Confronting Confrontation


Confronting Confrontation



THE RIGHT TO confrontation has precipitated a showdown between the English courts and the European Court of Human Rights. Article 6(3) of the ECHR provides that ‘everyone charged with a criminal offence’ has the right to ‘examine or have examined witnesses against him’. This basically means that the accused, or his lawyer, should have a chance to put questions to adverse witnesses.

In the combined applications of Al-Khawaja and Tahery v United Kingdom,1 the prosecution had introduced witness statements from witnesses who were not present at the defendants’ trials. Strasbourg held that there was a breach of the confrontation right because the convictions were based ‘solely or to a decisive extent’ on the evidence of absent witnesses, even if the witnesses were absent for good reason: in Al-Khawaja the witness (the complainant in a sexual assault trial) was dead; in Tahery the witness was too frightened to testify. The English courts found this position hard to accept. In Horncastle the UK Supreme Court and the Court of Appeal agreed that the ECtHR jurisprudence on confrontation should not be followed.2 So English hearsay law was in apparent conflict with the Convention: defendants could be convicted on hearsay evidence, seemingly in breach of Article 6.3 Al-Khawaja has now had a rehearing before the ECtHR’s Grand Chamber,4 and the second judgment makes some concessions to the English position by allowing exceptions to the ‘sole or decisive’ rule. Many commentators will be critical of Strasbourg’s abandonment of its hard line.5

Confrontation rights also have a lively recent history in the United States. The Sixth Amendment to the US Constitution provides that ‘in all criminal prosecutions, the accused shall enjoy the right … to be confronted with the witnesses against him’. In a 2004 decision, Crawford v Washington,6 the Supreme Court reinvigorated this clause. Departing from its previous case law,7 the Court held that even if there are indications that a statement of an absent witness is reliable, the statement would not usually be admissible under the Confrontation Clause. The Crawford approach is somewhat stricter than the European one—there is no exception, for example, for evidence which is not the ‘sole or decisive’ basis on which the accused is convicted.

There is plainly much at stake in debates about confrontation. Complete compliance with the confrontation right means that where a witness makes a statement to the police incriminating the accused and then dies, or cannot be found, or is no longer fit to testify, or, perhaps, is too scared to come to court, then her statement cannot be introduced at trial. While under the European approach there is the caveat that the evidence can be admitted if it is not ‘sole or decisive’, the practical effect is that the prosecution must fail where it depends on the evidence of an absent witness, even if a court provided with the evidence would consider the case to be proved beyond reasonable doubt. This chapter takes a close look at confrontation in order to see whether this result can be justified. The discussion is confined to absent witnesses, although in Europe the confrontation right has also generated a significant case law on anonymous witnesses.8 While concentrating on Article 6(3) of the ECHR, I refer to US doctrine to illuminate the scope and rationale of the right in Europe.


The confrontation right overlaps with, but is narrower than, the hearsay rule. A succinct definition of the hearsay rule is that ‘a statement other than one made by a person while giving oral evidence in the proceedings is inadmissible as evidence of any fact stated’.9 While the hearsay prohibition has been narrowed slightly by the Criminal Justice Act 2003, the rule still applies to a wide range of statements. Statements made to the police are covered, but so are statements made in informal conversations between friends and statements made in business records. In the United States it is clear that the confrontation right does not extend this far. Under Crawford, it applies only to ‘testimonial’ statements. While ‘testimonial’ has not been definitively defined, the broadest definition mooted in Crawford was ‘statements … made under circumstances which would lead an objective witness to believe that the statement would be available for use at a later trial’.10 Post-Crawford, the Court has held that laboratory reports by forensic scientists are testimonial,11 as are statements made by a complainant to the police shortly after an alleged incident of domestic violence.12 Statements made in an emergency telephone call to the police, however, were not classified as testimonial, because the operator would have been concentrating on dealing with the emergency rather than producing evidence for trial.13

Under the ECHR, things are less clear, but one important element to note is that in all of the cases where a violation has been found, the problematic evidence seems to have been statements made to investigative authorities—police officers or examining judges.14 Indeed, the Court commonly uses the word ‘deposition’ to describe the type of evidence to which the confrontation right attaches.15 Thus where a witness gives what a common lawyer regards as hearsay evidence, there may not be an issue under the Convention so long as the hearsay does not take the form of a witness statement. In AM v Italy, a child had complained to his parents that during a holiday in Italy he had been indecently assaulted by the applicant.16 Statements were taken from the child’s parents and from a psychotherapist who was treating him. The arguments before the ECtHR focused on the absence of confrontation of these witnesses, with the Court finding a breach of Article 6(3) because the applicant ‘did not have a chance to examine the witness statements [sic] that formed the basis of his conviction’.17 This suggests that had these witnesses been confronted, proceedings would have been Convention compliant, even though the witnesses would have been relating hearsay, ie what the child had told them.18

Some witnesses may make statements to the police that while helpful to the prosecution, do not directly incriminate the accused. An example would be a statement such as ‘I heard a gunshot at 11 o’clock’. In contrast, all of the cases in which the ECtHR has found a violation seem to have involved ‘accusatory’ statements, where the defendant accuses a specific person of a crime.19 In X v United Kingdom, various people who had filmed an incident in Northern Ireland at which the killing of two soldiers took place were allowed to give evidence in court anonymously.20 The evidence apparently involved the witnesses describing the making of their films and photos of the incident; they did not identify the applicant themselves. Among the reasons why the European Commission found that witness anonymity did not infringe the confrontation right in this case was that the evidence ‘did not implicate the applicant’.21 With so little to go on, it is impossible to say how significant this observation is, or how accusatory a statement would have to be before it required confrontation under the ECHR. To give but one example, an eyewitness who provides a description to the police of the person who attacked her might be said to implicate a defendant without accusing him.22 But while one can only speculate as to what the scope of the ECHR’s confrontation right is, there must surely be some limit to it. Otherwise the prosecution would not be able to rely on business records in a case where the original maker of the record was dead or could no longer be identified.

There are various reasons why the restricted scope of the confrontation right is significant. One is that it undermines the British courts’ principal argument in Horncastle. The courts noted that hearsay can sometimes be perfectly reliable, and thus a rule that it can never be the ‘sole or decisive’ element in a conviction is unnecessarily strict. But, if the Strasbourg Court sees confrontation as confined to statements made to the authorities, the English courts, by framing the argument in terms of hearsay, may be talking past Strasbourg.23 This is most evident when we consider a series of examples given by the Court of Appeal, and endorsed by the UK Supreme Court, which are intended to demonstrate that Strasbourg is being unreasonable.24 In one example a woman makes an emergency telephone call to the police and identifies the person who is attacking her. But this is not really a statement made in response to police questioning, and so might not be seen as a ‘statement’ or ‘deposition’ in the ECtHR’s terms. Nor, in all likelihood, would it be regarded as testimonial in the United States.25 Another example involves bank records in a fraud case; here it is even clearer that the records would not be regarded as requiring confrontation in either jurisdiction.26 The example that comes closest to raising a confrontation issue27 involves a witness who writes down the registration number of a car involved in a drive-by shooting, there being sufficient corroborating evidence (gunshot residue found in the car) to rule out the possibility of mistake. If the witness reports the number to the police, there is a good chance that the confrontation right under the ECHR and the Sixth Amendment would be engaged. But Strasbourg’s answer cannot quite be predicted with confidence, for in the terms introduced above the report of the number plate implicates but does not accuse. The example would be more on point if the witness claimed to recognise the car’s driver, and gave the police a specific person’s name. This is plainly accusatorial and does not quite deliver the intuition that the Court of Appeal wanted: that there is no value in confrontation when evidence is well corroborated.

Paradigmatically, confrontation involves the relevant witness testifying in the accused’s physical presence at trial, in full view of the accused and the fact-finders and with the accused being able to put questions.28 But departures from this paradigm may be permissible.29 The core of the right under the ECHR seems to be that the accused should have some opportunity to put questions to the witness: procedures whereby the witness is questioned pre-trial, outside the presence of the accused or his lawyer, but where the defence has the ability to influence the questions put to the witness, may comply with Article 6(3).30 Witness anonymity is permitted, so long as ‘counter-balancing’ measures are in place.31 Testimony via video-link would not raise an issue under the Convention.32

(a) Exceptions

Other than forfeiture (which is considered separately below), the confrontation right in the United States may be without exception.33 In particular, the Al-Khawaja situation, where confrontation is not possible because the witness died unexpectedly before trial, does not give rise to an exception. The position under the ECHR is similar.

However, in one important respect Strasbourg does take a more flexible approach than the US Supreme Court, because a statement made by a justifiably absent witness will be admissible under Strasbourg jurisprudence if it is not the ‘sole or decisive’ basis of conviction. This is a vague criterion—almost any evidence relied on by the prosecution might turn out to be the decisive ‘feather that tips the scales’—and it has not always been consistently applied by the ECtHR.34 In recent cases, however, the notion seems to be applied quite strictly. In Al-Khawaja the witness statement was reasonably well corroborated: another complainant had made a similar allegation against the defendant, and the absent witness had initially made her allegations to friends. But the statement was still considered to be decisive. In other recent cases, the Court glosses the sole or decisive test in terms of whether the statement was the only ‘direct’ evidence of guilt35 (as it was in Al-Khawaja). There remains some ambiguity, however, as to whether use of a decisive statement from an absent witness will inevitably make the trial unfair. In Al-Khawaja the UK Government argued that, because the applicant had had alternative means of challenging the witness statement, there were measures in place to ‘counter-balance’ the lack of confrontation of the sole and decisive witness statement. The ECtHR disagreed on the facts, but did not quite rule out the possibility of counter-balancing in appropriate cases.36

(b) Forfeiture

In the United States the only significant exception to the confrontation right is forfeiture: the defendant can lose the confrontation right through his own wrongful behaviour. This exception was considered in detail in Giles v California.37 While recognising a forfeiture exception, a majority of the US Supreme Court held that the mere fact that the victim had died at Giles’s hands was not enough to engage it. Forfeiture was construed narrowly, as applying only to conduct designed to prevent the victim from testifying.

It is not entirely clear how the ECtHR would deal with a case where the absence of a witness whose statement is ‘sole or decisive’ had been caused by a defendant. In Rachdad,38 the Court noted that the applicant had contributed to the difficulty in securing confrontation: he had apparently been abroad (allegedly to avoid arrest) when his trial took place, and by the time he appealed against his convictions, some five years later, only one witness attended court. A breach of Article 6 was still found. In Al-Khawaja and Tahery, the Court distinguished the English Court of Appeal’s judgment in Sellick,39 noting that lack of confrontation may have been justified in that case because the witnesses ‘were being kept from giving evidence through fear induced by the defendants’.40 This implies that forfeiture would be recognised in a strong case. However, the indications are that, like the US Supreme Court, Strasbourg would take a restrictive view of the doctrine. The Strasbourg Court recently ventured the following general observations on ‘waiver’:

[A] waiver must, if it is to be effective for Convention purposes, be established in an unequivocal manner, be attended by minimum safeguards commensurate with its importance, and should not run counter to any important public interest … . Moreover, before an accused can be said to have implicitly, through his conduct, waived an important right under Article 6, it must be shown that he could reasonably have foreseen what the consequences of his conduct would be.41


Neither the ECtHR nor the US Supreme Court has gone very far in developing an explicit theory of confrontation. Nevertheless, there is value in exploring what these courts have said about the confrontation right, before turning to the rather better developed accounts in the academic literature. At the outset, it is worth noting that there are two main ways in which confrontation—and indeed most procedural rights—can be theorised.

Confrontation might be seen as a right that is instrumental to fact-finding, because it promotes accurate verdicts. We can refer to such accounts of confrontation as ‘epistemic’. Alternatively, confrontation might be thought of as a ‘non-epistemic right’. On this view, there is value in confrontation even if it would not promote accuracy, perhaps as a way of respecting the defendant’s dignity. These are not mutually exclusive ways of thinking about confrontation: one may value confrontation for a mixture of epistemic and non-epistemic reasons. Nevertheless, the distinction is a useful way of analysing arguments about confrontation.

(a) Strasbourg’s Human Right

To date, most of the indications are that the ECtHR understands the value of confrontation in purely epistemic terms. Its only explicit statements about the importance of confrontation stress two considerations. First, where confrontation is lacking, a defendant is deprived of ‘any opportunity of observing the demeanour of [the] witness when under direct questioning, and thus of testing her reliability’.42 Secondly, in the case of an anonymous witness:

If the defence is unaware of the identity of the person it seeks to question, it may be deprived of the very particulars enabling it to demonstrate that he or she is prejudiced, hostile or unreliable. Testimony or other declarations inculpating an accused may well be designedly untruthful or simply erroneous and the defence will scarcely be able to bring this to light if it lacks the information permitting it to test the author’s reliability or cast doubt on his credibility. The dangers inherent in such a situation are obvious. Furthermore … [a trial court is] precluded by the absence of the said anonymous persons from observing their demeanour under questioning and thus forming its own impression of their reliability.43

An emphasis on reliability is also reflected in the observation that the evidence of an absent witness should be treated with ‘extreme care’.44 In cases where confrontation does not occur, the ‘sole or decisive’ test also indicates that unreliability is the key concern.

The only element of the Strasbourg case-law that might be hard to square with an emphasis on reliability is the apparent restriction of the confrontation right to some sort of testimonial evidence—statements made to the authorities, or perhaps a narrower class of ‘accusatorial’ statements. To the common lawyer familiar with the concept of hearsay, this focus might seem odd. The (epistemic) ‘hearsay dangers’ of veracity, perceptual ability, memory and ambiguity can affect any hearsay evidence and can be best explored through cross-examination of the declarant. Under a non-epistemic conception of the confrontation right, however, it might be thought that the defendant’s ability to challenge his accusers—those who denounce him to the authorities—is a way of respecting his dignity. However, the restricted scope of the right might also be explained on epistemic grounds. A rule that all testimony should be subject to direct challenge by the defendant risks being so broad as to be unworkable, encompassing even such things as business records. A focus on accusatorial statements to the authorities might then be a way of marking out a particular category of statement that is potentially outcome-determinative and where the risks of the witness having an axe to grind are pronounced (as in the earlier example where the witness to the drive-by shooting gives a name to the police). Like any rule,45 this specification will be imperfect, sometimes being over- and sometimes under-inclusive with regards to its rationale of admitting reliable evidence. But it might still draw the line in a sensible place.46

This rule-based defence of the confrontation right provides a response to the English courts’ claim in Horncastle that evidence subject to the confrontation right can nevertheless be reliable enough to be properly decisive. A strict rule might be seen as the most appropriate way of protecting the right from abuse by judicial discretion. From one perspective, it is not surprising that the ECtHR should approach confrontation in this rule-based way. Given that the Court generally defers to national judges’ assessments of evidence,47 it would have little influence over domestic courts’ application of the confrontation right were it to allow an exception for reliable evidence. The ‘sole or decisive’ test is probably the closest it can come to incorporating a reliability rationale in its confrontation jurisprudence. From another perspective, however, the adoption of a hard and fast rule sits less comfortably with the way the ECtHR relates to Member States. While the US Supreme Court justifies its own strict confrontation rule partly on the basis of distrust of judicial discretion,48 it is hard to imagine Strasbourg explicitly making such an argument. The European Convention on Human Rights is part of a very different political order to the US Constitution. Expressions of lack of faith in domestic judicial decision-making could undermine the ECtHR’s authority.49 This presents something of a dilemma for the Strasbourg Court. Its institutional position means that, insofar as the confrontation right is based on an epistemic rationale, it is pushed towards adopting a strict rule that is not entirely responsive to reliability concerns. Yet it is simultaneously pulled away from imposing a rule so strict—as an absolute rule of inadmissibility might be—that the rule’s discretion-limiting functions would be obvious.

(b) Crawford’s Confrontation Clause

The US Constitution is a foundational political document in a way that the ECHR is not. Because the US Supreme Court’s recent interpretations of the Confrontation Clause are heavily influenced by this broader context, its pronouncements are not very useful for understanding confrontation in England and Europe. The discussion in this Section can therefore be brief.

In its recent case-law, the US Supreme Court has veered towards an ‘originalist’ interpretation of the Sixth Amendment Confrontation Clause. Its concern has been largely with how the clause would have been understood at the time of the founding. From this perspective, the Court has felt little need to justify the confrontation requirement, giving the impression that the ‘purpose of confrontation is confrontation’.50 While the Court has suggested that the ultimate goal of confrontation ‘is to ensure reliability of evidence’,51 this rationale plays little role in the recent case-law. The reason for this is partly distrust of the judicial discretion that would be involved in creating exceptions for reliable evidence.52 It is also because, even if the right was originally written into the Constitution because of concerns about the reliability of un-confronted testimony, now that it is in the Constitution it should be honoured for its own sake.

(c) Police Tactics

When we move beyond judicial pronouncements, we find a richer set of theories of confrontation. One theme in the literature is that police interviewing tactics give us reason to distrust witness statements. As William O’Brian puts it: ‘statements that are created once litigation is anticipated or underway are inherently suspect’.53 Even where a witness has no axe to grind—recall the Court of Appeal’s number plate example in Horncastle—police questioning can be suggestive and manipulative. Margaret Berger, emphasising deterrence of police misconduct, has made a similar argument.54

There is certainly good reason to be concerned about police interviews with witnesses. Research in England and Wales found that ‘interviews were… highly interviewer driven, with a confirmatory bias’.55 One assessment concludes that ‘[o]fficers are apt to interview witnesses in ways that are wholly improper and ineffective’.56 However, a strong confrontation right may not be the best way to deal with the problem. To the extent that improper suggestion is the culprit, cross-examination may be ineffective if the police version of events has been internalised by the witness, especially as trials take place months or even years after the alleged crime, requiring witnesses to ‘refresh their memories’ from their police statements. Berger’s hope is that confrontation will encourage better police performance, but even if cross-examination is good at rooting out police malpractice, its influence on police conduct generally is likely to be limited, especially given the high rate of guilty pleas.57 Electronically recording witness interviews may be a more effective way of regulating police tactics, enabling an assessment of what transpired during interviews irrespective of the witness’s availability for cross-examination.58 By focusing on the dangers of presenting the fact-finder with unreliable evidence, O’Brian side-steps these particular criticisms. It is, however, not obvious that the product of police questioning is so unreliable that we are better off not admitting it at all unless there is confrontation—especially if electronically recorded evidence of an interview is available. I return to this point in the more detailed analysis of reliability below.

By drawing attention to the possibility of police abuse of the questioning process, O’Brian and Berger gesture at a slightly different value inherent in confrontation. The trial of Sir Walter Raleigh is often taken to be paradigmatic of the possibility of abuse in a system without confrontation. Raleigh was convicted of treason, largely on the basis of the testimony of Sir Thomas Cobham, who had apparently told his questioners that Raleigh had been part of a plot to overthrow James I.59