1. MR. PHILIP HAVERS QC (instructed by Treasury Solicitors, London SW1H 9JS) appeared on behalf of the Appellant.

2. MR. ANDREW NICOL QC (instructed by Lovell White Durrant, London EC1A 2DY) appeared on behalf of the Respondent.

3. Anthony-Noel Kelly is an artist who used body parts which he had acquired from a young man called Neil Lindsay, who was employed as a laboratory technician by the Royal College of surgeons, to make casts which he then exhibited at the London Contemporary Art Fair held in Islington in January 1997. Dr. Martin, H.M. Inspector of Anatomy, realised that the casts had been taken from body parts and investigations were commenced and the police involved. In April 1997, Kelly and Lindsay were arrested and admitted that Lindsay had provided some 20 or so parts and had been paid by Kelly. Both men were charged with theft of a number of body parts between June 1991 and November 1994. Kelly was further charged in the alternative with receiving, but that charge was not left to the jury.

4. Kelly had sought and had been granted access to the College in about 1992 to sketch exhibits and body parts there. He had not been content merely to sketch but had arranged to take parts away to make casts from them. He had disposed of some of them in a field in Kent and others had been kept in premises in Brixton. The police recovered some 37 parts from both these places.

5. The trial commenced at Southwark Crown Court before H.H. Judge Rivlin, Q.C. and a jury on 23 March 1998. The prosecution case had been completed and defence submissions made by Friday 27 March 1998. Those submissions, which were rejected by the judge, were, broadly speaking, that as a matter of law a human body part could not be stolen since it could not be owned by anyone. That left the only issue which the jury had to consider which was whether it had been proved that either defendant had been dishonest. Kelly asserted that he had honestly believed that the College had no further interest in the body parts and so he was entitled to take them and use them as he had done. He believed, he said, that the parts had been abandoned by the College and did not belong to anyone.

6. On Sunday 29 March the Observer published in its review section a full page article headlined in bold type :-

7. The article was set round a large photograph of one of the casts made by Kelly from the body parts allegedly stolen by him which was an exhibit in the trial. To the side of the article was an unsmiling photograph of Kelly and below it, again in bold type, were the words, extracted from the article:-

8. The article was written by Brian Masters who, as a footnote to the article records, “is the author of ‘Killing for Company: The Case of Dennis Nilson’ and ‘She Must Have Known: The Trial of Rosemary West’”. The article had been trailed on the front page of the newspaper in these words:-

9. The article was drawn to the attention of the judge on Monday 30 march after he had given his ruling on the submissions and it was clear that the trial was to continue. Mr Kelly intended to give evidence and to call character witnesses on his behalf. His counsel understandably took the view that the article involved an attack on Kelly’s character and at least by implication upon his honesty and veracity and thus there was prejudice which would require the discharge of the jury. It is apparent from the transcript before us that both counsel for the prosecution and the judge were of the same view so that if the jury or some of them had read the article and an application for discharge had been made, it would have succeeded.

10. The article has, of course, to be read as a whole. But it is pertinent to highlight some parts of it. The first paragraph reads:-

11. While the author disowns any desire to consider a matter which might be relevant to the decision to be made by the jury, he makes it clear that he is concerned in the article with Kelly’s personality, which he describes as ‘perverted’, and his interest in dead bodies or necrophilia. Thus, although it is said by Mr Andrew Nicol, Q.C. on behalf of the Respondent, no doubt correctly, that the subject of the article was necrophiliac art, as defined by the author, what is conveyed to the reader is that it is about Kelly and what motivates him to do as he does. The article discusses Kelly as a person and as an artist. It does so against the background of the use of the dead body by artists as contrasted with the obsession with dead bodies or their parts displayed by serial killers such as Dennis Nilson and Jeffrey Dahmer. The thrust of the article is that Kelly’s use of dead body parts falls in the latter rather than the former category.

12. The article describes how Kelly used the parts “to create works of art with which to decorate the walls of rooms intended for human habitation”, making the point that there was not much creativity in what Kelly did since he was merely producing inside-out copies of the raw materials. The author then defined what he meant by necrophilia. It was not intercourse with a corpse, but a general ‘love of the dead’ which could take many forms, ranging from a desire to be near or work with bodies to necrophagy or the killing of people in order to satisfy an obsession with dead bodies.

13. I should cite the sixth paragraph of the article in full. It reads:-

14. The author goes on to ridicule the analogy that he understood had been drawn in court between Kelly and Leonardo da Vinci, describing it as ‘absurdly false’. There follows a discussion of the positive treatment of the dead body by artists whom the author regards as truly great. This is contrasted with necrophilic art which is described as negative, static and stagnant: examples are said to be the likes of Salvador Dali and Rachel Whiteread. This leads to consideration of non-artistic necrophiles and a reference to serial killers, including Jeffrey Dahmer, of whom this is said:-

15. The concluding paragraph reads:-

16. By great good fortune, it turned out when enquiries were made on the Monday morning that only one juror had seen the article. The judge’s immediate reaction, having read the article, was to say that he would accede to any application, were one to be made, to discharge that juror. He went on to say this:-

17. He then went on to make the obvious point that if prosecuting counsel did explore “these areas, which he would be perfectly entitled to do”, the juror who had seen the article might say to herself “Well, there we are: that is what it is all about”.

18. The juror in question was asked, in the absence of the other jurors, whether she would be concerned that what she had read might colour her views. She said this:-

19. She went on to say that she would be able to put the article from her mind and promised not to mention its contents to any of her fellow jurors. It was thus very fortunate that the one juror who had seen the article should have shown such responsibility in her reaction to it. The defendants, particularly Kelly who was the most affected by the article, were anxious not to have to go through the ordeal of a trial again and were naturally concerned that, if the juror was discharged, the others (who of course knew that the court was concerned with an article published in the Observer) might think that there had been something gravely damaging to the defendants in it. So no application to discharge either the one juror or the whole jury was made. The judge, in directing the jury, took pains to say that the article itself did not contain some dark secret about the case or either defendant and reiterated his warning that the jury should try the case only on the evidence before them and must not seek out the article. The trial then continued.

20. The reaction of the judge is important. It is not determinative of whether there was a substantial risk that the proceedings would be seriously prejudiced, but this court should, in deciding whether contempt is made out, attach due weight to it. Simon Brown, L.J. said in one of the most recent cases in this field, A.G. v Birmingham Post and Mail Ltd [1999] 1 WLR 361 at p.371F:-

21. The only reservation I have is that the very immersion in the case may produce a degree of overreaction by counsel which influences the judge. Particularly in a case which has attracted much media interest (and it is such a case which is likely to suffer from ill-timed potentially prejudicial comments) the judge will be especially concerned that there should be no chance of adding to matters which might influence the jury in a fashion prejudicial to either side. But it is apparent that the judge’s immediate reaction, knowing the issues in the case and that Kelly’s honesty and so credibility were critical, was that the article was seriously prejudicial to him. That, as it seems to me, is an important matter for us to bear in mind in reaching our decision.

22. Kelly’s character and credibility were indeed critical and he called evidence to establish that he had a positively good character as opposed to merely relying on his lack of previous criminal convictions. The judge in his summing up referred to this evidence and made the point that whether his art was good or bad was irrelevant to whether he had acted dishonestly. Furthermore, as the judge said, “the fact that someone has a deep and genuine interest, a fascination, a passion, or an obsession with something, is not and cannot be a defence to taking someone else’s property dishonestly”. Kelly was put forward as a man passionately interested in art and a man of considerable integrity. He called a witness called Edwina Orr who said that she saw nothing morbid about his fascination with dead objects but “he is passionately serious about his art and he is a dedicated artist”. She described him as kind and warm-hearted and believed there was nothing bad in him. The jury also had read to them the statement of the surgeon who first introduced Kelly to the College. In it he said:-

23. Mr Nicol has argued that Kelly’s motive for taking the body parts was beside the point since it could not determine whether or not he had acted dishonestly. That, as a matter of logic, may be true. But it seems to me to be obvious that a jury would be likely to take a different view of the credibility and honesty of someone who was obsessed with dead bodies and derived a degree of comfort from owning them from that of someone who was a serious artist of genuine integrity.

24. I should finally set out passages in the summing up which direct the jury how to approach their task and which bear on the issues before us. I give the page numbers in the bundle from which we have been working:-

25. I must now consider the principles of law which we must apply. The application is brought under the so-called strict liability rule set out in Section 2 of the Contempt of Court Act 1981. It is not for a moment suggested that the publication of the article was intended to prejudice Kelly’s trial or that the Respondents did not act with caution and in good faith. The evidence before us shows that the editor was concerned about the possibility of contempt and sought legal advice which he relied on in publishing the article.

26. The test is that contained in s.2(2) which reads:-

27. In his skeleton argument, Mr Philip Havers Q.C. for the Attorney-General suggested that the course of justice might be impeded because the trial might have been aborted, but he did not before us (in my view rightly) pursue the argument. He accepted that there was in the circumstances of this case no point in seeking to rely on the risk of impediment independently of that of prejudice. Impediment would result from prejudice. Undoubtedly if the judge had concluded that the course of justice had been seriously prejudiced, he would properly have discharged the jury and ordered a new trial to be held when the prejudicial effect had sufficiently faded.

28. There have been a considerable number of authorities which establish the proper approach to and construction of s.2(2). In one of the more recent, A.G. v MGN Ltd [1997] 1 All ER 456, Schiemann L.J. set out the principles in some detail at pp.460-462. It is unnecessary to recite them in full. A brief summary of what is uncontroversial is that the risk must be judged at the time the publication is made and must be more than remote. The potential prejudice must be serious, an ordinary English word which must be given its proper weight. The court must be sure that the risk exists before convicting. The nature of the published material and its presentation, the timing of the publication, the likelihood of its coming to the attention of jurors or potential jurors, the likely impact on the jury and the ability of the jury to abide by any judicial directions which seek to neutralise any prejudice are all relevant considerations in individual cases. To quote Auld L.J. in A.G. v BBC [1997] EMLR 76 at 81:-

29. Recently the court, notably through Simon Brown, L.J., has been concerned that there should, so far as possible, be a uniformity of approach from the Court of Appeal when dealing with convictions allegedly unsafe because of media prejudice and this court when dealing with alleged contempt. While I recognise the desirability of such uniformity, it must always be remembered that the law of contempt is concerned with preventing a publication because it creates a substantial risk of serious prejudice whereas the Court of Appeal will be concerned with whether there has been such prejudice as renders the conviction unsafe. The fact that no actual prejudice has resulted from the publication cannot prevent it being a contempt within the meaning of s.2(2), although no doubt the lack of actual prejudice may be relevant in deciding whether there really was a substantial risk of serious prejudice. It may be, for example, that it transpires that no juror has read the article or that counsel for the potentially prejudiced party decides not to make an issue of it or the jury by its verdict shows that it has not been affected by the publication. In A.G. v Birmingham Post and Mail [1999] 1 WLR 361 at 369H, Simon Brown L.J. said this:-

30. My own approach would be slightly different. It seems to me that the prejudice required by s.2(2), which must be serious, is not of a lesser degree than that required to make good an appeal against conviction. To establish contempt it needs only be shown that there was a substantial risk that serious prejudice, which must in my view mean such prejudice as would justify a stay or appeal against conviction, would result from the publication. That such prejudice does not in the event result is nothing to the point. Thus uniformity of approach is achieved by requiring that the prejudice within the meaning of s.2(2) must be such as would be likely to justify at least a stay. But, since the risk has to be judged at the time of publication, the court will have to be satisfied that jurors will be likely to have seen it and the court will disregard any extraneous factors, such as decisions by counsel for tactical or other reasons not to raise the matter or to seek a fresh trial.

31. So in this case it was fortunate that only one juror had seen the article and that she professed not to have been influenced by it so that no application for discharge was made and no real prejudice occurred. But that cannot avail the Respondents, nor should it, if there was at the time of publication a substantial risk of serious prejudice. Thus, although I reach it by a slightly different route, I for my part would agree with Simon Brown L.J.’s test that ‘to create a seriously arguable ground of appeal is a sufficient basis for finding strict liability contempt’. This test is based on the assumption that neutralising directions have been given, but that no submission has been made to seek to persuade the judge to order a stay and that the judge has not of his own motion thought to do so.

32. I should say a word about neutralising directions. Not unnaturally, in all contempt cases such as this counsel for the alleged contemnor relies on a number of judicial observations about the ability of jurors to comply with their oaths and to decide cases solely according to the evidence. Mr Nicol is no exception. It is clearly right that this should be borne in mind but, as has been said, it cannot be taken too far because, if it is, there would be no need for any law of contempt. It seems to me that if, absent a special direction, the court is satisfied that serious prejudice would be likely to result, the giving of a special direction would not necessarily remove a substantial risk that such prejudice would occur. There is no research which enables the court to know how jurors react in this sort of situation. Thus, as it seems to me, the court should be slow to assume that a special direction will remove the risk. That it might is insufficient; the risk will remain. Thus I think it is helpful to consider whether a special direction is required to avoid serious prejudice. If it is, it is in my view likely that a substantial risk will remain even if that direction is given. But, again, whether that is so will depend upon the circumstances of the individual case.

33. It is pertinent to remember that s.2, indeed the 1981 Act as a whole, was made necessary by the decision of the European Court of Human Rights in the Sunday Times Thalidomide case that the decision of the House of Lords in A.G. v Times Newspapers [1974] A.C. 273 showed that the law of this country contravened the Convention. This was because the House of Lords decision gave too much weight to the protection of the administration of justice and too little to the protection of freedom of speech. Lloyd L.J. said in A.G. v Newspaper Publishing Ltd [1988] Ch. 333 at p. 382 (one of the Spycatcher cases):-

34. Thus in applying s.2(2) due weight must be given to the protection of freedom of speech. In Worm v Austria (1997) 25 EHRR 454 the European Court had to consider, in the context of contempt of court, the extent to which freedom of expression had to give way to the need to protect the administration of justice. The court had to consider the scope of Article 10.2 which, so far as material, reads:-

35. The relevant passages in the judgment are paragraphs 39 and 40, which read:-

36. “39. In the present case it was not contested that the applicant’s conviction was aimed at “maintaining the authority and impartiality of the judiciary” and that it thus pursued a legitimate aim under the Convention.

37. The phrase “authority of the judiciary” includes, in particular, the notion that the courts are, and are accepted by the public at large as being, the proper forum for the settlement of legal disputes and for the determination of a person’s guilt or innocence on a criminal charge; further, that the public at large have respect for and confidence in the courts’ capacity to fulfil that function.

38. Then paragraph 47:-

39. Finally, paragraph 50:-

40. “50. Restrictions on freedom of expression permitted by Article 10(2) “for maintaining the authority and impartiality of the judiciary” do not entitle States to restrict all forms of public discussion on matters pending before the courts.

41. There is general recognition of the fact that the courts cannot operate in a vacuum. Whilst the courts are the forum for the determination of a person’s guilt or innocence on a criminal charge, this does not mean that there can be no prior or contemporaneous discussion of the subject-matter of criminal trials elsewhere, be it in specialised journals, in the general press or amongst the public at large.

42. Provided that it does not overstep the bounds imposed in the interests of the proper administration of justice, reporting, including comment, on court proceedings contributes to their publicity and is thus perfectly consonant with the requirement under Article 6(1) of the Convention that hearings be public. Not only do the media have the task of imparting such information and ideas; the public also has a right to receive them. This is all the more so where a public figure is involved, such as, in the present case, a former member of the Government. Such persons inevitably and knowingly lay themselves open to close scrutiny by both journalists and the public at large. Accordingly, the limits of acceptable comment are wider as regards a politician as such than as regards a private individual.

43. However, public figures are entitled to the enjoyment of the guarantees of a fair trial set out in Article 6, which in criminal proceedings include the right to an impartial tribunal, on the same basis as every other person. This must be borne in mind by journalists when commenting on pending criminal proceedings since the limits of permissible comment may not extend to statements which are likely to prejudice, whether intentionally or not, the chances of a person receiving a fair trial or to undermine the confidence of the public in the role of the courts in the administration of criminal justice”.

44. It seems to me that, if they are applied as I believe they should be, the provisions of s.2(2) will not contravene the Convention.

45. Mr Havers referred us to a decision of the New Zealand Court of Appeal , Gisborne Herald v S-G [1995] 3 N.Z.L.R. 563 . The court was considering, in the context of the New Zealand Bill of Rights Act 1990, how the balance between freedom of speech and the need to ensure a fair trial should be approached. The court decided that so far as possible both rights should be accommodated but, where both could not be fully assured, it was appropriate to curtail freedom of media expression temporarily so as to guarantee a fair trial.

46. The decision is interesting but does not in my judgment assist us in approaching our task. It is based on the New Zealand legislation and noted the slightly different approach of the Canadian Supreme Court in Dagenais v Canadian Broadcast Corp. (1994) 94 C.C.C. (3d) 289 emphasising the equal status of both rights and that a ban on publication should only be imposed if such a ban was necessary to prevent a real and substantial risk to the fairness of the trial. It seems to me that the New Zealand approach which Mr Havers espouses tends in the forbidden direction of the House of Lords decision in A.G. v Times Newspapers.

47. I must now consider the application of the principles I have sought to summarise to the circumstances of this case. As the judge himself said, it was likely that jurors had wondered for themselves whether some morbid obsession might have motivated Kelly rather than high minded artistic considerations. That does not in my view make the article any less prejudicial. If jurors had those feelings, to give a respectable basis to them would serve to confirm them and to make those jurors less likely to respond favourably to any arguments to the contrary.

48. Both counsel sought to draw comfort from the reaction of the juror who had seen the article. Mr Havers underlined her observation that she ‘realised it was very damaging towards one of the defendants’. Mr Nicol relied on her reaction not to read it and her statement that she did not take it seriously. Why, he asked, should we assume that other jurors would have reacted differently if they had seen the article? I am afraid I take the view that all concerned were indeed very fortunate in the juror’s reaction; she was unusually responsible. I am sure that most would have read it (particularly if it had been drawn to their attention by someone who realised their involvement in the trial) and there would have been a real risk that they would have been influenced by it. Furthermore, her immediate realisation that it was ‘very damaging’ is important. The extract in bold print and the heading do nothing to dispel any sense of hostility to Kelly in the article. It seems to me that there was a real risk that what might linger in the jurors’ minds was the damage to Kelly inherent in the article. Thus I think that her reaction benefits Mr Havers rather than Mr Nicol.

49. I can understand why the legal adviser may have thought there was no breach of s.2(2). He may not have appreciated that the only issue in the trial which remained live was whether the defendants had been dishonest. Thus Kelly’s credibility was critical. He was putting himself forward, supported by witnesses, as a genuine artist whose use of body parts was solely for the purpose of his art. The article, taken as a whole, cast doubt on that. It raised the question whether he was not a man obsessed with dead bodies, bereft of emotion and having the same sort of motivation for his actions as named serial killers. This sort of obsession, apart from shedding doubt on the evidence he and his witnesses had placed before the jury, might lead the jury to think that he was the less likely to have cared whether the College had abandoned the bodies. The point made by Mr Nicol that Kelly’s fascination or obsession with corpses was manifest to the jury does not in my view provide an answer. It is no doubt true, but the nature of the obsession or fascination and the reason for it was likely to be a highly material consideration for the jury in deciding whether to reject his defence. To put it crudely, if the jury believed he was the sort of person described in the article they would surely consider that his motives were likely to have been base and that he did not deserve to be believed.

50. An editor of a newspaper will always be taking a risk if he chooses to publish an article which refers directly to a trial in progress (except insofar as it is a fair and accurate report of such a trial). If it turns out that the issues were not quite as he believed so that there was a substantial risk that the course of justice in the proceedings would be seriously prejudiced he has committed a contempt.

51. I have found the decision in this case not at all easy but, if it had been left to me, I would have decided that a contempt had been committed. What tipped the balance for me was the reaction of the trial judge and the likely effect of the article on the only issue left to the jury, namely whether Kelly had been dishonest.

52. Since writing my judgment, I have had the opportunity of considering that to be delivered by Sedley L.J. There is little between us on the correct approach, although it may be that my threshold is somewhat lower than his. I would attach less weight than he does to neutralising directions. I think that, if such directions are needed, it will usually be difficult to say that a substantial (which after all means only more than remote) risk of serious prejudice remains. However, I, like he, regard this as a case which is very close to the borderline and I have asked myself whether I can be sure, as opposed to thinking it probable, that there really was a substantial risk that the course of justice in the proceedings would be seriously prejudiced.

53. While I would have concluded that the balance comes down in favour of contempt being proved, I recognise that the scales would have only just tipped. This being so, I am persuaded that I should not press my views to the point of dissent from Sedley L.J’s conclusion and all that that would entail by way of rehearing and additional expense. What divides Sedley L.J. and myself is whether contempt has been proved to the required standard. That gives rise to no points of principle and, as it seems to me, if Sedley L.J. cannot be sure, it would be wrong for me to hold out for my point of view. Accordingly, although with considerable hesitation, I concur with his conclusion that this application should be dismissed.

54. This means that it is, strictly speaking, unnecessary to consider s.5 of the 1981 Act. However, the point has been argued before us and, in case the matter goes further, I should briefly set out my views upon it.

55. Section 5 provides:-

56. It is clear that an incident which results in a trial and so the trial itself may create the matter of general public interest. An example is found in the incursion into Buckingham Palace by Michael Fagan which was much discussed in the press so as to lead to contempt proceedings: see A.G. v Times Newspapers & others The Times 12.2.83. The matter of general public interest was the safety of the Queen and the nature of the intruder and the sort of man who had managed to enter the Queen’s bedroom was an important aspect of that matter. Accordingly, the prejudice to the legal proceedings against Fagan was truly incidental to the discussion. But here, although no doubt necrophilic art and Kelly’s place in it could be said to be a matter of public interest, the reality is that the article concentrated on Kelly and his activities gave rise to it. It was not, as it seems to me, a matter which arose because of but independently of the trial as was the safety of the Queen in Fagan’s case or the sanctity of human life in A.G. v English [1993] 1 A.C. 116. Thus the prejudice was not merely incidental to the discussion since the discussion was about Kelly’s actions which had led to the trial. Accordingly, in my judgment, if we had decided that the article breached s.2, s.5 would not have provided a defence.

57. The history of this case and the principles of law by which it has to be decided are fully set out in the judgment of Collins J. I will limit myself, therefore, to considering some of the more problematical questions of law which he has addressed before setting out my own conclusion on the alleged contempt.

58. Juries are daily told by trial judges that they are not expected to leave their experience and common sense behind when they enter the jury box or retire to the jury room. The unique virtue of the jury system is its reliance for a just verdict on a combination of the popular good sense of twelve people chosen at random with their conscientious attention to the judge’s directions on the law. Experience, for instance in the form of differential verdicts or of discerning questions from the jury room, shows repeatedly how well on the whole juries handle both elements of their task.

59. But for reasons which Parliament has found compelling, the detail of how juries individually or generally function is not open to research. In the Gisborne Herald case to which Collins J. has referred, Richardson J. (as he then was) at 574-5 spoke of the limited data in New Zealand on the effect of media publicity on juries, contrasting this with published research in Canada and the United States and commenting:

60. Mr. Havers QC for the Attorney-General has stressed the word “guarantee” in Richardson J’s formulation; but I do not believe that Richardson J meant to say that under the Bill of Rights Act 1990 (derived as it is from the International Covenant on Civil and Political Rights) freedom of press comment must yield whenever there is a risk of prejudice. On the contrary, he cites with approval the words of McLachlin J. in Dagenais v. Canadian Broadcasting Corporation (1995) 94 CCC (3 ^rd) 289, 370-1:

61. Although the material parts of the Human Rights Act 1998 are not yet in force, the history of the Contempt of Court Act 1981 makes it not only legitimate but necessary to construe and apply the latter so far as possible in conformity with the European Convention on Human Rights. The values which the Convention assigns to free expression and to fair trials are replicated in the New Zealand Bill of Rights and in the Canadian Charter of Fundamental Rights and Freedoms, making the jurisprudence of all three instruments – the one European, the other two governing common law jurisdictions – of potential value to us. It is therefore interesting to find substantially the same concept of a localised balance of rights and interests, rather than a defeasible priority of the claims of justice, informing the decision of the European Court of Human Rights in Worm v. Austria, from which Collins J. has given full citations. I say “rights and interests” because in Convention terms the Attorney-General is probably not a “victim” of any violation of Article 6: the right to a fair criminal trial is conferred by Article 6(1) only on the accused. What gives the Attorney-General his domestic standing is (as the European Court of Human Rights recognises) the public interest in the integrity of the system of justice, whether it is the defence or the prosecution which stands to suffer by the interference. The media for their part enjoy an express right under Article 10.

62. Although Mr. Havers has resisted the suggestion that we should have any regard to the question of proportionality before the 2 ^nd October 2000, (when Ministers have announced that they intend to bring the full Act into force and when Mr. Havers concedes that we shall have to do so), it seems to me for the reasons I have touched on to be both permissible and helpful, at least as a cross-check on judgment, to ask whether in the present case the preservation of Mr. Kelly’s right to a fair trial and the public interest in assuring it were such as to make it both necessary and proportionate in a democratic society that the Observer should not publish Mr. Masters’ article as and when it did. The jurisprudence not only of the European Court of Justice and the European Court of Human Rights but of the Canadian Supreme Court and the New Zealand Court of Appeal makes central use of this form of reasoning. If I take the liberty of repeating what I said in R. v. Home Secretary, ex parte McQuillan [1995] 4 All ER 400, 422, it is only because it has since been cited with approval by others:

63. To this must of course be added what Sir Thomas Bingham MR approved in R. v. Ministry of Defence, ex parte Smith [1996] QB 517, 554:

64. What is true of executive decisions is also true, mutatis mutandis, of the courts’ adjudications. It follows from the formulation in Ex parte Smith that my reference to the significance of the right which is at issue should be understood as a reference to the significance of the interference in each particular case rather than to a general ranking of rights in order of supposed importance

65. The avoidance of a double standard

66. The first resort of respondents to contempt applications tends to be to what Lawton J. (as he then was) said in the criminal appeal case of R. v. Coughlan (1976) 63 Cr. App. R. 33, 37:

67. Collins J. has set out what the judge said to the jury in the present case. As one would expect, no issue was raised on the appeal [1999] 2 WLR 384 in relation to prejudice because the trial judge had defused the situation. Equally plainly, contempt looks not to the retrospective risk that the jury has been prejudiced but to the antecedent risk that it will be.

68. This said, however, I respectfully share the concern expressed by Simon Brown LJ. in Attorney-General v. Unger [1998] 1 CrAppR 308, 318-9 and reiterated by him in Attorney-General v. Birmingham Post and Mail [1999] 1 WLR 361, 369, that the courts should not speak with a robust voice in criminal appeals and a sensitive one in contempt cases. Like Collins J. I would be inclined to hold that this is not to be done by postulating different degrees of prejudice for the two purposes but by adopting a single standard and recognising that it will operate differently in the two contexts. The standard in both contexts is a substantial risk of prejudicing the fairness of a criminal trial. The self-evident difference is that the risk has ordinarily to be gauged prospectively in a contempt case (and therefore without regard to the outcome of the trial) but retrospectively in a criminal appeal (where ex hypothesi there has been a conviction).

69. One way, therefore, to ensure that the court speaks with a single voice is to test an accusation of contempt by assuming:

70. If in such a situation an appeal on the ground of prejudice would not succeed, no more should the publisher be guilty of contempt. The prospective risk of serious prejudice cannot be any greater than the actual possibility, in the assumed situation, that it has occurred. By parity of reasoning, a case in which an appeal would in the assumed events succeed will ordinarily be a case where contempt is made out, provided always that the court is sure that the facts meet the test.

71. This formulation differs, I accept, from that of Simon Brown LJ. in Unger: it looks to the existence of grounds for allowing or dismissing an appeal against conviction rather than for granting or refusing leave to appeal. This is because, the test on appeal being now the safety of the conviction (Criminal Appeal Act 1968, s.2(1), as amended by the Criminal Appeal Act 1995), any substantial risk (cf. s.2(2) of the 1981 Act) that a conviction has been contributed to by a prejudicial publication will ordinarily make it unsafe. To reduce this threshold to the leave stage, requiring only an arguable case of risk, may be to set the threshold of contempt unduly low, at a level where there is not a demonstrable risk of prejudice but only an arguable case of it.

72. The reason why I believe my test nevertheless to be consistent with Simon Brown LJ’s view is that it makes the assumptions needed to produce parity between the two situations. Thus in the present case one asks: if the jurors had read the article but Judge Rivlin, though giving them a clear neutralising direction, had rejected an application to discharge them, would Mr. Kelly’s conviction be unsafe? If the answer is no, then assuming (see assumption (d)) that it is not for some adventitious reason, I do not see how publication can be said to have prospectively carried any greater risk of prejudice.

73. It does not follow, of course, that the outcome of an appeal determines whether there has been a contempt of court. Those responsible for the publication will have played no part in the appeal. If the Attorney-General concludes that it is in the public interest to seek their committal, both sides are entitled to a hearing free of presumptions.

74. One reason why I have considered at some length the appropriate tests in the present state of the law is that there is, it seems to me, no obvious outcome on the facts before us. The arguments and merits are closely balanced.

75. Two central factual elements which are clearly important are the “as and when” – the manner and time – of publication. Asked by the court where was the danger to free speech in deferring publication until the trial was over, Mr. Nicol QC rightly responded that the danger lay in the question. But the corollary of the editor’s freedom of judgment about when to publish is that the risk he must anticipate is much sharper if he elects to publish in mid-trial.

76. The manner of publication, by contrast, was relatively sober though not wholly unsensational, an educated and reasoned critique of a questioned art-form. If it was in the public eye because of Mr. Kelly’s trial, that made publication more, not less, legitimate (see Attorney-General v. Times Newspapers Ltd. (Times, 12 ^th February 1983), where the trial of a man found in Buckingham Palace raised legitimate concerns about the security of the Queen). Nevertheless, for the reasons given by Collins J. this is not a case in which section 5 affords an answer to a strict liability accusation.

77. The want of research data about the effect of media comment on juries requires the court to make certain assumptions. Both in criminal appeals and in contempt cased these are, broadly, that juries will seek conscientiously to abide by the trial judge’s directions, but that a point may come where this is more than they can fairly be expected to do. To this, however, must be counterposed the fact, which is central to the jury system, that jurors enter the jury box with ideas of their own, some of which may well have a legitimate bearing on their verdict. Thus, in the present case, if the jury which convicted Mr. Kelly had read the Observer article, two further things would have been of importance. First, the article offers a viewpoint which most or all of them will undoubtedly have entertained themselves and which some of them probably held. Secondly, they would have heard the judge’s direction about ignoring both media comment and personal animus, and have done their best to follow it.

78. One further consideration is of importance. The trial judge’s own reaction was unequivocal: if any of the jury had read or learned of the contents of the article they would have to be discharged. Plainly – and the decided cases confirm it – the initial response of the judge in charge of the trial is of high evidential relevance for this court. Equally plainly it cannot be determinative. Because the judge’s concern is that the integrity of an ongoing trial should be unassailable, he will rightly incline towards caution. This court’s task is, in the light of the trial judge’s response but with the benefit of discrete evidence and argument, to reach its own view.

79. In such circumstances, can this court say with the necessary confidence that there was at the date of publication a substantial risk that the course of justice in the case against Mr. Kelly would be seriously prejudiced if jurors were to read the article; in other words, a risk of bias sufficient to render an eventual verdict of guilty unsafe?

80. The principal element which seems to me capable of tipping the scales in the Attorney-General’s favour, given the opposing considerations I have so far considered, is the risk that the article would reinforce and give legitimacy and cogency to the feeling which many of the jurors must anyway have had about Mr. Kelly’s work. If it were so, the accusation of contempt would be readily made out, for there is no doubt that, whatever the author and editor and legal adviser thought was the issue at trial, the honesty of Mr. Kelly’s acts and his veracity in accounting for them were the key issues for the jury; so that anyone who took Mr. Masters’ view was not going to look benignly on the defence. The risk was thus undoubtedly one of serious prejudice to the course of justice.

81. But was it substantial? In the end, and not without anxiety, I have concluded that it is simply not possible to be sure that the risk created by the publication was a substantial risk that a jury, properly directed to disregard its own sentiments and any media comment, would nevertheless have its own thoughts or value judgments reinforced by the article to a point where they influenced the verdict. As a first cross-check, I doubt whether an appeal would have been allowed had the jury which convicted Mr. Kelly read the article. As a second cross-check, it seems to me that the threat from this article, published when it was, to the course of justice in Mr. Kelly’s trial was not sufficient to make either prior restraint or subsequent punishment a proportionate response in a society which, as a democracy, values and protects the freedom of the press

82. For these reasons the Attorney-General’s motion must in my judgment fail.