Pieter Dirksen

Harpsichordist - Musicologist - Organist

Books & Editions
Papers & Lectures
Online Essays
Thematic Overview
of Publications
Concert Reviews

The Earliest Version of Bach’s St. John Passion

Pieter Dirksen

When Johann Sebastian Bach was appointed to the post of Thomascantor in Leipzig in the spring of 1723, the only requirement was that he was to produce cantatas on a regular basis; he was not officially required to compose passions as well. It is not surprising that the citizens of Leipzig would sooner have chosen Telemann and Graupner (both of whom remained at their old jobs after receiving substantial raises in salary) before they settled for Bach. At that time, Bach had not yet established any significant reputation as a composer of church music, whereas Telemann and Graupner already had multiple full-year cycles of cantatas to their credit. In addition, Bach had not composed any new church music for some years, because Lutheran church music did not play a role at the court of Cöthen, where he was master of music from 1717 to 1723. Bach never lived down the title of ‘court music master’ once he came to Leipzig, and he did not seem to regret it very much. On the contrary, he himself never used the title of ‘Cantor’. So it was no idle gesture for the city council to have Bach sign a declaration in which he promised to produce pieces, ‘welche nicht opernhaft herauskommen, sondern die Zuhörer vielmehr zu Andacht aufmuntere’ [‘would not be of an operatic character, but rather would excite the listener to greater Piety’].

In comparison with many other North and Central German cities, Leipzig was definitely late in its establishment of a tradition of passion music. As we know all too well from Bach’s biography, it was a city which was primarily a conservative and orthodox community, long resistant to the performance of oratorio settings of the passions, i.e. passions based on the model of the Italian opera. This was not allowed until as late as 1717, and even then it was not in a principal church but rather in the smaller Neukirche; the first work to be performed was Telemann’s Brockes Passion. This was such a success that the church worthies were more or less compelled to continue it, and in 1721 and 1722 the aging Thomascantor Johann Kuhnau performed passions of his own composition in the Thomaskirche. These, however, were not large-scale concertante works, but rather passion settings with very simple choruses and songlike solo segments. Kuhnau’s successor, J.S. Bach, was in a position to build on this tradition, and he did so with the greatest alacrity. After all, a passion was far more independent of the liturgy than a cantata, not only because of its length, but also because of its dramatic character. And so, within a year of his appointment, on 7 April 1724, Bach’s passion music after the Gospel of John received its first performance. In contrast with Kuhnau, Bach, of course, did not compose a ‘churchly’, modest passion, but instead made use of every means available to him. He was wise enough to follow familiar traditions in his choice of text. In addition to the exact text of the passion from John, and a selection of appropriate chorale verses, he mostly selected his free texts (the arias and closing chorus) from verses drawn from Barthold Heinrich Brockes’s very popular contemporary retelling in verses of the passion (‘Der für die Sünde der Welt leidende und sterbende Jesus’ [‘Jesus suffering and dying for the Sins of this World’], 1713) familiar to the Leipzigers since 1717. An unknown text-writer was entrusted with the task of simplifying the somewhat turgid and circuitous language of Brockes’s poem for Bach. The text of no. 20 is a good example:


Dem Himmel gleicht sein bunt-gestriemter Rücken
Den Regen-Bögen ohne zahl
Als Lauter Gnaden-Zeichen schmücken;
Die (da die Sünd-Flut unsrer Schuld verseiget)
Der holden Liebe Sonnen-Strahl
In seines Blutes Wolcken zeiget.

[His blood-striped back the heavens so resembles
which, as sign of purest Mercy, are adorned with Rainbows numberless,
which (for it atoneth for our flood of sin), the Sunshine of Gracious Love
strikes from the clouds of his Blood]

Bach, St. John Passion:

Erwäge, wie sein blutgefärbter Rücken
In allen Stücken
Dem Himmel gleiche geht
Daran, nachdem die Wasserwogen
Von unsrer Sündflut sich verzogen,
Der allerschönste Regenbogen
Als Gottes Gnadezeichen steht!

[Consider, how his blood-stained back,
In all its parts
The heavens resembleth
Where, after the floods of our sins have ebbed away,
The fairest of rainbows shines, as a sign of God’s mercy!]


The St. John Passion has a special place in Bach’s oeuvre: it is the only one of his 'major' works which never achieved a definitive form. And moreover it went through a complicated series of revisions and alterations. After the first version of 1724, he performed it again a year later, reshuffling a number of choruses and arias so that the piece would still seem like a ‘new’ composition. This says a lot about Bach’s relatively slow rate of production (in contrast to what one often reads in the literature, he was anything but facile): he was apparently unable to produce a complete new passion within the timeframe of one year. And it also suggest something of the intense experience in those days of a single performance on Good Friday—a great difference from our contemporary tradition of numerous performances [in the Netherlands] over the course of several weeks. In 1728 or 1732, Bach returned to the 1724 version, but cut the added texts from the Gospel of St. Matthew (in no. 12 , the closing sentence ‘Da gedachte Petrus an die Worte Jesu und ging hinaus und weinete bitterlich’ [‘Then Peter bethought himself of the words of Jesus, and went out, and wept bitterly’], and no. 33: the rending of the Veil of the Temple), and as a result introduced several changes, resulting in a slimmed-down, ‘pure’ passion according to St. John.

In 1739, it is highly probable that Bach began work on a new score (Berlin State Library, Mus. Ms. Bach P 28 – see the illustration below), with which he returned to the initial version of 1724, and where, as was usual for him, he could not resist making numerous corrections and revisions. Unfortunately, he broke off this work after only 10 numbers, about a quarter of the way through the passion. Most likely he received notice from the town council while at work to the effect that there could be no performance of the passion that year, to which Bach coolly answered that ‘… er hätte ohnedem nichts darvon, und wäre nur ein Onus, er wolle es den Herrn Superintendenten melden, dass es ihm wäre untersagt worden, wenn etwa ein Bedencken wegen des Textes gemacht werden wolle, so wäre socher schon ein paar mahl aufgeführet worden’ (‘… in any case it was not important for him, and only a burden, and that he would tell the superintendent [i.e. the principal officer of the church] that if this were owing to objections about the text, that this passion had already been performed previously a couple of times [without objections]’). It would be characteristic for the temperamental Bach to have tossed aside his pen in the middle of no. 10 (measure 42) and put away the manuscript, since the Evangelist’s announcement ‘Jesus aber antwortete’ [‘And Jesus then said …’] remained uncompleted.

First page of the score of the St. John Passion, showing the beginning of the work (autograph, ca. 1739). Berlin State Library, Mus. Ms. Bach P 28, fol. 1r.

And so the interference of the Leipzig authorities succeeded in depriving us of a definitive version of the score, such as the one we have for the St. Matthew Passion, which was produced in 1736. Bach apparently lost interest in passions after 1739: when he performed the St. John Passion, after all, on another occasion in 1749, he had the 10 year old score completed by a student (recently identified as Johann Nikolaus Bammler, who entered the Thomasschule in 1737 and remained in Leipzig until 1750), but he made hardly any changes in it. Bach did not even bother to introduce (or have the copyist introduce) the improvements made in 1739 in the first 10 movements in the performing parts.

The original 1724 score has unfortunately been lost, and all that remains of the parts made for the first performance is a set of so-called ‘doublets’ or copies of the primary parts for ripieno singers and players (four vocal parts, two violin parts, and a continuo part). They survived as part of a larger collection of parts which document in their totality all of the versions of the St. John Passion, and contain numerous crossings-out, pasteovers, and restorations (Staatsbibliothek zu Berlin, Mus. Ms. Bach St 111). All of the soloists’ parts, i.e. those of the Evangelist and the other dramatic parts, the vocal parts of the arias and the obbligato instrument parts, must therefore be reconstructed from the later sources. Fortunately, we have a sure guide in the extra 1724 continuo part, which in fact documents the entire passion. From this it can be seen that the earliest version was indentical, in its major characteristics, to that of 1739/49. The principal source for nos. 11 through 40 consists of Bemmler’s copy made after the lost autograph score of 1724, augmented with information from the seven 1724 ripieno parts (to the extent that we can still identify the oldest readings in this material). These parts are the principal basis for the reconstruction of the first ten movements; for the oboe, viola, and solo vocal parts, as well as for the obbligato part of ‘Ich folge dir gleichfalls’ [no. 9] we must refer to readings from parts taken from the 1725 and 1728/32 versions. The surviving sources do not permit complete coverage for a reconstruction, but we are convinced that this amounts to a reasonably close approach to the oldest version.

The fact that Bach broke off his careful copying and revising after no. 10 has effectively deprived us of any valid ‘definitive’ version of the St. John Passion, and as a result, all four versions performed by Bach have an equal degree of historical validity. However, we only have complete information about the second (1725) and fourth (1749) readings. As for the third version (1728/32), we do know what was cut, but not what was added in its place (a matter of an aria and a sinfonia). This version, therefore, cannot be reconstructed.

We have more information about the first version, because this was compositionally largely identical to the last version, with some differences in detail. Of course we omit the amplifications in the first ten movements from 1739, as well as various other changes in the rest of the passion—as far as they can be determined. The palette ranges from minimal, barely noticeable variants, to a few significant changes such as the less streamlined figuration in the upper strings in the opening chorus, the different nature of the continuo part for no. 9 (‘Ich folge dir gleichfalls’), and the style of the chorales. The two chorales which can be found in the autograph portion of the 1739 score were adapted by Bach to conform to his later style by the addition of more passing notes and ornamental figures. We can also see something of the same type of revision in the aria ‘Von den Stricken meiner Sünden’ [no. 7]: Bach has added inimitable refinements to the alto solo here with ornamental figures and suspensions, making a ‘more elegant’ piece. Nevertheless, the more elemental 1724 part is musically just as acceptable.

We encounter a few important variants in the last part of the St. John Passion. The recitative ‘Und siehe da, der Vorhang im Tempel …’ [no. 33], an interpolation from the Gospel of St. Matthew, originated with the second version, from 1725, and at that time it replaced a much shorter recitative about the same event, drawn from St. Mark: ‘Und der Vorhang im Tempel zerriss in zwei Stück, von oben an bis unten aus’. Only the continuo part of this original 1724 recititative has survived, and the vocal part of our version is a reconstruction (made by the undersigned). A separate problem is formed by the arioso ‘Mein Herz, indem die ganze Welt’ [no. 34], where we do not know how it originally appeared; it is the only movement where the flutes have independent parts and obviously do not double any originally implied instruments (because these instruments are already in play). The four-voiced woodwind harmonies have here been reduced, hypothetically, to two voices.


In Bach’s first year as Thomascantor he did not yet have transverse flute players available, and so during his work on the St. John Passion he obviously would not even have anticipated these instruments. The instrumentation of two oboes, strings, and continuo is the standard scoring for most of the cantatas for Bach’s first year in Leipzig, resulting in a sound with a definite oboe sonority. The lack of flute sound, which we so strongly associate with Bach’s passions (the St. Matthew Passion in particular), is thus the most obvious element of our reconstruction. It is not entirely clear what took place on the occasion of the premiere on Good Friday 7 April 1724, but probably the two traverso parts were added at the very last minute. It is all too clear that Bach had not planned in any way for these instruments: they have a far less grateful role than they do, for example, in the St. Matthew Passion. In many of the choruses they are confined to doubling the oboes (as they do in the opening chorus) or the violins (in the various turba [crowd] choruses); or the tenor, at the upper octave. A comparison with the St. Matthew Passion, where the flutes were part of the score from the very beginning and play a central role in the orchestration, is particularly illuminating.

We find a very important indication on the first page of the 1739 score (see the illustration above): here Bach apparently unthinkingly copied out the title page of the 1724 version: ‘J.J. Passio secundum Joannem â 4 Voci. 2 Oboe. 2 Violini, Viola è Cont./di J.S. Bach’ [‘With the help of Jesus! Passion after St. John for four voices, 2 oboes, 2 violins, viola, and continuo/by J.S. Bach’]. It is understandable that the instruments which were only used for incidental numbers, such as the viola d’amore or the viola da gamba, are not mentioned here; moreover these were undoubtedly played as ‘extra’ instruments by one of the ensemble’s players. This would not apply for the flutes, which play continuously from the 1725 version onwards and therefore were part of Bach’s new plan. The missing indication for ‘2 Traversi’ on the title page of 1739 can only lead to the conclusion that the ‘original’ version was indeed composed without flutes.

In particular, the two arias which were scored with flutes from 1725 onward are particularly informative: both are written in keys where the player must constantly use complicated fingerings (cross fingerings) which sounds rather unconvincing on the traverso – indeed much more so than, for example, on the oboe. Certainly, during this early period of the instrument’s use, Bach generally wrote for it in easy keys like D major or E minor —a long way from the B-flat major of ‘Ich folge dir gleichfalls’ [no. 9] and even further away from the F minor of ‘Zerfließe, mein Herze’ [no. 35]. It can be seen from the parts that the (re-)instrumentation of these arias was not an unqualified success: the flute part had to be doubled, and even, in the case of ‘Zerfließe, mein Herze’ further supported in one of the performances by a muted violin. Although there is no concrete proof, the nature of these arias leaves no doubt as to their original instrumentation.
It is very improbable that Bach would have planned a major vocal work like the St. John Passion without a single aria with violin solo, and the obbligato par of the aria ‘Ich folge dir gleichfalls’ [no. 9] is far more idiomatic for solo violin. The figuration is strongly violinistic and recalls fast movements in similar metres for solo violin – see Example 1 below. Moreover, it appears that for reasons of the more limited range of the flute, an internal repetition of the instrumental ritornello (measures 40-48) was transposed up an octave.

Example 1
a ‘Ich folge dir gleichfalls’, BWV 245/9, obbligato part, bars 1-16

b Presto from Sonata I for solo violin, BWV 1001/4, bars 25-29

c Gigue from Partita III for solo violin, BWV 1006/7, bars 25-27

The highest instrumental voice of ‘Zerfließe, mein Herze’ [no. 35] is clearly conceived for the oboe. In view of the key, which can be considered far better playable by the oboe, it is also apparent that this part goes no higher than b-flat2 – certainly a rather low upper limit for the traverso. Bach always wrote his traverso parts up to at least d3 or e3. The sweet-toned thirds and sixths between the two obbligato instrumental voices come out far better in the homogeneous scoring for oboe and oboe da caccia.

Example 2
‘Zerfließe, mein Herze’, BWV 245/35, bars

These two reconstructed scorings result in an arrangement of the obbligato instruments for the arias which is logical and far more typical of Bach. In this context we omit ‘Mein teurer Heiland, laß dich fragen’ [no. 32], on the one hand because it is accompanied only by continuo, and on the other because it is, in the first place, a chorale setting and only in the second place a bass aria. The remaining seven 'true' arias are orchestrated as follows in our reconstruction (and always with continuo as well):

no. 7 (‘Von den Stricken’) - 2 oboes
no. 9 (‘Ich folge dir’) - violin
no. 13 (‘Ach, mein Sinn’) - strings [and oboes ad libitum]
no. 20 (‘Erwäge’) - 2 viole d’amore
no. 24 (‘Eilt’) - strings
no. 30 (‘Es ist vollbracht’) - viola da gamba
no. 35 (‘Zerfließe’) - 2 oboes [oboe & oboe da caccia]

A symmetrical principle can be very clearly discerned here: the first and last arias are written for two oboes, the second and the penultimate for a solo string instrument, and the innermost ‘ring’ for the four-part string ensemble. In the center of this structure, then, comes the unique appearance of the two viole d’amore, and it is not by chance that this aria at the same time is by far the longest of them all. This ‘hidden’ symmetry connects with other symmetries which have already been discovered within the St. John Passion, such as that of the opening and closing choruses, the tonalities used for the recitatives, and the ordering of the turba choruses in the second half of the passion.

Our reconstruction’s dominant characteristic is the omission of the flutes, which is further justified by the following considerations:

[1] There are various indications that during Bach’s Leipzig period, the number of performers in his liturgical compositions gradually increased. When Bach arrived in Leipzig in 1723, he initially had only a small number of musicians – to be exact, a small number of instrumentalists—at his disposal. Under his predecessor Kuhnau, the regular church music had gradually been moved to the sidelines, and the musical center of gravity had been transferred to the Neukirche and the Collegium Musicum, founded there in 1701 by Telemann with its student musicians. Slowly but surely, Bach was able to draw increasingly on this resource and make use of it for his church music performances, but not really until 1729 when he himself finally became the leader of the collegium. The St. John Passion, originating in the early Leipzig period, therefore probably received its premiere in a performance with only a few musicians.

[2] The omission of the flutes results in a whole new orchestral balance. In essence, the violin parts no longer need to be taken by more than one player each because the oboes now have a much more soloistic role (the surviving ripieno violin parts from 1724 may have been connected with Bach’s last minute decision to add flutes to the scoring). This is certainly evident in a structure like that of the opening chorus: without the doubling flutes, the oboes’ chains of dissonant suspensions sound thinner, but also more searing. What has become clearer than ever to us in this version is the chamber-music quality of the St. John Passion, particularly in contrast to its ‘big sister’, the St. Matthew Passion. One-on-a-part instrumental scoring is a natural consequence of this.

[3] Bach wrote the St. John Passion for a premiere in the Thomaskirche. At the last minute, four days before Good Friday, the town council decided to move the passion vespers to the other principal Leipzig church, the Nicolaikirche. We can read in the city annals that Bach agreed to this, with the proviso that the ‘Clav-Cymbel’ [harpsichord] located there be put in good repair. We can conclude, from the fact that Bach had already been giving cantata performances in the Nicolakirche during the winter immediately preceding, that the composer had assigned the harpsichord a special role for the performance of the St. John Passion: a role which differed essentially from that which it played in the cantatas. That can only have to do with the central function of the Evangelist’s recitatives, which Bach probably accompanied in person at the harpsichord. Naturally we have imitated that in our performance.


‘Zerfließe, mein Herze’ is a strikingly ‘emotional’ aria in the Italian operatic style (of course, as seen through Bach’s eyes), but there are many other ‘theatrical’, secular elements to be found in the St. John Passion. For example, the mighty opening chorus ‘Herr, unser Herrscher’ begins with a typically Italianate chain of dissonances in the oboes, and Bach uses various fashionable dance forms as a basis for the arias, such as the Passepied for ‘Ich folge dir gleichfalls’ [no. 9] the Sarabande for ‘Ach, mein Sinn’ [no. 13], and the Menuet ‘en rondeau’ for the final chorus [no. 39]. This latter piece, in particular, is interesting; homophonic, ‘galant’ closing choruses otherwise occur only in Bach’s secular cantatas (the St. Matthew Passion, which contains a very similar final chorus, excepted). The aria ‘Es ist vollbracht’ [no. 30], the meditation on Christ’s last words, is particularly theatrical in character: the opening and final sections with a viola da gamba and descending lines inspired by the French tombeau style, and in the center a rebellious stile concitato with the full string ensemble (‘Der Held aus Juda siegt mit Macht’). No (opera) composer could have had more vividly contrasting stylistic means at his disposal.

Whereas most of the arias of the St. John Passion are strikingly succinct, clearly in order not to impede the flow of the passion narrative, Bach emphatically departed from this concept in one place: in the inimitable arioso/aria pair, nos. 19-20. Here Bach made use of two viole d’amore, considered at that time as instruments of a definitely secular character and very rare in Bach’s church music. In the preceding arioso ‘Betrachte meine Seel’ [no. 19] the equally ‘secular’ lute is added to the ensemble. The text of these two pieces—a reaction to the scourging of Christ—is expressly mystic/pietistic with its cult of blood and wounds (in this way his ‘back, all red with blood’ becomes ‘the most beautiful rainbow in the world is revealed as God's own sign of blessing’) and it must have raised many an orthodox Lutheran eyebrow in Leipzig. The long-drawn aria ‘Erwäge, wie sein blutgefärbter Rücken’ [no. 20] is particularly pictorial: words in the text such as ‘consider’ (the opening motive), the ‘waves’ and the ‘rainbow’ are depicted in the exuberant figuration of the two ‘love’ violas. Clearly Bach paid little attention to his promise not to compose ‘operatically’: he saw the Good Friday vespers, with their semi-liturgical passion music, as a perfect opportunity to show off his compositional versatility and his ability as Capellmeister.

© Pieter Dirksen, 2004


Werner Breig, ‘Zu den Turba-Chöre von Bachs Johannes-Passion’, in Hamburger Jahrbuch für Musikwissenschaft 8 (1985), pp. 65-96.

Alfred Dürr, Die Johannes-Passion von J.S. Bach. Entstehung, Überlieferung, Werkeinführung (Kassel, 1988).

Martin Geck, Johann Sebastian Bach. Johannespassion BWV 245 (Meisterwerke der Musik 55; München, 1991).

Dagmar Hoffmann-Axthelm, ‘Bach und die Perfidia iudaica. Zur Symmetrie der Juden-Turbae in der Johannes-Passion’, in Basler Jahrbuch für historische Musikpraxis 13 (1989), pp. 31-54.

Konrad Küster, ‘Text und Musik in der Gesamtform von Bachs Johannes-Passion’, in Beiträge zur Bach-Forschung 9/10 (1991, pp. 20-28.

Hermann Melchert, Das Rezitativ der Bachschen Johannes-Passion (Wilhelmshaven, 1988).

Arthur Mendel, Kritischer Bericht, Neue-Bach-Ausgabe II/4: Johannes-Passion (Kassel, 1974).

Arthur Mendel, ‘Meliora et melioranda in the two versions of BWV 245/1’, in Bach Studien 5 (Leipzig, 1975), pp. 113-121.

Arthur Mendel, ‘Traces of the Pre-History of Bach’s St. John and St. Matthew Passions’, in Festschrift für Otto Erich Deutsch (Kassel, 1963), pp. 31-48.

Ulrich Prinz (ed.), J.S. Bach: Johannes-Passion, BWV 245. Vorträge des Meisterkurses 1986 und der Sommerakademie J.S. Bach 1990 (Schriftenreihe der Internationale Bachakademie Stuttgart 5; Kassel, 1993).

Hans-Joachim Schulze, ‘Bemerkungen zur zeit- und gattungsgeschichtlichen Kontext von Johann Sebastian Bachs Passionen’, in Johann Sebastian Bachs historischer Ort, ed. Reinhard Szeskus (Bach-Studien 10; Wiesbaden-Leipzig, 1991), pp. 202-215.

Peter Wollny, ‘Neue Bach-Funde’, in Bach-Jahrbuch 83 (1997), pp. 7-50.