King's Music

Bach's St. Matthew Passion

The Original Version - Peter Holman

Reviewing a new edition of the St John Passion last month, CB made the point that during the last thirty years we have become accustomed to the idea that major Baroque works such as Messiah and the St John Passion tend to exist in more than one version. A number of people (including myself) have performed the 1725 version of the St John Passion, attracted by the quality of the arias and chorale settings Bach added to the work, and perhaps worried by the fact that the version normally performed today is a hybrid. Bach began a revision of the work around 1739, but only got as far as no. 10, so after that the ‘standard’ version reverts to the original 1724 text. To do the 1725 version (or the 1724 version, for that matter) you need to use the unrevised versions of the first few numbers, though unfortunately they are only included in the Bärenreiter full and study scores, not in the vocal score or the set of parts.

It is less well known that there are two versions of the St Matthew Passion. The version normally performed today comes from a revision Bach made for a performance on Good Friday 1736, and survives in a beautiful fair-copy autograph score and a set of parts. As anyone who has taken part in a performance of the work knows, Bach specified his intentions in the later version with especial care. The instrumentation is worked out in precise detail, there are numerous slurs and expression marks, and the length of the continuo notes in the recitatives is specified exactly, contrary to the normal practice of the period.

It comes as something of a shock, therefore, to open the facsimile of the score copied by Bach’s son-in-law Johann Christoph Altnickol (1720-59); it is published by Bärenreiter as a supplement to Alfred Dürr’s Neue Bach-Ausgabe volume (Kassel, 1972). Altnickol must have made his copy of the St Matthew Passion after 1744, when he arrived in Leipzig to study at the University and became drawn into Bach’s circle, though it is likely that he had in front of him a source that was considerably earlier than the 1736 score. We know that the work was first performed on Good Friday 1727 and again in 1729, and it has been suggested recently that Bach began to compose it as early as 1725. Altnickol’s score seems to preserve this first version, and its almost complete lack of slurs, dynamics and expression marks suggests to me that the copy source was Bach’s original composing manuscript, made at a stage before he brought the work to performance. Unfortunately, Altnickol was none too accurate a copyist, so it is often difficult to distinguish between genuine variants and his errors.

The most striking difference between the Altnickol score and the 1736 version is that there is only one continuo line, written at the bottom of the system in the double choir movements. This seems to imply that in the original version the voices and upper instruments were deployed antiphonally around a single central continuo group. In the later version there are two separate continuo parts, both marked ‘organo’, presumably for the main St Thomas’s organ and a portable instrument. Perhaps this change of scoring reflected some change in the way Bach’s forces were laid out, though Bach scholars seem to be agreed that both the 1727 and the 1736 performances took place in the main choir and organ lofts at the west end of St Thomas’s, Leipzig. One change in the layout that does seem to have been made is that the chorale melody ‘O Lamm Gottes’ in the first chorus was sung in 1736 by ripieno sopranos supported by the organ of the ‘swallows nest’ gallery, across the nave from the main galleries. In the Altnickol score the chorale is marked ‘Organo’, is mostly doubled in octaves by the flutes and oboes, and is not underlaid after the first two phrases. This suggests that in the original version the chorale was played rather than sung, as in many movements elsewhere in Bach’s sacred music; the ‘organo’ was presumably the main instrument, not the one in the ‘swallows nest’.

Bach made a number of other major changes to the 1736 version. He added a second version of the ‘Passion Chorale’ (no. 17) at the beginning of the Garden of Gethsemane scene, and replaced the simple chorale ‘Jesum laß ich nicht von mir’ at the end of Part 1 with the concerted chorale setting ‘O Mensch, bewein’ (no. 29), which he had used to begin the 1725 St John Passion, and which probably came from a lost Weimar passion. In the Altnickol score ‘Ach, nun ist mein Jesus hin!’, the first number of Part II is a bass solo rather than an alto solo, while in the arias ‘Erbarme dich’ (no. 39) and ‘Gebt mir meinen Jesum wieder’ (no. 42) the violin soloists come from the second and first choirs respectively; in the 1736 version it is the other way round. Most striking, in Altnickol’s score the aria ‘Komm, süßes Kreuz’ and the preceding recitative (nos. 56 and 57) have lute rather than viola da gamba obbligato. The parts are almost identical, but the aria is a much more private contemplation of the Cross with a lute rather than a viol, requiring an intimate style of singing from the bass soloist. It is likely that there was no viola da gamba at all in the original version: there is no part for it in the recitative ‘Mein Jesus schweigt’ (no. 34), and in the following aria ‘Geduld’ (no. 35) the bass part is unmarked; it is still marked ‘Violonc. e org’ in the autograph score, but is allocated to ‘Viola da Gamba’ in the set of parts. Another interesting change of scoring is in the recitative ‘O Schmerz!’ (no. 19), where the parts marked ‘flauto’ (recorders) in the later version are labelled ‘2 Travers:’ (i.e. flutes) in the Altnickol score. Is this the only example, at a time when the recorder was rapidly being superseded by the flute, of a composer replacing flutes with recorders?

Bach also made many detailed changes to the 1736 version. Some of them consist of minor changes to melodic lines and harmonies, though in a few cases, such as ‘Geduld’ (no. 35) and the aria ‘Ich will bei meinem Jesu wachen’ (no. 20), the changes amount to recomposition. Many of the minor changes involve the addition of ornamentation to vocal and instrumental lines. For instance, it is not clear whether he intended the main idea of ‘So ist mein Jesus nun gefangen’ (no. 27a) to be sung and played entirely without appoggiaturas, as in the Altnickol score, or whether he expected them to be added by the performers. Similarly, the main solo violin theme of Altnickol’s version of ‘Erbarme dich’ has no slide on the first down beat, and no appoggiaturas in the first two bars. Again, it is not clear whether Bach intended the performers to add the 1736 ornaments in performance; a possible solution would be to add some or all of them to subsequent appearances of the theme.

Why perform the early version of the St Matthew Passion? Although editors and performers often prefer the final versions of composers’ works – the received version of Messiah is closer to Handel’s versions of the 1750s than to the original 1742 Dublin version – it has become common to explore and perform early versions of Bach, if only to throw light on his subsequent revisions. Thus in recent years we have had recordings of the first versions of the Brandenburg Concertos, the Orchestral Suites, the presumed original string or wind versions of the harpsichord concertos, and a number of the cantatas, including ‘Christ lag in Todesbanden’ BWV4 and ‘Himmelskönig, sei willkommen’ BWV182. So far as I can discover, the 1727 version of the St Matthew Passion has never been performed in this country, and it is possible that it has never been revived anywhere since the eighteenth century. If so, then the performance at Stoke by Nayland in Suffolk on 26 May needs no justification; I hope it will provide new insights into a great masterpiece.

CD reviews

Bach St Matthew Passion I: Deborah York, Magdalena Kozena, Mark Padmore, Peter Harvey SATB; II Julia Gooding, Susan Bickley, James Gilchrist, Stephan Loges SATB; Gabrieli Players, Paul McCreesh 161’ 32” (2 CDs) Archiv 474 200-2

Bach Sacred Vocal Works Various soloists, Monteverdi Choir, English Bach Soloists, John Eliot Gardiner 510’ 01” (9 CDs) Archiv 469 769-2 (rec 1985-8)

I’ve placed these two sets together because I happened to listen to the Matthew Passions on consecutive days and the Gardiner version is a good example of the earlier generation of early-music performance with which to compare the McCreesh. The most notable feature of the latter is the relationship of voice and instruments in the arias: I’m not degrading the importance of the conductor by saying that they sound as if he isn’t there and that the musicians are making chamber music together. The two outstanding voices are the ladies of Choir I, with Deborah York by no means outclassed by the more famous Czeck mezzo. The other six (+ the ripieno soprano Ulla Munch) are all fine; one would normally single out the evangelist, and Mark Padmore is excellent, but I don’t think I’ve heard a Bach Passion with a poor evangelist: if you’re no good, you don’t get the job!

McCreesh takes seriously Bach’s layout into two choirs. But Gardiner’s Anne Sofie von Otter starts off in Choir I and finishes in Choir II while his less-satisfactory Ann Monoyios moves in the opposite direction. A distinctive characteristic of the Matthew Passion is the lay-out for two ensembles; so it seems a bit perverse in a no-expenses-spared recording to ignore that, especially since the Gardiner disc has such clear stero spacing.

The feature of the McCreesh peformance that has attracted particular publicity is the use of just nine singers. If you have no preconceptions that the work is for chorus, it makes perfect sense of the score and surviving performance materials (except that implies additional singers for the Ancillae & Pilate’s wife, for Judas & Priest 1, and for Peter, Priest 2 & Caiphas – but earlier versions may have differed). In fact, I found that the small ensemble felt just right (apart from the opening chorus taking a while to settle). The part-writing was absolutely clear, without the need of individual lines of interest to be artificially brought out, and the balance with the orchestra sounded better than in Gardiner’s recording.

One great assets of the McCreesh recording is the use of proper organs, not the usual portables. Gardiner’s allocation of his three chamber instruments is another example of his lack of interest in the two-choir layout of the forces, with one for the evangelist, one for arias and a third chiefly for choruses and chorales. Bach may have originally performed the work with one, but McCreesh takes advantage of two fine new instruments by Marcussen in Roskilde Cathedral (the magnificent old instrument heard in previous recordings there is not used), and it is thrilling to hear a proper organ sound, especially in the dramatic sections of recitative, played confidently by James Johnstone.

The McCreesh recording shows the advantages of looking at the music, trusting what it seems to imply about how it should be performed, getting the best performers available, then working with them to present the music with understanding, sympathy and skill. The music can seem overpowering and heavy (especially the opening chorus). McCreesh’s lighter manner illustrates Komm rather than klagen; is it fanciful to detect a pastoral tone in the 12/8, reminding us of the last big feast of the Church year, Christmas, hinted by the words ‘als wie ein Lamm’? Gardiner takes nearly a minute longer for this movement, with four heavy beats per bar. Curiously, he is a few minutes shorter over the whole work, though I wouldn’t have guessed it. I’m not sure whether his analysis of the story into sections helps the listener. The McCreesh version flows better; it is also virtually a mid-price set, since it is on two discs rather than three. Buy it, forget about the single-voice controversy, and just enjoy it: I recommend it strongly.

The rest of the Gardiner package includes the Christmas Oratorio, St John Passion and Mass in B minor, nine discs with a fat booklet containing excellent notes but no translations. (An annoying feature of the McCreesh booklet, incidentally, is the way that the texts for the second disc are numbered as in the score rather given the track number). Things have changed since the late 1980s, and I suspect that the conductor himself might like the chance to re-record (though he is unlikely to change his mind about the vocal forces). But this set offers good performances and can be recommended to anyone wishing to buy such a packet: perhaps a useful present (along with the Bärenreiter scores) for a young relative studying music.

CB