Desktop versionMobile version
OpenEdition Books

A Musicology of Performance

 | 
Dorottya Fabian

4. Analysis of Performance Features

Full text

1Having focused on overall trends and the inter-relationship between HIP and MSP characteristics in the previous chapter, here I provide a systematically detailed account of performance features. My aims are, however, the same as throughout the book: 1) to investigate the claim that performances have become more uniform and homogeneous both technically and stylistically; 2) to examine how different performance features interact to create particular interpretations and aesthetic constructs; 3) how these relate to time and place as well as musical sensibilities and knowledge; and 4) what all this tells us about musical performance. Ultimately I aim to provide empirical evidence for the complex nature of performing western classical music and a model for an integrated analytical approach.

2The recordings testify to a fascinating palette of interpretative possibilities. It is quite staggering to contemplate how Bach managed to compose such pieces that speak to us almost exactly 300 years later with such directness and wealth of potential that all violinists wish to perform and record them, one generation after another. In the previous chapter I have quoted several violinists of varied background and ilk who all discussed, explicitly or implicitly, the emotional pull and instrumental challenge of Bach’s Violin Solos. They mentioned the honesty, awe, curiosity, puzzle and bewilderment these works represent for them. So what is their answer? How do they solve the problems? How do they respond to Bach’s invitation? What conversations may we, listeners, witness between composer and performer 300 years apart? How is the age of the musician reflected in his or her dialogue with the music? Are youthful players drawn to different aspects of the music then older ones? Are cultural pre-conceptions more dominant than personal age and psychological-musical maturity? The spectrum offered by Bach from timeless contemplative music to period-bound genre pieces is wide open.

3Musicians and listeners find the seemingly endless ways of engaging with the Solos unquestionably rewarding. As a listener one often has a sense that quite a few violinists must also enjoy facing the technical challenges inherent in them. Otherwise they wouldn’t be returning to the pieces and performing and recording them over and over again. The sheer sonority, physicality, virtuosity of certain movements is so imminent and the performance of them so abundantly brilliant that one is reminded of musicians confiding off-record, “Yes, wasn’t that great, playing so fast? I can do it!” –– echoing Horowitz, who, when asked once why he had played so fast, responded simply: “Because I can.” The reward is not exclusively the musician’s. Listeners are enchanted as well; otherwise the market would have long relegated the works back to the pile of forgotten music or the practice studio. I for one certainly love listening to this music and find something beautiful, interesting, or novel in most versions and even those I don’t much enjoy can envelop me in Bach’s sound world in ways that lift me into another sphere. It is therefore an exciting challenge for me to explain how they do it and why I may prefer one over another when most are truly wonderful and of exquisite standards.

4I have organized my analysis along performance features, moving from the most factual towards the more subjective measures. First I look at tempo choices, vibrato and ornamentation. This is followed by the discussion of rhythm and timing which entails an analysis of playing dotted rhythms and the topic of inégalité as well as the expressive timing of notes, including tempo and rhythmic flexibility. Subsequently I discuss matters that are even harder to describe in words, such as bowing (including the playing of multiple stops), articulation, and phrasing. Wherever possible I compare performance choices to musicological opinion, as found in both historical sources and modern pronouncements. The points I make are supplemented by further observations presented in tables, graphs and transcriptions as well as descriptions of particular moments in recordings. The more detailed or specific analytical observations and descriptions appear in boxed texts shaded grey for ease of navigation.

5It is a daunting task to provide an honest account when analysing forty-odd recordings of Bach’s Six Sonatas and Partitas for Solo Violin. Each complete set entails about two hours of music. Given my position regarding the problems of the current state of music performance studies that I outlined in chapters one and two, my conclusions cannot be based on a case study. The main challenge is, therefore, to find a balance between sampling a judiciously appropriate cross-section of movements and violinists for each performance feature under investigation without getting lost in detail. Perhaps an eclectic approach will help: At times, like in the case of tempo choice and ornamentation, I provide comprehensive coverage with numerous tabulated summaries and transcriptions. Other times, like in the case of vibrato, I illustrate my claims through measurement of specific moments in a select few movements. In both cases I aim to highlight the potential for misinterpretations and misrepresentations; how the data, masquerading as objective, can nevertheless be manipulated to support whatever argument the researcher wishes to make.

6Certain performance features, like the performance of dotted rhythms or multiple stops, self-determine what excerpts I have to focus on as only certain movements feature such material. What is of special interest here is the obvious slippage that readers will notice: ostensibly about dotting or the delivery of multiple stops, the discussion will actually shift to various other performance features, namely articulation, dynamics, timing, bowing, and tempo (again)—illustrating the workings of a complex, non-linear system of interactions that give rise to musical character, the affective-aesthetic dimension of the performance. Overall, it is probably inevitable that certain movements will feature more frequently because they offer more for discussion. In particular, the E Major Partita somehow emerged as a focal point, but the two Sarabandes, the A minor Andante and the G minor Adagio and Fuga also receive much attention. At the same time, the famous Ciaccona is conspicuous for its absence. It deserves a separate study.

4.1. Tempo Choices

  • 1 The “breathless tempi of much early music” is an opinion generally expressed rather casually, as in (...)

7Right here at the start and in relation to an apparently straightforward matter, a difficulty has to be noted: it is near impossible to generalize tempo trends as each movement provides something different (Tables 4.1-4.3). Furthermore, and perhaps contrary to expectations, period specialists have slower averages than MSP players, except in the slow movements, allemandes and courantes. Therefore, the once commonly held view that performances, especially HIP versions, of baroque music have become faster as we progress through the twentieth century is questionable.1

8But let us stop for a minute and take a closer look. First, how should one group violinists into MSP and HIP categories if there is such confluence between approaches as discussed in the previous chapter? Readers are invited to consult Table 4.1 to see the results of grouping violinists simply on the basis of specialist / non-specialist. Table 4.2 takes into account a larger pool of players including recordings made since the beginning of the twentieth century. The top row presents the same simple subdivision (specialist / non-specialist) while the bottom row re-configures the grouping by adding the HIP-influenced players to the specialist group and contrasts them to the “hard-core” MSP cluster. Finally, in Table 4.3 I present the average tempo values of all recorded performances selected for the present study against the pre-1977 average.

  • 2 Dario Sarlo studies over 130 recordings of the E Major Preludio in his forthcoming book, The Perfor (...)

9In each Table the results are slightly or strikingly different! This is extremely important to note because it highlights how easy it is to draw incorrect conclusions. Even if one is interested in overall trends only (Table 4.3), caution is warranted: only 19 out of 28 movements (68%) show a slight increase in tempo. It is noteworthy that the Fugues, the E Major Preludio, as well as the D minor Allemanda and Giga have become slower. Moreover, the size of the pool of recordings examined also has to be kept in mind. Many more recordings may exist and although I believe I have examined a fairly exhaustive portion of the most readily available versions, additional versions could change the results reported here.2 Furthermore, the perceptual difference between one or two metronome marks is unlikely to be significant, especially since these values are averages based on overall tempo calculated from duration.

  • 3 Dorottya Fabian, Bach Performance Practice, 1945-1975: A Comprehensive Review of Sound Recordings a (...)

10However honestly one reports averages, such a presentation hides what I believe to be the case: that tempo is a personal thing. My investigations over the years indicate that there are musicians who like to play fast and there are those who prefer it slower. Alternatively, some tend to play fast movements really fast and slow movements quite languidly (e.g. Ibragimova in the current data set), while others prefer less extreme tempo choices regardless of movement type (most HIP violinists). This lesson can be drawn from the examination of pre-1980 recordings as well and seem to hold true in other repertoires too.3

  • 4 R-squared is a statistical measure of how close the data points are to the fitted regression line. (...)

Table 4.1. Summary of tempo trends 1977-2010 (For change to be noted R2 = >0.001)4

Table 4.1. Summary of tempo trends 1977-2010 (For change to be noted R2 = >0.001)4

Table 4.2. Average MSP and HIP tempos across all studied recordings made since 1903 (Joachim). Violinists who were added to HIP are: Zehetmair, Tetzlaff (both), Tognetti, St John, Barton Pine, Gringolts, Mullova (1993, 2008), Faust, and Ibragimova

Table 4.2. Average MSP and HIP tempos across all studied recordings made since 1903 (Joachim). Violinists who were added to HIP are: Zehetmair, Tetzlaff (both), Tognetti, St John, Barton Pine, Gringolts, Mullova (1993, 2008), Faust, and Ibragimova

Table 4.3: Average tempos in recordings made pre and post 1978 (before or after Luca). MSP and HIP versions are combined

Table 4.3: Average tempos in recordings made pre and post 1978 (before or after Luca). MSP and HIP versions are combined
  • 5 The Standard Deviation (SD) is defined as the average amount by which scores in a distribution diff (...)

11A focus on individual differences in tempo choices makes overall trends recede and diversity emerge. Some basic statistics may assist us to better understand the extent of this diversity. Standard Deviations (SD) are useful in determining global relationships. They show that there are examples in every movement of every work where the tempo difference is above two or close to three SD.5 Almost any violinist, irrespective of age, may be performing above one SD in any given movement, indicating fairly wide-spread tempo choices. This is also true when HIP and HIP-inspired players are grouped together against “hard-core” MSP players as well as when tempo choices are pooled into groups of non-specialist versus period violinists.

12One overall observation that can be made reasonably safely, I believe, is that extreme tempos (i.e. SD values well above ±2) are more typical among MSP players: First and foremost Zehetmair, Tetzlaff, Kremer (in 2005) and Gringolts who all tend to play fast, but also Edinger and Gähler who tend to choose tempos at the slower end of the spectrum. In contrast, when compared to the entire pool of the studied recordings, only two HIP violinists stand out with frequently higher than ±2 SD scores: Monica Huggett and, to a lesser extent, Brian Brooks. Hers are often among the slowest whereas Brooks’ are among the faster versions. Other HIP violinists might deviate much in just one movement, like Beznosiuk who plays the C Major Allegro assai much slower than the norm (-2.04 SD). Among HIP-inspired violinist a similar example is Tognetti who takes the A minor Fuga very fast (2.84 SD) and the Double of the B minor Sarabande rather slow (-1.82 SD). Otherwise his tempos are close to the average found in recordings issued since the beginning of the twentieth century.

  • 6 Key sources for an understanding of the role of meter and its relationship to tempo are Johann Kirn (...)

13So what have we learned about tempo that was worth the trouble and informs the main goals and argument of this book? We gained three significant insights: First, there is great plurality in tempo choice among the examined recordings which may be overlooked when reporting only averages. This diversity seems to be relatively independent of the performer’s age and generation; it is related, rather, to individual preference. Nevertheless closer study may indicate the impact of advances in musicology and performance practice in terms of knowledge about baroque dances and determining the right tempo of these and other movements (e.g. through an understanding of the meaning of eighteenth-century time signatures and notation practices).6 Blanket statements that the performance of baroque music has become faster over time hide these important reasons and may lead to unwarranted explanations privileging broader cultural or social forces.

14Second, although the average tempo of HIP-inspired players is often closer to the average of HIP violinists, the range of tempos found in the former group is nevertheless more in line with the spread found among “hard-core” MSP players (evidenced in the reduction of SD values when these players are grouped with MSP versions). This, in turn, points to the third insight, namely the greater congruence of tempo (i.e. less diversity) among HIP versions.

  • 7 The New Bach Reader—A Life of Johann Sebastian Bach in Letters and Documents, ed. by Christoph Wolf (...)

15Period violinists tend to choose less extreme tempos in every movement. This means that the tempo contrast between faster and slower movements is also less significant in these versions, reflecting historically informed tempo choices. Current readings of historical treatises encourage moderate tempos across the board; the idea being that extreme tempos are a feature of nineteenth-century romantic conceptions of music. Perhaps this belief impacts primarily on the slower movements where musicians may feel that an overtly slow Adagio or Largo may foster a romantic emotion. The resulting faster pace of ostensibly slow movements may foster the public impression that performances have become faster. Even if this was upheld by an examination of a large corpus of recorded performance, a tendency for faster tempos in all movement types may have more to do with historical performance practice research than modernist aesthetics: playing Bach’s music fast is congruent with Philip Emmanuel Bach’s edict, reported by Forkel, that “in the execution of his own pieces [his father] generally took the time very brisk.” However, the sentence continues by stating that “besides this briskness [Bach contrived] to introduce so much variety in his performance that under his hand every piece was, as it were, like a discourse.”7 The more moderate speeds found in HIP versions often serve a more closely articulated performance that is richer in nuance and agogic-rhetorical detail.

  • 8 Schröder, Bach’s Solo Violin Works, p. 168.

16Finally, the subjectivity of tempo is also underlined by a comparison of recommendations found in the literature and actual practice. For the Preludio of the E Major Partita, for instance, Schröder recommends a “relaxed [crotchet] = 110, approximately.”8 Most violinists tend to opt for a faster tempo with the average being around 118 (see Tables 4.2 and 4.3c). Schröder’s own version is considerably slower; in fact it is the second slowest at 99 crotchet beats per minute (Beznosiuk’s tempo is 98 bpm).

4.2. Vibrato

  • 9 Most violin tutors discuss vibrato at length. Many modern studies could also be mentioned that prov (...)

17Apart from tempo and dynamics, vibrato is perhaps the most often studied parameter, especially in empirically orientated studies of vocal and violin performance.9 Being a very personal matter, a whole study could be dedicated to analysing and comparing the nature and characteristics of different violinists’ vibrato. Although of potential interest to other violinists, it is doubtful if written accounts and objective measurements provide meaningful information regarding this particular matter. Visual inspection of players in action, reflecting on resulting sound qualities and pondering their affective dimension may be more productive. Engaging with bodily, kinaesthetic actions and musical-emotional gestures are likely to be more useful than thinking in terms of vibrato width (depth) or rate.

18In this regard my data set shows that period and HIP-inspired players vary their vibrato more, often combining it with “hair-pin” swells (i.e. rapid crescendo on a single note) or messa-di-voce effects (i.e. adding or increasing vibrato in the middle of a rapid crescendo-decrescendo). Longer notes starting straight and being vibrated from half-way through are also quite common. Other times the vibrato might be tapered off into straight notes either in combination with a diminuendo or without. Quite clearly, for these violinists vibrato is an expressive device whereas “hard-core” MSP players use it as a part of basic tone production. Their well-regulated, near-continuous, and uniform vibrato indicates this (Figure 4.1a-c).

Figure 4.1a-f. Spectrograms of bars 5-6 of the D minor Sarabanda performed by (a) Hahn, (b) Shumsky, (c) Podger, (d) Tognetti, (e) Ibragimova and (f) Mullova in 2008. [Spectrogram parameters: High band Hz: 4800; Window size / Display width: 6 sec; Colour Spectrum Level Minimum: -80dB; Frequency resolution: 12.5 Hz.]

19Exceptions among MSP players include Buswell. He plays many notes non-vibrato, using it instead to intensify certain longer notes, such as melodic high-points. Mullova is an interesting case as her vibrato often starts from below; in other words with a downward fluctuation of pitch (especially in 2008). In her later recordings she limits vibrato to function only as decoration, often simply a fast single quiver on a short or long note. On the basis of visually inspecting the spectrograms of her recordings I believe that it might be more bow (and / or finger) pressure fluctuation that contributes to a vibrato effect rather than change in the width of the pitch’s frequency range achieved through finger or wrist movements. Her practice of lifted bowing (quick decay) allows more room for changes in dynamics (intensity of sound level) than for vibrato.

  • 10 To the extent something jarring may create negative emotions, this is corroborated by Juslin when h (...)

20Such a conflation of bowing and vibrato underscores the multiplicity of elements contributing to the acoustic-aesthetic stimulus we listeners perceive. Tonal colouring enriches the expressive content musical moments communicate. The greater the blend of elements from bowing, phrasing, timing, dynamics, tone colour, harmony and melody, the stronger the “Gestalt” experience: we are less inclined to say “oh, did you hear that vibrato?” or “what exquisite bowing!” and more inclined to respond globally: “ah, that was beautiful!” When one or another performance element comes to the fore our analytical mind may interfere with our holistic perception. We may relish the fact that we can pick out the cause of an effect, “ahah, she used vibrato here to highlight that note”; or “gosh, that was a wide vibrato!” But if such analytical recognition happens too often, the overall verdict may be that the performance is “idiosyncratic” or “mannered.”10 So it is really doubtful what purpose the attention to measuring vibrato (or indeed looking at performance features in an itemized fashion as I do here in this chapter) serves.

  • 11 Vibrato measurements were taken on thirty-eight selected notes in the A minor Sonata’s Andante (num (...)

21Furthermore, quantitative results are also silent about additional characteristics of approaches to vibrato use. For instance, quite a few violinists reduce their vibrato during repeats (e.g. Buswell, Huggett, St John, Mullova in 1993). Fischer is the only player who vibrates more notes in repeats but the depth is extremely narrow, making the vibrato sound more like a fast, enhancing quiver or extra vibration. Nevertheless, for those who are interested in quantitative measures I offer Table 4.4 and a brief commentary on the four overarching information that can be deduced from the data: 1) a steep decline in use of vibrato since around 1995; 2) a fairly standard vibrato rate; 3) a narrower vibrato depth among HIP and younger players; and 4) a lack of noteworthy change in vibrato style in subsequent recordings of the same violinist.11

Table 4.4. Vibrato rate, width (depth) and frequency of use measured on selected notes in different movements and averaged across each selected violinist. (Rate is expressed in cycles per second (cps); Width in semitones (sT). Frequency refers to occurrence of vibrato on the selected pitches). Standard Deviation (SD) indicates the evenness of each player’s vibrato (the smaller the number, the more regular the vibrato)

Table 4.4. Vibrato rate, width (depth) and frequency of use measured on selected notes in different movements and averaged across each selected violinist. (Rate is expressed in cycles per second (cps); Width in semitones (sT). Frequency refers to occurrence of vibrato on the selected pitches). Standard Deviation (SD) indicates the evenness of each player’s vibrato (the smaller the number, the more regular the vibrato)

Brooks’ vibrato is often very fast and extremely narrow; hard to measure at all let alone accurately. Therefore the frequency percent is an estimate based on where rate could be measured but width not and where width was measurable but not the rate (signal showing a thick line with fluctuating intensity).
* Gähler’s vibrato is basically continuous but when playing full chords it is not easily executed or measured. Measurements were possible for more than 60% of the selected notes. His vibrato tends to be slow but not too wide except on single longer notes. On these occasions it becomes faster

  • 12 Richard Miller, The Structure of Singing: System and Art in Vocal Technique (London and New York: S (...)

First, the steep decline in the use of vibrato since around 1995 is also indicated by the considerable drop in vibrato use in subsequent recordings of the same violinist (see Kremer 91% to 45.3%, Barton Pine 62% to 14%). The exceptions are Tetzlaff, who vibrated more frequently in 2005 and Kuijken, whose approach remained steady, around 60%, the highest among HIP versions.
Second, according to various researchers, a well-regulated vibrato has around six to six-and-a-half cycles per second (cps).
12 Table 4.4 shows most violinists’ vibrato rate to hover around this value. The vibrato speed of HIP players is slightly slower (except for Brooks). I believe this slower rate to be related to the idea of vibrato being an ornament because a slower vibrato tends to be more audible and to sound more like an expressive decoration or a change in tone colour. A fast vibrato sounds more like a quiver, at times a sign of tension, although it could also tend towards a shake or trill.
Third, vibrato depth (or width) is shallower in HIP (up to 0.15 of a semitone [sT]) than in MSP versions (up to around 0.3 sT). This may also be related to the aim of avoiding vibrato altogether: In a stream of straight notes a very shallow vibrato may create enough tonal nuances to signify certain musical moments.
Fourth, multiple recordings of violinists show no substantial change in their vibrato style. One may speculate that no change in vibrato width and rate in these multiple versions is evidence that the technical delivery of vibrato is a personal signature and when players control it unconsciously it reflects their ingrained technical style. Of course at their level of professionalism it is perfectly within their conscious control to vary vibrato for expressive effect—or to eliminate its use entirely.

22Table 4.4 also provides information about variety of vibrato. This can be deduced from the SD values. Standard Deviation values indicate the variability in the measurement of each player’s vibrato. According to this, the depth varies much less than the rate. Brooks, Perlman and Tetzlaff in 1994 have the least regular vibrato rate, having greater than 1 SD. The results for Zehetmair’s vibrato are also worth pointing out: they are extreme and indicate one aspect of his highly idiosyncratic style. He used vibrato infrequently and also varied it quite a bit. At significant high notes it tended to be wide and slow—very noticeable, indeed.

23Finally it is worth picking out some violinists from the roster for individual attention. This may help to unpack the interaction of vibrato with other performance features or to explain seeming anomalies in Table 4.4.

Looking at players with narrow vibrato width, Kachatryan should be mentioned first. When he is playing softly, it becomes almost impossible to measure due to lack of visible signals. Hence the range of percentage for him in Table 4.4: the lower number excludes notes where only rate could be measured. Brooks’ vibrato is also often extremely narrow and quite fast, making accurate measurements difficult. Therefore the percentage of frequency of use is an estimate only; a combination of where rate could be measured but depth not (i.e. signal showing a thick line with fluctuating intensity) and where depth was measurable but not the rate. Ehnes’ vibrato is audible primarily because it tends to be rather slow. At the same time it is quite narrow (and therefore hard to measure) and it tapers off on longer notes. Lev’s vibrato seems often to be a result of bow pressure fluctuation, a consequence of strong down-bows followed by quick decay. At other times it might be a short quiver creating emphasis. Barton Pine has no measurable vibrato in 2004 (only the D minor Sarabanda was available). Bowing (e.g. swell and reverse swell / down-bow) shows some fluctuation but as it is not vibrato but an effect of bowing, it is rather pointless to measure.
In Gringolts’ A minor Andante there is nothing measurable. What one may hear as vibrato seems to be an effect of his bowing (pressure, speed) that creates a kind of “gliding,” rapidly fluctuating sound. However, it is hardly possible to call that vibrato: not a single vibrato curve can be seen in the spectrogram, just some unsteadiness in the intensity of the signal. In his interpretation of the E Major Loure, quivers (or just a single quiver) are more audible when “leaning-on” with bowing (down-bows). Proper vibrato curves are still hard to detect and often turn out to be trills. The two notes where vibrato could be measured in the Loure were the B crotchet on the down-beat of bar 8 and the A crotchet on the forth beat of the same bar. The vibrato rate of the two notes was 6.4 and 7.1 cps, respectively, while the depth measures were 0.13 and 0.19 sT.
Among older MSP players, who are inclined to regard vibrato as integral to tone production, Poulet’s vibrato is quite slow but not too wide. It becomes more prominent on melodically or harmonically important notes. Shumsky’s vibrato is wider than the younger players’ and at times slower as well. Gähler’s vibrato is basically continuous but when playing full chords with his curved “Bach-bow,” it is not easily executed creating blurred signals. Measurements were possible for more than 60% of the selected notes. His vibrato tends to be slow but not too wide except on single longer notes. On these occasions it was faster or sped up after a slower beginning.

24All this speaks particularly to two of my main aims in this book. Firstly, it shows that even vibrato has again become less homogeneous during the past thirty years; it has started to sound similarly varied to what one hears on early twentieth-century recordings. This is true on several levels: at the level of contrasting MSP and HIP approaches to vibrato along a continuum of frequency of use and tone production, and at the level of varying vibrato “style” or quality. Secondly, it shows the interaction of performance elements contributing to tone production and musical expressivity.

  • 13 Beethoven Violin Sonatas with Sophie Mutter and Lambert Orkis on Deutsche Grammopon 457 623-2 (1998 (...)

25Diversity in vibrato is interesting first and foremost for how it is varied, not whether it is used at all. It is fascinating to realize that vibrato has again become part of expression rather than tone production. And not just in this special repertoire as might be objected. I have studied many recordings of Beethoven and Brahms violin sonatas as well, for instance, and can confirm that variation in vibrato for expressive tonal effects is on the rise among players as different as Anne-Sophie Mutter, Viktoria Mullova, Rachel Barton Pine, Isabelle Faust or Daniel Sepec.13

  • 14 Joseph Szigeti, Szigeti on the Violin (New York: Dover Books, 1979), especially pp. 38-43.
  • 15 Katz, Capturing Sound, pp. 85-98, esp. 93, 95-96.

26Why is variety in vibrato an important finding? Because a fundamental piece of evidence that supports the criticism of uniformity in performance is uniformity of sound—homogeneous, undifferentiated tone achieved through seamless bowing of equal pressure throughout and perfectly controlled, even vibrato, as exemplified by Hilary Hahn’s playing, for instance. Joseph Szigeti complained already in the 1950s that players have lost the “speaking quality” of their bowing, that too much emphasis has been put on right hand technique and power of tone. He pointed out that without varied bow strokes the musical characters are ironed-out.14 Alternatively people blamed the recording industry and oversized concert halls for needing big sounds that carry into the farthest corners and balance well with the might of large orchestras. We have seen, in chapter three, how famous teachers contributed to this development and Table 4.4 clearly shows that violinists of the Juilliard / Curtis School continue this tradition. The continuous vibrato, allegedly developed to counteract the depersonalized nature of sound recordings,15 became a vehicle for an idealized aesthetic of perfect control manifest in consistent tone. Perhaps it is a saturated market calling on musicians to differentiate themselves, perhaps it is the opportunity offered by smaller independent record labels, perhaps it is the musicians’ need to do something new, perhaps it is our pluralistic time that allows musicians to interpret, to play and not be bound by rules and the tyranny of perfection that resulted in the return of varied vibrato. Most likely it is a combination of all this that calls on violinists to question the nature of perfection by creating performances full of tonal shades, uneven bow strokes and variously vibrated notes.

4.3. Ornamentation

27Much has been written about ornamentation in baroque music and I focus only on those matters that relate to the broader concerns of this book. Pertinent among these is the aesthetic effect achieved. Therefore, first I will discuss some of the scholarly opinions that highlight essential aesthetic problems and then will aim to analyze the recordings in this context.

Problems of Aesthetics and Notation Practices

  • 16 Basic and detailed information on ornamentation practices, including the differences between nation (...)
  • 17 Bach himself prepared such a table for his eldest son (see “Explication unterschiedlicher Zeichen, (...)

28One of the most difficult problems of ornamentation and improvisation in baroque music is deciding not so much how but when and how much to do it.16 There are countless charts providing solutions to symbols of graces and many examples of figurative embellishments.17 However, in the end all sources, modern and historical alike, reiterate that it is a matter of good taste to know when and what, or how much, is appropriate. What’s more, it is obvious that the sensibility of this infamous bon goût changes with the passing of generations.

  • 18 Pier Francesco Tosi, Opinioni de cantori antichi sopra il canto figurato (1723), cited in Frederick (...)
  • 19 Johann-Joachim Quantz, On Playing the Flute [Versuch], trans. by E.R. Reilly (New York: Schirmer, 1 (...)
  • 20 Cited in Neumann, Performance Practices, p. 521.
  • 21 Ibid., p. 538, citing Quantz, Versuch, chapter thirteen, par. 9.

29Take Pier Francesco Tosi (1653-1732), for instance. Writing in his old age, when the much younger Farinelli (alias Carlo Maria Broschi, 1705-1782) and other popular singers are admired for their ability to lavish simply sketched melodies with passagi and roulades, he decries them as “modernists” who are lacking in true taste, scorning them for “their offences against the true art of singing.”18 At the same time, Johann-Joachim Quantz (1697-1773), a contemporary of Farinelli, enthusiastically praises the 1720s-30s when, in his view, the art of singing reached its greatest height.19 However, reading both Tosi and Quantz further, one encounters statements from the former: “whoever cannot vary and thereby improve what he has sung before is no great luminary”;20 and from the latter: “all-too-rich diminutions will deprive the melody of its capacity to ‘move the heart.’”21 So how do we know what is “all-too-rich” and what may be an “improvement” on what has already been sung?

30Among modern writers on the matter Neumann provides an open-ended basic framework when he distinguishes between first- and second-degree ornamentation:

  • 22 Neumann, Performance Practices, p. 529.

As a rule of thumb […] an adagio is skeletal if it contains no, or only very few, notes smaller than eighths; it has first-degree diminutions if it contains many sixteenth notes; it has second-degree diminutions if it contains a wealth of thirty-second notes or smaller values. The skeletal types were always in need of embellishment; the first-degree types may fulfill stylistic requirements in the lower range […] further ornamental additions are optional and often desirable on repeats; the second-degree designs were in no need of further enrichment but on repeat could be somewhat varied.22

  • 23 Butt, Bach Interpretation: Articulation Marks in Primary Sources (Cambridge: Cambridge University P (...)
  • 24 For instance, compare Birnbaum’s text (published in The New Bach Reader, pp. 338-348) with Giovanni (...)

31Bach’s notation practices tend to fall into the third category (i.e. second-degree design). Deciding on ornamentation in his music is often complicated as he was in the habit of notating out diminutions, even graces that others indicated with symbols (see fn. 17 above). John Butt considers it likely that this practice of Bach, that he notates everything that is to be played, might be the key reason for the lasting appeal and value of his music.23 Bach was chastised for it in his own time by Johann Adolf Scheibe (1708-1776) and defended by Johann Abraham Birnbaum (1702-1748) in a public debate that seems to rehearse the familiar problem of taste, the composer’s “honour” and the performer’s “prerogative” that are voiced also in other historical sources.24

  • 25 The New Bach Reader, p. 338.
  • 26 The New Bach Reader, pp. 346-347.

32Scheibe (1737) reproached Bach for writing out all the melodic embellishments (figures or divisions) and for not leaving space for the performer’s improvisation: “Every ornament, every little grace, and everything that one thinks of as belonging to the method of playing, he expresses completely in notes.”25 In Bach’s defense, Birnbaum observed that it was a fortunate situation when a score where embellishments are added by the composer was available, for he knew best “where it might serve as a true ornament and particular emphasis of the main melody.” Birnbaum considered it “a necessary measure of prudence on the part of the composer” to write out “every ornament ... that belongs to the method of playing.” He asserted that improvised embellishments “can please the ear only if it is applied in the right places” but “offend the ear and spoil the principal melody if the performer employs [embellishments] at the wrong spot.” To avoid attributing errors of melody and harmony to the composer, Birnbaum posited the right of “every composer […] to […] [prescribe] a correct method according to his intentions, and thus to watch over the preservation of his own honor.”26

The Role of Delivery

33The mid-twentieth-century Bach scholar and conductor Arthur Mendel (1905-1979) pointed out the crucial lesson in this debate, which turns the attention away from petty point scoring and the matter of taste toward the fundamental issue in twentieth-century Bach performance and playing baroque music in general. He suggested that Scheibe’s objection was perhaps due to the difficult rhythmic patterns that arise from written-out turns and other embellishments:

  • 27 Arthur Mendel (ed.), Bach: St John Passion—Vocal Score (New York: Schirmer, 1951), xxii.

Because of the essentially improvisatory character of trills, appoggiaturas, and other ornaments, the attempt to write out just what metric value each tone is to have can never be successful. I think this may be partly what Scheibe meant in criticizing Bach for writing out so much [...] The attempt to pin down the rhythm of living music at all in the crudely simple arithmetical ratios of notated meter is [hardly ...] possible.27

34In their concern for the text, the notated score of a composition, and its technically correct rendering, modern musicians are easily misled by the visual representation of music. Notes of equal significance in print will likely be played with equal importance. Recognizing the ornamental nature of Bach’s notation practices is a first step toward rendering rhythmic patterns and melodic groups with some freedom.

  • 28 Noted also by Jaap Schröder in ‘Jaap Schröder Discusses Bach’s Works for Unaccompanied Violin,’ Jou (...)
  • 29 Joel Lester, Bach’s Works for Solo Violin: Style, Structure, Performance (Oxford-New York: Oxford U (...)

35The G minor Adagio is a good example of this problem. The opening bars of it would likely have been notated as in the top system of Figure 4.2 by most other baroque composers, especially those of the Italian tradition.28 Instead, Bach wrote out a possible embellished performance version. Playing the notes rhythmically accurately is therefore a mistake, because spontaneous ornamentation is never rhythmically stable or exact. As Lester points out, “Thinking of the Adagio as a prelude built upon standard thoroughbass patterns can [help] the melody [be] heard not so much as a series of fixed gestures, but rather as a continuously unfolding rhapsodic improvisation over a supporting bass.”29

Figure 4.2. Bars 1-2 of Bach’s G minor Adagio for solo violin (BWV 1001). The top system is a hypothetical version that emulates the much sparser notation habits of Italian composers such as Corelli, who tended to prescribe only basic melodic pitches that the musicians were supposed to embellish during performance. The middle stave is Bach’s notation, reflecting one possible way of ornamenting the passage. His original slurs indicate ornamental groups to be performed as one musical gesture. The lowest stave shows the “standard thoroughbass” that players might conceptualize for a “rhapsodic improvisation” to unfold

4.1. Literal versus ornamental delivery of small rhythmic values in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bars 1-2. Six versions: Monica Huggett © Virgin Veritas, Oscar Shumsky © Musical Heritage Society, Shlomo Mintz © Deutsche Grammophone, Julia Fischer © PentaTone Classics, Sergey Khachatryan © Naïve, Alina Ibragimova © Hyperion. Duration: 2.08. See also Audio examples 5.22, 5.23 and 5.24.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.07

  • 30 For trends in earlier recordings, see my ‘Towards a Performance History of Bach’s Sonatas and Parti (...)

36Tracing flexibility (or lack thereof) in performances of the G minor Adagio throughout the course of the work’s recorded history provides an excellent window into the trajectory of Bach performance practice since the beginning of the twentieth century.30 Among the recordings under examination here, Huggett, Wallfisch, Barton Pine, Ibragimova, Mullova and others, including Schröder and Buswell, play with enough flexibility to create the impression of free ornamentation. Shumsky, Fischer, Khachatryan, Ehnes, Mintz, among others, provide much more literal and measured readings (Audio example 4.1 at Figure 4.2).

  • 31 Michael Spitzer and Eduardo Coutinho, ‘The Effect of Expert Musical Training on the Perception of E (...)

37The difference in performance style also impacts on the perceived affective dimension of the piece. This movement is clearly melancholic-meditative since it contains many dissonances, displays descending melodic tendencies, and is in the minor mode;31 a soliloquy that perhaps sounds less sad and personal when played in a measured style and sounds more as a self-reflective monologue passing through passionate outbursts and calming-tenderness when performed with fluid, rhapsodic freedom. Pondering the perceived affective dimensions of various versions is fascinating, especially since fairly small variations in particular details can lead to quite diverse overall impression. I will come back to this in the next chapter where I consider individual differences and affective response.

38Given Bach’s notation practices, it might seem enough for the contemporary musician and listener to adopt a quasi-improvisatory style of playing when faced with performing Bach’s written-out ornaments. However, there are also movements that are less clearly decorated, where the quantity and place of additional ornaments and embellishments are worth contemplating. Although Bach habitually wrote out more figuration than his contemporaries, including elaborate ornamental figures as we have just seen, as well as appoggiaturas and terminations of trills that need to be recognized and performed as such, he indicated graces such as trills, slides, and mordents relatively sparingly. In addition, surviving successive versions of pieces, for instance the Inventions and Sinfonias (BWV 772–801) or the reworking for lute of the E Major Partita for Solo Violin (BWV 1006a), show different degrees of ornamentation. According to contemporary records, Bach was also in the habit of improvising richly textured, polyphonic continuo parts instead of the apparently more common simple chordal style. So what is a performer supposed to do? What might be within “good taste”?

  • 32 Neumann, Performance Practices, p. 528. I would not even use the word “legitimate” since the range (...)
  • 33 Neumann, Performance Practices, p. 536.

39There seems to be “a fairly broad range of legitimately possible levels of ornamentation, extending from a desirable minimum to a saturation point.”32 Moreover, “informed” subjectivity is encouraged by Neumann when he recommends Georg Friedrich Telemann’s published embellishments to his Sonate metodiche for violin or flute (Hamburg: [n.p.], 1728) as “helpful […] to late Baroque diminution practice because they strike a happy balance between austerity and luxuriance.”33

  • 34 Quantz, Versuch, chap. 18, par. 76, as paraphrased from Neumann, Performance Practices, p. 536.

40As always, performances are judged ultimately for their expressive affect and, in this regard, Quantz’s reasons for his admiration of “Italian singing style [and] lavishly elaborated Italian arias” are perhaps the most useful guide. His praise is earned because they are profound and artful; they move and astonish, engage the musical intelligence, are rich in taste and rendition, and transport the listener pleasantly from one emotion to another.34 In the footsteps of Quantz we should feel liberated to speak of the subjective, to use metaphor, to allow our holistic and affective mind-body to sort out what “works.” Such a reporting goes against my scientifically trained brain but feels valid to my listening, music-loving persona. I aim to balance the two by recognizing that both the performer and the analyst are confronted with a multitude of musical puzzles and possibilities with no hard and fast rules, but only “sufficiently developed taste,” culturally and historically conditioned expectations, and subjective boundaries regarding what feels appropriate as aids and foundations for aesthetic judgment. Importantly, I hope to have demonstrated that none of this ought to be a moral issue, not even in the music of the “great” Johann Sebastian Bach. Performance is not about absolutes but about conviction and affect, nowhere more so than in relation to ornamentation and embellishment.

The Performance of Embellishments

  • 35 A striking exception to this generalisation is Stefano Montanari’s 2011 recording of the works (Ama (...)

41So what do we find in the recordings? In this repertoire the past thirty years seem to demonstrate that we have reached a stage where quite a few violinists dare ornament several movements quite lavishly, especially in most recent times. However, ornamentation happens less in the slow movements that are already embellished by Bach and more in the lighter dance movements. Furthermore, adding short graces and altering articulation, rhythm or dynamics, or, as we have seen, using vibrato to ornament special notes are more common than melodic embellishment.35 Nevertheless, there are also significant instances of sumptuous embellishing, even complete rewriting of bars and passages, and even in movements containing written-out figuration.

42Again we can note the difference between what the factual information tells us and what it hides. Table 4.5 lists the most ornamented movements and the violinists involved in order of the amount of ornamentation and / or embellishment observed. It is immediately apparent that the E Major Partita features prominently and, although slow movements are represented, none of the opening adagios of the sonatas is listed (to my knowledge only Montanari adds embellishments in the G minor Adagio and this 2013 recording was made after the designated three decades of 1980-2010 under discussion here). The table also makes it apparent that ornamentation is more frequently practiced by non-specialist than HIP violinists—only Luca, Huggett, van Dael, Wallfisch, Beznosiuk and Podger represent period instrumentalists, and only Luca’s name can be seen in the columns of the slow movements. Huggett adds beautifully haunting ornaments in the D minor Sarabanda but only during the repeat of the first half (bb.1-8) and in none of the other slow movements. It is also obvious from Table 4.5 that less than one third of the forty-odd violinists studied here add ornaments and only four (Luca, Mullova, Gringolts, and Faust) to a significant extent.

Table 4.5. Most embellished movements listed in order of amount of ornamentation. The named violinists add graces and embellishments extensively, decreasingly so as moving downward in each column. Others not listed here may also add a few graces here and there

Table 4.5. Most embellished movements listed in order of amount of ornamentation. The named violinists add graces and embellishments extensively, decreasingly so as moving downward in each column. Others not listed here may also add a few graces here and there
  • 36 Fabian, ‘Ornamentation in Recent Recordings.’

43What we cannot see from the table is the fact that the solutions of different violinists tend to vary widely in terms of type, place and frequency of added ornaments. I discussed more of this fascinating detail in a separate paper so here I only provide some additional observations, transcriptions and Audio examples.36 My current concern is rather to explore further the meaning and affective dimensions of ornamentation and embellishment.

44In the slow movements (including the E Major Loure) what I find most important to highlight is the sharp contrast between Gringolts’s almost constant ornamentation and the more selectively added melodic passing notes and occasional flourishes or compound graces in Luca’s, Tetzlaff’s, Tognetti’s, Mullova’s and Faust’s or Beznosiuk’s versions. Kuijken tends to add only trills and approggiaturas. Furthermore, compared to Gringolts, these versions maintain a greater sense of rhythm and basic pulse and they also manage to keep the original melody intact.

45In the A minor Andante, for instance, Gringolts bows very smoothly and swiftly, adding many twists and turns throughout. In contrast Luca’s embellishments are more selective, filling in or varying certain well-recognizable melodic motives in a pattern-like manner, keeping with the movement’s steady style of harmony and rhythm (Audio examples 4.2).

4.2. Ornamentation in J. S. Bach, A minor Sonata BWV 1003, Andante, extract: repeat of bars 1-11. Two versions: Sergiu Luca © Nonesuch, Ilya Gringolts © Deutsche Grammophon. Duration: 1.07.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.08

46The difference in affect is palpable: the Andante in Luca’s hand is a simple, direct, beautifully balanced little aria, an intimate song. In Gringolts’ interpretation it sounds more like an over-cultivated flower that delights with its intricacies and quasi magical unfolding (as if on a sped-up film showing the opening of a bud) but somehow leaves the heart detached; calling forth a captivated outside observer rather than an intent listener enchanted by the serene beauty of the song. Overall I find Gringolts’ embellishments and bowing curiously unusual and intriguing. It surprises me that my students are quite categorically dismissive of this performance. I have always found them to prefer Luca’s version and in the ensuing discussion I will eventually offer reasons for this fairly constant aesthetic response.

  • 37 Robert Maxham, ‘Bach Violin Partitas: No. 1 in b; No. 3 in E. Solo Violin Sonata No. 2 in a. Ilya G (...)
  • 38 Joseph Magil, ‘Bach Solo Violin Partitas 1+3, Sonata 2. Ilya Gringolts DG 315,’ American Record Gui (...)
  • 39 Maxham does when he refers to the “jazzy ornamentation” of the Tempo di Borea Double.

47Interestingly, while reviewers of Gringolts’ disk comment on its “swaggering individuality and abundant fantasy”37 and find that his “readings are remarkably elastic, and many figures are played with capricious fleetness,”38 they hardly mention ornamentation.39 Most surprisingly Robert Maxham opines that “movements like the Second Sonata’s Andante ([…]) come closer to the interpretive canon,” meaning, I guess, the playing of violinists like Szigeti and Milstein! Apart from the aural experience, my transcription of Gringolts’ lavishly embellished performance of the Andante’s second half shows that this can hardly be the case (Figure 4.3).

Figure 4.3. A minor Andante, bb. 12-24. Transcription of melodic embellishments in Gringolts’ performance during repeat

  • 40 Schröder, Bach’s Solo Violin Works, p. 171.

48Similar observations could be made in relation to the E Major Loure or B minor Sarabande as well. The Loure is a particularly good example of the idiosyncratic nature of Gringolts’ ornamenting. In light of Schröder’s recommendation that the movement should never sound busy and should always retain a sense of “quiet nobility,” Gringolts’ reading seems a-historical.40 It sounds busy, indeed. The near constant addition of notes, trills, slides, scalar figures and appoggiaturas requires him to play with very light, uneven bowing, creating continuous chiaroscuro effects as he glides up and down and across the strings. Nevertheless he maintains some of the key rhythmic elements and pulse of the Loure (Audio example 4.3).

4.3. Ornamentation in J. S. Bach, E Major Partita BWV 1006, Loure, extract: repeat of bars 5-20. Ilya Gringolts © Deutsche Grammophon [edited sound file, first time play of bars 12-20 eliminated to show only the repeats]. Duration: 0.58.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.09

49The special quality of his playing becomes particularly clear when compared to other versions (cf. Audio examples 4.7, 4.12, 5.1 and 5.2). Although in the Sarabande Gringolts’ embellishing contains a few more metrically well-defined moments (for instance b. 7), elsewhere it again tends to sound overelaborated (e.g. b. 11) as I will discuss in more detail below. Here too the overall effect is created by the “gliding” bow strokes and the many smooth, sliding filler notes that grace not just larger melodic leaps but stepwise motions as well (cf. Audio example 4.4 contrasting Mullova, Gringolts and Luca, below).

50While transcriptions convey well the differences between Luca’s and Gringolts’ ornamentation of the A minor Andante, with the B minor Sarabande the situation is different. Here the visual information can be deceptive because the transcription of Mullova’s decorations may seem just as lavish as Gringolts’ (Figure 4.4a, b). Furthermore, some of her solutions resemble those of Gringolts in terms of placement and shape or pitch content. Yet the delivery, the sound of their respective performances is very different, highlighting the difficulty in finding academically meaningful (and printable) presentations of critical observations—or rather, it underscores the difficulty in making the object of study (aurally perceived sound) readily available for analysis, for visual and verbal dissection.

  • 41 Ideally one should have a tie or slur between the two high Bs at either ends of bars 9-10 in the tr (...)
  • 42 J[ohn] D[uarte], ‘Bach Partitas—No. 1 in B Minor, BWV 1002; No. 2 in D minor, BWV 1004; No. 3 in E, (...)

51One of the similarities between the transcriptions is found in measure 17. Both musicians grace the E dotted crotchet with an upper neighbour motion, but Mullova plays the two semiquavers ornamentally (i.e. soft, light), like Luca, not melodically as Gringolts does. In the transcriptions I aimed to convey this by using smaller notes for Mullova and normal semiquavers for Gringolts. Similarly, she plays the upward runs in bars 32 and 10 before the beat, giving emphasis to the downbeat and not affecting the basic pulse. Gringolts, on the other hand, while using a practically identical type of embellishment in b. 10, plays it in a less dotted and rhythmical manner. It sounds more like a gracing of the high B that he introduces as anticipation at the end of the previous bar (Figure 4.5a).41 Through his constant diminution of rhythm and anticipation or delay of harmonic notes Gringolts loosens the sarabande pulse that Bach so clearly outlined with harmonic, melodic and rhythmic structures. Mullova’s performance remains close to the implications of the score and is metrically steadier, with a less altered melody line. As one reviewer put it, “she adorns the ‘open spaces’ of the Sarabande of the B minor Partita in a delightful fashion” and “with discretion.”42 Her embellishments fulfill their supposed historical function as they heighten the rhythmic-melodic-harmonic character of the music. Gringolts’ constantly flowing, light flourishes cover up these underlying structures. To better appreciate the differences and to exemplify Mullova’s own interpretative trajectory, I have included the same section in all three of Mullova’s commercial recordings (in chronological order), followed by Gringolts’ performance and then Luca’s discussed below (Figure 4.4c). Mullova’s 1987 recording (the first example heard in the audio) has no ornaments and thus establishes the reference for Bach’s score (Audio example 4.4: repeat of bb. 9-17 Mullova 1987, 1993, 2008 followed by Gringolts then Luca. For other, unembellished versions see Audio example 4.17).

4.4. Ornamentation in J. S. Bach, B minor Partita BWV1002, Sarabande, extract: repeat of bars 9-16. Five versions: Viktoria Mullova 1987 © Philips, 1992 © Philips, 2008 © Onyx, Ilya Gringolts © Deutsche Grammophon, Sergiu Luca © Nonesuch. Duration: 2.31.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.10

Figure 4.4a. B minor Sarabande—Transcription of embellishments in Gringolts’ performance of the repeats

Figure 4.4b. B minor Sarabande—Transcription of embellishments in Mullova’s 2008 performance of the repeats

52In comparison to these highly ornamented versions I offer a transcription of the more sparsely embellished performance of Luca, bars 9-17 of which are heard at the end of Audio example 4.4 (Tognetti, Tetzlaff and Holloway embellish the movement so sparsely that it is not worth commenting on; it is enough just to enjoy!). What is interesting about Luca’s performance is the way he also varies accenting, articulation and rhythm during the repeats (e.g. both the dotting in bb. 8, 16 and the paired slurs in b. 15 are heard only during repeat). Although the added decorations often fall on the downbeat, their shape, rhythm and delivery tend to function so as to give impetus to the second beat of the measure, which is traditionally accented in sarabandes (e.g. bb. 2, 3, 10).

Figure 4.4c. B minor Sarabande—Transcription of embellishments in Luca’s performance of the repeats

53The E Major Menuet is another movement worth considering briefly. Being in the French style some players (e.g. Schröder, Podger, St John) introduce slight dotting or lilted playing of the quavers. Podger also adds short graces, trills and mordents, during repeats. Given its simple structure and many repetitions (especially if Menuet I is played again as a da capo after Menuet II and with all repeats), it seems quite reasonable to expect HIP performers to add decorations. This is not really the case. Apart from Podger and Wallfisch it is again the HIP-inspired MSP violinist who ornament the most, in particular Isabelle Faust (cf. Table 4.5). She performs all repeats in the da capo as well and not twice the same way. What is wonderfully playful about her rendering is that she uses ideas and figures from different parts of the movement to vary other bars. It teases the ear and brings a smile on the listeners’ face, as I have often found in conference and class presentations. The Audio example offers one straight and two slightly lilted interpretations of the first 8 bars—with St John’s an instance of the latter. This is followed by Faust’s recording edited so that only repeats are played taken from her performance of Menuet I and the Da Capo of Menuet I (Audio example 4.5).

4.5. Ornamentation and lilted rhythm in J. S. Bach, E Major Partita BWV1006, Menuet I, four extracts: bars 1-8. Jaap Schröder © NAXOS; bars 1-8 with repeat. Rachel Podger © Channel Classics; bars 1-8. Lara St John © Ancalagon; embellished repeats from Menuet I and its Da Capo. Isabelle Faust © Harmonia Mundi [edited sound file to show repeats only]. Duration: 2.31.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.11

  • 43 In 1983 Kuijken also adds a trill with appoggiatura on the downbeat of bar 12 and a trill on the hi (...)

54I identified and discussed several similar findings in relation to the other ornamented movements in my 2013 paper (see fn. 17), to which the interested reader is referred. However, I need to make one amendment to the data presented in that paper. There I state that “except for Luca adding two graces in measures 6 and 21, Faust is the only violinist who embellishes the D minor Sarabande, although I have heard it embellished by others in live concerts” (p. 18). Since publication I have noticed that Huggett adds soulful embellishments during the first repeat while Beznosiuk and Barton Pine add grace notes and short ornaments in a few bars: Beznosiuk in bars 1, 2, 6 and 21; Barton Pine (only on the commercial disk from 2004) in bars 4, 6, 11-12 and 18-19.43 What Beznosiuk does is simpler but at times similar to Faust’s solutions in that he plays around with adding mostly appoggiaturas. Barton Pine adds an appoggiatura in bar 6 and light-fast turn-like graces and trill elsewhere. Overall Beznosiuk’s playing is not just less embellished but also much more reserved and measured than Faust’s or Barton Pine’s. In addition Huggett plays most of the semiquavers in the second half, but in particular the coda (bb. 25-29), as if she was improvising (Audio example 4.6).

4.6. Ornamentation in J. S. Bach, D minor Partita BWV 1004, Sarabanda, six extracts: repeat of bars 1-8. Two versions: Pavlo Beznosiuk © Linn Records, Isabelle Faust © Harmonia Mundi; repeat of bars 16-21. Two versions: Isabelle Faust, Pavlo Beznosiuk; repeat of bars 4-21. Rachel Barton Pine 2004 © Cedille [edited to show only repeats]; repeat of bars 1-8 and 25-29. Monica Huggett © Virgin Veritas. Duration: 5.05.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.12

55Otherwise enough has already been said here regarding the fundamental aesthetic issue: whether the added ornaments fit well with the pulse and overall musical character of the movement or go somewhat against these. Other details, such as the appropriateness of choosing upper or lower appoggiaturas, compound ornaments, and so on (discussed in my above mentioned paper) seem like moot points. Few listeners are acutely aware of eighteenth-century rules and even fewer would register such minor details when listening to the recordings. The speed with which such ornaments pass by is frequently quite fast. Accurate transcription is often only possible by repeated close listening and slowing down the recording to half or slower speed with the aid of computer technology: useful for creating accurate data but of little ecological value. So to conclude this exploration of ornamented-embellished examples I rather just recommend listening to the brilliant and playful passage work that varies statements (usually but not always the last) of the Gavotte and Rondeau’s theme (Figure 4.5; see also Audio example 5.6 illustrating Gringolts’ embellishing recurrences of the theme) and the many graces and melodic alterations in the Loure (Figure 4.6 and Audio example 4.7). In the Loure the differences in sound effect, in delivery, are hard to convey in transcription. The scale up to the top B in bar 1, for instance, is played quite differently by all 3 violinists who introduce it (van Dael, Wallfisch and Mullova). And this is so even though all three of them use the gesture to create a spritely effect, to point up the rhythm and create energy! Speed, bowing, timing, and articulation contribute to the clearly perceivable aural difference.

Figure 4.5. Gavotte en Rondeau, E Major Partita: theme and its major variants

Figure 4.6. Transcription of 7 different ornamentations during the repeat of bars 1-3 of the Loure, E Major Partita WITH audio 4.7

4.7. Ornamentation in J. S. Bach, E Major Partita BWV 1006, Loure, extract: repeat of bars 1-3. Seven versions: Ilya Gringolts © Deutsche Grammophon, Isabelle Faust © Harmonia Mundi, Viktoria Mullova 2008 © Onyx, Lucy van Dael © NAXOS, Elizabeth Wallfisch © Hyperion, Sergiu Luca © Nonesuch, and Richard Tognetti © ABC Classics. Duration: 1.53.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.13

Diversity—Once More

56Moving on from added ornaments and returning to questions of understanding the meaning of notation, one particular moment of the set needs our attention: the final two bars of the A minor Grave (Figure 4.7). This excerpt illustrates spectacularly the rich variety of interpretations and individual solutions currently available on record. At least eighteen different solutions can be identified among the thirty-one examined recordings (Table 4.6). Such diversity challenges the claims that today’s performances occupy a uniform scene with very few real individuals compared to the “golden age” of the early recording era.

  • 44 Neumann, Ornamentation in Baroque and Post-Baroque Music, pp. 519-520.
  • 45 Frederick Neumann, ‘Some Performance Problems in Bach’s Unaccompanied Violin and Cello Works,’ in E (...)

57How to render the notation of the two double stops in the penultimate bar is open to debate (Figure 4.7), although Neumann asserts, on the basis of considerable evidence and concurring with several other authors, that the wavy line Bach notated indicates vibrato.44 According to Neumann a combination of bow and finger vibrato should be used followed by a trill starting on the main note.45

Figure 4.7. Grave, A minor Sonata: facsimile of final bars

  • 46 Greta Moens-Haenen, ‘Vibrato,’ Grove Music Online. Oxford Music Online.
  • 47 Robin Stowell, The Early Violin and Viola (Cambridge: Cambridge University Press, 2001), p. 66.
  • 48 Neumann, ‘Some Performance Problems,’ p. 29.

58Older MSP versions tend to trill the top notes or just the D#, more recent and HIP versions (except for Huggett and Ibragimova) seem to opt for a kind of tremolo effect achieved through “bow vibrato.” This is a term used by several authors and is related to what Moens-Haenen calls “measured vibrato.”46 She explains it as playing with “controlled pressure changes of the bow” and states that the “beats should be strictly in time” because “measured vibrato is as a rule written out by the composer” (in eighth or sixteenth-notes that are indicative of speed). The resulting pulsating sound is said to emulate the tremulant stop on the organ. Other authors use the term “bow vibrato” and discuss its possible context and execution a little less categorically. They seem to agree that the wavy line here may be a sign for this tremolo-vibrato effect. Stowell cites Baillot’s treatise identifying “a wavering effect caused by variation of pressure on the stick”;47 Neumann claims that “in the Italo-German tradition vibrato was often done by bow pulsation,”48 whereas Ledbetter suggests vibrato or a “sort of bow vibrato.”

  • 49 Ledbetter, Unaccompanied Bach, p. 121.

Since [the pair of wavy lines] seems to have something to do with going up a semitone, one obvious possible interpretation is the vibrato (flattement) followed by glissando up a fret (coulé du doigt) common in French viol repertory. The flattement can be either a finger vibrato (striking lightly, repeatedly, and as closely as possible to the fixed finger with the finger adjacent to it), or it can be a wrist vibrato. Marais and others use a horizontal wavy line for this. […] In the Brandenburg Concerto the solo parts [from measure 95] are accompanied in the ripieno by a sort of bow vibrato (repeated notes under the slur), and that also is a possible interpretation of the wavy line. J.J. Walther (1676) uses repeated notes with a wavy line to indicate this.49

59Basically, the accelerating frequency of energy pulsation creates an effect that is similar to what Caccini (1602) and others of the early baroque period called the trillo (rapidly repeated pitch) and that is indeed very similar to the tremulant stop on the organ. Since in the A minor Grave the effect is not written out in small note values but simply indicated by a wavy line connecting two held notes, the choice between finger and bow vibrato remains open and thus the term “bow vibrato” seems more appropriate than “measured vibrato.” Bow vibrato is also distinct from “tremolo” which is achieved by rapid up-down bow movements rather than change in bow pressure.

60Other possibilities for variety include the shape and speed of trills and tremolos, choice of dynamics and articulation (e.g. slurring the dyads), and the decisions whether to add short or long, lower or upper appoggiaturas and / or terminations to the trill(s), to play the final octave with or without anticipation, and whether to emphasize either note of the final octave. Altogether approximately eighteen different solutions were found in thirty-one examined recordings; some more subtle than others. Table 4.6 provides a summary of the main choices without differentiating all the nuances that are clearly audible and make each version slightly different from the other. These nuances are illustrated through a selection of audio excerpts (Audio example 4.8).

4.8. Interpretations of the notation at the end of J. S. Bach, A minor Sonata BWV 1003, Grave, extract: bars 22-23. Eight versions: James Ehnes (© Analekta Fleurs de lys) trilling both notes in both dyads; James Buswell (© Centaur) trilling the top notes of the dyads; Miklós Szenthelyi (© Hungarton) strongly vibrating the first dyad and then trilling the top note; Lara St John (© Ancalagon) trilling the top note and then both notes; Pavlo Beznosiuk (© Linn Records) producing a trillo (bow-vibrato) on the first dyad followed by a trill on the second; Brian Brooks (© Arts Music) performing a less obvious bow vibrato; Benjamin Schmid (© Arte Nova) playing a crescendo with increasing vibrato leading to trill on second dyad followed by decrescendo; Alina Ibragimova (© Hyperion) playing without vibrato, creating a small mezza-di-voce on the first dyad then a trill with diminuendo on the second. Duration: 2.43.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.14

Table 4.6. Summary of solutions in the penultimate bars of the A minor Grave

Table 4.6. Summary of solutions in the penultimate bars of the A minor Grave

Brooks’s and Tetzlaff’s delivery seems to be a combination of bow and finger vibrato perhaps closer to tremolo.
* Fischer and Schmid create a crescendo-decrescendo over the duration of the two dyads (accompanied by increasing-decreasing vibrato) whereas Ibragimova performs a messa di voce proper on the first dyad alone without any vibrato and articulates the second dyad quite separate from the first. There is hardly any crescendo-decrescendo in Podger’s and Khachatryan’s versions; straight-held dyads with trill on D# although one can perhaps detect a very slight tremolo/vibrato in Podger’s

Ornamenting or Improvising?

  • 50 This has started to change as evidenced in courses and masterclasses in improvisation for classical (...)

61Our discussion of ornamentation started with the need to understand what Bach’s notations mean; to render the many small note values as if spontaneously unfolding improvisations over an imagined melody. Circling back to such broader issues, the contention that ornamentation should always be improvised is worth unpacking. So far I have argued that the performance should sound as if improvised. How are we to know, really, whether or to what degree an added ornament is pre-planned or spontaneous? Musicians intimately familiar with a particular style are able to play extempore, they just have not been encouraged much to do so until recently.50

62People, including musicians, used to say that ornamentation in recorded performance is not desirable because of the repeatability offered by the medium. The idea being that the listener will not be able to hear it as improvised, or as an added ornament, because it will recur unchanged in each playing of the record. It may even become annoying to hear it over and over again instead of “the music proper.” Christopher Hogwood apparently advised his players against ornamenting in the recording studio

  • 51 Timothy Day, A Century of Recorded Music: Listening to Musical History (New Heaven and London: Yale (...)

because he felt that risk-taking, ‘wild risks’ and ‘fantastic cadenzas,’ improvisatory élan, spontaneity and dangerous living which would certainly elicit cheers in a live performance ‘nearly always pall on repeated hearings.’51

63A debatable position, indeed! Yes, perhaps I hear some of these embellishments as if part of the composition—especially when they seem as fitting and enriching as in Mullova’s and Faust’s recordings. But I certainly don’t get tired of hearing them and due to the style of delivery, always hear them as embellishments.

  • 52 I am aware that this could be a result of post-performance editing (patching up two different play- (...)

64However, there is one recording where I think the listener is surely witnessing improvised (rather than potentially pre-meditated) graces: Richard Tognetti sometimes adds graces during first play and different ones during repeat.52 This habit and the way he plays the ornaments create an impression of ad hoc, possibly improvised-on-the-spot performance. Often ornaments are added at less obvious places as well and occasionally they sound jolted rather than smoothly integrated. For me this indicates a spontaneous gesture, not something fully thought through, practised and polished. Such instances are the added trills in bb. 7-8, 21, 30-31 during the first play of the B minor Sarabande rather than in the repeat, whereas during repeats there is only one embellishment on the last beat of bar 3. Another example is the added embellishment during the repeat of bar 1 of the Sarabande Double. Given the structure of the music, this gracing could become a pattern but it doesn’t; none of the other analogous bars has any ornamentation. All this suggests spontaneity rather than relying on premeditated solutions, even if it happens to be a coincidence resulting from post-production editing and selection of takes. If takes are so different, then there must have been spontaneity. That such improvisation can be witnessed in a studio recording goes some way against the much heard assertions that the studio inhibits players, that recordings are not performances and that the spontaneous nature of ornamentation is against the repeatable nature of sound recordings.

65In my experience repeated listening does not diminish the aesthetic value and spontaneous effect of the embellishments observed in these versions of Bach’s Six Sonatas and Partitas for Solo Violin. I also question the reality of people commonly listening to the same recording repeatedly, especially now, when there is an abundance of music at one’s fingertips to download instantly.

Summary

66Ornamentation is a performance feature that could be seen as independent from most other aspects. You either add a trill, grace note, melodic embellishment or you do not. Simple. End of story. No wonder that during the early decades of the early music movement an enormous emphasis was placed on discussing the importance of adding graces to create a “stylish” baroque performance. But what is “stylish”? The decision to add a trill rather than a mordent? To trill from above rather than from the note? To play a trill with or without termination? Or to correctly decide whether the appoggiatura should be short or long? If this were true, my discussion of the recordings would have completely missed the point!

67I did not engage with any of such fussy detail because the real issue in ornamentation, as I see it, is indeed style; the musician has to demonstrate bon goût not just through the choices of ornaments but through their placement and frequency. Most importantly, they have to perform the embellishments in a way that truly decorates the music, enhances its character, rhythmic, melodic, and harmonic potential. As soon as we start talking about these aesthetic issues it becomes all too obvious that ornamentation is also interacting with several other performance features in achieving its effect. It is not a simple matter, not an independent element but combines with articulation, rhythm, dynamics, timing as well as bowing and tone production. Throughout my commentary I repeatedly drew attention to the differences between what transcriptions can convey and what one actually hears. I tried to describe the sound effect, the perceived gesture rather than evaluate the content or type of an ornament and embellishment.

68The discussed examples confirmed an increasing liberty in performing these works—once regarded monumental and untouchable—that brings forth a playful attitude. Violinists as diverse as Luca, Mullova, Podger, Tognetti, Faust and Gringolts seemed especially to delight in manipulating the material through added trills and slides, appoggiaturas and other graces or even changes in melodic turns, filling in gaps between notes, ornamentally highlighting gestures and, ultimately, re-writing measures and entire passages, as in the Gavotte an Rondeau. Most importantly, the graces and flourishes, whether written out by Bach or added by the player were delivered “gesturally,” with a sense of play, abandon, and improvisatory freedom. In this regard Mullova, Faust and Luca have excelled especially, but all of them displayed a confident “ownership” of the pieces conveyed by a sense of exuberance, daring fervour and convincing personal authenticity.

  • 53 As recalled in Neumann, Performance Practices, p. 538.
  • 54 Tellingly reviewers are ambivalent although tending towards the positive: Maxham (‘Ilya Gringolts,’ (...)

69If we now reconsider Quantz’s opinion, namely that he saw a danger that all-too-rich diminutions will deprive the melody of its capacity to “move the heart” and recommended that instead of indulging too freely in diminutions, players should render a simple melody nobly, clearly, and neatly,53 then this examination should perhaps conclude with the verdict that Gringolts may have gone too far. His performance can still strike as appealing and musical—especially if somebody has never heard the pieces before—but not quite appropriate if we desire to preserve the style and compositional aesthetics of Bach as we currently understand these.54 In Gringolts’ performance, the lack of rhythmic definition and pulse and constantly shifting dynamics and timbre make his embellishments sound restless and over-elaborate, perhaps more rococo in style, bringing to mind the aesthetic reactions of seventeenth-century commentators who originally used the word barocco in a pejorative sense to denigrate an over-elaborate piece by comparing it to a misshapen pearl.

  • 55 Richard Taruskin, ‘On Letting the Music Speak for Itself: Some Reflection on Musicology and Perform (...)

70Although it is undeniable that the embellished versions represent a minority group among the thirty-odd recordings studied here, the variety and creativity of solutions found in them proves that there are violinists today who are not afraid of putting their personal stamps on Bach’s works. Interestingly, MSP violinists outnumber HIP players among those who frequently introduce sumptuous embellishments (Table 4.5). The sheer fact that embellished versions appear to proliferate as we pass through the decades (think also of Montanari’s 2013 recording not studied here) provides ground for hope that performers are leaving behind the modernist “Urtext-mentality” of the 1950s to 1980s period. As they reclaim their prerogative to bring compositions to live rather than just “letting [them] speak for themselves,”55 these violinists offer listeners diverse and individual interpretations. The choice is wide and the differences among them are as good as ever.

4.4. Rhythm

  • 56 Houle, Meter in Music.

71As I have mentioned earlier, there are several issues one should consider under the umbrella of rhythm. Broadly speaking the topic involves anything to do with the timing of notes and thus it is linked to articulation as well as bowing. In baroque music, rhythm is organized around metrical hierarchies and the underlying harmony.56 A projection of these through micro variations in rhythmic delivery may also impact on local tempo. And, as we have just seen while exploring ornamentation, performances that sound improvisatory are achieved through flexible-gestural rendering of notated rhythmic values. Speaking more specifically, the playing of dotted rhythms and the French convention of notes inégales are particularly noteworthy when baroque performance conventions are considered. First I will focus on these, leaving the broader, more qualitative matters of rhythmic flexibility for later. However, it will transpire that even the seemingly straightforward matter of playing dotted rhythms is impossible to discuss without reference to tempo and articulation or bowing. Neither close listening nor quantitative measurements can tell the full story on their own as cause and effect are not veridical.

Dotted Rhythms

  • 57 The literature on rhythmic alteration in baroque music is too large to be cited here. Comprehensive (...)
  • 58 For instance in relevant movements of the two Passions, the Brandenburg Concertos and the Goldberg (...)

72The performance of dotted rhythms in baroque music has received much attention during the second half of the twentieth century. Most researchers advocated double or over-dotting in pieces where such rhythms prevail. For long, the late Frederick Neumann was the only author voicing an opposition.57 Case studies of recorded performances generally confirmed over-dotting in Bach repertoire.58

  • 59 The dotting ratio in this discussion is expressed as percentage of the duration of the dyad. In oth (...)

73In the Violin Solos, three movements offer opportunity for studying the delivery of dotted rhythms: The D minor Corrente, B minor Allemanda, and the C Major Adagio. Table 4.7 provides the quantitative results showing slight under-dotting in the D minor Corrente and over-dotting in the other two movements.59

Table 4.7. Average dotting ratios in 3 movements where such rhythms prevail; selection of recordings, listed in order of release date

Violinist

Dm Corrente

CM Adagio

Bm Allemanda

Luca 1977

0.74

0.8

0.85

Kremer 1980

0.76

0.8

0.73

Ricci 1981

0.76

n/a

n/a

Kuijken 1982

0.71

0.86

0.75

Zehetmair 1983

0.70

0.85

0.78

Shumsky 1983

0.73

Mintz 1984

0.78

0.81

0.78

Schröder 1985

0.72

0.78

0.77

Perlman 1987

0.65

0.8

0.75

Mullova 1987

n/a

n/a

0.79

Buswell 1989

0.75

0.79

0.78

Ughi 1991

0.77

Mullova 1993

0.74

n/a

0.8

Tetzlaff 1994

0.77

0.78

0.77

Huggett 1995

0.69

0.79

0.71

Van Dael 1996

0.73

0.81

0.74

Poulet 1996

0.71

0.80

0.82

Rosand 1997

0.75

n/a

n/a

Wallfisch 1997

0.72

0.85

0.78

Hahn 1997

0.78

0.81

n/a

Gähler 1998

0.74

0.73

0.74

Ehnes 1999

0.8

0.81

0.75

Podger 1999

0.72

0.77

0.77

Schmid 1999

0.77

0.77

0.82

Barton Pine 1999

0.76

0.81

0.77

Matthews 1997 (2001 issue)

0.73

0.79

0.79

Brooks 2001

0.72

0.89

0.75

Gringolts 2001

n/a

n/a

0.76

Lev 2001

0.75

0.74

n/a

Kuijken 2001

0.71

0.86

0.76

Poppen 2001

0.76

n/a

n/a

Szenthelyi 2002

0.74

0.82

Barton Pine 2004

0.74

n/a

n/a

Holloway 2004

0.75

0.75

0.80

Kremer 2005

0.76

0.85

0.73

Tetzlaff 2005

0.78

0.82

0.73

Tognetti 2005

0.72

0.81

0.8

Fischer 2005

0.78

0.81

0.79

Beznosiuk 2007 (2011 issue)

0.66

0.76

0.74 (v. kerned)

Barton-Pine 2007

0.7

0.82

n/a

St John 2007

0.67

0.77

0.78

Mullova 2008

0.71

0.78

0.78

Faust 2009

0.73

0.79

n/a

Ibragimova 2009

0.70

0.84

0.75 (kerned)

Khachatryan 2010

0.74

0.78

0.81

Average

0.73

0.80

0.77

  • 60 Data is reported for the 1930s-1950s period in Dorottya Fabian and Eitan Ornoy, ‘Identity in violin (...)
  • 61 Samuel Applebaum, The Way they Play, Book 4 (Neptune City, N.J., Paganiniana Publications, 1975), p (...)
  • 62 Neumann, ‘Some Performance Problems,’ p. 25.

74Compared to earlier recordings and published views of some violinists, the tendency to under-dot in the Corrente is striking. It is also stronger among HIP (average length of dotted note = 0.716) than MSP versions (average = 0.742). Earlier selective measurements of recordings made between 1930 and 1970 showed over-dotting becoming the dominant practice by the 1950s, excepting Heifetz.60 In an interview Luca talks about over-dotting and suggests that the D minor Corrente “is most effective played very short on the sixteenths.”61 His delivery of both the dotted and the short notes is indeed very staccato. Nevertheless even his average dotting is slightly under-dotted (0.74). This overall trend towards under-dotting may reflect the view that dotted rhythms within fast triplet motion are to be assimilated into long-short swings because the notation of dotting in such context represents the eighteenth-century way of expressing what later became notated as crotchet-quaver with a triplet bracket over the two notes.62 (Audio example 4.9)

4.9. Dotting and accenting in J. S. Bach, D minor Partita BWV 1004, Corrente, extract: bars 1-6. Three versions. Sergiu Luca © Nonesuch, Ingrid Matthews © Centaur, Julia Fischer © PentaTone Classics. Duration: 0.30.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.15

75In contrast, both the B minor Allemanda and the C Major Adagio tended to be over-dotted. However, these movements also show differences between HIP and MSP versions in the extent of over-dotting. Compared to MSP violinists, HIP players used a lesser over-dotting in the Allemanda and a stronger over-dotting in the Adagio (cf. Table 4.7). Additionally, one could observe a progressively more staccato delivery of dotted patterns in the Allemanda and Corrente while the slower Adagio tended to be played legato. In this movement the dotted notes form multiple stops. The time needed to sound each note of the three or four-part chord has likely contributed to the elongated delivery.

  • 63 Fabian and Schubert, ‘A New Perspective.’
  • 64 Dorottya Fabian and Emery Schubert, ‘Musical Character and the Performance and Perception of Dottin (...)
  • 65 Nicholas Cook, Beyond the Score: Music as Performance (New York: Oxford University Press, 2014), p. (...)

76Mentioning articulation and tempo in relation to dotting is crucial as research shows their impact on the perception of dotting.63 Dotted rhythms sound sharper, more over-dotted when played in a detached style with “air” (i.e. silence) between the long and the short note. They also tend to sound more dotted when the tempo is fast. In other words, the desired effect or musical character can be achieved without much over-dotting, explaining the lesser average ratios found in the B minor Allemanda compared to the C Major Adagio. Listeners, and performing musicians are also listeners, “may believe they respond to the ‘dottedness’ of a rendering” but “their judgement seems to reflect […] a higher order construct,” a holistic perception of performance features that “incorporates tempo and articulation” as well as dotting.64 Musicians focusing on projecting a particular character or mood seem to intuitively adjust the interacting elements. But when trying to explain what to do, they may incorrectly emphasize one or the other component as seen in much of the above mentioned literature on dotting in baroque music. As Nicholas Cook noted, there are musical situations when “close listening may correctly identify the effect, yet fail to proceed from effect to cause.”65

  • 66 Research into jazz swing has shown that the ratio reduces as the tempo increases (i.e. the length o (...)

77This auditory illusion may also explain why we hear most recordings of the D minor Corrente to be well dotted in spite of the measured under-dotted ratios. The fast tempo most performers chose in this movement compensates for the smoother ratio between the dotted and short notes. In the flow of running triplets the airy (or “kerned”) delivery of the dotted notes and the staccato articulation of the short notes create a floatingly skipping effect.66 Importantly, the dotting ratio tends to be sharper on downbeats and most under-dotted on the last pairs before the return of the running triplets (i.e. last beat of bars 4 and 6, for instance). In other words a strong sense of pulse propels the skipping effect forward to the next flight of triplets (cf. Audio example 4.9).

  • 67 The term “kerning” was introduced by Schubert and Fabian to denote the gap or silence between notes (...)

78A comparison of Kremer’s two recordings of the B minor Allemanda illustrates the auditory illusion at play (Figure 4.8; Audio example 4.10). The more sustained style of the earlier version sounds less dotted even though the measured ratios produce identical average dotting in the two recordings. In the later version Kremer plays both the dotted notes and their short pair staccato, in a “lifted” way, allowing the decay of the tone to start on average around 66% of the notated nominal duration. This “kerning” is stronger in case of the long notes, when the decay starts already around 56%. Shorter note values are held on average to 76% of their written value.67 The different delivery shows up very clearly in the spectrogram of the audio file (Figure 4.8).

Figure 4.8. Adobe Audition screenshots of spectrograms showing bars 1-3 of the B minor Allemanda in Kremer’s two recorded performances. The 2005 version (bottom pane) shows more gaps between note onsets indicating staccato playing

4.10. Articulation and perceived dotting in J. S. Bach, B minor Partita BWV 1002, Allemanda, extract: bars 1-3. Two versions. Gidon Kremer 1980 © Philips, 2005 © ECM. Duration: 0.26.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.16

79Mullova is also an interesting case. She has 3 recordings of the B minor Allemanda (Figure 4.9; Audio examples 4.11). The earliest, 1987 version is slow, the bow strokes are long and even, giving it a sustained and broad feel. The 1992 version is more staccato but not much faster. It sounds a little more dotted overall, partly because of the shorter bowing and more detached style. The most recent, 2008 version is faster, even more staccato and sounds the most dotted. Inspection of spectrograms shows lots of kerning; both the dotted and the short notes are sharply articulated with clear gaps between note onsets (Figure 4.9). When one compares the measured dotting ratios of the three recordings the difference is very little, with the averaged dotting ratio found to be the least over-dotted in the most recent version: 0.79 in 1987, 0.8 in 1992, and 0.77 in 2008. It must be the faster tempo and staccato delivery that contribute significantly to nevertheless perceiving this last recording as the most dotted of the three.

Figure 4.9. Adobe Audition screenshots of spectrograms showing bars 1-3 of the B minor Allemanda in Mullova’s three recorded performances. The 1992 (middle pane) and 2008 (bottom pane) versions show increasing gaps between note onsets indicating progressively more staccato playing

4.11. Articulation and perceived dotting in J. S. Bach, B minor Partita BWV 1002, Allemanda, extract: bars 1-3. Three versions. Viktoria Mullova 1987 © Philips, 1992 © Philips, 2008 © Onyx. Duration: 0.42.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.17

80A slightly different case to the movements discussed so far (Table 4.7) is the E Major Loure. Dotted rhythms are frequent here too, but they rarely form a “running” pattern. Instead they occur in isolation but at prominent melodic or metrical moments. Furthermore, both the dotted crotchet—quaver and the dotted quaver—semiquaver combinations are common. Importantly, the Loure is perhaps the most diversely performed movement of the whole set of the Six Solos. As will be discussed in chapter five, the Loure was considered to be “the slow movement” of this partita for a long time. The many languid, lyrical, and melodically conceived interpretations evidence this approach. With the HIP movement in full swing by the 1980s, the Loure’s dance character has become recognized, giving impetus for a less lyrical and more rhythmical delivery in recent versions. However, long-held traditions die hard and performances of the Loure seem to occupy an open field: tempo often remains slow but rhythm more sharply shaped, or tempo faster but articulation legato and phrasing melodically orientated. What is interesting to note therefore is not so much the dotting ratio per se, but the interconnection of tempo, articulation, phrasing and dotting.

81Contrary to what we observed in relation to the D minor Corrente, B minor Allemanda and C Major Adagio, here the slower, more legato versions (e.g. Hahn, Khachatryan, Fischer) tend to be less over dotted, with the average dotting ratio being somewhere between 0.75 and 0.77. The more staccato versions of the Loure, whether faster or similarly slow, have a sharper average dotting ratio, around 0.80 (Audio example 4.12, see also Audio examples 4.7, 4.3, 5.1 and 5.2).

4.12. Tempo, articulation and musical character in J. S. Bach, E Major Partita BWV 1006, Loure, extract: bars 1-8. Four versions: Hillary Hahn © Sony, Miklós Szenthelyi © Hungaroton, John Holloway © ECM, Alina Ibragimova © Hyperion; repeats of bars 1-8: Jaap Schröder © Naxos. Duration: 3.13.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.18

82I was puzzled enough by this finding to look closer. First I found that the dotted quavers tended to be played less dotted than the dotted crotchets (except for Khachatryan and Fischer, among a few others). Score analysis revealed the reason (Figure 4.10): these dotted quavers in, for instance, bb. 4, 6-7 are preceded by a long written-out appoggiatura (or suspension) on the downbeat and so played with emphasis by most players, blurring the onset of the dotted note. Since the dotted notes are in essence the resolution of the dissonance, they are delivered softly and under the same bow as the suspension (note that these two notes are slurred throughout by Bach). Furthermore, since in terms of harmony these dotted quavers are concluding points, their semiquaver pair actually functions as an anacrusis to the following note. So it is not so much the dotted note that catches the ear but the short note as it leads quickly into the next main note. This explains why these patterns are not particularly over-dotted; rather they are articulated in a manner that highlights melodic direction and adds rhythmic momentum.

Figure 4.10. Loure, E Major Partita, bars 4-7

Rhythmic Alteration

83Apart from dotting, alteration of rhythm occurs in other context as well. In particular, a lilted-dotted manner of performing the quavers in the two E Major Menuets (e.g. Podger, Matthews, Tognetti, Holloway) and the B minor Sarabande (e.g. Luca and Mullova in 1992 during repeats) are noteworthy. Although the French convention called notes inégales was supposed to be used in pieces where strings of evenly notated quavers were common, the slightly lilted delivery of quaver pairs in these French dance movements is stylistically reasonable and certainly effective. In Menuet I of the E Major Partita there are complete bars with only quavers (bb. 4, 6) and the swinging interpretation gives them energy and a stronger sense of pulse. This lilting also adds variety and contrasts well with the shifted accentual pattern Bach creates through the prescribed slurs from bar 20 onwards. These indicate slurring the first three quavers in each figure, effectively going against the 3/4 metre and earlier paired quavers, although still highlighting most down-beats (cf. Audio example 4.5; see also Figure 5.3).

84In the B minor Sarabande the quavers that are played in a slightly swung manner by Mullova in 2008, for instance, tend to have a melodic pattern that resembles “sigh motifs” (b. 15, cf. Figure 4.4b; Audio example 4.4). In bars 21 and 23 Bach even marks them with the indicative slur (descending pairs of slurred notes). If a performer intends to project the sarabande’s characteristic pulse and the harmonic-melodic design of the movement, lilting the rhythm in these bars is almost inevitable. For instance, technical constrain caused by the triple stop crotchets in bar 15 together with the emphasis on the musical content of descending pairs of suspension-resolution quavers (“sigh motif”) in the melody will make the on-beat notes sound longer and stressed and the off-beat quavers lighter and shorter. So here we have an example when score and performance analyses go hand in hand rather than being in a hierarchical relationship. The cause and effect are reciprocal.

85As discussed earlier in relation to ornamentation, Luca’s interpretation of the B minor Sarabande defers from these more subtle and general rhythmic flexibilities because he plays several of the quaver patterns in a strongly dotted manner during repeats, especially in the first half of the movement. Dotting is first introduced in the prima volta bar 8 where he plays the first quaver short and then the next two as a dotted dyad. However, he doesn’t apply this alteration in a uniform manner. For instance, the similar gesture in bar 2 is played as written but in bar 4 it is dotted again. The first four quavers in bar 3 are played straight but the last two are dotted, and so on. This constant variation and fluidity lends an ornamental quality to these rhythmic alterations. Mullova also plays notes in a dotted manner in 1992 (in b. 8 prima volta, and the repeat of bb. 1, 7 (beats 2-3), 10 (beats 2-3), 16, 21 (beat 2), 23 (beats 2-3), 30 (beat 2)) as can be heard in Audio example 4.4.

86Finally, there are also those alterations that affect a whole group of notes. In these movements, like elsewhere, the difference in interpretation may largely depend on how flexible a violinist may shape certain gestures, how she or he may project the pulse of the dance, the harmonic architecture of the piece. When Menuet I is played in a “sturdy” style, with equal emphasis on all chords, the dance character is weakened because the 3/4 pulse is hard to perceive.

4.13: Sturdy style of evenly emphasized beats in J. S. Bach, E Major Partita BWV 1006, Menuet I, extract: bars 1-8. Rudolf Gähler (curved “Bach-bow”) © Arte Nova Classics. Duration: 0.13
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.19

87When notes are grouped to highlight metrical units, like, for instance, Tognetti’s or Holloway’s delivery of the slurred four semiquavers in bar 14 of the B minor Sarabande (heard in Audio example 4.17), then there is a different sense of phrasing and movement, as we can hear in many examples used throughout this book. Slightly stressing moments by holding over harmonically or metrically important notes, or the first note under a slur and then slightly hurrying the remaining notes within the same beat or until the next significant harmonic-metric moment occurs, creates subtle nuances and local ebb and flow.

Rhythm and Musical Character

88Beyond dotting ratios, formal notes inégales convention, and other quantifiable matters it is the character of these timing flexibilities that underlie the perceived differences between the studied recordings. It is not just how the dotted rhythms are delivered in the D minor Corrente that catches the attention. But whether the running triplets are grouped by metric units (i.e. by bar or pairs of bars) or played-out in a literal fashion; whether certain notes are accented through staccato or stressed by dynamic or timing accents, or not highlighted at all.

  • 68 Schmid performance shows some similarity but he seems to just stop the bow after the downbeat and t (...)

89Matthews’ version starts differently to many others because she accents not just the down-beat chord but also the first note of the first triplet group in both bars 1 and 2. Playing these notes staccato makes them sound almost linked to the previous crotchets; as if they were slurred to the down-beat. This feature of her interpretation is more striking and memorable than the question of how she delivers the dotted notes (cf. Audio example 4.9).68 St John’s E Major Menuet is similarly unique in that she really hurries the second beat (the quavers sound almost short-long) but takes the first beat rather leisurely, enabling a strong sense of down-beat but an unusual, slightly limping 3/4 pulse. This slight hurrying towards the end of bars is also perceptible in bars 4 and 6. The strings of six quavers of these bars in St John’s performance move rapidly towards the next bar, but not before a momentary hesitation on the first note in each group of six. The effect is assisted by minor shifts in dynamics. As discussed under ornamentation, an even more playful Menuet is created by Faust who makes many alterations both rhythmic and ornamental (cf. Audio example 4.5).

90Innumerable other examples could be mentioned, of course, that have unique features and I will discuss some more as I progress towards a holistic analysis of these recorded performances. Although many interactions of performance features have already been pointed out in this chapter, the topic of playing rhythm “gesturally,” grouping notes according to pulse, harmony or melodic contour is the first that cannot be separated out properly and is best discussed in terms of bowing, phrasing and articulation, the next big topic of this chapter.

4.5. Bowing, Articulation and Phrasing

Bowing and Timbre

91One of the most commonly noted differences between HIP and MSP violin playing rests with bowing. Older and MSP violinists use a broadly uniform, seamless bowing with mostly even portato strokes, (e.g. Shumsky, Poulet, Perlman, Hahn, Ehnes, Fischer, Khachatryan). HIP violinists tend to use shorter bow strokes and the uneven distribution of weight of their period bows makes the difference between up and down strokes more prominent. Accordingly, the performances of the younger and HIP-inspired players and especially those using a baroque apparatus display a less on-the-string bowing and less even tone. They use a great variety of strokes and a more articulated style with faster note decay. In these recordings one can hear more rhythmic projection, a greater emphasis on the dance character, on meter and pulse. This is achieved primarily through the more lifted bowing; the uneven bow-strokes delineate metric-harmonic groups. The effect can be emulated, to a certain degree, with a modern bow as demonstrated by Lev, Buswell, and more obviously by Zehetmair, Schmid, Tetzlaff , St John, and others (cf. Table 3.3). In the liner notes to her 2008 recording Mullova captured this shift from long, even strokes to a bouncier, metrically orientated bowing eloquently, as cited earlier (chapter three).

  • 69 Bowing observations were made by careful repeated listening and two experienced violinists with ter (...)

92Fingering choices involving lower positions and use of open strings are also more commonly observed in the recordings of specialist and HIP-inspired players.69 This impacts on timbre as lower positions involve more string crossing creating timbrel variety as each string has its own character. In contrast to HIP practice, the MSP style’s focus on melodic unity encourages shifting to remain on the same string and create a unified tone for the melodic phrase. These two opposing aesthetic views contribute to considerable aural differences. However, recording technology cannot be ignored when wondering about timbre. The acoustics of the recording venue, the placement and choice of microphones, as well as post-production editing such as the addition of reverbs can alter the overall sound to such a degree that it is simply not possible to make sound comparative judgements regarding the players’ actual tone. What is possible is to comment on certain local effects within the same recording. These are most likely achieved through bowing and thus reflective of interpretative decisions rather than technological artefact.

93Importantly, by now we have had at least three decades of training and performance with period apparatus, and many violinists use relatively longer baroque bows that enable Italianate cantabile playing as much as French-style “thundering” down-bows. It is therefore interesting to note the many nuanced differences in bowing among HIP players alone. Not surprisingly age seems to matter in this regard. Generally speaking, among HIP violinists Schröder seems to have used the most “conventional” bowing. Beznosiuk has also tended to use longer bow strokes, especially when performing multiple voices.

94To provide some specific information I discuss just one example, the beginning of the G minor Fuga (Audio example 4.14). The differences among HIP versions are quite palpable even though all period violinists play the opening in a detached style. Still, the length of bow strokes varies across players who also often stress different notes or stress the same ones but to a different degree.

4.14. Different bowings within HIP style in J. S. Bach, G minor Sonata BWV 1001, Fuga, extract: bars 1-6. Six versions: Sigiswald Kuijken 1983 © Deutsche Harmonia Mundi, Rachel Podger © Channel Classics, Elizabeth Wallfisch © Hyperion, John Holloway © ECM, Pavlo Beznosiuk © Linn Records, Monica Huggett © Virgin Veritas. Duration: 2.11.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.20

Kuijken uses the shortest, crispiest strokes in both of his recordings until b. 4, but in the thicker textures of bb. 4-6 the strokes sound longer. Luca’s are also short but sounding less staccato. Podger and Huggett play in a detached manner but with somewhat longer strokes. While Podger tends to emphasize, through longer strokes, the third beat of bars, Wallfisch often stresses the first as well as the third beats. Huggett tends to lean on more notes than anybody else (not necessarily in the first six bars). Holloway plays slower than others but his accenting is similar to Beznosiuk’s. They tend to emphasize the first and third crotchet beat of bars. Wallfisch arpeggiates the chords in bars 4-6, the others tend to play the triple stops in a fast bow-stroke.
Bars 35-42 also show differences within a basically detached style in each recording. Huggett and also Matthews always emphasize the main (bottom) note or part, whereas Schröder and Kuijken (2001) change this pattern in b. 38 to bring out the top line. Podger plays the top notes much shorter, with less resonance than the pedal note. Luca’s strokes are relatively longer, Wallfisch’s fairly off the string and Kuijken’s almost spiccato, especially in 1983.

95Variety of bowing and articulation could also lead to major differences in interpretation. The performance solutions of bars 35-42 of the G minor Fuga are strikingly varied because of the ambiguities in Bach’s notations. Table 4.8 lists the main types of delivery. Indicative transcriptions are provided in Figure 4.11. The first thing to note is the lack of obvious preference for particular solutions among HIP or MSP musicians.

96The score has triple stopping between bb. 35-37 which is followed by double stops over a D pedal in bb. 38-41. Most commonly the first two and a half bars are played in an arpeggiated manner while the double stops over the D pedal as written or with playing the D after every double stop (forming semiquaver groups). Other solutions include playing the minims as chords (Kremer 1980, Gähler, Szenthelyi, Luca, Mullova) or arpeggiating the entire section (e.g. Barton Pine, Tetzlaff, and most HIP violinists). Van Dael, Huggett, Holloway and Tognetti introduce a different arpeggiation at bar 38: by re-articulating the top note and repeating the pedal D at the end of each group they create sextuplet patterns slurred in threes.

Table 4.8. Summary of basic differences in executing bars 35-41 of G minor Fuga

Table 4.8. Summary of basic differences in executing bars 35-41 of G minor Fuga

* Crescendo usually starts in the second half of b. 39, or in b. 40; there is a subito p in b. 38 after which loud dynamics resume

Figure 4.11. Bach’s original scoring of bars 35-41 of the G minor Fuga and transcriptions of performed interpretations (cf. Table 4.8)

4.15. Different interpretations in J. S. Bach, G minor Sonata BWV 1001, Fuga extract: bars 35-42. Seven versions: Ingrid Matthews © Centaur, Rachel Podger © Channel Classics, Jaap Schröder © NAXOS, Sergiu Luca © Nonesuch, Elizabeth Wallfisch © Hyperion, Sigiswald Kuijken 1983 © Deutsche Harmonia Mundi, John Holloway © ECM. Duration: 3.09.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.21

Within these basic similarities among groups of violinists there were many important differences. A comparison of Podger’s and Matthews’ recording, for instance, shows Podger to approach the section with virtuosity creating a richly sonorous recording. In contrast, Matthews’ reading has a dreamy quality, especially from bar 38 onwards. Here she drops the dynamics to very soft levels and starts the new arpeggiation pattern slowly and gradually, giving it a tentative feel (cf. first two examples in Audio example 4.15).
Others have also played around with the dynamics of this episode. Quite a few violinists (e.g. Kremer, Tetzlaff) chose to drop the volume to
p or pp in b. 38 either gradually from bar 37 or 36 or quite abruptly so as to coincide with the new figuration in b. 38. A crescendo often ensued from half-way through b. 40. Perlman, in contrast, played the whole passage forte while others started mf and gradually built up the volume to the beginning of b. 42.

97As mentioned earlier, and illustrated through Audio example 4.15, variety also involved matters like stressing certain harmonies, holding over down-beats, accenting every half-bar, playing the pedal note louder or softer than the other notes, adding slurs to the double stops from b. 38 onward, or suddenly slowing and then slightly speeding tempo to mark bars 38 or 42. But of course the most striking difference comes from whether the passage is arpeggiated or not and if so how.

  • 70 Neumann, ‘Some Performance Problems,’ p. 27.

98Neumann states that “Bach invariably writes ‘arpeggio’ or ‘arp’ when he wants chords so treated,” for instance in the D minor Ciaccona (bb. 89-120).70 These bars of the G minor Fuga are not marked like that. If Neumann is right about Bach’s notational practices, it becomes questionable whether the arpeggiation one hears in the recordings has any historical validity. More pertinently, the question arises when such a playing tradition may have started. The recorded history of the works indicates no arpeggiation in this episode until the 1980s. In my collection of over 60 versions, proper arpeggiation is found only in those listed in column 1 of Table 4.8. Except for Sándor Végh (1971), Arthur Grumiaux (1961) and Emil Telmányi (1954), who play the passage as written (i.e. like Luca et al.), all others tend to play it as semiquavers, although some start this at b. 38 only.

  • 71 Ibid.

99Interestingly, a few paragraphs later Neumann acknowledges the existence of passages where “arpeggiation is not ornamental in nature, but necessitated by the technical limitations of the instruments.”71 Given that these measures are played as written by at least eight violinists on record (although Telmányi and Gähler are using a curved bow), it remains arguable what might be considered a “technical necessity.”

Multiple Stops

100Looking at the delivery of multiple stops more generally it is perhaps not surprising to find that MSP violinists tend to play four-note chords as three plus one or two plus two notes, giving emphasis to the melodic pitch, particularly if it is the top note. The alternative among these versions is a rapid delivery, making the notes sound more chord-like. In such cases a noticeable sound quality results, especially in the fugues: the greater bow pressure and stronger attack of older players create a “whipping” effect (Kremer, Shumsky). In more recent recordings of younger violinists the bowing is lighter and the sound less forceful (Mullova, Faust, Barton Pine, St John, Gringolts, Ibragimova). Of course some of these violinists use gut strings and / or a baroque bow which may assist in achieving the effect because of the lower tension in both the strings and the bow hair. The lower bridge of the baroque violin also makes string crossing easier. HIP players (less so Kuijken) and those influenced by historical performance practice research are more likely to arpeggiate three and four-note chords to a varying degree. Tognetti and Tetzlaff, for instance, tend to break chords faster while Huggett and Wallfisch slower. Depending on the musical context, they highlight the bottom, the top, or none of the notes. Finally, in polyphonic textures with the subject in the lowest voice some violinists play the chords from top note down (e.g. A minor Fuga bb. 40-43 and 92-93 in Kremer 2005).

101A good example to look at more closely is the famous opening chord of the G minor Adagio. Its delivery illustrates at least three basic differences in approach: Huggett and Holloway play the opening chord with emphasis on the low G followed by a moderately slow arpeggiation and then a pause on the high G. Wallfisch and others (especially HIP violinists but Zehetmair, Tognetti, Mullova, Ibragimova, etc. as well) also arpeggiate but faster and in one gesture, holding out (or not) the top G at the end. The third type of delivery aims to make it sound like a chord, generally with a fast break between the bottom and top two notes (e.g. Perlman, Shumsky, Buswell, Ehnes, Fischer) and holding out the top dyad. St John’s delivery sounds more like one plus one plus two. Matthews plays the bottom two notes then the Bflat and eventually the top G; these last two notes are delivered in a slowly arpeggiated manner (Audio example 4.16, for further versions see Audio examples 4.1, 5.22, 5.23 and 5.24).

4.16. Performance of multiple stops in J. S. Bach, G minor Sonata BWV 1001, Adagio, extract: bar 1, chord 1. Four versions: James Ehnes © Analekta Fleurs de lys, Ruggiero Ricci © One-Eleven, Ingrid Matthews © Centaur, John Holloway © ECM. Duration: 0.18.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.22

102Turning now to the interaction between bowing, articulation and dynamics in relation to multiple stops, I again consider the Sarabande from the B minor Partita. It provides a good example of how these features combine to create differences within broadly similar interpretative approaches (Table 4.9; Audio example 4.17).

Table 4.9: Summary of modifications in the B minor Sarabande within the two essentially differing approaches

Table 4.9: Summary of modifications in the B minor Sarabande within the two essentially differing approaches

4.17. Increasingly lighter bowing, less legato articulation and stronger pulse in J. S. Bach, B minor Partita BWV 1002, Sarabande, extract: bars 9-17. Eleven versions: Ruggiero Ricci © One-Eleven, James Ehnes © Analekta Fleurs de lys, Sergey Khachatryan © Naïve, Gerard Poulet © Arion, Julia Fischer © PentaTone Classics, Benjamin Schmid © Arte Nova, Richard Tognetti ©ABC Classics, John Holloway © ECM, Lara St John © Ancalagon, Ingrid Matthews © Centaur, Elizabeth Wallfisch © Hyperion. Duration: 5.46.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.23

103Overall, MSP players tend to emphasize the melody through a sustained, over-legato style, where the top notes are held over while the chords of the next note are played softly and unobtrusively (e.g. Ehnes, Khachatryan, Perlman, Shumsky). Fast chords, soft and short lower notes can also be observed in many other versions, including some HIP, where the melody is foregrounded (Matthews, Holloway, Barton Pine). But as one can hear in the Audio examples, there are many subtle variations within this general legato style.

104A real difference is achieved only by adopting a radically different bowing style, one that is more lifted and uneven, resulting in a constant fluctuation of timbre and dynamics. In these versions there are slight swells up to the melody note and then a decay or “reverse swell” before the next multiple stop. This bowing creates a little accent on the lower pitches at the beginning (whether played arpeggio or as fast chord / “sliding-up” to the top note) and thus integrates the harmony more (Huggett, Schmid, Ibragimova, Zehetmair). The gaps between the chords make the melody sound less legato while the fluctuation of the dynamics enhances the sense of pulse. At times there seems to be a slight pushing forward of tempo; just moving on rather than speeding up. As we have seen earlier (Audio example 4.4), Mullova’s three recordings show the trend from the “fast chords, even tone, legato melody, over-held top notes” style of playing to a lighter, more detached articulation, arpeggiated or almost arpeggiated delivery of multiple stops, constantly fluctuating dynamics, and stronger projection of pulse. As was noted earlier, she alternates the rhythm by adding dotting (like Luca) in her 1993 recording and delivers increasingly flexible, expressive, and ornamented versions in 1993 and 2008.

Phrasing and Dynamics

105A telling instance of interacting performance features is the phenomenon that differences in phrasing can be captured by observing the use of dynamics. Terraced dynamics was quite common, especially for repeats or between the A and B materials of binary form movements (e.g. Khachatryan in D minor Giga). Zehetmair also used terraced dynamics but for sections within the A and B parts of these movements. On the other hand, violinists like Faust or St John, would start these movements soft and play them gradually louder, more “defined,” especially for the repeats. Fugues also tended to be structured through dynamics. Apart from large-scale crescendos, violinists, especially MSP musicians, quite commonly played episodes or certain fugal statements softer or louder than the surrounding material (e.g. Shumsky, Kremer, Ehnes in the G minor Fuga). HIP players used less contrastive dynamics; their volume generally stayed within a narrower band than those of older MSP violinists. Dynamics could also reflect the thickness and register of the musical texture with higher or denser material tending to sound louder. The prescribed terraced dynamics in movements such as the A minor Allegro, E Major Preludio, and the gigues were generally observed by all studied performers to a greater or lesser extent probably contributing to a potential sense of uniformity of interpretation in certain movements.

106An aspect of dynamics even more clearly linked to phrasing is the convention of so-called phrase-arching; whether a performer plays louder and faster as they progress towards the middle of the phrase and then softer and slower as they conclude it, and to what extent (cf. Table 3.3).

  • 72 Cook, Beyond the Score, pp. 176-223.

107Nicholas Cook dedicated a whole chapter to this issue in his recent book.72 Stravinsky railed against and ridiculed such performances. In his Harvard Lectures in 1939 he jibbed:

  • 73 Igor Stravinsky, Poetics of Music in the Form of Six Lessons (1939), trans. by Arthur Knodel and In (...)

The sin against the spirit of the work always starts with the sin against its letter and leads to the endless follies […] Thus it follows that a crescendo, as we all know, is always accompanied by a speeding up of movement, while a slowing down never fails to accompany a diminuendo.73

108Such phrasing was commonly noted in the recordings of MSP violinists, especially Shumsky, Ricci, Perlman, Buswell, Mintz, Ehnes, Hahn, Fischer, Khachatryan, Tetzlaff (esp. in 2005) but also in the less radically HIP versions, for instance Beznosiuk and, to a lesser extent, Schröder.

  • 74 Cook, Beyond the Score, p. 156 referring to a study by Luke Windsor et al., ‘A Structurally Guided (...)
  • 75 The metric rubato one often hears in early recordings (when the tempo fluctuates in the melody only (...)

109The most noticeable differences with regards to dynamics and phrasing in the current dataset stem from decisions regarding the length or “identity” of a phrase. This finding is in line with the empirical research that asserts the existence of a combination of tempo curves in expressive performance, “with different weightings at different metrical levels, together with certain individual gestures corresponding to specific structural events.”74 As I have stated repeatedly, older and MSP violinists, especially those who play slow movements rather legato, tend to phrase in longer units while HIP and younger players have shorter arches progressing by single bar, half-bar, or pairs of bars. The difference in affect is profound. For contemporary sensibilities schooled on HIP, the phrase-arch approach sounds “romantic,” especially when combined with vibrato tone, slower tempo, heavy bow pressure and legato articulation. By contrast, the more rapid, quasi chiaroscuro ebb and flow in performances that articulate shorter musical gestures sound “rhetorical,” as if somebody is speaking or presenting an argument. The former may come across as sentimental, a pleading to and manipulation of emotions; or perhaps aggressive, forceful and demanding. The latter may strike as a genuine first-person story-telling, impassioned and authentic, or intimate and self-orientated. It does not force a reaction; it simply conveys a compelling account.75

110Focusing on musico-technical elements contributing to these potential affective differences, recordings of the Largo from the C Major Sonata may serve us well to illustrate the point. The readings of Shumsky, Poulet, Mintz, Perlman and Hahn, among others, exemplify the legato approach where cadence points are subdued into ever continuing melodic lines. Tempo is relatively steady but the dynamics definitely have an arching profile, gradually building with each subsequent sub-phrase to climactic high notes and dropping only towards major structural moments, like b. 8 and b. 16 (Figure 4.12 bottom two panels). In contrast, Wallfisch, Zehetmair, Huggett and others articulate almost every harmonic figuration and make audible many temporary cadence points (Figure 4.12 top 3 panels; Audio example 4.18).

111Violinists in the first group “bow-over” rests; e.g. bars 1-2, where the melody line is basically sustained in one continuous flow, in spite of the rests between its segments. In contrast, Wallfisch, Zehetmair and Huggett seem to relish the silences and create pauses even where there is none notated. In their reading, the four groups of semiquavers in bar 3 each has its own little dynamic and tempo arch and there is a little rallentando and pause on beat 3 of the next bar (cadence marked with trill in the manuscript). Typically they separate the sequential repetitions of bars 4-5 and create another mini cadence on the down-beat of bar 6. The next sequential repetition (bars 6-7) is again separately articulated and the second statement is not linked to the four semiquavers on the second beat of bar 7.

4.18. Articulated / “rhetorical” versus legato / “romantic” style in J. S. Bach, C Major Sonata BWV 1005, Largo, extract: bars 3-8. Five versions: Elizabeth Wallfisch © Hyperion, Monica Huggett © Virgin Veritas, Thomas Zehetmair © Teldec, Gerard Poulet © Arion, Oscar Shumsky © Musical Heritage Society. Duration: 4.14.
To listen to this extract online scan the QR code or follow this link:
http://dx.doi.org/​10.11647/​OBP.0064.24

Figure 4.12. Comparison of tempo and power (dynamics) in bars 3-8 of the Largo from the C Major Sonata performed by (from top to bottom) Wallfisch, Huggett, Zehetmair, Poulet and Shumsky. Beat level tempo is indicated by the smoother line near bar and beat numbers. Power (dynamic) is indicated by the volatile slopes. Note the longer and clearer power arches in the bottom two panels as opposed to the near constant shifts in the more locally articulated versions above. Note also Shumsky’s fairly even tempo line. I used Sonic Visualiser to prepare this analysis

112To various degrees, Wallfisch and the others also delineate the elision of the quavers cadence on beat 1 of bar 8 and the new start of the opening melody—the four semiquavers on the same beat. What is important to note, too, is that Wallfisch and others (e.g. Ibragimova) manage to keep the line going even though their articulation is very detailed and their phrasing locally nuanced. They manage to avoid the pitfall of creating a fragmented and over-accented reading that lacks flow.

113Compared to this locally articulated phrasing, violinists listed in the first group (i.e. Shumsky et al.) tend to create one large phrase from the beginning to bar 8, even though some of them play the first two bars more separated than the others. Their over-legato approach geared towards a long-spun melody becomes most obvious from bar 4 onwards. One hears their performances as pushing ahead, using the momentum of the pairs of sequences. Continuously increasing intensity, vibrato and volume, they reach a climax on the high A of bar 7 where they slow down to emphasize the trilled cadence resolving on beat 1 of bar 8. They foster the perception of released tension by softening the dynamics.

4.6. Conclusions

114In this chapter I continued to examine the similarities and differences between HIP and MSP versions taking a more detailed approach focusing on specific performance features. I wanted to see the degree of homogeneity present across these recordings and whether particular characteristics could be linked to time and place or violin school. Most importantly my aim was to gain a textured understanding of the complex nature, the “thousand layers” of music performance; the interactions of technical and musical, physical, acoustic and bodily features contributing to the process of emergent stylistic transformations. As this is the overarching aim of the entire book, I will draw only partial conclusions here. Additional information will come to light in the remaining chapters.

115The analyses of performance features confirmed the more general points made in the previous chapter: MSP violinists, especially of the older generation and those associated with the Juilliard School tended to use similar bowing and more vibrato than HIP and HIP-inspired violinists. Beyond this very broad generalization however, it was difficult to categorize players and performances. The most important overlap between MSP and HIP was found in bowing. The overlap reflected generational differences: older HIP players (especially Schröder) differed less from MSP violinists in their bowing while younger MSP players (except for Hahn) tended to emulate to a greater or lesser degree bowing typical of period practice. Investigation of tempo showed a greater alliance among HIP-inspired violinists with MSP customs in that they too tended to choose more extreme tempos than HIP violinists. On the other hand, ornamentation proved to be a strong indicator of HIP influence, with several MSP players considerably outperforming their period specialist colleagues in embellishing and ornamenting many movements of the Six Solos, especially since the mid-2000s.

116Once the discussion moved into accounting for features such as rhythm, timing, articulation, dynamics and phrasing, the interaction of performance elements became palpable and made the previous attempts at categorizing styles rather futile. The masking effect of articulation and tempo on the perception of dotting was re-confirmed; the complex and interdependent impact of bow-strokes, articulation and dynamic nuance together with its contribution to diverse aesthetic affect was repeatedly noted. The detailed descriptions endeavoured to convey in words the myriads of audible differences across the selected recordings and explain what may constitute the complex of these differences; they endeavoured to identify “lines of flight” and “multiplicities” inhabiting the “rhizome” of solo violin Bach performance at the turn of the twentieth and twenty-first centuries.

  • 76 I am aware of only one or two recordings of the complete set made prior to 1950 (e.g. Yehudi Menuhi (...)
  • 77 Richard Turner, ‘Style and Tradition in String Quartet Performance: A Study of 32 Recordings of Bee (...)
  • 78 Bruno Repp, ‘Diversity and Commonality in Music Performance: An Analysis of Timing Microstructure i (...)
  • 79 Performance (mostly tempo) trends in symphonic repertoire have been studied by Bowen, among others. (...)

117In the face of the evidence one wonders why the notion of homogeneity in later twentieth-century performance is such a pervasive impression. Is it because there are so many more recordings available now that it is harder to pick out the salient ones compared to the “golden age” when only a handful of violinists made recordings and rarely of solo Bach?76 Or could it be that solo performance, especially string and voice, tend to be more individualistic than piano, orchestral, or ensemble playing, and that these may have been the basis for drawing conclusions regarding homogeneity? Or is it simply the performance of baroque music, including Bach’s that has changed so much providing opportunity for such diversity? These are realistic scenarios for potentially different conclusions, but research into string quartet practice also seems to point towards diversity77 and important variations have been noted among pianists performing nineteenth-century music.78 In my experience as intermittent listener to concertos (for instance Mozart’s violin concertos), diversity seem harder to evidence in that repertoire whether one focuses on the soloists alone or the total experience. But in orchestral music I hear considerable differences between the interpretations of various conductors, like Simon Rattle, Claudio Abbado, Philippe Herreweghe or Nikolaus Harnoncourt, in repertoire as diverse as Berlioz, Mahler, Beethoven, Mendelssohn, and Brahms.79 It could be that writers all too readily base their opinion on formative listening experiences during earlier times, perhaps ending by the 1980s. The lack of systematic attention to and analyses of current performances has allowed the casual impression to be reinforced by commercial propaganda or reviewers’ choices that often favour the circulation of particular versions, and not always the most interesting ones. The striking differences between two almost concurrently released recent versions of the Bach Solos (Ibragimova and Faust, both recorded in 2009) should really make commentators rethink the validity of claiming uniformity in contemporary music performance. Whereas Ibragimova uses extreme tempos (very fast allegros and rather languid slow movements), adds no embellishments, plays with no vibrato and fairly softly almost all the time, Faust provides a varied, energetic reading full of bounce and a multitude of ornaments and embellishments. To me the first sounds stylish but “calculated”; the latter spontaneous and more “naturally” engaged and engaging. Expanding the variety, just a year later Khachatryan’s set was issued, one that diverts only very lightly from the standard MSP approach, most typical of the 1970s and 1980s.

  • 80 Leech-Wilkinson, The Changing Sound of Music.
  • 81 Nicholas Cook, ‘Bridging the Unbridgeable? Empirical Musicology and Interdisciplinary Performance S (...)

118It is of course reasonable to posit that a notated musical composition defines the boundaries of its possible execution to a greater or lesser extent, contributing to potential homogeneity of performances. After all, we are in the habit of saying “performance of this or that piece” for some reason! The expression implies that a performance (or recording) is supposed to be a representation of the work. As discussed in chapter two, representational thought is restrictive because it requires the conceptualization of a piece of music to be a “faithful” copy of a supposed “original.” Exposure to a hundred years of sound recordings has started to make us aware of how loose these boundaries might be and musicologists and theorists like Leech-Wilkinson80 and Cook81 have started unpacking the issue of representation and the nature of creativity in performance. Importantly, Deleuzian philosophy, to which I will refer increasingly in the final part of this book, encourages thinking of music as something without a stable entity; “music as process of becoming.”

119Nevertheless the perceived or very real boundaries conveyed by the score are upheld by our current training and associated value systems to a certain degree (quite a high degree in some cases as witnessed in the playing of Ehnes, Edinger, and Hahn, for instance). The question then is: how many radically different readings one can get of a piece? Or rather, to what extent readings must differ from each-other before one starts talking about homogeneity in performance? Only when we have many more detailed studies available of diverse repertoires, artists and periods shall we be in the position to answer these questions. For now, it seems cautionary to note the variety and beauty that the studied violinists bring to their interpretations, deepening and broadening our understanding of these seminal pieces composed by Bach.

120The final point then is not to deny the existence of trends, but to emphasize their limited scope in helping to map twentieth-century developments of western music performance. To put it another way, analysis of individual interpretations seems more productive for an understanding of music as performance and creativity in performance, while establishing and examining trends maybe useful when thinking of music as culturally embedded communication.

Notes

1 The “breathless tempi of much early music” is an opinion generally expressed rather casually, as in Bernard D. Sherman (ed.), Inside Early Music: Conversations with Performers (New York: Oxford University Press, 1997), p. 292. Taruskin noted both the acceleration of tempo choices over time and some slower than historically documented tempos chosen by HIP performers, although in relation to orchestral or ensemble music (Richard Taruskin, Text and Act: Essays on Music and Performance (New York: Oxford University Press, 1995), pp. 134-135, 214-217, 293-294, 232, etc.)). From his discussions it transpires what other, less often cited writers also note, namely that some early twentieth-century musicians often chose much faster tempos than what we are accustomed to today. See Vera Schwarz, ‘Aufführungspraxis als Forschungsgegenstand,’ Österreichische Musikzeitschrift, 27/6 (1972), 314-322. Among the 55 versions of the E Major Preludio in my collection the fastest was recorded in 1904. The performer is nineteenth-century virtuoso Pablo Sarasate. Just how much faster than anybody else he plays is indicated by his Standard Deviation score: 3.27 (see fn. 5 for an explanation of Standard Deviation values). Sarasate’s E Major Preludio recording remains the fastest among Dario Sarlo’s larger sample set as well (Dario Sarlo, The Performance Style of Jascha Heifetz (Farnham: Ashgate, 2015)). In Sarlo’s collection the next fastest is by Deych (1999) and then Brooks (2001), followed by Szigeti in 1908 and Wallfisch in 1997. Another nineteenth-century violinist, Eugène Ysaÿe is also noted for his incredibly speedy interpretation of the third movement of Mendelssohn’s Violin Concerto. See Dorottya Fabian, ‘The Recordings of Joachim, Ysaÿe and Sarasate in Light of their Reception by Nineteenth-Century British Critics,’ International Review of the Aesthetics and Sociology of Music, 37/2 (2006), 189-211.

2 Dario Sarlo studies over 130 recordings of the E Major Preludio in his forthcoming book, The Performance Style of Jascha Heifetz (Farnham: Ashgate, 2015) and refers to James Creighton’s Discopaedia of the Violin, 1889-1971 (Toronto: University of Toronto Press, 1974). This indicates that “by 1971 there were at least 320 complete and partial recordings from the solo works” (Sarlo, Heifetz, 2015, chapter eight).

3 Dorottya Fabian, Bach Performance Practice, 1945-1975: A Comprehensive Review of Sound Recordings and Literature (Aldershot: Ashgate, 2003), pp. 97-124. Another issue is the perception of tempo. As Harnoncourt pointed out many years ago, well-articulated music will always sound faster. Nikolaus Harnoncourt, Baroque Music Today: Music as Speech (Portland, Oregon: Amadeus Press, 1988), p. 52. Musik als Klangrede, trans. by M. O’Neill (Salzburg: Residenz, 1982).

4 R-squared is a statistical measure of how close the data points are to the fitted regression line. It is a measure of variance explained. See Timothy C. Urdan, Statistics in Plain English (New Jersey: Lawrence Erlbaum Associates Inc., 2005), p. 155. I calculated a simple linear regression of tempo choices over time.

5 The 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. I did not include a tabulation of SD values here but the information can be provided upon requests.

6 Key sources for an understanding of the role of meter and its relationship to tempo are Johann Kirnberger’s treatise, Die Kunst des reinen Satzes in der Musik (The Art of Strict Composition in Music, 1774, especially Book II, part IV, pp. 105-153) and George Houle, Meter in Music, 1600-1800 (Bloomington: Indiana University Press, 1987). The tempo of dance movements are discussed in many recent books on baroque music, including David Ledbetter, Unaccompanied Bach: Performing the Solo Works (New Haven: Yale University Press, 2009) and Jaap Schröder, Bach’s Solo Violin Works: A Performer’s Guide (London: Yale University Press, 2007).

7 The New Bach Reader—A Life of Johann Sebastian Bach in Letters and Documents, ed. by Christoph Wolff (London and New York: Norton, 1998), p. 436.

8 Schröder, Bach’s Solo Violin Works, p. 168.

9 Most violin tutors discuss vibrato at length. Many modern studies could also be mentioned that provide a historical context for the aesthetic ideals behind its use and discussion of individual characteristics. The reader is pointed to three key publications that report, among others, vibrato in recorded violin performances: David Milsom, Theory and Practice In Late Nineteenth-Century Violin Performance: An Examination of Style in Performance, 1850-1900 (Aldershot: Ashgate, 2003); Mark Katz, Capturing Sound: How Technology has Changed Music (Berkeley and Los Angeles: University of California Press, 2004) and Daniel Leech-Wilkinson, The Changing Sound of Music: Approaches to Studying Recorded Musical Performance (London: CHARM, 2009).

10 To the extent something jarring may create negative emotions, this is corroborated by Juslin when he posits that “[a] possible effect of emotion on aesthetic processing is that positive emotions can lead a listener to process the music more ‘holistically,’ whereas negative emotions will him or her to process it more ‘analytically.’” Patrik N. Juslin, ‘From Everyday Emotions to Aesthetic Emotions: Towards a unified theory of musical emotions,’ Physics of Life Reviews, 10 (2013), 235-266 (pp. 256-257). I thank Daniel Bangert for alerting me to this recent paper by Juslin.

11 Vibrato measurements were taken on thirty-eight selected notes in the A minor Sonata’s Andante (number of violinists = 29) and the D minor Partita’s Sarabanda (n = 9) movements or the E Major Partita’s Loure (n = 9) when there was no recording of the D minor Partita by a given violinist. Frequency of vibrato use was calculated by noting how many of these notes were played with vibrato. The software program Spectrogram 14 was used to measure vibrato. The program displayed a spectrogram of the audio file using the following parameters: High band Hz: 4200; Window size / Display width: 4 sec; Frequency resolution: 18.7 Hz; Colour spectrum range: 0 to -80 dB. To calculate vibrato speed / rate the number of undulations were counted relative to the note duration in milliseconds. Vibrato depth / width was calculated from the frequency (Hz) difference between the inner bottom and top edges of undulations (as read off the screen by manually placing the pointer to the relevant position). All depth measurements were carried out twice to check for possible errors and adjust if necessary.

12 Richard Miller, The Structure of Singing: System and Art in Vocal Technique (London and New York: Schirmer Books, 1986). See also Leech-Wilkinson, The Changing Sound of Music.

13 Beethoven Violin Sonatas with Sophie Mutter and Lambert Orkis on Deutsche Grammopon 457 623-2 (1998); Viktoria Mullova and Christian Bezuidenhout on Onyx 4050 (2010); Rachel Barton Pine and Matthew Hagle in concert on Chicago’s WFMT Radio (2005); Isabelle Faust and Alexander Melnikov on Harmonia Mundi 902025-27 (2009); Daniel Sepec and Andreas Staier on Harmonia Mundi 901919 (2006). However, a cursory listening to about ten recordings of Mozart’s Violin Concerto No. 5 made since the 1980s did not show much diversity in tone (or in other features, for that matter). For instance, the nominally HIP version of Andrew Manze (Harmonia Mundi 2006 HMU 807385) has a fairly vibrato-less tone while Hilary Hahn’s 2015 recording (Deutsche Grammophon 0289 479 3956 6 CD DDD GH) displays all the hallmarks of MSP and evenly regulated, continuous vibrato.

14 Joseph Szigeti, Szigeti on the Violin (New York: Dover Books, 1979), especially pp. 38-43.

15 Katz, Capturing Sound, pp. 85-98, esp. 93, 95-96.

16 Basic and detailed information on ornamentation practices, including the differences between national styles are readily and abundantly available in modern publications. The best known are Arnold Dolmetsch, The Interpretation of the Music of the Seventeenth and Eighteenth Centuries (London: Novello, 1949/1915 [R1969]); Robert Donington, The Interpretation of Early Music (London: Faber, 1963 [rev. edn 1989]); Putnam Aldrich, Ornamentation in Bach’s Organ Works (New York: Coleman-Ross, 1950), Frederick Neumann, Ornamentation in Baroque and Post-Baroque Music (Princeton: Princeton University Press, 1978). Therefore, I restrict discussion to issues that are relevant to my project at hand.

17 Bach himself prepared such a table for his eldest son (see “Explication unterschiedlicher Zeichen, so gewisse Manieren artig zu spielen, andeuten” in Clavierbüchlein für Wilhelm Friedemann Bach, ca. 1720; manuscript in School of Music Library, Yale University) that has often been reproduced and transcribed. Examples of figurative embellishments can be found, for instance among Handel’s scores (performance copies he created for some of his singers) or, famously, Roger’s 1710 Amsterdam edition of Corelli’s violin sonatas Op. 5. In my discussion, I aim to distinguish between the French practice of agréments and the Italian melodic embellishment. The former comprises short grace notes either notated in small font or indicated by symbols, such as trills, mordents, appoggiaturas, slides and turns. Melodic embellishments are basically diminutions or divisions where written notes are broken up into smaller denominations, providing figuration on the melody and harmony. To confuse matters further, Bach often uses a combination of signs and normal font in his notation practices, although he often omits the sign itself while still notating out a part of the grace. A common example is a long note followed by two short notes in a suffix melodic shape. Even without the trill sign on the long note, the two short notes may in effect be the written-out ending of a trill. Similarly, appoggiaturas are often written out, masking the context and erroneously inviting added appoggiaturas (for starting a trill, for instance). I have shown such examples in my paper ‘Ornamentation in Recent Recordings of J. S. Bach’s Solo Sonatas and Partitas for Violin,’ Min-Ad: Israel Studies in Musicology Online, 11/2 (2013), 1-21, available at http://www.biu.ac.il/hu/mu/min-ad/. Literature on ornamentation in Bach’s music is vast. For a comprehensive review of such specialized literature see Fabian, Bach Performance Practice, chapter five.

18 Pier Francesco Tosi, Opinioni de cantori antichi sopra il canto figurato (1723), cited in Frederick Neumann, Performance Practices of the Seventeenth and Eighteenth Centuries (New York, Schirmer 1993), p. 521.

19 Johann-Joachim Quantz, On Playing the Flute [Versuch], trans. by E.R. Reilly (New York: Schirmer, 1975/1752), chapter eighteen, par. 58.

20 Cited in Neumann, Performance Practices, p. 521.

21 Ibid., p. 538, citing Quantz, Versuch, chapter thirteen, par. 9.

22 Neumann, Performance Practices, p. 529.

23 Butt, Bach Interpretation: Articulation Marks in Primary Sources (Cambridge: Cambridge University Press, 1990), pp. 207-208.

24 For instance, compare Birnbaum’s text (published in The New Bach Reader, pp. 338-348) with Giovanni Bononcini’s complaint in the Preface to the publication of his Sonate da Chiesa a Due Violini (Venice: [n.p.], 1672): “because today there are some [performers] so little informed of that art of tasteful embellishment that in singing or in playing they want with their disorderly and indiscreet extravagances of bow or of voice to change, indeed to deform, the compositions (even though these were written with every care and conscientiousness) in such a manner that the authors have no choice but to beg those singers and players to content themselves with rendering the works plainly and purely as they are written” (cited in Neumann, Performance Practices, 570). François Couperin (Preface to Pièces de Clavecin, Book 3) is also on record demanding performers to be faithful to his notation whereas Michel de Saint Lambert, like Scheibe, defended the rights of the performers “to add new ornaments [and] leave out those that are prescribed [or] to substitute others in their stead” (Principes du Clavecin [1702], p. 57, cited in Neumann, Performance Practices, p. 514).

25 The New Bach Reader, p. 338.

26 The New Bach Reader, pp. 346-347.

27 Arthur Mendel (ed.), Bach: St John Passion—Vocal Score (New York: Schirmer, 1951), xxii.

28 Noted also by Jaap Schröder in ‘Jaap Schröder Discusses Bach’s Works for Unaccompanied Violin,’ Journal of the Violin Society of America, 3/3 (Summer 1977), 7-32; and Ledbetter, Unaccompanied Bach, p. 95ff.

29 Joel Lester, Bach’s Works for Solo Violin: Style, Structure, Performance (Oxford-New York: Oxford University Press, 1999), p. 38.

30 For trends in earlier recordings, see my ‘Towards a Performance History of Bach’s Sonatas and Partitas for Solo Violin: Preliminary Investigations,’ in Essays in Honor of László Somfai, ed. by László Vikárius and Vera Lampert (Lanham, Maryland: Scarecrow Press, 2005), pp. 87-108. For a short summary of how the opening chord is played by various violinists in the current selection of recordings see section “Multiple Stops” later in this chapter.

31 Michael Spitzer and Eduardo Coutinho, ‘The Effect 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. I will discuss this movement and diverse affects generated by different interpretations in chapter five.

32 Neumann, Performance Practices, p. 528. I would not even use the word “legitimate” since the range is so broad the word loses its meaning, if it is a useful concept / adjective at all.

33 Neumann, Performance Practices, p. 536.

34 Quantz, Versuch, chap. 18, par. 76, as paraphrased from Neumann, Performance Practices, p. 536.

35 A striking exception to this generalisation is Stefano Montanari’s 2011 recording of the works (Amadeus Elite Paragon, 2012 DDD AMS 108/109-2 SIAE) that I first heard when revising this chapter. He embellishes practically every movement, including the opening Adagios, fast finales and fugues, and not just sparingly.

36 Fabian, ‘Ornamentation in Recent Recordings.’

37 Robert Maxham, ‘Bach Violin Partitas: No. 1 in b; No. 3 in E. Solo Violin Sonata No. 2 in a. Ilya Gringolts (vn). Deutsche Grammophon B0000315-02 (58:56),’ Fanfare, 27/2 (November 2003), 111-112.

38 Joseph Magil, ‘Bach Solo Violin Partitas 1+3, Sonata 2. Ilya Gringolts DG 315,’ American Record Guide, 66/6 (Nov 2003), 75-76.

39 Maxham does when he refers to the “jazzy ornamentation” of the Tempo di Borea Double.

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

41 Ideally one should have a tie or slur between the two high Bs at either ends of bars 9-10 in the transcriptions and place the ornament over the bar line under this tie / slur, as per dotted curve.

42 J[ohn] D[uarte], ‘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.

43 In 1983 Kuijken also adds a trill with appoggiatura on the downbeat of bar 12 and a trill on the high G in bar 16. In the 2001 recording he graces the second beat (B of the upper voice) in bar 10, does the same in bar 12 as in 1983. The trill in bar 16 is not properly executed, sounds more like just a fast lower appoggiatura.

44 Neumann, Ornamentation in Baroque and Post-Baroque Music, pp. 519-520.

45 Frederick Neumann, ‘Some Performance Problems in Bach’s Unaccompanied Violin and Cello Works,’ in Eighteenth-Century Music in Theory and Practice: Essays in Honour of Alfred Mann, ed. by Mary Parker (Stuyvesant, NY: Pendragon Press, 1994), pp. 19-48 (pp. 29-30).

46 Greta Moens-Haenen, ‘Vibrato,’ Grove Music Online. Oxford Music Online.

47 Robin Stowell, The Early Violin and Viola (Cambridge: Cambridge University Press, 2001), p. 66.

48 Neumann, ‘Some Performance Problems,’ p. 29.

49 Ledbetter, Unaccompanied Bach, p. 121.

50 This has started to change as evidenced in courses and masterclasses in improvisation for classical musicians. David Dolan et al., ‘The Improvisatory Approach to Classical Music Performance: An Empirical Investigation into its Characteristics and Impact,’ Music Performance Research, 6 (2013), 1-38. See also Aaron L. Berkowitz, The Improvising Mind: Cognition and Creativity in the Musical Moment (New York: Oxford University Press, 2010).

51 Timothy Day, A Century of Recorded Music: Listening to Musical History (New Heaven and London: Yale University Press), p. 158 citing Christopher Hogwood in James Badel, ‘On Record: Christopher Hogwood,’ Fanfare, 9/2 (November-December 2005), 90.

52 I am aware that this could be a result of post-performance editing (patching up two different play-throughs where the embellishments may have been the same in repeats but different in the two takes). However, I was actually present when these recordings were made and I remember that Tognetti did vary the graces he was adding during first plays and in the repeats.

53 As recalled in Neumann, Performance Practices, p. 538.

54 Tellingly reviewers are ambivalent although tending towards the positive: Maxham (‘Ilya Gringolts,’ 111-112) states “Almost everybody will find something to choke on—but only for a few seconds”; and concludes by saying “it could easily turn out to be either the best or the worst recording of the piece ever made.” David Denton on the other hand finds the quality of playing “quite superb” but a “harsh element introduced in the E Major Preludio and the rather cut up of the following Loure” are “not welcome.” In this very short review Gringolts’ presentation of the Gavotte is acknowledged to be “humorous” and the two Menuets “charming.” (David Denton, ‘Bach Partitas no. 1 in B minor BWV 1002 and no. 3 in E Major BWV1006, Sonata no. 2 in A minor BWV1003. Ilya Gringolts (violin) Deutsche Grammophon 474-235-2,’ Strad, 114/1363 (November 2003), 1269.

55 Richard Taruskin, ‘On Letting the Music Speak for Itself: Some Reflection on Musicology and Performance,’ Journal of Musicology, 1/3 (July 1982), 101-117.

56 Houle, Meter in Music.

57 The literature on rhythmic alteration in baroque music is too large to be cited here. Comprehensive recent summaries include Stephen Hefling, Rhythmic Alteration in Seventeenth- and Eighteenth-Century Music (New York: Schirmer, 1993) and Neumann, Performance Practices. A detailed review of this literature is provided in Fabian, Bach Performance Practice, pp. 169-185.

58 For instance in relevant movements of the two Passions, the Brandenburg Concertos and the Goldberg Variations as reported in Fabian, Bach Performance Practice, pp. 170-179 and Dorottya Fabian and Emery Schubert, A New Perspective on the Performance of Dotted Rhythms,’ Early Music, 38/4 (2010, November), 585-588.

59 The dotting ratio in this discussion is expressed as percentage of the duration of the dyad. In other words a literal (or mechanical) dotting is achieved by a ratio of 0.75:0.25 between the dotted and the short note. I calculated these ratios from note on-sets and off-sets. Note on-sets and off-sets were identified by aural and visual inspection of audio signals using spectrographic displays in the software programs Adobe Audition 1 and Sonic Visualiser 3.2 (“Nevermore spectral transform” option of the “Mazurka” plug-in). The window size for visualization in Adobe Audition 1 was between 0.5 and 2.7 milliseconds which provided magnification that countered the possibility of errors occurring within the range of human perception. (Roughly speaking the threshold is 20 milliseconds [mS], but below 200 mS sounds can be lost depending on their level of intensity; see Stanley Gelfand, Hearing: An Introduction to Psychological and Physiological Acoustics, 2nd edn (New York and Basel: Marcel Dekker, 1990)). In the D minor Corrente note on-sets and offsets were identified in bars 3-4 and 6. In the B minor Allemanda the first 7 bars were studied but only dotted dyads were measured (i.e. dotted figures where the short note was replaced by a series of shorter values were not included). In the C Major Adagio bars 1-4 and 6 were studied (there are no dotted rhythms in bar 5). The identification of note on-sets and note off-sets in each studied recording was prepared twice, once each by the author and a research assistant to ensure accuracy. All marking points were confirmed aurally through the clicking device in Sonic Visualiser. In particularly difficult-to-hear recordings the audio files were slowed down to half speed (e.g. Huggett, B minor Allemanda). Sonic Visualiser is a free computer program for audio analysis, see http://www.sonicvisualiser.org/ or Cannam, Chris, Landone, Christian, and Sandler, Mark, ‘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.

60 Data is reported for the 1930s-1950s period in Dorottya Fabian and Eitan Ornoy, ‘Identity in violin playing on records: interpretation profiles in recordings of solo Bach by early twentieth-century violinists,’ Performance Practice Review, 14 (2009), 1-40 (p.22).

61 Samuel Applebaum, The Way they Play, Book 4 (Neptune City, N.J., Paganiniana Publications, 1975), p. 261. I thank Daniel Bangert for this reference.

62 Neumann, ‘Some Performance Problems,’ p. 25.

63 Fabian and Schubert, ‘A New Perspective.’

64 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, 177-203 (p. 198).

65 Nicholas Cook, Beyond the Score: Music as Performance (New York: Oxford University Press, 2014), p. 144.

66 Research into jazz swing has shown that the ratio reduces as the tempo increases (i.e. the length of the dotted note becomes shorter with respect to the beat with increased tempo) until it plateaus around 200 beats per minute to a more or less even 0.5:0.5. Anders Friberg and Andreas Sundström, ‘Swing Ratios and Ensemble Timing in Jazz Performance: Evidence for a Common Rhythmic Pattern,’ Music Perception, 19/3 (2002), 333-349.

67 The term “kerning” was introduced by Schubert and Fabian to denote the gap or silence between notes. See Emery Schubert and Dorottya Fabian, ‘Perception and Preference of Dotting in 6/8 patterns,’ Journal of Music Perception and Cognition, 7/2 (2001), 113-132.

68 Schmid performance shows some similarity but he seems to just stop the bow after the downbeat and then accent the first note of the triplets.

69 Bowing observations were made by careful repeated listening and two experienced violinists with tertiary music degrees acted as research assistants to aid the interpretation of the aural analyses. Fingering was not analyzed in detail. The use of avoidable open strings was interpreted as evidence for fingering choices favoring lower positions, as was common during the baroque period. Comments regarding dynamics are also based on aural analyses. References to various dynamic levels should be understood as relative to the specific recording under discussion or occasionally in comparison with other versions. In this latter case the amplitudes of the recordings were normalized for the purpose of comparison.

70 Neumann, ‘Some Performance Problems,’ p. 27.

71 Ibid.

72 Cook, Beyond the Score, pp. 176-223.

73 Igor Stravinsky, Poetics of Music in the Form of Six Lessons (1939), trans. by Arthur Knodel and Ingolf Dahl (New York: Knopf, 1960), p. 128.

74 Cook, Beyond the Score, p. 156 referring to a study by Luke Windsor et al., ‘A Structurally Guided Method for the Decomposition of Expression in Music Performance,’ Journal of the Acoustical Society of America, 119 (2006), 1182-1193.

75 The metric rubato one often hears in early recordings (when the tempo fluctuates in the melody only and at the bar level, with frequent ritenutos but few longer accelerandos and rallentandos) shares certain similarities with the “rhetorical” HIP approach. However a detailed comparative account is beyond the scope of the current study. For a discussion of metric rubato and its manifestation in recordings of artists born and trained well within the nineteenth century see Richard Hudson, Stolen Time: The History of Tempo Rubato (Oxford: Clarendon Press, 1994); Leech-Wilkinson, The Changing Sound of Music or; Neal Peres Da Costa, Off the Record: Performing Practices in Romantic Piano Playing (New York: Oxford University Press, 2012); and Dorottya Fabian, ‘Commercial Sound Recordings and Trends in Expressive Music Performance: Why Should Experimental Researchers Pay Attention?,’ in Expressiveness in Music Performance: Empirical Approaches across Styles and Cultures, ed. by Dorottya Fabian, Renee Timmers, and Emery Schubert (Oxford: Oxford University Press), pp. 58-79.

76 I am aware of only one or two recordings of the complete set made prior to 1950 (e.g. Yehudi Menuhin 1934-1936; George Enescu late 1940s or early 1950s), a few complete sonatas (e.g. Adolf Busch, Joseph Szigeti, Jascha Heifetz) and the D minor Partita (e.g. Borislav Huberman 1942, Adolf Busch 1929). However, most of the recordings from that period were of individual movements. James Creighton, Discopaedia of the Violin, 1889-1971 (Toronto: University of Toronto Press, 1974) lists more than what I have but does not provide dates for the recordings.

77 Richard Turner, ‘Style and Tradition in String Quartet Performance: A Study of 32 Recordings of Beethoven’s Op. 131 Quartet’ (PhD Thesis Sheffield, University of Sheffield, 2004).

78 Bruno Repp, ‘Diversity and Commonality in Music Performance: An Analysis of Timing Microstructure in Schumann’s “Träumerei’’,’ Journal of the Acoustical Society of America, 92/5 (1992), 2546-2568; Nicholas Cook, ‘Methods for Analysing Recordings,’ in The Cambridge Companion to Recorded Music, ed. by Nicholas Cook et al. (Cambridge: Cambridge University Press, 2009), pp. 221-245.

79 Performance (mostly tempo) trends in symphonic repertoire have been studied by Bowen, among others. José A. Bowen, ‘Tempo, Duration, and Flexibility: Techniques in the Analysis of Performance,’ Journal of Musicological Research, 16/2 (1996), 111–156.

80 Leech-Wilkinson, The Changing Sound of Music.

81 Nicholas Cook, ‘Bridging the Unbridgeable? Empirical Musicology and Interdisciplinary Performance Studies,’ in Taking it to the Bridge: Music as Performance, ed. by Nicholas Cook and Richard Pettengill (Ann Arbor: Michigan University Press, 2013), pp. 70-85.

List of illustrations

Title Table 4.1. Summary of tempo trends 1977-2010 (For change to be noted R2 = >0.001)4
URL http://books.openedition.org/obp/docannexe/image/1860/img-1.jpg
File image/jpeg, 212k
Title Table 4.2. Average MSP and HIP tempos across all studied recordings made since 1903 (Joachim). Violinists who were added to HIP are: Zehetmair, Tetzlaff (both), Tognetti, St John, Barton Pine, Gringolts, Mullova (1993, 2008), Faust, and Ibragimova
URL http://books.openedition.org/obp/docannexe/image/1860/img-2.jpg
File image/jpeg, 408k
Title Table 4.3: Average tempos in recordings made pre and post 1978 (before or after Luca). MSP and HIP versions are combined
URL http://books.openedition.org/obp/docannexe/image/1860/img-3.jpg
File image/jpeg, 252k
Caption Figure 4.1a-f. Spectrograms of bars 5-6 of the D minor Sarabanda performed by (a) Hahn, (b) Shumsky, (c) Podger, (d) Tognetti, (e) Ibragimova and (f) Mullova in 2008. [Spectrogram parameters: High band Hz: 4800; Window size / Display width: 6 sec; Colour Spectrum Level Minimum: -80dB; Frequency resolution: 12.5 Hz.]
URL http://books.openedition.org/obp/docannexe/image/1860/img-4.jpg
File image/jpeg, 720k
Title Table 4.4. Vibrato rate, width (depth) and frequency of use measured on selected notes in different movements and averaged across each selected violinist. (Rate is expressed in cycles per second (cps); Width in semitones (sT). Frequency refers to occurrence of vibrato on the selected pitches). Standard Deviation (SD) indicates the evenness of each player’s vibrato (the smaller the number, the more regular the vibrato)
Caption † Brooks’ vibrato is often very fast and extremely narrow; hard to measure at all let alone accurately. Therefore the frequency percent is an estimate based on where rate could be measured but width not and where width was measurable but not the rate (signal showing a thick line with fluctuating intensity).* Gähler’s vibrato is basically continuous but when playing full chords it is not easily executed or measured. Measurements were possible for more than 60% of the selected notes. His vibrato tends to be slow but not too wide except on single longer notes. On these occasions it becomes faster
URL http://books.openedition.org/obp/docannexe/image/1860/img-5.jpg
File image/jpeg, 520k
Caption Figure 4.2. Bars 1-2 of Bach’s G minor Adagio for solo violin (BWV 1001). The top system is a hypothetical version that emulates the much sparser notation habits of Italian composers such as Corelli, who tended to prescribe only basic melodic pitches that the musicians were supposed to embellish during performance. The middle stave is Bach’s notation, reflecting one possible way of ornamenting the passage. His original slurs indicate ornamental groups to be performed as one musical gesture. The lowest stave shows the “standard thoroughbass” that players might conceptualize for a “rhapsodic improvisation” to unfold
URL http://books.openedition.org/obp/docannexe/image/1860/img-6.jpg
File image/jpeg, 80k
Title Table 4.5. Most embellished movements listed in order of amount of ornamentation. The named violinists add graces and embellishments extensively, decreasingly so as moving downward in each column. Others not listed here may also add a few graces here and there
URL http://books.openedition.org/obp/docannexe/image/1860/img-7.jpg
File image/jpeg, 132k
Caption Figure 4.3. A minor Andante, bb. 12-24. Transcription of melodic embellishments in Gringolts’ performance during repeat
URL http://books.openedition.org/obp/docannexe/image/1860/img-8.jpg
File image/jpeg, 108k
Caption Figure 4.4a. B minor Sarabande—Transcription of embellishments in Gringolts’ performance of the repeats
URL http://books.openedition.org/obp/docannexe/image/1860/img-9.jpg
File image/jpeg, 232k
Caption Figure 4.4b. B minor Sarabande—Transcription of embellishments in Mullova’s 2008 performance of the repeats
URL http://books.openedition.org/obp/docannexe/image/1860/img-10.jpg
File image/jpeg, 232k
Caption Figure 4.4c. B minor Sarabande—Transcription of embellishments in Luca’s performance of the repeats
URL http://books.openedition.org/obp/docannexe/image/1860/img-11.jpg
File image/jpeg, 200k
Caption Figure 4.5. Gavotte en Rondeau, E Major Partita: theme and its major variants
URL http://books.openedition.org/obp/docannexe/image/1860/img-12.jpg
File image/jpeg, 316k
Caption Figure 4.6. Transcription of 7 different ornamentations during the repeat of bars 1-3 of the Loure, E Major Partita WITH audio 4.7
URL http://books.openedition.org/obp/docannexe/image/1860/img-13.jpg
File image/jpeg, 208k
Caption Figure 4.7. Grave, A minor Sonata: facsimile of final bars
URL http://books.openedition.org/obp/docannexe/image/1860/img-14.jpg
File image/jpeg, 92k
Title Table 4.6. Summary of solutions in the penultimate bars of the A minor Grave
Caption † Brooks’s and Tetzlaff’s delivery seems to be a combination of bow and finger vibrato perhaps closer to tremolo.* Fischer and Schmid create a crescendo-decrescendo over the duration of the two dyads (accompanied by increasing-decreasing vibrato) whereas Ibragimova performs a messa di voce proper on the first dyad alone without any vibrato and articulates the second dyad quite separate from the first. There is hardly any crescendo-decrescendo in Podger’s and Khachatryan’s versions; straight-held dyads with trill on D# although one can perhaps detect a very slight tremolo/vibrato in Podger’s
URL http://books.openedition.org/obp/docannexe/image/1860/img-15.jpg
File image/jpeg, 120k
Caption Figure 4.8. Adobe Audition screenshots of spectrograms showing bars 1-3 of the B minor Allemanda in Kremer’s two recorded performances. The 2005 version (bottom pane) shows more gaps between note onsets indicating staccato playing
URL http://books.openedition.org/obp/docannexe/image/1860/img-16.jpg
File image/jpeg, 168k
Caption Figure 4.9. Adobe Audition screenshots of spectrograms showing bars 1-3 of the B minor Allemanda in Mullova’s three recorded performances. The 1992 (middle pane) and 2008 (bottom pane) versions show increasing gaps between note onsets indicating progressively more staccato playing
URL http://books.openedition.org/obp/docannexe/image/1860/img-17.jpg
File image/jpeg, 300k
Caption Figure 4.10. Loure, E Major Partita, bars 4-7
URL http://books.openedition.org/obp/docannexe/image/1860/img-18.jpg
File image/jpeg, 48k
Title Table 4.8. Summary of basic differences in executing bars 35-41 of G minor Fuga
Caption * Crescendo usually starts in the second half of b. 39, or in b. 40; there is a subito p in b. 38 after which loud dynamics resume
URL http://books.openedition.org/obp/docannexe/image/1860/img-19.jpg
File image/jpeg, 276k
Caption Figure 4.11. Bach’s original scoring of bars 35-41 of the G minor Fuga and transcriptions of performed interpretations (cf. Table 4.8)
URL http://books.openedition.org/obp/docannexe/image/1860/img-20.jpg
File image/jpeg, 244k
Title Table 4.9: Summary of modifications in the B minor Sarabande within the two essentially differing approaches
URL http://books.openedition.org/obp/docannexe/image/1860/img-21.jpg
File image/jpeg, 208k
Caption Figure 4.12. Comparison of tempo and power (dynamics) in bars 3-8 of the Largo from the C Major Sonata performed by (from top to bottom) Wallfisch, Huggett, Zehetmair, Poulet and Shumsky. Beat level tempo is indicated by the smoother line near bar and beat numbers. Power (dynamic) is indicated by the volatile slopes. Note the longer and clearer power arches in the bottom two panels as opposed to the near constant shifts in the more locally articulated versions above. Note also Shumsky’s fairly even tempo line. I used Sonic Visualiser to prepare this analysis
URL http://books.openedition.org/obp/docannexe/image/1860/img-22.jpg
File image/jpeg, 298k

Buy

Print version

Loading

Unavailable