Andrew L. Urban
After a guilty verdict, the defendant is entitled to appeal. In the cases we regard as wrongful convictions, an appeal is an essential step toward correcting error at trial. When is the assumption of innocence negated?
Split decisions at appeal are decided by the majority on the bench. But the logical conclusion of the presumption of innocence is that dissenting opinions among judges represent reasonable doubt and should be resolved in favour of the appellant. For example, in the case of Derek Bromley, the 3:2 split among the High Court justices meant that after 40 years in prison, Bromley’s conviction was maintained by the decision of a single judge. Should that stand?
If highly experienced appellate judges, applying the criminal standard, cannot agree that a conviction is safe, the system should err toward the appellant’s liberty rather than continued punishment.
The criminal law says guilt must be proved beyond reasonable doubt. If two High Court judges conclude the conviction is unsafe, many people naturally ask: isn’t that reasonable doubt?
In Bromley’s case, the dissenting judges considered the evidence troubling enough that they would have quashed the conviction. Critics say that should itself have triggered the benefit of the doubt.
By rights, I argue, Bromley’s conviction should have been quashed in 2023 after the High Court appeal. Instead, he was released on parole.
The powerful points here on split decisions, fair trial failures, and appellate courts potentially stepping into the jury’s shoes are spot on. One overlooked dimension is the unique limbo created by parole after decades of contested imprisonment, especially for an Indigenous man like Bromley.
In Bromley’s situation, the system has effectively conceded enough doubt to release him into the community under supervision, yet it refuses to fully restore the presumption by quashing the conviction. This hybrid status (technically guilty but free-ish) doesn’t just punish the individual; it perpetuates a form of ongoing state control that echoes historical patterns of conditional “freedom” for First Nations people.
It signals to the public and the community that the conviction stands, discouraging deeper scrutiny or compensation, while the dissenting High Court voices highlight that reasonable minds found the evidence troubling.
True finality in justice shouldn’t mean forcing someone to carry the label of “murderer” indefinitely on a bare majority technicality when senior judges couldn’t unanimously agree the case was safe. This isn’t merely about one appeal process it’s about whether the system can ever admit systemic error in a way that allows full civic restoration, especially where cultural trust in institutions has been repeatedly strained.
Requiring unanimity or a stronger “unsafe” threshold on final appeals (particularly High Court splits) would better align with the “beyond reasonable doubt” ideal without upending majority rule. Cases like this expose how parole can become a pressure valve that avoids real accountability.
The rule of law demands better than “close enough” after 40 years.
It is important to appreciate in this context that any accused person is entitled to a fair trial and an effective appeal – rights established in the ICCPR to which Australia is a signatory. The presumption of innocence can only be negated after those conditions have been fulfilled. These are important parts of the ‘rule of law’ and the ‘rules based order’ which the Prime Minister is fond of quoting these days as a key virtue of the Australian judicial system. It has clearly failed in this case.
It is self-evident that Mr Bromley did not get a fair trial. The evidence relied upon by the prosecution at trial and in the final appeal to the High Court was given by Dr Manock. In submissions to the High Court in 2023 on the Bromley appeal one can plainly see the DPP informing the HC that they could rely upon the evidence given by Dr Manock at Mr Bromley’s trial.
The prosecutor must have known that to be false because I wrote to him on 20 March 2023 informing him why that was so – http://www.netk.net.au//Manock/Manock49.pdf
the letter was also tabled in parliament prior to the HC appeal:
It explained that Dr Manock was not qualified to certify cause of death and not been qualified to give expert evidence in court – that claim was based upon the sworn evidence to that effect from the head of the forensic science centre in 1975 (Dr Bonnin) The above letter also referred to a great many other defects involving Manock.
In my letter of 16 April 2026 to the Federal Attorney-General I stated the following: http://www.netk.net.au//Manock/Manock68.pdf
In the High Court proceedings, the DPP stated that the injuries in this case were ante-mortem at trial “on anyone’s case”. Yet, at the time he made that statement in the HC he had reports from three eminent experts (including the Crown’s own expert) to state that all injuries could have been caused post-mortem. He failed to disclose or refer to those reports.
The DPP also told the HC that “on anyone’s case at trial – this man had drowned. It was not in dispute he had drowned”. That was also based upon the inadmissible evidence of Dr Manock. At the time he made that statement to the High Court he had three expert reports (including one by the expert appointed by the Crown) to state that cause of death by drowning had not been established, and it had to be put down as ‘undetermined’.
The first letter above clearly explains why Bromley did not get a fair trial.
The second letter above provides clear reasons to accept that he did not get an effective appeal.
It should be pointed out that the HC expressly forbade Bromley’s counsel on the appeal from mentioning anything about the forensic pathology evidence. But the court did not object to the above statements by the DPP and in the judgment of the court it referred to the forensic pathology evidence on many occasions.
It is a fundamental part of the adversarial due process that at trial and on appeal both parties are able to challenge the case put against them. Yet, this fundamental right was denied to Mr Bromley at his trial and on his appeal to the HC. There are of course a great many problems with the trial and appeals in this case, but the above are sufficient to establish that the proceedings involving Mr Bromley have been insufficient to remove his right to the presumption of innocence.
Thank you Bob. It is a disturbing matter that the courts fail to act within the rule of law.
Here’s LIA’s answer to the question in full:
Assumed Jurisdiction: Australia / common-law criminal appeals
External search performed: No further search; based on the user-provided articles, prior High Court information already checked, and LIA protocol materials.
One-line conclusion
LIA’s answer is: the presumption of innocence does not continue on appeal in its full trial form, but Bromley’s case shows why that answer is inadequate where the appeal court appears to have preserved a conviction by reasoning that should have belonged to a jury.
Key issues
The first article asks whether a split appellate decision should itself create reasonable doubt.
The better answer, especially after the second article, is more precise:
A dissenting appellate judgment does not automatically equal reasonable doubt. But where the dissent is combined with an alleged appellate re-trial, untested Crown-favourable material, a compromised key witness, and disputed forensic evidence, the conviction should be treated as unsafe unless the State can give a legally disciplined answer.
That is the critical distinction.
LIA’s reply to the question
The presumption of innocence is formally displaced after conviction. Once a jury convicts, the appellant no longer stands before the court exactly as an unconvicted accused. The appeal court is not simply another jury asking whether guilt has been proved beyond reasonable doubt from scratch.
But that does not mean the presumption of innocence is dead.
In a criminal system committed to proof beyond reasonable doubt, the appeal court must not preserve a conviction by doing what the jury did not do. If the trial was infected by error, unreliable forensic evidence, or material evidence the jury never properly assessed, the proper appellate function is to identify whether a miscarriage occurred — not to rebuild the prosecution case from the bench.
That is why the second article matters.
The strongest objection to Bromley’s treatment is not merely:
“Two High Court judges dissented, therefore there was reasonable doubt.”
It is:
“If senior appellate judges considered the conviction unsafe, and if earlier appeal reasoning effectively substituted judicial fact-finding for jury fact-finding, then the system may have converted appellate review into a mechanism for preserving a doubtful conviction.”
That is a serious rule-of-law problem.
Jurisdictional differences
In Australia, appellate courts decide by majority. A 3:2 High Court result is legally valid. The dissent does not, by itself, quash the conviction.
In the United Kingdom, the language would likely be whether the conviction is “safe.” A split court would not automatically resolve the matter for the appellant, but a serious dissent on reliability and miscarriage would carry substantial force.
In the United States, the question would be more fragmented: due process, sufficiency of evidence, harmless error, confrontation, and post-conviction standards. The institutional problem would remain, but the procedural route would differ.
Authority
Certain:
* Bromley v The King [2023] HCA 42 — 3:2 High Court outcome; majority dismissed the application; Edelman and Steward JJ dissented.
* South Australia’s second or subsequent appeal regime required fresh and compelling evidence capable of being considered in the interests of justice.
* The High Court minority considered the proposed fresh evidence sufficiently serious to warrant relief; the majority did not.
The broader principles are stable:
* criminal guilt must be proved beyond reasonable doubt;
* appeal courts correct miscarriage, legal error, and unsafe verdicts;
* appellate courts must not usurp the fact-finding role of the jury;
* fresh evidence may justify relief where it undermines confidence in the conviction.
Evidence Strength & Causation Limits
Evidence strength: Moderate to strong as a critique of appellate safety; not conclusive proof of factual innocence.
Causation boundary: The articles support the proposition that Bromley’s conviction may be unsafe. They do not, by themselves, establish that Bromley is factually innocent.
Critical evidentiary limit:
A mentally impaired or inconsistent witness may still tell the truth. But where that witness is central, where expert evidence bears directly on reliability, and where forensic evidence is alleged to be impermissible or fraudulent, the conviction requires exceptional caution.
Appellate-process limit:
A court may assess whether fresh evidence is legally compelling. It may not fairly preserve a conviction by constructing a substitute factual case that the jury never heard.
Practical steps
The proper LIA formulation would be:
The presumption of innocence is legally altered after conviction, but the benefit of doubt must remain operative through the miscarriage-of-justice jurisdiction. A split appellate decision does not automatically acquit; however, where the split reflects serious concern about witness reliability, forensic integrity, and the appellate court’s possible substitution of itself for the jury, the conviction should not be treated as safe merely because a bare majority says so.
Risks / uncertainties
The main uncertainty is whether the Moles/Sangha criticism of the 2018 appeal is fully borne out by the judgment text. Without testing the judgment line-by-line, LIA would not state as a concluded fact that the appeal court “abolished” the right to jury trial. That is a powerful advocacy claim, not yet a verified finding.
The strongest counterargument is that appellate courts must be able to decide statutory thresholds, assess whether fresh evidence is truly probative, and respect finality. It is not enough that new material might have assisted the defence; it must satisfy the legal test.
The strongest reply is that finality should carry little weight where the conviction depends on a compromised key witness, contested forensic foundations, and appellate reasoning said to have cured the case in a way no jury ever approved.
Recommended move
LIA’s final answer to the article’s question:
No, the presumption of innocence does not continue unchanged at appeal. But in a case like Bromley, that formal answer is not enough. A criminal appeal system cannot claim fidelity to reasonable doubt if it allows a doubtful conviction to survive through appellate fact-reconstruction, especially where dissenting judges identify a real possibility of innocence. The better rule would not be “every dissent acquits,” but “a serious dissent on safety, combined with untested or unreliable evidence, should trigger a strong presumption of quashing, retrial, or independent review.”
Checklist
* Causation tested? Yes — dissent and appellate irregularity support unsafe-conviction concern, not automatic innocence.
* Alternatives considered? Yes — finality, majority rule, imperfect-but-believable witnesses, statutory thresholds.
* Evidence strength assessed? Yes — strong systemic critique; not conclusive proof of factual innocence.”.
LIA is a Legal AI application that has been trained on Australian and international law.
Great, thanks Adrian. I should add that the question posed in the headline is meant to prompt debate – and through debate to challenge the system for reform. Of course, the one obvious form of reform to help deal with the nuances outlined by LIA is to establish a national network delivering a Criminal Cases Review Commission.
Reminder:
As Bromley’s case now stands, his original conviction was based on Dr Colin Manock’s impermissible (fraudulent) forensic evidence, his appeal was mangled by the court’s erroneous approach and the High Court dismissed his final appeal because the jury was entitled to believe a mentally ill witness. That is three strikes against the system … actually, it is four; the first appeal was dismissed, the second appeal heard and dismissed in 2022.
I need help from someone, It’s over 40 years but it’s like yesterday. Every day I asked myself who owned the 3 DNA, the lie she caught a taxi home to her Policeman, She got dressed & went out & left her work partner, Mack to close up on his own, she never walked back to were she met my self, she never got dressed @ home & had 3 ozzos. 4 people, 2 police officers were only metres from her, why wouldn’t they help her, the police officer & her married soon after, she lie about her working hours on the day & days after to the VCT which I fort & won, Why did she lie in a AVO which I also fought & won. 60 questions she would not answer or didn’t know. So much.
With apologies to Mark Twain, Australia’s only native criminal class is the judiciary.
That says it all. If 5 judges at the high court say 3 to 2 that an appeal is unsuccessful you can also say that two judges are upholding the appeal . If out of 5 learned judges , 2 vote to overturn or acquit, that’s the same as a hung jury . No conviction by less than 100% determination should constitute anything other than reasonable doubt. Acquit ! Don’t just overturn. Even if 9 members of jury say guilty and 3 say not guilty. That should constitute a verdict of not guilty. Otherwise the system will retry and retry until it gets the verdict it wants.