As is now generally agreed, so-called ‘authentic’ performances of baroque music are modern in conception and sound. The point was argued convincingly during the 1980s by Laurence Dreyfus, Richard Taruskin and others (Dreyfus, ‘Early Music Defended against Its Devotees: A Theory of Historical Performance in the Twentieth Century’, The Musical Quarterly 69/3 (1983), 297–322; Taruskin, Daniel Leech-Wilkinson, Nicholas Temperley and Robert Winter, ‘The Limits of Authenticity: A Discussion’, Early Music 12/1 (1984), 3–25; Nicholas Kenyon, ed., Authenticity and Early Music: A Symposium (Oxford: Oxford University Press, 1988)). Radically separated from nineteenth-century performance traditions, the reconstructed ‘old’ sound was exciting and new, and, for me at least, there have been occasions when these ‘authentic’ performances have been revelatory, shedding light on music I thought I knew (although, to be fully confessional in terms of the era in which it all began, I also loved the Swingle Singers and Wendy (Walter) Carlos's Switched-On Bach).

To be clear: there can be little question that in the past fifty years intensive scholarly research and, to at least an equal extent, practical experimentation have successfully recaptured important elements of former performing practices, and also that these practices have injected an invigorating element into a wide range of concert music. Still, early-music performance largely remains a modern interpretation for a modern audience. Since we have neither the will nor the way to recreate the context of performances from an earlier era, there is no way authenticity can be anything other than a construction, something that becomes especially clear to me as I speed down the road listening to one of my ‘authentic’ recordings. In fact, when I'm travelling with Handel (whom I imagine in the front passenger seat), I worry what would be the most disturbing to him: the speed, the music coming from nowhere, or the sound itself. Would he recognize his own work in that context?

Despite advantages and insights generated by the early-music movement, there have also been losses. Taruskin bemoans an absence of the musicianship and commitment (the ‘authenticity’) that embodied great performances of baroque music in the romantic tradition and reaffirmed the continuing importance of a composer and his music in the present, singling out as an example the performance of Bach's Brandenburg Concerto No. 5 at the Salzburg Festival in 1950 with Wilhelm Furtwängler, piano, Josef Niedermayr, flute, and Willi Boskovsky, violin (Taruskin, ‘The Pastness of the Present and the Presence of the Past’, in Authenticity and Early Music, ed. Kenyon, 157). Geoffrey Burgess, in his Editorial in the last issue of this journal, considers various ways to get beyond what can be stifling ‘historical correctness’ by delving more deeply into historical practice and ‘creating music that moves our audiences’; he points to more autonomy for musicians involving performance from original scores or ‘diplomatic transcriptions. . .[using] minimal editorial intervention’, reconnection with the social context of early music away from formal concert halls, and renewed interest in rhetoric and gesture. My concern here is rather with the loss of a performing tradition that the modern ‘authenticity’ movement has largely eradicated: deliberate adaptations and reorchestrations that brought music from earlier periods into line with contemporary practice in order that it might retain currency. Aside from a few examples like Mozart's reorchestrations of Handel, which are esteemed because of the value ascribed to the adaptor, there are many others of historical interest that are neglected or forgotten. Just as we are interested in the past, so was the past interested in the past, and their use of early music is as important a view into their present as our ‘authenticity’ movement is to ours.

Let us consider for a moment a slippery syllogism based on the definition of ‘authentic’ performance as modern: if ‘authentic’ performances are based on contemporary practice, and inauthentic adaptations are based on contemporary practice, then ‘authentic’ performances are (or equal) inauthentic adaptations. Of course, if A=B, then B=A, so it must be true not only that ‘authentic’ performances are inauthentic adaptations, but also that inauthentic adaptations are ‘authentic’. It is this formulation (however false the syllogism may be) that I want to pursue: that adaptations of older music to the music of the present are (importantly) authentic to the period in which they are created.

I first became interested in this topic many years ago when I came into contact with late eighteenth-century adaptations of Purcell (Ellen T. Harris, Henry Purcell's ‘Dido and Aeneas’ (Oxford: Oxford University Press, 1987), 130–147, and Harris, ‘King Arthur's Journey into the Eighteenth Century’, in Purcell Studies, ed. Curtis Price (Cambridge: Cambridge University Press, 1995), 257–289). In addition to various idiosyncratic differences among the adaptations, there were also generic changes that brought Purcell's style ‘up to date’: the formal flexibility and metrical proportions apparent in the elision of Purcell's movements were eliminated in favour of clear cadential structures, double bars at ends of movements, and the alteration of rhythmic note values where necessary. Rhythmic and harmonic structures considered irregular were smoothed over, Scotch snaps reversed or evened out and successive non-harmonic notes eliminated. New orchestral accompaniments were added, phrase structures regularized, long melismas texted and declamatory airs (considered recitative) cut. And, hey presto, Purcell's theatrical music was transformed into a late eighteenth-century number opera. In writing about these adaptations, I suggested it ‘might even be refreshing’ to hear Purcell's works in this way, endorsing Christopher Spencer's view of restoration adaptations of Shakespeare, including Nahum Tate's King Lear (with its happy ending): that adaptations can be enjoyed as new works apart from the original, tell us about the period of the adaptation and shed new light on the original (Five Restoration Adaptations of Shakespeare (Urbana: University of Illinois Press, 1965), 32). My enthusiasm was not widely shared. As far as I know, there have been no attempts to recreate and perform the eighteenth-century versions of Dido and Aeneas or King Arthur.

But why not? We've had an ‘authentically’ performed Dido and Aeneas set on a long white bench at the lip of the stage in costumes of no specific time period with a trio of countertenors as the Sorceress and her two witches (Barrie Kosky, director; Frankfurt Opera, 2010; Los Angeles Opera, 2014), and another performed by the Akademie für Alte Musik Berlin and Vocalconsort Berlin that was choreographed by Sasha Waltz (Staatsoper Unter den Linden, Berlin, 2005; Sadler's Wells Theatre, London, 2007) in which the dancers began completely submerged underwater in large tanks, individual characters were portrayed by up to three different people and the costumes were ‘a postmodern ragbag of tunics, crinolines, riding hats and Wellington boots’ (Judith Mackrell, The Guardian, Friday, 16 March 2007, accessed online 15 July 2015). An ‘authentic’ concert performance of the eighteenth-century adaptation of Dido for the Academy of Ancient Music might be a revelation in such company.

Nineteenth-century adaptations of Handel's music fascinate me for similar reasons. My first exploration of this topic resulted from the appearance at the Houghton Library at Harvard University of some manuscript copies of cantatas by Handel in the hand of Pauline Viardot (‘Viardot Sings Handel (With Thanks to George Sand, Chopin, Meyerbeer, Gounod, and Julius Rietz)’, in The Fashion and Legacy of Nineteenth-Century Italian Opera, ed. Roberta Montemorra Marvin and Hilary Poriss (Cambridge: Cambridge University Press, 2010), 9–29). What really stood out to me were Viardot's curatorial interest and extraordinary musicianship. It was Viardot, of course, who purchased the autograph of Mozart's Don Giovanni in 1855 after the Österreichische Nationalbibliothek in Vienna, the Königliche Bibliothek in Berlin and the British Museum in London declined the opportunity. She also owned an autograph of Bach's, and the Abbé Fortunato Santini gave her a manuscript of arias from his collection.

In 1859 Hector Berlioz adapted Gluck's Orphée for Viardot in the title role – another adaptation that has retained some currency today. ‘J'ai perdu mon Eurydice’ became possibly the aria she performed most often in recital, as it gave her the opportunity to emphasize the striking contrasts within the rondo form and to vary the repetitions of the refrain. Berlioz wrote a detailed description of her performance of the aria (‘Viardot sings Handel’, 22). For the opportunities they offered in terms of contrast and variation, rondos also became a focal point of her interest in Handel. Massenet reorchestrated ‘Lascia ch'io pianga’ for her by 1848, Gounod prepared a reorchestration of ‘Verdi prati’ around 1850 and Julius Rietz presented her with a copy of ‘Where shall I fly? – See the dreadful sister rise’ from the Chrysander edition of Handel's Hercules in 1859, the same year as the premiere of Berlioz's Orphée. We have multiple recordings of Handel's ‘Arias for Cuzzoni’ sung by Lisa Saffer (Harmonia Mundi 907036, 1991) and Simone Kermes (Berlin Classics 16422, 2009), among others, a ‘Tribute to Faustina Bordoni’ with Vivica Genaux (Deutsche Harmonia Mundi 88691944592, 2012) and arias for Durastante sung by Lorraine Hunt (later Lorraine Hunt Lieberson; Harmonia Mundi 907056, 1992). Why not arias and rondos by Handel and Gluck as sung by Pauline Viardot and reorchestrated by Berlioz, Massenet and Gounod, with additional repertoire from her edition of arias from the ‘classical period’, École classique du chant (1860), for which she provides a keyboard accompaniment and detailed performance instructions? The preserved score of Mendelssohn's ‘revival’ of Bach's St Matthew Passion, which includes, among other alterations, very significant cuts and changes of harmony and orchestration, has been published (Bärenreiter), performed several times and recorded (Naïve 20001, 2002); Mendelssohn's charming reorchestration of Handel's Acis and Galatea is now easily available, as well, in a published score (Carus) and recording (Carus 83.420, 2009). Given Viardot's equal concern with recapturing music from the past, a recording of arias as sung and edited by her could offer additional insight into what might be called the ‘authenticity’ movement of the nineteenth century.

I have never been a fan of the Mozart reorchestrations of Handel. I grant that they are historically important, but they seem neither here nor there to me: not Handel and not Mozart either. Christoph Wolff has argued that van Swieten was the guiding force behind Mozart's arrangement of Messiah (‘Mozart's Messiah: “The Spirit of Handel” from Van Swieten's Hands’, in Music and Civilization, ed. Edmond Strainchamps and Maria Rika Maniates (New York: Norton, 1984), 1–14). Indeed, the goal seems to have been to keep as much as possible of what Handel wrote: according to David Schildkret, forty per cent of the numbers ‘retain Handel's original instrumentation’ (‘On Mozart Contemplating a Work by Handel: Mozart's Arrangement of Messiah’, in Festa Musicologica: Essays in Honor of George J. Buelow, ed. Thomas J. Mathiesen and Benito V. Rivera (Stuyvesant: Pendragon, 1995), 132). With the exception of a few well-known alterations – ‘Rejoice greatly’, for tenor; ‘The trumpet shall sound’, with trombones (although, as Wolff points out, the German biblical phrase about the last judgment calls for trombones (die Posaunen) rather than trumpets); and the replacement of the final aria, ‘If God be for us’, with a recitative – Mozart sticks close to the original. He rather judiciously adds winds throughout, but also, in general, leaves continuo-accompanied movements and passages as they are.

It is on this point of the ‘filling-up’ of the harmony that late nineteenth-century reorchestrations differ significantly from Mozart's. A good example, if not the prime example, of this practice is the score published in 1884 by the great Halle lieder composer Robert Franz with the title Der ‘Messias’. Oratorium von G. F. Händel. Unter Zugrundlegung der Mozart'schen Partitur mit den Nöthigen Ergänzungen, herausgegeben von Robert Franz (The Messiah. Oratorio by G. F. Handel. Founded upon Mozart's Score, with the Necessary Completions, edited by Robert Franz). In his enthusiastic, extended and somewhat dyspeptic defence of this adaptation against Franz's critics, Ebenezer Prout hammers home the point clearly stated in the title: Franz begins with Mozart and adds ‘the necessary completions’ (The Monthly Musical Record 21 (1891), 73–77, 97–100, 105–111, 122–125, 150–153). A comparison of the scores of Handel, Mozart and Franz confirms this statement, but also reveals their many additional differences.

Franz's score gives us the first, great adaptation of Messiah in the post-keyboard-continuo era. Therefore, in terms of filling up the harmony, where Handel writes simple recitative and Mozart leaves it as is (to be filled in by the fortepiano or harpsichord), Franz provides a simple harmonization for string quartet. In arias where Handel includes continuo passages and Mozart leaves them as is, Franz generally provides a four-part harmonization for clarinets and bassoons. Prout regularly praises the judiciousness of Franz's additions, as in the aria ‘But who may abide’:

In a few places Franz eliminates Mozart's additions, as in the recitative, ‘Behold I tell you a mystery’, where Mozart's brass parts are deleted.

As with the additional accompaniments, Franz retains most of Mozart's added counterpoint and goes on to introduce further instances. For example, he not only keeps the counterpoint added by Mozart to the long tenor runs in ‘Ev'ry valley’, but also, after reassigning ‘Rejoice greatly’ to soprano, adds an independent viola line in counterpoint with the voice where Mozart had the viola double the bass – as well as a cadenza of sixty-two notes that would have suited Pauline Viardot perfectly.

A significant difference between the three scores occurs in the use of instrumental doubling in the choruses. Handel originally composed the work for strings only. By 1745 he had added oboes and bassoons as doubling instruments to strengthen the sound, and, as found in the surviving parts, he frequently has the oboes double the soprano line in the choruses. In contrast, as also indicated in surviving parts, Mozart doubles the choral alto, tenor and bass lines with trombone ‘in the style of Viennese church music’ (Schildkret, ‘On Mozart Contemplating a Work by Handel’, 141; alas, the decision of Andreas Holschneider, ed., Der Messias, W. A. Mozart Neue Ausgabe sämtlicher Werke, series 10/28/1, volume 2 (Kassel: Bärenreiter, 1960) not to include the trombones in the critical score but only in the Preface has apparently led to them being omitted in many recordings). Franz, as far as I can tell from the published score, included no standard doublings in the choruses. The differences become clear in the chorus ‘And the glory of the Lord’: where Handel doubles the soprano with oboe, Mozart doubles the alto, tenor and bass with trombone, and Franz, without adhering to any regular choral reinforcement, doubles the chorus with clarinets and bassoons in the passage (‘shall be revealed’) where Handel's original accompaniment was for continuo alone (bars 17–33), thus following his pattern of using these instruments to replace the continuo. My sense from reading the scores is that the doublings provide as good an indication as any of the different sound quality at the heart of the three versions.

How I would like to hear a good performance of Franz's Messiah today. Some sense of its lush sound can be garnered from the lovingly performed Messiah (reorchestrated but not specifically in Franz's version) recorded in 1927 with the BBC Symphony Orchestra and Choir conducted by Sir Thomas Beecham (GEMM CDS 9456). (This is not the so-called ‘Beecham Messiah’ of 1959, a recording of a newly commissioned orchestration by Sir Eugene Goossens with its famous, or infamous, as you will, cymbal crash on the opening chord of the ‘Hallelujah Chorus’, which version, however, also has its delights.) What may be hard to fathom, but fits perfectly with our revised sense of ‘authentic’ performances today, is that the Franz version of Messiah, based, like the modern early music movement, on ‘profound scholarship’ and ‘artistic honesty’, was considered the most authentic version of its day:

The New York Times waxed enthusiastically about the Franz Messiah on multiple occasions. The first mention, on 26 September 1885, comes in an anonymous review of a performance in Worcester, Massachusetts: ‘The new orchestration of Robert Franz was used for the first time in America, and some novel effects were produced. A pyrotechnic cadenza was introduced in the soprano air, ‘Rejoice greatly’, which Mrs. Barton sang with thrilling effect’ (all cited articles from The New York Times viewed at <www.nytimes.com> in TimesMachine on 23 July 2015). The following day brought a review of one ‘A. de G. S’. (with a dateline of 28 August) of the premiere of Franz's Messiah in Birmingham, England: ‘Much interest was expressed in this rendering because it was known that Herr Richter intended introducing the revised score of Robert Franz for the first time in England, it being Richter's first conductorship of Handel's masterpiece in any country, and, also, it was the nearest and closest reproduction of Handel's original performance that it was possible to obtain. . .’ (my italics).

Modern orchestras have largely been closed out of performances of the great baroque oratorios by the early-music movement. Although some symphony orchestras have succeeded in replicating the sound of early music on modern instruments, I don't believe they will ever reclaim baroque repertoire by restraining the power of the instruments in their hands, especially as ever-more-attenuated performances that have no basis in historical practice achieve critical acclaim (see, for example, the ‘chamber’ Messiah performed by Solomon's Knot with eight singers providing all the choruses and solos and only fourteen players). But today's orchestras could compete by instituting their own ‘authenticity movement’ and offering rich and full-throated performances of nineteenth-century adaptations. Although I had no takers on performances of eighteenth-century Purcell, I'm hoping I won't be the only person who would find a beautiful performance of Franz's Messiah very appealing.