W. A. Mozart (1756–1791)

“All good things come in threes” – this phrase came to mind as I sat down to address the following text, having already posted twice on the Henle blog about Mozart’s famous “Alla Turca” piano sonata in A major: Post number 1 dealt with the sensational Budapest discovery of the Mozart sonata’s part-autograph and its editorial consequences ultimately leading to our new, revised edition. Post number 2 unravelled for the first time the previously misinterpreted “repeat” instructions on Mozart’s last autograph page of the “Rondo Alla Turca”. And now number 3: Turning up in the meantime has been a copyist’s copy from Mozart’s time (!), so far completely unknown.

Today’s post is, in a nutshell, about the new editorial consequences – actually very few – that this source has had for the Henle Urtext (HN 1300), and about how erroneously this source finding was evaluated in another publishing house – namely, Bärenreiter Verlag.

This was one of my most exciting experiences in recent years as a Henle editor: At the end of 2016 I was invited by the Munich auction house Zisska & Lacher to evaluate a completely unknown copyist’s copy of the A-major sonata. I thoroughly examined the source and, describing it in detail, consulted the musicologist Dexter Edge, the world’s leading expert on the complex topic of “18th century copyists”, about the question of dating the undated source. He actually recognised the anonymous Viennese hand and reliably dated the manuscript to the “early 1780s at the latest”. For its catalogue (pp. 13–15, lot 34), the auction house wrote up its detailed source appraisal based on our preliminary work. The copy was ultimately sold to the USA, its owner (Christopher J. Salmon) then generously made colour scans of the source available to me.

I introduced the exciting source to my Mozart colleagues at the Prague Mozart Congress in the summer of 2018, describing it in detail with several illustrations and a meticulous source evaluation; the oral presentation has meanwhile been published (Zu einer bislang unbekannten zeitgenössischen Abschrift von Mozarts Klaviersonate A-Dur KV 331, in: Mozart Studien 27, Vienna, 2020, pp. 193–213).

Suffice it to say here that this newly discovered copyist’s copy is definitely relevant and important from the viewpoint of the Urtext editor, though it is unfortunately not as spectacular as was initially anticipated. Why? Because it was copied either directly from Mozart’s autograph, or, more likely, from another copy of the autograph, and contains no entries by a third hand (for instance, by the composer himself). But since Mozart’s autograph is known to be extant only in part (see my blog post number 1), the copy quite respectably replaces the autograph’s missing sections. “Quite respectably” indicates that the copy includes writing errors and lots of inaccuracies – as evidenced by the meticulous comparison with the available autograph sections. The source value of the new copy can also be summarised as follows: Had Mozart’s autograph of K. 331 survived intact, the new source would have no editorial significance. But as it is, the source is very helpful in approximately replacing the lost sections.

I have incorporated all the insights to be gained from the new source within the newly-revised Urtext edition © 2021 (HN 1300). Moreover, I have delineated all the readings and anomalies within a very detailed “Critical Report” for downloading.

The few new (!) readings resulting from the copyist’s manuscript, which are certainly of interest to every pianist, have been included in the main music text of the new Urtext edition or in footnote annotations. These are summarised here for my blog readers (C = copy, A = autograph, FE = first edition):

Mvt. Measure Passage Comment
1 1 Opening C: “sotto voce” (A lost) (see illustration 1),
FE “p” instead of “sotto voce”.
Text adopted from FE, C in the footnote comment, for no engraver alters such “on his own initiative”.
1 26 top, 3rd quaver,

bottom note of the chord

C: d1 (A lost),
FE: e1.
Text adopted from C; FE, commented on in the footnote and in the critical apparatus, because assumed is an engraving error in FE.
1 29 top, 4th from last semiquaver C: b1 (A lost),
FE: g sharp1.
C evaluated as writing error (A would probably have had g sharp1), because the chord differs from, for example, m. 11; reading from C in the footnote.
1 48 bottom, final note C with additional lower octave (A lost) (see illustration 2),
FE: only e1.
Text adopted from C (A would probably have an octave), FE, commented on in the footnote and in the critical apparatus, because presumed in FE is an engraving error.
1 54 top, 9th note C: d1 (A lost),
FE: e1.
Text adopted from C (A would probably have d1), FE, commented on in the footnote and in the critical apparatus, because presumed is an engraving error in FE.
2 65 Entire measure Measure lacking (!) in C (A lost); (see illustration 3)
FE with certainly correct text. Couldn’t (!) C transmit here an early version of A? (Would still be musically possible).
2 82/83 and 83/84 Bottom Ties over the bar line only in C (A lost); commented on in the critical apparatus
3 1 Heading C: “Allegrino”, no “Alla Turca” (A lost, would probably be identical with C);
FE: “Alla Turca”, which we adopted together with “Allegrino” (with footnote comment and detailed commentary in the critical apparatus
3 42 Top C with slur on the semiquavers of the 1st and 2nd crotchet (A lost);
FE no slurs.

Illustration 1: Excerpt from C, start of the first movement.

Illustration 2: Excerpt from C; 1st movement, m. 48.

Illustration 3: Excerpt from C; 2nd movement, mm. 62-66, lacking m. 65.

But now before you transfer these indications to your textually outdated music edition, I’d recommend purchasing the new Henle Urtext edition HN 1300 © 2021, with the completely correct music text, a detailed preface and a concise critical apparatus.

Even though the editor of the Bärenreiter edition of the A-major sonata (BA 11816, © 2020) views the issue otherwise – and is, in my opinion, utterly wrong – it must finally be emphasised that professional source criticism cannot ignore the following facts: The extant manuscripts of the three piano sonatas, K. 330–332 (autographs and the new copyist’s manuscript of K. 331), document the sonatas’ emergence as “stage 1”. Available with Artaria’s publication of the textually revised first edition (Vienna, 1784) is “stage 2” within the genesis of this work and hence the “definitive version” or Mozart’s “last word”. For:

  • Mozart verifiably authorised the engraver’s model for Artaria, hence the first edition as such.
  • Textual changes in the print as compared to the autographs of K. 330 and 332 are in part significant and, from a stylistic viewpoint, must be attributed solely to Mozart (in the A-major sonata, Mozart’s improvements, as compared to stage 1, are ultimately marginal).

For (1): Source Criticism and Text Criticism

The engraver’s models containing Mozart’s corrections for K. 330–332 are lost. As compared with the autographs or manuscript copies, the textual differences in the first edition, some of which are striking, must inevitably come from the “missing link”, that is, the engraver’s models. The lack of this “missing link” is the norm in almost every Mozart edition. For all source criticism, evaluating the first edition as a source is solely about verifying the authorisation of the source(s), usually to be based on the plausibility of circumstantial evidence in the absence of supporting documents.

In the case of K. 330–332 the matter is very simple and clear: Mozart himself sent or brought his manuscript engraver’s models to Artaria: “Now I have given the 3 sonatas for solo piano…to Artaria for engraving” (letter of 12 June 1784 to his father in Salzburg). Better evidence of authorisation is hardly possible.

That the first edition contains striking engraving errors is, of course, not an argument at all against its value as a source, merely evidence that Mozart was not interested in proofreading the galley proofs – which, by the way, is also notoriously the case in the Mozart edition. And even if, purely hypothetically (!), the textual differences in the first edition as compared to the autograph were to go back to a third party – naming Josepha Aurnhamer is speculative in the Bärenreiter editor’s preface – then this must have occurred with the consent of Mozart, the author, for, as I gladly repeat: He himself “gave them to Artaria for engraving”, thus authorising the textual status of Artaria’s print.

This source evaluation does, of course, have consequences: If there is no obvious error in the first edition or a gross inaccuracy, demonstrated by comparing the (part) autograph and/or the copyist’s copy, the printed version reveals the better music text on which the Urtext is to be based; the manuscript version is then simply outdated as preliminary (stage 1).

To give just one example: If the first edition shows dynamic markings that are STILL lacking in the autograph, then these must be owing to Mozart’s intervention in the (lost) engraver’s model, philologically speaking, thus authorising such additions (e.g., 1st movement, mm. 28–30; 2nd movement, mm. 19 and 20).

Copy, 1st movement, mm. 27-30 without dynamic markings.

First edition, 2nd movement, mm. 27-30 with dynamic markings.

Nevertheless, Mozart’s autograph and, in part, the copyist’s manuscript remain of the highest source value. For with the aid of this stage 1, which has been surmounted more or less, we can through meticulous comparison clarify and correct the print’s ostensible and confirmed errors. (And there are also spectacular examples of this in the case of K. 331, just to recall the wrong note in measure 3 of the minuet). The “definitive version”, authoritative for us editors, is for certain in the first-edition print. (A source reproduction of both stages – as is curiously the case with Bärenreiter – is not only confusing to every pianist, but simply a gross editorial mistake.)[1]

For (2): Style Criticism.

No further support is required for the completely unequivocal source evaluation and its premises presented under (1), but text- and style-critical arguments also document the higher source value of the 1784 first edition vis-à-vis the manuscript transmittal. This applies only marginally to the A-major sonata (most notably in the third movement’s very distinctive “Alla Turca” heading), though serious differences and emendations can be found in the first editions of the two sister sonatas in C major, K. 330, and in F major, K. 332, vis-à-vis their manuscript predecessors. Only two notable passages can be cited as evidence of Mozart’s subsequent revision, notated in the lost engraver’s model for Artaria:

  • 330: The wonderful music in the first-edition’s closing measures of the slow movement (from m. 61 with upbeat to m. 64) is totally missing in the autograph. Who could have composed such touching, epilogue-like music if not Mozart himself?

    First edition K. 330: End of the second movement.

  • 332: The written-out ornaments of the repeated sections (mm. 21–40) in the slow movement cannot be traced back to any other composer than Mozart himself.

    K. 332: Comparison between the autograph and the first edition.

To summarise once again: Mozart’s autograph of the A-major sonata (stage 1) was probably copied (perhaps several times) for pedagogical purposes. One of these Viennese copies has now surfaced and is accessible, thanks to its present owner. Before going to press, Mozart revised to a greater or lesser extent the texts of all three piano sonatas, K. 330, 331 and 332, and sent the engraver’s models he had accordingly prepared (stage 2). A reputable Urtext edition of the sonatas therefore draws on the manuscript and printed sources to create the optimal text, edited on the basis of the authorised “definitive version” (= first edition, Artaria, 1784), i.e., with their engraving errors and superficial details emended as a consequence of a strict textual comparison and annotated in the “Critical Report”.

What are your thoughts on this topic? Please feel free to express them in the Comments Section of this blog post.


[1] That the editor does not mention a word of my published research results on the new copyist’s copy in his edition, otherwise appearing to be so immensely scholarly, is, compared to that, just in bad style.

This entry was posted in copy, Monday Postings, Mozart, Wolfgang Amadeus, new source, piano solo, Piano Sonata K. 331 (W.A. Mozart), Urtext and tagged . Bookmark the permalink.

4 Responses to »“Latest news about Mozart’s piano sonata in A major, K. 331”«

  1. Wojciech says:

    The editorial interventions in the 1st and 3rd movements could be made even without the reference to the Viennese copy, just by analyzing the harmonic content of the theme in the 1st mov. and comparison of the analogous spots in the 3rd. The exception is perhaps m.82 in the 1st mov. – the single note e1 rather than octave could have been made on purpose. The octave right after d sharp 1 is simply unconvenient to play and Mozart knew he was aiming at wider audience with different proficiency levels. As for the dating of the whole piece – to me the composition time before 1780 seems highly unlikely. In 1779-80 Mozart was busy composing Zaide (its plot set in the Orient) in which there appears not a note of “Turkish” stylization. Furthermore, I have observed that the trio from Act 2 of Zaide with its lilting, siciliano – like movement might be regarded as a stub for the variations theme. It thus seems logical that the two pieces could have been written in temporal proximity. The well known fact that Mozart quoted Gluck`s La rencontre imprevue (overture) also poses problems – it`s difficult to establish when he got to know the piece. My guess is that this was in the early 1781 during his Munich stay (Gluck`s opera was revived there just two years earlier and Mozart could thus have access to the score). To sum up – to me the most probable date is 1781.

    • Wolf-Dieter Seiffert says:

      Dear Wojciech,
      thank you so much for your impressive comment. You know your Mozart well.
      I do not really agree with the beginning of your comment where you try to deduce editorial decisions from analyzing the given music text. Urtext editors should exclusively rely on the best source-text given and respect it, no matter if music analyses or technical problems may come to subjectively “more convincing” solutions.
      Regarding the dating of K. 331, I think I agree with you: most probably 1781 – before he wrote K. 330, 332 (my argumentation would go into the direction of Barbara Ployer being Mozart’s gifted pupil at that time and the quite conspicuous soprano-clef notation for the right hand in K. 330—332). Thank you for indicating Zaide (Trio 2nd act) as a “stub” for the variation theme. I have to think about this furthermore.

      • Wojciech says:

        Dear Mr. Seiffert
        Thank you very much for your comment. May I share with you a few of my discoveries concerning K.331? Let`s begin with the mysterious “Allegrino” indication. This was clearly obsolete in the 1780s (and perhaps is a reason why Artaria changed it to Allegretto in the 4th impression of K.331 – they probably received a number of questions concerning this marking and couldn`t provide an explanation). The rarity of its use is borne out by the lack of corresponding entry in the most compehensive lexicon of the era (Koch 1802). I know that Allegrino was used by composers of preceding generations like
        G.B. Sammartini and J.Ch. Bach (Trio Sonata “Notturno”) – Mozart of course could know the term through Bach. But my discovery concerns Michael Haydn and his String Quintet G major (P. 109, MH 189, also known as Divertimento or Notturno) written in Salzburg c. 1773. Its minuet bears the indication Allegrino and the minuet`s trio is marked Allegretto. It thus seems obvious that the two terms are not synonymous – the character of the music in this particular piece suggests a slightly slower tempo for the Allegretto. What`s more interesting, the melodic line of the minuet is similar to the earlier mentioned Terzetto from Zaide (“O selige Wonne”). When we look into the Mozart`s autograph of Zaide we can see that he gave here a rare additional indication sotto voce, exactly as in the K.331 theme`s first version. Thus the connecting line between the three pieces becomes obvious. The only other example of Allegrino after that date I found in Joseph Eybler`s String Quintet in A major op.6 n.2 (published in Vienna by Traegg in c. 1803) – finale marked Rondo Allegrino. Mozart and Eybler were friends in Vienna – perhaps they played together M. Haydn`s Quintet? As for the designation alla turca – I`m now pretty convinced that it was an editorial addition because Artaria was simultaneously preparing J. Haydn`s Cembalo Concerto in D major (with its 3rd movement Rondo all`ungharese) – Haydn`s work has a slightly lower plate number and both titles were advertised together (Wiener Zeitung 25.8.1784). If Mozart wanted such a detailed indication he would first use a more general one (such as Rondeau en Polonaise. Andante in K.284). By the way – it was almost certainly Artaria`s editor`s decision to put both segments of this indication together – Allegrino alla turca (and so it appears in the second edition by Schott in Mainz). That they were separated in the first print was clearly the engraver`s decision who couldn`t fit it together because of the notes` stems, flags and beams above the staff. Thus to my mind there`s no need to separate it in the modern editions.

        • Wolf-Dieter Seiffert says:

          Dear Wojciech, once again my admiration. You are really “deep in” this field. Let me send you an internal email. My blog article deals about the revised urtext edition of K. 331, – based on sources not on speculations. Best, Wolf-Dieter

Leave a Reply

Your email address will not be published. Required fields are marked *

The maximum upload file size: 5 MB. You can upload: image, audio, video, document, other. Links to YouTube, Facebook, Twitter and other services inserted in the comment text will be automatically embedded. Drop file here