The Easiest Way To Improve Your Lyrics

Probably my single best piece of songwriting advice.

Lyrics are the first thing most people notice, so more than any other aspect of a song, they really need to be perfect. Many great pieces of music have been dragged down by poorly executed lyrics, and while there are, of course, many factors to a great lyric, there's one often-overlooked aspect that regularly causes unprepared songwriters to lose control: The shapes of the words themselves. It sounds simple, and honestly it kind of is, but trust me, really understanding those subtle nuances of sound is an absolute game-changer, and once you know how to think about it, it's one of the easiest ways to make your lyrics better.

And thanks as well to Henry Reich, Jon Hancock, Gene Lushtak, Eugene Bulkin, Oliver, Adam Neely, Dave Mayer, OrionWolfie, David Bartz, CodenaCrow, Arnas, Caroline Simpson, Michael Alan Dorman, Blake Boyd, Charles Gaskell, Tom Evans, David Conrad, Ducky, Nikolay Semyonov, Kenneth Kousen, h2g2guy, Andrew Engel, Peter Brinkmann, naomio, Alex Mole, Betsy, Tonya Custis, Walther, Graeme Lewis, Jake Sand, Jim Hayes, Scott Albertine, Evan Satinsky, Conor Stuart Roe, ZagOnEm, רועי סיני, Brian Miller, Thomas Morgan, Serena Crocker, Adam Ziegenhals, Mark, Amelia Lewis, Justin St John-Brooks, DialMForManning, Andrew Wyld, JD White, Graham Orndorff, gunnito, Foreign Man in a Foreign Land, Kyle Kinkaid, Tom, William Christie, Joyce Orndorff, Isaac Hampton, Mark Mitchell Gloster, Andy Maurer, William Spratley, Don Jennings, Cormag81, Derek Hiemforth, Bryan, Mikeyxote, Milan Durnell, Dan Whitmer, Thel 'Vadam, FAD3 Chaos, Michael Morris, Bill Owens, Martin Romano, George Burgyan, Marc Testart, Carlfish, Matthew Soddy, Flavor Dave, Alin Nica, DraconicDon, Megan Oberfield, morolin, An Oni Moose, Ken Birdwell, Blue 5alamander, Cliff Hudson, JayneOfCanton, Ethan Savaglio, Robert Bailey, Deirdre Saoirse Moen, juneau, Sina Bahram, Ira Kroll, Patrick Minton, Justin Katz, Roahn Wynar, Chuck Dukhoff of The Stagger Lee Archives, Bob D'Errico, Robert Shaw, Donald Murray, David Shlapak, JD, Rennie Allen, Travis Briggs, Claire Postlethwaite, Greyson Erickson, Matt Deeds, Jordan Nordstrom-Young, Strife, Brian Covey, Miles_Naismith, Jay Harris, Sean Murphy, JasperJackal, Tommy Transplant, Wolfgang Giersche, Olaf, Colleen Chapman, ParzivaLore, Gil, d0d63, Jon Purdy, Ken Brown, Colin Kennedy, The Mauses, christopher porto, Billy Abbott, William Wallace, Karel P Kerezman, Ted Trainor, mightstill, Nick Loh, Randy Thomson, rpenguinboy, Antarct, Mikaela, Erika Lee, Vinayak Nagaraj, sandra zarbatany, Aenne Brielmann, Emma Finch, Eric Berg, Michael Dmytriw, Brian Gregory, Klaus, and robert johnston! Your support helps make 12tone even better!

Also, thanks to Jareth Arnold!

[tone]

Some additional thoughts/corrections:

1) Honestly if you remember discourse around Unconditionally, there's a good chance it's not about the pronunciation thing, it's about the… questionable staging choices she made while performing it at the AMAs, but that's why I said "initial" conversation in the intro. That aspect of the song's controversy wasn't relevant to my broader point here so I left it out.

2) Technically iambic pentameter isn't a poetic form, it's a metrical pattern used in various poetic forms, but for the point I'm making that distinction isn't relevant enough to justify a more precisely structured sentence.

3) Encore is actually also an example of a more advanced form of prosodic dissonance, emphasizing the second syllable of the title word in order to draw out the apparent rhyme with the rest of the line. If he'd instead accented the first syllable, the rhyme structure wouldn't work. (First-syllable accent is the pronunciation I'd expect, but I couldn't find any interviews of him saying it so I don't know if this a dialect thing or a musical choice.)

[mitori]

I solved the issue of writing good lyrics by just making instrumental music instead 👍

[Fewkulele]

One thing I love about Weird Al Yankovic: His prosodic alignment is frequently better than the artists he's parodying, often making his lyrics more catchy/memorable than the originals.

[akmadsen]

Linguist here, mainly specialized in phonetics. Schwa [ə] is defined as a mid-central vowel (referring to height and place of articulation) and while there's room for _some_ variation, it's not at all at the level you describe around 5:50. Schwa isn't a blanket term for unstressed vowels but rather a specific vowel. In broader phonetic transcriptions its use may be somewhat lax, but that's really the case for _all_ symbols in a broad phonetic transcription since we're doing away with much/all of the allophonic variation in exchange for more consistent and clear transcriptions. Generally speaking though, [ə] is a pretty well-defined vowel sound. It's true that it's typically unstressed in transcriptions of English but that's in part because linguists will, by convention, try to use some of the close-by mid-central vowels for stressed syllables to avoid ambiguity. It's not that schwa is inherently incapable of being stressed.

[joeldcanfield_spinhead]

The three things every aspiring songwriter I work with does, which are also the three things I first I warn against:
1. Speak like Yoda do not
2. PuTTING the acCENT on the wrong syLLABle
3. Mixing rhymes: none, with internal/imperfect, with perfect. Pick one, any one, and stick.

[robertmyers]

Since I grew up in the 70s, my go-to example for this is Stevie Nicks putting the stress on "washes" in the second syllable in the song "Dreams" ("when the rain wa-SHES you clean you'll know"). The lesson from back then was that you had to be Stevie Nicks to get away with it and you could really only do it once.

Great topic and thanks for the term "prosodic dissonance" to finally give me the right way to talk about it.

[TheHopperUK]

I have a real fondness for lyrics that don't sound too forced to fit the meter. 'Human' by the Killers has an awkward line in the chorus, but the verses are fantastic, so natural. 'Sometimes I get nervous when I see an open door' is so good.

[guystreamsstuff]

I'm in france and I've sung in many heavy metal bands here that have tried to write lyrics in english and prosodic dissonance is a really common problem. "But it rhymes!" is not enough, it sounds weird and unenglish.

[socialcontracttheory]

"My dad's favorite band, Counting Crows"
jesus, that made me feel old

[bigpicturehero]

When the prosodical stress matches melodic or rhythmic stress, the music speeds up. When it doesn't, things slow down and become less singy and more talky. Pop and folk and rock take advantage of this all the time, but rap takes it to the next level. A rapper's flow is essentially their personal style of going on and off the beat.

[Packbat]

14:50 I appreciate the cleverness of the self-deprecation there but I think you're doing a really good job with your videos and I really appreciate them.

[Intabih]

UM-brella, umbr-EL-la, UM-brel-LAH-ELLA-ELLA.

[mizoik]

Love me some linguistics x music,

Edit:
GATHER HERE LINGUISTS WE ARE TALKING ABOUT SCHWA

[actuallyasriel]

I first noticed this when I got into Vocaloid music and noticed that a lot of English covers have pretty bad misalignment, and it's been a bugbear ever since. And you basically just summarized it in the first two minutes. :P

[ericherde]

5:38 Common misconception. Schwa actually can be stressed. The misconception arose because, in RP and SSB, any schwas that are stressed are actually pronounced as a slightly different vowel, the strut vowel. Because schwa is never stressed in the most privileged dialects of English, the textbooks claim that it’s never stressed at all. However, many dialects of English, including standard American, do not have the schwa/strut distinction and do have stressed schwa (if you hold your mouth the same way for the ‘u’ of strut and the ‘a’ of affect, then your dialect is one of these, and that ‘u’ is a stressed schwa).

However, it is true that unstressed vowels often sound closer to schwa than their stressed counterparts, at least in English. For more info, see Dr. Geoff Lindsey’s video on the topic.

[notoriouswhitemoth]

A stressed schwa is in fact possible, it shows up a lot in Welsh

[zozzy]

There's also something to be said about secondary stress - something like "poisoníng" from War Pigs still stresses a syllable with secondary stress, and Katy could have chosen to do something similar for the word "Un-cən-DISH-ən-Əl / -ə-LEE, " putting metrical emphasis on the first or last syllable (if she had wanted to). Given the right context, it could be something like "unconditional-LY" to the rhythm of "so phenomenal-LY" from the refrain to Can't Stop the Feeling! by Justin Timberlake, or "UN-conditional" like the hook from Unbelievable by EMF.
Also, something additional I think works in Katy's favor is that the stressed -TION rhymes with Un-, which does take at least secondary stress.

[SplotchTheCatThing]

To me lyric writing has always been kinda like assembling a jigsaw puzzle -- sketch out a few lines so that I know what I'm trying to say, then figure out what parts of them aren't quite fitting together, and find a way to change them so that it slots neatly. And then repeat for the next few lines. Just like with every other kind of writing, unless you're actively aiming for a sense of spontaneous improvisation, the most important part of it is revision.
And beyond that, I've always found it helpful to start writing verses from the middle, and begin the writing process with disjointed lines instead of consecutive ones. That way I can fully create my rules, and figure out where my wriggle room is, before I have to start writing to those rules.

[harryleblanc]

Yes! This is one of my pet peeves. English (like every language) has a natural rhythm in its prosody, and a good lyricist leverages it to strengthen the song, rather than ignoring it.

[spky_guy]

5:30 schwa can very much be stressed, in fact, you stress them in your dialect all the time; this is a myth derived from a misinterpretation of RP Phonology, where schwa really is never stressed.

But, while there's nothing about schwa which means it can't be stressed, it is a very common reduction of any other vowel, since it's basically made by just letting your tongue rest motionlessly in your mouth. It's the easiest sound to produce, and thus, it appears often in reduced unstressed positions.

I'd recommend anyone interested watches Dr Geoff Lindsey's video on the matter :)