dissenting: Most of the time, the words we choose have some imprecision in conveying our intended meaning (assuming we’ve thought enough about the matter to have a clear intended meaning).
An oft-used example in law schools is a statute forbidding vehicles in a park. Certainly cars can’t be driven there, but what about a golf cart? Or a Segway? It depends upon what vehicle means.
This sort of imprecision is especially important when we give juries the instructions they must follow. The instructions must be sufficiently precise that they accurately tell the jurors the key concepts that must guide their deliberations. Anything less can erode a defendant’s right to be found guilty only when guilt of all elements of the offense has been proved beyond a reasonable doubt. In addition, an imprecise definition of an element of the offense could lead to convicting a defendant because the jury understood the definition to require less culpability than the law actually requires. If a defendant were convicted after such a misunderstanding, key principles of our criminal-justice system are undermined— including the principle that all of us should have fair notice of what’s forbidden by our criminal laws.
This is such a case: the defendant may well have been convicted under a standard requiring less culpability than the law requires.
Let’s start with a precise discussion about the required elements of child endangerment. Are we talking about merely being negligent? Negligence results in civil liability for damages and requires only a failure to do something that a reasonably prudent person would do. Or does criminal liability for child endangerment require something more?
*26The statute, K.S.A. 21-3608(a), requires a defendant to have intentionally and unreasonably put a child in a situation in which the child’s life, body, or health “may be injured.” (Emphasis added.) But a statute defining criminal conduct is unconstitutional if it is too vague. Our Supreme Court held in State v. Fisher, 230 Kan. 192, 195, 631 P.2d 239 (1981), that the term “may” in the child-endangerment statute means “something more than a faint or remote possibility; it means that there is a reasonable probability, a likelihood that harm to the child will result.” (Emphasis added.) The court noted that “may” can have many different meanings and ultimately chose a relatively limited definition so that the statute would not be too vague. Our Supreme Court thus interpreted “may” in the child-endangerment statute to mean “a reasonable probability, a likelihood that harm to the child will result.”
Its use of the term “likelihood” is significant. The Fisher court relied on a Colorado Supreme Court decision that had interpreted “may” to mean “reasonable probability” in a similar Colorado statute. 230 Kan. at 194-95 (citing People v. Hoehl, 193 Colo. 557, 568 P.2d 484 [1977]). But “reasonable probability” is itself ambiguous. For example, does something with a “reasonable probability” of occurring happen more — or less — frequently than something with just a “probability” of occurring? Thus, the Kansas Supreme Court combined the phrasing to add some additional precision: reasonable probability, a likelihood that harm toill result.
Accordingly, in State v. Sharp, 28 Kan. App. 2d 128, 13 P.3d 29 (2000), our court reversed a child-endangerment conviction when the jury instruction simply substituted “might” for “may.” The panel in Sharp concluded that there was a real possibility that the jury would not have convicted the defendant had it been told that the State must show not merely that harm to the child “might” result but instead had been told that conviction required that the defendant’s acts caused “a reasonable probability, a likelihood that harm to the child will result.” 28 Kan. App. 2d at 135.
So what does this phrase actually mean? One dictionary notes that “likely” can mean “[possessing or displaying the qualities or characteristics that make something probable,” but also that it can mean “[wjithin the realm of credibility; plausible.” American Her*27itage Dictionaiy 1014 (4th ed. 2006). Another dictionaiy says something is “likely” when it is “probable” or has “a better chance of existing or occurring than not.” Webster’s Third New International Dictionary 1310 (1993). “Probability” can be a neutral term: we can say either “the probability of success is zero” or “the probability of success is 100%.” The American Heritage Dictionary’s first definition of “probable” is “[l]ikely to happen or to be true,” as in “War seemed probable in 1938,” or “The home team, far ahead, is the probable winner.” But the same dictionary also defines “probable” as “[l]ikely but uncertain; plausible.” American Heritage Dictionaiy 1397.
There’s quite a bit of play, then — or potential imprecision — in the meaning of “a reasonable probability, a likelihood.” But we have at least four factors that lead me to conclude that the phrase means “more likely than not" when we interpret this criminal statute. First, when the Kansas Supreme Court took the definition of “a reasonable probability” from a Colorado case, it added “a likelihood” as a synonym for the meaning the court was identifying. Fisher, 230 Kan. at 195. Likely (or likelihood) and probable (or probability) are often defined with respect to each other, and each has an indeterminate meaning. But the combination of likely and probable suggests that we should be using ordinary, prototypical meanings, not alternate definitions. It would be odd for the court to have referenced less common alternative meanings for each of the two terms in defining this concept. The first definition for probable in the American Heritage Dictionaiy is “likely to happen,” combined with two examples that demonstrated a high probability that the event would occur. When used together, the terms probability and likelihood suggest that the event must be more likely than not to occur. See State v. Tice, 686 N.W.2d 351, 353-55 (Minn. App. 2004) (interpreting child-endangerment statute that malees it a crime to place child “in a situation likely to substantially harm” that child to mean that the harm must be more likely than not to result from the act based on use of the term “likely”). Second, the limitation on the meaning of the statutory term “may” came in Fisher on a void-for-vagueness challenge. The key test in such cases is that the statute must give “adequate warning” as to *28what conduct is illegal. State v. Richardson, 289 Kan. 118, 124, 209 P.3d 696 (2009). Choosing a nonprototypical meaning for these terms would be inconsistent with that goal. Third, the rule of lenity also calls for narrow interpretation of criminal statutes so that citizens have fair notice of the conduct that may subject them to imprisonment or other punishment. State v. Boyer, 289 Kan. 108, 113, 115, 209 P.3d 705 (2009); State v. Reese, 42 Kan. App. 2d 388, 390, 212 P.3d 260 (2009). The values that stand behind that rule would be undermined if we chose less common meanings that expanded potential liability. Last, the discussion in Fisher ends with “a likelihood that harm . . . will result.” (Emphasis added.) 230 Kan. at 195. That too suggests that we are dealing with an event that more likely than not actually would occur due to the action taken.
In sum, the Kansas Supreme Court has interpreted the child-endangerment statute to require that the defendant’s conduct cause “a reasonable probability, a likelihood that harm . . . will result.” 230 Kan. at 195. The terms “a reasonable probability, a likelihood,” and “will result” should be understood in this context to require that harm to the child was more likely to occur than not from the defendant’s conduct.
But here is the critical question in this case: Would a jury have understood that the State had to prove that the situation in which the defendant placed the infant made it more likely than not that she would be injured? The judge’s instruction merely told the jury that the defendant’s conduct had to have “placed [the child] in a situation in which there was a reasonable probability that [the child’s] life, body or health would be injured or endangered.” That left the jury to guess about what “reasonable probability” meant. The jury could easily have understood the imprecision of that term to allow conviction even when the probability of harm was well less than even odds.
In some circumstances, a synonym for “probability” is “chance.” Jurors could easily have understood that they could convict the defendant if there was a “reasonable chance” or a “decent chance” that harm would result from putting an infant in a car seat with only the top strap fastened. In calculating how to interpret the *29single term “reasonable chance,” the jurors would have applied a different standard than the one our Supreme Court supplied in Fisher.
Perhaps some would argue that I have read too much into the words our Supreme Court used in Fisher, but this was our Supreme Court’s unanimous and authoritative interpretation of the statute. This statute, moreover, was twice reenacted with this same language after Fisher and before Cummings’ acts led to this infant’s death. See L. 1992, ch. 298, sec. 36; L. 1993, ch. 291, sec. 59. When the legislature reenacts a law after our Supreme Court has interpreted it, it has long been understood that the legislature implicitly accepts the court’s construction of that statute. Colver v. McInturff, 112 Kan. 604, 608, 212 P. 908 (1923); see 2B Singer and Singer, Sutherland Statutes and Statutory Construction § 49.9 (7th ed. 2008). Moreover, as I have already explained, the natural reading of the phrases the court used in Fisher is that the probability of the harm occurring must bé more likely than not. As the Fisher court said, the statute’s purpose “is to protect children, and to prevent their being placed where it is reasonably certain that injury will result,” a situation “ In which injury appears likely and probable.’ ” 230 Kan. at 199.
Is this much ado about nothing? Would a better instruction have resulted in a different verdict? After all, because the defendant made no objection at trial to the jury instructions, we must affirm unless we are firmly convinced there is a real possibility that the jury would have rendered a different verdict had it received a proper instruction. See State v. Colston, 290 Kan. 952, 976, 235 P.3d 1234 (2010). Given the way this case was presented to the jury, I am firmly convinced that there is a real possibility of a different verdict here.
The prosecutor’s closing argument essentially turned the question of probability into one of foreseeability:
“Was it a reasonable probability that that could happen? .... [R]ead the warning labels that are clearly displayed on that car seat. R is specific about the danger of strangulation. It is reasonably foreseeable[;] you could believe if they put a warning label on the seat that it could happen. Clearly, somebody foresaw that it could happen.”
*30This is a tragic case. The jury found that the defendant’s act of placing this 13-month-old girl into a car seat for a nap — rather than a crib — was unreasonable. But doing something unreasonable is only one element of this crime: the act must also be one that causes a reasonable probability, a likelihood that harm to the child will result. Yet Cummings may well have believed that the probability of the child’s death was anything but likely; she testified that she had put K.H. and her own daughter to sleep in that car seat on many other occasions without it resulting in a child’s death or injury. Based on her testimony, the jury could well have concluded that there was only a small chance that harm to the child would result but that — because of the seriousness of the risk — there was a “reasonable probability” or a “decent chance” of harm. But the act must be both unreasonable and one that causes a likelihood -that harm to the child will result for the criminal offense of child endangerment to have been committed.
If only the first of those two tests — unreasonable conduct — is met, there still is civil liability for negligence. When we cause injury to another by failing to act as a reasonably prudent person would and the harm is reasonably foreseeable, there’s civil liability for negligence. That might well occur in a great many situations involving children. But only when there’s a likelihood that harm would result does the negligent act become the crime of child endangerment.
In my view, the jury should be instructed that something is a reasonable probability or a likelihood only if the chance of it occurring is more likely than not. The majority suggests that because the defendant has only asked for the specific words our Supreme Court used in Fisher — “a reasonable probability,” “a likelihood,” and “something more than a faint or remote possibility” — we need not consider whether these terms mean more likely than not. But the defendant argues not only that the jury should have been told that “a likelihood” was required but also that the jury should have been told that “a likelihood that harm to the child will result.” In combination, a reasonable juror would understand that the language urged by the defendant on appeal would mean that the chance of harm must have been more likely than not. As I have *31explained, I think that’s the most reasonable interpretation of the words used in Fisher, which contains the same words that the defendant claims should have been used. Accordingly, the defendant’s argument necessarily contends that the verdict would have been different had the jury understood that the act had to be more likely than not to cause the harm that occurred.
Whether we consider daycare providers or parents, this essential aspect of the crime of child endangerment must not be lessened. At some time in a child’s life, just about every parent does something that — in hindsight — the parent realizes wasn’t reasonably prudent in terms of the child’s safety. Perhaps you let your child ride a bike without a helmet, or gave up on the argument with your 17-year-old son about putting on his seatbelt. Or you let your young child play with a sharp stick or even play in a normally calm creek. Most of the time, no one is injured. But the potential risks can be serious.
A jury could easily think that there’s a “reasonable chance” or a “decent chance” of injury in a high-risk, low-probability situation like the ones I’ve just described. Arguably, Cummings’ case is similar. And with the hindsight bias we all have when we know how badly things turned out in a case like the one now before us, jurors can be tempted to interpret the words liberally toward finding that a crime occurred. Otherwise, since they are not told that there can also be civil liability, they may worry that no one will be responsible for a tragic death. But under Fisher, we should not put people in prison for child endangerment unless it’s more likely than not that the child will be hurt by the action taken.
Before closing, I must address three additional points. First, the majority has relied in part on State v. Daniels, 278 Kan. 53, 72-73, 91 P.3d 1147, cert. denied 543 U.S. 982 (2004). I do not find that case useful in analyzing the claim Cummings has made here. The issue before the court in Daniels was whether the evidence was sufficient to convict the defendant of child endangerment, not whether the jury instruction was proper. In the court’s one-paragraph resolution of the issue of evidentiary sufficiency, the court referenced the change in the pattern jury instruction following Sharp. The issue in Daniels was not whether that instruction was *32erroneous. Second, I should note that K.S.A. 21-3608(a) has two words — “or endangered” — that I have not yet discussed. As written, the statute malees it a crime to intentionally or unreasonably put a child in a situation in which the child “may be injured or endangered.” But in construing the statute so as to survive a void-for-vagueness constitutional challenge, the Fisher court essentially read the words “or endangered” out of the statute: “The phrase ‘or endangered’ adds little, if anything to the statute; if a child is endangered, it may be injured; it is the likelihood of injury against which the statute speaks.” 230 Kan. at 199-200. Third, the 2010 legislature recodified the entire Kansas Criminal Code, and the new statute on child endangerment makes it illegal to unreasonably and intentionally put a child in a situation in which the child “may be endangered.” L. 2010, ch. 136, sec. 78. In the new code, the legislature used endangered, not “injured or endangered.” Whether the legislature sought to negate the Fisher interpretation of the earlier Kansas statute is not at issue in Cummings’ case; her acts occurred well before the new statute was enacted.
Was it more likely than not that an infant strapped into this infant carrier with only the top strap fastened would be harmed? Or, to use the words in Fisher, was this a situation “in which injury appears likely and probable,” one in which “it is reasonably certain that injuiy will result”? 230 Kan. at 199. Those questions are for the jury to decide based on a proper instruction. I would reverse die district court’s judgment and remand the case for a new trial with a proper jury instruction.