This is a story about knowing when to leave well alone, about a composition that remained unfinished not because of writer’s block or an untimely composerly demise but because Charles Ives chose to finish his Concord Sonata by wilfully leaving it unfinished. We ought, Ives tells us, to think again about the boundaries of a work’s ‘completeness’.

Traditionally, Gramophone Collection pieces chew the critical cud, then, 3,000 words later, climax with an announcement of the recording to own. This Collection opens with disappointing news, though. There is no, could never be an ‘ideal’ Concord Sonata; indeed, any recording that aimed to rationalise Ives’s open-plan unfinishings, that came sealed with a coat of classical gloss, would, by definition, discount itself from the get-go. In his autobiographical Memos, Ives states that his sonata ‘is the only piece which every time I play it, or turn to it, seems unfinished’; moreover, ‘I may always have the pleasure of not finishing it’. What are we supposed to read into these words? Is the published score there for the performing? Or might there be scope for embellishing, extending or reimagining the basic text? In 1943, Ives turned up at a midtown Manhattan recording studio to record what, in modern-day parlance, amounted to demo recordings of Concord Sonata extracts. What clues lie within these time capsules? By committing himself to disc, even for promo purposes, was he ‘finishing’ his piece? In short: how to play the sonata when everything Ives said about it undermined the conventional trade-off between notation and sound, suggesting an afterlife beyond the page – a life that can only properly begin as the last notes fade to silence, and resonate through the imagination.

From Emily Dickinson to John Coltrane, true-to-its-roots American art has instinctively questioned definitive endings, the Western ideal of ‘finishing’: the perfect cadence of drawing a conclusive line under material. When Coltrane cut his first recorded version of ‘My Favourite Things’ in 1961 – duration 13 minutes – his evolving relationship with everyone’s favourite Sound of Music show-stopper was only just beginning. Captured live in Japan, five years on, Coltrane leads Rodgers and Hammerstein’s unassuming pop-song construct through a 60-minute spacewalk around the cosmology of sound itself. Meantime, Cage and Feldman were keeping their art in touch with the narrative margins by defying the convention of omniscient narrator. Pieces stopped without necessarily ending. And, as Jack Kerouac typed out his Beat road novel On the Road on a scroll rather than sheets of paper, the symbolism spoke volumes. The form is open-ended and cyclical; a road travelled is also a new beginning.

But artists communicate by floating their ideas towards our field of perception; art requires a frame, no matter how ambiguous, illusive or internally contradictory. Ives might have subtly renegotiated the relationship between notation and sound but, as composer, there was one convention he could never escape. He was reliant on interpreters; and potential interpreters relied on Ives framing his ideas – even ideas that he was determined would remain open-ended – with a score.

Its title-page helpfully lists the sonata’s four movements – ‘Emerson’ (page 1), ‘Hawthorne’ (page 21), ‘The Alcotts’ (page 53), ‘Thoreau’ (page 59) – and the ‘finished’ score of Ives’s Piano Sonata No 2, Concord, Mass, 1840-1860 looks reassuringly familiar at first. The only apparent deviation from normal publishing practice is lengthy scrolls of explanatory text parachuted in between each movement. These quotes, it turns out, are borrowed from Essays Before a Sonata, in which Ives sets out the back-story of his composition, originally published in tandem with the first edition of his score in 1920. He needed listeners to understand that the philosophical and musical core of his sonata were one. The Concord, we learn, is a philosophical investigation into sound and language: the appropriateness, or perhaps not, of music aspiring to express anything other than sound. His opening salvo – ‘How far is anyone justified, be he an authority or a layman, in expressing or trying to express in terms of music (in sounds, if you like) the value of anything, material, moral, intellectual, or spiritual, which is usually expressed in terms other than music?’ – is well worth trying to get your head around. In this sonata, Ives gives voice to that idea.

Unsolvable paradoxes

Philosophy spurs dormant and unused brain cells into action. Language and modes of expression shift. And, looking closer at the score of the Concord Sonata, Ives’s engagement with the writings of the New England transcendental philosophers – each movement named appropriately: after Ralph Waldo Emerson, Nathaniel Hawthorne, Amos Bronson Alcott and Henry David Thoreau – has palpably bled into the fabric of his piece. In Ives’s Piano Sonata No 1, finished in 1910, unbarred streams of consciously extending, compressing, collapsing phrases already sound like a music impatient with its loveless marriage to notational convention. The Concord Sonata moves that impatience to the next level, the whole composition but a lucky escape away from fracturing at the structural seams.

In his First Sonata, Ives doggedly rebooted his narrative with rapid and seemingly ad hoc edits; material heard in complex modes of transformation-juxtaposition slamming into the cathartic structural relief of verse-and-chorus responses. By 1920 that impulse had become super-refined, no longer an ‘effect’, now an integrated way of working with structure. As Ives sustains his invention for page after page without ruling a single bar-line – dropping indications into the mix like ‘rush it!’ or ‘very fast, in a kind of reckless way’ – notation is lent a new function. He is psyching pianists up to respond spontaneously, making sure there can be no comfort zones for lazy refuge, no possibility of ‘learnt’ spontaneity. Ives wanted his notation to sound as live as the moment you’re hearing it, and any pianist approaching the Concord Sonata must deal with an ultimately unsolvable paradox. The more accurately you play – the more truthfully you can zone into the music’s substructures – the more spontaneous and liberated the Concord Sonata is going to sound. But that accuracy cannot be prissy or judgemental; the piece must be allowed its unfinished, unanswerable questions. New England philosophy earthed Ives’s composition in preoccupations that moved the sonata principle outside itself. On no account, please, a European sonata that speaks with an American accent.

Another interpretative problem there for the solving – what to do with those ad-lib cameo parts for, during the first movement, solo viola and, at the very end of the piece, solo flute? Should pianists wrap the suggestion of ‘alien’ instrumental voices back into the work’s fantasy? Or take Ives at his word and include them? Different pianists take different views. The sonata archetype usually implies a self-contained musical universe; Ives’s music, though, was rarely self-contained. The sound of his music was the sound of quotation and allusions to hymns, ragtime, spirituals, marches et al, a characteristic inventory of sources, all feeding into the Concord Sonata alongside another form of borrowing too embedded to be considered mere surface ‘quotation’: the Concord Sonata is infatuated with the first four notes of Beethoven’s Fifth Symphony. It underpins harmonies and governs the trajectory of melodic contours. It sits inside every gestural pore as a centring truth that, during a particularly volatile moment, might reassert order or act like a pivot-point that abruptly twists the narrative 180 degrees. In Essays Before a Sonata, Ives compares Emerson to Beethoven: they are both ‘invaders of the unknown’. The Fifth Symphony’s opening call to action is ‘above the relentlessness of fate knocking at the door, above the greater human-message of destiny, and strives to bring it towards the spiritual message of Emerson’s revelations, the Soul of humanity knocking at the door of the Divine mysteries’.

Concord takes flight

As Charles Ives and his wife arrived at Mary Howard’s 49th Street recording studio on the morning of April 24, 1943, Howard (Toscanini’s recording engineer of choice) already knew the drill. During a 1969 interview, she described how Ives’s visits would invariably begin with the bad boy of American music busting through her rickety elevator doors yelling, ‘If they don’t know anything about music, well, all right, I’ll tell ’em!’

Ives had given Howard a list of likely performers to mail his demo discs to – ‘but it wasn’t a long list because there weren’t that many people who played his music’ – while making it clear that he wasn’t interested in taking copies home himself. During his performance of ‘The Alcotts’, the Concord’s still-point slow movement, Ives projects a unified and serene faux-choral tone. The opening phrase invokes way-back spirituals. A Celtic inflection momentarily stains the harmony tartan; then Ives gives the section marked ‘Slower and quietly’ a swinging-on-the-front-porch ‘homey’ lilt. Beethoven’s falling minor third becomes a portal through which to sample the world’s folk music. But there are also unsettling, distant shadows. With the opening theme reappearing as a midpoint fff climax, Ives is OCD-careful to make sure that harmonically conflicting grace notes, marked p, can punch above their diminutive weight. A climax with caveats; a climax not allowed to resolve or properly finish.

Ives’s playing is heartfelt but objective, a yin-meets-yang quality that wise performances embrace. The passage caricatured above as ‘Celtic’ is fleeting and Ives resists over-dramatising the moment, smudging it instead into the texture, implying that folk music of whatever ethnic source obeys the same universal harmonic and structural truths. During three brief extracts from ‘Emerson’, Ives hands pianists a timbral blueprint for the base sound he imagined: bangy attack, boozy rhythmic freedom; this is not the time or the place for consciously refined, ‘pretty’ playing.

John Kirkpatrick in 1945 understood this; German pianist Aloys Kontarsky in 1962 didn’t. Kirkpatrick’s recording was the first complete Concord Sonata on record, made by the only pianist in the catalogue who actually worked with Ives himself. And listening to the stinging lash of Kirkpatrick’s robustly articulated rhythms, his knack of sustaining simultaneously unfolding layers of event – his fingers panning between different strata as the harmony changes – this sounds like instinctively sympathetic, piece-specific playing. The sonata opens with a kaleidoscopic bang: all the material Ives will use in his piece spinning in flux, nailed into place over a network of references to Beethoven’s Fifth. Ives proceeds by gradually winding the energy levels down over the four movements, the flute theme emerging just before the end feeling like a resolution. But is it? As the closing bars return us to the harmonic world of the opening, but somehow not as aHammerklavier-like memory, Kirkpatrick taps into a deeper structural sleight of hand. The Concord Sonata’s temporal perspective is in fact distorted – it begins with a climax and signs off with a resolution that could also be a new beginning.

Kirkpatrick sounds fantastical throughout but includes neither viola or flute parts; Kontarsky includes both but sounds painfully literal. At 35 minutes, this is the fastest Concord around; and, while Kontarsky makes mincemeat of Ives’s technical challenges, he’s oddly in denial that the piece is anything unusual. The result is casual and colourless.

Supersonic Concord

In 1972 Roberto Szidon’s expansive view of the first two movements takes the work’s Beethovenian heritage as a starting point. The fugal fragments that Ives embedded into the first movement are like shards chipped from late Beethoven piano sonatas, Szidon’s driven and relentless trawl through the second movement constantly finding new horizons in the manner of a symphonic first movement. But this steer serves him less well during his cold and aloof ‘The Alcotts’, while ‘Thoreau’ needs him to take his foot off the accelerator.

Nina Deutsch’s 1976 performance has its technical vulnerabilities compared to Szidon but her understanding of the Concord’s inner poetry (she studied with Kirkpatrick) feels infinitely more authentic. Her ‘Emerson’ is no-nonsense and her warm-blooded ‘Alcotts’ demonstrates how tepid Szidon is. Louise Bessette’s by-the-book Romanticism in 1987 skates over the Concord’s surface, transforming ‘Hawthorne’ into throwaway Debussy. ‘The Alcotts’ is mechanistic, its climactic section lacking control. Gilbert Kalish, also from 1976, is bizarrely muted, inhabiting a squeezed middle-ground of dynamics and gesture.

Marc-André Hamelin has recorded the piece twice – like Kirkpatrick, although the latter’s 1968 remake has never been reissued – and Hamelin’s 1988 version has polarised opinion. The opening of ‘Emerson’ displays Ives’s wares with all the technical pizzazz of Kirkpatrick and Hamelin gives the contrasting 7/4 passage scope to breathe. The problem is, of course, that the Concord Sonata is exactly the sort of piece that gives Hamelin his technocratic jollies and here you pays your money and takes your choice; there’s no viola or flute (‘this sonata ain’t big enough for the three of us’ is the message) and Hamelin’s technique can take him places other pianists can’t reach. But sometimes this feels like an airbrushed photo of a great performance. Donna Coleman a year later takes an opposite view, hers being perhaps the most individual performance here. At 18 minutes – a whole six minutes longer than Kontarsky – her ‘Emerson’ is by far the longest and she rejects the idea of the Concord Sonata as corybantic blowout. Her relentlessly dark and brooding view reinvents Ives’s intentions; this boldness might not be everyone’s cup of tea but she makes an eloquent case.

Throughout the 1990s, new versions flowed forth. John Jensen, recorded in 1989, based his performance on a scholarly rethink of the available sources but his playing is bland and cautious. Nor do Robert Shannon and Per Salo, both recorded in 1990, bring much to the table: Shannon is worthy but unexciting while Salo, recording for a Danish label more famous for issuing neo-bebop, is taken down by an insipid-sounding piano and a curious lack of direction.

But Easley Blackwood, recorded in 1991, puts us back on course. Blackwood is also a composer and fuses a composerly nous for knowing what goes where with a crazed virtuosity and a soulful bout of whimsy, a quality lacking from too many Concord Sonata performances. The range of Blackwood’s timbral palette is deeply satisfying: the rough and tumble of ‘Emerson’ is met with a big-boned, craggy tone that during ‘The Alcotts’ migrates towards a delicate, barrel-organ-like wheeze. ‘Thoreau’ – with the flute part included – reaches a profound stillness.

Nicholas Zumbro’s 1992 performance is a composite from various live concerts. ‘Emerson’ is deliberate and has expressionist undertones that don’t feel altogether apt for ‘Hawthorne’; however, ‘The Alcotts’ is all sweet and tender. From 1995, René Eckhardt and Alexei Lubimov are chalk and cheese – or rather cheese and chalk. If Hamelin offers glossy photography, Eckhardt mails a picture postcard. Lubimov, like Donna Coleman, pursues a very personal path. He shifts the weight of the sonata towards a monumental view of ‘The Alcotts’, de‑emphasising the usual supremacy of ‘Emerson’ by taking a more measured perspective. Sophie Cherrier makes an especially lovely job of the flute solo.

Fresh perspectives

In 1999 Philip Mead re-energised ideas about the Concord Sonata with an off-the-leash recording that leaves the thrills multiplying. The booming roar of his playing gets slightly compromised by an inadequate recording rig but this beast can’t be tamed. Pow! The opening passage mines the coalface of Ives’s textures, Mead needing you to hear all the interconnections of material. Bash! ‘Hawthorne’ is forcibly removed from anything to do with Debussy, Mead needing you to understand the cumulative impact of Ives flitting between helter-skelter arpeggios and the grounding of ragtime and marching-band gestures. And then rest. ‘The Alcotts’ is kindly without descending towards sentiment.

Pierre-Laurent Aimard and Ciro Longobardi, both in 2004, also offer broadly ‘modernist’ views. Aimard, though, already feels time-locked, mainly because he micromanages the material and kills the impetuous nature of the Concord Sonata stone dead. Longobardi, by contrast, feels liberated by Ives’s notation – precisely the right approach. His ‘Alcotts’ is clearly modelled after Ives’s own recording; ‘Emerson’ in particular sits on a slipstream of ebbing and flowing sound that only the moment seems to define.

Four further recordings – Steven Mayer (2002), Hamelin again, Roger Muraro (both 2004) and Jeremy Denk (2010) – locate the Concord as deriving from 19th-century Romanticism. Hamelin’s second attempt feels middle-aged and stuffy; for sure his technique is exemplary but his ‘Alcotts’ now sounds glib and maudlin. Denk’s Lisztian Concord Sonata is an intelligently thought-through account that, curiously for an American pianist, is determinedly European. Denk abhors rough corners and this clean-cut account is one for those who might find Mead’s or Longobardi’s efforts excessive. Muraro manages to make the Concord Sonata sound dull – quite an achievement, as he drizzles caster sugar over ‘The Alcotts’. Of these ‘mainstream’ views, Mayer stands out. He is motivated by detail: every reference to Beethoven’s Fifth is cannily nudged to the surface but without sounding like playing inside inverted commas. There’s no viola or flute but Mayer evokes them expertly, persuading me that, just perhaps, the knowing suggestion is better than the reality.

An ultimate Concord?

As trailed at the beginning, basically an impossible question to answer. John Kirkpatrick must be the starting point, after which the field is pretty much open. With Kirkpatrick’s second recording in vinyl-only purdah, Hamelin’s first recording is the moment that the tradition of Concord interpretation steps up a gear. I admire the idiosyncratic natures of Donna Coleman, Easley Blackwood and Alexei Lubimov; they get the Ivesian spirit, leaving Eckhardt, Muraro et al floundering. Then again, it could be argued that their approaches, Blackwood being the honourable exception, miss the larger picture. What’s really needed is some electronic mash-up of Mead and Mayer; but, for its sheer visceral excitement and bloody-mindedness, the version I’d take off my shelf more readily than the others is Mead’s.

A footnote: issued in 2011, Michael Tilson Thomas and the San Francisco Symphony’s recording of Henry Brandt’s four-decades-in-the-making orchestration of theConcord Sonata is an intriguing prospect at first but simply doesn’t fly. Ives’s implausible demands are dragged into the realm of the plausible. Another fable about knowing when to leave well alone.

Recommended recordings

A slice of history

John Kirkpatrick 

(Soundmark)

Kirkpatrick is our only primary source; he worked with Ives, heard Ives play, and his kaleidoscopic take on the Concord Sonata stings.

A technical blueprint

Marc-André Hamelin

(New World)

Hamelin’s first recording is polite and polished – overly so perhaps – but his technical might is relentlessly revealing.

A mainstream marvel

Steven Mayer

(Naxos)

Mayer’s hyper-detailed account is free of dogma and ideological point-scoring. Some may want an aesthetically edgier view but Mayer’s sincerity is persuasive.

The top choice 

Philip Mead

(Metier)

Philip Mead’s performance is loud, rude and jammed with idiosyncratic corners. Charles Ives would have been hugely appreciative and admiring.

Selected discography

Date / Artists / Record company (review date)

1943 Ives (excs) / New World NW80642

1945 Kirkpatrick / Soundmark (available from Amazon)

1962 Kontarsky, Plümacher, Schwegler / Wergo WER6940-2

1972 Szidon, Stangler, Sonntag / DG 477 5439GTA2 (6/72R)

1976 Deutsch / Vox CDX5089

1976 Kalish Nonesuch AW71337

1987 Bessette / CBC MVCD1041

1988 Hamelin / New World NW80378 (9/89)

1989 Coleman / Et’cetera KTC1079 (11/90)

1989 Jensen / Music & Arts CD630 (11/90)

1990 Shannon / Bridge BCD9036

1990 Salo / Kontrapunkt 32046 (2/91)

1991 Blackwood / Cedille / CDR90000 005

1992 Zumbro / Concert Artist CACD9005 (nla)

1995 Eckhardt, Pameijer / Brilliant Classics 9135

1995 Lubimov, Verney, Cherrier / Apex 0927 49515-2 (5/97R; 5/03)

1999 Mead / Metier MSVCD92037

2002 Mayer / Naxos 8 559127

2004 Aimard / Warner 2564 60297-2 (6/04)

2004 Longobardi / Limen CDE06-CD07

2004 Hamelin / Hyperion CDA67469 (11/04)

2004 Muraro / Accord 480 0862

2010 Denk / Think Denk Media TDM2567

This article originally appeared in the November 2012 issue of Gramophone.

Explore Charles Ives's life and music...

Charles Ives: lonely American giant

'He plunged ahead solely on the basis of his ear, his stamina, his conviction, his talent and his need to create' (John McClure, Gramophone, April 1967)... Read more