1. A Sleepy Prelude

One evening, about fifteen years ago, I grabbed the red-brick, brick-sized Faber edition of W. H. Auden’s Collected Poems from my bookcase. I was deep in a post-PhD hinterland, adjuncting at three universities while working a full-time admin job. Any free time should have been spent on publishable work that might improve my prospects. But I was exhausted that evening. I idly flipped through Auden, paused on the poem ‘Lullaby’, and picked up a dusty guitar with my other hand. In another life, as a teenager and undergrad, music had been my whole identity, but I’d put that part of myself away to get through my literary PhD, and had made no attempt to bridge those passions. I’d never written a song based on someone else’s poem. I hadn’t written any kind of song for years. But I started strumming and humming, reaching for an itch I’d long tried not to scratch.

I knew Auden as a formalist, with tidy poems that inspired so many musical collaborations in his lifetime. A poem called ‘Lullaby’ – how hard could it be to sing? I soon realised its structure might be more than I’d bargained for. Here’s the first of four stanzas:

Lay your sleeping head, my love,
Human on my faithless arm;
Time and fevers burn away
Individual beauty from
Thoughtful children, and the grave
Proves the child ephemeral:
But in my arms till break of day
Let the living creature lie,
Mortal, guilty, but to me
The entirely beautiful.

For starters, ten lines is painfully long for a cohesive musical verse. ‘Classical’ composers get away with more complex text settings, but I wanted something that sounded like a simple folk song, a plausible lullaby. Rhyme and metrical patterns seemed like obvious places to start, but ‘Lullaby’ was a riddle on both fronts. The poem’s four-stress line is fairly steady, but has seven syllables, so that most lines start and end on a stress. (‘Tyger, tyger, burning bright’ is another example of this equivocal metre, which poetry nerds might scan as either acephalous iambic or catalectic trochaic.) Auden also commits to an intricate rhyme scheme, with some rhyme pairs reaching from four lines away. The only end-rhymes close enough together to have any chance of being heard when sung are lines 2 and 4, then 8 and 9. But already in this first stanza, ‘arm/from’ and ‘lie/me’ are typical of the very slant rhyme throughout the poem, which any singer’s pronunciation would struggle to make audible.

And yet, the whoosh of these formal challenges opened something in me – something in my ear and voice as I reconsidered ‘Lullaby’ as a musical puzzle. Fumbling with chords and melodic possibilities, a lifetime of musical knowledge rushed to meet the technical understanding of poetic form I had acquired since my music school days. In that whoosh, I experienced poem-setting as a form of very close, deeply embodied reading. The intimate, physical process drew out features of the poem I couldn’t have grasped on the page. Holding Auden’s syllables in my mouth and lungs – in the more conscious, muscular way required by singing – made new subtleties literally palatable.

Music is a temporal form, and a lot of these were features you might only perceive in their duration or return. I always think of long vowels as the pizza dough of singing, and in that opening stanza, I felt the long A of the first syllable (‘Lay’) pull through the second line (‘faithless’), and across the pseudo-rhyme of ‘grave’ between ‘away’ and ‘day’. This string of A sounds was interwoven with the long U at the start of the second line (‘Human’), head-rhymed with ‘Proves’ and echoing in ‘beauty’ and the stanza’s last word, ‘beautiful’. Consonants, on the other hand, are the pizza toppings in singing, and Auden’s alliteration brought such a crunchy and gooey mix of textures, with the rush of fricative Vs and Fs in the teeth and lips (love/faithless/fevers/grave/Proves/ephemeral, etc.), amid nasal Ms (human/arm/time/ephemeral/mortal), all building towards the Ls and Ts in the final lines (‘Let the living creative lie … Mortal, guilty … entirely beautiful’, swilling from the back to the tip of the tongue.

I could go on, or say as much about the tricky syntactical structures that revealed themselves as I tried to divide Auden’s ten-line sentence (in all four stanzas) into singable, but still coherent musical phrases. The point isn’t that I’d discovered some secret to what Auden was thinking when writing ‘Lullaby’. These intricacies are unconscious for any good poet, and I hope it’s clear that I’m retrospectively reconstructing a musical composition process that was mostly intuitive on my part as well. But just as Auden trained his ear to the sonic and subsonic dimensions of language, I couldn’t help but listen to his words in a new, more sensitive way. ‘Lullaby’ is a poem about finding a moment for tenderness amidst the ravages of time. The speaker accepts the sleeping lover’s mortality and vulnerability, and the sense that the love and care between them cannot last forever only makes that moment of safety more precious. I felt all of this more deeply through the care Auden’s poem demanded from my singing. Every poem makes its own demands for setting, but these always deepen my relationship with the text.

Anyway, here’s what I came up with for ‘Lullaby’, if you want to listen. In the rest of this essay, I want to think through the creative and critical implications of poem-setting, moving beyond this sense of it as a formal puzzle, toward a better sense of the ethical relationship it engenders between the poet and their musical interpreters.

2. Between the Critical and Creative

From those first moments messing around with Auden’s ‘Lullaby’, the process of making ‘poemsongs’ – as I unimaginatively call them – has become an obsessive habit, demanding a careful balance of creative and critical thinking. The creative thrill of working out a new setting lies partly in the problem-solving and decision-making as you feel your way through the text. But as I’ve hopefully shown, poem-setting is also clearly a mode of formal analysis. At every level, it is an interpretative act, which different composers might take in very different directions, foregrounding different aspects of the text in their musical ‘reading’, with varying degrees of creativity or criticality.

Treating text-setting as a creative-critical practice helps address a binary debate, in some music scholarship, over whether it is a creative or critical act. Some commentators, such as the composer Martin Boykan, dismiss the idea that musical text settings can be ‘some sort of exercise in literary criticism’. He argues against those who suggest ‘the appropriate musical “reading” of the text is the primary criterion of the composer’s success.’ Rather, he suggests it is better to view such songs as ‘a new work of art that happens to use the same words as the poem.’[1] To bolster his case, Boykan cites no less an authority than Arnold Schoenberg, who in his 1912 essay, ‘The Relationship to the Text’, also mocks music critics who complain that a ‘composition does not do justice to the words of a poet.’[2]

Boykan raises a reasonable concern here. Because the text comes first, it’s hard not to view the musical setting as a ‘secondary’ creative process. As he says, the trouble stems partly from ‘comparing’ the song and source text. Yet, such comparisons inevitably arise with any adaptation across art forms. With a film based on a beloved novel, for example, many viewers will treat its faithfulness to the book as their ‘primary criterion’, along similar lines to Schoenberg’s complaint about doing ‘justice’ to the original. The same is true of translations, where perceived fidelity is often a measure of a translator’s ‘success’. Boykan is wary of such parallels, however, writing that ‘the relationship between musical affect and poetic meaning is so cloudy and uncertain that I doubt it is useful to think in terms of translation at all’.[3]

But why not? Many scholars of adaptation and translation would argue that these should also be regarded as ‘new works of art’ in their own right. Boykan’s reservations seem grounded in a belief that the critical potential of text-setting somehow undermines its creative standing. Others, however, don’t see these in conflict. The composer Matt Barber says outright that ‘setting poetry to music can be a form of literary criticism and analysis.’ While I looked to Auden’s deceptive formalism for my first foray into poem-setting, Barber says the linguistic obstacles of poetry like Wallace Stevens’s were part of the attraction for his own song cycle, To the Roaring Wind (2011). Without pitting the poetry or music against each other, Barber is adamant about the need for balance: ‘The reality is that a composer who takes up Stevens has to find a way to embody the music of the poetry without hiding it, all the while adding something that isn’t merely an add-on.’[4] Elliott Carter – another great setter of Stevens’s work – agrees about the enjoyable challenges of the process. He initially found Stevens’s poems ‘challenging to set to music because of their quick changes of character, use of irony and employment of unusual words’.[5] Stevens’s complex language might bring unique challenges, but with any poet, we can reimagine the composer’s job as hearing or reading the ‘music’ already present in the poetry.

Barber and Carter’s characterisations of poem-setting show a clear give-and-take between the creative and critical aspects of musical ‘interpretation’, without any sense of these critical dimensions undercutting the creative. As they both suggest, this dialectic begins with the choice of a poem. I also like a source text with some resistance, and agree with Carter that the game is less enjoyable if a text too readily anticipates its melody or meaning. Auden might have seemed like an easy first choice, but I’ve enjoyed other structural challenges – like trying to keep the expanding verse of Elizabeth Bishop’s ‘Visits to St. Elizabeths’ from spiralling out of control, or stretching the two lines of Dickinson’s ‘In this short life…’ into a song of respectable length. Other times, I’m drawn to a syntactical riddle, trying to find a simple song-form for the thirteen unpunctuated lines of Lucille Clifton’s ‘blessing the boats’ or the ending of James Joyce’s Ulysses (my only prose setting, if that’s allowed). Every poem is a new jigsaw, straining my critical understanding of its form and meaning and my creative decisions with the music.

3. Between Writing and Performance

One difficulty when gauging the creative and critical dimensions of poem-setting is that most writing on the subject, or on musical text-settings generally, focuses on ‘classical’ modes. Everything I said above about formal features being drawn out by the process is true regardless of style or genre, of course – and I hate invoking simplistic binaries like ‘classical’ versus ‘popular’ music. But there is at least one crucial difference between these two modes, which has practical implications. In classical settings, broadly speaking, a composer writes something down for someone else to perform. Whether for a soloist or a choral ensemble, the piece first takes written form. Because of this, a ‘classically’ composed text-setting has a separate existence from any performer’s interpretation, comparable to a poem’s existence on the page, separate from any potential readers. In popular music, on the other hand, and in popular genres of text-setting, the composer and performer – at least initially – are more likely to be the same person, and the composition itself might exist entirely in their muscle memory.

Maybe this sounds insignificant. But I’m arguing that the differently embodied nature of singing one’s own poem-setting – the physical interaction with the text through one’s own mouth and throat and lungs – adds an important layer of creative-criticality. However adeptly you might have fitted a poem’s metrical and phonetic quirks to a musical structure, however sensitively you might have matched its content and meaning to musical style or specific melodic and harmonic shapes, the act of taking someone else’s words into your voice and breath and body deepens the interpersonal dimensions of verse-setting in a very physical way.

Firstly, performing it yourself adds another layer of interpretation. In his study of rhythm and metre in German lied, Yonatan Malin compares the process to reading a poem aloud:

Musical settings also specify the rhythm of performance to a much greater degree [than poems typically do]; they are in this sense particular “readings” of the poems in time. Even with agreement about poetic meter and further levels of accentuation, there are a variety of ways that one can read a given poem.[6]

Any performer, like any reader, makes countless conscious and unconscious decisions when interpreting a text. Certain words or phrases might be given different inflections, alongside other practical considerations. Every time I sing Auden’s ‘Lullaby’, I have to decide how to pronounce words like ‘glacier’ in my American accent. In turn, these interpersonal negotiations give poem-setting in popular modes a unique ethical urgency. With ‘Lullaby’, for instance, I have to consider how the meaning of a queer love poem written within particular political and biographical contexts will change when voiced through my body and lived experience.

Among verse-settings in popular modes over the past century or so, many examples would illustrate this tension between the intellectual puzzle the process offers and the embodied realities any text conjures. In this hypothetical but much-needed history of popular poemsong, one could do worse than start with Billie Holiday’s ‘Strange Fruit’. Although its enduring cultural impact is rightly linked to Holiday’s haunting performance, the lyrics began life as a poem by Jewish schoolteacher Abel Meeropol. He published it in a teacher’s union magazine as ‘Bitter Fruit’ in 1937 – the same year Auden wrote ‘Lullaby’, incidentally. Meeropol then set it to music himself soon after, and Holiday recorded her definitive version in 1939.

Meeropol was reportedly moved to write ‘Strange Fruit’ after seeing photographs of Black lynching victims in the American South, and the hanging bodies that his unforgettable metaphor compares to rotting fruit. Such provocative conceits are arguably more typical of poetry on the page, where ‘the most heterogeneous ideas are yoked by violence together’ – as Samuel Johnson said of Metaphysical poets like John Donne – leaving the reader to contemplate their relationship. Pulled along by Holiday’s stark performance, the effect is relentless in ‘Strange Fruit’. The absence of personal pronouns removes the grounding perspective we normally get from popular song lyrics, and the three-verse structure denies listeners the usual feeling of release. These features might belie its poetic origins, but Billie Holiday made the song her own in performance (and bore all the personal and political consequences of doing so). In that regard, ‘Strange Fruit’ unquestionably fulfils Martin Boykan’s ideal synthesis; and viewing Billie Holiday’s version as a ‘new work of art’, separate from Abel Meeropol’s poem ‘Bitter Fruit’ – separate in her performance even from his musical setting – shows the extent to which embodiment can alter a text’s meaning and effect. A text inspired by horrific photos from a distance becomes something new in the body and voice of Holiday’s lived experience.

Musically, ‘Strange Fruit’ is a jazz ballad, but the process remains in the classical mode, with the composer handing off the piece for a performer’s interpretation. A more nuanced history than I have space for would show that in the course of the twentieth century, these roles and modes increasingly collapsed, particularly as folk songwriters became more interested in repurposing existing texts. Pete Seeger’s adaptations of ‘Guantanamera’ and ‘Suliram’, respectively from Cuban and Indonesian sources, are mid-century cases in point, raising complex questions not only about cross-cultural translation but also about any act of musical interpretation.

4. Towards an Ethics of Poem-Setting

We can push these questions further by acknowledging that every musical text-setting is an act of ‘appropriation’. By definition, it involves using someone else’s work and words for one’s own purposes. While I don’t believe this makes text-setting categorically ‘problematic’ as a practice, I do think the creative-critical framing of this practice might include both formal and ethical considerations. Adapting a poem with music can have profound and positive effects on our appreciation not only of its sonic and rhythmic structures but also of the historical, cultural, and personal contexts embedded in the borrowed language. At the same time, as I’ve said, taking another person’s words into your own voice is a deeply embodied undertaking, with implications and responsibilities that extend beyond the intellectual thrill of experiencing those words anew.

So, I’ll finish this essay by comparing a couple of more recent poemsongs and another example of my own, which show these ethical dynamics in practice. Patty Griffin’s ‘Up to the Mountain (MLK Song)’ is one striking illustration. Immediately in her title, putting ‘MLK Song’ in brackets both clarifies and complicates the lyrics’ authorship. The words of the song itself are then paraphrased rather than taken directly from Martin Luther King Jr’s ‘I’ve Been to the Mountaintop’ speech, about leading his listeners to the Promised Land without expecting to make it there himself. In her soaring gospel setting, Griffin adapts King’s poignant metaphor, from a speech delivered in Memphis in April 1968, the day before his murder. On one hand, this is a white songwriter appropriating a Black man’s words, while also drawing on a musical genre strongly associated with Black identity and history. On the other hand, Griffin’s ‘Up to the Mountain’ is also a reverential tribute to a figure of national importance for all Americans.

With any poemsong, it matters who wrote the words, and it matters who uses these words to make something new, who takes another’s voice into their own. These things matter ethically and politically, but also for the new meanings created in this interpersonal exchange. In that sense, poem-setting as a practice bears the same responsibilities and benefits from the same serious reflection on these matters as other acts of either creative or critical interpretation. Patty Griffin seems aware of this, first giving the song to the Black gospel legend Solomon Burke, who knew King personally, to record his version of ‘Up to the Mountain’ in 2006, before releasing her own the following year. The song has since become a popular showpiece on programmes like American Idol, where every new singer adds their own layers of meaning.

Another song from 2006 offers a trickier comparison. Kris Delmhorst’s ‘Since You Went Away’ appears on her album Strange Conversation, a collection of folky settings sourced from a range of poets – including Browning, Byron, Whitman, St. Vincent Millay, Cummings, and others. More than any individual poem-setting, Delmhorst’s album project was a catalyst for my own first attempts at poemsong, showing how varied and unexpected musical approaches could give new life to well-loved poems. Delmhorst also made a meaningful creative choice by including her version of the Harlem Renaissance poet James Weldon Johnson’s ‘Sence You Went Away’ on an album based on poems by mostly white poets. (Strange Conversation also ends with an adaptation from Rumi.)

Yet this particular setting is complicated by Delmhorst’s translation of the poem’s ‘Negro dialect’ – as Johnson and other poets, such as Paul Laurence Dunbar, described their use of regional Black American vernacular. In the title of Delmhorst’s version, ‘Since You Went Away’, only a single letter changes. But the lyrical translation goes beyond spelling in her recording. The first stanza of Johnson’s poem, ‘Sence You Went Away’ reads:

Seems lak to me de stars don’t shine so bright,
Seems lak to me de sun done loss his light,
Seems lak to me der’s nothin’ goin’ right,
Sence you went away.

On ‘Since You Went Away’, according to the lyrics on Delmhorst’s website, she sings:

Seems like to me the stars don’t shine so bright,
Seems like to me the sun has lost its light,
Seems like to me there’s nothing going right,
Since you went away.

I want to be clear that I’m not including this example to single out Delmhorst for ‘misappropriation’, but rather because her considered approach raises such difficult and important questions. Should she be singing this poem at all, we might ask, if it requires changing its language to something she feels able to sing? I don’t believe there is a simple answer. James Weldon Johnson was a prolific lyricist himself, collaborating with his composer brother, J. Rosamond Johnson, on successful work for vaudeville and Broadway. Their own dramatic, classical-style setting of ‘Sence You Went Away’ was recorded by a number of artists, including the great Paul Robeson. The published lyrics for their 1913 version reproduce the poem’s dialect and spellings, but the sheet music bears the title ‘Since You Went Away’. It’s perhaps impossible to know if this reflects the Johnson brothers’ wishes or their publisher’s.

Above, I argued for treating poem-setting as a form of ‘translation’ (pushing back against Martin Boykan’s resistance to this comparison). Going further, I’d suggest that ideas from translation scholarship – rather than text-setting scholarship – help reframe these tough questions. For starters, rather than focusing on choices made in individual settings, translation theory helps us appreciate how every translation involves a degree of ‘violence’. This violence, as Lawrence Venuti puts it in his influential study, The Translator’s Invisibility (1993):

… resides in the very purpose and activity of translation: the reconstitution of the foreign text in accordance with values, beliefs and representations that pre-exist it in the target language, always configured in hierarchies of dominance and marginality, always determining the production, circulation, and reception of texts.[7]

This applies equally to musical settings, where the ‘foreign text’ of the poem is reconstituted according to values inherent in the ‘target language’ of the music. Other scholars add nuance to this sense of ‘violence’. In ‘The Politics of Translation’ (1993), Gayatri Chakravorty Spivak emphasises the positive sense of a ‘fraying’ between self and other, if the translator can open themselves to the text’s alterity. In The Location of Culture (1994), Homi Bhabha places a similar emphasis on the ‘inter’ part of these interpersonal and intercultural encounters, as a space where shared new meanings can emerge. Contrary to Boykan’s sense, quoted above, that ‘the relationship between musical affect and poetic meaning is so cloudy and uncertain that I doubt it is useful to think in terms of translation at all’, I would argue that treating it as a translation is precisely what has the best chance of illuminating this cloudiness and uncertainty.[8]

I feel all of these complexities with every poem I attempt to set. Music is my only language besides English, and thinking of poem-setting as translation feels both apt and helpful. Alongside the formal pleasures I described at the start of this essay, I’m fascinated and driven by these ethical challenges – not only fitting a text to much in a technical sense, but opening myself to the ‘fraying’ of an interpersonal encounter. The stakes are small: I’m not a professional musician, and posting these little ‘poemsongs’ on my SoundCloud or performing them on the odd occasion is unlikely to raise much notice. For me, they really are just different ways of reading and spending time with poems I love. But even alone in my kitchen (the most echo-y room in the house), my body takes up these tangled relations every time I sing them.

That’s not to say I’ve always been satisfied with my creative-critical solutions. A few years ago, I set Langston Hughes’s poem, ‘The Negro Speaks of Rivers’. I’m not alone in adoring this poem as long as I can remember, first on the page, then after first hearing recordings in Hughes’s magnificent voice as an adolescent. I know that adoration grew further upon learning that this poem – which is about the interconnectedness of all rivers and the interconnectedness of Black global history – was conceived on a train journey when Hughes was just 17 years old, crossing the Mississippi River near my hometown of St. Louis. But I know these selfish personal connections could never override the difference between Hughes’s and my lived and embodied experiences. I was conscious of those differences while composing music for his poem, and in every private play-through, I’m trying to open myself to that alterity, in Spivak’s sense, trying to feel the poem as deeply as I am able.

My Hughes song is the one setting I have never posted online or performed for others. My reservations go beyond debates about who has the ‘right’ to sing certain texts or in certain musical genres. The word ‘Negro’ doesn’t appear in the poem, and there’s no other language that needs ‘translating’ in the way that Kris Delmhorst adapted James Weldon Johnson, although other signifiers inevitably take on different inflections in my singing. Structurally, Hughes’s free verse wasn’t easy to set either. Here’s the whole poem:

I’ve known rivers:
I’ve known rivers ancient as the world and older than the flow of human blood in human veins.

My soul has grown deep like the rivers.

I bathed in the Euphrates when dawns were young.
I built my hut near the Congo and it lulled me to sleep.|
I looked upon the Nile and raised the pyramids above it.
I heard the singing of the Mississippi when Abe Lincoln went down to New Orleans, and I’ve seen its muddy bosom turn all golden in the sunset.

I’ve known rivers:
Ancient, dusky rivers.

My soul has grown deep like the rivers.

The nine prose-like sentences, in wildly varying lengths with no regular metre or rhyme, are much better suited to ‘through-written’ classical settings (of which there are some great ones). In Venuti’s terms, I was conscious of the ‘violence’ I was inflicting on Hughes’s lines to make them fit a simpler song structure, as cutting the nine sentences into four regular verses with a bridge. Again, the process offered new formal insights. The repeated phrase ‘I’ve known rivers’ became an anchor point, and the only repeated line – ‘My soul has grown deep like the rivers’ – took on a different emphasis in the two different melodic phrasings I gave it. Long vowels were pivotal as ever, in useful contrast here with the soft ‘rivers’ at the end of several lines, or the grittier linguistic texture of proper names when Hughes starts naming world rivers. Again, though, the crossword-like puzzle was only one layer of the process.

I wouldn’t want to overstate either the technical or ethical difficulty of setting ‘The Negro Speaks of Rivers’. Nor would I claim any special competence in using musical settings to connect more deeply with so many poets whose lives and voices differ from my own. As with most creative and critical acts, poem-setting involves a complex interpersonal exchange. As the English composer Lennox Berkeley once wrote: ‘One has only to think what a composer has to do to a poem; he has to destroy or at best modify its natural rhythm.’ For this reason, he says, ‘it remains a risky undertaking on which one hesitates to embark.’[9] But I keep thinking about ‘setting’ as the name for this risky practice. At one level, it suggests a physical operation, with words or syllables being carefully ‘set’ or attached to notes like gemstones to a ring. It also makes me think of my grandmother having her hair ‘set’ every Tuesday, or of a person being otherwise set in their ways. Alongside that sense of rigidity, however, there’s something more open-ended about the idea of each poemsong as a unique spatial ‘setting’, or conceptual location in which these strange interactions play out. These imagined spaces also make room for reflecting on the critical understanding and creative decisions involved. When it works, exploring these settings leads to encounters outside our immediate experience – and, at best, a strange but meaningful conversation with another poet.

[1] Boykan., p. 124.
[2] Martin Boykan, ‘Reflections on Words and Music,’ Musical Quarterly 84, no. 1 (2000), 123-136, p. 123.
[3] Boykan, p. 123.
[4] Interview with Bart Eeckhout, ‘”Windblown Ceremonies in Search of Vocalissimus: An Interview with Matt Barber on His Composition To the Roaring Wind.” Wallace Stevens Journal, vol. 43 no. 2, (2019) 152-167, p. 154.
[5] Qtd. Brenda Ravenscroft, ‘Metamorphoses: Elliott Carter’s Musical Settings of Stevens’s Poetry’, Wallace Stevens Journal, 43, no.  2 (Fall 2019), pp. 234-238: 234.
[6] Yonatan Malin, Rhythm and Meter in German Lied (Oxford University Press, 2010), p. 4.
[7] Lawrence Venuti, The Translator’s Invisibility: A History of Translation, 2nd edition (Routledge, 2008), p. 18.
[8] Boykan, p. 123.
[9] Lennox Berkeley, Lennox Berkeley and Friends: Writings, Letters and Interviews, ed. Peter Dickinson (Boydell Press, 2021), p. 143.