TY - JOUR AU - Schiff,, David AB - Abstract The advent of reliance on guilty pleas creates a disparity between the legitimations of the criminal justice system and the current practices of the system that they are meant to legitimate. Despite the minority of convictions following criminal trials as a percentage of the overall conviction rate, it is trial, and especially jury trial, that still provides the focus for our legitimations. This article explores what legitimations are being, and might be, offered for reliance on guilty pleas, and how those legitimations could reorientate our legitimations for the criminal justice system more generally. In particular, it suggests how the presentation of the procedures of the modern criminal justice system as giving priority to the truth of convictions or the avoidance of wrongful conviction misrepresents that system, whose legitimation can be better represented by reference to the concepts of autonomy and the autonomous exercise of rights. [W]e have known instances of murder avowed, which never were committed; of things confessed to have been stolen, which never had quitted the possession of the owner … It is both ungenerous therefore, and unjust, to suffer the distractions of fear, or the misdirected hopes of mercy to preclude that negative evidence of disproof, which may possibly, on recollection, be in the power of the party; we should never admit, when it may be avoided, even the possibility of driving the innocent to destruction.—William Aukland, 17711 1. Guilty Pleas and the Professed Aims of the Criminal Justice System We have come a long way since the sentiments expressed in the quote above were put into practice and guilty pleas positively discouraged.2 Now they are very much encouraged, and reliance on them offers a significant challenge to current attempts to advance justifications for the practices that make up our system of criminal justice. If one looks at standard works, such as Andrew Ashworth and Mike Redmayne’s The Criminal Process, the criminal justice system is vindicated through a focus on the trial: ‘The function of criminal procedure is to regulate and facilitate the preparation of cases for trial.’3 In turn, the values attributed to the criminal trial are combinations of accurate fact finding, fairness4 and human rights.5 Within this matrix of practices and values, fairness and human rights typically operate as restraints, whereby trial is ‘not just about accurate fact-finding’.6 However, such attendant values should not be understood merely as restraints, since: ‘Respect for rights should be seen as a concomitant aim of criminal process … an objective to be obtained while pursuing that aim.’7 Similarly, Ian Dennis, in The Law of Evidence, writing of the need for criminal verdicts to be ‘legitimate’, argues for the pursuit of fact finding accuracy to be accompanied by ‘moral authority’.8 The latter, which he equates with respecting the defendant’s autonomy and human rights, is presented as a restraint on the truth finding process, for: A factually inaccurate or doubtful conviction can never be legitimate … But a decision may be factually correct [he gives the example of a confession obtained through torture] and yet lack the elements of moral authority and expressive value necessary for the further [legitimating] functions of the verdict.9 The priority given to accurate fact finding, restrained by these additional requirements of moral authority, is justified in turn by the moral wrongfulness of an unjust conviction.10 A parallel emphasis is to be found in the introduction to a recent volume of essays entitled The Integrity of Criminal Process. The editors discuss the ‘trade-offs [which] must be made in adjudication between truth-finding and other normative considerations’, namely the ‘familiar procedural values such as “fairness”, “due process”, “natural justice” or “judicial legitimacy”’, on the assumption that these values operate to restrain what would otherwise be ‘a dogmatic insistence on factual rectitude’.11 Current practices, however, run contrary both to this focus on trials and to the rationales offered in reliance on trial procedure. Two-thirds of convictions involving serious crimes in the UK are obtained through guilty pleas rather than jury trials.12 For all crime prosecuted by the Crown Prosecution Service (CPS), the figure is 91.6%.13 In the United States, the situation is more extreme, with around 97% of criminal cases being resolved by guilty plea.14 Not only are trials a minority method for producing convictions, but the sentence discounts offered to defendants15 ‘encouraging’ or ‘persuading’ them to plead guilty and forgo trial, as Ashworth and Redmayne acknowledge, may undermine attempts to provide an overall justification for criminal procedure in terms of rights that contribute to accurate, fair and human-rights-based decisions.16 These discounts, in their various forms (pleading guilty to a lesser charge, pleading guilty to obtain a reduced sentence on the charge proceeded with and pleading guilty to an agreed version of facts likely to secure a lower sentence), are, they admit, in tension with their attempt to present criminal procedure as directed towards the pursuit of accurate verdicts accompanied by the protection of fundamental rights. They regard these incentivised guilty pleas as potentially running contrary to ‘the presumption of innocence, the privilege against self-incrimination, the right not to be discriminated against in the exercise of article 6 rights, and the right to a fair and public hearing’.17 There is something suspect in constructing a framework for evaluation of a social practice that condemns the majority of the outcomes produced by that practice.18 This is particularly so when it is accompanied by an acceptance that the practices being criticised (guilty pleas) could not be replaced by the procedures (trials, ideally before a jury) that are being used to generate the framework for critique. In the UK, as in the United States, there is general acceptance that guilty pleas, encouraged by reductions in sentence, are a long-term feature of our present arrangements,19 on the basis that if the alternative is full trial, it is simply too expensive, too time-consuming or in other ways unmanageable.20 From this perplexing start, which has suggested a disparity between the focus for our legitimations and the practices that they are meant to legitimate, we will first consider the manner in which our judiciary justify the safety of guilty pleas incentivised through the offer of discounted sentences. Whilst lip service is paid to the reliability of guilty pleas as evidence of factual guilt, the principal basis for upholding convictions obtained through guilty pleas is by reference to defendants’ agency in voluntarily agreeing to their convictions. As we will show, however, voluntariness offers the judiciary scant means or sure footing for combating measures that place ever greater pressure on defendants to forgo their right to trial.21 We will then consider other justifications for our practice of reliance on guilty pleas, and whether the values that are used to legitimate them might provide a more accurate normative account of the current criminal justice system in general, rather than those which focus on trials. In other words, this article will suggest a possible reversal of approach, with justifications for guilty pleas, rather than trials, being offered as underpinning the legitimacy of the criminal justice system. 2. Pleading Guilty to Obtain Sentence Reductions—the Internal View The practice, in the course of adjudicating on individual appeals, of supervising and justifying convictions obtained via guilty pleas, like the practice of supervising and justifying convictions obtained after full trial, is the responsibility of the criminal appeal courts, the most important of which in the UK is the Court of Appeal Criminal Division. Despite the large body of critical writing, especially in the United States, that questions both the accuracy and fairness of convictions obtained via guilty pleas,22 ironically the Court of Appeal treats these convictions as far safer than those arising from jury trials, as evidenced by the limited rights to appeal available to those convicted by their plea. It is rare for a guilty plea conviction to be successfully appealed, although it may be appealed where it is either a nullity or unsafe, or where ‘a clear injustice has been done’.23 These grounds are not independent of each other. A plea will be treated as a nullity if it can be shown to be either equivocal or involuntary. Where it is not, it will be treated by the Court of Appeal as an acknowledgement of guilt by the defendant,24 giving rise to a definite presumption that the conviction is safe. That presumption operates in a manner close to25 a bar on appeals, given the difficulties of rebutting it. Evidence exonerating a convicted person would suffice, but irrefutable proof that a convicted person could not have committed a crime is a relatively rare event.26 In the absence of such evidence, any appeal focuses on the quality of the guilty plea, to see if there is any reason why it should not be treated as a voluntary confession of guilt. To quote Taylor on Criminal Appeals: an appellant who wishes to challenge a plea must show either that her plea was involuntary, or equivocal (ie she pleaded guilty without understanding the nature of the charge or without intending to admit that she was guilty of what was alleged), or based on fundamentally mistaken advice.27 Compared with an appeal against conviction after trial, the scope for an appeal based on new evidence following a guilty plea is severely restricted, as the evidence, unless showing or approaching exoneration, has to indicate that the plea was not intentional and voluntary. New psychiatric evidence may reveal that the appellant did not understand what they were admitting to, and it may also support appeals based on defences going to mental capacity which were not raised at trial, such as a defence of diminished responsibility.28 But other kinds of new evidence will not alter the status of the guilty plea as a voluntary admission of guilt, or the tremendous weight placed upon such a ‘voluntary admission’: [It is] highly relevant to the issue of whether the conviction was ‘unsafe’ that the appellant had been fit to plead, had known what he was doing, had intended to plead guilty, and had done so without equivocation after receiving expert advice.29 In S v Recorder of Manchester, Lord Morris acknowledged the role of guilty pleas as an alternative to evidence functioning as a basis for conviction: ‘Guilt may be proved by evidence. But also it may be confessed.’30 The justification for this approach to the safety of convictions following guilty pleas was articulated by Lord Hughes in R v Asiedu.31 In that case, a defendant appealed against his conviction following a guilty plea at his retrial, on the basis that the plea was entered without the benefit of evidence that ought to have been disclosed to the defence, which evidence threw doubt on the credibility of the prosecution’s principal scientific witness. In his discussion of the relevance of the guilty plea, Lord Hughes made the following three statements: A defendant who pleads guilty is making a formal admission in open court that he is guilty of the offence … once he has admitted such facts by an unambiguous and deliberately intended plea of guilty, there cannot then be an appeal against his conviction, for the simple reason that there is nothing unsafe about a conviction based on the defendant’s own voluntary confession in open court.32 Of course a defendant who is confronted by a powerful case may have difficult decisions to make whether to admit the offence or not. He will of course be advised that if he does plead guilty that fact will be reflected in sentence, but that general proposition of sentencing law does not alter his freedom of choice in the absence of an improper direct inducement from the judge … He will always have it made clear to him that a plea of guilty, should he choose to tender it, amounts to a confession. Only he knows the true facts, which usually govern whether he is guilty or not and did so here.33 … the trial process is not a tactical game. A defendant knows the true facts; he ought not to admit to facts which are not true, whatever the evidence against him, and this will always be the advice he is given. If he does admit them, the evidence that they are true then comes from himself, whatever may be the other evidence advanced by the Crown.34 In the first of these statements, Lord Hughes asserts that there is nothing unsafe about a voluntary guilty plea. Whilst unsafe includes matters going to a fair trial as well as the truth of a defendant’s guilt, this assertion is still a claim that a guilty plea is ordinarily determinative of the issue of factual guilt. Indeed, the plea so determines the factual guilt of the accused that the standard of wrongdoing required nevertheless to make it unsafe is stated to be the same as that identified in Mullen,35 namely an abuse of process that could not be remedied by a trial or retrial.36 The credibility claimed for the plea in Asiedu does not seem to have been altered by the view expressed on the defendant’s separate appeal against sentence, though clearly shared by this court, that the defendant was ‘a liar on an epic scale’.37 In the second statement, Lord Hughes acknowledges the factors that may have influenced the defendant to plead guilty, but counters these with a reminder that the choice to plead guilty is a choice of the defendant, and that such a plea will, at law (whatever its credence outside the legal system), be treated as a confession of guilt. Whilst the second statement stresses the defendant’s freedom of choice, the third focuses on the defendant’s responsibility for the situation arising from his plea. The defendant ought not to admit facts that he knows to be untrue in order to seek a tactical advantage (here, a reduction in sentence). The implication is that, having been the author of his own misfortune by telling deliberate untruths, he cannot now complain about the conviction that results. Of the three justifications, the claim that a voluntary guilty plea determines a defendant’s factual guilt is the most problematic. Or at least, it is if one interprets this as a statement about pleas as a form of evidence. As already pointed out, the defendant in Asiedu was known to be an inveterate liar, whose story altered in order to seek tactical advantages over the course of the investigation and the two trials, attempting, in turn, to avoid conviction, be charged with a lesser offence and seek a lower sentence. His testimony as a witness against his co-defendants would clearly be suspect, and there was no reason to expect his own testimony, tailored to whatever he saw as his own advantage, to be any less so. As such, if his plea represents a confession, it is a rather unreliable one. More generally, the history of acknowledged miscarriages of justice is littered with false confessions.38 And whilst voluntary confessions are ordinarily regarded as a strong form of evidence when examined within a trial, they are nevertheless not assigned that kind of weight, as a form of evidence, in the quotations here. There are different reasons than evidential ones, which come into focus with guilty pleas. The second statement contains a somewhat oblique acknowledgement of the major reason why guilty pleas should be regarded as a suspect form of evidence: the sentence discount ordinarily available to those who plead guilty. Asiedu was seeking all three versions of the discount—a lesser sentence for the offence charged, a lesser charge and an agreed version of the facts to be placed before the judge, intended to downplay the seriousness of his involvement. Whilst his co-defendants had already received life terms for conspiracy to murder without prospect of parole for 40 years, Asiedu was told that if he pleaded guilty his maximum sentence would be a fixed term of 40 years for the lesser offence of conspiracy to cause explosions likely to endanger life, with the possibility of further reductions on the basis of mitigating circumstances. His eventual sentence of 33 years meant that he could hope for parole after half that term, and would have a right to be freed after 22 years.39 Deciding whether to plead guilty when this may amount to spending some 23.5 years less in prison is, as Lord Hughes admits, ‘a difficult decision’. It is the kind of decision that could lead an innocent person to plead guilty. It is, as affirmed by the judgment in this case, not a decision that can be remedied if, at a later date, evidence is revealed that weakens the prosecution case, even if, as occurred in Asiedu, that evidence is something that ought to have been disclosed to the defence prior to the plea (unless the non-disclosure is the kind of deliberate wrongdoing that constitutes an abuse of process justifying the proceedings being permanently stayed). So, not only are guilty pleas that are rewarded by sentence discounts suspect as a form of evidence,40 they also operate to relieve the criminal justice system from undertaking the task of reviewing the rest of the evidence against the accused.41 There will be no assessment of the prosecution case to see if, with or without this confession, there could be guilt beyond reasonable doubt. Nor can there ordinarily be an appeal if matters come to light which indicate that the prosecution case was weaker than the defendant was led to believe, such as undisclosed or new evidence, or poor advice on the strength of the prosecution case. The first consequence operates to reinforce the second—as there is no trial or hearing at which the prosecution put their case, there is no record that can be reviewed on appeal. 3. The Ethics of Guilty Pleas—an External View If there is questionable justification for treating guilty pleas as a safe basis for conviction (that is, guilt beyond reasonable doubt), and even less justification for treating confessions associated with guilty pleas and obtained through the promise of reduced sentences as overwhelmingly reliable evidence of guilt, how should we understand them? The better explanation is that they are, in simple terms, speech acts—they execute a performance. The person who pleads guilty voluntarily and knowingly changes their normative position. By this act, they alter their status from a defendant to a convicted person. Thereafter, and as a result of this voluntary act (and not its status as reliable evidence), they will be treated as a person whose conviction has been established beyond a reasonable doubt. As Lord Hughes rightly notes in Asiedu, the defendant had been offered a choice. He chose, echoing HLA Hart, to exercise a power constructed via a power-conferring rule, and thereby altered his status, just as would occur if he had voluntarily made a will or entered into a marriage.42 Having executed a speech act43 that he has been informed will have this effect, he cannot now disclaim his new status, any more than the groom or bride who reluctantly undergoes or regrets a marriage can claim to be single.44 From the perspective of the criminal process as a means of establishing the truth of a person’s guilt, allowing such a speech act to operate, especially one accompanied with an incentive for it to be exercised, is a potential mistake. And this potential mistake is not remedied (though it may be ameliorated) by ensuring that the defendant enjoys full disclosure of the available evidence45 or receives competent advice on the strength of that evidence and any available defences, and the likelihood of acquittal.46 Nor does it really worsen matters if defendants directly, or through their lawyers, enter into a process of negotiation with the prosecution to establish the exact basis on which their rights to a trial can be exchanged for immediate but discounted sentences. As long as defendants are aware that they can exchange trials for discounted sentences (and in our system it is part of their lawyers’ duties to make them aware of this), the decision whether to plead guilty or go for trial will be a response to factors that have no intrinsic evidential value. In addition, ex post judicial decisions to treat guilty pleas as evidence of remorse, entitling defendants to lesser sentences, inevitably feeds back to defendants, who can then be influenced to plead guilty in expectation of lower sentences, without any formal bargains between prosecution and defence, or any promise from judges.47 If defendants are innocent, their decisions to plead guilty include the value that they place upon the loss of their status as citizens innocent of the crimes of which they are accused. The advice that lawyers are required to give their clients prior to a plea (only to plead guilty if they are in fact guilty) assumes a world in which innocent persons prioritise the maintenance of that status over a huge loss of welfare (suffered by both themselves and their families) if they were to receive a harsher or longer sentence. Along with the implications of longer sentences for those persons’ welfare, another crucial personal circumstance is those persons’ attitudes to risk, with the risk averse being more likely to plead guilty to receive the sentence discount. Conventional wisdom claims that those who commit crimes are likely to be less risk averse than those who do not,48 so discount levels need to be set high if sufficient numbers of guilty persons are to be persuaded to forgo the possibility that a trial might lead to an acquittal.49 All other factors being equal, this suggests that an innocent person may well plead guilty in order to obtain a reduced sentence in the face of a weaker case against them than would occur with a guilty person. According to the ethicist Ruth Grant: The system is meant to ascertain the facts according to the evidence and to assign the punishment proportionate to the crime for the sake of protecting the public. [However] These purposes—truth and justice—cannot be served by a bargaining process.50 For Grant, bargaining away one’s right to a civil trial may be just, as settlement within a civil procedure can be seen as a Pareto superior situation: if only two parties’ welfare is at issue, why is it unjust if both agree to a situation (settlement) that each of them regards as preferable to the costs and risks of a full trial? One wants civil settlements, like any contract, to be the result of a voluntary agreement reached by parties who have equal access to information and the ability to form rational judgments about risk. In the context of civil litigation, this requires each side to have an informed understanding of the likelihood of succeeding at trial, which has implications for their respective access to legal advice, discovery and other pretrial procedures, etc. The community (represented by the state) has little or no independent direct interest in getting at the truth of the facts in dispute. However, if criminal procedure is intended to identify and punish those who have committed crimes, then it should neither encourage distortions to fact-finding procedures such as sentence discounts for pleading guilty nor treat guilty pleas obtained under such conditions as significant additional evidence of defendants’ guilt. Wrongful convictions not only punish the innocent; they typically end all efforts to find the persons actually guilty. For Grant, the focus on the voluntariness of guilty pleas (a common focus of UK and US appeal courts and a large US academic literature on plea bargaining) is an ethical mistake.51 If the purpose of criminal procedure is to establish truth, allowing parties to contract out of the truth-seeking process is a mistake that is not avoided, or even reduced, by making sure that they are put into a position whereby they can make a correct assessment of the risks of going to trial and the relative advantages, to themselves, of pleading guilty. Ronald Dworkin, in an essay widely cited in discussions of the normative basis for criminal procedure,52 considers the ground on which a society can compromise its commitment not to convict the innocent. He requires a proper recognition and weight to be given to avoidance of the moral harm that occurs when someone innocent is convicted. This moral harm is different and additional to any physical or psychological harm suffered by an individual who is wrongly convicted. It refers to the injustice of the conviction—the injustice that occurs when someone is dealt with other than in accordance with substantive law. Dworkin argues that the procedures offered to protect against the harm of wrongful convictions should not be inconsistent with the rights purported to be given to persons under substantive law. In particular, one cannot claim to regard wrongful conviction as a serious moral harm and then offer procedures that give no protection against its occurrence, as would be the case, for example, if one allowed every decision on procedure to be the result of fresh ad hoc cost–benefit calculations of what served the public interest. But above this minimum he concedes that the assessment of the appropriate level of protection to be offered against the risk of wrongful conviction is a matter for legislation, a term that he applies to both statutes and the antecedent announcement by courts of the rules of criminal procedure. Such legislative decision must, however, treat all persons with equal respect, by which he means that it must offer all persons the same level of protection against the risk of suffering the moral harm of wrongful conviction. His essay is usually cited to support claims that the criminal process must be structured to avoid wrongful convictions, and to refute any suggestion that wrongful acquittals are to be given the same moral weight as wrongful convictions, on the basis that the first does not involve a moral harm whilst the second does.53 If we apply Dworkin’s analysis to guilty pleas and sentence discounts, it is not clear that these practices are illegitimate. The level of protection offered to all defendants is the same. Whatever level of protection is offered by a full trial, all defendants have an equal right to insist upon it. They can only be denied trial if they voluntarily surrender this right by pleading guilty. So, there is no problem with equal respect and concern. One can disagree with the level of protection offered against wrongful convictions, and claim that there should be fewer guilty pleas and more trials, but this is the sort of judgment that, for Dworkin, can be settled by legislation. The only part of Dworkin’s analysis that comes close to criticising the several varieties of guilty plea systems is his concern that the level of protection afforded against wrongful conviction might be set at a lower level than is warranted because the risk of such moral harm is unequally distributed. A majority who cannot see themselves facing prosecution might settle for lower-cost, less-accurate criminal procedures than they would require if they were members of a more vulnerable minority. In our society, the risk of facing prosecution is not equal, but differs by reference to such factors as class, race and sex. For Dworkin, this would not matter unless the difference in the risk of prosecution amounts to an increase in risk that is ‘great for individuals’. He does not regard the different vulnerability of individuals, based on the different implications for them of suffering criminal penalties (having families, being scared of prison, etc), as relevant, as these are not peculiar to any specific minority group. He concludes: So even in the real world majoritarian decisions that fix a particular level of accuracy in criminal decisions in advance of particular trials, through the choice of rules of evidence and other procedural decisions, can be faulted for serious unfairness only if these decisions discriminate against some independently distinct group in one or another of the ways just canvassed. It is not enough, to make these decisions unfair, that they put one rather than another value on moral harm of different sorts, so long as this valuation is consistent and unbiased.54 Thus, the vulnerabilities that make individuals likely to choose to plead guilty in order to receive lower penalties than they would receive if convicted at trial are not for Dworkin—at least, not in this essay—a factor that alters the appropriateness of this procedure, with its lowering of accuracy and saving of costs. Individuals are treated equally. There may, however, be an issue arising from the different impact of the guilty plea procedure on black defendants, if this group (as evidence seems to show) more commonly refuse to plead guilty and, as a result, suffer from longer periods of imprisonment.55 However, this is not a result of less resources being put into preventing this group from suffering from wrongful conviction, but perhaps the converse—their disposition to refuse to plead guilty means that more resources are being put into avoiding wrongful convictions with this group than with the general population.56 Moving on from Dworkin and questions of moral harm, what of the status of guilty pleas as voluntary acts? They are constrained by the fear of additional punishment, with prison of course considered as an archetypal example of punishment. But is the guilty plea coerced? Normally, one would only describe something as coerced if the harm that influences a person to make a particular choice is impermissible.57 Difficult choices—including choosing between alternatives that one would rather avoid—are not by reason of being undesired, or even harmful, thereby coerced choices. Nor does one appropriately describe a choice as ‘coerced’ when what is being offered is an opportunity that one would not otherwise have. And the basis for assessing the nature of a choice is the situation faced by the chooser at the time that they choose.58 Applying these principles, one needs to begin by considering the position of defendants who are not going to be offered the opportunity to plead guilty and thereby obtain lesser sentences. So long as there is evidence that justifies bringing prosecutions, our defendants cannot complain that they are suffering a wrong in being prosecuted. The prospect of punishment if convicted may be painful to them, but provided the punishment that is expected will be one that is appropriate to the crime and thus deserved, then, again, they cannot complain of a wrong. The accusation that plea bargains are rendered involuntary due to pressure is often based on the claim that the sentences applied to those who are found guilty at trial will be excessive and undeserved. So, for example, mandatory sentences create situations which, viewed from outside the legal system, must sometimes seem harsher than may be warranted (as the culpability of those who meet the conditions of any single offence will vary for various reasons in their particular circumstances). But, from inside the legal system, any sentence prescribed by an authoritative legal source, such as the legislature, cannot easily be described as ‘inappropriate’ or ‘unduly harsh’.59 For persons offered the opportunity to forgo their right to trial in exchange for a sentence less than that which would occur if convicted after trial, the introduction of these opportunities does not transform the prospect of the sentences that could otherwise occur after trial into a form of coercion. Nor does this analysis alter if the difference between the sentence likely after trial and that offered in exchange for a guilty plea is large, or even gargantuan. The fact that the difference between these choices creates a significant incentive to choose one of them does not make the choice coerced, and thereby involuntary. Even if the attractiveness of one over the other makes the selection of the preferred option a ‘no brainer’, the choice is still not thereby either coerced or involuntary. We all make regular ‘no brainer’ choices in favour of things we prefer, without these constituting coerced or involuntary decisions. This understanding of what makes a choice coerced forms the basis of the House of Lords analysis of plea bargaining in McKinnon v United States.60 There, the court considered whether it would be an abuse of process for the defendant, a computer hacker, to be extradited to the United States, where he would be subjected to pressure to plead guilty. The US prosecuting authorities had outlined the deal that he could expect. On pleading guilty, he would be likely to receive a sentence of 3–4 years, with a promise to support his application for repatriation after six months to one year, which would most probably be granted, allowing him to serve the balance of his sentence in the UK. If he insisted on his right to trial and was found guilty, the US authorities would increase their estimate of the damage done by his hacking to justify a sentence of between 8 and 10 years and would oppose any application for repatriation, with the result that, with only the possibility of a 15% remission of sentence for good behaviour whilst in prison, he would spend all of this period in the US, most probably in a high-security prison. Lord Brown, speaking for all of the Lords, could see no undue pressure in the disparity between the alternatives. The threat represented by the sentence at trial was lawful—it was simply what a person prosecuted for this offence in the United States would be likely to receive if all of the relevant facts were proved at trial. If a description of what was likely to happen at trial contained nothing that was unlawful, then a promise of something less than what conviction at trial would merit in exchange for a guilty plea could not constitute undue pressure. Duress sufficient to render a guilty plea involuntary is here equated with the threat of an illegal action.61 The limited nature of the protection against incentives to plead guilty offered by the insistence that guilty pleas should be voluntary is further evidenced by the US Supreme Court decision in Bordenkircher v Hayes.62 There, the prosecutor had offered the defendant the prospect of a 5-year sentence for an offence that carried a maximum penalty of 10 years. The defendant, who had two previous felony convictions, refused the offer. The prosecutor responded by threatening, if the offer was not accepted, to bring a further charge under the state’s habitual criminal statute, which provided for any person convicted of a third felony to receive an automatic life sentence. The US Supreme Court took the view that if defendants could lawfully be persuaded to plead guilty through promises of more lenient treatment, it was not possible to prevent prosecutors from threatening to pursue a more severe penalty if the defendant refused to plead guilty. The promise and the threat were simply different sides of the same coin. Provided there was evidence to justify bringing a more serious charge or a more reprehensive version of the facts involved in the original charge, prosecutors were entitled to use their discretion to frame the charges to ‘persuade’ a defendant to plead guilty. The threat (5 years if you plead guilty or life if you are convicted at trial) was not a denial of the defendant’s right to trial, nor a reason to undo the convictions of those who elect to plead guilty in response. What, then, are the implications of these courts’ views of coercion for reform proposals aimed at restricting the use of guilty pleas, or some of their potentially undesirable effects? For example, in order to avoid persons with weak cases against them pleading guilty in response to the pressure of sentence discounts, one might wish to restrict the right to plead guilty to those cases where there is strong or overwhelming evidence of guilt. By denying the guilty plea procedure to those who have weak, or at least weaker, cases against them, we could perhaps decrease the likelihood that the discounts given for guilty pleas will result in convictions of persons who might, had they insisted on their right to trial, have been acquitted (thus reducing the number of wrongful convictions likely to result from guilty pleas). But then we have problems of perceived injustice in sentencing. For we could well end up with a situation where persons who have the strongest evidence against them systematically receive lower sentences and those with weaker cases against them are denied the same ‘opportunities’.63 4. What Values Are Being Promoted in the Leading Decisions on Guilty Pleas? McKinnon and Bordenkircher are perfectly good decisions, if we are content to allow defendants to convict themselves through their own voluntary guilty pleas. By contrast, the leading decisions claiming to identify ‘improper pressure’ are less well reasoned. Whilst they are articulated in terms of the need to protect the voluntariness of defendants’ choices whether or not to plead guilty, these decisions are, we believe, more accurately understood as preserving the authority of judges to determine appropriate sentences.64 In Turner,65 the guilty plea conviction was quashed on the basis that the defendant believed that he had been told the sentence that the judge was minded to impose if he pleaded guilty, as well as the harsher sentence likely to be imposed if he was convicted by the jury. This belief was said to deprive the defendant of his ability to make a free choice of whether or not to plead guilty. In the more recent case of Goodyear,66 the Court of Appeal, reversing Turner, decided that defendants should be able to ask the judge for an advance indication of the maximum sentence likely to be given in response to a guilty plea: ‘We do not see why a judicial response to a request for information from the defendant should automatically be deemed to constitute improper pressure on him.’67 But having acknowledged that receiving information regarding the maximum sentence to be awarded for a guilty plea was not duress, the Goodyear court went on to deny defendants other information relevant to their decision to plead: the sentence that would follow if they insisted on their right to trial and were later found guilty.68 As in Turner, their decision was justified by reference to the ‘improper pressure’ to plead guilty that resulted from this information. One normally assumes that increased information about the likely consequences of making choices is something that improves the position of autonomous actors who wish to make decisions in their own best interests.69 Thus, in the situations facing the defendants in Turner and Goodyear, learning what the judge was minded to do should be considered an advantage, rather than something obviously ‘improper’. In place of advice as to the sentences that would follow from the two forms of conviction, advice that, even at its best, retained elements of guesswork, defendants would know with certainty what would follow from their decisions to plead guilty or to opt for trial. Yet, in these cases, it was the extra element of certainty, which was described as a form of duress, that made the decision to plead guilty ‘involuntary’. The court in Goodyear sought to distinguish Turner on the basis that a ‘different culture’ existed in 1970. But this different culture (new sentencing guidelines formalising the discounts for guilty pleas and new pretrial procedures that provide express opportunities for pleading guilty) did not alter the duress or lack of it that had arisen in Turner; it simply made it impossible to claim that judges should always remain aloof from the processes that lead to defendants exchanging their right to trial for a sentence discount. The justification offered for both of these decisions assumes that information, or at least some of the information received from the judge on any sentence discount, has a particularly deleterious effect on defendants’ autonomy, regardless of the surrounding circumstances, such as the amount of discounts normally given for guilty pleas, the guidelines that direct the giving of such discounts or the manner in which such information will be incorporated into the advice given by the defendants’ counsel. Why does it undo the voluntariness of defendants’ choices (their freedom to choose) for defendants to learn with certainty the maximum sentence that can be imposed on them if convicted at trial, so long as they have access to lawyers who can advise on the likelihood that their trials will result in them actually receiving these maximum?70 And why should judges’ advice on the maximum discount that could be achieved be more important than the size of that discount, either as a percentage of the maximum or as an absolute number of additional years in prison? The Goodyear court also claimed that a judge could only inform the defendant (via counsel, if the defendant was represented) of the maximum sentence that would follow a guilty plea if the judge were asked for this information. Again, this was justified in terms of duress. It was not duress for defendants to be given this information where it was solicited, nor was it duress for a judge to ask, even in open court, if defendants’ counsel had raised the possibility of learning this information with their clients. But the extra step of informing defendants without being asked would amount to duress, giving rise to a possible successful appeal, on the basis that the guilty plea was not voluntary.71 What, we surmise, was really happening in Turner, and even more clearly in Goodyear, is that the courts were using the semantics of the need to protect the voluntariness of a defendant’s decision to plead guilty in order to preserve judicial sentencing discretion. If one wishes to maximise the sentencing discretion of the judiciary, then one needs to keep them out of plea bargaining. If the judges do not make promises, they (and any judge assigned to the case later) cannot be held to them. Goodyear indications will only serve their purpose if indications once given can be relied upon. Accordingly, and at least save exceptionally, indications thus given are binding in as far as they go, hence the need for circumspection before they are given. Particular caution is warranted where a Goodyear indication is sought in the case of a specified offence which might attract an extended sentence or a sentence of imprisonment for public protection.72 This allows judges to refuse to implement discounts promised by others (prosecutors and the police) that do not do justice to the severity of the crimes for which defendants are convicted. In Turner, the court attempted to keep the judiciary completely free from the restraints that would follow from their involvement with plea negotiations, whilst in Goodyear the judiciary sought to ensure that any plea bargain would not restrict their freedom to impose sentences after trial (where more information relevant to sentences was likely to have become apparent). Both Turner and Goodyear are understandable in terms of a judicial desire to retain sentencing discretion, but neither are plausible in terms of their claim to identify kinds or levels of duress that render the ordinary incentivised guilty plea ‘involuntary’. A caveat needs to be entered here. Although the limitations placed on judicial involvement in Turner and Goodyear may not make sense as claims about what makes defendants’ pleas voluntary, this does not mean that such limitations are unwelcome. They still offer impediments to the increasing use of sentence discounts, with their distorting effects on the accuracy of convictions. A system like that in the United States, where prosecutors bargain with defendants over the charge, the discount on the charge and the facts, is made easier by the involvement of judges in the process.73 The majority of US plea bargains are given effect as a recommendation to a judge of the sentence agreed with the defendant. Judges are not bound by the recommendation, and defendants cannot change their pleas even where they receive a higher sentence than the recommendation. But the judiciary’s awareness, or at least belief, that their ability to process criminal cases depends upon these bargains being struck creates a situation in which judges are de facto bound by what has been agreed in all but the most exceptional cases.74 Otherwise the ability of prosecutors to secure those plea bargains will be undermined. In this manner, judicial involvement in US plea bargaining forms part of a system in which prosecutors are commonly able to offer defendants up to a 75% reduction in their sentence if they plead guilty. One might therefore conclude that, whatever the reasons for it, not allowing judges to become bound by the agreements reached between prosecutors and defendants is a good thing (if one wants safe convictions) whatever the quality of the semantics used to justify it. 5. Reconsidering the Legitimacy of the Criminal Justice System in View of Guilty Pleas Ian Dennis, in his discussion of what makes criminal justice legitimate, draws on the necessity to treat defendants as autonomous agents: ‘law as a special means of subjecting human conduct to the governance of rules necessarily adopts a specific conception of moral personality, one which respects the autonomy of those subject to the legal order’.75 This respect for autonomy, and for the role of the defendant as a choosing actor, rather than a passive object, is linked by him to more general ideas of natural justice. The idea … is that the best justification of natural justice lies in the moral requirement of respect for the autonomy of the party affected by a decision; this leads to using a process which gives him a role in it and can thus explain to him the basis on which the application of the rule or policy to him is justified.76 A similar emphasis on defendants’ autonomy, and their active engagement in the procedures that determine their guilt or innocence, can be found in Herbert Packer’s classic description of what constitutes ‘due process’ rather than ‘crime control’: Finally, there is a complex of assumptions embraced within terms like ‘the adversary system’, ‘procedural due process’, ‘notice and an opportunity to be heard’, ‘day in court’, and the like. Common to them all is the notion that the alleged criminal is not merely an object to be acted upon, but an independent entity in the process who may, if he so desires, force the operators of the process to demonstrate to an independent authority (judge and jury) that he is guilty of the charges against him. It is a minimal assumption. It speaks in terms of ‘may’, not ‘must’. It permits but does not require the accused, acting by himself or through his own agent, to play an active role in the process; by virtue of that fact, the process becomes or has the capacity to become a contest between, if not equals, at least independent actors. Now, as we shall see, much of the space between the two models [of social control and due process] is occupied by stronger or weaker notions of how this contest is to be arranged, how often it is to be played, and by what rules.77 With guilty pleas, the autonomy of defendants is particularly acute. Whilst many of the decisions attributed to defendants are made by their lawyers on their behalves, the plea is one that the courts insist must be the defendant’s own: ‘The accused, having considered counsel’s advice, must have a complete freedom of choice whether to plead guilty or not guilty.’78 Those who seek to provide a normative account of the criminal justice system through a focus on the trial, and in particular the jury trial (the kind of trial that those accused of the most serious crimes have, or have a right to demand), present the rights afforded defendants, and the values of autonomy that they embody, as restraints on the pursuit of factually accurate verdicts. Rights are presented as mechanisms that give priority to the avoidance of erroneous convictions over erroneous acquittals. Where they cannot be explained in terms of the avoidance of erroneous conviction, they are justified in terms of the need to ensure that defendants are only convicted through mechanisms regarded as fair. And, in the case of human rights, these may be seen as additional requirements that need not contribute to either accurate convictions or ideas of fairness. Whatever the strengths of such normative accounts of criminal justice with their focus on the trial, this kind of theorising fails to represent those parts of the system that lie either side of the trial: pleas and appeals. In these procedures, respect for defendants’ rights, and autonomy, does not operate as a supplement to the pursuit of accurate verdicts, but as an alternative. In the context of appeals, the conviction may be factually correct or it may not be, but the defendant has had their rights, and having made their decisions either through their lawyers or, in the case of the plea, in person, they cannot now complain that the result of the exercise of those rights has been a conviction.79 The focus on rights and autonomy provides a semantic that operates to restrict the review of evidence undertaken on appeal from trial verdicts, and eliminates almost any possibility of a review of evidence in the case of convictions obtained via a guilty plea. Defendants cannot, on appeal, seek to provide evidence or make arguments that were available to them at trial. They cannot disown their decisions made on the advice of their lawyers unless they can show that advice to have been flagrantly incompetent. And, if they choose not to exercise their right to trial, they cannot now disclaim the meaning that law gives to a guilty plea: that these defendants are factually guilty. The semantics of autonomy provides the courts of appeal with a mechanism for not overturning convictions. It enables them to have a workable relationship with the first instance courts and the criminal procedures that they supervise. The usefulness of such semantics, to courts of appeal, can be seen if one considers the implications of dispensing with the guilty plea, or requiring appeal courts to investigate the safety of convictions obtained through guilty pleas motivated by sentence discounts. With over 90% of convictions being the result of guilty pleas, the appeal courts (and in particular the Court of Appeal Criminal Division) are in no position to declare most, and in practice nearly all, voluntary guilty pleas unsafe. Nor are they going to be able to review the safety of every conviction obtained in the hope of securing a sentence discount. In turn, magistrates’ courts and the Crown Court are not going to be able to manage their current workloads without accepting guilty pleas.80 Also, it is not possible to retain both our current mitigation practices and eliminate sentence discounts for guilty pleas. Whilst it may be naive to believe that all those who plead guilty are genuinely remorseful, it would still appear harsh to treat those who proclaim remorse and accept guilt at the earliest opportunity in the same way as those who insist on their innocence up until the moment that they are found guilty. Recognising this difference through sentence discounts will continue to offer an incentive to plead guilty that diminishes the safety of convictions obtained through guilty pleas. In this situation, the presentation of criminal justice as a system that gives priority to the avoidance of wrongful conviction through procedural fairness and rights, or supplements that commitment with the functioning of human rights, itself misrepresents the criminal justice system. Whilst there are features that support claims that wrongful conviction (convicting the factually innocent) is prioritised over wrongful acquittal, the overwhelming majority of convictions are produced and retained through operations and communications that exhibit a commitment to autonomy that is given priority over factual accuracy. In Dworkin’s essay on the ethics of criminal procedure, he argued that one could critique attempts to reduce the level of protection offered against wrongful conviction where this represents a diminution of the commitment to avoid convicting the innocent embodied within the historically established practices. This, he admitted, is a conservative principle; but one could also dispense with practices, even ancient ones, that are ‘islands of inconsistency that could not be brought within any justification that attaches the level of importance to the injustice factor in the mistaken conviction that is necessary to explain the rest of the law’.81 To continue his metaphor to us raises the question: what is the sea and what are the islands? The answer is now clear. The overwhelming majority of convictions, achieved though guilty pleas and incentivised by sentence discounts, if justifiable, are justified both in theory and practice by reference to the concepts of autonomy and the autonomous exercise of rights, at the expense of a commitment to the truth of the conviction or the avoidance of wrongful convictions.82 Acknowledgement We wish to thank our colleagues Kate Malleson and William Wilson, and the two OJLS referees for their insightful comments. Footnotes 1 William Aukland, Principles of Penal Law (2nd edn, London 1771), quoted in Albert Alschuler, ‘Plea Bargaining and Its History’ (1979) 13 L & Soc’y Rev 211, 216–17. 2 For a succinct explanation of why they were discouraged, see John Langbein, The Origins of Adversary Criminal Trial (OUP 2003) 18–20. 3 Andrew Ashworth and Mike Redmayne, The Criminal Process (4th edn, OUP 2010) 23. That said, we recognise that the criminal trial has been subjected to many recent challenges, described briefly as ‘The Trial Under Attack’ by the editors in Antony Duff and others (eds), The Trial on Trial. Volume 1. Truth and Due Process (Hart Publishing 2004) 3–8. Despite those challenges, the editors endorse the significance of a focus on the trial: ‘Just as the meaning of pre-trial investigation and procedure is determined by the nature of the criminal trial, so normative evaluation of pre-trial investigation and procedure is determined by a normative theory of the criminal trial’ (12). For their comprehensive account, see Antony Duff and others (eds), The Trial on Trial.: Volume 3. Towards a Normative Theory of the Criminal Trial (Hart Publishing 2007). 4 ‘The twin objects of the criminal trial are accurately to determine whether or not a person has committed a particular criminal offence and to do so fairly’: Ashworth and Redmayne (n 3) 23. 5 Ashworth and Redmayne (n 3) 48. 6 Ashworth and Redmayne (n 3) 25. 7 Ashworth and Redmayne (n 3) 48. There are many who urge in an analogous manner, eg the argument for fairness to be more than a restraint on the pursuit of truth in Laura Hoyano, ‘What Is Balanced on the Scales of Justice? In Search of the Essence of the Right to a Fair Trial’ [2014] Crim LR 4, esp 26; and, in the context of criminal appeals, Nick Taylor and David Ormerod, ‘Mind the Gaps: Safety, Fairness and Moral Legitimacy’ [2004] Crim LR 266, 267–8. 8 Ian Dennis, The Law of Evidence (6th edn, Sweet & Maxwell 2017) 56, para 2-025. 9 Dennis (n 8) 57, para 2-025. 10 Dennis (n 8); here, Dennis (para 2-026) relies on Ronald Dworkin, ‘Principle, Policy, Procedure’, first published in 1981, then updated in Ronald Dworkin, A Matter of Principle (Harvard UP 1985) ch 3, discussed below in sections 3 and 5. 11 Jill Hunter and others (eds), The Integrity of Criminal Process: From Theory into Practice (Hart Publishing 2016) 5. There are innumerable examples of comparable approaches to the widespread application of these sentiments, which focus on the trial, in the academic literature. Whilst there are differences in the scope that commentators afford to rights and fairness as restrictions on the pursuit of truth, there is no suggestion in the literature that rights might operate as an alternative to the pursuit of a factually accurate verdict—essentially the issue explored in this article. 12 In England and Wales, cases disposed of at Crown Court via guilty pleas to all charges rose from 56% in 2001, peaked at 71% in 2009, and now appear to have concentrated at around 66–67%. See Criminal Court Statistics Bulletin: July to September 2017 Main Tables, Table C3 accessed 20 September 2018. 13 In 2016–17, the CPS secured 493,331 convictions, representing 83.9% of all cases (magistrates’ courts 84.8%, Crown Court 78.9%), with 76% of its cases being disposed of by guilty pleas. The percentage of its convictions obtained via guilty pleas is therefore 91.65%. See Crown Prosecution Service Annual Report and Accounts 2016–2017 (July 2017) 4–6 accessed 20 September 2018. This figure has been confirmed to us in an email from the CPS, which concludes: ‘I would point out that the schemes which the CPS operates, such as Transforming Summary Justice and Better Case Management, are designed to ensure good legal decision making and effective case management, in turn often leaving defendants with little choice but to plead guilty in a large number of cases.’ 14 See US Sentencing Commission, Sourcebook of Federal Sentencing Statistics (2016) Table C accessed 20 September 2018. This is the figure for felonies prosecuted in federal courts. The percentage of convictions obtained through plea for state felony cases is slightly lower, at 94%. At least, this is the figure for 2006, published by the Bureau of Justice Statistics, Department of Justice, NCJ 226848 accessed 20 September 2018. On the early common law and more recent US history of plea bargaining and guilty pleas, see Alschuler (n 1). 15 The guidelines issued by the Sentencing Council in 1994 introduced a tariff of sentence reductions which started at one-third off the normal sentence where the defendant pleads guilty at the first opportunity (to plead in court proceedings, not to confess; see below), falling to a one-tenth reduction where the guilty plea occurs immediately prior to the start of trial. Once trial has started, it is a matter for the judge’s discretion what reduction up to 10%, if any, to allow for a change of plea. The current guidelines retain this approach. See Reduction in Sentence for a Guilty Plea: Definitive Guidelines (Sentencing Council 2017) accessed 20 September 2018. The reduction for pleading guilty operates independently from any personal mitigation, such as a confession to the crime when first questioned by the police. See Definitive Guidelines (ibid) Key Principles, 4. Both Lord Justice Auld, in his Review of the Criminal Courts of England and Wales (Lord Chancellor’s Department 2001), and the Runciman Royal Commission on Criminal Justice, Report (Cm 2263, 1993) expressed concern at the level of cracked trials (that is, trials of those pleading not guilty listed, but not proceeded with, whether through change of plea or for some other reason) and recommended the increased use of sentence discounts to persuade the guilty to plead guilty earlier. Mike McConville and Luke Marsh dispute the claim that cracked trials could be reduced by these incentives, with consequent saving of resources, having reviewed the evidence on the reasons why guilty pleas arise during or just before the start of trial: Mike McConville and Luke Marsh, Criminal Judges: Legitimacy, Courts and State-Induced Guilty Pleas in Britain (Edward Elgar 2014) 93–4, 98–120. 16 Ashworth and Redmayne (n 3) 26. 17 Ashworth and Redmayne (n 3) 312. 18 Penny Derbyshire made a similar point about attempts to justify criminal justice by reference to the characteristics of jury trial, when the number of these trials is dwarfed by those which take place before magistrates: Penny Derbyshire, ‘The Lamp That Shows That Freedom Lives—Is It Worth the Candle?’ [1991] Crim LR 740. The counterargument is set out by Duff and others, Volume 1 (n 3) 12: ‘Given that the trial is one of the central ways in which the rights of the accused are adequately protected, there is at least good reason to consider other aspects of the criminal justice process against the standards set out by the criminal trial, properly understood and properly theorized.’ They further suggest (esp 14–15) that the role of that consideration is not only to provide a normative theory of the dominant practice, but to offer critique and condemnation if: (i) the normative theory of trial shows that trial procedures are essential to justify conviction and punishment; (ii) it is essential that defendants have the opportunity to challenge the evidence against them; or (iii) the criminal justice process is to be understood as a communication with defendants about the wrongfulness of their conduct. 19 As acknowledged succinctly by Kennedy J in Lafler v Cooper 566 US 156, 170 (2012), ‘criminal justice today is for the most part a system of pleas, not a system of trials’. 20 By 1970, the US Supreme Court had accepted that plea bargaining was ‘inherent in the criminal law and its administration’, and not constitutionally invalid: Brady v United States 397 US 742, 751–2, 1970. In the UK, if the current right to jury trial is maintained, Ashworth and Redmayne doubt the affordability of even the introduction of a judicial oversight of guilty pleas so as to ensure that only those with weak defences plead guilty. Their proposal is to reduce the maximum level of sentence discount to 10% on the basis that, in their judgment, this would reduce the incentive to plead guilty to a tolerable level: Ashworth and Redmayne (n 3) 317. A case for introducing a significant judicial examination of the evidence prior to the acceptance of any guilty plea in the United States is set out by Candace McCoy, ‘Plea Bargaining as Coercion; The Trial Penalty and Plea Bargaining Reform’ (2005) 50 Crim Law Q 67. 21 Whilst our reliance on guilty pleas as a percentage of convictions has increased in line with US figures, the extent of the discounts required to produce these figures has remained modest in comparison with US plea bargaining practices. However, a proposal to increase sentence discounts for guilty pleas (from one-third to one-half) was made as recently as 2010: Ministry of Justice, Breaking the Cycle: Effective Punishment, Rehabilitation and Sentencing of Offenders (Cm 7972). This proposal was not shelved because of its effect on the potential safety of convictions, but principally because the government felt that it might result in sentences, particularly for sex crimes, that the public would regard as too lenient. See Julian Roberts and Ben Bradford, ‘Sentence Reductions for a Guilty Plea in England and Wales: Exploring New Empirical Trends’ (2015) 12 Journal of Empirical Legal Studies 187. The authors of that article found that 63% of those pleading guilty enjoyed a sentence reduction—by comparison with what the sentencing judge regarded as appropriate for conviction after trial—of one-third (196). 22 For the United States, see eg Albert Alschuler, ‘The Changing Plea Bargaining Debate’ (1981) 69 Cal L Rev 652, contending that plea bargaining, despite reforms, is an inherently unfair and irrational process; Stephanos Bibas, ‘Plea Bargaining Outside the Shadow of Trial’ (2004) 117 Harv L Rev 2464, arguing that the plea bargaining process requires many reforms since it ‘adds another layer of distortions that warp the fair allocation of punishment’ (2468); Stephen Schulhofer, ‘Plea Bargaining As Disaster’ (1992) 101 Yale LJ 1979, concluding that plea bargaining should be abolished. For the UK, see McConville and Marsh (n 15). 23Boal (1992) 95 Cr App R 272 (CA) 278. 24Chalkley and Jeffries [1998] 2 Cr App R 79, 94. 25 The guilty plea does not deprive the Court of Appeal of jurisdiction to hear an appeal: Lee (1984) 79 Cr App R 108. In the earlier case of Peace [1976] Crim LR 119 (refusal to quash a conviction following a guilty plea, even though evidence that the defendant could not have committed the offence had led to his receiving a pardon), a voluntary plea was virtually treated as an absolute bar to an appeal. 26 The classic UK example of exoneration is Stefan Kiszko, who had been convicted and served 16 years in prison before his release in 1992. His infertility should have eliminated him from the investigation of the sexual assault/murder case from the start, as the perpetrator had left traces of live semen on the victim. The US Innocence project, located at Cardozo Law School accessed 20 September 2018, uses DNA evidence to establish the innocence of those convicted. As at April 2018, they claim to have achieved this in 354 cases over 25 years. To understand the additional difficulties in demonstrating exoneration versus showing a conviction to be ‘unsafe’, see Paul Johnson and Robin Williams ‘Post-conviction DNA Testing: The UK’s First “Exoneration” Case?’ (2004) 44(2) Science & Justice 77 accessed 20 September 2018. One should also have regard to the interpretation given to s 133 of the Criminal Justice Act 1988, dealing with compensation for those who show, beyond a reasonable doubt, that they did not commit the crimes for which they were convicted. On this we await the Supreme Court’s decision in R (on the application of Hallam) v Secretary of State UKSC 2016/0227 (on appeal from the Court of Appeal [2016] EWCA Civ 355). 27 Paul Taylor (ed), Taylor on Criminal Appeals (2nd edn, OUP 2012) para 9.03. 28 Such appeals can be based on rulings by the trial judge or advice provided by defendants’ lawyers that the facts that they wished to raise would not amount to a defence at law. For an example of an appeal based on such advice, referred to the Court of Appeal by the Criminal Cases Review Commission but ultimately unsuccessful, see R v Evans [2009] EWCA Crim 2243. 29 Taylor (n 27) para 9.03, quoting the judgment in Lee (n 25). 30 [1971] AC 481 (HL) 501. 31 [2015] 2 Cr App R 8. 32Asiedu (n 31) para 19, emphasis added. 33Asiedu (n 31) para 31, emphasis added. 34Asiedu (n 31) para 32, emphasis added. 35Mullen [1999] 2 Cr App R 143. Lord Hughes also relied on R v Togher [2001] 1 Cr App R 33. 36 This is not to suggest that the failure to grant leave to appeal in this case was wrong. As the defendant had pleaded guilty following a retrial, the Court of Appeal was in the unusual position of having had the full benefit of those earlier proceedings to help them consider whether the non-disclosure complained of would have made any difference at trial. 37R v Manfo Kwaku Asiedu [2009] 1 Cr App R (S) 72, para 8 (sentencing appeal). 38 Timothy Evans, Lattimore, Leighton and Salih in relation to the death of Maxwell Confait, Judith Ward, the Guildford Four, the Birmingham Six, etc. For background reasons, see Gisli Gudjonsson, The Psychology of Interrogations and Confessions: A Handbook (Wiley 2003); Brian Cutler, Keith Findley and Timothy Moore, ‘Interrogations and False Confessions: A Psychological Perspective’ (2014) 18 Canadian Criminal Law Review 153; Saul Kassin and others, ‘Police-Induced Confessions: Risk Factors and Recommendations’ (2010) 34 Law & Hum Behav 3. 39R v Manfo Kwaku Asiedu (n 37) para 4. 40 Ashcroft and Redmayne argue that inducements akin to those offered through sentence reduction would, if offered by the police in order to secure a confession, render that confession inadmissible. They regard reduction of one-third of a sentence to be equal to the financial inducements discussed in JB v Switzerland [2001] Crim LR 748. Ashcroft and Redmayne (n 3) 109–12, 314–15. 41 Whilst there must be an allegation of facts which, if proven, would constitute the offence charged, the evidence that substantiates those allegations will not be put to proof. 42 For a succinct statement of this, see HLA Hart, The Concept of Law (2nd edn, OUP 1994) 27–8. 43 The classic work is that of John Searle, Speech Acts: An Essay in the Philosophy of Language (CUP 1969), but it is unnecessary for the purposes of the argument here to go into the many complications. 44 In the United States, the role of guilty pleas as speech acts voluntarily accepting the status of convicted persons has in part been separated from their role in confirming the truth that offenders have committed the offences charged. Defendants can make an Alford plea. Here, the defendant asks the trial judge to accept a guilty plea despite publicly refusing to acknowledge responsibility for the underlying crime, a plea which would be unacceptable as equivocal in English law: North Carolina v Alford (1970) 400 US 25. For an analysis of the speech acts executed by a jury verdict, see Hock Lai Ho, A Philosophy of Evidence Law: Justice in the Search for Truth (OUP 2008) 12–32. 45 Whilst the duty of disclosure operates at the time of the plea, full disclosure is not assured, nor easily remedied if the duty is breached, as Asiedu demonstrates. The disclosure duty applies to evidence available at the time of the plea, which, if the defendant is to receive a full one-third discount, must be at the ‘earliest opportunity’, normally the plea hearing. There will always be cases where more disclosable evidence will become available later, and full information about the strength of the prosecution case may well not actually be known until the end of the prosecution’s evidence at trial, by which date, for example, the credibility of the witnesses may have become more apparent. 46 Incorrect advice going to whether defendants’ factual claims, if accepted, would constitute a defence will allow those who plead guilty to appeal later on the basis that their guilty pleas were mistaken, although they will have to show that the defences that had been overlooked would quite probably have succeeded, or that there would have been a reasonable prospect of them succeeding. This is a demanding test: see eg R v Sadighpour [2013] 1 Cr App R 20. Advice merely going to the strength of the defence (eg the credibility of witnesses) will not allow an appeal against a guilty plea conviction to succeed. Even if it did, the error in this kind of advice would need to be such that no reasonable lawyer would have recommended a guilty plea in light of this evidence; something akin to what was described in R v Ullah [2000] 1 Cr App R 351, 358 as ‘a matter of very serious misjudgement’. Research shows that at least some defendants are advised to plead guilty in cases where they might reasonably expect an acquittal. In the 1970s, John Baldwin and Mike McConville included a reappraisal of 1000 sets of committal papers by a former chief constable and retired justices’ clerk. These appraisers were able to predict acquittals with a fair degree of accuracy (80%). In one-fifth of the cases where defendants were advised to plead guilty by counsel, these independent assessors took a different view of the possibilities for acquittal: John Baldwin and Mike McConville, Negotiated Justice: Pressures on Defendants to Plead Guilty (Martin Robertson 1977) 75. There appears to be some variation in the willingness of barristers to recommend trial rather than guilty pleas, with London barristers more orientated towards trial than those practising outside the capital. See Peter Tague, ‘Tactical Reasons for Recommending Trials rather than Guilty Pleas in Crown Court’ [2006] Crim LR 23. 47 This creates ‘implicit bargaining’: see Milton Heumann, ‘A Note on Plea Bargaining and Case Pressure’ (1975) 9 L & Soc’y Rev 515. 48 See Russell Covey, ‘Reconsidering the Relationship between Cognitive Psychology and Plea Bargaining’ (2007) 91 Marq L Rev 213. 49 This is the view of Stephanos Bibas (n 22). The actual situation may be far more complex, with a host of personal dispositions affecting the decision to plead guilty. Any suggestion that guilty pleas are only entered by the guilty, or that they represent a rational discounting of the relative likelihood and benefits of going to trial versus pleading guilty, are belied by the complexities of decision making revealed by cognitive psychology research. See Chad Oldfather, ‘Heuristics, Biases, and Criminal Defendants’ (2007) 91 Marq L Rev 249; Lucian Dervan and Vanessa Edkins, ‘The Innocent Defendant’s Dilemma: An Innovative Empirical Study of Plea Bargaining’s Innocence Problem’ (2013) 103 J Crim L & Criminology 1. 50 Ruth Grant, Strings Attached: Untangling the Ethics of Incentives (Princeton UP 2012) 101. Grant’s attribution of a truth-finding role to criminal justice differs from those, like Michael Gorr, who acknowledge the injustice of punishing the innocent, but do not recognise criminal justice as a process for identifying truth. For Gorr, the purpose of the criminal justice system is not truth, but deterrence. When pursuing this goal, the wrong suffered by innocents wrongly convicted under plea bargaining pressure can be compared to the wrong suffered by innocents following wrongful conviction at trial. In the former case, a larger number suffer a shorter wrongful sentence. The loss of confidence in the reliability of each conviction (Grant’s truth) does not matter. One simply quantifies the respective quantities of injustice incurred in the pursuit of deterrence and, if the quantity of wrongful injury inflicted by plea bargaining is less, with greater deterrent effect (more guilty persons convicted), it is justified. Gorr treats confidence in the truth of individual convictions as something that only affects deterrence if the public learn of the distorting effects of guilty pleas in individual cases: Michael Gorr, ‘The Morality of Plea Bargaining’ (2000) 26 Social Theory and Practice 129. Similarly, Richard Lippke considers that plea bargaining, where the incentives are not too robust, can be seen as an ethical practice on the basis that American trials may actually be worse truth-seeking mechanisms (due to unequal access to resources and the scope for influencing juries through lawyering tactics). See his final chapter, ‘Plea Bargaining and Getting at the Truth’, in Richard Lippke, The Ethics of Plea Bargaining (OUP 2011). He does not believe that UK trials or continental procedures fail as truth-seeking mechanisms to the same extent. And, following Lippke’s argument, if, like Josh Bowen (‘Punishing the Innocent’ (2008) 156 U Pa L Rev 1117), one believes that the majority of innocents are recidivists arrested and charged by a police force that operates mainly by rounding up ‘the usual suspects’, who will incur disproportionately high costs in proceeding to trial with a low likelihood of being acquitted, then plea bargaining will generally benefit those innocents. 51 See Robert Scott and William Stuntz, ‘Plea Bargaining as Contract’ (1992) 101 Yale LJ 1909; Frank Easterbrook, ‘Plea Bargaining as Compromise’ (1992) 101 Yale LJ 1969; Frank Easterbrook, ‘Criminal Procedure as a Market System’ (1983) 12 JLS 289; Thomas Church, ‘In Defense of Bargain Justice’ (1979) 13 L & Soc’y Rev 509. 52 See n 10. 53 Conceiving of victims as persons with rights to have substantive law applied against those who have committed crimes against them, and thereby suffer from the moral harm of injustice where this does not occur, complicates this equation. 54 Dworkin (n 10) 88. 55 Roger Hood found that suspects with an Afro-Caribbean background tend to plead not guilty more frequently than others and, when convicted, tend to receive longer sentences largely because they have forfeited the discount for pleading guilty: Roger Hood, Race and Sentencing: a study in the Crown Court (Clarendon Press, 1992) 125. 56 There is also the issue of whether, even if there are inequalities in the pressures to submit guilty pleas, those who choose voluntarily to plead guilty can thereafter, in Dworkin’s terms, claim that a moral harm has (i) occurred and (ii) been inflicted on them. 57 See Conrad Brunk, ‘The Problem of Voluntariness and Coercion in the Negotiated Plea’ (1979) 13 L & Soc’y Rev 527. For a helpful discussion of the nature of coerced choice in general, see Japa Pallikkathayil, ‘The Possibility of Choice: Three Accounts of the Problem with Coercion’ (2011) 11(16) Philosophers Imprint 1. 58 For the argument presented here, another helpful discussion of the nature of coerced choice is Joshua Dressler, ‘Duress’ in John Deigh and David Dolinko (eds), The Oxford Handbook of Philosophy of Criminal Law (OUP 2011) ch 11. 59 For an example of arguments challenging the voluntariness of plea bargaining, based on the claim that harsh post-trial sentences create undue pressure, see H Mitchell Caldwell, ‘Coercive Plea Bargaining: The Unrecognized Scourge of the Justice System” (2011) 61 Cath U L Rev 63. John Kleinig acknowledges that an excessive discrepancy between the sentences applied at trial and following a plea may put ‘undue’ pressure on defendants, rendering their pleas involuntary: John Kleinig, ‘The Nature of Consent’ in Franklin Miller and Alan Werthheimer (eds), The Ethics of Consent (OUP 2010) ch 1, 15. But if one assumes that the sentence at trial is not something that a defendant is entitled to avoid, and the reduction for a plea is a ‘pure’ offer and not an offer coupled with a threat, then he would only regard the plea as involuntary if the ‘baseline conditions in which the offer is made are inhumanely unacceptable’. He admits that ‘others’ (by which he means other moral philosophers) would not go this far, preferring to accept choices made under pressure without illegitimate threats as voluntary, but unconscionable: Kleinig (ibid) 16. 60 [2008] UKHL 59. 61 Lord Brown gave the facts of the Canadian Supreme Court decision in USA v Cobb [2001] 1 SCR 58 as an example of what he considered to be ‘a wholly extreme case’ that could still constitute undue pressure to plead guilty. If the standard for undue pressure is illegality, this decision offers little comfort to those who fear that developments within the English legal system might further increase the incentives to plead guilty. There, Canadian citizens who refused to cooperate with extradition to the United States had been threatened by a US judge with the maximum sentence allowed to him under the law; and the US prosecutor, interviewed on Canadian TV, had said that conditions in the prisons where the defendants could expect to serve their sentences following non-guilty pleas would include homosexual rape. These are not, in effect, threats to impose something that is illegal; homosexual rape appears to be a fact of life in federal US prisons, and the judge’s threat to impose the maximum sentence allowed by law was no different from what was being threatened by the US prosecutors in McKinnon. 62 434 US 357 (1978). 63 Distortions similar to this hypothetical situation already occur along lines of race: see n 55. Lord Justice Auld, when considering this distortion as part of his examination of the procedure for inducing guilty pleas, thought that he could see no bias or discrimination. Treating guilty pleas as the actions of autonomous individuals, he concluded that individual members of this group had to accept the consequences of their actions. See Lord Justice Auld’s Review (n 15) 440. 64 On this point see the High Court of Australia case Barbaro v The Queen [2014] HCA 2 accessed 20 September 2018. 65R v Turner [1970] 2 QB 321. 66R v Goodyear [2005] EWCA Crim 888. 67 ibid para 49. 68 In restricting the information that could be indicated to the maximum sentence to be imposed in response to an immediate guilty plea, the court followed the recommendations contained in the Royal Commission on Criminal Justice Report (n 15) ch 7, paras 50–1, rather than the approach adopted in Lord Justice Auld’s Review (n 15) 434–4. The latter suggested that: ‘The judge should be entitled, formally to indicate the maximum sentence in the event of a plea of guilty at that stage and the possible sentence on conviction following a trial …’ 69 For example, Gerald Dworkin identifies the external requirements for autonomous action with the absence of ‘manipulation, deception, the withholding of relevant information, and so on’. See Gerald Dworkin, ‘Autonomy and Behavior Control’ (1976) 6(1) The Hastings Center Report 23, 25; and generally, Gerald Dworkin, The Theory and Practice of Autonomy (CUP 1988). 70 Brunk (n 57) 546 argues that judicial involvement in plea bargaining could make the plea involuntary if the defendant would expect to be punished by the judge, for refusing to enter a guilty plea, by having a more severe sentence imposed after conviction at trial than would normally be applied to a defendant who had pleaded not guilty. If this argument justifies the finding of improper pressure in Turner, it certainly cannot justify the position adopted in Goodyear. 71 Ashworth and Redmayne (n 3) 317 also believe that plea bargaining, which involves ‘the authoritative figure of the judge’, ‘imposes additional pressure’ to plead guilty. But, as noted in Goodyear, increased information, coming from the party who will decide the actual sentence, does not increase the pressure to plead guilty (except, perhaps, for those defendants who are inclined to discount their counsels’ estimates of the differential between sentence following plea versus conviction at trial). 72R v Newman (Shane) [2011] 1 Cr App R (S) 68, para 15. This argument could be developed significantly in relation to appeals against sentence (particularly references for unduly lenient sentences) that the appeal courts have had to grapple with, which have resulted from plea bargains when an indication of sentence was provided. 73 See Nick Vamos, ‘Please Don’t Call It “Plea Bargaining”’ [2009] Crim LR 617. 74 There is an alternative formal process, whereby the agreement only binds the defendant if the judge agrees to it, allowing the plea to be withdrawn if the judge does not. According to Vamos (ibid) 625, this is only used in a minority of plea bargains. 75 Dennis (n 8) 38, para 2-011, quoting Gerry Maher, ‘Natural Justice as Fairness’ in Neil McCormick and Peter Birks (eds), The Legal Mind: Essays for Tony Honoré (Clarendon Press 1986) 114. 76 Dennis (n 8), again quoting Maher at 116. 77 ‘Two Models of the Criminal Process’ (1964) 113 U Pa L Rev 1, 8–9. 78Turner (n 65) 326. 79 Hence the necessity for particular types of new evidence, either concerning the mental state of the accused when they made their plea or of indubitable exoneration; this is the only evidence on appeal available to those who have pleaded guilty, since they cannot identify procedural errors, either prior to or in their trials, that they might otherwise have wished to identify to contest their convictions. 80 In support of this contention, consider some of the basic information of changing practices given in the Report by the Comptroller and Auditor General for the Ministry of Justice, ‘Efficiency in the Criminal Justice System’ (HC 852, March 2015–16): ‘Backlogs in the Crown Court increased by 34% between March 2013 and September 2015, and waiting time for a Crown Court hearing has increased by 35% (from 99 days to 134) since September 2013’ (Summary, para 8); ‘The number of cases entering the system is reducing, but they are becoming more complex and resource-intensive. There has been a 6% fall in cases going to the Crown Court in the last year compared to the previous 12 months … However, there has been a 12% rise in sex offence cases in the Crown Court in the last five years (from 9,178 in 2010–11 to 10,309 in 2014–15) … Prosecutions for other serious offences are also increasing … These cases can involve complex evidence, and trials with multiple defendants. The average length of a Crown Court trial increased from 11.5 hours in 2010–11 to 14.6 hours in the year to September 2015 (27%). The increase in trial length means that it would cost £44 million more to hear the same number of cases in the year to September 2015 as in 2010–11’ (Part One, 1.7) accessed 20 September 2018. 81 Dworkin (n 10) 91. 82 As with the evolution of guilty pleas, so too trial practices have evolved, with their different possibilities for miscarriages of justice. See Richard Nobles and David Schiff, ‘Trials and Miscarriages: An Evolutionary Socio-Historical Analysis’ (2018) 29 Crim LF 167. © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) TI - Criminal Justice Unhinged: The Challenge of Guilty Pleas JF - Oxford Journal of Legal Studies DO - 10.1093/ojls/gqy036 DA - 2019-03-01 UR - https://www.deepdyve.com/lp/oxford-university-press/criminal-justice-unhinged-the-challenge-of-guilty-pleas-aNdqMtwuCF SP - 100 VL - 39 IS - 1 DP - DeepDyve ER -