL CODE COMMENTARY 5.01 pt.I, at 381 n.92 (1985) (muddying the
waters as to whether it tracks the fact/law distinction despite the way Model Penal Code 5.01(1)
clearly functions); see also Alexander, supra note 4, at 49 n.41.
28
Model Penal Code 5.01(1)s usage of the term circumstances is equivalent to my term
condition, and neither should be confused with Model Penal Code 1.13(9)s narrower reference
to circumstances [elements]. Model Penal Code 5.01(1)s reference to circumstances as he
believes them to be includes conduct elements and result elements as well as circumstance elements.
Thus, a person who is arrested just as he is about to shoot to kill a person who, as it turns out, is
already dead is guilty under Model Penal Code 5.01(1)(c), despite the fact that the circumstances
about which he is mistaken is the result element of killing.
2008] IMPOSSIBILITY ATTEMPTS 533

requires, something that no actor can ever possess who makes a mistake of law. 29
As a consequence, because Element One bases an actors liability upon what he
believes the circumstances to be, and because Element Two functions to
eliminate liability for actors who make mistakes of law, actors are liable for
impossibility attempts under MPC section 5.01(1) if, and only if, they make
mistakes of fact.
We shall see shortly that the law/fact test has several strengths.
Unfortunately, just as Jaffe hijacked the intent test, the law/fact test can be
hijacked by a bogus test that masquerades under comparable terms and, yet, is as
inconclusive as Jaffe. Indeed, the bogus test is Jaffe in disguise.
The bogus test mimics the law/fact test by employing the comparable terms
legal and factual impossibility. Like the law/fact test, the bogus test treats
legal impossibility as a defense to attempt liability, and factual impossibility as no
defense. However, the bogus test measures law and fact differently than the
genuine test. Rather than measure an actors liability by the kind of mistake he
makes as between law and fact, the bogus test measures his liability by whether it
is fact or law that precludes the act he intends from being a completed crime.
Another of Sandys cases, People v. Dlugash, 30 illustrates the bogus way in
which the law/fact test can be hijacked in the name of legal and factual
impossibility. Although New York had replaced its prior tests of impossibility
with MPC section 5.01 by the time Dlugash arose, Dlugash reviews New Yorks
prior effort to resolve impossibility attempts by distinguishing between legal and
factual impossibility. Dlugash starts by saying that legal impossibility is based
upon a mistak[e] of law, while factual impossibility is based upon a mistak[e]
of fact. However, Dlugash immediately contradicts itself by implicitly taking the
position that what determines whether an attempt is factually or legally impossible
is which of the two acts we discussed abovenamely, act #1 or act #2an actor is
adjudged to have intended. Thus, Dlugash refers to the shooting of a stuffed decoy
in the mistaken belief it is a living deer as an instance of legal impossibility
reasoning, apparently, that (i) the shooter must intend to shoot the target he
actually shoots, for otherwise he would not succeed in hitting the target, (ii) the
target he shoots is a stuffed decoy, and (iii) what renders it impossible for him to
commit the crime of poaching by shooting a decoy is the law that states that
poaching is the shooting of living animals. At the same time, Dlugash refers to
the futile picking of an empty pocket as an instance of factual impossibility
reasoning, apparently, that (i) the pickpocket, believing that the pocket might

29
See John Hasnas, Once More unto the Breach: The Inherent Liberalism of the Criminal
Law and Liability for Attempting the Impossible, 54 HASTINGS L.J. 1, 9, 3233 (2004). But see DUFF,
supra note 2, at 211 (arguing, mistakenly I believe, that it is legally possible to convict an actor under
current attempt statutes whose mistake is one of law, e.g., to convict an actor of attempted sexual
intercourse with a girl who is under 16, who, acting with the purpose of violating the law, has sexual
intercourse with a girl who he knows to be 17 in the mistaken belief that the age of consent is 18).
30
363 N.E.2d 1155 (N.Y. 1977).
534 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

contain valuables, intends to pick a pocket containing valuables, and (ii) what
renders it impossible for him to pick the pocket is the fact that the pocket is empty.
Needless to say, this understanding of factual and legal impossibility
suffers from the same fallacy as Jaffe: the fallacy of assuming that, as between acts
#1 and #2, actors in impossibility attempts intend to commit only one, when in
reality they intend to commit both. Because they intend to commit both, every
impossibility case is both an instance of factual impossibility and legal
impossibility, depending upon how it is characterized. Thus, instead of describing
the attempted poaching as a legal impossibility and the attempted pick pocketing as
factual impossibility, Dlugash could have said the opposite. It could have
characterized the poachers act as an instance of act #2, and the pickpockets as an
instance of act #1, thereby describing the former as factual impossibility and the
latter as legal impossibility. Consequently, this kind of inquiry into factual and
legal impossibility replicates the conclusory and random way the Jaffe court
inquired into what actors intended. 31
A better test of legal and factual impossibility focuses upon the kind of
mistake, as between law and fact, that causes what he actually does not to be an
offense. 32 The law, for the purposes of the test, consists of a full specification of
the act-types that the state officially declares to be punishable. 33 The facts
31
Commentators also fall into this trap. See, e.g., Alexander, supra note 4, at 3435
(referring to a mistake in thinking a gun is loaded as factual and referring to a mistake in thinking
that the owner has not consented to the transfer of property as legal, despite that the latter is not a
mistake with respect to which he would have benefited from having a good lawyer, but, is a mistake
with respect to which he would have benefited from information from having a good private
investigator).
32
Larry Alexander denies that any nonarbitrary line exits between law and fact in
impossibility cases. See id. at 45. He argues that, If [the] application [of a law] is part of [its]
meaningand consider whether one knows the meaning of a law if he cannot identify any actual
extension of it in the worldthen factual mistakes are legal ones, and factual impossibility is a
species of legal impossibility. Id. at 52. Despite my admiration for Alexanders command of legal
theory, I think Alexander confuses application of law qua a states full specification of the act-types
it prohibits and application of the law qua actual act-tokens thereof. Yes, every criminal event is
an act-token of an act-type that the state has declared to be prohibited. But mistakes regarding what
act-types are prohibited (law) differ from mistakes regarding whether conduct is an act-token thereof
(fact). Both mistakes can result in a person thinking that he has violated the law when he has not, but
the sources of the mistakes differ. Both mistakes can also, confusingly, be said to be mistakes of
fact. But one is a mistake about the fact of what act-types the state has declared to be prohibited,
and the other is a mistake about whether conduct possesses the empirical features of an act-token
thereof.
33
Kenneth Simons would divide all such mistakes of law into two categories: (1) mistakes
regarding the governing law that declares conduct to be an offense, which Simons treats as
exculpatory in attempt cases, and (2) mistakes regarding elements of an offense, which, he says,
some commentators would treat as inculpatory in attempt cases. See Simons, supra note 24, at 457.
Much as I respect Simons, I agree with Alexander that no analytical line can be drawn between
mistakes regarding a governing law and its elements. See Alexander, supra note 4, at 3940, 43,
57 (discussing Lady Eldon in his hypotheticals 1.a. and 1.b.). Every person who makes a
mistake regarding an element of an offense also, necessarily, makes a mistake about the governing
law.
2008] IMPOSSIBILITY ATTEMPTS 535

consist of the empirical features that determine whether conduct is an act-token of

what is acknowledged to be a prohibited act-type. It follows, therefore, that no
middle ground exists between law and fact, 34 and there are no mixed 35 mistakes
that consist of neither one nor the other. An actor makes a mistake of law in an
impossibility case and, hence, has a defense, if he is in need of the services of a
good lawyerthat is, if, although he knows what he is empirically doing, he
mistakenly believes that the state has officially declared acts of that type to be
punishable. An actor makes a mistake of fact in an impossibility case, and, hence,
has no defense, if he is in need of the services of a good private investigatorthat
is, if, although he knows what act-types the state officially declares to be
punishable, he mistakenly believes that his conduct is an act-token thereof.
The best critique of the law/fact test is the one that Sandy levels in his
Comment in Hypothetical Law Review against the position he previously
expounded in The Case of Lady Eldons Lace. The problem, as Sandy admits in
Comment, is that, while the categories law and fact generally correspond
with shared intuitions about which attempts are exculpatory and inculpatory,
respectively, the match is not perfect: mistakes of fact can occur that observers
commonly regard as exculpatory, and mistakes of law can occur that observers
commonly regard as inculpatory.
To illustrate a mistake of fact that observers commonly regard as exculpatory,
consider the following version of a case that Sandy poses in his Comment, which I
elaborate as follows: 36

Midwestern Voodoo
Mildred, a gullible, fifty-five-year-old house-bound woman in North
Dakota, deeply resents her ex-husband and his trophy wife in Palm
Beach, Florida. Mildred reads in the National Examiner and on the

34
Larry Alexander hypothesizes four cases that he thinks fall in the middle. Alexander, supra
note 4. However, the four reduce essentially to two, and neither is an instance of a middle ground.
Thus, an actor who knows that the state declares hunting on days with a red flag to be an act-type of
poaching, but who, being colorblind, sees a red flag where there is actually a green flag, makes a
mistake of fact because, while he knows what act-types are prohibited, he mistakenly thinks that his
conduct possesses one of the empirical features that, if present, constitutes what he knows is the act-
type. In contrast, a woman who imports French lace knowing that she possesses French lace but
misreading the statutory prohibition on Flemish to be a prohibition on French lace is making a
mistake of law, because while she knows the empirical features that her conduct actually possesses,
she does not know that the state has declared conduct with such features to be a prohibited act-type.
35
For commentators who believe that a third category of mixed mistakes of law and fact
exists, see Ira P. Robbins, Attempting the Impossible: The Emerging Consensus, 23 HARV. J. ON
LEGIS. 377, 39497 (1986); Fernand N. Dutile & Harold F. Moore, Mistake and Impossibility:
Arranging a Marriage Between Two Difficult Partners, 74 NW. U. L. REV. 166, 184201 (1979);
FLETCHER, supra note 21, at 686 (describing the application of law as being intermediate between
law and fact). However, they are mistaken. There is no middle ground between its being a mistake
of law or fact. See Simons, supra note 24, at 458.
36
KADISH & SCHULHOFER, supra note 5, at 598.
536 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

Internet about lethal, Haitian-origin Voodoo rituals that novices can

practice in their home, provided that they purchase a recommended
booklet, do the exercises, and possess DNA-containing tissue of their
target. Mildred, who possesses a lock of her ex-husbands hair that he
gave her when they were dating, buys the book and becomes convinced
that it is worth a try. Mildred reads that, although success is not
guaranteed, a good chance always exists that Voodoo will succeed. She
is relieved to learn that the method leaves no clues and does not require
her to be in the presence of her husband. She is also comforted to learn
that responsibility does not fall solely on her own shoulders, for she will
not succeed unless the spirit world independently agrees that her target
deserves his ill fate. Mildred practices a ritual designed to cause her ex-
husband to die of disease. When it does not work, she confesses to a
friend who, in turn, notifies the police who arrest her for attempted
murder.

My students disagree about whether all instances of Voodoo are exculpatory.

However, when they are asked to pass judgment on Mildreds case as if they were
prosecutors, they agree that, whether or not Mildred deserves Gods punishment
for endeavoring to kill an innocent person, the state would be abusing its power if,
given Mildreds mistake of fact, the state officially declared her to be an
attempted murderer and made it part of her public record. 37
Now consider mistakes of law that my students regard as being just as
inculpatory as most mistakes of fact. Again, the best illustration is one of Sandys,
i.e., his inventive hypothetical involving Mr. Fact and Mr. Law, which I elaborate
as follows:

Mr. Fact and Mr. Law

Mr. Fact and Mr. Law both set out independently to get a jump on the
bow-hunting season by sneaking out a day before the season begins. For
technical reasons, the exact date of the states bow-hunting season for
deer tends to change from year to year, but this date this year is Friday,
October 15. Ironically, Messrs. Fact and Law each make a mistake that
results in their stalking and killing deer on what they mistakenly believe
to be the day before the hunting season but is actually the first day of the
season itself (Friday, October 15). Mr. Fact makes the factual mistake of
thinking, Today is Thursday, October 14. Mr. Law makes the legal
mistake of thinking, The season begins on Saturday, October 16.

37
Cf. FLETCHER, supra note 21, at 166 ([O]ne case in which virtually everyone agrees that
there should be no liability . . . is 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 an incantation
to banish ones enemy to the nether world.).
2008] IMPOSSIBILITY ATTEMPTS 537

While each is butchering his deer carcass, he is each approached by a

game warden who intends to congratulate him. Instead, Mr. Fact and
Mr. Law both confess, thinking they have been caught red-handed while
hunting out of season.

My students disagree about whether attempted hunting out of season ought to

be a crime in the first place. However, they agree with Sandy and other scholars 38
who have commented on the case that in terms of culpability, Mr. Fact and Mr.
Law are equally blameworthy.
In sum, the law/fact test has two defects. First, it fails to explain and predict
all shared intuitions about impossibility attempts. Second, like Fletchers intent
test, even when it does explain shared intuitions, it fails to go further and also
justify themthat is, it fails to ground them in normative principles of just
responsibility. 39

C. What Kind of Objective Act an Actor Performs

The third test, which is typically attributed to Arnold Enker, 40 focuses on the
objective act 41 that an actor performs. Several federal courts have adopted

38
See, e.g., Hasnas, supra note 29, at 3132; Dutile & Moore, supra note 35, at 184. Cf.
HYMAN GROSS, A THEORY OF CRIMINAL JUSTICE 20911 (1979) (some mistakes of law are just as
blameworthy as analogous mistakes of fact).
39
Cf. Alexander, supra note 4, at 64 (Fletchers test has no . . . normative basis).
40
See Arnold N. Enker, Impossibility in Criminal AttemptsLegality and the Legal Process,
53 MINN. L. REV. 665 (1969). Enker himself acknowledges the prior work of Graham Hughes. Id. at
684 (citing Graham Hughes, One Further Footnote on Attempting the Impossible, 42 N.Y.U. L. REV.
1005 (1967)).
41
Enkers test is not the only legitimate test that is predicated on the objective nature of an
actors conduct. Other objective tests have also been advanced, see infra note 104, including by
Antony Duff. Duff argues that an actor is not guilty of an impossibility attempt unless (1) the actor
possesses a mental state of purpose regarding circumstance elements of the offense, and (2) the
actors action, in the circumstances as they would appear to the reasonable observer in the agents
situation, is at all plausible as a means of trying to actualize the [agents] further intention. DUFF,
supra note 2, at 20809, 211, 22629, 23234, 37883, 39899. The problem with Duffs test is that
it would acquit actors who are widely regarded as culpable. Thus, it would acquit an actor of
attempted murder who, in an effort to recover insurance on airplane cargo, plants a bomb that he
believes will kill the planes passengers but that fails to detonate because of a faulty fuse, see id. at
20809. And it would acquit an actor of attempted rape who has sexual intercourse at a party with a
woman whom he believes and hopes is unconscious from drink but who actually just died. See id. at
229.
538 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

variations of Enkers test, 42 and Sandy, despite his earlier doubts about it,43
eventually embraced it also. 44
Enkers test is best understood by what it is a reaction to. Enkers test is a
reaction to what Enker regards as an ominous trend toward subjectivism in
criminal law, that is, a trend toward basing an actors criminal liability less upon
his actually engaging in objective conduct that the state specifically prohibits,
and more on his being subjectively willing to do so. Enker worries that if
overzealous prosecutors are freed from the obligation to prove objective conduct,
they will be tempted to invoke questionable evidence of criminal intent, such as
jailhouse informers, accomplice testimony, and prior crimes, to convict innocent
defendants who lack an objective basis for rebutting false attributions of criminal
intent. 45
Enker directs his criticism to criminal attempts because, in contrast to other
inchoate offenses which specify some objective prohibited conduct on an actors
part, criminal attempt statutes can be satisfied by any objective conduct that is
consistent with an actors criminal intent. Among all attempts, Enker further
focuses on a certain class of impossibility attempts, namely, where an actor makes
a mistake of fact regarding the existence of circumstance elements of an offense 46
(e.g., that sexual intercourse is without consent). Enker argues that the latter
attempts present a particularly high riskand, ultimately, an unacceptable riskof
convicting the innocent.
To make the point, Enker contrasts this latter class of impossibility attempts
with all other attempts. When an actor is charged with a non-impossibility
attempt, Enker says, the actor will have engaged in objective conduct that, though
it is not itself a crime, is nevertheless suspicious. Thus, when an actor is charged
with attempted murder for having lain in wait to shoot a victim, he will have
engaged in the objective and suspicious conduct of (i) concealing himself, (ii) in
the proximity of a putative victim, (iii) while armed with a weapon. Similarly,
Enker says, when an actor is charged with an impossibility attempt based upon a
mistake regarding a conduct element or result element of an offense, he will also
have engaged in objective conduct that is suspicious. Thus, when an actor is
charged with attempted murder for trying to shoot a person with a gun that
jammed, he will have engaged in the objective and suspicious conduct of (i)
pointing a gun at a putative victim, and (ii) pulling the trigger. In contrast, Enker

42
See United States v. Oviedo, 525 F.2d 881, 885 (5th Cir. 1976); United States v. Everett,
692 F.2d 596, 600 (9th Cir. 1982); United States v. Innella, 690 F.2d 834, 835 (11th Cir. 1982).
43
See KADISH & SCHULHOFER, supra note 5, at 596.
44
Id. at 60001. For commentators who have adopted similar tests, see FLETCHER, supra note
21, at 14657 (a manifest criminality test of liability for attempts to commit crimes of social harm);
Robbins, supra note 35, at 33943, 398412, 41719; Thomas Weigend, Why Lady Eldon Should Be
Acquitted: The Social Harm in Attempting the Impossible, 27 DEPAUL L. REV. 231, 26673 (1977).
45
See Enker, supra note 40, at 670, 682, 68788, 692.
46
Id. at 669, 679.
2008] IMPOSSIBILITY ATTEMPTS 539

says, no such suspicious, objective conduct obtains in the event of impossibility

attempts based upon factual mistakes regarding circumstance elements. Thus,
when Lady Eldon is charged with attempting to import French lace based upon her
transporting English lace in the mistaken belief it was French, her objective
conduct is not only lawful (as is always the case with criminal attempts), but also
unsuspicious. For it consists of the innocuous conduct of (i) transporting, (ii)
English lace, (iii) that is non-dutiable. 47
For these reasons, Enker rejects the Model Penal Code, which renders actors
culpable for all mistakes of fact in impossibility cases. Instead, he proposes that
courts continue to do what he claims many courts have been doing all along,
namely, generally acquit actors who engage in impossibility attempts based upon
factual mistakes regarding circumstance elements. Specifically, he advocates
dividing all attempt cases into two sets, which he calls legal and factual
impossibility, respectively, each consisting of two subsets:

Legal Impossibilityconsisting of (1) impossibility cases based upon

mistakes of law, and (2) impossibility cases based upon mistakes of fact
regarding circumstance elements; 48

Factual Impossibilityconsisting of (1) non-impossibility attempts,

and (2) impossibility attempts based upon mistakes of fact regarding
conduct and result elements. 49

Enker subjects all instances of legal impossibility to a per se rule, treating

all of them as non-culpable. With respect to instances of factual impossibility,
Enker entrusts it to the courts to decide on a case-by-case basis whether the
features of an actors objective conduct sufficed to corroborate the allegation of
criminal intent. 50
The strength of Enkers thesis is that it highlights a legitimate concern about
impossibility attempts, namely, that where an actors objective conduct does not
corroborate his alleged intent, risks increase that he will be wrongly convicted
based upon what false witnesses claim to have been his criminal intent.
Nevertheless, the deficiencies in Enkers thesis are several.
First of all, Enker mistakenly assumes that the risks of false conviction are
distinctive to attempt cases, when, in reality, they exist to the same extent in non-
attempt statutes that take the form, It is a crime for a person to do X with intent to
commit offense Y. Yet despite the fact that commission of X itself is neutral on
the probability of an actors committing offense Y, no jurisdiction requires that an
actors objective conduct in committing X corroborate his intent to commit offense
47
Id. at 677.
48
Id.
49
Id. at 669.
50
Id. at 698703.
540 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

Y. Instead, jurisdictions rely on the adversary process and on the presumption of

innocence to ferret out false testimony of intent. Consider, for example, a statute
that prohibits assault with intent to kill. Although particular assaults may
possess features that are circumstantial evidence of intent to kill, assaults are not
themselves evidence of intent to kill. Yet no jurisdiction requires as a matter of
substantive criminal law that the objective circumstances of an actors assault
corroborate his intent to kill.
Second, Enker mistakenly assumes that because risks of false conviction
correlate with impossibility attempts based upon factual mistakes regarding
circumstance elements, they are necessarily present in such cases. To illustrate the
mistake, consider the following case, which Enker would classify as legal
impossibility and, hence, non-punishable: an actor, having spent thousands of
dollars refitting his car with a false container, fills it with packets of white power
that he is observed purchasing at a great cost and is later stopped at the border
only to discover that the packets contain innocuous talcum powder. The actors
mistake is about the existence of a circumstance element of the offense, i.e., that
the material he is transporting is cocaine. Yet his objective conduct, i.e., the
construction of the hidden container, the concealing of packets, and the enormous
sum he paid, all strongly corroborate his criminal intent.
Third, Enker mistakenly assumes that evidence of objective conduct is the
only way to decisively corroborate the existence of criminal intent. Criminal intent
can be decisively provedand sometimes better provedby evidence in the form
of wired, recorded, and videotaped conversations. Consider the following case,
which Enker would classify as legal impossibility and, hence, non-punishable:
the police, as part of a sting operation, videotape a suspect, A, as he talks in detail
about wishing to purchase a kilo of cocaine; the police, having bugged A, also
record his telephone conversations with potential purchasers; on the day agreed
upon, A tenders the undercover agent $50,000 in return for a kilo of white power
to A that turns out to be talcum power, whereupon the police arrest A for
attempting to purchase cocaine. Again, contrary to what Enker assumes, this is a
case in which the actors objective conduct does corroborate his intent to purchase
drugs because no one would tender $50,000 for innocuous powder. However, As
videotaped and recorded conversations are even better proof of As intent than his
objective conduct, and would stand alone to support a conviction. 51
Fourth and finally, Enker mistakenly assumes that because it is conceptually
difficult in law to frame an actors liability for a mistake of law, it is impossible. 52
As a consequence, Enker groups all mistakes of law together and exculpates all
who make them, despite the fact that Mr. Law is generally regarded as being just
as blameworthy as Mr. Fact.

51
John Hasnas, who otherwise agrees with Enker, admits that an actors statements can be as
probative of his intent as evidence of objective conduct. See Hasnas, supra note 29, at 68 n.186, 73.
52
Enker, supra note 40, at 669 n.13.
2008] IMPOSSIBILITY ATTEMPTS 541

Some federal courts purport to adopt Enkers test, but, in reality, they depart
from it significantly. 53 While they agree with Enker that all mistakes of law are
exculpatory, they ignore Enkers distinction regarding mistakes of fact about
circumstance elements and, instead, subject all mistakes of fact to the same
analysis that Enker recommends for what he calls factual impossibility.
Specifically, they subject all mistakes of fact to a case-by-case analysis to
determine whether an actors objective conduct, unaided by evidence of his
criminal intent, suffices to corroborate the states allegation of criminal intent. The
Fifth Circuit in United States v. Oviedo puts it this way:

[W]e demand that in order for a defendant to be guilty of a criminal

attempt, the objective acts performed, without any reliance on the
accompanying mens rea, mark the defendants conduct as criminal in
nature. The acts should be unique rather than so commonplace that they
are engaged in by persons not in violation of the law. 54

In addition, neither Oveido nor the courts that claim to follow it actually
practice what the quotation in Oveido literally preachesor what Enker would
presumably wish them to practice. The quotation in Oviedo purports to require
courts to examine an actors objective acts without any reliance on the
accompanying mens rea, 55 a requirement consistent with Enkers suspicion of
mens rea evidence. In practice, however, courts interpret objective acts to
include the very thing that Enker would exclude, namely, hearsay statements in the
form of an actors admissions of what he intends. Thus, in deciding that Mr.
Oviedos objective acts corroborated the states allegation of criminal intent, the
Fifth Circuit took into account not only what he physically did but also what a state
undercover agent reported he said. 56
In sum, Arnold Enkers test has four deficiencies set forth above. In contrast,
the test that the federal courts mistakenly attribute to him has only one deficiency,
i.e., that it aggregates all mistakes of law and exculpates them all. However, in
order to avoid the other three deficiencies, the federal courts have largely
transformed Enkers test from a controversial rule of substantive criminal law into
an innocuous, sufficiency-of-the-evidence rule to the effect that persons not be

53
See cited cases supra note 42.
54
United States v. Oviedo, 525 F.2d 881, 885 (5th Cir. 1976). See also United States v.
Innella, 690 F.2d 834 (11th Cir. 1982).
55
Oviedo, 525 F.2d at 885 (emphasis added).
56
Id. (Here we have only two objective facts. First, Oviedo told the agent that the substance
he was selling was heroin . . . .). Accord United States v. Innella, 690 F.2d 834, 835 (11th Cir.
1982) (In the present case Innellas objective acts were unequivocal. His words . . . were consistent
with an attempt to purchase a controlled substance.).
542 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

convicted of an impossibility attempt based upon mistakes of fact unless their

conduct and statements support the existence of criminal intent on their part. 57

IV. A PROPOSAL

The key to the impossibility puzzle lies in examining why the law/fact test so
often gets it right. If we can answer that question, we will have identified a test
that explains and predicts all our widely shared intuitions. Then, having identified
the test, we inquire into whether the test is normatively justified. Before we do so,
however, it may be useful to examine several related paths that others have
traveled.

A. Traveled Paths

1. Future Threat

Why, then, does the law/fact test so often get it right? When my students are
asked that question, they give an answer that, I think, comes close to being correct
and yet falls short. The reason, they say, is that mistakes of fact and law correlate
with the respective presence and absence of future danger. When defendants make
mistakes of fact that are widely intuited to be inculpatory, they reveal themselves
to be future dangersthat is, they reveal themselves to constitute future threats to
interests the law seeks to protect. When defendants make mistakes of law that are
widely intuited to be exculpatory, they reveal themselves not to constitute such
future threats. 58 Thus, my students say, consider the paradigmatic cases that are
widely regarded as inculpatory and exculpatory, respectively, namely (1) where an
actor shoots to kill but misses, and (2) where an actor engages in adultery
mistakenly thinking it is a crime. Actor 1 (whose mistake is one of fact) is guilty
57
See Robbins, supra note 35, at 418 (suggesting that Oviedo is essentially a sufficiency-of-
the-evidence test). If the federal court test falls short of a sufficiency-of-the-evidence test, it is only
because, while they take into account everything an actor says in the course of his conduct, they may
exclude what he says afterwards, such as to confidants or the police. For commentators who
embrace the latter approach, see Hasnas, supra note 29, at 7177; Weigend, supra note 44, at 266
73.
Some commentators criticize the Model Penal Code for failing to make its corroboration
requirement for interrupted attempts applicable to non-interrupted impossibility attempts as well. See
Enker, supra note 40, at 682; Robbins, supra note 35, at 411, 42230. I believe they misunderstand
the purpose of the Codes corroboration requirement. Its purpose is not to provide objective evidence
that actors charged with attempt possess criminal intent. (If that were its purpose, it would, indeed,
be a failing to refrain from requiring it in all attempt cases). Rather, it is to provide objective
evidence of a certain kind of intent in a certain kind of attempt casesnamely, objective evidence in
interrupted-attempt cases that an actor was sufficiently committed to his criminal enterprise to wish
to carry through with it.
58
For commentators who assume that references to threats are references to future threats and
dangers, see Kevin Cole, The Voodoo We Do: Harm, Impossibility, and the Reductionist Impulse, 5 J.
CONTEMP. LEGAL ISSUES 31, 55 (1994); Weigend, supra note 44, at 26162.
2008] IMPOSSIBILITY ATTEMPTS 543

of an attempt because he reveals himself to be someone who may be willing to

shoot again. In contrast, Actor 2 (whose mistake is one of law) is not guilty of
attempt because the only thing he reveals himself willing to do again is something
that the state regards as lawful.
This hypothesis has some force to it. However, just as threats for purposes of
impossibility are not a function of what an actor would do if his error were
corrected in time to act, 59 so, too, threats for impossibility purposes are not a
function of what an actor would do in the future. This is so for two reasons. First,
it would transform the social practice of blaming from an exclusively backward-
looking judgment of what an actor has done into a judgment that consists in part of
what he will do in the future. The scholarship of impossibility attempts is rife with
references to the conflict between subjectivists (who purportedly believe that
actors ought to be judged entirely on the basis of the criminality of what they
subjectively intend to do) 60 and objectivists (who believe that actors ought to be
judged in significant part on the criminality of what they objectively do). 61
Objectivists have several concerns, including Arnold Enkers previously-discussed
concern that, if actors are judged entirely on the basis of their intent, they will be
falsely convicted of things they did not at all intend. In part, however, objectivists
possess a more profound jurisprudential concern about the justice of blaming itself.
They are rightly concerned that if guilt becomes a function of future
dangerousness, actors will be blamed for something that has nothing to do with
blamenamely for what actors are predicted to do in the future. 62
Second, the future-threat hypothesis fails to account for actors whose mistakes
of fact are widely intuited to be inculpatory but who are not future threats.
Consider, for example, a devoted and otherwise law-abiding son who, in order to
put his aged, terminally-ill and near-comatose mother out of her misery, holds a
pillow over her face to smother her breathing, only to discover afterwards that she
had already died of natural causes. Most people would say that the son is guilty of
attempted homicide, despite the fact that, realistically, he is no future threat to
anyone. Or consider an actor who, while attempting to break into a safe that turns
out to be empty, is himself shot by the homeowner and rendered a quadriplegic.
Most people would say that he is guilty of attempted larceny, despite the fact that
he is not an future threat to anyone.

59
See text accompanying supra notes 2123.
60
Significantly, even the purest subjectivists, i.e., those who drafted Model Penal Code
5.01(1), balk at punishing everyone with a guilty mind and willingness to act on it, because in accord
with what I call the third element of criminal responsibility, they further require that actors also be
threat[s] to the interests the substantive statute at hand seeks to protect. See MODEL PENAL CODE
2.12 (1962) and text accompanying infra notes 7684.
61
See Antony Duffs masterful exposition of the divide between subjectivists and objectivists
in Duff, supra note 2, at chs. 68, 10. See also FLETCHER, supra note 21, at 13984; Lawrence
Crocker, Justice in Criminal Liability: Decriminalizing Harmless Attempts, 53 OHIO ST. L.J. 105763
(1992); Hasnas, supra note 29, at 2530.
62
See DUFF, supra note 2, at 381.
544 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

2. Ex Ante Threat

Lawrence Crocker agrees with my students that the key to impossibility

attempts is the threat 63 of harm that an actors conduct presents. In contrast to
my students, however, Crocker argues that the measure of a threat is not the future
danger that an actor presents when he is viewed ex post, but the objective risk 64
that he imposes on the rights of others when his conduct is viewed ex ante.
Crocker concedes that objective risks cannot be measured from the ex ante
viewpoint of an omniscient observer because from an omniscient viewpoint, the
probability of harm in impossibility cases is always zero. 65 He also concedes that
risks cannot be measured from the ex ante viewpoint of the actor himself because
in addition to the latters viewpoint being subjective, the probability of harm from
an actors viewpoint is always high in impossibility cases. 66 Instead, he argues
that the measure of objective risks in impossibility cases is the ex ante viewpoint
of an idealized 67 observer. Whether an actor has a defense of impossibility,
Crocker says, depends upon whether an idealized observer, viewing the actors
conduct ex ante, would adjudge there to be a risk.
Crocker is in good company in thinking that the answer lies in the ex ante
viewpoint of an idealized observer. 68 However, as Larry Alexander points out, 69
all such approaches suffer from a common failing. They require that idealized
observers be hypothetically invested with a certain quantum of knowledge
knowledge that is neither omniscience nor whatever the actor himself possesses.
Yet there is no non-arbitrary standard for affixing that quantum of knowledge.
[The idealized-observer] approach . . . is indeterminate through and through. Its
application will perforce be completely arbitrary and manipulable. 70
To illustrate, consider Crockers approach. Crocker proposes that an
idealized observer be vested with all the knowledge, free of mistakes, that an
expert in physics and engineering would possess who used her personal powers of
observation to study the relevant courses of conduct from their outset, but who
possessed no instruments to boost her powers of observation, that is, no
microscope, no instruments of chemical analysis, no transit, no x-ray, no wind
gauge. 71 With such an observer in mind, now consider an actor who, having

63
Crocker, supra note 61, at 1062 (quoting Hyman Gross).
64
Id. at 1111.
65
Id. at 1100.
66
Id. at 1099.
67
Id.
68
See, e.g., DUFF, supra note 2, at 38283 (advocating a reasonable person approach);
FLETCHER, supra note 21, at 14657 (advocating an ideal-observer approach for attempts to commit
harm-based offenses).
69
Alexander, supra note 4, at 6567.
70
Id. at 67.
71
Crocker, supra note 61, at 1100.
2008] IMPOSSIBILITY ATTEMPTS 545

waited on a frigid morning to kill a romantic rival, seizes his target, places a loaded
pistol to the targets head, and repeatedly pulls the trigger, hearing it click, only to
realize to his fury that, because of the frigid cold, the gun is jammed and will not
fire. Is the actor guilty of attempted murder? It depends, Crocker would say, on
what an expert in physics and engineering who had observed the gun from the day
of its manufacture till the day in question, but who lacked sophisticated
microscopes and x-rays, would have known about the gun. If the expert
would have known about the defect even without resorting to the microscopes and
x-rays, the actor is guilty of attempt. If the expert would not have known about the
defect without resorting to such tools, the defendant is not guilty of attempt. The
problem with Crockers test is that it not only produces counterintuitive results,
e.g., acquitting our guy with a jammed gun, but it also lacks any grounding in
norms of criminal responsibility. There is no justification for making an actors
criminal responsibility depend upon what an expert would have known when the
actor himself is less than an expert.
Sandy himself eventually embraced another idealized-observer test, namely,
Ira Robbinss reasonable-person test. 72 Robbins hypothesizes someone who is
otherwise identical to the actor but whose inferences about what he observes are
those of a reasonable 73 person, i.e., a person of average or normal
understanding. 74 An actor whose attempt fails because of a mistake is guilty of
attempt, Robbins says, only if an average person who is otherwise identical to the
actor would have made the same mistake. Thus, Robbins would say, Mildred the
Voodoo practitioner is not guilty of attempt because an average person who wished
to kill her ex-husband and who had read an article about Voodoo in the National
Inquirer would not have inferred that Voodoo would work. In contrast, an actor
who shoots to kill but misses is guilty of attempt if an average person would have
realized that the actor was mistaken in thinking that the target was in the line of
fire.
Robbinss test may appear to be more congenial than Crockers because
Robbins invokes the familiar and reassuring language of reasonableness.
Despite appearances, however, Robbinss test is just as arbitrary as Crockers. For
one thing, Robbins uses reasonableness differently than it is commonly used in
criminal law and torts. Reasonableness in criminal law and torts is not an
empirical measure of which facts actually obtain, e.g., whether Voodoo is in fact
effective. It is a normative measure of what kinds of conduct, thinking, and
emotions are normatively appropriate to such facts as obtain or are believed to

72
See Robbins, supra note 35, embraced in KADISH & SCHULHOFER, supra note 5, at 60001
& n.12. For an earlier version of the same test, see Jerome Elkind, Impossibility in Criminal
Attempts: A Theorists Headache, 54 VA. L. REV. 20, 35 (1968).
73
Robbins, supra note 35, at 441.
74
Id. at 442 n.326 (citing Minnesotas statute favorably).
546 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

obtain. 75 Furthermore, although Robbinss test resolves Voodoo cases just as

Crockers does, it leaves other cases radically indeterminate. Consider a person
who shoots but misses because he mistakenly thinks that his target is in his line of
fire or within range. How can one decide whether an average person would know
that target is in a line of fire or within range? Finally, like Crocker, Robbins does
not ground his test in principles of criminal responsibility. There is no justification
for making an actors criminal responsibility depend upon whether an average or
normal person would have made the same mistake. 76

B. A New Path

I believe that another path exists, a path that is consistent with the objectivist
concern that actors be blamed for what they have done, and not for what they are
predicted to do. To explore it, let us return to our prior hypotheticals regarding the
son who mistakenly thinks he is putting his beloved mother out of her misery, and
the safecracker who becomes a paraplegic in the course of theft. What explains the
widely shared intuition that, despite the fact that neither is a future danger, both are
guilty of attempt, while a person who commits adultery in the mistaken belief that
adultery is a crime is not guilty?
The answer, I believe, is that although none of these actors is presently a
threat, the son and safecracker revealed that they were threats while Mr. Adulterer
revealed that he was not a threat. In the language of MPC section 2.12, Mr.
Adulterer did not actually cause or threaten the harm or evil sought to be
prevented by the law defining the offense, 77 while the son and safecracker did.
The resulting testthat a person is guilty of an attempt to commit offense X if and
only if he reveals himself to have been a threat to the interests that offense X seeks
to protectnot only explains the cases that the law/fact test succeeds in
explaining, it also explains the cases that the law/fact test fails to explain. Thus, it
explains the widely shared intuitions that Mildred in Midwestern Voodoo is not
guilty of attempt, and it explains why Mr. Law is just as blameworthy as Mr.
Fact.
Nevertheless, the proposed test raises at least two concerns. First, the test
raises a concern about the meaning of threat. The concern can be framed as
follows:

75
See Peter Westen, Individualizing the Reasonable Person in Criminal Law, 2 CRIM. L. &
PHIL. (forthcoming 2008).
76
See Simons, supra note 24, at 485.
77
MODEL PENAL CODE 2.12 (1962) (The Court shall dismiss a prosecution if, having
regard to the nature of the conduct charged to constitute an offense and the nature and the attendant
circumstances, it finds that the defendants conduct . . . (2) did not actually cause or threaten the harm
or evil sought to be prevented by the law defining the offense . . . .).
2008] IMPOSSIBILITY ATTEMPTS 547

A Question Regarding Threat

Threats are a subset of risks, and risks, in turn, are probabilities of
specified events, including specified harms and evils. As such, risks are
epistemic in nature, that is, they are based upon the probabilities from
zero to one (0 to 1) that arise from facts as known to specified
observers. 78 It follows, therefore, that it is coherent to speak of future
threats and risks, provided that one is referring to the incomplete
knowledge that specified observers possess ex ante. And it is coherent to
speak of past threats and risks regarding harms and evils that, we know
ex post, actually occurred and, hence, we know ex post possessed
probabilities of 1. And it is coherent to speak of past threats and risks
regarding harms that, we know ex post, did not occur, provided that we
are referring to the incomplete knowledge that specified observers
possessed ex ante. But it is incoherent to speak of actual past risks and
threats regarding harms that we know ex post did not occur if we mean to
be referring to the knowledge that we possess ex post. For once we
know that the harm did not occurwhich is, indeed, what we know in all
impossibility caseswe realize that, rather than their ever being an
actual threat or risk of the harm, its probability was 0.

This concern has force because it derives from a commonplace understanding

of threats and risks as epistemic concepts. However, it is a fallacy to assume this
epistemic understanding of threats and risks exhausts their meaning. We often use
threats and risks differently. We use them to refernot to probabilities of
future harm based upon limited knowledge of the totality of factors that render the
harm certain to occur or certain not to occurbut to a likelihood of harm under
counterfactual events that we can imagine occurring. That is, we use threats and
risks of harm to refer in retrospect to how easily counterfactual events could
have obtained that, had they obtained, would have produced harm that, thankfully,
did not occur. 79
To illustrate, imagine that a CEO is told that an off-duty custodian happened
to notice a disgruntled ex-employee behaving strangely; that the custodian took the
initiative to notify security officers; and that the latter arrested the intruder just as
he was about to enter the CEOs office with a bomb and a suicide note. Upon
receiving the report, the CEO says, Wow, that is scary. What a close call! The

78
See Larry Alexander, Insufficient Concern: A Unified Conception of Criminal Culpability,
88 CAL. L. REV. 931, 936 (2000).
79
Lawrence Crocker and John Strahorn argue that an actor cannot be a threat to a person who,
unbeknownst to him, is already dead, because, they say, to be a threat is to be a risk to someone who
still possesses interests. See Crocker, supra note 61, at 1097, 110304; John Strahorn, The Effect of
Impossibility on Criminal Attempts, 78 U. PA. L. REV. 962, 97079 (1930). I use threat to refer to
harms that ex-post observers believe would have befallen persons whom the state seeks to protect
under counterfactual events that observers believe could have obtained, regardless of whether they
are the persons whom the actor at hand intended to harm.
548 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

guy was a real threat. Obviously, by scary the CEO is not referring to his fears
about the future because he knows that the intruder is no longer a threat. Nor is he
referring to past fears, given that he was unaware of the threat at the time. Instead,
he is referring to a fear he presently experiences as he retrospectively contemplates
the real threat he feels he narrowly dodgednamely, the terrible harm he would
have suffered if even a single circumstance had been slightly different in ways he
fears could easily have occurred.
This brings us to the second concern about the proposed test. The second
concern goes to its normative justification. What is the normative justification for
confining liability for impossibility attempts to persons who were threats to the
interests the law at issue seeks to protect? The justification, I believe, is that it is
an appropriate limitation on state power, over and above the requirement of a
guilty mind and manifest willingness to act on it.
State power is not the same as divine power. 80 Being omniscient about what
is in peoples hearts, God presumably does not have to wait until people act on
their guilty minds in order to know whether and how much punishment they
deserve. Nor, being perfect in His righteousness, is God constrained in
administering punishment by anything beyond what people deserve. In contrast,
state officials who administer earthly punishment are neither omniscient about
what is in peoples hearts nor free of wrongdoing themselves. Because the state
can never know exactly what is in a persons heart, the state conditions an actors
criminal liability upon proof of not only a guilty mind, but also a manifest
willingness to act on it. And because the state can neither know exactly how much
punishment an actor deserves nor ever be completely righteous, the state ought to
require something further before imposing the burden and disgrace of purposeful
condemnation and suffering. It ought to require as a condition of criminal
responsibility that an actors conduct actually affect the people of the state by
either infringing or threatening to infringe interests that they seek to protect by
means of the statutes at hand. It ought to require that an actors conduct matter to
its citizens by unnerving 81 them. Otherwise it is playing God. 82
This third element of criminal responsibility (over and above the requirements
of a guilty mind and willingness to act on it 83 ) is a stealth requirement that, being

80
For the argument that divine punishment differs from mundane punishment, see Hasnas,
supra note 29, at 5152.
81
FLETCHER, supra note 21, at 472.
82
DUFF, supra note 2, at 344. Even if we could make sense of this model of a cosmic ledger
(and I am not sure that we can), is it an appropriate model for the human activity of blaming
wrongdoers? Such a model is appropriate (if it is appropriate at all) to a detached, god-like observer
of the human scene. We, however, are not such detached observers: we are participants in the human
scene. We are agents, patients, and interested observers . . . acting, thinking and responding to each
other within a human social life.
83
See GROSS, supra note 38, at 223 (arguing that impossibility cases are a window into a
third dimension of culpability beyond requirements of a bad act and guilty mind). John Hasnas
also argues that impossibility cases reveal a third requirement of criminal responsibility over and
2008] IMPOSSIBILITY ATTEMPTS 549

nearly always satisfied, tends to pass undetected. The requirement is always

satisfied in the event of completed crimes, given that actors who commit
completed crimes inflict the very harms or evils the state seeks to prevent. And it
is nearly always satisfied in cases of inchoate crimes, 84 given that actors who
commit inchoate crimes nearly always reveal themselves to have been threats to
interests the state seeks to prevent. However, the requirement reveals itself in rare
impossibility cases, e.g., Midwestern Voodoo and Mr. Adultery, in which actors
neither inflict harms or evils nor threaten interests that the state seeks to prevent.
Mildred the voodooist did not threaten the states interest in life because, given the
publics view of voodoo and its knowledge of Mildreds motivations in resorting
to voodoo, no one is likely to believe she would have killed her ex-husband under
any counterfactual circumstances they fear could have obtained. Nor does the
adulterer threaten any interest the state seeks to protect by means of punishment.
Given that adultery is not a crime, and given that the public has no intention of
making it a crime, no one is likely to believe the adulterer would have committed a
crime under any statute it fears the state might have enacted.
I have focused thus far on a test that I believe explains and justifies widely
shared intuitions about who should, and should not, be punished for impossibility
attempts. However, if I am right, it has implications for the law of attempt
generally. For it means that when we punish criminal attempts, we are punishing
persons not only for what they did, i.e., manifest a willingness to act on their guilty
minds, but for something they clearly did not do. We are punishing them for what
we believe they would have done under counterfactual circumstances that we know
did not exist but that we fear could have occurred.

V. THE DISTINCTIVE CHALLENGE OF PUNISHING

ACTORS FOR MISTAKES OF LAW

We have seen that the law/fact test, like MPC section 5.01(1), is deficient for
two reasons: (1) it fails to explain why some mistakes of fact are widely regarded
as being exculpatory, e.g., Midwestern Voodoo; and (2) it fails to explain why

above proof of a guilty mind and willingness to act on it. However, rather than claiming that the
requirement consists of conduct by an actor that actually infringes or threatens to infringe interests
that the state seeks to protect, Hasnas argues that it consists of objective behavior to prevent corrupt
and overzealous state officials from falsely convicting innocent actors of possessing guilty minds
they never possessed at all. See Hasnas, supra note 29, at 4554.
84
By inchoate crimes, I mean crimes of the form, It is an offense to X, with an intent to
commit offense Y, including impossibility attempts. For a different usage, see Larry Alexander &
Kimberly D. Kessler, Mens Rea and Inchoate Crimes, 87 J. CRIM. L. & CRIMINOLOGY 1139, 1139
(1997) (confining inchoate crimes to ones in which there is still time [for an actor] to desist and
renounce, thus excluding impossibility attempts).
550 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

some mistakes of law are widely regarded as being blameworthy, 85 e.g., the
mistake by Mr. Law, even if they are also commonly thought to be unpunishable. 86
My proposal addresses the first problem by invoking language from another
section of the Model Penal Code, i.e., MPC section 2.12, 87 though my proposal
goes further than MPC section 2.12 by explaining what it means to have
threatened interests that a statute seeks to protect, and why a threat-based test is
normatively justified. The tests effect is to reduce the class of factual mistakes
that are inculpatory by making it an element of attempt that actors manifest
themselves to be threats to interests that statutes at hand seek to prevent.
I shall now examine whether and how my proposal addresses the problem
presented by inculpatory mistakes of law. I believe that my proposal explains
whyand howan actor can be rightly punished for an impossibility attempt
based upon a mistake of law. Indeed, to show in detail how my proposal deals
with mistakes of law and fact, I frame it as a statute that draws in part on what the
Model Penal Code separately states in MPC sections 5.01(1) and 2.12 (see the
Appendix to this article).
I must emphasize, however, that I do not advocate that my statutory proposal
regarding mistakes of law be formally adopted. To be sure, the universal rule,
which MPC section 5.01(1) codifies, that mistakes of law are exculpatory,
produces some injustice because it exculpates actors like Mr. Law who are as
blameworthy as Mr. Fact. However, such injustices are too rare and attenuated to
justify the difficulties and controversies that a statute would generate that punishes
actors for mistakes of law. 88 Accordingly, I tender my hypothetical statute solely
to illustrate the proposal in concrete terms.
My proposal regarding mistakes of law will be controversial because the
paradox it addresses seems irreconcilable. The paradox consists of the tension
between Sandy Kadishs implicit position in Comment in the Hypothetical Law
Review, on the one hand, and Fernand Dutile and Harold Moores response, on
the other. Sandys insight regarding Mr. Fact and Mr. Law is that the two
actors are equally blameworthy. In contrast, Dutile and Moore argue that because
Mr. Facts mistake presupposes the law as it is and Mr. Laws mistake presupposes
the law as it is not, no common legal norm exists by which they can be equally
punishable.
I believe a middle ground exists between these two positions. However, I
confess that I am less confident about the validity of my proposal than about the
force of the opposing positions it seeks to reconcile. Accordingly, to acknowledge

85
See supra note 37. Cf. Fletcher, supra note 13, at 59 (The simple intuition that no one can
attempt a crime that does not exist will not suffice.).
86
PAUL ROBINSON, CRIMINAL LAW 687 (1997); A.P. SIMESTER & G.R. SULLIVAN, CRIMINAL
LAW: THEORY AND DOCTRINE 315 (2d ed. 2003); ASHWORTH, supra note 26, at 406.
87
MODEL PENAL CODE 2.12 (1962) (see supra note 77).
88
See Cole, supra note 58, at 55 ([Given] the practical problems in punishing inculpatory
legal mistakes, . . . the defense of pure legal impossibility seems like a good idea.).
2008] IMPOSSIBILITY ATTEMPTS 551

the difficulty of the problem, I shall explore it in the form of an imaginary dialogue
between Sandy and Dutile/Moore. I take the liberty of attributing to Sandy the
logic that I believe is implicit in his hypothetical regarding Messrs. Fact and Law,
despite the fact that Sandy does not himself pursue it in his Comment. I attribute
to Dutile and Moore the strongest arguments I can marshal in defense of their
position.

An Imaginary Dialogue

KADISH. I drafted the hypothetical regarding Messrs. Fact and Law

because I believed it illustrates something that, though rare, can
nevertheless occurnamely, instances of impossibility in which an
actor who makes a mistake of law and a comparable actor who
makes a mistake of fact are equally blameworthy.

DUTILE/MOORE. We agree with you, Sandy, that Messrs. Fact and

Law are equally blameworthy. Indeed, we have gone on record in
saying so. 89 But that is not the question. The question is, not
whether they are equally blameworthy, but whether any criminal
prohibition exists that they equally violated.

KADISH. With due respect, Im puzzled that you would distinguish the
two questions. The question whether actors are equal in
blameworthiness is not distinct from whether a prohibition exists
that they both violated. To the contrary, to say that Messrs. Fact
and Law are equal in their blameworthiness means that a norm
must exista rule in the form of a norm, It is wrong to do Xby
which each actor is measured and found to be identical to the other.
That is what it is to be equally blameworthy. 90

DUTILE/MOORE. Nice point. Youre right that Messrs. Fact and

Laws equal blameworthiness presupposes a common prohibitory
norm that they both violated. However, it need not be a legal norm.
It can be a moral norm. Indeed, we have said all along that Messrs.
Fact and Law are equally culpable from a moral rather than a
legal standpoint. 91 The fact that a moral norm exists that Messrs.
Fact and Law both violated does not mean that a comparable legal
normor comparable criminal offenseexists.

89
Dutile & Moore, supra note 35, at 196 (From a moral standpoint, they are equally
culpable.).
90
See generally PETER WESTEN, SPEAKING OF EQUALITY: AN ANALYSIS OF THE RHETORICAL
FORCE OF EQUALITY IN MORAL AND LEGAL DISCOURSE 5992 (2001).
91
Dutile & Moore, supra note 35, at 196.
552 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

KADISH. I agree that moral blameworthiness is not necessarily

coextensive with legal blameworthiness. Considerations can obtain
that make it appropriate for society to embrace a moral norm
without embracing a comparable criminal norm, and vice versa.
However, the implicit norm that you now concede Messrs. Fact and
Law must jointly have violated in order to become equally
culpable cannot possibly be a moral norm apart from law. For
there is no such thing as a hunting season in morals. The hunting
season, like the criminal prohibition that gives it effect, is entirely a
function of the laws effort to regulate collective action. The
implicit norm that Messrs. Fact and Law jointly violated is
quintessentially a legal norm, not a moral one.

DUTILE/MOORE. We must disagree. There is, indeed, a moral norm

that Messrs. Fact and Law did both attempt to violatenamely, the
moral duty of all persons to obey the law. The duty to obey the
law is a moral duty rather than a legal duty because the law confines
itself to punishing persons for disobeying particular laws, rather
than punishing them for disobeying law in general or for attempting
to do so. Measured by the moral norm to obey the law, Messrs. Fact
and Law are equally blameworthyalbeit morally rather than
legallyof attempting to disobey the law.

KADISH. Actually, I agree that all persons have a duty to obey the law
in general. I agree that the duty is moral, rather than legal, in
nature, 92 and that Messrs. Fact and Law both violated that moral
dutythereby rendering themselves equally blameworthy in that
respect. However, that cannot be the respect in which Mr. Fact and
Mr. Law are widely regarded as being equally blameworthy. For
that is a sort of blameworthiness that all attempters share in
common, regardless of whether their attempts are based on mistakes
of fact or law. Measured by that norm, a person who engages in
sexual intercourse with a married woman in the mistaken belief that
adultery is a crime would be widely regarded as equally
blameworthy as a person who shoots to kill in the mistaken belief
that his gun is loaded, for both persons attempt to disobey the law.
No, something distinctive must exist regarding Messrs. Fact and
Law that renders them equally culpable, despite the fact that one of
them made a mistake of fact and the other a mistake of law.

92
For disagreement about whether legislatures should make it an offense in law to attempt to
break the law, compare DUFF, supra note 2, at 15659 (no), with Fletcher, supra note 13, at 59
(yes).
2008] IMPOSSIBILITY ATTEMPTS 553

Whatever it is, it cannot consist of their both having violated a

moral norm to obey the law, for the blameworthiness of one of
them, at least i.e., Mr. Fact, consists predominantly of his having
also violated a legal norm.

DUTILE/MOORE. Perhaps you are right that the implicit prohibition

that renders Messrs. Fact and Law equally blameworthy is not a
moral prohibition. But that only shows how profound the
problem is. For the fact remains that there is no single lawno one
criminal prohibitionthat Mr. Fact and Mr. Law could have
attempted to violate. After all, by virtue of hunting on October 15,
in the mistaken belief that he was doing so on October 14, Mr. Fact
intended to violate the law as it stood, i.e., a law that made it an
offense to hunt before October 15. In contrast, by virtue of
intentionally hunting on October 15, Mr. Law intended to violate a
law that existed only in his imagination: he intended to violate a
non-existent law that made it an offense to hunt before October 16.
Therefore, despite your plausible claim that by virtue of Mr. Fact
and Mr. Law being equally culpable, they must have violated the
same law, we know that they did not attempt to violate the same
law.

KADISH. With due respect, again, I think youre overlooking the single
criminal offense that Messrs. Fact and Law both violated. The
criminal offense they both violated is not the one that makes it a
crime to hunt before October 15. They both fully complied with
that law. Rather, the criminal offense they both violated is the one
that makes it a crime to attempt to commit an offenseand,
specifically, attempting to hunt out of season. 93

DUTILE/MOORE. Now we fear that it is you, Sandy, who are

overlooking something. Hunting out of season may be the short
title to an offense in a states penal code. But it cannot possibly be
the offense itself because it is fatally incomplete. It fails to specify
what every actor needs to know in order to comply with the law,
i.e., the precise calendar dates before and after which it is illegal to
hunt. To be complete, the offense must either specify those
calendar dates explicitly or incorporate them by reference from
elsewhere in the law. And it is entirely a formal matter, not a
substantive one, whether the offense specifies the dates or
incorporates the dates by reference. 94

93
Sandy makes this very argument. See KADISH & SCHULHOFER, supra note 5, at 599600.
94
See Alexander, supra note 4, at 6869.
554 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

KADISH. A nice point, and I must agree with you.

DUTILE/MOORE. Thank you. But notice what now follows from it.
Once you specify the offense that each actor attempted to commit,
you are forced to allege that the overall offenses of attempt that
Messrs. Fact and Law allegedly committed are different. Mr. Fact
must be alleged to have attempted the genuine offense of Hunting
before October 15, while Mr. Law must be alleged to have
attempted the imaginary offense of Hunting before October 16.
As a result, no single offense of attempt exists that Messrs. Law and
Fact can both be alleged to have committed.

KADISH. I agree with you that, if we are obliged to charge Mr. Law
with a non-existent offense, then no single offense of attempt exists
that Messrs. Law and Fact have both committed. I also agree that
charging Mr. Law with attempting to commit a non-existent offense
would be as unjust as charging Mr. Adulterer with the non-existent
offense of adultery. Howeverand I must emphasize thisI do not
believe that we are obliged to charge Mr. Law with attempting to
commit a non-existent offense. Indeed, the contrasting case of Mr.
Adulterer shows why. The reason that Mr. Adulterers mistake of
law is exculpatory is that he reveals himself to have been no threat
to interests that the state seeks to protect by means of a statute at
hand. In contrast, the reason that Mr. Laws mistake is inculpatory
is that he reveals himself to have been a substantial threat to the
interests that the state seeks to protect by means of the substantive
statute at hand, i.e., the statute that makes it an offense to hunt
before October 15. Mr. Law reveals that he would have committed
an offense under counterfactual circumstances that the society at
large fears could easily have obtained, namely, counterfactual
circumstances in which he either discovers his mistake on October
13 or the state elects to start the season on October 16.

DUTILE/MOORE. Lets be sure we understand you. Are you saying

that Mr. Law could be charged with attempting to violate the
hunting law that now existsthat is, attempting to hunt before
October 15when hunting on October 15 was precisely what he
intentionally did?

KADISH. Exactly.

DUTILE/MOORE. Now we are the ones who are puzzled. Indeed, we

hardly know where to begin. But for starters, how can you say that
2008] IMPOSSIBILITY ATTEMPTS 555

Mr. Law is guilty of attempting to do something that he clearly did

not want to do, i.e., to hunt before October 15?

KADISH. We must guard against allowing the lay meaning of words,

including attempt, to govern their technical meanings as terms of
art. For lay persons, to attempt X is to try to bring X about. But
that is not its technical meaning in most jurisdictions in criminal
law. Consider an actor who, to recover insurance, places a bomb on
an airplane, believing that it will also kill the pilot but not wishing
toonly to learn that an alert attendant defused the bomb. The
Model Penal Code and many jurisdictions would regard the actor as
guilty of attempted murder, despite the fact that he did not want to
kill the pilot.

DUTILE/MOORE. Okay, we agree that the problem cannot be solved

by consulting dictionaries. But surely you agree that a person is not
guilty of an attempt to commit offense X unless he had a guilty
mind and willingness to act on it.

KADISH. Yes, I agree.

DUTILE/MOORE. And surely you will also agree that possessing a

guilty mind for purposes of an attempt to commit offense X
means, at least, possessing mens rea regarding the material elements
of offense X. In the words of the Model Penal Code, it means
possessing the mens rea that is otherwise required for commission
of the crime. 95

KADISH. I agree that the Model Penal Codeand every jurisdiction of

which Im awarerequires that an actor charged with an attempt to
commit offense X possess mens rea regarding the material elements
of offense X. But I do not agree that it is what justice requires.
Indeed, as long as Messrs. Fact and Law are understood to be
equally culpable, it cannot be what justice requires. Remember,
we have seen that to say they are equally culpable presupposes a
legal prohibition that they both violated. Yet the prohibition cannot
be one that requires mens rea regarding the material elements of
hunting before October 15, because while Mr. Fact intended to
hunt before October 15, Mr. Law did not.

DUTILE/MOORE. Okay, but now the ball is squarely in your court.

You admit that an actor is not guilty of attempt to commit offense X

95
MODEL PENAL CODE 5.01(1) (1962).
556 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

unless he possesses a guilty mind. Yet you deny that the guilty
mind consists of mens rea regarding the material elements of
offense X. What could an actor charged with attempt possibly have
a guilty mind about, if not about the material elements of offense X?

KADISH. Good question. The answer is this: he could be charged with

having a guilty mind about the material elements of the crime he
mistakenly thinks he is committing. 96

DUTILE/MOORE. Uh-oh, we were afraid that you were going to say

thatafraid for your sake, that isbecause it puts you in the
unhappy position of flouting the principle of legality. That is, it
puts you in the unhappy position of presupposing the existence of a
criminal offense that the legislature has not enacted.

KADISH. I am glad you invoke the principle of legality because I, too,

used to think that it constituted a barrier to what I propose. 97
However, you must remember that my proposal grounds Mr. Laws
liability not on the actual existence of a non-existent offense, but on
Mr. Laws subjective intentionsin this case, his intentions
regarding elements that would be an offense if it existed. I propose
to base an actors attempt liability on two things: (1) on what the
actor believes the material elements to be regarding the offense he
believes he is committing, and (2) on his possessing the
intentionality regarding them that criminal attempt ordinarily
requires regarding such material elements. Thus, Mr. Laws
liability would be based on his intentionally engaging in what he
believes to be the material elements regarding what he believes to
be the offense he is committing, i.e., hunting before October 16.

DUTILE/MOORE. Perhaps you are right. Perhaps your proposal does

not presuppose the actual existence of a non-existence offense and,
hence, does not violate legality per se. But there is a problem
nonetheless. By decoupling the material elements of the offense for
which an actor possesses mens rea from the material elements of the
crime that he is charged with attempting, you remove all restraints
on which crimes an actor can be charged with attempting. Consider
Mr. Law. Once Mr. Law possesses mens rea regarding the material
elements of an imaginary offense, he could be charged with
attempting to commit anything, say, murder. Or consider Mr.
Adulterer. Mr. Adulterer has mens rea regarding the material

96
See Alexander, supra note 4, at 67 (considering but rejecting this possibility).
97
See KADISH & SCHULHOFER, supra note 5, at 59798.
2008] IMPOSSIBILITY ATTEMPTS 557

elements of an imagined offense. Does that mean that he could be

charged with attempting to hunt before October 15? That is
absurd. 98

KADISH. Your argument would be correct if the sole requirements for

attempt liability were (1) a guilty mind, and (2) a manifest
willingness to act on it. However, those are not the sole
requirementsor, at least, not under my proposal. An actor is not
guilty of an attempt under my proposal unless, in addition to
possessing 1 and 2, he was a threat to interests that the statutory
offense at hand seeks to prevent. This third element of criminal
responsibility greatly restricts the range of offenses that actors who
make mistakes of law can be charged with attempting. Thus, it
would prevent prosecutors from trying Mr. Law with attempted
murder, because no evidence exists that Mr. Law threatened
interests that murder statutes seek to protect. And it would also
prevent prosecutors from trying Mr. Adulterer for hunting out of
season. But it would not prevent prosecutors from trying Mr. Law
for attempting to hunt before October 15 because Mr. Law was a
threat to the interests that the latter offense seeks to protect.

DUTILE/MOORE. Were not surprised to see you invoke the language

of threat, because MPC section 2.12 invokes it, too. However,
you must admit that you and MPC section 2.12 are using the term
threat in a strange way. You are using it to refernot to risks of
future harm, or past risks as measured ex ante by observers with
limited information, or past risks as measured ex post by observers
with unlimited information regarding harms that have since
materialized. You are using it to refer to the strange creature of past
risks as measured ex post by observers with unlimited information
regarding harms that have not materialized. Given its strangeness,
dont you think that threat is too vague a term on which to base an
actors liability?

KADISH. To be honest, I find myself vacillating. Sometimes I find

myself agreeing with the framers of MPC section 2.12 that threat
is sufficiently clear. At other times, I find it impermissibly vague.
In any event, however, we can easily supplement the term threat
with a provision that specifies precisely what we mean in this
context. We can specify that an actor who possesses the mens rea I
have described and who manifests a willingness to act on it is guilty
of an attempt to commit an offense if based upon his conduct and

98
See Cole, supra note 58, at 5255 (making this reductio ad absurdum).
558 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

state of mind, the triers of fact, representing informed citizens of the

jurisdiction, conclude that he would have committed the offense
under counterfactual circumstances they fear could have obtained.

DUTILE/MOORE. Phew, now we have the opposite problem. Now

youre asking that we base something as real as criminal liability
upon something as hypothetical as counterfactual assumptions.
That is asking a lot.

KADISH. It may seem that Im asking a lot, but actually Im not. First,
were all accustomed to making counterfactual determinations in
criminal law because we do so every time we determine but-for
causation. Second, my claim regarding counterfactuals is not
something new that is being superimposed on existing norms in the
area of attempt. It merely spells out what the legal term threat
already means in provisions as commonplace as MPC section
2.12(2).

DUTILE/MOORE. Perhaps so, but as you know, counterfactual tests are

said to be notoriously indeterminate. In order to apply them, one
must possess criteria to decide precisely which historical facts are
hypothetically altered and which are not. Yet such criteria are often
lacking. 99

KADISH. You are right that counterfactual tests can be problematic,

depending upon how precisely counterfactuals must be stated.
However, notice how my proposal differs from most counterfactual
tests. Mine does not require consensus on what historical facts are
and are not changed. Mine is a psychological test that incorporates
by reference whatever counterfactual determinations people happen
to fear could occur.

DUTILE/MOORE. Still, we are not yet convinced that your proposal is

sufficiently protective of defendants. How can you be sure that an
actor who does one thing would have done something elseand a
crime, at thatunder circumstances that you now know never
existed?

KADISH. Of course, I cannot be sure. No one can. But that is a

problem that will forever plague us as long as attempt is a crime,
regardless whether it is an impossibility attempt or an interruption
attempt. To convict a person of an attempt is to convict him of

99
See Alexander, supra note 4, at 6364.
2008] IMPOSSIBILITY ATTEMPTS 559

something we believe he would have done under circumstances that

we now know never existed.

DUTILE/MOORE. Maybe so. But, nevertheless, attempts based upon

mistakes of law possess a distinctive feature that increase the risk of
false conviction if jurors are allowed to speculate regarding what an
actor would have done. After all, an actor who makes a mistake
of law is consciously doing something that the state regards as
innocuous, while the conduct he is charged with attempting may be
quite heinous. An actor who is willing to violate what he
mistakenly thinks is the law may not be willing to do something
heinous. How can we trust jurors to know what actors in this area
would have done?

KADISH. You raise a valid point and one that must be addressed.
However, I think it can be addressed by adding a proviso to the
effect that an actor is not liable for attempt based upon a mistake of
law unless his principal goal would have been served as well by
means that would have sufficed to commit offense X. A proviso of
that kind would protect a person who makes the mistake of law of
thinking that importing a food staple like sugar is a crime but would
hardly think of doing anything as heinous as importing products that
are actually contraband. Yet it would not protectnor should it
an actor like Mr. Law, whose principal goal would be served as well
by hunting on October 14.

DUTILE/MOORE. Well, that is admittedly an improvement. However,

it does not address all our concerns. We are also concerned about
actors who mistakenly engage in conduct that is less onerous or less
likely to be detected by the police than what they would actually
have to do to commit a crime. If jurors possess open-ended
authority to decide what an actor would have done, they may
convict actors who are willing to do safe and easy things, but are not
willing to do the onerous and detectable things that are actually
needed to commit a crime.

KADISH. You may be right that Im too optimistic about the jurys
ability to make the counterfactual judgments that inquiries into
threats require. However, if so, you yourselves have put your
fingers on a missing element that, if added to my proposal, may
solve the problem. We can amend the proposal to specify that an
actor is guilty of an impossibility attempt only if the means he
mistakenly believed would culminate in a crime are neither
substantially harder for the police to detect nor substantially easier
560 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

or safer for persons to execute than means that suffice to commit

offense X. Does that satisfy you?

DUTILE/MOORE. Maybe. We will have to think about it. But we will

say this, Sandyyou have gotten us thinking about something we
assumed to be unthinkable.

VI. CONCLUSION

My interest in impossibility attempts began with Sandys casebook essays,

The Case of Lady Eldons French Lace, and Hypothetical Law Review
Comment. Sandys essays present students of impossibility with three
challenges: (1) to explain the widespread intuition that certain impossibility
attempts are culpable and certain others are not; (2) to explain the widespread
intuition that impossibility attempts based on mistakes of law, e.g., Mr. Law, are
sometimes just as culpable as their counterpart attempts based upon mistakes of
fact, e.g, Mr. Fact; and (3) to account for cases where people have conflicting
intuitions about whether impossibility attempts are culpable.
I believe that the three puzzles have a common solution. The solution lies in
what I have argued is a third, stealth requirement of criminal responsibility over
and above the two requirements of a bad act and guilty mind. The third
requirement is that an actor not be punished unless citizens of the jurisdiction that
enacted the criminal statute at issue regard the actors conduct as a threat to
interests that the statute seeks to prevent. Whether conduct is a threat is a matter
of citizen psychology. An attempt to commit offense X is a threat if, and only if,
citizens of the jurisdiction that enacted the statute making X a crime perceive it as
such. In turn, citizens regard conduct as such a threat when they are convinced
that an actor would have committed offense X under counterfactual circumstances
that they fear could have obtainedor, more accurately, when they fear its
commission sufficiently to feel he should be punished for his guilty mind and
willingness to act on it. 100
This psychology of threats explains widespread intuitions that certain
impossibility attempts are culpable and others are not. Actors who intentionally
shoot to kill but miss are widely regarded as culpable because, in addition to their
possessing guilty minds and a willingness to act on it, their conduct leaves people
widely believing that they would have killed under counterfactual circumstances
that people fear could easily have obtained, i.e., the bullet trajectories being
slightly different. In contrast, superstitious voodooists in the United States who

100
This test captures what, I believe, George Fletcher means in arguing that an impossibility
attempt is not culpable unless the actors conduct manifests aptness, Fletcher, supra note 21, at
150, what R.J. Spjut means in arguing that an actor is not culpable unless he had a real prospect of
success, R.J. Spjut, When is an Attempt to Commit an Impossible Crime a Criminal Act, 29 ARIZ. L.
REV. 247, 255, 278 (1987), and what Antony Duff means in arguing that an attempt is not culpable
unless it is a serious attempt. DUFF, supra note 2, at 383.
2008] IMPOSSIBILITY ATTEMPTS 561

stick pins in dolls with intent to kill are widely regarded within the United States as
nonculpable because, despite possessing guilty minds and willingness to act on it,
their conduct leaves citizens of the United States widely not convinced that they
would have killed, except under counterfactual circumstances that citizens of the
United States do not fear could have obtained, i.e., voodoo pins being lethal.
Similarly, actors within States where adultery is not a crime who engage in
adultery in the mistaken belief that it is a crime are widely regarded as nonculpable
because their conduct leaves citizens in such States widely unconvinced that the
adulterers would have committed a crime, except under counterfactual
circumstances the citizens do not fear could have obtained, i.e., that they would
declare adultery to be a crime.
The psychology of threats also explains the widespread intuition that actors
who make mistakes of law are sometimes just as culpable as their counterparts
who make mistakes of fact. Consider Sandys Mr. Fact and Mr. Law who
both endeavored to go deer hunting on the day before the hunting season began
but, being mistaken about dates, unwittingly ended up hunting during hunting
season. Mr. Fact knew the hunting season legally began on October 15 but
mistakenly thought he was hunting on October 14. Mr. Law knew he was actually
hunting on October 15 but mistakenly thought the season legally began on October
16. The reason that observers regard the hunters as equally culpable is that in
addition to being equally willing to hunt out of season, the hunters both revealed
that they would have committed the crime of hunting before October 15 under
counterfactual circumstances that observers regard as equally capable of having
obtained. Thus, just as Mr. Fact could easily have learned before October 15 that
he was making a mistake of fact about the dates of the week, Mr. Law could
equally well have learned before October 15 that he was making a mistake of law
about the date of the hunting season. And if they had learned of their mistakes
before October 15, they would presumably both have hunted before October 15.
Finally, and most importantly, the psychology of threats explains why people
often possess conflicting intuitions about whether impossibility attempts are
culpable. Thus, just as people differ from age to age and from culture to culture in
their assessments of retrospective threats, people within common cultures can
differ, too, in the way they assess retrospective risks. 101 They differ because they
differ about how much they fear that counterfactual circumstances could have
obtained, and about how fearful they must be in order to adjudge actors culpable of
criminal attempt.
Consider, for example, the following hypotheticals that I discuss with my
students: (1) actor A intentionally shoots to kill a target but misses; (2) actor B
intends to shoot to kill, but his gun is defective; (3) actor C intends to shoot to kill,

101
See GROSS, supra note 38, at 206 (Whether conduct will be thought to threaten harm
depends on general expectations [among a jurisdictions residents] about harmful outcomes under
particular circumstances. These expectations may vary, and so then must judgments regarding
attempt liability.).
562 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

but the police have replaced the target with a dummy and, in any event, the target
has already died from unrelated causes. If my students are any measure, people in
our culture differ in their judgments of culpability. All my students regard actor A
as culpable; fewer regard actor B as culpable; and many fewer regard actor C as
culpable. The difference, I suspect, lies in how fearful the attempts appear to them
in retrospectthat is, how close students believe the actors came to succeeding,
and how close they believe actors must come to be punishable. Actor A came the
closest because the outcome depended at the last moment solely upon his agency
and upon his exercising it a few millimeters in one direction rather than another.
Actor C was the farthest from success because police, who were in control all
along, knew well before the last moment that he would not succeed. Actor B is in
between because, while the outcome of his conduct was not solely a function of his
own agency (as it was for actor A), nevertheless, unlike actor Cs conduct, actor
Bs conduct was such that no one knewor was likely to be capable of knowing
whether he would succeed until after he pulled the trigger.
These differences in psychology seem to vary based upon the gravity of
offenses and the mental states of actors. Consider the following sexual offenses
cases that differ mostly in heinousness: (1) actor A, a twenty-one-year-old disk
jockey at a tenth-grade prom, has sexual intercourse with a girl whom he is told is
underage but who actually came of age six months beforeresulting in As being
charged with attempted statutory rape; and (2) actor B is a state prisoner whose
intended victim is his prison cellmate, V, who, having complained to prison
authorities that B was forcing him to engage in sodomy, cooperates in a sting
operation with guards who maintain the cell under protective surveillance as he
submits to Bs sexual demands under the pretense of fearing for his life, resulting
in Bs being charged with attempted forcible sodomy. 102 Again, if my students are
a measure, people are more likely to find B culpable than A, apparently because
attempted forcible sodomy is the graver offense. Yet, significantly, students with
misgivings about convicting A of attempted statutory rape tend to diminish in
number if they are told that, rather than merely believing that the grade girl was
underage, As very purpose was that the girl be underagesay, in order to fulfill a
wager with a fellow disk jockey that A could seduce an underage girl before the
evening ended.
This psychological fact (that people can differ with one another in assessing
retrospective risks for purposes of punishment) is relevant to which institutionas
among legislatures, judges, and juriesought to pass judgment on retrospective
risks. Some authorities argue that legislatures and judges ought to make final
determinations as matters of law regarding which risks are and are not culpable.
Thus, they advocate tests that base attempt liability upon actors having present
and apparent ability to succeed (as opposed to being intrinsically incapable of
succeeding), and upon putative victims and objects of criminal activity being

102
See People v. Bink, 84 A.D.2d 607 (N.Y. App. Div. 1981).
2008] IMPOSSIBILITY ATTEMPTS 563

present (as opposed to being missing or dead). 103 The problem with such
tests is that they impose procrustean views regarding the fearfulness of
retrospective risks where, in reality, peoples views are variable, e.g., requiring that
an actor who has sexual intercourse with a woman whom he believes and hopes is
alive but is dead be acquitted, despite the fact that many regard such actors as
culpable. Given the diversity of views regarding retrospective risks, it seems
advisable to leave it to triers of fact, charged with representing informed 104 citizens
of the jurisdiction which makes X a crime, to decide whether they believe the
accused is guilty of attempt to commit offense Xby deciding whether the
accused would have committed offense X under counterfactual circumstances they
fear could have obtained.

APPENDIX: A HYPOTHETICAL STATUTE

A person is guilty of an attempt to commit an offense, X, 105 that the state

declares to be punishable by at least three years in prison, if

(1) while possessing purpose or belief regarding the circumstance

and result elements of a crime, 106

(2) he purposely does or omits to do anything which, under the

circumstances of law 107 or fact as he believes them to be, is an act or
omission constituting a substantial step in a course of conduct that is
planned to culminate in the crime; 108 and

103
See DUFF, supra note 2, at 8389 (reviewing such rules). Duff goes on to argue that his
proposed objective test of impossibility attempts duplicates what the former objective tests are best
understood to be designed to achieve. See id. at 20636, 37884.
104
This requirement, that triers of fact act as representatives of citizens who are informed, is
designed to address Sandys objection to basing the impossibility defense on uninformed popular
fears, e.g., that HIV can be spread through spitting. See KADISH & SCHULHOFER, supra note 5, at 599
n.10. See also DUFF, supra note 2, at 381 (the impossibility defense cannot justly be made a function
of uninformed, popular fears); GROSS, supra note 38, at 22123 (same).
105
The purpose of this introductory clause is two-fold: (1) to provide that an actor is guilty of
an attempt only if he attempts to commit what the state has declared to be an offense; and (2) to
provide that an actor is guilty of an attempt to commit an offense only if the offense is punishable by
at least three years in prison.
106
Section 1 distinguishes a crime, from offense X. The term offense X refers to an act-
type that the state declares to be prohibited. The term a crime includes two kinds of act-types: (1)
act-types that the state has declared to be prohibited, and (2) act-types that an actor mistakenly
believes the state has declared to be prohibited.
107
Section 2 extends liability to actors who make mistakes of law, thereby making it
possible to punish actors like Mr. Law.
108
Section 2 also collapses into a single clause, similar to 5.01(1)(c), what the Model Penal
Code now places in three clauses: 5.01(1)(a), (b) & (c). It does so by recognizing that the class of
defendants who are guilty of violating 5.01(1)(c) necessarily includes the class of defendants who
violate 5.01(1)(a) & (b). Cf. MODEL PENAL CODE COMMENTARY 5.01(1) pt.I, at 305 n.17 (1985)
564 OHIO STATE JOURNAL OF CRIMINAL LAW [Vol 5:523

(3) by virtue of his conduct and state of mind, he shows himself to

have been a substantial threat to the interests that the law declaring X to
be an offense seeks to protect. 109
________

[Alternative Provision]
[(3) based upon his conduct and state of mind, the triers of fact,
representing informed 110 citizens of the jurisdiction, conclude he would
have committed offense X under counterfactual circumstances they fear
could have obtained.] 111
________

(4) If the persons conduct is based upon a mistake of fact or law,

his conduct and state of mind shall not be deemed to be a substantial
threat to the interests the law declaring X to be an offense seeks to
protect, unless the means he mistakenly believed would culminate in a
crime are neither substantially harder for the police to detect nor
substantially easier or safer for persons to execute, than means that
suffice to commit offense X. 112

(citing a Hawaii statute that similarly collapses into a single provision what the Model Penal Code
separately does in 5.01(1)(b) and 5.01(1)(c)).
109
This is designed to exclude from liability for attempt persons who make mistakes of fact,
such as Mildred the Midwestern Voodooist, and persons who make mistakes of law, such as Mr.
Adulterer. The Model Penal Code, through 2.12(2), excludes actors like Mildred. However, rather
than doing so directly in its definition of attempt, the Model Penal Code first includes actors like
Mildred within its definition of attempt (MODEL PENAL CODE 5.01 (1962)) and, then, excludes them
under the separate provision of 2.12(2).
110
This requirement, that triers of fact act as representatives of citizens who are informed,
prevents the impossibility defense from turning upon uninformed popular fears, e.g., that HIV can be
spread through spitting. See supra note 104.
111
This Alternative Provision achieves the same thing as the standard formulation. If the
standard formulation seems more acceptable, it is only because by employing the language of
threat, the standard masks its effect. To say an actor revealed himself to have been a threat
means he revealed himself to be someone who would have jeopardized the interests that the law
defining crime X is designed to safeguard under a counterfactual situation of a kind that society
knows did not obtain but fears could have obtained.
112
This explains why an actor who makes the mistake of fact in thinking that hemophiliacs
will bleed to death if they are slightly nicked is not guilty of murder for inflicting a minor paper cut
on a hemophiliac enemy with intent to kill him; because there is no reason to believe that, when the
paper cut fails, the defendant would be willing to resort to the kinds of alternative methods that, to be
effective, leave more incriminating trails of evidence. It may also distinguish Lady Eldon, who
makes the mistake of law of thinking that there is a duty on French lace, from the professional
smuggler, who makes the mistake of law of importing German brandy in the mistaken belief that
there is a duty on such brandy: lace is easy to smuggle and hard to detect and, hence, there is no
reason to believe that, just because she is willing to risk smuggling French lace, Lady Eldon would
also be willing to smuggle the kinds of items on which there is a duty (e.g., Turkish brandy).
2008] IMPOSSIBILITY ATTEMPTS 565

(5) If the persons conduct is based upon a mistake of law, his

conduct and state of mind shall not be deemed to be a substantial threat
to the interests the law declaring X to be an offense seeks to protect,
unless his principal goal would have been served as well by means that
would have suffice to commit offense X. 113

113
This provision explains why a woman who makes the mistake of law of importing sugar in
the mistaken belief that there is a duty on sugar is not culpable, while the professional smuggler who
imports German brandy mistakenly thinking there is a duty on German brandy may be culpable of,
say, attempting to smuggle Turkish brandy: the womans goal is to possess a staple of life for
personal cooking and, hence, is not a goal that is likely to be furthered by smuggling anything which
there is or is ever likely to be contraband or dutiable; while the professional smugglers goal is to
exploit a black market and, hence, is a goal which is likely to be furthered by smuggling something
that is contraband or dutiable (i.e., Turkish brandy).