Group agency and legal proof; or, why the jury is an 'it'.

• Author: Pardo, Michael S.
• Position: IV. Theoretical Implications: Evidence Theory and Paradoxes of Proof through Conclusion, with footnotes, p. 1824-1858

4. THEORETICAL IMPLICATIONS: EVIDENCE THEORY AND PARADOXES OF PROOF

   Juries, as group epistemic agents, reach factual conclusions about whether certain propositions have been proven. Little has been said thus far about the details of the epistemic conclusions reached by jurors. In turning to this question, this Part argues that the conception of juries as group agents contains important lessons for understanding legal evidence and proof. Moreover, for reasons explored below and in Part V, the significance is not only of theoretical import--a number of practical insights follow.

   Some basics of the proof process are uncontroversial. Juries provide factual conclusions on each of the elements of a crime, civil claim, or affirmative defense. (109) These conclusions are whether the party with the burden of proof on an element has proven it to the applicable standard of proof. The three common standards of proof are: "a preponderance of the evidence" (applicable in most civil cases and for affirmative defenses in some criminal cases); (110) "beyond a reasonable doubt" (constitutionally required for the elements of crimes and applicable for some affirmative defenses); (111) and, to a lesser extent, "clear and convincing evidence" (applicable on some issues in civil and criminal cases). (112) Jurors rely on the admissible evidence to make these determinations, and trial and appellate courts review the sufficiency of the evidence and the reasonableness of jury conclusions in light of these standards. (113)

   Theoretical accounts of legal proof take the foregoing as a starting point. Two leading theoretical accounts, probabilistic and explanatory, conceptualize the nature of juror factual conclusions, and the inferences leading to them, in light of the proof standards. The first conceives of the standards as probabilistic thresholds and jury conclusions as probabilistic judgments; (114) the second conceives of the standards as explanatory thresholds and jury conclusions as explanatory judgments. (115) As explored in prior literature, the probabilistic conception faces a number of conceptual problems. (116) The most famous (or notorious) has come to be called the "conjunction paradox," or the "conjunction problem." (117) The explanatory conception, by contrast, avoids the conjunction paradox. (118) The prior debates within evidence law on these competing conceptions have largely focused on, or assumed, decisions by a single or unified decision maker, rather than examining the analytical issues that may arise from the group aspects of proof. (119)

   This Part demonstrates that group aspects give rise to additional conceptual problems for the probabilistic account and provide further vindication for the explanatory account. Some of these theoretical implications follow for reasons analogous to the original conjunction paradox, and thus my analysis first draws some lessons from that paradox before introducing three new conceptual problems that arise because of group aggregation. The discussion proceeds in three sections: Section A outlines the two conceptions of proof; Section B discusses the original conjunction paradox; and finally, Section C introduces three conceptual problems at the group level.

   Before turning to these issues, however, a brief word on the practical implications may help to center the theoretical discussions to follow. Whenever jurors or judges draw inferences from evidence and conclude whether a fact has been proven, they are presupposing some conception of what is required of applicable legal standards as well as what does or does not follow from the evidence. (120) The theoretical accounts attempt to "make[] explicit what is implicit in these practices." (121) Moreover, the accounts also allow us to examine whether and how the practices and rules fit with or deviate from the goals of the proof process specifically, and civil and criminal litigation more generally. (122) Thus, the theoretical discussions are practically important precisely because they purport to tell us something important about how the law is implemented in individual cases. (123) This is true not only with regard to trial outcomes. It also applies to cases that never make it to trial or are reversed on appeal, (124) cases that settle or are never filed or charged, (125) and "primary" (that is, non-litigation) behavior (126)--because each depends on some assessment of whether potential evidence is sufficient to warrant a legal resolution. The practical implications thus ramify throughout the law. (127)

   1. Two Conceptions of Proof

      The probabilistic and explanatory conceptions of proof each provide accounts of the nature of evidence, the standards of proof, juror inferences, and what it means for the evidence to satisfy a standard. (128) In outlining the two accounts, the discussion will focus on general aspects of each to the extent necessary to ground the analytical discussions to follow below. It will, therefore, gloss over some nuances in the accounts. (129)

      The probabilistic conception relies on standard probability theory and its axioms to conceptualize the proof process. (130) Standards of proof, under this account, express probabilistic thresholds, and jurors determine whether the probability of each element, given the evidence, surpasses the threshold. (131) Conventionally, "preponderance of the evidence" is taken to mean "proof greater than 0.5," "clear and convincing evidence" to mean "greater than 0.75," and "beyond a reasonable doubt" to mean "greater than O.9." (132) The thresholds are points on a scale between 0 and 1, in which 1 means certain truth, 0 means certain falsity, and 0.5 means complete uncertainty or indifference. (133) Thus, in a civil case under the preponderance standard, when the plaintiff's claim is made of two elements, A and B, the jury determines whether the probability of A is greater than 0.5 and whether the probability of B is greater than 0.5. (134) If the jury concludes that each element is proven beyond 0.5, then the verdict will be for the plaintiff; if the jury concludes that the probability of either element is 0.5 or below, then the verdict will be for the defendant.

      The explanatory conception relies on a theory of cognition based on the idea of "inference to the best explanation" and variations on it. (135) Under this conception, the standards of proof express explanatory thresholds, and jurors evaluate the plausibility of alternative explanations of the evidence and disputed events. (136) Under the preponderance standard, the process most closely resembles one of inference to the best available explanation. (137) Specifically, jurors evaluate which of the competing explanations better explains the evidence and events, and then determine whether that explanation includes the elements of the claim. (138) For example, in a civil case under the preponderance standard, when a plaintiff's claim involves two elements, A and B, jurors evaluate whether the better explanation includes A and B. (139) As with the probabilistic conception, higher proof standards require higher thresholds, but under this account the standards require greater explanatory thresholds. (140) Under the "clear and convincing" standard, the plaintiff's explanation must be clearly and convincingly better than the defendant's explanation, and include the elements. (141) Under the "beyond a reasonable doubt" standard, there must be a plausible explanation consistent with guilt (in other words, that includes each of the elements of the crime) and no plausible explanation consistent with innocence (that is, the explanation does not include one or more of the elements). (142)

   2. The Conjunction Paradox

      The conjunction paradox arises because under the probabilistic conception, the law appears to ignore an elementary aspect of probability theory. (143) Namely, the probability of two independent propositions is the conjunction or multiplication of each proposition. (144) Return to the example of a civil case with two elements, A and B. When a jury finds each element proven to 0.6, the plaintiff wins. This is a consequence of the fact that the standard of proof applies to each individual element. (145) The plaintiff wins even though--assuming independence among the elements--the likelihood of the plaintiff's claim is, because of the product rule, only 0.36 (0.6 x 0.6). Thus, plaintiffs win under this conception even when their claims appear not to be more likely true. (146) The problem is exacerbated by the addition of elements. (147)

      The conjunction issue generates additional perverse implications for the probabilistic account. Here are five of the more prominent problems. First, suppose a second plaintiff brings an identical claim involving elements A and B. A jury finds A is proven to 0.9 and B is proven to 0.5. Unlike the plaintiff in the first case, the plaintiff in the second case loses even though this plaintiff's claim is more likely to be true in light of the evidence (0.45 versus 0.36). Under this conception, whether parties deserve to win no longer appears to track the strength of their claims in light of the evidence--surely a problematic consequence for a system interested in accurate outcomes.

      Second, applying the probabilistic threshold (for example, "beyond 0.5") to the case as a whole also causes problems. First, it is inconsistent with the law, and thus fails as an explanatory account of the proof process. Second, its normative implications, if followed, would create other difficulties. For example, a party's proof requirements would depend on the number of formal elements. (148) The proof for each element would appear to exceed "beyond a reasonable doubt" and quickly approach certainty with only a modest number of elements.

      Third, the foregoing scenario has assumed independence among the elements, but this often will not be the case. This creates additional difficulties for the probabilistic account, because...