1. The literature on coherentism in law is extensive. See, among others, Ronald Dworkin, Law's Empire (1986); MacCormick, Neil, Coherence in Legal Justification, in Theorie der Normen [Theory of Norms] 37 (Krawietz, Werneret al. ed., 1984)Google Scholar; Neil MacCormick, Legal Reasoning and Legal Theory (corrected ed. 1994); Bernard Jackson, Law, Fact, and Narrative Coherence (1988); Aulis Aarnio et al., On Coherence Theory of Law (1998); Alexander Peczenik, On Law and Reason (1989); Alexy, Robert & Peczenik, Alexander, The Concept of Coherence and Its Relevance for Discursive Rationality, 3 Ratio Iuris 130 (1990)CrossRefGoogle Scholar; Hage, Jaap, Law and Coherence, 17 Ratio Iuris 85 (2004)Google Scholar. For a critical review of the main coherence theories of law and adjudication, see Amalia Amaya, The Tapestry of Reason: An Inquiry into the Nature of Coherence and its Role in Legal Argument (forthcoming 2013).

2. For holistic approaches to evidence evaluation, see Abu-Hareira, M., An Early Conception of Judicial Fact-Finding, Jurid. Rev. 79 (1986)Google Scholar; Allen, Ronald J., The Nature of Juridical Proof, 13 Cardozo L. Rev. 373 (1991)Google Scholar; Allen, Ronald J., Factual Ambiguity and a Theory of Evidence, 88 NW. U. L. Rev. 604 (1994)Google Scholar; and Pardo, Michael, Juridical Proof, Evidence, and Pragmatic Meaning: Towards Evidentiary Holism, 95 NW. U. L. Rev. 399 (2000–2001)Google Scholar. For narrative models of legal proof, see MacCormick, supra note 1; MacCormick, Neil, The Coherence of a Case and the Reasonableness of Doubt, 2 Liverpool L. Rev. 45 (1980)CrossRefGoogle Scholar; Neil MacCormick, Rhetoric and the Rule of Law (2005), at 214–237; Jackson, supra note 1; and Bert C. van Roermund, Law, Narrative and Reality (1997). Narratives also play a critical role in some psychological models of legal fact-finding. See Hastie, Reid & Pennington, Nancy, A Cognitive Theory of Juror Decision-Making: The Story Model, 13 Cardozo L. Rev. 519 (1991)Google Scholar; and Willem A. Wagenaar, Peter J. van Koppen & Hans F. M. Crombag, Anchored Narratives: The Psychology of Criminal Evidence (1993).

3. Prominent contemporary coherence theories of epistemic justification include Laurence BonJour, The Structure of Empirical Knowledge (1985); Keith Lehrer, Theory of Knowledge (2000); Davidson, Donald, A Coherence Theory of Truth and Knowledge, in Subjective, Intersubjective, Objective 137 (2001)CrossRefGoogle Scholar. Coherence is also claimed to be pivotal in moral rather than epistemic justification; see John Rawls, A Theory of Justice (1999); Alan H. Goldman, Moral Knowledge (1988); and Michael DePaul, Balance and Refinement: Beyond Coherence Methods in Ethics (1993). For coherentist accounts of the justification of practical statements, see Henry Richardson, Reasoning about Final Ends (1994); Susan Hurley, Natural Reasons: Personality and Polity (1989); and Thagard, Paul & Millgram, Elijah, Deliberative Coherence, 108 Synthese 63 (1996)Google Scholar.

4. On the psychological plausibility of coherence models of evidential reasoning in law, see Hastie & Pennington, supra note 2; Wagenaar, van Koppen & Crombag, supra note 2; Simon, Dan, A Third View of the Black Box: Cognitive Coherence in Legal Decision-Making, 71 U. Chi. L. Rev. 511 (2004)Google Scholar. For a summary of the main empirical findings to the effect that Bayesianism fails to describe accurately the inferences of legal decision makers, see Sacks, Michael J. & Thompson, William C., Assessing Evidence: Proving Facts, in Handbook of Psychology in Legal Contexts 338–339 (Carson, David & Bull, Ray eds., 2d ed. 2003)Google Scholar; Bradley D. McAuliff et al., Jury Decision-Making in the Twenty-First Century: Confronting Science and Technology in Court, 306–307. For a critique of Bayesian epistemology on the grounds that it involves a high degree of idealization, see Richard Foley, Working without a Net: A Study of Egocentric Epistemology 41 (1993); and Alvin Plantinga, Warrant: The Current Debate (1993), at 137–162. Theories of legal reasoning that enjoy psychological plausibility have an advantage over those that idealize away from normal human cognitive capacities insofar as they are better situated to regulate and improve legal practice. The main tenet of naturalized approaches to legal epistemology is that issues about how legal fact finders may reason, given their cognitive resources, are relevant to answering questions about how they ought to reason. On naturalized epistemology and evidence law, see infra Section VIII.

5. See Haack, Susan, A Founherentist Theory of Empirical Justification, in The Theory of Knowledge (Pojman, L. ed., 1999)Google Scholar.

6. See Olsson, Erik, Making Beliefs Coherent, 7 J. Logic, Language & Info. 143 (1998)CrossRefGoogle Scholar; Olsson, Erik, Cohering With, 50 Erkenntnis 273 (1999)CrossRefGoogle Scholar.

7. On the relation between coherence and truth, see Section VII below.

8. For a proposal as to how one may accommodate the role that emotions play in legal reasoning within a coherentist approach to evidential reasoning in law, see Paul Thagard, Hot Thought: Mechanisms and Applications of Emotional Cognition (2006), at 135–156. That coherence approaches to evidential reasoning in law may be easily modified to make room for emotions does not imply that the law's treatment of emotions be itself coherent. For an argument to the effect that conflicting conceptions of emotion are at work in (criminal) law, see Kahan, Dan M. & Nussbaum, Martha C., Two Conceptions of Emotion in Criminal Law, 96 Colum. L. Rev. 269 (1996)CrossRefGoogle Scholar.

9. See Amaya, Amalia, Legal Justification by Optimal Coherence, 24 Ratio Iuris 304 (2011)CrossRefGoogle Scholar.

10. It is important to clarify that within the category of “fact finder” I mean to include all legal agents who are involved in the process of legal fact-finding and not exclusively those in charge of fact-finding at trials.

11. See Thagard, Paul & Verbeurgt, Karsten, Coherence as Constraint Satisfaction, 22 Cognitive Sci. 12 (1998)CrossRefGoogle Scholar. See also Paul Thagard, Coherence in Thought and Action (2000), ch. 2.

12. Thagard has slightly modified the principles of explanatory coherence since their original formulation in Thagard, Paul, Explanatory Coherence, 12 Behav. & Brain Sci. 479 (1989)Google Scholar. Here I present them as Thagard states them in Thagard, Coherence, supra note 11, at 43.

13. Coherence, unlike logical consistency, is a matter of degree. This seems to be, as Douven puts it, one of our most basic intuitions about the notion of coherence. See Douven, Igor & Meijs, Wouter, Measuring Coherence, 156 Synthese 405 (2007), at 406CrossRefGoogle Scholar.

14. Three main kinds of coherence problems may be distinguished: pure, foundational, and discriminating coherence problems. A “pure” coherence problem does not favor the acceptance of any particular set of elements. A “foundational” one selects a set of elements as self-justified. A “discriminating” coherence problem favors a set of elements, but their acceptance still depends on their coherence with other elements. Even though the abstract definition of coherence in terms of constraint satisfaction is nondiscriminating, in the sense that all elements are treated equally, Principle E4 makes Thagard's notion of explanatory coherence discriminating. This principle allows that some kinds of information be treated more seriously than others, i.e., propositions describing observations and experimental results, while remaining within a coherentist framework, for the acceptance of the favored elements is not guaranteed, but they may be rejected if they fail to cohere with the entire set of propositions. See Thagard, Coherence, supra note 11, at 70–72. On the classification of coherence problems, see Thagard, Paul, Eliasmith, Chris, Rusnock, Paul & Shelley, Cameron, Knowledge and Coherence, in Common Sense, Reasoning, and Rationality 104 (Elio, R. ed., 2002)CrossRefGoogle Scholar.

15. R v. Cannings [2004] EWCA (Crim) 1.

16. See Grimm, Stephen R., On Intellectualism in Epistemology, 120 Mind 705 (2011)CrossRefGoogle Scholar.

17. See Christopher Hookway, Scepticism 139 (1990); Fantl, Jeremy & McGrath, Matthew, Evidence, Pragmatics, and Justification, 111 Phil. Rev. 67 (2002)CrossRefGoogle Scholar; Fantl, Jeremy & McGrath, Matthew, Pragmatic Encroachment, in The Routledge Companion to Epistemology 558 (Bernecker, Sven & Pritchard, Duncan eds., 2011)Google Scholar. For views according to which whether a subject knows something to be the case depends on his practical situation, see Fantl, Jeremy & McGrath, Matthew, On Pragmatic Encroachment on Epistemology, 53 Phil. & Phenomenological Res. 558 (2007)CrossRefGoogle Scholar; Jason Stanley, Knowledge and Practical Interests (2005); Hawthorne, John & Stanley, Jason, Knowledge and Action, 105 J. Phil. 571 (2008)CrossRefGoogle Scholar; Keith DeRose, The Case for Contextualism (2009); Jeremy Fantl & Matthew McGrath, Knowledge in an Uncertain World (2009).

18. Consider the train cases in Fantl & McGrath, Evidence, supra note 17, at 67–68.

Intuitively, in Train Case 1, you are epistemically justified in believing that the train stops in Foxboro, but in Train Case 2, even though you have the same evidence, you are not justified in believing this proposition. Differences in pragmatic factors (e.g., the costs of being wrong) can make a difference to justification. Similar cases can be found in Cohen, Stewart, Contextualism, Skepticism and the Structure of Reasons, 13 Phil. Persp. 57 (1999)Google Scholar; DeRose, Keith, Contextualism and Knowledge Attributions, 52 Phil. & Phenomenological Res. 513 (1992)CrossRefGoogle Scholar; Stanley, supra note 17; Fantl & McGrath, Knowledge, supra note 17; and Jacob Ross & Mark Schroeder, Belief, Credence, and Pragmatic Encroachment, Phil. & Phenomenological Res. Published online March 22, 2012.

19. It is critical to note that the recognition that our non-truth-related goals are relevant to epistemic justification does not amount to conflating epistemic justification with pragmatic justification. One may impose a pragmatic condition on justification (e.g., that to be justified in believing that p requires the rationality of acting as if p) while denying that epistemic justification for p requires pragmatic justification for p. See Fantl & McGrath, Evidence, supra note 17, at 83–84.

20. See Pardo, Michael, The Gettier Problem and Legal Proof, 37 Legal Theory 45–46 (2010)Google Scholar.

21. On the kind of pragmatic factors that bear upon the justification of fact finders’ beliefs, see infra Section VI.

22. On “deliberative coherence,” see Thagard, Coherence, supra note 11, at 127–132; and Thagard & Millgram, supra note 3.

23. The incorporation of pragmatic elements does not, however, call into question the coherentist commitments of the theory, as the acceptance of these elements depends, as much as that of any other element, on their coherence with all the other elements. See infra note 30.

24. Thagard also suggests that reasonable doubt might be viewed as an additional constraint on the maximization of coherence. In his view, this constraint would require that the hypothesis concerning guilt must be substantially more plausible than those concerning innocence. See Thagard, Hot Thought, supra note 8, at 366. As for the presumption of innocence, Thagard suggests that one could modify ECHO—a computational implementation of the principles of explanatory coherence—in several ways so as to account for it. My suggestion, as shown below, is to modify instead the principles of coherence, more specifically, the principle of data priority.

25. In most legal systems, the presumption of innocence, the standard of proof, and admissibility rules are the most important constraints for defining the normative institutional framework within which the explanatory evaluation of hypotheses about the events being litigated proceeds. There are, however, other rules of evidence and procedure—as well as constitutional doctrines—that might further constrain coherentist reasoning in forensic contexts. For instance, in some legal systems there are corroboration rules that would prevent the acceptance of an element unless another one is also accepted. Or, as another example, some legal systems have rules of weight that may assign some priority to some elements, in addition to the priority given to propositions describing evidence as well as to the innocence hypothesis, in the coherence calculation. I thank Ho Hock Lai for helpful suggestions on this issue.

26. Hastie & Pennington, supra note 2, at 158.

27. Thagard, Paul & Kunda, Ziva, Making Sense of People, in Connectionist Models of Social Reasoning and Social Behavior 3 (Read, S.J. & Miller, L.C. eds., 1998)Google Scholar.

28. On visual abduction, see Thagard, Paul & Shelley, Cameron, Abductive Reasoning: Logic, Visual Thinking and Coherence, in Logic and Scientific Methods 413 (Chiara, M.L. Dallaet al. eds., 1997)CrossRefGoogle Scholar.

29. These criteria of coherence are not merely meant to be a safeguard against error; but the constraint-satisfaction approach to coherence aims to be a theory of “epistemic” justification. The distinctive feature of epistemic justification, as opposed to other kinds of justification, is its essential connection with truth. One's beliefs are epistemically justified only if and to the extent that one has reasons to think that they are likely to be true; see BonJour, supra note 3, at 5–8. Thus the constraint-satisfaction approach to the epistemic justification of evidentiary judgments in law is meant to provide an account of truth-conducive criteria, i.e., criteria such that if one accepts all and only those beliefs that satisfy these criteria, one has reason to believe that they are true (or at least approximately true); see Thagard, Paul, Coherence, Truth, and the Development of Scientific Knowledge, 74 Phil. Sci. 28 (2007)CrossRefGoogle Scholar. The constraint-satisfaction approach aims not merely at providing reasons to reject a belief as false but at giving reasons to believe it is true. In short, the constraint-satisfaction approach to coherence, insofar as it is a theory of epistemic justification, is at the service of the dual objectives of obtaining truth and avoiding error; see Lehrer, supra note 3, at 20–21. For a critique of the claim that hypotheses that maximize coherence in Thagard's sense are likely to be true, see Millgram, Elijah, Coherence: The Price of the Ticket, 97 J. Phil. 82 (2000)Google Scholar; and Erik J. Olsson, Against Coherence: Truth, Probability, and Justification (2005), at 162–170.

30. It might be argued that rather than resulting in a modified version of coherentism, these adjustments ultimately turn the theory into something other than a coherence theory of justification. They do not. The main tenet of coherentism—and what makes it a distinctive theory of justification—is the claim that coherence generates justification. There are different roles that coherence may play in justification. Coherence may play a negative role, i.e., incoherence may defeat justification. It may also play a positive role, i.e., coherence may be a source of justification. Here two cases may be distinguished: coherence may be claimed to be a “conditional” source of justification, i.e., it may justify, but only given justification from other sources; or it may be claimed to be an “unqualified” source, i.e., it may generate justification where no justification existed before. Some versions of foundationalism give coherence a negative role; others also take coherence to be a conditionally basic source, but all deny that coherence is an unqualified source. In contrast, coherentism claims that coherence has the potential to generate justification “from scratch.” It is acceptance or rejection of the claim that coherence is an unqualified source that makes a theory of justification either foundationalist or coherentist. See Audi, Robert, Foundationalism, Coherentism, and Epistemological Dogmatism, 2 Phil. Persp. 407 (1988), 431–432CrossRefGoogle Scholar; and Robert Audi, The Structure of Justification (1993), at 144–145. See also Shogenji, Tomoji, The Role of Coherence in Epistemic Justification, 79 Australasian J. Phil. 90 (2001)CrossRefGoogle Scholar. The modifications proposed here are meant to give an account of the practical and institutional dimensions of evidential reasoning in legal contexts but leave the main tenet of coherentism untouched. I thank an anonymous referee for this objection.

31. For an explanationist approach to legal reasoning about facts, see John R. Josephson & Susan G. Josephson, Abductive Inference: Computation, Philosophy, Technology (1994); Abimbola, Kola, Abductive Inference in Law: Taxonomy and Inference to the Best Explanation, in The Dynamics of Judicial Proof 337 (MacCrimmon, M. & Tillers, P. eds., 2001)Google Scholar; and Allen, Ronald & Pardo, Michael, Juridical Proof and Best Explanation 27 Law & Phil. 223 (2008)Google Scholar; Amaya, Amalia, Inference to the Best Legal Explanation, in Legal Evidence and Proof: Statistics, Stories, Logic 135 (Kaptein, H., Prakken, H., & Verheij, B., eds., 2009)Google Scholar.

32. William Lycan, Judgment and Justification (1988), at 129. See also Lycan, William, Explanation and Epistemology, in The Oxford Handbook of Epistemology 412–414 (Moser, P. ed., 2002)Google Scholar.

33. On the generation and selection of explanatory hypotheses, see Peter Lipton, Inference to the Best Explanation (2d ed. 2004). On the relevance of the context of pursuit to explanatory reasoning, see Sintonen, Matti & Kikeri, Mika, Scientific Discovery, in Handbook of Epistemology 214–218 (Niiniluoto, I., Sintonen, M. & Wolenski, J. eds., 2004)Google Scholar.

34. The term “base of coherence” is Raz's. See Raz, Joseph, The Relevance of Coherence, 72 B.U. L. Rev. 273 (1992)Google Scholar.

35. The term “contrast set” is borrowed from Josephson. See Josephson, John, On the Proof Dynamics of Inference to the Best Explanation, in The Dynamics of Judicial Proof 293 (MacCrimmon, M. & Tillers, P. eds., 2001), at n.18Google Scholar.

36. See Klein, Peter & Warfield, Ted A., What Price Coherence? 54 Analysis 129 (1994)CrossRefGoogle Scholar (arguing that a belief set may be rendered more coherent not only by subtracting beliefs but also by adding beliefs).

37. The label “reinterpretation” is Conte's. See Maria Elisabeth Conte, Condizioni di Coerenza [Conditions of Coherence] (1999) (Italy), at 88.

38. These three ways in which coherence may be enhanced in the course of legal decision making are broadly inspired by the belief revision operations that are distinguished in the belief revision literature. See Peter Gärdenfors, Knowledge in Flux (1988). See also Olsson, Erik J., Making Beliefs Coherent, 7 J. Logic, Language & Info. 143 (1998)CrossRefGoogle Scholar; and Olsson, Erik J., A Coherence Interpretation of Semi-Revision, 63 Theoria 105 (1997)CrossRefGoogle Scholar. For an application of belief revision theory to legal epistemology, see Amalia Amaya, 15 Formal Models of Coherence and Legal Epistemology, Artificial Intelligence and Law 429 (2007).

39. See Oscar Wilde, The Importance of Being Earnest; R v. Smith [1915] 11 Cr. App. R. 229.

40. See Wagenaar, van Koppen, & Crombag, supra note 2, ch. 11.

41. This is the so-called “objection from the bad lot” or “problem of underconsideration” raised by van Fraassen against models of inference to the best explanation in science. See Bas C. van Fraassen, Laws and Symmetry (1989), at 143. For a detailed discussion of this problem and its relevance to legal fact-finding, see Amaya, supra note 31.

42. As I argue below, such belief would be, nonetheless, justified iff it could be accepted as such by an epistemically responsible fact finder. An important distinction needs to be made between the conditions under which a fact finder is justified in believing that p and the conditions under which the belief that p is justified.

43. See Simon, supra note 4, at 520–549.

44. See Baehr, Jason, Evidentialism, Vice, and Virtue, 78 Phil. & Phenomenological Res. 545 (2009), at 549–552CrossRefGoogle Scholar.

45. For a defense of this responsibilist version of coherentism, see Amaya, Amalia, Coherence, Justification, and Epistemic Responsibility in Legal Fact-finding, 5 Episteme 306 (2008)CrossRefGoogle Scholar. Theory. Similarly, Baehr argues that evidentialism needs to be modified by incorporating a condition according to which “justification supervenes on a belief that fits a person's evidence only if this person has exercised certain intellectual virtues in the formation and maintenance of this belief”; Baehr, Jason, Four Varieties of Character-Based Virtue Epistemology, Southern J. Phil. 469 (2008), at 484–485Google Scholar. For a statement and defense of this responsibilist version of evidentialism, see Baehr, Evidentialism, supra note 44. BonJour's account of a priori justification also endorses a similar responsibility-based constraint according to which a person is a priori justified in believing a given claim just in case he has rational insight into the necessity of this claim, provided that such insights have been arrived at on the basis of “reasonably careful reflection,” which is incompatible, argues BonJour, with dogmatism, bias, and other intellectual vices. See Laurence BonJour, In Defense of Pure Reason (1998), at 110–115.

46. The suggestion that epistemic responsibility is an important concept in a coherence theory of justification may be found—even if not elaborated on—in some of the coherentist literature. BonJour argues that epistemic responsibility is at the core of the notion of epistemic justification; BonJour, supra note 3, at 8. Lehrer similarly argues that trustworthiness, which is the keystone of his coherence theory of justification, is a matter of intellectual virtue; see Lehrer, Keith, Discursive Knowledge, 60 Phil. & Phenomenological Res. 637 (2000), at 648–650CrossRefGoogle Scholar; and Lehrer, Keith, The Virtue of Knowledge, in Virtue Epistemology: Essays on Epistemic Virtue and Responsibility 200 (Fairweather, A. & Zagzebski, L. eds., 2001)Google Scholar. Epistemic virtues, understood as broad cognitive abilities or powers rather than as traits of character one might be held responsible for, are also claimed to be relevant to a coherentist epistemology; see Sosa, Ernest, The Coherence of Virtue and the Virtue of Coherence, in Knowledge in Perspective: Selected Essays in Epistemology 192 (1991)CrossRefGoogle Scholar. In addition, some proponents of coherence theories of moral justification as well as practical deliberation emphasize the relevance of features of the agent that make for responsible epistemic behavior to attributions of justified belief; see DePaul, supra note 3, at 174; Goldman, supra note 3, at 183; and Hurley, supra note 3.

47. I thank an anonymous referee for raising this objection.

48. For a discussion of cases of intellectual deficiency as counterexamples to responsibilist accounts of epistemic responsibility, see Alston, William, Concepts of Epistemic Responsibility, 68 Monist 57 (1985)CrossRefGoogle Scholar.

49. On the relevance of memory to coherence-based reasoning, see Samet, Jerry & Schank, Roger, Coherence and Connectivity, 7 Linguistics & Phil. 57 (1984)CrossRefGoogle Scholar.

50. For an argument to the effect that one can have unjustified beliefs that are epistemically blameless, see Pryor, James, Highlights of Recent Epistemology, 52 Brit. J. Phil. Sci. 95 (2001), at 114–116CrossRefGoogle Scholar.

51. For an argument to the effect that just as the notion of “what a virtuous person would do” is insufficient to explain rightness, the notion of a belief should be justified on grounds independent of epistemic virtue, see Crisp, Roger, Virtue Ethics and Virtue Epistemology, 41 Metaphilosophy 22 (2010)CrossRefGoogle Scholar.

52. For a defense of a pure virtue epistemology that makes the concept of justified belief derivative from the concept of a virtue, see Linda Zagzebski, Virtues of the Mind (1996).

53. Thus the requirement of epistemic responsibility is not claimed to be a defining feature of justification. Rather, it is meant to play a constraining role: justification is a matter of coherence, but for reasons of coherence to generate justification, they have to operate against the background of epistemically responsible agency. Similarly, Baehr articulates and defends a modified version of evidentialism according to which justification is a matter of the possession of good epistemic reasons, provided that if a person's evidential situation significantly involves her agency, the person functions in a manner consistent with intellectual virtue. On this approach, as in the approach defended here, virtuous functioning is not a justifying factor or a necessary condition of justification but plays a background or constraining role. However, Baehr's reasons for giving virtuous activity a constraining role in a theory of epistemic justification rather than making it a necessary condition of justification are not related either to cases in which epistemic responsibility and epistemic justification seem to come apart (e.g., cases of intellectual deficiency) or to a rejection of the view that the notion of justified belief is derivative from that of intellectual virtue. Instead, his proposal is motivated by cases of “brute” or “passive” justification, i.e., cases in which justification results primarily from the mere brute functioning of the subject's cognitive mechanisms. In such cases, argues, Baehr, virtuous agency is not involved in any way, and this shows that intellectual virtue cannot be a necessary condition of justification. See Baehr, Evidentialism, supra note 44, at 558–561.

54. See Kornblith, Hilary, Justified Belief and Epistemically Responsible Action, 92 Phil. Rev. 33 (1983)CrossRefGoogle Scholar. The relation between epistemic responsibility and epistemic justification is a central issue in contemporary epistemology. For a useful survey of the different views on this topic, see Pryor, James, Highlights of Recent Epistemology, 52 Brit. J. Phil. Sci. 95 (2001)CrossRefGoogle Scholar.

55. See Feldman, Richard, Epistemic Obligations 2 Phil. Persp. 236 (1988)Google Scholar; Feldman, Richard, Epistemological Duties, in The Oxford Handbook of Epistemology 362 (Moser, P. ed., 2002)CrossRefGoogle Scholar.

56. For a defense of a virtue approach to the notion of epistemic responsibility, see James A. Montmarquet, Epistemic Virtue and Doxastic Responsibility (1992); and Lorraine Code, Epistemic Responsibility (1987).

57. There are different views in the literature regarding the traits of character that mark off intellectually virtuous behavior and how to classify them. See Montmarquet, James A., Epistemic Virtue, 96 Mind 482 (1987)CrossRefGoogle Scholar; Cooper, Neil, The Intellectual Virtues, 69 Philosophy 459 (1994)CrossRefGoogle Scholar; Robert C. Roberts & W. Jay Wood, Intellectual Virtues: An Essay in Regulative Epistemology (2007).

58. Alternatively, one could endorse an aretaic approach and define all deontic concepts in terms of virtues. Linda Zagzebski provides an exemplary account of how this might be done. See Zagzebski, supra note 52, at 232–259. For an attempt to define the epistemic responsibility of legal fact-finders exclusively in aretaic terms, see Amaya, Coherence, Justification, and Epistemic Responsibility in Legal Fact-finding, supra note 40, at 311–14.

59. This account is not meant to be exhaustive but merely illustrative of the kinds of duties and virtues that are relevant to evidential reasoning in law.

60. See Feldman, Epistemological Duties, supra note 55, at 367–369.

61. See Hall, Richard J. & Johnson, Charles R., The Epistemic Duty to Seek More Evidence, 35 Am. Phil. Q. 129 (1998)Google Scholar. Critically, the duty to seek out more evidence requires fact finders to search also for second-order evidence, i.e., evidence about the merits of one's evidence. On the epistemic significance of second-order evidence, see Christensen, D., Higher-Order Evidence, 81 Phil. & Phenomenological Res. 185 (2010)CrossRefGoogle Scholar; and Feldman, Richard, Respecting the Evidence, 19 Phil. Persp. 95 (2005)CrossRefGoogle Scholar. It is also important to notice that at least in the context of legal fact-finding, issues concerning the sufficiency of one's first-order as well as second-order evidence not only are epistemological but also involve an important moral and political component. See Alex Stein, Foundations of Evidence Law (2005), at 118–133.

62. This assumes a more active view of jurors than the one that is currently in place in most jury systems. For an argument in support of the view that triers of fact ought to be allowed to play a more active role, see Jackson, John, Analyzing the New Evidence Scholarship: Towards a New Conception of the Law of Evidence, 16 Oxford J. Legal Stud. 309 (1996), at 326CrossRefGoogle Scholar.

63. There are different views about the kind of epistemic duties we have. A narrow conception of our epistemic duties is that they are exclusively duties to have a particular cognitive attitude (belief, disbelief, or suspension of judgment) toward a particular proposition. A broader conception is that they are duties to engage in behavior that will maximize one's number of true beliefs. A broader conception, I submit, best characterizes legal fact finders’ epistemic duties.

64. Margolis, Joseph, The Locus of Coherence, 7 Linguistics & Phil. 3 (1984), at 23CrossRefGoogle Scholar.

65. DeRose, Keith, Contextualism: An Explanation and Defense, in The Blackwell Guide to Epistemology 187 (Greco, J. & Sosa, E. eds., 1999), at 190Google Scholar.

66. In suggesting that coherentism ought to be contextualized, I depart from other accounts according to which contextualism is conceived as an account of justification alternative (and superior) to coherentism. For a defense of such a view, see Henderson, David K., Epistemic Competence and Contextualist Epistemology: Why Contextualism Is Not Just a Poor Person's Coherentism, 91 J. Phil. 627 (1994)CrossRefGoogle Scholar.

67. See Duran, Jane, A Contextualist Modification of Cornman, 16 Philosophia 377 (1986)CrossRefGoogle Scholar (arguing that a contextualization of the justifying set is necessary for the coherence theory to be descriptively adequate and psychologically plausible). However, this is just one of the dimensions along which coherence theories need to be contextualized, the other two being the degree of coherence required for justification and the relevant contrast set.

68. Of course, different views about which parameters shift with context will yield different contextualized versions of coherentism. For example, Schaffer argues that what shifts with context is not the threshold of justification but rather the set of relevant alternatives. An application of this contextualist view to coherentism would result in a contextualized version of coherentism according to which it is only the contrast set, rather than the degree of coherence required for justification, that varies with context. See Schaffer, Jonathan, What Shifts? Thresholds, Standards, or Alternatives?in Contextualism in Philosophy: Knowledge, Meaning, and Truth 115 (Preyer, Gerhard & Peter, Georg eds., 2005)Google Scholar.

69. The relevance of the importance of being right about a certain matter for the severity of the standards of justification is often pointed out in the contextualist literature. See Mark Timmons, Morality without Foundations: A Defense of Ethical Contextualism (1999), at 211; Annis, David B., A Contextualist Theory of Epistemic Justification, 15 Am. Phil. Q. 213 (1978), at 215Google Scholar; Fogelin, Robert J., Contextualism and Externalism: Trading in One Form of Skepticism for Another, 10 Phil. Issues 43 (2000), at 48CrossRefGoogle Scholar; Lewis, David, Elusive Knowledge, 74 Australasian J. Phil. 549 (1996), at 556CrossRefGoogle Scholar; and Michael Williams, Problems of Knowledge: A Critical Introduction to Epistemology (2001), at 161. But see Dretske, Frederick I., The Pragmatic Dimension of Knowledge, 40 Phil. Stud. 363 (1981), at 375CrossRefGoogle Scholar (arguing that the importance of what is justified should not affect what counts as an adequate justification).

70. Lillquist, Erik, Recasting Reasonable Doubt: Decision Theory and the Virtues of Variability, 36 U.C. Davis L. Rev. 85 (2002–2003)Google Scholar (arguing that a flexible standard of reasonable doubt that varies depending on the costs of error associated with each case, should be preferred to a fixed standard of reasonable doubt requiring the same level of certainty across all criminal cases).

71. Cf. Stein, supra note 61 at 179 (arguing that the reasonable doubt standard should apply equally in all cases, regardless of the severity of the offense on trial, as there is a distinct “moral harm” in all wrongful convictions that, unlike the bare harm that an erroneous conviction produces, stays invariant across different crimes). On the claim that the violation of the right not to be convicted if innocent constitutes a special kind of harm (i.e., a “moral” harm) over and above the bare harm a person suffers through punishment, see Ronald Dworkin, A Matter of Principle (1985), at 80.

72. See Robert Fogelin, Walking the Tightrope of Reason: The Precarious Life of a Rational Animal (2003), at 102. For some defenses of the claim that expertise and one's occupation are factors that bear upon the question of how severe the standards of justification should be, see Timmons, supra note 69, at 211; and Annis, supra note 69, at 213.

73. Timmons argues that whether a belief is justified varies in ways that depend upon one's goals and purposes. See Timmons, supra note 69, at 192–193.

74. The relevance of methodological constraints to attributions of justified belief is pointed out by Williams. See Williams, supra note 69, at 160.

75. Id. at 160. See also Williams, Michael, Is Contextualism Statable? 10 Phil. Stud. 80 (2000), at 84Google Scholar.

76. Sinnott-Armstrong, Walter, What's in a Contrast Class? 62 Analysis 75 (2002), at 76CrossRefGoogle Scholar.

77. Williams argues for the relevance of dialectical features to attributions of justified belief. See Williams, Problems, supra note 69, at 161.

78. See, e.g., Lewis, supra note 69, at 559. According to Lewis's “rule of attention,” those possibilities that one is in fact attending to in a context cannot be properly ignored.

79. This view has its roots in Peirce's writings and it also enjoys some popularity now, particularly among the advocates of “inferential” views of contextualism. For instance, Annis says that for justification it is necessary to meet only those objections that are an expression of a “real” doubt. See Annis, supra note 69, at 214. See also Williams, Problems, supra note 69, at 161.

80. The label of “threshold of termination” is borrowed from discourse theorists Beaugrande and Dressler, who define it as the place at which “the comprehension and integration of the text is deemed as satisfactory.” See Robert A. De Beaugrande & Wolfgang U. Dressler, Introduction to Text Linguistics (1981), at 43.

81. See Fumerton, Richard, A Critique of Coherentism, in Theory of Knowledge: Classical and Contemporary Readings 247–250 (Pojman, L. ed., 2d ed. 1999)Google Scholar.

82. See Michelle Taruffo, La Prova dei Fatti Giuridici: Nozioni Generali [The Proof of Legal Facts: General Notions] (1992) (Italy), at 285; Twining, William, Necessary but Dangerous? Generalizations and Narrative in Argumentation about “Facts” in Criminal Processes, in Complex Cases: Perspectives on the Netherlands Criminal Justice System 80 (Malsch, M. & Nijboer, J.F. eds., 1999)Google Scholar; Menashe, Doron & Shamash, Mutal E., The Narrative Fallacy, 3 Int'l Comment. Evidence article 3 (2005)Google Scholar. The risks involved in holistic processing are quite real, as empirical work has shown. See Simon, supra note 4.

83. See Williams, Michael, Coherence, Justification, and Truth, 34 Rev. Metaphysics 243 (1980)Google Scholar.

84. See Simon, Dan, A Psychological Theory of Legal Decision-Making, 30 Rutgers L.J. 1 (1998), at 88–89Google Scholar.

85. See Williams, Coherence, Justification, supra note 83, at 249.

86. See Laurence BonJour, The Structure of Scientific Knowledge (1985), at 89–92; Shogenji, supra note 30, at 94; John L. Pollock, Contemporary Theories of Knowledge (1986), at 73.

87. For a statement of the isolation objection, see John L. Pollock, Knowledge and Justification (1974), at 27–28. For some responses, see Pollock, Contemporary, supra note 86, at 76–77; Williams, Coherence, Justification, supra note 83, at 249–252; BonJour, Structure, supra note 3, at 139–143.

88. Thagard takes the principle of data priority to provide a solution to the isolation objection as directed against coherence theories of epistemic justification. See Thagard, Coherence, supra note 11, at 73.

89. See BonJour, Structure, supra note 3, at 107–108.

90. See Gianformaggio, Letizia, Legal Certainty, Coherence, and Consensus: Variations on a Theme by MacCormick, in Law, Interpretation and Reality 402 (Nerhot, P. ed., 1990), 430CrossRefGoogle Scholar.

91. A similar point is made by Nance. See Nance, Dale A., Naturalized Epistemology and the Critique of Evidence Theory, 87 Va. L. Rev. 1551 (2001)CrossRefGoogle Scholar, at 1586 n.107.

92. One could meet the objection that a coherence theory of justification does not help us achieve the goal of establishing the truth at trial by endorsing a coherence theory of truth. Indeed, a version of the coherence theory of truth is defended by some proponents of coherentist approaches to evidence and legal proof, such as Jackson, Law, Fact, supra note 1, and van Roermund, supra note 2. However, an antirealist conception of truth seems to be incompatible with many procedural and evidential institutions; see Damaška, Mirjan, Truth in Adjudication, 49 Hastings L.J. 289 (1997–1998)Google Scholar. In addition, the coherence theory of truth faces serious objections; see Ralph C.S. Walker, The Coherence Theory of Truth: Realism, Anti-Realism, Idealism (1989). These objections may also be directed against any attempt to analyze the truth of evidentiary statements in terms of coherence. Thus the challenge for coherence theories of evidence and legal proof is to show that coherentist standards of justification lead us to accept beliefs about the facts under dispute that are likely to be true in a way that does not involve a definition of truth as coherence.

93. See BonJour, Structure, supra note 3, at 171. See also Thagard, Coherence, supra note 11, at 78–80; and Thagard, Coherence, Truth, supra note 29.

94. See Lehrer, Theory of Knowledge, supra note 3. See also Lehrer, Keith, Coherence and the Truth Connection: A Reply to My Critics, in The Current State of the Coherence Theory 253 (Bender, J. ed., 1989)CrossRefGoogle Scholar; Lehrer, Keith, Justification, Coherence, and Knowledge, 50 Erkenntnis 243 (1999)CrossRefGoogle Scholar.

95. See Davidson, supra note 3.

96. See Kvanvig, Jonathan, Truth Is Not the Primary Epistemic Goal, in Contemporary Debates in Epistemology 285 (Steup, Matthias & Sosa, Ernest eds., 2005)Google Scholar. Similarly, truth does not seem to be the only goal that is relevant in moral rather than epistemic deliberation. See DePaul, supra note 3.

97. See William Twining, Rethinking Evidence (2d extended ed., 2006), at 75–80; Mirjan R. Damaška, Evidence Law Adrift (1997), at 120–124; Summers, Robert S., Formal Legal Truth and Substantive Truth, 18 Law & Phil. 499–500 (1999)Google Scholar; and Damaška, Truth in Adjudication, supra note 92, at 289.

98. See Hurley, supra note 3.

99. See Richardson, supra note 3.

100. For a naturalistic approach to evidence law, see Allen, Ronald J. & Leiter, Brian, Naturalized Epistemology and the Law of Evidence, 87 Va. L. Rev. 1491 (2001)CrossRefGoogle Scholar. For a critique, see Nance, supra note 91. See also Redmayne, Mike, Rationality, Naturalism, and Evidence Law, 2003 Mich. St. L. Rev. 849 (2003)Google Scholar; with a reply by Ronald J. Allen and Brian Leiter in the same issue.

101. See Ziff, Paul, Coherence, 7 Linguistics & Phil. 31 (1984)CrossRefGoogle Scholar.

102. See Hastie & Pennington, supra note 2; Wagenaar, van Koppen, & Crombag, supra note 2; and Simon, Third View, supra note 4.

103. See Holyoak, Keith J. & Simon, Dan, Bidirectional Reasoning in Decision-Making by Constraint Satisfaction, 128 J. Experimental Psychol.: Gen. 3 (1999)CrossRefGoogle Scholar; Holyoak, Keith J.et al., The Emergence of Coherence over the Course of Decision-Making, 27 J. Experimental Psychol.: Learning, Memory & Cognition 1230 (2001)Google Scholar; Simon, Dan & Holyoak, Keith J., Structural Dynamics of Cognition: From Consistency Theories to Constraint Satisfaction, 6 J. Personality & Soc. Psychol. Rev. 283 (2002)CrossRefGoogle Scholar; Simon, Dan, Snow, Chadwick J. & Read, Stephen J., The Redux of Cognitive Consistency Theories: Evidence Judgments by Constraint Satisfaction, 86 J. Personality & Soc. Psychol. 814 (2004)CrossRefGoogle ScholarPubMed.

104. Two main approaches to virtue epistemology may be distinguished: virtue responsibilism and virtue reliabilism; see Axtell, Guy, Recent Work on Virtue Epistemology, 34 Am. Phil. Q. 1 (1997)Google Scholar. Virtue responsibilists understand virtues as personality traits or qualities of character that are analogous to the moral virtues. See Code, supra note 56, Montmarquet, Epistemic Virtue, supra note 56, and Zagzebski, supra note 52. In contrast, proponents of “virtue reliabilism” understand virtues as reliable cognitive abilities or powers. See Sosa, supra note 46; Ernest Sosa, Apt Belief and Reflective Knowledge (2007); John Greco, Putting the Skeptics in Their Place (2000); John Greco, Achieving Knowledge: A Virtue-Theoretic Account of Epistemic Normativity (2010). It is the responsibilist version of virtue epistemology that seems to me to be more promising as a framework for analyzing issues of knowledge and justification in law. For a discussion of the main versions of virtue epistemology, see Greco, John, Virtues in Epistemology, in The Oxford Handbook of Epistemology 287 (Moser, P. ed., 2002)CrossRefGoogle Scholar. For some work that aims at developing a virtue approach to the epistemology of legal proof, see Amaya, supra note 58, and Lai, Ho Hock, Virtuous Deliberation, in Virtue, Law, and Justice (Amaya, A. & Ho, H. L., eds., 2012)Google Scholar.

105. There is nonetheless an important body of work on discovery by evidence scholars. See, among others, Schum, David A., Marshalling Thoughts and Evidence during Fact-Investigation, 40 S. Tex. L. Rev. 401 (1999)Google Scholar; and Tillers, Peter & Schum, David A., A Theory of Preliminary Fact-Investigation, 24 U. C. Davis L. Rev. 931 (1990–1991)Google Scholar.