Why Mahler sometimes needs to be updated

Why Mahler sometimes needs to be updated


norman lebrecht

January 25, 2022

We reported yesterday that Universal Edition had issued a heavily revised edition of the fourth symphony, taking in the composer’s final corrections from 1911.

It had previously issued a similarly upgraded edition of the fifth and second symphonies.

One of our readers describes the difference:

In one of the first years of this century, I was spending a few days at the Bregenzer Festspiele. For those who don’t know, Bregenz is the summer home for the Wiener Symphoniker, just as the Philharmoniker goes to Salzburg. They both perform the same functions: play some concerts and mostly serve as pit bands.

On this particular trip I was invited to a Symphoniker concert which announced only Mahler’s Sinfonie Nr. 5. I thought there must be something else. Surprise encores? (Actually, the one and only time I heard an encore after a Mahler symphony was a sad, sad event with Barenboim/Chicago at Carnegie Hall.)

As the concert approached I learned we were going to hear the world premiere of a new critical edition prepared by Dr Reinhold Kubik of the Internationale Gustav Mahler Gesellschaft.

For about 40 minutes, Dr Kubik stood in front of the orchestra and gave a fascinating lecture about his work on the new edition. There were over 700 corrections made from consulting every available source, including Mahler’s own marked-up scores and archival material from all around Europe, including (and I find this especially fascinating) the timings of each movement when Mahler conducted it himself, including performances at the Concertgebouw.

We were treated to “before” and “after” examples of the symphony. Some of them were jaw-droppingly astounding.

We then had an intermission and the premiere of the complete new edition.

I met Kubik afterwards and discussed his projects at IGMG and the performance we just heard. I remarked that I thought the adagietto was played a bit too fast, and Kubik explained to me why he though it too slow. Using archive sources like the Concertgebouw timings, it was determined that Mahler performed it between six and seven minutes. The performance that evening had taken a bit over nine minutes. I asked about some of the legendary recordings, naming in particular HvK and the later Bernstein, both of which drag it out to about 11 minutes. Kubik said, “Anything over 10 minutes is perverse” and sent me to the Bruno Walter NYP broadcast from 1947 which clocks in at 07:35. A good example of how not to lead it can be found on James Levine’s RCA recording (Philadelphia Orchestra) which lasts over 12 minutes.

Since then, when I hear Nr 5 I can actually tell if the orchestra is using Dr Kubik’s new edition, or an older one. These things do matter, especially if you are lucky enough to have someone illustrate the significant changes for you.

I believe the Rattle/Berliner recording was the first to use the new edition. Even so, his adagietto was played at about the same tempo as the Symphoniker had in Bregenz.

I look forward to hearing the “new” Nr. 4.



  • Mahler is genius says:

    The main issue in that case is: if you are a conductor and are going to present Adagietto for an audience, will you do it in 7 minutes or in 10? Which is the best for the music itself? I highly doubt that the fast rendition is good for emotional connections. It’s like a tradition – playing this movement slow.
    You may say – well, then you’re not a good conductor – first, do what written, and then what you feel.
    But as we all are different human-beings I will insist on hearing the beautiful (perhaps not as Mahler wanted) version of Adagietto. Which, of course, could last 7 minutes if it’s convincing enough.

    • Anon says:

      Hermann Scherchen took 15 minutes. Is that the record?
      Philadelphia Orchestra, 1964. https://www.youtube.com/watch?v=5zV1Th46MyA
      Over twice the time taken by Mengelberg, Concertgebouw, 1926, https://www.youtube.com/watch?v=2HQpJdORX6w
      I prefer it when it is not so slow that the melodic line disappears.

      • Kenny says:

        I believe Haitink/BPO for Philips is 16′. There’s a book with all this stuff, which I can’t find right now, but it’s way out-of-date now given the fast and furious — to choose a phrase at random — new recordings come out.

    • John Borstlap says:

      Here is a superb take on the Adagietto by a young, truly gifted conductor, not too slow, not too fast, and with an entirely natural flow which is the most important aspect:


      • Peter San Diego says:

        A wonderful find, Mr. Borstlap! Many thanks. The Adagietto should be a love letter, not a dirge, and this conductor gets it just right. Frans Aert-Burghgraef seems worth going out of the way to hear. (I also note that Edo de Waart thinks highly of him.)

      • Kenny says:

        I’d like it a lot better if that were a violin and not a viola. And how is this a “rehearsal”? What’s being rehearsed? (Is that an excuse for his wearing a T-shirt and jeans?) Sounds like the rehearsals have already happened, and successfully.

  • John Borstlap says:

    All of this simply shows that music cannot be notated in a fixed way, as language can.

    Also, it is questionable whether a ‘perfect’ notation is possible at all, or a ‘most reliable’ one as ‘fixed’ by the composer, in various rounds of corrections. Since he can be mistaken as well as the correcting musicologists, who is going to decide? Still, the composer is the best source, with the caveat that his insights in his own music may be flawed.

    A telling example of this latter case is how Stravinsky looked at his early Firebird music when he was old. At the end of the score, there are a number of tutti chords with an accent, which – in the style of the original music and with the background of Rimsky’s music – clearly mean: heavy, substantial blocks of sound. When Stravinsky revised the music, he turned these chords into short, sharp staccati and he performed them thus himself, in his later days. And they don’t sound right at all, stick out of the aesthetics of the music. This was because S had developed from a late romantic composer towards a modernist, ‘objective’ one and heavy chords sounded much too dramatic and ‘kitschy’ to him. So, he was wrong with revising and performing the end of the piece. Another telling Stravinsky example: his own conducting of the ballet Apollon Musagètes (recording) is wooden, heavy, downbeat and ‘matter of fact’, and the ‘soul’ of the music: its dreamy, elegiac and elegant style, and sinuous melodic lines, are lost: bad interpretation. The recording of St Martin in the Fields under Marriner gives the music its radiant beauty.

    Another example: Debussy’s Nocturnes (1894), one of the most sophisticated scores in the repertoire with an abundance of refinement and differentiation, gave immense trouble for the composer in terms of minimal corrections. Every time when he conducted them himself, he changed little things and was never happy with the result. When, around 1912, a conductor asked him which version he would consider best, he gave him his own marked conducting score which was covered with numerous pencil marks in 3 different colours, and said: “Look here, I no longer know, just do what you think is best.” When going-down to minute details, differences no longer have a real meaning as to interpretation, the more detailed, the more possibilities with equal meaning.

    Wagner had trouble with his tempo indications, and learned the hard way: you cannot really fix a tempo with metronome numbers or a detailed description, and hence his very general instructions which leave much room to interpretation, which also depends upon type of players, acoustics, relation between stage and pit, etc. The same margin of practical circumstances with concert music.

    Brahms found-out that the more details he wrote down, the more misunderstandings, so he resorted to notate very economically, leaving much room for interpretation. He approved of different interpretations of his music.

    • David K. Nelson says:

      And of course one of old Igor’s other motivations in revising the scores (apart from making them easier for him to conduct) to his most popular ballets was to try to put them back under copyright. But there was nothing he could do about the proliferation of scores and parts that were still P.D.

      I am wandering into territory where perhaps I have no business being, but this matter of tempo and what is “correct” has long interested me, since I seem to be attracted by tempo extremes, whether it be what Herman Scherchen does with the Eroica Symphony (that last movement!) or what Bernstein does with the last movement of the Pathétique Symphony.

      Just maybe composers do not know best. Some of the finest and most beautiful moments in Stokowski’s recordings of L’ascension by Messien come when he simply disregards the tempo and dynamic markings, for example, to extraordinary effect.

      Although many these days seem to want to decry the move away from gut strings, the one thing that can be said about “modern” strings is that they hold pitch longer, which in addition to the fact that they don’t break as easily is why musicians, however reluctantly, made the switch. This difference in turn can encourage slower tempos. Put another way, you don’t want sagging pitch to ruin the final measures of the Adagietto, and one way is to get on with it.

      If it is objected that today’s period instrument orchestras hold pitch perfectly, well, I’d say 1) yes, on recordings to be sure and 2) even “live,” they are now playing in climate controlled halls which itself is a very 20th century development. When Brahms was writing his violin concerto Joachim did express reservations about how certain solo passages would sound in an overly-warm hall. I am sure he was thinking of some first movement double stops which make that concerto the challenge it is.

      I wonder too if larger and more resonant halls do not encourage (or at least, do not punish) slower tempos.

      And the last of my speculations is the role of recordings where you can listen and compare and find possibilities you never dreamed of contained on the recording you just heard for the first time. Mahler had a tempo in mind for his Adagietto, probably arrived at sitting at the piano, but what if he had actually heard other tempos, a whole range of other opinions and decisions and extremes? Bernstein was a child of the recording age in ways that Bruno Walter was not.

      Additionally, recordings allow for tempos that might induce distracting audience noise and restlessness in the hall. John Borstlap mentioned Debussy’s Nocturnes. Every concert performance I have heard of them was marred (to my ears) by audience noise, even if it was no worse a level of audience noise that I’d hardly notice during a Tchaikovsky Symphony. Recordings let me hear the Nocturnes the way I prefer. I do think audience noise or the risk of audience noise can affect a conductor’s choice of tempo. For whatever reason, after X minutes of music some people just start getting restless or noisy.

    • “…This was because S had developed from a late romantic composer towards a modernist, ‘objective’ one …”

      I recall being told in school that the real reason he made these changes was to have something to put a new copyright on.

      • John Borstlap says:

        Yes, but that is only part of the truth. He made revisions and corrections in the early scores because he wanted a better notation, including the Sacre (the notation of the Danse Sacrale never satisfied him). Petrushka’s later edition is with less instruments so that the average orchestra can play it; the original score of 1911 is for an unusually large orchestra. Most revisions concern a ‘thinning-out’ of the textures.

  • Monsoon says:

    “Some of them were jaw-droppingly astounding.”

    Pretty infuriating to make this claim but then not provide an example.

    “Since then, when I hear Nr 5 I can actually tell if the orchestra is using Dr Kubik’s new edition, or an older one. These things do matter, especially if you are lucky enough to have someone illustrate the significant changes for you.”

    If someone has to “illustrate the significant changes for you,” then I question just how significant they really are.

    As for Mahler’s preference for the adagietto to be played fairly quickly, that’s been well known since forever.

  • phf655 says:

    During Mahler’s lifetime it was the custom and style for the strings to play with Portamento (basically, sliding from note to note). The strings were probably still made of gut, not steel as used today, which resulted in a different sonority. Portamento is no longer tolerated, except in very limited cases, and the sound made on gut strings doesn’t carry well in larger halls. None of this is mentioned in the discussion of the proper tempo for the adagietto. In short, recreating original performance style is very difficult, often futile, and may result in something that doesn’t appeal to contemporary sensibilities.

    • John Borstlap says:

      If contemporary sensibilities don’t relate to older music, they should be changed, developed, refined, and adapted, not the other way around. Music is a living thing and will appeal to anybody who is still living her/himself, but works should not be adapted to the lack of understanding of the listener.

      When math is taught at school, it is not adapted to the ignorance of the pupils. The same with language. These things have to be learned, effort has to be invested, and the rewards come later. But of course, in an egalitarian society where such common sense values of adulthood are eroding, culture is often brought down to the level of the lazy ignorati.

  • Peter says:

    It is my understanding that Mahler intended the Adagietto to be a love song to Alma, and not a funeral dirge. If one were to imagine the melody being sung, that might explain why the “faster” tempo works better for one’s breaths between the phrases.

    While it’s wonderful to have editions that bring us closer to Mahler’s intentions, the costs of the score and rental of the parts from the publisher is so expensive that I think many orchestras will stick to the older editions that they either already own or can purchase more inexpensively.

    • Barry Guerrero says:

      Who says a love song can’t be slow, or that a funereal dirge can’t come closer to a walking tempo? It’s subjective silliness with no real facts to back it up, other than the tempo descriptions Mahler wrote in his scores.

  • True North says:

    I would think that most of the edits in the new critical edition are to do with orchestration rather than anything structural. Mahler was notoriously insecure about his orchestration. I recall reading in the Malte biography about an episode where he insisted that if after his death someone could orchestrate his music better, it was their duty to do so. If Mahler had lived to be 100 years old, he would no doubt have continued tinkering with these minor details for several decades more. And to what benefit? There will never be a “perfect edition” that finally reveals the grand vision we have somehow been missing all along. What matters is the commitment of the orchestra and the conductor to the music on their stands … regardless of edition.

    • John Borstlap says:

      The problem is that Mahler was an extraordinarily gifted orchestrator. Also the orchestration of his earliest scores is superb. But he developed, got more & more complex and subtle, and yes, then you are never really happy. So ‘insecurity’ is a rather generalized notion with Mahler and with some nuances to be made.

      Also there are some surprising aspects of his scoring, for instance the 1st mvt of nr IX is ultra-differentiated, but the next mvt is almost rude in its peasant style, painted with the biggest brush while the first mvt is extremely refined and differentiated. The subtleties of the 1st mvt cannot all be rendered adequately, simply because they are too complex and lead to suppression of details instead of clarifying them. You can hear that conductors struggle with this movement.

  • Rich C. says:

    My rule of thumb: Slow movements sound slow no matter how fast they are taken, and fast movements sound fast no matter how slow they are taken.

  • GM says:

    As a point of fact the International Mahler Gesellschaft critical edition of the Fifth Symphony (edited Kubik, 2002) is issued by the work’s original publisher Edition Peters, who also handle the Sixth Symphony (2010, originally published by C.F. Kahnt) and the Kindertotenlieder.

  • DML says:

    I remember Richard Rodney Bennett saying to me that he didn’t mind what interpretation performers gave to his music as long as they played it at the correct tempo, ie his careful metronome indications.

    • norman lebrecht says:

      RRB knew his business and was very nice about it.

    • John Borstlap says:

      As said above, metronome marks are dangerous business. In different halls, with different players, and at different times of the day, tempo may appear differently. It is not a scientifically fixable thing. Music of different types has its own intrinsic right tempo, but dependent upon circumstances.

      The problem with Beethoven’s metronome numbers is well-known; they are too fast and probably he read the number at the wrong side of the little block.

      Wagner stopped giving metronome numbers when he found-out they could not indicate the right tempo at every performance.

      Ravel notated everything meticulously including metronome numbers but was often unhappy with performances, in spite of all his indications been followed literally. His own conducting was dull and unimaginative.

      Debussy made a mistake with his metronome number at the beginning of Iberia (too fast) with the result that a ‘loyal’ performance sounds cramped and hurried for that movement.

      Etc. etc….

      But for works which does not depend on musical interpretation but on precise sound bites, exact metronome numbers may be the right thing, although they may not mean very much (Boulez, Xenakis, etc.)

  • Alexanderh says:

    I don’t think the editorial changes justify another edition. The reason behind it is, with no doubt, a desperate effort by UE to continue generating income from a composer that is already in the public domain (let’s not forget this new edition will be copyrighted for years).

    UE used to be a good music publisher, but now with Scodo I have no respect for them: composers are paying Universal/Scodo 15-50 EUR/month for the “privilege” of being published on under demand. Composers have to surrender Universal 30% of all benefits after taxes and any expenses by UE. For Universal it means zero risk and a guaranteed income, and for the composer the possibility that, if their piece is performed or the score is sold, they get some income; Universal always wins: if the piece is sold they get 30% after any expense, and if it isn’t they still get the subscription from the composer. People paying for being published… I don’t get it!

    But as I said, zero respect for UE, and I will buy my scores from publishers other than UE whenever possible.

    • Franz1975 says:

      It is certainly shameful. I see lots of composers claiming “being published by Universal Edition”. Thanks to Scodo being published by Universal does not mean anything important anymore.

  • Max Raimi says:

    As my good friend Miles Hoffman once remarked, “Composers don’t write urtexts–they write music.” I know that I often revisit the music I write, and consider revising it. I never look at the same piece twice with exactly the same ideas as to what are the ideal solutions to myriad problems. Perhaps the revisions are an improvement, perhaps not. Mendelssohn spent years fussing with the “Italian” Symphony, and would have ruined it for all time if his earlier version hadn’t turned up, a fascinating saga that the inimitable Gerard McBurney explored in one of his “Beyond the Score” presentations with the Chicago Symphony years ago.
    Having not heard the “new” Mahler, I am in no position to evaluate whether it constitutes an improvement. But the old Mahler is pretty good.

  • Matthias says:

    You have to think there is a lot of interpretative leeway when it comes to the Mahler 5 Adagietto, since Gustav Mahler himself marked it as “sehr langsam” – “very slow”, but didn’t conduct it that way. So who’s right – Mahler the conductor or Mahler the composer?

    In any event, historical research should be used to inform performances, not to limit the possibilities of artistic expression.

  • Barry Guerrero says:

    Color me perverse, as 9.5 to 10.5 minutes works best for me. I don’t like the ultra-fast Adagiettos, regardless of what Mahler’s ‘intentions’ might have been. But here’s something to harp about (pun intended).

    The word ‘langsam’, and even ‘sehr langsam’, are applied several times in the Adagietto. Yet, people scream bloody murder if you dare to go slower than Walter or Mengelberg (I don’t like either rendition, personally). Yet – YET – the slow of Mahler 6 is marked “andante moderato”, and the word ‘langsam’ isn’t written even once! Where’s the outrage when conductors turn the slow movement of M6 into a Brucknerian adagio?

  • Anthony Sanderson says:

    I understood Mahler could adapt scores to suit the hall he was conducting in, so I wonder whether there really can be a universal score for every venue. However, these projects all seem very interesting.

  • Barry Guerrero says:

    So just as an f.y.i., here’s what is actually in the Philharmonia pocket size score of the Adagietto movement from Mahler 5. The head tempo description is “sehr langsam”, this is quickly followed by “molto rit.”, and then the words “a tempo (molto Adagio)”.

    At bar 8, we pick up the tempo a bit: “Nicht schleppen ([in German] somewhat more flowing than at the beginning)”. Then bar 21 “rit.”; bar 23 “Wieder ausserst langsam”; bar 28 “etwas draengend” (somewhat rushing ahead); bar 30 “fliessand”, followed by yet another slow down: “zuruckhaltend”.

    At rehearsal figure 2 (bar 39), we come to what is the ‘development’ section – a contrasting “B” section. Here Mahler marks “Fliessender” (more flowingly). Here there are indications to push ahead through this developmental area: “etwas draengend” (just before the key change into Gb major), then “Fliessand”. However, through the most climactic bars of this passage, Mahler simply writes “espress.” (espressivo).

    When we get to the recap (which sounds like the beginning), Mahler again writes “Molto Adagio” (Tempo I). Then before the climax at the end, Mahler writes to go even slower: “rit.”, followed by “Noch langsamer” (slower still). The ending bars do have “draengend” written above, but that’s also tempered by his calling for “viel Bogen wechseln” in the descending quarters notes in the strings (change direction of the bow frequently). He also writes “breit”, which means “broad”. The ending measure has a hold marked over it, followed by instructions to go “attacca” into the finale.

    • John Borstlap says:

      I think from all of this can be concluded that Mahler did not love his wife at all, but only an idea of her.

    • Peter says:

      Anyone who can read a score can see what you’ve just written. But how about investing in a well-researched book about Mahler, his symphonies, and their background such as Constantin Floros’ “Gustav Mahler: The Symphonies.”
      Floros writes: “In the score he used, the conductor Willem Mengelberg made some notations that are of inestimable value for a deeper understanding of the Adagietto. He wrote: ‘This Adagietto was Gustav Mahler’s declaration of love for Alma! Instead of a letter, he sent her this manuscript without further explanation….’ ”

      Mengelberg knew Mahler well, so I doubt what he wrote was subjective silliness. Music is more than just tempo directions and dynamic indications.

      • Barry Guerrero says:

        . . . to which I would add, a love song to one’s wife doesn’t necessary mean to pick up the tempo. Eleven minute renditions of the Adagietto have been recently denigrated as being ‘funereal’. Yet, Chopin’s famous funeral march from his second sonata usually goes along at a slow walking tempo. Regardless, I’m simply reporting what’s written on the page. What that means to you, or doesn’t mean to you, is of no interest to me. The opening and the return of the “A” section both have “Molto Adagio” written over them. Clearly there’s a disconnect. What tempo modifiers there are, go no farther than “fliessand” and “fliessender”. Nowhere does it say, “please deliver this movement home in less than 9 minutes, would you!”. Regardless of whether Walter and Mengelberg are both ‘correct’, I think it sounds rushed. Yes, that part is subjective. So is the notion that a love song to one’s wife can’t be slow. Ever listened to Elvis Presley’s “Love Me Tender”? . . . that doesn’t exactly move along. Have ever looked at the notes that Mengelberg wrote over the first bars of his own score to Mahler 4? . . . they’re ridiculous.

        If you can’t get through those four bars without tripping over yourself, you’re in trouble. But more to the point than any of this, is that the very same people who cry ‘bloody murder’ when a conductor dares to go over 10 minutes with the Adagietto (even though the score provides them with plenty of justification), are the very same people who say NOTHING when the slow movement to Mahler 6 is made into a Brucknerian styled Adagio (which it’s not).

        The word ‘langsam doesn’t appear even once in that particular movement. Conductors sometimes stretch it out to 18 or 19 minutes (the norm runs between 14 and 16 minutes, with some conductors going as ‘flowingly’ as 12.5 minutes). No one says anything! Regardless, this whole topic of the supposedly correct tempo for the Adagietto is not a new one by any means. Gilbert Kaplan addressed it quite thoroughly back in the middle 1980’s. I met Gilbert Kaplan sometime in the 1990’s and have a lot of respect for his work. I just don’t agree that a 10 or 11 minute performance of the Adagietto is too slow, or is ‘funereal’ sounding . . . . a subjective reaction? Yes. It’s ‘beyond the notes’ (which anybody can do).