4.1.1. Introduction: Discussion

The concept of musical gesture is usually linked to the experiential or performative aspect of music. In his online lecture series on the subject,[1] Robert Hatten defines “Gesture” as a “movement that may be interpreted as significant” (a definition which closely mirrors the OED’s definition: “A movement expressive of thought or feeling”). He adds that “[m]usical gesture presents more challenging problems, since it must often be inferred from notation and an understanding of performance practice and style” (Hatten 2001, lecture 1). Zohar Eitan goes further, claiming that musical gestures

may be mapped directly onto analogous processes in extra-musical domains. In particular, they may be mapped onto expressive extra-musical patterns, such as bodily gestures or vocalizations. (Eitan 2003, 217; see also Eitan 2000)

Both Eitan and Hatten see gesture as affecting performance and experience more directly than the thematic and harmonic categories of conventional analysis. Similarly, John Rink sees gesture as central to the performer’s conception of the musical work:

Whereas analysts concentrate on musical structure, performers attend primarily to musical ‘shape’, which is analogous to structure but tends to be more dynamic through its sensitivity to momentum, climax, and ebb and flow, comprising an outline, a general plan, a set of gestures unfolding in time. (Rink 1990, 323; cf. le Huray 1990, 19; Eitan 2003, 217)

For these writers, thinking of music in terms of gesture facilitates the appreciation and projection of unity and continuity in analysis, listening and performance alike. It makes it easier for performers to bring out local directionality, creating in the listener certain expectations on how the music will proceed (cf. Cohen 1994, 32-34) and thereby intensifying the sense of forward momentum and (especially when expectations are frustrated) drama.

For Hatten, “gesture” is

a holistic concept, synthesizing what theorists would analyze separably as melody, harmony, rhythm and meter, tempo and rubato, articulation, dynamics, and phrasing into an indivisible whole. [...] For performance, these overlapping strands must be further melded into a smooth, and at some level undivided, continuity. That melding is achieved most efficiently by means of an apparently natural, human gesture. Performers strive to create a shaping and shading of each phrase that is more than the sum of the motivic and harmonic units of which they are composed. (Hatten 2001, lecture 1)

Nonetheless, gestural analysis focuses on short musical events – motifs, figures or short phrases. The sense of unity is forged through a recognition of the gesture’s internal continuity and coherence, and of the interconnections between gestures. This enables performers to recognise – and project – seemingly disparate and distinct “motifs” as manifestations of the same “gesture”.

However, gestural projection of unity often depends on the performers’ willingness to abandon the facile continuity of an unbroken legato. As Hatten observes,

Performers knowledgeable about historical performance practice [...] are more likely to project articulations and subtle details that realize characteristic gestures in a way that is stylistically consistent with their implied expressive meaning and ongoing development. Romantically-schooled pianists are less likely to adjust to these stylistic constraints, perhaps due to differences in the modern piano, a bias toward unbroken continuity of (melodic) line, and/or a bias toward pitch-generated structural motives. (Hatten 2001, lecture 4)

The demand for detailed articulation is particularly emphatic in discourse on the analysis and performance of Baroque music. According to David Schulenberg,

the chief distinction between Baroque and later expression may be that in [the former] the signs are small figures in the surface, while in later music the signs take the form of larger music processes, such as the extended crescendo or the prolonged dissonance. (Schulenberg 1992, 105)

This “atomistic” way of thinking is common to several different approaches to the analysis and performance of Baroque music in general, and Bach’s in particular. Among other things, it is embedded into highly detailed lexical-symbolic analyses of Bach’s music. These analyses can be traced back at least as far as Albert Schweitzer’s Bach monograph (Schweitzer 1966), in which the author sought to assign specific extra-musical meanings to recurring motifs and figures in Bach’s music. Later scholars attempted to develop similar theories on more historically credible grounds, arguing 17th- and 18th-century treatises on musical rhetoric reveal a coherent doctrine of the affections (Affektenlehre) and of meaningful musical figures (Figurenlehre); on this basis, they claimed to reconstruct a “dictionary” of musical “words”. Arnold Schering’s pioneering work in this field (e.g., Schering 1941) was refined and expanded by scholars like Hans-Heinrich Unger (1941) and Arnold Schmitz (e.g., Schmitz 1950, 1970). The wide acceptance of their ideas is attested by their inclusion in the entries on musical-rhetorical figures in the 1955 and 1997 editions of the German music encyclopaedia Die Musik in Geschichte und Gegenwart (Schmitz 1955, Krones 1997; see also Blume 1975, 111-117) and in the recently-published Bach-Lexicon (Hartmut Grim, in Heinemann 2000, 35-40, 192-194). This theory inspired the emergence of the rhetorical approach to the performance of Bach’s music.

Much of the discourse on localised detail projection in Baroque performance thus focuses on the analogy between music and speech. The speech-act, however, can be considered a specific type of gesture (Eitan 2000). Historical performers also emphasise, in theory and practice alike, the realisation of dance rhythms and dance-like features, which also relate to physical, bodily gesture.[2]

In recent years, explicit references to the embodied, gestural dimension of Baroque rhetoric and dance gained prominence in verbal discourse on Baroque performance. Bradley Lehman defines musical gestures as follows:

Musical gestures are contrasts of character within a composition, from phrase to phrase and section to section: recognition and expression of great diversity within default continuity. [...] Gestural playing (or singing) is multi-dimensionality. The performer allows the articulation, accentuation, even (somewhat) the tempo to be different on every few notes if that is the natural shape of the lines. Everything is dynamic, fluid, in flux. (Lehman 2004)

Bruce Haynes’s conception of gestural performance is broadly similar. Like Schulenberg, he claims that the phrase as most musicians define it today – a unit lasting several bars – was foreign to Baroque discourse on music:

Melodies in Baroque pieces tend to be complicated, with twists and turns, and this is because their basic structural unit is smaller than the Romantic phrase. (Haynes forthcoming, chapter 11)[3]

Haynes rejects the “long-line phrase”, which creates “a sostenuto effect” by employing “a single breath or bow stroke” throughout the length of the phrase. Instead, he advocates a gestural style, focusing on the shaping of local units, such as short figures, individual notes and the all-important silences between the notes. “The long-line”, he writes,

was designed to promote a legato ambiance, broad movements, and one important point per phrase; gestures, by contrast, promote a series of silences, quick changes of character, and ever-changing detail – a sound kaleidoscope. [...] when each gesture is given its special character, its individual dynamic and rhythmic shape, Baroque lines gain life and logic, while phrases hold the gestures together and give them continuity and coherence. (Haynes forthcoming, chapter 12)

While Lehman’s and Haynes’s explicit emphasis on gesture is an innovation, the approach they advocate has characterised much Baroque performance since the advent of historical performance in the 1960s – and even more prominently since the 1970s and 1980s (for a detailed survey, see Fabian 2003, esp. 205-248; cf. Golomb 1998, forthcoming). Indeed, Haynes’s approach closely resembles Nikolaus Harnoncourt’s (1988, 25, 41ff and passim), even though Harnoncourt does not employ explicitly gestural terminology. Verbal discussion of gestural performance was thus born out of the wish to explain, advocate and intensify a style of performance which had already been established.

Previous arguments in favour of this style relied primarily on rhetorical theory. Several performers – most notably Harnoncourt – endorsed a rhetoric-as-semantics approach, based on a firm belief in the viability of Figurenlehre.[4] Others (e.g., Gustav Leonhardt) adopted a more circumspect, rhetoric-as-speech approach. These musicians argue that Baroque music follows the patterns of speech, and should be articulated accordingly.[5] While rejecting the semantic specificity entailed in Figurenlehre theories,[6] they emphasise the direct, emotive character of Baroque figures and consequently advocate their detailed realisation in performance. The analogy between music and speech entails a rejection of uniform intensity and uninflected, undifferentiated phrasing, and encourages flexibility and attention to detail. This connection between detailed articulation and the arousal of affections is also commented on in several German Baroque treatises (Butt 1990, 19-24). It thus arguably provides a firmer basis for performative expression than arcane conceptions of rhetoric-as-semantics (see also Butt 1990, 12-15, 1991, 84-85; Gustav Leonhardt in Sherman 1997, 196; Koopman 2003, 44-45).

On the other hand, the rhetoric-as-semantics approach did not necessarily result, either in theory or in practice, in gestural performance. Rhetorical or gestural performance was presented from its inception as a necessary correction to the inflexibility and uniformity of the “terraced dynamics” approach (e.g., Wenzinger 1968; cf. Lehman 2004). However, early Figurenlehre theorists often supported the terraced style. Arnold Schering, for example, argued that it is very important to decode Bach’s symbols for listeners’ benefit (Schering 1941, 71-72), but he exhorted performers not to emphasise the musical motifs that supposedly embody these symbols.[7] Even when directly alluding to the concept of musical gestures, he argued that it should not result in gestural performance. Instead, he claims that what is present in the music (+) need not, indeed should not, be emphasised in performance (-); this may be labelled, in schematic abbreviation, as a +/- approach:[8]

Bach’s affect is always identical with a particular melodic gesture. This gesture is so clear that it can always be understood [...] A performer who over-emphasises such clear gestures through exaggerated emotional expression is similar to an actor who accompanies every movement with exaggerated facial and bodily expressions. (Schering 1936a, 188; my translation)[9]

This advocacy of literalism and restraint, and explicit rejection of detailed localised gestures, is also reflected in several traditional performers’ opposition to the gestural detail in HIP (historically-informed performance) interpretations. Helmuth Rilling, for example, writes:

the proponents of the “historical” approach direct their attention too much to microstructure. Short individual notes or groups of notes that are separated after a tie emphasize momentary events and distract one from more important interrelationships. It seems to me that this might be a possibility for small-scale movements, but it is an encumbrance for complexes of large dimensions. (Rilling 1985, 14)

Writers like Hatten and Haynes, by contrast, believe that proper “attention to microstructure” can enhance the projection of “important interrelationships”.

4.1.2. Introduction: Outline

In this paper, I propose to explore some of these issues by examining recordings of the First Kyrie from Bach’s B minor Mass. This is one of Bach’s large-scale movements, and therefore provides a particularly good case-study for examining Rilling’s contention that locally-gestural performances disrupt overall architecture.

The paper is divided into two parts. Part One focuses on the gestural realisation of figures and motives within the movement’s fugal subject, and its potential to enhance or disrupt continuity. Part Two examines how the shaping of local gestures can affect the movement’s overall trajectory, and also whether and how the gestural imagery applied to the movement as a whole (conceiving it in terms of one or two large-scale gestures) can be realised in performance. Both parts rely on my analysis of the First Kyrie, which can be accessed in a [ separate window ].

Within each part of the article, I distinguish between “interpretations in theory” (verbal analyses and commentaries) and “interpretations in practice” (performances); this distinction is akin to Jerrold Levinson’s distinction between CIs (Critical Interpretations) and PIs (Performative Interpretations). Levinson is rightly sceptical of the possibility of one-to-one mapping between the two types of interpretation, though he does not rule out the viability of comparing between them:

When we hear a striking PI of a familiar piece, the question we put to ourselves as interpreters of such interpretations should be not, ‘what CI does that PI embody or convey?’ [...] but instead ‘What CIs might such a PI support or reflect?’ An insightful PI might prompt one to arrive at a new CI, or allow one to confirm the validity of a CI already proposed, or induce one to question a CI regarded as authoritative, and so on, but it cannot itself unambiguously communicate a CI. (Levinson 1993, 57)

As Levinson suggests, relationships between verbal and performative reception are not easy to establish or to interpret. In some cases, a verbal interpretation or analysis might have clear performative implications, and one can point to specific performances that realise these implications. Even in these cases, however, one cannot always assume that the performer was familiar with the analysis (or vice versa). A similarity between a performance and an analysis (or, for that matter, between two performances) might point to convergence – two musicians arriving independently at a similar view of how the music should be understood or performed; and the performers’ view of the music (insofar as they express it in words) might not correspond to the analyst’s view.

Even when the CI and PI can be safely attributed to the same musician, the exact correlation and causation are not easy to establish. It is possible that the performer’s words might represent a post-hoc justification for performative choices, rather than representing the thought processes that shaped the performance. In other cases, the performer’s words and the musical choices documented in the recording are difficult to reconcile with each other.

These points should be kept in mind when reading the analyses contained in this paper. In making comparisons between different interpretations, I have made every effort to cite available evidence for the performers’ intentions. However, I draw attention to distinct correlations between verbal and performative discourse (or between different performances) even when there is no evidence for direct connections between writer and performer (or two performers). At the end of the article, I will try to gauge the possible significance of some of these correlations.

4.2. Part One: Figures and Gestures within the Subject

Fugal subjects are often regarded as anchors of stability: Deryck Cooke (1959, 8) compares them to “a brick or a block of stone [...] something of no importance in itself, only useful as raw material to be built into a structure”. Consequently, it is often assumed that they should maintain a steady character:

Since the basic figure of the subject remains constant, the phrasing of that figure should also remain constant. Thus, throughout a fugue, or any composition built on constant motives, the phrasing for the motives remains unchanged. (Tureck 1960, II, 20)

At their most rigid, realisations of such prescriptions reflect a non-developmental conception: the subject is consistently phrased in an internally stable, unyielding manner. This approach is linked to the “terraced” style, and, at least in its extreme manifestations, stands at odds with the rhetorical-gestural style. It is especially problematic when applied to the First Kyrie’s subject. Cooke himself cites this subject among the exceptions – cases where “the thematic material of polyphony is itself expressive, even highly expressive” (Cooke 1959, 9). Several factors – inner polyphony, harmonically open-ended character, wide range, chromaticism – contribute to its complexity and intensity alike.

4.2.2. Interpretations in Theory: The Internal-Polyphony Approach

The writers discussed so far marginalize the subject’s inner polyphony; they can thus describe the subject as passionate, yet purposeful and devoid of internal conflict. By contrast, Ernst Kurth, in his 1917 treatise Grundlagen des linearen Kontrapunktes (Foundations of Linear Counterpoint), claimed that inner tensions – generated by internal polyphony – are a central feature of Bach’s style, and contribute considerably to the expressive power of Bach’s music (Kurth 1991, 72 and passim).

Kurth’s theory was applied to Bach’s choral fugues by his student Eugen Thiele.[12] In his references to the First Kyrie, Thiele draws attention to several figures within the subject (Thiele 1936, 23, 26, 28) – albeit more on a structural than an expressive front. However, he ends up analysing the subject as a single rising gesture – an archetypal example of intensification through combined linear (melodic) and harmonic means (“Ausdruckssteigerung durch linear-harmonische Mittel”; ibid, 37).

More recent writers place a stronger emphasis on the subject’s internal polyphony. John Butt, for example, describes the lower G-F# figure – which he terms a “sigh” figure – as “the most significant component of the opening harmony”, central to the movement’s motivic structure and expressive import alike. He points out that it “acts as a recurring ‘pedal’ in the fugue subject and also constitutes the climax of the phrase in b. 7” (Butt 1991, 87). Part of its effect, however, depends on its performative realisation: when properly articulated, with an accent on the off-beat G, it “rubs against the meter – it’s a metrical and melodic dissonance” (in Sherman 1997, 180; see also Butt 1990, 30).

Other writers view the subject as a series of connected figures. Stauffer (1997, 55-56), relying in part on Blankenburg’s and Butt’s analyses, enumerates five of them: Repercussio (“Kyrie”);[13] a double-layered “wedge” consisting of Blankenburg’s Gradatio in the upper register and Butt’s “sigh” figure in the lower register; a “chromatic digression”; and Blankenburg’s Exclamatio.

Nikolaus Harnoncourt, applying a rhetoric-as-semantics approach, discerns yet another figure. In the notes to his 1968 recording, he writes:

The rhythm which pulsates through the entire “Kyrie” I can probably be understood as a very intensive musical and rhetorical gesture of supplication: “Herr, erbarme dich unser” (Lord, have mercy on us). (Harnoncourt 1989, 191; for a different translation, see Harnoncourt 1968, 11)

In the notes to his 1986 recording, he offers a more detailed explication of the gesture’s character and its performative implications alike:

It makes quite a difference if every player realizes that a figure which occurs throughout the Kyrie is a gesture of supplication, and the fact that this has been recognized as such in Western music for many centuries is probably connected with physical imagery. If one is urgently asking for something one drops to one’s knees, tugs at garments, and this gesture of supplication has an element of tugging, even when translated into music. (Harnoncourt 1986, 39)

While Harnoncourt ascribes this gesture to the upper part (ibid, 42; see fuller quotation in Section 4.2.4. below), one could argue that his imagery is equally applicable to the “sigh” figure in the lower register. The Gradatio’s “tugging” potential is stronger, since its downbeats are placed on the strong beats; the sigh’s downward trajectory, however, seems more redolent of supplication, and its attempt to “hold back” the Gradatio’s progression is also consistent with the “tug in the garment” imagery.

In any case, the gesture is heard more clearly outside the subject, in the ritornelli’s F2. Here, the “tug in the garment” is isolated through the insistently mono-rhythmic texture, and (in the stronger beats) played out by the entire orchestra; furthermore, most of its appearances have a downward melodic trajectory even in the upper register.

Harnoncourt’s imagery is reminiscent of Terry’s, insofar as both writers discern a gesture of supplication. Terry’s gesture, however, is ascending, continuous and flowing (cf. Dickinson 1950, 192; Mellers 1980, 164), whereas Harnoncourt’s is descending, halting and hesitant. Furthermore, Terry’s gesture depends on a continuous realisation of the subject as a whole, whereas Harnoncourt’s demands an internal disruption within the subject.

The disruptive potential of realising figures as performative gestures becomes evident in the example below, which brings together the various figures discerned in the subject. This illustration conflates the analyses of Blankenburg, Butt, Harnoncourt and Stauffer, as quoted above. With one exception (Harnoncourt’s “tug in the garment”), all of the figures cited in this example are cited in Stauffer’s analysis:

One would suspect that an attempt to realise all these figures as performative gestures would sound disturbingly fragmentary. The “sigh” figure and the Gradatio interfere with each other’s progression – as Stauffer’s “wedge” imagery implies, they pull in opposite directions. However, an examination of the movement’s recorded performances only partly confirms this expectation.

4.2.3. Interpretations in Practice: Non-Gestural and Anti-Gestural Shaping

This section focuses on the subject’s initial appearances: the opening statement by the first flute and first oboe (bars 5-9) and the tenor’s entry commencing the first fugal exposition (bars 30-33). In these appearances, the subject is sufficiently exposed to allow a listener to detect most details in its shaping. Two things, however, must be mentioned at the outset:

1. The subject never appears entirely in isolation; its character is partly determined by the shaping of other strands in the texture.
2. In most performances, the subject is not shaped identically in all appearances; the statements discussed here are not necessarily representative of their respective renditions.

Within the subject, constituent units can be distinguished by dynamics and articulation alike. The least disruptive option, however, is to avoid these distinctions and perform the subject sempre legato with little or no dynamic inflection. This smoother approach is prevalent among symphonic conductors (e.g., Karajan 1952, 1974; Jochum 1957, 1980; Lehmann; Maazel; Giulini 1972, 1994), as well as pre-1980s modern-instrument Bach and Baroque specialists like Münchinger, Rilling (1977), Corboz and Marriner.[14] In some of these performances, this locally uninflected approach is allied with an attempt to project the movement’s overall structure.

This approach obviates inner conflicts in the subject, even when its internal polyphony is observed. If the upper-register line is not shaped with a Gradatio-like upward trajectory, a slight separation of the lower “sigh” figure does not alter the basic affect: since there is no rising gesture, neither the “sigh” nor the chromatic digression are felt to disrupt it.

The most consistent representative of this approach is Karajan:[] in both his commercial recordings (1952, 1974), the subject is shaped with very little inflection or distinction between components. Most other performances are more varied. For example, Jochum 1957 [] features divisions into short legati in the vocal statement, and a slight crescendo in the Gradatio figure. The “sigh” is slightly distinguished by being sung more piano than its already-soft surroundings. The narrow dynamic range and the gentle, barely-perceptible separation between the legato phrases, however, largely prevent the emergence of distinctive gestures.

Before the advent of HIP, the main alternative to this uninflected approach was the sternly articulated approach, exemplified by conductors like Ramin, Mauersberger, Richter (1961, 1969a, 1969b) and Klemperer (1961, 1967). Ramin and Mauersberger (respectively, erstwhile Kantors of the Leipzig Thomaskirche and the Dresden Kreuzkirche) were primary representatives of the “Leipzig” or “Saxony” school of Bach performance. In the 1930s-1950s, they were regarded as the vanguard of “historical performance”. Their performances were endorsed by musicologists like Arnold Schering and Wilibald Gurlitt, who themselves adopted a strictly x/- approach to Bach performance (see note [8]); this lent an aura of authenticity to their austere aural image of Bach’s music.[15] Their shaping of the First Kyrie’s subject was typical of their general cultivation of uniform intensity, terraced dynamics and literalism. Karl Richter studied with both Ramin and Mauersberger;[16] Klemperer employed Ramin as continuo player and musicological adviser in 1929-1932, and his views on Bach performance were clearly influenced by Ramin’s (Klemperer 1982, 48, translated in Klemperer 1986, 67; Heyworth 1996, 317-319).

The most extreme representative, in this particular case, is Rudolf Mauersberger:[17]

Here, the subject’s constituent elements are separated, but they all seem to have the same character. The “sigh” is distinctly isolated from the higher quaver pairs; the latter, however, are not connected to each other. Due to equalised accentuation and static dynamics, there is little sense of movement. The sharply-etched, static character of each figure is at odds with any gestural approach; some might even consider it anti-gestural.

While this characterisation holds true, in my view, for Richter’s performance of this movement, this is only hinted at by his initial shaping of the subject. The two instrumental statements [] sound meticulously weighted, note-by-note. The vocal shaping [] is initially more expansive: legato articulation in groups of four; individual notes clearly enunciated, but without Mauersberger’s insistent aspiration. The most flexible shaping, however, is reserved for non-subject material (especially in E1 and R2).

In all these interpretations, the detailed yet rigid shaping of the subject is allied with a similarly rigid shaping of the movement as a whole. Klemperer, on the other hand, associates a similar treatment of the subject with a globally-directional shaping of the movement.

4.2.4. Interpretations in Practice: Historical Performance – Rhetoric and Dance

For all the contrasts between them, the “smooth” and “statuesque” approaches both avoid explicit gestural shaping. Performative realisation of the gestures discussed above (the Gradatio, the “sigh” and the “tug in the garment”) only began in earnest with the advent of HIP.

Renditions of the First Kyrie’s subject reveal two contrasting trends within HIP: a mildly-detached, lightweight approach, and the projection of internal tensions. The most extreme realisation of the former tendency, however, comes from a non-HIP recording – Schreier 1982. This reading is closer to the Leipzig tradition, not only in its genealogy [18] but also in its specific musical elements. In all instrumental statements (including the bass in bar 22), it combines insistent staccati on quavers with almost equally incisive articulation in the surrounding texture; dynamics are almost uninflected:

The choral articulation is less incisive, but equally rigid.

Schreier’s reading is reminiscent of Mauersberger’s in its dynamic and articulatory rigidity. The light textures, fast tempo and more incisive articulation, however, are more typical of the then-emerging HIP style, as is the treatment of as a separate figure.

A similar if milder approach can be found in Parrott, Schreier 1991, Eby, Koopman and Fasolis. The articulation is gently detached;[19] the effect is closer to dance-like elegance than to aggressive, harshly-accentuated staccato. This does not depend on articulation alone: an impression of lightness arguably involves lighter texture, avoidance of heavy emphases or harsh downbeat accentuation, and some degree of dynamic and/or tempo flexibility.[20] Thus, Schreier 1982 generates heaviness through rigid dynamics and accentuation, which makes it difficult to describe the resulting interpretation as “gestural”. The dance-like character of the other recordings is more easily associated with human movement. In Parrott’s case, the lightweight effect is more pronounced in the isolated figure (F2, bars 19-21 and simile) than within the subject (where an emphasis on the figure balances the lighter effect in the upper register).

In both Parrott and Koopman ([], []), vocal statements of the subject are less clipped and more dynamically flexible than instrumental statements:[21]

All the above-mentioned performances retain the forward placing of the subject (vis-à-vis other strands in the texture), at least in instrumental statements. This adds to the sense of lightness; the figure’s buoyancy would probably have been softened by a fuller surrounding texture.[22]

More rhetorically-inclined performers draw out several of the figures that have been read out of (or into) the subject in a manner more consistent with the verbal analyses cited in the previous section. Surprisingly, perhaps, the “sigh” figure is not frequently treated as a metric dissonance; even when the slurring of this figure is observed, there is usually little or no emphasis on its first note, and therefore little suggestion of off-beat accentuation. There are occasional hints in several performances (including the pre-HIP Scherchen 1950 and Shaw 1960); in other cases, the figure’s first note is emphasised when the subject is stated in the bass (bars 22-23), but not in other statements (e.g., Rifkin, Gardiner, Leonhardt, Harnoncourt 1968 and 1986, Rilling 1988). The most consistent emphasis on this figure can be found in Hengelbrock’s choral statements.[] [23]

The Gradatio receives more consistent attention from HIP performers (primarily those of rhetorical inclination). Wenzinger (1968, 42) equated Gradatio with crescendo (however, cf. Bartel 1997, 220-225, 267). There are two ways to realise this identification in the Kyrie subject: three separate crescendi; and linked, continuous crescendi, creating a single gesture out of the three disparate pairs. The former approach is reminiscent of Stauffer’s description of the subject as a “wedge”; the latter is reminiscent of Terry’s and Blankenburg’s analyses.

Leonhardt’s performance demonstrates the first approach. In the instrumental statements, dynamic gradation is implied by articulation: a light upbeat followed by an accented, tenuto, downbeat from woodwinds and bass alike.[24] This latter affect contributes to the Gradatio’s prominence by submerging the “sigh” figure.

Leonhardt’s vocal statements are phrased in the standard four-note pattern. Dynamics play a more distinctive role here: a subtle internal echo allowing prominence to the Gradatio’s two-note crescendo.

A similar pattern occurs in Herreweghe 1988,[] albeit with more continuous support from the bass and a clearer tendency towards overall dynamic construction. In the vocal statements, Herreweghe submerges the “sigh” figure by joining it together with the downbeat; the emerging pattern is , rather than the more conventional .

The vocal statements in Herreweghe 1996 provide the clearest illustration of the Gradatio as a single crescendo (and also vividly brings out Blankenburg’s Exclamatio figure).[25] As such, they illustrate Hatten’s claim that gestures are “not necessarily continuous sound, but [can consist of] continuity of shape, curve, motion across silence” (Hatten 2001, lecture 2).

Similar patterns can be found in Jacobs, Hengelbrock [], Koopman [], Brüggen, Christophers, and Jeffrey Thomas. The last mentioned, however, seems closer to the older tradition of viewing the subject as a single rising gesture, especially in the vocal statements:[26]

While not all vocal entries are shaped in precisely this manner, the sense of continuity – within and beyond the phrase – is always palpable. I was unable to find another performance that projects a similar trajectory.

Thomas Hengelbrock projects the Gradatio figure beyond (but not through) the “sigh” figure (and more clearly in vocal than in instrumental statements). However, his shaping of this figure, and of the subject as a whole, is deliberately hesitant. Hengelbrock shapes the vocal entries as a series of legato pairs. This articulation is by no means unique to Hengelbrock (cf. Brüggen’s instrumental statements, Gardiner’s vocal statements). But Hengelbrock’s tempo is slower, his emphases heavier, the breaths separating the pairs more extended, and the dynamic contrast between higher and lower registers more clearly distinct. Thus, his “sighs”, unaccented though they are, still act as interruptions to the Gradatio.

The same factors are also present in Harnoncourt’s 1986 version.[27] Harnoncourt’s tempo ( = 50) is not as slow as Hengelbrock’s ( = 46). On the other hand, Harnoncourt’s emphases and accentuations are heavier, and the separations between phrases often longer. The “tug in the garment” imagery focuses on a particular figure, but its spirit affects the performance even when that figure is absent: there is a sense of something being dragged backwards, tugged in the opposite direction to its purported motion. Thus, the impression of constant, deliberate interruption is even stronger here than in Hengelbrock’s reading.

In the orchestral statements, this effect is further intensified through the independent shaping of the bass. Earlier (note 24 above), I cited examples wherein the bass line supports the subject’s trajectory; here the two are at odds with one another. Harnoncourt does not ignore the strong-weak metric division; but he shapes the bass as an independent melody, whose trajectory only partly coincides with the subject’s. Here, different gestures are employed simultaneously; consequently, as one part of the texture seems to strive forward, another seems to drag backwards.

The effect is somewhat softened in R1’s F2, when the “tug in the garment” figure appears in isolation. Here, Harnoncourt follows the markings in the 1733 Oboe parts in bars 19-20.[28] As he explains,

In the introduction to the first Kyrie there is a continuing insistence, and at the same time the symmetry is quite obvious. Bach creates an upper part out of the quavers which are not slurred and continues it, and this upper part employs the gesture of supplication. The symmetry of the lower parts is slightly impaired by the slurs. Bach seems to have felt that was going too far in the long run. (Harnoncourt 1986, 42)

For that he reason, he contends, Bach added the slurs marked as “Oboe only” in the example below:[29]

this is precisely where the strings play the gesture of supplication, and this is suddenly resolved by brilliant inconsistency in that Bach writes slurs in two places where they would not be expected to occur. (Harnoncourt 1986, 44)

Harnoncourt preserves this inconsistency in his performance, articulating the “tug” gesture () as separate notes in the strings simultaneously with tied quavers in the woodwinds, without accenting the first notes in these slurred pairs. The effect is a softening of a passage which, in many other performances, emerges as more insistent than the rest of the ritornello; the sense of hesitation, however, remains.

In the vocal statements in E1, the sense of hesitation and instability is created primarily through the voices’ detailed articulation and deliberate accentuation of the subject, combined, as before, with the slow tempo.

Across the movement, however, it is the mutual questioning of parts which is predominant.

Harnoncourt’s and Hengelbrock’s questioning manner is an exception; the clearer trajectories of Thomas and Herreweghe (1996), the elegance of Parrott and Koopman, and the calmer renditions of Rifkin, Leonhardt, Herreweghe (1988) and others are more common. It is also these influences that filter beyond HIP into later mainstream renditions. Several pre-1980s performances by Bach and Baroque experts (e.g., Münchinger, Rilling 1977, Marriner, Corboz) are characterised by a flowing, dynamically-narrow approach, with little articulatory inflection. Later readings, however (e.g., Rilling 1988, esp. instrumental statements; Schneidt; Beringer; Abbado; Rilling 1999; Ozawa’s vocal statements) reveal HIP influences through their lighter textures and more detailed articulation (short legati, gently-detached non legato). A few performances (e.g., Biller, Ozawa’s orchestral statements) even approach dance-like elegance.

4.2.5. Interpretations in Practice: Summary

Two prominent tropes in the First Kyrie’s verbal reception (the single rising gesture and the Gradatio) focus on the subject’s overall trajectory. Only in recent years, however, did performers begin to project this trajectory in a clearly audible fashion. HIP musicians, in particular, seem to have explored the performative potential of these tropes, enhancing the sense of movement and directionality which is closely linked with the performative realisation of gestural potential. Few pre-HIP renditions reveal a concern with the subject’s shape; for the most part, they focus on projecting its general character, allowing (at most) a few localised dynamic nuances.

HIP musicians – and those under their influence – produce at least two types of gestures: localised, dance-like elegant movement; and rhetorically-influenced gestures, with a clear trajectory (whether continuous or interrupted) which is reminiscent of expressive or symbolic interpretations of the subject. The most detailed among them (Harnoncourt and Hengelbrock) seem intent on the realisation of several gestures, in close conjunction or even simultaneously, in different strands of the texture. This approach seems to realise the misgivings expressed by Helmuth Rilling (quoted above): while giving coherence to each gesture, they seem content to allow them to clash with each other, threatening to undermine the movement’s overall structure. In Part Two, I will examine the interaction between localised gesture and overall shaping.

4.3. Part Two: The Structure and Shape of the Fugue

Mono-thematic fugues are sometimes cited as the ultimate examples of Baroque Unity of Affect; fugal subjects are viewed, in this context, as the unchanging building-blocks of the fugue as a whole (Neumann 1953, 101-105; Tureck 1960, II, 20; Cone 1968, 70-71). According to this conception, the shaping of the subject determines the character of the entire fugue. In performances of the First Kyrie, this claim might be applicable in terms of affect: how lightly or portentously the subject is performed reliably presages the performance’s general character. But when it comes to predicting the overall shape, a myopic discussion of the subject can be misleading. In particular, several performances which seem utterly un-gestural when one focuses exclusively on their shaping of the subject reveal a keen interest in realising gesture on a larger scale.

4.3.1. Interpretations in Theory

The First Kyrie arguably displays a strong equivalence between the micro- and macro-structure. Terry, for example, seems to consider the movement as a whole – not just its subject – as a single continuous gesture:

First the Tenors, then the Altos, then the Sopranos, and lastly the Basses raise the threnody, which swells with increasing urgency until it reaches its tremendous climax, eight bars from the end, upon the entry of the vocal Basses. (Terry 1924, 33)

Dickinson applied Terry’s imagery of the subject even more explicitly to the movement as a whole, describing it as “a most dramatic image of humanity stretching out to heaven in two large gestures of increasing urgency” (Dickinson 1950, 192). Similarly, Stauffer (1997, 57) directly linked the subject’s complexity with the movement’s “growing urgency” and “rising strength and momentum”.

The two-gesture imagery is more common than Terry’s single-gesture approach. The movement is clearly divided in two parts (see Overview and Analysis in [ separate window ]), the second of which is commonly regarded as more intense. E2 builds up in a composed crescendo, proceeding upwards from the bass (balanced, however, by continued thematic activity in the orchestra’s higher registers: the choir gradually merges into the oboes and violins).[30]

Harmonically, too, E2 features a greater build-up of tension. In E1, the alteration of Dux and Comes is straightforward, and each tonic statement of the subject is firmly supported in B minor. E2, on the other hand, is tonally more active and volatile. There are strong cadences, in the bass, on F#-minor (bars 85/6, 91), E minor (bars 97/8), and A major (bars 93/4, 100), as well as B minor. With one exception, however (the bass’s entry at the beginning of the exposition), these cadences are not aligned with the subject’s entries, which consequently do not receive the same harmonic support as their E1 counterparts. The second soprano’s tonic entry is avoided altogether: instead, it enters in E minor (bar 97). Instead of marking clear points of demarcation, the aforementioned cadences underpin continued activity elsewhere in the texture. This lack of clear cadences across the texture prevents closure and maintains momentum.

In gestural terms, this invites two different interpretations: the location of local dramatic gestures, or the quest for large-scale gestures, of the type suggested in Dickinson’s commentary.

It is possible to locate at least two powerful local gestures in E2: the bass’s entry at the beginning of E2, and the first soprano’s entry at the transition to R3 (bar 102). The latter represents the culmination of a gradual rise-from-the-depths and consequently might be treated as a dramatic climax. Both gestures, however, have a clear beginning but no clear ending; and the soprano entry is not easily treated in isolation. It is questioned both harmonically (a tonic entry of the subject underpinned by a strong applied dominant) and texturally (the entry is disguised by the second soprano’s sustained F#). This increases the dramatic tension at this point, but makes it harder to demarcate a local gesture. There is no sense of closure at this point; tension is maintained throughout the ritornello’s Vordersatz. Full resolution in the tonic is only attained at the ritornello’s Epilog (that is, at the end of the movement).

The sense of cumulative tension in E2 can thus lead to interpreting the rise-from-the-depth as a single gesture, spanning the entire exposition and possibly culminating beyond it, in the transition to R3’s Fortspinnung.[31] While it is easy enough to describe this gesture in writing (and defend this description by emphasising the lack of clear caesuras within and after E2), realising it in performance is another matter. This is gesture writ large – even in the fastest performances, E2 lasts about a minute and a half, and as the transition to R3 is seamless, tension usually continues to unfold after it as well. Musical gesture is commonly associated with the lifespan of human gesture (this is certainly the case in the theories of Hatten, Eitan, Lehman and Haynes); it connotes something compact and discrete that could be encompassed within the span of a normal human breath. A 20-bar fugal exposition is too long to count as a single gesture, by this definition.

As noted above, rhetorical or gestural thinking has led in some cases to a dissection within the subject. The projection of internal divisions and conflicts on this local level would arguably make it even harder to construct entire sections of the movement as single gestures. On the other hand, an emphasis on the Gradatio element in the subject can facilitate the creation of linked gestures. Local gestures can be connected into a continuous chain; overall shape can be projected through – rather than against – smaller units. Different performers’ responses to this challenge will be discussed in the next section.

4.3.2. Interpretations in Practice

E2’s basic shape allows for a spectrum of performative realisations. Some musicians attempt to shape it as a single, continuous rise in tension, culminating in bar 102 or beyond. Others offer a less linear shape, with several ebb-and-flow patterns within E2, rendering the climax at or around bar 102 less obvious; or maintain low tension through most of E2, offering only a single crescendo at bars 99-102. Finally, there are many performances that do not delineate strong patterns of tension and resolution, but rather project an almost uniform level of intensity throughout. Of interest in this context, however, is not only the shape given to the section, but also – indeed, primarily – the means used to project each vision, and particularly the relationship between the shaping of localised gestures (such as those discussed above) and of the section and movement as wholes.

In some cases, there is an inverse correlation: strong localised shaping is allied with a lack of overall projection, whereas smooth, uninflected shaping of the subject is linked with an attempt to project entire sections of the movement as large-scale gestures. This is especially evident in pre-HIP readings, dating mostly between 1950 and 1980 (see note 14). Two extreme examples of such inverse correlations are Mauersberger and Karajan 1952. Mauersberger focuses on chiselling out local details at the expense of a static reading of the whole, whereas Karajan seems to ignore local details in favour of overall sweep.

Mauersberger achieves textural clarity through meticulous articulation and cohesive tonal production. His dynamics, however, are virtually static. Karajan 1952 is characterised by predominantly legato articulation, with little or no separation between phrases; his texture is treble-dominated. On the other hand, he strongly projects the movement’s overall shape. E2, in particular, is shaped as a single, inexorably rising gesture: a series of crescendi, connected by brief passages of dynamic stasis. There is a slight articulatory emphasis on the first soprano’s forte entry in bar 102; otherwise, the exposition is shaped almost exclusively by dynamics, and the orchestra plays a more dominant role than the choir.

An even more interesting contrast emerges from a comparison of two ostensibly less extreme examples: Jochum 1957 and Richter 1961. In both performances, attention is devoted to projecting shape and texture alike. Nonetheless, their conceptions remain radically different.

Richter proves more interventionist than his Saxon mentors and colleagues. Günther Ramin and Kurt Thomas, though not as dynamically uniform as Mauersberger, do not attempt to project anything similar to Karajan’s arch of rising tension; Richter goes one step further, seemingly seeking to prevent this arch from arising. His shaping of E1 is comparatively nuanced – his is among the few performances where this section is shaped with greater detail and global directionality than the movement’s second part.[32] The movement’s second part is built in a much stricter, terraced manner – even in his live 1969 recording, which is generally more flexible than his two studio recordings.

Terraced rigidity is already clear in the Interlude, where the four phrases are mutually distinct and internally uniform. The basses then enter sforzando – there is no diminuendo in the end of the Interlude; the subject is shaped more meticulously, resembling the instruments’ weighting of the subject in R1, rather than the chorus’s more expansive approach in E1. This powerful gesture places the climax at the beginning of the second part. There is nowhere to go from this point, unless dynamics are reduced to allow for a new crescendo, an option which Richter avoids. There is a small-scale crescendo in bars 99-104, and a slight lightening of intensity (softer dynamics and timbre) on R3’s F1; but another crescendo through F2 leads back into an intense rendition of R3’s Epilog. These small modifications notwithstanding, Richter’s shaping of the movement creates an unyielding, monumental impression.

Eugen Jochum’s vision of the First Kyrie, in both of his recordings, is more organic and developmental. The beginnings of all main structural sections are highlighted. There are clear demarcations at the end of each ritornello; the end of the Vordersatz in both choral ritornelli serves as a high point of tension – especially in R3. Tension is maintained, however, even in moments of comparative relaxation. Thus, the dynamics drop at the beginning of each Fortspinnung; but Jochum preserves momentum by highlighting inner strands in F1. The more homophonic F2 serves as a crescendo leading into the Epilog, where the bass’s entry is emphasised in all three ritornelli.

While several passages in Jochum’s performances can be regarded as “high points” (the transition into R2,[33] the beginnings of the Epilogs), the most highly charged passage is in bars 81-112 (E2 and beginning of R3). The contrast between Jochum 1957 and Richter 1961 in this passage is revealing. Whilst Richter opens E2 with firm confidence, Jochum begins it misterioso: the basses enter piano, partly submerged, veiled and barely shaped (after a diminuendo in bar 80). The mists are only gradually lifted. The level of local activity increases as the section proceeds: other voices shape the subject more actively, especially in dynamics. When the tension threatens to flag in bars 95-96 (which feature more melodic flow and less strongly-pronounced cadences), Jochum brightens the sonority and highlights the bass, with its more active melody. The first soprano’s oft-submerged entry in bar 102 is clearly highlighted, but the crescendo continues to mount afterwards. Only at the beginning of bar 112 – when harmonic resolution is reached – does the performance attain a degree of relaxation.

Overall, then, Jochum projects the entire sequence (bars 81-112) as a single uninterrupted rise of tension. Several other performances follow a similar trajectory (e.g., Scherchen 1950, 1959; Karajan 1952; Klemperer 1961, 1967; Maazel; Rilling 1977;[34] Giulini 1972, 1994), although they usually cover a narrower range of dynamics and colours.[35] Jochum’s uninflected account of the subject, which virtually ignores its internal gestural potential, is consistent with this aim: emphasis is placed on local details only in rare cases, and these can probably be accounted for in terms of the quest for a single, overall shape.

All the performances enumerated above seem to project a +/+ approach (see note [8]): they discern a clear drive towards a climax (or a climactic area) in Bach’s score, and see it as their interpretive prerogative (or obligation) to bring it out in performance (however, see Section 4.4., paragraphs 13-14). They also share several other characteristics: slow tempi; heavy textures; shaping more through dynamics than through articulation. In most cases, the subject in E2 is treated as part of the overall crescendo, not isolated from its surroundings.[36] Most of them do not attempt (or, at any rate, do not achieve) a high level of clarity; the most notable exception is Klemperer 1967, followed by Rilling 1977, Giulini 1972 and Jochum (1957 and especially 1980).

In their almost total erasure of local gestures, however, these performances exemplify an approach which, from a historical-performance perspective, is antithetical to the character of Baroque music generally, and Bach’s in particular. Bruce Haynes, for example, recognises that the “long-line phrase” is often used “to clarify structure”.[37] However, he contends that, in Baroque music, the long line is superimposed upon the music, and acts as

a bit of a “Procrustean Bed,” since it rarely fits the smaller structural units, the Figures and gestures, of which the music is made up. The result are many gratuitous, meaningless crescendos and diminuendos that are misleading to the ear, since they don’t confirm the logic of the music – in fact, they often conflict with its meaning. (Haynes forthcoming, chapter 12)

This characterisation is not entirely suited to readings of the First Kyrie – the performances mentioned above pursue a single trajectory, rather than myriad “crescendos and diminuendos”; they do, however, “iron out” local gestural potential. The mere separation of figures from the overall line, however, does not in itself constitute a gestural approach, as clearly illustrated in Mauersberger’s performance.

In the first period under discussion, however (1950-1980; see note 14), distinctive treatment of figures did seem at odds with large-scale shaping: most of the performances which brought out local details seemed to ignore the movement’s shape. In this respect, too, Mauersberger seems to demonstrate an x/- approach (see note [8]): he seems preoccupied with texture above all. It is harder to apply the x/- label to Richter and Ramin: their forceful bass entries at the beginning of E2 constitute a strong, intentional-sounding gesture – born, perhaps, out of an attempt to reconcile a dramatic conception of the movement with a belief in terraced dynamics and a suspicion of gradual build-up (cf. Schweitzer 1966, I, 362-363).[38] Nonetheless, they largely share Mauersberger’s marginalization of shape in favour of texture, and the isolation of localised figures does seem to come at the expense of overall development.

In later performances related to this tradition – primarily Schreier 1991 and Biller – there is a greater willingness to bring out the movement’s shape. Their meticulously-separated phrases, however, impede the sense of flow, as does their insistence on isolating and highlighting the subject. This latter tendency is especially evident in Schreier 1991, where the alto, first soprano and second soprano entries in E2 are preceded by diminuendi. Schreier thus retreats precisely where other conductors seek to intensify the crescendo.

Schreier’s approach reflects one of the consequences of viewing fugal subjects as anchors of stability (see above): the notion that expositions should be kept relatively stable, whereas episodes with non-subject material (as well as non-subject passages within expositions) could be treated more flexibly. As noted above, this approach implies that the subject should be consistently highlighted in performance, and its phrasing and character should remain unaltered throughout. This conception is illustrated – almost caricatured – in Georg Solti’s performance. Here, the subject is consistently rendered in harsh, unyielding dynamic and articulation, and all its appearances are emphasised; as in Schreier 1991, most subject entries are preceded by diminuendi, and are thus further highlighted. Non-subject material is performed more flexibly, but the subject’s rigidity precludes the projection of directionality within the subject and across fugal expositions.

Other performances feature a more subtle subject/non-subject distinction (see also note 36). In his 1977 version, Rilling projects a continuous E2 crescendo, with no distinction between thematic materials. In 1988 and 1999, however, most forward movement occurs on non-subject material (some of it simultaneous with the subject), without halting the crescendo. Part of the explanation is that the 1977 performance is dominated by dynamic modifications, whereas in 1988 and (especially) 1999, articulation plays a stronger role – especially within the subject (see also below). It is still easier, however, to raise the overall dynamic level in legato passages, which is how Rilling treats most stepwise passages and rhythmic figures like and even in the 1999 version.

A detailed gestural treatment of the fugal subject, however, can affect the projection of larger shapes. Many HIP readings combine a gestural treatment of units within the subject (particularly the projection of moderate local directionality in the Gradatio figure) with a tendency towards partial build-up – local patterns of ebb-and-flow. In these performances (e.g., Leonhardt, Herreweghe 1988, Max, Christophers), the transition to R3 functions, at most, as a local point of arrival, rather than as the culmination of an extended, cumulative rise in tension.[39]

As I noted above, Herreweghe’s 1996 performance features the most distinct shaping of the Gradatio figure as a single rising gesture, at least in E1. This does not translate, however, to a clear trajectory for the movement as a whole. Herreweghe has frequently stated his belief that Bach’s movements are constructed in accordance with the rules of classical speech, with a clear division into six parts:

1. Exordium (introduction);
2. Narratio (presenting the issue);
3. Propositio (presenting the speaker’s thesis);
4. Confirmatio (presenting the main arguments supporting the thesis);
5. Confutatio (refutation of opposing arguments);
6. Peroratio/Conclusio (conclusion).[40]

He reiterates this position with reference to the First Kyrie, stressing in particular the importance of recognising the Confutatio: “in a speech, the confutatio is done with more tension than when one is first exposing one’s theme” (Herreweghe, in Sherman 1997, 282), which presumably means that gestures should be drawn out more emphatically.

However, it is difficult to deduce, from either of his recordings, where Herreweghe believes the First Kyrie’s Confutatio is located.[41] Instead of a concentrated area of higher tension, Herreweghe 1996 features constant patterns of ebb-and-flow within sections, sometimes different ones for different voices (projected primarily through dynamics). The beginning of R2 and R3 are among the moments of higher tension, followed in each case by a subtle relaxation at the beginning of the Fortspinnung. Disjunctions between simultaneous phrase boundaries are clearly exposed – not least in E2, where the focus of attention constantly shifts between different strands of the texture, and is usually drawn away from the subject. Forward momentum – which is already present in the Interlude, with its recurrent if delicate accentuation of the “tug” gesture – is associated with a discreet underlining of small crescendi and accentuations in different parts (e.g., the sustained notes and figure in the alto, bars 93-94; the bass’s cadential “Kyrie ele-” – – in bars 97/8; the second soprano’s sustained notes in bars 99-102).

These figures, and others, are always shaped in a manner consistent with the standard metric accentuation, in terms of accents and dynamics alike, and gently nudge the music forward. In this sense, however, there is no privileging of specific moments or sections; if anything, this activity increases, with stronger accentuation, in the final ritornello, especially in the Fortspinnung. The resulting performance is strongly directional on the level of individual phrases, but does not project a clear overall trajectory for the movement as a whole: local gestures facilitate the sense of flow, but they are clearly given precedence over global shaping.[42]

For a performance of this movement that realises Herreweghe’s stated ideals (clear patterns of tension and contrast, vocal-led lyricism and conductorial restraint), one might turn to René Jacobs’s recording.[43] In this version, all ritornello entries emerge as significant events, and the ritornelli themselves are shaped in an analogous manner (cf. my discussion of Jochum’s performance above): full sonority at Vordersatz; relaxation at beginning of Fortspinnung (or towards end of Vordersatz); insistent unanimity on F2; diminuendo conclusion for Fortspinnung; a fuller and calmer Epilog, concluded with another diminuendo. With the exception of F2’s entries (bars 19, 62, 116), there are no sudden jolts: sections are clearly marked, yet connected and prepared.

Jacobs builds up to a clear climax at bar 102. The crescendo only begins on the alto entry in bar 88. Notwithstanding a slight diminuendo at bars 95-96, the overall trajectory is clear from this point, and the first soprano’s entry (and its orchestral doubling) in bar 102 is strongly highlighted – a gesture all the more marked by the lack of emphasis on the subject in previous entries.

This dramatic gesture notwithstanding, the performance on the whole maintains a peaceful atmosphere. Since all ritornelli end in diminuendo, heightened tension is always followed by relaxation (in R3, there is also a diminuendo towards the Vordersatz’s cadence, in bars 109-111). This is another feature that differentiates Jacobs from the dramatic-organic performances described in the previous section, which usually end with a dramatically affirmative forte.

In both Herreweghe 1996 and Jacobs, gestures are subtly underlined, gently demarcated and allowed to join into hyper-gestures; indeed, Jacobs’s reading arguably approaches the traditional shape. They do not, however, present strongly discrete gestures, of the type which Rilling (quoted above) considered “an encumbrance for complexes of large dimensions”.

A stark illustration of this disruptive potential of localised gestures can be found in Harnoncourt’s 1986 performance. In his writings, Harnoncourt claims that truly valuable music is there “to open [listeners’] eyes, give them a good shaking, even to frighten them” (Harnoncourt 1991, 11), and that performers should maximise these effects. This is intimately linked with his promotion of rhetoric-as-semantics as the key for Baroque music, and his advocacy of a performance style in which each figure is clearly marked and independently shaped (see note 4). In polyphonic textures, such a detailed approach can lead to subtle or blatant clashes within the texture. For Harnoncourt, this is a beneficial outcome; in his view, revealing the music’s complexity, its dialogues and its inner tensions, is more important than mere textural clarity (Harnoncourt 1988, 43-45; 1986, 35-36). In this respect, his approach to Bach’s polyphony is reminiscent of Ernst Kurth’s theories (see above).

This approach clearly informs his 1986 rendition of the First Kyrie, which is dominated by deliberate hesitation and metrical dissonances (see also Section 4.2.4 above). Having established a clear metre, Harnoncourt constantly brings out short gestures which begin on weak beats or between beats, accenting their first notes:[44] this almost always happens in just one or two strands in the texture, and therefore rubs against more regular patterns elsewhere. In a sound stage of equal yet alternating balance (within the orchestra, and between orchestra and choir), this creates a halting effect; overall directionality is consistently questioned.

Harnoncourt appears to drive E2 towards a climax. However, this effect is constantly compromised. The orchestral parts are shaped with much local detail, creating a continuity between the end of the Interlude and the beginning of E2 and maintaining the familiar pattern of metric dissonances. As the choir’s phrasing becomes more distinct (the bass’s entry is partly submerged), more opportunities for clashes emerge. Within E2, most progress towards the climax is achieved on non-subject material, especially in bars 93-96 and 99-102. The short legati in these bars (such as the figure in bars 99-101) emerge as something of a relief. By this point, the orchestra is doubling the choir (and therefore no longer has independent phrase boundaries); strands in the texture do not impede each other, allowing the emergence of clearer overall directionality. The subject itself disrupts this, however, as it maintains its familiar, deliberately disjoint shape.

Finally, at bar 102, instead of allowing the culmination of the crescendo, Harnoncourt places a subito piano for all but the first soprano. This distinctive gesture, affecting the entire texture, assists in clarifying the soprano’s entry, but it frustrates the expectations created by the previous build-up: instead of having a full, confident Vordersatz, R3 has to commence with its own build-up.

Harnoncourt’s utilisation of the disruptive potential of intensified gestures is quite in keeping with his general aesthetics. A similar attitude, albeit with different detailed means, informs Thomas Hengelbrock’s reading of this movement. Hengelbrock describes this movement as an expression of mourning:

The combination of the individual musical elements (sighing motifs, funeral march rhythms, use of out-of-scale notes and large jumps, chromatically intensified exploitation of the thematic span up to the ninth etc.) decode this movement as a funeral chorus, an “actus tragicus” of unprecedented magnitude. (Hengelbrock 1997; cf. Schering 1936b, 10-11)

Hengelbrock’s emphasis on simultaneously-occurring elements is reflected in his performance. He and Harnoncourt bring out many of the same gestures. Hengelbrock, however, employs longer stretches of legato, fewer and lighter accentuations (thus facilitating flow, despite employing a slower tempo than any other HIP conductor), and fewer metric dissonances (Hengelbrock’s local dynamic peaks are usually located on strong beats).

In E2, Hengelbrock focuses attention on the choir. He brings out the subject’s Gradatio, thereby facilitating clear directionality (although other, more continuous figures, mostly occurring between subject statements, are still the primary motivators in the crescendo). However, he clearly separates the Gradatio from the lower sigh, creating an internal stop-and-go effect. His overall shape is an intermittent crescendo similar to Jacobs’s and Jeffrey Thomas’s: a rise in dynamics in bars 90-94 and 97-102 (esp. 99-102), with an interrupting relaxation in 95-96. The soprano entry in 102 is clearly brought out, and gradual relaxation only commences at bar 108. The Fortspinnung is somewhat softer, F2 gently insistent. The final Epilog is expansive (bars 120-123 dominated by an intermittent crescendo in the tenor), with a diminuendo (and pause) in the last two bars.

Though more flowing than Harnoncourt’s, Hengelbrock’s reading is more heavily accentuated than Jacobs’s, and his rendition of the E2 crescendo therefore contains a greater degree of struggle. In Karajan 1952 and Jochum (1957, 1980) – and, more recently, Hickox (see analysis below), Rilling 1999 and Abbado – one senses the arrival of the climax as the dramatic yet inevitable culmination of a continuous striving upwards. In Hengelbrock’s performance, the feeling is that the goal has been reached with considerable strain and struggle.

Even in Harnoncourt’s and Hengelbrock’s readings, the constructive-continuous potential of gestures is demonstrated in their shaping of non-subject material (such as the figure). Rilling’s assumption (quoted above) that a detailed approach to articulation hinders large-scale shaping seems to rely on the precedent of performances like Jochum’s and Karajan’s, where local detail is indeed sacrificed for the shaping of a larger whole. However, other performances (e.g., Gardiner, Brüggen, Hickox, Jeffrey Thomas, Abbado – and, indeed, Rilling 1999) illustrate that a detailed gestural approach, replete with incisive articulation and clear demarcation of gestural components, can actually contribute to overall shaping of large sections.

Hickox’s version provides a particularly striking illustration. Like Karajan, Jochum and others, Hickox shapes E2 as a large-scale crescendo. However, while these performers avoid local gestures, Hickox constantly brings them out. The sense that individual lines are “complicated, with twists and turns” (Haynes forthcoming, chapter 11, quoted above) is strongly projected; yet Hickox also demonstrates how local gestures can be used to support the projection of large-scale shape.[45]

Even in this performance, the most dynamically active moments in his E2 crescendo are on non-subject material; it is probably easier to raise overall dynamics in legato passages. However, Hickox does use the insistent articulation on the subject, as well as the shaping of other voices, to ensure that tension does not flag at any point. He accentuates both the “Kyrie” () and the Gradatio figures within the subject, and seizes on the gestural potential of repeated motifs elsewhere in the texture – in particular, cadential figures and recurrences of the “tug” gesture in the orchestra.

Hickox’s performance thus alternates between progress-through-accentuation (especially when the subject is present) and progress-through-dynamics (especially when the subject is absent). His emphatic gestures (accentuations of individual notes, culminations of small crescendi) coincide, for the most part, with strong beats, and therefore assist in creating forward progression; they are less discrete than Harnoncourt’s, more connected to and leading into each other (the faster tempo and lighter texture further enhances this effect). In this case, then, local gestures assist in ensuring textural clarity and enhance and facilitate the shaping of large-scale patterns of tension and relaxation.

4.4. Summary

In Part Two of this paper, I discerned two conflicting uses of the term “gesture”. Gestural theory, as developed by Hatten, Eitan, Haynes and others, focuses on small phrases; but the word “gesture” has been used to describe larger units in several analyses of the First Kyrie; and it could be argued that the metaphor is useful to describe what conductors like Jochum and Karajan attempt to do in this movement. There is a definite sense of goal-oriented movement in their shaping of E2, which is equivalent to Dickinson’s gestural imagery for this passage and which is not captured by other metaphors I can suggest (e.g., an arch of rising tension). The “gesture” metaphor captures the feeling (experienced, at least, by this listener) that there is an attempt to perform the entire passage in a single, unbroken breath.[46]

All this notwithstanding, there is probably much truth in Bruce Haynes’s claim (personal communication) that “the word ‘gesture’ should be limited to short Figures and motifs; larger units should have a different name for the sake of clarity”. By describing performances like Jochum’s and Karajan’s as “gestural”, one focuses on their apparent attempt to project goal-oriented movement on a larger scale – but one also conceals the locally uninflected character of their phrasing.

Haynes himself acknowledges the fear – also expressed by musicians like Helmuth Rilling – that gestural performance might have a tendency “to atomize, to break up and lose overall comprehensibility without a common thread” (Haynes forthcoming, chapter 12). His own view, however, is the “logical sequence of a series of gestures” (ibid; my emphasis) can actually facilitate the projection of overall directionality (see also the quotes from Hatten in Section 4.1.1 above).

My examination of the First Kyrie’s recorded performances reveal that a gestural approach can have both a disruptive and a constructive effect on larger patterns of tension and release. Even the disruptive effect can be intentional; in Harnoncourt’s 1986 performance, for example, the halting effect is consistent with the conductor’s “tug in the garment” imagery, as well as with his approach to musical performance in general and to the rendition of polyphonic textures in particular.

However, localised gestures – whether “disruptive” or “constructive” – are relatively recent phenomena in recordings the First Kyrie, and seem to reflect the distinctive impact of HIP approaches. Mauersberger and Karajan (1952) represent two poles among “traditional” (pre-HIP) musicians: Mauersberger focused on chiselling out local details at the expense of a static reading of the whole, Karajan ignored local details in favour of overall sweep. It is only after 1980 that one can point to performances of this movement that combined local gestural detail with a projection of the overall shape. In this sense, the performance history of the First Kyrie can be seen as an illustration of larger trends in the performance of Bach’s music (and, indeed, Baroque music generally).

This article focused on one possible explanation for gradual emergence of a gestural performances: the impact of the figurative-rhetorical approach, in analysis and performance alike. However, one must not discount other, complementary explanations. The emergence of rhetoric-as-speech in performance clearly had as much to do with direct practical exploration as with the adoption of a theoretical approach. Gustav Leonhardt stated recently that his style is based more on his direct experience with old instruments than on theoretical study and reflection (in Sherman 1997, 203). Fabian (2003, 207 and passim) argues that rhetoric-as-speech has been revived by performers before it received serious scholarly examination. Their musical (and, in some cases, organological) insights have led them to recognise – and realise in sound – key musical features that were missed earlier, and their performances might well have influenced scholarly research on the subject.

Another impact of the historical performance movement has been the use of smaller choral and orchestral forces; and this, too, might account for many developments, especially in the performance of a complex polyphonic movement like the First Kyrie. Most of the performances recorded before 1980 employed boys’ choirs or large amateur choirs, sometimes consisting of 100-300 singers. This probably made it harder to achieve textural clarity and detailed phrasing in the chorus.

Helmuth Rilling acknowledges this in a recent interview (Parrott & Rilling 2000, 39). He estimates that he has gradually reduced the size of his choir from 40 singers (already a small choir by 1960s-1970s standards) to about 24 singers. He relates this to a rise in professional standards, claiming that he now has stronger, better-trained voices at his disposal, and that this enables him to achieve greater clarity without losing strength. It is interesting to note, in this context, that Rilling had already advocated detailed phrasing, localised dynamic inflections and varied articulations in his book on the B minor Mass (Rilling 1979), and yet he rarely follows his own advice in the 1977 version. The 1988 and 1999 recordings, on the other hand, realise the 1979 recommendations much more fully – partly, one suspects, because the smaller choir made it easier for him to achieve his stylistic ideals.

Such factors should make researchers wary of assuming that recorded performances invariably represent the performers’ aesthetic ideals. This question is particularly vexing when examining the work of conductors (who act through other musicians – choral singers, orchestral players – making it difficult to ascribe specific features to the conductors themselves); there are also difficulties in ascertaining intentionality in recordings, particularly studio recordings, given the powerful impact of the technical team (producers, sound-engineers, etc.) on the final product. In this paper, I worked solely on the basis of these final products, and tried to deduce as much as I could from them.

In some contexts, this problem can be circumvented: as long as the argument does not strongly rely on attributing the interpretation preserved in the recording to the musicians, the recording might be treated “as is” (cf. Johnson 1999, 198). Thus, it makes little difference whether the structural cohesiveness I ascribe to the shaping of the First Kyrie in René Jacobs’s recordings (see Section 4.3.2 above) arises from the musicians’ planned interpretation or from the record producer’s choice of takes. In this particular case, my aim has been to exemplify one option of shaping this movement, and no interpretive-historical significance is attached to the identity of the musicians responsible. Therefore, the important question is whether my analysis convincingly reflects the interpretation as documented in the recording.

On the other hand, questions of attribution do affect my claims regarding the lack of overall shaping in some of Herreweghe’s performances (see Section 4.3.2 above). Since I partly relate this aspect to Herreweghe’s general approach to interpretation (see note 42), it should be mentioned that the lack of cohesion might be the result of editing, rather than conductorial intention. The fact that Herreweghe claims to have taken an active part in the editing process is important, but not necessarily decisive.

Overall, it has not been my intention to suggest that all performances are, invariably and reliably, realisations of a particular theory or of a detailed, pre-set analytic interpretation. To be sure, there are cases where one can show a clear link between theory and practice (the most notable example, among the performers discussed here, is Harnoncourt’s 1986 performance). In other cases, however, such relationships are problematic (e.g., Herreweghe, Rilling) or non-existent: many performers simply do not give detailed verbal accounts of their views on the music or on their role of performers. In such cases, recorded performances, unreliable though they might be, supply the only evidence of their artistic intentions.

This has some bearing on my analyses, particularly with regard to the shaping of E2. In Part Two, I noted that several performances drive this section towards a climax at or around bar 102. However, only one performer (Rilling 1984, 8) has explicitly supported a performance which “presses forward to climaxes”, and spoke of the transition from E2 to R3 as movement’s dramatic peak. The similarity (in this respect and others) between several “symphonic” readings of the First Kyrie (see Section 4.3.2 above) can be attributed to stylistic connections (several performers emerging from similar backgrounds, in terms of performance traditions and repertoire alike).

The partial resemblance, especially in terms of the movement’s overall shaping, between these performances and several recent HIP versions (see especially my discussion of Jacob’s performance in Section 4.3.2. above) cannot be accounted for in a similar manner, but the influence of the symphonic style – and of some of the ideas connected with it, such as the quest for “organic” or “architectonic” shaping for large-scale movements – might still have played a part. It is likely that the musicians responsible for the later performances were acquainted with the older approach (through attending concerts, hearing recordings, or even taking part in performances), but one need not postulate the direct influence of a specific performance or performer. That these performers (and, indeed, recent modern-instrument musicians – such as Abbado, Ozawa, and Rilling in his recent recordings) have also adopted a more detailed gestural-figurative approach might reflect the confluence of several different factors: the influence of period instruments (whose sound has made an impact on many modern-instrument players as well); the use of smaller and increasingly-professional choirs (plus the rising professional standards of period-instrument players); the direct impact of rhetorical theories; and these theories’ indirect impact, as performers who are not necessarily versed in musical rhetoric nonetheless emulate the resulting performance style.

All these factors, and others, have indeed contributed to an increasingly detailed, gestural approach. It should be noted, however, that the use of the term “gesture” itself to denote this style is relatively recent. In several analyses and commentaries on the First Kyrie (especially from the 1920s-1950s), the word “gesture” was used to denote large-scale patterns of tension and resolution; whereas the style which I describe as “gestural” is a later 20th-century phenomenon, and is more often referred to as rhetorical or speech-like.

Rhetorical performance is sometimes associated with an atomistic approach, the dissection of movements and phrases into discrete, separate units. Gestural discourse points to the possibility that the emphasis upon, and inflection of, local units can enhance overall continuity, directionality and expressivity.[47] As I attempted to demonstrate, HIP and HIP-influenced performers (some of whom have consciously employed rhetorical terminology) realised this potential before the introduction of gestural terminology into verbal discourse on performance. Gestural discourse can, however, assist researchers in reaching a better understanding and appreciation of these performers’ interpretations.

4.5. Chronological Discography

Günther Ramin 1950 (incomplete)
Thomanerchor Leipzig, Dresdner Kreuzchor, Stadt- und Gewandhausorchester Leipzig/ Günther Ramin. Thomas-Kirche, Leipzig; July 28, 1950. Original catalogue number: Cantate Dokumentation 640 223; 1 LP, issued after 1950.

Hermann Scherchen 1950
Wiener Akademie Kammerchor, Wiener Symphoniker/ Hermann Scherchen. Vienna, Mozartsaal; October 1950. First catalogue number: Westminster WL 50-37/38/39; 3 LPs, issued 1951. CD re-issue: Universal Victor, Japan MVCW-14015-6; 2 CDs, issued 1998.

Herbert von Karajan 1952
Chor der Gesellschaft der Musikfreunde, Wien; Orchester der Gesellschaft der Musikfreunde, Wien (choruses), Philharmonia Orchestra (arias and duets)/ Herbert von Karajan. Musikvereinsaal, Vienna (choruses) & London (arias and duets); October 26-November 5, 1952. First catalogue number: EMI-Angel 3500 C (35015-6-7); 3 LPs, issued 1954. CD re-issue: EMI Classics 5 67207 2 5; 2 CDs, issued 1999.

Fritz Lehmann 1953
Berlin Chamber Choirs, Berlin Symphony Orchestra/ Fritz Lehmann. First catalogue number: Vanguard Bach Guild BG 527/28; 2 LPs, issued 1953 or 1954.[48]

Kurt Thomas 1955
Choir of the Dreikönigskirche, Frankfurt, Collegium Musicum Orchestra [Frankfurt]/ Kurt Thomas. First catalogue number: L’Oiseau Lyre OL 50094/96; 3 LPs, issued 1955.

Eugen Jochum 1957
Chor des Bayerischen Rundfunks, Symphonie-Orchester des Bayerischen Rundfunks/ Eugen Jochum. Munich; December 1957. First catalogue numbers: Epic (S)C-6027/Fontana CFL1028-9; 2 LPs, issued 1958. CD re-issue: Philips Duo 438 379-2; 2 CDs, issued 1993.

Rudolf Mauersberger 1958
Dresdner Kreuzchor, Staatskapelle Dresden/ Rudolf Mauersberger. Dresden Kreuzkirche; September-October 1958. First catalogue number: Eterna 8 20 074/76; 3 LPs, issued c. 1960. CD re-issue: Berlin Classics 0091712BC; 2 CDs, issued 1996.

Hermann Scherchen 1959
Wiener Akademie Kammerchor, Orchester der Wiener Staatsoper/ Hermann Scherchen. Vienna Konzerthaus, Mozartsaal; April & June 1959. First catalogue number: Westminster WST-304; 3 LPs, issued 1960. CD re-issue: Deutsche Grammophon 471 253-2; 2 CDs, issued 2002.

Robert Shaw 1960
Robert Shaw Chorale & Orchestra/ Robert Shaw. Manhattan Center, New York; June 6, 7, 9, 12-17, 1960. First catalogue number: RCM Victor LM 6157 (mono) LSC 6157 (stereo); 3 LPs, issued 1961. CD re-issue: RCA Victor Living Stereo 09026 63529 2; 2 CDs, issued 1999.

Karl Richter 1961
Münchener Bach-Chor, Münchener Bach-Orchester/ Karl Richter. Munich, Musikhochschule, February & April 1961. First catalogue number: Archiv Produktion ARC3177-79; 3 LPs, issued 1961. CD re-issue: Archiv Produktion 427 155-2; 2 CDs, issued 1989.

Otto Klemperer 1961 (incomplete)
Philharmonia Orchestra and Chorus/ Otto Klemperer. Kingsway Hall, London; December 4-9, 1961. First catalogue number: Testament SBT 1138; 1 CD, issued 1999.[49]

Lorin Maazel 1965
Rias-Kammerchor, Rundfunk-Symphonie-Orchester, Berlin/ Lorin Maazel. Berlin, September 1965. First catalogue number: Philips SPM3-581 (mono), SPS-3-981 (stereo); 3 LPs, issued 1966. CD re-issue: Philips 426 657; 2 CDs, issued 1990 (currently unavailable).

Otto Klemperer 1967
BBC Chorus, New Philharmonia Orchestra/ Otto Klemperer. Kingsway Hall, London; September & October 1967. First catalogue number: EMI-Angel SC-3720; 3 LPs, issued 1968. CD re-issue: EMI CMS 7 63364 2; 2 CDs, issued 1990.

Nikolaus Harnoncourt 1968
Wiener Sängerknaben & Chorus Viennensis, Concentus Musicus Wien/ Nikolaus Harnoncourt. Casino Zögernitz, Vienna; April & May 1968. First catalogue number: Telefunken Das Alte Werk 3-Tel. SKH-20; 3 LPs, issued 1968. CD re-issue: Teldec Das Alte Werk 4500-95517-2; 2 CDs, issued 1994.

Karl Richter 1969a
Münchener Bach-Chor, Münchener Bach-Orchester/ Karl Richter. Main Auditorium of the Bunka-Kaikan, Tokyo; 9 May 1969. First catalogue number: Archiv Produktion 453 242-2; 2 CDs, issued 1996. Also available as part of the 10-CD set “Bach: Sacred Masterpieces”; Archiv Produktion 463 701-2, issued 2000.

Karl Richter 1969b
Münchener Bach-Chor, Münchener Bach-Orchester/ Karl Richter. Klosterkirche Dießen, Ammersee; September 1969. Televised film version directed by Arne Arnbom. ZDF-Unitel. Not commercially available. See also http://www.unitel.de/ucatalog/concert/61_3.htm.

Karl Münchinger 1970
Wiener Singakademiechor,[50] Stuttgarter Kammerorchester/ Karl Münchinger. Sofiensaal, Vienna; May 1970. First catalogue number: London 1287; 2 LPs, issued 1971. CD re-issue: Double Decca 440 609-2; 2 CDs, issued 1994.

Carlo Maria Giulini 1972
New Philharmonia Chorus & Orchestra/ Carlo Maria Giulini. St. Paul’s Cathedral, London; 10 July 1972. First catalogue number: BBC Legends BBCL 4062-2; 2 CDs, issued 2001.

Herbert von Karajan 1974
Wiener Singverein, Berliner Philharmoniker/ Herbert von Karajan. Philharmonie, Berlin; September & November 1973, January 1974. First catalogue number: Deutsche Grammophon 2709049; 3 LPs, issued 1974. CD re-issue: Deutsche Grammophon 459 460-2; 2 CDs, issued 1999.

Helmuth Rilling 1977
Gächinger Kantorei Stuttgart, Bach-Collegium Stuttgart/ Helmuth Rilling. Stuttgart; April 1977. First catalogue number: CBS 79 307; 3 LPs, issued 1977. CD re-issue: CBS Odyssey MB2K 45615; 2 CDs, issued 1989.

Neville Marriner 1977
Chorus and Academy of St. Martin-in-the-Fields/ Neville Marriner. London; November 1977. First catalogue number: Philips 6769002; 3 LPs, issued 1978. CD re-issue: Philips 416 415-2; 2 CDs, issued 1986 or earlier.

Michel Corboz 1979
Ensemble Vocal de Lausanne, Ensemble Instrumental de Lausanne/ Michel Corboz. Temple de Lutry, Switzerland; October 1979. First catalogue number: Erato STU71314; 3 LPs, issued 1980. CD re-issue: 0630-13732-2; 2 CDs, issued 1996.

Eugen Jochum 1980
Chor des Bayerischen Rundfunks, Symphonie-Orchester des Bayerischen Rundfunks/ Eugen Jochum. Herkulessaal, Munich; March & April 1980. First catalogue number: EMI-Angel DS-3904; 3 LPs, issued 1980. CD re-issue: EMI Double Forte 5 68640 2; 2 CDs, issued 1995.

Peter Schreier 1982
Rundfunkchor Leipzig, Neues Bachisches Collegium Musicum, Leipzig/ Peter Schreier. Paul-Gerhardt-Kirche, Leipzig; November 1981, February 1982 First catalogue number: Ariola-Eurodisc 301 077-445; 3 LPs, issued 1983. CD re-issue: Berlin Classics 002123BC; 2 CDs, issued 1997.

Joshua Rifkin 1982
Bach Ensemble/ Joshua Rifkin. Rutgers Presbyterian Church, New York; 31 December 1981 – 11 January 1982. First catalogue number: Nonesuch 9 79036-2; 2 CDs, issued after 1982.

Andrew Parrott 1984
Taverner Consort & Players/ Andrew Parrott. St. John Smith’s Square, London; 4-5 & 10-15 September, 1984. First catalogue number: EMI Reflexe 7 47293 8; 2 CDs, issued 1985.

John Eliot Gardiner 1985
Monteverdi Choir, English Baroque Soloists/ John Eliot Gardiner. All Saints’ Church, Tooting, London; February 1985. First catalogue number: Archiv Produktion 415 514-2; 2 CDs, issued 1985.

Gustav Leonhardt 1985
Collegium musicum van de Nederlandse Bachvereiniging, La Petite Bande/ Gustav Leonhardt. Haarlem; 13-19 February 1985. First catalogue number: EMI/Deutsche Harmonia Mundi CDC16951-8; 2 CDs, issued 1985.

Nikolaus Harnoncourt 1986
Arnold-Schönberg-Chor, Concentus Musicus Wien/ Nikolaus Harnoncourt. Konzerthaus, Vienna; April 1986.[51] First catalogue number: Teldec Das Alte Werk 6.35716; 2 CDs, issued 1986.

Philippe Herreweghe 1988
Chorus and Orchestra of the Collegium Vocale Ghent/ Philippe Herreweghe. Minderbroederskerk, Ghent; April 1988. First catalogue number: Virgin Veritas VCD 7 90757-2; 2 CDs, issued 1989.

Helmuth Rilling 1988
Gächinger Kantorei Stuttgart, Stuttgarter Kammerorchester/ Helmuth Rilling. Kirche der Karlshöhe, Ludwigsburg; May 1988. First catalogue number: Intercord INT 885 855; 2 CDs, issued 1988.

Frans Brüggen 1989
Netherlands Chamber Choir, Orchestra of the 18th Century/ Frans Brüggen. Vredenburg, Utrecht; March 1989 (live). First catalogue number: Philips 426 238-2; 2 CDs, issued 1990.

Anders Eby 1990
Mikaeli Chamber Choir, Drottningholm Baroque Ensemble/ Anders Eby. St. John’s Church, Stockholm; January 22-24, 1990. First catalogue number: Proprius PRCD 9070/1; 2 CDs, issued 1992.

Georg Solti 1990
Chicago Symphony Chorus and Orchestra/ Georg Solti. Orchestra Hall, Chicago; 25, 26 & 28 January 1990 (live). First catalogue number: Decca 430 353-2; 2 CDs, issued 1991.

Robert Shaw 1990
Atlanta Chamber Chorus, Atlanta Symphony Orchestra/ Robert Shaw. Symphony Hall, Atlanta, Georgia; March 5-7, 1990. First catalogue number: Telarc CD-80233; 2 CDs, issued 1990.

Peter Schreier 1991
Rundfunkchor Leipzig, Staatskapelle Dresden/ Peter Schreier. Lukaskirche Dresden; January 1991. First catalogue number: Philips 432 972-2; 2 CDs, issued 1992.

Hanns-Martin Schneidt 1992
Münchener Bach-Chor, Münchener Bach-Orchester/ Hanns-Martin Schneidt. Philharmonie im Gasteig, Munich; March 21, 1992 (live). First catalogue number: Calig CAL 50929/30; 2 CDs, issued 1994.

Richard Hickox 1992
Collegium Musicum 90/ Richard Hickox. St. Jude’s Church, London; 11-13 & 15-16 June 1992. First catalogue number: Chandos Chaconne CHAN 0533; 2 CDs, issued 1992.

Hermann Max 1992
Rheinische Kantorei, Das kleine Konzert/ Hermann Max. Zeughaus, Neuss; 1-6 June 1992. First catalogue number: Capriccio 60 033 2; 2 CDs, issued 1993.

Jeffrey Thomas 1992
American Bach Soloists/ Jeffrey Thomas. St. Stephen’s Church, Belvedere, California; June 15-18, 1992. First catalogue number: Koch International Classics 3-7194-2; 2 CDs, issued 1992.

René Jacobs 1992
RIAS-Kammerchor, Akademie für alte Musik, Berlin/ René Jacobs. Jesus-Christus-Kirche, Berlin-Dahlem; September 1992. First catalogue number: Berlin Classics BC 1063-2; 2 CDs, issued 1993.

Ton Koopman 1994
Amsterdam Baroque Orchestra and Choir/ Ton Koopman. Wallonne Church, Amsterdam; March & May 1994. First catalogue number: Erato 4509-98478-2; 2 CDs, issued 1995.

Harry Christophers 1994
The Sixteen Choir and Orchestra/ Harry Christophers. St. Augustines, Kilburn, London; April 1994. First catalogue number: Collins Classics 70322; 2 CDs, issued 1994.

Carlo Maria Giulini 1994
Chor des Bayerischen Rundfunks, Symphonie-Orchester des Bayerischen Rundfunks/ Carlo Maria Giulini. Herkulessaal, Munich; June 2 & 3, 1994 (live). First catalogue number: Sony Classical S2K 66 354; 2 CDs, issued 1994.

Karl-Friedrich Beringer 1994
Windsbacher Knabenchor, Deutsche Kammerakademie Neuss & Trompetenensemble Läubin/ Karl-Friedrich Beringer. Münster zu Heilsbronn; July 18-23 1994. First catalogue number: Hänssler CD 98.959; 2 CDs, issued 1994.

Philippe Herreweghe 1996
Chorus and Orchestra of the Collegium Vocale Ghent/ Philippe Herreweghe. Abbaye aux Dames, Saintes; July 1996. Catalogue number: Harmonia Mundi France HMC 901614.15; 2 CDs, issued 1998.

Thomas Hengelbrock 1996
Balthasar-Neumann-Chor, Freiburger Barockorchester/ Thomas Hengelbrock. Evangelische Kirche Gönningen; 4-10 October 1996. First catalogue number: Deutsche Harmonia Mundi 05472 77380 2; 2 CDs, issued 1997.

Diego Fasolis 1997
Coro della Radio Svizzera, Sonatori de la Gioiosa Marca/ Diego Fasolis. Cattedrale di San Lorenzo, Lugano, Switzerland; June 3, 1997. First catalogue number: Arts Authentic 47525-2; 2 CDs, issued 1998.

Claudio Abbado 1999
Swedish Radio Chorus, Solisten der Berliner Philharmoniker/ Claudio Abbado. Großes Festspielhaus, Salzburg; March 29 and April 4 1999 (live). First catalogue number: Universal 109 374-2; 2 CDs, issued 2002.[52]

Helmuth Rilling 1999
Gächinger Kantorei Stuttgart, Bach-Collegium Stuttgart/ Helmuth Rilling. Stadthalle Sindelfingen; March 1999. First catalogue number: Hänssler Edition Bachakademie, vol. 70 (CD 92.070); 2 CDs, issued 1999.

Georg Christoph Biller 2000
Thomanerchor Leipzig, Gewandhaus Orchester Leipzig/ Georg Christoph Biller. Thomaskirche, Leipzig; 28 July 2000 (live). First catalogue number: Philips 465 949-2; 2 CDs, issued 2001. Also available on DVD (Image Entertainment ID0667EIDVD).

Seiji Ozawa 2000
Tokyo Opera Singers, Saito Kinen Orchestra/ Seiji Ozawa. Naganoken Matsumoto Bunko Kaikan, Japan; 29 August – 4 September 2000. First catalogue number: Philips 468 363-2; 2 CDs, issued 2001.