Skip to Main Content
Alpha

Help us to improve this service by completing our feedback survey (opens in new tab).

Ajinomoto Sweeteners Europe SAS v ASDA Stores Ltd

[2010] EWCA Civ 609

Case No: A2/2009/1683
Neutral Citation Number: [2010] EWCA Civ 609
IN THE COURT OF APPEAL (CIVIL DIVISION)

ON APPEAL FROM THE HIGH COURT OF JUSTICE

QUEEN’S BENCH DIVISION

MR JUSTICE TUGENDHAT

[2009] EWCA 1717 (QB)

Royal Courts of Justice

Strand, London, WC2A 2LL

Date: 02/06/2010

Before :

LORD JUSTICE SEDLEY

LORD JUSTICE RIMER

and

SIR SCOTT BAKER

Between :

AJINOMOTO SWEETENERS EUROPE SAS

Appellant

- and -

ASDA STORES LIMITED

Respondent

Mr James Price QC and Mr Matthew Nicklin (instructed by CMS Cameron McKenna Llp) for the Appellant

Mr Andrew Caldecott QC and Mr Manuel Barca (instructed by Kempner Robinson Llp) for the Respondent

Hearing dates: 11 and 12 May 2010

Judgment

Lord Justice Sedley :

1.

In defamation cases, both civil and criminal, there has for centuries been a rule that the question libel or no libel is to be answered in respect of a single meaning. This is unproblematical where there is only one thing that the words can sensibly mean, but it can be highly problematical where reasonable people might put more than one construction on the words read in their proper context. By the early 17th century, in order to discourage suits and prosecutions (the latter almost all private and open to abuse – hence the concomitant development of the action for malicious prosecution), the courts had decided that such words were always to be construed in the most inoffensive sense reasonably possible – in mitiori sensu: see Mitchell v Brown 3 Inst. 167. This practice – it can hardly be dignified as a doctrine – was superseded after the Restoration by a rule that the impugned words were to be understood “according to the general and natural meaning, and agreeable to the common understanding of all men” (Townshend v Hughes (1676) 2 Mod. 159); or, as more vividly put by Levinz J in Naben v Miecock (1683) Skin. 183, “in their natural, genuine and usual sense and common understanding, and not according to the witty construction of lawyers, but according to the apprehension of the bystanders”.

2.

This, however, was another version of the single meaning rule. It still declined to recognise that two or more bystanders can perfectly reasonably make different sense of the same words. It simply shifted the preferred meaning to the middle ground. By virtue of Fox’s Libel Act of 1792, the function of deciding meaning passed to the jury in both civil and criminal libel cases, but still subject to two things: that the judge must determine the outer limits of meaning, and that, within those limits, the jury’s verdict of libel or no libel must relate to what they find to have been the meaning, not a meaning, of the words.

3.

The fiction that there is a single reasonable reader, so that words, duly taken in context, have only one meaning, has remained embedded in the law of defamation. In his classic judgment in Slim v Daily Telegraph [1968] 2 QB 157, endorsed in Charleston v News Group Newspapers [1995] 2 AC 65, Diplock LJ said:

“… the argument between lawyers as to the meaning of words starts with the unexpressed major premise that any particular combination of words has one meaning which is not necessarily the same as that intended by him who published them or understood by any of those who read them but is capable of ascertainment as being the 'right' meaning by the adjudicator to whom the law confides the responsibility of determining it.

Where, as in the present case, words are published to the millions of readers of a popular newspaper, the chances are that if the words are reasonably capable of being understood as bearing more than one meaning, some readers will have understood them as bearing one of those meanings and some will have understood them as bearing others of those meanings. But none of this matters. What does matter is what the adjudicator at the trial thinks is the one and only meaning that the readers as reasonable men should have collectively understood the words to bear. That is 'the natural and ordinary meaning' of words in an action for libel.

Juries, in theory, must be unanimous upon every issue on which they have to adjudicate; and since the damages that they award must depend upon the defamatory meaning that they attribute to the words, they must all agree upon a single meaning as being the 'right' meaning. And so the unexpressed major premise, that any particular combination of words can bear but a single 'natural and ordinary meaning' which is 'right,' survived the transfer from judge to jury of the function of adjudicating upon the meaning of words in civil actions for libel."

4.

In the circumstances it is surprising – it might even be thought gratifying - that neither of the very experienced leading counsel appearing in the present appeal could recollect a case in which a judge had had difficulty in directing a jury in accordance with the single meaning rule, or in which a jury had evinced difficulty in applying it.

This case

5.

Aspartame is a sugar substitute which has attracted a measure of controversy but is a lawful food additive. The claimant is the European limb of one of the world’s major producers and suppliers of it.

6.

The defendant is a large supermarket chain, now a subsidiary of Walmart, which sells its own brand of health foods. Some of these carry a flower design, on one petal of which is written “No hidden nasties”, together with a legend typically reading “No artificial colours or flavours and no aspartame”.

7.

Three possible meanings were ascribed to this packaging by the claimant:

A: That aspartame is harmful or unhealthy.

B: That there is a risk that aspartame is harmful or unhealthy.

C: That aspartame is to be avoided.

The defendant averred that it meant

D: That these foods were for customers who found aspartame objectionable.

8.

In a meticulously reasoned judgment on the preliminary issue of meaning ([2009] EWHC 1717 (QB); [2010] 1 QB 204) Tugendhat J rejected meaning A outright. It accordingly leaves the stage. He found that a substantial number of consumers would derive meaning B from the packaging. He also found that a substantial number would derive meaning D from it. Meaning C, he found, added nothing to meaning B, and it too accordingly leaves the stage.

9.

The judge then applied to meanings B and D the single meaning rule. This he took to require that “The court should not select one bad meaning where other non-defamatory meanings are available”. His reasoning rested on article 10 of the European Convention on Human Rights and on the particular regard for it enjoined by s.12(4) of the Human Rights Act 1998. He accordingly took the single meaning to be D. On this finding, if it is to stand, the claimant loses.

10.

So far I have deliberately said nothing about the cause of action. If it had been libel, James Price QC, for the claimant, would have contended that the judge had not only reverted to the old ‘mitiori sensu’ criterion rather than arrive at the informed preference which the law now requires but, in founding his choice on article 10, had accorded free expression inappropriate priority over the property right contained in article 1 of the First Protocol. But his principal case rests on the fact that this was not a libel action at all but a claim for malicious falsehood, a tort in which, he contends, the single meaning rule does not operate.

11.

For the defendant, Andrew Caldecott QC accepts that no authority explicitly requires us to apply the single meaning rule to a malicious falsehood claim, but he contends, first, that no authority prevents it (which Mr Price for his part accepts) and, secondly, that there are solid jurisprudential reasons for maintaining parity of legal reasoning between the two classes of claim.

Does the single meaning rule apply to malicious falsehood?

The appellant’s case

12.

Mr Price’s argument starts from the intellectual high ground of Diplock LJ’s judgment in Slim. Even in libel the rule is an anomaly, explained but no longer excused by either history or practice. If we were starting from scratch, Mr Price suggests, nobody would try to justify such a rule, whether for judges sitting alone or for judges directing juries. The judge would, as he or she now does, eliminate unreasonable meanings as a matter of law and leave the rest to the jury. If they find that a significant segment of reasonable readers would think the words defamatory of the claimant, then it should be sufficient for a verdict of libel that a meaning of the words will have damaged the claimant’s reputation. The damages will then reflect the inferred extent of the libel. Exactly the same would apply, but with narrative reasons, to a judge sitting alone.

13.

Had Tugendhat J done that here, his finding that the packaging conveyed both meaning B and meaning D would have been as far as he was required to go. Insofar as it was then found at trial that meaning B would have adversely influenced customers – a finding which would give credit, so to speak, for the comparable reasonableness of meaning D – but no further, and subject to proof of malice and falsity, the claim would be made out.

14.

To the argument that it would be mistaken in principle to draw such a line between libel and malicious falsehood, Mr Price replies that these are torts of different kinds: libel is an injury to reputation and thus a personal tort; malicious falsehood is an injury to property and thus an economic tort. In at least one of the other economic torts, the tort of passing off, Lightman J in Clark v Associated Newspapers [1998] 1 WLR 1558 held that it was enough to establish that one of the possible reasonable meanings of the publication would mislead a substantial number of people; but he also held that a single meaning was required for the statutory tort of false attribution of authorship.

15.

In Vodafone v Orange [1997] FSR 34 Jacob J, invited by consent to find the single meaning of words sued on for malicious falsehood, said:

“As a comparative stranger to this branch of the law I find the “one meaning rule” strange, particularly for malicious falsehood. Without authority, I should have thought it would be enough to satisfy the criterion of falsity for the plaintiff to prove that the defendant made a statement which was false to a substantial number of people. That, for instance, is the position in passing off (a tort also concerned with false representations): for that tort it is enough to show that the representation fools some of the people, even if not most of them.

The reason for the libel rule in part relates to the entitlement of jury trial for libel (as Diplock LJ explained in Slim). Save in exceptional circumstances the right jury trial remains for libel and slander (see section 69(1) of the Supreme Court Act 1981) but there is no such right in relation to malicious falsehood. So it by no means follows that the historical reason for the rule in libel should apply to malicious falsehood. Another reason for the rule relates to the function of a jury in awarding damages for defamation: unless one has settled on a particular meaning one cannot judge the extent of the defamation. But in malicious falsehood damages are rather different: they are essentially compensatory for pecuniary loss as for most other torts. So again it does not seem necessarily to follow that the libel rule should apply to the tort. However, as I say, the parties were agreed that I should proceed on the basis that I am a notional jury identifying the single meaning of the words complained of.”

16.

To this Mr Price adds the laconic opinion of the Supreme Court of South Australia in Entienne v Festival City Broadcasters [2001] SASC 60, [41]:

“To insist upon an innocent interpretation where any reasonable person could, and many reasonable people would, understand a sinister meaning is to refuse reparation for wrong that has in fact been committed.”

17.

He also calls in aid what Lord Nicholls said in Charleston v News Group Newspapers [1995] 2 AC 65, 73-4:

“Newspaper readers will have seen only the banner headline and glanced at the picture. They will not have read the text of the accompanying article. In the minds of these readers, the reputation of the person who is the subject of the defamatory headline and picture will have suffered. He has been defamed to these readers. The newspaper could have no cause for complaint if it were held liable accordingly. It has chosen, for its own purposes, to produce a headline which is defamatory. It cannot be heard to say that the article must be read as a whole when it knows that not all readers will read the whole article.

To anyone unversed in the law of defamation that, I venture to think, would appear to be the common sense of the matter. Long ago, however, the law of defamation headed firmly in a different direction. The law adopts a single standard for determining whether a newspaper article is defamatory: the ordinary reader of that newspaper.”

If this would, but for historical accidents, be the commonsense approach to defamation, why abandon common sense, Mr Price asks, when the claim is for malicious falsehood?

18.

There is even so, as Mr Price acknowledges, a measure of practical as well as jurisprudential discomfort in having two different tests of meaning in what may be single lawsuit for libel and malicious falsehood, albeit one necessarily tried without a jury, since malicious falsehood carries no such right. The judge may, in the present state of the law, be required to adopt a single meaning for the one tort and a range of meanings for the other. But such complexity, Mr Price suggests, as well as being quite manageable, will be in good company, for there are already different burdens of proof (in malicious falsehood the claimant must prove falsity; in libel the defendant must prove truth), different requirements in relation to malice, and a need to prove different kinds of harm.

The respondent’s case

19.

To Mr Price’s persuasive argument Mr Caldecott has advanced a formidable response. The single meaning rule, he submits, while artificial, affords a balancing mechanism of sufficient significance that to abandon it would bring an unpredictable range of consequences in its wake. The balance it strikes is between the need to protect reputation and the use of factitious complaints to inhibit free expression, and between both of these and the entitlement of the public to be informed. It moves the law into the middle ground between the author’s intent, a test highly favourable to defendants, and multiple meanings, a test which puts all the cards in the claimant’s hand. Having ruled out untenable meanings, the court proceeds by fixing on the most tenable one and trying out the question whether it is libellous. This is practical and fair.

20.

Moreover, Mr Caldecott submits, the possible variety of meanings is factored into the single meaning rule itself. Neill LJ, whose expertise in this field is unrivalled, said in Hartt v Newspaper Publishing plc (26 October 1989, CA, unreported), at 15:

“The hypothetical reasonable reader … must be treated as being … someone who does not, and should not, select one bad meaning where other non-defamatory meanings are available.”

21.

Moreover, Lord Nicholls, giving the opinion of the Privy Council in Bonnick v Morris [2003] 1 AC 300, said at §21:

21. ….The “single meaning” rule adopted in the law of defamation is in one sense highly artificial, given the range of meanings the impugned words sometimes bear: see the familiar exposition by Diplock LJ in Slim v Daily Telegraph Ltd [1968] 2 QB 157, 171-172. The law attributes to the words only one meaning, although different readers are likely to read the words in different senses. In that respect the rule is artificial. Nevertheless, given the ambiguity of language, the rule does represent a fair and workable method for deciding whether the words under consideration should be treated as defamatory. To determine liability by reference to the meaning an ordinary reasonable reader would give the words is unexceptionable.

22.

Notwithstanding that the Privy Council went on to disapply the principle in relation to the Reynolds defence, Mr Caldecott can at least rely on this as an antidote to the bane of Lord Nicholls’ remarks in Charleston (where, indeed, Lord Nicholls had gone on to accept that, given the huge variety of ways in which mass circulation papers were read and interpreted, it was the very crudeness of the single meaning yardstick that, from the point of view of the law, justified it).

23.

He can also place some reliance on the report of the Faulks Committee on Defamation (1975), which cited the trenchant comments of Diplock LJ in Slim but, acknowledging that these were powerful criticisms, concluded nevertheless that the defects were “inherent in a system which prescribed the criterion as the interpretation of the ordinary reader.” The report (§102) went on:

“We hold strongly that this is the correct test since it most nearly approximates to the actual meaning or meanings which the actual readers of the publication as a class are likely to have placed upon the words complained of.”

If this makes sense in libel, it is submitted, it makes equal sense in malicious falsehood.

24.

Next, Mr Caldecott asks how, in a malicious falsehood action, a court is to determine the issue of dishonesty in relation to multiple meanings; and how, if there is not to be a wide chilling effect, are the authors of words to anticipate what they can say?

25.

Lastly, Mr Caldecott attacks Mr Price’s endeavour to segregate malicious falsehood as an economic tort. He draws attention to a series of reported cases in which malicious falsehood has been deployed in relation to non-commercial wrongs: theft from an employer (Joyce v Sengupta [1993] 1 WLR 337), the content of an election leaflet (Quinton v Peirce [2009] FSR 17), sexual allegations (Khodaparast v Shad [2000] 1 WLR 618) and privacy (Kaye v Robertson [1991] FSR 62); as well as in conjunction with a libel claim in order to circumvent an offer of amends (Tesco v Guardian Newspapers [2009] EMLR 5).

26.

He points out, too, that malicious falsehood lacks the element of strict liability which characterises the economic tort of passing off. Like libel when qualified privilege is challenged, it focuses on the publisher’s state of mind. It also, at least in a case like the present one which concerns the exercise by consumers of their freedom of choice, depends centrally on Asda’s article 10 right and the correlative rights of the public. If in addition the property right protected by article 1 of the First Protocol is engaged, it is engaged on both sides. Why, given that all these interests are in play, should a minority of buyers determine what the majority are to be allowed to read before choosing what they will buy? And if the damaging and innocuous meanings are equally balanced, why should the court not opt for the one that preserves free expression? After all, Mr Caldecott says with some force, where an innocent meaning is also available, the burden of proving malice in relation to the damaging meaning becomes very hard to discharge.

Conclusions

27.

The choice we are presented with is constrained by an immovable object, the single meaning rule in libel. Nothing we decide can have any impact on this: as Diplock LJ said in Slim, it has passed beyond redemption by the courts. What we have in the first instance to decide is whether, given its presence, it ought to be equally applied to the tort of malicious falsehood.

28.

In my judgment not all of Mr Price’s reasons for holding back bear scrutiny. In particular I do not find at all helpful his endeavour to segregate malicious falsehood as an economic tort (which is not a term of legal art) distinct, for that reason, from defamation. A great many fact situations are able to be litigated through more than one cause of action, and the choice of them is as often tactical as jurisprudential. I find much more realistic Mr Caldecott’s approximation of the two torts with which we are concerned on the ground that both concern the protection of reputation, albeit one protects the reputation of persons and the other the reputation of property, typically in the form of the goodwill of a business.

29.

Both also attract similar human rights considerations because both involve a value protected by ECHR article 10(1). The brief debate about the presence of a countervailing article 8 value in libel and its possible (but to my mind doubtful) replication in relation to malicious falsehood by the property right in article 1 of the First Protocol is far less relevant than the controls built into both torts by article 10(2).

30.

But when all this is said, the two are not so close as to be variants of a single tort, as libel and slander might be said to be. For historical reasons they have developed with different characteristics; they make different demands on the parties; and they offer redress for different things. This is the gap across which it is sought to make the single meaning rule jump.

31.

In my judgment the powerful reasons advanced by Mr Caldecott for doing this are outweighed by one dominant fact: the rule itself is anomalous, frequently otiose and, where not otiose, unjust.

32.

The anomaly is very nearly common ground. With the help of counsel’s scholarship we have seen how a pragmatic practice became elevated into a rule of law and has remained in place without any enduring rationale. It is frequently otiose, as counsel’s own experience testifies, because in the great majority of defamation cases the choice between libel and no libel, by the time the case goes to a verdict, is an either-or choice.

33.

But where it is capable of being applied, as it is in the present claim, the rule is productive of injustice. On the judge’s unchallenged findings, the meanings which reasonable consumers might put on Asda’s health-food packaging include both the damaging and the innocuous. Why should the law not move on to proof of malice in relation to the damaging meaning and (if malice is proved) the consequential damage without artificially pruning the facts so as to presume the very thing – a single meaning - that the judge has found not to be the case?

34.

I do not accept that doing this will make trials of malicious falsehood claims unwieldy or over-complex. This is not because these claims are always tried by a judge alone: the experience of common law judges is that juries are on the whole very good at assimilating and applying sometimes complicated directions. It is because it makes the trial of the issues fairer and more realistic. Instead of (as here) denying any remedy to a claimant whose business has been injured in the eyes of some consumers on the illogical ground that it has not been injured in the eyes of others, or alternatively (and Mr Caldecott’s case necessarily involves this) giving such a claimant a clear run to judgment when in the eyes of many customers the words have done it no harm, trial of plural meanings permits the damaging effect of the words to be put in perspective and both malice and (if it comes to it) damage to be more realistically gauged.

35.

For these reasons I would hold that the single meaning rule is not to be imported into the tort of malicious falsehood.

36.

This being so, the second issue – the proper content of the single meaning rule – does not arise. I admit that this a relief. Mr Price’s fallback position, crudely put, might well have involved replacing both the original ‘mitiori sensu’ rule and the more recent ‘medii sensu’ rule with a ‘pejori sensu’ rule that would have been equally productive of unfairness. But much more would have come into issue, as in the event it does not.

37.

For the reasons I have given I would allow this appeal and invite counsel’s written submissions as to the appropriate consequential order.

Lord Justice Rimer:

38.

The common law has not yet determined whether the single meaning rule in defamation has any place in the economic tort of malicious falsehood. The issue before Tugendhat J and now before us is whether any dispute as to meaning of the statement complained of in a malicious falsehood claim should be resolved by the application of that rule; or whether the court should recognise that a statement can reasonably mean different things to different readers and decide the case accordingly.

39.

Tugendhat J held that a substantial number of consumers would understand the statements on the defendant’s packaging as conveying meaning B, one that is damaging to the claimant; and also that a substantial number would understand them as conveying meaning D, an innocuous one. Having also held that the single meaning rule should apply as much in malicious falsehood as in defamation, he decided that meaning D was that single meaning, a conclusion requiring the removal of meaning B from further consideration and spelling the end of the claim.

40.

That conclusion is said by the claimant to be unprincipled and unjust. The single meaning rule in defamation is the product of an accident of history resulting in a fiction that assumes that the reasonable man will understand a particular statement in only one way - its supposed single natural and ordinary meaning. Like most legal fictions, it is artificial and has something of the absurd about it. Whatever merit it may have for defamation law, there is, says the claimant, no sound basis for importing it into the different tort of malicious falsehood.

41.

The potential for injustice in the present case flows from the fact that, before discarding it as legally irrelevant, the judge made the finding he did as to meaning B. If the case were allowed to go to trial and the claimant were able to prove that such meaning was false, uttered with malice and calculated to damage it, why should it not be entitled to damages for the injury which the falsehood will have caused it? More importantly – and this is the primary remedy the claimant wants – why, if it can prove its case, should it not be entitled to have the defendant restrained by injunction from doing that which it wants to do, namely (presumably for its own commercial benefit) to continue to publish a falsehood that will continue to damage the claimant in the eyes of a substantial body of consumers? The result, however, of the application by the judge of the single meaning rule is that that body of consumers is removed from the court’s radar. The court instead satisfies itself with the fiction, contrary to its own finding, that the entire consuming public will interpret the defendant’s packaging as bearing a single innocuous meaning.

42.

Those considerations impress me that there ought to be no scope for the application of the single meaning rule in the tort of malicious falsehood. In his responsive submissions, however, Mr Caldecott submitted that there are good policy reasons for doing so, and Sedley LJ has summarised them. Cogent though Mr Caldecott’s submissions were, they did not persuade me. At their heart lies the assertion that the single meaning rule in defamation achieves a fair balance between the claimant’s right to protection against defamatory publications and the defendant’s right to freedom of speech. That was how the judge rationalised the rule as it applies to defamation, describing it as a control mechanism. He recognised that malicious falsehood claims raise different questions but considered that a balance must equally be struck in such claims between the defendant’s right to freedom of expression and the claimant’s property rights sought to be protected. In his view, the single meaning rule would achieve that balance.

43.

If the single meaning rule does achieve a fair balance in defamation law between the parties’ competing interests, that would appear to be the result of luck rather than judgment; and how the measure of such claimed fairness might be assessed may anyway be questionable. The application of the rule can also be said to carry with it the potential for swinging the balance unfairly against one party of the other, resulting in no compensation in cases when fairness might suggest that some should be due, or in over-compensation in others. No doubt it would keep the common law tidy if the single meaning rule were also applied in malicious falsehood claims, particularly because there will be cases in which a claim might be brought either in defamation or malicious falsehood. The common law has, however, never worried about tidiness. It has always been more concerned with meeting the justice of the particular case and developing itself accordingly. If the single meaning rule did not exist, I doubt if any modern court would invent it, either for defamation or any other tort. If the resolution of the present claim has to be forced into the artificial straitjacket of that rule, it will, I consider, carry with it the potential for the production of an injustice. The court ought only to risk the suffering by the claimant of such injustice if there are compelling policy reasons why the single meaning rule, itself an anomaly, ought to prevail in malicious falsehood claims as in defamation. I am not persuaded that there are any sufficient such reasons, not least because an essential ingredient of the claim is proof of malice on the part of the defendant. In a case in which the court has found that the statement complained of would be read by many in a damaging sense, the single meaning rule should not be allowed to bar a claim that the defendant has maliciously disparaged the claimant’s goods.

44.

For these reasons, as well as those given by Sedley LJ, I would hold that the single meaning rule does not apply in malicious falsehood claims. I too would allow the appeal.

Sir Scott Baker:

45.

I have read with admiration the masterly judgment of Sedley L.J. and agree with his analysis and reasoning. We heard compelling arguments from both sides but in the end I am persuaded that the single meaning rule has no place in the tort of malicious falsehood.

Ajinomoto Sweeteners Europe SAS v ASDA Stores Ltd

[2010] EWCA Civ 609

Download options

Download this judgment as a PDF (240.1 KB)

The original format of the judgment as handed down by the court, for printing and downloading.

Download this judgment as XML

The judgment in machine-readable LegalDocML format for developers, data scientists and researchers.