¶36 I am in full agreement with the majority's analysis in this case, and specifically with its conclusion that the district court's failure to conduct a rule 403 analysis of the prior bad acts evidence was prejudicial error. I agree with the majority that, in this case, the prior bad acts evidence was deployed in such a way as to make it nearly impossible for the jury to avoid drawing a propensity inference, and that the evidence should have been excluded on that basis. I write separately, as I did recently in State v. Murphy , 2019 UT App 64 , 441 P.3d 787 , to again express reservations about the manner in which the doctrine of chances (the Doctrine) is currently being used in Utah.
I
¶37 My first concern is a big-picture one: I wonder whether it could ever be appropriate for the Doctrine to be applied to admit prior acts evidence to rebut a defendant's claim that he acted in self-defense. Lane does not raise this issue, but I think it would be worthwhile for a future litigant to raise it, so that a Utah appellate court can weigh in on the question after full briefing.
¶38 As described by our supreme court, the Doctrine is "a theory of logical relevance that rests on the objective improbability of the same rare misfortune befalling one individual over and over." State v. Verde , 2012 UT 60 , ¶ 47, 296 P.3d 673 (quotation simplified), abrogated on other grounds by State v. Thornton , 2017 UT 9 , 391 P.3d 1016 ; see also State v. Lopez , 2018 UT 5 , ¶ 52, 417 P.3d 116 (stating that doctrine of chances cases "involve rare events happening with unusual frequency"). At root, the Doctrine is simply "probability reasoning." Verde , 2012 UT 60 , ¶¶ 50, 53, 296 P.3d 673 ; cf. Hopt v. People , 120 U.S. 430 , 440, 7 S.Ct. 614 , 30 L.Ed. 708 (1887) (referring to the "doctrine of chances" as a tool used to "establish a probability").
¶39 Because the Doctrine is a probability-based construct, it has been widely applied to admit prior bad acts evidence in cases in which the accused's defense is that the allegedly criminal act in question occurred by accident or random chance rather than by design. See Murphy , 2019 UT App 64 , ¶ 54, 441 P.3d 787 (Harris, J., concurring) (citing cases). 7 In such cases, the prosecution may be allowed to introduce evidence of previous incidents involving the defendant in order to demonstrate the extreme statistical improbability that the allegedly criminal act occurred solely by accident or random chance. See, e.g. , United States v. York , 933 F.2d 1343 , 1350 (7th Cir. 1991) (stating that "[t]he man who wins the lottery once is envied; the one who wins it twice is investigated"), overruled on other grounds by Wilson v. Williams , 182 F.3d 562 (7th Cir. 1999). That is, where the defendant's claim is that "the event in question was an accident," the Doctrine can apply to rebut that claim, as our supreme court explained in Verde : "Propensity inferences do not pollute this type of probability reasoning," because "[t]he question for the jury is not whether the defendant is the type of person who, for example, sets incendiary fires or murders his relatives." 2012 UT 60 , ¶ 50, 296 P.3d 673 (quotation simplified). Instead, "[t]he question is whether it is objectively likely that so many fires or deaths could be attributable to natural ca[u]ses." 8 Id. This evidence "tends to prove a relevant fact without relying on inferences from the defendant's character," and is therefore not impermissible propensity evidence. Id. ¶ 51. In the context of rebutting a claim of mistake or accident, "[i]t is that objective unlikelihood [of repeated similar misfortunes] that tends to prove" that actions were brought about by "human agency, causation, and design" rather than by accident or random chance. Id. ¶ 50 (quotations simplified).
¶40 A doctrine like this-based on probability reasoning and on the statistical unlikelihood of repeated occurrences of rare, random events-would seem to lose much of its logical coherence if applied in contexts where the underlying acts in question are not random at all, but instead are based on human volition. Applied in such contexts, it would seem to become very difficult-if not entirely impossible-to separate the permissible "probability" inference from the impermissible "propensity" inference. I explained in Murphy that I fear this problem might exist in cases in which the Doctrine is applied to admit prior bad acts for the purpose of rebutting a defendant's claim that the complaining witness is lying. See 2019 UT App 64 , ¶¶ 57-59, 441 P.3d 787 (Harris, J., concurring). I see the potential for this same problem in cases in which the Doctrine is applied to admit prior bad acts for the purpose of rebutting a claim of self-defense.
¶41 In cases like this one, in which a defendant stands accused of a violent act but claims he acted in self-defense, we may be less likely to believe the defendant's claims if presented with evidence that he has made this claim before, whether successfully or unsuccessfully. But the reason we are less likely to credit the defendant's claim in this context has little to do with probability and a lot to do with the easily drawn inference that the defendant might be the type of person who commits violent acts. The fact that he has been previously involved in violent acts is not usually something that is based on randomness or fortune (like winning the lottery or being struck by lightning). It is based on a whole host of factors, most of which involve non-random, purposeful decisions on the part of the defendant and others. Specifically, becoming involved in violent acts involves human decision-making, and a person's state of mind when he commits those acts-e.g., whether the person acted in self-defense-is also volitional rather than random.
¶42 That is, in many instances, the reasons a person is involved in incidents resulting in violent acts, and the reasons a person forms a particular mens rea while doing so, are not probability-based, and therefore I wonder about the wisdom of trying to apply a probability-based doctrine in this context. The fact that Person A is much more likely than Person B to be involved in a violent scrape and then claim self-defense would seem to have a lot more to do with propensity or with other non-random environmental factors than it does with simple mathematical probabilities. See, e.g. , Paul F. Rothstein, Intellectual Coherence in an Evidence Code , 28 Loy. L.A. L. Rev. 1259 , 1262-63 (1995) ("The essence of this probable guilt argument is that there is a disparity between the chances, or probability, that an innocent person would be charged so many times and the chances, or probability, that a guilty person would be charged so many times. If there is such a disparity, however, it is only because a guilty person would have the propensity to repeat the crime. If it were not for the propensity to repeat, the chances, or the probability, that an innocent person and a guilty person would be charged repeatedly would be identical. Hence, the argument hinges on propensity and runs afoul of the first sentence of Rule 404(b)."). At a minimum, it seems that the variables involved in running a metaphorical probability calculation in this context may be too numerous to make the calculation meaningful in any given case.
¶43 In my view, even assuming the soundness of Verde 's underlying probability logic, see supra ¶ 39 note 8, and even assuming there may exist scenarios in which that logic could be usefully applied in a self-defense (or other volitional) case, the entire exercise is a nonstarter unless two threshold conditions can be met. First, the party asking the court to admit prior bad acts evidence pursuant to the Doctrine should be able to clearly articulate what the event of "rare misfortune" is that triggers the Doctrine's application. See Verde , 2012 UT 60 , ¶ 47, 296 P.3d 673 . Where the Doctrine is applied to rebut a claim of mistake or accident, this is usually easily accomplished: the event of rare misfortune is, say, the death of a bride in a bathtub, or the mistaken taking of a horse. See id. ¶¶ 48-49. In the self-defense context (as in the fabrication context, see Murphy , 2019 UT App 64 , ¶¶ 57-59, 441 P.3d 787 (Harris, J., concurring)), it is often difficult to articulate what that event is, as illustrated in this case. Is the event of rare misfortune that Lane was previously involved in fights? Is it that Lane was previously involved in fights for which he claimed that he acted in self-defense? Or is it that Lane was previously involved in fights in which he employed a knife? I cannot tell, and (even upon questioning at oral argument) neither can the State. None of these options involve random events of chance. As in this case, if it is difficult to clearly identify the event of "rare misfortune," it raises the likelihood that the evidence of prior acts is not coming in for permissible probability purposes but, instead, is coming in for impermissible propensity purposes. Moreover, without clear identification of the event of "rare misfortune," it becomes difficult to determine whether the "four foundational requirements," which are prerequisites to the application of the Doctrine, have been satisfied. See Verde , 2012 UT 60 , ¶¶ 57-61, 296 P.3d 673 (listing materiality, similarity, independence, and frequency as the "four foundational requirements" of the Doctrine).
¶44 Second, the party asking the court to admit prior bad acts evidence pursuant to the Doctrine should be able to clearly articulate both (a) the purposes for which the evidence can permissibly be used and (b) the purposes for which the evidence cannot permissibly be used. If these purposes cannot be articulated in a way that a lay juror can readily understand, that is a good clue that the Doctrine is being misapplied. Again, this case is a good example. The jury was instructed that it could "consider [the prior bad acts] evidence, if at all, for the limited purpose of self-defense," but that the "evidence was not admitted to prove a character trait of the defendant or to show that he acted in a manner consistent with such a trait." I confess that I do not know what this instruction means. No mention at all is made of any probability-based inference that might be permissibly drawn with regard to evidence properly admitted pursuant to the Doctrine. No meaningful guidance is given regarding the purposes for which the evidence may, and may not, be used. I cannot imagine lay jurors having any idea what to make of an instruction like this, and if the jury is not clearly instructed, the risk of jurors resorting to impermissible propensity inferences is too great.
¶45 All of which leads me not only to conclude that the Doctrine was misapplied in this case, but also to wonder whether the Doctrine could ever be properly applied in a self-defense context. No Utah appellate court has yet held that application of the Doctrine to cases in which the defendant claims self-defense is proper. 9 Some other courts have applied the Doctrine to allow prior acts evidence in this context, see, e.g. , State v. Monroe , 364 So. 2d 570 , 571-73 (La. 1978), but those cases are rare, and it is therefore far from established that the Doctrine applies in self-defense cases. I urge parties in future cases to raise and fully brief this issue, instead of-as the parties did here-simply assuming that the Doctrine applies in this context.
II
¶46 My second set of concerns has to do with the manner in which the Doctrine was specifically applied in this case. That is, assuming that the Doctrine could be meaningfully applied to admit relevant, non-character prior acts evidence in the self-defense context, the Doctrine was misapplied in this case in several material ways.
¶47 First, as the majority ably describes, the district court did not conduct a separate rule 403 analysis, a step that is " 'essential to preserve the integrity of rule 404(b).' " See supra ¶ 20 (quoting Verde , 2012 UT 60 , ¶ 18, 296 P.3d 673 ). Even if a court concludes that, under governing case law, the Doctrine can logically apply, and even if it concludes that the Doctrine's "four foundational requirements" for application are met, see Verde , 2012 UT 60 , ¶ 57, 296 P.3d 673 , the court still must analyze the evidence under rule 403 to ascertain whether the probative value of the admissible part 10 of the evidence is substantially outweighed by the danger of unfair prejudice, including the danger of the jury drawing an impermissible propensity inference. The district court failed to take this important step.
¶48 Second, as I have already mentioned, the instruction given to the jury was inadequate, and did not meaningfully assist the jury in navigating its way through a logical and metaphysical minefield. "A complete, properly worded limiting instruction has two prongs. The negative prong forbids the jury from using the evidence for the verboten purpose. In contrast, the affirmative prong explains how the jury is permitted to reason about the evidence." Edward J. Imwinkelried, Criminal Minds: The Need to Refine the Application of the Doctrine of Objective Chances as a Justification for Introducing Uncharged Misconduct Evidence to Prove Intent , 45 Hofstra L. Rev. 851 , 873 (2017). The instruction given in this case was conclusory, and informed the jury that it could not draw a character inference but could use the evidence for "self-defense." This is precisely the sort of instruction that commentators have rightly criticized. See id. at 873-74, 876 (offering as an example of an "inadequate" instruction one where, "[a]fter stating the negative prong of the instruction, in the affirmative prong the judge ... give[s] the jury only the guidance that they may use the evidence for the purpose of proving 'intent,' "
and noting that this sort of instruction "can lead the jury into improper character reasoning"). Assuming that, on the facts of this case, it were possible to articulate purposes for which the evidence could and could not be used, those purposes needed to have been spelled out in much more detail than they were.
¶49 Third, I am concerned about the manner in which the district court analyzed the "frequency" factor. See Verde , 2012 UT 60 , ¶ 61, 296 P.3d 673 . The point of this factor is to ensure that the event of "rare misfortune" in question has been visited upon the defendant "more frequently than the typical person." Id. ¶¶ 47, 61 (quotation simplified). Assuming that one can pinpoint what the event of rare misfortune is in this instance, and that one can meaningfully apply probability (rather than propensity) reasoning to a situation involving several levels of human volition, our case law then requires the court to compare this defendant to a "typical person" to ascertain whether the event occurred to the defendant with greater frequency. In this case, the court's complete analysis on this point was as follows: "Here, Defendant has been involved with three serious assaults in four years. Even given his chronic homelessness and the higher frequency of assault surrounding shelters, the rate of Defendant's involvement in these assaults is not mere accident." I find this analysis lacking. The court did not take any evidence to establish the profile of a "typical" resident of that part of Salt Lake City, or any evidence intended to establish a baseline regarding the number of physical altercations per year in which such a resident might typically be involved. Under these circumstances, I see no reasoned basis for the court's intuition-level conclusion that a person living in that part of the city becoming involved in one fight every fifteen months is necessarily "frequent." Bound up in that analysis are various assumptions by the court-arrived at without evidence-of what living conditions are like for homeless citizens of Salt Lake City. This is an instance where the court, in my view, needed to take additional evidence-from experts, if necessary-to arrive at a sound conclusion about whether the number of assaults in which Lane was involved was atypical for a resident of that part of town.
III
¶50 But I question whether our courts should even be asked to engage in inquiries like that, given the bigger problems I see with the application of the Doctrine to admit prior acts evidence in cases in which a defendant claims that he acted in self-defense. Because of my various concerns about the district court's admission, pursuant to the Doctrine, of Lane's prior assaults, I share the majority's view that Lane was not afforded a fair trial, and therefore I concur in the majority's disposition. I also urge litigants in future cases to raise and brief issues they might see with application of the Doctrine, in this or other contexts, in order to enable the Doctrine's application in Utah to be reexamined in an appropriate case.
The defense of mistake or accident can be raised with regard to either actus reus or mens rea . In the famous "Brides in the Bath" case, the defense was that there had been no actus reus , and that the three brides had each died by accident while bathing. See State v. Verde , 2012 UT 60 , ¶ 49 n.20, 296 P.3d 673 (citing Rex v. Smith , 11 Crim. App. 229, 84 L.J.K.B. 2153 (1915)), abrogated on other grounds by State v. Thornton , 2017 UT 9 , 391 P.3d 1016 . In the case of Dean Wigmore's famous hypothetical about a hunter who shot at his companion three times, the hunter necessarily concedes the existence of an actus reus , but defends the case on the grounds that he did not intend to shoot. See 2 John Henry Wigmore, Evidence in Trials at Common Law § 302, at 241 (James H. Chadbourn ed., 1979). In these examples, however, the underlying defense is the same: it was a mistake or an accident.
It bears noting that the underpinnings of even this logic have been credibly (albeit impliedly, without mentioning or citing to Verde ) called into question. See, e.g. , State v. Vuley , 2013 VT 9 , ¶¶ 19-22, 193 Vt. 622 , 70 A.3d 940 (holding that the Doctrine cannot be used, even in its probabilistic sense, when applied to "human action" rather than to truly random events, because "[i]nferring from the implausibility of all occurrences being accidents that any particular occurrence was not an accident necessarily involves reasoning based on propensity," and that "it would be an inference based on propensity to say that, because a man has intentionally killed a wife, he is therefore more likely to have intentionally killed this wife"). For the purposes of this opinion, however, I assume that the logic of paragraphs 49-51 of the Verde opinion is sound (even though it may not be), and point out additional flaws in Verde 's rickety structure that I believe may exist even if its underlying logic is sound.
The matter was discussed at some length in State v. Labrum , 2014 UT App 5 , 318 P.3d 1151 , but this court ultimately stopped short of deciding whether the Doctrine could be employed for this purpose because it determined that the prior bad acts evidence was admissible on another ground. Id. ¶¶ 29-31. To date, our supreme court has not addressed the issue, although it has generally espoused a remarkably broad view of the Doctrine's applicability, holding that it applies in other contexts also involving non-random volitional acts, including to rebut defenses of fabrication, see Verde , 2012 UT 60 , 296 P.3d 673 , and consent, see State v. Lowther , 2017 UT 34 , ¶ 25, 398 P.3d 1032 . For the reasons set forth herein and elsewhere, see State v. Murphy , 2019 UT App 64 , ¶¶ 45-65, 441 P.3d 787 (Harris, J., concurring), my view is that these decisions may merit reexamination.
Propensity evidence has great probative value, which is in part why our rules of evidence ban it. See David P. Leonard, The New Wigmore: A Treatise on Evidence: Evidence of Other Misconduct and Similar Events § 1.2, at 6-7 (2009) (stating that propensity evidence is excluded "not because it has no appreciable probative value, but because it has too much"). In conducting an appropriate rule 403 balancing in this context, the "probative" side of the equation should include only the value of any admissible probability inferences, and should not include the value of any impermissible propensity inferences (which should be assessed on the "prejudice" side of the equation).