A. Two pure models of mens rea

Anglo-American criminal law’s general part is traditionally divided between different “theories” of liability. ^{Footnote 1} These include theories in which: (1) one can be liable for some bad result R when one’s acts caused R to come about or when one’s omissions failed to prevent R; or (2) one can be liable for trying to cause R, even when R does not come about (either at all, or at least not because of that trying). The first of these is a theory of a principal for a completed crime; the second is a theory of a principal for the inchoate crime of attempt.

These two theories provide analogies that illuminate two models of mens rea for accomplice liability. On the first analogy, accomplices are likened to principals; indeed, they are seen as but a kind of principal who by his own acts or omissions causes or fails to prevent some criminally prohibited result. On this model of complicity, the mens rea for accomplices should be exactly the same as it is for other kinds of principals, no more and no less. This means that for crimes requiring a mens rea less than that of purpose (or specific intent) such as knowledge, recklessness, or negligence, accomplice liability on this model only requires the lesser mens rea as well. On the second analogy, by contrast, accomplices are likened to attempters in that neither attempters nor accomplices are thought to cause the criminally prohibited result. Lacking causation, the thought is, only serious culpability suffices for liability. So like the mens rea of attempt, accomplices must have the most culpable mens rea of purpose (or specific intent). This, even though the crime aided (or attempted) itself has some other and lesser mens rea requirement.

There is a third pure model of liability that has historically played a role in shaping the mens rea doctrines of accomplice liability, and that is the model of vicarious responsibility. The analogy here is to a “partnership in crime,” an agreeing by one party to another party’s criminal acts. This is based on the conspiracy theory of liability. The idea is that merely agreeing — without procuring or aiding — is enough so long as the conspirator shares the criminal objectives of the principal. The mens rea this model dictates is the same as that dictated by the attempt model, namely, a true purpose to achieve the criminal result (even if less than purpose suffices as mens rea for the underlying offense).

We shall consider only the first two models in our analysis. We ignore the third model, partly because it no longer accurately describes majority Anglo-American criminal law (although it once did). Additionally, the model is normatively undesirable: no one is morally blameworthy for another’s actions just because he agreed that such actions should be done when such agreement in no way aids or procures the criminal acts. It is thus a hopeless enterprise to seek a mens rea that can make one blameworthy on this vicarious basis, because there is none.

Only two pure models, thus, will organize the analytic, descriptive part of our analysis. We shall first consider the completed-crime-as-a-principal model.

B. The mens rea for complicity if accomplices are simply one kind of principal

The answer to our question becomes pretty straightforward if accomplices are in reality but one kind of principal. For then the mens rea for accomplice liability should be no different than it is for principals generally. These are puzzles here, ^{Footnote 2} but these are puzzles about mens rea generally. On this view there is nothing uniquely puzzling about the mens rea of accomplice liability.

Some would dispute this last conclusion. They urge that even on the completed-crime-as-a-principal model there must be two mens reas for accomplice liability: there must be a mens rea requirement with respect to the accomplice’s act of aiding or procuring; and there must be a secondary and separate mens rea requirement with respect to the result the one aided causes to come about. If intent is the mens rea requirement, for example, there must be an intent to aid, and there must be an intent that the act aided succeeds in causing the result that makes that act criminal.

Yet this is a mistake. The law could require two mens reas for any crime, not just those of accomplices. It could, for example, require an intent regarding the actor’s means as well as his end; it could require for murder that the accused not only intend the result, death of his victim, but also that he intend a certain means of accomplishing that death, say lying in wait or poisoning. But the law only rarely requires such intended means as well as intended ends. And in any case, there is no special necessity to do so just because the intended means are the actions of another person.

The completed-crime-as-a-principal model is not the model for mens rea of accomplices of either of the two sources of the general part of the criminal law in America, the common law or the Model Penal Code. This is because standard criminal law doctrine does not conceive of accomplices as causers of criminally prohibited states of affairs. ^{Footnote 3} Rather, the principals accomplices procure or aid are the causers of such states of affairs. Accomplices are held liable for the results principals cause, not as co-causers with such principals, but only “derivatively,” that is, by virtue of standing in a certain relationship to those who did cause such bad results.

That accomplices do not cause the results caused by their principals has long been based on two ideas. One of these ideas is that often an accomplice’s acts of aiding or procuring are not necessary for the criminal result to obtain. In such cases those acts are not in law a “cause-in-fact” of the harm. The second idea is that principals act as “free, informed, voluntary” actors whose acts both start fresh causal chains and cut off the antecedent causal contributions of other earlier actors. By long established proximate cause doctrines, ^{Footnote 4} principals are thus typically intervening (or “superseding”) causes so that accomplices are not in law legal (or “proximate,” “direct”) causes of some harm.

Despite the prevalence of these general doctrines of proximate causation, there are nonetheless four situations in which the common law holds defendants who are in-fact causes as principals even though they contributed to some criminal result only through the acts of others. The first of these are cases in which the principal actor does not temporally intervene between the accomplice’s acts of aid and the harm the law prohibits. These are the simultaneous co-causer cases. For example, A holds the victim while P stabs her with a sword. In such cases, both simultaneous co-causers of the death are held as principals even though one killed only through the acts of another.

A second class is that known as the “innocent agent” kind of case. If A dupes P into doing some criminal act, or A coerces P into doing it, or A manipulates some body part of P into causing a bad result, or A puts P into a state of shock, hypoglycemia, hypnosis, or insanity, so that P’s body is involved in causing some criminal result, A is held to have used P like he might use any other tool to accomplish his own criminal ends. In cases of such innocent agency by P, both the common law ^{Footnote 5} and the Model Penal Code ^{Footnote 6} uniformly hold A as a principal for causing or failing to prevent some result. The mens rea required for convicting A is accordingly just the mens rea of the substantive crime that A used P to accomplish.

As the late Sandy Kadish has shown in persuasive detail, ^{Footnote 7} the innocent agent cases are misnamed. For it isn’t the innocence of the agent that induces the law to hold the aider as a principal and not as an accomplice in these cases. Rather, that innocence is a proxy for something else, what the law calls “proximate” or “direct” causation. On the long established legal doctrines above described, A proximately causes some harm H through the actions of P only when P’s acts do not constitute an “intervening cause;” and P’s acts, if innocent, are usually not intervening causes, whereas if they are not innocent but culpable, they usually are intervening causes. The limitations on when the law adopts the first model, accordingly, are really drawn in terms of causation, not in terms of the guilt or innocence of the putative principal.

A third situation in which the common law holds an actor who acts through another as a principal, are cases of omission. If P has a duty and an ability to save another, and then fails to do so when little risk or inconvenience is involved, P is liable for the other’s death or injury by having failed to prevent it. Yet omissions, even if intentional, are in law ineligible to serve as intervening causes. ^{Footnote 8} Thus, if A induces P to omit to save another, causes P to omit (by tying P up, say), or aids P in some way in P’s refraining from doing anything, A is liable as a principal for the death or injury of the person that P failed to save. Analogously, if A like P is under a duty to save the victim, when A omits to hand P the life-saving rope, A is also liable as a principal for his own failure to prevent the victim’s death or injury.

Fourth and finally, agents can be legally culpable actors and yet not constitute intervening causes if they are merely negligent or reckless in their culpability. ^{Footnote 9} Thus, when P negligently fires the defective boilers of his steamship, his negligent action does not necessarily prevent earlier, culpable actions of others from being causes of the explosion and the resulting deaths of the passengers on the ship. ^{Footnote 10} Both the owner who ordered the ship to depart, and the captain who ordered the engineer to fire the boilers, may in law be causes of the death by explosion along with the engineer who actually fired the boilers. All are equally culpable, so none is even a “relatively innocent” agent. Yet each can be (seriatim rather than simultaneous) co-causers of the harm nonetheless.

For the many cases that fall into these four categories, the common law thus treats defendants who cause criminal harms through the action of others as principals. The mens rea required of such defendants in such cases is unexceptionable — it is simply the mens rea of the substantive offense. It is only those whom the law regards as accomplices — people who are only derivatively liable because the principals they aid or procure are the “free, informed, voluntary actors” that constitute intervening causes — that raise the possibility of there being a mens rea requirement for complicity different than the mens rea requirement of the substantive offense aided. To find a model for such an independent mens rea, the common law turned to the second pure model of mens rea we earlier distinguished, the model based on the mens rea of attempt. We describe that second model in the next section.

C. Modeling the mens rea of complicity on the mens rea of attempt

Pursuing the analogy to attempt requires us first to be clear about what the mens rea requirements for attempts are. Attempts are secondary crimes in the sense that one is never guilty of “attempt” simpliciter; one is only guilty for attempting some action that has itself already been made criminal. That might suggest that the mens rea for attempts would itself be along the lines of the first model above, namely, that whatever mens rea was required for the crime attempted would also be the mens rea of criminal attempts. Anglo-American criminal law has not gone this way, however. Rather, at common law, attempts were said to be “specific intent” crimes irrespective of whether or not the crime attempted required specific intent for its mens rea.

Specific intent crimes standardly require true purpose with regard to the acts attempted to be done and with regard to the results attempted to be achieved; mere predictive belief, even to a near certainty, is not enough. ^{Footnote 11} The man who shoots at the light in a tent, believing it to be almost certain that he will also hit a baby behind the light, does not have the specific intent to kill required for attempted murder (even though had he hit and killed the baby he would have had mens rea sufficient for murder). ^{Footnote 12}

To understand what the law of attempts (and later, the law of complicity) means by “specific intent,” or “purpose,” requires that we take a brief excursion into the differing elements that make up the actus reus of any crime. No criminal code prohibits simple bodily movements such as wiggling one’s toes, for the reason that such acts are of no moral moment. It is only when such voluntary movements of the human body instantiate act-types like killing a policeman, rape, breaking and entering, and the like, that morality takes an interest in such actions. It is thus only in terms of such complex act-types that criminal actus reus prohibitions are framed.

To understand what the mens rea provisions require with respect to such act-types requires that one first get a handle on the different elements that make these act-types the types they are. Both the Model Penal Code ^{Footnote 13} and the common law ^{Footnote 14} that preceded it divide such elements into three categories: results (of voluntary acts); circumstances (in which the causing of such results might take place); and conduct (that the voluntary act instantiates on some occasion). We need to assess a mens rea requirement like that of purpose with respect to each such element separately because whether purpose applies to that act-type, and even what “purpose” means, varies by the kind of element involved..

Both the common law ^{Footnote 15} and the Model Penal Code ^{Footnote 16} often distinguish so-called “result crimes” from so-called “conduct crimes.” Homicide is said to be a result crime because the crucial element of the actus reus, death, must be the result of the voluntary act of the accused. By contrast, theft, rape, and breaking and entering, are said to be conduct crimes because an act-type (taking, raping, breaking, entering) that defendant’s voluntary act instantiates is what is prohibited, not a result. In our long-settled view, ^{Footnote 17} this distinction between conduct and result crimes is wholly artificial and unsustainable, because all act-types prohibited by the criminal law contain result elements. Indeed, it is just these results — movement of someone else’s property, penetration of a woman’s body, breakage of a window or door, being inside another’s building — that make a voluntary act an instance of the conduct prohibited by theft, rape, and breaking and entering statutes.

Still, since the conduct/result distinction is conventionally recognized by both the common law and by the Model Penal Code, we will honor the convention to the extent of dealing with mens rea regarding conduct separately from mens rea regarding results. For the mens rea of attempts, that separate discussion can be brief: the specific intent required for attempt liability is the same vis-à-vis conduct as it is for results. Both require true purpose (that he engage in such conduct or that he cause such result) before liability attaches. In attempted murder, for example, one must intend the result-element (of death of the would-be victim) and equally intend the supposedly different conduct-element (to kill him), to have the specific intent required for conviction.

This matter is further complicated by attempt law’s different treatment of the third standard element of actus reus, what are called “circumstance elements” of offenses. Circumstance elements are states of affairs that must exist for the accused to be guilty of some crime but which are not caused to exist by the accused’s criminal act. If the crime involved is attempted murder of a police officer, for example, the fact that the victim was a police officer is classified as a circumstance element. Surprisingly, perhaps, the mens rea of attempts does not require specific intent with regard to the circumstance elements of the crimes attempted; rather, the only mens rea required regarding these elements is whatever mens rea is required by the underlying offense. ^{Footnote 18} If the substantive crime of murder of a police officer, for example, only requires negligence with respect to the fact that the victim is a police officer, then the crime of attempted murder of a police officer only requires negligence regarding that circumstance element too.

Given this differential treatment of result and conduct elements from circumstance elements, “the” attempt mens rea model is itself something of a mixed bag. Nonetheless, if we apply this model to the mens rea for complicity, we get the dominant common law version of mens rea for complicity. On this model, complicity requires true purpose with regard to the result or conduct elements of the offense aided, even when that latter offense does not require purpose with regard to those elements. One thus is an accomplice to possession of a controlled substance, for example, only if the aid one renders (such as giving the location of a seller) is given with the purpose that the sale takes place and possession be achieved. One’s purpose, in other words, must not only be to give aid to someone in his doing of a criminal act, but also that the person aided succeeds in the doing of that act. This is often referred to in the cases as the accomplice having to have a “stake in the enterprise.” ^{Footnote 19}

This attempt-inspired model is the dominant common law view of the mens rea of attempts. Such a view typically carries with it the supposed “corollary” that one cannot be an accomplice to crimes that have only negligence or recklessness as their mens rea, such as involuntary manslaughter. The thought here is that to be an accomplice one would have to give aid with the purpose that the principal aided cause the death of the victim, yet the underlying crime aided (such as involuntary manslaughter) can only be done by one who is negligent or reckless with respect to death. The common mantra has been that “one cannot try (or intend) to be negligent/reckless.” ^{Footnote 20}

It is true that even prior to the Model Penal Code there had been seeming exceptions to this legal corollary of the common law. In the McVay case, ^{Footnote 21} for example, the court said that there could be accomplice liability for the ship owner because he was culpable (i.e., negligent) in the way required for conviction of the offense he aided (involuntary manslaughter). Yet, the court failed to notice that the owner’s liability did not require him to be held as an accomplice. For the engineer was not an intentional killer, but only a negligent one — the engineer’s act of firing the boiler did thus not necessarily constitute an intervening cause. This means the owner’s order, no less than the captain’s, might well be considered a proximate cause of the death of the passengers. In this event, the mens rea for the principal offense, mere negligence, was thus sufficient to secure the owner’s conviction because he was, in truth, but another principal along with the ship’s captain and engineer.

There is a puzzle about the common law’s adoption of the attempt model of mens rea as the mens rea for complicity, and that is that accomplices, unlike attempters, are punished for the completed crime. This is typically a much more severe punishment than that given to mere attempters. While it is true that in complicity cases the harm the law prohibits has occurred, this can be true in attempts too, so long as the defendant (as in complicity cases) is held in law not to be a proximate cause of that harm.

This problem does not bedevil the third pure model of liability we discussed early on, that of vicarious responsibility. On this third model, the accomplice by his aid has somehow adopted the principal as his agent or partner, and this identifying-with-the-principal’s-act-and-purpose is thought to make it fair to punish the accomplice for a harm that he did not cause. Yet in our view, there is no circumstance in which one is rightly blamed vicariously, no matter how explicitly one agreed, promised, ratified, or whatever, the criminal acts of another. All of us admittedly have valid normative powers to create certain kinds of moral and legal obligations, a power we exercise in the making of requests, promises, and commands; we also have normative powers to change wrongful behavior into permissible behavior via our consent. ^{Footnote 22} But we have no normative power to make ourselves blameworthy for others’ breach of their obligations merely by consent, promise, or other form of assent. We can heroically take upon ourselves another’s punishment; yet such acts are heroic just because we do not deserve such punishment, not originally and not by any voluntary assumption. The upshot of putting aside this third model of liability is that the common law is left with the attempt model to explain its mens rea requirements; and on such a model the puzzle about punishment for results remains.

D. The “impure” (or mixed) model of mens rea for accomplices

The best way to view the Model Penal Code’s mens rea requirements for complicity is as a combination of two of the pure models we have just discussed. The Code achieves this compromising combination by dividing its mens rea requirements into two components, “primary” mens rea and “secondary” mens rea. ^{Footnote 23} While both mens rea are required in order to be an accomplice, primary mens rea concerns the accomplice’s acts of aiding, while secondary mens rea deals with the criminal acts of the principal that are being aided. For the primary mens rea requirements, the model is the common law’s attempt model: specific intent (or “purpose”) is required. For the secondary mens rea requirement, the model is the mens rea of the principal model: the accomplice only needs the mens rea sufficient for conviction as a principal of the crime aided, which, as mentioned, is often less than true purpose. Let us consider primary and secondary mens rea in detail, because on their face the two mens rea requirements seem to contradict each other.

An accomplice has primary mens rea if he aids “with the purpose of promoting or facilitating the commission of the offense.” ^{Footnote 24} Both this language and the official commentary behind it make plain that this generally requires an accomplice to have what we shall call strong primary mens rea regarding the commission of a crime by another. Strong primary mens rea is the requirement that the accomplice has the purpose that the act he is aiding will have all the material elements that make the act an offense. ^{Footnote 25} So if the underlying offense is sale of drugs and the accomplice supplies information about the location of the drugs to the seller who then acquires and resells them to a buyer, ^{Footnote 26} the supplier of information must have as his purpose in giving the information that the drugs be acquired and resold by the person to whom he gave the information. The accomplice must have a “stake in the sale” by the principal, in the language of the common law on this point. Mere knowledge that the person who received information about the drugs will then acquire and resell them is not enough.

The Model Penal Code’s strong primary mens rea is of course nothing other than the common law view of mens rea, one requiring specific intent regarding all conduct and result elements of the crime aided or attempted. Yet the Code introduces secondary mens rea in order to avoid the legal corollary of the attempt model for complicity mens rea of the common law. (The corollary was that one could not be an accomplice to crimes of negligence or recklessness because one cannot purposely be negligent or reckless.) To reverse this corollary, at least concerning “result crimes,” the Model Penal Code allows that “when causing a particular result is an element of an offense . . . the kind of culpability, if any, with respect to that result that is sufficient for the commission of the offense,” is also sufficient mens rea for accomplice liability. ^{Footnote 27} The Code adopts, in other words, the underlying-offense model of mens rea as sufficient for accomplices, despite the seeming requirements of more in its primary mens rea provision.

The McVay case discussed before was intended by the Code drafters to illustrate this secondary mens rea requirement. ^{Footnote 28} When the owner of the steamship ordered his ship’s captain to fire up the defective boilers (which later exploded, killing passengers), he did not in fact have as his purpose that anyone die as a result of his order. He did not, thus, have specific intent with regard to the result — death — or the conduct — killing. But the Code’s secondary mens rea requirement now says that the ship’s owner does not need to have as his purpose that the passengers die in order to be an accomplice to some form of homicide; it is enough that he has only the culpability with respect to their death needed by the ship’s captain as a principal, which (at common law for involuntary manslaughter) is gross negligence. If the ship’s owner should have known of the danger and ordered the ship to sail anyway, that inadvertent carelessness is enough for accomplice liability under the Code.

Yet what happens to primary mens rea in cases such as this? Here the Code moves to what we shall call “weak primary mens rea.” Weak primary mens rea still requires purpose (or specific intent) but the object of that required purpose is different. The accomplice now need not intend that his principal succeed in perpetrating the underlying offense with all of its elements; rather, the accomplice need only intend that the principal engage in “conduct causing such result,” in the language of the Code. ^{Footnote 29} On the facts of McVay: the ship owner need not intend the death of the passengers, the result element of the underlying offense of involuntary manslaughter; he need only intend, for example, that the ship captain fire up the boilers (since it was this conduct, among others, that causally contributed to the deaths).

This weakening of primary mens rea saves the Code from formally contradicting itself in its primary and secondary mens rea formulations. ^{Footnote 30} Yet avoidance of formal contradiction is small consolation if the Code’s weakening of primary mens rea allows the conviction of just the people the Code sought to protect by requiring purpose rather than knowledge as primary mens rea, namely, those suppliers of goods and services who foresee but do not intend the criminal uses to which their goods and services are put by those to whom they supplied them. We pursue two questions below to show that this is indeed the result of the Code’s weakening of primary mens rea.

The two questions are these. First, focusing for now only on so-called “result crimes,” aren’t suppliers of goods and services who merely know (or worse, merely should know) of the criminal uses to which those goods and services will be put, accomplices in the crimes aided by receipt of such goods and services? If so, this would be an application of the Code that goes against its clearly expressed intent. Second, can this unacceptably extensive liability of accomplices be cabined to “result crimes”? Or does it not extend to all crimes, so that no work is left to do for strong primary mens rea? And does not a universal requirement of only weak primary mens rea result in a too extensive liability for accomplices generally? We take up each question in turn.

Regarding the first question, we should focus exclusively on crimes in which everyone would agree that there is a result element. Involuntary manslaughter is the Code’s example, so let us stick with that one. Crucial to understanding the Code’s requirement of weak primary mens rea is understanding the reference of “conduct” when we are told that accomplices must have as their purpose the promotion of “conduct causing such result.” ^{Footnote 31} Let us work through the possibilities.

First, since conduct is one of the three possible elements of actus reus in the Code, ^{Footnote 32} maybe “conduct” here refers to an element of the underlying offense. With crimes like “abuse of a child where the abuse results in death,” or “driving in a manner dangerous to the public when that driving results in death,” the conduct could be abuse of a child or dangerous driving, separate from the result element of death. The required purpose would then be to promote or facilitate such abuse or such driving, not necessarily to promote death. But most crimes are not so defined; most crimes are like homicide, where the only conduct that is an element of the offense is conduct like killing that contains the result element, death, in it; one cannot intend such “conduct” and not intend such results as death. Weak primary mens rea would become strong primary mens rea for crimes like homicide under this construction, depriving the secondary mens rea requirement of any application.

Second, the Code Commentary’s example of felony-murder ^{Footnote 33} suggests that perhaps “conduct” refers to an element of some crime other than the one for which accomplice liability is sought. Thus, in a bank robbery where a teller is killed, an agent who has as his purpose to aid the robbery could be held as an accomplice to reckless (“depraved heart”) murder so long as he was reckless with respect to such death. In such a case, primary mens rea is satisfied by the purpose to aid the robbery, which would be the purpose to aid the conduct in the relevant sense. Yet this construction is not available for more typical cases like McVay, since in such cases there is no crime other than manslaughter for which the conduct could be an element.

Third, the conclusion of the preceding discussion — that “conduct” need not refer to any element of any offense — thrusts one out onto what we shall call the uncharted sea of conduct that is not an element of any crime. What makes the sea so broad, open-ended, and uncharted are two features. One is the embarrassment of riches with which any act can be described (once one is freed from use of some particular statutory description as an element of an offense). The ship engineer’s act of firing up the boilers, for example, could be equally well described in terms of other of its properties; thus, it was an act of: moving one’s fingers and arms; lighting the fire-box; heating the water in the boilers; getting up steam; exploding the boilers; following the owner’s order; and so on. Any individual act has an indefinitely large number of equally good descriptions. ^{Footnote 34} And that act will always be intended only under some of those descriptions and not others; ^{Footnote 35} so it matters what description is the “conduct” that is required for weak primary mens rea.

The second feature of actions relevant here is not the prolixity in description of any given act; rather, it stems from the fact that most actions like “getting up a head of steam” or “firing up the boilers” are a compound of different acts. ^{Footnote 36} If one settles on “firing up the boilers” as the right description of the principal’s conduct here, surely this is typically done not by one act but by a series of acts: one fills the boiler with water, fills the fire box with coal, presses the igniter switches, moves the bellows, opens the steam valves, and so on. Most actions we care about are compound events, much like “breaking a rack” in billiards is a compound of many striking events between individual billiard balls. ^{Footnote 37}

So what does the weak primary mens rea requirement of “purpose with respect to the conduct of the principal causing the death of the passengers” come to in a case like McVay? The worry here is not (or not only) indeterminacy. It is rather that the net for accomplice liability now sweeps too wide, even by the Code’s own professed intent. Suppose A₁ is a supplier of water for the boilers, A₂ is a supplier of the coal, and they both deliver the goods that they sell to the ship’s captain in McVay. Or suppose that A₃ is a repairman who inspects and makes repairs on the firebox before each departure. And suppose that A₁, A₂, and A₃ should have noticed the defects in the boilers, but they didn’t (thus satisfying the secondary mens rea for common law involuntary manslaughter, which is negligence). Don’t A₁, A₂, and A₃ each also satisfy weak primary mens rea for accomplice liability? Does each not have as the purpose with which he supplies his goods and services, that the engineer do at least some of the acts that are encapsulated in “fire up the boilers?” After all, some of those acts of the engineer include filling the boilers with water and the coal bins with coal, and maintaining a working firebox? And do not A₁, A₂, and A₃ intend to aid him in those acts, all part of his conduct of firing up the boilers?

The upshot is that the Code’s weakening of primary mens rea regarding results so expands the liability of accomplices that it captures just the suppliers it said it was not going to capture. Indeed, it goes beyond those cases to capture suppliers who do not even know of the relevant dangers but only should know.

We said there were two problems. The first we have now concluded. The second exacerbates the first. For now consider why the reductio just concluded doesn’t extend to all accomplice liability under the Code, not just to accomplices to so-called “result crimes.” In our view, mentioned earlier, all crimes are result crimes. That is, all crimes contain result elements. The supposed examples of so-called conduct-crimes — crimes like theft, rape, breaking and entering — turn out to use verbs of action that are causally complex, that is, contain result elements. Just as to kill is to cause death, to hit is to cause contact, to maim is to cause disfigurement, so to steal is to cause movement of property of another, to rape is to cause penetration, to break is to cause breakage, and to enter is to cause presence within. Once one sees that there are no pure conduct crimes — crimes whose actus reus contains no result elements — then one should also see that the reductio above applies to all accomplice liability.

Moreover, even if (contrary to fact) there were crimes whose elements consisted only of conduct but no results, conduct is systematically treated like a result under the Code. This is true for attempt liability, for example, where purpose (or specific intent) is required for results and conduct, willy nilly. ^{Footnote 38} An attempted rapist must have as the purpose with which he assaulted his victim, the “conduct” of rape, as much as an attempted killer must have as the purpose with which he assaulted, the result of death. Likewise under the general mens rea provisions of the Code, ^{Footnote 39} conduct and results are treated the same when it comes to specifying what purpose means; only circumstances are treated differently. So even if conduct were analytically distinguishable from results, it is nonetheless morally and legally indistinguishable from results for purposes of assessing culpability. Conduct and result elements should also be treated the same here. If there were conduct crimes, complicity mens rea regarding such conduct should be construed to be the same as it is for results; namely, only underlying offense mens rea plus weak primary mens rea is required. A true purpose that such conduct take place is not required; suppliers of goods and services would thus still be accomplices, despite their lack of any stake in the enterprise of those whom they supply.

The third kind of element under the Model Penal Code (in addition to result and conduct elements) is circumstances. To our knowledge, there are no purely “circumstance crimes.” Be that as it may, circumstances are commonly elements of result or conduct crimes. Rape, for example, is causing the result, penetration, by one’s voluntary acts in the circumstance that there is no consent. Perhaps at least here the Code preserves strong primary mens rea for accomplices, rather than allowing liability to turn on the secondary mens rea of the underlying offense and the weak primary mens rea of intending to aid something?

Why drafters of the Code explicitly take no position on the mens rea needed for accomplice liability for the circumstance elements of the underlying offense, is a well-known puzzle. As noted earlier, both the Code and the common law that preceded it do not leave this question open in the mens rea of attempts. For attempts, the mens rea regarding circumstances sufficient for conviction of the underlying offense is sufficient for liability for attempting such an offense as well; a specific intent regarding such circumstance elements, in other words, is not required.

Yet with the mens rea for complicity, the Code creates a “deliberate ambiguity as to whether the purpose requirement extends to circumstance elements of the contemplated offense or whether, as in the case of attempts, the policy of the substantive offense on this point should govern.” ^{Footnote 40} The stated rationale for this ambiguity is to leave it to the courts to decide on a statute-by-statute basis, analogizing complicity to the statute-by-statute basis on which this matter is governed for conspiracy. Still, the Code’s preferred position is that the “purpose to promote or facilitate the commission of the crime” (the required primary mens rea for conspiracy as well as complicity) exists “although the actor did not know of the existence of a circumstance . . . when knowledge of the circumstance is not required for the substantive offense.” ^{Footnote 41}

If this preference is taken seriously, then the mens rea regarding circumstance elements joins the mens rea regarding result and conduct elements: concerning all three possible elements of the actus reus of the underlying offense, ^{Footnote 42} only the mens rea of that offense is required for accomplice liability, as long as weak primary mens rea is present. The upshot is that in all accomplice cases, knowing, reckless, or negligent suppliers of goods and services to criminals are made liable by the Code, because weak primary mens rea is too weak to save them. This is too broad a liability; it is also plainly not what the Code intends.