The problem

Most people are quite shocked when they discover that indistinct covert recordings are routinely accompanied by a police transcript. They worry that police might be ‘biased’ and ‘hear what they want to hear’. It’s actually quite a lot worse than that – in several ways that I will introduce quickly here and discuss in more detail later.

But let’s get one thing straight from the start

This is not about police being ‘biased’ in that negative sense. For the record, in my dealings with police over the past 20 years, I have encountered only a tiny few who intentionally misuse transcripts of indistinct recordings.

The problem is not the police; it is the law

The problem isn’t that police are biased, but that law makers (despite their impressive skills of rhetoric) have poor understanding of the nature of speech, and how speech perception works.

You'll find our quick but fascinating movie at the bottom of this page
In case you missed it, this 90sec video gives an impression of the power of priming

Police are human. All humans hear indistinct audio in a way that is unconsciously influenced by their assumptions and expectations. That is not a ‘bias’ you can turn on or off at will.

Defence transcripts are just as much affected by this kind of ‘priming’ as police transcripts are.

The only reason to focus on police is that police transcripts are used routinely, admitted in criminal trials on a weekly basis. That’s because police transcripts are given a special status by the law, based on the erroneous idea that detectives’ ability to hear words where others can’t comes from listening many times.

The role of knowledge: a double-edge sword

The reason police hear indistinct covert recordings more confidently than others do is not that they listen many times. If it was, it would be far more cost effective to hire professional transcriptionists to listen many times.

The real source of detectives’ ability is that they have expectations, derived from their knowledge of the case, about what they might hear. And it is important to recognise that this contextual knowledge does give police genuine insight, which, used with great care, in light of good understanding of phonetic science, should be taken appropriate advantage of in the legal process. After all, police knowledge is what helps them catch crooks – and we all want them to do that.

The problem is, when it comes to transcripts, knowledge and expectations are a double edged sword, which can mislead as easily as assist.

The experience of being misled by unreliable assumptions is indistinguishable from the experience of being assisted by reliable knowledge (a fascinating fact that we go into a lot more elsewhere on this site, especially under CAUSE in the top menu).

Police transcripts are bound to have errors

For these and other reasons, it is inevitable that police transcripts of indistinct covert recordings will have substantial errors, omissions, and other issues, regarding both what is said and who is saying it. These errors are most likely unrecognised by the officers making the transcripts.

The issue is how to let the courts benefit from the reliable aspects of a police transcript while avoiding being misled by the unreliable ones. That’s an important and substantial issue, which makes a valid and useful project for cooperative research between phonetic science and the law.

Defence lawyers, even defendants, are bound to be influenced by the police transcript

The problem is that our current law, developed on the basis of common knowledge, allows** evaluation of police transcripts to be left as a matter for the jury. Which effectively means it is a matter for defence lawyers.

That takes no account of the well-established findings of phonetic science that show how strongly an inaccurate transcript can influence what you hear in indistinct audio, and how hard it is to ‘unhear’ what you have once heard, even when you know for sure it can’t be right (check some examples under CAUSE above).

It also takes no account of how difficult and time consuming it is trying to check transcripts of lengthy indistinct recordings. For both these, and other, reasons, defence teams are highly unlikely to notice how much the police transcript is influencing their perception.

The upshot is: our current legal practice is not fair on the jury, let alone on the defendant.

Maybe this brief background is enough to help you see the crux of the problem: police transcripts are highly likely to include misleading inaccuracies of various kinds; and these inaccuracies are highly unlikely to be detected by the defence or the jury.

Where to from here?

You might now like to look at the Problem in a bit more detail. Or The Cause gives examples (some rather entertaining) to show how speech perception really works (very different from the assumptions of ‘common knowledge’ relied on by the law). With that under your belt, you’ll be ready to delve into the Case Study, which outlines in detail how reliance on a police transcript resulted in potential wrongful conviction in a real murder trial. And when you are ready for more – we have an extensive Learn more section. All of that available from the top menu.

Or check out the right sidebar for some quick introductions and overviews.

**It is important to note that our law does not insist that evaluation of transcripts be left to the jury. There are plenty of precedents for judges intervening, for a range of reasons and in a range of ways. The point is that the law allows evaluation to be left to the jury, and legal practice makes this common.

Sign up here for occasional news about Forensic Transcription (no spam!)