Chapter 1
Analysing Dylan Songs


ERIC CLAPTON once said about Dylan: ‘His way of playing anything is totally hybrid. It doesn’t make sense musically to the scholar. [. . . ] At first listening, everything he does is just real hopeless. Then you look back and realise it’s exactly right.’ As a scholar I take this as a challenge: If something is ‘exactly right’, but still doesn’t make sense to the scholar, it is either the scholar’s sense or the scholar’s analytical tools that are inadequate. I take the liberty of disregarding the first possibility (although that is probably the commonest cause for scholarly not-being-made-sense-to-ness) and concentrate on the second: the problems inherent in musical analysis of music of Dylan’s kind.

The Object

The practice of musical analysis is closely connected with developments in the genres and styles of music making of the end of the eighteenth and throughout the nineteenth century. This is the so-called classical-romantical period with composers like Mozart, Beethoven, Schubert and Wagner. It is also the period when the concept of art took the shape that is familiar to us today.1

An analysis presupposes an object of analysis. A central idea in traditional analysis is the double notion of a unified musical work with an internal development. The work may be the product of a composer’s free and creative mind, but it is laid down once and for all in the score, and there ends the liberty. Any realization of the work must be compliant with the score – add a note, and you have, strictly speaking, a different work. There is really no room for improvisation, except within very limited boundaries, and even these are usually not exploited: in the place where the pianist is supposed to improvise an exuberant cadenza, most performers are content with playing the cadenzas that Mozart himself, or Beethoven, wrote. This is of course lamentable from the point of view of musical creativity, but the analyst applauds: it gives him a clearly defined object of analysis, from the first note to the last.

Contrast this with the situation of a Dylan song: what is the object of analysis of, say, ‘The times they are a-changin’ ’? The published ‘score’, usually with some remarkably silly piano arrangement of Dylan’s guitar strumming? The album version, tabbed and uploaded to some tab site? Clearly both alternatives are inadequate, to say the least. My first encounter with ‘Forever Young’ was of this kind. I saw it in a song book before I had ever heard it. I looked through the melody, the chords, tried to imagine how it might sound – and rejected it as a fairly uninteresting post-motor-cycle-accident, pre-Blood-on-the-tracks-song. When I later, almost reluctantly, bought the album, this song was a shock of emotional intensity, in this case even reinforced by the tension between the two versions. And you can probably take any Dylan song the same way: It doesn’t look much on the paper – whatever power there is lies somewhere else.

This is one reason why a transcription in any form cannot do the song, as a musical work, justice. The other is, of course, that no matter how meticulously you note every single detail of one particular performance, the next time you hear it, it will be different, either because Dylan has rearranged the song, or simply because of the improvisational character of popular music in general and Dylan’s music making in particular. With an object of analysis which cannot be objectified, since it changes all the time, there is really no other alternative than to endorse Paul Williams’ approach: to treat every single performance of a song as an independent work of art. The performance is the object.

We might have settled with this, but upon closer look, it is still too simple an explanation. I used to be attracted to Jeff Todd Titon’s approach to the improvisational character of blues2 . He assumes that most of the early blues songs were improvised on the spot; that words and music were assembled while singing, from the singer’s storehouse of phrases, situations, turns, descriptions, according to some specific pattern, but without a fixed text that is repeated exactly from performance to performance. Even when singing the ‘same’ song several times in a row, it is improvised from scratch each time, and minor (and even major) differences occur between the versions (he actually tested this).

I used to think of Dylan’s performances in the same way, given the huge mass of text and the re-workings of some of the texts, until I realized that the variations are too small to really fit the model. The texts are actually memorized in a next to exact form, and the different versions of a song like ‘Tangled up in Blue’ are for the most part conscious re-writings, not improvisations. There may be exceptions to this, such as the constantly changing lyrics to ‘Simple Twist of Fate’ during the 1984 tour and of ‘Ain’t Gonna Go to Hell for Anybody’ some years earlier, but whereas they may well conform wit Titon’s blues model in their outward appearance, and to some extent also in their cognitive foundation, they are still not improvisations taken from a storehouse of building blocks. Rather, they are literary sketches, worked out in front of a live audience instead of in the poet’s solitary chamber, but they are literary all the same.

Now, the same goes for the music. This may seem odd since what is usually considered to be a distinctive trait of Dylan as an artist is his way of constantly changing the musical setting of his songs. The melody is probably the most unstable element, but even the tempo, the rhythm, the instrumentation change all the time. But still: the songs are always recognizeable as such. Usually the chords and harmonies are intact, and variations are within the normal limits of the genre. However different two performances of ‘Just Like a Woman’ may be, they are versions of the same song all the same, and not only two different works of art which happen to have the same words attached to them. There is always something that is preserved, through all the variations.

This simple observation opens up a large field of interesting questions concerning Dylan’s relation to the different styles that have influenced him (blues, English folk ballads, The Beatles and rock ’n’ roll) on the one hand, and to the different genres of music production (the orally transmitted folk/blues, the performed, score-based classical music and the commercial popular music, transmitted through electronic media). To make this long story short: Dylan relates strongly to improvisational musical traditions, but also to traditions with an ultimately defined ‘work’: defined either by what the composer has decreed, as in classical music, or by what is considered commercially most efficient.

In this sense, Dylan’s songs are not improvisational once they have reached an album:3 he seldom deviates strongly from and always relates closely to the ‘official’ versions. This means that although each performance may be considered an independent ‘work of art’, it is still meaningful to treat the group of works that can be subsumed under the label ‘Just Like a Woman’ as one single work of art with many realizations, much in the same way as fifty prints from the same plates are individual works but at the same time representations of the same work.

Paul Williams has proposed a view similar to this. In the introduction to his books Bob Dylan: Performing Artist he relates an anecdote about a backstage meeting with Dylan. He was going to make a comparison between the different Dylan versions and a series of lithographs by Picasso, worked on over the course of six weeks. He writes:

At first Dylan protested that he wasn’t interested in that kind of art at all, but he looked at the page and seemed to be pulled in. Staying with his initial (it seemed to me) anti-intellectual stance, he pointed to the second earliest of the drawings and proclaimed it as the best: ‘He should have stopped at that one.’ Then, looking closer: ‘Oh, but I see why he had to keep going . . ..’

But the problems still remains: what is it about a performance that makes it worthy of the label ‘work of art’? What are the criteria, and are they the same for Dylan as for, say, Pavarotti or U2? Are ‘works of art’ from different media or styles comparable at all? This discussion presupposes another discussion: that about what a work of art is in the first place. One possible definition may be: ‘A structure made up of elements that are considered apt for reflection or contemplation, and conceived and presented in a way that stimulates this.’ That seems to be what we do with art: we enter a different state (of mind or place) to expose ourselves to something that we allow to influence us, emotionally or intellectually.4 That implies several things: the ‘structures apt for reflection’ are not objectively given, but open to interpretation on different cultural or historical contexts – in other words: they are dependent upon style. A work of art can only be efficient in some kind of relation – including the revolutionary – to a style.5 In periodes of stylistic change this relation tends to be explicit, whereas in most cases it is implicit. In Dylan’s case there are a number of such implicit styles involved: the blues background, the folk, the standard rules of European/American harmony, the development of rock, traditions of voice treatment and the relationship with the text, including the relationship between song and spoken language, etc. To be able to appreciate a performance fully, as a work of art, it is in fact necessary for the listener to relate it to the stylistic systems he/she finds relevant, on all different levels, from the individual style (Dylan’s own style), through genre style (rock/blues/folk in general), and maybe even to some kind of ‘meta-style’ of song in general. Seen this way, every performance can be seen as a contribution to an ever-ongoing debate – what can be done within this genre? where are the limits? which crossovers are possible or artistically interesting?

This brings us back to the definition given above, and the second implication that can be derived from it: that art-status presupposes a volontary act by the listener, both by allowing it to influence us, and regarding the stylistical references we make, whether these are explicitly volontary (I choose to regard Dylan as a blues singer even when he sings ‘Emotionally Yours’, because I find it rewarding), or unconscious (I don’t consciously realize it, but my appreciation of a Dylan song must derive from a whole lot of different things I’ve heard and appreciated before), or for lack of knowledge (I don’t know enough about the folk movement in the early 60s to really be able to understand that possible relation fully).

Most of this happens unconsciuosly – we don’t ponder a performance and then decide to let it hit us in the stomach with a feeling that changes our life. Rather, the choices have been made beforehand: we choose to like a certain kind of ‘screaming’ or whining when it seems that it can be rewarding. I seriously doubt that eighteenth-century Viennese, however sophisticated and developed their taste may have been, would have understood anything of any Dylan song. Even (half) the audience at Newport in 1965 and during the following tour with the Band chose not to be moved or touched by performances that are now classical precisely because of their emotionality.

The harmony

The other problem with the concept of the work is its inherent idea of organicism. According to the organicist theory, a work of art is, very simply stated, to be judged as an organism, where all the parts are related to each other in much the same way as the limbs of the body are. This is one of the basic aesthetical principles behind classical music. Even harmony is arranged in a similar manner: all the available chords are related to the tonic, the fundamental harmony of the piece, in a coherent and logical way, which determines the transition from one chord to another.

This system found its way from classical music into the realm of popular music through the hymns and the popular tunes of the Stephen Foster-type, and the military marching bands. They even managed to influence the field-hollers and laments of black Missisippi workers to eventually produce the twelve-bar blues, and from here it spread to today’s popular music genres. But in these new settings, most of the underlying aesthetic presuppositions of the Western classical harmonic system were originally not present. The aesthetics behind the rural blues is almost the exact opposite: instead of striving for a coherent organical whole with beginning, middle and end, it rather emphasises timelessness, through ostinato patterns (the twelve-bar blues pattern itself can be considered one) and through the predominance of the tonic and the fifth. Wilfrid Mellers describes two songs by Pete Williams, an early blues singer, as ‘still basically hollers in which speech is translated into pentatonic tumbling strains, with the guitar providing an ostinato accompainment with no sense of harmonic progression.’6 Harmonic progression, which is so essential to western music, has virtually no place at all in blues music, and the dominant chord, the means par exellence by which harmonic progression is achieved, is also a foreign flower in the landscape of the blues. It is remarkable how the dominant chord often sounds awkward – if it is clearly sounded at all – in much rural blues, mainly due to the sharp contrast between the blues’ flat pentatonic seventh, and the sharp leading note of the dominant chord.

Analysing an Idea

Knowing what a pie is doesn’t make anyone a baker. If the foregoing may be taken as a statement that Dylan’s music can be analysed, it still remains to show how that can be done. This is not entirely easy, and one major obstacle is that there is no firmly established analytical tradition for this kind of music. The tools and methods of musical analysis which are used today rely heavily on the work of the founders of the discipline in the nineteenth century, for better or for worse. They were certainly clever and skilled academics, but their material and theoretical background was limited, and modern musicology has only reluctantly, and most often half-heartedly, realized that theories based upon works by Mozart may work exellently for analysing works by Mozart, but may be worthless for other genres. Popular music is one of these, but there are even examples from within the ‘established’ field of musicology: Most, or all, early (pre-Bach) music defies analysis with the traditional methods, in much the same way as a Dylan song does. It quickly becomes evident that the established models or methods for musical analysis – Schenker analysis, functional analysis, thematic analysis etc. – are all derived from and therefore applicable only to music from the classical-romantical era. The moment one crosses this border, in one direction or another, there is a tendency that the treatment of the music itself stops at the descriptive level, or it is abandoned altogether, in favor of a sociological approach. Counting measures and determining keys can be done with any kind of music. Determining what is really contained in those measures, is a completely different matter.

Not only the tools and methods are received for free within the traditional fields of analysis; so is the goal of the analysis. When undertaking a Schenker analysis of a Schubert sonata, it is obvious what the goal is: to find the underlying tonal structure of the piece and its relation to the ‘surface level’ of the sounding music. This goal rests on a number of implicit presuppositions (that there is such a tonal structure; that it stands in a certain relation to the ‘surface’; etc.). These are basically presuppositions about the underlying systems of musical ‘meaning’ at work in the piece. Since these belong to the foundation of both the music in question and of the analytical method, they can more or less be taken for granted. (organicism, harmony). This is not the case when analysing a Dylan song or a Palestrina motet: there are no such analytical shortcuts available, because of the discrepancy between the underlying ideas of the music and the established analytical methods.

Even though there is a growing tradition of popular musicology, each new analysis will still tend to begin more or less from scratch and make up its own goals and methods along the way. We will therefore start out each separate analysis with very elementary questions: Why does the music sound the way it does? Which are the desicions and choices the artist has made, and why has he made them; what is his ‘goal’? Which are the means involved in reaching that goal? Which are the choices he has not made himself but that have been made for him, because they are so integrated in his musical background? Which are the underlying conceptions or systems of meaning that come out of this background? How does he relate to the different traditions (blues, folk, gospel)? Possibly also: What are the choices he has decided not to make, and why? A further development of this, as a kind of ‘meta-level’ would be the question of the cultural/aesthetic implications inherent in these choices.

Two basic presumptions are worth mentioning. One is that the music under consideration is interesting because of the effect it has on the listener, and that this effect has causes that can be explained or at least related to or deduced from specific characteristics of the music itself. Musical effect is not the result of ‘one heart speaking directly to another’ or ‘the forces of cosmos (or God) channeling energy through one individual in a mystical and ineffable manner’ or ‘the simultaneous contact with heaven and earth’ – it stems from a certain use of dissonance, agogics and rhythms, phrasing, texture, tempo, breathing, voice tessitura, instruments, interplay between musicians and a whole lot of other factors which may be difficult to pinpoint, and the totality of which may never be fully grasped, except in the immediate and intuitive manner of the listener, but which nevertheless are at the basis of the experience.

The other is that the musician does not have complete control over the making of an artwork. In one end is the influence of the musical traditions, which may not be conscious on the part of the musician himself. On the other end is the listener, who in the sense outlined above is the real ‘creator’. This means that what Dylan himself has to say about his music may be interesting but not necessarily essential for the understanding of his music.

On the other hand, this may be a too structuralistic understanding of music, which partly misses the point that music may also function as communication, not the least for Dylan. It is not an ‘empty’ structure to be filled by the receiver, it is an intended sturcture with a specific content from Dylan’s side, which may or may not be perceived by the listener, but which still should be treated as an essential part of the ‘artwork’. In other words, to the extent that musical elements are also expressive and communicative elements, it will be of importance to the analysis what is being communicated, so that although our emphasis lies on the study of the music, any song analysis which doesn’t take into account the lyrics of the song will be incomplete.

*  *  *

These are the some of the elements that will show up again and again in the articles that follow. I have seen it both as my task and as a pleasure to use whatever knowledge I have of aesthetics and traditional music analysis and theory, and apply that to whatever knowledge I have of Dylan’s music. If I have a project, it is to demonstrate how long-ranging ideas are: when I discuss ancient theories of the harmony of the spheres in connection with Dylan’s music, it is not only because I happen to enjoy reading ancient texts, it is also – and primarily – because I think they are relevant. We may not share the beliefs and the world view of the Ancients, but our beliefs and views have been formed out of whatever we as a culture have received and incorporated from what has come before us, and there is no easy way of deciding where that influence stops and something entirely new takes its place.