Why Writing “Alleged” Isn’t a Safe Harbour
The use of the word “alleged” has long been misunderstood as a protective device in defamation law.
Many publishers and individuals believe that by prefacing an accusation with “it is alleged”, they insulate themselves from liability. This assumption is incorrect.
Courts consistently emphasise that liability turns on the meaning conveyed to the ordinary, reasonable reader, assessed in the full context of the publication, not on the superficial use of qualifying words.
Australian case law illustrates this clearly. In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, the Court of Appeal confirmed that while “allegations” may in common usage distinguish between untested assertions and facts, the mere inclusion of the term is not conclusive. White JA said at [12]:
It is important to stress that the word used in the broadcast was “allegations”. The oft quoted dicta of Lord Devlin in Lewis v Daily Telegraph Ltd [1964] AC 234 that “[a] man who wants to talk at large about smoke may have to pick his words very carefully if he wants to exclude the suggestion that there is also a fire” (at 285) demonstrates why the use of the word “allegations” is not conclusive. The court must still ask whether the overall publication conveys a defamatory imputation capable of lowering the plaintiff’s reputation.
Similarly, in Bazzi v Dutton [2022] FCAFC 84, the Full Federal Court reiterated that the decisive issue is the impression left on an ordinary, reasonable reader or viewer. The qualifier “alleged” does not automatically neutralise the sting if the context implies guilt or condones the accusation. Rares & Rangiah JJ sait at [25]:
The question for decision is whether an ordinary reasonable reader, listener or viewer would understand that the publication conveys the alleged imputation. The test uses the objective standard of reasonableness, based on how that hypothetical person would understand the matter complained of.
Rares & Rangiah JJ said at [47]:
… it is the general impression created in the mind of the ordinary reasonable reader of a publication that determines whether it conveys one or more imputations of and concerning a claimant. As Lord Devlin said in Lewis [1964] AC at 285 ‘it is the broad impression conveyed by the libel that has to be considered and not the meaning of each word under analysis’.
The principle is longstanding. In Lewis v Daily Telegraph Ltd [1964] AC 234, the House of Lords warned that publications reporting allegations, particularly those connected to criminal conduct, may nonetheless convey to readers that there is substance to the claim. Lord Reid said:
A statement that someone is suspected of fraud or is being investigated in connection with fraud must be defamatory because it tends to make people think the worse of him. So would a statement that he was alleged to be guilty of fraud. Both are allegations of suspicion and both may be unjustified. Both mean that there is smoke and both tend to suggest that there is fire. But in neither case is it permissible to say ‘therefore it is proved that there is fire’. The law is that nothing short of proof of guilt will justify a statement that a man is guilty of an offence, and nothing short of proof of reasonable grounds of suspicion will justify a statement that he is suspected.
The court emphasised that it is not sufficient to state that something is an allegation merely; what matters is whether the publication as a whole suggests that the plaintiff is guilty.
Australian courts have repeatedly adopted this reasoning, treating republication of allegations as tantamount to endorsing them, unless clearly and unequivocally disclaimed.
Statute law reinforces this approach. Under s 25 of the Defamation Act 2005 (QLD) and its uniform counterparts, truth is a defence, but a publisher who repeats an allegation must prove the truth of the underlying imputation, not merely that it was alleged.
Likewise, s 30 of the Act provides a defence of statutory qualified privilege, but this depends on the publisher acting reasonably and in the public interest; using the word “alleged” is only one factor among many, and never determinative.
Other doctrines underscore the risks. The so-called “bane and antidote” principle recognises that inserting a disclaimer or denial after a defamatory sting rarely cures the harm.
Accordingly, the law is clear: “alleged” is not a safe harbour.
Whether in print, broadcast, or social media, liability depends on the overall impression.
If that impression conveys guilt, misconduct, or other defamatory imputations, the mere presence of the word will not prevent a claim.
Publishers must rely instead on substantive defences, such as truth, honest opinion, or qualified privilege, each of which has precise statutory or common law requirements that cannot be satisfied by cautious wording alone.
The Ordinary Reasonable Reader and Overall Impression
Understanding how courts interpret meaning is central to defamation law.
Liability does not turn on technical wording or a publisher’s intention, but on the impression the publication leaves on the ordinary, reasonable reader.
This principle shapes how allegations, qualifiers, and imputations are assessed, and it underpins both the common law authorities and the statutory framework.
The Lewis v Daily Telegraph authority on meaning and impression
Defamation law in Australia and the United Kingdom has consistently affirmed that the meaning of a publication is judged through the lens of the ordinary, reasonable reader.
The classic authority is Lewis v Daily Telegraph Ltd [1964] AC 234, where the House of Lords held that the key question is what impression the publication leaves on such a reader, not the technical language used by the publisher. Lord Reid said at p 259:
The ordinary man does not live in an ivory tower. He is not inhibited by knowledge of rules of law or evidence. He can and does read between the lines in the light of his general knowledge and experience of worldly affairs. … But the court must not give the words a meaning which they cannot reasonably bear. … It would be an inference upon an inference to deduce guilt of fraud from the fact that someone is suspected of fraud.
The Court explained that jurors must consider the “broad impression” conveyed, taking into account the tendency of ordinary people to read between the lines and draw reasonable inferences. Lord Devlin said at p 285:
… it is the broad impression conveyed by the libel that has to be considered and not the meaning of each word under analysis.
The case involved a report that a company was under investigation by fraud authorities. Although couched as an allegation, the publication was held to be capable of conveying the defamatory meaning that the company was actually guilty of fraud.
Australian courts have repeatedly applied this principle. In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, the Court of Appeal emphasised that while the word “alleged” may indicate that accusations are unproven, the ordinary understanding of readers does not always stop there.
If the overall tenor of the broadcast implied guilt, the imputation was still defamatory. White JA said at [13]:
I do not accept that a reasonable viewer not avid for scandal would assume from the fact that employees had complained of bullying, that the complaint of bullying was justified, as distinct from some employees being unhappy and being prepared to make a complaint, whether justified or not.
Likewise, in Bazzi v Dutton [2022] FCAFC 84, the Full Federal Court considered whether a tweet branding a federal minister a “rape apologist” conveyed a defamatory imputation.
The Court confirmed that qualifiers and context must be assessed objectively: the decisive issue was what the hypothetical reasonable reader or viewer would infer, rather than the publisher’s intent. Wigney J said at [62]–[63]:
… The broad and overall impression that the ordinary reasonable reader would have gleaned from the tweet is that Mr Bazzi was passing on content from an article in The Guardian about something that Mr Dutton had said, along with his comment or observation on that topic. … The ordinary reasonable reader … would consider the tweet as a whole …
This line of authority underscores that liability cannot be escaped by careful wording alone. Courts will examine the whole publication, including headlines, emphasis, and tone, to determine its sting.
The “ordinary reasonable reader” is neither defamation lawyers parsing technicalities nor a naïve reader incapable of inference, but a person who brings general knowledge and common sense to bear in interpreting what they read or hear.
Pleading natural & ordinary meaning vs (true/false) innuendo
In defamation proceedings, plaintiffs traditionally plead that the matter conveys either its natural and ordinary meaning or a true or false innuendo.
The natural and ordinary meaning arises from the words themselves as understood by the reasonable reader, without special knowledge.
A true innuendo requires extrinsic facts known to readers to convey a defamatory sting. In contrast, a false innuendo arises where an implication is drawn even though no special facts are required.
However, courts have made clear that these labels do not control the outcome.
The practice of pleading imputations in Australia stems from procedural fairness: the defendant must know the case it has to meet.
Yet, as the High Court noted in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50, the publication must always be read as a whole and judged by its impact on the ordinary reader. McHugh J said at [26]:
A reasonable reader considers the publication as a whole. Such a reader tries to strike a balance between the most extreme meaning that the words could have and the most innocent meaning. The reasonable reader considers the context as well as the words alleged to be defamatory. … The bane and antidote must be taken together.
The categorisation of a pleaded imputation as “ordinary meaning” or “innuendo” is less important than whether it is reasonably capable of being conveyed.
The Uniform Defamation Acts, including the Defamation Act 2005 (QLD), require courts to determine whether the matter complained of is capable of conveying the imputation alleged.
Section 25 (truth) and section 30 (qualified privilege) both presuppose that the defamatory meaning has first been identified.
As the Court explained in Harvey v John Fairfax Publications Pty Ltd [2005] NSWCA 255, whether a pleaded imputation is capable of arising is a question of law for the judge; whether it was actually conveyed is a question of fact for the tribunal. Hunt AJA said at [192]:
It is for the judge to decide whether the words complained of are capable of conveying the pleaded imputation. If so, it is then for the jury to decide whether, in fact, the publication did convey that imputation to the ordinary reasonable reader.
The High Court in Trkulja v Google LLC (2018) 263 CLR 149 reaffirmed this, holding that courts must reject “strained, forced or utterly unreasonable” interpretations, but allow for the ordinary reader’s tendency to draw inferences. Kiefel CJ, Bell, Keane, Nettle and Gordon JJ said at [33]:
It is well settled that the question whether particular matter is capable of conveying a defamatory meaning is a question of law. The court must rule out meanings which are ‘strained, or forced, or utterly unreasonable’… But the court should proceed on the basis that the ordinary reasonable reader is a person who can read between the lines in the light of that reader’s general knowledge and experience of worldly affairs.
Thus, the categorisation of imputations is a tool for pleading, not a substantive shield.
The ultimate inquiry remains: what did the publication, in context, convey to the ordinary, reasonable reader?
Saying “It Is Alleged…” vs Asserting Facts — Does It Change the Sting?
A common misconception among publishers is that prefacing a claim with the word “alleged” insulates them from defamation liability.
The law is clear that this is not the case.
Courts consistently emphasise that the decisive question is what the ordinary, reasonable reader would take away from the publication when read as a whole.
Merely inserting the term “alleged” does not guarantee that the sting of the imputation is neutralised.
In Lewis v Daily Telegraph Ltd [1964] AC 234, the House of Lords considered whether reporting that a company was “under investigation” for fraud conveyed a defamatory sting.
The court held that although the article technically described an allegation, the ordinary reader could still infer that the company was, in fact, guilty of misconduct. Lord Devlin said at p 285:
… it is the broad impression conveyed by the libel that has to be considered and not the meaning of each word under analysis.
This principle has been absorbed into Australian law. The High Court in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50 warned that the republication of accusations generally carries with it the adoption of the defamatory sting, unless it is clearly and fairly qualified. Gleeson CJ said at [26]:
The law has long taken the view that a person who republishes defamatory material is to be treated, in general, as accepting responsibility for the defamatory character of what is repeated. The repetition of an accusation, without more, is taken to convey the same imputation as the original publication.
Gleeson CJ also said at [27]:
Of course, circumstances may exist in which the manner of reporting, and the context, make it clear that the publisher is not adopting the truth of the allegation but is merely reporting its existence. But unless that is made clear, the ordinary reasonable reader is likely to treat the repetition as carrying with it the same defamatory sting.
Recent appellate authority reinforces this. In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, White JA noted that while “allegation” in common parlance distinguishes between untested claims and established facts, it is not conclusive. White JA said at [12]:
It is important to stress that the word used in the broadcast was ‘allegations’. … the use of the word ‘allegations’ is not conclusive. However, it should not be overlooked that the ordinary understanding of that word does not always imply guilt or that the underlying substance is true. In common parlance it is often used to distinguish between untested or unproved assertions, which may or may not have foundation, from established facts and circumstances.
The Court emphasised that the word does not invariably prevent an imputation of guilt from being conveyed; what matters is the overall tenor of the broadcast and the weight that ordinary viewers would likely give to the accusation as opposed to the qualifier.
Similarly, in Bazzi v Dutton [2022] FCAFC 84, the Full Federal Court considered a tweet calling a federal minister a “rape apologist.” The Court observed that whether such a statement is defamatory is determined objectively by reference to the ordinary reader, not the subjective intent of the publisher.
The presence of qualifying language, or lack thereof, was not determinative. Instead, the focus was on the impression conveyed when the words were read in their social media context, where brevity and emphasis can magnify defamatory imputations.
The statutory framework is consistent with this approach. Under the Defamation Act 2005 (QLD) and its uniform counterparts, a publisher who repeats a defamatory allegation cannot rely on proving merely that an allegation was made.
Section 25 (truth) requires proof of the truth of the underlying imputation, while section 30 (qualified privilege) requires the publisher to demonstrate that publication was reasonable in the circumstances.
Neither provision recognises the word “alleged” as a free-standing defence.
Courts are also sceptical of attempts to rely on disclaimers or denials to counterbalance defamatory accusations, the so-called “bane and antidote” principle.
In Charleston v News Group Newspapers Ltd [1995] 2 AC 65, the House of Lords held that the effect of a publication must be considered as a whole, but noted that qualifiers or explanations are often overwhelmed by sensational headlines or images. Lord Bridge said at p 70–71:
… it is necessary to take into account the context in which the words were used and the mode of publication. … the bane and antidote must be taken together. … Whether the text of a newspaper article will, in any particular case, be sufficient to neutralise the defamatory implication of a prominent headline will sometimes be a nicely balanced question for the jury to decide and will depend … on the manner in which the whole of the relevant material is set out and presented.
Lord Nicholls said at p 77–78:
This is not to say that words in the text of an article will always be efficacious to cure a defamatory headline. It all depends on the context, one element in which is the lay-out of the article. Those who print defamatory headlines are playing with fire. The ordinary reader might not be expected to notice curative words tucked away further down in the article. The more so, if the words are on a continuation page to which a reader is directed.
Australian courts have applied this principle to hold that readers may accord more weight to the defamatory element than to the subsequent qualification, particularly where presentation or emphasis reinforces the damaging sting.
Accordingly, the law draws no sharp line between saying “it is alleged” and asserting facts. The critical issue is the context, emphasis, and overall impression.
If an ordinary reader would reasonably take away a belief that the plaintiff is guilty of misconduct, the publication will be defamatory despite the presence of the word “alleged”.
Repetition and Adoption of Third-Party Allegations
One of the most firmly established principles of defamation law is that a publisher who repeats a defamatory allegation made by another is generally treated as having adopted it.
The law does not recognise a “neutral reporting” rule at common law. Instead, liability attaches because the ordinary reasonable reader often infers that repetition carries an endorsement of the underlying accusation.
The High Court in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50 made this point clear. McHugh J said at [139]:
The bane and antidote must be taken together. A reasonable reader considers the publication as a whole. Such a reader tries to strike a balance between the most extreme meaning that the words could have and the most innocent meaning. The reasonable reader considers the context as well as the words alleged to be defamatory.
The case involved the publication of extracts from a speech made in Parliament, which suggested corrupt conduct. The publisher sought to argue that it was merely reporting what had been said by a third party.
The High Court rejected this, stressing that defamation law requires the publication to be considered as a whole.
If, when read in its entirety, the material leaves readers with the impression that the plaintiff is guilty of misconduct, the publisher is liable as though it had made the allegation itself.
The judgment confirmed that merely attributing an allegation to another source is not a shield. Gleeson CJ observed at [26]–[27] that a publication cannot escape liability by couching the words as someone else’s statement if the “sting” is nonetheless conveyed.
This approach reflects the ordinary reader’s perspective: a defamatory accusation is no less damaging when presented as a quotation or report, unless the reporting clearly undermines or neutralises it.
The case law underscores how context dictates liability.
In Charleston v News Group Newspapers Ltd [1995] 2 AC 65, although the defendants relied on the presence of disclaimers within the article, the House of Lords held that the entire publication must be assessed.
If the overall impression created is defamatory, disclaimers or qualifications may not suffice.
Similarly, Australian courts applying Rivkin have emphasised that while context might occasionally neutralise the effect of an allegation, this will be rare.
The general rule remains that repeating an allegation is tantamount to making it.
Accordingly, reporting “what others said” is not a safe course. Unless the publication clearly distances itself from the allegation and undermines its defamatory sting, the ordinary, reasonable reader will likely treat it as adopted.
Rivkin remains the leading Australian authority, reminding publishers that the law considers the overall impression of the publication, rather than the technical form of words used to attribute the accusation.
The “Bane and Antidote” Principle — Why Disclaimers Rarely Cure
Courts have long recognised that the impact of a defamatory publication cannot be sanitised by adding a denial, disclaimer, or qualifying term such as “alleged” or “rumoured.”
This is encapsulated in the “bane and antidote” principle, which requires the publication to be read as a whole but acknowledges that the antidote will rarely remove the sting of the bane.
The classic English authority is Charleston v News Group Newspapers Ltd [1995] 2 AC 65, where the House of Lords held that an article alleging actors’ involvement in pornography was defamatory despite disclaimers buried in the text. Lord Bridge said at [69] to [70]:
The article was accompanied by doctored photographs implying two actors’ involvement in pornography. Although the body of the article contained explanations disclaiming that implication, the plaintiffs contended — and the House accepted — that the defamatory sting still arose because of the presentation.
Even where qualifying words exist, they may be overwhelmed by sensational headlines, photographs, or context. The ordinary reader, influenced most by the initial impact, may never absorb the purported antidote.
Australian courts have consistently applied this approach.
In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, the Court of Appeal noted that the ordinary understanding of the word “allegations” does not invariably imply guilt, but equally it does not prevent an imputation of guilt from being conveyed.
Where the sting of the broadcast suggested wrongdoing, the use of “alleged” was insufficient to neutralise its defamatory effect.
White JA’s reasoning (as noted above) reflects the judicial concern that readers or viewers place more weight on the accusation itself than on the qualifier.
The High Court in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50 further confirmed that context is everything. If the publication, read as a whole, suggests the publisher has adopted the allegation, disclaimers will not suffice.
Similarly, the Court in Trkulja v Google LLC (2018) 263 CLR 149 warned that strained interpretations must be rejected, but technical caveats cannot displace reasonable inferences drawn by readers.
This approach reflects a pragmatic view of human behaviour. Readers and viewers typically engage with headlines, images, and prominent statements; qualifications tucked away in the body of the publication may be ignored or discounted.
The law, therefore, treats such disclaimers with scepticism, recognising that the damage to reputation arises from the overall impression left by the publication.
In practice, this means that adding “it is alleged” or “rumoured” does not immunise a publisher. If the ordinary, reasonable reader would still understand the publication as asserting guilt or serious sexual misconduct, the sting remains.
The bane and antidote principle ensures that publishers cannot rely on superficial caveats to escape responsibility for defamatory meaning.
When “Alleged” Appears in Reports of Official Proceedings
The word “alleged” assumes a distinct meaning when used in the context of official proceedings.
While it cannot by itself shield a publisher from liability, the law provides a limited statutory defence for fair and accurate reports of matters of public concern.
This protection is narrowly framed, requiring accuracy, attribution, and substantial fairness, and it applies only where the publication genuinely reflects what occurred in the recognised proceeding.
Elements of the fair & accurate report defence
The law recognises a limited statutory protection for publishers who report on official proceedings, even where the report includes defamatory allegations.
Under s 29 of the Defamation Act 2005 (QLD) and its uniform counterparts, it is a defence if the defendant proves that the matter was published as, or was contained in, a fair report of proceedings of public concern.
This defence is grounded in the public interest in open justice and transparency of government functions.
To succeed, three key requirements must be satisfied
The following three (3) requirements must be satisfied:
- Accuracy
- Attribution
- Substantial fairness
We will explain these in more detail below.
Accuracy — The report must fairly reflect what occurred in the proceeding. Courts require that the words used do not distort or sensationalise the underlying event.
In Hoser v Harrison [2024] FedCFamC2G 436, the Federal Circuit and Family Court emphasised that what matters is whether the imputation pleaded is reasonably capable of being conveyed by the published matter, judged against the actual content of the proceeding. At [191], the Court stated:
It is not sufficient that the pleaded imputation must be stated with sufficient precision. The pleaded imputation must be one that is reasonably capable of being conveyed by the matter published. Whether a pleaded imputation is reasonably capable of being so conveyed is a question of law; and although such question permits only one answer, whether any particular pleaded imputation is reasonably capable of being conveyed by the publication is ‘a question about which reasonable minds may sometimes differ’.
Attribution — The report must make clear that the allegations are those made in the official forum, not the publisher’s own assertions.
This requirement reflects the reasoning in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50, where the High Court stressed that attribution alone will not save a publication if the overall impression is one of adoption (as extracted above).
Nonetheless, attribution is an essential element for the statutory defence because it informs the reader that the source of the allegation is a judicial, parliamentary, or other public authority.
Substantial fairness — The report must not present a one-sided or misleading version.
In Morgan v John Fairfax & Sons Ltd (No 2) (1991) 23 NSWLR 374, the court recognised that reasonableness in reporting is assessed by reference to all circumstances, including whether the publication presents a balanced account.
This principle is now reflected in s 30(3) of the Defamation Act 2005 (QLD), which requires the court to consider multiple factors in evaluating reasonableness. It says:
(3) In determining for the purposes of subsection (1) whether the conduct of the defendant in publishing matter about a person is reasonable in the circumstances, a court may take into account the following factors to the extent the court considers them applicable in the circumstances—
(a) the seriousness of any defamatory imputation carried by the matter published;
(b) the extent to which the matter published distinguishes between suspicions, allegations and proven facts;
(c) the nature of the business environment in which the defendant operates;
(d) whether it was appropriate in the circumstances for the matter to be published expeditiously;
(e) any other steps taken to verify the information in the matter published.
It is important to note that this defence applies only where the reporting is faithful to the proceeding itself. The mere use of the word “alleged” is not enough to meet these statutory requirements; the defence depends on substance, not formulaic language.
What does (and doesn’t) qualify as a “report” of proceedings
The protection of s 29 is confined to matters of “public concern” expressly defined to include judicial proceedings, parliamentary debates, and other recognised forums.
A “report” in this context means a communication that conveys, in substance, what transpired in those proceedings.
The High Court in Trkulja v Google LLC (2018) 263 CLR 149 reaffirmed that the publication must be assessed by what the ordinary reasonable reader would understand.
Where a media outlet paraphrases or sensationalises allegations, it may no longer be a “report” of proceedings but rather an independent defamatory publication.
Likewise, if a headline or presentation conveys that the publisher endorses the allegation, the protection of s 29 is likely to be lost.
In contrast, faithful and accurate reporting of what was said in court, Parliament, or another official proceeding — even if defamatory — may attract protection.
The statutory defence exists to ensure the public can be informed of what transpires in matters of public concern.
But the boundaries are tight: embellishment, omission of critical context, or reliance on the word “alleged” alone will not suffice.
Defences and How “Alleged” Interacts with Them
The main statutory and common law defences in defamation law—truth, honest opinion, and qualified privilege—operate independently of any cautious wording like “alleged.”
Each defence requires strict proof: establishing the truth of the underlying sting, demonstrating that a statement is genuinely opinion supported by proper material, or showing that publication was reasonable in context.
Courts have repeatedly confirmed that the mere use of “alleged” cannot satisfy these substantive requirements.
Truth/Justification — proving the underlying sting, not just “alleged”
The defence of truth, also referred to as justification, is codified in s 25 of the Defamation Act 2005 (QLD) and its uniform counterparts.
It provides that it is a defence to the publication of defamatory matter if the defendant proves that the defamatory imputations carried by the publication are substantially true.
The critical point is that the publisher must prove the truth of the underlying sting, not merely that an allegation was made.
This principle has been settled for decades. As the House of Lords explained in Lewis v Daily Telegraph Ltd [1964] AC 234, (above), a defendant cannot escape liability by showing only that an inquiry or allegation existed; the truth that must be proved is the defamatory meaning conveyed to the reasonable reader.
Australian courts have adopted this stance. In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, the Court of Appeal highlighted that merely proving the existence of “allegations” was insufficient.
The publisher had to justify the substance of the imputations reasonably conveyed by the broadcast. The decision illustrates that the word “alleged” does not alter the operation of s 25: the defence of truth requires demonstration that the sting of the imputation itself was accurate.
Honest Opinion — opinion vs disguised fact and the need for proper substratum
The statutory defence of honest opinion is set out in s 31 of the Defamation Act 2005 (QLD). It applies where the defendant proves that:
- The matter was an expression of opinion rather than a statement of fact.
- The opinion related to a matter of public interest; and
- The opinion was based on proper material.
Courts have consistently warned that publishers cannot disguise assertions of fact as opinion simply by adding qualifiers such as “it is alleged.”
In Bazzi v Dutton [2022] FCAFC 84, the Full Federal Court considered whether a tweet branding a minister a “rape apologist” was opinion or fact.
As already discussed above, the Court reaffirmed that the dividing line turns on what the ordinary reasonable reader would perceive.
If a statement purports to be a fact, attaching the label “alleged” will not transform it into opinion.
The requirement of a proper substratum is critical. In Barbaro v Amalgamated Television Services Pty Ltd (1989) 20 NSWLR 493, the New South Wales Court of Appeal emphasised that an opinion must be based on facts which are either true, protected by privilege, or otherwise defensible.
Thus, where a publication presents an allegation as the factual foundation for an opinion, the publisher cannot rely on the mere existence of the allegation — the underlying facts must themselves be justified or privileged.
Qualified Privilege — reasonableness and context requirements
Qualified privilege exists both at common law and in statute. The statutory form is contained in s 30 of the Defamation Act 2005 (QLD), which provides a defence where the defendant proves that the recipient had an interest in receiving the information, the matter was published in the course of giving that information, and the conduct of the publisher was reasonable in the circumstances. Section 30 states:
(1) There is a defence of qualified privilege for the publication of defamatory matter to a person (the “recipient”) if the defendant proves that—
(a) the recipient has an interest or apparent interest in having information on some subject; and
(b) the matter is published to the recipient in the course of giving to the recipient information on that subject; and
(c) the conduct of the defendant in publishing that matter is reasonable in the circumstances.
The High Court in Lange v Australian Broadcasting Corporation (1997) 189 CLR 520 recognised an extended qualified privilege at common law for communications about government and political matters. The Court said at 571–572:
The common law of Australia should now be developed, consistently with the Constitution, so as to recognise an extended qualified privilege to publish defamatory matter about the conduct of government and political matters that affects the people of Australia. This qualified privilege extends to communications to the public on government and political matters. However, the privilege is not absolute: the publication must be reasonable in the circumstances.
However, reasonableness remains a central touchstone. In Roberts v Bass (2002) 212 CLR 1, the Court confirmed that recklessness or lack of diligence in verifying allegations could defeat the defence. Gaudron, McHugh and Gummow JJ said at [75]:
The protection of the privilege is lost if the defendant publishes the defamatory matter recklessly, without considering or caring whether it is true or false. The privilege is also lost if the defendant fails to give consideration to the reasonableness of the communication in the circumstances.
Then said at [77]:
A defendant cannot rely upon the privilege if he or she did not believe that the defamatory material was true or was indifferent to its truth or falsity. Equally, a failure to take proper steps to verify the accuracy of serious allegations may deprive the defendant of the protection of the privilege.
The statutory requirements of s 30(3) of the Act direct the court to consider a range of factors, including the seriousness of the imputation, the steps taken to verify its accuracy, and whether the matter contained the substance of the plaintiff’s side of the story.
The use of the word “alleged” does not satisfy these requirements.
As Morgan v John Fairfax & Sons Ltd (No 2) (1991) 23 NSWLR 374 made clear, the overall conduct of the publisher is scrutinised. Reasonableness requires more than cautious terminology; it demands a balanced and careful approach to reporting serious allegations.
Taken together, the defences of truth, honest opinion, and qualified privilege all demonstrate that “alleged” has no independent protective force.
What matters is whether the publisher can prove the truth of the imputation, establish that a statement is genuinely opinion supported by a factual substratum, or show that the publication was reasonable and proportionate in the circumstances.
Case Law Quick Summary
Australian and English case law provides clear guidance on how courts treat the use of “alleged” in defamation.
Authorities such as Lewis v Daily Telegraph, Rivkin, Bailey, and Bazzi demonstrate that liability hinges on the overall impression conveyed, rather than the cautious language chosen.
Repetition of third-party allegations and reliance on qualifiers rarely protect a publisher, with courts consistently warning that “alleged” is no substitute for a substantive defence.
Ordinary meaning and impression (Lewis and progeny)
The starting point for determining defamatory meaning is the ordinary, reasonable reader test. In Lewis v Daily Telegraph Ltd [1964] AC 234, the House of Lords held that a publication alleging a company was “under investigation” for fraud could convey to readers that the company was guilty, even though technically framed as an allegation.
The case established that courts must consider the impression left on ordinary readers, who may “read between the lines” and infer wrongdoing.
This approach has been absorbed into Australian law. In Trkulja v Google LLC (2018) 263 CLR 149, the High Court emphasised that the question of capacity—whether material is capable of conveying a defamatory imputation—is a matter of law, and that strained or unreasonable interpretations must be rejected.
However, if the ordinary reader could reasonably infer guilt or misconduct from the words, liability may follow regardless of the publisher’s qualifications.
Repetition/adoption and context (Rivkin)
The High Court confirmed in John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50 that a publisher who repeats another person’s defamatory statement is generally treated as adopting it.
The publication must be assessed “as a whole,” and attribution alone will not immunise liability if the sting is effectively conveyed.
Gleeson CJ noted that readers do not draw fine distinctions between repeating and endorsing: if the overall impression suggests the plaintiff is guilty, the publisher is responsible.
The strictness of the repetition rule reflects public policy against the spread of rumour.
Without it, defamatory matter could be endlessly republished with impunity. Only where the report falls within a statutory defence—such as a fair and accurate report of judicial or parliamentary proceedings can liability be avoided.
“Alleged” is not a shield — recent judicial treatment (Bailey; Bazzi)
Contemporary Australian appellate courts have squarely addressed the limits of cautious terminology.
In Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352, the Court of Appeal held that although the word “allegations” may in ordinary usage distinguish between unproven and established facts, its inclusion is not determinative.
The broadcast was still capable of conveying imputations of guilt. White JA observed that the weight of the defamatory material may eclipse the effect of the qualifier.
In Bazzi v Dutton [2022] FCAFC 84, the Full Federal Court considered whether a tweet calling a senior politician a “rape apologist” conveyed a defamatory imputation.
The Court reaffirmed that the decisive question is the impression on the ordinary reader or viewer, not the subjective intent of the publisher.
The presence or absence of the word “alleged” was immaterial: the focus remained on whether the defamatory sting was in fact communicated.
These recent authorities demonstrate that “alleged” is not a shield. The courts treat it as a minor contextual factor, not a determinative legal defence.
Table of Key Cases on the Use of “Alleged” in Defamation
| Case | Court/Year | Precedent / Principle Established |
|---|---|---|
| Lewis v Daily Telegraph Ltd [1964] AC 234 | House of Lords (UK) | Established that liability depends on the broad impression left on the ordinary, reasonable reader. Even stating that someone is “under investigation” or “alleged” to have committed wrongdoing may still convey guilt. Disclaimers or cautious wording do not prevent a defamatory meaning if the overall implication is one of guilt. |
| John Fairfax Publications Pty Ltd v Rivkin [2003] HCA 50 | High Court of Australia | Confirmed that repeating a defamatory allegation is generally treated as adopting it. A publisher cannot escape liability by attributing allegations to others unless the context clearly and fairly distances them from the claim. Introduced the “bane and antidote” principle: disclaimers rarely neutralise the sting. |
| Charleston v News Group Newspapers Ltd [1995] 2 AC 65 | House of Lords (UK) | Reinforced the “bane and antidote” principle: a publication must be assessed as a whole, but disclaimers or clarifications buried in the text are often insufficient to cure defamatory headlines, images, or prominent statements. |
| John Fairfax Publications Pty Ltd v Harvey (Harvey v John Fairfax Publications Pty Ltd) [2005] NSWCA 255 | New South Wales Court of Appeal | Clarified that whether words are capable of conveying an imputation is a question of law for the judge. In contrast, whether they actually conveyed that meaning is a matter of fact for the jury to determine. Reinforced the centrality of the “ordinary reasonable reader” test. |
| Trkulja v Google LLC (2018) 263 CLR 149 | High Court of Australia | Reaffirmed that courts must exclude strained or unreasonable interpretations but must allow for the ordinary reader’s tendency to read between the lines. Confirmed that liability is determined by what an ordinary, reasonable reader would infer, not by technical or subjective interpretations. |
| Bailey v WIN Television NSW Pty Ltd [2020] NSWCA 352 | New South Wales Court of Appeal | Held that while “allegations” may in standard usage signal unproven claims, this is not conclusive. What matters is the overall tenor of the publication. The sting of guilt or misconduct may still be conveyed despite the qualifier. |
| Bazzi v Dutton [2022] FCAFC 84 | Full Federal Court of Australia | Confirmed that the decisive test is objective: what would the ordinary reader infer? The presence of words like “alleged” is not determinative. Context and overall impression decide whether a defamatory imputation is conveyed. |
| Barbaro v Amalgamated Television Services Pty Ltd (1989) 20 NSWLR 493 | New South Wales Court of Appeal | Clarified the defence of honest opinion: an opinion must be based on a proper substratum of facts. Labelling something “alleged” does not turn a factual claim into an opinion. |
| Morgan v John Fairfax & Sons Ltd (No 2) (1991) 23 NSWLR 374 | New South Wales Court of Appeal | Explained that “reasonableness” is essential to the statutory qualified privilege defence. Use of “alleged” alone does not establish reasonableness; balance, verification, and fairness in reporting are required. |
| Hoser v Harrison [2024] FedCFamC2G 436 | Federal Circuit and Family Court of Australia | Stressed that an imputation must be reasonably capable of being conveyed by the published material, judged against the actual content of proceedings. Reinforces accuracy and fairness requirements under statutory reporting defences. |
Conclusion — Key Risk Warnings and Safe-Use Principles for “Alleged”
The case law and legislation converge on a single principle: the word “alleged” offers no automatic protection against liability in defamation.
Courts will look past superficial qualifiers to determine what meaning the ordinary, reasonable reader would take from the publication.
Meaning and impression: As Lewis and its progeny show, if the sting of guilt or misconduct is conveyed, the use of “alleged” does not prevent liability.
Repetition and adoption: Following Rivkin, publishers who repeat third-party allegations are generally deemed to adopt them, unless they fall within narrow statutory reporting defences.
Modern authority: Bailey and Bazzi confirm that “alleged” carries no independent protective force; the overall impression remains the test.
For publishers, this yields several safe-use principles:
- Never rely on “alleged” alone — it does not alter the legal tests for defamatory meaning or available defences.
- Check the sting — ask whether the publication, read as a whole, still suggests guilt or serious misconduct.
- Rely on substantive defences — truth (s 25), honest opinion (s 31), qualified privilege (s 30), or fair report (s 29) require proof of facts, context, or reasonableness, not cautious wording.
- Consider presentation — headlines, emphasis, and images often overpower disclaimers; courts assess the entire publication.
In short, the law warns against treating “alleged” as a safe harbour.
Liability turns on substance, not form. Publishers must rely on established statutory and common law defences, supported by accuracy, fairness, and reasonableness, rather than the illusory protection of a single word.
FAQ – Why ‘Alleged’ Isn’t Enough
Here are quick, plain-English answers to common questions about why “alleged” isn’t a magic shield in defamation, focusing on the ordinary, reasonable reader and the overall impression of a publication.
It’s general info, not legal advice, and meant to help you spot risks and use proper defences (truth, honest opinion, qualified privilege, fair report) rather than relying on cautious wording alone.
What does “alleged” actually do in defamation?
Saying “alleged” doesn’t give you a free pass. Courts look at what the ordinary, reasonable reader would take from the piece as a whole. If the overall impression still suggests guilt or serious misconduct, liability can arise. “Alleged” is just one contextual factor and, by itself, won’t neutralise a defamatory sting or satisfy any defence.
Why is the “ordinary reasonable reader” test so important?
Because meaning is judged by how a typical reader, using common sense, interprets the publication as a whole. It’s not about legal technicalities or the publisher’s intention. If the broad impression implies guilt, qualifiers like “alleged” won’t save it. This lens underpins both case law and the statutory defences across the Uniform Defamation Acts.
What did Lewis v Daily Telegraph decide about allegations?
It confirmed that reporting someone is “under investigation” or “alleged” to have done wrong can still convey guilt to ordinary readers. The court stressed the “broad impression” over word-by-word parsing. Put bluntly: pointing to smoke may still imply fire. You can’t justify the publication by proving only that an investigation or allegation existed.
How did Bailey v WIN Television treat the word “allegations”?
Bailey recognised that “allegations” often signal untested claims, but said that label isn’t conclusive. What matters is the overall tenor. If the broadcast, read as a whole, implies the allegations are factual or the person is guilty, the imputation can still be defamatory. Viewers often give more weight to the accusation than the qualifier.
What’s the key takeaway from Bazzi v Dutton?
The court applied an objective test: what would the ordinary reader or viewer infer? It’s the overall impression that counts, especially in short, punchy formats like tweets. Qualifiers don’t automatically blunt the sting. If the context implies endorsement of the accusation or condones it, the publication can carry the defamatory imputation.
Does repeating someone else’s claim protect me?
No. Repetition usually equals adoption. If you republish an accusation, readers will generally treat it as yours unless you clearly and fairly distance yourself and present it in a way that genuinely neutralises the sting. Merely attributing the claim to a third party is rarely enough if the overall impression still suggests guilt.
What is the “bane and antidote” principle?
It’s the idea that you assess the publication as a whole (bane and antidote together), but disclaimers or denials often won’t cure a damaging headline, image, or emphasis. A sensational presentation can overshadow careful caveats tucked away later. Courts are sceptical that small-print qualifiers undo the reputational harm caused upfront.
Do headlines and images matter?
A lot. Headlines, teasers, and pictures can significantly impact reader takeaway. Even if the body text includes balancing words, a sensational headline or striking image can leave an impression of guilt. Courts assess the entire presentation, and if the overall impression is defamatory, qualifiers further down the page may not be sufficient to save the publication.
Is “alleged” a defence under the Defamation Act?
No. There’s no freestanding “alleged” defence. Substantive defences are truth (s 25), honest opinion (s 31), qualified privilege (s 30), and fair report of proceedings of public concern (s 29). Each has strict elements. The word “alleged” alone does not satisfy any of them and won’t fix an otherwise defamatory publication.
What does the truth/justification defence actually require?
Truth requires proving the underlying sting of the imputation is substantially true—not merely that an allegation or inquiry existed. If the ordinary reader would take your piece to assert guilt, you must prove that guilt. Showing only that “someone alleged it” or that “police investigated” isn’t enough to justify the defamatory meaning.
When does honest opinion apply if I use “alleged”?
Honest opinion protects opinions, not disguised assertions of fact. The opinion must relate to a matter of public interest and be based on proper material. Dressing up a factual claim with “alleged” won’t turn it into opinion. If it reads as a statement of fact, the defence fails unless you can meet the truth requirements.
What is qualified privilege and how does “alleged” fit in?
Qualified privilege protects reasonable communications made to recipients who have a legitimate interest in receiving them. Reasonableness is the touchstone: seriousness of the imputation, verification steps, balance, and whether you included the other side. Simply sprinkling “alleged” through a story doesn’t make it reasonable if the overall conduct and content fall short.
What counts as a fair & accurate report of proceedings?
The fair report defence covers faithful, accurate, and substantially fair reports of recognised proceedings (e.g., courts, Parliament). It’s narrow. You must clearly attribute, avoid distortion or sensationalism, and reflect what actually happened. “Alleged” language doesn’t itself trigger protection; substance, accuracy, and fair presentation do.
Does social media change the test?
The legal test remains the same, but context is crucial. Short formats, headlines and embedded previews can magnify the sting. Courts consider the impression on the ordinary reader when viewing the post as a whole. A brief post that implies guilt isn’t rescued by “alleged” if the overall takeaway still suggests wrongdoing.
Practical tips to reduce defamation risk when using “alleged”?
Don’t rely on the word. Check the sting of the piece as a whole; verify serious claims; distinguish clearly between suspicion, allegations and proven facts; include meaningful balance; attribute accurately; and consider layout, headline and images. If you’re repeating others, make your non-adoption crystal clear—or better yet, meet a proper defence.
