[Originally published 9.12.09]
Lest anyone say that linguists do not bother themselves with burning issues, the kind which is relevant to the daily lives of each and every one of us: David Crystal, the respected and renowned English linguist, relays in his blog the story of a discussion which recently cropped up in a BBC radio show regarding a line from the song Live and Let Die by Sir Paul McCartney and Wings. While not exactly a pressing matter, this allows us to examine a number of tools the linguist uses to analyze a phonetic question. For starters, let us have a look at the song's first verse as it appears in the liner notes:
When you were young
And your heart was an open book,
You used to say "live and let live".
But if this ever changin' world
In which we live in
Makes you give in and cry,
Say "live and let die"
The emphasized line is not entirely grammatical due to the repeated "in". In order to reconcile this abomination, the BBC hosts suggested that perhaps McCartney' had sung In which we're livin' and had dropped the final g, as is so often done in spoken English. However, Crystal claims that this explanation is unsatisfactory and that McCartney would never have done that; in other words, the line is sung as it appears in the liner notes. His reason: McCartney is a Liverpudlian and the local dialect, Scouse, does not omit the final g. Crystal gives the Beatles' All My Loving as another example of a song where the final g is kept. So if we follow Crystal, McCartney sang In which we live in and not In which we're livin'.
Now, I am in no way Crystal's match in matters English, but a few problems arise from his analysis:
- If we go by simple listening, I am not sure which of the two versions I can "hear". One could have had a clear hint about the rest of the sentence owing to the word we or we're, but it is very difficult to tell which of the two options is the right one due to British English's non-rhotic character. For your convenience, the video:
- If we take All My Loving as a benchmark, I definitely hear the g being dropped (or rather, do not hear the g at all). During the entire song, over and over and over again. And then once more in the refrain, which McCartney sings by himself. Here is the song:
I hear one thing but Crystal hears another, plus he has a dialectal analysis on his side. I turned to the wonderful (and free) acoustic analysis software Praat. Praat means 'talk' in Dutch, and the software can show the soundwaves constituting an audio file and even identify phonetic attributes such as pitch and amplitude (loudness). I first tried checking All My Loving to see if the elusive g is indeed there. The division into letters and syllables is mine, and what we are interested in is the light blue bar which Praat identifies by itself. This is the pitch, the tone, given as frequency in Herz. Trusting the annotation Praat creats automatically is a risky habit, but it seems to be valid in this case; note how it corresponds to the darker parts of the wave. Either way, the higher the blue bar is, the higher pitch. The lower the blue bar, the lower the pitch (i.e. a deeper voice).
What we can see is that at the end of this line, there is no additional, lower sound after the n. Presumably that is where the g would have been, but this is not the case. It is worth emphasizing once more that Praat's analysis is far from perfect, and one could also claim that the g is so deep that it does not even appear on this scale. But if that is the case, it means the consonant had been dropped.
Assuredly, this is by no means conclusive evidence, but I am starting to feel better about my suspicion regarding g-dropping in McCartney's songs. Moving back to the song that started this whole ruckus, we want to check two things: on the one hand, does McCartney sing live in or livin'/living, and on the other hand, does he sing we're or we. We concentrate on the light blue bar once again:
It seems to me like we can rule out the existence of a g with some certainty, but that still does not solve the problem with the sentence itself. It might be In which we live in and it might be In which we're livin'. So we check the third word of this line, sprinkling a few red dots this time around:
These red dots signify the formants, the resonance of the sound, a tricky concept which I will try to explain briefly. They are useful in our case, since according to this breakdown I must divide the vowels in the spectrogram in a way which implies two vowels, and in effect a non-rhotic we're.
A formant can be coneptualized as the frequency of the resonance in our mouth. Every vowel creates some resonance and it is usually possible to identify 4 different formants in ordinary speech. Roughly speaking, just like light is comprised of a spectrum of frequencies (in effect, the colors), so is a sound a spectrum of frequencies. The most prominent frequencies, shall we say, are the formant, and the beauty is that you can differentiate different vowels by their formants. If we think about the oral cavity again, it is possible to tell the vowels apart according to what happens to air inside the oral cavity when they are pronounced. For a more robust explanation, consult your favorite phonetics textbook or even Wikipedia.
What is all this good for? Formants are ordered along horizontal lines according to their frequency. The first formant is the lowest bar, the second is the one above it, and so on. In our spectrogram we want to ignore the noise and take note of the two thickest arrays of red dots, one just above the light blue line and one directly on top of it. In the segue which I marked between the e and i (while abusing notation in the image), the upper formant rises a bit, and this is precisely the difference between the two vowels! Furthermore, in the e, the formant just beneath the two we have discussed begins a downwards decline. This is a feature of the English non-rhotic r.
All things considered, it would be a bit silly to posit the existence of a diphtong merely based on this analysis, yet it does seem as though there is a we're, rather than a long we. The last two analyses lead me to conclude that McCartney did in fact originally try to sing In which we're livin'.
The Beatle himself says he does not remember what that line was supposed to be like, but in his opinion, the first version (In which we live in) "is cuter." David Crystal, an expert on English, makes use of his acquaintance with various dialects, but I disagree with him based on a quick and dirty phonetic analsis. On the one hand we have a dialectical claim, on the other hand we have more "technical" phonetic tools, and both methods assist the analysis, in the most part because neither can provide conclusive evidence. So much for my findings. If any of our readers are phoneticians, make yourself heard!
[Acknowledgments: Thanks and apologies to Justin Fitzpatrick, who taught me to use Praat but had probably hoped I would use it for more serious research. He also tried to teach me about formants with limited success]
But, but… there isn't even supposed to be a g there. It's just supposed to be an engma instead of a regular n. As far as I know, no English speaker pronounces an "entire" n and then an "entire" g at the end of a word.
It also says so in scholarly papers, in a clarifying paragraph at the beginning, that g dropping doesn't actually have any dropping.
(And another point, just something that seems odd – they claim there isn't dropping in the Liverpudlian dialect, but exactly one line before the line in question we have changin'… What's going on here?
I'm not sure your method is correct. There isn't supposed to be a change in pitch when pronouncing the g, because both consonants are pronounced together, even when the g isn't dropped.
There's another way to decide: Take a place where you know there's no g (that is, the syllable in at the end of a word. Best if it's stressed the same way) and see if it's similar in terms of formants and frequency to the pronunciation at the end of this line. I'd also suggest taking for comparison a place where you have ing at the end of a word, but according to what you've described, there's no such place.
I don't know anything about phonetics and frequency analysis, but if we look at the relevant literature (wink), I'm pretty sure Guns N' Roses perform the song with In the world in which we live in.
(from the album Use your Ilussion)
I thought he was dead.
@YP, Well, I write for Dagesh Kal, I don't write for David Crystal's blog. I agree that some of his assertions are bizarre, but all I can say is that when I spoke with Scousers, I never noticed whether or nor they pronounced a g in the coda. Probably because I'm no phonetician. From Crystal's post:
On the contrary, -ing is often said with the -g sounded as well, in that part of the world. While it's always possible to 'drop the g' in rapid colloquial speech, as it is in any accent, this is unlikely in the more forceful articulation of a song whose beat is relatively slow. I don't recall other Beatles songs with -ing endings – such as 'All My Loving' – reducing the final consonant.
@O-2, If both consonants are pronounced at the same time and there is no difference between them, then the g had already dropped. If there is a difference between the two, it ought to be reflected in the height of the sound. Concerning in, you can indeed find that syllable in the same line, but the differences in pronunciation are pretty big.
@O, I'm guessing Guns N' Roses followed the official liner notes.
@A, Beatles-Paul is dead, but Wings-Paul is something else entirely :-)
And I, unlike @O, am sure that Guns N' Roses sing In which we're livin, not to mention Paul himself.
In any case, this blog is wonderful in itself, and a post which combines The Beatles and language is a real cause for celebration. Thank you!