· topic

Principles of Songwriting

In January 2006, Dan Connolly wrote Arpeggio in D, which is both a piece of music that he wrote, and a short essay on songwriting. I chatted with him a bit on IRC about writing music, and commented that I was planning on replying to Arpeggio in D with some of my own songwriting notes. So I want to talk a bit here about how I approach the art and craft of songwriting.


The milieu that I grew up in used music as a social device. You get identified by the kind of music that you like at first, and then when you grow a bit older and more instrumental, you might end up forming a school or college band with somebody. This seems to have been the case across countries for many decades: lots of artists got started out in teenage bands. This was indeed the more popular route when I was growing up, but there was also the more middle class route of learning classical instruments, and being involved in a choir or orchestra.

Before I seriously learned any instrument, I had a romantic vision of being able to play something well. Perhaps I was influenced by Lisa's "saxamophone", but I tended to see an instrument as a means of expression first and foremost, and thought it would be really cool to be able to make music on an instrument in a fashion as versatile as whistling. On the other hand I realised that an instrument shapes your way of thinking about music. On a piano you're limited by the number of fingers you have, but you can reach pretty much any given chord in an octave. On a guitar, you're not only limited by fingers but by strings, and it's much harder to reach some chords, depending on your tuning.

At any rate, my primary instrument is the six stringed electric tremolo guitar: a regular guitar, for me. I play a Strat, more often than not, somewhat influenced by Hank Marvin and Jimi Hendrix. It's ironic, then, that my actual playing is more heavily influenced by Jimmy Page, who is generally considered to be from the Gibson stable. These are just technical details, but they're important because the instrument is often a primary focus on the craft side of things. A passion for technical details reflects well in the results.

It was difficult for me to learn the guitar at first because my fingering technique tended to be a bit weak. I compensated for this by detuning the guitar an entire step (that is, two semitones, the equivalent of two frets), to DGCFAD, and then using a capo on the 2nd fret. This has become somewhat of a trademark for me. I often see it used by acoustic guitar players, since acoustic guitars tend to be much more tensely strung and thus harder to play, but I think it's unusual on an electric guitar. I'm not sure if I'd recommend it to others, especially learners. It's probably a better idea to start learning on an acoustic guitar, uncapoed.

Again, this is a detail of the craft. Its bearing on motivation is that my songwriting habits are dictated and constrained by my primary choice of instrument. I don't say constrained in the sense of regretting it, because I think that the guitar has proven itself the most versatile instrument of the century. A jazz drummer friend of mine, William Loughborough, attributes this to the fact that a guitarist can bend notes whereas a pianist cannot: and a guitarist can play slide, but a pianist can't glissando smoothly. This is important, he believes, from a microtonal standpoint. When you listen to the music of Robert Johnson, for example, if you have a good ear you'll find that it is strongly microtonally influenced. He's only using the chromatic scale as a guide. There may be other factors to why the guitar is so popular too.

If you have a guitar, then you can do one of three things with it: not play it, play an existing piece, or play a new piece. You can do both of the last things simultaneously if you're adapting a piece. Obviously this article is concentrating on how to play a new piece, how to make your own songs up, but it's also important to think about what existing pieces you play. Nobody plays the guitar entirely as a tabula rasa: no song comes ex nihilo. Indeed, the motivation for songwriting often comes from wanting to emulate some musical hero or heroine.

As I mentioned, I generally pattern my guitar playing after Jimmy Page, though he's far from my only influence as you'd expect. When I learned how to play the guitar, I invariably learned how to play things that Page got to first. One of the biggest benefits of this was learning alternate tunings. The regular Spanish tuning for guitar, EADGBE, is very versatile in that it lets you play in many keys quite easily. But it doesn't have the sonorous droning quality, a kind of consistency that most other tunings give. My favourites are DADGAD, D modal, DGDGBD, open G (Keith Richards uses this one almost exclusively), and CACGCE. I use others too.

This spoke to my desire of versatility. Even with many years of practice on the guitar, I'm not all that versatile in technique, but I still aim for that in songwriting. On the other hand, my "aims" and my motivation for songwriting are more personal: I didn't take that route of classical music playing, and so I'm not composing written music and nor do I necessarily want to become a master of technique on my chosen instrument. But nor do I really have much of that social aspect of playing in a group and wanting to perform and eventually record albums or make money, because I'm just interested in music for the sake of music; for fun, for me.

Learning to Write

Most people appear to pick up the writing of songs once they've been engaged with an instrument for some time, but I wanted to write songs before I knew how to play guitar well. I wrote lyrics, which I guess amounts to just poetry with a beat, but I turned out to be better at writing music, which surprised me a little.

Before you can learn to write songs, you have to decide what constitutes a song for you. For me, the definition is quite muddy. One of the things that crystalises my view of writing songs is an old interview with Dylan, in which he picks up a flute and plays it for the interviewer, who's concentrating on lyrics. He asks the interviewer if that isn't music, saying that music doesn't need lyrics. I don't think songs necessarily shouldn't have lyrics, just that they can get by without them if need be. Much classical music gets by without them. Some folk music gets by without them. I don't let the normative pop format become a constraint.

As I say, some people come into songwriting because they mess about on an instrument and find something that sounds cool. Then they try to fashion it into something resembling what, for them, constitutes a song, which is often based on popular music. This happened to me to some extent, and still happens, but I was also very interested in learning how to write songs. I wanted to know what other people were doing, whether there were any guilds for it, whether there was some received knowledge, some things to read, techniques that are well known. So I set about looking for them. When that didn't satisfy me particularly, I set about listening to the shapes and structures of songs that I liked. That turned out to be much more informative.

Since I was approaching songwriting from the colloquial, non classically educated viewpoint, there was always the lure of learning what's called music in traditional curricula. That is, how scales work, what terms like arpeggio and dominant seventh and mixolydian mean. Because I tend to absorb things quite readily, I learned a lot of basic music theory just in passing, though I didn't specifically seek it out in any structured way, and I don't find it particularly interesting. But however you acquire the information, you do need to have a working knowledge of how music works.

Consider The Beatles: they're the most popular band in history, but they didn't know how to write manuscript music. They were asked in the sixties how they communicate what they're writing...

Is it true none of you can read or write music?
Paul: None of us can read or write music. The way we work is like, we just whistle. John will whistle at me and I'll whistle back at him.

They had a good working knowledge, of course, and George Martin filled in all the classical bits that they wanted. Again, it all depends on what sort of music you want to write, and your motivation.

I'm not sure how much of songwriting is learned, and how much is a natural skill. Some people seem to be able to write wonderful songs on almost their first go, but most take a while to mature to a point where people are happy with them. On this I can only give my estimation of the public artists that I know, and a few friends whose songs I've overheard. And of myself, of course: I really like some of my very first songs, even some that I didn't like as much at the time.

But I think that there are particular ways that you can go about songwriting that it's useful to know about, and I always enjoy reading people's observations on the process, even if what they say seems very divorced from what they do. For example, Paul Simon once gave what I think is a great bit of advice for lyricists, saying that if ever you're stuck, it's a good idea to start a song with a truism, and lead off from there. But when I thought about his greatest songs I figured that he didn't tend to do that, with the notable exception of Homeward Bound, which was written on Widnes railway station (sadly now demolished, with a dual carriageway in its place). The starting place of a song, anyway, is a good thing for us to start considering here.

Traditions and Themes

If you're just experimenting on the guitar and find something interesting, that's one way to write a song. But what if you want to write a song about a particular thought, feeling, event, or whatever? Or what if you just want a song that's in some way meaningful, and aren't even sure what theme to write upon?

In almost all music, love is the starting place. Something like their first fifteen of The Beatles's number one hits were love songs (was "Nowhere Man" their first non-lovesong #1?). Of course, you might say that all songs are about love at some level. Or you might say that none of them are. Or, you might say...

It's a proven fact: Most people who say I love you don't mean it. Doctors have proved that. So love generates a lot of songs. Probably more so than a lot. Now it's not my intention to have love influence my songs. Any more than it influenced Chuck Berry's songs or Woody Guthrie's or Hank Williams'. Hank Williams, they're not love songs. You're degrading them songs calling them love songs. Those are songs from the Tree of Life. There's no love on the Tree of Life. Love is on the Tree of Knowledge, the Tree of Good and Evil. So we have a lot of songs in popular music about love. Who needs them? Not you, not me.

Bob Dylan, 1991

I don't know it's a good idea to want to write a song about something if you don't know what that thing is. That sounds like you're kidding yourself, trying to overcome something. If you have to make something interesting, then it probably won't be; but if you're trying to overcome some mental hurdle, then you might get something there. My mental hurdle has never been "what do I write?" but "how do I write it?".

Both Paul Simon and Bob Dylan liked Martin Carthy's "Scarborough Fair" enough that they decided to do versions of it, both of which went on to be quite famous, and Carthy is famous now too. Carthy is British, and it seems that the folk scene in Britain in the '60s was really hopping. I think it's because the '60s had that culturally progressive and yet rooted tone, and folk music gave room for huge creativity on top of a solid and sturdy background foundation. The traditions of folk music are tried and trusted, and though a lot of it is ephemeral and cultural dreck, the really genuine stuff seems to stay accepted as genuine for longer than anything else. And because it's folk music, if you become a folk musician, you're almost expected to take some old song and make it your own, embellish it, play it to others. It's like evolution of the species, and it produces good things.

I like some folk music from the point of view of coming to it after I started writing songs, and finding that some of my motifs were already out there and had been for hundreds of years. That was exciting, because it was like rediscovery. But I can't really say that it influenced my songwriting particularly on a conscious level; the conscious connection only came years later. And as to what works on a subconscious level, that's hard to say.

Thoughts and Feelings

One of the things that I worry about having written a song is that nobody else will hear it in quite the same way. I do wonder whether other songwriters, especially commercial ones, worry about this too. Some people seem to set out deliberately to make them sound different to different people, but I don't know that for sure. I worry about it because if I like something that I've produced, I don't want the context that I have about a song to be lost, any more than I'd want you to show this article to someone one random sentence at a time. Things make less sense out of context.

How we think and feel about a song changes over time. I hear people say that they get bored with particular songs, which seems the most common way in which people's attitudes towards a song change. On the other hand, I remember someone apologising for how The Sex Pistols sound now, saying that they sounded absolutely amazing in their day because of how they usurped what went before. Now they just sound weedy, it was conceded, because that idea of usurping what went before took hold, and now we have all kinds of extremely intense music. But music is only going to be able to get intense to a limit, before it becomes noise. It's usually been a thing where people disagree over where that noise barrier is... when rock 'n' roll first came out it was considered subversive and noisy, whereas in actual fact it's a very logical progression from skiffle and rhythm and blues.

Even if you try to avoid it, if you're sufficiently talented I think all of your songwriting is going to have a personal, even autobiographical element to it. In fact, why avoid it? Unless you're writing songs for a radio advertisement or something, your songs are yours. You could have a computer generate you some MIDI according to some parameters if you want a completely soulless invention. It's easy to make harmony according to rules, to follow scales, to put together something that sounds "okay" and formulaic. If you're really talented, you can even make something sound remarkable that is yet formulaic. But as a writer I'm not into that, any more than I'm going to write a markov chained novel. Even if I wrote a random novel, it'd be my "randomness", and because of that there'd actually be patterns all over the place.

So we come to the first piece of advice: avoid formulae. Maybe it's the only piece of advice, really, with lots of supplementary clauses and corollaries. If you become robotic in what you're writing, you're no longer actually a songwriter. Perhaps that's okay, but this is an essay about songwriting. But herein lies the tension, because harmony in music, what sounds good, is all about formulae. There are only twelve notes in the chromatic scale, and they're all slightly out of tune in the even tempered tuning. That's what we use: we bust harmony out of these dozen slightly shaky tones. In a way, there's not much room for manoeuvre there, but there again it has served western music for centuries now, from Bach's Well Tempered Claiver onwards.

My definition of art is something that allows the expression of thought and emotion due to being crystalised in some path through a sufficiently huge search space such that the choices reinvoke some aspect of the thoughts and emotions that went into it to begin with. What this means is that the more complex the art form, the more arty it is. I think that this is borne out by the fact that painting is considered to be a prototypical art, whereas literature is less so. There are many more combinations of paintings that one can make than there are books; even if we're dealing with conceptually huge numbers here it seems an obvious thing: painting uses flecks of pigment which have the whole colour range to be manifest in, whereas literature has the larger words, of which there are about 600,000 in English, and of which people use at maximum about 20,000.

Music seems to fit on the more artistic end of all the artistic media, and yet twelve tones on the chromatic scale is a lot less than 600,000 words in English. The missing element is the fact that music is about not just melodies but chords, textures, timbres, harmonies, and so on. A song as written tends to be quite an abstract thing of words and chords and maybe a melody, but it can be interpreted in any one of countless ways. Bands tend to do many quite varied takes of a song if they're feeling experimental. Some don't, but the point of note is that the ability is there: music really is a wide search space to go through, and as such is on the more prototyical end of the art media spectrum.

So you have to remember that there are songs in the abstract sense of having some kind of musical score, even a very rudimentary one, and then there are recorded performances of a song, which are static and don't have any alternate realisation. These are realisations (or serialisations, or representations) of the original song, which itself might have manfiested itself in a particular recording. Often when I've settled on some outline of a song, I'll record it straightaway, and then that recording will provide the main inspiration for subsequent versions, whether I try to derive away from the original or not.

All of this, for me, constitutes the song, but I try not to put borders around it particularly otherwise one simply ends up with Plutarch's Ship of Theseus problem. It's not important. What is important depends on why you're writing songs in the first place. If you're recording an album, then your aim is to have the most compelling master possible for your album. If you're a pub performer, then you will want to work on gradually honing your performance of whatever song gets the crowd going. If you're writing songs for yourself, you just want to go with whatever's natural.

From a songwriting point of view, the ontological problem of a song is not a problem at all unless you let it be one. If you don't, for example, reincorporate and renew your old material into your current work where you think it fits, then... well in that case there would only be Tinker, Tailor, Soldier, Sailor, and no The Song Remains The Same as we know it. So the second piece of advice, certainly related to not being formulaic, is don't constrain yourself. Whilst that might even include not being constrained by the chromatic scale, I think there's some mileage in it yet.

The Writing

With all that in mind, how do you actually write songs? What steps do you take to go about it? It gets a bit hard to tender advice here. I'm not sure what drives the creative process beyond the technical aspects: I've already mentioned several times that finding something on the guitar at random is often a way that people start to write a song. But finishing it is another matter, even if you have that pseudo-random seed. I often hear the next part of a song naturally in my head. Sometimes if I think I'm just mimicing something else that I heard, I'll try to use a different chord somewhere, and then that'll take me in a whole new direction.

So techniques are often closely related to the initial parts of inspiration that you get for writing a song. Kashmir started out as a DADGAD tuning practice, and evolved into something else entirely. It's the "something else entirely" that's the natch. Shakespeare took lots of old sources for his plays, very often simply revising an old play as in the case of King Lear and Hamlet, but through his magical alchemy he came up with translunary brilliance. That's probably the thing that just can't be taught. Geniuses have it in copious amounts, and maybe we all have it a bit. For nuturing it, I think you need to put yourself in remarkable, bustling, busy, tranquil, amazing, thoughtful, reflective, progressive, dizzying, spellbinding, driving environments. Looking at the really good musicians, they all seem to have been doing things at odd cultural times. The same thing seems to be true for great scientists and writers: Shakespeare wasn't the only amazing Elizojacobean playwright. You need to pick up that something in the air.

The most useful tip I can think of in this domain is take notes. Folk songs persist because of oral tradition; it's a strong thing to train the mind that way. If you come up with a little ditty, remember it. Use it later; you don't have to use it when you come up with it. If need be, write it down. I have the bad habit of recording a song and thinking that's enough to preserve it, and then later going back and finding out that I've forgotten how to play it. I usually work it out again, which can often teach me interesting things along the way, but be careful.

Most of all, people who want to get good at things have to practice them. Well, as I've already discussed, in songwriting that isn't always true, but if you want to write good songs, you actually have to do the writing at some point. The nice thing about songs is that they don't go away. Once you've written five songs, you have five songs. Even if you're feeling a bit creatively low for a time, you can still think "well, I have five nice songs". That's the nice thing about art in general. But it's easy to get into the mindset of always wanting to produce extremely high quality art about where you're at at the moment. I often hear the best artists say that: they say they're capturing where they are at the moment. That harks back to the whole point about art being so contextual. In good art you need both the contextual element and the transformative element.

So in summary: start with pseudo-random seeds, immerse yourself in interesting environments, seek a translunary perspective, preserve what you write, and write often. That way you'll probably find that songwriting becomes an organic, natural experience. If you're doing all that and still straining, it might be because that approach to songwriting isn't for you. This isn't some universal rulebook, just some hints and tips at what might work.

Bringing it all Together

Songwriting is a powerful art and a difficult craft. But I think it's made less expressive and more difficult by the inherent cultural preconceptions we have about what makes a song, and what music is for. If you can shed those, and you don't have to be particularly revolutionary or even clever about it, you can already come a long way towards making good songs, for some value of what I consider good.

It's impossible to be entirely culturally independent from contemporary music ideas and values, and hard to be even moderately so. Moreover it's not a good thing to seek in many cases, since to do so would be like Ramanujan catching up with the previous 100 years of mathematical discoveries on his own, in a tenth of the time. Whilst he proved himself a genius in the meantime, the ground which he covered had already been trod. Someone should've told him.

But still, art is what's different, not what's the same. If someone creates a huge marble tablet and spreads a load of milk over it... I've seen that in some museum of contemporary art, and I think that perhaps it was art. Lots of people disagree with me, but it was interesting, bizarre, and different. I don't exactly like postmodernist art myself, but I do at least pay it the nominal courtesy of recognising it to be so. If someone did the same marble and milk installation again somewhere, however, whilst knowing about the original, I don't think that'd be art. Perhaps if they constructed it to be a thousand times larger that'd be engineering art, but it wouldn't be art in the sense that the original has it.

Art is also the mirror of nature, so I hear, but if so then it's a slightly Carrollesque mirror. A photograph can be art, it's true, but only in the sense that it was chosen out of billions upon billions of possible photographs that could have been taken and weren't; otherwise any random photograph could be considered art, and any photograph taken by a machine would be considered art. Music might have tribal drumming or the song of the nightingale as its antecedent, but what we make is different now, it's a unique phenomenon in nature. Birds don't paint, even though they do sing. A good human song might sound like a bird in its beauty, but a tape recording of a birdsong is not a human art. In like fashion, imitative music without the transformative element is not going to be considered good music.

So again, avoid formulae, and shun constraints. Use whatever you find. Beatrice Harrison played cello with the nightingales in her garden, which is a beautiful thing even in the concept. If you just missed a note on the guitar but it sounds good, that's an avenue. Then if you want to really develop the knack for songwriting, immerse yourself in it, writing songs in a flurry of activity, even if you don't produce anything good to start with. Stimulate the creative urge in whatever means you can. Make sure that none of your output is lost, as much as possible, and don't be afraid to reuse it. Overall, seek to capture something special, even if you can't put your finger on how you did it. Once you've made a song, it's there to keep and to share. And most of all, don't create without enjoying what you've done!

Sean B. Palmer, 2006-12-22