‘I think there is a great tragedy in Mahler, and a great tragic sense!’
In an interview with Herbert von Karajan, Richard Osborne, one of the most important of Karajan’s biographers, asked the conductor about the prophetic note in the first two movements of Mahler’s 5th and the entire 6th symphonies. Karajan answered:
“The collapse of a culture and the foresight of everything that was coming is here certainly […]. I think there is a great tragedy in Mahler, and a great tragic sense. His gaiety, though, is often made. After the first two movements, after such great torture, all you can do is let the music flow.” 
These remarks are hardly compatible with the widely-held image of Karajan as a brilliant architect of the most subtle timbral differentiation, emphasizing a fluent, delicate interpretation and crossing the partly extreme expressive gulfs which are so prominent in Mahler’s Symphonies. Rather he seems to be aware of the devastating impact especially of the 6th and 9th (and partly also the 5th) Symphonies, works which he conducted during the last 16 years of his life and especially intensively in the late 1970s and 1980s.
In fact we will see that there are two significant sides to Karajan’s Mahler-Interpretation: on the one hand a distinctive tendency toward an infinite melody (in the meaning of Wagner’s “Unendliche Melodie”) and subtle links, with the fragility and contradictions of multiple musical idioms in his works on the other hand. We will hear examples of both sides, but before I come to that point I firstly have to consider one aspect which is often claimed in Karajan biographies like that of Richard Osborne.
In Gerhard R. Koch’s article Beauty rather than truth. A passing affair: Karajan’s examination of Mahler [Lieber die Schönheit als die Wahrheit. Eine vorübergehende Affäre: Karajans Auseinandersetzung mit Mahler], one of few articles focussing on Karajan as a Mahler-conductor, the title already suggests a highly prejudiced perspective. I want to mention two of his viewpoints, which seem to me symptomatic for a good number of discussions regarding Karajan’s relationship to Mahler:
- For Koch it is a clear fact that Mahler’s extremes of satire and grotesque effects on the one hand and the desperate Outbreak on the other were a strange world for Karajan, for whom “composition was primarily an art of timbre [Klangkunst].” 
- The result of this is that Karajan only began to explore Mahler’s symphonies and songs relatively late in his career, confirmed by a somehow misleading statement by Karajan, who in 1978 mentioned: “I steered clear of Mahler: well considered, because I had not the palette of timbres which I needed to strike the specific Mahler-Sound […] in Mahler the line between the sublime and the ridiculous is very thin … but then I discovered the distinct veiled timbre.”
To prove this, he mentions that Karajan’s exploration of Mahler began in 1973 with the Fifth, followed in 1974 by the Song of the Earth, 1977 by the Sixth, and 1979 by the Fourth Symphonies. In between he performed and recorded the Rückert– and Kindertotenlieder with Christa Ludwig. Finally (as the highlight of his Mahler-exploration) he recorded Mahler’s Ninth twice, one in the studio between 1979 and 1980 and one live in 1982).
Unfortunately, prejudice and partly wrong, partly inconsistent information lead to a fatal alliance which seems to support a certain ideology rather than the real facts. The fact is that Karajan conducted Mahler’s works for the first time in 1955, a few years before the start of the great Mahler-explosion in 1960, and that most interestingly in the United States:
March 12th, 1955: Chicago Orchestra Hall (Berlin Philharmonic Orchestra, Carol Brice, Mezzosoprano: Songs of a Wayfarer
March 22nd, 1955: [same performers] Syracuse, N.Y.: Boston University Lincoln Audi-torium: Songs of a Wayfarer
On February 28th, 1960, he conducted the same work (Wayfarer Songs) with the same performers for the first time in Europe (Berlin, Concerthall of the University of Music). The next day, for the first time he conducted Mahler’s Song of the Earth, which he performed no less than 8 times by 1973 (Koch’s “point of departure” for Karajan’s “Mahler-affair”). And it is perhaps by chance, but a remarkable fact, that Karajan’s last Mahler performances also took place in the United States: the Ninth Symphony with the Berlin Philharmonic Orchestra in New York and in Pasadena, CA.
Before I discuss selected passages from Karajan’s interpretation of Mahler, the explanation of a few further points is required: one to understand the specific position of Karajan’s recordings in general, and one which deals with a few methodological aspects. Although considering recordings of music as “meaningful cultural objects” becomes more and more an important tool for musicological research, recordings are still – as Nicholas Cook pointed out – “a largely untapped resource for the writing of music history, the focus of which has up to now been overwhelmingly on scores.” 
However, to analyse recordings is in no way a generally accepted strategy, especially if one makes use of formalised analytical approaches. Although there is not enough time to discuss this issue in detail, I would like to present one powerful critic against the empirical analysis of recordings. Richard Taruskin writes in his famous book Text and Act: Essays on Musical Performance (1995) that “turning ideas into objects, and putting objects in place of people, is the essential modernist fallacy – the fallacy of reification, as it is called. It fosters the further fallacy of forgetting that performances, even canned performances, are not things but acts.”  Taruskin’s pejorative use of “canned performances” (any studio-productions or recorded CDs) lastly would end the disuse of mechanically reproduced music as a tool for musicological research in general. The following aspects should, however, be considered:
- If we follow Taruskin’s view, wouldn’t we exclude important aspects referring to the text (e.g. aspects of historically informed performance practice, which necessarily must establish a close correlation between text and act)?
- How should composers conducting their own works be categorized (like Wagner, Mahler, Strauss, Strawinsky, Penderecki, Boulez, and so on to name just a few)? Is their interpretation of their own music (act) an instance beyond or separated from the score (text)?
Analysing the interpretation of works by Mahler, we must be aware that a distinction of text and act is very problematic or – as I would say – even impossible. We know that Mahler permanently changed details of his scores as result of performances. In the case of the First Symphony, Mahler made additions, corrections, subtle changes in the orchestration In the score for all of the five performances until the score was printed in 1900. The result of (these?) performances even caused changes during the printing process: Text and Act constitute an inseparable correlation.
If we focus on Karajan’s interpretation, we must consider that for him “the canned performance” was in a way the ideal one, because only here could he make use of subtle balances between the instrumental groups, which were for him, as he mentioned in an interview about his recordings of Schoenberg’s Variations for Orchestra, impossible to realize with such a degree of differentiation in the concert hall. The recording as the final and ideal step in the realization of the text was undoubtedly a paradigm for Karajan’s aesthetic foundation of conducting.
A) Process of Recording (production):
- Karajan’s often extremely intensive recording sessions differ in one important aspect from nearly all other conductors. His influence on recordings didn’t end with the recording sessions and the detailed control (“Abhören”) in the studio, but with the process of “mastering”. What does this mean? Normally we differentiate three major steps within the process of “sound shaping” [Klangbildgestaltung]
- the special characteristics, position, and balance of the microphones, namely:
main microphones [“Hauptmikrophonie”]: recording of the entire ensemble (choir, orchestra, chamber ensemble ….)
spot microphones [“Stützmikrophonie”]: recording of single groups of instruments (e.g. flutes, trombones, percussion….)
Surround ambience microphones [“Raummikrophonie”]: creating virtual acoustic environments or selected spatial effects
B) Process in the Studio (post-production 1):
“mixing”: bringing together the selected analogue or digitally recorded soundtracks to a whole (→ “stem”: summary signal [“Summensignal”), the balancing of the different dynamic levels, reverberation
C) Process in the studio (post-production 2):
“audio-mastering”: general characteristics of the sound (the balancing of single instrumental groups, improvement of the “mixing”.
Normally the conductors’ exertion of influence during a CD-Production ends with B). Karajan is a rare example of a conductor who supervised the recordings up to and including the audio-mastering, as you will see in the following CD-example (a talk between the sound designer of the Deutsche Grammophon Gesellschaft, Günter Hermanns, and Karajan).
The analysis of the following examples is based on four aspects:
- The balance of the instrumental groups
- Tempo and Dynamics
- Phrasing, expression (and its relation to the score)
- The position of a single interpretation within a framework of historical performance practice.
These aspects are of course basics for musical interpretation, although they by far cannot claim completeness. It is for instance nearly impossible to consider aspects which change from performance to performance, like the specific acoustical environment, the relationship between the performers and the audience, or expressive visual gestures by the performers which may or may not influence the interpretation of music.
Nevertheless, as Nicholas Cook pointed out in his article A Bridge Too Far? Musical Performance Between the Disciplines (2011), a methodologically open and undogmatic approach towards musical interpretation which integrates cultural and historical as well as empirical tools and methods is indispensable for the analysis of musical interpretation. Multiple hearings of the music itself, especially for measuring the timing data, was, until about 10 years ago, the most widely used methodological tool for studies in musical interpretation. “With the development of more sophisticated software for working with recordings, however, such as Sonic Visualiser, tempo graphs and other analytical visualisations can be incorporated into the playback environment.” 
In a way, seeing and hearing can – to certain extent – interact and “link measurement to the aural experience”, although the instance of hearing, in my opinion, is of utmost importance. Thus I doubt to a certain degree Nicholas Cook’s opinion that sonic visualisation may “guard against the ear’s tendency to hear what it expects to hear […].”  By all the possible shortcomings, the hearing experience is the central issue in our discussions about musical interpretation, and it doesn’t seem to be a purely subjective approach towards music experience, which is considerably deficient in comparison with empirical methods. Thus I will insistently defend the primacy of the hearing experience, although that doesn’t entirely exclude the use of empirical tools like Sonic Visualiser.
In my first example I will concentrate on the dynamic balance of instrumental groups, comparing two recordings of a certain segment in the 1st movement of Mahler’s 9th Symphony, bars 308ff. which constitute in a way the climax of the entire movement (bars 313ff.). It is preceded firstly by a striking allusion to a passage in the 1st movement of Das Lied von der Erde (Das Trinklied vom Jammer der Erde [The Drinking Song of Earth’s Sorrow]), culminating in the words “Du aber Mensch, wie lang lebst denn Du? (But you, O man, for how long do you live?”):
This allusion is followed by an intense collapsing passage (“Einsturzpassagen” as Adorno defined it), culminating in the main rhythm of the movement, played with most intense energy (“mit höchster Kraft”) in the trombones. There are several aspects to consider when hearing this passage:
- it is the violent recapitulation of the initial rhythm of the movement (Bernstein describes it as a “heartbeat”)
- there is a strong tritone-tension (F-sharp ↔ C) at the culmination
- It is the only time that the tam-tam plays ff (12 tam-tam-strokes in pp follow a few bars later, then there is no more tam-tam heard for the rest of the symphony.)
Most conductors (including Abbado) clearly focus all energies on the trombone-rhythm, whereas the tam-tam-stroke is considerably quieter. Bernstein also emphasizes the trombones, but highlights/emphasizes a short strong tam-tam-beat which is immediately dampened again (Folie xx: Bernstein’s Conducting Score, New York Philharmonic Orchestra, p. 47, Leon Levy Foundation). Karajan follows instead a quite different strategy. He gives the tam-tam the most prominent position, carving the trombones out of it, thus bringing to light the singularity of the one and only fortissimo tam-tam-stroke as a symbol of death and – especially in combination with trumpets and trombones – of utmost horror (as described by Berlioz in his Treatise on Orchestration). Although Karajan’s interpretative approach seems to be unusual, it come closest to that of one whose familiarity with Mahler can hardly be put into question and who conducted the world premiere of the 9th Symphony in June, 1912: Bruno Walter.
However, Karajan’s position within a framework of interpretative concepts is hardly to bring in harmony with the cliché (stereotype?) of a brilliant and perfect sound-designer, so often mentioned in the “Karajan Discourse”, rather it opens up a kind of interpretation of(?) the score of Mahler’s 9th which is somehow validated by Mahler’s autograph itself: here we can see that the tam-tam-stroke was part of the original conception (written in black ink), whereas the rhythmic pattern in the trombones (written in blue pencil) comes from a later step of scoring (a kind of revision which Mahler usually wrote in after completing the preliminary orchestral draft.
The Adagietto from Mahlers 5th Symphony is doubtless one of the most popularized pieces among Mahler’s symphonies. As film-music for Luchino Visconti’s Death in Venice (1971) it became a piece widely separated from its symphonic context, but closely affiliated with the unity of love, death and commemoration: a unity not least emphasized by Leonard Bernstein, who conducted the Adagietto not only as part of the memorial service for his teacher Serge Koussevitzky in 1951, but also during the funeral ceremony for Robert Kennedy in 1968. Undoubtedly, the peculiar reception of the Adagietto is rather ambivalent. Dismissed by Adorno as “shallow sentimentality” and even cautiously criticized by Richard Strauss, whose positive response to the symphony was only “slightly dimmed by the little Adagietto”, the connotation with the sphere of death and parting (as emphasized for instance by Bernstein) has a significant effect on the musical interpretation, especially on the setting of the tempo. Whereas the Adagietto for instance in Willem Mengelberg’s performance from 1926 only lasts a little longer than 7 minutes, in Karajan’s recording from 1973 it lasts nearly 12 minutes (quite the same duration as in Bernstein’s performance with the Vienna Philharmonic Orchestra → Handout 4).
Analysing the part of the First Violins, one can clearly see that all conductors mentioned in Handout 4 considerably reduce the tempo during the three eighth-notes in bar 2, thus hardly focussing it as an upbeat, but as a unity in itself, which is more or less separated from b. 3 (especially in Solti’s and Bernstein’s interpretation). Whereas Karajan, Solti, and Bernstein make extensive use of Rubati, Mengelberg keeps the tempo from b.3 onwards (with a few exceptions) in a range of Tempo 96 – 108 per eighth-note. In Abbado’s performance from 2004 we can see an increase of tempo from b. 5 onwards until the next entrance of the main theme in the Celli, b. 10. Another remarkable result of this comparative analysis of interpretations is the fact that Mengelberg, who was in close contact with Mahler and after the latter’s death was aside from Walter THE authority in reference to Mahler-interpretation, avoids the breath marks indicated in b. 3, 4, 6, and 7 (Mahler would have called it “Luftpause”) in favour of a continuous flow of the melody. Karajan emphasized even more the breakless melodic unity, more or less hiding the breath marks. A different situation can be found in Solti’s recording. Although he extends the first eighth note-group in b.2 considerably, he keeps its upbeat function clearly in b. 3 and 6 and thus comes quite close to Mahler’s indication of the breath marks.
The most striking difference between Mengelberg’s performance and all the later ones mentioned in Handout 4 is the extensive use of portamenti. As early as 1905 some violinists criticized the increasing influence of operatic pathos in violin playing, especially the “intolerable vibrato” in combination with a badly executed portamento as the “mortal enemy” of proper way of playing the instrument (“der Todfeind jedes gesunden Musizierens” ). Clive Brown rightly summarizes that “the prominent portamento which was still common during the early decades of [the 20th] century has now gone out of fashion. Where it is still used, it is executed much more discreetly because it is generally performed with a faster movement of the left hand and a decrease in bow pressure.”  Thus an expressive device which was allowed in the late 18th century only “from time to time” for the solo player (e.g. in Johann Friedrich Reichardt’s treatise Üeber die Pflichten des Ripien-violinisten [On the Duties of the Violin-Ripieno Player] from 1776) and a few decades later decried by Antonio Salieri, criticizing a violinconcerto by Antonio Lolli as imitating “now a parrot, now a dog and now a cat”, became more and more popular, especially in the time of Louis Spohr and the violinist Pierre Joseph Rode (~ 1810 – 1820).
However, again a few decades later in Manuel Garcia’s A New Treatise on the Art of Singing (London 1855) as well as in Charles de Beriot’s Methode de violon divisee en trios parts from 1858, the use of portamento was restricted, as both of them suggested “different speeds and intensities of portamento”, namely an increasing motion in the higher rather than in the lower part, and again about 50 years later, e.g. in the Kunst des Violinspiels by Hermann Schröder (1902) the portamento (as a speciality of the French School of Violin-Playing) was denounced as a “perverted mannerism”. Nevertheless, the portamento was quite common in orchestra performances until at least the late 1920s and Mahler’s detailed instructions for its execution, e.g. in the last movement of the Ninth Symphony.
For Karajan, as for many other contemporary conductors, portamento was something to avoid as much as possible. For Karajan – as can be seen in a rehearsal of the 2nd movement of Mahler’s 5th Symphony – portamento was closely associated with light music, and thus limited to a minimum (more or less a light slur at the very end of the indicated note), whereas Mahler clearly wanted a continuous slur from one note to the next, as can be seen in Ex. 6, b. 6, 8, and 9 from the final Adagio of Symphony Nr. 9. Regarding the execution of portamento in historically informed performance practice here is in definite conflict with an aesthetic paradigm shift, the latter going back not more than about 80 or 85 years. If you compare for instance Karajan’s and Bruno Walter’s interpretations of the beginning of this movement one can easily hear the meticulous realization of Mahler’s indications in Walter’s recording from 1938.
Although Karajan’s performances and recordings of Mahler’s symphonies by no means represent an unambiguous concept of musical interpretation, it seems to be much more a prejudice than reality to reproach him for an unbroken and smooth (or – between the lines – superficial) timbral palette. A responsible evaluation is still lacking in a thorough analysis of his interpretations. The 25th anniversary of his death in 2014 will hopefully be a stimulus for further and more detailed investigations into this matter; however, the outline of my contribution could possibly focus on a few, but – in my opinion – central facets of Karajan’s approach to Mahler.
First of all, Karajan demonstrates that the aspect of timbre (so often disparaged as of minor relevance for musical interpretation) is indeed one of the most important cues for the expressive content of Mahler’s music. Quite often it is this quality of his scores which demonstrates the tragic sense in his music (as claimed by Karajan) most prominently. In this respect, Karajan’s recordings represent not only an unreservedly con-vincing interpretation of the expressive impact of Mahler’s music, but also one which demonstrates a remarkable affinity towards the historical roots of Mahler interpretation. On the other hand, Karajan’s reading of the Adagietto provides us at first glance with a disposition of the tempi emphasizing a quite slow and heavy melodic flow, often ignoring the breath marks. However, we must consider the special position of this movement within the context of the symphony: a symphony for a large orchestra, and here – in the Adagietto – a movement for strings and harp only.
There is hardly anything comparable within the scope of his other symphonies, but one could find parallels in his songs, for example in segments and especially the end of the Rückert song “Ich bin der Welt abhanden gekommen”. In fact, the problem of widely different tempi adopted for the Adagietto is closely connected with the concept of a “song without words” on the one hand (Mengelberg, Walter), or a meditative farewell, emotionally charged by a remarkable amount of tempo changes (Karajan, Bernstein) on the other. Thus the intentional vocal line (as expressed by no small measure by the breath marks) is transformed in Karajan’s interpretation to a more ballad-like development of a narration, focussing at this early stage of the Adagietto on vocal/melodic continuity in favour of the expressive content of each word or syllable: a kind of interpretation, which has – maybe not by chance – a certain affinity to the realm of renunciation and resignation (as prominently represented for instance in the film-music for Luchino Visconti’s Death in Venice, dating also from the early 1970s). Indeed, the question of whether the Adagietto is a kind of “instrumental sister” to “Ich bin der Welt abhanden gekommen”, or a piece of its own individual right is discussed controversially among Mahler-scholars.
Karajan’s decision, however, seems to be clear: for him it is quite the same expressive quality as in the Rückert-song mentioned before: withdrawal into a world of farewell and loneliness, as well as a tragic sense so prominently exposed in Karajan’s understanding of Mahler’s music.— Peter Revers
 Richard Osborne, Karajan – A profile and an interview, in: Grammophone, April 1978, p. 1687.
 Gerhard R. Koch, Lieber die Schönheit als die Wahrheit. Eine vorübergehende Affäre: Karajans Auseinandersetzung mit Mahler, in: Jürg Stenzl / Lars E. Laubhold (Hrsg.), Herbert von Karajan (1908 – 1989) : der Dirigent im Lichte einer Geschichte der musikalischen Interpretation, S. 87. Also Peter Uehling claims that Mahler and the Viennese school are a repertoire never conducted by Karajan before 1972 (Peter Uehling, Karajan – Eine Biographie, Reinbeck bei Hamburg 2006, S. 157.
 Nicholas Cook, Methods for analysing recordings, in: The Cambridge Companion to Recorded Music, ed. by Nicholas Cook, Eric Clarke, Daniel Leech-Wilkinson and John Rink, Cambridge etc. 2009.
 Richard Taruskin, Text and Act: Essays on Musical Performance, New York 1995, S. 24.
 Nicholas Cook, A Bridge Too Far? Musical Performance Between the Disciplines, in: Heinz von Loesch und Stefan Weinzierl (Hrsg.), Gemessene Interpretation, S. 8.
 Andreas Moser, Vom Vortrag, in: Joseph Joachim / Andreas Moser, Violinschule in 3 Bänden, Berlin 1905, S. 34.
 Clive Brown, Bowing Styles, Vibrato and Portamento in Nineteenth-Century Violin Playing, in: Journal of the Royal Musical Association, Vol. 113, No. 1 (1988), S. 1119ff.)