The impulse to gather songs together is probably as old as song itself. Doing so allows us to contemplate particular themes, topics or stories, or the output of certain poets and composers. But what seems at first like a straightforward activity has, in the form of the song cycle, come to represent an altogether more intriguing complex of creative and aesthetic issues. According to Charles Rosen, what began as a modest genre for the unambitious amateur became a major endeavour that in weight and seriousness rivalled opera or the symphony (the string quartet and the piano sonata might be better referents) – a triumph all the more remarkable in that its basis remained the simple lyric, and that its ascendancy has not been challenged.1 The purpose of this book is to examine the reasons behind the song cycle's rise in status in the late eighteenth and early nineteenth centuries, to discuss its period of great achievement, and to ponder its continued attraction for musicians and audiences during the last century or so. It will not only consider a wide repertoire, but also treat the song cycle as a barometer of music's changing role in culture and society in a number of countries and across many generations, in particular through its relationship to other arts (especially literature) and its participation in the formation of individual and national identities.
But is the song cycle really a genre? Unlike the sonata or symphony, it can boast no conventional pattern of movements; there need only be more than two songs. It might seem that the order in which individual elements are placed should be honoured, and that only the whole work should be taken into account; but to do so would contradict centuries of performance practice. Similarly, one might expect the group to be connected in some way; perhaps that, as a cycle, beginning and end will conjoin. While there are famous song cycles that do just that, there are others (just as famous) that do not. These conundrums, among others, explain why attempts to define the song cycle often end with a throwing up of hands: by describing it as ‘the most maddening of genres’ or claiming that ‘the only requirement is a demonstrable measure of coherence’.2
However a song cycle is structured, the ‘cycle’ element is loaded and the genre's stature is dependent on it. The obvious association with nature – with life cycles, the passing of seasons, and so on – especially during the second half of the nineteenth century conjured up organic metaphors, with attendant ideas of truth, unity, wholeness, progress, even purity (sometimes referred to under the umbrella term organicism). No matter how quaint those words might sound these days, it is vain to ignore the fact that part of the song cycle's allure is that it can be seen to adhere to those ideals. Yet at the same time the genre has proved itself capable of questioning our investment in those values. I’ll return to this point in more detail shortly; before going any further, though, it will be helpful to give a brief preview of how the song cycle came into being.
Although almost every European country had its own song-writing tradition by the start of the nineteenth century, those of Austria and Germany have outweighed all others in terms of influence. It is there that we must begin. The rise of the song cycle cannot be separated from significant changes in the way in which, around that time, music's relationship to words was conceived. Its emergence would not have been possible without the lyric poem, one of the central forces of literary Romanticism in the late eighteenth century. Since classical Greece, the lyric had been associated with an idea of singing. But what was still more important for the German Romantics was its emphasis on the ‘lyric-I’. One important strand of their poems was typically written in the first person, and typically dealt not so much with characters and plots as with personal expression. Such focus on inner motivations encouraged connections between the poem's voice and that of the author – an association of profound consequence for the interpretation of musical settings of those texts. The Aristotelian view that poetry should be concerned with mimesis – imitation – was discarded by Romantic writers such as August Wilhelm Schlegel and Johann Gottfried Herder; in its place came poetry as a language of feeling. According to Herder, poetry was ‘the music of the soul. A sequence of thoughts, pictures, words, tones is the essence of its expression.’3 It was a short step from likening poetry to music to endowing music with poetic significance. Increasingly, Romantic writers sensed that the art of tones was a superior vehicle for conveying the inexpressible, and endowed music – particularly instrumental music – with power beyond language's grasp.
It is important to recall, though, that music's new, elevated status was not automatically extended to vocal genres. At the end of the eighteenth century, singing (apart from that taking place in the opera house or in church) was mostly an amateur pursuit – an entertainment for domestic settings, among family and friends. Such Hausmusik had to be accessible, and arrangements of ‘traditional’ folk songs were very popular. Settings of more sophisticated poetry, such as Johann Wolfgang von Goethe's, also tended towards the folkish: simple and singable. These songs (or Lieder, as the Germans call them) were usually written for voice and piano (sometimes voice and guitar), and were mostly strophic, meaning that each verse of a poem (however different in lyrical or narrative content) was sung to the same music. This strophic musical form was used because the words were thought more important than their musical treatment; emphasis was on the poet first, the expressive powers of the singer second, the composer a distant third.
Two social factors need to be taken into account. First, the Lied arose at the same time as, and catered for, a new musical audience: the educated middle classes, who were gradually supplanting the aristocracy as the main patrons of the arts. This shift necessitated a greater commercial awareness on the part of composers and their publishers, who now needed to advertise their wares in a competitive marketplace. Such commodification of music had a direct impact on the development of the song cycle. Because Lieder, like the poems they set, were relatively small-scale, they tended to get published in collections. Gradually – taking their cue from poetic cycles, which emerged from similar circumstances – various terms were borrowed to mark such groupings: Reihe (series), Kranz (ring), Zyklus (cycle) or Kreis (circle). These terms might be decided before the poems or songs were composed, or they might be applied afterwards. But in large part they were granted with an eye to making a sellable identity rather than out of any abstract concern to establish a new genre.
The second factor to bear in mind is the close relationship, even in the earliest examples, between Lieder and expressions of national identity. It was in particular the folk-like (the German adjective is volkstümlich) aspect to these simple strophic songs, setting German words, which appealed as an expression and definition of national spirit. As this is a point to which we'll return, I will say no more here, apart from to observe that a similar nationalist urge appears in the song cycle traditions of almost every other European country, perhaps because of a tendency to set poems in the native language, often with reverential reference to details of the local landscape.
The umbrella term ‘song cycle’, when used to describe collections at the beginning of the nineteenth century, thus encompasses many different types. Often they were based around a poetic topic: the seasons or months of the year; a collection of flowers or colours; the experience of wandering; a sequence of emotional states; or an examination of particular imagery, such as the forest. The scholar Ruth Bingham has proposed three basic conceptual shapes for such ‘topical cycles’ (the circle, the spoked wheel and the spiral), but in few cases does much emphasis seem to have been placed on their precise order – they function more like catalogues, simply collecting diverse items under a common title.4 A more self-conscious approach to the order of songs developed through the cycle's interaction with narrative literary forms, such as the Liederspiel (a play including songs), and through choosing to set poems included in novels, such as Goethe's Wilhelm Meisters Lehrjahre (Wilhelm Meister's Apprenticeship, 1795). Because these latter songs were excerpted from larger-scale narratives, they often make no sense as stories in themselves (Bingham calls them ‘external-plot cycles’). But the pull of order was great. Poetic cycles were, as the nineteenth century wore on, increasingly constructed to follow their own ‘internal plots’, encouraging listeners to think of them as closed, complete forms whose order should be respected.
The assertion of underlying coherence was crucial in the upward cultural mobility of the song cycle: it implied that such a collection was worthy of more serious contemplation, in line with larger-scale musical genres, such as symphonies or – more pertinently – cycles of lyrical piano pieces. This impulse to ‘higher forms’ – evident most clearly to us in Beethoven's one song cycle and in several by Schubert – transformed Lied composition. Increasingly, composers moved away from simple strophic patterns to what is often referred to as through-composition, allowing the music to take its own course rather than being bound by repetitions, and hence aspiring to a more direct response to the meaning of a poem. That formal freedom was matched by a new approach to vocal writing: one which, where appropriate, abandoned tuneful, volkstümlich melody in favour of flexible, expressive lines that worked more closely with the piano. The piano's increasing prominence was in keeping with Romantic privileging of instrumental music: soon it tended to define the character of a particular Lied and to suggest an independent interpretation of the text. Another critical change in approach was composers’ increasing concern with establishing musical coherence, in terms of treatment of themes and harmony. Early examples include Beethoven's An die ferne Geliebte (1816, of which more below) and Carl Maria von Weber's Die Temperamente beim Verluste der Geliebten (The Four Humours at the Loss of the Beloved, 1816). However, and despite the pre-eminence of these composers, it was not until the late 1830s, with the ballad cycles of Carl Loewe, that critics started explicitly to connect the concept of a cycle with its arrangement of keys.5
In other words, the Lied – and, with it, the song cycle – began during the early years of the nineteenth century to become established as a work in its own right, rather than as mere musical accompaniment to a poem. And inevitably that process gradually took it out of the realm of amateur performance. As will be discussed often in this volume, the relationship, never entirely simple and often fraught, between compositional practice and performance practice has had a profound effect on the way in which ideas about the song cycle have been formulated. Once Lieder began to be sung by professionals, and particularly once those professionals began to sing complete cycles in public venues, composers’ ambitions soared. The once modest song cycle became a potent vehicle for lofty ideals: for exploring new methods of text setting and modes of musical representation, new vocal and instrumental techniques, new harmonic territories. In this sense, the song cycle interacted more actively with changing conceptions of musical modernity than did any other genre: to trace the history of music through works such as Schumann's Dichterliebe (1840), Musorgsky's Sunless (1874), Schoenberg's Das Buch der hängenden Gärten (1909) and Pierrot lunaire (1912), and Boulez's Le Marteau sans maître (1955) is to engage with generation upon generation of the avant-garde.
Modernism has never solely been about technical innovations, of course. The song cycle has also continued to attract composers because, through its close relationship with poetry, and its tendency to focus on a single vocal soloist (both as ‘author’ and ‘performer’), it encourages experimentation with notions of self-expression and what might be called constructions of the subject – both of them modernist (and postmodern) preoccupations. At the same time, though, that solitary subject could only be constructed in relation to society; for every daring, radical song cycle there is another that reflects the more conservative outlook of the middle-class audience. This point needs emphasis. Devoting an entire book to the song cycle obviously implies that it is an established genre; but it is very far from the monolith one might assume, a point hinted at by the fact that there has, so far as I could find, never been a volume like this before in English.6 Although there are many song cycles that nicely illustrate the ‘musically composed’ model, there are many more that are arranged more loosely, by topic or simply by poet: the different types have continued side by side, and all can be referred to as cycles. To put this another way: it will become apparent during the course of this book that the idea of the song cycle has often been more important than whether the cycle itself is a coherent, cyclical structure.
In order to do justice to the varied aesthetic principles behind this complex genre, I do not intend to chart a history of the song cycle, but to make ports of call at what seem significant moments. By way of an introduction, the remainder of this chapter will delve a little deeper into the challenges posed by the concept of the song cycle, by considering some problematic examples. I start where many have started, with what is often called the first ‘proper’ song cycle: Beethoven's An die ferne Geliebte. Following that are two examples by Schubert, ones whose questionable cyclic status raises some interesting issues about the way in which we define cycles more generally: the Abendröte cycle (1819–23), and Schwanengesang (1828). Then come some responses to perhaps the quintessential Romantic example, Schumann's Dichterliebe (1840), which should help clarify what is at stake in attempting to define how cycles work, and perhaps start to answer the much harder question of why they matter. And finally we move into the twentieth century, to three examples that demonstrate how, despite radical reconfigurations in musical language, an idea of the song cycle has continued to capture musicians’ imaginations: from Francis Poulenc's Tel jour, telle nuit (1936–7) to Serge Gainsbourg's Histoire de Melody Nelson (1971) and A Grand Don't Come for Free by The Streets (2004).
The first known definition of the song cycle, by music historian and librarian Arrey von Dommer, dates from the early 1860s:
Two main points emerge, both of significance for our understanding of the song cycle: first, the emphasis on coherence and comprehensiveness; second, the idea of diversity within that unity – that individual poems can stand alone, but also work as components of a larger entity.Liederkreis, Liedercyclus. A coherent complex of various lyric poems. Each is closed in itself, and can be outwardly distinguished from the others in terms of prosody, but all have an inner relationship to one another, because one and the same basic idea runs through all of them. The individual poems present different expressions of this idea, depicting it in manifold and often contrasting images and from various perspectives, so that the basic feeling is presented comprehensively.7
Dommer refines his definition with further commentary about the music's technical features, again stressing the idea of parts making up a whole. Typically, he says, the music is strophic. While the melody, musical form and key usually change between poems, the latter are frequently bound together through ritornelli or other instrumental interludes. Although Dommer makes no reference to possible narrative structures, he concludes by claiming that there is little to distinguish the song cycle from the solo cantata, apart from the fact that the former has no recitatives and that it replaces arias with Lieder. The song cycle Dommer seems to have in mind is Beethoven's An die ferne Geliebte, in which all the songs apart from the last are strophic (with only minor variations in the vocal melody but with the accompaniment varied between strophes), and in which instrumental transitions unify the parts.
In a letter to his publisher Steiner, Beethoven referred to his ‘Liederkreis an die Entfernte’, his song cycle to the distant one, and the published score was subtitled ‘Ein Liederkreis’.8 This seems to have been the first time a composer referred to a collection of songs as a cycle (Kreis), which is one of the reasons why it has often been taken as the first example. Other defining features are that the six interconnected poems are by a single poet, Alois Jeitteles, and that they sustain a single narrative voice. The theme is separation; the vocal ‘character’ (or persona) hopes that his love songs will travel over mountains and valleys to reach the ear and heart of his beloved. He asks the clouds, brook, birds and winds to carry his message. But as the month of May comes round again, and he watches a swallow building its nest, he realises that they will never be reunited: all he can do is sing longingly to his beloved, in the hope that the distance between them may recede.
Lost love, distant landscapes and consolation in nature are constant themes in nineteenth-century song cycles. But it is not these that are usually cited as evidence for An die ferne Geliebte being their progenitor. Rather, it is a musical trait. Although Beethoven's vocal writing is fairly straightforward, in keeping with the folk-like idiom of contemporary Lieder, it displays a greater concern with formal coherence than any previous example. Because all the songs are linked by the piano, their independent performance is discouraged. What is more, the music of the opening returns at the end, causing An die ferne Geliebte to begin and end in the same key, and thus consolidating the sense of harmonic completion. Opening and closing in the same key might seem unremarkable; after all, it was common practice for instrumental works at the time. However, composers tended not to approach the song cycle with similar concern for tonal patterning, probably because they took their lead from the words rather than ‘purely musical’ concerns. Although individual songs might not be particularly adventurous in harmonic terms, the tonal relationships between them could often be quite wide-ranging (close relationships would be – in descending order – by degrees of a fifth or fourth, or by a change of mode (major to minor or vice versa); modulations of a step upwards or downwards, or by a tritone, are particularly distant). Usually, songs would jump from key to key, as if along stepping-stones. Beethoven's innovation in An die ferne Geliebte was in using the piano transitions not only to change the mood and melody, but to modulate, thereby smoothing over the potentially jarring effect of, for example, hearing G major followed by A♭ major (the keys of the second and third songs). If you will, he replaced the stepping-stones with a bridge. In other words, this is a cycle not only by virtue of its text, but also by virtue of the consistency and cross-references within its musical design.
However, there are several reasons to query whether An die ferne Geliebte should be thought the first or the archetypal song cycle; Dommer chose it for his 1865 definition for reasons we might disagree with. To be sure, citing Beethoven as creator of the model song cycle provided the ‘new’ genre with an auspicious heritage. But An die ferne Geliebte is an anomaly, both for its composer (he wrote nothing else like it) and in the sense that its influence was not immediate. Moreover, in the time between An die ferne Geliebte and Dommer's definition, the scale and range of the song cycle had expanded considerably. Strophic forms were no longer de rigueur; instead, more varied, often through-composed, structures were favoured. Instrumental interludes between songs were unusual and vocal writing on the whole had become less folk-like. Dommer's reference to the accompaniment's ability to ‘portray and paint the situation in a characteristic way, and also supply, in regard to the expression, what the voice must leave unfinished’, is good for Beethoven, but was taken much further by later composers. Indeed, while Beethoven casts a long shadow over almost every musical genre in the nineteenth century, song seems to have slipped his grasp; or, as Schumann put it (maybe with his future achievements as a glint in his eye), song was the only genre to have made significant development since Beethoven.9
The composer who contributed most to this post-Beethoven phase was Franz Schubert. His development of Lieder as an art form has already been mentioned; his achievements with song cycles were no less profound. His most famous cycles, Die schöne Müllerin and Winterreise, have a chapter devoted to them later in this book; for that reason, we shall avoid them here, and approach the question of definition using examples that have somehow resisted categorisation: the Abendröte cycle to poems by Friedrich Schlegel and, perhaps more surprisingly, Schwanengesang.
Abendröte (Sunset) is unusual in almost every way. Few composers set Schlegel's lyrical poetry, and Schubert's decision to do so in 1819 probably reflects the sophisticated literary circles in which he moved. Schlegel's Gedichtzyklus (cycle of poems) consisted of two sets of ten poems, each set prefaced by an untitled prologue and ending with a sonnet entitled ‘Der Dichter’ (‘The poet’).10 They are bound by mood more than anything else. The first half describes the beauty of the natural world, with poems to the mountains, the birds, the river, the rose, the butterfly, the sun, the wind, and the boy and shepherd who attend to them. In the second half night has fallen. The wanderer walks serenely through darkness, again surrounded by natural wonders – the moon, nightingales, waterfalls, flowers, stars and the night breeze. A lovelorn maiden is the sole other human presence, before the poet steps forth to conclude the cycle.
Schubert produced only eleven songs, although the front page of the fair-copy manuscript indicates that he intended to write a complete cycle under the name Abendröte, the title of Schlegel's first prologue. Perhaps Schubert was inspired by An die ferne Geliebte to attempt a musically unified cycle when he began his Schlegel settings but abandoned the idea.11 One obstacle to considering Abendröte as a cycle is that the songs are for different types of voice; while all can be sung by a female, the tessitura or ‘cruising altitude’ of ‘Die Rose’ is very high, and the end of ‘Abendröte’ rather low. Another, more considerable, problem has to do with how to order the songs, as Schubert gave no real guidance. If they are given in Schlegel's order, we are left with gaps, which discourages thinking of them as a whole. More importantly, that grouping produces a sequence of slow songs in major keys, and thus lacks the variety we usually expect in a cycle.
What might seem trivial – not least because it ignores the often-startling harmonic journeys of the individual Abendröte songs – raises an important issue.12 Although, after An die ferne Geliebte, a cycle's status is often ‘proved’ in analytical terms by its consistency of harmony and mood, an example that is too uniform risks failing to keep one's interest. For example, Schubert's Winterreise was initially criticised because of its monotonous tone. While we might now disagree with that judgement (the monotony of Winterreise is surely crucial to its effect), something similar might have been in the back of Schubert's mind as he composed what we now know as his next cycle, Schwanengesang, which seems almost deliberately to seek out contrast and diversity.
Indeed, if Schubert's Abendröte fails as a cycle because of its incompleteness – it provides neither a coherent narrative nor a sense of musical progression – then so too must Schwanengesang, composed in 1828 and published posthumously in 1829. The title, ‘Swansong’, was devised by the publisher Tobias Haslinger, with a view to selling it as Schubert's ‘last work’.13 Haslinger's role raises some crucial issues about the extent to which we should abide by what, for better or worse, are known as the composer's intentions: it is quite possible that Schubert never meant these songs to be presented as a cycle. Yet this has not prevented several attempts to prove that they should be taken as one, indicating the extent to which the term ‘cycle’ is somehow thought to bestow added value.
Schwanengesang is made up of two sections, devoted to different poets. It begins with seven songs to poems by Ludwig Rellstab; then come six Heinrich Heine settings; and then, to end, is a setting of Anton Seidl's ‘Die Taubenpost’. The final Rellstab song begins the second section, meaning that the poets are not neatly grouped together, and the ordering within the groups, on the surface, makes little sense. Moreover, the two sections are different in form and character. The Rellstab settings have tended to be overlooked in favour of the Heine ones, which are typically treated as a cycle in their own right. (The reason for this might have less to do with Schubert's music than with the reputations of the two poets, Heine being by far the more famous and respected.) Rellstab's poems last between three and six stanzas; Heine's are mostly in three brief stanzas. When setting Rellstab, Schubert adapted his music to the varied structures of the poems; thus some songs are almost exactly strophic, while others are essentially through-composed. With Heine, though, Schubert complemented the simplicity of the text's structure with ternary (or sometimes binary) forms. Neither do the two sections connect on harmonic terms; the whole cycle may begin and end in G major, but only because of Haslinger's inclusion of ‘Die Taubenpost’. Otherwise, there are some unexpected leaps from one song to the next, almost to the point where they seem designed to obscure potential connections between them.
It has been argued that when the Heine settings are rearranged in the same sequence as in the poet's Die Heimkehr a poetic and musical narrative emerges, thereby forming a ‘single coherent work’.14 But this assumes (on shaky evidence) that there is a distinct narrative to the poems in the first place: the eighty-eight poems of Die Heimkehr are merely grouped into loose sub-cycles, around topics such as the sea and returning home. Schubert only once chose successive poems from Heine's collection, otherwise interleaving songs from different sub-cycles. While there may be associative links between the poems, and perhaps an emotional trajectory, no specific story is told. With regard to musical coherence, the matter is always more complicated. There are some characteristic melodic contours that recur; but if the order in Schubert's autograph manuscript seems deliberately to obscure key relationships between the songs, rearranging them to form a more ‘logical’ sequence has little to recommend it.15 What is more, Schubert was prone to transposing songs to suit particular singers, which puts a question mark over any attempt to fix a particular meaning for any given key or key relationship. He does seem to have used tonality to evoke a sense of distance from or closeness between certain songs (for example, the progression from the reasonably hopeful ‘Ständchen’ to the rejected grief of ‘Aufenthalt’ is conveyed through the awkward shift from D minor to E minor). But these local connections suggest a concern for immediate contrast rather than abstract design.
It is striking that recently the Heine settings have tended to be considered individually or in pairs, rather than as a group.16 Moreover, as with all Schubert's cycles, they are often performed independently. In part this probably has to do with their length and, for want of a better word, weight; while there are several complete recordings of Schwanengesang, the prospect of singing the whole work live, from beginning to end, is daunting. Yet Haslinger's decision to present these songs as a cycle has, by default, meant that they are known as such, proving the lasting power of labels – and of publishers.
Indeed, it is apparent from the Beethoven and Schubert examples given above that any attempt to define the song cycle as a genre is at the mercy of titles, taking the inclusion of Liederkreis or Schwanengesang as our guide rather than any formal features. We also need to decide how seriously we should take a composer's arrangement – is it within the scholar's or performer's rights to reorder Abendröte or Schwanengesang, in order to enhance its claims to being called a cycle? From another angle, we need to decide the extent to which a cycle has to be cohesive, and how much variety is acceptable. Harmonic relationships between songs seem to matter (particularly to scholars, who tend to have the score in front of them), but it would be rather short-sighted to make connections only on ‘purely’ musical terms. All these questions become still more urgent when we consider one of the most famous of all song cycles, Schumann's Dichterliebe.
Dichterliebe, like An die ferne Geliebte, deals with lost love, makes explicit tonal connections between songs, and returns to earlier musical material towards its end. These are some of the reasons behind its status as the quintessential song cycle, a work against which all others tend to be measured; however, the full story cannot be addressed without taking into account broader contexts: the relationship between music and text, historical circumstances, aesthetic values and performance history; and, critically, consideration of the method of scrutiny as much as the object scrutinised. It thus seems worth spending some time here with the different approaches taken to Dichterliebe: to do so is a means by which to introduce several key issues that will resurface throughout this book.