Pioneering Haydn On Record
by Christopher Howell

An occasional series dedicated to commercial and live recordings of Haydn symphonies up to and including those of the first two complete cycles by Ernst Märzendorfer and Antal Dorati

Symphony no. 47 in G major Hob. I/47 (1772)

A pdf version of this article can be downloaded for offline reading and printing here.

Symphony no. 47 comes a decade after the last two works I have discussed, Symphonies 22 and 33. If Haydn had stopped here, this would be remembered as a mature masterpiece. The main subject of the first movement is essentially a fanfare on one note for the horns, with a brief answering phrase from the strings. The major-second dissonance when the second horn enters must have sounded very strange to 18^th century ears. Even today it sounds strange in a way that the dominant seventh with which Beethoven opened his First Symphony, and which caused such ructions in its day, and on which Haydn arrives in bar 7 before resolving on the tonic, no longer does. This horn fanfare, now piano, dominates the whole of the first subject section in a way we associate more with early romantic composers such as Schubert. Proto-Schubertian, too, is the second subject in which the violins scamper in triplets over a treading bass, followed by a calm chorale-like theme from the oboes. The development works out both first and second subject material at exceptional length for a symphony of this period. Then, when the recapitulation arrives, it is suddenly in the minor key. This is brushed away almost immediately by the recapitulation of the second subject, in the major key, after which the movement closes with reassuring normality.

The second movement, in which the strings are muted, is in variation form, but in a rather different guise from the variation movements with which Haydn was to delight his later audiences. These normally have a theme in “binary” form, that is to say in two sections, each of which is repeated. No repeats are indicated in this movement and the theme is in ternary form. The first part is for strings only, and listeners perhaps assumed this was to be a movement for just strings, as was common among the earlier symphonies. Instead, the wind enter for the second section. The ternary form is completed by a reprise on the strings, but “upside down”, as it were. The lower strings play the melody previously played by the violins, while the violins play what was originally the bass line. There now follow three transparently constructed variations – the melody and bass line simply become more florid each time. First in semiquavers, then in semiquaver triplets, then in demisemiquavers. A fourth variation is a simple restatement of the theme with fuller orchestration, but four repeated chords lead to a dramatic pause, after which Haydn meditates on the four repeated notes (echoes of the first movement motif?) before finishing with a further serene reference to the theme.

The “Menuet al roverso” has caused this symphony to be named occasionally “The Palindrome”. The second half of the minuet is simply the first half played backwards, likewise the second half of the trio. Haydn, in fact, did not even write these second halves out but, since 18^th century musicians were evidently more adept at playing backwards than we are today, Robbins Landon has thoughtfully provided a realization in his edition[1]. Listeners may not necessarily notice the trick that is being played on them if the music sounds equally attractive and melodious whichever way round it is played (which it does). But Haydn did his best to help them not to miss it in the trio. This begins with a phrase for horns, another for oboes, a third for strings and finally one for oboes and strings. So the listener who fails to notice that the notes are being played in reverse order would be likely to realize at least that the colour scheme is reversed.

The finale opens with a long-breathed theme over a repeated-note bass (another reference to the opening motif of the symphony?). Before long this has burst into a fierce episode dominated by acciaccaturas, anticipating Mozart’s “Turkish” music – though maybe the inspiration comes from Hungarian folk music. This then turns out to be one of Haydn’s monothematic sonata structures – when the time has come for a second subject, the first theme re-enters as calmly as if butter would not melt in its mouth. But in some ways, the most remarkable feature of the movement is the first eight bars of the development, where Haydn dwells on the three sharp chords that had rounded off the exposition – a device much loved by the young Beethoven.

There appear to have been four recordings of this symphony before the arrival of the two complete cycles. By far the earliest was by the Vienna Chamber Orchestra under Franz Litschauer, a Haydn Society issue (HSLP 1026) set down around June 1950. Litschauer’s contribution to this series was invariably distinguished. Using a small orchestra, he does not have a harpsichord, so far as I can hear, but the bassoon can be heard adding its piquant voice to the bass line at various points. The outer movements are unhurried but vital, and have their first repeats. The second movement is gracefully phrased, though the BNF transfer available on YouTube has the first two notes missing, and the minuet goes at a very nice pace. I take it that the third repetition of the second part of the trio is an editing error. A distinguished performance.

Next came Antonio Janigro, whose recording with the Zagreb Radio SO was set down in August 1963 as part of his set of the “Sturm und Drang” Symphonies (Vanguard VRS-1107/VSD-2146). Not that this mainly sunny symphony has particularly “Sturm und Drang” features.

A lot of my introductory comments were made on the basis of the Litschauer recording. Janigro presents a radically different work. In the first movement, the horn fanfare motif is rapped out smartly at a much faster tempo and Janigro’s tautly controlled conception sees that the music never steps outside the classical world. This is finely done after its own light. I shall have to think about this on the basis of the performances that follow. Janigro’s second movement is supremely lovely, a calmly Gluckian tour of the Elysian fields. The minuet is slower and more stylized than Litschauer’s. My description of the opening theme of the finale as “long-breathed” does not fit Janigro’s performance, in which it is whizzed along like the finale of an opera buffa. If Haydn intended the repeated notes in the bass line to recall the opening motif of the first movement, they certainly do not at this tempo. As with the first movement, it is finely done after its own light. My impression so far – and I await further interpretations – is that Litschauer gives us a great symphony, Janigro gives us a charming one. First repeats in the outer movements.

Raymond Leppard and the English Chamber Orchestra set down their version on 1^st December 1968 (Philips 6527 096), a disc also containing Symphonies 22 and 39. In my article on Symphony 22, I expressed surprise that such a well-regarded disc in its day, under a fondly remembered conductor, had never been issued on CD, and expressed my gratitude to a fellow member of a discussion group who provided an excellent transfer. In the event, I was not overwhelmed by no. 22.

Leppard is swifter still than Janigro in the first movement of no. 47, but with a peremptory tone that better expresses Haydn expanding and challenging traditional moulds. In the finale, on the other hand, he takes a tempo midway between Litschauer and Janigro. This avoids opera buffa connotations and Leppard again adopts a peremptory manner that emphasizes the Sturm and Drang element. In these two movements, in which the first repeats only are played, Leppard seems preferable to Janigro, if not to Litschauer. Reservations concern the middle movements. The staccato upbeat to the Un poco adagio, cantabile second movement sets the tone at once. In place of Janigro’s Elysian fields and Litschauer’s gentle songfulness, Leppard goes for conversational elegance. This is attractive prose rather than poetry. A very distant harpsichord can be detected in this movement. If a conductor feels the harpsichord is necessary – fortunately, it does not seem to be – could he not insist with the engineers that it is properly audible? The minuet goes with a very nice lilt, but my reservation here concerns a curious point. Robbins Landon, in his realization of the roverso passages, takes the view that, as with any other minuet, the two sections of the minuet are to be played twice, likewise those of the trio, the repeats being dropped in the da capo (some HIP performers do not even agree with this latter). That is to say, the minuet is played frontwise twice, then backwards twice, and so too the trio. Litschauer and Janigro do this. Leppard evidently considers that the roverso is itself the repeat, so he plays it frontwise just once, then backwards just once, the trio following suit. Did it not strike him that the movement was unrealistically short this way?[2] Swings and roundabouts between Janigro and Leppard, then. 

The Hungarian Chamber Orchestra led (not conducted) by Vilmos Tátrai recorded their version on 5-6 July 1971 (Hungaroton SLPX 11576). On other occasions I have commented that this group, led by their founder and distinguished first violin, plays more like an expanded string quartet than a reduced symphony orchestra. The result in the outer movements of this symphony is perhaps too comfortable, suggestive of a group of friends settling down for a cosy musical evening. Mellow enjoyment is certainly on offer, but is this enough?

On a practical level, their first movement is slightly slower than Litschauer’s. I have complained that Janigro and Leppard seemed on the speedy side, so what did Haydn write? Nothing, it would seem. The Robbins Landon score has Allegro in square brackets, indicating that it is an editorial suggestion. Since there is evidently no tempo indication in the source, if you think Moderato – or for that matter Presto assai –more suitable, you are not necessarily wrong. For what it is worth, the old Breitkopf edition had Allegro without explanation or apology. More significantly, IMSLP has available two arrangements, one for piano trio, the other, already referred to in footnote 2, for piano solo, both published in the 1780s, well within Haydn’s lifetime. These indicate Allegro. So for Haydn’s contemporaries, it was implicit that a first movement of this type would be played Allegro. Another significant point is that it is an Allegro in four, indicated with a C, without the vertical cut that would make it an Allegro in two. On this basis, Litschauer and Tátrai should be about right. However, I find that Litschauer gives it a symphonic dimension while Tátrai, agreeable is his performance is, comes across as a little placid.

The finale is a different question. It is marked Presto assai and with a cut C – so in two, not four. Looked at this way, Janigro’s opera buffa performance has its justification, though I continue to find it too frothy. Tátrai is slightly faster than Litschauer here, but once again I find him too comfortable. Perhaps Leppard is closest to the mark, though Litschauer’s tempo allows us to hear the popping bassoon on the repeat notes at the opening, something we do not hear on the others, if it is there at all[3].

Tátrai’s second movement shows his group at its best, joining Janigro in the Elysian fields. His minuet is dainty but very attractive. On balance, I would prefer this very nice performance over Janigro or Leppard, but I am not sure that Litschauer has yet been surpassed. First repeats only in the outer movements.

Märzendorfer is pretty swift in the first movement, but manages to give the idea of a striding four-in-a-bar. The movement unfolds steadily and unhurriedly despite the fast tempo. Like Leppard, he captures the ebullience of Haydn breaking away from formal conventions. His rasping oboes add to the “modern” feel but, while I have welcomed them in other symphonies, I had doubts about their intonation throughout this one. Märzendorfer is in the Elysian fields in the second movement with Janigro and Tátrai, but the acerbic oboes are unwelcome in this context. Märzendorfer’s minuet has a delightful spring and, against all odds, his oboes play sweetly in the trio. So it looks as though Märzendorfer deliberately had them play within a millimetre of overblowing elsewhere. He takes Haydn’s Presto assai in cut time literally in the finale, yet just avoids the frothiness of Janigro, emphasizing the Sturm und Drang elements as did Leppard. A splendid interpretation, but a splendid performance only if the oboes are not an issue.

Dorati, at a similar tempo to Märzendorfer’s, likewise creates a striding four-in-a-bar. Furthermore, he succeeds in giving the horn fanfare motif a threatening tone, contrasting strongly with the second subject material which scampers along delightfully. It is as though Haydn is registering the effects of war’s alarms upon the pleasant life in the fields. A genuinely imaginative interpretation. Dorati also expresses very well the calm beauty of the second movement and his minuet has delicate grace without the simpering and curtseying that can disfigure his interpretations of these movements. The finale goes at a fairly swift pace and avoids digging in too harshly in the Sturm und Drang episodes, keeping it all light and joyful, but not frothy as with Janigro. I might prefer Leppard or (oboes permitting) Märzendorfer in this movement but overall I am delighted to find a Dorati performance that I can enjoy almost without reservations. I shall sometimes go back to Litschauer, but the Dorati is of course much better recorded.

Since all six performances discussed agree in playing only the first repeats in the outer movements, a word about second half repeats is in order. Sonata form originally grew from the binary form of typical baroque dance movements, in which the two halves are much the same length. For as long as this remained the case, it was logical to repeat both halves. As the sonata structure grew, the second part expanded to make what we now call the “development” and “recapitulation”. Haydn himself, later on, ceased to indicate second part repeats, evidently recognizing that the second part had by now become too long and elaborate to justify repetition. Moreover, very often the second part, without repeat, would be as long as the first part with it, so there was also proportion to be considered. Symphony 47 must be one of the earliest in which the second part of both outer movements is fully developed in this way. Hence, I imagine, the unanimous decision of all six artists not to play the second part repeats in this symphony. Doubtless some more recent performances play them, if a listener wishes to try the effect.

Christopher Howell © 2026

[1] The old Breitkopf edition provided no such realization, leaving the performers to master the art of playing backwards.

[2] A piano arrangement by “Mr. Tindal”, published by Bland of London around 1780, has the following note: “N.B. This Minuet is to be played twice thro’ in the usual manner, then twice backward. The Trio is to be played in the same manner.” So Haydn’s contemporaries were in no doubt about how to play it.

[3] A bassoon should be automatically added to the bass line of Haydn’s symphonies, excepting maybe movements for strings only, through to the time when he started writing independent parts for that instrument. Not many performances include it.