Christopher Pyne: the c-word or the g-word?

16 May 2014

Social media claims Christopher Pyne dropped the ‘C’ word in parliament on Wednesday, but he says the word was ‘grub’. (SMH)

Huge interest the last day or two here in Oz as to whether Christopher Pyne, a right-wing politician, swore at a fellow politician in parliament.

What if this audio were evidence in a murder trial? Can proper phonetic analysis resolve debates as to what was really said, or is it just a ‘battle of the ears’? Can this storm in a teacup shed light on serious problems in the use of indistinct covert recordings in criminal trials?

Here’s the video clip and some of the social media coverage

  • The Sydney Morning Herald (Note: many of the clips on youtube have been altered in some way so listen to this one.) If link doesn’t open, copy and paste

Here’s the audio

The full clip:

The sentence:

The word in isolation:

Here’s the analysis

The entire pre-vocalic (before the vowel) part of the word is completely masked by a cough. Note that a cough can easily give the impression of a velar (back of the throat) consonant, like /k/. Here’s a waveform and spectrogram with the cough and the last part of the word highlighted.


A couple of things to notice about the end of the word:

  • all formants (horizontal dark bands) trend downwards through the vowel – suggesting the following consonant is labial (made with the lips, like /p, b, m) ; if it were alveolar (like /t ,d, n, s) we would expect F2 to bend upwards from its low position in the vowel
  • the consonant burst centres around 3kHz – again suggesting a labial; for an alveolar we would expect a burst above 5kHz

These observations do not confirm ‘grub’ as a definite, accurate transcription (especially when we cannot see or hear the first part of the word). They do suggest ‘grub’ as a far more likely transcription than the c-word. (Note similar findings by Lochlan Morrissey at FullySic.)

Add to these observations the fact that no one in parliament at the time reacted to Mr Pyne’s epithet in a way that suggested he had ‘dropped a c-bomb’. Sadly, one parliamentarian calling another a ‘grub’ attracts no more than the mildest notice, but using the c-word would surely raise a comment from those nearby (that’s why it is called a ‘bomb’). The whole c-word drama arose from someone listening to the audio after the fact.

So why are so many people so sure they hear the c-word?

First, this is another classic demonstration of the power of priming. Note that priming doesn’t require explicit suggestion of the word. Note also that our ears are generally primed to hear rude words. I quite often get startled by the f-word, only to realise the person had innocently said ‘If I can just …’.

Second, it is a classic case of the persistence of priming. It is genuinely hard to un-hear something you have ‘heard with your own ears’ – even when what you think you heard is demonstrably inaccurate – especially if (and ‘hearers’ of the c-word will hate me for saying this but if you can get over that, it is actually really interesting) what you think you heard fits a preconceived notion you hold. It is striking how many people who claim to really hear the c-word are those that just don’t like or trust Christopher Pyne – a similar observation was made in responses to the Trayvon Martin/George Zimmerman audio.

What if this audio were evidence in a murder trial?

This is a matter of passing interest, but similar issues really do affect criminal trials on a weekly basis.

Thanks to radio 2UE Breakfast for bringing this topic to my attention. You might like to listen to this short interview with their John Stanley and Garry Linnell about it (shared with permission).

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