Following Mozart's Così fan tutte: A Compositional History (Woodbridge: Boydell, 2008), Ian Woodfield has made an equally penetrating study of Mozart's previous opera. Readers may be forgiven for wondering at a reference (xiii) to ‘three’ Mozart operas ‘so far investigated’: Così, the Prague Don Giovanni and the Vienna Don Giovanni. Previous discussions (my own included) have indeed identified distinct Prague and Vienna versions of Don Giovanni, but they can hardly be said to be two different operas. Indeed, much of Woodfield's absorbing and meticulous study of extant sources tends towards merging, if not quite blending, these versions into a kind of Don Giovanni soup in which the choice of ingredients is left to the local chef or impresario.

Undoubtedly, this flexible model of what we understand by ‘Mozart's Don Giovanni’ represents a more realistic view of late eighteenth-century compositional and theatrical practice than does any attempt to fix the opera in a definitive form from which all variants are a departure. It has long been understood that what was performed may have varied considerably during the Vienna run (May–December 1788, after which it disappeared from the Burgtheater repertory in Mozart's lifetime), but the reasons for such changes (notably the omission of the scena ultima) remain unclear. If Woodfield is right, and there were versions that may have been considered but not actually staged, he has added a biographical layer to what was already a difficult, albeit productive period for Mozart. Woodfield also reviews sources that may have been beyond the reach of Mozart and Da Ponte, if not of their co-creators, the singers, designers and managements of the Prague and Vienna theatres. The relations between sources demonstrate two-way transmission between Vienna and Prague as well as export of scores from Vienna copying houses to other theatres, some intended for singspiel performance (in German, with dialogue replacing simple recitatives). Vienna numbers are found in material used for performance in Prague in 1791, when Mozart returned there; and although no source gives ‘any clue as to the musical text performed’, it remains possible that Mozart was ‘an active participant’ in the post-1788 evolution of Don Giovanni (129–130).

This leaves areas of uncertainty, and areas ripe for hypothesis: Woodfield admits that what he produces are precise questions, rather than answers. Nevertheless the existence of grey areas offers considerable insight into how the operatic culture of the time viewed a ‘text’ that perhaps became fixed only in the nineteenth century. What would seem aberrations to pure-minded adherents of Werktreue turn out to be normal practice, and conceivably had the composer's consent, if not his positive endorsement. The relatively common practice of placing Elvira's Vienna aria (‘Mi tradì’) in Act 1, as a response to Leporello's catalogue, makes some sense except that Mozart would not normally have had two arias for the same character in the same act and the same key (but see Woodfield's discussion of the D major version of ‘Mi tradì’, 71–73). In the 1826 New York libretto (possibly sanctioned by Da Ponte), the aria (without recitative) precedes the quartet.

Woodfield acknowledges prior work from other scholars, notably Hans Ernst Weidinger and Dexter Edge, and future researchers will find his concise introduction to their work helpful. He pushes the story on by full presentation of source evidence, often in tabular form, including thirty-one pages (158–188) of ‘page-break analysis’ by which the derivation of copies is deduced (Appendix 2; Appendix 1 details another useful method, ‘error transmission’). It would require much time and a substantial research grant to check these data, and one simply has to trust the author (which I gladly do).

The battery of evidence is formidable, but it still leaves space not only for a ‘theory of transmission’ but also for a strong hint of an ideological position about operatic texts. Woodfield offers a credible picture of ‘organized chaos’, as composer and librettist struggled to complete an opera while copyists were already at work. The creative team includes the singers; we should not consider the Vienna additions as offerings merely to placate them, or view them as a departure from, or indeed a move towards, ‘some kind of abstract, aesthetically ideal version’ (2). Edge has already noted that Ottavio's Prague aria (‘Il mio tesoro’), written for Baglioni, was copied into the performance material for Vienna, which has singers' names inserted (53); while this does not mean that the Vienna tenor Morella actually sang both arias, as most tenors do today, the possibility cannot be excluded, at least in rehearsal. The aria composed for Morella, ‘Dalla sua pace’, was sometimes transferred to Act 2 in (roughly) the place normally occupied by ‘Il mio tesoro’.

How many versions, therefore, can be distinguished? Woodfield both clarifies the position and then (for he is a realist) blurs it again. He presents the familiar differences in parallel columns (xvi–xvii), headed simply ‘Prague’ and ‘Vienna’, but eventually identifies ‘four distinct stages in the Viennese conception’ (31). Of these, Vienna 1 is Prague plus the new numbers (unwieldy). The second (source missing) is ‘a hypothetical revision’ of the ordering in Act 2 (31). Vienna 2a and 2b are versions likely to have been considered satisfactory for performance in the particular circumstances of Vienna in 1788: these are covered by the table of extant sources (5–6). But then, as mentioned above, the Vienna additions were sent to Prague, presumably with Mozart's consent (and Da Ponte's? His point of view we seldom consider; perhaps he had none). And the Vienna copying house was preparing scores for other centres, which adapted them to their own circumstances, for that particular season.

Eventually, after Mozart's death, came moves to rationalize the textual situation in the Breitkopf und Härtel full score of 1801 (136–140), apt for an age developing its belief in Werktreue: an object for Lydia Goehr's ‘imaginary museum’ rather than a resource for the theatre, and a brake on further evolution of the hybrid versions flourishing in the 1790s. Printed texts enshrine a work into versions we recognize today as ‘Prague’ and ‘Vienna’ tout court. The Neue Mozart Ausgabe score (1968) relegates the two wonderful arias and the buffo duet written for Vienna to the appendix, suggesting that only one version has true authenticity. Its much later Kritische Bericht (2003) presents a modified statement about sources – one that, of course, Woodfield modifies still further. The inescapable implication is that the concept of a complete and authentic edition is or ought to be doomed. It does not quite follow, however, that all the hard work was in vain, or that the Gesamtausgabe should be considered an aberration. There is a market for printed scores, and purchasers of masterpieces want a stout binding, not a pile of manuscript fascicles such as was prepared for the theatres. These Woodfield aptly compares to a pack of cards, ready to be reshuffled to suit the circumstances of a particular place in a particular season. Perhaps in future electronic scores may be produced that users can shuffle around; but for print there was no alternative to making an ordered selection.

Even Mozart's most extensive alterations, to Idomeneo in 1786, do not amount to a new concept: the plot is unchanged and the main effect is on nuances of characterization – also affected by different singers' approaches to identical texts. Some versions of Don Giovanni acted in the composer's and librettist's lifetimes were outside their control (most obviously the singspiel versions), and knowledge of these richly informs reception history. Probably undertaken with no intention to slight the original, they document what seemed theatrically presentable in an irrecoverable time and place; this does not afford them status as a template for later interpretations. The modern theatre is not the eighteenth-century theatre; layers of meaning have accumulated that require access to a text we can ascribe to definite, even if multiple, authorship. We may like to imagine ‘out there, somewhere, ontologically, is the “real” Don Giovanni … enshrined mostly in Mozart's score and the printed libretto’; versions traceable to Mozart have ‘privileged status, in relation to which all performances and all other kinds of interpretations are secondary or subsidiary’ (Richard Will, ‘In Search of Authentic Mozart’, News of the National Humanities Center, Fall/Winter 2010, 1; <http://nationalhumanitiescenter.org/newsletter2010/nhcnewswinter2011.pdf> (4 July 2011)). This is surely right, but following Edge, Tyson and others, Woodfield makes tellingly clear the pertinence of that ‘mostly’; and he points to the irony of performances today going ‘authentic’ just as ‘the academy’ is beginning to take a more flexible view of such texts. We are indebted to him for presenting the ingredients that make up the early forms of Don Giovanni but we should not regard it as intrinsically wrong to adopt a version of nearly identifiable authorship rather than remixing the Don Giovanni soup for every modern production; we can safely leave that to the stage director.