Mahler: Symphony No. 5. BBC Symphony Orchestra, Pierre Boulez live 28 August 1968. Nuova Era “Reprints” 2326 (1989)

Mahler: Symphony No. 5. BBC Symphony Orchestra, Pierre Boulez. Nuova Era “Reprints” 2326 (1989), no barcode

Recorded live, 28 August 1968, at the Proms



Boulez the other way around
Originally posted on, 25 December 2011

For a long time, Mahler by Boulez was only a tantalizing and frustrating plea by esteemed French critic (in the French magazine Diapason) and Mahler biographer Henry-Louis de La Grange. Invariably, whenever a new recording came out, he’d finish his review, or so it seems, by a comment to the effect that, “yes, it is a fine version, but ah, compared to Boulez whom I heard conduct it live with the BBC or the London Symphony Orchestra in 197 and something, it is nothing”.

So of course, one waited for Boulez to record Mahler like the Jews await the Messiah or the muslims awaited for Muhammad to receive the revelation. Needless to say, when it finally happened, with his cycle for DG, it was slightly disappointing. Some of it was great, some of it was plain, and some of it was even downright wrong-headed (I’ll substantiate that in my review of Boulez’ 7th, whenever I write it). Also, by then, much Mahlerian water had flown under the listeners’ bridges, a whole new generation of Mahler conductors had emerged in the late 1970s and 1980s, Levine, Abbado, Inbal, Tennstedt, Bertini, Rattle, Sinopoli, Chailly, and the older generation was revisiting Mahler, with Bernstein in particular embarking upon his second cycle for DG. So what Boulez had to say wasn’t as novel, original and indispensable is it may have seemed to de La Grange when he heard it live two or three decades before.

But that is making the bold assumption that Boulez in the 1990s performed Mahler as he had in the late 1960s and 1970s. In fact, on the evidence of this live concert with the BBC Symphony Orchestra from 1968, it is striking the extent to which he didn’t. So the concert is interesting for giving us a chance to hear Boulez’ Mahler as Henry-Louis de La Grange might have heard it, and pined for it. The sound is acceptable given that it is an aircheck, the brass and the tutti are clear and present, but there is a kind of soft rattle around the string tone. There are various sonic and execution glitches, but the audience is remarkably inaudible, except at the end of the movements. Clearly this is not to be taken as anything competitive with the studio recordings of then and now. But as a documentation of Boulez’s Mahler in the late 1960s, it’ll do.

Interpretively, it is indeed, for its vintage, a very personal and original approach. Boulez’ funeral march opens in a deliberate manner and with a burdened gait that had no precedent in the studio (although Scherchen had conducted it in concert in Philadelphia in 1964 in a very similar way, documented on Tahra TAH 422 – link will open new tab to my review). To give an idea, Boulez reaches the fast section in 5:53. Bruno Walter, in the 1947 premiere recording from Columbia, had taken 4:41, and Bernstein, in New York in 1963, 5:09. And it is not just a question of tempo, but also the tolling weight that Boulez invests in the regular marching rhythm. Interestingly, a year later, in 1969, Sir John Barbirolli made a famous studio recording for EMI that was very similar in its opening pace (5:50), and Bernstein, in his 1987 remake with the Vienna Philharmonic for DG, went even further – e.g. slower: 6:15. On the other hand, when he finally made it in the studio, in 1996, almost thirty years later, also with the Vienna Philharmonic and for DG, Boulez reverted to a more “normal” pace (5:16).

But bringing off the first movement is not just a matter of choosing an opening tempo: it is also about the tempo relationship with the “fast” section, and whether one favors stong contrast or a sense of organic unity. There Boulez remains unique by adopting a very, almost excruciatingly even, slow tempo, slower even than Walter’s, one that doesn’t allow for very much of Mahler’s “leidenschaftlich” (passionately) and “wild” (no translation needed) to unfold, but that is at least very consistent with his opening one. Still, Boulez establishes strong contrast, but through dynamics (it really erupts fortissimo) rather than speed, and if the passage lacks passion, it is certainly vehement. Returns the opening tempo, and Boulez’ final section, at 11:00, starts beautifully, with ghostly strings, although the downside is that the more passionate surge that follows remains very pedestrian, and the great “klagend” climax at 12:27 misfires entirely: no impact, wrong balances, flutes and no brass.

The second movement displays the same strong interpretive personality. There, Boulez plays the maximum contrast of tempo, launching in not only with the “grösster Vehemenz” (greatest vehemence) asked by Mahler, but even with a fury verging on frenzy, with brass biting like enraged pitbulls. But then comes the slow section with its reminiscence of the first movement, and there, in keeping with his first movement tempi, Boulez adopts a very held-back pace. It is in fact Scherchen that he comes closest to here (in his 1952 studio recording for Westminster). In these slow passages, as in the first movement, Boulez dangerously treads the line between deliberate and sluggish – and determinedly crosses it at the “più mosso subito” at 7:47. I guess you can call it “più mosso” (literally: more moving), in view of what preceded it, but still, more “mosso” than very slow doesn’t make it “mosso”. Here it sounds like one of those deliberately clumsy Ländlers so typical of Mahler. Well at least it is original, if not altogether convincing.

In all these choices of tempo, and especially his taste for the strong contrasts, one can question Boulez’ sense of architecture and organic development, but nowhere more than when returns the theme of the first movement’s fast section, at 9:47 (measure 392), where Boulez is now (because of his adoption of a very fast opening tempo of the second movement, of which this section is a continuation) MUCH faster than in the first movement.

In keeping with his furious approach to the second movement, Boulez plays the Scherzo’s wild fury and sardonic grimace rather than its playful charm, an approach that I like very much although many, including the esteemed Mahler critic on Musicweb-international Tony Duggan, might find cause to object. Duggan contends, with some justification, that the Scherzo shouldn’t be the continuation of the second movement, but a bridge between the “negative” of the first part and the “positive” of the third part. Still, as convincing as Duggan’s argument may seem intellectually, emotionally there is great thrill in Boulez’ approach, and sometimes, in music, emotion must take precedence over intellect.

Although the effect is more furious with Boulez, it is interesting that, timing-wise, his Scherzo should be so close to Bruno Walter’s (B 15:22, W 15:10) – and Walter is to this day, according to the table of timings in Peter Fülöp’s magisterial Mahler Discography, the fastest Scherzo ever recorded, among more than 170 versions – because, in the Adagietto again, it is to Walter that Boulez comes closest (B 7:35, W 7:42). This means that it is one of those very flowing versions, true to Mahler’s original conception (according to his early champion Willem Mengelberg) of the Adagietto as a tender and passionate declaration of love to Alma and not the (equally beautiful) “Death in Venice” meditation on the passing of things that it became under later conductors, starting (in the studio) with Bernstein in 1963 (11:00) and continuing… with Boulez himself in his studio recording (10:56).

Like his second and third movements, Boulez’ finale is all thrill and excitement. It starts like Walter’s, which is brisk but not as brisk as Bernstein in 1963. But very soon Boulez whips it up to even higher voltage and hardly ever lets go, to the point that it sounds a bit too inflexible at times, with little concern for Mahler’s “grazioso” passages. In the end Boulez tosses it off in 13:22 (to Bernstein’s 13:43 and Walter’s 14:08), which makes it one of the fastest finales ever recorded. The impression is of joy (“Allegro giocoso” is Mahler’s indication, together with “frisch”, lively) almost verging at times on a kind of breathless urgency and even grim fury – the mischievous joy of a Till Eulenspiegel, maybe.

I don’t know if this is what Henry-Louis de la Grange heard in the late 1960s or 1970s, but if it is I can understand – embellishments of memory helping – why he would have raved about it. Nobody had conducted Mahler’s 5th like that before – and I doubt that anybody did since; Boulez certainly didn’t thirty years later. About everything is disputable in this performance: Boulez’ choice of tempi, his tempi relationship and his sense of architectural coherence. But one thing can’t be disputed: it is a strong interpretive personality that is at play here, one for which I can see only Hermann Scherchen as a precedent (even Mitropoulos was more disciplined), in the readiness to sacrifice architectural coherence to the thrill, excitement and heat of the music. Boulez has the reputation of being a technically precise, “intellectual” but “cold” and “emotionally detached” conductor. Well, in 1968, it was the other way around.

[Addendum from January 2023] Same performance was reissued in 1991 on Arkadia CDGI 754.2 barcode 8011571754027, paired with a live of the 9th from 6 June 1971, see listing on, and again in 1993 by the label Entreprise, “Documents” LV 901/02 barcode 8011662900050 (see listing on, paired with a live of the 9th from…22 October 1972 (as I write there is an offer on for the ludicrous price of… 988.76 USD, and even 1186,51 USD on – I kid you not!) I don’t know how the sonics compare.

The 1990 release on Hunt CD 718 (no barcode) is another live performance with the BBC Symphony Orchestra, from 4 August 1970.


The dating “4 August” isn’t on the CD nor in Fülöp’s discography:, I found it here and the post seemed informed enough on the discography of Boulez for me to trust it:

Here are a few of the Mahler recordings that have been available on
various labels. I’m especially enthusiastic about the BBC SO
performances of the 2nd and the 8th:

Mahler: Symphony No. 2
Felicity Palmer; Tatania Troyanos
BBC SO et al; Boulez
London, August 27, 1974
Originals SH855/6; Enterprise ENT LV 915/6

Mahler: Symphony no. 3
Yvonne Minton
BBC SO et al; Boulez
London, 1974
Artists Live Recordings FED 024.25

Mahler: Symphony No. 3
Yvonne Minton
New York Philharmonic et al; Boulez
New York, October 23, 1976
New York Philharmonic Mahler set: NYP-9801

Mahler: Symphony No. 5
BBC SO, Boulez
August 28, 1968
Nuova Era 2326; Arkadia CDGI 754

Mahler: Symphony No. 5
BBC SO, Boulez
London, August 4, 1970
Hunt CD 718

Mahler: Symphony No. 6
BBC Symphony Orchestra
London, Royal Albert Hall, August 4, 1972
Live Documents CD LV 995: Artists FED 032

Mahler: Symphony No. 8
BBC SO et al; Boulez
London, Royal Albert Hall, July 25, 1975
Artists Live Recordings FED041/2

I was alerted by the table of timings contained in Fülöp’s discography that Boulez must have radically changed his conception of the first movement between 1968 and 1970: it runs 14:04 in 1968 vs 12:12 in 1970 (with Boulez’ later DG studio recording clocking at 12:52), a considerable difference. I don’t have the Hunt CD but it is, at the time of writing, available on YouTube. In 2017, Saint-Laurent Studio of Québec published another 1970 performance of Boulez, this one from Cleveland on March 5, “Boulez vol. 10” YSL 0425 T (the link is to the listing on St-Laurent’s website), and that one too has been posted on  YouTube. First movement there times 12:00. Limiting my comments to that first movement, of the two 1970 performances, Cleveland is by far the best, the most forward-moving and tension-filled, as well as disciplined and sonically present. While, in London, Boulez maintains his very held-back tempo in the fast middle section (he does so also in Vienna in 1996, but with far more tonal beauty, so that the effect isn’t as agonizingly slow), he is signficantly faster and “back to normal” in Cleveland; but in both performances, yes, the funeral marches are less funeral and more flowing than in 1968 (as an illustration, the fast section was reached in 5:05 in London, 5:02 in Cleveland, vs 5:53 in 1968 (5:17 in Vienna).

In an interview made by Universal Edition in 2011 and also posted on YouTube, Boulez recalls (at 6:00) that the Fifth was the first Mahler Symphony he conducted, with the BBC, “in 1965 or 1966 I don’t remember”, and with an ironic smile he adds “I don’t remember exactly the performance, I don’t think it was really the peak of the performance I have ever done, because that was the first time, simply”. I conjectured, and then, finding an online archive of the Proms, I was able to verify  that, with memory blurred, Boulez was in fact referring to this 1968 concert (much gratitude the BBC for making it available! Where you learn that first part of the concert was Mozart’s Piano Concerto No. 27 with Clifford Curzon, and that Boulez was supposed also to conduct a new piece by… Sir William Walton! Boulez, Walton???? But that fortunately for both, “Capriccio burlesco… was not completed in time” – I wouldn’t be surprised if that were diplomatic language to say that Boulez didn’t want to conduct it… I could also corroborate the date of August 4 for the 1970 BBC performance). And the interpretive characteristics of that performance certainly fit the notion of “a first discovery”. But let me defend the 1968 Boulez performer against the 2011 Boulez critic: warts and all, there was possibly more personality in the “discovery” than in the later, more “mainstream” versions.

Comments are welcome