Open Book Publishers logo Open Access logo
  • button
  • button
  • button
GO TO...
book cover

5. Affect and Individual Difference: Towards a Holistic Account of Performance

© Dorottya Fabian, CC BY

Having focused on specific performance features in the previous chapter, I now turn to global matters in relation to specific moments of particular movements and increasingly enlist Deleuzian terms and thinking.1 First I look for differences within the loose boundaries of MSP and HIP and then across multiple recordings of the same violinists. This is followed by a comparative exploration of musical character and affect in selected examples, concluding with a consideration of idiosyncratic versions and listeners’ reactions.2 Essentially, I argue that a more holistic approach to performance analyses brings to light the multitude of interactions at play that contribute to the overall effect of a performance. This might be best grasped subjectively paying attention to both measurable and felt features. I show that the seeming chaos of individual differences caused by the relative and diverse contributions and non-linear interactions of performance features can be illuminated, if not fully explained, by using a Deleuzian lens. Deleuze and Guattari offer a language to account for stabilizing (territorializing) and diversifying (de-territorializing) functions of technical and aesthetic traits and assist in pinpointing moments or performance solutions that push interpretations along the style vector from MSP to HIP; they help describe the “in between.” Just like in the previous chapter, here too the detailed or more specific descriptions appear in boxed texts shaded grey for ease of navigation.

5.1. Differences within the MSP and
within the HIP Styles

The Loure, the Gavotte en Rondeau, and the two Menuets of the E Major Partita have already received considerable attention especially in relation to ornamentation, rhythm and bowing. This provides a good foundation for further discussions of these movements. The focus now is on general interpretative differences within the two respective style categories rather than across them.

The Loure

According to Stowell, Schröder, and Ledbetter, several eighteenth-century French authors regard the loure to be a slow gigue, a dance in 6/4 metre with a heavy pace.3 The German contemporaries of Bach—Mattheson and Walther—also concur regarding the slow tempo and the “arrogant and proud nature” of the dance.4 Schröder further claims that the dotted rhythms underline the “fiery” jumps and complex steps of the dancer. However, the movement should never sound busy and should always retain a sense of “quiet nobility.”5 Somewhat contrastingly Ledbetter finds Wendy Hilton’s description apt for Bach’s loure: “a unique blend of gently expressed nobility, tenderness and tranquillity.”6

Schröder recommends a moderate speed of crotchet = 96, “which allows the violinist to combine the pesante character with featherweight ornaments as well as distinguishing the light, arpeggiated upbeat chords from the strong ones on down-beats.”7 In contrast, the tempo Stowell suggests is crotchet = 80. He emphasizes the movement’s “carefully balanced phrases [...] clear harmonies, and ornamented melody” contributing to a “languid” character.8

MSP Interpretations

As mentioned earlier, violinists tend to perform the E Major Loure as a slow, lyrical movement although quite a few of the more recently recorded versions show signs of a more dance-like approach. Still, the “fiery, proud and arrogant” character highlighted by Schröder is rare. Perhaps Zehetmair’s version could be described as “fiery” due to its fast tempo and sharply defined rhythm. But the fast tempo seems to go against notions of “quiet nobility” and “languid character.” Along the spectrum of the two basic approaches to interpreting this Loure—defined as lyrical and melodically orientated versus dance-like and rhythmically orientated—there is a great variety of grades and differences. Both approaches can be slower or faster, lighter or heavier overall, flowing or sturdy, more or less legato, detached or staccato, dotted to different degrees, ornamented or not, played on the string with quite intense bowing or with light, lifted bow strokes, and so on.9

Broadly speaking one might say that the MSP versions more commonly feature the melodically orientated approach and that these tend also to be fairly legato, with phrasing aided by arching dynamics (crescendo-decrescendo). The best examples are Hahn, Poulet, Shumsky, and Ehnes, but Tetzlaff’s second recording may also be mentioned (the earlier is also fairly legato but lighter and less vibrato).

Teasing out general differences among these MSP versions, it is worth noting that Hahn’s and Tetzlaff’s more recent interpretations focus on beauty of tone and melody. They both take lyricism to a level that may be perceived as emotive (cf. the intense vibrato on longer notes and extremely soft dynamics in Tetzlaff’s), whereas the versions of Szenthelyi, Perlman, and Mintz project a gentle pulse that counteracts such an impression.10 Szenthelyi plays with strong vibrato but less legato bowing which helps create some forward momentum. Ibragimova’s version is rather different. She provides an articulated and more varied reading but she plays rather slow and soft and the performance becomes languid, making the music sound repetitive (cf. Audio example 4.12 and 5.1).

Differences can be observed in the approach to repeats. Hahn, Mintz, Fischer, Shumsky, and Poulet do not differentiate between first play and repeat. Others often do. Tetzlaff varies ornamentations in 2005. Lev, Perlman and Tetzlaff in his earlier recording use softer and / or less fluctuating dynamics during repeats. Lev adds further variation by playing more legato.

Further specific differences among MSP versions involve the use of dynamics and tempo fluctuation for the shaping of phrases. Many start the movement moderately loud and build towards a climax in b. 8. However, what happens in the next four measures tends to differ: For instance Perlman further intensifies through crescendo and vibrato whereas Zehetmair releases the tension and reduces dynamics. Barton Pine also relies on dynamics to aid phrasing while Kremer’s recordings have considerable tempo fluctuations.

Differences can also be noted in the large-scale phrasing of the second half of the Loure (bars 12-24). Several violinists tend to maintain the momentum to about bar 15 and then provide a brief relaxation of intensity and dynamics. A building up of tension and crescendo then starts in b. 19 leading to a climax in b. 20. Zehetmair, in contrast, accelerates to b. 19 and starts this passage f, broadening the tempo in b. 20 with a concurrent diminuendo and strong rall to the dotted crotchet chord and then continuing pp from the up-beat to b. 21. A similar trajectory with less flexibility and contrast can be observed in Shumsky’s performance as well. The last four bars (21-24) again vary regarding whether the tension is maintained to the end or tapered off to a soft conclusion. Those starting b. 21 p often create a little swell of dynamics in b. 22 before the final decrescendo and rall from beat 5 of b. 23 (Audio example 5.1).

5.1. Phrasing in J. S. Bach, E Major Partita BWV 1006, Loure, extract: bars 12-20. Three versions: Christian Tetzlaff 1994 © Virgin Veritas, Thomas Zehetmair © Teldec, James Ehnes © Analekta Fleurs de lys. Duration: 2.16.

To listen to this extract online follow this link:

Some versions display a more localised, moment to moment variation. Tognetti, for instance, plays with dynamic nuance, achieved through stress and subtle-rapid changes in bow pressure and speed on the note or two-note level. Zehetmair changes dynamics to shape sub-phrase after sub-phrase (i.e. every two bars or so), while in his earlier recording Tetzlaff creates many little dynamic arches (crescendo-decrescendo) during the first play of each half of the movement. In the later version Tetzlaff’s choice of agogic stresses and execution of trills often differ between first play and repeat (e.g. stress on downbeat B in b. 22 is only in first play; written out F on downbeat in b. 15 is played more like a longer appoggiatura in the first play but more as written during repeat; from b. 22 to the end of the movement is softer in repeat).

Another source of diversity is the manner in which chords are delivered. Among MSP players Lev, Tetzlaff, Tognetti, Barton Pine, and most of the younger violinists tend to play chords lightly, at times slightly arpeggiated. Older performers, especially Perlman, Kremer, and Shumsky, play with heavier strokes aiming to make the notes sound together. In spite of this broad generalization chords are generally played in diverse ways in most versions, depending on the musical context of the chord.

HIP Interpretations

Some of these characteristics typical of MSP recordings can be noticed in HIP versions as well, in particular Kuijken’s and Schröder’s but also van Dael’s. Of these, Kuijken’s is the most legato (with longer, fairly sustained bow strokes) and the least rhythmical, especially in 2001. Schröder and van Dael project the dance character in the first half of the movement but van Dael’s tempo proves too slow to keep the momentum in the second half and Schröder slows down when the texture includes more double and triple stops. At times his bowing sounds weighty on the string and his strokes are longer in the second half with fewer metrical stresses. He adds an accented appoggiatura to almost every down-beat and half-bar beat during the first repeat and many in the second half. However, because of their regularity and uniformity they foster the impression of standardization rather than spontaneous embellishing (cf. Audio example 4.12).

The other HIP violinists tend to present a more rhythmically orientated reading. As I have shown in chapter four, some of them also add many ornaments (just like their younger MSP colleagues, such as Tetzlaff, Tognetti, and Gringolts, as well as Mullova and Faust). Nevertheless there are also many noticeable differences among these HIP versions as well.

For instance, Wallfisch plays rather slower and very staccato, leaving gaps between notes. She leans on and holds the dotted notes while playing their short pairs as well as most third and sixth beats very short, with kerning. Her performance does not become mechanical because of subtle variations. The passage between bars 19 and 22, for instance, is played much less sharply articulated. In b. 20 the paired slurs are gently projected while the notes on beat three are shaped to form part of the cadence (chord) on beat four. Podger chooses a similar tempo but plays more legato and in a less overtly dotted, leaping manner. She does not add ornaments and plays the repeats as the first time round. Huggett’s performance is similar to Podger’s as it is more legato than Wallfisch’s and she does not add ornaments either. On the other hand Huggett leans on the longer notes and suspensions like Wallfisch does. Luca’s interpretation is more flowing and faster and in that it is similar to Huggett’s. However, the added ornaments during repeat enhance the dotted character of the movement and he also plays some of the figures more staccato creating a very different overall effect (Audio example 5.2).

5.2. Comparison of HIP performances in J. S. Bach, E Major Partita BWV 1006, Loure, extracts: bars 19-22. Two versions: Elizabeth Wallfisch © Hyperion, Rachel Podger © Channel Classics; bars 17-22. Monica Huggett © Virgin Veritas; repeat of bars 12-24. Sergiu Luca © Nonesuch. Duration: 2.15.

To listen to this extract online follow this link:

Importantly, Luca’s ground-breaking playing finds its true descendants in the recordings of Mullova and Faust some 30 years later. These are not just straightforward next generation “improvements” on the past. They are new constellations; new “multiplicities.” The lively energy, rich subtleties and luxuriantly varied and inventive ornamentations of Mullova’s and Faust’s performances are indicative of a liberated, non-literalistic performing style of baroque music becoming common as the new millennium enters its second decade. It is a testament to Sergiu Luca’s visionary musicianship to have introduced-anticipated such “lines of flight” in the midst of the Urtext-orientated, “authentistic” 1970s.11

The Gavotte en Rondeau

The basic differences in performing the Gavotte relate to dynamics (whether fairly uniform or varied, and to what extent and how), bowing (whether lighter or heavier; shorter or longer), and to the shaping of the character of both the rondo theme and the various episodes (whether the contrasts in texture and thematic material are emphasized or downplayed and how each section is articulated and phrased). It is easy to see, then, that there could be a multitude of solutions and combinations of choices and this is indeed the case. A rough generalization is summarized in Table 5.1. For closer detail, and to enable insight into homogeneity or diversity of performance practice within each overarching stylistic categories I offer comments on two HIP and two MSP versions recorded within a relatively short period of time: Wallfisch (1997) versus Huggett (1995) and Gringolts (2002) versus Lev (2002).

Table 5.1. A generalized summary of performance characteristics found in recordings of the E Major Gavotte en Rondeau movement. Date is provided if one version by the same artists differed from the other.







Lighter (varied)



Slower (slightly)

Contrasted (varied)










Matthews (terraced)







Barton Pine





Mullova (esp. ’08)





St John



van Dael



Barton Pine





Hahn (episodes)





Kremer’05 (theme)








Shumsky (in linear motion)

St John



van Dael



Kremer’05 (episodes)



Shumsky (in double stops of theme and episodes)






St John

Wallfisch Zehetmair
















van Dael

Barton Pine



(a little)

Khachatryan Faust (somewhat)



Mullova (more in ’08)


St John


van Dael











Matthews (mostly)



Schröder Shumsky



HIP: Wallfisch and Huggett

Elizabeth Wallfisch (1997) and Monica Huggett (1995) both play at a moderate tempo with mostly light, short bow strokes and subtle dynamic variations. Both articulate the score in great detail. Their performances are nevertheless very different.

Wallfisch mostly accentuates notes and figures by stressing the initial note (e.g. first quaver in bb. 53-57), by playing ‘off-notes’ very staccato and / or lightly (e.g. bb. 12-13 second group of four quavers or bb. 72-73 the second double stop in each quaver dyad), and by slight dynamic contrasts (bb. 24-26.1: f, b. 26.2-7: p, bb. 28-30.1: f, bb. 30.2-32: mp; bb. 82-83: f, bb. 84-85: mp, bb. 88-89: pp). She plays the rondo theme fairly flowingly, without any particular caesura after the minim in b. 2 or b. 6 (Audio example 5.3).

5.3. Articulation in J. S. Bach, E Major Partita BWV 1006, Gavotte en Rondeau, extract: bars 9-36. Elizabeth Wallfisch © Hyperion. Duration: 0.50

To listen to this extract online follow this link:

In contrast, Huggett’s version relies on tempo fluctuation and the timing of notes more than anything else. The theme itself is presented in a fairly straightforward manner but each episode includes subtle but constant shifts in tempo usually at the bar level, creating what amounts to rubato in the classical sense of the term: robbed time given back straight away.12 In the first episode, for instance, there is a slight accelerando on the upward figures and a ritenuto on the other patterns. These fluctuations tend to go across the bar line as the upward figures occupy the second half of each bar, except in b. 11. Furthermore, both the rits and accelerandos may involve only two to three notes in an obvious or pronounced manner, so the effect is more to do with flow, shaping and articulation than tempo per se (see bb. 10-15, Audio example 5.4).13

5.4. Timing and tempo fluctuations going across bars in J. S. Bach, E Major Partita BWV 1006, Gavotte en Rondeau, extract: bars 9-16. Monica Huggett © Virgin Veritas. Duration: 0.15

To listen to this extract online follow this link:

Huggett’s performance of the second episode is particularly noteworthy. She creates a little rit. in b. 28 as the figuration on beat one outlines the tonic. This is followed by a speeding up on beat two to lead into the next bar which is a repeat of the episode’s opening gesture, this time in the dominant. However, by playing the second beat of this bar ritenuto Huggett allows the listener to momentarily relish the dominant for its own sake, for during the remainder of the episode Bach teases perception with the potential of making it the new tonic. Indeed, Huggett’s performance engages with this teasing full on by alternatively accelerating (b. 31), slowing (b. 33) and holding back (b. 35) to draw out harmonic tensions and expectations. The final dominant preparing the tonic cadence at the end of the episode (bb. 38-39) is again played rapidly as if in a hurry to round it off at last and to end the game with a corresponding rit. (Audio example 5.5).

5.5. Articulation and timing in J. S. Bach, E Major Partita BWV 1006, Gavotte en Rondeau, extract: bars 24-40. Monica Huggett © Virgin Veritas. Duration: 0.30

To listen to this extract online follow this link:

MSP: Lev and Girngolts

The two non-specialist violinists chosen for comparison are actually exhibiting HIP tendencies in their performances of the Gavotte en Rondeau. So this discussion identifies further layers of diversity and complexity within HIP “territory” indicating “rhizomic” tendencies.

Lara Lev seems to play around more with tempo fluctuations whereas Ilya Gringolts more with dynamic shifts. He plays the theme through while Lev breaks it up with slight pauses at the minim double stops. Her detailed articulation of episode one creates an impression of slightly slower tempo and she uses similar speeding up and slowing down strategies to Huggett in episode 3, especially from b. 53 onward. In contrast Gringolts plays these bars in even tempo but with constantly fluctuating dynamics at the bar and half-bar level (Audio example 5.6).

5.6. Embellished rondo theme and comparison of MSP styles in J. S. Bach, E Major Partita BWV 1006, Gavotte en Rondeau, extracts: bars 40-72. Ilya Gringolts © Deutsche Grammophon; bars 48-64. Lara Lev © Finlandia Records. Duration: 1.09.

To listen to this extract online follow this link:

In the final episode their dynamics are fairly similar although Gringolts has a greater contrast between loud and soft with pp in bars 74, 80, 84, and 88. Other differences include a crescendo by Gringolts as opposed to a descrescendo by Lev in bar 86; flexible (Gringolts) versus in-tempo (Lev) performance of bars 88-89; crescendo / accelerando followed by decrescendo in bars 90-1 in Lev’s version and even dynamics with rallentando starting in the second half of bar 90 in Gringolts’ performance. Lev plays the final rondo statement softly, Gringolts forte.

Menuet I-II

Before looking for diversity, it is worth pointing out overall conventions in performances of these paired dance movements, often but not always, delivered as Menuet I, Menuet II, and Menuet I da capo. These conventions are largely upheld by both HIP and MSP violinists.

The two menuets are generally contrasted so that the first one is livelier and louder while Menuet II is softer and lyrical, more legato. Whether HIP or MSP, the opening measures of Menuet I are usually played forte and with heavier (e.g. Gähler, Kremer, Szenthelyi, Shumsky) or lighter accents (e.g. Mintz, Zehetmair). Those opting for a heavier accenting and slower tempo tend to play it that way all the way to the repeat sign while others (e.g. Shumsky, Tognetti) might play the bars with quavers (bb. 4 and 6) lighter or more staccato and with greater flow. The standard interpretation of the second half is to contrast the multiple stopping of bb. 9-18 with the linear quaver motion of bb. 19-26; the latter being softer, lighter, and inflected; usually more staccato but following the slur marks of the score. Szenthelyi and Hahn are perhaps the only violinists who play these quavers almost legato. In Barton Pine 2007, Schröder, Podger, Kuijken (both), Wallfisch and van Dael’s performances the quavers are lilted, creating a dotted effect, as I showed in chapter four under rhythmic alteration (cf. Audio example 4.5).

Clear four-bar phrases are the hallmark of most performances of Menuet II. Some articulate the slurred pairs of quavers more strongly than others (e.g. Perlman, Zehetmair, Shumsky) and in a few recordings (e.g. Huggett, van Dael) the figuration and slurring of bars 21-24 are emphatically articulated allowing for a shift of accents across the bar line (resulting in a sequence of quasi 5/4 + 4/4 bars). Phrasing is aided primarily either by gentle crescendo-decrescendo dynamic arches and tempo rubato (e.g. Luca, Tetzlaff, Barton Pine, Perlman, Huggett, Kuijken) or by agogic timing and rhythmic inflections (e.g. Gringolts, Tognetti, Lev, Shumsky). It is noteworthy that the names listed in brackets usually include both older and younger, MSP as well as HIP violinists.

The first half of Menuet II is often softer and more “reserved,” especially in the first four bars. Greater dynamic fluctuations can be observed in the second half even though the whole movement might end in pp (e.g. Barton Pine). Mintz represents an exception by finishing Menuet II forte. The slurred pairs are highlighted in most versions through gentle accents or slight elongation of the first notes of each pair (except by Wallfisch who instead plays the second notes very short and lifted). The pairing of slurred notes—something HIP considers common practice—is the least obvious in the recordings of Hahn (MSP), Schröder (HIP) and Kuijken (HIP).

Within these general observations there is, of course, almost infinite variety. What might be instructive to discuss further are some striking individual features. Analysis of Kremer’s two recordings and a comparison of Gringolts’ and Huggett’s version provide interesting insights into the complex interactions of performance elements as well as varied personal approaches; both contributing to subtle but significant interpretative differences.

Kremer, in his 1980 recording, creates a stronger than usual contrast within Menuet I by playing bars 18-26 and 29-31 in a “featherweight” soft style while otherwise maintaining a detached, marcato bowing. Menuet II is also unique in that the repeat has a greater dynamic range than the first play: there are stronger accents in bb. 17-20 and a stronger crescendo to b. 29; the f is louder in bb. 29-30 followed by p both times.

In the 2005 recording he plays both movements rather unevenly. Although bowing is fairly light, notes with triple stops as well as certain down-beats are forcefully accented. In Menuet I he creates a bar by bar change of f-p-f at the beginning of the second half which he did not do in 1980. He accents strongly other beats as well and there is a subito pp in b. 21 after which he plays faster, especially during the repeat.

Most of the interpretative strategies are already present in his earlier version, but in 2005 they are exaggerated. This is particularly true of the way he differentiates contrasting thematic materials in the first half of Menuet II through tempo fluctuations illustrated in Figure 5.1. With regards to interaction among performance features, it should be noted that the tempo fluctuation in Kremer’s 1980 recording is less noticeable partly because the dynamics are softer and bowing lighter in the “drone” passage while in 2005 these bars are played louder and less legato, creating a broad sound underscoring the slower tempo and the musette-like, rustic quality of the music (Audio example 5.7).

5.7. Similarities between subsequent recordings of J. S. Bach, E Major Partita BWV 1006, Menuet II, extracts: bars 1-16. Gidon Kremer 1980 © Philips; bars 1-16 and repeat of bars 1-4. Gidon Kremer 2005 © ECM. Duration: 0.52.

To listen to this extract online follow this link:

Figure 5.1. Summary of tempo fluctuation in Kremer’s two performances of bars 1-16, Menuet II (E Major Partita) showing the averaged metronome value for each segment in the 1980 recording and the 2005 version.

Tempo fluctuation can also be observed in the other two versions selected for closer analysis. Both Huggett and Gringolts contrast to Kremer because the flexibility in these recordings is more to do with timing inflections (i.e. rhythmic rubato) than change of tempo between the contrasting groups of bars. Importantly, Huggett’s and Gringolts’ performances also differ from one-another. They show contrasting strategies regarding what aspect of the music to bring out.

Gringolts highlights metric units; Huggett melodic-harmonic goals. Gringolts lays emphasis on downbeats (even in bars 21-22) by slightly delaying them and hurrying to the end of the measure. In contrast, Huggett follows melodic patterns across bar-lines creating shifted metric accents. Interestingly, Huggett’s more fluctuating and “terraced” dynamics actually follow the metric units rather than the melodic ones that she shapes so emphatically through tempo. She also strongly accentuates particular downbeats, more than Gringolts (e.g. bb. 24 and 25). These differences are shown in Figures 5.2 and 5.3 (Audio examples 5.8, 5.9).

Figure 5.2. Comparison of beat-level timing data of bars 5-9 of Menuet II (E Major Partita) in Gringolts’ and Huggett’s recording (first play) showing different interpretative approaches. Huggett phrases in pairs of bars (accelerating in the first and slowing in the second); Gringolts articulates each bar (accelerating to the middle and then slowing).

5.8. Contrasting interpretative strategies in J. S. Bach, E Major Partita BWV 1006, Menuet II, extract: bars 5-9. Two versions: Ilya Gringolts © Deutsche Grammophon (highlights metric units); Monica Huggett © Virgin Veritas (highlights melodic-harmonic goals). Duration: 0.16.

To listen to this extract online follow this link:

Figure 5.3. Comparison of beat-level timing data of bars 21-25 of Menuet II, E Major partita in Gringolts’ and Huggett’s recording (first play) showing different interpretative approach. Huggett tends to delay the second or third beat highlighting melodic contour or harmonic cadence. Gringolts tends to delay the downbeat and thus gives emphasis to the metric unit of the bar.

5.9. Contrasting interpretative strategies in J. S. Bach, E Major Partita BWV 1006, Menuet II, extract: bars 21-25. Two versions: Ilya Gringolts © Deutsche Grammophon (highlights metric units); Monica Huggett © Virgin Veritas (highlights melodic-harmonic goals). Duration: 0.14.

To listen to this extract online follow this link:

The audible differences between Gringolts’ and Huggett’s performances are significant yet so fine-grained that explaining the nature of the difference in “assemblage” requires considerable analytical attention. Both of them play unevenly, flexing the timing of notes creating local nuance and tempo variation. Only close listening and software assisted analysis help reveal their different “molecular” lines that diversely “deterritorialize” the ostensibly common HIP “territory.”14

In this section we have seen several examples of considerable diversity in “strata” formation (“thickening”) as performance features (“assemblages”) interacted within both the MSP and the HIP “territories” of the “milieus” of performance style. These differences underscore the complex, heterogeneous and dynamic inter-relationship of bowing, accenting, tempo, dynamics, timing, phrasing, articulation and ornamentation; at times leading away (“deterritorializing”) from MSP, other times weakening (“deterritorializing”) HIP, or moving towards a “nomad,” idiosyncratic “multiplicity.” To engage more with possible “rhyzomic” tendencies I now turn to the examination of multiple recordings of the same violinists.

5.2. Multiple Recordings of Violinists

There are five violinists who made more than one recording of the works (or certain sonatas or partitas) during the period under discussion (Table 5.2). Only one of them is a period specialist (Sigiswald Kuijken).

Table 5.2. Violinists who made more than 1 recording between 1977 and 2010.


Recording 1

Recording 2

Recording 3


Complete set 1980

Complete set 2005


Complete set 1983

Complete set 2001


B minor Partita 1987

3 Partitas 1993-4

Complete set 2008


Complete set 1994

Complete set 2005

Barton Pine

Complete set (concert) 1999

Gm and Dm 2004

Complete set (concert) 2007

It can be an interesting challenge to account for the differences among subsequent recordings of the same repertoire by the same artist. Such an exercise makes it rather explicit that we tend to find what we are looking for. The complex nature of performance allows for a multitude of observation points and some elements may be less prone to differ over time than others. If we focus on these we are likely to note similarities while features that might be more nuanced, hidden or harder to grasp may change more. The verdict also depends on where one draws the line: how big a difference counts as change? When does a “molar line” become a “molecular line” or even a “line of flight”? Are we looking for overarching characteristics, expressive-affective qualities or specific technical components and solutions? How many or what kinds of “molecular lines” and “lines of flight” do we need for transformation to occur; to arrive at a different “multiplicity,” “assemblage,” “territory”? Are we focusing on possible nuanced changes within one performer’s interpretation of a piece in the context of this same performer’s oeuvre or in comparison with the composition’s performance history?

Musicians tend to lay emphasis on their changing sensibilities and insights when asked to give reasons for re-recording a composition. Yet systematic comparative analyses more often result in identifying similarities than radical differences. Exceptions that come to mind are Glenn Gould’s 1955 and 1980 recordings of Bach’s Goldberg Variations, Rubinstein’s Chopin interpretation from the 1930s and from after the 1960s, or Leonhardt’s 1953 recording of the Goldberg Variations compared to the later versions from 1965 and 1973. What transpires, generally, is that subsequent recordings take similar ideas further; the underlying interpretative choices become more obvious with the passing of time, perhaps because the musician is more comfortable or confident about their view of the pieces; the message they wish to convey. It is plausible that frequent performing of works ingrains certain solutions that are hard to change radically.15 This seems to be the case here, too, especially in relation to the recordings of Kuijken, Barton Pine and Tetzlaff. However, some specific as well as conceptual differences can also be observed if one focuses less on the measurable and more on the affective aspects of these recorded performances.

Gidon Kremer

Among the violinists with multiple recordings of the works, Kremer seems to have changed his interpretation the most, if we discount for the moment the fundamental change between Mullova’s early B minor Partita recording and the two later recordings (Table 5.2). Kremer’s second version (2005) is much more strongly accented and articulated than his earlier playing (1980); he often chooses a slower tempo and his vibrato is narrower and less continuous. However, even he displays basic similarities in artistic approach.

We have already seen how the considerable tempo fluctuation observed in the 2005 version of the E Major Menuet II was, actually, just a stronger projection of the same idea regarding structural delineation of four-bar groups through contrasting tempo (Figure 5.1). Other examples include the A minor Grave and Fuga. Figure 5.4 clearly shows the strong similarities between the two recordings’ dynamics profiles, indicating similar conceptualisation of large-scale form. In both versions of the Fuga movement Kremer plays the episodes of bars 111-124 and 206-220 softly. Both times he creates a crescendo from bars 45 to 61 (the return of fugal subject) and 232 to 271 (gradual “climb” of thematic material from lower to higher register with wider leaps and thicker texture). Similarly, having reached a climax in bar 166, he then starts a gradual decrescendo towards b. 206 in both recordings.

Figure 5.4. Dynamic profiles of Kremer’s two recordings of the A minor Fuga.

Kremer plays the A minor Grave in a fairly intense, sustained (legato) style in both versions, with a clear distinction between louder multi-voiced moments of climax and softer-lighter linear and ornamental gestures. These dynamic contrasts are more extreme in the later recording but the interpretative principle remains essentially the same. In 2005 we can hear that the accent on the dotted F# in bar 8 is stronger but both times it is followed by a rapid decrease in dynamics in b. 9 before the new build-up of volume starts leading to a climax in b. 12, with the broadening and ritenuto being again more obvious in 2005. In the second half of b. 21 the low notes are played forte while the upward flourishes piano in both recordings but in the later version Kremer creates a stronger gesture by adding a breath (‘Luft-pause’) followed by subito piano before the second forte G (Audio example 5.10).

5.10. Comparison of subsequent recordings by same violinist in J. S. Bach, A minor Sonata BWV 1003, Grave, extracts: bars 5-9. Two versions: Gidon Kremer 1980 © Philips, 2005 © ECM; bars 19-21. Two versions: Gidon Kremer 1980, 2005. Duration: 3.22.

To listen to this extract online follow this link:

In a DVD documentary Kremer claimed that twenty years of life experience

cannot simply vanish, they return in music. I expect that my own understanding of Bach and my own abilities have changed. But not so much that you could say an entirely different person is playing. I am still the same Gidon Kremer; with a different violin, in a different church but with the same music.16

At this point the film cuts excerpts from the two recordings of the Ciaccona next to each other and they sound rather similar. Much greater differences can be observed in other movements. Later in the same film he makes the important point:

I also believe that we always remain in a composer’s debt. This will never change. Generations of interpreters will come and go but the music of Bach, Mozart and Schubert will remain as will that of certain contemporaries such as Shostakovich. And it will always be a mystery as to how it should be performed.17

Such statements are not fashionable nowadays in certain academic circles that question the existence of “works” and the composers’ privileged position and authorship but illustrate well the attitude that contributes to Kremer’s reputation as a musician with “great integrity.”18

Rachel Barton Pine

Other violinists with multiple recordings show an even greater tendency to take their interpretation further in the same direction as before. The performances of Barton Pine are essentially similar. The general character of the movements tends to be the same across the versions even though surface differences can be noted. Overall, her 1999 broadcast recording sounds more “individual” or unique than her 2007 concert performance, but the latter is perhaps more polished and blends more with the performance tradition of the works. The metrically based gestures and more detailed articulation of the earlier version make that interpretation closer to HIP. Her commercial disk from 2004 (that contains only two of the six works) also shows a nice balance of HIP and MSP characteristics.

She plays the A minor Grave much more lightly and flowingly than Kremer, already in 1999. Some of her phrasing strategies and dynamics are similar to Kremer’s but less extreme (e.g. p in bb. 4 and 12; slightly accented, louder high F#s in b. 8). By 2007 the relaxed flow is even more prominent because many dotted notes are simply stressed rather than dotted; played as if a starting note of an ornament (e.g. b. 9 beat 2) and the dynamics remain predominantly mf and mp. Noticeable expressive gestures are similar to the 1999 version. On both occasions she plays the F# in b. 8 louder, slightly stressed; the last two semiquavers of the first two beats in b. 19 staccato; and the low A and G in b. 21 (beats three and four) clearly separated. The phrase arches are also similar, highlighting the same moments of climax and repose (Audio example 5.11).

5.11. Comparison of subsequent recordings by same violinist in J. S. Bach, A minor Sonata BWV 1003, Grave, extracts: bars 5-12. Rachel Barton Pine 1999 © Chicago WFMT 98.7; bars 9-12. Rachel Barton Pine 2007 © The Artist; bars 19-23. Two versions: Rachel Barton Pine 1999, 2007. Duration: 3.25.

To listen to this extract online follow this link:

She performs the A minor Allegro in virtuoso style on both occasions even though some details differ: Both are fast and have light detached bowing but in 1999 more notes are played staccato (e.g. in echo bars, in b. 33 during repeat, the un-slurred semiquavers in bb. 48-50). The flourish linking back to the repeat of the first half is more pronounced in 1999 (only a 3-note scale in 2007). She plays bars such as 17, 27, and 42 fast but with well-defined rhythm in 1999. There are also more stressed notes and slight “breathing pauses” (luft pause) that articulate the large-scale form and delineate sections, figurations, harmonic goals. The 2007 performance has a consistent level of dynamic while the echo effects are more pronounced than in 1999 (e.g. p in b. 56).

Other minor differences include the flourish at the end of the E Major Gigue. It is performed only in 1999. On the other hand, she plays Menuet I with slightly lilted inégalité only in 2007. Her playing seems to have more flow in 2007, probably because the shorter musical units are integrated into longer melodic lines (e.g. no echo effect in bars 22-3 of E Major Menuet II but rather played as part of one long phrase). But she adds more ornaments in 1999.

Barton Pine’s tendency in the 2007 version towards a smoother, perhaps slightly more “mainstream” phrasing is already noticeable in her commercial disk from 2004 which includes the G minor Sonata and the D minor Partita only (compare, for instance, the D minor Allemanda or G minor Presto movements from 1999 and 2004). An exception is the D minor Sarabanda. Here articulation, phrasing, dynamics and bowing remain very much HIP-like (not to mention the vibrato-less tone) and it is the 2004 version that has added ornaments (bb. 6, 11-12, 18-19, cf. Audio example 4.6). She remains one of the few violinists (along with Tetzlaff and Mullova) who regularly perform the complete set in concert. A recent such occasion was on 23 August 2014 when she performed all six pieces along with other German baroque solo violin compositions in a marathon concert at the Ravinia Festival.19 Her commercial recording of the complete set is scheduled for release in March 2016.

Christian Tetzlaff

The changes in Tetzlaff’s two recordings show an overall increase in expressivity, at least in certain movements. Most noticeable is a larger range of dynamics in 2005, resulting in greater local detail and more obvious shaping of larger-scale units. These differences tend to be noted by commercial reviews as well. The earlier version is described as “intelligent, carefully considered […] musically imaginative without any effusions of Romantic sentiment”20 while the second recording as having “a remarkable air of spontaneity, the result of a pervasive rubato […].”21

According to an interview in Strad, at the time of the first recording Tetzlaff found it “most important not to make [the Sonatas] sound like contrapuntal exercises.” He relished their “speaking, dramatic quality.”22 He considered the Partitas to be “quite different—a secular counterpart, with lots of dances.” Tellingly, he added “But for a long time no one would really dance in this music,” hinting at his allegiance to HIP.23

Tetzlaff’s current views on the Bach pieces are explored in a 2012 interview for the New Yorker which seems to confirm that nowadays he aims for a more personally involved and intimate reading of the pieces.24 In the article Tetzlaff is attributed seeing “the cycle as Bach’s ‘personal prayer book,’” and the interviewer claims that “Tetzlaff’s mystical side comes out most strongly when he speaks of Bach.” He reports a concert of the Solos in Dresden during which the “slower movements were almost uncomfortably introverted,” Tetzlaff exposing “layer after layer of vulnerability, creating an atmosphere of naked confession.” And in conversation Tetzlaff confides:

Bach’s music confronts the player and the audience in a very personal situation, in a very alone way.25 And I try at that moment to put away pretensions—in levels of violin playing, pretensions of being a strong man, of being invulnerable—and instead say, ‘This is where all of us have common ground.’ Most of the time, we try to tell ourselves ‘I am confident’ or ‘I am doing well.’ But then, in a moment alone at home, you feel how close you are to some kind of abyss. […] Music, even at terrible moments, can make you accept so much more—accept your dark sides, or the things that happen to you. Maybe it’s just because you see that this is a common trait for all of us. You see that we are not alone. […] And that’s what the concert situation is about for me, when I am sitting in the hall and also when I am playing myself. It’s about communication—I almost want to say ‘communion.’ As a player, you really don’t interpret anymore. You listen, together, with the audience.

I read this article long after I had completed my analyses of the recordings and cited it at length because it is a very personal confirmation of the observations I was struggling to put into scholarly language. A reviewer of this second recording has actually found that while his “interpretations have deepened” some listeners “may find them fussy.”26

The fugues of the A minor and G minor sonatas are good examples of the changes towards greater expression and detail in Tetzlaff’s second recording. Both movements are played rather softly in 1994 with an easy flow and forward momentum, but there are only few and very discretely shaped cadence points that might aid the listener in hearing or identifying the different structural sections and phrases. Furthermore, Tetzlaff shapes and groups melodic-harmonic-rhythmic gestures much more obviously in 2005 and throughout these fugal movements, not just at the beginning. He creates many more “mini goals” towards which the music is moving and then arrives, mostly achieved through surges in dynamics and increased stresses on crucial harmonic or melodic notes. At times he also uses slight ritardandos to highlight moments of arrival. The cadence points in bars 18, 45, 73 of the A minor Fuga are much more audible in the 2005 recording than in the earlier version. Comparing the ending of the two performances one can notice that Tetzlaff starts a crescendo in measure 240 on both occasions but in 1994 it seems aborted, it does not develop. In 2005, on the other hand, it leads to bars 250-251 and restarts in 252 with a climax in b. 257 and cadence in 262 followed by a decrescendo in 268-269. He creates a new rise and fall between bars 269-280 with a final crescendo leading to the climax of the fioritura in 286-287 and the final two bars. Furthermore, the slurred pairs of quavers and various groups of semiquavers are much more strongly articulated in 2005. He shapes them by leaning on the first note in each group and playing the rest faster and lighter (Audio example 5.12).

5.12. Comparison of subsequent recordings by same violinist in J. S. Bach, A minor Sonata BWV 1003, Fuga, extract: bars 240-289. Two versions: Christian Tetzlaff 1994 © Virgin Veritas, 2005 © Virgin Classics. Duration: 2.46.

To listen to this extract online follow this link:

The greater expressiveness can also be observed in the G minor Adagio. Both readings have an air of improvisation but the flexibilities in the 2005 version are more linked to harmony and melody and less to the bar and metre. This makes the later performance sound freer, more gestural. However, as I have shown in earlier discussions (chapter four), at times this surge in musical expression leads to quite romantic sounds, especially when vibrato is used more prominently, when both tone and dynamics exploit extreme ranges of ppp and when bowing is more sustained serving a longer-spun melody, such as in the A minor Andante or the E Major Loure. It is perhaps this change in the underlying artistic approach from ostensibly HIP to more subjectively grounded that Tetzlaff has referred to when stating in an interview, “I have new ideas about them [i.e. the Bach Solos] and I would love to record them again.”27

The greater expressivity notwithstanding, Tetzlaff’s two recordings also show many crucial similarities, especially in surface detail. Importantly, when these surface details differ, the gesture or shaping tends to become stronger (i.e. more noticeable) rather than going in the opposite direction. But doesn’t this contradict what I have just claimed, namely a change in the underlying artistic conception? The complexity of music performance becomes quite obvious when one wishes to tease out these differences and similarities. What is the underlying conception and what is surface / technical detail? When does an effect become a cause? Which elements (“lines of flight”) create a sense of different aesthetics (new “multiplicity”) and which seem to make the same point simply more audible (thickening “molar lines”; territorializing the “assemblage”)? When do we cross an aesthetic threshold by making a gesture stronger? Harnoncourt said many years ago that “when we emphasize one specific aspect, another specific aspect is weakened until it disappears. We do not just have more and more expression.”28

Contemplating these issues illuminates what Deleuzian language might formulate as “expression is not simply an effect of material relations, but in turn acts on those relations.”29 If a performance is considered to be an “assemblage,” a “multiplicity of heterogeneous elements” unified by the co-functioning of the role the components play and whether these contribute to stability (“territorialisation”) or to loss of identity (“deterritorialisation”),30 then pondering thresholds of interpretative solutions, technical details, affect, aesthetics and style is an exercise where Deleuzian thinking may be harvested usefully.31 It is an interrogation of the processes of interactions, of transformation, the process of performing; the moment that we listeners perceive holistically, in its totality. The analyst, on the other hand, while trying to describe the moment and account for the perceptual experience, is stuck in the domain of words. By the time the interactive elements are dissected and the phenomenon described, the whole is somewhat lost and we seem to be left with scattered debris, interesting bits and pieces that once belonged to the object of our wonder and to our sensed experience. More importantly, perhaps, by the time the analysts has accounted for the elements contributing to the experience, the perceived moment has long passed and the multiplicity of heterogeneous elements has already configured (transformed into) a different assemblage. Therefore it is important to emphasize that the analytical observations refer to more global impressions, they are not trying to pin down moments, to explain moment to moment causes and effects. My aim is to explore the overall nature of each interpretation; I consider the totality of a movement or piece as the “assemblage” and “territory” even when highlighting specific moments (bars, beats, notes) in the recorded performances. When I look for similarities and differences I am thinking of “molar” and “molecular” lines that thicken or thin the “territory” as represented by the whole.

In that sense, when investigating similarities between Tetzlaff’s two versions I am able to note only one major difference between his two interpretations of the B minor Allemanda, for instance: the considerably greater dynamic differentiation of the first repeat in 2005. The repeats of both halves are played softer in the later recording, but only the first half is performed louder than mf the first time, so the contrast is most striking in that section. Otherwise the light, quasi legato dotting, the “flowing-continuous” style is the same in both recordings, and similar notes or note groups are stressed on both occasions and in both first plays and repeats. This single difference in dynamics does not have transformative or “fracturing” power. The two recordings of this movement portray an essentially identical “territory,” but perhaps with slight variation in “assemblage.” At the same time, interpretative gestures become somewhat stronger in the later version: The elongation of each first triplet note in b. 15 is more pronounced and cadence points are highlighted by greater Rits (e.g. bb. 8-9 and, especially, 18-19) in 2005. For me these represent “molecular lines” as they “deterritorialize”; they contribute to the process of transforming the later recording into a more expressive reading or “territory” (Audio example 5.13).

5.13. Comparison of subsequent recordings by same violinist in J. S. Bach, B minor Partita BWV 1002, Allemanda, extract: bars 15-19. Two versions: Christian Tetzlaff 1994 © Virgin Veritas, 2005 © Virgin Classics. Duration: 0.51.

To listen to this extract online follow this link:

Greater dynamic contrast between first play and repeat is the most striking difference between Tetzlaff’s two versions of the A minor Andante as well. Both recordings provide a gentle, lyrical reading with the first play utilizing “reverse swells” (leaning on main notes followed by release (decrescendo)) to highlight appoggiaturas while in the repeat these are played more like accented notes. Both “assemblages” have similar phrasing and fluctuating dynamics, with the later version using greater contrasts and a more forward-moving and flexible tempo. Crescendos and diminuendos are stronger and broader, creating a more passionate and dynamically varied performance. Bars 16-17 are played louder, with more intensity, while bb.21-23 are softer, Tetzlaff keeping the dynamic nuances within the pp-mp region. The accents on the G-D-F triple stop in b. 13 and the first and third semiquavers of the final beat of b. 23, the Rit in bb. 8-9 as well as the Rall in b. 25 are much stronger in the later version but do not represent new interpretative decisions. They are simply thicker “molar lines” contributing to stability; to a clearer delivery of interpretative intent. In neither version are the paired slurs in b. 15 emphasized; on both occasions Tetzlaff plays them in a straightforward legato manner (Audio example 5.14).

5.14. Comparison of subsequent recordings by same violinist in J. S. Bach, A minor Sonata BWV 1003, Andante, extract: bars 15-19. Two versions: Christian Tetzlaff 1994 © Virgin Veritas, 2005 © Virgin Classics. Duration: 0.43.

To listen to this extract online follow this link:

Sigiswald Kuijken

Throughout chapters three and four I repeatedly claimed that Kuijken’s two versions are surprisingly less HIP in many respects than one would expect and what critical reception tends to assert. This is an impression I have developed through long hours of comparative listening and it is only true to a certain extent; or relatively speaking. When I first heard it in the 1980s (and before I heard Luca’s), it sounded very different to those I grew up with (e.g. Grumiaux, Milstein, Szeryng). However, close listening to Kuijken’s first version in comparison with other period violinists’ recordings (especially that of Luca and later Podger, Huggett and many others) made me more aware of the range of possibilities and the scope of historically informed playing in this music. I was eager to hear his second version when it came out in 2001. I thought Kuijken would make his interpretation more radical, given the extra twenty years of experience and stylistic developments as well as the evidence of his recordings with La Petite Bande, for instance. However, overall there is not much difference between the two, at least not in terms of basic conception and approach (cf. Table 3.3 for rating of performance features in both but also Tables 4.4 [vibrato] , 4.7 [dotting], 4.8 [Gm Fuga]). In the next section of this chapter I will discuss some interesting and telling differences that close listening and analysis can reveal (cf. Audio examples 5.17 and 5.21). These tend to show a return to less detailed, more even or literal playing, in other words, towards more “authentistic” style, just like we saw when comparing performances of Rachel Barton Pine. But here I rather briefly summarize the main similarities.

Tempo, phrasing, bowing and movement characters are fairly similar throughout the two recordings but particularly noticeable in the C Major Largo and Allegro assai, the D minor Allemanda and Sarabanda and the G minor Siciliana and Presto. The later recording has a different ambience—much more reverberant acoustics—that often calls forth a different initial impression, one that feels “heavier,” more laboured, especially in the fast finale movements. This feeling is supported by the somewhat slower tempos. At times bowing is also broader (more sustained) and the articulation and phrasing more legato. But these are just slight differences in degree, not in kind. The consistent dynamics, the similarity of accents, added ornaments (e.g. b. 12 in D minor Sarabanda), and phrasing across the two recordings impress the casual listener more than these slight differences that one notices when listening more attentively.

Viktoria Mullova

If Barton Pine, Tetzlaff and Kuijken seem to have toned down the HIP elements of their playing in their later recordings, the trajectory of Mullova’s performances illustrates the opposite path. I have already shown ample evidence of her playing becoming increasingly similar to how period violinists perform baroque music. Not just similar, but in many respects at the cutting edge, leading the way.32 This is revealed in her phrasing, bowing and ornamentation, in particular. The deliberate and radical change in her approach emerged in the early 1990s and in that she is similar to Leonhardt, for instance, whose 1953 Goldberg Variations recording is nothing like his 1965 or 1973 versions.

Mullova once said she was pleased that Philips agreed not to release her recording of the Bach Solos from the 1990s because she really disliked them.33 She must have been thinking of the three Sonatas only because the three Partitas have been issued and already show the HIP-Mullova in the making. Interestingly, although the Gramophone reviewer finds her playing on this disk “a breath of fresh air” and notes that “Her dance movements are light on their feet (helped by her quickness of bowing through chords) and her nuancing of tone and volume is full of subtlety—warm and expressive but with no trace of romanticism,” he nevertheless compares her disk with Kagan’s and Shumsky’s MSP versions.34 As if the HIP qualities were not recognized as such, given Mullova’s reputation at the time as a virtuoso soloist of romantic concertos! Such slippage helps maintain false views of trends and under-informs readers. Still, the reviewer “tentatively add[s] the disk to [his] ‘treasure island’ collection” and decides to wait “hopefully for the sonatas to join it.” Although he had to wait for more than ten years, John Duarte would probably agree that it was worth it. What Mullova achieves on the complete set recorded in 2007-2008 for the Onyx label is not so much the radical change from her interpretation of the B minor Partita in 1987—this is already witnessed on the 1993-1994 disk of the three Partitas—but mastery of period technique and complete freedom of play that comes with full command and “ownership” of style.

The 1992-1993 recordings can still sound a little mechanistic due to consistency of pulse and accents. Several movements of the B minor Partita are played at a considerably slower tempo than in 1987 and with more staccato bowing; both contributing to a somewhat laboured and artificial effect; a self-conscious attempt at HIP. By 2008 the bowing is shorter and bouncier; accents more varied, delineating shorter or longer segments; tempo a little faster (similar to 1987) assisting greater flow; phrasing is freer; ornamentation more abundant. Movements like the D minor Giga have no time to become mechanistic because there is a constant ebb and flow of dynamics—partly due to varied bowing—and accentual detail.

The three recordings of the Tempo di Borea movement from the B minor Partita illustrate these observations well. The 1987 version has a moderate speed (minim = ca. 77 beat per minute) and a detached style of bowing. Accents fall on each crotchet creating a 4/4 rather than cut C () pulse. The quavers are played evenly. Typically for performances of this movement, the second half includes contrasting dynamics where the more linear measures are played softer and with shorter bow strokes. The 1992 version is considerably slower (ca. 67 bpm) and much more staccato. This combination eventually makes the movement sound rather choppy. However, the 4/4 pulse is weaker, at times the accents hinting at . The approach to dynamics is similar to the earlier recording. The 2007-2008 recording is the fastest (ca. 82 bpm), has fewer accents and the notes and bars are much more strongly grouped bringing forth the cut C pulse. Dynamics fluctuate primarily as a result of bowing, although the linear bars of the second half are played softer on this occasion as well. In this version the performance has not simply regained some of the flow and richer tone of the 1987 version but took the interpretation to a different level. The flow is of a different kind. It is a “line of flight” that projects and follows the metre. Because the accenting often seems to be simply a side-effect of differences between up- and down-bow strokes, the pulse does not become mechanistic or repetitive-predictable; it does not “territorialize” but rather “moves between.” Bow pressure and speed contribute to a chiaroscuro effect enriching the dynamic palette. In hindsight, the staccato bowing in the 1992 version seems like an artificial and abortive attempt (a rhizomic molecular line getting lost) at what the 2007-2008 recording achieved through mastery of baroque bowing (Audio example 5.15).

5.15. Comparison of subsequent recordings by same violinist in J. S. Bach, B minor Partita BWV 1002, Tempo di Borea, extract: repeat of bars 39-50. Three versions: Viktoria Mullova 1987 © Philips, 1992 © Philips, 2008 © Onyx. Duration: 0.57.

To listen to this extract online follow this link:

In sum, the program notes to a concert of Beethoven’s sonatas by Mullova and Kristian Bezuidenhout justifiably claimed that “Her recent recording of Bach’s Solo Sonatas and Partitas represents a significant milestone in Viktoria’s personal journey into this music.”35 Tim Ashley enthused already in 2000 after a concert showcasing the E Major, G minor and D minor works that “In Viktoria Mullova—whose technical perfection combines with matchless, uncompromising interpretative subtlety—they [the Bach Solos] find perhaps their ideal interpreter. […] To hear Mullova play Bach is, simply, one of the greatest things you can experience […].”36

5.3. The Holistic Analysis of Interpretations

The previous section has already engaged with issues that underscore the complexity of analysing music performance. Here I focus on these matters by teasing out further the subtle differences (“territorializations” and “deterritorializations”) between multiple versions recorded by the same violinists. Bringing the various elements together and weighing up their relative contribution to the “multiplicity” each performance assembles lead to addressing the interpretations in terms of affect and / or musical character. In my consideration of the complex interactions of both compositional and performance features prompting the potential affective response, I often resort to metaphor and formulate arguably subjective judgements. In some instances “objective” measurements of any one of these elements would simply give undue emphasis to something that does not act alone but in tandem with several other (non-measurable) elements, creating in-the-moment effects and transformations of meaning. In these cases aural analysis, or what might be called “musicological” or “close / focused listening,”37 better equips the researcher to ponder what one perceives than quantified data. It also enables ecologically valid propositions. In other instances software assisted analysis is important to clarify perception. It can “guard against mistakes such as when an accent is attributed to dynamics [but] was in fact the result of an agogic (temporal) emphasis.”38 I continue comparing primarily the multiple recordings of Kuijken, Tetzlaff, Kremer, Barton Pine and Mullova because these can serve as good test cases for both approaches. The differences are usually rather subtle and often quite hard to put into words: The analyst is forced to face the challenge of interrogating the interaction of performance features—the “relations of causality” to use Deleuzian language—and to grasp each reading holistically as well as in the complexity of their detail. This way the truism that each performance is different may open up to scholarly explanation.

“Subjective” Aural Analysis: The D minor Giga

Take the various versions of the D minor Giga, for example. Focused listening would observe the projection of pulse, the presence of accents, tempo, bowing, articulation, and delivery of the marked terraced dynamics. The overall effect may be virtuosic or dance-like or perhaps a bit of both. These could then be regarded as either tending towards MSP (virtuosic) or towards HIP (dance-like). In this regard the general observation is that the more recent versions of Barton Pine, Kuijken, and Tetzlaff tend to be less detailed and more virtuosic.

Barton Pine plays a little faster in 2007 and with fewer accents and agogic stresses than earlier. The pulse is still very perceivable and there are obvious phrases but because of the faster tempo, more evenly flowing, “uninterrupted” stream of semiquavers it sounds more virtuosic, overall. Alternatively, I could say that the 1999 version has more ebb and flow in terms of dynamics as well as tempo and the agogic stresses are part of that ebb and flow, making the performance sound more phrased and detailed while being virtuosic. Her 2004 version is closer to the earlier than the later reading and I would be hard pressed to pinpoint any perceivable difference apart from variation in the ambiance of the recording (the 2004 version seems to have a thinner tone, a more “distant” sound, perhaps due to microphone placement). In her three recordings then it is tempo that is perhaps the “thickest molar line” “territorializing” the virtuosic effect. The molecular lines of ebb and flow, dynamics, and agogic stresses “deterritorialize” but do not transform the virtuosic character to something else, such as dancing (Audio example 5.16).

5.16. Comparison of subsequent recordings by same violinist in J. S. Bach, D minor Partita BWV 1004, Giga, extracts: bars 1-5. Two versions: Rachel Barton Pine 1999 © Chicago WFMT 98.7, 2004 © Cedille; bars 1-11. Rachel Barton Pine 2007 © The Artist. Duration: 1.03.

To listen to this extract online follow this link:

Kuijken’s two versions of the D minor Giga are not about virtuosity overall. The 1981 recording projects a strong pulse and groups shorter units and phrases. Bowing is light, terraced dynamics are observed. It sounds like a standard HIP reading; bouncy and dance-like, with easy flow and clear architecture based on the piece’s harmonic structure. The 2001 version has a more relaxed tempo (ca. 74 bpm in contrast to ca. 85 in 1981) but also much less detail. It quickly starts sounding mechanistic. It is not that there are fewer accents but that the accents are uniformly executed without the ebb and flow of dynamic shades and subtlety of bowing. Perhaps a slightly faster tempo would push the performance towards the virtuosic type, but the rather routine delivery of semiquavers and regular accents would probably counter-act anyway. Together with the slower tempo, these instead create a performance that evokes the old-fashioned, “sewing-machine” or clockwork-like MSP style of the 1950s and 1960s (Audio example 5.17).

5.17. Comparison of subsequent recordings by same violinist in J. S. Bach, D minor Partita BWV 1004, Giga, extract: bars 1-10. Two versions: Sigiswald Kuijken 1983 © Deutsche Harmonia Mundi, 2001 © Deutsche Harmonia Mundi. Duration: 1.02.

To listen to this extract online follow this link:

Tetzlaff’s two versions further support some of my claims about Kuijken’s 2001 recording in the above two sentences. Tetzlaff’s playing is virtuosic on both occasions. In 1994 he uses relatively short bow strokes, plays with light pulse, and light tone. From about bar 9 onwards he starts to phrase in shorter units and introduces clearer accents. The performance does not sound mechanistic because of the near constant ebb and flow of dynamics and shadings of tone through bowing. In contrast, the 2005 recording takes virtuosity a step further by increasing tempo (from ca. 83 to ca. 89 bpm). The section between bars 3 and 9 is quite smoothly virtuosic with even, rapid flow. Although obvious downbeat accents and clearer phrasing are introduced from bar 10, the overall impression becomes mechanistic because, perhaps due to the fast tempo, the accents are uniform and routine; there seems to be no time to shape the tone, the phrase, the harmonic or melodic unit (Audio example 5.18).

5.18. Comparison of subsequent recordings by same violinist in J. S. Bach, D minor Partita BWV 1004, Giga, extract: bars 1-12. Two versions: Christian Tetzlaff 1994 © Virgin Veritas, 2005 © Virgin Classics. Duration: 1.05.

To listen to this extract online follow this link:

In the four versions played by Kuijken and Tetzlaff, tempo, accenting, bowing, dynamics and shades of tone interact in ways that create diverse “multiplicities.” The interaction of tempo and regular accenting form “molar lines” causing thickening in the “strata.” Slower tempo combined with regular accents creates “assemblages” belonging to the “territory” frequently described as “sewing-machine-style” or modernist-authentistic HIP (Kuijken 2001).39 Assemblages where the combination of regular accents and faster tempo are the key stabilizing features “territorialize” virtuosic readings (Tetzlaff 2005). The impact of these “molar lines” is weakened, the territories of virtuoso and authentistic HIP are deterritorialized when bow-strokes and pulse are varied and become “multiplicities” in themselves. They make room for the “in between molecular lines” of ebb and flow, rapidly shifting dynamics and the “betweeness” of tonal shades (Tetzlaff 1994).

The two other violinists who made more than one recording of the piece during the period are Kremer and Mullova. Contrary to those discussed above, in their case it is the later, more recent recording that is more detailed (and less categorizable, i.e. more “deterritorialized”).

Kremer’s two recordings have similar tempos (the earlier being faster at ca. 85 compared to ca. 81 bpm in 2005) and overall approach. However, in the later recording Kremer uses stronger and more frequent accents, often quite harsh and sharp. These tend to sound ugly (e.g. tied E in bb. 14-15; G natural in b. 19) because they lack purpose; they are not properly integrated into the flow of the music. These accents interrupt an otherwise even and virtuosic rattling off of uniformly controlled passage work (Audio example 5.19). These “lines of flight” break away from the normative but the transformation is not for the better; just leading to randomness.

5.19. Comparison of subsequent recordings by same violinist in J. S. Bach, D minor Partita BWV 1004, Giga, extract: bars 12-20. Two versions: Gidon Kremer 1980 © Philips, 2005 © ECM. Duration: 0.52.

To listen to this extract online follow this link:

The crucial role of tempo when interacting with accenting, phrasing and dynamics is highlighted again when comparing Mullova’s two versions. Her 1993 recording of the D minor Giga has a relaxed tempo (ca. 74 bpm). The pulse is perceptible and the phrases are pointed out. She observes the terraced dynamics. However, the regularity of accents tends to make the music sound mechanistic, especially in bars 7-15 and 25-30. This is all the more obvious when one compares it to her 2008 recording. Here she plays faster (ca. 85 bpm), with similarly short bowing but stronger and more frequent accents and stresses. In fact she articulates shorter units of various lengths. This helps counteract any sense of uniformity because the faster tempo and constantly shifting tone, dynamics and accentual detail combine to draw attention to melodic-harmonic directions, pulse, ebb and flow (Audio example 5.20).

5.20. Comparison of subsequent recordings by same violinist in J. S. Bach, D minor Partita BWV 1004, Giga, extract: bars 6-16. Two versions: Viktoria Mullova 1993 © Philips, 2008 © Onyx. Duration: 1.10.

To listen to this extract online follow this link:

While Tetzlaff’s faster tempo seemed to limit the opportunity for phrasing and shaping and contributed to a more mechanistic impression, Mullova’s faster tempo seems to serve flow and aids the moulding of musical units. The “multiplicity” of each performance feature (not just their interaction with other elements) is shown here. At times they territorialize, other times they deterritorialize to a greater or lesser extent; at times a feature may be the thickest molar line, while other times it works in tandem with other elements. In Mullova’s 2008 recording the performance features seem to be completely rhizomic and non-categorical. To describe their full nature, function and interactions is beyond words, they are “in between,” constantly shifting and breaking away; they have to be heard.

“Objective” Measures:
The A minor Grave and G minor Adagio

Software assisted analysis is also helpful in clarifying what is going on in a performance. Listening to Kuijken’s two versions of the A minor Grave I noted that the earlier recording sounded freer, more improvisatory while the more recent interpretation more measured but still flexible. I guessed that perhaps there was less tempo fluctuation in the second recording and the tone also sounded more even, indicating less tonal inflections with the bow. Bowing felt more legato, with fewer swells and gaps between groups of notes or at phrase ends. Nevertheless changes in dynamics were quite clearly audible but perhaps serving a conception of the music on a larger-scale; conveying greater order and progressing in bigger chunks.

Trying to mark up the sound files for detailed observations in Sonic Visualiser immediately clarified how much more measured the 2001 recording was, indeed.40 Marking beats were reasonably straightforward as most flexibility unmistakably fitted with metric units. In contrast, working with the earlier sound file was quite hard. The kind of flexibility Kuijken adopted in 1981 followed gestural content and had little to do with pulse or metre. When I was marking perceived phrase boundaries I rarely found myself noting downbeats, or beginning of beats, but much more often the second or third beat of a bar or the off-note of the downbeat or another beat. It is this lack of pulse that lands the 1981 performance a “preluding” character, an improvisation that sounds as if freely following the fancy of the player—a fantasia (Audio example 5.21).

5.21. Comparison of subsequent recordings by same violinist in J. S. Bach, A minor Sonata BWV 1003, Grave, extract: bars 1-8. Two versions: Sigiswald Kuijken 1983 © Deutsche Harmonia Mundi, 2001 © Deutsche Harmonia Mundi. Duration: 2.27.

To listen to this extract online follow this link:

Figure 5.5. Beat-durations and dynamics in Kuijken’s two recordings
of the A minor Grave bb. 5-8

In Figure 5.5 the tempo lines for bars 5 to 8 of each recording show how much more volatile the 1981 version is. The power curves confirm that the use of dynamics is also different. The bigger and more frequent slopes in the 2001 version indicate a greater role of standard crescendos—decrescendos in aid of grouping the notes of metrical units. The relatively more arch-like dynamic line of the 1981 version hides its many rapid oscillations within a more gradually changing range. Perhaps the frequency of data capture or the algorithm that Sonic Visualiser uses to measure power would need to be changed to register the dynamic nuance that results from varied bow speed and pressure as Kuijken’s fingers and bow glide through the rapid notes, crossing strings to remain in the lower positions and creating a series of typical baroque chiaroscuro effects.

In the earlier version the intricate rhythms of the notated score are played with fluidity and an exclusive focus on harmonic and melodic goals (rather than metre / pulse). Although I can illustrate and evidence my observations by providing graphs of tempo and dynamic fluctuations, these quantified measurements certainly do not tell the entire story. Rather, they simply signal aspects of the performance that subsume all the other key contributing elements: tone, shades, bowing, as well as the fundamental decision that governs the interpretation, namely whether metre rules or freedom of melody and harmony.

Another potential limitation of measurements can be illustrated by presenting the graphs of phrase durations in Kuijken’s two recordings of the A minor Grave (Figure 5.6). In this case the data show up the similarities between the recordings, namely that both are flexible and “improvisatory.” However, given the discussed perceptual experience of close listening that could be demonstrated by graphing beat durations (Figure 5.5) such a graph is somewhat misleading and certainly hides important information regarding the fundamental differences between the two versions: the alternative approach each utilizes to create flexibility and the diverse effects these changes achieve. If only the graph at Fig 5.6 is offered, homogeneity in performance might be evidenced, completely missing and thus denying the existence of crucial musical and affective variances and transformations.

Figure 5.6. Comparison of phrase durations in Kuijken’s two recordings of the A minor Grave misleadingly showing similarities in tempo flexibilities

Analysis of Kuijken’s two recordings of the G minor Adagio provides similar results. The later recording sounds improvisatory and reasonably flexible with some dynamic nuances and swells on notes but closer inspection shows a much more measured delivery of ornamental groups (e.g. bb. 6-7) that is anchored in the underlying metre / pulse. The earlier recording is much more improvisatory because of metrical freedom, greater flexibility of tempo rubato, and constant shifts in tone and bowing. The ornamental groups are often hurried or rushed-over capriciously as they lead to melodic high- or low notes, meditative-exploratory gestures—for instance melodic re-starts, repetitions, wayward, indecisive turns, momentary harmonic goals that continue on unexpectedly—and cadence points.

Perception of Affect

In a recent study, Spitzer and Coutinho referred to my 2005 paper on the performance history of Bach’s solo violin works and lamented the fact that it “has not touched on the affective dimension.”41 They found this regrettable especially because of “an entrenched school of thought,” promoted among others by Daniel Leech-Wilkinson,42 “that holds that emotional expression in music is mostly influenced by performance style, rather than the music’s acoustic features or formal structure.” They embarked on an initial investigation to “ascertain whether the acoustic features of musical emotion […] converge with analytical findings […]” and used Kremer’s 2005 recording of the G minor Sonata to test their case.43

The listeners in Spitzer and Coutinho’s study found Kremer’s 2005 recording of the Adagio movement to express “sadness, sorrow, melancholy, tenderness” but also “tension” and “being moved.”44 Spitzer and Coutinho did not analyze Kremer’s performance in detail. Instead their reasoning for affective response was largely based on score analysis linked to contemporary psychological investigations of emotional expression in music.45 They claimed that the G minor Adagio’s “structural features suggest sadness”:

a slow tempo, minor-mode key, narrow intervals, legato articulation, variability of texture, preponderance of descending melodic contours, and high level of dissonance, especially involving semitone appoggiaturas (an ornament which “leans on” the main note a step above). In musical semiotics, such appoggiaturas are historically associated with pianti, or crying, figures (Monelle, 2000), as if the musical contour were iconically representing the sound of a sobbing voice.46

They also noted that

David Huron has made the connection between sadness in music and the “detailed-oriented thinking” of “depressive realism” (Huron, 2011, p. 48). Our analysis of the Adagio reveals it to be highly fragmentary in its texture and tonal structure, and this ‘atomistic’ quality could be related to the detail-oriented quality of its sad affect.47

“In the Adagio,” Spitzer and Coutinho assert on p. 42, “the ideas evolve fluidly from measure to measure.” And on p. 43 they claim that “The atomization of the schema in the Adagio suggests the lack of goal—lethargy—connected with sadness or depression.” All these points can be useful as one considers differences among recorded performances.

Figure 5.7. Score of G minor Adagio, bars 1-12

When I listen to Kremer’s 2005 version I can hear a flexibility that is nevertheless somewhat jolted, not at all fluid and certainly contributing to a fragmented, atomistic quality. The bars with demi-semiquaver and smaller note values have considerably more fluency than bars with only semiquavers (e.g. b. 6). These are played in a rhythmically literal way, without a sense of pulse, direction or shape. Whether this results in lethargy is arguable. Some of the notes—for instance the triple stops in bars 3 and 4 (especially the first and last) but also the double stops in bar 7 and, quite surprisingly, the E tied quaver in the middle of the same bar—are harshly accented, in a way that does not make any musical sense to me even in the context of Kremer building a climax to the cadence point at the end of bar 8. Together with the increased bow-pressure, considerably swelling volume and broadening of tempo, in my view these accents create an “angry” or “tense” effect rather than lethargic depression. But when Kremer follows these attacks with soft and mellow, almost caressing sounds, as in bars 9-10, I can hear why the overall rating of emotional expression might be “melancholy,” “sadness” and “sorrow.” For me, however, an inescapable impression is the fragmented nature of this reading. It just does not flow enough to take me to a sublime, affective state, but keeps jarring and thus falling into self-aware / self-conscious bits, making the technical components all too obvious (Audio example 5.22).

5.22. Phrasing in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 1-8. Gidon Kremer 2005 © ECM. Duration: 1.27

To listen to this extract online follow this link:

In contrast, the relatively more even style of Perlman’s interpretation (apparently also rated in a follow-up study by Spitzer48), for instance, conveys a strong sense of unity (Audio example 5.23).

5.23. Phrasing in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 1-8. Itzhak Perlman. © EMI Classics. Duration: 1.37

To listen to this extract online follow this link:

I propose that the main contributing elements are his sustained style of bowing and fairly uniform dynamics. In Perlman’s playing the difference between flexible gestures and more measured moments is mild and well integrated. Unity and stability are achieved through continuity of phrasing and tone. The overall effect, however, is less personal. It speaks to me more as if somebody was making an appeal for some common cause; it sounds “public”; an objectified performance of “the music” rather than a personally invested communication.49 The additional affective element that I believe contributes importantly to such a perception is best grasped by stating what the performance is lacking. In psychological terms it is lacking what Huron might link to the “detailed-oriented thinking” of “depressive realism.” In terms of emotional expression it is lacking everything that characterizes the flip-side of the sustained style: it does not “breathe,” let alone “sob,” “yearn,” “plead” or “implore.” In Perlman’s rendering there is no sense of vulnerability, of conveying upheavals of the soul.

In light of Spitzer and Coutinho’s propositions, we can now briefly revisit Kuijken’s two versions of the opening movements from the G minor and A minor Sonatas. Perhaps the more measured and thus less flowing and more “atomized” later versions are closer to conveying “dogged” depression that feels hopeless and circular. The more freely flowing, improvisatory, melodic-harmonic goal orientated earlier versions, on the other hand, express heartfelt sadness that nevertheless has hope to heal. The way the performer conveys a sense of free musical fancy that seemingly obeys only the passions of his soul carries within the seeds of consolation and redemption—just like an uncontrolled, cathartic grieving-crying, saying out loud, has the potential of letting go, of accepting, of moving on.

Is all (or any) of this in the music, in the performance? Or do we, listeners project our emotions or personal predispositions onto what we hear? Surely all three, as countless empirical studies and philosophical arguments have shown.50 My point here is more to note: firstly, that compositions carry a range of affective potential; secondly, therefore subtle differences in performance can shift (“de-” and “re-territorialize”) the affective meaning communicated; and thirdly, that listeners may prefer performances that convey moods that more closely match their own psychological predispositions. When I find Kuijken’s later recording more “mechanistic,” perhaps I react against possible affective qualities that may evoke “depressive realism.” Whether the explanation of these subjective reactions sounds plausible is open to debate. But it is undeniable that aspects of these differences among the performances can be quantified. They certainly exist, how we interpret them is another matter.51

As noted above, how people identify the performance as well as compositional elements that contribute to overall aesthetic or emotional effect is something music psychologists have been investigating for several decades by now. I hope my discussion makes it apparent that I am not so much interested in the question whether basic emotions are conveyed or experienced when listening to music performance. Rather, I am trying to find a language to describe the affect—the pre-conscious, felt meaning—and to pinpoint the potential performative and bodily reasons for it. I have also collected listeners’ aesthetic response to selected movements performed by a range of violinists and in the next section I will draw on these as I comment on some of the most idiosyncratic versions available on record: those made by Huggett and Zehetmair.

5.4. Idiosyncratic Versions and Listeners’ Reactions

So far we have explored recent performance trends along the categories of MSP and HIP. We have also observed individual signatures and personal interpretative trajectories. At all times the discussion ended with the manifestation of problems regarding classifications, the establishing of boundaries for “territories.” The differences in degree (de-territorializing “molecular lines”) turned out to be as important as differences in kind (break away “lines of flight” causing transformation and moving towards re-territorialization). In this final section I explore how we might discuss essentially “rhizomic” interpretations that are so idiosyncratic that they seem to fit what Deleuze and Guattari may call the “nomad” because they are “open-ended.”52 The two performances that I believe fit this “category” are the recordings of Zehetmair (1983) and Huggett (1996).

Thomas Zehetmair

Zehetmair’s idiosyncratic style is hinted at by his tempo choices which often show more than 2 Standard Deviation (SD) from the average of over sixty studied recordings, especially in faster movements (D minor Corrente and Giga, B minor Corrente, E Major Gavotte en Rondeau).53 Other times they are more than 1 SD from the norm (first 3 movements of the G minor Sonata; first and last movements of the A minor Sonata, first and third movement of the C Major Sonata, B minor Tempo di Borea, D minor Sarabanda, E Major Loure, Menuet I and Bourée). His tempo is 4.16 SD from average in the A minor Andante! If one compares his tempos to only MSP violinists the results can be even more extreme than when the calculations are made within the pool of HIP and HIP-inspired violinists Table 5.3).

Table 5.3. Comparison of Zehetmair’s tempo SD scores in the E Major Partita when compared to the combined group of 34 MSP and 12 HIP-inspired violinists (columns A) as opposed to the group of 12 HIP and 12 HIP-inspired players (columns B).




Menuet I

Menuet II

















Zehetmair SD scores















His use of vibrato is also highly unusual. As noted in chapter four (Table 4.4), Zehetmair’s vibrato is often slow and wide. He obviously uses it to give emphasis to certain melodically or harmonically charged notes because few notes are vibrated but when they are it is very noticeable. However, his idiosyncratic style is primarily a result of the interaction of volatile phrasing, unpredictable bowing and accents, sudden changes in dynamics and extreme fluctuations of tempo. His delivery of notated ornamental groups is often extremely rapid and light with a gliding bow stroke. The effect is not particularly “decorative,” and at times reminds one of Gringolts’ recording (which is also highly idiosyncratic but has already received much attention to be interrogated here).

Given the earlier discussion of the G minor Adagio, I exemplify my claim regarding Zehetmair’s distinctively personal style by commenting on his performance of this movement. Overall he projects a tri-partite structure of the movement with bars 9-14 representing the contrasting middle section of a quasi ABA form.54 Such an outline of large-scale structure is quite unique among the recorded performances studied here. Most sound more like a through-composed movement.

Zehetmair starts off in the “melancholy” improvisatory style with gentle dynamics, legato bowing and a relaxed tempo. He takes the upward slide to the Eflat in bar 2 rather fast and light and not so much accents the Eflat but lets it ring out creating a slight gap (silence) before moving on with the last G-Bflat double stop quaver of the bar. This indeed lends the gesture a feel of “new beginning”; much more so than Kremer’s in 2005, lending support to Spitzer and Coutinho’s score analysis.55 The ornamental figure in bar 3 is shaped through tempo flexibility and bowing so that the first note of the fourth beat (Bflat semiquaver) gains an agogic stress by being held a little longer (score at Figure 5.7). From bar four onwards there are strong dynamic contrasts linked to dissonances and their resolution as well as particular melodic groups. There is a constant and sharp fluctuation of dynamics that is created through a combination of accents, swells, and reverse swells, faster and slower bow-strokes. The effect is further fostered by rapid acceleration and deceleration that was already noticeable in bar 3. At the end of bar 5 all this suddenly comes to a halt; tempo slows, rhythm becomes measured, and dynamics drop to subito ppp (Audio example 5.24).

5.24. Phrasing in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 1-8. Thomas Zehetmair © Teldec. Duration: 1.10

To listen to this extract online follow this link:

The non-vibrato precisely controlled tone has an otherworldly, “from a distance” effect. This only lasts till the second beat of bar 6, however, at which point the more volatile style returns. He repeats this vibrato-less ppp effect at the turn of bars 18-19, where the musical content is similar to bb. 5-6. But first we get a passionate, impetuous outburst between bars 10 and 12 during which Zehetmair plays with urgency, pressing ahead with tempo, tone and dynamics. There is nothing melancholic about these bars, even sadness is forgotten. One particularly individual solution is the staccato, almost thrown bowing of the three demisemiquavers (G-Bflat-D) leading to the second beat of bar 10. This creates enormous energy and changes the affective landscape instantaneously (Audio example 5.25).

5.25. Phrasing in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 9-14. Thomas Zehetmair © Teldec. Duration: 0.39

To listen to this extract online follow this link:

But what might be the emotion communicated? What kinds of hermeneutical metaphors come to mind as I listen to it holistically? I began my description in the previous paragraph by noting that the performance starts off similarly to many other versions. So melancholy and sadness seem obvious and reasonable associations. But the sense of immanent volatility conveyed through constant shifting of tone, tempo and intensity together with a sense of brewing impulsiveness that eventually bursts to the fore in bar 10 undermine such a simplistic association. Is it then a soliloquy going through a variety of emotions? Or is it perhaps a dialogue between a desperate, passionate, sad but angry soul and a calming consoler? Imagining two voices in “ardent dialogue” makes sense to me as it explains the rapid shifts in intensity, dynamics and tempo—it feels more plausible, more in line with how the performance unfolds than hearing it, for instance, as the hysterics of a grieved protagonist.

Zehetmair’s approach to Bach’s Six Solos for Violin is certainly not MSP. His bow strokes, articulation, rhythmic projection may be inspired by HIP but he takes these HIP characteristics out of their “territory” and infuses them with subjectivity. In other words, he interprets HIP “rules” liberally. The shorter, more lifted bow strokes, the locally nuanced and “airy” articulation mix freely with longer legato lines, sustained and heavier bowing. Vibrato-less tone alternates with strongly vibrated, emotive timbres; rapidly delivered fast and homogeneous passages and movements alternate with closely articulated, highly differentiated ones that dance and bounce. The ostensibly “molar lines” that create the HIP or MSP sound mix together in such fine grain of multiplicities that something entirely differently is born, especially if one takes into account the set as a whole.

Whether this “nomad” recording is considered appealing or mannered depends on the listener’s preferences; I certainly was positively stunned by the power of its individuality when I first heard it around the turn of the millennium, more than fifteen years after it was recorded. However, when I collected subjective responses from participants with varied musical background to various renditions of selected movements of the Bach Solos,56 Zehetmair’s received the most conflicting comments ranging from “Lovely performance, beautiful phrases and line of the melody” to “Terrible performance. Wrong phrasing, not musical; the performer doesn’t understand the music at all.” Some considered it to be by a “self-indulgent, ill-informed performer” playing in a “very performer-centred” way. Comments also tapped into the idiosyncratic nature of Zehetmair’s reading: “[The] exaggerated phrasing, that is very extreme crescendos and decrescendos in very short spaces of time, made it come across as a less emotionally genuine performance.” And: “Quite flexible and spontaneous. It makes it interesting. A bit too fast though and rough at times.”

The unique blend of MSP and HIP in Zehetmair’s performance was also remarked upon: “Even though this performance wasn’t on a period instrument, I felt that the performer showed quite a good understanding of Baroque style and phrasing.” Another participant considered it to “almost capture [her own] concept of how this piece should sound. The only things lacking are the right kind of instrument, i.e. a baroque setup violin, and less vibrato and better ornaments. Affectually, it is pleasing. Intonation could be a little nicer too.” However, others felt that “While [it is] somewhat historically informed, it is not expressive i.e. does not try to exaggerate the affect of the piece, and this exaggeration is an integral part of baroque musical oration. Such plain performance is a hallmark of the modern style which emphasises faithfulness to the score, and to the mastery of the composer at the expense of the performer’s vital interpretative authority.”

Monica Huggett

Apart from Gringolts, the other violinist who tends to provide the most highly idiosyncratic readings of the works is Huggett. In many respects she is the opposite of Zehetmair. Not only is she a period specialist using historical apparatus but she also tends to play slowly while Zehetmair’s playing is usually fast, as discussed above (Table 5.3). When compared to the approximately sixty recordings in my collection issued since 1903, the slow tempos of Huggett’s readings are particularly noteworthy in the nominally faster movements, such as the finales for the Sonatas and the E Major Preludio, but also the fugues. She tends to play slower in the dance movements of the Partitas as well, but less markedly (Tables 5.4 and 5.5). It should also be noted that her tempos are even slower relative to the pool of HIP and HIP-inspired violinists. For instance in such a comparison her SD score for the E Major Preludio is -1.37, for the E Major Gavotte en Rondeau it is -2.15 and for the E Major Tempo di Borea it is -1.61.

Table 5.4. Standard Deviations of Monica Huggett’s tempo choices in Sonata movements compared to the average of approximately 60 recordings made since 1903. High negative SD values indicate considerably slower tempos.

Gm Fugue

Gm Presto

Am Fugue

Am Allegro

CM Fugue

CM Allegro assai









Table 5.5. Standard Deviations of Monica Huggett’s tempo choices in Partita movements compared to the average of approximately 60 recordings made since 1903. Negative SD values indicate slower than average.

Bm Allemanda

Bm Borea

Borea DL

Dm Allemanda

Dm Giga

EM Preludio

EM Gavotte

EM Tempo di Borea

SD, Huggett






- 0.87



I should come clean and admit that my “relationship” with the recording is ambivalent. I like it, essentially, but then I also find it frustrating. It is full of beautiful and original detail; relished dissonances, broken chords, highlighted harmonic or melodic moments, ornaments (written out or added), clear polyphony, gorgeous violin timbre recorded well for sonorous effect, and even vitality in certain sections. Yet there is also a sense of laboriousness, lack of flow, angular and erratic phrasing; as if all this magnificent detail and precision were a huge struggle (which I am sure is not)! Both these characteristics of her version are remarked upon by reviewers and also the participants in the above mentioned experiment.

Reviewers find that “her fussiness with phrasing often disrupts the flow of the music”;57 that her “constant stop-go approach undermines the rhythmic integrity of (many) movements”;58 and that “Her rhythmic flexibility (very marked in the Chaconne) may upset some traditionalists, but it gives her readings a thoughtfully spontaneous air.”59 In contrast, Stowell notes “her remarkable intonation, technical precision and tone quality” and considers “rhythmic flexibility” her “greatest interpretative asset.” In his view Huggett’s “shaping and articulation of phrases generally allow the music to unfold naturally and with a sense of spontaneity.” However, even he mentions that “some may […] dislike her occasional ‘stop-start’ approach.”60 In the accompanying booklet Huggett notes that “the bass must always be given extra attention so that the balance between harmony and melody is not tipped too far in melody’s favour.”61 According to Jacobson, this “explains her meticulous way of dwelling on the lower notes” in many passages but can lead to an entire movement being “repeatedly thrown off course by the apparent establishment of a whole new tempo unrelated to what went before.” He also finds it “odd” that she embellishes only in a few repeats. Jacobson’s overall opinion is the same as mine: it is “hard to arrive at a balanced critical conclusion.”62

Participants in the listening experiment liked the “serenity” of Huggett’s playing but also noted the fragmentation and sense of effort: “[the] piece seemed to be performed in chunks punctuated by the performer breathing! but a sense of time could still be established.” Or: “although [it is] regular, the performer allows breathing and small variations within the beat. It makes it expressive without being overly expressive”; and “[the] tempo felt more painful, as if it was more work to play the notes, and the harmonies.” One participant specifically noted the uniqueness of Huggett’s style: “The performer plays this piece with their own style,” while another conveyed her approval by stating, “flexible within phrases to allow more expression. I like that. Also, time for breaths between phrases.”

Given my earlier discussion of various performances of the G minor Adagio, it would make sense to choose this movement for a detailed discussion of Huggett’s unique style. Indeed, her playing of it is very different to all I have analysed so far. Yet describing it may not necessarily be worth an attempt. It is a hybrid between the detailed, articulated, improvisatory style and the more measured and fragmented versions (Audio example 5.26).

5.26.Phrasing in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 4-8, 13-15. Monica Huggett © Virgin Veritas. Duration: 1.24

To listen to this extract online follow this link:

In contrast, picking a fast movement for study should provide pertinent insights. In these her attention to articulation and the underlying harmony makes her readings especially characteristic and unique (just like in the E Major Gavotte en Rondeau, as discussed earlier; see Audio examples 5.4 and 5.5). So for an interrogation of her performing style I turn to the E Major Partita, yet again. I discuss Huggett’s performance of its opening Preludio movement in relation to many other versions because this helps to illuminate how idiosyncratic her playing is.

The E Major Preludio

This movement has an improvisatory character based on broken chords. The even semiquaver figuration has traditionally been interpreted in a sleek virtuoso style. Such clockwork-like perpetual motion (moto perpetuo) reading is also typical among the recordings studied here, most obviously those by Kremer, Ehnes, Hahn, St John, Brooks, Ibragimova, Kuijken, Lev, and Tetzlaff.

5.27. Virtuosic style in J. S. Bach, E Major Partita BWV 1006, Preludio, extract: bars 1-33. Lara St John © Ancalagon. Duration: 0.38

To listen to this extract online follow this link:

In his forthcoming exhaustive monograph on Jascha Heifetz’s performance style, Dario Sarlo provides an interesting discussion of recorded interpretations of the E Major Preludio.63 He examines the threshold of what may count as the “Italian virtuoso moto perpetuo” style (where rapid figuration is persistently maintained)64 compared to the “French improvised preluding style” which Sarlo labels “expressive.” To arbitrate performance style he examines tempo and tempo fluctuation, as can be stipulated from differences in calculated metronome estimates based on durations of larger sections. Through score analysis and a review of analytical literature, Sarlo identifies eight sections summarized in a table, reproduced here from his pre-publication manuscript as Table 5.6. He then compares the duration of each section in eleven recorded performances. By calculating the percent deviation from the overall average metronome value he is able to conclude whether the performance is closer to the virtuoso moto perpetuo style (i.e. less fluctuation) or the expressive improvisatory style (higher percentage deviation).

Table 5.6. Eight structural subdivisions of the E Major Preludio
according to Sarlo (forthcoming)




% of piece



Theme, bariolage








Theme, bariolage




Build-up to harmonic climax




Build-up to and from harmonic climax




Dominant progression




Final dominant






The eleven recorded performances investigated by Sarlo on this occasion include three of those studied here: Kremer 2005, Wallfisch and Huggett. Out of all eleven, he found the highest fluctuation in Huggett’s performance (9%). He listed Wallfisch’s fluctuation as 6.2% while Kremer’s as only 4.2%, one of the lowest and similar to the classic, stereo-typical moto perpetuo style exhibited by Sarasate in a recording from 1904.

Sarlo’s data provide neat empirical evidence for differences between a smooth, “clockwork-like” performance and one that has more agogic details and tempo fluctuations. He asserts that the former falls into the moto perpetuo style because of the fast delivery.66 But does moto perpetuo have to be fast? Can a performer “persistently maintain rapid figuration” while playing at a moderate tempo? Or how slow can it be before it transforms into something else? Put another way, how fast can be an “expressive” version before the tempo prohibits playing around with agogic accents and timing? Would a more holistic or comprehensive examination of performance characteristics substantially enrich the process of labelling? Would it refine our understanding of differences between performances or the range along the two ends of the spectrum from moto perpetuo (alias virtuosic) to “expressive”? To answer some of these questions I re-examine these three recordings and a few others that are good cases in point. But first a brief overview of my data set seems useful.

All versions since Luca’s recording in 1977 tend to project the underlying harmonic progressions to a greater or lesser degree, outlining large-scale structural units.67 Smaller units and more localised events are highlighted by agogic stresses, accents and inflections. These are introduced to various degrees by most players with Gringolts, Huggett, Schröder, and van Dael providing the least clockwork-like and most detailed readings. Huggett, Gringolts and Matthews employ considerable tempo modifications as well, for instance starting the bariolage section at bar 64 slowly and then speeding up, or playing the echo measures faster than the more independently articulated notes of the louder complementary bars. Huggett’s, van Dael’s, and especially Schröder’s version may sound somewhat laboured due to the slower tempo and uneven delivery that often outlines bar by bar grouping of notes. Gringolts’ faster tempo gives his interpretation a more flowing quality. Among the more virtuosic versions Podger and Matthews mark arrival points and new figurations by accenting or slightly stressing notes (e.g. every downbeat between bb. 79-98) whereas Wallfisch tends to accent most downbeats except in the bariolage sections (bb. 17-29; 67-79) and when the music seems to be moving in pairs of bars. Wallfisch performs the Preludio very fast resulting in a rather rough sound (Audio examples 5.28; see also Audio examples 5.29 and 5.30 discussed at Figures 5.8 and 5.9).

5.28. Expressive versus virtuosic “moto perpetuo” style in J. S. Bach, E Major Partita BWV 1006, Preludio, extracts: bars 1-17. Jaap Schröder © NAXOS; bars 29-63. Rachel Podger © Channel Classics; bars 101-138. Lucy van Dael © NAXOS. Duration: 2.44.

To listen to this extract online follow this link:

The marked dynamics (echo effects) are observed by all players but some make the contrast stronger (e.g. Kremer 1980 and Shumsky less obviously) while others tend to remain within a relative dynamic range of mf-f (e.g. Wallfisch, Barton Pine, Mintz, Poulet, etc.). The most homogenous dynamics are observed in Brooks,’ Kuijken’s, and Ehnes’ recordings. Podger seems to achieve piano by using shorter bow strokes. She marks slurred notes and louder sections by longer strokes.

It is important to further explore the violinists’ varied approach to dynamics. This contributes significantly to the impression of diverse clusters among these recordings. Kuijken, Huggett, Zehetmair, Kremer, Holloway, Podger, Faust, and St John use a smaller range of dynamics. St John’s performance is softer and lighter than the others while Kremer’s and Podger’s are relatively louder overall. In all these versions there are also longer periods with fairly consistent dynamics. Although terraced dynamics are also found, the contrast in these recordings is not necessarily strong. Wallfisch’s version differs because of her sharply accented f and p dynamics. In spite of the narrow range, these violinists all use crescendos and diminuendos as well, primarily to shape longer segments.

Kremer and Barton Pine create a strong decrescendo during the bariolage sections (to bars 29 and 79, respectively) while Zehetmair and Kuijken play these at consistent dynamics. Schmid creates a crescendo from b. 63 to 73 and then a decrescendo to b. 79. Zehetmair’s tempo in these sections is unsteady with slight acceleration. His tempo settles after the accented new figuration in bars 29 and 79. In contrast to the relatively uniform dynamics of Kuijken, Zehetmair, Kremer, Holloway, Podger, St John and Faust, Barton Pine utilizes fluctuating dynamics on a larger dynamic scale and more frequently. For instance she creates many crescendos and diminuendos between bars 79 and 109 (starting with a diminuendo to b. 79, then crescendo to 89, a renewed crescendo to 93 then diminuendo to 97 and a crescendo with rallentando to 109). Otherwise her tone and bowing is very similar to Kremer’s in both of her concert recordings and her tempo only slightly slower in 1999 than Kremer’s (the duration of the performances are 3:20 for Barton Pine versus 3:12 for Kremer).

The overview of general characteristics indicates that distinguishing between moto perpetuo and “expressive” styles cannot be limited to a study of tempo fluctuations. Close listening draws attention to three important components of the differing approaches: Firstly, whether the structural points identified by Sarlo are signified in some way, for instance by temporal accents or dynamics; secondly, whether the performer highlights other moments more obviously; and thirdly, how dynamics are used throughout. A further issue is the degree of tempo fluctuation but from a perceptual point of view. How perceptible is it? Are the reported percentages of deviation clearly audible? Does one hear Kremer’s tempos less fluctuating than Wallfisch’s, for instance? Is the difference statistically or perceptually significant?

As Table 5.7 shows, the performers selected for closer study (Wallfisch, Kremer 2005, Huggett, Zehetmair, Schmid, Holloway, St John, Mathews, Podger, Faust, both Kuijken, and both Barton Pine) rarely highlight in an obviously audible manner the eight structural points identified by Sarlo on the basis of score analytical literature.68 The most commonly highlighted ones are the beginning of sections two (b. 33), six (b.109), and eight (b. 130). These violinists much more typically accent new figurations, whether through dynamics or temporal elongation (e.g. bb. 7, 9, 29, 42, etc.). At times one feels that the agogic (temporal) accents also serve technical control, providing just a touch of extra time to adjust-alter bowing, position or secure pitch. In other words, the physicality of performance is very much in the fore during most versions of this movement.

Table 5.7. Summary of aural analysis of selected recordings of the E Major Preludio according to the 8 sections identified by Sarlo.

Violinist, duration, comment

b. 1-32

b. 33-58

b. 59-82

b. 83-89

b. 90-108

b. 109-22

b. 123-29

b. 130-38



Very angular; stops & starts with every figuration

Dynamics fairly consistent

Decrescendo from bb.17 to 28

New start in b.29

Big Rit. to b.43;

Bowing is varied (shorter/longer)

Certain bars accelerate

Bariolage uneven dynamically as well

b.79: accelerates to b.82 (with crescendo-decres.)

bb.83-89 accelerates (with crescendo-decres.)

b.90: new start;

Tenuto on second beats in bb. 102-3

Agogic and p

Scale up in b.117 accelerates, then tenuto and rallentando

Last beat in b.129 has Rit..

Trills in bb.134, 135

b.136: in tempo

b.138: slight Rit..



Dynamic accents to mark new figurations or downbeats

Dynamics as score;

New start in b.29

b.33 marked;

Stresses also in bb. 34, 35, 42, etc.

Some fast accents (e.g. b.63); terraced dynamics; bariolage starts f; b.79: p

b.83 marked;

slight pause before moving to b. 90

Every downbeat marked (bb.90, 91, 92);

Slight pause for b.107

Slightly slower from b.109;

Most flexible tempo in this section

Not marked

Slight tenuto on high E (b.130;

Fast pick-up of tempo)

Kremer 2005


Considerable decrescendo during bariolage

Crescendo bb.29-32

Crescendo bb. 33-39;

Accents on second beats of bb. 39-40;

Score dynamics

Decrescendo during bariolage (to b.79)

Accented notes (each beat) Crescendo: bb.79-94

Crescendo: bb.79-94;

Decrescendo: bb.94-7

Accented b.108, b.119

Slows down to b.129

Broadens a little to end

Barton Pine 1999


Similar to Kremer (tone, bowing) and her own 2007.

Decrescendo to b.29; crescendo to b.33

New figurations articulated

Decrescendo to b.79 followed by crescendo

Crescendo to b.93

Decrescendo to b.97;

Crescendo to 101

Slows to end of b.108

Restart in b.109 and again in b.119

Slight Rit. to bb.129-130

Trill in bb.134, 135

Barton Pine 2007


Similar to 1999, more relaxed

Decrescendo to b.29;

Crescendo to b. 42;

Crescendo to b.42; agogic stress in b.43

Fluctuating tempo

Score dynamics + constant cresc-decresc.

Decrescendo to b.79

Crescendo to b.89

Crescendo to 93; decrescendo to b.97;

Crescendo to 109

Slowing to (with crescendo) 109

Crescendo from b.122 to b.129; Ritenuto over bb.129-30

Ritenuto over bb.129-30

Trill in b.135

Table 5.7, cont. Summary of aural analysis of selected recordings of the E Major Preludio according to the 8 sections identified by Sarlo.

Violinist, duration, comment

b. 1-32

b. 33-58

b. 59-82

b. 83-89

b. 90-108

b. 109-22

b. 123-29

b. 130-38

Kuijken ‘83


Articulated; fluctuating timing/tempo; score dynamics;

Smoother from b.16;

Shorter bow for p bars;

b.29 marked

Consistent to b.43;

Score dynamics;

b.55: f

Score dynamics to b.67 then consistent;

Crescendo starts in b.79

Accent b.90;

Slows to b.94; consistent dynamics;

Light accents in b.103-4

Consistent playing

Slight Rit. to b.134 from b.131.

Trill in bb. 134 & 135

Kuijken 2001


Homogenous, consistent dynamics, few accents, even bowing

bb.1-4: short bows

score dynamics (narrow range); bariolage even w. consistent dynamics

Very few accents (e.g. bb.39-40), little contrast between f and p

Score dynamics (bb.61-66);

Bariolage consistent;

New start in b.79

Perhaps the most moto perpetuo – “persistently maintained,” even though it is NOT fast, but rather undifferentiated.

Slight Rit. to b.134

Trill in bb.134 & 135



Relaxed, fairly consistent

Staccato eighths; score dynamics; slight accelerando at b.17; slight diminuendo to b.30

Softer, short bow; b.43 terraced dynamics; strong accents in bb.51-2

Not marked.

cresc. to b.63; decresc. bb.73-8; cresc. from b.79

Smooth dynamics; few accents

Few accents

p start; Terraced dynamics

Descending lines p; ascending lines f

Slight tenuto in b.129;

Rit. to b.134




p bits slightly faster;

Bariolage accelerates at beginning

b. 29: new start

Agogics in bb.39-40

Fluctuates a lot but smoothly;

Decrescendo bb. 67-78

Fairly consistent dynamics

Accelerates slightly at 90;

Tenuto agogic in b. 102

Agogic accent in b.109

Slight tenuto for 130;

Rit. in bb.133 to 134

Trill in b.135



Score dynamics

Accelerando at b.17;

Steady tempo, consistent dynamics from b.29

Accent in b.39; score dynamics from b.45

Accelerando from 67 (uneven bariolage);

Accented b.79 then crescendo

Crescendo to b.93

Crescendo to b.93;

decresc. to b.101

then cresc. b.102 (ff)

Decrescendo to b.109

More legato bowing; fluctuating tempo; many short crescendos

Broader b.123;

b.129 not really marked

Trill in b.135



Agogics in bb.3, 5

Bariolage has slight decrescendo

Accelerando from b. 29

Accelerando from b.33

Agogic in b.48 each f and p also in bb.55, 57

Very gradual decrescendo in bariolage (67-78)

Accelerando bb.79-82

Accelerando bb.83-7

Agogic in bb.90, 91, 92

Slight agogic in bb.102, 103 (on second beats)

Agogic in b.130. Improvised Cadenza in bb.134-5



f and p are not accented, dynamic change almost gradual;

bariolage has slight crescendo

no marking of b.29

Smoothly linked b.33; light stress on low notes (bb.45, 46, etc.)

Terraced dynamics are not accented;

Stressing 4 notes of down-beats (bb.55, 56)

Bariolage is smooth, fairly consistent dynamics

Slightly stressed b.83; slightly stressed 2nd sixteenth note in bb.87-8

Slight agogic in bb.90, 94

Bottom notes staccato

Slight accent on repeated slurred notes (bb.102-3 second beat)

Not marked

Tenuto high G (b.118) then p

Accented low notes (bb.119-21)

Very slight stress on first 2 semiquavers (b.123)

Slight Rit. at end of b.129

Slight tenuto on E in b.130

Trill in b.135

St John


Rapid blur of sixteenth notes beyond first note in bar;

Bariolage rushes ahead and has crescendo-diminuendo to b.29

Most 1st notes in bb. 43-55 stressed

f in b.55 stressed

second beat of b.59 accented;

Terraced then consistent dynamics; diminuendo bb.74-9

Soft until b.91

Crescendo bb.92-4;

Broadening bb.94-6

Softer from b.97;

Slows from third beat b. 107 (accent on third beats bb.107-8)

Rit. to b.109; tempo little slower thereafter;

Fluctuating dynamics bb.113-4, then p;

Tempo fluctuates most from here on

Broadens a little to downbeat (4note groups) bb.123-4

Tenuto in b.130

Rit. to b.134

Trill in b.135

Diminuendo in b.138

pp end



Small dynamics contrast between f and p

Slight tenuto on E (b. 9)

Tempo little unsteady during bariolage, consistent dynamics

p from b.29

Smooth, few accents (e.g. second beat bb.39-41; low notes in bb.43-51 esp. in f measures;

Smooth contrast of terraced dynamics

Small range terraced dynamics;

bb.67-78: consistent dynamics

Slight Rit. to b.79

Very slightly stressed down beats throughout

Very slightly stressed down beats throughout;

Slight Rit. to b.109

Slightly stressed b.109;

bb.119-22: Bass notes stressed

Light Rall. into and Rit. on E in b.130

Consistent dynamics

Trill in b.135

Among violinists using dynamic accents, Wallfisch stands out because of the frequency and strength of her accenting. She uses bowing and dynamics to strongly accent most new figurations and many downbeats, for instance in bb. 33-35 and 90-92. At the speed of her playing these seem to enhance virtuosity and do not impact negatively on the “persistent maintaining of rapid figuration.” Kremer accents fewer new figurations and less sharply. However, he, and to a lesser extent Holloway and also Kuijken, among many others, accents the second beats in bars 39-41, creating a momentary shift in the metrical pattern. The other performer who uses primarily dynamic accents is St John. Although Kuijken’s 1981 recording is fairly detailed with clearly articulated groups, agogic stresses, and terraced dynamics, he hardly plays any accents in his second version from 2001. This version has more consistent dynamics and is also twenty seconds slower than the earlier one! Strong temporal (agogic) accents are more common in the recordings of Holloway and Matthews as well as Huggett (Audio examples 5.29 and 5.30 discussed at Figure 5.8 and Figure 5.9).

Another noteworthy difference in relation to the use of accents, whether dynamic or temporal, lies in their effect or function. Wallfisch accents single notes, rapidly and sharply. They are “stabbed.” Podger and Faust, on the other hand briefly lean on groups of notes, not just emphasizing them but shaping and integrating them into the flow of the music. This adds ebb and flow and thus perhaps makes the performance less strictly moto perpetuo but still squarely in the virtuoso style. This shaping-grouping of two to four notes at the beginning of certain bars has a different effect to both the short and sharp dynamic accent and the longer temporal stress on a single note. The former rarely impacts on perceived tempo because it is simply a fast stab. The latter arrests the flow of the music much more than when the temporal elongation involves the entire metric unit (here this means four semiquavers) because of the distribution of extra time. When two to four notes are elongated and thus grouped (as opposed to a single “bass note”), this functions as a shaping mechanism; and shape assists flow as it creates a sense of direction.

Although I would categorize both Kremer and Barton Pine’s performances as in moto perpetuo style, Barton Pine’s might impress more like “virtuosic” because of the dynamic variety. Kremer’s is more etude-like because of its homogeneity. St John’s performance is also unquestionably virtuosic, and not just because of the very fast tempo. Her light bowing, rapid accents and fluctuating dynamics within a narrow (and fairly soft) range all contribute to this impression. Her tempo tends to rush ahead somewhat in the bariolage sections but otherwise it sounds steady. It noticeably fluctuates only from bars 109 onwards, in preparation for the climax. The rapidly delivered, diminuendo final bars confirm the virtuoso approach.

As for the other versions, practically all of them are shaped as moto perpetuo, except Huggett’s and perhaps Schröder’s, Matthews’ and Holloway’s. Why? Because they all tend to play in a smooth style using dynamics and dynamic accents to highlight moments rather than temporal (agogic) accents and locally nuanced tempo fluctuations as Huggett does. Even Holloway‘s tempo fluctuations are more smoothly executed and over longer periods of time. Schröder’s and Matthews’s versions are closer to being deemed “expressive” rather than moto perpetuo. He plays rather slowly with fairly frequent accenting and detailed articulation (cf. first item in Audio example 5.28). She also creates stresses frequently (e.g. bars 3, 5, 7), and always with temporal elongation. However, otherwise the tempo in Matthews’ recording is fairly fast and steady except for two sections: She accelerates from bar 29-32 and then from 33-36 and also at the analogous bars of 79-82 and 83-87. Even so, there are extended periods, notably the two bariolage sections, where the music moves smoothly and evenly, with consistent dynamics, tempo, and bowing (cf. first item in Audio examples 5.29 and 5.30, respectively).

Huggett, on the other hand, constantly stops and restarts the flow, articulating figurations bar by bar or as seem to fit her reading of harmonic and melodic motions. There are no sections longer than two to four bars that might have any fluency and consistency. Apart from tempo fluctuation and agogic timing she also varies her bowing much more than others. So even when the tempo remains reasonably steady her change to more staccato bowing, for instance, interferes with and counters the “persistent maintaining of rapid figuration.” In sum, the degree and extent of her tempo fluctuation and the frequency of stopping and accelerating make her performance individual and unusual to the extreme (cf. second item in Audio examples 5.29 and 5.30).

These aural impressions are not easily shown through measurements. When bars and beats are marked up the results indicate greater fluctuation than what my perceptual “threshold” would notice. This is particularly true when beat-level tempo is visualized (see Wallfisch or Kuijken in Figures 5.8b and 5.9b). However, the listening mind automatically adjusts minor differences and fluctuations—it receives the performance as a Gestalt with no regards for such micro variations. In Figures 5.8 and 5.9 two excerpts are illustrated in four of the versions studied: the opening bars of the movement (Figure 5.8) and a longer section between bars 29 and 63 that includes several agogic accents as well as dynamic contrasts (Figure 5.9). The smoother tempo curves of Kuijken and Wallfisch are quite obvious when bar-level tempo fluctuation is graphed. Wallfisch’s strong accents are seen at beat level but her tempo graph is reasonably smooth when bar-level fluctuations are presented. Although the beat-level mapping shows constant fluctuation for Kuijken and Wallfisch, even here their versions are smoother than Matthews’ or Huggett’s, confirming the moto perpetuo style (Figures 5.8b and 5.9b). The differences between Matthews’ and Huggett’s “expressive” style also show up with Huggett’s graph demonstrating more frequent and deeper fluctuations both at bar and beat levels.

Figure 5.8a. E Major Preludio, bars 1-12. Bar level tempo fluctuation in four recordings (Matthews, Huggett, Kuijken 2001 and Wallfisch).

Figure 5.8b. E Major Preludio, bars 1-12. Beat durations in 4 recordings
(Matthews, Huggett, Kuijken 2001 and Wallfisch).

5.29. Expressive versus virtuosic “moto perpetuo” style in J. S. Bach, E Major Partita BWV 1006, Preludio, extract: bars 1-12. Four versions: Ingrid Matthews © Centaur, Monica Huggett © Virgin Veritas, Sigiswald Kuijken 2001 © Deutsche Harmonia Mundi, Elizabeth Wallfisch © Hyperion. Duration: 1.20.

To listen to this extract online follow this link:

Figure 5.9a. E Major Preludio, bars 29-63. Bar level tempo fluctuation in 4 recordings (Huggett, Matthews, Kuijken 2001 and Wallfisch).

Figure 5.9b. E Major Preludio, bars 29-63. Beat durations in 4 recordings (Huggett, Matthews, Kuijken 2001 and Wallfisch).

5.30. Expressive versus virtuosic “moto perpetuo” style in J. S. Bach, E Major Partita BWV 1006, Preludio, extract: bars 29-62. Four versions: Ingrid Matthews © Centaur, Monica Huggett © Virgin Veritas, Sigiswald Kuijken 2001 © Deutsche Harmonia Mundi, Elizabeth Wallfisch © Hyperion. Duration: 3.53.

To listen to this extract online follow this link:

Clearly, Sarlo is right when he asserts “a link between playing the piece more slowly and playing it less persistently.” The two least persistent versions in the above sample are also the slowest (Huggett 101 beat per minute; Matthews 102 bpm).69 However, quite a few of other slower versions deliver the figuration just as “persistently” and regularly as the fastest versions (St John 142 bpm; Wallfisch 140 bpm), perhaps even more so. Figures 5.8 and 5.9 show that Kuijken’s slowish 2001 version (109 bpm) has a fairly even tempo, generally steadier than Wallfisch’s.70 The fast speed of St John’s interpretation impresses with its whirl-like virtuosity, partly because of its hushed, light tone quality that makes it sound even faster. The stabbed accents and rough tone in Wallfisch’s version create a performance that sounds like being on edge, “just made it.” In contrast, the relaxed, hardly differentiated second version of Kuijken, the more even tone and consistent dynamics of Kremer or Schmid (both players’ tempo is around 129 bpm) make these versions more than comply with the idea of “persistent delivery.” Kuijken’s 2001 recording is so homogenous and steady that it easily takes the palm for being “clockwork-like” moto perpetuo, even if not that virtuosic.

Although I set out to show how idiosyncratic Huggett’s interpretation is, I ended up discussing other versions more. Individuality is often more obvious against the relief of the common. By commenting on similarities and differences within that common backdrop Huggett’s radically different way of playing emerges. Not only that, but we can also see the range of diversity within even such straightforward musical character as the moto perpetuo. And, if that would not be fascinating enough, the complexity of music performance also manifests, yet again. The empirical evidence measurement of tempo fluctuation provided for the distinguishing of performance styles in the E Major Preludio (between moto perpetuo and “expressive”) turned out to be but the tip of the iceberg. To tease out the constituent performance elements contributing to a perception of “rapid figuration [being] persistently maintained” we had to examine delivery of accents (dynamic or temporal), use of dynamics and bowing, as well as tempo choice and relate them to each-other. This helped to see how perception of tempo stability and overall effect of persistency are formed.

5.5. Conclusions

In this chapter I recaptured differences between the broad and increasingly loosened MSP and HIP categories primarily through close analysis of four movements from the E Major Partita, multiple recordings of the same violinists, and an examination of affect and musical “meaning” in the D minor Giga, A minor Grave and G minor Adagio. The final section focused on two sets of idiosyncratic versions, one each of nominally MSP or HIP-inspired (Zehetmair) and HIP (Huggett) . Here I drew upon the opinions of general and musically trained listeners as well to support my view of these recordings being highly individual. Throughout the chapter my aim was to engage with the performances at a “holistic” level and to show the multitude of interactions at play. At times this enabled me to point out the usefulness of Deleuzian thinking and terminology when attempting to explain differences in kind and degree or to distinguish between kind and degree.

One of the main conclusions of this chapter stems from the comparisons of subsequent recordings made by the same violinists. The findings refine Daniel Leech-Wilkinson’s71 and Eitan Ornoy’s72 observations that artists tend to develop their approach to compositions early in their career; subsequent recordings by the same musician take the pieces further in the same direction. Although by and large this was found to be true in the current data set as well, complicating factors have also come to light. Slight differences were noted between the earlier and later recordings of Tetzlaff who increased his vibrato both in terms of frequency and prominence and created more extreme dynamic contrasts in places where the earlier recording displayed only modest contrast. However, some of these differences, although ostensibly representing the same idea (e.g. contrast), actually led the overall character of the playing towards different aesthetics and affect. The difficulty of unpacking cause and effect or the hierarchy of similarity and difference became transparent, questioning the feasibility of the task.

More significant differences were noted in Kremer’s two versions. The 2005 release was found to be rather wayward and uneven in tempo and more forcefully articulated, perhaps as a somewhat individualized understanding of HIP principles developed through collaborations with Harnoncourt and others in performances of post-baroque repertoire. Kremer radically decreased the use of vibrato, chose more extreme tempos and dynamics and delivered more powerful accents. Nevertheless basic interpretative choices remained intact even in his two recordings, as exemplified in discussion of the A minor Grave and Fuga, as well as the E Major Menuet II, for instance.73 To recall Kremer himself again, as cited earlier, “I am still the same Gidon Kremer; with a different violin, in a different church but with the same music.”

The only violinist who has radically changed her approach to performing Bach’s Six Sonatas and Partitas for Solo Violin is Viktoria Mullova. In her case we witness a total transformation from MSP to “beyond” HIP, or what I would currently regard the ideal (HIP) Bach performance—now, at the beginning of the second decade of the new millennium.

Although it could be argued that the music sets the boundaries of possible expressive gestures and phrasing as well as affective communication, the findings of this chapter provided further confirmation of the results of previous chapters: the interpretative differences among players show a broader range than what one finds within the two versions of any one violinist (except for Mullova in the current data set). Importantly, this chapter demonstrated that differences within MSP and within HIP versions can be considerable. This is particularly noteworthy in relation to MSP which has been much criticised for lacking individuality. Close examination underscored my position that generalizations are not particularly useful. They hide more than what they reveal. Furthermore, what might be typical in one movement of a given recording might not be true for another. Therefore I compared specific instances (e.g. D minor Giga, E Major Preludio, Loure, Gavotte and Menuet) and examined the level of uniformity in these particular cases. Although I offered some cautiously formulated general conclusions about particular violinists’ performance styles, more often than not the results prompted questions and highlighted the problem of dealing with complex dynamical systems.

Performance is non-linear and complex, rather than complicated. Its analysis does not deal with simple steps but the simultaneous irruption of many variables. I aimed to show that idiosyncratic performances are extreme examples of complexity because they are “unpredictable” and full of deterritorializing “lines of flight.” As Latour notes, “The more disorderly the message, the higher its information content […] [and] the less able the receiver is to predict” what will happen next. “A message high in information is one low in predictable structure, and therefore high in ‘entropy.’”74 However, there is tension between predictability and information content: unpredictability may eventually become randomness. When it does, the listener may feel disengaged; the music, the performance does not make sense anymore, it may seem mannered, tasteless, out of style, may be lacking flow, coherence. Some of the comments from the listening study imply that the unique styles of Zehetmair and Huggett are verging on this thin line of tension between “meaning” and idiosyncrasy.

The structure of any system must have some meaning; it “must somehow ‘represent’ the information important to its existence.” In other words a performance has to make musical sense whether in terms of a particular convention such as HIP or MSP or in terms of compositional style, or musical character. And because the “notion of ‘distributed representation’ [means that] the elements of the system have no representational meaning by themselves, but only in terms of patterns of relationships with many other elements”75 (i.e. patterns of relationships involving bowing, articulation, tone, tempo, timing, dynamics and so on), these unique versions often fall outside easy categorizations. In Huggett’s and Zehetmair’s recordings the “patterns of relationships” among the performance elements are porous, malleable, unpredictable, and at times seemingly incongruent. Their rich information content makes them particularly interesting to study and extremely challenging to “evaluate.”

This chapter’s more holistic explorations of recorded performances increasingly called upon subjective reflections and the contemplation of affective response. Affect is a nameless sensation that precedes cognition and recognized emotion. It is a bodily reaction; a felt transition from one state to another. It can be argued that aesthetics may be thought of as this capacity of sensing affectively. Since music performance unfolds in sound and penetrates the body in its entirety at once through the auditory system, we experience it holistically and affectively. In contrast to language and the written world music is first “comprehended” through the body, empathetically. Cognition, analytical dissecting and meaning-making occur only afterwards. Explaining the parts, however, rarely leads to a full understanding of the whole, underscoring the importance of allowing for the use of metaphoric language in an attempt to convey the felt experience. Musical complexity remains resistant to meaningful theoretical formulations or models. This is good because theories incline towards closure and thus stifle creativity; they are “the gravestones of musical invention.”76 If we want music performance to thrive, to be relevant and meaningful for ever new generations of listeners and musicians alike, we should not strait-jacket performers into theoretical constructs of normative performance rules and conventions. Nor should we deny or be silent about the listeners’ (including analysts’) holistic-affective experience, even if it is hard to find scholarly language for it.

1 Gilles Deleuze and Felix Guattari, A Thousand Plateaus: Capitalism and Schizophrenia (London: Bloomsbury Academic, 2013).

2 As I state in the conclusions of this chapter, for me “affect” is a nameless sensation that precedes cognition and recognized emotion. It is a bodily reaction; a felt transition from one state to another.

3 Jaap Schröder, Bach’s Solo Violin Works: A Performer’s Guide (London: Yale University Press, 2007), p. 171; Robin Stowell, The Early Violin and Viola (Cambridge: Cambridge University Press, 2001), p. 119; Ledbetter, Unaccompanied Bach: Performing the Solo Works (New Haven: Yale University Press, 2009), p. 170.

4 Stowell, The Early Violin, p. 119.

5 Schröder, Bach’s Solo Violin Works, p. 171.

6 Ledbetter, Unaccompanied Bach, citing Wendy Hilton, Dance and Music of Court and Theatre (Stuyvesant, NY: Pendragon Press, 1997), pp. 407, 437.

7 Schröder, Bach’s Solo Violin Works, p. 172.

8 Stowell, The Early Violin, p. 119.

9 In an unpublished interview with Daniel Bangert, John Holloway explains why the Loure is a “very good example of needing to put information together in order to arrive at a credible interpretation.” He states that violinists were “brought up with the idea that because the E Major Partita is how it is and that a piece with a lot of movements has to have a slow movement somewhere, [so] the Loure had to be the slow movement.” This influenced bowing and rhythmic delivery fostering a legato and literalist rendering. In contrast, familiarity with French baroque dances and Muffat’s teaching on bowing promotes a completely different approach. There are quotations from the interview and further discussion of Holloway’s (and other’s) view on bowing and interpretative approaches in Daniel Bangert, ‘Doing without thinking: processes of musical decision-making in period instrument performance’ (PhD Thesis, The University of New South Wales, 2012). For comments on the Loure see especially pp. 58-59. Holloway’s performance of bars 1-8 of the Loure can be heard in Audio example 4.12.

10 John H Planer discusses definitions of sentimentality in relation to musical performance in ‘Sentimentality in the Performance of Absolute Music: Pablo Casals’s Performance of Saraband from Johann Sebastian Bach’s Suite No. 2 in D Minor for Unaccompanied Cello, S. 1008,’ The Musical Quarterly, 73/2 (1989), 212-248. He mentions “emotional exaggeration” and refers to M. H. Abrams (among others) who defines sentimentality as “an excess of emotion, an overindulgence in the ‘tender’ emotions of pathos and sympathy.” However, Planer notes Abrams’ warning, that “excess or overindulgence is relative to the author and period” and therefore it is better to define it “by the use of clichés and commonplaces to express feeling.” Planer, ‘Sentimentality,’ p. 214 referring to M. H. Abrams, ‘Sentimentalism’ in A Glossary of Literary Terms, 4th edn (New York: Holt, Rinehart and Winston, 1981), p. 175.

11 Term coined by Taruskin in 1988. See Taruskin, Text and Act: Essays on Music and Performance (New York: Oxford University Press, 1995), p. 99.

12 Richard Hudson, Stolen time: The History of Tempo Rubato (Oxford: Clarendon Press, 1994).

13 Similar strategy is observed in the third episode but here the fluctuation of tempo is also reflected in subtly rising and falling dynamics, especially from b. 53 onwards. The first half of each bar is slightly slower while the second half is faster and crescendo, culminating in the emphasized (and held back) high notes of b. 59. In the other episodes the tempo fluctuation is less pattern-like but similarly enlisted to highlight the small motivic cells that make up sub-phrases and decorate harmonic progressions.

14 These terms are introduced by Deleuze and Guattari (A Thousand Plateaus) and I have explained them in chapter two.

15 Dario Sarlo provides contrary examples when he argues that both Nathan Milstein and Joseph Szigeti have mellowed their virtuosic “moto perpetuo” approach to the E Major Preludio by playing it slower in later years. He also cites Szigeti confirming his changed aesthetic idea about the piece (The Performance Style of Jascha Heifetz (Farnham: Ashgate, forthcoming), chapter nine)). One might also think of Nikolaus Harnoncourt’s two recordings of Bach’s Mass in B minor; the 1967 version heralding the new HIP style while the 1986 version turning the clock back, or rather, taking HIP in a different direction.

16 Gidon Kremer, Back to Bach, A film by Daniel Finkernagel and Alexander Lück. EuroArts (2055638), 2006, 0.39:23-0.39:50.

17 Kremer, Back to Bach, at 0.55:32.

18 An example of academia probing more deeply and practically (rather than theoretically) into the issues of interpretative styles and the contribution of performers is the Study Day organized by Daniel Leech-Wilkinson in May 2014 at Kings College London to explore “The Construction of Musical Performance Norms.” The call for papers invited talks on “the ways in which norms of musical performance are tacitly agreed, taught, and maintained and on how they might be challenged.” It suggested topics such as “teaching performance norms; ethics of performer obligation, the role of critics, agents, managers and producers in assuring performance norms; the examination and adjudication of performance style; performer anxiety in relation to perceived obligations and expectations; the implications of past recordings and beliefs about tradition; escaping, subverting or changing norms; the politics of performance; performance norms in relation to race, class and gender; the function of music as comfort or critique; music and utopia” and emphasized that “[c]ontributions need not be confined to western classical music.” See Performance-studies-Network ( Paraphrasing, reworking and “re-imagining” compositions are of course as old as our written records. These practices have again become popular especially among jazz musicians (e.g. Keith Jarrett, Jacques Loussier, Uri Caine, The Bad Plus) and some classical musicians (e.g. Red Priest, and Joe Chindamo, among many others). For a philosophical exposition and history of the “work” concept see Lydia Goehr, The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music, rev. edn (New York: Oxford University Press, 2007). A few years ago I have also conducted a study to examine the possibility and implications of performing Schumann’s Träumerei according to baroque performance conventions. See Dorottya Fabian, Emery Schubert, and Richard Pulley, ‘A Baroque Träumerei: The performance and perception of a violin rendition,’ Musicology Australia, 32/1 (2010), 25-42.

19 At 2pm she played the first two sonatas and first partita interspersed with a piece each by Baltzar and Pisendel. This was followed with Part 2 of the concert at 8pm when she delivered the remaining Bach Solos together with compositions by Westhoff and Biber. eNewsletter received from on 21 August 2014. See also Voices, Chicago Sun-Times blog 20 August, 2014, available at
[last accessed October 2015].

20 Robin Stowell, ‘Reviews: CDs—Bach Sonatas and Partitas for Solo Violin, Christian Tetzlaff (violin) Virgin VCD 45089 2,’ Strad, 106/1261 (May 1995), 541-542.

21 Duncan Druce, ‘Reviews: Bach 3 Sonatas and 3 Partitas BWV1001-BWV1006, Christian Tetzlaff vn, Hänssler Classic CD98 250 (130 minutes DDD),’ Gramophone, 85/1021 (August 2007), 71.

22 Herbert Glass, ‘A Reasoning Romantic—Profile: Christian Tetzlaff,’ Strad, 106/1259 (March 1995), 260-265.

23 Idid.

24 Jeremy Eichler, ‘String Theorist: Christian Tetzlaff Rethinks How a Violin Should Sound’ [Profiles], The New Yorker (27 August 2012), 34-39. All citations from p. 39. Tetzlaff also discusses his views of the Bach Solos in an interview with Edith Eisler, ‘Christian Tetzlaff from Bach to Bartók,’ Strings (February-March 1999), 50-56 (Here and elsewhere italics in original unless otherwise stated).

25 Apparently Tetzlaff believes the incorrect Italian grammar on the title page of Bach’s autograph score from 1720 (Sei Solo instead of Sei Soli, i.e. Six Solos) is actually a “spiritual double-entendre, since Sei Solo can also be rendered as ‘You are alone.’’’ (Eichler, ‘String Theorist,’ p. 39). This interpretation is also mentioned by Elizabeth Wallfisch in ‘Masterclass: Bach’s Solo Violin Sonata in G Minor,’ Strad (July 2007), 64-67. She highlights that Bach’s first wife died the same year this beautiful autograph was penned. The suggestion that Sei Solo might mean “you are alone” may have been inspired by Helga Thoene’s papers retold in the liner notes to Christoph Poppen and The Hilliard Ensemble’s Morimur compact disk as these emphasize the tragedy Bach experienced when, upon returning from Karlsbad to Cöthen in 1720 he found his wife of thirteen years, Maria Barbara dead and buried. Thoene shows that the manuscript paper used by Bach for the fair autograph copy of the Six Solos originates from near Karlsbad. She speculates that therefore it is likely Bach prepared the copy in the wake of his wife’s death. Helga Thoene, ‘Geheime Sprache—Verborgener Gesang in J. S. Bch’s “Sei Solo a Violino,”’ in Morimur CD ECM New Series 1765, 461895-2 (München: ECM Records, 2001), pp. 19-28 (p. 20). Thoene’s theory gained considerable following among violinists (while hardly any among Bach scholars), probably because of Poppen’s disk or because she herself is a violin professor at the University of Düsserldorf. I have heard Simone Standage speak (in an early music performance masterclass held in Sopron, Hungary around 2002) of Thoene’s suggestion that the D minor Ciaccona is a “tombeau for Maria Barbara” as if it was a historical fact. See also Helga Thoene, ‘Johann Sebastian Bach Ciaccona: Tanz oder Tombeau—Verborgene Sprache eines berühmten Werkes,’ Köthener Bach-Hefte, 6 (Köthen: Historisches Museum Köthen/Anhalt, 1994), 15-81; and Helga Thoene, “‘Ehre sey dir Gott gesungen”— Johann Sebastian Bach Die Violin-Sonata G-Moll BWV 1001, Der Verschlüsselte Lobgesang,’ Köthener Bach-Hefte, 7 (Köthen: Bach-Gedenkstätte und Historisches Museum Köthen/Anhalt, 1998), 1-113.

26 Joseph Magil, ‘Bach: Solo Violin Sonatas and Partitas [Christian Tetzlaff],’ American Record Guide, 70/4 (July/August 2007), 69-70 (p. 70).

27 Anthony Tommasini, ‘A Violin Virtuoso and Total Bach,’ New York Times, 28 April 2000, p. E4, available at

28 ‘Podiumdiskussion: Zur Situation der Aufführungspraxis Bachscher Werke 1978,’ in Bachforschung und Bachinterpretation heute: Wissenschaftler und Praktiker im Dialog, ed. by Reinhold Brinkmann (Kassel: Bärenreiter, 1981), pp. 185-204 (p. 196).

29 Jason Read, ‘Review Essay—The Full Body: Micro-Politics and Macro Entities,’ Deleuze Studies, 2/2 (2008), 220-228 (p. 227). I thank Ellen Hooper for this reference.

30 Read: ‘Review Essay,’ p. 221. Here I used Read’s spelling of territorialization and deterritorialization.

31 I am grateful to Ellen Hooper for our thought-provoking discussions of such approaches to performance analysis.

32 Joseph Magil does not agree. In his review he claims that “Viktoria Mullova conveys a bland studiousness. As usual, she plays like a student who has just learned a work and isn’t yet sure what to make of it. Her attitude toward period performance practice is that it is a set of rules to be followed rather than a key to unlocking any of the music’s mysteries.” One wonders what CD he listened to, but the publication clearly lists the Onyx album! See ‘Guide to Records—Bach: Violin Sonatas & Partitas; Pauset: “Kontrapartita,”’ American Record Guide, 72/5 (September 2009), 61.

33 Andrew Palmer, ‘Viktoria Mullova: The Individualist,’ in Violin Virtuosos, ed. by Mary VanClay and Stacey Lynn (San Anselmo: String Letter Publishing, 2000), pp. 47-57.

34 John Duarte, ‘Bach Partitas—No. 1 in B Minor, BWV 1002; No. 2 in D minor, BWV 1004; No. 3 in E, BWV 1006, Viktoria Mullova vn, Philips 434 075-2 PH (77 minutes DDD),’ Gramophone, 72/853 (June 1994), 80.

35 Concert presented on 12 March 2011 as part of the 2010/2011 Camerata Musica International Artists series in Cambridge (UK), available at

36 The Guardian, Monday 10 April 2000, available at

37 “Musicological listening” is a term used in Nicholas Cook, Music, Imagination and Culture (Oxford: Clarendon Press, 1992), p. 152. However, “close” or “focused listening” is more commonly used; e.g. Nicholas Cook, Beyond the Score: Music as Performance (New York: Oxford University Press, 2014); Daniel Leech-Wilkinson, The Changing Sound of Music: Approaches to Studying Recorded Musical Performance (London: CHARM, 2009), chapter eight, paragraphs 20-21,

38 Cook, Beyond the Score, p. 143.

39 Taruskin introduced the term “authentistic” as a descriptor of what, in my view, Dreyfus called “sewing-machine” style. See Taruskin, Text and Act (esp. pp. 99-143) and Lawrence Dreyfus, ‘Early Music Defended against its Devotees: A Theory of Historical Performance in the Twentieth Century,’ The Musical Quarterly, 69/3 (1983), 297-322.

40 Sonic Visualiser is an open access computer program developed at Queen Mary University for viewing and analysing the contents of music audio files available at or Chris Cannam, Christian Landone, and Mark Sandler, ‘Sonic Visualiser: An Open Source Application for Viewing, Analysing, and Annotating Music Audio Files,’ in Proceedings of the ACM Multimedia 2010 International Conference (Florence: October 2010), 1467-1468.

41 Michael Spitzer and Eduardo Coutinho, ‘The Effects of Expert Musical Training on the Perception of Emotions in Bach’s Sonata for Unaccompanied Violin No. 1 in G Minor (BWV 1001),’ Psychomusicology: Music, Mind, and Brain, 24/1 (2014), 35-57 (p. 36).

42 Daniel Leech-Wilkinson, ‘The Emotional Power of Musical Performance,’ in The Emotional Power of Music: Multidisciplinary Perspectives on Musical Arousal, Expression, and Social Control, ed. by Tom Cochrane, Bernardino Fantini, and Klaus R. Scherer (Oxford: Oxford University Press, 2013), pp. 41-54.

43 I have shown the impact of performance features on affective and aesthetic response in relation to recordings of the D minor Sarabanda; Dorottya Fabian and Emery Schubert, ‘Baroque Expressiveness and Stylishness in Three Recordings of the D minor Sarabanda for Solo Violin (BWV 1004) by J. S. Bach,’ Music Performance Research, 3 (2009), 36-55. See also fn. 51 below.

44 Spitzer and Coutinho, ‘The Effects of Expert Musical Training,’ p. 48, Figure 10.

45 In relation to the Adagio movement of the G minor Sonata Spitzer and Coutinho refer to David Huron, ‘A Comparison of Average Pitch Height and Interval Size in Major— and Minor-Key Themes: Evidence Consistent with Affect-Related Pitch Prosody,’ Empirical Musicology Review, 3 (2008), 59-63; Alf Gabrielsson and Erik Lindström, ‘The Role of Structure in the Musical Expression of Emotions,’ in Handbook of Music and Emotions: Theory, Research, Applications, ed. by Patrik Juslin and John Sloboda (Oxford: Oxford University Press, 2010), pp. 367-400; Patrik Juslin, ‘Emotional Communication in Music Performance: A Functionalist Perspective and Some Data,’ Music Perception, 14/4 (1997), 383-418; and Sarha Moore, ‘Interval Size and Affect: An Ethnomusicological Perspective,’ Empirical Musicology Review, 7/3-4 (2012), 138-143.

46 Spitzer and Coutinho, ‘The Effects of Expert Musical Training,’ p. 37. The internal citation refers to Raymond Monelle, The Sense of Music: Semiotic Essays (Princeton: Princeton University Press, 2000).

47 Spitzer and Coutinho, ‘The Effects of Expert Musical Training,’ 40. The internal citation refers to David Huron, ‘Why is Sad Music Pleasurable? A Possible Role for Prolactin,’ Musicae Scientiae, 15 (2011), 146-158.

48 Michael Spitzer, ‘Affektive Shapes and Shapings of Affect in Bach’s Sonata for Unaccompanied Violin No. 1 in G Minor (BWV 1001),’ in Music and Shape, ed. by Daniel Leech-Wilkinson and Helen Prior (New York: Oxford University Press, forthcoming). I had no opportunity to access this paper while completing this manuscript.

49 Jonas Kaufmann and Helmut Deutsch have recently discussed the issue of concentration and conscious control while performing highly emotional music. In this interview Kaufmann evokes “Karajan’s famous remark about ‘controlled ecstasy.’ Everyone, myself included, should have the impression that I am abandoning myself completely to the emotion that I’m depicting, but a final controlling authority ensures that I don’t damage my voice or become overexcited.” (See, ‘You can’t simply carry on as usual afterwards—Jonas Kaufmann and Helmut Deutsch in conversation with Thomas Voigt,’ Schubert: Winterreise, Jonas Kaufmann [tenor], Helmut Deutsch [piano]. Sony Classical 88883795652 (USA: Sony Classical Records, 2014), CD Booklet, pp. 7-12 (pp.11-12)). Perhaps in Perlman’s performance this “final controlling authority” is given too much room and nobody, perhaps not even the violinist, has the impression that he is abandoning himself to the expressive power of the music.

50 Handbook of Music and Emotions: Theory, Research, Applications, ed. by Patrik Juslin and John Sloboda (Oxford: Oxford University Press, 2010). Peter Kivy, Sound Sentiment: An Essay on the Musical Emotions (Philadelphia: Temple University Press, 1989).

51 I have investigated “emotion” or what I prefer to call “musical character” empirically in relation to thirty-three recordings of Variation 7 from Bach’s Goldberg Variations. This investigation revealed five different clusters of “affect” but the ninety-eight participants chose a variety of specific (synonymous) words available within the emergent five categories to describe the performances. This demonstrated considerable subjective differences within broad agreements. See Dorottya Fabian and Emery Schubert, ‘Musical Character and the Performance and Perception of Dotting, Articulation and Tempo in Recordings of Variation 7 of J.S. Bach’s Goldberg Variations (BWV 988),’ Musicae Scientiae, 12/2 (2008), 177-203.

52 Deleuze and Guattari contrast “representational thought” which is analogical and structured with “nomadic thought” that “moves freely in an element of exteriority. It does not repose on identity; it rides difference” (Brian Massumi, ‘Translator’s Foreword: Pleasures of Philosophy,’ in Deleuze and Guattari, A Thousand Plateaus, pp. x-xi). It is the open-endedness and freedom from categories (representation) that make me label Zehetmair and Huggett’s recordings of the Bach Solos “rhizomic” and “nomad.” Gringolts’ version might also fit such a description but as it has already received much attention I am not discussing it here any further.

53 As explained at fn. 5 in chapter four, Standard Deviation (SD) is defined as the average amount by which scores in a distribution differ from the mean. It shows how much variation there is from the mean. Generally three standard deviations account for 99.7% of the studied data. One SD accounts for about 68% of the data set while two SD about 95%. When SD is close to 0 this indicates that the data points are very close to the mean. In the current study negative values indicate a deviation slower than the mean score while positive values are faster than average.

54 This is so even though there are “in-between” cadential gestures in Zehetmair’s performance (e.g. bars 3-4, 13-14, and 20-21). Like Kuijken in 1981, his shaping of phrases follow harmonic and melodic goals with several overlaps or elisions of ending-beginnings that fall on metrically stronger points.

55 Spitzer and Coutinho, ‘The Effects of Expert Musical Training,’ p. 40.

56 These included the D minor Sarabanda, the A minor Andante and the E Major Loure. However, due to randomization not all participants heard all versions of all movements. Most of the cited comments refer to Zehetmair’s performance of the D minor Sarabanda.

57 Joseph Magil, ‘Bach: Solo Violin Sonatas and Partitas’ [Monica Huggett], American Record Guide, 61/4 (July 1998), 88-89.

58 Bernard Jacobson, ‘CD Review: Bach: Sonatas and Partitas for Solo Violin—Monica Huggett (vn) (period instrument), VIRGIN 7243 5 45205 2 5(152:33),’ Fanfare, 22/1 (September 1998), 116-117 (p. 116).

59 Lionel Salter, ‘Reviews: Bach 3 Sonatas and 3 Partitas, BWV10010-06, Monica Huggett vn, Virgin Classics CD 545205-2 (152 minutes),’ Gramophone, 75 (January 1998), 78.

60 Robin Stowell, ‘Reviews: CDs—J.S. Bach: Sonatas and Partitas for Solo Violin BWV 1001-6. Monica Huggett (violin) VIRGIN VERITAS 7243 5 45205 2 5,’ Strad, 109/1297 (May 1998), 539.

61 Monica Huggett, ‘A Performer’s View of Bach’s Sonatas and Partitas for Solo Violin,’ J. S. Bach: Sonatas and Partitas BWV 1001-1006. Virgin Veritas 7243 5-45205-2 5 (Holland: Virgin Classics, 1997), CD booklet, pp. 12-13 (p. 13).

62 Jacobson, ‘CD Review,’ 116-117.

63 Sarlo, Heifetz (forthcoming), chapters six-ten.

64 Sarlo uses the definition provided by Michael Tilmouth, ‘Moto perpetuo,’ in Grove Music Online. Oxford Music Online (Oxford—New York: Oxford University Press, 2007-2015), available at

65 Sarlo, Heifetz, Table 10.2 (as presented in the original MS Word manuscript in 2014, before typesetting and publishing).

66 To be fair, in discussing his results Sarlo refines the point when he states: “There is clearly a link between playing the piece more slowly and playing it less persistently. Based upon the evidence, it is not only the total durations of these recordings that differs with the virtuosic early 1900s recordings, but also the approach to persistent figuration. […] Unlike many of the other slower-paced period instrument performances, Wallfisch takes a different approach to the Prelude, since hers is the fifth fastest of all 136 recordings. It is therefore not simply in duration that Wallfisch diverges from the other period performances—the more persistent nature of the Wallfisch figuration, identified with a lower fluctuation percentage, suggests a closer link to the virtuosic and moto perpetuo approach of Sarasate […] than to the period performances of […] Huggett.” Sarlo, Heifetz (forthcoming), chapter ten.

67 Sarlo identifies eight sections: bb. 1-32, 33-58, 59-82, 83-89, 90-108, 109-122, 123-129, and 130-138 (cf. Table 5.6). As I will show performers do not necessarily highlight these boundaries but often chose other moments.

68 This discrepancy between analyses and performance illustrates how “irrelevant” music analysis literature can be for performance analysis (and for performers) once the attention is to describe rather than to prescribe what performers are (not should be) doing, even if the analysts might behave as if they had the “upper hand.” This of course has implications for the tertiary training of performers and also for practice-led research discussed in chapters two and six.

69 In my entire collection Schröder (99 bpm) and Beznosiuk (98 bmp) are the only two that are slower than Huggett and Matthews. The next slowest is Ehnes (104 bmp) and then Szeryng (105 bpm).

70 Because I calculated beat per minute tempos from duration it might be useful to provide the base durations I used when working with these Preludio recordings. Huggett: 4:04, Matthews: 4:04, St John: 2:55, Wallfisch: 2:58, and Kuijken 3:47.

71 Daniel Leech-Wilkinson, ‘Recordings and Histories of Performance Style’ in The Cambridge Companion to Recorded Music, ed. by Nicholas Cook et al. (Cambridge: Cambridge University Press, 2009), pp. 246-262.

72 Eitan Ornoy, ‘Recording Analysis of J. S. Bach’s G minor Adagio for Solo Violin (excerpt): A Case Study,’ JMM: Journal of Music and Meaning, 6 (Spring 2008), section 2, available at

73 In earlier sections of the book my comments on Kremer’s performances of other movements also noted many similarities between his two recordings.

74 Bruno Latour, Pandora’s Hope: Essays on the Reality of Science Studies (Cambridge, MA: Harvard University Press, 1999), p. 7.

75 Latour, Pandora’s Hope, p. 11.

76 Richard Toop, ‘Against a Theory of Music (New) Complexity,’ in Contemporary Music: Theoretical and Philosophical Perspectives, ed. by Max Paddison and Irène Deliège (Aldershot: Ashgate, 2010), pp. 89-98 (p. 97).