Bach: The Sonatas for Harpsichord and Viola da gamba, BWV 1027 – 1029

Media Review

2023-08-13 — Original posting

Bach: Viola da gamba Sonatas BWV 1027-1029, Boulanger/De Pasquale, CD cover

Table of contents

Johann Sebastian Bach, ca. 1722
Johann Sebastian Bach, ca. 1722

Bach’s Sonatas for Harpsichord and Viola da gamba

This posting is about the three Sonatas for Harpsichord and Viola da gamba, BWV 1027 – 1029 by Johann Sebastian Bach (1685 – 1750). Bach apparently wrote these sonatas in Leipzig. A manuscript only survived for the first one, BWV 1027. The time of composition is not known exactly. However, the general assumption is that the sonatas are compositions from the late 1730s or from the early 1740s. I won’t discuss the works in detail here. Let me just give the list of the movements for the tree sonatas:

No.1 in G major, BWV 1027

  1. Adagio (12/8)
  2. Allegro ma non tanto (3/4)
  3. Andante (4/4)
  4. Allegro moderato (2/2)

No.2 in D major, BWV 1028

  1. [Adagio] (3/4)
  2. [Allegro] (2/4)
  3. Andante (12/8)
  4. Allegro (6/8)

No.3 in G minor, BWV 1029

  1. Vivace (4/4)
  2. Adagio (3/2)
  3. Allegro (6/8)

The Recordings — An Overview

I currently have three recordings of the above sonatas on CD:

Two legacy recordings on LP (my basement archive):

The latter two are not in the scope of this comparison post. Instead, before discussing performances, I’m using these to present a bit of personal history with recordings of Bach’s Sonatas for Viola da gamba:

A Personal Look Back

In my high school years, music (LP) purchases were usually rather coincidental: in hunter-gatherer manner, I collected whatever I came across (and was affordable). One typical example is from my second stay in the city of Paris in spring 1971. We (a colleague of mine and myself) were staying in cheap hotels in the Quartier Latin, exploring the city all on foot. In one of these long strolls I ran into a student’s LP shop on Boulevard Saint Michel (not too far from the river Seine).

Early LP Purchases

I ended up spending hours in that shop, satisfying my hunger for music by listening to LPs on end (i.e., until the shop personnel stopped me). I suspect that the shop’s repertoire was restricted to cheap and/or local labels. However, I clearly enjoyed the music nonetheless. And, of course, as the prices were very affordable, I did purchase a few LPs:

The selection of these LPs was by no means the result of a critical evaluation. The audio quality on the shop’s hand-held earphones was horrible. However, I just liked the music (could not get enough of listening to Bach’s Concerto for Three Harpsichords in C major, BWV 1064!). You might ask: why Ruggero Gerlin? Well—because that’s what they had in the LP rack…

André Navarra & Ruggero Gerlin, 1964

Bach: Sonatas for Viola da gamba — Navarra, Gerlin (LP Cover)
Bach: Sonatas for Viola da gamba — Navarra, Gerlin (LP Cover, back)

Bach: “Sonates pour violoncelle et clavecin“, BWV 1027 – 1029

André Navarra, Cello
Ruggero Gerlin, Harpsichord

Musidisc RC 796 (LP, stereo); 1964


LP Contents

The sonatas in this recording are labeled as “Sonatas for Cello and Harpsichord“:

  • No.1 in G major, BWV 1027 [14’15”]
  • No.2 in D major, BWV 1028 [6’13” (I/II) + 9’26” (III/IV) = 15’39”]
  • No.3 in G minor, BWV 1029 [14’14”]

The Artists

The artists are the cellist André Navarra (1911 – 1988) and the harpsichordist Ruggero Gerlin (1899 – 1983). I always wondered why the LP cover mentions the conductor (and violinist) Roland Douatte (1921 – 1992) as Artistic Director. Was this just a prominent name that was expected to help increasing the LP sales?

A Legacy Interpretation

The recording and the comments on the LP cover are probably indicative of the Bach reception in the 1960s in general, possibly in France in particular. For one, the text (see below) claims that the sonatas date to different periods, originating from Cöthen, and possibly the first Leipzig years. The current Wikipedia entry on the sonatas indicates otherwise, based obviously on newer research.

Moreover (and more importantly), the LP text simply calls these “Sonatas for Cello and Harpsichord”. The word “viol” or “viola da gamba” doesn’t even appear:

Johann Sebastian BACH, a true musical genius by reason of his highly important works takes part in all periods and in all styles.

His cello and harpsichord sonatas we are glad to have recorded for you express excellently Bach’s manner when writing for such instruments.

The striking rendering by André NAVARRA as a cellist and by Ruggero GERLIN as a harpsichord player enhances still further the beauty of such works.

The first two sonatas comprise the same movements in identical order. The part of the harpsichord is written out completely. It is difficult to execute since a three-voiced polyphony determines the melodic intervals and calls for constant fingering changes. The harpsichord part is, therefore, rather seldom in hand.

In the last sonata, the distinctive language is borrowed from Vivaldi’s concertos. For Bach, the fugal imitative style remains a natural means of expression, but he tends more and more to keep only its indispensable elements: the principle of harmonizing a theme or subject by superposition of a second theme—or counter-subject—, thus avoiding identical reexposition and through the perpetual renewal giving a more lively impression. From one sonata to the next, there is undoubtedly an evolution towards an aesthetic which is freer from the traditional polyphonic concert and leads to the concerto, where one part of the structure will play a dominating role. This progression is particularly evident due to the fact that this sonatas were published in chronological order of their composition: the first, in G major, dates from the Weimar period (1715–1717), the second from Cöthen (1717–1719), and the third either from his last years at Cöthen or the first Leipzig years.

(Typos corrected by the author).


Ruggero Gerlin was a well-known harpsichordist around the middle of the 20th century. He was a pupil of last century’s pioneer of the harpsichord, Wanda Landowska (1879 – 1959). One can sense this background from his playing—but even more so from his choice of instrument: not the monstrosities that Landowska was playing, but most likely a Neupert instrument, model “Bach”. Despite the model name, these have very little in common with the instruments that were in use at Bach’s time. Their sound—well, horrible, synthetic, characterless—at least to my ears now.

Sound, Experience

I only had vague memories of how that recording sounded. However, I just found a video recording with the two artists from 1964, from a performance in Bordeaux, around the time of the LP recording. The video covers the third movement of BWV 1029. Needless to say that André Navarra’s cello playing and sonority has very little in common with the sound / sonorities of a viola da gamba. A great cellist of his time, undoubtedly—but his interpretation clearly is one of the past.

The traditional, legato, vibrato-rich cello sonority dominates the performance—fortunately: anybody with even just a moderate affinity towards historically informed performances (especially on period instruments) will likely find the harpsichord sound rather horrible (see also above)

Needless to say: I haven’t listened to that LP in decades, and I don’t have the least desire to unearth that sound again now…

I should state, though, that when I selected that recording, I hadn’t yet experienced the beauty and the richness of historically informed performances!

Edmund Kurtz & Frank Pelleg, 1966

Bach: Sonatas for Viola da gamba — Kurtz, Pelleg (LP Cover)
Bach: Sonatas for Viola da gamba — Kurtz, Pelleg (LP Cover, back)

Bach: The “Sonatas for Cello and Harpsichord”, BWV 1027 – 1029

Edmund Kurtz, Cello
Frank Pelleg, Harpsichord

Monitor Records — Monitor Collectors Series (LP, stereo); 1966


I don’t even have traces of memory why and how in January 1988 I added this second recording of the Sonatas BWV 1027 – 1029. I can only guess that it must have been a gift, because at that time, I had long discovered historically informed performances in general, but most definitely for baroque and older music. Actually, this recording turns out to be from 1966, i.e., from around the same time as the recording with André Navarra / Ruggero Gerlin.

LP Contents

Also here, the sonatas in this recording are labeled as “Sonatas for Cello and Harpsichord“:

  • No.1 in G major, BWV 1027 [13’45”]
    • I. Adagio [4’04”]
    • II. Allegro ma non tanto [3’32”]
    • III. Andante [2’42”]
    • IV. Allegro moderato [3’12”]
  • No.2 in D major, BWV 1028 [15’30”]
    • I. Adagio [2’00”]
    • II. Allegro [4’18”]
    • III. Andante [4’40”]
    • IV. Allegro [4’15”]
  • No.3 in G minor, BWV 1029 [16’25”]
    • I. Vivace [5’55”]
    • II. Adagio [6’05”]
    • III. Allegro [4’10”]

The Artists

The artists here are

  • Russian-born Australian cellist Edmund Kurtz (1908 – 2004), a pupil of Julius Klengel (1859 – 1933), and
  • Frank Pelleg (1910 – 1968), composer, conductor, pianist and harpsichordist. Pelleg was born in Prague. He emigrated to Palestine in 1936. In this recording, Frank Pelleg played a harpsichord by William de Blaise (1907 – 1978)—an instrument which followed the pattern / concept of the instruments by Neupert.

As indicated, that’s another cello interpretation. Here, at least, the LP cover mentions the viola da gamba as the instrument that Bach assigned these works to. The text refers to Leopold, Prince of Anhalt-Köthen (1694 – 1728), for whom (according to the text) Bach wrote these compositions:

(…) From all accounts, he [Ludwig] was a tolerably good player of the clavier and violin, and quite an expert on his favorite instrument, the six-stringed viola da gamba. A fair number of Bach’s solo sonatas were created for such princely performance, and it seems safe to assume that the first interpreters of the three Sonatas here recorded were Prince Leopold and Herr Kapellmeister Bach. (In modern performance, the scores are most often played on the cello, which superceded the viola da gamba, even in Bach’s own lifetime ).

Significantly, cello and clavier converse as equals in all three Sonatas. Most other composers in a similar situation would surely have felt it the better part of valor to flatter their bosses with flamboyant solo parts, set against a discreet keyboard accompaniment. It speaks well, both for the Prince’s humility and Bach’s musical integrity, that such a procedure was not followed here.

Themes are set forth by both instruments, and both carry forward the polyphonic threads—generally with the keyboard developing two separate lines, and the cello adding a third.

(Typos corrected by the author).


I have definitely not listened to that recording for at least 40 years. At most, I have played it once or twice between 1988 and 1992. Needless to say that I have no memory of that interpretation. And as with the interpretation by Navarra & Gerlin, I don’t have any desire to unearth that performance (again) now…

The Recordings on CD

The following table lists timings (seconds) and ratings (1 .. 5) for each of the recordings, per movement. A description of the media and their content is found in the sections below, together with comments about each of the interpretations.

Bach: Sonatas for Harpsichord & Viola da gamba, BWV 1027 – 1029: timing & ratings table (© Rolf Kyburz, all rights reserved)
Bach: Sonatas for Harpsichord & Viola da gamba, BWV 1027 – 1029: timing & ratings table (© Rolf Kyburz, all rights reserved)

As stated, the movement times are in seconds (minutes in the summary line). For the last movement in each sonata, the timing in the list is without trailing blank time. Blue indicates a slower performance (longer duration), red a faster / shorter performance. The color coding in the rating columns should be obvious.

A key observation with the timings / tempo is that with one exception (the Andante in the Sonata No.1, BWV 1027), the interpretation by Anner Bylsma / Bob van Asperen (violoncello piccolo / organ) is more fluid / faster. This is not significant insofar as the difference in instrumentation almost inherently leads to different tempo choices. The two interpretations with the original instrumentation (viola da gamba / harpsichord) are very much in line with each other. The Andante in the Sonata No.1, BWV 1027, once more is the only exception here.

I decided also to cover the extraneous pieces in the recordings with Anner Bylsma / Bob van Asperen and Lucile Boulanger / Arnaud De Pasquale, including even comparisons with “external” references. In order to retain clarity for the reader, I am using a lightly tinted background for these performance comments.

Jordi Savall & Ton Koopman, 1977

Bach: Viola da gamba Sonatas BWV 1027-1029, Savall/Koopman, CD cover

J.S. Bach: The Sonatas for Viola da gamba, BWV 1027 – 1029

Jordi Savall, Viola da gamba
Ton Koopman, Harpsichord

Virgin Classics 7243 5 62065 2 6 (CD, stereo ADD); ℗ 1978 / © 2002; digital remastering © 1989 EMI Electrola
Booklet: 8 pp. en/fr

Bach: Viola da gamba Sonatas BWV 1027-1029, Savall/Koopman, CD, UPC-A barcode
amazon media link

CD Contents / Track Listing

The CD exclusively contains the three Sonatas for Harpsichord and Viola da gamba, BWV 1027 – 1029:

  • No.1 in G major, BWV 1027 [13’28”]
    • I. Adagio [3’59”]
    • II. Allegro ma non tanto [3’46”]
    • III. Andante [2’26”]
    • IV. Allegro moderato [3’17”]
  • No.2 in D major, BWV 1028 [14’57”]
    • I. Adagio [1’52”]
    • II. Allegro [4’02”]
    • III. Andante [4’50”]
    • IV. Allegro [4’13”]
  • No.3 in G minor, BWV 1029 [15’26”]
    • I. Vivace [5’30”]
    • II. Adagio [5’59”]
    • III. Allegro [3’56”]

Artists & Instruments

None of the artists needs to be introduced here. Both have been central to the historically informed performance scene over the past decades:

The Catalan conductor, composer, and gambist Jordi Savall i Bernadet (*1941) not only plays a key role as conductor (primarily for baroque music), but he is also one of the doyens among Europe’s viola da gamba players. I have written about him in both functions (conductor, gambist). Here, he performs on a 7-string 1697 viola da gamba by Barak Norman (1651 – 1724), London.

The Dutch Organist, harpsichordist and conductor Ton Koopman (*1944) has been a frequent “guest” in my blog, be it as organist, at the harpsichord, or as conductor—see my earlier reviews. In this recording, he performs on a 1747 harpsichord by Sébastien Garnier.

Recording Date and Location

The recording took place on 1977-03-01:03 at Claydon House, Buckinghamshire, England, U.K.


Comments on the Performance

Sonata No.1 in G major, BWV 1027

I. Adagio (12/8): Ton Koopman solves the challenge of the long notes by discreetely / inconspicuously repeating the note once of twice. Jordi Savall’s playing is very expressive and highly differentiated—expectedly excellent. Jordi Savall is of course able to make long notes grow without the use of prominent vibrato (which he does use, but selectively, and very subtly). Unfortunately, the recording puts the viola da gamba too much into the foreground, which makes the harpsichord sound pale, lacking color, volume and brightness / presence.

II. Allegro ma non tanto (3/4): A better tempo (very slightly more vivid) than Boulanger / De Pasquale, but on the other hand, the latter offer the better sonority & sound.

III. Andante (4/4): The most fluid of the interpretations, rarely a slight excess in “pushing forward”, lacking some calm. Does not “breathe” quite as much as with Boulanger / De Pasquale.

IV. Allegro moderato (2/2): Can’t quite match the sonorous richness of the most recent recording, nor the beauty of its agogics and articulation.

Rating: 4.5 / 4 / 4.5 / 4 = 4.25

Sonata No.2 in D major, BWV 1028

I. [Adagio] (3/4): The harpsichord sonority can’t compete with the one in Arnaud De Pasquale / Lucile Boulanger’s recording. Also, from today’s perspective, it seems unintelligible why Jordi Savall applies a fairly nervous vibrato to highlighted notes in the descant.

II. [Allegro] (2/4): My main, minor objections here are to do with sonority and balance. As for the latter: the harpsichord is too much in the background, and therefore lacks clarity / transparency and presence. Also the viola da gamba sonority in this recording can hardly compete with that in newer recordings.

III. Andante (12/8): An Andante in quavers, which makes it hard to feel the basic 12/8 pace. However, the dominance of semiquaver motifs precludes a substantially faster pace. However, even for an Andante in 3/8 time, the pace feels rather slow, almost dragging. Is this movement really such a sad lament?

IV. Allegro (6/8): A relatively narrow spatial soundscape: the viola da gamba prominently in the center, the harpsichord on the right-hand side, in the rear. Too bad this doesn’t allow the polyphony, the interaction between the voices to come to full bearing. By now, the recording technique seems inadequate for the quality of the interpretation / performance.

Rating: 4 / 4 / 4 / 4 = 4.0

Sonata No.3 in G minor, BWV 1029

I. Vivace (4/4): Despite all the merits that this recording may have (many still regard it the reference for these compositions): time has moved on! After 45 years, the recording has a hard time competing with recent ones, such as with Lucile Boulanger / Arnaud De Pasquale. First and foremost in terms of recording technique (acoustics, balance, spatiality of the soundscape), but also (even!) in terms of agogics and Klangrede (Harnoncourt, 1983).

II. Adagio (3/2): Compared to Lucile Boulanger’s, Jordi Savall’s vibrato often a tad nervous, and intensity and warmth are not even close to the former’s 2011 recording. The acoustic setting (in particular for the harpsichord) does not help, either.

III. Allegro (6/8): An excellent interpretation, no doubt, and not as pushed as Bylsma / van Asperen. However, in terms of richness and differentiation in sonority and soundscape, spatiality, it can not compete with the recording by Lucile Boulanger / Arnaud De Pasquale (featuring about the same tempo).

Rating: 4 / 4 / 4 = 4.0

Anner Bylsma & Bob van Asperen, 1990

Bach: Viola da gamba Sonatas BWV 1027-1029, Bylsma/van Asperen, CD cover

J.S. Bach: The Sonatas for Viola da gamba, BWV 1027 – 1029
J.C.F. Bach: Sonata in A major, BR-JCFB B 1

Anner Bylsma, Violoncello piccolo
Bob van Asperen, Organ

Sony Classical / VIVARTE SK 45945 (CD, stereo); ℗ / © 1990
Booklet: 32 pp. en/de/fr

Bach: Viola da gamba Sonatas BWV 1027-1029, Bylsma/van Asperen, CD, UPC-A barcode
amazon media link

CD Contents / Track Listing

The main content of the CD consists of the three Sonatas for Harpsichord and Viola da gamba, BWV 1027 – 1029. The CD includes these Sonatas in inverse order:

  • No.3 in G minor, BWV 1029 [13’45”]
    • I. Vivace [4’53”]
    • II. Adagio [5’31”]
    • III. Allegro [3’21”]
  • No.2 in D major, BWV 1028 [13’00”]
    • I. Adagio [1’41”]
    • II. Allegro [3’50”]
    • III. Andante [4’07”]
    • IV. Allegro [3’21”]
  • No.1 in G major, BWV 1027 [12’32”]
    • I. Adagio [3’36”]
    • II. Allegro ma non tanto [3’14”]
    • III. Andante [2’43”]
    • IV. Allegro moderato [3’00”]

In addition, the recording includes the Cello Sonata in A major, BR B 1 / Wf X:3, the No.30 from the collection “Musikalisches Vielerley” (1770) by Johann Christoph Friedrich Bach (1732 – 1795), Johann Sebastian Bach’s fifth son (a.k.a. “Bückeburger Bach“, after the city of Bückeburg, where he spent most of his life). The Sonata consists of three movements:

  • Sonata in A major, BR B 1 / Wf X:3 [12’07”]
    • I. Larghetto [4’12”]
    • II. Allegro [4’15”]
    • III. Tempo di Minuetto [3’48”]

The Artists

The late Dutch cellist Anner Bylsma (1934 – 2019) was one of the pioneers in historically informed performance on the cello. His career took off in 1959, when he won the first prize at the Pablo Casals Competition in Mexico, where obviously the jury was directed by Pablo Casals (1876 – 1873). He was the first one to record Bach’s six Cello Suites, BWV 1007 – 1012 on a period instrument (in 1979).

The Dutch harpsichordist, organist and conductor Bob van Asperen (*1947 in Amsterdam) studied harpsichord with Gustav Leonhardt (1928 – 2012), and pipe organ with Albert de Klerk (1917 – 1998), both at the Amsterdam Conservatory.

The Instruments

Anner Bylsma
Anner Bylsma

The booklet includes an extensive interview with Anner Bylsma, in which the artist explains his view on Bach’s Sonatas for Viola da gamba. He referred to the fact that at Bach’s time, the viola da gamba was rapidly falling out of fashion, and Bach ceased to write for the instrument during the second half of his life. Rather, for special effects (with very few exceptions) he resorted to the violoncello piccolo or the Viola pomposa—which by themselves would soon fall into oblivion.

Questioning the Original Versions?

Bylsma claimed that when using a cello, these sonatas “don’t sound / resonate”, neither with a piano, nor with a harpsichord. He even went as far as stating that the sonatas face the same problem when performed on viola da gamba and harpsichord. A manuscript only exists for BWV 1027, of which there are also copies (voices only) of a version for two traversos and basso continuo, BWV 1039, and from these copies, Bylsma suspects that there may have been an earlier original for two violins and basso continuo.

The second sonata (BWV 1028) only exists in a copy from 1753, which Bylsma does not find convincing at all (the bass frequently moving above the viola da gamba voice). He suspect that this may originally have been a trio sonata for organ, which somebody later arranged for viola da gamba and harpsichord.

And with the third sonata (BWV 1029), Bylsma ponders whether it originally was an orchestral work, given the strong allusions to / similarities with the Brandenburg Concerto No.3 in G major, BWV 1048. Bylsma states that also this sonata sounds weak (pale)—in comparison to BWV 1048.

Bylsma’s Conclusions

Bob van Asperen
Bob van Asperen

In Bylsma’s opinion, the bass viola da gamba (basse de viole) sounds “a little too low” for these compositions. Therefore, he decided to use a 5-string violoncello piccolo (an instrument from Tyrol, around 1700), which features a brighter sound. Also, the strongest / most prominent string on this instrument is the top one (the tuning is C-G-d-a-e’). In addition, he uses this argumentation to shift certain passages up by an octave. He finds the violoncello piccolo better suited for sonata playing than the basse de viole. Not all artists will agree with Bylsma on this—see also below.

Along the same lines, Bylsma opted for an accompaniment with organ (Bob van Asperen performs on a trunk / chest organ by Fama & Raadgever, Utrecht, 1985). Bylsma states that this offers better balance between the voices than a harpsichord. And, unavoidably, he argues that the harpsichord can impossibly make the very long, resting notes in BWV 1027 keep resonating. Of course, Bylsma is primarily a cellist…

Recording Date and Location

The recording of this CD took place in May 1990 at Hervormde Kerk in Bennebroek, The Netherlands.


Comments on the Performance

In order to avoid confusion, I’m commenting Bach’s Sonatas in numerical order, not in the sequence on the CD.

Sonata No.1 in G major, BWV 1027

I. Adagio (12/8): I can’t follow Anner Bylsma’s argumentation for using an organ for accompaniment. Bach explicitly mentions Cembalo on the front page of the manuscript. Yes, the chest organ can of course hold the long notes (such as the long D in the first two bars). However, now, these notes are too prominent, and of course completely static. In contrast, Anner Bylsma lets the long notes evolve with the help of subtle dynamics and a gently growing vibrato. The latter of course is a typical cello vibrato that one will never (or rarely) hear on a fretted instrument such as the viola da gamba.

The violoncello piccolo / organ combination may have some few advantages—but overall, it does not convince me here. For example: Anner Bylsma uses vibrato, the organ tone is of course totally flat—a discrepancy that is not present in the original configuration.

II. Allegro ma non tanto (3/4): Here, the alternative instrumentation plays out really well! The chest organ (a nice flue register!) is more (but not too!) prominent, an ideal partner to the violoncello piccolo, which in Bylsma’s hands has a sonority that resembles that of a viola da gamba. Beautiful!

III. Andante (4/4): The advantage of the organ is, that it accentuates the difference between the slurred quaver quadruplets (organ) and the quaver pairs on the viola da gamba. A minor drawback here is the slight lack of agogic tension.

IV. Allegro moderato (2/2): The tempo often feels pushed—moderato? And unfortunately, the organ tends to play itself into the foreground, often almost covering the violoncello piccolo.

Rating: 3.5 / 4.5 / 4 / 3.5 = 3.9

Sonata No.2 in D major, BWV 1028

I. [Adagio] (3/4): Calm, serene, peaceful. Unfortunately, the recording (or was this the artists’ intent?) makes the piece sound like an organ solo with cello accompaniment. The violoncello piccolo occasionally sounds a bit squeaky. This may be due to the microphone placement, or the sound engineering in general. However, I could also imagine that Anner Bylsma deliberately attempted to make the violoncello piccolo sound like a viola da gamba?

II. [Allegro] (2/4): Nice playing, nice music. One minor observation is that in the second half, the tempo subtly appears to run away, which defeats the tension.

Text Alterations

More important, though: as Anner Bylsma has doubts about the original version, he treated the material rather freely. His main objection being that the right-hand keyboard voice frequently moves above the string instrument (which may be more hurtful with the selection of an organ in lieu of the harpsichord), he freely switches the two voices, wherever he feels like doing so. I find this questionable. Why should the two voices not compete with each other?

III. Andante (12/8): Again, for major segments, Anner Bylsma swaps the two top voices. Moreover of that, he shifts entire passages up by an octave. The result leaves me with mixed impressions, especially in comparison with the other, “proper” interpretations…

IV. Allegro (6/8): Anner Bylsma postulates the origin of this composition being a trio sonata for organ. Here, however, especially in the middle section, where the keyboard part is taking a life on its own, the performance rather reminds me of the Sinfonia from Bach’s Cantata BWV 29, “Wir danken dir, Gott, wir danken dir” which essentially is a scarcely orchestrated version of the Partita No.3 for Solo Violin in E major, BWV 1006. Needless to say that also here, Anner Bylsma did not refrain from altering the text. The result: beautiful music, but not Bach’s Sonata BWV 1028.

Rating: 4 / 4 / 3.5 / 4 = 3.9

Sonata No.3 in G minor, BWV 1029

I. Vivace (4/4): Along the lines of my comments to the sonatas .in G major and in D major: why issue a CD labeled “Sonatas for Viola da Gamba“—and then altering the keyboard instrument (which Bach explicitly mentioned at least for BWV 1027), and—moreover—not even sticking to the viola da gamba that the CD title refers to? And there are again text alterations, which rather make this an “arrangement for violoncello piccolo and organ”. Yes, it is still beautiful, baroque music, very well-played, but…

II. Adagio (3/2): Here, Anner Bylsma leaves the text alone, but the single flue stop for the descant, and even more so, the matte, dull bass line don’t do justice to the composition, whether it is Bach’s original or not.

III. Allegro (6/8): The fastest tempo, once more. This may indeed be a “proper” Allegro in 6/8 time, with two beats per bar. However, with the dominating semiquavers, and even more with the semiquaver triplet passages, this performance feels breathless, too driven, pushed to the limit. Ornaments are often superficial, and there is no chance to indulge in agogics and differentiation—neither for the artists, nor for the listener. Allegro translates to joyful. This performance is neither joy nor fun. And the restless descant voice on the organ (even though it’s a simple flue register) soon feels obtrusive.

Rating: 4 / 3.5 / 3 = 3.5

Johann Christoph Friedrich Bach: Sonata in A major

The Sonata in A major, BR B 1 / Wf X:3 is the only “extraneous” (non-Johann Sebastian Bach) composition in this comparative review. Here, however, there is nothing I can compare it to, as it is my only recording of that composition. So, let me keep this short:

I. Larghetto: Beautiful music, which combines baroque elements (articulation, ornamentation) with aspects of the upcoming empfindsamer Stil (sentimental style). It still is mostly baroque, though, truly pleasant in melodies and harmonies. In Anner Bylsma’s interpretation, Interestingly, there is a fermata in the penultimate bar that leads the artist to fill in a cadenza. Anner Bylsma’s cadenza is short and light, but very fitting and beautiful.

II. Allegro: Very nice, joy- and playful music, with occasional reflective harmonic turns. Very nice playing by Anner Bylsma. Bob van Asperen’s accompaniment on the chest organ is discreet, almost inconspicuous, leaving the dominant role to the often virtuosic cello part. The continuo part in the manuscript consists of a single, ciphered bass line.

III. Tempo di Minuetto: Again very pleasant music. A minuet in A-B-A form, where the longer B part is not explicitly labeled as Trio.

Rating: 4.0

Lucile Boulanger & Arnaud De Pasquale, 2011

Bach: Viola da gamba Sonatas BWV 1027-1029, Boulanger/De Pasquale, CD cover

J.S. Bach: The Sonatas for Viola da gamba, BWV 1027 – 1029; Sonata in D minor, BWV 1023; Toccata in C minor, BWV 911

Lucile Boulanger, Viola da gamba
Arnaud De Pasquale, Harpsichord

Alpha 161 (SACD, stereo); © 2011
Booklet: 36 pp. fr/en

Bach: Viola da gamba Sonatas BWV 1027-1029, Boulanger/De Pasquale, CD, EAN-13 barcode
amazon media link

CD Contents / Track Listing

The main content of the CD consists of the three Sonatas for Harpsichord and Viola da gamba, BWV 1027 – 1029:

  • No.1 in G major, BWV 1027 [14’00”]
    • I. Adagio [4’12”]
    • II. Allegro ma non tanto [3’52”]
    • III. Andante [2’38”]
    • IV. Allegro moderato [3’18”]
  • No.2 in D major, BWV 1028 [15’10”]
    • I. Adagio [2’00”]
    • II. Allegro [4’09”]
    • III. Andante [4’40”]
    • IV. Allegro [4’21”]
  • No.3 in G minor, BWV 1029 [15’30”]
    • I. Vivace [5’34”]
    • II. Adagio [6’03”]
    • III. Allegro [3’53”]

In addition, the recording includes Bach’s Sonata for Violin and Basso continuo in E minor, BWV 1023, composed around 1714 – 1720. Lucile Boulanger transcribed this for viola da gamba and harpsichord, transposing it to D minor. The Sonata comprises three movements:

  • Sonata in D minor, BWV 1023 [12’39”]
    • I. [Preludio] (4/4) [1’32”]
    • II. Adagio ma non tanto (3/4) [3’48”]
    • III. Allemanda (2/2) [4’21”]
    • IV. Giga (12/8) [2’58”]

Moreover, Arnaud De Pasquale performs a piece for harpsichord solo by Bach:

The sequence of the works on the CD is BWV 1023 — BWV 1027 — BWV 911 — BWV 1028BWV 1029.

The Artists

Lucile Boulanger (2021, © Richard Dumas)
Lucile Boulanger (© Richard Dumas)

Lucile Boulanger

The French gambist Lucile Boulanger (*1986, see also took up the viola da gamba at the age of five. Her main teachers on the instrument included Jérôme Hantaï (*1961) in Cergy and Christophe Coin (*1958) at the CNSMD in Paris, where she completed her education in 2009. After winning prizes at several international competitions, she now pursues an active career as chamber musician and as member of notable ensembles, see her Web biography. In the early ‘2000s Lucile Boulanger also worked as voice actress.

Lucile Boulanger performs with (baroque) formations such as the Ensemble Correspondances, led by Sébastien Daucé (*1980) the Ensemble Pygmalion under Raphaël Pichon (*1984), and Les Talens Lyriques under the direction of Christophe Rousset (*1961). And it almost goes without saying that the gambist is also performing as member of consorts of viols, such as the Ricercar Consort with Philippe Pierlot (*1958), the ensemble L’Achéron led by François Joubert-Caillet (*1982), and Musicall Humors with Julien Léonard.

Arnaud De Pasquale

Arnaud De Pasquale (© Arnaud De Pasquale)
Arnaud De Pasquale

Lucile Boulanger’s duo partner in this recording is the French harpsichordist, pianist and organist Arnaud De Pasquale. The latter grew up in Poitiers, in an environment of baroque music. He started learning the harpsichord at age five. He did his initial studies at the local conservatory, then moved to the CNSMD in Paris. There, De Pasquale studied with Olivier Baumont (*1960) and Blandine Rannou (*1966). This was followed by master classes with Pierre Hantaï (*1964), Blandine Verlet (1942 – 2018), Skip Sempé (*1958), and others.

Also Arnaud De Pasquale performs in a number of baroque formations, such as the Ensemble Correspondances, led by Sébastien Daucé, the Ensemble Pygmalion with Raphaël Pichon, and the Collegium Vocale Gent under Philippe Herreweghe (*1947). And, of courese, he also pusues an international career as soloist on the organ, the harpsichord, and the piano.

The Instruments

Viola da gamba

As a gambist, Lucile Boulanger has quite a different view about the instrumentation of Bach’s Sonate à Cembalo è Viola da Gamba. In a promotion video for her latest CD release (which I may discuss in this blog), she does state, though, that the viola da gamba is an intimate instrument, best suited for chamber music and small venues, rather than big concert halls.

This may be what Anner Bylsma referred to when he claimed that the instrument “does not sound”. Indeed, in the rare instances where Bach included the viola da gamba in cantatas or his passion oratorios, it is for very few, very special (introverted, reflective) moments / scenes. Definitely, in Lucile Boulanger’s recording on Bach’s sonatas for viola da gamba, nothing could be farther from the listener’s experience than a statement such as “it does not sound”!

As the Sonata in D major, BWV 1028 requires an A, (“low A“), the artist selected a 7-string basse de viole (tuned A, — D — G — c — e — a — d’). Her instrument is from 2006, by François Bodart in Andenne, Belgium, built after a German instrument from 1699, by Joachim Tielke (1641 – 1719), Hamburg (one of the principal makers at Bach’s lifetime). Her bow is by Craig Ryder, Paris, 2011, after a model from around 1760.


Arnaud De Pasquale performs on a two-manual harpsichord made in 1979 by Philippe Humeau in Barbaste. This is a replica of an instrument now located in the Musée des Pyrenées in Lourdes. It is thought to be by Johann Heinrich Silbermann (1727 – 1799), Strasbourg, a member of the famous family of organ builders. His instruments were well-known in their time. Still, no other harpsichords by Johann Heinrich Silbermann have survived to this day.

Recording Date and Location

The recording took place on 2011-06-06:09 at the Chapelle Notre-Dame du Bon-Secours, Paris 14.


Comments on the Performance

Sonata in D (E) minor, BWV 1023

As this sonata is not central to the scope of this review, I’m discussing it further down below.

Sonata No.1 in G major, BWV 1027

I. Adagio (12/8): The most beautiful sonority among the three, excellent acoustic balance & partnership, especially between the viola da gamba and the right hand on harpsichord. My main quibble: the tempo is a bit (too) slow, which occasionally makes the performance feel static. And when the artists slow down a little, then the flow appears to stop.

II. Allegro ma non tanto (3/4): See above—and also here, the pace is maybe a little too careful: at least in a direct confrontation (e.g., with Savall / Koopman), the pace initially feels rather like an Andante. Yes, it’s “ma non tanto“—however, the basic annotation still is Allegro. That said: after a few bars / lines, one feels “in” the interpretation, and the focus turns towards the beautiful playing and sonority!

III. Andante (4/4): Beautiful playing—and here I like the tempo, the flow, the pronounced agogics (harpsichord), the subtle tempo variations, the “breathing” of the music.

IV. Allegro moderato (2/2): This interpretation wins in terms of agogics and careful, detailed and harmonious articulation (and dynamics on the viol).

Rating: 4.5 / 4.5 / 5 / 4.5 = 4.6

Sonata No.2 in D major, BWV 1028

I. [Adagio] (3/4): Solemn, calm, harmonious, rich in agogics and sonority. Arnaud De Pasquale’s subtly arpeggiating touch further contributes to the gentle atmosphere in this interpretation. In comparison to Jordi Savall, Lucile Boulanger’s playing stands out with it’s much more harmonious, subtle and selective use of vibrato. Also, the artists here are less strict in sticking to Bach’s notation: there are frequent (but fitting, small, non-intrusive) extra ornaments. Beautiful!

II. [Allegro] (2/4): Excellent once more. When listening to this, Anner Bylsma’s reasoning for freely swapping the upper two voices seems incomprehensible. The sonority is very good and well-balanced. One little quibble, which I have difficulties nailing down to a specific cause: to me, there is a, occasional, slight trace of rigidity (or is it relentlessness?) in the rhythmic foundation—should there be a tad more agogic swaying / breathing?

III. Andante (12/8): The tempo is almost in line with Jordi Savall’s interpretation—though just a tad more fluid. This lets the music stay earnest, contemplative—but keeps it from feeling like a sad lament, thanks to some rhythmic swaying. And unlike Anner Bylsma, I don’t see the slightest need to resort to octave shifts and voice swapping.

IV. Allegro (6/8): Same here! I don’t see any problem with the right-hand keyboard voice often moving above the string instrument—it actually enters a vivid dialog with the latter, full of life, enthralling even!

Rating: 5 / 4.5 / 4.5 / 5 = 4.75

Sonata No.3 in G minor, BWV 1029

I. Vivace (4/4): Full of lively agogics, dynamics, and Klangrede (Harnoncourt, 1983), and a marvelous, rich soundscape with the harpsichord voices in an intense dialog with the viol. The pace is noticeably slower than with Anner Bylsma / Bob van Asperen—but thanks to the richness in ornamentation, agogics, articulation, dynamics, etc. it never feels sluggish.

II. Adagio (3/2): Lucile Boulanger demonstrates that Anner Bylsma’s vibrato is not needed at all for the piece to feel expressive, warm, and intimate. Indeed, her voice is a touching, intense recitativo accompagnato. Yes, the gambist does occasionally use vibrato, in order to let long notes harmoniously evolve. However, that is a very subtle, discreet effect, inconspicuous, and often resembling a slow flattement. And Arnaud De Pasquale’s accompaniment is completely in line with the viol, in terms of Klangrede / agogics, and careful, gently arpeggiated articulation. And this marvelous, full bass sonority—unbeatable, I’m tempted to say!

III. Allegro (6/8): Listening to this after Anner Bylsma / Bob van Asperen—and I wouldn’t hesitate stating that the harpsichord (especially this marvelous instrument, and particularly in this excellent recording) is infinitely better than the organ for this music! And, of course, the tempo isn’t pushed here, but leaves room for agogics, differentiation in articulation and phrasing. Interestingly, the pace does not feel all that much slower

Rating: 5 / 5 / 5 = 5.0

Sonata in D (E) minor, BWV 1023

I. [Preludio]

Beautiful, this solo part! It’s as if Bach had it written with the viola da gamba in mind! Free preluding (which is what this is) is very much in the realm, the domain of the viola da gamba. I actually claim that in this transcription, the solo sound better than in the original: warmer, richer, and (in a way) “speaking more directly”. The latter presumably because (with the transposition and the move to a lower octave) the sound of the basse de viole is close to the human voice?

The one issue with this interpretation is that adding a string instrument for the continuo makes little sense, as it would obscure (or compete with) the solo. Here, the bass consists of a single, long, “naked” drone in C,—and on the harpsichord, this tone would havce long gone over the 29 bars of the Preludio. Arnaud De Pasquale resorts to repeating that tone (octavated, or with the addition of the 16′ stop) once or twice per bar. He tries not to make this sound too uniform by “picking the appropriate spots”—nevertheless, the result is rather prominent and feels “a bit odd”.

II. Adagio ma non tanto

A beautiful, expressive and intense lament, full of Klangrede—like a recitativo accompagnato. It is again profiting from the proximity to the human voice. The viola da gamba in its element! Arnaud De Pasquale’s setting of the ciphered bass is harmonious, natural, supportive, and very much in line with the melody voice. There is actually no need for doubling the bass line with a stringed bass instrument.

III. Allemanda

Also here, the interpretation sounds as if Bach had written it for the viola da gamba. The solo profits from the naturally expressive “language” of the instrument—even though the composition does not permit many of the instrument’s specific abilities, such as rich arpeggiando, or polyphony.

IV. Giga

Very nice, full of momentum, lively, with all the syncopes, etc.—and again, the harpsichord’s beautiful, warm bass register is more than just sufficient for the basso continuo. And, as in all the other movements in this recording, the acoustics, the sound / sonority, the acoustic balance are all excellent!

Rating: 4 / 5 / 5 / 4.5 = 4.6

Two Recordings of the Original Violin Version

In a separate blog post, recently updated and expanded, I have compared recordings of Bach’s Sonatas for Harpsichord and Violin, BWV 1014 – 1019, including the recordings with

The Sonata in E minor, BWV 1023 is included with these recordings. For details on the media see the aforementioned blog post. BWV 1023 was not part of that comparison, and it isn’t within the main subject in this post, i.e., the sonatas for harpsichord and viola da gamba. Nevertheless, I wanted to add a few, sketchy remarks on how Lucile Boulanger’s transcription and her performance with Arnaud De Pasquale compares with these performances of the original version:

Andrew Manze, Richard Egarr, Jaap ter Linden

I. [Preludio]: Interestingly, the artists did not add the viola da gamba to maintain the drone E. Rather, Richard Egarr repeats the E—sparingly, and not in regular intervals. With this, to the listener, the drone is often (essentially) absent from the Preludio, except for the occasional (rare) reminder. To me, this is about as unsatisfactory as Arnaud De Pasquale’s frequent and prominent repetitions.

II. Adagio ma non tanto: In a direct confrontation, the viola da gamba clearly overshadows the violin, as it avoids the excess poignancy of the gut strings on the violin. Andrew Manze’s playing is very expressive and rich in dynamics and agogics—but it can’t “talk” to the (my / the listener’s) heart as directly as the viol.

III. Allemanda: Beautiful, expressive playing—sadly, the harpsichord is often overshadowed by the string instruments. Despite Andrew Manze’s Klangrede, in a direct confrontation with the viola da gamba interpretation, the (solo) sonority sometimes feels a bit austere.

IV. Giga: To me, the best movement in this interpretation—rich in agogics and expression, full of momentum, with playful (and natural) extra ornaments in the violin part.

Rating: 4 / 4.5 / 4.5 / 4.5 = 4.4

Rachel Podger, Trevor Pinnock, Jonathan Manson

I. [Preludio]: Rachel Podger’s Preludio is more fluent than Manze’s or Lucile Boulanger’s. And here Jonathan Manson discreetely maintains the drone E. It is not as “static” as on an organ, but rather flexible, maintaining some “life”. My one quibble: in comparison with the basse de viole, Rachel Podger’s violin often sounds rather (too) poignant, and with the rich agogics, the articulation in fast passages is not always quite clear.

II. Adagio ma non tanto: See above. Sadly, Rachel Podger’s solo is hampered by her tendency towards prominent, exaggerated belly notes. And where she uses vibrato it also tends to be heavy and excessive.

III. Allemanda: Here again, Rachel Podger performs at the clearly most fluid pace—is this still appropriate for an Allemanda? Well, the movement is in split time, so Rachel Podger may have a point here. And in this fast(er) movement, belly notes and excess vibrato are far less prominent. Of course, there’s still the artist’s tendency towards occasionally exaggerated dynamics, such as accents, highlighted notes, and the like.

IV. Giga: My main (critical) remarks here are again about the frequent, exaggerated belly accents / dynamics (and their predictability). Also, the violinist’s occasional extra ornaments can’t match Andrew Manze’s.

Rating: 4 / 3 / 4 / 3.5 = 3.6

Toccata No.2 in C minor, BWV 911

The Toccata performance is a very welcome opportunity to enjoy the beauty of Arnaud De Pasquale‘s Silbermann replica harpsichord:

Toccata / Introduction: Excellent playing in all aspects—articulation, clarity, touch, agogics, phrasing, ornamentation, Klangrede. And the full, rich sonority of the harpsichord is simply marvelous!

Adagio: The tempo is my one and only quibble with this interpretation. It is distinctly slower than the piano performances (briefly discussed below)—and yet, to me, this feels like Andante rather than Adagio. It may be Adagio in crotchets (4/4)—but somehow, the quaver beat appears to prevail, which leads to a certain unrest / unease (at least relative to the expectation). I suspect that the pace is simply too slow for the listener to sense the crotchet beat—making it feel too fast, paradoxically….

Fuga: As excellent as the introduction. In comparison with piano interpretations, it is particularly striking how on the harpsichord the stylus phantasticus segments makes the cadenza in bars 83 – 85 (or the final bars) stand out from the rest of the fugue.

Rating: 4.5

Comparing with Piano Interpretations?

Interestingly, Arnaud De Pasquale’s recording is the only one on the harpsichord in my collection, and one of only three in all the Toccatas. The others are Gustav Leonhardt (1928 – 2012) playing Toccata No.6 in G minor, BWV 915, and Ton Koopman (*1944) with Toccata No.7 in G major, BWV 916.

I do have two performances of the Toccata No.2 in C minor, BWV 911, performed on the modern piano, though:

The confrontation with the harpsichord performance makes me realize how much better the harpsichord is at building tension within smallest motifs, even within single notes. With this, an interpretation on the piano unavoidably sounds fairly flat (sorry, pianists out there… !). The difference between the two instrument generations is most pronounced in the stylus phantasticus (as in the introduction, or in the short cadenzas), which really, really calls for a harpsichord (or a baroque organ, maybe).

Fair Comparison?

I can’t and don’t want to compare these performances to Arnaud De Pasquale’s on the harpsichord. The characteristics of the instruments are so vastly, fundamentally different that one way or another, a comparison would be unfair. Let me just briefly characterize these performances:

Walter Gieseking‘s performance now is one of the past, for sure. In his hands, the entire Toccata sounds like a finger study. His tempo is always motoric and faster than what even pianists do today, moreover, he has a tendency to accelerate further within a movement. The articulation is often superficial, and devoid of agogics, the dynamics fairly flat.
A historic document at best, not really commendable.

Claire Huangci is a pianist that I hold in very, very high esteem (certainly among the pianists of her generation). Indeed, her articulation is incredibly careful and detailed, she applies agogics and very diligent dynamics and phrasing, and her ornamentation is truly baroque (as much as possible on a modern instrument).

In short: one of the best interpretations imaginable on a concert grand (Yamaha CFX). With the exception of the introduction (where her tempo is the same as Arnaud De Pasquale’s), Claire Huangci’s tempo is between Gieseking’s and that in the harpsichord performance. Excellent, highly recommended—but still not a substitute for interpretations on harpsichord…

Additional Media References

Walter Gieseking Plays Bach

Bach: Partitas BWV 825-830, Gieseking, CD cover

J.S. Bach: The Partitas, BWV 825 – 830; Toccata BWV 911, Keyboard works BWV 811, 831, 906, 944, 971

Walter Gieseking, Piano

Music and Arts Program of America, CD 947 (2 CDs, mono); ℗ 1996
Booklet: 12 pp. English
Hardcopy CD currently not available

Bach: Partitas BWV 825-830, Gieseking, CD, UPC-A barcode

Claire Huangci — Bach: The Toccatas

Bach: Toccatas BWV 910-916, Huangci, CD cover

J.S. Bach: The Toccatas, BWV 910 – 915; Toccata & Fugue in D minor, BWV 565 (arr. Ferruccio Busoni)

Claire Huangci, Piano

Edel Kultur / Berlin Classics 88547002016 (CD, stereo); ℗ / © 2021
Booklet: 24 pp. de/en

Bach: Toccatas BWV 910-916, Huangci, CD, UPC-A barcode
amazon media link


I can keep my conclusions from this review short:

  • A very strong recommendation in favor of the 2011 recording by Lucile Boulanger and Arnaud De Pasquale. That recommendation also applies to people who like and maybe already have the recording by Jordi Savall and Ton Koopman
  • The recommendation also extends to the included recording of the Sonata BWV 1029 in Lucile Boulanger’s own transcription for viola da gamba and harpsichord, that I actually prefer over the two recordings of the original violin version mentioned above.
  • The aforementioned CD also includes Arnaud De Pasquale’s interpretation of Bach’s Toccata BWV 911. This also is fantastic / highly recommended. That judgement is not the result of a direct comparison with other period instrument performances. However, I did briefly discuss two performances on modern piano, among which I can strongly recommend the interpretation by Claire Huangci. I do of course prefer the harpsichord version, but for people who don’t enjoy or don’t want to delve into the sound of the harpsichord, Claire Huangci’s recording of all of Bach’s Toccatas may come to the rescue!


Harnoncourt, Nikolaus (1983). Musik als Klangrede: Wege zu einem neuen Musikverständnis. Salzburg: Residenz Verlag. ISBN 978-3-7017-0315-9

AboutImpressum, LegalSite Policy | TestimonialsAcknowledgementsBlog Timeline
Typography, ConventionsWordPress Setup | Resources, ToolsTech/Methods/Pics/Photography

Leave a reply—comments are welcome!