Analysis Blog Historical Performance Musicology

Historical Performance: Brahms Op. 117 No. 1 by Adelina de Lara (1951)

The recordings made by Adelina de Lara, Ilona Eibenschütz and Etelka Freund are interesting case studies to consider. These are women whose professional performing careers were “curtailed by marriage” (Kim, 2012, p.52). This removal from performing life kept, to some degree, their performing practices away from the changing musical culture of the 20th century. Therefore, the recording I shall consider, and those of Eibenschütz and Freund, are 1950s recordings, but bear little resemblance to the prevailing performance practice of that time. Instead, their recordings offer more of an unadulterated insight into the styles and traits of 19th century performance, especially that imparted by Clara Schumann.

Adelina de Lara (1872-1961) studied with Clara Schumann from 1885-1891. This included a preliminary year studying with Fanny Davies, Marie and Eugenie Schumann (Adelina de Lara, 2010, 0:00-0:09; de Lara, 1945, p.145). Clara Schumann, in her diary, described de Lara as “one of my best pupils” (Adelina de Lara, 2010, 1:10-1:32). This is also attested to by Edith Heymann (1872-1960) who listed de Lara as one of Clara’s “most famous pupils” (Various Artists, 2015).

The practice of dislocation (a subset of “expressive asynchrony” (Dodson, 2011, p.59)), is a played separation of the hands (usually, the left hand (L.H.) before right hand (R.H.)) for vertically aligned notes in the score (Costa, 2012, p.45). This is a pervasive practice for solo piano recordings and piano rolls of the early 20th century. The technique is particularly used for “music of a slow and expressive character” to punctuate metrically significant beats or to emphasise particularly expressive passages (Costa, 2012, p.46; Dodson, 2011). The L.H. may precede the R.H. as if it were a grace note by pre-empting the beat, or, landing on the beat and so displacing the R.H. after the beat. These metrical discrepancies are dependent on the context of the musical passage and the rhythmic perception of the listener. Dislocation further serves to prioritise the melody; however, when the melody is in the L.H. (as in Chopin’s B-minor Prelude Op.28 no.6), Vladimir de Pachamnn maintains the ordering of L.H. before the R.H. (Dodson, 2011). Thus, the ordering of a dislocation is ‘bass then treble’, rather than ‘accompaniment then melody’.

The practice of dislocation is prevalent among many of the recordings available from the students of Clara Schumann. In the first intermezzo of Brahms’s Op. 117, de Lara most notably uses dislocation in the tenor melodies of the A section (bars 6-8 and 16-18) and bass notes on beat 4 throughout the B section. In Figure 1 below, the dislocations used are to prioritise the tenor melody through maintaining temporal distance from the bass voice an octave lower. This passage contains a pair of hairpins “<>” which Kim (2012, p.48) terms as “accent-type” hairpins. These symbols are used as an indication of “more” being required, whether that is vibrato, arpeggiation or dislocation. De Lara substantially increases her use of dislocations at this point in the piece. Therefore, her use of dislocation may be to prioritise the inner melody and respond to textual symbols.

Figure 1: Bars 6-8 of Adelina de Lara’s performance of Brahms Op. 117 no. 1 Intermezzo in E-flat major showing: unnotated arpeggiations (blue), dislocations (red grace notes), rhythmic alterations (green).

Carl Friedberg (1872-1955), an eminent pupil of Clara Schumann, also made a recording of this Intermezzo in 1953 (Friedberg, 2003). His recording is one perhaps more palatable to a contemporary audience; however, the ascending leap across bar 6-7 of E-flat to B-flat, contains similar dislocations. This consistency of dislocations at a certain point across performers and recordings may point towards a stylistic performing practice; rather than egregious, affected liberties taken by aged performers. Other than these more obvious usages of dislocation by de Lara, she also uses dislocation on the first notes and last notes of phrases of the A and A’ sections. This usage brings emphasis to the lilting ‘crotchet quaver’ bass rhythm; aligning with Brahms’s description of Op. 117 as “three cradle-songs of my sorrows” (Swafford, 1998, p.587). Furthermore, Brahms included the words, “Schlaf sanft, mein Kind, schlaf sanft und schön! Mich dauert’s sehr, dich weinen sehn” above the Intermezzo (Brahms, 1892, p.3). This translates to, “Sleep softly my child, sleep softly and well! It hurts my heart to see you weeping” (Scott, 2014, p.190). Therefore, the lilting, rocking quality of accompaniment is justly emphasised by de Lara.

A widespread technique used by performers of the early 20th century (at the very least) is that of unnotated arpeggiation. This is the spreading or rolling of a chord which may appear in three varieties: a chord spread from the lowest note to highest note in succession, a simultaneous spreading of left and right hands respectively, or the playing of the lowest note (or notes) before a block chord in the R.H. (closer to a dislocation) (Costa, 2012, pp.101-102). Unnotated arpeggiation is another subset of Dodson’s (2011, p.59) “expressive asynchrony”. Costa (pp.102-103) links pianistic arpeggiation with its keyboard ancestors, the harpsichord, virginal, spinet and clavichord. Its common usage and effects are that of softening the attack of chords, emphasising certain voices within the chord or drawing attention to a particular harmony (Costa, 2012, p.102). Unnotated arpeggiation, like dislocation, is used most noticeably in slow, expressive music.

Throughout the intermezzo, de Lara uses arpeggiation in all varieties and for all reasons listed above. In beat 6 of bar 11 she draws attention to the colourful secondary dominant, F7 by spreading the chord (see Figure 2). This method of emphasising poignant harmonies occurs also in bar 15 bb. 1-2 with the arpeggiation of the A half-diminished chord. De Lara’s frequent use of unnotated arpeggiations in bar 12 draws attention to the Brahms’s intricate counterpoint whilst softening dense chords. Costa (2012, p.187) highlights de Lara’s arpeggiation on the second dotted crotchet beat of bar 12, which produces an agogic emphasis. Once more, the use of two hairpins in bar 12 may have provoked de Lara’s increase in frequency of unnotated arpeggiations.

Figure 2: Bars 11-15 bb. 1-3 showing: unnotated arpeggiations (blue), dislocations (red), rhythmic alterations (green).

In bar 13, there is use of notated arpeggiations. The notated arpeggiations throughout this intermezzo only appear when the hand is required to play an interval greater than an octave. However, de Lara plays the notated arpeggiations in a much more present way than the unnotated arpeggiations. This may be observed in bars 13-14, where the notated arpeggiations are slower and the notes of the L.H. are more distinctly separate than the following and preceding unnotated arpeggiations. During the notated arpeggiations of bar 13, de Lara uses the variety of arpeggiation which pre-empts the beat so that the R.H. melody remains metrically strong. The R.H. chords are not spread but are placed at the apex of the arpeggiation to emphasise the melody.

In bars 38-41 of the slower A’ section, expressive asynchrony occurs more often than synchronicity.  Fanny Davies (1861-1934), one of Clara Schumann’s renowned students who trained Adelina de Lara for a preliminary year before starting with Clara (Adelina de Lara, 2010, 0:00-0:09) is discussed by The Observer in May 8th, 1927. Remarking on Davies’s playing of Brahms, “She spreads the chords instead of crashing them, and leads the melodies gently by the hand” (quoted in Scott, 2014, p.150). Indeed, upon listening to her 1928 recording of Robert Schumann’s A minor concerto, Op. 54, her purposeful unnotated arpeggiations of bars 12-19 soften the chords and draw attention to the most important notes (Davies, 1992).

There are many forms of rhythmic alterations employed by early 20th century pianists. Costa terms one form, “the rhythmic alteration of melody notes while essentially preserving the metrical regularity of the accompaniment” as “metrical rubato” (2012, p.189). Roudet (2014, p.75) terms this same technique as “linear rubato”. Both attribute this, in part, to the bel canto Italian singers “who had continued to apply sprezzatura and perform in the tradition of seconda practica” (Roudet, 2014, p.75). The combination of many forms of metrical and expressive asynchronies occurring simultaneously in de Lara’s recording make it difficult to assign a compartmentalised method of discussing each form of rubato. Instead, I shall deal with the more noticeable moments of direct rhythmic alterations, where the musical text has been altered by the performer.

Expressive asynchronies often manipulate the identity of the beat, pushing or pulling the idea of consistent metricity across, or even within, bars. Bars 8 and 45 of de Lara’s recording feature such strong dislocations and arpeggiations that the quaver crotchet rhythm of the B-flat octaves more closely resemble a demisemiquaver crotchet rhythm. The most prevalent use of rhythmic alteration by de Lara is that of altering the pair of semiquavers in A and A’. She lengthens the first semiquaver and the shortens of the second semiquaver of the pair consistently throughout the intermezzo (I have drawn attention to this where the pair of semiquavers appears in my previous and following figures). The swung rhythm further adds to the lilting quality of this lullaby (Scott, 2014, p.201).

Figure 3 highlights the variety and regularity of rhythmic alterations and expressive asynchronies which occur from bars 16-20. The <> hairpins from bars 18-20 are simultaneously that of the “lingering” and “closing” types (Kim, 2012, p.56). The lingering usage of the pair of hairpins is that of “lingering at the expressive peak”; in this case, bar 20 beat 1 (p.48). On this beat, there is a dislocation of a bass octave and treble chord which emphasises: the melodic high point, the harmonically interesting B-flat half-diminished chord, and metrical strong point. This harmony is lingered on, which I have indicated by a tenuto mark. Many of the uses of dislocation and unnotated arpeggiation have already been discussed. However, bar 20 poses difficulty in the nomenclature of L.H. techniques used. The rhythmic interaction between bass and treble in this bar could be in response to three stimuli: the close of the A section, the diminuendo hairpin and the rit. molto performance direction. The “closing” hairpin used is a means of slowing to bring closure to a phrase or section (Kim, 2012, p.48). The slower tempo indicated by rit. molto, lends more time for expressive asynchrony, which features more heavily in slower music (Costa, 2012, p.190). This rhythmic alteration may also link with Dodson’s (2009, p.90) findings that Paderewski reserved “his most radical revisions to the notated rhythms for passages of special significance in relation to the large-scale form”.

Figure 3: Bars 16-20 showing: unnotated arpeggiations (blue), dislocations (red), rhythmic alterations (green), tenuto (pink), mistake (gold).

Early recordings, particularly that of the solo piano repertoire, feature heavy and frequent modifications to tempo. Such modifications, if heard by performers today, may likely be viewed as erratic or indulgent (Costa, 2012, p.252). However, tempo modification for the purposes of expression are a resource frequently employed by performers who were directly acquainted with the 19th century composers of whom they were playing. Adelina de Lara alters employs various modifications to tempo at the level of the bar, the phrase, the phrase group and the section (Scott, 2014, 220-226). This is in response both to textual prompts and conveying the “meaning of the music”; a concept close to Clara Schumann’s pedagogy (de Lara, 1945, p.145).

The values of Figure 4 and 5 below are an imperfect method of measuring musical rubato as phrases frequently begin on the anacrusis, beat 6, of each bar. Thus, measuring phrase length, rather than bar length, may give a more truly reflective musical insight. However, bar length offers a regular value with which to compare, rather than phrases which may be inconsistent lengths. Bar length has been measured through manually plotting the beginnings of each bar using Audacity. I have rounded each value measured in seconds to the nearest tenth of a second and each percentage to the nearest integer. Figure 4 visually shows the bar number and corresponding bar length. A general trend towards slower tempos can be observed as the piece progresses. This graph also demonstrates the variability of tempo throughout the piece through the many, and large, peaks and troughs. Furthermore, it is evident that the minore B section (bars 21-37) contains the greatest variability of tempo.

Figure 4: Graph displaying “Bar Length (s)” on the y axis and “Bar Number” on the x axis. Higher values indicate slower tempi of bars.

From the tempo analysis of this recording, the three sections of this piece are distinctly different (see Figure 5). The A section is characterised by a brisk but flexible walking tempo, mostly in response to written indications (as in bars 15-16 and 20). The B section offers the greatest range of tempo with bars as short as 2.6 seconds (bar 23) and as long as 6.2 seconds (bar 29). Also, the B section has the largest mean percentage change between bars, 15%. This shows the temporal freedom and impassioned character of the B section in contrast with the comparatively more homogenous A and A’ sections. This may be in response to the textual direction at the beginning of the B section, “pp sempre ma molto espressivo” (Brahms, 1892, p.4) and frequent use of <> hairpins above the R.H. melodic peaks. These hairpins are usually expressed as lingering-type hairpins by de Lara in this B section. Each expressive peak is frequently joined with large emphasis placed on the dislocated bass. These strong bass notes may have had a different sonic quality in the late 19th century due to discrepancies between the straight-strung Viennese pianos and the cross-strung modern Steinway (Winter, 2004, p.38; Costa, 2018; Norris, 2011). The loudest dynamic written by Brahms throughout this whole intermezzo is p (aside from the rf at bar 53) and the predominant dynamic indicated in the B section is pp. De Lara, however, crescendos up to f (at the very least, mf) in bars 29-32. Thus, the initial marking of “molto espressivo”appears to have been prioritised over “pp sempre”.

Section and barsTempo indicationMean bar length (s)Range of bar lengths (s)Mean change between bars (%)
A (1-20)Andante Moderato3.32.9-4.4 (1.5)10
B (21-37)Più Adagio3.82.6-6.2 (3.6)15
A’ (38-57)Un poco piu Andante5.34.1-6.8 (2.7)12
Figure 5: Table showing the sectional differences of tempo across Adelina de Lara’s recording.

Similar tempo modifications can be heard by Ilona Eibenschütz (1872-1967) in her recording (1950s) of Brahms’s Op. 76 no. 4 Intermezzo in B-flat (Eibenschütz, 2012). However, Eibenschütz received blunt criticism from Clara Schumann regarding her tendency to go “too quickly over everything” (Clara Schumann quoted in Scott, 2014, p.156). According to Scott, Eibenschütz’s playing may have been more of a reflection on late Brahms’s rather than Clara’s (p.158). Eibenschütz’s recordings at an old age thus may have been “exaggerations of exaggerations” (p.158). However, her late recordings are not outside of the bounds of discourse compared to her compeers. As Crutchfield notes (quoted in Costa, 2012, p.306), “Friedberg’s performance and Eibenschütz’s are as different as night and day, but night and day in the same city”.

The consideration of de Lara’s recordings, along with other early 20th century recordings, offer insight into the techniques and values of those performers. It may be inferred that practices which may be viewed as departing from the score (unnotated arpeggiation, unmarked tempo and dynamic modifications) are instead based on a different interpretation of that score. Even the most fundamental of musical symbols, a hairpin or a vertically aligned chord can connote a different meaning, and therefore a different response from the 19th century performer rather than the 21st century performer. Therefore, the performer’s faithfulness to the text must be based on a comprehension of that text. Comprehension of a text may be an impossibility due to the relativistic, subjective nature of that text. Thus, the concept of textual fidelity may be more bleakly summarised as, “fidelity to an incomplete understanding of the text based on limited sources and cultural biases”. Nevertheless, the consideration of historical sources gives the 21st century performer the option to integrate, or disregard, various methods of interpretation in line with their desired performance aesthetic (Costa, 2012, pp.xxv-xxvii).

Kennaway (2011, p.82) notes the discrepancy of the 19th century critic speaking louder and more clearly (due to a physical artefact remaining) than the 19th century performer (of which no artefact remains). This leads to the dilemma of whether modern performers should play “as the clear majority of musicians at the time actually played, or as critics would have liked them to play” (p.82). Further to this, the consideration of a modern audience must be made. Winter (2004, p.19) notes, “issues of interpretation cannot ultimately be separated from the issue of for whom one is interpreting”. The response of a modern audience to a recording by Carl Reinecke may elicit laughter and ridicule, rather than connecting with and moving the audience. Thus, the historical meaning of the music has not been translated to a modern audience in order to convey the same effect intended by the composer (a problematic concept) in relation to their audience. According to Clara Schumann, textual fidelity is based on understanding the meaning of the music (Adelina de Lara, 2010).  Furthermore, one could posit the viability of a performance which considers the meta meaning behind why historical techniques were employed rather than what techniques were used. For example, an unnotated arpeggiation may be used to prioritise a melody which may instead be carried out through careful voicing.

It is difficult, therefore, to convey an entirely accurate connection between score and performance through an analysis of this recording. However, what may be inferred are the individual traits of this performance and its similarities and differences with performances by other performers born in the 19th century. The consideration of recordings made by other students of Clara Schumann offer a range of performance styles which may allow for inferences to be made of 19th century (or earlier) performance practice (Kim, 2012, p.48). These inferences may allow an individual interested in 19th century performance practice to go through a threefold process: emulation of recordings, internalisation of those techniques used, then application of those techniques onto other works where no early recording exists (Costa, 2015).

This method of the rather unfortunately initialised, ‘Recording Informed Performance’ (RIP) has been carried out by the pianist, Anna Scott. Through close study and analysis of the recordings by de Lara, Eibenschütz, Friedberg and others, she has replicated their performances (Leech-Wilkinson, 2017). This process has informed her playing of repertoire in a conventional performance style as well as a Recording Informed style. As Costa (2015, 45:30-46:13) and Anton, his student, notes, when practising these 19th century techniques, it is difficult to return to interpretations which more closely align with a 21st century performance aesthetic.

It may be concluded that the study of early 20th century recordings reveals insights into the performance practice of the 19th century. This performance practice is unlike the current methods of performance which prevail in the concert hall or are aspired to in competitions. In part, the expressive decisions made by 19th century performers were based on an alternative interpretation of the score than we may hold today. Gaining deeper insight into various methods of interpretation and alternative modes of expression offer untrodden paths to the performer of today; it allows one to gain a fresh outlook on familiar repertoire. In turn, this may invigorate listeners and stimulate generations of performers to come.


Adelina de Lara [gullivior]. (2010, Nov 18). Adelina de Lara – “Clara Schumann and her teaching” – Reminiscences and examples [Video file]. Retrieved from

Barolsky, D.G. (2007). The performer as analyst. Society for Music Theory, 13 (1), 1-13. Retrieved from

Brahms, J., (1892). Drei intermezzi: Von Johannes Brahms, Op. 117. Berlin: Simrock.

Costa, N.P.D. (2012). Off the Record. New York: Oxford University Press.

Daniel Barenboim [Daniel Barenboim]. (2017, Jun 9). Daniel Barenboim & Christoph Waltz on the fidelity to a text | Parallels & paradoxes part 1/4 [Video file]. Retrieved from

Davies, F. (1992). Piano Concerto in A minor, Op. 54: Allegro affettuoso – Andante espressivo – Allegro. On Fanny Davies plays Schumann [CD]. Sussex: Pearl.

de Lara, A. (1945). Clara Schumann’s teaching. Music and Letters, 26 (3), 143-147.

de Lara, A. (1991). Intermezzo, Op. 117, No. 1 in E Flat Major (Brahms). On Pupils of Clara Schumann – Fanny Davies, Ilona Eibenschütz, Adelina de Lara [CD]. Sussex: Pearl Records.

Dodson, A. (2009). Metrical dissonance and directed motion in Paderewski’s recordings of Chopin’s mazurkas. Journal of Music Theory, 53 (1), 57-94.

Dodson, A. (2011). Expressive asynchrony in a recording of Chopin’s prelude no. 6 in B minor by Vladimir de Pachmann. Music Theory Spectrum, 33 (1), 59-64. doi: 10.1525/mts.2011.33.1.59.

Dodson, A. (2012). Solutions to the “Great nineteenth-century rhythm problem” in Horowitz’s recording of the theme from Schumann’s Kreisleriana, op. 16, no. 2. Music Theory Online, 18 (1), 1-12. Retrieved from

Eibenschütz, I. (2012). Intermezzo, Op. 76, No. 4 in B-Flat Major. On Behind the Notes – Brahms Performed by Colleagues & Pupils [CD]. New York: Arbiter Records.

Friedberg, C. (2003). Intermezzo, Op. 117, No. 1 in E-Flat Major. On Carl Friedberg: The Brahms/Schumann tradition [CD]. West Chester: Marston.

Kennaway, G. (2011). Do as some said, or as most did? – A Foucauldian experiment with nineteenth-century HIP. Current Musicology, 92 (4), 79-101.

Kim, D.H.S. (2012). The Brahmsian Hairpin. 19th-Century Music, 36 (1), 46-57.

Lawson, C. & Stowell, R. (1999). The historical performance of music. Cambridge: Cambridge University Press.

Leech-Wilkinson, D. (2015). Challenging performance. Retrieved from

Leech-Wilkinson, D. (2017). Challenging performance. Retrieved from

Neal Peres Da Costa [ AustralianHaydn]. (2017, Oct 11). Dr. Neal Peres da Costa discusses Beethoven [Video file]. Retrieved from

Neal Peres Da Costa [ABC Classic]. (2018, Jul 30). Neal Peres da Costa talks about his Streicher piano [Video file]. Retrieved from

Neal Peres Da Costa [sydneycon]. (2015, Nov 4). About music lecture – Neal Peres Da Costa [Video file]. Retrieved from

Norris, D.O. (Presenter). (2011, Jun, 09). Brahms: Piano pieces (op. 118) [Discovering Music]. Retrieved from

Roudet, J. (2014). Frederic Chopin, Clara Schumann, and the singing piano school. In W. Brooks (Ed.) Ohne worte: Vocality and instrumentality in 19th-century music (pp. 65-107). Ghent: Leuven University Press.

Scott, A. (2014). Romanticizing Brahms: Early Recordings and the Reconstruction of Brahmsian Identity (Doctoral thesis). Retrieved from

Swafford, J. (1998). Johannes Brahms. London: Macmillan.

Various Artists. (2015). Recaptured by pupils & Colleagues. [Recorded by Various Artists]. [CD]. New York: Arbiter Records.

Winter, R.S. (2004). Orthodoxies, paradoxes and contradictions: Performance practices in nineteenth-century piano music. In R. L. Todd (Ed.) Nineteenth-century piano music (2nd ed.) (pp. 16-54). New York: Routledge.

By Connor Gaydon

I am a Music graduate and piano teacher for 5+ years. I have created a range of resources to broaden and deepen the knowledgebase of the Improving Pianists community.