Max Martin has famously claimed that lyrics have to sound right, not make sense, and there's things like Ariana Grande singing "now that I've become what I really are", and plenty of very memorable English lyrics written by non-native speakers. On the other hand, if you're doing lyrical hip-hop or a Texas country song with a story and a plot, those do need to make sense.
Metalheads of a certain age remember "if you are a false, do not entry". Now that is cool, very memorable, and sounds better than "don't enter if you are false" any day. But, of course, it relies on its meaning, which comes across even with the bad grammar. Correcting it just robs it of its unintentional humor and also its vaguely Shakespearean rhythm.
In defense of meaning, I remember this from Chesterton:
So that's what I think... meaning is important, even if it's only what little part of the meaning comes across from barely intelligible lyrics. But there are also real aspects to the sound of phonemes, and accents and rhythms matter a lot. But I also think I only understand a tiny percentage of how all this works, and we could probably understand these things much better.
Metalheads of a certain age remember "if you are a false, do not entry". Now that is cool, very memorable, and sounds better than "don't enter if you are false" any day. But, of course, it relies on its meaning, which comes across even with the bad grammar. Correcting it just robs it of its unintentional humor and also its vaguely Shakespearean rhythm.
In defense of meaning, I remember this from Chesterton:
And then, this part is why I still remember something I last read over 20 years ago:Oddly enough, it is in the rational lines of Virgil or Milton, much more than in the extra-rational lines either of the Merediths or the Sitwells, that we feel the final mystery of song; the something that instantly gives delight and escapes from definition; the thing of which we say: "I cannot tell, for the life of me, why that is so good as it is." I cannot tell, for the life of me, why the line "Like Teneriffe or Atlas unremoved" is as good as it is. Yet it is perfectly straightforward; it merely mentions a cape and a mountain, and adds the somewhat superfluous information that they are not removed. The only thing I am quite sure about is that the sense depends on the sound and the sound depends on the sense. It actually would not sound the same, if another meaning were expressed by the same sound. It actually would not mean as much, if other words expressed the same meaning. It would be easy enough to try the experiment in a rough and ready way. It is obvious that, if we wrote "Like Beachy Head or Snowdon unremoved," it would not be within a thousand miles of the thing; though Beachy Head is a cape and Snowdon is a mountain. What is not quite so obvious is that the converse is also true. It might be too lightly inferred that the mere noise of the names is alone majestic. It might be even suggested that the down-rushing dactyl of "Teneriffe" has some faint echo of words like "terrible" or "towering," and that the sound is the secret. But it is not so, though the alternative experiment might be a little more elaborate to construct.
But can we analyze this more scientifically? I very vaguely remember running across some scientific research which claimed to find that across languages and cultures, people find names with "o" and "u" vowels more fitting for round creatures, and "e" and "i" vowels for spiky, pointy creatures. Now, that makes sense. Some vowels have lower formants than others (oversimplifying: warmer EQ), and also your mouth has to make a rounder shape to produce them. So, there is probably something to it. Some consonants also sound harsher than others, though I don't know if the universality of that has been researched.Let us have a stab at it, as Mr. P. G. Wodehouse's young man said when asked if he would be a reasonable being. Thackeray mentions somewhere, in one of his essays, that in some old cookery-book or book of etiquette he had come on the fact that men in the eighteenth century drank a wine called Teneriffe, apparently an alternative to port or Madeira. Thackeray says, I think, that it sounds like having to swallow the Matterhorn. But if it were something quite familiar, like port wine, it would sound like any other detail of the dinner-table. As for the word "Atlas," we have only to knock out the capital letter, and it means a commonplace work of reference, an ordinary book of maps. Now, suppose somebody were writing a very mild and jog-trot domestic poem in decasyllabics, rather like those poems in which Cowper celebrated the tea-urn or the cat. And suppose the particular passage explained how somebody's after-dinner table was left in a litter by negligent servants; books and wine and everything in a hugger-mugger-- His pipe and napkin, like his spectacles, Like snuff and toast and pen and ink or books, Like teneriffe or atlas, unremoved. It would not make the same noise. It actually would not sound in the ear, as a matter of mere acoustics, the same. The fact of talking only about two trivial objects would, in fact, alter the actual impact of the sound upon the ear and the nerves. Nobody would be looking for a great sonorous effect, and nobody would find it. The fact that the two objects are mountains, mysterious and remote and legendary mountains, does enter irrevocably into the merely physical process; and it is the largeness of those mountains that fills the lungs and the ear.
So that's what I think... meaning is important, even if it's only what little part of the meaning comes across from barely intelligible lyrics. But there are also real aspects to the sound of phonemes, and accents and rhythms matter a lot. But I also think I only understand a tiny percentage of how all this works, and we could probably understand these things much better.


