Friday, 5 October 2012

sometimes /præps/ is /pæps/

Richard Dawkins
credit: Shane Pope
In his recommendable blog #423 of the 4th of October, in which Jack Windsor Lewis writes about /r/ elision, he mentions a BBC radio interview between Jim Al-Khalili and Richard Dawkins, professor of evolutionary biology and author of books such as The God Delusion and The Selfish Gene, probably two of his best-known books. The interview was broadcast by BBC Radio 4 on the 4th of September.

Fairly at the beginning of this interview Al-Khalili asks Richard Dawkins: "With these two books do you think you hit a nerve or need among the general population?" To which Richard Dawkins replies: "Two different nerves I suppose. I mean The Selfish Gene is about biology, is about evolutionary biology and although at time I thought I was just laying out what was a kind of orthodoxy and it turns out that perhaps that wasn't quite right."

What is interesting here and what Jack refers to in his blog is the way Dawkins pronounces the word highlighted in my transcript - "perhaps".

Listen to the section "[...] and it turns out that perhaps that wasn't quite right."

There's no /r/-sound audible.

The 'big three' do not mention this informal variant pronunciation, although they do record /præps/. You needn't imitate the r-less variant, but should be prepared to stumble across it.


  1. I make this p'haps. The h is strong enough to come over the loudspeakers, not just earphones. Also still disyllabic. To get praps, you would first have to drop the h and let the linking r kick in (non-rhotic accents), no problem for h-droppers like me, but even h-defenders would have to suppress the h in order to arrive at praps. Now I suspect that praps has already become an established alternative pronunciation (lexicalized if you like), recognized by pronunciation dictionaries.

  2. @Sidney: Thanks for your detailed comment. I bow to the acuity of your ears. I also hear a fairly strong aspiration and my first idea was identical with your proposal; but then I listened to it over and over again and somehow the other variant became more and more convincing to me (you are probably familiar with this phenomenon).
    In short - I wouldn't rule out that it's (with a syllabic /p/) rather than .

  3. I've had a look at the recording in Praat now. Terrible quality as expected, probably been through some severe MP3 compression during its lifetime.

    There's no vowel of per, so any initial syllabicity is carried by the aspiration of the initial p. In short, it's a straight attempt at a standard nonrhotic pronunciation, but with a voiceless first syllable.

    The word breaks down as follows:

    1. p occlusion: not apparent in the extract.

    2. p aspiration (where the extract starts): very weak, just audible, about 0.03s, faintly voiced (inaudible) 3 pulses.

    3. h hiss, strong, strongly voiced, about 0.06s, 4 pulses.

    4. a vocoid segment, as strong as the other vowels in the extract, about 0.08s, 5 pulses.

    5. p occlusion, about 0.03s.

    6. p burst and/or aspiration, unvoiced and very weak and sounds t-like at first (suggesting this is also s hiss); becoming stronger and more th-like, still unvoiced (s assimilated to the oncoming th of that). Entire unvoiced hiss sequence of p-s-th about 0.07s.

  4. Thanks for taking the trouble to analyse it!

  5. One final thought. These online MP3 recordings are degraded to an unknown degree. Their purpose is to let you hear the content of a radio program without overtaxing your internet connection. They're not really good enough for judging niceties of prounciation, you can never be sure whether some fleeting sound feature was produced by the speaker or by the compression algorithm.

  6. Despite all actual or assumed corruptions and distortions by the mp3 compression algorithm I think that the pron /pæps/ (with only weak aspiration of the initial /p/) is not at all unlikely in very relaxed speech.