Some Notes on Editing Bruckner’s Second Symphony

William Carragan
Contributing Editor, Anton Bruckner Collected Edition, Vienna

Published in The Bruckner Journal, March, 2009

William Carragan - Some Notes on Editing Bruckner’s Second Symphony - Austrian 1 Florin 1872


My double edition of the Second Symphony of Anton Bruckner had its start in a paper I was writing in 1985-6, devoted to the methods of revision Bruckner used in his finales. I undertook that research to give myself more background for the work I was doing on my completion of the Ninth Symphony. The final movements always seem to be the ones most altered when Bruckner looked over his work critically, and I felt that valuable clues to Bruckner’s compositional methods could be obtained through studying and analyzing the revisions. Of course one has to start with the earliest versions, and in 1986, all of the earliest versions appeared to be published except that of the Second. Indeed the only version of the Second in legitimate print at the time was one by Leopold Nowak dating from 1965, which closely paralleled the composite edition brought out by Robert Haas in 1938. As for Haas’s edition, it came out at the time of the Anschluss, and was the first one to be offered after the court decision which held that only if the Collected Edition scores were sufficiently different from those under copyright by Universal Edition, as the successors of Bruckner’s original publishers, could they be granted a new copyright. It seemed to me that Haas took every opportunity to make his edition different from the Universal score which followed the first printing of 1892. He unabashedly melded the versions of 1872 and 1877, producing a handsome enough piece, but one which did not express Bruckner’s ideas at any time in the revision history of this symphony. It represents a highly adventurous expression of the role of an editor, something that would never be tolerated today.

This characteristic of the Haas edition became apparent to me, as a researcher, only slowly. Haas’s rare critical report, a copy of which came to me through the courtesy of Leopold Nowak and David Aldeborgh, is not prepared to the standard of his earlier reports on the First, Fourth, and Fifth, all of which I have perused in detail. It does contain reproductions of the material of the symphony not present in his edition, but with enough errors and presented in a confusing enough way so as to make the preparation from it of a pure edition of the symphony, as it stood at any time in the revision period of 1872-1892, utterly impossible. For example, he mentioned the Finale as containing 806 measures, but only with the greatest difficulty can one account for 805 of them, without knowing what the other one could be or where it should lie, or indeed with little certainly that the 805 measures were correct. It was in this atmosphere that I wrote to Vienna asking for a film of Mus.Hs. 19.474, which at the time was regarded as the composition score of the early version. This is the one in which Bruckner at a late date wrote “Alte Bearbeitung” across the top of the first page. After a bit of discussion, Hofrat Nowak asked me to prepare a new edition of the symphony for the Collected Edition, knowing, as many others did as well, that he had not dealt fully with the problems of the Haas edition in 1965; indeed, as I later learned, he made only 23 changes in it.

The Study

In the next five years I did a great deal of work on the symphony, spending a sabbatical in Vienna, St. Florian, and Kremsmünster working on the sources. The scores, 19474 and three copy scores, are all preserved at Vienna except for some detached leaves at the abbey of Kremsmünster, but an extensive set of manuscript orchestral parts for the Second is in the library of the convent of St. Florian where Bruckner spent so much happy time throughout his life. These parts were copied in 1872, and became the materials for the performances of 1873 and 1876 under Bruckner’s direction. As is shown later in this discussion, 19474 could not have been the score from which these parts were copied; there must at one time have been an earlier composition score which the copyist used. Judging from other composition scores of Bruckner’s, this score, unknown today, was almost certainly much messier than the generally neat 19474; perhaps it was discarded when Bruckner moved into the Belvedere in 1895. Thus the unaltered regions of the orchestral parts became the best evidence for the content of the original composition score, and the prime sources for the early version became not only 19474, but also the manuscript orchestral parts which had not been closely studied by either Haas or Nowak. As I sat in the old music collection high atop the Albertina in Vienna, and in the grand and peaceful Room 13 at St. Florian, one misconception about this symphony after another fell away, and little by little, the history of the symphony emerged. And it was very different from what had been presented to the public. Indeed one of the main reasons for that was a serious and clumsy mis-sequencing of the scores by Haas, as seen in his work. I still remember the amazement I felt when I realized that had happened, and that everything was different from what Bruckner enthusiasts had thought for decades.

Through my research I was able to determine with almost complete accuracy the state of each movement of the symphony at its first conception in 1872, at the unproductive rehearsal later in 1872, at the first performance in 1873, at the second performance in 1876, after a revision to the second movement carried out immediately after the 1876 performance, after the revision of 1877, and at the publication of 1892 which was followed by further performances. I was eventually asked by Hofrat Nowak and by Dr. Herbert Vogg of the Musicological Press of the International Bruckner Society to prepare two versions of the symphony, an early one and a late one. It seemed that the best thing was to provide two versions which represented the true beginning and the true end point of Bruckner’s involvement with this attractive and vigorous composition. However, there were several problems. It has been customary in the Collected Edition to use only manuscript sources, although Nowak went against that idea in a few places, particularly with the utterly essential ritardando markings in the finale of the Seventh. I kept an open mind on this topic, and it became clear through manuscript evidence that Bruckner was deeply invested in the preparation of the print, just as he was with the first printing of the Fourth. Corrections seem to have been quietly made for the print in the galleys, which sadly do not survive. However many of these important corrections were not entered in the error-laden scores later used by Haas and Nowak for their editions. One of the most important and essential corrections was the repair of an omission of a measure near the end of the finale, which was re-entered as a paste-over in the Stichvorlage, Mus.Hs. 6035, a copy score originally dating from 1872 but heavily modified in 1877, with Bruckner’s own aged (1892) hand on the original paper behind it stating “Ein Takt fehlt.” Reflecting the reservation that many people still feel against the first published editions, I kept the date of 1877 for my new score of the late version, but provided in it as extras the important variants of 1892, particularly the extended ending of the first movement and the sonorous rescoring of the last measures of the finale. Thus the late version of the symphony, under my editorship, took shape with the inclusion of Bruckner’s very last, or nearly last, thoughts.

The early version required a completely different kind of thinking. It was particularly important not to invade the process of composition–instead, to distinguish operationally between composition and revision. Here the parts were an important guide; each one of the 35 parts reflected individually in its revisions, overwritings, blottings, scrapings, gougings, paste-overs, rebindings, and wholesale deletions, the entire history of the symphony from 1872 to 1877. The first three movements were copied during the composition of the symphony, being finished just as the finale neared completion. Then, perhaps at a later date, the finale parts were copied. During the copying of the first three movements, the order was Kopfsatz-Scherzo-Adagio in each part, as they are labeled. It is clear that this was the order of the first conception of the symphony, not just an artificial sequencing of composition. When the Finale was completed at a time later in 1872 and then added to the parts, the movements in the parts were reshuffled so that they were in the order Kopfsatz-Adagio-Scherzo-Finale. At that same time, two other changes were made, a difficult-to-hear revision in some wind passages in the first movement, and a change in the figuration of the first violins in part 5 of the Adagio from groups of six to groups of five. (These were not the same groups of five as are present in the versions of 1877 and 1892.) A bit later, the horn was removed from the end of the Adagio and the clarinet and violas were substituted for it; this was done by the copyist Carda in time for the 1873 performance. Other changes for 1873 included a violin solo and intensified wind parts in part 5 of the Adagio, a new, shorter passage (labeled in manuscript “Neuer Satz”) to replace some of the most interesting writing in the development of the Finale, and a cut near the end. Every one of these 1873 changes reflects the heavy hand of revision. Except for the violin solo and its accompanying wind enhancements, each one represents a retreat from a bold, fresh concept, in which Bruckner was beginning to formulate his new, Viennese, way of doing things. Nonetheless, in 1990 and early in 1991, I felt that it would be a good idea to make these variants easily accessible to performers so that the edition would have the widest scope possible.


Then two things happened. The first was the performance and recording of the new edition by the Bruckner Orchester Linz under the direction of Kurt Eichhorn, a distinguished conductor from Munich who was a favorite guest artist of the orchestra, and who had previously prepared a very creditable Seventh with them. I was present for this process which stretched over two weeks in March and April of 1991, with performances in Vöcklabruck and Linz and many subsequent recording sessions. One of the most pleasant occurrences of these weeks was when a delegation of the orchestra players came to me during a session break, saying “Diese Symphonie ist super!” I think they were used to the massive Fifth and Eighth, and wondered how the slenderer Second could be worthwhile for them, and were very pleased to find that it was. At these sessions we prepared two versions of the symphony, that of each movement in its first conception, and that of the 1873 premiere. The first concept was strongly preferred by the orchestra. Particularly in the fifth part of the Adagio, the violinists implored me to use the groups of six in the edition rather than the groups of five. I remonstrated, “The fives are slower,” but they rejoined that the sixes were easier and flowed better. I asked Heinz Haunold, the excellent concertmaster and violin soloist, how he felt, and he agreed. He could tell what we all knew: the violin solo at that point of the movement effectively prevented the orchestra from rising to the great climax which was written in the score, and which was later to be emulated in most of the other slow movements that Bruckner wrote. In other words, the 1873 variant not only had the character of an evident retrenchment from an earlier, bolder concept, but it also contained a fatal trap for the performers of the symphony.

The second thing to happen was my last meeting with Hofrat Nowak, six weeks before his death, at the Sanatorium Rekawinkel outside Vienna. According to Frau Paula Nowak this was the last musicological discussion he had with anyone. He received me with his usual courtly dignity, and was most interested in the new edition of the Second, delighted that it had already been performed, broadcast, and recorded. High among the recommendations he made to me at that time was that I should always write about the symphony so that what I said could be understood. He knew well Haas’s dense and imbricated self-referencing, which make his accounts very difficult and frustrating to use. But even more important, he told me not to prepare too complex an edition, that is, not to be afraid to consign some of the material to the critical report. He said that if one gave people too much, they would just be addled; “verwirren” was the wonderful word he used. Coming from dealing with exactly that problem at Linz, I realized the rightness of what he said, and that I had to choose between the 1873 revision and the 1872 first concept for the edition, and not try to present them both. The choice was inescapable: 1872. Since then, from the performances and recordings that have been made of it, I have received many, many appreciative comments of approval for that choice, especially for the placement of the Scherzo before the Adagio which people find revelatory.

Input from others, then and now

Bruckner’s friend Johann Herbeck was always ready to give Bruckner advice, and especially at the time of revision of the Second and Third in 1877 seems to have given him a lot of it. But we have no true idea of what that advice was. Herbeck’s close and insistent concern with the Second was documented by Ludwig Herbeck in a biography of his father, in which he said that Bruckner was notable in his ability to resist Herbeck’s strenuous requests. As is becoming more and more clear through research by many people, we must learn to drop the idea of Bruckner as one who was indecisive and easily led. He was, in fact, a typically obstinate and self-willed countryman, and at the same time a composer of formidable education and great self-confidence in putting pen to paper. His bold and easy-to-read manuscripts bristle with character and are masterpieces of firm decision and resolve. As for Herbeck, Haas claimed in his critical report of 1938 that Herbeck himself wrote the notes of the notorious violin solo into Bruckner’s fair-copy manuscript now known as 19474. But indeed I could see that the solo was actually entered by the copyist Carda, in a combination of pen and pencil in his well-known hand, which is not at all similar to either the pen or pencil script of Herbeck of which we have copious amounts in the Austrian National Library. Furthermore, as has become clear from the Bruckner-Levi correspondence concerning the Eighth, Bruckner’s normal reaction to criticism may have been to rethink or redo something, but not necessarily along the lines that had been suggested. Quite aside from the fact that Levi made no specific suggestions to Bruckner about the Eighth, Bruckner’s alterations of that symphony in the period 1887–1890 were all his own rapidly-evolving ideas, and we should expect that the same was true in 1877 concerning the Second.

One of the irritating aspects of the Haas/Nowak editions of the Second Symphony (they are so similar that they hardly need be distinguished) is the encouragement they give to conductors to make their own choice about this or that passage, each conductor making a unique version of the symphony to his or her liking. No other symphonist says to his performers “do this, or do that, you choose;” one does not see such an element of choice in the work of any composer until the aleatoric writing of Karlheinz Stockhausen. And surely a good edition represents the best convictions reached by the editor as to the validity and consistency of every part of the piece, thrusting aside personal prejudices and doing the best that can be done to reflect the composer’s thought. Thus there can ultimately be little or no choice for the performer. In my two volumes I have included a few variants of movement endings, where they can do no harm, but, for example, presenting the slow movement’s violin solo would also have meant presenting the thickened wind parts, the short form of the clarinet/viola ending used in 1873, and somehow finding room in the book for the finale’s lovely but superfluous and eventually-rejected “Neuer Satz” and directions for other cuts and changes. Anybody would be greatly confused looking at something like that, and each conductor would struggle to make some sort of system of choices with chaos being the result, worse than before. In the same way, the horn and clarinet endings of the slow movement should not be included in both scores as alternatives; the horn ending belongs clearly and simply to 1872 and not to 1873 or any later version, and the long-form clarinet ending, which is the one people know, belongs only to 1877 and 1892. Let people take their choice as to this version or that, and then play the symphony!

There is a similar situation of an early variant in the Third Symphony. Nowak brought out the 1873 version, principally based on the flawless copy score given to Wagner in 1874 and it is wonderful. However there is also an 1874 version, which one can base on the Bayreuth score’s twin, considerably revised in the  year following the composition, now catalogued as Mus.Hs. 6033 in Vienna. This score has many interesting modifications dating from later in 1874 but not after that. On commission from a Japanese conductor and orchestra, I prepared performing materials for that version of the Third Symphony, and it was performed and recorded at that time. 1 It too is wonderful, in some ways possibly stronger than 1873, though that point is arguable. In addition it is possible that an early “ghost” variant could be made of the Seventh Symphony, including all the texts of the semi-erasures in the composition score, not just the few that Haas selectively entered (as at measure 123 of the first movement in the low winds). But these variants certainly do not need to be published as scores, any more than the 1873 Second needs to be. That is what critical reports are for, and that for the Second will be produced in due time, though it will contain no great revelations. The great revelations are in my already-published scores! The well-known, excellent criteria defining “version” and “variant” which were devised many years ago by Juan Cahis are of great help here in determining what to publish.

An astonishing discovery

One of those great revelations is a discovery I made on December 12, 1990, in the room in St. Florian in which Bruckner used to live, and in which I was then living. That was the matter about the trumpet parts very near the end of the first movement. Everyone who knows the Second remembers that at the peroration in the Haas edition, the trumpets play for one more measure than the rest of the orchestra and are cut off by another orchestral chord, but that in the Nowak edition, the extra notes are followed by a rest measure inserted in 1877, with silence on beat 1 over the drum roll, as if the cavalry charge had fallen off a cliff. All of the 1877 source scores agree on this point; even 19474 has these notes by virtue of a repeat mark. But the parts do not have these trumpet notes. In the parts, the trumpets play when the rest of the orchestra plays, and at no other time. And the parts have to be correct, because they contain the music that Bruckner heard in 1873 and 1876 when he conducted it. If he had wanted the extra trumpet notes, he would have had the two trumpeters put them into the parts, as he had players put in many other reminders in their own handwriting. Thus it is clear that the parts, which are correct because they were performed, were copied from some other source than 19474 with its erroneous repeat mark. That can only be the missing composition score. I can remember standing by the table in that dimly-lit room in which Bruckner himself had slept, overwhelmed by the meaning of this discovery and by the potential for further unexpected revelations about this fascinating piece.


It is improper and mistaken to conclude from the relative absence of performance markings in the early versions that Bruckner initially preferred a steady, monumental, unchanging performance style, and then in later versions, perhaps under the influence of others, put in elaborate indications for many nuances which should only be used in performances of the first publications. First, to aver that a plain score requires a plain performance is to assert a negative, which is a logical error. Second, we know from the baroque period that performance then required a great deal of input from the performers, often going far beyond what was written on the page, although the elaborate writing of Bach and Couperin sets serious limits to what can be added to their complex scores. Third, the Bruckner early versions keep coming along in rather plain form, symphony by symphony, even though revisions to earlier symphonies had already introduced quite a bit of interpretative detail in them. Most decisively, in the case of the Second, we have the parts which were used in two performances not long after the symphony was written, in which Bruckner had Carda add many expressive indications, with even more of them being added by the musicians under instruction from the podium. Clearly Bruckner wanted a performance full of nuance and interest. Let us remember that the critics were somewhat bemused by Bruckner’s composition, but quite enthusiastic about his conducting. That would not have happened if his interpretive skills had not caught their interest.

What this means is that the conductor has the responsibility of conducting the early version with a similar range of expressiveness to that which is explicitly called upon in the late version. And at this point, one needs to realize that both versions are really the same symphony. Thus I decided to include tempo, expression, and articulation markings from the later version in the early version as well, being sure to enter them in such a way that their origin is clear to the reader. This was done painstakingly after a great deal of thought. Of course the conductor does not need to follow these markings. But with their inclusion, he or she has available Bruckner’s best thoughts on the matter, which are very detailed and self-consistent and require almost no editorial supplementation. I feel that the presence of these marks is one of the strongest aspects of my scores. The performances which have already derived from them have been full of vigor, interest, and eloquence.

In my two volumes I have prepared faithful accounts of the symphony in its earliest form and in its latest form. Everything else, whatever its character, is transition. With respect to other editions, surely the edition previous to mine which presents the most accurate account of the symphony as it was in Bruckner’s mind at some specific time is the 1892 first printing. The 1938 edition of Haas gives us some nice music that is not in the 1892 print, but on his own terms which are decidedly not Bruckner’s. It is not a way station between 1872 and 1877; it is a fabrication that occupies no position in the history of the symphony. It doesn’t matter how pretty or artistic it is; in ways which strongly affect the music, it is a fabrication, a lie. And sadly, the Nowak version of 1965 doesn’t help; in a way, it almost ratifies the Haas. Perhaps Nowak didn’t correct things at that time for the reason that a fragmentary but crucial manuscript which could have validated a more faithful reading was not in the Library. Haas had seen it in 1938, when it was in private possession, but because analysts typically promise not to divulge the whereabouts of manuscripts in private possession, it was not available to Nowak and indeed had disappeared. Fortunately, by the time I came on the scene, the Library had been able to find and purchase it, and its meaning for the edition was obvious to me. In this way my two volumes, faithful to 1872 and 1877/1892, resolve the problems of the structure and content of the Second and provide a new and exciting way to appreciate this “super” symphony. Conductors who continue to use Haas and Nowak, saying such things as “that’s what the players know” with reference to some of the most distinguished orchestras in the world, are simply perpetuating a falsehood, and putting their own preferences ahead of Bruckner’s. Surely we owe this great composer more respect and loyalty than that.


I am grateful to Ken Ward for giving me the opportunity of making this journalistic contribution, and to David Aldeborgh, Findlay Cockrell, Inez Maria Haettenschwiller, Benjamin Korstvedt, and Paul Nudelman for valuable suggestions concerning both analytical content and musicological politics.