The dark side of mondegreens: how a simple mishearing can lead to wrongful conviction

We are all familiar with mondegreens. Even if you don’t know the word, you’ll have had the experience. Mondegreens are funny mishearings — famously of song lyrics (think: “there’s a bad moon on the rise” misheard as “there’s a bathroom on the right”).

Similar things occur with everyday speech (it’s what turns for all intents and purposes into for all intensive purposes or an all-staff email into an all-star female).

Hearing things that aren’t really there even extends to random stimuli. Some of you might recall the scandal of the McDonald’s cussing Minion toys. They were speaking Minionese (nonsense prattle), but outraged parents heard swearwords: “well I’ll be damned” and “what the fuck”. Hundreds of others who clicked on the link heard them too.

And the more they listened, the clearer these obscenities became. Once you’re told what to listen for, the brain is primed – and it’s pretty much impossible not to hear it.

Mondegreens can be hilarious. But they have a darker side too.

Revealing the darker side of mondegreens

These days, many criminal trials feature evidence in the form of covert recordings (conversations lawfully captured by police without the knowledge of the speakers).

Covert recordings can be of extremely poor quality — and, as with those incomprehensible song lyrics, it’s often hard to make out what is being said without the assistance of a transcript. But what if the transcript is misleading? What if it “assists” listeners to “mondegreen” an innocent remark into an admission of guilt?

Sound far-fetched? You might be surprised to learn that it has happened numerous times.

Amazingly enough, it is actually encouraged by everyday practice in our criminal justice system.

How can that be?

A High Court ruling made in 1987, and used since as a precedent in countless trials, allows indistinct covert recordings to be accompanied by a transcript.

That makes perfect sense – until you realise who creates the transcript.

Police provide their own transcripts, in the role of a so-called “ad-hoc expert” – someone with no real expertise but deemed to have “specialised knowledge” arising from familiarity with a particular case.

Transcription of indistinct audio is a highly specialised skill in which police have no training whatsoever. Even with the best of intentions, police transcripts are liable to be inaccurate and misleading.

Nevertheless, current Australian practice allows police transcripts to be given directly to the jury.

There are safeguards. The judge has to instruct jury members that they should listen carefully and form their own opinions of what is said, using the transcript only as an aid. The problem is that is not humanly possible.

Everyday experience, or an encounter with those potty-mouthed Minions, confirms research findings going back at least to the 1920s: listeners are highly susceptible to being primed by suggestions as to what they might hear. It’s perfectly possible to listen carefully yet confidently hear words that were never spoken.

It’s a special case of the suggestibility well known in legal contexts — but so irresistibly powerful that lawyers themselves get primed by misleading transcripts, even when accurate hearing would benefit their own clients.

The law, with no consultation of phonetic science, has developed a process in which detectives’ knowledge of the case “assists” them to create misleading transcripts. Their transcripts then “assist” lawyers, judges and juries to interpret the audio.

Everyone believes they are “listening with their own ears”. Mondegreens go unnoticed. The consequences can be catastrophic.

The risks of wrongful conviction

At Forensic Phonetics Australia, you can hear numerous examples of real covert recordings used in trials.

Read the case study of the murder trial where the only “direct” evidence was a single phrase in a police transcript. The phrase has since been shown, first, to be inaccurate and, second, nevertheless to exert a strong priming effect on listeners, who confidently use their hearing to draw conclusions about guilt.

Too late for that prisoner though. He recently died in prison, part-way through a 30-year murder sentence.

Or there’s the case of the young mother who spent two years in custody based on a few isolated words detectives claimed to hear in the background of another conversation. And there are plenty more to explore.

These kinds of things sound unthinkable to citizens in a modern democracy, but they go unnoticed by lawyers familiar with the practices developed since the 1987 High Court ruling.

It’s time for reform

Clearly, juries need a transcript to help them interpret poor-quality covert recordings. And, clearly, police insights should assist the process of creating the transcript. But just giving police transcripts direct to the jury? That’s not fair to jurors, let alone defendants.

We need a proper process for handling covert recordings – one that ensures a reliable transcript accompanies indistinct audio.

That won’t happen just from lawyers applying precedents. We need a full consultative process like the one that recently produced the excellent National Standards for Working with Interpreters in Courts and Tribunals.

A 30-year anniversary is an auspicious time to change a system that allows one person’s mondegreen to be another person’s jail sentence.

This piece was written with the assistance of Dr Helen Fraser, who runs Forensic Phonetics Australia. Dr Fraser is running a new forensic transcription experiment. Feel free to test your own hearing at this link: please participate in the find-the-phrase experiment.