"The Difference of Differential Punishment" - Sayid R. Bnefsi (Fairfield University)
Criminal Law and Philosophy, 2025
By R.A. Duff (review of the paper by Sayid R. Bnefsi)
In many jurisdictions, someone who tries but fails to commit a crime is liable to conviction for a criminal attempt, but will be punished less severely than they would have been punished had the attempt succeeded. Legal theorists continue to devote substantial time and energy to the question of whether, and why, this common practice is justified; Sayid Bnefsi offers a new and intriguing argument in favour of the practice, at least in the context of attempted murder.[1]
The central claim is that murder and attempted murder involve ‘essentially distinct wrongs’, so that the successful murderer commits two distinct criminal wrongs, whilst the unsuccessful would-be murderer commits only one; and we properly punish one who commits two wrongs more severely than we punish one who commits only one of those wrongs. What then are these two wrongs? According to Bnefsi attempted murder involves the wrongness-making condition of the offender’s extreme indifference to a person’s interest in not being killed, which is manifested by taking affirmative steps to take that person’s life. Murder, by contrast, involves the wrongness-making condition of the offender’s reckless indifference to the occurrence of a killing during the commission of a crime in which risking a killing is an inherent part of the crime. Attempted murder thus involves ‘an indifference to the person-relative legal interest that others do not act with the intent to take their life’; murder ‘injures the person-neutral legal interest that all individuals have against killings during unlawful activities in which killings are inherently at risk’.
Bnefsi grounds his argument on three doctrines in American criminal law: the felony murder rule, under which it is murder to cause death in the course of committing any of a specified range of felonies; the transferred intent doctrine, under which someone who attempts to kill V1, but instead accidentally kills V2, is guilty of murdering V2, and perhaps of attempting to murder V1; and the rule that one can be convicted of attempting to commit a crime even if it was in fact impossible for the attempt to succeed. Thus, for instance, someone who accidentally causes death in the course of committing a felony commits the second kind of wrong, but not the first: that is why they are not guilty of attempted murder if, luckily, they kill no one. If I accidentally kill V2 in attempting to kill V1, however, I commit both kinds of wrong, and can therefore be punished for both. So too, Bnefsi argues, when I attempt to kill V1 and succeed in killing V1 I commit both kinds of wrong, and can properly be punished for both; but if my attempt fails I commit, and can properly be punished for committing, only the first kind.
The doctrines to which Bnefsi appeals, especially felony murder and transferred intent, are of course controversial, as Bnefsi recognises, but he does not here question them: indeed the felony murder rule, for instance, ‘cannot be rejected’. His argument is thus not an argument from first principles about how any system of criminal law should deal with attempted crimes. It is, rather, a rational reconstruction of American criminal law: an attempt to make coherent normative sense of this existing system. I’m inclined to agree that this is the best way to set about the normative theorising of criminal law—though such an approach does not preclude rejecting some aspects of the system that is being theorised, as many would reject such doctrines as felony murder.
In this short note, I’ll comment briefly on three aspects of Bnefsi’s argument. The first is his claim that the felony murder and the transferred intent rules are inconsistent with the ‘merger doctrine, that ‘the lesser offense of attempt is absorbed by the greater one’ of actually completing the crime.[2] He rejects this doctrine, since his claim is that the attempt and the completed crime are two distinct wrongs, which therefore cannot simply merge; but these two rules are not in fact inconsistent with this doctrine—if we understand this doctrine in line with what Bnefsi describes as its ‘prevailing rationale’, that the ‘successful commission of the target crime logically includes an attempt to commit it’.[3] If I accidentally cause someone’s death in the course of committing a felony, or accidentally shoot V2 in attempting to kill V1, these are murders, but do not amount to ‘successful commissions of the target crime’ of murder: for murder was not my target, nor does the death mark the success of my enterprise. They are indeed, despite Bnefsi’s claim that ‘there is no such thing as unintentional murder’, unintentional murders: for I do not intend to cause the death of the person I actually kill. Thus the merger doctrine does not imply that these kinds of—unintended, non-successful—murders logically include attempted murders.
Second, Bnefsi’s account of what makes attempted murder wrong—the agent’s ‘indifference to the person-relative legal interest that others do not act with the intent to take their life’—seems vulnerable to counter-examples. Suppose that D plants a bomb in a building, set to explode at a time when, he believes, the building will be occupied, intending thereby to kill whoever is in the building; he neither knows nor cares who they will be; by chance, the building is empty when the bomb explodes. Or D, out in the woods with his shotgun hoping to kill someone (anyone), sees movement behind a bush and, thinking it is a human being, fires at it; but it is in fact a deer. Such actions are, surely, criminal attempts at murder,[4] but do not attack or violate any ‘person-relative interest’: there is no one who can complain ‘You tried to kill me’. Perhaps, then, these attempts violate a ‘person-neutral’ interest: in that case, either we must say the same of all attempts, if we want to give a single account of what makes attempts wrong; or we must recognise that different kinds of attempt are wrong in different ways.
Third, Bnefsi limits his argument to murder and attempted murder (for the felony murder rule applies only to murder). But we must ask whether this kind of argument can be extended to other offences than murder, to explain why in general someone who unsuccessfully attempts to commit a crime should be punished less severely than they would have been punished had they succeeded. If it cannot be extended, then Bnefsi must presumably argue either that in other cases attempts should be punished just as severely as they would have been had they succeeded, which would make murder a striking exception to that general rule; or that some other account must be given of why in other cases attempts should be punished less severely—in which case we might wonder why that account, whatever it is, should not simply apply to murder as well. However, I think that an argument like Bnefsi’s could in fact be mounted for a very wide range of offences —for all those that can be committed without an intent to commit them.
For one example, consider criminal damage. I commit criminal damage if I damage property belonging to another either intending to do so or being reckless whether I do so.[5] Commission of the complete offence therefore does not ‘logically include’ an attempt to commit it, since attempt requires an intention to commit the complete offence:[6] if I damage another’s property recklessly rather than with intent, it is not true that I have attempted (successfully) to damage their property. Thus it can be true of someone who commits the complete offence that they would not have been guilty of attempted criminal damage had no damage actually been caused (if they caused damage recklessly rather than with intent); and someone can act in a way that would make them guilty of criminal damage if they actually caused damage without being guilty of criminal attempt if they actually cause no damage. These are the very features of the felony murder rule to which Bnefsi appeals in distinguishing murder from attempted murder. So should we then say that attempted criminal damage and criminal damage involve ‘essentially distinct wrongs’, because they violate ‘categorically different legal interests’? The answer, I think, is both ‘Yes’ and ‘No’.
‘Yes’, because we can say that criminal damage itself includes ‘essentially distinct wrongs’. It can be committed by attacking another’s property with the intention of damaging it, and also by acting in a way that recklessly endangers another’s property: but attacks and endangerments are distinct kinds of wrong;[7] thus an attempted criminal damage, which requires an intention to damage, is a different kind of wrong to reckless criminal damage, which involves no such intent. But also ‘No’, because attempted criminal damage involves the same kind of wrong as would be committed if the attempt succeeded: that is, the actual damage to another’s property completes the wrong that the attempter tries to perpetrate. The same is true of felony murder. Murder itself includes ‘essentially distinct wrongs’: the paradigm wrong committed by someone who attacks a victim intending to kill them; and the distinct wrong committed by one who causes death in the course of a felony that endangers life and manifests ‘extreme indifference to the value of human life’.[8] Once we understand this, we can see that the way in which an attempt to commit crime X and the completed commission of X can involve ‘essentially distinct wrongs’ gives us no reason to punish the failed attempt less severely than it would have been punished had it succeeded: for the successful completion of the crime involves the same kind of wrong as a failed attempt; and whilst the non-intentional commission of the complete crime (reckless criminal damage, felony murder) does involve a different kind of wrong to that involved in a failed attempt, it cannot be the case that someone commits both these kinds of wrong at the same time—that someone both intends to commit X and commits X non-intentionally.
Two final points are worth noting briefly. First, one could argue that a failed attempt and a successful commission of the target crime do involve different wrongs, and so merit differential punishments: for a successful murderer commits the wrong of murdering another, and has the victim’s death to answer (and be punished) for, while a failed would-be murderer fails to commit that wrong, although they do commit a lesser wrong of the same kind. That is a classic argument for punishing attempted and successful crimes differentially: it is a controversial argument that Bnefsi’s argument would, if successful, bypass; I suspect that he cannot bypass it. Second, some would deny that attacks and endangerments constitute ‘essentially distinct wrongs’: for the basic, single kind of wrong, they would argue, consists in endangering others’ interests in a way that displays a culpable indifference to those interests. Alexander and Ferzan deploy a version of this kind of argument,[9] and Bnefsi might share their view, since he analyses both of his ‘essentially distinct wrongs’ in terms of indifference. However, the natural implication of that view is that we should not have a law of attempts, defined in terms of intention, at all. Instead, we should have a general law of inchoate crimes, which criminalises anyone who acts with the mens rea required for the complete offence, and whose conduct comes sufficiently close to completing the offence: thus if one who intentionally damages another’s property, and one who damages it recklessly, are both to be convicted of the same complete offence of criminal damage, we should also convict of the same offence of ‘inchoate criminal damage’ both one who tries but fails to damage another’s property and one who recklessly endangers another’s property but does not in fact damage it.[10] We would then need to ask whether and why we should punish an inchoate offence less severely than it would have been punished had it been completed; but an appeal to ‘essentially distinct wrongs’ would not, I think, help us answer that question.
[1] Sayid Bnefsi, ‘The Difference of Differential Punishment’, Criminal Law and Philosophy, Online First, https://doi.org/10.1007/s11572-025-09755-w.
[2] J Dressler, Understanding Criminal Law, 5th edn (LexisNexis, 2009), 381; quoted by Bnesfi, n. 11.
[3] Bnesfi, at n. 18, again quoting Dressler (n. 2 above).
[4] In the second case, perhaps also in the first, conviction would depend on the doctrine that factual impossibility does not preclude conviction for an attempt—a doctrine that Bnefsi rightly accepts.
[5] See the English Criminal Damage Act 1971, s. 1; and the American Model Penal Code § 220.3 (on ‘Criminal Mischief’).
[6] See the English Criminal Attempts Act 1981, s. 1; Model Penal Code § 5.01.
[7] See R A Duff, Answering for Crime (Hart, 2007), ch. 7.
[8] Model Penal Code § 210.2(1)(b). The transferred intent rule requires different treatment, which I don’t have the space to discuss here.
[9] See L Alexander and K K Ferzan, Crime and Culpability: A Theory of Criminal Law (Cambridge University Press, 2009).
[10] See further R A Duff, Criminal Attempts (Oxford University Press, 1996), chs 5.3, 13.2-3.
