Cream Holdings Ltd. & Ors v Banerjee & Anor

This is an interesting, difficult and important appeal concerning the prior restraint of media publications. It raises in stark form the question of the true construction and application of s12(3) of the Human Rights Act, 1998, more particularly the meaning in the subsection of the word “likely”.

Section 12(3) provides:

“No such relief [relief affecting the exercise of the Convention right to freedom of expression] is to be granted so as to restrain publication before trial unless the court is satisfied that the applicant is likely to establish that publication should not be allowed.”

Does “likely” in this context mean more likely than not or something less than that (and, if so, what)? Somewhat suprisingly this has not yet been clearly and authoritatively decided despite a number of recent high profile cases involving injunction proceedings against newspapers. Hitherto perhaps it has not proved necessary to reach a clear view on the point. On this appeal it is.

The respondents, the claimants in the action, are a group of companies which began in 1992 with a nightclub in Liverpool and have since expanded and diversified their business. In addition to running clubs, they stage large events not only in the United Kingdom but worldwide. They also carry on a substantial business licensing and franchising their brand name and logo and merchandising clothes, records and other items. Nothing turns on the distinctions between the individual companies within the group and, like the judge below, I shall refer to them collectively as “Cream”.

The first appellant (Ms Banerjee) is a qualified chartered accountant who, having previously worked for a firm of accountants (Kidsons - now merged with Baker Tilly) dealing with Cream’s business, in February 1998 became directly employed by Cream as their in-house accountant. That employment ended in February 2001, Ms Banerjee’s subsequent claim for unfair dismissal being settled at an Employment Tribunal hearing in February 2002.

The second appellants publish the Daily Post and Liverpool Echo, Liverpool’s leading newspapers with a strong reputation for investigating stories of local public interest. For convenience I shall refer to them simply as the Echo.

Following the settlement of her claim against Cream, Ms Banerjee went to the Echo with a number of allegations of financial irregularities by Cream, allegations which to some extent were supported by documentary material (mostly copied) which she had taken without permission. On 13 June 2002 the Echo published a story using some of those materials, a story of corruption alleging payment of a substantial sum by a Cream director (Mr Davenport) to a senior local council officer (Mr Wadkin) in connection with a licence which Cream needed from the council to stage a particular event. The Echo had earlier told the council of the allegation and they in turn had told the police who, on the day before the article was published, had arrested the two men. By letter to Cream on the day of publication the Echo invited Mr Davenport’s immediate response to a number of questions which the letter raised in connection with that story, stating in addition that “the Echo is investigating a number of other allegations which have been raised” and inviting Cream to address various listed questions concerning those allegations too.

It was in these circumstances that Cream immediately decided to seek injunctive relief to prevent publication of any further allegations. They made no complaint about the initial article, recognising both that it was too late to do so and that in any event there was a public interest in exposing corruption. That first article, however, had expressly said that “there is no evidence any other Cream directors knew of the arrangement” and Cream contended that there was no justification for the publication of further material involving the clearest breaches of confidentiality on Ms Banerjee’s part.

The Echo agreed not to publish any further article pending the hearing of Cream’s application for interlocutory relief and both sides then filed extensive evidence, some by way of open statements, some dealing with confidential matters.

The application came before Timothy Lloyd J on 3 July 2002 and on 5 July he granted an injunction restraining the appellants until trial or further order from publishing, disclosing or using certain confidential information as defined in a confidential schedule save to certain specified bodies. He refused, however, to make an interlocutory order for the return of the documents. He handed down two judgments, one described as “open”, the other as “private”.

On this appeal (brought with the permission of Robert Walker LJ) Miss Victoria Sharp QC for the appellants advances two central contentions. First, she submits that the judge below applied the wrong test under s12(3), that of “a real prospect of success” rather than “more likely than not”; secondly, she submits that whatever is the true meaning of “likely” in the subsection, the judge’s own findings on the evidence were such that he could not reasonably have been satisfied that Cream was likely to succeed at trial on the merits given what she submits is “the overwhelming public interest in publication in this case”.

Although it is unusual in this court to state one’s conclusions so early in the judgment, I propose to do so here because they necessarily dictate whether the confidentiality in much of the material before us needs to be preserved pending trial and what therefore can be said at this stage in an open court judgment. These then, stated in the most summary form, are my conclusions:

Given that in the result the interlocutory injunction should in my opinion remain in force until trial, it follows that only the first of these four conclusions can be fully explained in an open court judgment; the other conclusions necessarily depend on the detailed facts of the case which cannot now be openly revealed and which therefore I propose to deal with only most cursorily.

It is convenient first to look briefly at the four cases which have thus far touched most closely on the point now arising. This is best done chronologically.

In Douglas -v- Hello! Limited [2001] QB 967, Keene LJ in paragraph 149 of his judgment recorded Mr Tugendhat QC’s argument for the claimants as follows:

“…. Mr Tugendhat’s submission is that the phrase ‘likely to establish’ does not mean ‘more probable than not’, because that interpretation in certain circumstances could bring it into conflict with the European Convention on Human Rights itself by giving priority to article 10, the right to freedom of expression, over article 8, the right to respect for private and family life. Such an automatic priority, it is said, would not be in conformity with the Convention. Since section 3(1) of the Human Rights Act requires the court to construe legislation in a way which is compatible with the Convention rights ‘so far as it is possible to do so’, section 12(3) itself must be read in a way which avoids giving precedence to article 10 rights. It is argued on behalf of the claimants that the words ‘likely to establish’ in that subsection should be taken to mean ‘not fanciful’ or ‘on the cards’.”

Keene LJ then turned to consider that argument in paragraphs 150-153 of his judgment (paragraphs with which Brooke LJ and Sedley LJ both expressly agreed, Brooke LJ stating in terms that he had “nothing to add” to them):

“150

For my part, I do not accept that there is any need for conflict between the normal meaning to be attached to the words in section 12(3) and the Convention. The subsection does not seek to give a priority to one Convention right over another. It is simply dealing with the interlocutory stage of proceedings and with how the court is to approach matters at that stage in advance of any ultimate balance being struck between rights which may be in potential conflict. It requires the court to look at the merits of the case and not merely to apply the American Cyanamid test. Thus the court has to look ahead to the ultimate stage and to be satisfied that the scales are likely to come down in the applicant’s favour. That does not conflict with the Convention, since it is merely requiring the court to apply its mind to how one right is to be balanced, on the merits against another right, without building in additional weight on one side. In a situation such as the one postulated by Mr Tugendhat, where the non-article 10 right is of fundamental importance to the individual, such as the article 2 right to life, the merits will include not merely the evidence about how great is the risk of that right being breached, but also a consideration of the gravity of the consequences for an applicant if the risk materialises. The nature of the risk is part of the merits, just as it would be at trial when the balance had to be struck. That is as relevant at the interlocutory stage as it would be at trial. But that does not require any strained interpretation of section 12(3).

151

Certainly section 12(3) is making prior restraint (i.e. before trial) more difficult in cases where the right to freedom of expression is engaged than where it is not. That is not a novel concept in English law. As was said by Laws J. in R -v- Advertising Standards Authority Ltd, ex parte Vernons Organisation Ltd [1992] 1 WLR 1289, 1293:

‘there is a general principle in our law that the expression of opinion and the conveyance of information will not be restrained by the courts save on pressing grounds. Freedom of expression is as much a sinew of the common law as it is of the European Convention for the Protection of Human Rights and Fundamental Freedoms.’

152

Perhaps more to the point, the jurisprudence of the European Court of Human Rights is generally hostile to prior restraint by the courts. Prior restraints on publication are not prohibited by the Convention, as the European Court made clear in The Observer and the Guardian -v- United Kingdom [1991] 14 EHRR 153, the ‘Spycatcher’ case, but in that same case it went on to say, at p 191, paragraph 60:

‘On the other hand, the dangers inherent in prior restraints are such that they call for the most careful scrutiny on the part of the court. This is especially so as far as the press is concerned, for news is a perishable commodity and to delay its publication, even for a short period, may well deprive it of all its value and interest.’

153

It is impossible to accept that a statutory provision requiring a court to consider the merits of the case and to be satisfied that the balance is likely to be struck in favour of the applicant before prior restraint is to be granted is incompatible with the Convention. It follows that no strained reading of the language of section 12(3) is needed to render it compatible with Convention rights. The wording can be given its normal meaning. Consequently the test to be applied at this stage is whether this court is satisfied that the applicant is likely to establish at trial that publication should not be allowed. Even then, there remains a discretion in the court. ”

Sedley LJ, besides expressly agreeing with Keene LJ’s judgment on the point, stated his own conclusions as follows:

“134

… A newspaper, say, intends to publish an article about an individual who learns of it and fears, on tenable grounds, that it will put his life in danger. The newspaper, also on tenable grounds, considers his fear unrealistic. First of all, it seems to me inescapable that s. 12(4) makes the right to life, which is protected by article 2 and implicitly recognised by article 10(2), as relevant as the right of free expression to the court’s decision; and in doing so it also makes article 17 (which prohibits the abuse of rights) relevant. But this in turn has an impact on s. 12(3) which, though it does not replace the received test (or tests) for prior restraint, qualifies them by requiring a probability of success at trial. The gauging of this probability, by virtue of s. 12(4), will have to take into account the full range of relevant Convention rights.

135

How is the court to do this when the evidence – viz that there is and that there is not an appreciable risk to life - is no more than evenly balanced? A bland application of s. 12(3) could deny the claimant the court's temporary protection, even if the potential harm to him, should the risk eventuate, was of the gravest kind and that to the newspaper and the public, should publication be restrained, minimal; and a similarly bland application of s. 12(4), simply prioritising the freedom to publish over other Convention rights (save possibly freedom of religion: see s.13), might give the newspaper the edge even if the claimant's evidence were strong. I agree with Mr Tugendhat that this cannot have been Parliament's design. This is not only, as he submits, because of the inherent logic of the provision but because of the court's own obligation under s. 3 of the Act to construe all legislation so far as possible compatibly with the Convention rights, an obligation which must include the interpretation of the Human Rights Act itself. The European Court of Human Rights has always recognised the high importance of free media of communication in a democracy, but its jurisprudence does not – and could not consistently with the Convention itself – give article 10(1) the presumptive priority which is given, for example, to the First Amendment in the jurisprudence of the United States' courts. Everything will ultimately depend on the proper balance between privacy and publicity in the situation facing the court.”

In the course of the argument before us, Sedley LJ said that he for his part had intended by that judgment to indicate that the test of likelihood set by s12(3) was in his then view one of more likely than not. He pointed out that it had not in fact been necessary for the court there to set the standard.

The next authority in point is the decision of Sir Andrew Morritt VC in Imutran Limited -v- Uncaged Campaigns Limited [2001] 2 All ER 385. Having regard to the dates respectively of the hearing and the judgment in that case, it is plain that the Vice-Chancellor received a copy of this court’s judgment in Douglas -v- Hello! Limited between the two. He dealt with s12(3) as follows:

“17.

Counsel for the defendants submitted that the requirement of likelihood imposed a higher standard than that formulated in American Cyanamid -v- Ethicon Ltd [1975] AC 396. I did not understand this to be disputed by counsel for Imutran. He submitted that whatever the standard was his case satisfied it. Theoretically and as a matter of language likelihood is slightly higher in the scale of probability than a real prospect of success. But the difference between the two is so small that I cannot believe that there will be many (if any) cases which would have succeeded under the American Cyanamid test but will now fail because of the terms of section 12(3). Accordingly I propose to apply the test of likelihood without any further consideration of how much more probable that now has to be. See Douglas -v- Hello! Limited …”

In the result the Vice-Chancellor concluded: “that Imutran is likely to establish at the trial of the action that publication of its confidential documents should not be allowed.”

I come next to Theakston -v- MGN Limited [2002] EWHC 137 in which Ouseley J, having referred to Keene LJ’s judgment in Douglas -v- Hello! Limited, continued:

“19

Mr Tugendhat for the claimant submitted that section 12(3) should be applied as if the test were not discernibly different from the familiar interlocutory test whether the case has a real prospect of success. He referred me to the later judgment of Sir Andrew Morritt V-C in Imutran …

I am not sure that this [paragraph 17 of the Vice-Chancellor’s judgment in Imutran] fully supports Mr Tugendhat’s submission. In any event, I have some difficulty in seeing how the approach required by section 12(3) can be other than that the claimant must show that it is more probable than not that he will succeed in obtaining an injunction at trial. I cannot envisage, as a matter of ordinary English, an injunction which is likely to be granted but more probably than not will be refused. If Parliament had intended the relevant test to be whether the claimant had a real prospect of success, it would have used that familiar legal phrase. I consider that it intended to impose the discernibly more rigorous requirement which it did in this particular context of freedom of expression.”

Ouseley J , however, then went onto state that it mattered not for the purposes of the application before him which test applied since in any event he considered it more probable than not that the claimant would succeed at trial in obtaining an injunction so that on either approach interlocutory relief could and should properly be granted.

The fourth and final authority is A -v- B plc [2002] 3 WLR 542 in which this court, having expressed in paragraph 8 of its judgment (given by Lord Woolf CJ) the “hope that as a result of our decision the citation of authorities on this scale will be regarded as unnecessary and not accepted by judges of first instance who have to hear these applications”, then set out extensive “guidelines which are intended to assist the judiciary and the parties to deal with the majority of these applications in a more proportionate manner”. For present purposes I need cite only paragraphs 11(iii) and 12 of the judgment:

“(iii)

As to the word ‘likely’ in section 12(3) useful guidance is provided by Sir Andrew Morritt VC in Imutran Ltd -v- Uncaged Campaigns Ltd [2002] FSR 20. He said of section 12:

[Paragraph 17 of the Imutran judgment is then set out as in paragraph 19 above, save only that the mention of Douglas -v- Hello! Limited at the end is omitted]

There is no conflict between section 12(3) and the Convention. (See Douglas -v- Hello! Ltd [2001] QB 967; Keene LJ paragraph 150)

…

12.

In the above paragraphs we have attempted to assist courts as to how they should go about the task of holding the balance between the conflicting rights when hearing these applications. We are suggesting that frequently what is required is not a technical approach to the law but a balancing of the facts. The weight which should be attached to each relevant consideration will vary depending on the precise circumstances. In many situations the balance may not point clearly in either direction. If this is the position, interim relief should be refused.”

Paragraph 11(iii) is but one of fifteen subparagraphs containing the A -v- B guidelines. The appeal there turned not on the meaning of “likely” but on whether the information in question (statements by two women concerning a footballer’s extra-marital sexual activities) was properly to be regarded as confidential. The meaning of likelihood, Miss Sharp tells us, was not even argued in that case and nor, indeed, was Imutran cited by counsel; rather it was the court itself which decided to base one of its guidelines on that authority.

One problem with this series of cases is that both the Vice-Chancellor in Imutran and this court in A -v- B appear to have regarded the views expressed in Douglas -v- Hello! Limited as entirely consistent with the Imutran approach, ie with a test, as Mr Tugendhat later put it in Theakston, “not discernibly different from the familiar interlocutory test of whether the case has a real prospect of success”. I confess to some difficulty with this. Not merely did Sedley LJ in Douglas -v- Hello! Limited speak (in paragraph 134) of “a probability of success at trial” (an expression which to my mind clearly supports his recollected conclusion on the point, as too does his use of the phrase in paragraph 135 - “no more than evenly balanced”), but Keene LJ’s judgment too seems to me consistent only with the balance of probability test. True, the reference at the beginning of paragraph 150 of Keene LJ’s judgment to “the normal meaning to be attached to the words in section 12(3)” (and the reference in paragraph 153 to “no strained reading of the language [being] needed”) may be said to beg rather than answer the critical question. Having, however, just outlined in the preceding paragraph Mr Tugendhat’s contended-for lesser meaning (“not fanciful” or “on the cards”), paragraph 150 was surely rejecting that contention and finding that the fuller meaning (“more likely than not”) would not conflict with the Convention. I find it difficult to read paragraph 150 as holding that the lesser meaning would not conflict, ie as saying merely that Mr Tugendhat’s argument was immaterial because in any event the ordinary meaning of the words supported his contended for construction. This attempt to construe Keene LJ’s judgment apart, it seems to me (as clearly it seemed to Ouseley J in Theakston) that the more obvious meaning of the words in s12(3) is “more likely than not” and that this must be what Keene LJ meant when he spoke of “the normal meaning”.

To my mind, therefore, we are left in this position. This court has twice expressed a view on the issue. In neither case did its view affect the outcome of the appeal and in both it can probably be regarded as obiter. The views are probably inconsistent although not recognised as such by the second court. There are also two conflicting decisions at first instance (the second of which, Theakston, was referred to by this court in A -v- B on another point, without comment on its rejection of the Imutran approach). Against this background, I conclude that we have no alternative but to decide the issue for ourselves.

I can take this quite shortly. I seems to me plain that the present issue was simply never raised below and certainly was never fully developed. Given that the entire hearing took place within a day that is perhaps less surprising than it might otherwise have been. In paragraphs 18 and 19 of his open judgment the judge referred to A -v- B and the guidelines there set out:

“These guidelines start with the normal guidelines applicable to the grant or refusal of an interim injunction. They refer to and assist in the application of section 12. They discuss the tension between the right to freedom of expression and the importance of the free press and the right to privacy. They refer, obviously enough having regard to the identity of the claimant in that case and the nature of the allegations, to the relevance of a public position of a claimant, or of the claimant having courted public attention or put himself forward to the public with a particular image. They refer to and discuss the Press Complaints Commission Code of Practice. Both counsel made detailed submissions to me by reference to these guidelines, which have assisted me in my task.”

I pass to paragraph 36:

“I must now apply the relevant legal principles to the facts. In some respects the facts alleged by [Miss Banerjee] are denied by [Cream], in other respects they are not. It seems to me that I am entitled to consider the detail of what is said and not said in the evidence, but clearly I cannot form a view as to who is likely to be found at any trial to be telling the truth where the evidence is contested. In order to establish its entitlement to an injunction Cream has first to show that it is likely, in the terms of section 12(3) of the Human Right Act, to be able to establish at trial that a permanent injunction should be granted to restrain publication. Since confidentiality is admitted, … the question is whether Cream is likely to be able to succeed at trial despite the defendants’ defences based on public interest and the Public Interest Disclosure Act.”

The judge then observed that “all these cases are necessarily fact sensitive” and that he must accordingly address the detailed facts (the subject of his private judgment to which reference was then made). Returning then to his public judgment, he said in paragraph 39 that there seemed to him a number of “seriously arguable issues both ways”, in particular as to whether the defences would succeed at trial and as to whether disclosure is justified in the public interest on the basis of correcting a false public image, and continued in paragraph 40:

“Having regard to the interpretation of the word ‘likely’ in the Human Rights Act section 12(3), approved in A -v- B, it seems to me that Cream has established the necessary likelihood of a permanent injunction. I do not say it is more likely than not, but there is certainly a real prospect of success.”

In short, the judge plainly directed himself in accordance with what this court in A -v- B described as the “useful guidance” provided by the Vice-Chancellor in Imutran and did so, moreover, without it having been made clear to him that he was thereby rejecting the different approach for which Miss Sharp was suggesting that Douglas -v- Hello! Limited stands.

It is not disputed that the word “likely” is capable of a range of definitions. So much, indeed, is clear from a glance at either Words and Phrases Legally Defined 3^rd Edition (1989) or Stroud’s Judicial Dictionary of Words and Phrases 6^th Edition (2000). Without, of course, attempting any comprehensive survey of the cases, it is, I think, instructive to consider one or two of them in a little detail.

This court in Bailey -v- Rolls-Royce (1971) Limited [1984] ICR 688 had to consider the meaning of the word likely in s72(1) of the Factories Act 1961: “A person shall not be employed to lift, carry or move any load so heavy as to be likely to cause injury to him.” Each member of the court concluded that “likely” in that context meant “more probable than not”. May LJ said that “there was clearly a risk that injury might occur, but I do not think that I can say that it was ‘likely’, or ‘probable’, or ‘more probable than not’”. Slade LJ agreed and said “this construction of the phrase is, I think, one which not only accords with the natural meaning of the words according to ordinary English usage, but also with what may be presumed to have been the intention of Parliament”. Stephenson LJ also agreed that the words in the section refer to “injury which is ‘probable’, and ‘probable’ can be expanded, or extended, and possibly clarified, to mean ‘more probable than not’.”

In a number of contexts, however, the word has been differently construed. Section 1 of the Children and Young Persons Act 1933 makes it an offence for anyone having care of a child to wilfully neglect the child “in a manner likely to cause him unnecessary suffering or injury to health”. In R -v- Sheppard [1981] AC 394 Lord Diplock said, at p405:

“The section speaks of an act or omission that is ‘likely’ to cause unnecessary suffering or injury to health. This word is imprecise. It is capable of covering a whole range of possibilities from ‘it’s on the cards’ to ‘it’s more probable than not’, but having regard to the ordinary parent’s lack of skill in diagnosis and to the very serious consequences which may result from failure to provide a child with timely medical attention, it should in my view be understood as excluding only what would fairly be described as highly unlikely.”

Professor John Smith, in his commentary appended to the brief report of R -v- Wills [1990] 2 Crim LR 714 in which Lord Lane CJ applied Lord Diplock’s dictum in Sheppard albeit acknowledging it was obiter, said this:

“Of course the words of a statute must be construed in their context but Lord Diplock’s interpretation of ‘likely’ seems very strained. If it only excludes what is ‘highly unlikely’ it includes what is merely ‘unlikely’; ie, the result is ‘likely’ to occur although it is unlikely (but not highly unlikely) to do so. With respect, that does not seem to be a possible meaning: ‘likely’ cannot include ‘unlikely’. On the contrary, a synonym for ‘likely’ is ‘not unlikely’.”

Again in a very different context, that of the legislation governing pre-action disclosure, Rix LJ recently said in Black -v- Sumitomo Corpn [2001] EWCA Civ 1819, 1584:

“70

The application has to be made by ‘a person … likely to be a party to subsequent proceedings’ against ‘a person … likely to be a party to the proceedings’ …

71

Of course, in one sense it might be said that a person is hardly likely to be party to subsequent proceedings whether as a claimant or otherwise unless some form of proceedings is itself likely to be issued. Two questions, however, arise. One is whether the statute requires that it be likely that proceedings are issued, or only that the persons concerned are likely to be parties if subsequent proceedings are issued. The other is whether ‘likely’ means ‘more probably than not’ or ‘may well’. As to the first question, in my judgment the amended statute means no more than that the persons concerned are likely to be parties in proceedings if those proceedings are issued. …

72

As to the second question, it is not uncommon for ‘likely’ to mean something less than probable in its strict sense. It seems to me that if I am wrong about the first question, then it is plain that ‘likely’ must be given its more extended and open meaning (see Lord Denning MR in Gunning’s case [Gunning -v- United Liverpool Hospitals’ Board of Governors [1973] 1 WLR 586]) because otherwise one of the fundamental purposes of the statute will have been undermined. If, however, I am right about the first question, the second question is of less moment. Even so, however, I am inclined to answer it by saying that ‘likely’ here means no more than ‘may well’. Where the future has to be predicted, but on an application which is not merely pre-trial but pre-action, a high test requiring proof on the balance of probability will be both undesirable and unnecessary: undesirable because it does not respond to the nature and timing of the application; and unnecessary, because the court has all the power it needs in the overall exercise of its discretion to balance the possible uncertainties of the situation against the specificity or otherwise of the disclosure requested.”

More recently still, in Venables & Thompson -v- News Group International (transcript, 4 December 2001), Dame Elizabeth Butler-Sloss P was concerned with an alleged breach of a court order enjoining newspapers against the publication of “any information likely to lead to the identification of the [claimants’] past, present or future whereabouts”.

Paragraph 20 of her judgment reads:

“I bear carefully in mind, as I must, the great importance of clarity in the language of injunctions which may lead to committal proceedings. The words ‘likely to lead to’ were drafted by counsel in the injunction proceedings, amongst them [counsel for the newspaper groups] … It was not then suggested to be ambiguous or uncertain. The words are now challenged as unclear or imprecise. I have been directed to the dictionary meaning. Collins Dictionary of the English Language gives a number of definitions of the word ‘likely’. They include: tending or inclined, apt; probable; having good possibilities of success. I have also had my attention drawn to reported decisions on the use of the words ‘likely’ and ‘probable’. Those decisions and others point to the fact that the word ‘likely’ may be used in different contexts in different ways. There will undoubtedly be cases in which the use of the word ‘likely’ will be properly equated with ‘probable’. In other cases it has been used to mean something which could or might well happen. Words have to be seen in the context of the sentence and in the context of the situation in which they are used. In my judgment a detailed consideration of the decisions cited to me does not help me to come to a conclusion in the present case. The purpose of the injunctions is to protect the lives and personal safety of these boys. [In my earlier judgment] I used the phrase ‘real possibility that their lives would be at risk’. The use of the word ‘likely’ in the order is not to be equated with statistical probability that it will lead to the identification of the boys or their whereabouts but to the real risk, the real danger, the real chance that it may lead to that dangerous situation.”

The final, and perhaps most illuminating, of the authorities to which I think it helpful to refer is In re Harris Simons Construction Limited [1989] 1 WLR 368. Hoffmann J there was concerned with s8(1) of the Insolvency Act 1986 which gives the court jurisdiction to make an administration order if it “(a) is satisfied that a company is or is likely to become unable to pay its debts” and “(b) considers that the making of an order … would be likely to achieve” one or more of the purposes specified in s8(3). Peter Gibson J in an earlier case had said:

“As I read s8 the court must be satisfied on the evidence put before it that at least one of the purposes in s8(3) is likely to be achieved if it is to make an administration order. That does not mean that it is merely possible that such purpose will be achieved; the evidence must go further than that to enable the court to hold that the purpose in question will more probably than not be achieved.”

Hoffmann J, having cited that passage from Peter Gibson J’s judgment, continued as follows:

“He therefore required that on a scale of probability of 0 (impossibility) to 1 (absolute certainty) the likelihood of success should be more than 0.5. I naturally hesitate to disagree with Peter Gibson J., particularly since he had the benefit of adversarial argument. But this is a new statute on which the judges of the Companies Court are still feeling their way to a settled practice and I therefore think I should say that in my view he set the standard of probability too high. My reasons are as follows. First, ‘likely’ connotes probability but the particular degree of probability intended must be gathered from qualifying words (very likely, quite likely, more likely than not) or context. It cannot be a misuse of language to say that something is likely without intending to suggest that the probability of its happening exceeds 0.5, as in ‘I think that the favourite, Golden Spurs at 5-1, is likely to win the Derby’. Secondly, the section requires the court to be ‘satisfied’ of the company’s actual or likely insolvency but only to ‘consider’ that the order would be likely to achieve one of the stated purposes. There must have been a reason for this change of language and I think it was to indicate that a lower threshold of persuasion was needed in the latter case than the former. The first of the sentences I have quoted from the judgment of Peter Gibson J. suggests that he did not take this variation into account. Thirdly, some of the stated purposes are mutually exclusive and the probability of any one of them being achieved may be less that 0.5 but the probability of one or other of them being achieved may be more than 0.5. I doubt whether Parliament intended the courts to embark on such calculations of cumulative probabilities. Fourthly, as Peter Gibson J. said, section 8(1) only sets out the conditions to be satisfied before the court has jurisdiction. It still retains a discretion as to whether or not to make the order. It is therefore not unlikely that the legislature intended to set a modest threshold of probability to found jurisdiction and to rely on the court’s discretion not to make orders in cases in which, weighing all the circumstances, it seemed inappropriate to do so. Fifthly, the Report of the Review Committee on Insolvency Law and Practice (1982), (Cmnd. 8558), para. 508, which recommended the introduction of administratorship, said that the new procedure was likely to be beneficial

‘only in cases where there is a business of sufficient substance to justify the expense of an administration, and where there is a real prospect of returning to profitability or selling as a going concern.’

Elsewhere the report speaks of an order being made if there is a ‘reasonable possibility’ of a scheme of reconstruction. I think that this kind of phraseology was intended to be reflected in the statutory phrase ‘considers that [it] would be likely’ in section 8(1)(b).

For my part therefore, I would hold that the requirements of section 8(1)(b) are satisfied if the court considers that there is a real prospect that one or more of the stated purposes may be achieved. It may be said that phrases like ‘real prospect’ lack precision compared with 0.5 on the scale of probability. But the courts are used to dealing in other contexts with such indications of the degree of persuasion they must feel. ‘Prima facie case’ and ‘good arguable case’ are well known examples. Such phrases are like tempo markings in music; although there is inevitably a degree of subjectivity in the way they are interpreted, they are nevertheless meaningful and useful.”

These cases, at one in recognising that the meaning of the word “likely” depends always on the context in which it is found, suggest too a number of considerations tending to point one way or another towards the intended meaning. Hoffmann J in In re Harris Simons Construction, for example, had regard amongst other things to whether court had to be “satisfied” of the relevant likelihood or only to “consider” it, the consequences of finding likelihood established (there to found the court’s discretion), and what I may loosely call the travaux preparatoires.

I turn next to the Parliamentary material, researched at our request and provided to us following the conclusion of the hearing. Subsection 3 of s12 came to be introduced into the Human Rights Act at a very late state in the legislative process, the third reading of the Bill in Committee in the House of Commons. The Home Secretary, Mr Straw, said on 2 July 1998:

“Subsection (3) provides that no relief is to be granted to restrain publication pending a full trial of the issues unless the court is satisfied that the applicant is likely to succeed at trial. Among concerns expressed about the Bill’s possible impact on the freedom of the press, there was concern that interim injunctions … might be granted simply to preserve the status quo with a view to a full hearing of the application later. However, by that time the story that was to be published might no longer be newsworthy. As I said earlier, time and again the convention jurisprudence reinforces the freedom of the press against, for example, the assertion of rights under article 8. One example of that is part of the judgment of the European Court of Human Rights in the 1991 ‘Spycatcher’ case. Dealing with the issue of interlocutory relief, the court said:

‘News is a perishable commodity and to delay its publication for even a short period may well deprive it of all its value and interest.’

Given that, we believe that the courts should consider the merits of the application when it is made and should not grant an interim injunction simply to preserve the status quo ante between the parties.

… [W]e believe that the new clause would protect a respondent potential publisher from what amounts to legal or legalised intimidation. We have already discussed the difficulty of getting interlocutory relief. It will be very difficult to get it unless the applicant can satisfy the court that the applicant is likely to establish that publication should not be allowed. That is a much higher test than that there should simply be a prima facie case to get the matter into court.”

The appellants, contending for the balance of probability test, suggest that this is the more natural and ordinary meaning of the word “likely” in this context. Unlike Hoffmann J’s 5-1 Derby favourite, litigation is usually only a two horse race. It is one thing to say of a horse carrying odds of 5-1 against that it is likely to win when more probably it will not; quite another to say that of one of two parties to litigation. As Miss Sharp asked rhetorically: “What would a losing client think if, having lost, his lawyer then sought to justify previous advice that he was likely to win by saying he had meant only that there was a real prospect of winning?”. The appellants point too to the court having to be “satisfied” of the relevant likelihood under s12(3), not merely having to consider it - see Hoffmann J’s second reason in In re Harris Simons Construction.

Miss Sharp further relies upon the long-established principle in the closely related field of defamation law that where the defendant contends that the words complained of are true and swears that he will plead and seek to prove the defence of justification, the court will not grant an interlocutory injunction unless, exceptionally, it is satisfied that the defence is one which cannot succeed. This is often called the rule in Bonnard -v- Perryman ([1891] 2 Ch 269) and it appears to extend also to the defences of privilege and fair comment. In defamation therefore, it is even harder to obtain interlocutory relief than were the claimant facing the suggested balance of probability test in s12(3). So much the more likely, submits Miss Sharp, that Parliament was intending by s12(3) to introduce a test at least as stringent as that rather than the lower and less precise threshold test of a real prospect of success, a test so low, indeed, that a failure to meet it would in any event render the claim vulnerable to strike out by summary judgment under CPR 24.2 and amounts to little more than was previously required by the American Cyanamid approach - see American Cyanamid -v- Ethicon Limited [1975] AC 396, 407:

“The court no doubt must be satisfied that the claim is not frivolous or vexatious; in other words that there is a serious question to be tried.”

Above all, however, the appellants emphasise and rely upon the Human Rights context in which this clause falls to be construed, that of freedom of expression and, moreover, Parliament’s obvious concern to strengthen the law against prior restraint. The ECtHR in Observer Limited and Guardian Newspapers Limited -v- United Kingdom (1991) A216 14 EHRR 153 (the 1991 Spycatcher case to which Mr Straw referred when moving the new clause at third reading) stated in paragraph 59 of its judgment two major principles established by its earlier jurisprudence:

“(a)

Freedom of expression constitutes one of the essential foundations of a democratic society; subject to paragraph (2) of Article 10, it is applicable not only to ‘information’ or ‘ideas’ that are favourably received or regarded as inoffensive or as a matter of indifference, but also to those that offend, shock or disturb. Freedom of expression, as enshrined in Article 10, is subject to a number of exceptions which, however, must be narrowly interpreted and the necessity for any restrictions must be convincingly established.

(b)

These principles are of particular importance as far as the press is concerned. Whilst it must not overstep the bounds set, inter alia, in the ‘interests of national security’ or for ‘maintaining the authority of the judiciary’, it is nevertheless incumbent on it to impart information and ideas on matters of public interest. Not only does the press have the task of imparting such information and ideas; the public also has a right to receive them. Were it otherwise, the press would be unable to play its vital role of ‘public watchdog’.”

The court continued in paragraph 60:

“For the avoidance of doubt, … the Court would only add to the foregoing that Article 10 of the Convention does not in terms prohibit the imposition of prior restraint on publication, as such. This is evidenced not only by the words ‘conditions’, ‘restrictions’, ‘preventing’ and ‘prevention’ which appear in that provision, but also by the Sunday Times judgment of 26 April 1979 and its Markt Intern Verlag GmbH and Klaus Beerman judgment of 20 November 1988. On the other hand, the dangers inherent in prior restraints are such that they call for the most careful scrutiny on the part of the Court. This is especially so as far as the press is concerned, for news is a perishable commodity and to delay its publication, even for a short period, may well deprive it of all its value and interest.”

Professor Colin Munro, summarising that and other more recent Strasbourg cases in his valuable article Prior Restraint in the Media and Human Rights Law in the Scottish journal Juridical Review (2002), stated at paragraph 23:

“There cannot as yet be said to be a ‘bright-line’ rule against judicial prior restraint in ECHR law. However, it is clear that prior restraints are viewed as pernicious and that, to be upheld as justifiable, their use will have to be viewed as appropriate, proportionate and absolutely necessary.”

Ouseley J was right, submits Miss Sharp, to have concluded in Theakston (see paragraph 21 above) that Parliament “intended to impose the discernibly more rigorous requirement [than merely a real prospect of success] which it did in this particular context of freedom of expression.”

The respondents submit that the reasoning in Theakston is unconvincing: as other authorities on the meaning of the word “likely” amply establish, there is no reason why, as a matter of ordinary English, something should not be likely to happen even though more probably it will not. As for Ouseley J’s second reason, that Parliament would have used the phrase “a real prospect of success” if that was what was intended, Mr Bartley Jones QC advances the countervailing argument that if Parliament had intended the test to be that of more likely than not, then that is the test which would have been stipulated. Indeed Mr Bartley Jones adopted a suggestion made by Arden LJ in argument that, if what was intended was the balance of probability test, there would have been no need for any reference to likelihood at all: the subsection would simply have read “… unless the court is satisfied that the applicant will establish …” (not “… unless the court is satisfied that the applicant is likely to establish”), the word “satisfied” itself implying the balance of probability test. That proposition I would reject, not least as being irreconcilable with Hoffmann J’s highly persuasive judgment in In re Harris Simons Construction.

To my mind a better argument for the respondents is one which mirrors the fifth of Hoffmann J’s reasons in In re Harris Simons Construction, namely that s12(3) merely sets out a pre-condition to be satisfied before the court has the power to grant interlocutory injunctive relief and that even when this threshold is crossed the court can still refuse to make the order sought. Echoing Hoffmann J, Mr Bartley Jones submits that “it is therefore not unlikely that the legislature intended to set a modest threshold of probability to found jurisdiction and rely on the court’s discretion not to make orders in cases in which, weighing all the circumstances, it seemed inappropriate to do so”.

A further submission made by the respondents is that it would inevitably be very difficult for the claimant to cross the higher threshold test of balance of probabilities without in fact the evidence being heard. This submission in turn reflects reasoning to be found in earlier authority - see paragraph 72 of Rix LJ’s judgment in Black -v- Sumitomo Corporation (set out in paragraph 35 above):

“Where the future has to be predicted, … a high test requiring proof on the balance of probability will be both undesirable and unnecessary …”

What to my mind, however, is substantially the strongest of the respondents’ arguments is that, to construe s12(3) as imposing the higher threshold test contended for by the appellants is necessarily to rank the right to freedom of expression above all possible countervailing rights (privacy, proprietary rights in confidential material and so forth) so that, sometimes at least, to give effect to s12(3) so construed will be to interfere disproportionately with other basic rights. Postulate a claimant seeking to protect his right to privacy who, if publication were to go ahead, would suffer catastrophic, irreversible and uncompensatable harm (whether to his career, his family life or whatever). Assume that at the interlocutory stage he can establish only a 40% prospect of being able at trial to make good his right to privacy. Assume too, however, that an expedited trial could take place in a month’s time without the proposed delay in publication meantime causing the defendants any appreciable loss, the story not being of the “perishable” kind.

On the appellants’ construction no interim injunction could be granted. Were the claimant, however, then to succeed in his claim at trial and establish that publication had involved an unlawful and unwarranted intrusion into his private life, surely he would have legitimate grounds for complaint, not least as to the courts’ inability to grant him the interim protection he had sought.

In support of this argument the respondents rely upon another part of the court’s guidelines in A -v- B, at paragraph 11(xii):

“On the difficult issue of finding the right balance, useful guidance of a general nature is provided by the Council of Europe Resolution 1165 of 1998 … in these terms:

‘11. The Assembly reaffirms the importance of every person’s right to privacy, and of the right to freedom of expression, as fundamental to a democratic society. These rights are neither absolute nor in any hierarchical order, since they are of equal value.’”

There is no inconsistency between A -v- B guideline (xii) and what, for example, Lord Steyn said in Reynolds -v- Times Newspapers Ltd [2001] 2 AC 127, 207 as to the right to freedom of expression being a constitutional right and that “[b]y categorising this basic and fundamental right as a constitutional right its higher normative force is emphasised”. It is one thing to say, as indeed I myself said in Al-Fagih -v- HH Saudi Research & Marketing (UK) Ltd [2001] 2 EMLR 215, in a passage repeated with approval by this court in Loutchansky -v- Times Newspapers Ltd (Nos 2-5) [2002] QB 783, 803, that the media’s right to freedom of expression, particularly in the field of political discussion “is of a higher order” than “the right of an individual to his good reputation”; it is, however, another thing to rank it higher than competing basic rights. See too in this regard what Sedley LJ said in paragraph 135 of his judgment in Douglas -v- Hello! Limited,set out in paragraph 17 above.

Section 3 of the Human Rights Act legislation (including therefore s12(3)) is to “be read and given effect in a way which is compatible with the Convention rights”. This in my judgment requires that s12(3) will in all cases be able to be applied compatibly with and as between competing human rights. It seems to me essentially this argument which Mr Tugendhat QC was advancing in Douglas -v- Hello! Limited as described in paragraph 149 of Keene LJ’s judgment there - see paragraph 15 above.

There is an important distinction between the American Cyanamid approach and that now required by s12(3) even if the threshold test created by the latter is, as I would hold, that of a real prospect of success convincingly established. The distinction is that under American Cyanamid the court is concerned only to find a serious issue to be tried, not to resolve either conflicts of fact or difficult questions of law so as to gauge the merits of the claim. I accordingly have some difficulty with the Vice-Chancellor’s views expressed in Imutran - see paragraph 19 above. It seems to me that there will indeed be a number of claims for injunctive relief which now will fail when earlier they would have succeeded: they will fail because the court is required by s12(3) actually to consider their merits (so as to reach a judgment as to the prospects of their eventual success) and cannot grant relief unless satisfied on cogent evidence that the claim does indeed have a real prospect of succeeding at trial notwithstanding the defendant’s ex hypothesi conflicting right to freedom of expression. To construe s12(3) as I would propose is by no means to rid it of all force and effect. Nor is the comparative imprecision of the test of real prospect of success fatal to its adoption - see the final paragraph of Hoffmann J’s judgment in In re Harris Simons Construction set out in paragraph 40 above.

What Mr Straw said in Committee on 2 July 1998 (see paragraph 41 above) is, I would suggest, consistent equally with both contended-for constructions. On either side’s approach courts in future will not “grant an interim injunction simply to preserve the status quo ante” and the new test will be “much higher … than that there should simply be a prima facie case to get the matter into court”.

Although, as I have already indicated (see paragraph 25 above), it seems to me that the more obvious meaning of the words in s12(3) is “more likely than not to establish” (rather than “has a real prospect of establishing”), the latter meaning is undoubtedly an available one on the language used, the synonym for the word “likely” being, as Professor Smith suggested (see paragraph 34 above), “not unlikely”.

Recognising, as of course I do, the importance of the right to freedom of expression - “the lifeblood of democracy” as Lord Steyn described it in R -v- Home Secretary (ex parte Simms) [2000] 2 AC 115, 126 - an importance, indeed, to which s12(4) of the Human Rights Act itself requires us to have “particular regard”, and accepting too, as I do, that the higher threshold test for which the appellants contend represents the more natural meaning of the words used, it is tempting to construe subsection (3) in this way. I have concluded, however, that it would be wrong to do so given the impact that such a construction would have on competing Convention rights as I have sought to explain in paragraphs 51 and 52 above. In short, I have come to accept the argument which Mr Tugendhat appears unsuccessfully to have advanced, first in Douglas -v- Hello! Limited and later, notwithstanding the assistance he then had from the Vice-Chancellor’s judgment in Imutran, in Theakston.

Although the reasoned rejection of that argument by both Keene LJ and Sedley LJ in Douglas -v- Hello! Limited (see respectively paragraphs 16 and 17 above) for the most part seems compelling, the one feature of the case which I would respectfully suggest neither judgment directly addresses is the different situation arising at the interlocutory stage from that which will exist at trial. Of course it is possible at the interlocutory stage for the court to look ahead and reach a judgment as to the prospects of success at trial and clearly, in determining those prospects, the court will have regard to the possible risks and consequences of interfering with the respective rights which each party is seeking to vindicate. That does not, however, take account of the possibility that in certain cases, although the claimant at the interlocutory stage looks more likely eventually to lose than to win, there may be compelling reasons why nevertheless his ultimate chance of victory, odds against though it be, should be preserved rather than pre-empted at the interlocutory stage.

That is the possibility I have sought to illustrate above and, of course, on the appellants’ contended-for construction, the court is forbidden to have regard to it. It is this which has convinced me that s12(3) must after all be construed to provide the lower threshold test. That is not, of course, to say that, whenever the test is satisfied, the court will grant interlocutory relief. On the contrary, the lower threshold test merely gives the court a discretion in a greater number of cases. Often the court will not think it right to exercise that discretion in favour of prior restraint unless it is indeed satisfied that the claim will more probably than not succeed at trial. As, moreover, is stated in paragraph 12 of the judgment in A v- B (see paragraph 23 above):

“In many situations the balance [between conflicting rights] may not point clearly in either direction [in which event] interim relief should be refused.”

There is, of course, no inconsistency between this paragraph and the test of real prospect of success. How could there be, given, as already indicated, this courts adoption of that lower test in paragraph 11(iii) of the same judgment. I reject Miss Sharp’s submission to the contrary.

I return briefly to the facts of the present case about which, for reasons already explained, I can say but little in this open judgment.

The judge below, having concluded in paragraph 40 of his judgment (see paragraph 29 above) that “there is certainly a real prospect of success”, then turned to the other considerations arising. In reaching all his conclusions he expressly took account of the specific factors referred to in s12(4) of the Act. Finding that damages would not be an adequate remedy for either side, he passed in paragraph 44 to the balance of convenience:

“Here, if the story is published, but should not have been, Cream is likely to suffer an irreparable loss of an unquantifiable nature. By contrast, if the publication is restrained it will delay the second defendant’s story but subject to the time taken to get to trial it would not necessarily preclude its publication altogether. It seems to me that for these reasons the classic balance of convenience test favours the granting of an interim injunction, possibly with directions for a speedy trial.”

Turning then to the importance of the Convention right to freedom of expression, and again reminding himself of the A -v- B guidelines, he posed the question:

“Can the grant of an injunction, even for what may be a few months until a speedy trial, be justified as an interference with freedom of expression and in particular with that on the part of the press?”

That question he finally answered in the affirmative. He did so above all because of Ms Banerjee’s “undoubted obligation of confidentiality” which, he accepted, having regard to her employment as an in-house chartered accountant, placed her under “a high-level duty”. The injunction did not, I should make plain, prevent her reporting any alleged financial irregularities to any criminal or regulatory authority. Rather it banned press publication. As already stated, the breach of confidentiality involved in the proposed publication is undoubted. The question at trial will be whether the appellants can establish that publication is nonetheless justified in the public interest, either as the disclosure of wrongdoing or the correction of a false public image.

In the light of the confidential material before us and the judge’s findings upon it as set out in his private judgment, I take the view that he was entitled to form the judgment that Cream had convincingly established a real prospect of success at trial. Not every judge would necessarily have reached that same conclusion; indeed, I myself might well not have done so. That, however, is not the test. The test is whether, the judge having correctly directed himself, as I am satisfied he did, on all issues of law, his judgment falls outside “the generous ambit within which a reasonable disagreement is possible” - see Lord Fraser’s speech in an analogous context in G -v- G (Minors: custody appeal) [1985] 1 WLR 647, 652. In my judgment it did not. Having, since writing this section of the judgment, read both Sedley LJ’s and Arden LJ’s confidential as well as public judgments, I need add no more than that this remains my view. I direct that these confidential judgments shall be available only to the parties and their advisors and to the regulatory authorities excluded from the order of Timothy Lloyd J below unless and until the injunction granted by him is discharged or expires.

I would accordingly dismiss this appeal.

By way of final reflection I add just this. Whatever pre-condition to the grant of interim relief is set by s12(3), were the subsection always to be applied literally it is difficult to see how in many cases publication could properly be enjoined even for as long as is necessary for a proper interlocutory decision to be arrived at. Take the present case. Had the Echo not agreed to delay publication for the three weeks it took to arrange for a one-day hearing, it may be doubted whether the court could properly have forbidden the publication. And what if the judge needs time to consider his conclusion and reserves judgment? Certainly injunctions in these circumstances would be impossible if the higher test were adopted. Rather there appears some need for flexibility. I do not, however, include this reflection within my reasoning; after all, in many cases it is only as a result of the media’s voluntary action in bringing their proposed publication to the claimant’s attention that any possibility of prior restraint arises in the first place.

It seems to me, as it does to Simon Brown LJ, that the natural meaning of the word ‘likely’ in the immediate context of s.12(3) of the Human Rights Act is ‘more probable than not’. As Ouseley J said in Theakston -v- MGN Ltd [2002] EMLR 398, it is not easy to envisage, as a matter of ordinary English, an injunction which is likely to be granted but which more probably than not will be refused. It seems to me no easier as a matter of logic to say that it is likely that an injunction will be granted but more likely that it will not be.

To say this is not to cast doubt on the many instances in which, in other contexts, the statutory purpose has been held to require one of the milder meanings of ‘likely’ also to be found in the dictionary. A good example – and I cite it for a reason to which I shall be coming - is the decision of this court in Black -v- Sumitomo [2002] 1 WLR 1562, a case on pre-action disclosure by a person likely to be a party to eventual proceedings, where at paragraph 72 Rix LJ said:

“As to the second question, it is not uncommon for ‘likely’ to mean something less than probable in its strict sense. It seem to me that if I am wrong about the first question, then it is plain that ‘likely’ must be given its more extended and open meaning (see Lord Denning MR in Dunning’s case), because otherwise one of the fundamental purposes of the statute will have been undermined. If, however, I am right about the first question, the second question is of less moment. Even so, however, I am inclined to answer it by saying that ‘likely’ here means no more than ‘may well’. Where the future has to be predicted, but on an application which is not merely pre-trial but pre-action, a high test requiring proof on the balance of probability will be both undesirable and unnecessary: undesirable, because it does not respond to the nature and timing of the application; and unnecessary, because the court has all the power it needs in the overall exercise of its discretion to balance the possible uncertainties of the situation against the specificity or otherwise of the disclosure requested.”

When therefore in Douglas -v- Hello!, paragraph 134, I paraphrased s.12(3) as requiring a probability of success at trial, it was in the ‘strict sense’ described by Rix LJ that I was using the word. The other two members of the court, Brooke LJ and Keene LJ, used the statutory word ‘likely’ without expanding it. Our decision did not depend on the precise register of the word. Having now heard full argument on the present appeal, I am satisfied for the reason I have given that in s.12(3) the word in its immediate context does means probable in its strict sense.

But the overarching question which emerged late in the argument before us, and which I now think is crucial, is whether the process of construction stops there, or whether s.3 of the Human Rights Act itself requires us to read the word ‘likely’ down from the meaning conferred by its immediate context in order to make s.12(3) Convention-compliant. There is no reason whatever to suppose that Parliament, in mandating us so far as possible to read all legislation and give it effect compatibly with the Convention rights, did not mean to include the Human Rights Act itself in that obligation.

Suppose, first of all, a case where it is manifest that if an interim injunction is not granted the claimant may be seriously and irreparably harmed, while if it is granted the defendant will suffer little if at all. Where s.12 does not apply – that is, in all but freedom of expression cases – the fact that the legal or factual basis of the claim at the interlocutory stage appears weak is a factor, but by no means a conclusive one, in the court’s decision as to how best to hold the situation until trial. The same is true of a case where the most the judge can find is that the chances of final success are even, as in my reading of paragraph 40 of his open judgment, Lloyd J found in the present case. That, it seems to me, is as it should be: not only is a modest risk of a catastrophic event something against which the court ought to be able temporarily to guard a claimant; every lawyer knows that by the time of trial the claimant’s prospects of success may have grown, sometimes through the disclosure of the defendant’s own documents, to a very strong probability. It takes a great deal, in my judgment, to justify a rule which denies such a claimant the possibility of the temporary protection of the law without which success at trial will be hollow.

The case of Venables -v- News Group Newspapers Ltd [2001] EMLR 255 furnishes a disturbing illustration. The President was able to grant the appropriate injunctions without regard to s.12(3) because her decision was final, not interlocutory. But suppose that an emergency application had had to be made for a temporary restraint on the publication of the claimants’ identities and whereabouts pending trial, and that the evidence of a threat to their lives and safety consequent on publication had not been strong, so that one could not say that they were probably going to succeed at trial. Despite the gravity of the possible consequences of publication and the inconsequentiality of a moratorium until trial, the President’s hands would have been tied by s.12(3). Her eventual decision in the claimants’ favour would have been entirely abortive.

To the interpretation of statutes by close reading, which has historically been the first principle of judicial construction, s.3 has now added a second and more radical principle which seeks to give effect to a new Parliamentary intent: that the meaning to be ascribed to its words, whenever they were enacted, is so far as possible to be a meaning which respects Convention rights. One has only to look at a dictionary to see that the word ‘likely’ in s.12(3) is capable of bearing what Rix LJ calls a more extended and open meaning than literal probability. But is it capable of bearing it in the context of s.12(3), or does such an extension go beyond what is legally and linguistically possible? And if it is permissible, is it requisite?

I am satisfied that ‘likely’ in s.12(3) is quite capable of bearing its milder meaning. While the logical difficulty remains of something being likely but its opposite being more likely, the practical usage of the word to convey something which is reasonably or realistically possible has ample judicial support in other contexts; and the full present context is larger than s.12(3) or s.12 itself, for it includes s.3.

That is the first question. The second, a problematical one, is whether ‘likely’ is to be read down uniformly or only in cases where a Convention right other than the right of free expression is also at stake. Section 3 suggests the latter, however asymmetrical and unsatisfactory the consequent two-track system. But I do not think we need to answer the question now, because the present case in my judgment does involve a countervailing Convention right.

The expression ‘Convention right’ is defined by s.1(1) to mean the rights and fundamental freedoms set out in, among others, articles 2 to 12 “as read with articles 16 to 18” of the Convention. These provisions include, in paraphrase:

So far as concerns article 6, there is in my view nothing fair about a hearing which establishes a right to the protection of the law which has meanwhile been rendered completely nugatory, not because there was not a suitable case for interim relief but because the court’s hands were tied.

If, as has been held to be the case, a corporation is capable of enjoying privacy rights under article 8, the disclosure of its affairs – especially shameful ones – is on the face of it a breach of respect for its privacy; though the more shameful the matter disclosed, the less likely it is to survive the application of article 8(2). It must also be borne in mind that article 8(2), unlike the three succeeding articles, is specific in targeting public authorities – the court, therefore, but not the present defendants – as those whose interference with the privacy of others must be justified. For these reasons I would regard this limb of the claimants’ arguments as difficult to bring home.

Most directly in point is the restriction by article 10(2) of freedom of expression in order to protect confidential information. But is this protection a Convention right within the meaning of s.1(1), or is it simply a right which, if and to the extent that it exists in national legal systems, is allowed in a proper case to cut down the Convention right of free expression? We have not had full argument on the point, but my present view is that it is the latter. It sits among other qualifying tests – national security, public safety and so forth – which cannot intelligibly be called Convention rights. They are all, I would rather say, Convention values. The same must apply to the maintenance of the authority of the judiciary, which also features in article 10(2).

Article 17 does, however, seem to me to matter. It brings the article 6 right to a fair trial into the same frame as article 10(1), and by doing so denies the right of free expression a clear run where it comes into conflict with the right to a fair trial. It does the reverse too. A proper balance has in each case to be found between the two.

What is the effect of these conclusions on the present case? It is in my judgment that the strict meaning of ‘likely’ in the immediate context of s.12(3) is required by Parliament to give way to a wider meaning which will allow a just balance to be found between public information and the needs of justice. For reasons which I have given above, such a balance is not compatible with the probabilistic threshold of a better than even prospect of success, below which interim relief, however necessary for a fair trial, cannot be granted. It is compatible with a meaning of ‘likely’ which corresponds with a realistic possibility of eventual success in preventing publication. That is therefore the meaning which we are required by s.3 to give it.

I have read with great interest the draft of Arden LJ’s judgment which reaches the same conclusion by a different and more pragmatic route. Save where it takes issue with the materiality of s.3 of the Human Rights Act, I have no quarrel with it at all. If the statutory road were not laid out before us, as Simon Brown LJ and I consider it to be, I would readily take Arden LJ’s equitable path to the conclusion which she neatly expresses: the judge must be satisfied that there is no obvious reason why the claim should not succeed.

The only caveat I would wish to enter to the judgment of Simon Brown LJ on this aspect of the appeal concerns his suggestion in paragraph 61 that the court will often decide in its discretion to grant prior restraint only when it is satisfied that eventual success is more probable than not. I do not, with great respect, share this view. The fact that what is being sought is prior restraint of a publication which article 10 is designed to protect will be one of the factors in the appraisal of likelihood of success at trial. There is a risk of double-counting if it is used again as a ground for withholding relief.

On the foregoing footing the approach of Lloyd J in his open judgment to the grant of relief cannot be faulted. What remains is the question whether, although he applied a proper threshold test, he erred in law when he concluded, in his closed judgment, that there was likely to be no public interest justification for the disclosure which, pruned of its speculative elements, is the essential story which the first defendant has given the second defendant and the second defendant wants to publish.

As to this, I consider, with respect, that Lloyd J was wrong. Reverting to Arden LJ’s phrase, there was in my view an obvious reason why the claim for an injunction ought not to succeed. In this public judgment I can only give my reasons delphically, but I will add a short closed judgment since its content may be relevant at the trial of the action.

It will be apparent from the facts which are already known that the proposed disclosure is essentially true – otherwise the action would be in defamation. It is also known that the origin of the disclosure is the first appellant’s employment with the respondents as an in-house accountant. For reasons which I will set out in my closed judgment, I consider that the principal matter which the Echo wants to publish is incontestably a matter of serious public interest – certainly by comparison with the mildly pornographic material which the defendants in A -v- B and C succeeded in investing with this quality.

It is submitted for the claimants that there is nevertheless a real possibility that they will succeed at trial in preventing publication because the newspaper’s information derives directly from a breach of confidentiality on the part of the first defendant. So it does: it has never been denied. Moreover, on authority the newspaper can be no better placed than its informant; indeed, even without a delinquent informant – if, say, the evidence had been found in the street – the quality of confidentiality attaches to it. But it is also a principle of law that no confidence attaches to iniquity.

How are these two principles to be reconciled? It is said by the claimants that to let the second trump the first is to give carte blanche to any employee to go straight to the press with damaging disclosures about his or her employers; and if employees, why not independent professional advisers? The answer lies in the concept of graduated responses for which the Employment Rights Act 1996 as amended by the Public Interest Disclosure Act 1998 now provides a template. Lloyd J derived assistance from it. But what it seems to me he failed to recognise was that the first defendant had exhausted these responses. The detail, regrettably, has to be relegated to my closed judgment.

In these circumstances I see no real possibility of the claimants’ succeeding at trial. They could do so only on the footing that the duty of confidentiality was inviolable, and in law it is well established that it is not. The argument from consequences has to be taken seriously, but there are several answers to it. So far as concerns employees, there is now a statutory framework which the employee will step outside at his or her peril; but this the first defendant did not do. So far as concerns auditors, lawyers and others, there is a strong code of professional standards the breach of which may result in the loss of their livelihoods; but it is not suggested that the first defendant faces any such charge. The essential story is in my view one which, whatever its source, no court could properly suppress.

I would therefore discharge the injunction and allow publication forthwith.

I agree with the judgments of Simon Brown and Sedley LJJ on the meaning of the word “likely” in section 12(3) of the Human Rights Act 1998 but respectfully disagree with them on the reasons for that conclusion.

Section 12 of the 1998 Act provides in material part as follows:-

“12(1) This section applies if a court is considering whether to grant any relief which, if granted, might affect the exercise of the Convention right to freedom of expression.

…

(3)

No such relief is to be granted so as to restrain publication before trial unless the court is satisfied that the applicant is likely to establish that publication should not be allowed.

(4)

The court must have particular regard to the importance of the Convention right to freedom of expression …”

Section 12 of the 1998 Act applies independently of whether any party to the litigation is a public authority. It is common ground that the appellants are entitled to the Convention right of freedom of expression. The appellants accept that the respondent, if not entitled to the Convention right of privacy, is entitled to rights which are equivalent to that Convention right.

The present case is an action for breach of confidence. If a breach of confidence occurred, the second appellant (“the Echo”) was aware of all the relevant circumstances. This is not a case where defamation is alleged and where the Echo has indicated its intention to prove justification at trial. In such a case, the court will not grant an interim injunction to restrain publication unless it is clear that the plea of justification is bound to fail: Bonnard -v- Perryman [1891] 2 Ch 269, Holley -v- Smith [1998] QB 726. Nor is this a case where there is a strong case for publication in the public interest of the alleged confidential material of the nature that was held to exist in Lion Laboratories Ltd -v- Evans [1985] QB 526.

English law attaches considerable importance to freedom of expression. In R -v- Secretary of State for the Home Department, ex parte Simms [2000] 2 AC 115 at 126 Lord Steyn explained that:

“… freedom of speech is the lifeblood of a democracy. The free flow of information and ideas informs political debate. It is a safety valve: people are more ready to accept decisions that go against them if they can in principle seek to influence them. It acts as a brake on the abuse of power by public officials. It facilitates the exposure of errors in the governance and administration of justice in this country.”

Strasbourg jurisprudence also attaches considerable importance to freedom of expression: see The Observer and The Guardian -v- UnitedKingdom (1991) 14 EHRR 153, at paragraphs 59 and 60, which are set out in paragraphs 44 and 45 of the judgment of Simon Brown LJ. So important is freedom of expression that the necessity for any restriction on it must be “convincingly established”. As Laws J said in R -v- Advertising Standards Authority Ltd, ex parte Vernons Organisation Ltd [1992] 1 WLR 1289 at 1293, an interim injunction (which constitutes prior restraint on publication) is not prohibited by the Convention. As the European Court of Human Rights said in Market Intern Verlag GmbH and Klaus Beerman -v- Germany, 20 November 1988,

“… even the publication of items which are true and describe real events may under certain circumstances be prohibited: the obligation to respect the privacy of others or the duty to respect the confidentiality of certain commercial information and examples.” (paragraph 35)

However, the question whether to grant such an injunction calls for the most careful scrutiny, especially if the defendant is the press and the news is perishable (The Observer and The Guardian -v- UK).

The key requirement of the Strasbourg jurisprudence is that the need for a prior restraint must be “convincingly established”. Strasbourg jurisprudence does not, however, go so far as to say that the applicant must show that he will succeed at trial in establishing the need for restraint. Accordingly, in my judgment, it would not be contrary to the Convention right under article 10 to grant an interim injunction where the applicant cannot show that he will succeed at trial. Of course, the court would have to give the matter careful scrutiny but a temporary restraint may in some circumstances be necessary in a democratic society so that the applicant can demonstrate at trial that he has an obligation of confidence which the defendant owes to him.

Simon Brown LJ has cited a passage from an article by Professor Colin Munro entitled Prior Restraint in the Media and Human Rights Law in the Scottish Journal, Juridical Review (2002) at paragraph 23. Professor Munro opines that a prior restraint must be “appropriate, proportionate and absolutely necessary.” This quotation refers to only some of the qualifications on freedom of expression recognised by article 10(2) of the Convention. Other qualifications include national security, the prevention of disorder or crime, the reputational rights of others and the prevention of disclosure of information received in confidence. Where, however, reliance was placed for these matters, it would have to be shown that the restriction was proportionate.

Prior to the commencement of the 1998 Act, the threshold test which the respondent would have had to have satisfied as against the appellants in an action for breach of confidence would have been that there was a serious issue to be tried. If the court was so satisfied, it would have gone on to consider the balance of convenience (or balance of justice as Sir John Donaldson MR referred to it in Francome -v- Mirror Group Newspapers Ltd [1984] 1 WLR 892). The threshold test of serious issue to be tried was established in American Cyanamid Co -v- Ethicon Ltd [1975] AC 396. Lord Diplock explained the reason for the threshold test being cast in these terms in this way:-

“My Lords, when an application for an interlocutory injunction to restrain a defendant from doing acts alleged to be in violation of the plaintiff’s legal right is made on contested facts, the decision whether or not to grant an interlocutory injunction has to be taken at a time when ex hypothesi the existence of the right or the violation of it, or both, is uncertain and will remain uncertain until final judgment is given in the action. It was to mitigate the risk of injustice to the plaintiff during the period before that uncertainty could be resolved that the practice arose of granting him relief by way of interlocutory injunction; but since the middle of the 19th century this has been made subject to his undertaking to pay damages to the defendant for any loss sustained by reason of the injunction if it should be held at the trial that the plaintiff had not been entitled to restrain the defendant from doing what he was threatening to do. The object of the interlocutory injunction is to protect the plaintiff against injury by violation of his right for which he could not be adequately compensated in damages recoverable in the action if the uncertainty were resolved in his favour at the trial; but the plaintiff’s need for such protection must be weighed against the corresponding need of the defendant to be protected against injury resulting from his having been prevented from exercising his own legal rights for which he could not be adequately compensated under the plaintiff’s undertaking in damages if the uncertainty were resolved in the defendant’s favour at the trial. The court must weigh one need against another and determine where ‘the balance of convenience’ lies.

In those cases where the legal rights of the parties depend on facts that are in dispute between them, the evidence available to the court at the hearing of the application for an interlocutory injunction is incomplete. It is given on affidavit and has not been tested by oral cross-examination. The purpose sought to be achieved by giving to the court discretion to grant such injunctions would be stultified if the discretion were clogged by a technical rule forbidding its exercise if on that incomplete untested evidence the court evaluated the chances of the plaintiff’s ultimate success in the action at 50 per cent or less, but permitting its exercise if the court evaluated his chances at more than 50 per cent.

The notion that it is incumbent on the court to undertake what is in effect a preliminary trial of the action on evidential material different from that on which the actual trial will be conducted, is, I think, of comparatively recent origin, though it can be supported by references in earlier cases to the need to show ‘a probability that the plaintiff is entitled to relief’ (Preston -v- Luck ((1884) 27 Ch D 497 at 506) per Cotton LJ) or ‘a strong prima facie case that the right which he seeks to protect in fact exists’ (Smith -v- Grigg Ltd ([1924] 1 KB at 659) per Atkin LJ). These are to be contrasted with expressions in other cases indicating a much less onerous criterion, such as the need to show that there is ‘certainly a case to be tried’ (Jones -v- Pacaya Rubber and Produce Co Ltd ([1911] 1 KB 445 at 457) per Buckley LJ) which corresponds more closely with what judges generally treated as sufficient to justify their considering the balance of convenience on applications for interlocutory injunctions, at any rate up to the time when I became a member of your Lordships’ House.

An attempt had been made to reconcile these apparently differing approaches to the exercise of the discretion by holding that the need to show a probability or a strong prima facie case applied only to the establishment by the plaintiff of his right, and that the lesser burden of showing an arguable case to be tried applied to the alleged violation of that right by the defendant (Donmar Productions Ltd -v- Bart ([1967] 2 All ER 338 at 339, [1967] 1 WLR 740 at 742) per Ungoed Thomas J, Harman Pictures NV -v- Osborne ([1967] 2 All ER 324 at 336, [1967] 1 WLR 723 at 738) per Goff J). The suggested distinction between what the plaintiff must establish as respects his right and what he must show as respects its violation did not long survive. It was rejected by the Court of Appeal in Hubbard -v- Vosper—a case in which the plaintiff’s entitlement to copyright was undisputed but an injunction was refused despite the apparent weakness of the suggested defence. The court, however, expressly deprecated any attempt to fetter the discretion of the court by laying down any rules which would have the effect of limiting the flexibility of the remedy as a means of achieving the objects that I have indicated above. Nevertheless this authority was treated by Graham J and the Court of Appeal in the instant appeal as leaving intact the supposed rule that the court is not entitled to take any account of the balance of convenience unless it has first been satisfied that if the case went to trial on no other evidence than is before the court at the hearing of the application the plaintiff would be entitled to judgment for a permanent injunction in the same terms as the interlocutory injunction sought.

Your Lordships should in my view take this opportunity of declaring that there is no such rule. The use of such expressions as ‘a probability’, ‘a prima facie case’, or ‘a strong prima facie case’ in the context of the exercise of a discretionary power to grant an interlocutory injunction leads to confusion as to the object sought to be achieved by this form of temporary relief. The court no doubt must be satisfied that the claim is not frivolous or vexatious; in other words, that there is a serious question to be tried.

It is no part of the court’s function at this stage of the litigation to try to resolve conflicts of evidence on affidavit as to facts on which the claims of either party may ultimately depend nor to decide difficult questions of law which call for detailed argument and mature considerations. These are matters to be dealt with at the trial. One of the reasons for the introduction of the practice of requiring an undertaking as to damages on the grant of an interlocutory injunction was that ‘it aided the court in doing that which was its great object, viz abstaining from expressing any opinion upon the merits of the case until the hearing’ (Wakefield -v- Duke of Buccleuch ((1865) 12 LT 628 at 629)). So unless the material available to the court at the hearing of the application for an interlocutory injunction fails to disclose that the plaintiff has any real prospect of succeeding in his claim for a permanent injunction at the trial, the court should go on to consider whether the balance of convenience lies in favour of granting or refusing the interlocutory relief that is sought.”

It is thus clear from the American Cyanamid test that on an application for an interim injunction the court should not normally seek to form a view on the merits of the applicant’s case at trial.

As Keene LJ said in Douglas -v- Hello! Ltd [2001] QB 9679 at paragraph 150, section 12(3) requires the court to focus on the situation at trial. This follows from the requirement to be satisfied that the applicant is likely to establish at trial that publication should not be allowed. This is a different test from the American Cyanamid test. The exercise which the court is required to perform is an altogether different exercise from that which it would have performed under the American Cyanamid test.

The question then is what is the meaning of the word “likely” in section 12(3) of the 1998 Act? The parties are agreed that the court must evaluate the meaning of “likely” in its context. As Lord Nicholls said in Re H [1996] AC 563:

“In everyday usage one meaning of the word ‘likely’, perhaps its primary meaning, is probable, in the sense of more likely than not. This is not its only meaning. If I am going walking on Kinder Scout and ask whether it is likely to rain, I am using ‘likely’ in a different sense. I am inquiring whether there is a real risk of rain, a risk that ought not to be ignored.”

The context of section 12(3) is that of an evaluation at an interim stage of the applicants’ prospects of success at trial. I agree with Simon Brown LJ that the meaning of the word “likely” in this context has not yet been clearly or authoritatively decided by this court. In Douglas -v- Hello! Ltd, the majority view was that expressed by Keene LJ at paragraphs 150 to 153 of his judgment. In my judgment, Keene LJ does not in these paragraphs decide the meaning of the word “likely”. He discusses the meaning of that word without deciding the level or degree of likelihood required. That question was not an issue in the case.

Likewise in A -v- B&C [2002] 3 WLR 542, the question of the degree of likelihood required by section 12(3) was not an issue. However, this court gave the following guidance on this question:-

“iii)

As to the word “likely” in section 12(3) useful guidance is provided by Sir Andrew Morritt VC in Imutran Ltd -v- Uncaged Campaigns Ltd [2002] FSR 20. He said of section 12:

‘17. Counsel for the defendants submitted that the requirement of likelihood imposed a higher standard than that formulated in AmericanCyanamid. I did not understand this to be disputed by counsel for Imutran. He submitted that whatever the standard was his case satisfied it. Theoretically and as a matter of language likelihood is slightly higher in the scale of probability than a real prospect of success. But the difference between the two is so small that I cannot believe that there will be many (if any) cases which would have succeeded under the AmericanCyanamid test but will now fail because of the terms of s.12(3). Accordingly I propose to apply the test of likelihood without any further consideration of how much more probable that now has to be’

There is no conflict between section 12(3) and the Convention. (See Douglas -v- Hello! Ltd [2001] QB 967; Keene LJ paragraph 150).”

Accordingly, on this appeal the court must decide the degree of likelihood required by section 12(3) of the 1998 Act. I agree with Simon Brown LJ that no assistance on this point is to be derived from the passages from Hansard with which the court was provided in response to its request for this material. Counsel were right to take the view that those passages did not meet the requirements for admissibility of such material laid down in Pepper -v- Hart [1993] AC 593.

As the parties have agreed, the meaning of the word “likely” in a statute or a rule or a document takes its colour from its context (see Champion -v- Gwent Chief Constable [1990] 1 WLR 1, HL). Thus in procedural rules relating to pre-action disclosure, “likely” has been held to mean “may well” rather than “more likely than not” (Black -v- Sumitomo Corporation [2002] 1 WLR 1562).

In the context of section 12(3), the court is deciding the prospects of a party at a time when the rights of the parties and the fact of their violation may be uncertain. This is one of the points made by Lord Diplock in the American Cyanamid case. Parliament has not enacted that the court should make this assessment on the pleadings or on the material then before it. Those would be words of limitation which are not there (no doubt for good reason) and cannot, in my judgment, be read in. Thus section 12(3) does not simply reverse the rule laid down by the House of Lords in the American Cyanamid case and reinstate that applied by the Court of Appeal. As the passage I have cited shows, the Court of Appeal in that case took the view that it had to be satisfied that, if the case went to trial on no other evidence than that before the court at the hearing of the application, the applicant would be entitled to judgment for a permanent injunction in the same terms as the interim injunction sought. Section 12(3) has not reinstated that test.

It is necessary to analyse the word “likely” a little further. The word “likely” denotes a relationship between a conclusion and the evidence. It is addressed to the degree of certainty or confirmation with which that conclusion can be stated on that evidence. There is a separate question, which has troubled the law since the time of Bacon, if not before, as to the standard of proof with which that degree of certainty must be established. The law has indeed devised different scales of proof. At one end of the spectrum the tribunal must be sure beyond all reasonable doubt that a given set of facts has been proved on the evidence. At the other end of the scale, the court may be directed to accept a certain factual state as true unless it is satisfied that that state of facts did not exist. In the context of interim injunctions, leaving aside section 12(3), the rule which the court has devised is that the court has to be satisfied that there is a serious issue to be tried. The reason why that rule was devised was to avoid the need to make an assessment of the prospects of success in a situation where there would be likely to be a significant degree of uncertainty.

In the context of section 12(3) the standard of proof is the normal civil standard, namely, on a balance of probabilities. Lord Nicholls explained the meaning of this expression in Re H:

“The balance of probability standard means that a court is satisfied an event occurred if the court considers that, on the evidence, the occurrence of the event was more likely than not. When assessing the probabilities the court will have in mind as a factor, to whatever extent is appropriate in the particular case, that the more serious the allegation the less likely it is that the event occurred and, hence, the stronger should be the evidence before the court concludes that the allegation is established on the balance of probability.”

It is still necessary to determine what “the event” is in section 12(3) for the purpose of ascertaining whether the court can be satisfied as to it on a balance of probabilities.

On this question I, like Simon Brown LJ, have derived considerable assistance from the judgment of Hoffmann J in Re Harris Simons Construction Ltd [1989] 1 WLR 368. The facts are set out in paragraph 38 of the judgment of Simon Brown LJ and I need not repeat them. They show that the court was concerned as to the likelihood of an event to occur in the future. Simon Brown LJ has also set out the relevant part of the judgment of Hoffmann J (see paragraph 39, above). In that passage Hoffmann J makes a number of points which are pertinent in this case too. He held that the word “likely” in the provision of the Insolvency Act 1986 with which he was concerned depended on the context, and for that context Hoffmann J found support in the CorkReport which contained the recommendation for administration orders subsequently implemented by what is now in the Insolvency Act 1986. The relevant context for the meaning of the word “likely” can also be found in the statute or document in question itself and in my judgment, as I explain below, the context can in this case be found within the four corners of section 12 itself. Hoffmann J also analysed closely the textual indications in section 8. He drew attention to the fact that the court was directed to “consider” whether the relevant event was likely and that this requirement contrasted with the requirement in a further paragraph of the same sub-section, that the court should be “satisfied” as to the company’s present or future insolvency. That same contrast does not exist in section 12. Section 12(1) uses the words “is considering” but in subsection (1) those words mean “ is deliberating” not “has formed a view”. In Re H, the House of Lords held that the word “likely” did not mean more probable than not, even though the event in question was one of which the court had to be “satisfied”. Of that expression, Lord Nicholls said:

“The expression ‘if the court is satisfied’… envisages that the court must be judicially satisfied on proper material. There is also inherent in the expression an indication of the need for the subject matter to be affirmatively proved. If the court is left in a state of indecision the matter has not been established to the level, or standard, needed for the court to be ‘satisfied’.”

The further textual indications to which Hoffmann J referred was the fact that the provision in question was merely a threshold criterion and the court would still have to exercise its discretion to make an administration order. He also drew attention to the fact that Parliament would not have intended the court to embark on the calculation of cumulative possibilities which was relevant because the section refers to the likelihood of a number of events which can exist at the same time.

Finally, Hoffmann J made the point that even though “real prospect” may lack precision compared with a chance of 0.5 on the scale of probability, it was still a meaningful test.

It is useful to contrast the decision of Hoffmann J in Re Harris Simons Construction Ltd with the subsequent decision of Vinelott J in Re Primlaks (UK) Ltd [1989] 1 BCLC 734. Vinelott J had to decide as between the decision of Peter Gibson J, from which Hoffmann J had differed, and that of Hoffmann J. At 740 to 742, Vinelott J held:

“I prefer the approach of Hoffmann J. I would not improve on his, to my mind, compelling reasoning if I were to attempt to paraphrase it. I think, however, that in deference to the views expressed by Peter Gibson and Harman JJ and the able argument that has been addressed to me by Mr Higham, counsel on behalf of BCCI, I should add a few words.

The case of BCCI in essence is that the word ‘likely’ when ordinarily used means ‘probable’ or ‘more likely than not’. That is the sense in which the word is used in para (a) of s 8(1): ‘is or is likely to become unable to pay its debts’ and it must be given the same reading in para (b): ‘would be likely to achieve one or more of the purposes mentioned below’.

The fallacy in that argument lies in the assumption that ‘likely’ has a clearly definable, ordinary meaning. Like many words in everyday use, it carries with it, as it were, a penumbra of different shades of meaning:

‘… the wide spectrum that it covers makes it the more necessary, if precise conclusions are to be founded upon its use, to place it in a setting.’

(See per Lord Wilberforce in Gartside -v- IRC [1968] 1 All ER 121 at 134, [1968] AC 553 at 617 where, as Lord Wilberforce pointed out, the word ‘interest’ was used in different senses in different parts of the estate duty legislation.) The word ‘likely’ is most commonly used to indicate a degree of probability above 0.5 on the scale of 0 to 1, though frequently as Hoffmann J points out, where it is necessary to indicate the degree of probability attached by the speaker, it will be qualified by the addition of ‘quite’ or ‘very’. However, it is not always used in a sense that indicates a belief on behalf of the speaker that a state of affairs is probably true or would probably come about. One of the senses given to the word in the Oxford English Dictionary is ‘having an appearance of truth’, the sense in which an account of an event or an explanation may be said to be likely to be true. Even when used to describe a possible future event, it is not always used to convey the speaker’s belief that it will probably happen. So, for example, a sanitary inspector might refer to a foul and ill-ventilated drain as a ‘likely source of infection’ without wishing to commit himself to the view that if uncorrected infection will probably result; he may have in mind only that the state of the drain is such that if its condition is uncorrected there will be a real risk of infection.

Paragraph (a) of s 8(1) sets out a condition that must be met before the court can enter into an inquiry as to whether an administration order would serve any useful purpose. The court must be satisfied that the company is or is likely to become unable to pay its debts. Clearly in this context, the test prescribed must be whether a company currently able to pay its debts as they fall due will probably be unable to pay them in the future. It would be unjust to a company’s creditors to impose on them the regime of an administration order so as to improve and, perhaps, expand the company’s business if the probability is that the company will be able to pay its debts as they fall due.

Paragraph (b), although in form it sets out a further condition that must be met before the court comes to exercise its discretion, in substance has a different function. The difference in the function (as Hoffmann J observes) is reflected in the use of the word ‘considers’. If the test in para (a) is satisfied, the court has to consider whether an administration order would be likely to achieve one of the stated purposes and if it would whether, in the exercise of its discretion, it should appoint an administrator. In that context to give the word ‘likely’ the force of ‘more likely than not’ would, in my judgment, stultify the Act and achieve no useful purpose. The court has to weigh the prospect that one or more of the stated purposes would be achieved against the interests of creditors and others who might be adversely affected by it. The question must always be, if there is a real prospect that one or more of the stated purposes would be achieved, is that prospect sufficiently likely in the light of all the other circumstances of the case to justify making the order?”

The example which Vinelott J gives of the sanitary inspector is a telling example and is comparable with that of a hike on Kinder Scout given by Lord Nicholls in Re H above and with the facts in Venables & Thompson -v- News Group International [2001] EMLR 255, cited by Simon Brown LJ. Vinelott J also demonstrates that when the discretion is exercised whether to make an administration order on the basis that the threshold of a real prospect that one or more of the stated purposes would be achieved, the question which the court has to ask is whether that prospect is sufficiently likely in all the circumstances to justify the making of the order. In other words, when the court comes to exercise the discretion, weight is to be given to the degree of likelihood of the event which is the subject of the threshold criterion. That point is instructive here too since on the basis that section 12(3) establishes merely a threshold criterion at the level of “real prospect of success”, nonetheless the court in the exercise of its discretion whether to grant an injunction should take into account that degree of likelihood as to which it is satisfied as to that prospect.

I have already made the point that “likely” denotes a relationship between a conclusion and the evidence on which that conclusion is based. In other words, in order to determine whether something is “likely”, the evidence in relation to which the test is to be met must be specified.

For instance, if one says “X is a smoker and his life is likely to be shortened as a result” one has to ask whether this statement is being made in relation to a particular individual and his medical history or whether it is being made on the basis of statistical information. But section 12(3) specifies no such test. The exact basis against which a judge must be satisfied that there is likelihood as to the claimant’s success at trial must be implicit. He must make that judgment on the basis of evidence of which the court has been informed at the interim stage and on the assumption that that is the totality of the evidence. That is often an unrealistic assumption. There are very likely to be further relevant documents produced on disclosure. The judge cannot under section 12(3) speculate as to these matters: as Lord Nicholls said in Re H., to “be satisfied” the judge must be judicially satisfied on proper material. Such an assessment at the interim stage is likely, however, to be inaccurate. In those circumstances, in my judgment, Parliament cannot have intended that the judge should be satisfied that the applicant would be more likely than not to succeed in establishing that publication should be restrained.

Moreover, I have mentioned thus far only the possibility that the information available at the interim stage is incomplete. There is also the active possibility that evidence at the interim stage is disputed. It is impossible for the judge to take a view at the interim stage about whether evidence will be accepted at trial unless there is a clear conflict with an irrefragible piece of evidence, such as an authentic, contemporaneous document. There is no way in which the judge can make a judicial evaluation of evidence in the absence of those exceptional circumstances. Furthermore, if “likely” in section 12(3) means “more probable than not”, and the issue requires the resolution of a difficult point of law, then again, the test laid down in section 12(3) is inapt. A point of law which cannot be decided at the interim stage is either right or wrong. The judge may have an uninformed view about which way the point would be decided but he cannot make any judicial evaluation of it until it has been fully argued. A view in advance of argument of how a point should be decided may vary from one judge to another. I do not consider that Parliament would have laid down a test of prospects of success being more likely than not in those circumstances.

Accordingly, in my judgment, section 12(3) requires the court to envisage the outcome of trial of the action. The court must consider the merits of the legal and the factual case and decide, by reference to the merits of those cases, whether the applicant has a real prospect of success in establishing that publication should be restrained at trial. Accordingly, the judge will have to be satisfied that there is no obvious reason why the claim should not succeed. I do not consider that section 12(3) requires that the applicant must show that he is more likely than not to succeed in this way at trial. It is understandable that as this is a new exercise on any basis Parliament did not give it a more familiar legal label.

It follows that the court can be satisfied of the likelihood in this sense of both sides winning. As I have said, in the exercise of discretion no doubt greater weight should be given to the applicant’s case if the degree of likelihood can be said to be higher than “real prospect of success” but it is not a bar to the grant of an injunction that only a real prospect of success is shown.

In A -v- B&C, the Court of Appeal held that the judgment of Sir Andrew Morritt VC in Imutran -v- Uncaged Campaigns Ltd [2002] FSR 20 provided useful guidance. That passage has already been set out in the judgment of Simon Brown LJ. In formulating his conclusion, Sir Andrew Morritt VC drew an analogy with the American Cyanamid test. As I have explained, the test established by that case of “serious issue to be tried” does not normally require or entitle the court to make any assessment of the merits. In practice, however, there may be a fine line between a judge reaching the view that a party has no real prospect of success on the merits and saying that there is no seriously arguable issue. However that may be, I agree with Simon Brown LJ that it is better to put the American Cyanamid test firmly on one side when applying section 12(3) of the 1998 Act.

I do not consider that the example of counsel’s advice on the prospects of success to be of assistance. When counsel advises a client “you are likely to succeed at trial”, it is implicit that counsel is giving advice on the basis of material then before counsel. If at trial that basis is shown to be incorrect, and counsel was not at fault for not foreseeing this, no blame could be attached to counsel. However that may be, it is clear from the context in which the question is being asked that the client wishes to know whether it is more probable than not that he will succeed.

There are further points which throw light on the meaning of “likely” in section 12(3) and which can be gleaned from the context of section 12 as a whole. As Simon Brown LJ points out, if the degree of certainty required by “likely” in subsection (3) is “more likely than not”, there are serious adverse practical implications for the administration of justice. It would not be possible for the court to grant an injunction pending an appeal, or over an adjournment of an interim application, or pending speedy trial if (as may very well happen) it could not satisfy itself that the applicant was more likely than not to succeed at trial.

In addition, cases of this kind will often involve more than one issue of fact or law. In Re Harris Simons Construction Ltd, Hoffmann J referred to the permutations that could result from the requirement that the court considers whether one or more of the statutory objectives would be likely to be fulfilled. He held that a further reason supporting his construction of the word “likely” was that Parliament could hardly have expected judges to make complex calculations as to the percentage chance of any permutation of statutory objectives being achieved on the facts. The same point applies in relation to section 12(3). In some cases, there is only one issue of fact or law in dispute on which the applicant has to succeed in order to win the application. But it is more usual for there to be several such issues, whether of fact or law. It depends on the nature of the case whether the applicant has to succeed on only one of those issues or whether the applicant has to succeed on two or more, or all, of such issues. In order to form a view as to whether the applicant is more likely than not to succeed in obtaining relief in a case where the applicant has to establish several issues, sometimes connected and sometimes free-standing, the court would have to establish the percentage chance of those issues succeeding in combination. The court cannot take simply the chances of success on each issue in isolation since there has to be factored in the additional risk consequent on the need to succeed on other issues as well. By analogy with the point made by Hoffmann J in Re Harris Simons Construction Ltd, it seems to me unlikely that Parliament intended the court to undertake some potentially complicated calculus of variations of this nature.

In my judgment, in addition to context, there are important textual indicators in section 12(3) that “likely” does not mean “more probable than not”. No adverse inference flows from the use of the words “is satisfied” in section 12(3), as opposed to the provision before Hoffmann J in the Harris Simons Construction Ltd case, because there is no contrast in this case between “is satisfied” and “considers” (see above). Furthermore, if the meaning is “more likely than not”, it is difficult to see why the court “must have regard to the importance of the Convention right to freedom of expression” in the opening words of subsection (4) (set out above) since it is bound to have taken that into account in deciding whether or not the applicant will succeed at trial.

Moreover, as I have explained, the provisions of section 12(3) were clearly intended to do more than simply reinstate the test rejected by the House of Lords in the American Cyanamid case. Moreover, if that had been intended, words of limitation would have been needed and the words “is likely to” would have been unnecessary (compare the ante-penultimate paragraph of the citation, above, from the speech of Lord Diplock in the American Cyanamid case).

I do not myself think that this is a case which turns on section 3 of the Human Rights Act 1998. I refer here to the arguments set out in paragraphs 54 to 60 of the judgment of Simon Brown LJ and paragraphs 72 to 82 of the judgment of Sedley LJ. I accept that if section 12(3) were, on its true interpretation without reference to section 3, to involve a breach of another person’s Convention right at the interim stage, it would have to be interpreted so as to avoid that conflict. However, in that event, there seems to me to be no reason why it could not be so construed by simply adding a proviso to deal with that situation (for example, by adding the words “except where to do so would be incompatible with a Convention right of the applicant”) rather than imposing a qualification of general application.

As I have explained above, the Strasbourg jurisprudence makes it clear that exercise of the Convention right of freedom of expression should not be restrained unless the need for the restriction is “convincingly” established. This was echoed by Lord Nicholls in Reynolds -v- Times Newspapers Ltd [2001] 2 AC 127 at 200 when he said that the need for restriction must be “convincingly established by a compelling countervailing consideration and the means employed must be proportionate to the end sought to be achieved.” Accordingly, even though the judge only has to be satisfied as to a real prospect of success on the merits under section 12(3), he must satisfy himself as to compliance with this jurisprudence before exercising his discretion to grant an interim injunction. This seems to me to follow from the opening words (set out above) of section 12(4). I would not wish myself to attempt to reformulate this jurisprudence, as Simon Brown LJ has (as I read it) done by saying that “Often the court will not think it right to exercise that discretion in favour of prior restraint unless it is indeed satisfied that the claim will more probably than not succeed at trial” (judgment, paragraph 61, fifth sentence). I do not myself consider that the Strasbourg jurisprudence can be so neatly encapsulated. In my judgment, it permits, for example, the court considering an interim injunction to take into account factors such as the length of any delay before judgment at trial. Under our domestic jurisprudence, the judge will, in any event, have to consider the adequacy of damages and the balance of convenience. This is established by the American Cyanamid case in passages which I have not cited. On the other hand, as I have said above, weight is to be given to the degree of likelihood of the applicant’s prospects of success.

I now turn to the question of how the judge applied section 12(3) in the present case. In my judgment, the judge clearly considered that section 12(3) of the 1998 Act required him to ask whether the claimant had a real prospect of success and in this he applied the guideline given in A -v- B&C based on the Imutran case. He specifically said that he was not taking into account contentions which would be made at trial, but which were not in evidence before him and he disclaimed any intention of forming a view on disputed questions of fact. Accordingly, in my judgment, he could have been doing no more than forming a view as to the real prospect of success of the claimant at that stage. He did not find that it was more likely than not that the claimant would succeed at trial. In addition, the way in which the judge dealt with these issues in his private judgment shows that he did indeed consider the merits.

The appellants do not contend that, if there is any truth in their suppositions, that the appropriate regulatory authorities would not take appropriate action in this case on the information which the first appellant says that she has. So far as prospective investors in the respondent are concerned, the respondent’s case is that its financial position is sufficiently disclosed in its published financial statements. The transactions in those financial statements have (we are told) been agreed upon by all the shareholders on full disclosure and thus as a matter of law (there being no question of ultra vires or prejudice to creditors) are binding on the respondent. That case has yet to be tested at trial. The appellants’ information is not perishable in the short term, and the judge suggested the possibility of a speedy trial. The fact that the regulatory authorities can be informed is sufficient to protect the public interest at this interim stage even if the respondent has a public image which were to need to be corrected. In all the circumstances, I do not consider that the judge’s decision can be said to have been perverse.

I am not, however, impressed by the argument of Mr Bartley Jones QC, for the respondent, that if the regulatory authority to whom disclosure has already been made could not lawfully make disclosure to the public then neither could the appellants. In my judgment, this would not follow.

In a confidential judgment, I give further reasons for my conclusion that the exercise by the judge of his discretion cannot be disturbed by this court.

Accordingly, I would dismiss this appeal.