For at the common law . . . his fault was not to be wrung out of himself, but rather to be discovered by other means, and other men.
William Blackstone, Commentaries
If you were sitting down today to set out the principles of a good system of criminal justice, with a blank sheet of paper and all the wisdom of hindsight at your disposal, you would probably start, as I would, with the principle that nobody is to be convicted of anything unless the court is sure of their guilt. You would probably go on, as I would, to say that it is for the prosecutor to prove the case for conviction and not for the accused to prove his or her innocence. You might be surprised to learn that this particular heritage of the freeborn Briton is barely two centuries old; but the European Court of Human Rights has made it clear that the presumption of innocence is today as much a part of Continental inquisitorial systems as it is of accusatorial systems like ours.
At this point one might pause. What is to be allowed to contribute to the proof of guilt? Previous convictions, for example? Few things can point more tellingly to the likelihood of guilt than the fact that the accused has committed a similar crime half a dozen times before. The reason such evidence is generally excluded is not that it is irrelevant: it is that it is so relevant that it is likely to eclipse everything else in the case. But because of the real possibility that it is only the defendant’s record that has caused him or her to be singled out for suspicion and prosecution, or to eliminate mistake on the prosecutor’s part by showing method on the defendant’s, from time to time such evidence is admitted, and we might want to adopt both the rule and the exceptions. A rigid inclusionary or exclusionary rule would inevitably create injustices.
Then how about the accused person’s silence, whether at interview or in court? The law now allows this, too, to contribute to the proof of guilt, provided juries are given strict warnings first to eliminate any innocent explanation for the silence and then to ensure that there is other credible evidence of guilt. This development has been intelligibly contested by advocates of civil liberties, but its best justification is that it probably does no more than corral within safe bounds something that the common sense of juries has always led them to do.
Meanwhile, still writing on the blank sheet, one would have to turn to things the accused himself has said which point to his guilt. (Both for convenience of syntax and in recognition of reality, my paradigmatic defendant is a man.) An admission of guilt is about as significant as evidence gets. But, like a string of previous convictions, it can mislead. It may have been made in fear or distress in order to put an end to an ordeal; it may have come from a compulsive confessor; it may have been made in the hope of securing bail or facing a reduced charge. Well within the lifetimes of many in the United Kingdom, it may have been extracted by brutality or simply fabricated. So one would certainly put into the system the safeguards now spelled out in the codes appended to the Police and Criminal Evidence Act (1984) to ensure that police interviews are conducted without oppression or improper inducement, that live recordings of them are made and that courts have the power to exclude admissions improperly obtained.
But why only a power? Why not a duty to exclude such evidence? A duty of exclusion seems to follow straightforwardly enough, not least because it will deprive police misconduct of any reward. The problem is the unauthorised phone tap or raid or random search, perhaps undertaken mistakenly rather than maliciously, which turns up damning evidence of serious crime. What principle forbids a society to use such evidence to prosecute wrongdoers? The easy answer – the rule of law – turns back on itself once it is accepted that the detection and prosecution of crime are also part of the rule of law. The answer arrived at not only by appellate courts throughout the common law world but also by the European Court of Human Rights is that there is no principled answer. If you want a principle, it has to be either the common law’s historic view that evidence is evidence no matter how it is obtained – a licence and an encouragement to the authorities to break the law – or the bald exclusionary principle adopted in 1961 by the United States Supreme Court and since then under almost constant siege. Both the United Kingdom and the European Court of Human Rights have settled into an uneasy position between the two poles, recognising that while in some cases the breach of legality will be so marginal as not to matter or so serious that it cannot decently be overlooked, in others it has to be painfully weighed against the importance of the evidence it has produced. The persisting difficulty is that there is no legal calibration of the scales.
But the common law itself always made one crucial exception: the rule that it did not matter how evidence had been obtained did not apply to confessions. Here judges historically have taken it on themselves to exclude unfairly obtained admissions of guilt; and this self-conferred power has been raised to a higher-order principle by a statutory requirement in the Police and Criminal Evidence Act to exclude confession evidence which the Crown cannot prove to have been obtained in circumstances which cast no serious doubt on its reliability. I doubt, in the light of the bitter judicial experiences of recent decades, whether one would want the model system to retreat an inch from this position.
So we have reached a position in relation to routine police interviewing in which self-incrimination is acceptable because – and only because – it cannot be used in court unless it has demonstrably occurred in risk-free conditions. Those conditions include voluntariness, but the suspect may be volunteering an explanation because he has been warned that an adverse inference may be drawn from unexplained silence, and to that measured extent there is pressure to speak. For my part I do not find this morally or ethically repugnant, and as a trial judge I encountered no evidence (apart from the still unresolved problem of suspects whose solicitors advise them without good reason to remain silent) that it worked injustice.
With self-incrimination now strictly monitored where it matters most, in the police station, you might wonder what is left to worry about. The answer is quite a lot. Let us go back for a moment to the ideal system. We can agree that nobody should have to account for themselves simply to satisfy an inquisitive official, and therefore that nobody’s refusal to do so should be taken to connote that they have been up to no good. This much we can ascribe to the fundamental right to be let alone; but that right has nothing directly to do with self-incrimination: it is dealt with by the larger and different right of silence. We may also be able to agree that, in cases where officialdom has good grounds for suspecting you of an offence and tells you what these grounds are, not only what you say but what you don’t say in response may be relevant at trial.
But a moment’s reflection suggests that there is more to life in a developed democracy than this. A large number of private activities regulated by the state, albeit with the back-up of criminal sanctions, depend on the honesty and self-discipline of those concerned: financiers handling large sums of other people’s money, or drivers who like a drink. Such people from time to time may be required to account for themselves either to public officials or to opponents in litigation. What is to happen when such a person, compulsorily answering entirely legitimate enquiries from someone who up to that point had no particular reason to suspect them of crimes, or whose suspicions lacked proof, makes an incriminating admission? In bare principle, you could take one of three attitudes. You could say that both the question and the answer are writ in water because there is a fundamental principle that people cannot be required to incriminate themselves. Or you could recognise that the regulatory regime has a legitimate need for answers to such questions but prohibit the use of these answers in court. Or you could decide not only that there is a legitimate need for answers but that if these answers afford proof of criminality they should be put before a jury like any other evidence.
The first of these approaches, the total exclusion of incriminating questions, has the virtue of universality and of apparent simplicity. I say ‘apparent’ because in practice there are few questions to which an incriminating answer is not possible, and it is frequently only the person being questioned who knows whether the answer will in fact incriminate him. It also has the vice that, an old and never very watertight vessel, it is today leagues adrift from its anchorage and listing badly. The third approach, total admissibility, dovetails with the central purpose of particular legal regimes backed by criminal sanctions: to be able to prosecute people who abuse a privileged position. Its downside, at least where the equivalent of PACE procedures is not in place, is the risk of oppression and malpractice in pursuit of admissions. The second approach – you can ask the question but can’t use the answer – has the vices of both and the virtues of neither. It does nothing to protect the innocent from oppression or therefore from unjust administrative sanctions: to do that, strong procedural controls are needed. But by shutting out every forensic use of incriminating answers obtained under legal compulsion, however careful and controlled the procedure by which they have been obtained, it protects the guilty from conviction. Even so, I shall have to return to this eccentric proposition because, remarkably, it is the impasse into which the modern law of human rights in Europe has been driven.
To see how this has come about it is necessary to unravel a long skein of history. As often happens, a significant part of this history consists of the participants’ own beliefs about it. Blackstone held that the privilege against self-incrimination was an inalienable part of the Englishman’s heritage: ‘No man is to be examined to prove his own infamy.’ It is found as a judicial maxim as early as 1568, stated plainly by Chief Justice Dyer on behalf of the Court of Common Pleas. The redoubtable John Lilburne, at his trial for high treason in 1649, said to the court: ‘By the laws of England I am not to answer questions against or concerning my selfe,’ and Justice Keble reassured him: ‘You shall not be compeld.’ Dalton’s Countrey Justice in 1618 claimed it as a maxim of the common law, and Barlow’s Justice of the Peace in 1745 asserted that, by keeping the accused out of the witness box, the maxim showed the law of England to be a law of mercy. In fact there was nothing peculiarly English about it: it was a widespread legacy of the medieval jus commune and had roots deep in the law of the early Church in Europe and the near East. In today’s world it has become a shield which protects corporations from having to divulge their own wrongdoing to the state, by whose licence they exist, and an elephant trap for public agencies trying to combat crime. How has this happened?
It was in 1898 that English juries became able for the first time to hear an accused person’s evidence in his own defence. Until that date the common law had considered anyone accused of felony incompetent as a witness on his own behalf. He could speak from the dock in order to question the Crown’s witnesses, call his own, make an unsworn statement and argue his case to the jury. Many defendants had made brave and effective use of these limited rights, though many more had watched their fate unfold in frightened silence. But until the 18th century defendants in criminal trials, however wealthy, were on their own: not even the great men whose downfall is recorded in the early State Trials volumes had lawyers representing them, although they might be allowed to consult counsel if points of law arose. If they did not speak nobody spoke for them, so that a right of silence meant little or nothing in court; and with this came the pressure to speak and make admissions, as often as not under direct questioning from the court. Indeed, until the development during the 18th century of the modern concept of the burden of proof, criminal procedure was essentially a dialogue between the accused, albeit unsworn, and the court. The Attorney-General and John Lilburne had some bad-tempered exchanges: ‘Do not interrupt me Mr Lilburne.’ ‘I pray you then do not urge that which is not right nor true, but notoriously false; for if you persevere in’t, I will interrupt you.’
Defence counsel were first allowed into the criminal process by the Treason Act of 1696, a measure which followed more than three decades, first of judicial revanchism for the regicide of 1649, and then of anti-Popish show trials and Judge Jeffreys’s Bloody Assize. Defence counsel were to be admitted in treason cases only; but in a preamble that anticipates Article 6(3) of the European Convention on Human Rights, the 1696 Act spelled out the need for the accused to have ‘just and equal means of defence of their innocencies’. It is likely to have been the same notion that from the 1730s began to persuade judges, prompted perhaps by the increasing use of counsel to conduct prosecutions, to allow defendants, too, to have counsel. By the end of the 18th century, it has been estimated, counsel was appearing for the defence in about one trial in three at the Old Bailey, and the first celebrity defence advocates were emerging.
The continued infiltration of lawyers from the civil into the criminal justice system as the 19th century unrolled produced an extraordinary paradox: accused persons were still not permitted by the common law to give evidence, and their remaining role in the proceedings – questioning witnesses, arguing law, addressing the jury – was now assumed entirely by their advocates. A French observer remarked as early as 1820 that in consequence ‘in England, the defendant acts no kind of part: his hat stuck on a pole might without inconvenience be his substitute at the trial.’ In fact, for many decades judges continued to make the accused himself rather than his counsel address the jury until, under pressure from the Bar, Parliament legislated in 1836 to give defence counsel this right. It was as the 19th century closed that the accused was first permitted to testify in his own defence, and Parliament by the same measure decreed that if he did so he lost his privilege against self-incrimination: if he entered the witness box he could be made to say on oath whether or not he was guilty – but he could not, except in special situations, be asked about his previous convictions (as I have said, that would have been over-incriminating). This, of course, is still the law; and it was joined during the late 20th century by a series of other situations in which Parliament made it clear – or thought it had – that the public interest in the exposure of incriminating facts overrides the personal privilege of withholding them.
While, at trial, the prosecution was expected – as it still is – to prove guilt without the accused’s help, in the pre-trial phase, when accused persons were held in appalling prison conditions from which few could reach out to secure evidence in their own defence, the law afforded the accused no privilege whatever against self-incrimination. By an Act of 1555 known as the Marian Statute, anyone arrested for felony was to be taken promptly before a Justice of the Peace, whose duty was to take down in writing anything, including things said by the accused, which was ‘material to prove the felony’. In cases of treason and other high felonies, the same task was carried out by the law officers of the Crown or the Privy Council. This was not the inquisitorial juge d’instruction system in which all evidence, pro and con, has to be investigated – it was a search for incriminating evidence from accusers and accused. Voluntary confession was naturally encouraged. The clerk of the peace’s vivid record of the trial of the Lancaster ‘witches’ in 1612, The Wonderfull Discoverie of Witches in the Countie of Lancaster (1613), illustrates the importance attached to unforced confessions: Elizabeth Device ‘made a very liberall and voluntarie Confession, as hereafter shall be given in evidence against her, upon her Arraignment and Triall’; and passing sentence, Justice Bromley said: ‘very few or none of you, but stand convicted upon your own voluntarie confessions and Examinations . . . What persons of your nature and condition ever . . . had more liberty given to plead or answer to every particular point of Evidence against you?’ There was, however, no formal inhibition on the threats or inducements to confess which might be held out to the defendant. The common law courts did not themselves use torture, but in state cases, until its virtual banning by statute in 1641, the Crown’s prerogative was used to authorise torture in order to secure confessions. The entire record of accusation and interrogation was sent up to the assize court and – for the better part of two centuries – read out to the jury. It was against this that the accused had to do the best he could without legal assistance.
It was during the 18th century that judges began to insist on oral testimony from those Crown witnesses who were available; but it was not until Jervis’s Act of 1848, by which time defence counsel had become a dominant feature of the trial process, that it became a requirement of the law that every accused person must be told at the start of his pre-trial examination that he was under no obligation to answer questions, and warned that any answers he gave might be used against him at trial.
Jervis’s Act marks the final transformation of a resonant aphorism, historically much honoured in the breach, into a sanctified principle of English law. It had by then been incorporated not only in the fifth of the Amendments made in 1791 to the Constitution of the United States, but in discrepant forms in many of the antecedent American state constitutions, and had been treated in each case, as it still is, as a self-evident civil right. Like the separation of powers, which first Montesquieu and then Madison found it useful to discern in the British system of government, it was less a fact than an idea whose time had come. But to say this is not to say that the idea had come from nowhere. It was an idea with a very long, though not an entirely pure, pedigree.
The leading mid-20th-century scholar of the Fifth Amendment, Leonard Levy, considered that the privilege (as it has become in modern Europe, though Levy regards it as a right) against self-incrimination was an Anglo-Saxon legal device designed to stem the oppressive effects of the inquisitorial processes of the Continental Church, which disfigure the history of the later Middle Ages. Writing as he was in the long evening of the McCarthy era, in which the American Supreme Court had let the Fifth Amendment be drained of much of its content – as had happened in previous decades to the First – and supported as he was by Wigmore’s great treatise on the law of evidence, Levy’s account is the one conventionally adopted in modern common law judgments. But recent scholarship, with fuller access to early sources, has called this account in question. This revised version has, in its turn, been contested, not least because of its unaccountably benign view of the inquisitorial proceedings of the medieval Church, but it does shed important light on the source and diffusion of the notion that nobody should be required to incriminate himself.
The maxim nemo debet prodere seipsum – literally, ‘nobody is required to betray himself’ – was taken by medieval scholars to have its origin in the writings of the fourth-century ecclesiast St John Chrysostom. What St John in fact wrote, according to Gratian (the only surviving source), was: ‘I do not say to you that you should betray yourself in public nor accuse yourself before others, but that you obey the prophet when he said: “Reveal your ways unto the Lord.”’ This is some way from saying that nobody should be made to confess, but like much other sanctified text it did service as the source of a succession of medieval assertions of a privilege against self-incrimination.
But medieval Church practice mocked the principle. Le Roy Ladurie’s celebrated study of the Church annals of Montaillou found the future Pope Benedict XII of Avignon, Jacques Fournier, presiding between 1318 and 1325 as Bishop of Pamiers over an inquisition court which interrogated on oath anyone denounced for Albigensian heresy, using as much physical torment as was needed to produce a confession. In addition to being held in fetters in a tiny cell and fed on black bread and water, the modes of pressure to confess included excommunication, which would have had the effect of closing off recourse to confessional privilege, and – in state-promoted cases – torture. For Fournier’s victims, including the five who were burnt at the stake, as for the thousands of others who were tortured in the course of the medieval Church’s inquisitions, the maxim nemo debet prodere seipsum did not have a great deal of significance. Joan of Arc was formally admonished at her interrogation ‘that she must answer and tell the truth about such things as touch her trial; and that it was essential that she should do so, since the doctors’ – lawyers – ‘were of this opinion’. Ecclesiastical lawyers held the maxim not to apply to charges of heresy, nor to other charges of grave criminality, nor to accusations based on reputed criminality (fama) rather than proven acts, nor to cases where the proof was considered strong. In point of law it is nevertheless apparent that it was into the Church’s own doctrines that the maxim against self-incrimination was first introduced. Although these proceedings were in origin accusatorial, manuscript records suggest that the privilege was not infrequently invoked on examination of suspected recusants under oath, despite the mass of exceptions. And, importantly, the maxim survived in the popular mind.
In 16th-century England the ecclesiastical courts, from the time of the Reformation a limb of the Crown, and the specially created Court of High Commission, set up with wide jurisdiction in ecclesiastical causes, made use of the ex officio oath, which was taken to lie in their inherent powers, accompanied sometimes by torture, to expose apostasy and heresy (not always distinguishable from sedition and treason). No privilege against self-incrimination was known to them. The courts of common law, despite their own dubious practices, responded with writs of prohibition to limit the power and effectiveness of the ecclesiastical courts; and it was probably in this process that the common law first claimed, and in due course was accorded, the credit for devising what was in fact a much older privilege. But it is a matter of debate whether the courts of common law, which until the eve of the Civil War themselves tolerated the use of torture when occasion required, regarded the privilege against self-incrimination as the reason for interfering with the ecclesiastical courts. The major recorded challenge to the High Commission’s use of the ex officio oath, the case of Maunsell and Ladd, heard by the King’s Bench in 1607, failed to secure the issue of a writ of habeas corpus. The argument from principle of Nicholas Fuller, one of the radical lawyers of his day, not only failed to convince a majority of the five judges that there was an overriding privilege against self-incrimination, but also resulted in his being prosecuted and jailed for his arguments by the High Commission itself. Popular belief came in the early 17th century to ascribe the privilege to Magna Carta and the Petition of Right, but it was not until well after 1640 that it began to be articulated with a note of reverence by the secular courts, and then in a system in which it meant little at trial, and nothing at the investigatory stage where it was most relevant. Even so, the reasons repeatedly given for it are of relevance.
First, it was reasoned by the canonists, we are all sinners: nobody would be safe if the secular state could demand as of right that individuals own up to crimes for which there was no other evidence. There remains, it seems to me, an important truth in this. It is why I have suggested that in the ideal system there can be no question of a roving right of inquiry on the part of the state: it would turn the rule of law into something closer to a reign of terror. The problem is answered, however, not by a blanket prohibition on incriminating questions but by a strong precondition that such questions may be asked only in situations prescribed by law and legitimately calling for an answer.
Second, the medieval schoolmen reasoned, there was an important divide between the private confession of sins in church and the public excoriation of crime before the ecclesiastical courts. It is intelligible that such a distinction should be adopted by a theocracy which demanded, as it still does, the unstinting confession of sins, criminal or not, and the doing of penance for them as a condition of spiritual salvation; and the legal insularity of penitential admissions continues to be both a reality and a real problem. It may well be, in fact, that the primary meaning of the maxim was that the confession of sins should not amount to the confession of crimes.
Third, however, it was considered invidious to place suspects in what a modern American judge has called the cruel trilemma of perjury (if they lied), contempt (if they stayed silent) and conviction (if they owned up), especially in legal systems in which perjury was a common and serious charge. Whether the avoidance of psychological pressure really was a consideration in the cruel systems of inquiry, prosecution and punishment which the medieval Church operated when its authority was threatened, or whether it was a rationalisation of systems which deprived the accused of an equal voice with his accusers, is less important to my purpose than the moral problem it presents today.
What also matters, of course, is the received axiom itself, endowed with classical and theological respectability, which in our era has acquired independent life and vigour. To the historical evidence that the roots of the axiom are both longer and older than the Anglo-Saxon legal tradition one can add the striking contemporary fact that in two of the leading cases in the European Court of Human Rights, judges from 24 countries have concurred in reading the right of silence and the privilege against self-incrimination into the guarantee of a fair hearing contained in Article 6(1) of the Convention. They described these as ‘generally recognised international standards which lie at the heart of the notion of a fair procedure’. Such striking testimony to the ubiquity of the principle in legal cultures as removed from each other as those of Britain, Finland, Turkey, Poland, Spain, Slovakia and Switzerland seems to speak convincingly in favour of at least this much of the revisionist thesis. But what has given it an iconic modern status which historically it never enjoyed?
The privilege against self-incrimination was not going to become, as it did become, the boast of Georgian England without attracting the caustic eye of Jeremy Bentham. Bentham, as usual, gave no quarter in his assault, and his beady-eyed advocacy of enforced confession as an engine of truth overlooked the extent to which it might be an engine of cruelty and falsehood. But there is a real sting, even if his vocabulary is no longer acceptable, in his characterisation of the suspect’s moral dilemma as ‘an old woman’s reason’ and of the notion that the accuser should not look to the accused for evidence as ‘a foxhunter’s reason’. Those whose interests are served by the exclusion of self-incriminating evidence, Bentham asserted, are ‘evildoers of all sorts’ and ‘lawyers of all sorts’. At least he was prepared to distinguish between the two. But he was presciently right about the lawyers. The development of defence advocacy, as Georgian and Regency barristers made their lucrative way into the criminal process, and the accompanying enunciation of formal rules of presumed innocence, strict proof and – another novelty – silent defendants, as the police forces took over the role of private prosecutors, made the 19th century a criminal lawyers’ heyday. When, in the late Victorian era of penal reform, the accused was at last given a voice equal to that of his accusers, the price he was required to pay – an entirely logical one – was the qualified forfeiture of his assumed common law right not to incriminate himself. It has since become an axiom of criminal practice that the defence case stands at its highest at the moment when the Crown closes its case, for since 1898 the accused has faced a new and equally cruel dilemma: to give evidence and risk being cross-examined to perdition, or to stay silent and risk the jury drawing the inference that he is hiding something. In criminal investigations the privilege still holds good – but with the important rider, since 1994, that declining to answer may legitimately lead to the drawing of adverse inferences in court. Recent Home Office research indicates that the proportion of suspects refusing to answer questions has fallen since the 1994 enactment from 23 per cent to 16 per cent, but that the proportion making incriminating admissions has remained the same. The privilege is not, however, confined to criminal investigations: since well before Jervis’s Act it has been applied to disclosure of documents and facts in civil proceedings and to answers capable of leading not to prosecution but to forfeiture.
So i come back to a present in which two legal imperatives confront each other. One is the need for regulatory and legal systems to be able in specified situations to insist on answers to awkward questions and, if the answers warrant it, to use them to prosecute their authors. The other is the European Convention on Human Rights, binding on the United Kingdom as a treaty since 1950 and now patriated as a governing element of our domestic law, which has been held by the European Court of Human Rights in Strasbourg to forbid, as contrary to the guarantee of a fair trial, any use of exacted answers to convict the person giving them.
It has to be observed first how far we have come from the jus commune and the concerns which gave rise to the axiom. We are not looking at a threat to give the authorities a roving commission of inquiry into people’s private lives, though the axiom remains a necessary barrier to that possibility. We are not considering invading the secrecy of the confessional. We are, it is perfectly true, looking at regimes which pose the ‘cruel trilemma’ of perjury, contempt or conviction – but they are specific regimes to which nobody has to sign up unless they are prepared to accept the regulatory system that goes with them. It is this that is the critical difference between such regimes and the uninvited inquisitor.
In his powerful dissent in the European case of Ernest Saunders, Judge Martens, the Dutch judge, spelled this out very clearly; but not persuasively enough for the majority, for whom the maxim nemo debet prodere seipsum appears to have possessed a talismanic quality. They accepted that it was permissible for answers to have been demanded from Saunders to the incriminating questions put to him by DTI inspectors who were inquiring under statutory powers into illicit practices employed in the Guinness takeover bid for Distillers, and for those answers to be used for both administrative and investigative purposes. They accepted, too, that the privilege against self-incrimination did not extend to materials taken by compulsion from the suspect – documents, blood samples and so forth – because, they said, this was how the privilege was ‘commonly understood in the legal systems of the Contracting Parties to the Convention and elsewhere’ (a curious echo of the medieval body of exceptions which swamped much of the principle). But, the majority said, the principle remained intact where the effect of the compulsion was the giving of evidence against oneself. They would not, they said, decide whether the privilege was absolute or whether it could properly be infringed in particular circumstances, but they went on, in a brief passage which gives little hint of the depth of issues below its surface, to say that the court
does not accept the Government’s argument that the complexity of corporate fraud and the vital public interest in the investigation of such fraud and the punishment of those responsible could justify such a marked departure as that which occurred in the present case from one of the basic principles of a fair procedure . . . The public interest cannot be invoked to justify the use of answers obtained in a non-judicial investigation to incriminate the accused during the trial proceedings.
That sounds very much like the enunciation of an absolute right, yet it was made by judges whose own jurisdictions all recognise a welter of exceptions to the principle. It also perhaps reflects the fact that the UK’s first line of defence had not been one of legal principle but an endeavour to distinguish the damaging but purportedly self-exculpating answers given by Saunders to the DTI inspectors from truly self-incriminating answers. For comprehensible reasons, since the Strasbourg approach is heavily fact-oriented, the question of principle was argued only as a fall-back.
Having recognised the legitimacy of enabling regulatory regimes to insist on having answers to relevant questions, the court took what I have picked out as the second, compromising course and decided that although it was legitimate to ask an incriminating question it was illegitimate to use the answer to it in court. That, it seems to me, is the worst of all possible worlds – a world in which the best possible proof of criminality may be on the record and cannot be used because it has not been more circuitously and less reliably obtained.
How serious is the problem? How many areas of public administration and personal activity does it affect? A good impression can be obtained by looking at the mopping up operation conducted by the United Kingdom in the wake of its defeat in the Saunders case. The Youth Justice and Criminal Evidence Act 1999 amends 11 important pieces of primary legislation passed between 1982 and 1992, together with their Northern Ireland counterparts, most of them designed to detect financial malpractice before innocent people lose their savings or investments. The amendment in each case takes the broad form of forbidding the use of the information obtained by statutory investigation in any subsequent criminal proceedings except on a charge of giving false information to the inquiry. One has only to look at the amended provisions to see what a swathe has been cut through the financial regulatory system: general investigations into insurance companies; documents obtained from insurance companies; documents and evidence produced to inspectors conducting investigations into companies; insolvents’ statements of affairs; statements made by directors facing disqualification; answers given to inspectors investigating building societies’ affairs; investigations of persons carrying on investment businesses; investigations into insider dealing; information required from and investigations into banking institutions; statements required by the Director of the Serious Fraud Office; powers for assisting overseas regulatory authorities; inspections required by the Friendly Societies Commission; statements required in Scotland by a nominated officer.
Recent post-Saunders legislation has likewise been tailored. The Financial Services and Markets Act 2000 gives the Financial Services Authority major powers of investigation and – if the investigation warrants it – powers to impose conditions on providers’ conduct, to take administrative or civil proceedings to obtain redress for people who have lost money because of providers’ misconduct, to censure and impose financial penalties on them, to withdraw their authorisation or approvals, and lastly and importantly to prosecute for crimes such as money laundering. The Act gives the FSA four objectives in the deployment of these powers: to maintain market confidence, to promote public awareness, to protect consumers and to reduce financial crime. But in deference to Saunders it forbids the FSA to rely on statements obtained under its statutory powers of investigation, and at a stroke obstructs one of Parliament’s own explicit objectives, the reduction of financial crime. What sense does this make?
The tide has been stemmed in the UK, for the present at least, by a decision taken by the Privy Council, the final court of appeal on devolution issues, in the case of Brown v. Stott. A visibly drunk woman, who had told police officers in Dunfermline that a nearby car was hers, was required by them under s.172(2)(a) of the Road Traffic Act 1988 to say who had just been driving it. On the basis of her answer – ‘It was me’ – she was prosecuted for drunk driving. The sheriff was disposed to let in the evidence of her admission, but the High Court of Justiciary, driven principally by the decision in Saunders, held the answer to be inadmissible. The Privy Council took the opposite view. In essence the Judicial Committee considered Saunders either to be distinguishable, or its reasoning to be too uncertain to be followed. They held that the s.172 power to demand an incriminating answer was a proportionate and therefore a legitimate response to a major social problem. The decision insists, as I would also wish to do, that the privilege against self-incrimination is purposive, not doctrinal, and that its legitimate use is a question of the proportionality of means to ends, not of rigid rules.
Treating the privilege as a rigid rule has not only given us the anomalous decision in Saunders. It has arbitrarily limited the principle by deciding that it should not be applied to possessions and intimate samples which, because they do not consist of spoken or written words, somehow escape the doctrine altogether. The Court of Justice of the European Communities, despite its policy of protecting recognised human rights, has declined to include the privilege in the rights it considers to be protected by Article 6. And the privilege can be seen in almost parodic form in the consequential endeavours to give corporations the same protection as human beings from the consequences of self-incrimination. While the United States, Canada and – marginally – Australia have resisted this curious teleology, the courts of the United Kingdom have adopted it. An early and debatable Court of Appeal decision, which held that corporations enjoyed in full the privilege against self-incrimination, was adopted without argument four decades later in the House of Lords because of the historical accident that, both parties before the House being corporations, neither had an interest in disturbing a decision which accorded them a coveted privilege. Yet what answer is there to what the US Supreme Court said almost a century ago in Hale v. Henkel? ‘The corporation is a creature of the state . . . It would be a strange anomaly to hold that a state, having chartered a corporation to make use of certain franchises, could not . . . inquire whether they had been abused, and demand the production of the corporate books and papers for that purpose’ – or, it follows, if it could not then prosecute the company or its officers when their answers and their papers showed that they had committed crimes.
Some of the jurisprudence of the European Court of Human Rights itself has become trapped in the same thicket. Since it is only to criminal proceedings that Article 6, and therefore the inferential right not to incriminate oneself, relates, some effort has been devoted to distinguishing criminal from regulatory proceedings. This, it seems to me, is an added misfortune. Many internal systems of regulation, both statutory and consensual, carry powers far more draconian than any court – not infrequently including the power to deprive people for good of their livelihood. Why such systems, by being called regulatory or disciplinary, should not be expected to assure the fundamental procedural rights of those subject to their determinations I do not understand, and I do not believe that the common law, which gauges the needs of fairness substantively and not formally, understands it either. What the formal distinction has done is to provide a limited but illogical limit to a generalised privilege against self-incrimination; but with it have gone the formal protections against oppression which are the genuine up-side of a legal privilege now accorded to a teenager charged with dropping a fast-food carton in the street but not (so far as the ECHR is concerned) to a person facing extradition on a charge carrying life imprisonment. The substantive baby has gone out with the procedural bathwater.
One way of cutting through the present tangle is, no doubt, to regard the maxim as historically little more than humbug on the lips of torturers and bullies, and to forget it. To do so would be not only to ignore its potential restraint of worse abuses but to abandon as much of its moral content as remains a real force for good. We ought to be able to do better than our ancestors. A measured and proportionate approach can respectably form part of the good legal system I began by considering, not least because it conforms to the requirements of the European Convention on Human Rights.
There are two irreducible propositions to which legal methods for obtaining evidence must conform. One is that the use of force or fear is morally unacceptable. The other is that no intrinsically unreliable evidence should ever be used to secure a conviction. The two are linked in that it is oppression which characteristically renders confessions unreliable, as we know to our cost. But each also has its own rationale. I have explained why I entirely accept that it would be oppressive to allow the state to exact information from people at will. The question is whether it is necessarily oppressive to allow the state both to demand information from people undertaking particular activities and to use it in court if it incriminates them.
There are two possible exits from the approval by the European Court of Human Rights of only the first, not the second, part of this exercise. One is to take the logical next step and let the answers be used in court. But the other – the first of the responses I began by describing – is to conclude that, precisely because the use of enforced answers to secure convictions is the logical sequel of permission to obtain them, the ideal system should not allow incriminating questions to be asked in the first place. This, in fact, is the position that has now been taken in Strasbourg on Ireland’s counter-terrorism legislation. If, however, the detection and prosecution of crime is a worthwhile social purpose, it is necessary to ask: why not? What both history and morality show to be serious answers to the question are that individual liberty ceases to have meaning if the state can demand answers at will, and that the use of violence or fear to obtain answers is both unacceptably degrading and evidentially counterproductive. But these answers leave significant space: they allow room in particular for a democratic polity to specify by legislation activities which are not only to be hedged by criminal sanctions, but also to carry a controlled power to insist on truthful answers to awkward questions as a necessary and proportionate element of effective regulation. Where such a need for regulation backed by investigative powers and criminal sanctions is established, to make the resultant answers inadmissible in evidence has more in common with snakes and ladders than with a serious legal system. It protects the guilty from conviction without meeting any of the historical or moral purposes of the privilege.
But precisely because obligatory answers necessarily involve an invasion of autonomy, there has to be a cogent case for each such measure. And because there is already an element of oppression in the mere compulsion to answer, strong protections (of a kind now familiar from Code C to the Police and Criminal Evidence Act 1984) are necessary to ensure that the dignity of the individual is not further invaded. Subject to such controls, however, as the European Court of Human Rights has already recognised, every society is entitled to penalise people who decline to co-operate with a regulatory system to which they have voluntarily subjected themselves. And it is this – the voluntariness of subjection to a particular regime – which seems to me the final element of the pattern in the good legal system. There will be many activities which, although theoretically voluntary (owning goods or homes, for example, or using public services), are in practice a condition of life. But there are others (running a public limited company, or handling large sums of other people’s money, or – more marginally perhaps – driving a car) on which a society is entitled to impose a condition of co-operation with a regulatory regime, backed by criminal sanctions. People are free to take it or leave it; but if they take the plums they take the duff. If you want to give it a legal name, it is waiver. It’s what occurs every time someone takes a job as a shop assistant or bus conductor: they will be required to account for the money they handle, and they will face prosecution if their account reveals that they have been stealing it.
That seems to me to be both the logic of Brown and the illogicality of Saunders. The control which remains in place is not only the will of a democratic legislature, essential though that initially is. It is the United Kingdom’s treaty obligation to observe the European Convention on Human Rights. Legislation which permits the use at trial of obligatory answers to incriminating questions must still pass the test of proportionality, a test much subtler than the bare balance of interests or a broad sense of fair play. In the present context it has to start from the high historic and cultural premium placed on the privilege; it needs to recognise that the privilege at its most basic remains a shield against kinds of oppression and risks of injustice which are absolutely unacceptable; and in this light it needs to ask whether the particular incursion nevertheless meets a legitimate and necessary objective by the least invasive means available. Such means will need to include clear procedural protections, though logically they cannot include a caution to the effect that the examinee need not answer. While it may be sufficient for some compulsory answers to be used for regulatory purposes only, it will in my view be justifiable only in exceptional cases to exclude the use of incriminating answers for the very purpose their name suggests – incrimination. In our system there is the final filter of a judicial power to exclude evidence which, though admissible, will make the trial unfair. But because this is a power without clearer guidelines for its exercise than the sense of fair play – something that will necessarily vary from judge to judge – it is initially no substitute for measured legislative provision.
Regulatory systems, of course, are only an aspect of the larger process of the detection and prosecution of crime. In the world of which we have been citizens since the events of 11 September, the looming question is whether – in the absence of any prior regulatory bargain between the individual and the state – the prevention and detection of terrorism is by itself sufficient to justify the exaction of answers from suspects or from those believed to be concealing information; and if so, on pain of what sanctions. It is not very long since the Israeli High Court, under its able Chief Justice Aharon Barak, outlawed torture as a legitimate expedient in all circumstances. The European Court of Human Rights has now added lesser sanctions such as imprisonment to the prohibited expedients for extracting information. But in Europe, at least, there is a difference between the two: torture and inhuman treatment are unconditionally outlawed by Article 3 of the Convention; imprisonment is not. It may be that here too the question is going to become one of the proportionality of ends to means, a topic on which the jurisprudence developed in Strasbourg furnishes important guidance. It is when you stand on the edge of an abyss that it becomes very important not to lose your balance.