LORD BRIGGS:
Introduction
1.
This appeal from the Court of Appeal of Trinidad and Tobago raises the
well-known conundrum in the common law of defamation, namely the extent to
which (if at all) two or more different statements made upon different
occasions by the same defendant may be aggregated for the purpose of giving
rise to a cause of action in defamation, when none of those statements would do
so, viewed on its own.
2.
The conundrum arose from two decisions of the English Court of Appeal, Grappelli
v Derek Block (Holdings) Ltd [1981] 1 WLR 822 and Hayward v Thompson
[1982] QB 47, decided six months apart, in both of which Lord Denning MR gave
the leading judgment. The first (Grappelli) appeared to decide that in
no circumstances could this be done. The second (Hayward) appeared to
reach the contrary conclusion, upon the basis of a distinction drawn by Lord
Denning which has not found favour thereafter, with academics, practitioners,
or courts around the world which habitually apply the common law.
3.
Readers of this judgment who hoped that it might finally resolve this conundrum
will be disappointed. The Board does not consider that there is an entirely
satisfactory conceptual solution to the problem but, for the reasons which
follow, we have concluded that, on the particular facts of this case, the
appeal can be satisfactorily resolved without doing so. It is a feature of the
common law of defamation that neat conceptual solutions do not always provide
satisfactory answers to the endlessly varied fact-sets with which judges and
(in some jurisdictions) juries have to wrestle, for the purposes of achieving
an outcome which properly accords with justice and common sense.
The Facts
4.
At about 5.30 pm on 17 August 2007 there occurred in the residential
district of Wallerfield, in the town of Arima, Central Trinidad, a shooting incident
in which four male occupants of a car, and a woman in her home nearby, were all
shot dead by police officers: (“the Wallerfield shooting”).
5.
The second defendant in these proceedings, Trinidad Express Newspapers
Ltd (“Express Newspapers”), who publish the Daily and Sunday Express in
Trinidad, reported the Wallerfield incident at the time. Then, in December
2008, a little more than a year later, the second defendant published an
article headed “Fatal Blunder - Report reveals innocent Wallerfield five killed
in police mistake” and a further editorial article headed “A clear call for
justice” on the following day. They will be referred to as “articles A and B”.
The gist of them was that an unidentified person within the Trinidad Police
Service had put together a police assassination squad (also unidentified) with
a view to obtaining revenge against the supposed killer of a well-known female
drugs dealer with whom he was connected, and that the killer had been believed
to be one of those travelling in the car, although all those killed in the
Wallerfield shooting were in fact innocent citizens. Article B encouraged its
readers “to pay close attention to the proceedings of the ordered inquest when
it does come up in the Arima Magistrates Court, the better to ascertain whether
justice is being served”.
6.
Neither article A nor B identified any of the police officers alleged to
have been involved in planning or executing the shooting. They were, however,
as the court later held, thoroughly defamatory in nature, although they did not
defame any particular identified individuals.
7.
The inquest into the Wallerfield shooting was held in May and June 2009,
some six months after the publication of articles A and B. On 5 June, while the
inquest was ongoing, the second defendant published a further article (“article
C”) headed “Forensic expert testifies in Wallerfield shootings”, explaining
that the inquest was to ascertain the circumstances of the 17 August 2007
incident, and describing the testimony of the pathologist called to give expert
evidence. Article C identified by name nine police officers whose conduct was
said to have come into question following the shootings. They are the claimants
in these defamation proceedings.
8.
The inquest concluded on 30 June 2009, with findings that the officers
involved were not culpable in any way for the deaths of any of the five people
killed during the Wallerfield shooting. On 5 July 2009 the second defendant
published a further article (“article D”) headlined “Self-preservation important”.
The first line of the article was “kill or be killed”. It named the nine
officers again, reporting that they had all been cleared by the inquest, and
that the court had ruled that the actions of the officers were in no way
negligent against the four men killed and further ruled the death of the young
woman hit in the cross-fire was a misadventure. Article D reported that the
officers had been vindicated by the court, but it also reported statements by
relatives of the deceased and one of their representatives suggesting a sense
of continuing injustice on their part about the outcome of the inquest.
The Proceedings
9.
The ensuing defamation proceedings by the named police officers against
Express Newspapers and its general editor were issued in May 2010 and tried
before Seepersad J, without a jury, in March 2013. In his reserved judgment
dated 24 May 2013 dismissing the claim, he held that references to “police
officers” in articles A and B could not have been understood at the time to
refer to the claimants, although they were plainly defamatory in nature, and
that articles C and D could not be relied upon to identify the claimants as the
subjects of the defamatory content of articles A and B. The judge had also,
during the trial, ruled against the admission of evidence that the claimants
were in fact identified as the subject of articles A and B at the time of their
publication, because the facts supportive of that conclusion had not been
pleaded.
10.
The Court of Appeal (Bereaux, Moosai and Pemberton JJA) unanimously
allowed the claimants’ appeal. They concluded that articles C and D were
admissible for the purposes of identifying the claimants as the subject of
articles A and B, and that the judge had been wrong to exclude evidence that
the claimants were in fact identified as such at the time of the publication of
articles A and B. The defendants’ cross-appeal (about which nothing has turned
before the Board) was at the same time dismissed.
Grounds of Appeal
11.
There are the following three grounds of appeal before the Board:
i)
That material subsequent to publication of an alleged defamatory
statement cannot be prayed in aid for the purposes of founding a cause of
action based upon that statement, because a cause of action in defamation (if
there is one) is complete at the moment when a statement is published.
ii)
That, if subsequent material can in principle be relied upon for that
purpose, the Court of Appeal was wrong to reverse the finding of the judge that
articles C and D were not available for that purpose, mainly due to the passage
of time between them and the publication of articles A and B.
iii)
That the Court of Appeal was wrong to reverse the judge’s ruling about
the admissibility of evidence of contemporaneous identification of the
claimants.
It is convenient to deal with ground three first, before
addressing grounds one and two, which need to be considered together.
Ground Three
12.
The claimants sought at trial to introduce (without prior warning)
evidence from the claimants themselves that they had received telephone calls
shortly after the publication of articles A and B suggesting that the callers
had identified them as the subject of those articles. No attempt was made by
the claimants to call any person who made that identification. The judge ruled
that identification of that kind (extraneous to the defamatory statements
themselves) had to be pleaded with particularity, and had not been.
13.
The Court of Appeal regarded that analysis by the judge as “plainly
wrong” although there had been no ground of appeal making any such assertion.
Their view was that the admissibility of the claimants’ evidence was a matter
of weight, rather than one to be resolved on the pleadings.
14.
In the Board’s view the judge made no error of law in relation to this
question. He was better placed than the Court of Appeal or (for that matter)
than the Board to judge whether the absence of the requisite pleading rendered
the introduction of that evidence unfair to the defendants, and his conclusion
that it did cannot be faulted. Furthermore, there having been no ground of
appeal to the Court of Appeal on that issue, it was not one which the Court of
Appeal ought to have decided against the judge.
Grounds One and Two
15.
After a careful analysis of the relevant authorities (referred to below)
the judge concluded that the law did permit reference to subsequent statements
by the defendant for the purposes of identifying the (otherwise unnamed)
subjects of an earlier defamatory statement in certain circumstances, but that
articles C and D could not be used for that purpose in the present case. His
main reasons were, first, that there had been too long a lapse of time between
December 2008 (when articles A and B were published) and June and July 2009
(when articles C and D were published) for readers of articles C and D to make
the necessary connection between the police officers there named, and the
unnamed subjects of articles A and B. Secondly he found that articles C and D
were mainly concerned with reporting of the inquest, and made no sufficient
reference back to articles A and B or their content, to justify their use for
the purpose of identifying the claimants as their subjects.
16.
For its part the Court of Appeal concluded, again, that the judge had
been “plainly wrong” in that analysis sufficient to permit an appeal on what
was essentially a question of fact and evaluation, because the lapse of time
between the two groups of articles, although important, was insufficient to
displace a sufficient nexus between them about a notorious incident attracting
large public attention in a small country. In the view of the Court of Appeal
readers of articles C and D would connect the police officers therein named
with the defamatory content of articles A and B.
17.
In the Board’s view, the central question in this main part of the
appeal is not a question of law as such, since both the judge and the Court of
Appeal directed themselves by reference to a sensible (and similar) summary of
the relevant principles. Rather, the outcome turns on whether the Court of
Appeal ought to have departed from the judge’s factual and evaluative findings
about that issue, and to have come to a different and opposing evaluation of
its own. Nonetheless, and in order to set that question in context, it is
necessary to say something briefly about the conundrum raised by the relevant
authorities.
The Authorities
18.
The claimants in the Grappelli case (referred to above) alleged a
slander by the defendant concert promoters by telling the managers of certain
concert halls that forthcoming concerts by the claimants (the first of whom was
a famous violinist) had been cancelled because Mr Grappelli was very seriously
ill and might never tour again. There was nothing defamatory in that statement,
but publicity two months later of the claimants’ concert programme was alleged,
if aggregated with the original statement, to suggest by way of innuendo that
Mr Grappelli’s assertion of illness as the reason for cancelling his earlier
appearances had been dishonest. The Court of Appeal held that the question
whether a statement (whether libel or slander) gave rise to a cause of action
in defamation had to be determined at the moment of its publication, and not
later. At [1981] 1 WLR 822, 825B, Lord Denning said:
“I would go by the principle,
which is well-established, that in defamation - be it libel or slander - the
cause of action is the publication of defamatory words of and concerning the
plaintiff. The cause of action arises when those words are published to the
person by whom they are read or heard. The cause of action arises then: and not
later.”
At p 831 Dunn LJ said:
“Like Lord Denning MR, I would
prefer to deal with this on principle. I agree that a publication is an
essential part of the cause of action; that once there is publication the cause
of action is complete, and there is no room for the doctrine that the cause of
action can, so to speak, be allowed to be inchoate or lie dormant until such
time as some fact emerges which would transform an otherwise innocent statement
into a defamatory one.”
19.
The principle that a cause of action in defamation arises (if at all)
when the relevant statement is published (ie read by a recipient) is firmly
grounded in the common law of defamation, and was recently affirmed by the
Supreme Court of the United Kingdom in Lachaux v Independent Print Ltd (Media
Lawyers Association intervening) [2019] UKSC 27; [2019] 3 WLR 18, although
not with this conundrum in mind. It was a case about the impact upon the common
law of English legislation not replicated in Trinidad and Tobago.
20.
An early test of the universal applicability of the simple concept laid
down in the Grappelli case was provided, only six months later, by the
facts of the Hayward case. This arose from the Norman Scott affair in
which Mr Scott’s dog was shot by a person who later claimed that he had been
hired to assassinate Mr Scott, who had previously been in a homosexual
relationship with Mr Jeremy Thorpe, the then leader of the Liberal Party. On 9
April 1978 the Sunday Telegraph published an article suggesting that one of two
persons involved in the Scott affair was a wealthy benefactor of the Liberal
Party, but not naming him, and that they had both played a part in funding the
botched attempt to kill Mr Scott. Exactly a week later, on 16 April, the same
newspaper published a second article naming the claimant Mr Jack Hayward as the
party benefactor.
21.
Mr Hayward’s claim was tried before a jury and succeeded. Evidence was
tendered to show that the first article, on its own, sufficiently identified Mr
Hayward as a subject although without naming him. But the trial judge also
directed the jury that it could have regard to the second article for the
purpose of identifying Mr Hayward as the person defamed by the first article.
It was that direction to the jury that led to the appeal, squarely based upon
the Grappelli case which, it was submitted, rendered that direction
erroneous in law.
22.
The Court of Appeal disagreed, although each of the three members of the
court gave different reasons for doing so. Lord Denning said that the second
article was admissible as proof that the defendant had intended to refer to the
plaintiff in the first article. He said, “if the defendant intended to refer to
the plaintiff, he cannot escape liability simply by not giving his name”. Sir
George Baker said that the Grappelli case could be distinguished upon
the basis that it prohibited the use of subsequent material to render an
originally innocent statement defamatory, but not the use of the subsequent
material to identify the subject of an earlier defamatory statement. He said
that the first article had set readers a kind of crossword puzzle as to the
identity of the benefactor, which had been answered by the second article, so
that both were defamatory in conjunction. He accepted a submission that where
two statements formed part of a saga or series, they may both be looked at for
the purposes of identifying a cause of action in defamation. Sir Stanley Rees
agreed with Sir George Baker’s first reason for distinguishing the Grappelli
case. He said that where the original statement was defamatory, the second
statement identifies the person defamed and is published by the same party, the
two statements could be relied upon in conjunction. He noted that a second
statement with a sufficient reference back to the first could amount to a
re-publication of the first statement. Both Sir George and Sir Stanley also
regarded the first article as having been sufficient on its own to identify the
plaintiff as its subject, albeit only to a narrow class.
23.
Lord Denning’s intention-based reason for distinguishing the Grappelli
case has not stood the test of time. It was doubted by Eady J in Chase v
News Group Newspapers Ltd [2002] EWHC 2209 (QB), and by the Irish Supreme
Court in Bradley v Independent Star Newspapers Ltd [2011] IESC 17. The
leading text books note that it is contrary to established authority to the
effect that a defendant’s subjective intention is relevant to defamation only
upon an issue of malice: see for example Gatley on Libel and Slander 12th
ed (2013), para 31.27, and the earlier decision of the House of Lords in E
Hulton & Co v Jones [1910] AC 20, 24 per Lord Loreburn LC.
24.
Intention aside, the basis of principle upon which the Court of Appeal
in the Hayward case properly departed from its earlier decision in the Grappelli
case has never been satisfactorily resolved, although subjected to intense
judicial analysis. In addition to the Chase and Bradley cases,
there is also Baltinos v Foreign Language Publications Pty Ltd (1986) 6
NSWLR 85, and Fairfax Media Publications Pty Ltd v Pedavoli [2015] NSWCA
237, both from New South Wales.
25.
The Board’s view is that the rigorous exclusionary principle laid down
in the Grappelli case, namely that a subsequent statement by the same
defendant can never be aggregated with an earlier one, although consistent with
the conceptual basis of defamation at common law, goes too far. As has been
frequently observed it accords neither with reason, justice nor common sense.
The facts of the Hayward case, where the two statements were published
only a week apart, the first set a puzzle which the second answered and where
both were part of a saga or series, put that beyond doubt. Therefore the first
ground of appeal, which seeks to uphold that exclusionary principle in full,
fails.
26.
It is unnecessary for the purposes of determining this appeal for the
Board to resolve with any precision the question how the exception to the
exclusionary principle is to be framed. It may well be that there are several
different conceptual routes to its identification on different facts. It is
sufficient to say that the authorities on this question demonstrate that, for
two statements made by the same person, but published at different times, to be
aggregated for the purpose of giving rise in conjunction to a completed cause
of action in defamation, there must in the mind of the reasonable reader be
created a sufficient nexus, connection or association between the two of them,
so that (where one is defamatory and the other identifies the subject) there
comes a moment in time at which, in the mind of that reader, the claimant is
identified as the subject of the defamatory accusation. That moment in time
will generally be the time of publication (ie reading) of the second statement.
27.
That nexus or connection between the two statements may be established
by varying means. The defendant may, in the first statement, have invited the
reader to await further information in a later statement. The two statements
may be part of a single saga or series. The second statement may sufficiently
refer back to the first statement so as to incorporate it by reference, or its
contents as a legal innuendo, in the second statement. But these are not legal
categories. They are merely examples of ways in which, as a matter of fact, a
claimant may prove the requisite nexus or connection between the two
statements. Where this issue arises in defamation proceedings, it is
essentially a question of fact or evaluation, whether for the jury (in the
jurisdictions where the jury is still used) or for the judge, as in Trinidad
and Tobago. If the judge has correctly directed himself in law, then the
question for an appellate court is not whether, with the same direction, it
would have reached a different conclusion, but whether the decision of the
judge was plainly wrong, or outwith the boundaries of reasonable divergence of
view: see, in a slightly different defamation context, Stocker v Stocker
[2019] UKSC 17; [2019] 2 WLR 1033, paras 58-59 per Lord Kerr of Tonaghmore, and
his citation from the judgment of Lord Reed JSC in McGraddie v McGraddie
[2013] UKSC 58; [2013] 1 WLR 2477, paras 3-4.
28.
Returning to the present case, the Board can detect no error of law in
the judge’s self-direction about how he should decide whether articles A and B
could be aggregated with articles C and D for the purpose of completing the
claimants’ cause of action. At para 52 he asked himself whether the reasonable
reader of the two groups of articles would “make a sufficient nexus” between
them, and therefore between the content of articles A and B and the claimants.
He concluded that at the time of the publication of articles C and D where the
claimants were named, “the reasonable avid reader would more likely than not
have forgotten the tenor and purport of (articles A and B)”.
29.
Likewise, the Court of Appeal directed itself, at para 35 of the
judgment of Pemberton JA that:
“There must be a sufficient nexus
between the two sets of publications.”
They reached a different conclusion about that factual
question from the judge, and concluded that the judge was plainly wrong in his
conclusion.
30.
It is not, in the Board’s view, enough for an appellate court to
conclude that a judge was plainly wrong in reaching a factual conclusion merely
because they disagree with it. Nor is it sufficient that the judge may have
expressed his reasons for his conclusion in a relatively summary form so that,
because it does not address the full contrary reasoning of the appellate court,
it is thereby vitiated by being unreasoned. In the present case, the judge
alighted primarily upon the passage of time (in excess of six months) between
the two groups of articles as his main reason for concluding that the
reasonable readers of articles C and D would not have made an association in
their minds with articles A and B but rather would have forgotten those earlier
articles. The Court of Appeal’s principal reason for its opposite conclusion
was its perception that the underlying incident, namely the Wallerfield
shooting, was itself one which was likely to have remained in the minds of the
reasonable reader in a small country with only 1.3m people. The judge did not
mention the size of the population of Trinidad and Tobago in his reasoning, but
it seems to the Board entirely unlikely that he failed to have regard to it, or
considered the matter as if the reasonable readership were inhabitants of the
United Kingdom or London, which the Court of Appeal used by way of contrary
examples.
31.
It seems that the Court of Appeal’s conclusion may have stemmed from
their particular reading of the judge’s reasoning. They asked themselves whether
the judge was “plainly wrong … when he found that the subsequent articles …
were incapable of identifying the officers as the persons referred to …”
(para 34, their emphasis). That appears to reflect one sentence in the judgment
where the judge found “as a fact” that the subsequent publications were “not
capable of identifying the claimants …” (para 50, his emphasis). Had that been
the true basis of the judge’s reasoning, the Board would agree that it would
have been too narrow. However, in the Board’s view that sentence must be read
in the context of the following paragraphs in which the judge made clear that
his conclusion was properly based on his evaluation of the relevant factors.
32.
The Board is entirely unable to conclude, at this second level of appeal,
whether the judge or the Court of Appeal were right about this essentially
factual question. They are, although directly opposed, both conclusions which a
reasonable judicial mind could properly reach, directing itself correctly in
law and paying appropriate attention to the facts as found by the judge (none
of which have been in dispute at the appellate stage). There are factors such
as the passage of time and the different focus between the two groups of
articles that support the judge’s view, and factors such as the notoriety of
the incident in a small society and the invitation in Article B to await the
inquest which support the conclusion of the Court of Appeal. It follows that
the Court of Appeal should not have concluded that the judge was “plainly
wrong”. For that reason alone, this appeal ought to be allowed.
33.
The Board wishes to add however, (although this is not a point which was
advanced in either of the courts below), that if it had been considering the
matter afresh it would have been significantly influenced, in concluding that
the judge’s finding that there was a lack of a sufficient connection or nexus
between the two groups of articles was correct, by a perception that articles C
and, in particular, D were in their tenor exculpatory of the claimants. Article
C referred to expert pathological evidence which tended to support the police
officers’ case, and article D was mainly engaged in reporting, in considerable
detail, the finding of the inquest that the officers’ involvement in the shootings
had been entirely innocent, not merely of deliberate killing, but even of
negligence. It is true that article D made reference to continuing feelings of
injustice on the part of relatives of those who had been killed but, taken as a
whole, it cast no doubt on their reputation, and clearly recorded the decision
of a competent tribunal that they were innocent without adverse editorial
comment of its own.
34.
Where it is sought to use the identification of the claimants in a
later-published statement as identifying them as the subjects of an earlier
defamatory statement, the later statement must be read as a whole. The question
then is, do the reasonable readers of the two statements, if aggregated in
their minds, come away with a perception that the common maker of those
statements is, by the time of the second of them, asserting matters defamatory
of the claimants. If the effect of the second statement (or group of
statements) is to take away the defamatory sting in the first, then the
aggregation may well not be defamatory taken as a whole.
35.
Nonetheless the Board makes these observations for completeness, rather
than by way of decision, which is that this appeal should be allowed, because
the judge was not plainly wrong in his answer to the question whether the two
groups of articles could be aggregated.
