¹ For the relevance of character in the context of preventive means see for example Duff, RA ‘Pre-trial detention and the presumption of innocence’ in Ashworth, A et al. (eds) Prevention and the Limits of the Criminal Law (Oxford: Oxford University Press, 2013) pp 127–130Google Scholar; Wasserman, DT ‘The morality of statistical proof and the risk of mistaken liability’ (1991) 13 Cardozo Law Review 935Google Scholar. For character in the context of criminal liability see eg Duff, RA ‘Choice, character, and criminal liability’ (1993) 12 Law and Philosophy 345CrossRefGoogle Scholar; Tadros, V Criminal Responsibility (Oxford: Oxford University Press, 2005) ch 2Google Scholar; Bayles, MD ‘Character, purpose, and criminal responsibility’ (1982) 1 Law and Philosophy 5CrossRefGoogle Scholar. For character in sentencing see eg Bottoms, AE and Brownsword, R ‘The dangerousness debate after the Floud report’ (1982) 22 British Journal of Criminology 229CrossRefGoogle Scholar; Hirsch, A von and Ashworth, A Proportionate Sentencing: Exploring the Principles (Oxford: Oxford University Press, 2005) especially ch 4CrossRefGoogle Scholar; Roberts, JV and von Hirsch, A (eds) Previous Convictions at Sentencing: Theoretical and Applied Perspectives (London: Hart Publishing, 2010)Google Scholar.

² For the history of the use of character evidence to prove guilt and for a detailed account of the reforms in the Criminal Justice Act 2003 see M Redmayne Character in the Criminal Trial (2015) chs 5 and 7. The Criminal Justice Act 2003, Part 12, chapter 5 reformed the law on sentencing of ‘dangerous offenders’. For a brief history and an account of the reform see Redmanye, ch 12.

³ The paper does not discuss good character evidence, but it could lead to the conclusion that there is no sufficient ground for excluding it.

⁴ But see Picinali, F ‘Generalisations, causal relationships and moral responsibility’ (2016) 20 The International Journal of Evidence & Proof 121CrossRefGoogle Scholar.

⁵ Advocates of broad admissibility include Redmayne, above n 2; Spencer, JR Evidence of Bad Character (London: Hart Publishing, 2nd edn, 2009)Google Scholar; Culberg, D ‘The accused's bad character: theory and practice’ (2009) 84 Notre Dame Law Review 1343Google Scholar. Advocates of narrow admissibility include Tandy, R ‘The admissibility of a defendant's previous criminal record: a critical analysis of the Criminal Justice Act 2003’ (2009) 30 Statute Law Review 203 at 207–208CrossRefGoogle Scholar; S Parsons ‘The Criminal Justice Act 2003 – do the bad character provisions represent a move towards an authoritarian model of criminal justice?’ in (2007) Mountbatten Yearbook of Legal Studies 181.

⁶ Redmayne, above n 2.

⁷ Redmayne, above n 2, at p 93 mentions the unreported decision in Cole from 1810 as the point of emergence of the rule of exclusion.

⁸ Redmayne, above n 2, p 94.

⁹ Ibid.

¹⁰ Redmayne, above n 2, p 128, relying on Cross, R Evidence (London: Butterworths, 5th edn, 1979) pp 378–393Google Scholar.

¹¹ Ibid.

¹² [1952] 2 QB 911.

¹³ Spencer, above n 5, at para 1.72.

¹⁴ R v Smith (1916) 11 Cr App R 229.

¹⁵ This can be distinguished from having one's physical fingerprint found in the scene of crime, which is not necessarily associated with any transparently erroneous behaviour that was sure to turn the defendant into a suspect of crime.

¹⁶ DPP v Boardman [1975] AC 421.

¹⁷ Redmayne, above n 2, p 130.

¹⁸ CJA 2003, s 101(1)(a).

¹⁹ CJA 2003, s 101(1)(b).

²⁰ CJA 2003, s 101(1)(f).

²¹ CJA 2003, s 101(1)(g).

²² CJA 2003, s 101(1)(c).

²³ CJA 2003, s 101(1)(e).

²⁴ Picinali, above n 4, 129.

²⁵ For example, in the bus companies example which is a variation on the American case of Smith v Rapid Transit Inc 58 NE 2d 754 (Mass 1945). See the discussion in Wells, GL ‘Naked statistical evidence of liability: is subjective probability enough?’ (1992) 62 Journal of Personality and Social Psychology 739CrossRefGoogle Scholar; or the rodeo stadium example in Cohen, LJ, The Probable and the Provable (Clarendon Press 1977) pp 74–75CrossRefGoogle Scholar.

²⁶ Notably, legal practice does not yet indicate a problem, as naked statistical evidence is hardly ever relied on to prove propensity.

²⁷ R v Hanson [2005] EWCA Crim 824, [2005] 2 Cr App R 21.

²⁸ Ibid.

²⁹ Redmayne, above n 2, p 167; Choo, AL-T Evidence (Oxford: Oxford University Press, 4th edn, 2015) pp 263–264CrossRefGoogle Scholar.

³⁰ Redmayne, above n 2, p 169.

³¹ Redmayne, above n 2, pp 169–172.

³² This point is further clarified below.

³³ Enoch, D and Fisher, T ‘Sense and “sensitivity”: epistemic and instrumental approaches to statistical evidence’ (2015) 67 Stanford Law Review 557 at 587–592Google Scholar.

³⁴ It is possible that the police held a sample of the defendant's DNA due to previous involvement in crime; in this case, the DNA evidence actually indicates previous misconduct, and despite the different inference from such evidence, the meaning of the error is similar to its meaning in cases of ‘bad character evidence’. This may explain some of the controversy over DNA databases.

³⁵ This unrealistic assumption helps to demonstrate that the ambivalence does not necessarily have to do with the probative value of the evidence. The more realistic scenario, in which bad character evidence is only part of the evidence that leads to false conviction, is discussed below.

³⁶ See Redmayne, above n 2, pp 16–21.

³⁷ For lack of space, I do not engage here directly with all the vast criminological literature about recidivism; nor will I engage with the empirical literature on bias in section 3 of the paper. Rather, I adopt Redmayne's thorough analysis of those in n 2 above.

³⁸ Redmayne, above n 2, p 24 (Table 2.7).

³⁹ Redmayne, above n 2, pp 23–24.

⁴⁰ For an elaborate critical discussion of the ways to analyse the probative value of evidence of previous misconduct see Redmayne, above n 2, pp 35–41.

⁴¹ See Picinali, above n 4, at 122–123. For critical discussion see Pundik, A ‘Against racial profiling’ (2017) 67 University of Toronto Law Journal 175 at 188–189CrossRefGoogle Scholar.

⁴² What matters for the purposes of the current argument is that there are (indeed many) defendants who will have to fight these forces, and that some of them will have triumphed over them.

⁴³ Where the evidence does not contribute to the false conviction, it is not clear that the false conviction is profoundly tragic. Yet it is hard – if not impossible – to identify such cases; furthermore, the fact that there are some cases in which the evidence does contribute to the false conviction is the significant one when making general admissibility decisions.

⁴⁴ Thus, a clean record does not mean that one has not, or does not, commit crime. Nor does it mean that one does not face similar challenges and temptations as those who have convictions.

⁴⁵ This seems to be the assumption in the CJA 2003.

⁴⁶ Redmayne, above n 2, p 25.

⁴⁷ Subsection (e).

⁴⁸ As we shall see below, the nature of the other evidence might make a difference where it indicates that the defendant endorses some of the circumstances of the offence at issue.

⁴⁹ Levanon, L ‘The law of police entrapment: critical evaluation and policy analysis’ (2016) 27 Criminal Law Forum 35CrossRefGoogle Scholar.

⁵⁰ Ibid.

⁵¹ Redmayne, above n 2, pp 26–27.

⁵² In any case, what matters for current purposes is the need to avoid rare but profoundly tragic false convictions of reformed defendants. After all, a rule of exclusion based on the nature of errors is based on the assumption that it is justified to exclude some perfectly good evidence (eg all the evidence of previous misconduct that would not end up leading to errors) in order to avoid some errors.

⁵³ According to Aristotle's definition of a tragedy, it ‘is an imitation of an action that is serious, complete and of a certain magnitude; … in the form of action, not of narrative; through pity and fear effecting the proper purgation of those passions’: The Poetics of Aristotle Part VI (3^rd edn, edited with critical notes and a translation by SH Butcher, 1902).

⁵⁴ The Aristotelian account has been criticised as overly narrow, in the sense that it does not even capture all types of Greek tragedies. See Noyes, GR ‘Aristotle and modern tragedy’ (1891) 13 Modern Language Notes 6 at 11–12Google Scholar. See further Woodruff, P ‘Aristotle on character in tragedy, or who is Creon? What is he?’ (2009) 67 The Journal of Aesthetics and Art Criticism 301CrossRefGoogle Scholar; Battin, M Pabst ‘Aristotle's definition of tragedy in the poetics’ (1975) 33 The Journal of Aesthetics and Art Criticism 293CrossRefGoogle Scholar. Yet Aristotle's account was probably not meant as an exhaustive definition of a tragedy. See Noyes, ibid, at 7.

⁵⁵ See The Poetics of Aristotle, above n 53, Part IX; Frede, D ‘Necessity, chance, and “what happens for the most part”’ in Rorty, A (ed) Essays on Aristotle's Poetics (Princeton, NJ: Princeton University Press, 1992) pp 197, 205 and 212Google Scholar. The extent to which hamartia involves (limited) guilt is controversial. See Stinton, TCW ‘Hamartia in Aristotle and Greek tragedy’ (1975) 25 The Classical Quarterly (New Series) 221CrossRefGoogle Scholar.

⁵⁶ The Poetics of Aristotle, above n 53, Part IX ff.

⁵⁷ The Poetics of Aristotle, above n 53, Part IX.

⁵⁸ Ibid.

⁵⁹ The Poetics of Aristotle, above n 53, Part XI.

⁶⁰ Ibid.

⁶¹ The Poetics of Aristotle, above n 53, Part XII.

⁶² The Poetics of Aristotle, above n 53.

⁶³ The Poetics of Aristotle, above n 53, Part XIV.

⁶⁴ The Poetics of Aristotle, above n 53, Part XV.

⁶⁵ Ibid.

⁶⁶ ‘for pity is aroused by unmerited misfortune, fear by the misfortune of a man like ourselves’: The Poetics of Aristotle, above n 53, Part XIII.

⁶⁷ The Poetics of Aristotle, above n 53, Part XIII.

⁶⁸ See Barstow, M ‘Oedipus Rex as the ideal tragic hero of Aristotle’ (1912) 6 The Classic Weekly 2CrossRefGoogle Scholar.

⁶⁹ There is disagreement in the literature on the hamartia in Oedipus Rex, but many trace it to this killing that is later revealed as patricide, that also opens the door for him marrying his biological mother. See Hyde, I ‘The tragic flaw: is it a tragic error?’ (1963) 58 The Modern Language Review 321CrossRefGoogle Scholar; Golden, L ‘Hamartia, Atë, and Oedipus’ (1978) 72 Classical World 3CrossRefGoogle Scholar.

⁷⁰ At least under the assumption that Oedipus’ hamartia was indeed the ignorant killing of the person at the crossroads.

⁷¹ The Poetics of Aristotle, above n 53, Part XI.

⁷² Compare with Redmayne, above n 2, pp 128–129.

⁷³ Kant, I Critique of the Power of Judgment (Guyer, P (ed), translated by Paul Guyer and Eric Matthews; The Cambridge Edition of the Works of Immanuel Kant (Cambridge: Cambridge University Press, 2000)) §60CrossRefGoogle Scholar.

⁷⁴ For an introduction see Levinson, J (ed) Aesthetics and Ethics: Essays at the Intersection (Cambridge: Cambridge University Press, 1998)CrossRefGoogle Scholar; Glowacka, D and Boos, S (eds) Between Ethics and Aesthetics: Crossing the Boundaries (Albany: State University of New York Press, 2002)Google Scholar; Symons, S (ed) The Marriage of Aesthetics and Ethics (Leiden: Brill, 2015)CrossRefGoogle Scholar. For the Kantian view see Guyer, P Kant and the Experience of Freedom: Essays on Aesthetics and Morality (Cambridge: Cambridge University Press, 1996)Google Scholar.

⁷⁵ Kant, above n 73.

⁷⁶ See Kant, above n 73. Kant's account is complex and deserves more extensive and accurate exploration than can be provided here. See Guyer, P ‘Feeling and freedom: Kant on aesthetics and morality’ (1990) 48 The Journal of Aesthetics and Art Criticism 137CrossRefGoogle Scholar.

⁷⁷ This is not to suggest that such unfairness is more or less significant than other forms of unfairness that might arise in the context of the criminal justice process, such as instances where the state is complicit in misleading the court about the available evidence.

⁷⁸ It is only fair that some good choices should shield from certain arbitrary forces; it is only fair that some bad choices should expose to certain arbitrary forces; and there is nothing unfair in accepting that luck has its say, so that sometimes good choices fail to shield and bad choices do not expose to such arbitrary forces. Justice does not demand that we do anything to alter these realities. For additional suggestions about the value of choice see Scanlon, TM What We Owe to Each Other (Harvard University Press, revised edition, 2000) pp 251–255Google Scholar.

⁷⁹ I suggest this argument without losing sight of the fact that the ancient Greeks were interested in personality and fate rather than in the human power of choice.

⁸⁰ Compare with Darwall's, S account of recognition respect in ‘Two kinds of respect’ (1977) 88 Ethics 36CrossRefGoogle Scholar.

⁸¹ The demeaning nature of the mistake is also reflected in the particularly high impact on the defendant's self-esteem. See Redmayne's analysis of labelling, above n 2, pp 88–90.

⁸² For presentation and analysis of the relevant mock jury studies and real trial studies see Redmayne, above n 2, pp 53–59.

⁸³ Redmayne, above n 2, p 60.

⁸⁴ Ibid, p 55, citing such findings with respect to the impact of dissimilar convictions in S Lloyd-Bostock ‘The effects on juries of hearing about the defendant's previous criminal record: a simulation study’ [2000] Criminal Law Review 734.

⁸⁵ Ibid.

⁸⁶ See McEwan, J The Verdict of the Court: Passing Judgement in Law and Psychology (London: Hart Publishing, 2003) pp 169–170Google Scholar; Spencer, above n 5; Lord Justice Auld's position in Great Britain: Lord Chancellor's Department A Review of the Criminal Courts in England and Wales (2001) ch 11, paras 118–120. See also L Laudan and Allen, RJ ‘The devastating impact of prior crime evidence and other myths of the criminal justice process’ (2013) 101 Journal of Criminal Law and Criminology 493, 515 and 522Google Scholar.

⁸⁷ Redmayne's proposed solution seems to reflect the assumption that the lack of knowledge with respect to prejudicial inferences is such a non-starter, that it must be set aside. See Redmayne, above n 2, chs 2 and 12.

⁸⁸ R Aboodi et al ‘Deontology, individualism and uncertainty: a reply to Jackson and Smith’ (2008) The Journal of Philosophy 259, 270.

⁸⁹ For a comprehensive analysis of uncertainty and theories of resolving uncertainties see Walker, VR ‘Theories of uncertainty: explaining the possible sources of error in inference’ (2001) 22 Cardozo Law Review 22Google Scholar.

⁹⁰ For a clear introduction see Peterson, M An Introduction to Decision Theory (Cambridge: Cambridge University Press, 2009) ch 3CrossRefGoogle Scholar. To go further, see eg Giang, PH ‘Decision making under uncertainty comprising complete ignorance and probability’ (2015) 62 International Journal of Approximate Reasoning 27CrossRefGoogle Scholar and the references therein.

⁹¹ John Rawls’ difference principle is closely related to the maximin principle. See Rawls, J A Theory of Justice (Harvard University Press, revised edn, 1999) p 132 ffGoogle Scholar.

⁹² These assumptions are not in conflict with the presumption of innocence. For a discussion see Levanon, L ‘Sexual history evidence in sexual assault cases: a critical reevaluation’ (2012) 62 University of Toronto Law Journal 609CrossRefGoogle Scholar.

⁹³ See the discussion above.

⁹⁴ Sanchirico, CW ‘Character evidence and the object of trial’ (2001) 101 Colum L Rev 1227CrossRefGoogle Scholar.

⁹⁵ Enoch and Fisher, above n 33, and Enoch, D et al. ‘Statistical evidence, sensitivity, and the legal value of knowledge’ (2012) 40 Philosophy and Public Affairs 197CrossRefGoogle Scholar.

⁹⁶ The case of a defendant who committed another comparable offence beforehand is mostly theoretical, but if proved, it is possible that this too should open the door to ‘bad character evidence’.