What Should a Performer Decide When the Score Does Not Say Everything?

Building a Piano Piece · Season Three: From Score to Interpretation · Article One

If a score already contains pitch, rhythm, dynamics, slurs, articulation, tempo words and pedal marks, why can two reliable pianists sound so different in the same work?

One common answer is that the score does not contain the music and the performer must add feeling. This makes the problem too permissive. A score does not fix the duration of every note in milliseconds, actual sound level, pedal depth, voice balance or local timing. Yet the unwritten is not empty space. Genre, period, instrument, notational practice, form, harmony and internal repetition constrain the choices that can be persuasive.

The opposite answer—“be faithful and add nothing”—cannot literally be carried out. Selecting an average tempo, determining the overlap between two tones, foregrounding one note in a chord and breathing in a long phrase are interpretive decisions. A supposedly uninterpreted performance generally follows conventions that have become invisible to the player.

The performer needs neither unlimited freedom nor the pretence that no choice exists. A responsible order is possible: establish the text, identify historical and structural constraints, propose an audible solution, and test the result in performance.

First ask who wrote the mark on the page

The score on the desk is not automatically the composer’s manuscript. It may combine a composer’s text, first-edition engraving, later correction, editorial addition, pedagogical fingering, performer pedalling and house-style standardisation. Identical print colour gives these elements an appearance of equal authority that they do not possess.

An Urtext edition is not an unmediated autograph. Editors compare manuscripts, proofs, first editions, composer-corrected copies, secondary copies and documents. They establish relations among sources, select where sources conflict, and report variants. Editorial policies for the C. P. E. Bach complete edition explicitly discuss variants in pitch, rhythm, duration, articulation, ornament, dynamics, tempo and slurring. They also acknowledge that no critical edition is simply identical to one autograph or to a facsimile of all sources. The guidelines describe an argument, not the automatic recovery of an infallible text.

The Chopin National Edition similarly explains that it draws on autographs, editions corrected with the composer’s participation and pupils’ copies carrying Chopin’s annotations. Its performance commentary addresses ornaments, pedal and differences between historical and modern pianos. The edition’s official account shows that an authentic-text project can still encounter different compositional stages, publishing territories and teaching contexts.

When a fingering or pedal appears, therefore, the first question is not “Must I obey?” It is whether the mark belongs to the composer, a relevant source, a critical editor or a modern pedagogical layer. If the edition does not distinguish them, comparison with a source-critical edition often resolves the problem before an appeal to personal taste begins.

Divide score information into three layers

The textual layer: facts not casually altered

Pitch, basic rhythm, voices, repeats and explicit composer-supported dynamics, articulation and tempo relations constitute the text of the edition being used. When sources contain genuine alternatives, the performer should know which is being adopted rather than mixing them accidentally.

Textual faithfulness does not turn notation into mechanical sound. A crotchet establishes relative duration, not milliseconds. A p creates a dynamic relation, not a decibel value. A slur proposes grouping or articulation that still has to be realised at a tempo, on an instrument and inside syntax.

The convention and context layer: unwritten but historically bounded

Ornament starts, dance accent hierarchy, Classical articulation, Romantic voice displacement and forms of rubato may require treatises, instruments and records close to the work. C. P. E. Bach’s Essay on the True Manner of Playing the Clavier covers fingering, ornaments and accompaniment and remains an important eighteenth-century keyboard source. A new critical text and English translation also clarify its publication history.

A German keyboard treatise first published in 1753 cannot automatically determine Debussy, nor can every statement be transferred indiscriminately to every genre of J. S. Bach. Historical sources provide bounded evidence.

The interpretive layer: a solution for which the performer is responsible

Average tempo and its local changes, foreground and background voices, dynamic trajectory, note overlap, pedal timing, difference on a repeat and proportion of climax require a solution. Calling a solution personal does not exempt it from explanation. What relationship in the work does it answer? Does it conflict sharply with relevant context or instrument? Can the result be heard? Can the player reproduce it?

What the score leaves open is rarely one isolated parameter

Tempo changes movement, articulation and pedal together. At a faster rate, the same written value has a shorter physical duration, and staccato need not preserve the same proportion. Resonance accumulates differently, perhaps requiring less pedal, and movement is reorganised. Experiments on piano rate have found changes in temporal variability, key velocity and finger motion as tempo increases. Controlled-task research cannot select the tempo of a sonata, but it establishes that tempo is not an independent dial.

Dynamics similarly cannot be realised by tracing printed marks up and down. Research modelling expressive dynamics across piano and orchestral corpora finds relationships among notated dynamics, pitch and structural features, while even sophisticated models do not exhaust human interpretation. The authors note that timing, loudness, articulation and timbre communicate structure and affect.

A long printed crescendo therefore asks more than “become louder”. The player determines its point of departure, rate of growth, contribution of each voice, relation to motion, and release after arrival. Each decision changes the conditions of the others.

A five-stage interpretive experiment

1. Freeze the present text

Record the edition and its source character. Distinguish composer-supported marks, editorial additions and personal pencil by separate signs. Where an important variant exists, choose one for the current experiment and preserve the alternative for comparison.

2. Make an unembellished baseline

At a controllable tempo, perform a short passage with accurate pitch, rhythm and basic notated dynamics and articulation, without deliberately copying a pianist. The recording is not an “objective version”. It exposes the unacknowledged choices already being made.

3. Map structural constraints

Mark phrase, harmonic arrival, repetition and change, voice transfer, metrical weight and registral turn. Beside every planned performance change, write a reason: “lengthen before the cadence because the suspension and phrase both resolve”, not “more feeling here”.

4. Create two distinct, defensible versions

Version A might preserve temporal continuity while using dynamics and voicing to communicate structure. Version B might introduce clearer timing at major boundaries while narrowing the dynamic range. Do not alter every parameter simultaneously; otherwise the cause of the audible difference cannot be identified.

5. Remove the explanation, blind-listen and cold-test

Randomise filenames and decide which version communicates the intended relation before reading the notes. Reconstruct the selected solution next day without replaying it. A choice that exists only immediately after imitation or discussion has not become a stable interpretation.

Bach’s C-major Prelude: decisions remain in a sparsely marked score

Much of the C-major Prelude, BWV 846, from Book One of The Well-Tempered Clavier consists of continuous broken-chord figuration. A public-domain score is available through IMSLP. Without adding a note, a performer still selects basic tempo, harmonic direction within and across bars, bass–upper-voice relation, articulation of the recurring pattern, timing near cadences, and the use of pedal on a modern piano.

“Baroque objectivity” and “modern-piano Romanticism” are not adequate answers. Play without pedal to hear the harmony, then compare versions with pedal renewed at harmonic boundaries, with extremely shallow resonance, and without pedal. Compare a small dynamic arch in every bar with a longer trajectory across several harmonies. Retain the solution that keeps harmonic movement audible without turning every bar into an identical breath.

No recording can complete the judgement. Recordings offer possibilities; the text, harmonic route, decay of the modern piano and the player’s control must still meet again.

Three forms of overreach

Textual overreach changes a note, removes a repeat or adopts an editorial addition while claiming composer authority. The remedy is to name source and choice.

Historical overreach turns one treatise or pupil reminiscence into the single practice of an era. The remedy is to limit author, date, place, genre and instrument and seek corroborating or conflicting evidence.

Capacity overreach imagines a persuasive solution that can only occasionally be realised, or requires loss of pulse, sound and physical control. Interpretation must enter a playable range; tempo or technique may need revision rather than naming loss of control freedom.

Where does final responsibility lie?

Scores, critical editions, historical sources and recordings narrow the field without producing sound on the performer’s behalf. Performance research describes uses of timing, loudness, articulation and timbre without deriving the unique sound of one bar from data.

The unwritten should therefore be decided in this order: identify the textual layer; build constraints from period, genre, structure and instrument; create comparable solutions; select through listening, repeatability and independent cold testing; revise when new evidence or a new instrument changes the result.

Interpretation does not fill a blank left by the score. It makes proportionate choices among incomplete forms of evidence and accepts responsibility for them in audible performance.

Primary sources and further reading

• Editorial guidelines for the C. P. E. Bach complete edition
• Sources and performance commentary of the Chopin National Edition
• Critical text and English translation of C. P. E. Bach’s Essay
• Modelling expressive dynamics in classical piano and orchestral performance
• Effects of tempo on piano timing, key velocity and finger movement

Continue reading: Building a Piano Piece series page

If you would like to bring these ideas about listening, understanding, and practice to the keyboard, you might try ScoreFlow, an app I developed to make score reading and daily practice flow more naturally together.