Note: This article first appeared in George Martin┤s book ┤Making Music┤Ł in 1983, which is now deleted and long out of print
The way I wrote songs ten years ago was very different from my current method of working. After years of writing, I have only in the last two years developed a system. I used to work
in bursts rather than slowly and steadily, but now I find that by putting in time every day - if I am in a writing period - I can get more done than if I just sit around and wait for a song to happen. One of the benefits of working like this is that one can see the entire germ of a song develop from start to finish. Working steadily, it generally takes me about four to six weeks to complete a song, but if I am not working steadily it can take from four to six months.
The writing process
I work with my guitar and a legal pad and use about 50 pages to develop a song. I get going fairly early in the morning, because my mind is sharp, and start by dating the pad and putting down personal comments, such as how I am feeling that day, so that it becomes a diary of sorts. Slowly, a song will begin to emerge although sometimes it will stagger along, day after day, making no progress at all. The first page might have all sorts of lines that will never be used, but as I turn the pages, a little thought might come forward and suggest potential for development.
These notebooks are the clearest record I have of how my songs developed lyrically, I don't keep musical records, although I wish I had for the earlier material, but because they were written on scraps of paper, there isn't much left of them today.
I try and put as much information as possible into a song at the very beginning, and my first step ii to write out the lyrical structure, If, for example, the song is a simple A, A, B, A - verse, verse, bridge, verse - and the first verse has eight lines, I will write down one, two, three, etc., and put in the title of the song as the eighth line, if that is where I want it to fall.
I already know that the next verse is going to have the same structure, but the title probably won't go at the eight spot because that would be predictable. Instead, the song might end with the title, so I will put it in as the last line and maybe put it in the second verse as well, knowing that it will be thrown out later on.
If the title is interesting and evocative, I will use it as inspiration or include it within the song's structure by pencilling it in certain places. For example, there's a song I am working on at the moment called Train In The Distance in which one line repeats - "everybody loves the sound of the train in the distance, everybody thinks it's true". Because a train in the distance is an interesting metaphor for hopes and expectations, it will be used throughout the song. Sometimes however, the title is simply a title, as was the case with Late In The Evening. If I could have found another title, I would have: I would have done anything not to have a song called "Late in the Evening", but that was just the way it fell and I didn't try to fight it.
If a theme appears in the first verse, I will make a note of it. I don't notate or write on music paper, but I am always thinking about the structure of the song. For example, I might take a phrase that was in the opening, or the title, and use a part of it later on, or invert it, or use it in the bridge. It might also be used in an unexpected way, because the listener is usually so familiar with song structures that most of the time he or she knows, consciously or subconsciously, that a certain verse will be repeated. This works both for and against the songwriter; there is pleasure in receiving what you expect but the weakness is that it can be boring. However, in general my aim is to deliver what people expect because, if a familiar structure is set up and then changed to something unfamiliar, it can be disturbing or annoying to the listener. If a writer arbitrarily sets up a structure and then breaks it without cause,
the song tends to ramble and become incoherent; it is also more difficult to write a song like that.
One of the most satisfying achievements - if you can make it work - is when a song follows a circular route and ends up back where it should have been at the beginning, but one plane
higher, like a spiral. You cover ground and when you return home it's familiar, but you should have covered enough ground so that when you return to the beginning it's not the same in feel, title or melody,
Bridge Over Troubled Water
The way Bridge Over Troubled Water developed reflects my old style of writing. It was written in 1969; I had been listening to a lot of gospel music which had a great influence on me, especially the gospel quartets. A lot of pop singers grew out of these quartets which were actually the precursors of "doo-wap" rock 'n 'roll. They had their heyday in the late 1940s and early 1950s with groups like The Dixie Hummingbirds, The Swan Silvertones, and Sam Cooke and The Soul Stirrers.
One song that particularly impressed me was Oh Mary, Don't You Weep by The Swan Silvertones. There's a brief moment in the song when the guy sings a scat line and says something to the effect that he will be someone's bridge over deep water, so I'd have to say that Bridge Over Troubled Water actually came from that.
While working on the song, I was fooling around with two different approaches. One was sort of folkish and, although 1 knew it was not going to be kept in the song, it filled up some
space. The other approach had to do with diminished chords. I wrote the song on guitar, but knew I was going to use some kind of gospel piano and, similarly, although I wrote it in the key of G, I knew it was going to end up in the key of D or E-flat.
I don't remember much about actually writing the song - it just seemed to happen suddenly. The lyrics and melody grew simultaneously and the first two verses were written pretty quickly for me, because I'm generally a slow writer. Bridge Over Troubled Water was a very simple melody, although harmonically it wasn't straight ahead.
Song About The Moon
A song which is not put out yet, called Song About The Moon, is typical of my present method of writing. Musically, it began with the old Sam Cooke song, If You Ever Change Your Mind. I was playing along with the record one day and started making chord substitutions, changing the harmony and then the melody, At a certain point I stopped and said, well, this is no longer a Sam Cooke song; it's different, and I like it.
I kept singing the title of my song, and hated it, I kept thinking: I wish that phrase would get out of my head, it means nothing to me and it's leading me nowhere. It must have stayed around for months, and I would sit and scratch on my pad "Song About The Moon" and nothing would happen. Then, one day, it just began to flow and finished itself.
I was sitting in my apartment, looking out over the park and I wrote, "If you want to write a song about the moon, walk along the craters of the afternoon." That was good because it superimposed the topography of the moon on to the topography of New York City. I saw the craters in my mind's eye as if they were right there in the middle of the streets and buildings, So I wrote, "If you want to write a song about the moon walk along the craters in the afternoon." "Of the afternoon" implied that the afternoon had craters, which was too arty.
Having used the word "moon", there was the most typical rhyme going - "moon", "tune" - so I had to be very careful, Next came, "when the shadows are deep and the light is alien". Now, because you rarely hear the word "alien" in a song, your ears have to tune in, which is good, and there were the two elements - the moon and the streets - linked as well.
Then came "and gravity leaps like a knife off the pavement"; the two things were still linked, but the violence of the city was also implied. Next came, "and you want to write a song about the moon, you want to write a spiritual tune", and then I did some humming and "presto, song about the moon". In other words, what I am saying in the song is that, if you want to write, make a magical leap: you don't have to work on it; I'll show you how easy it is - just say "presto, song about the moon"; and there it is.
In the second verse I decided to use the same form of "if you want to write a song about, . ."; but what did I want to write? I always use the same subject, my favourite, and the only one which is really important in my view: the heart. So I wrote, "if you want to write a song about the heart". Because of the structure of the first verse, this beginning to the second implied that a long involved thing about the heart was coming up. Rather than do that, and create what I call the "if ever I should leave you" syndrome - the type of song, which makes you think "oh god, we have to wade through all the seasons now to find out what's going to happen" - I jumped back and wrote, "if you want to write a song about the heart, think about the moon before you start". In other words. I came back to the first verse to imply the danger of writing about the heart.
Next came, "because the heart howls like a dog in the moonlight and the heart can explode like a pistol on a June night". I was thinking about Jean Harris who had killed her lover, which was a big news story at the time, By following this with, "so if you want to write a song about the heart, and it's ever longing for a counterpart, write a song about the moon", the song was tied up because the title had been used for the last line of the verse, making the whole thing symmetrical. Because I was invoking the moon in relation to the heart, it now had the potential for being something more I still didn't know where I was going, but I had established that the moon was more than just a cliche.
I was beginning to think that the idea of the song was becoming too complex and inaccessible; that you had to know too much to like it. I kept asking myself, what was I actually talking about? What was the point of the song? Although I was enjoying it, I figured that I should be able to say what I had to say simply; it couldn't take this much to say what I had to say. Musically, I was keeping the song simple, without any flat nines or jazz chords.
Next came a little bit in the middle - "laughing boy, he laughed so hard he fell down from his place; laughing girl, she laughed so hard tears roll down her face" - but I don't know why I wrote it. When I referred back to the structure I realized I had a new element - the middle part - but I still had the moon and heart to think about and somehow they had to be tied together.
For the last verse, I wrote, "if you want to write a song about a face (tears rolling down her face), that you really can't remember but you can't erase, wash your hands in dreams and
lightning. " Then I started to use imagery that wasn't simile or metaphor; I just went straight into it: "Toss your hair at whatever is frightening. If you want to write a song about a face, if you want to write a song about the human race, write a song about the moon." And that's how it ended.
The song is about love, which is what I set out after, and so love, which is one of the most clich├ęd things around, became an acceptable subject for a song.
When I first began writing, I was really just doing college literary stuff and hadn┤t yet discovered my own voice or set of symbols. I now see that I tend to use a certain group of words
and ideas, mostly related to the body - the heart, bones, hands. I was very critical of my work in the beginning, but not sophisticated enough to know that what I was doing was naive. I hadn't lived enough or been exposed to enough to realize that what I was saying wasn't new. It was new to me because I hadn't said it before, but it wasn't new to the world.
I write almost exclusively from personal experience With other songwriters I can sometimes see where a thought came from, but why a writer will choose a particular phrase, why it struck him in such a way that he knew there was a song there, is inexplicable.
I think most songs should be written in the vernacular. There is, however, some good news and some bad news about this the good news is that it's simply the way we all speak; and the bad news is that it's ... simply the way we all speak! So there is a problem that has to be dealt with. To get around it, I will sometimes break the vernacular by using a word that wouldn't normally appear in a song, as I did with "alien" in Song About The Moon.
Odd as it may sound, I don't think poetry lends itself to song. We are all so accustomed to hearing popular songs sung the way we speak that poetry simply doesn't sound natural to our ears.
Editing your own words is a significant part of writing songs: the writer must edit all of the time. On the other hand, I'm more interested in hearing what people feel and in their mistakes than in their editing ability.
I had a title of The Late Great Johnny Ace going around in my head while I was still working on the Still Crazy album. At one time I was going to write a play about Johnny Ace and John Kennedy; then when John Lennon was murdered he became the third Johnny Ace. I decided I still wanted to write a play, but would try first to edit it down, to write it as a song and put all the thoughts into the song, When I had finished, I said, well, I guess it's all been done in the song, so I don't have to write the play.
The music I love is the music of my early adolescence guitar and vocal, four-part harmony groups on the corner and Southern rock 'n 'roll which I used to listen to on the radio.
If you play five melody lines for someone with a certain amount of taste and musical sophistication, they will probably pick the most musical. What I'm interested in, however, is what comes out of someone's heart when they sit down at their instrument or use their voice to pour it out.
With my own music I have discovered that it is the stuff that comes from the heart and not the clever things that works best. If I listen to my old songs, I can see the intellectual choices that I imposed on the music, which is the trap of using technique, as the chisel to create the piece. You chisel into your creative impulse with a technique and you might, for example, decide that the chisel will be 10/8 time becoming 6/8, as I did in How The Heart Approaches What It Yearns. I did it because the melody and the chords were so simple that I thought I'd better make them tricky, but trickiness can get in the way, which is why I don't think computer thinking is a substitute for creativity.
In developing the musical side of a song, I pay attention to the song's essence: whether, for example, its a rhythm pattern or a melody. Late In The Evening is a rhythm song and has to
do with words and rhythm rather than the melody, which is essentially the blues. If the rhythm pattern is the essential piece of information, then my first reaction is that the rhythm must be varied for interest. My second response is that, if the rhythm pattern is the thing I want to vary, it might be varied with the melody.
So there are these three elements I must deal with - melody, rhythm, harmony: the natural inclination of a song decides which I put first. The natural inclination of a song, however, implies a deviation from it, For example, I have a rhythm song and I am looking for some direction. I might look towards harmony or melody, or I could look for another rhythm direction, but my first instinct is to look in one of the other two areas. If I have a simple harmonic structure and a particular rhythm that I like, I would rather keep that rhythm going and vary the harmonic structure, by changing key or something similar. Likewise, if I have an unusual, unpredictable harmonic structure, I might find relief by changing the time signature, or using a simpler harmonic structure in the bridge, but in another key.
All of these elements have lyrical implications. The lyrics and the music have to work synergistically for the song to be good; if the song makes a radical change harmonically, there has to be a change lyrically as well. It can be, ironically, opposed to the initial change, but there has to be some recognition that something else is happening; it's a marriage, not just two elements wandering off any which way they like.
Sometimes I will write a song that feels natural from a guitar point of view but in a key that I can't sing in, so I'll transpose it or tune the guitar down a tone, as I did with The Late Great
I have gone through different phases in my music writing. There was a time when I used a little exercise - incorporating all of the 12 notes in the chromatic scale - to get me going. In
serial, or 12-tone, writing the writer had to use all of the notes before he could repeat a note, so the centre key was blown away. but what I was doing was more for the fun of it, I got the idea from Carlos Jobim. I used to analyze his music, and one day I realized that he was using every note in the scale. He came into the studio once and I said I admired his music and asked how he had decided to use every note in the scale; he had no idea what I was talking about: it had somehow happened naturally with him. However, I used my exercise as a means of getting the song going, and it can be a great help as long as it isn't the motivating force.
I used this technique for a while, but I don't any longer because I am going back to simpler melodies. Originally I moved away from the simple songs because I thought they were too simple. Now that I am going back to them, I have to keep reminding myself that it's all right to write a simple song.
The most beautiful tunes are the simplest, but they have all been written, at least that's how it feels when you're writing one clich├ę melody after another. Before, I was trying to write melodies that never previously existed, and within the context of popular music that's almost impossible because we are talking about vocal music to be sung by untrained voices, so the range isn't that great; you can't make really big leaps in the singing.
There are piano clich├ęs as well as guitar clich├ęs, and I don't advocate one instrument over the other for popular music. The guitar dictates a certain kind of melody, and pianistic songwriters are different from guitar writers. Even though I use a guitar, the way I compose is closer to piano because I am working with bass lines and leading tones a lot, and I know
where I am going harmonically; I am not just strumming. My hand always wants to go to certain chords that I love, and certain hand positions just lay themselves out perfectly on the
guitar. For example, if I can get an open E string on the bottom, I will be happy, so I try to get into the key of E if I can, or A, or something similar.
Song writing today
A little while ago I was watching a music show on television. A lot of the ideas seemed silly, but then I thought, they might be silly now, but in a few years these ideas are going to be much more clever and sophisticated. It's like any artist's early work: in the beginning it is usually naive, and slowly it becomes more sophisticated. Today, someone with some talent might be involved in video, which might strike many people as unusual. But in the future, people won't find it remarkable at all.
An interesting problem for many songwriters is what happens to them when they become too good for commercial taste. The age of the average listener - the record buyer - tends to remain the same, or perhaps drift up a bit and so the writer finds himself in a situation where, as he grows older, he is dealing with a much younger audience. I once read an interview with Hoagie Carmichael a year or so before he died, in which he was asked what he would write today if he were still writing. He answered that he was still writing, but nobody wanted to hear his music: in fact, those were some of his best songs.
Each new group of 16 or 17 year-olds that wants to listen to, or make, music has to start at the beginning, and I guess that's the fun of it. But this also means that people who are more mature are not going to be able to speak to them because they're not interested in the pain of growing up as a 16 year-old. For example, I am just interested in the pain of growing up at my time in life, so there inevitably comes a point at which the record market ceases to exist for certain artists, something that has already happened to a lot of people from the 1960s and 1970s; they are simply not around any more.
The pity of this is that in the popular arts we tend to discard people and ideas very easily. The public doesn't expect a popular artist to mature, and consequently they never become anything other than aging popular artists. I read a review of the work of a famous songwriter from the 1970s a little while ago and the critic was basically saying that the writer was a leftover and it was time for him to move on and make way for the next guy. I thought this was terrible, because there usually comes a time when the public stops paying attention, and it is usually at that point that the songwriter has just learned how to say what he or she wants to say.
The songwriters who tell us what they are thinking and feeling give us all a way of knowing what our generation thinks. It's very important that they keep saying it; it's not very important
that it should be in the charts. I find that the more popular I become, the more difficulty I have in getting people to look at my work seriously. I feel like saying, look, there's so much that's inhuman on this planet, and I'm saying something, so listen and tell me what you think. That, in my view, is what writing songs is all about; and that is the response that a song should generate.