T HE ESSENCE OF CULPA BILI T Y

25

I. Unpacking Recklessness

Because an understanding of recklessness is central to our argument, this is where we begin our analysis.3 T he Model Penal Code defines “recklessly” as follows:

A person acts recklessly with respect to a material element of an offense when he consciously disregards a substantial and unjustifiable risk that the material element exists or will result from his conduct. The risk must be of such a nature and degree that, considering the nature and the purpose of the actor’s conduct and the circumstances known to him, its disregard involves a gross deviation from the standard of conduct that a law-abiding citizen would observe in the actor’s situation.4

This formulation is substantially the same as the formulations of recklessness in federal and state criminal codes and judicial decisions.5

This formulation of recklessness makes it appear that an actor, to be deemed reckless, must consciously disregard a risk that satisfies two independent criteria: the risk must be “substantial”; and the risk must be “unjustifiable.” We believe that this appearance is mistaken with respect both to the purpose behind the formulation and to a normatively attractive conception of the culpability displayed through recklessness. In our view, the “substantiality” prong of the definition should be eliminated.6

Consider the following hypotheticals:

1.Driver is accompanied by Passenger, who suddenly displays the symptoms of a severe heart attack. Driver speeds down city streets in order to get Passenger to a hospital in time to save his life. In doing

3T his section draws from Larry Alexander, “Insufficient Concern: A Unified Conception of Criminal Culpability,” 88 Cal. L. Rev. 931 (2000).

4 Model Penal Code § 2.02(2)(c) (1985).

5See U.S. Sentencing Guidelines Manual § 2A1.4 cmt. 1 (1998); Ariz. Rev. Stat. Ann. § 13–105(9)(c) (West 1998); Cal. Penal Code § 450(f) (West 1999); N.Y. Penal Law § 15.05(3) (McKinney 1999); In re Steven S., 31 Cal. Rptr. 644, 652 (1994); Joshua Dressler, Understanding Criminal Law 144 (4th ed., 2006).

6Ultimately, we do advocate retaining the portion of the Model Penal Code’s recklessness definition that requires that the risk be a “gross deviation” from a risk that a law-abiding actor would take. See infra III.B. We believe, however, that one may impose what may be viewed as an “insubstantial” risk, but that risk may be so unjustified that its imposition is still a “gross deviation” from the risk a law-abiding person would take.

26	T HE CULPA BLE CHOICE

so, Driver creates risks to bystanders’ lives, limbs, and property of magnitude R.

2.Daniel loves to set off dynamite on city streets just for the thrill of watching the dynamite explode. In doing so, he creates risks to bystanders’ lives, limbs, and property of magnitude R/100.7

3.Deborah likes to take Sunday drives. Although she is a very careful driver, in taking Sunday drives she creates risks to bystanders’ lives, limbs, and property of magnitude R/10,000.

4.Demented likes to expose others to risk. He has concocted his own version of Russian roulette in which, when he pulls a switch, he creates a risk of magnitude R/100,000 that an innocent person will die or be severely injured.

Now if the substantiality criterion is truly independent of the justifiability criterion, then there must be some level of risk imposition below which the actor cannot be reckless. Suppose that level is just above R/100. In case 1, Driver does satisfy the substantiality criterion but may escape the judgment of recklessness by failing to satisfy the (lack of) justification criterion. In other words, Driver may have justifiably imposed risk R on bystanders because his purpose in doing so was to save the life of Passenger. Deborah, the Sunday driver in case 3, also escapes the judgment of recklessness, in this case because R/10,000 is below the risk threshold, or so we are assuming.

The problem with the independent criteria approach, however, surfaces with cases 2 and 4. Both Daniel and Demented have imposed risks that by hypothesis fail the substantiality criterion, yet the intuition is virtually unshakable that they are acting culpably recklessly toward bystanders. This is perhaps clearer in the case of Demented, who subjects others to risk simply because he enjoys it. Imposing (unconsented to) Russian roulette on others, no matter how high the ratio of empty to loaded chambers, seems a clear case of culpable conduct. What is doing all the work here is, of course, the justifiability criterion, which also explains the intuition about Daniel, namely, that the mere thrill of

7Our usage of “R” or “R/100” alludes to the actor’s subjective assessment of the relative frequency that the harm will occur. Our view of risk is discussed later in this section.

T HE ESSENCE OF CULPA BILI T Y

27

a dynamite blast does not justify imposing more than a de minimis risk on others.

What emerges from consideration of these cases is the proposition that the level of risk that one may permissibly impose on others is dependent on the reasons one has for imposing that risk. The reason Driver has for imposing risks on bystanders is a weightier one than the reason Deborah, the Sunday driver, has for doing so. Thus, we deem it permissible for Driver to impose a higher risk on bystanders than Deborah may impose. Daniel’s reason – the thrill of explosions – may justify even less risk imposition than Deborah is permitted. And Demented’s reason – imposing risks on others for its own sake – cannot justify even the slightest increase in risk.

The conclusion that one should draw from these examples is that recklessness consists of imposing unjustifiable risks on others. The level of risk imposed will bear on its justifiability but is not itself an independent criterion of recklessness. Even very tiny risk impositions can be culpable if imposed for insufficient or misanthropic reasons, just as very large risk impositions can be nonculpable if supported by weighty reasons. It certainly may be the case that there is a level at which the risk becomes so minute that it may not be worth devoting precious resources to criminalizing the conduct or prosecuting the actor; but this practical argument does not undermine our conceptual claim that an actor’s culpability hinges entirely on the unjustifiability of the risk. Imposing unjustifiable risks to others’ legally protected interests is culpable behavior because it displays insufficient concern for others’ interests.

A second question about recklessness is whether “risk” should be assessed subjectively or objectively. Consider the following case: David wants to get home in time to watch the Lakers game on television. He accelerates until his speedometer reads ninety miles per hour, a speed that he believes creates a very substantial risk of death, serious bodily injury, or property damage to other drivers, passengers, and pedestrians. In fact, his speedometer is broken, and he is going only fifty-five miles per hour, a reasonable speed given the road and traffic conditions. Is David acting recklessly?

There are two approaches one might take to this case. Under an “objective” approach, we could distinguish between the “risk imposed” and the “risk the actor believes that he is imposing.” The risk imposed

28	T HE CULPA BLE CHOICE

serves as an actus reus, whereas the risk the actor believes he is imposing speaks to his culpability.8 Under this view, David would not be reckless because the actual risk imposed is not of a magnitude that would render one negligent if imposed inadvertently. Or to put the point another way, were David to be aware of the actual risk he is imposing, he would be neither negligent nor reckless, because that level of risk is justifiable. Notably, David would still be culpable in that he would be attempting to act recklessly; but he would not in fact be reckless.

In contrast, we might take a “subjective” approach: an actor is reckless if he believes he is imposing a level of risk that would be unjustifiable (given his reasons) regardless of the actual risk he is imposing. On this approach, David is driving recklessly at fifty-five miles per hour because he believes he is driving ninety miles per hour.

As we discuss in Chapter 5, we believe that attempts and successes should be regarded as equally culpable and, because they are equally culpable, equally deserving of punishment. It therefore ultimately matters little whether the objective or the subjective approach to recklessness is chosen. On either approach, David is equally culpable and therefore equally blameworthy and punishable.

Still, as a conceptual matter, we reject distinguishing attempted recklessness from recklessness. We believe the correct approach is the subjective one because it avoids the troubling notion of an objective risk on which the objective view depends. Risk is an essentially epistemic concept. Risk is always relative to someone’s perspective, a perspective that is defined by possession of certain information but not other information. In law, when we say that there is a “risk” of x’s occurrence, we are using “risk” in the sense of relative frequency. That is, any given reference class will yield a relative frequency for an event’s occurrence. However, one may formulate the reference class widely or narrowly, thus changing the relative frequency. To ask what the risk is that John will be hit by lightning, we can give accurate answers if we say that there is a one in a billion chance that a person gets hit by lightning, a one in a

8See, e.g., Paul H. Robinson, “Prohibited Risks and Culpable Disregard or Inattentiveness: Challenge and Confusion in the Formulation of Risk-Creation Offenses,” 4 Theoretical Inquiries in Law 367 (2002); George P. Fletcher, The Grammar of Criminal Law: American, Comparative, and International, vol. 1: Foundations 8.4.2 (2007).

T HE ESSENCE OF CULPA BILI T Y

29

million chance that a person gets hit by lightning in the area in which John lives, a one in a thousand chance that a person gets hit by lightning on a golf course in the area where John lives, a one in one hundred chance that a person gets hit by lightning on a golf course during a rain storm in the area where John lives, and a one in one chance that John got hit by lightning on a golf course during a rain storm this Tuesday in the area where John lives. All of these probability assessments are correct within their given reference classes. In contrast, with full information, there is no need to resort to a probability. For God – who possesses complete information about everything – risk does not exist. For God, all events have a probability of either one or zero (leaving aside quantum events).9 So even though Albert was playing golf with John – and thus all the relative frequency accounts (except the last one) applied – in actuality Albert did not get hit by lighting, and thus, though he was “risked” to different degrees depending upon the reference class, there was no harm at all.10

Thus, the “objective” approach creates the following quandary. We must be able to assign a risk to an activity that is different from both whatever risk the actor perceives and the risk God perceives (one or zero). Thus, this approach requires that we construct an artificial perspective containing some but not all information. There is obviously an indefinite number of such possible perspectives, each one generating a different risk. Depending on how narrowly or widely one defines the reference class, the relative frequency will change. But there is simply no nonarbitrary way for us to select among reference classes. Nor does it make any sense to us why the culpability of an actor should hinge not upon what the actor knows, or what God would know, but upon what some other individual’s perspective might be. One who drops a bowling ball from the top of a building to measure the force of gravity for himself, and who believes there are people below whom he is

9 See Robert E. Goodin and Frank Jackson, “Freedom From Fear,” 35 Phil. & Pub. Aff. 249, 256 (2007). We use the discussion of “God” as a shorthand method for illustrating that with complete information, risks disappear. T here are only harms and misses. Thus, our discussion should not be read as entailing any sort of position on determinism and the like.

10For further argument as to why risks are not themselves harms, see Stephen R. Perry, “Risk, Harm, and Responsibility,” in Philosophical Foundations of Tort Law (David G. Owen, ed., 1995).

30	T HE CULPA BLE CHOICE

putting in extreme danger, is reckless. This is true despite the fact that his companion believes the risk is greater than he does; the building’s doorman would have estimated the risk to be slightly lower; and a window washer, with a better view below, knows that there are very few people below so that dropping the ball is unlikely to injure anyone. It simply makes no sense to allow the actor’s liability to hinge not upon what he knows, or God knows, but upon the arbitrary selection among the friend, doorman, and window washer for the correct perspective for assessing “objective risk.”

Or, to take a different example, suppose A and B are each driving cars on a two-lane highway and are approaching a blind curve. They both believe that if they stay on their side of the highway, the probability of causing injury or death when they go around the curve is very small. Likewise, they both believe that if they veer over into the left lane when rounding the curve, the probability of causing injury or death (by striking an oncoming vehicle) is quite high. Moreover, actuarial tables support their beliefs – very often an oncoming car will be entering the curve from the opposite direction, but rarely will there be a stalled vehicle or pedestrian in the right hand lane. Nonetheless, C, a bystander situated at a good vantage point, can see that this time there are no vehicles approaching from the opposite direction, but there is a small child just around the curve in the right lane. Thus, C estimates the risk of injury or death of veering into the left lane to be zero but the risk of remaining in the right lane to be virtually one. If A were to remain in the right lane, we would not deem him culpable. On the other hand, if B were to veer into the left lane, we would deem him reckless – not attempting to be reckless, but reckless full stop. Yet, from C’s perspective, A would be creating a huge risk, and B none. If any risk is “objective” then, it is the risk C assesses. Yet C’s assessment, like God’s, should be immaterial.11

11T here are other examples to illustrate our view that risk is a perspectival, relative-to- information-possessed, epistemic notion, not an ontic one (if we leave quantum physics, which is irrelevant to our concerns, aside). Consider a lottery that sells 1 million tickets for $1 each and pays out $1 million to the holder of the winning ticket. It should not matter whether the lottery officials have drawn the winning number before the lottery tickets are sold or draw it afterward, so long as they keep the number secret and do not trade on their knowledge if they draw it before the tickets are sold. So suppose they do draw the winning number before selling the tickets, and the winner will be the buyer of ticket 1436. The buyer of 1436 will, at the time of the purchase, assess the value of the ticket to be $1,