1. Childhood.

The parents of Johann Sebastian were Johann Ambrosius Bach (11) and Maria Elisabeth Lämmerhirt (1644–94), daughter of a furrier and town councillor in Erfurt, Valentin Lämmerhirt (d 1665). Another Lämmerhirt daughter became the mother of Bach’s cousin Johann Gottfried Walther, suggesting that Lämmerhirt blood was perhaps not unimportant for the musical talents of the Bach family’s greatest son. Elisabeth’s elder half-sister Hedwig Lämmerhirt was the second wife of Ambrosius Bach’s uncle, Johann Bach (4), organist of the Predigerkirche in Erfurt. Ambrosius, who had been Stadtpfeifer in Arnstadt and from 1667 violinist in the Erfurt town music, worked in Eisenach from 1671 as Hausmann (director of town music) and also as a musician at the ducal court of Saxe-Eisenach. Elisabeth and Ambrosius were married in Erfurt on April 8, 1668 and had eight children, five of whom survived infancy: Johann Sebastian, the youngest, as well as three sons (nos. 22, 71, and 23) and a daughter, Maria Salome. The date of Johann Sebastian’s birth, March 21, 1685, was carefully recorded by Walther in his Lexicon, by Sebastian himself in the family genealogy, and by his son as the co-author of the obituary. It is supported by the date of baptism (March 23; these dates are old-style) in the register of St Georg. His godfathers were Johann Georg Koch, a forestry official, and Sebastian Nagel, a Stadtpfeifer from nearby Gotha. The house of his birth no longer stands; it is not the handsome old structure (Frauenplan 21) acquired by the Neue Bachgesellschaft in 1907 as the ‘Bachhaus’ and established as a Bach Museum. He would have been born in the house in the Fleischgasse (now Lutherstrasse) that Ambrosius Bach bought in 1674 after gaining Eisenach citizenship.

After the time of the Reformation all children in Eisenach were obliged to go to school between the ages of five and 12, and (although there is no documentary evidence of it) Sebastian must have entered one of the town’s German schools in 1690. From 1692 he attended the Lateinschule (as had Luther, also an Eisenach boy); this offered a sound humanistic education. At Easter 1693 he was 47th in the fifth class, having been absent 96 half-days; in 1694 he lost 59 half-days, but rose to 14th and was promoted; at Easter 1695 he was 23rd in the fourth class, in spite of having lost 103 half-days (perhaps owing to illness, but probably also to the deaths of his parents). He stood one or two places above his brother Jacob, who was three years older and less frequently absent. Nothing more is known about his Eisenach career; but he is said to have been an unusually good treble and would have sung under Kantor A.C. Dedekind at St Georg, where his father was in charge of instrumental music before and after the sermon and where his relation Johann Christoph Bach (13) was the town organist.

His early musical education is matter for conjecture; presumably his father taught him the rudiments of string playing and notably the violin, Ambrosius’ primary instrument. The Stainer violin listed later among the instruments in Bach’s estate was probably an inheritance from his father. According to Emanuel, he had no formal tuition on keyboard instruments until he went to Ohrdruf. But even though he later described Johann Christoph as ‘a profound composer’, no doubt he was impressed by the latter’s organ playing and most likely received some instruction from him well before he turned ten years of age.

Elisabeth Bach was buried on May 3, 1694, and on November 27, Ambrosius married Barbara Margaretha, née Keul, the daughter of a former mayor of Arnstadt. Aged 35, she had already been twice widowed. Her first husband had been a musician, Johann Günther Bach (15), and her second a pastor, Jacobus Bartholomaei (both marriages had taken place in Arnstadt), and she brought to her third marriage two little daughters, Catharina Margareta and Christina Maria, one by each of her earlier husbands. A month before Ambrosius's own second marriage, on October 23, 1694, he and his family had celebrated the wedding of the eldest son Johann Christoph (22) in Ohrdruf. The music on that occasion was provided by Ambrosius Bach, Johann Christoph Bach (13), and Johann Pachelbel from nearby Gotha, as well as other friends and family members. This was probably the only occasion on which the then nine-year-old Sebastian met Pachelbel, his brother’s teacher. In all likelihood, Ambrosius commissioned for the occasion from his Eisenach cousin the wedding concerto “Meine Freundin, du bist schön” (text and surviving performing parts are in Ambrosius’s hands). But barely three months after re-marrying, on February 20, 1695 and two days before his 50th birthday, Ambrosius Bach died after a long and serious illness. On March 4, the widow appealed to the town council for help; but she received only her legal due, and the household broke up. Sebastian and Jacob were taken in by their elder brother Johann Christoph, organist at Ohrdruf.

Both were sent to the Lyceum. Jacob left at the age of 14 to be apprenticed to his father’s successor at Eisenach, Johann Heinrich Halle; Sebastian stayed on until 1700, when he was nearly 15, and thus came under the influence of an exceptionally enlightened curriculum. Inspired by the educationist Comenius and directed by the reform-minded young rector Johann Christoph Kiesewetter, it embraced religion, reading, writing, arithmetic, music, history, and natural science. Sebastian entered the fourth class probably about March 1695, and was promoted to the third already in July of the same year: on July 20, 1696 he was listed first among the seven new boys and fourth in the class; on July 19, 1697 he was first, and was promoted to the second class; on July 13, 1698 he was fifth; on July 24, 1699 second, and promoted to the first class, in which he was fourth when he left the school on March 15, 1700 (new style) and went to Lüneburg. For most of his time in Ohrdruf he performed as choral scholar under the experienced Kantor Elias Herda.

In the obituary Emanuel stated that under his brother Christoph’s guidance his father ‘laid the foundations for his playing of the clavier’. He also told the story of how Christoph would not allow his brother to use a certain manuscript that contained keyboard works ‘by the most famous composers of the day – Froberger, Kerll, Pachelbel’; how Sebastian copied it by moonlight, demonstrating ‘this very passion to improve himself in music’; how the brother took the copy away from him; and how he did not recover it until Christoph died in 1721. The manuscript has not survived but a good idea of its contents can be obtained from a comparable manuscript collection of works by Froberger, Kerll, Pachelbel, and some other composers compiled in 1692 by another Pachelbel pupil, J.V. Eckelt. Above and beyond this anecdotal information, the rapid progress made under the tutelage of his brother is documented by the so-called Weimar Tablature (discovered in 2005) that contains in the hand of the 13- to 14-year-old student D. Buxtehude’s most ambitious, large-scale chorale fantasia Nun freut euch lieben Christen gmein buxwv214. It provides the first reliable evidence of the unusually high level of proficiency and technical competence of the young organist and his primary choice of instrument.

St Michael’s, the late Gothic hall church where Christoph was organist, was destroyed by fire in 1753 and subsequently rebuilt. It had contained a good-sized organ that was examined by Pachelbel in 1693, but its deficiencies were never completely remedied during Christoph’s tenure. Sebastian no doubt picked up some of his expert knowledge of organ building while helping his brother with repairs. Besides, he probably got to know the famed organ builder Georg Christoph Sterzing who maintained his workshop in Ohrdruf; Bach’s connections to Sterzing are documented into the 1730s.

No documentary evidence exists to establish when Bach started to compose, but it is reasonable to suppose that it was while he lived in Ohrdruf – not least because other contemporaries, and his own sons in due course, began composing original music well before reaching the age of 15. The earliest organ chorales in the Neumeister manuscript provide plausible examples of pieces composed before and around 1700. They are characterized by sound craftsmanship, observance of models provided mainly by Pachelbel (his teacher’s teacher) but also by Buxtehude, and everywhere by the sense of an endeavour to break away from musical conventions and find independent solutions.

2. Lüneburg.

According to the school register, Sebastian left Ohrdruf on March 15, 1700 ‘ob defectum hospitiorum’ (‘for the lack of “hospitia”’); he had lost access to the free table and his brother could not take care of him without this essential support. The problem seems to have been solved by Elias Herda, Kantor at the Lyceum, who had been educated at Lüneburg. No doubt it was he who arranged for Sebastian to go north; probably he similarly helped Georg Erdmann, a fellow pupil of Sebastian’s, three years older, who left the school just before Bach (for the same reason). According to the obituary they travelled together and must have reached Lüneburg before the end of March. Both were entered in the register of the Mettenchor (Matins choir) by April 3, 1700 and probably sang in it within a matter of days for Holy Week and Easter, which fell on April 11.

St Michael’s Church in Lüneburg had two schools associated with it: a Ritteracademie for young noblemen, and the Michaelisschule for commoners. There were also two choirs: the ‘chorus symphoniacus’ of about 25 voices was led by the Mettenchor, which numbered about 15, and was limited to poor boys. Members of the Mettenchor received free schooling at the Michaelisschule, up to 1 thaler per month according to seniority, their keep, and a share in fees for weddings and other occasions (Bach’s share in 1700 has been put at 14 marks). From the arrangement of the pay-sheets it has been deduced that they were both trebles. Bach was welcomed for his unusually fine voice; but it soon broke, and (according to the obituary) he spoke and sang in octaves for eight days. After that he may or may not have sung, but no doubt he made himself useful as a keyboard accompanist or string player. As the last extant pay-sheet is that for May 29, 1700, no details are known; but it is clear that the school was short of instrumentalists at just this time.

At school, Bach’s studies embraced orthodox Lutheranism, logic, rhetoric, Latin and Greek, arithmetic, history, geography, and German poetry. The Kantor was August Braun, whose compositions have disappeared; the organist, F.C. Morhard, was a nonentity. The organ was repaired in 1701 by J.B. Held, who had worked at Hamburg and Lübeck; he lodged in the school and may have taught Bach something about organ building. There was an exceptionally fine music library, which had been carefully kept up to date. If Braun made good use of it, Bach must have learnt a good deal from the music he had to perform; but his chief interests probably lay outside the school. At St Nicholas’s Church was J.J. Löwe (1629–1703), distinguished but elderly. The St John’s Church was another matter, for there the organist was the renowned Georg Böhm (1661–1733), originally from Thuringia and certainly an additional attraction for Bach’s northern journey. He apparently lost no time to connect with Böhm, for the Weimar Tablature contains an autograph copy (dated 1700) of J.A. Reinken’s elaborate and demanding chorale fantasia An Wasserflüssen Babylon. The fact that this piece was copied from a manuscript in Böhm’s possession, as Bach specifically indicated, suggests an early, close, and apparently lasting relationship, a proposition supported by the fact that in 1727 Bach named Böhm as his northern agent for the sale of Partitas nos. 2 and 3.

Bach went more than once to Hamburg, some 50 km away, most likely encouraged by Böhm who, at the time, served as harpsichordist of the Hamburg Opera, one of the directors of which was Johann Adam Reinken. He may well have visited the opera, then directed by Reinhard Keiser, whose St Mark Passion he performed during the early Weimar years and again in 1726. He possibly also met with his cousin Johann Ernst (25), who was about this time studying in Hamburg, but he apparently was primarily interested in the city’s organ scene and in particular the organist of St Katharinen, Reincken, whose influence on the young Bach as both theorist and practitioner is difficult to overestimate. Marpurg’s familiar anecdote makes the point neatly: how Bach, returning almost penniless to Lüneburg, once rested outside an inn; how someone threw two herring heads out on the rubbish heap; how Bach – a Thuringian, to whom fish were a delicacy – picked them up to see if any portion were edible; how he found that they contained two Danish ducats, and was thus able not only to have a meal, but also ‘to undertake another and a more comfortable pilgrimage to Herr Reincken’.

J.A. Reincken (1643–1722), organist of St Katharinen since 1663, was a leading figure of the north German school. Böhm may have advised Bach to hear him; and his showy playing, exploiting all the resources of the organ, must have been a revelation to one brought up in the reticent tradition of the south. As for the organ itself, Bach never forgot it; in later years he described it as excellent in every way, said that the 32′ Principal was the best he had ever heard, and never tired of praising the 16′ reeds. Reincken might have given him a copy of his Hortus Musicus (1688); Bach’s reworkings of three sonatas with fugal movements from this collection (the keyboard pieces bwv954, 965, and 966) are likely to have been made relatively soon after 1700 and provided him with his first encounter with permutation fugues. In 1775, replying to Forkel, C.P.E. Bach reported that Sebastian early on became ‘a pure and strong fuguist’ through his own efforts. Permutation fugues in the manner of Reincken played a major initial role in his practice of fugal composition.

The market-place in Lüneburg had been graced since the end of the 17th century by a palace used for the visits of the Duke of Celle-Lüneburg and his court; the principal ducal residence and seat of government lay in Celle, some 80 km to the south. The duke, married to Eléonore d’Olbreuse, a Huguenot of noble birth, was a pronounced francophile and maintained an orchestra consisting largely of Frenchmen, which played in both Celle and Lüneburg. Thomas de la Selle, dancing-master at the Ritteracademie next door to Bach’s school in Lüneburg, was also a member of the Celle orchestra. Emanuel Bach knew that his father was often able to hear this ‘famous orchestra’ and thus to become acquainted with French taste. It cannot be ruled out that Bach occasionally helped out as an instrumentalist when the court orchestra played in the ducal residence in Lüneburg.

The date of Bach’s departure from Lüneburg is not known, but we may suppose that he completed his final school year after two years and graduated at Easter 1702. It seems unlikely that he remained in Lüneburg for any length of time after that, for he left without hearing Buxtehude and took extraordinary pains to do so in winter 1705–6. He probably visited relatives in Thuringia after Easter 1702. All that is definitely known is that in the fall of 1702 he competed successfully for the vacant post of organist at St Jacobi in Sangerhausen (the organist was buried on July 9), but the Duke of Weissenfels intervened and had Johann Augustin Kobelius, a somewhat older man, appointed in November of 1702. Bach is next heard of at Weimar, where he was employed at the court as a musician for the first two quarters of 1703, most likely at the instigation of, and as assistant to, the ailing court organist Johann Effler, a close friend of the Bach family. The court accounts have him down as a lackey, but he described himself as a ‘Hofmusikant’ (court musician) in the Ursprung. Bach was employed at the minor Weimar court, that of Duke Johann Ernst, younger brother of the Duke Wilhelm Ernst whom Bach served from 1708 to 1717.

Of the musicians with whom Bach now became associated, three are worth mentioning. Georg Christoph Strattner (c1644–1704), a tenor, became vice-Kapellmeister in 1695, and composed in a post-Schütz style. Johann Paul von Westhoff (1656–1705) was a fine violinist and had travelled widely, apparently as a diplomat, and is said to have been the first to compose suites for unaccompanied violin (1683). Johann Effler (1634–1711) was the court organist: he had held posts at Gehren as successor to Johann Michael Bach (14) and Erfurt (where Pachelbel was his successor) before coming in 1678 to Weimar, where about 1690 he moved to the court. A sickly man by the early 1700s, he may have needed to hand over some of his duties to Bach, and probably did something of the kind, for a document of July 13, 1703 at Arnstadt, where Bach next moved, describes Bach as court organist at Weimar – a post that was not officially his until 1708.

3. Arnstadt.

The Bonifaciuskirche at Arnstadt had burnt down in 1581 and was subsequently rebuilt in 1676–83; it then became known as the Neue Kirche, and so remained until 1935, when it was renamed after Bach. In 1699 J.F. Wender contracted to build an organ, which by the end of 1701 had become usable; on January 1, 1702 Andreas Börner was formally appointed organist. The organ was complete by June 1703, and was examined before July 3; there were more examiners than one, but only Bach was named and paid, and it was he who ‘played the organ for the first time’. The result was that on August 9 the Arnstadt burgomeister Martin Feldhaus, who – as brother-in-law of Johann Michael Bach (14) – had personally supervised the work of the organ builder, offered Bach the post over Börner’s head; at the same time, ‘to prevent any such “collisions” as are to be feared’, Börner was given other work. Bach accepted the post ‘by handshake’ on August 14, 1703. The exact date of his move to Arnstadt is not known, nor is his address. As his last board and lodging allowance was paid to Feldhaus, he probably spent at least that year in either the Golden Crown or the Steinhaus, both of which belonged to Feldhaus. Considering his young age, and local standards, he was well paid; and his duties, as specified in his contract, were light. Normally, he was required at the church only for two hours on Sunday morning, for a service on Monday, and for two hours on Thursday morning; and he had only to accompany hymns. He thus had plenty of time for composition and organ playing, and he took as his models Bruhns, Reincken, Buxtehude (all northerners), and certain good French organists. There is also circumstantial evidence (see below, Geyersbach affair) that he took part in musical activities of the small Arnstadt court capelle of the counts of Schwarzburg.

Bach was in no position to put on elaborate music at Arnstadt. The Neue Kirche, like the other churches, drew performers from two groups of schoolboys and senior students. Only one of these groups was capable of singing cantatas; it was supposed to go to the Neue Kirche monthly in the summer, but there does not appear to have been a duty roster. The performers naturally tended to go to the churches that had an established tradition and friendly organists; and Bach had no authority to prevent this, for he was not a schoolmaster and was younger than many of the students. Further, he never had much patience with the semi-competent, and was apt to alienate them by making offensive remarks. One result was his scuffle with the student J.H. Geyersbach (b 1682). On August 4, 1705, on his way home from the castle, he and his cousin Barbara, elder sister (aged 26) to his future wife, fell in with six students who had been to a christening feast; one of these was Geyersbach, who asked why Bach had insulted him (or his bassoon), and struck him in the face with a stick. Bach apparently was wearing his court uniform and drew his sword, but another student separated them. Bach complained to the consistory that it would be unsafe for him to go about the streets if Geyersbach were not punished, and an inquiry was held. The consistory told Bach that he ought not to have insulted Geyersbach and should try to live peaceably with the students; further, he was not (as he claimed) responsible only for the chorales but was expected to help with all kinds of music. Bach replied that if a music director were appointed, he would be willing enough.

Bach, unimpressed, asked for four weeks’ leave, and set off for Lübeck – ‘what is more, on foot’, says the obituary, adding that he had an overwhelming desire to hear Buxtehude. Dates and distance cast some doubts on his straightforwardness. He left Arnstadt about October 18, and was therefore due to be back, or well on his way back, by about November 15; he would thus have been unable to hear even the first of Buxtehude’s special services, which were given on various dates from November 15 to December 20 and in 1705 included the two oratorios Castrum doloris and Templus honoris buxwv 134–5. It is conceivable that Bach was not merely a member of the audience but likely a participant in the large ensemble Buxtehude assembled for the two major works; after all, he had to finance his extended trip and pay his Arnstadt substitute. Perhaps, like Mattheson and Handel before him, he also went to see if there was any chance of succeeding Buxtehude, and was put off by the prospect of marrying Buxtehude’s daughter, aged 30; in any case, by 1705 there was a rival in the field. However that may be, he stayed almost three months at Lübeck, and was absent altogether for about 16 weeks, not returning to Arnstadt until shortly before February 7, 1706, when he communicated.

On February 21 the consistory asked Bach why he had been away for so long; his replies were deemed unsatisfactory and barely civil. They next complained that his accompaniments to chorales were too elaborate for congregational singing, and that he still refused to collaborate with the students in producing cantatas; further, they could not provide a Kapellmeister for him, and if he continued to refuse they would have to find someone more amenable. Bach repeated his demand for a music director, and was ordered to apologize within eight days. From the next case that the consistory heard that day it seems that there had been actual ‘disordres’ in the church between Bach and the students. There is no evidence that Bach apologized, and the consistory dropped the matter for eight months. They brought it up again on November 11, 1706, and Bach undertook to answer them in writing. They also accused him of inviting a ‘stranger maiden’ to make music in the church, but for this he had obtained the parson’s permission. The woman singer in question cannot have been his cousin and future wife, for she had long been resident in Arnstadt and therefore would unlikely have been described as a stranger.

Neither Bach nor the consistory took further action; no doubt they saw that the problem would soon solve itself. Probably Bach had come back from Lübeck with exalted ideas about church music, requiring facilities that Arnstadt could not provide. His ability was becoming known; on November 28 he helped to examine an organ at Langewiesen. Forkel said that various posts were offered to him; and with the death of Johann Georg Ahle, on December 2, a sufficiently attractive vacancy seemed to have arisen.

4. Mühlhausen.

Ahle had been a city councillor of the Free Imperial City of Mühlhausen, organist of St Blasius, and a composer of lesser rank than his father Johann Rudolf Ahle, also of Mühlhausen. Musical standards had fallen during his tenure of office, but the post was a respectable one and various candidates gave trial performances. One was to have been Johann Gottfried Walther, the future lexicographer; he sent in two compositions for February 27, 1707 (Sexagesima) but withdrew after being told privately that he had no hope. Bach’s candidacy was promoted in particular by the organ builder Wender of Mühlhausen. He played at Easter (April 24) and probably performed Cantata bwv4. At the city council meeting on May 24 no other name was considered, and on June 14 Bach was interviewed. He asked for the same salary that he was receiving at Arnstadt (some 20 gulden more than Ahle’s); the councillors agreed, and a contract was signed on June 15. At Arnstadt his success became known; his cousin Johann Ernst (25) and his predecessor Börner applied for the Neue Kirche on June 22 and 23. He resigned formally on June 29, and presumably moved to Mühlhausen within a few days. Bach’s responsibilities in Mühlhausen also included the Brückenhof Church of the convent of Augustinian nuns where there was an organ by Wender without pedals; his principal duty there was to play for weekly and special services, alternating with the organist of the Marienkirche.

On August 10, 1707 Tobias Lämmerhirt, Bach’s maternal uncle, died at Erfurt. He left Bach 50 gulden, more than half his salary, and thus facilitated his marriage to Maria Barbara (b Oct 20, 1684), daughter of Johann Michael Bach (14) and Catharina Wedemann. The wedding took place on October 17 at Dornheim, a village near Arnstadt; the pastor, J.L. Stauber (1660–1723), was a friend of the family and himself married Regina Wedemann on June 5, 1708. Pupils began to come to Bach at about this time, or perhaps even earlier. Johann Martin Schubart (1690–1721) is said to have been with him from 1707 to 1717, and Johann Caspar Vogler (1696–1763) to have arrived at the age of ten (at Arnstadt), to have left for a time, and to have returned from about 1710 until 1715. These two became his immediate successors at Weimar; from their time onward he was never without pupils.

Sometime in 1707, he was commissioned to compose Cantata bwv150 in honor of the Mühlhausen town councillor Dr. Conrad Meckbach, possibly on the occasion of the latter’s 70th birthday on April 19. It was probably in the fall of the same year that Bach wrote Cantata bwv131; this was clearly intended for a penitential service, perhaps connected with a disastrous town fire of May 30. It was not Bach’s own Pastor Frohne who commissioned this cantata, but Pastor Eilmar of the Marienkirche – a fact whose possible significance will be seen later. On February 4, 1708 the annual change of council took place, and Cantata bwv71 was performed at the Marienkirche (with a repeat-performance at St Blasius one week later). It must have made an impression, for the council printed not only the libretto, as was usual, but also the music. Bach next drew up a plan for repairing and enlarging the St Blasius organ; the council considered this on February 21 and decided to act on it. Cantata bwv196 may have been written for Stauber’s wedding on June 5, if not his own in the previous year. At about this time Bach played before the reigning Duke of Weimar, Wilhelm Ernst, who offered him Effler’s post at his court. On June 25 Bach wrote to the council asking them to accept his resignation.

No doubt the larger salary at Weimar was an attraction, particularly as Bach’s wife was pregnant. But it is clear, even from his tactful letter to the Mühlhausen councillors who had treated him well, that there were other reasons for leaving. He said that he had encouraged ‘well-regulated church music’ not only in his own church, but also in the surrounding villages, where the harmony was often ‘better than that cultivated here’. He had also gone to some expense to collect ‘the choicest sacred music’. But in all this, members of his own congregation had opposed him, and were not likely to stop. Some people no doubt disliked the type of music that he was trying to introduce. Further, Pastor Frohne may have distrusted his organist; an active Pietist, he was at daggers drawn with the orthodox Pastor Eilmar of the Marienkirche – Bach had begun his Mühlhausen career by working with Eilmar, the librettist of bwv131; they had become intimate enough for Eilmar and his daughter to be godparents to Bach’s first two children.

The council considered Bach’s letter on June 26 and reluctantly let him go, asking him only to supervise the organ building at St Blasius. However badly Bach may have got on with his congregation, he was evidently on good terms with the council. They paid him to return and perform a cantata at the council service in 1709 as well as in 1710 (all trace of these works is lost). In 1735 he negotiated on friendly terms with the new council on behalf of his son Johann Gottfried Bernhard (47). He is not known to have been paid for supervising or opening the St Blasius organ, but he may have done so.

5. Weimar.

When he announced his resignation from Mühlhausen after a tenure of only 11 months, Bach said that he had been appointed to the Duke of Weimar’s ‘Capelle und Kammermusik’. However, Weimar documents show that on July 14, 1708, when his ‘reception money’ was paid over, he was called ‘the newly appointed court organist’, and that he was almost always so called until March 1714, when he became Konzertmeister as well. Effler, who most likely had engineered his succession, was pensioned off on full salary (130 florins); on December 24, 1709 he received a small gift as ‘an old sick servant’, and he died at Jena on April 4, 1711.

Bach reportedly wrote most of his organ works at Weimar, and the duke took pleasure in his playing. His salary was from the outset larger than Effler’s (150 florins, plus some allowances); it was increased to 200 from Michaelmas 1711, 215 from June 1713, and 250 on his promotion in 1714. On March 20, 1715 it was ordered that his share of casual fees was to be the same as the Kapellmeister’s. Moreover, he seems to have had a fair amount of spare time, in which, for instance, to cultivate the acquaintance of Telemann while the latter was at the neighbouring court of Eisenach (1708–12). Together with the violinist Pisendel he copied a concerto in G of Telemann’s (twv52: G2), probably during Pisendel’s visit to Weimar in 1709.

Six of Bach’s children were born at Weimar: Catharina (bap. Dec 29, 1708; d Jan 14, 1774); Wilhelm Friedemann (45) (b Nov 22, 1710); twins (b Feb 23, 1713; both died in a few days); Carl Philipp Emanuel (46) (b March 8, 1714); and Johann Gottfried Bernhard (47) (b May 11, 1715). The various godparents show that Bach and his wife kept in touch with relations and friends from Ohrdruf, Arnstadt, and Mühlhausen, besides making fresh contacts at Weimar; it is noteworthy that Telemann was godfather to Emanuel and apparently was present at his baptism in March of 1714.

On March 13, 1709 Bach, his wife, and her older sister Friedelena (who died at Leipzig in 1729) were living at the house of Adam Immanuel Weldig, a falsettist and Master of the Pages. They probably stayed there until August 1713, when Weldig gave up his house, having secured a similar post at Weissenfels. Like Telemann, Weldig was godfather to Emanuel; Bach (by proxy) to a son of Weldig’s in 1714. Weldig’s house was destroyed in 1944; where Bach lived before and after the given dates is not known.

Since July 29, 1707, Johann Gottfried Walther (the lexicographer) had been organist of the Stadtkirche; he was related to Bach through his mother, a Lämmerhirt, and the two became friendly. On September 27, 1712 Bach stood godfather to Walther’s son. Forkel told a story of how Walther played a trick on Bach, to cure him of boasting that there was nothing he could not read at sight. Their relations remained close; Bach later served as the Leipzig sales agent for the Musicalisches Lexicon (1732), in 1735 he negotiated on Walther’s behalf with the Leipzig publisher J.G. Krügner, and Walther’s references to Bach in his letters to Bokemeyer carry further evidence of their friendship. From one such letter it seems that during his nine years at Weimar Bach gave Walther some 200 pieces of music, some by Buxtehude, others compositions of his own.

Of Bach’s pupils, Schubart and Vogler have already been mentioned. Other private students include the page Adam Friedrich Wilhelm von Jagemann who took keyboard lessons in 1711/12. Johann Gotthilf Ziegler (1688–1747) matriculated at the University of Halle on October 12, 1712, but before that he had studied with Bach for a year or so, and had been taught to play chorales ‘not just superficially, but according to the sense of the words’; Bach’s wife stood godmother to his daughter in 1718, and in 1727 Bach employed him as agent, in Halle, for Partitas nos. 2 and 3. P.D. Kräuter of Augsburg (1690–1741) set out for Weimar in March 1712, and stayed until about September 1713. Johann Lorenz Bach (38) probably arrived in autumn 1713; he may have left Weimar by July 1717. Johann Tobias Krebs (1690–1762) studied with Walther from 1710, with Bach from about 1714 until 1717. Johann Bernhard Bach (41) worked with his uncle from about 1715 until at least March 1719, alongside Samuel Gmelin (1695–1752), who appears to have left in 1717. Cornelius Heinrich Dretzel of Nuremberg (1697–1775) may have been briefly with Bach as well.

The specification of the organ in the castle chapel, published in 1737, does not represent the organ that Bach left in 1717. Extensive alterations were made in 1719–30. Still less does the specification represent the organ that Bach was faced with in 1708, for he himself made even more extensive alterations in 1713–14. The organ is said to have been built by Compenius in 1657–8. It was overhauled in 1707–8, and a Sub-Bass added, by J.C. Weishaupt, who carried out further maintenance work in 1712. A contract for alterations had however been signed on June 29, 1712 with Heinrich Nikolaus Trebs (1678–1748), a Wender pupil, who had moved from Mühlhausen to Weimar in 1709. Bach and he had worked together on a new organ at Taubach in 1709–10, opened by Bach on October 26, 1710; in 1711 he gave Trebs a handsome testimonial, and in 1713 he and Walther became godfathers to Trebs’s son. Bach and Trebs collaborated again about 1742, over an organ at Bad Berka. Trebs’s new organ was usable during 1714; he had done 14 days’ tuning by May 19 and was paid off on September 15. Of this rebuild nothing is known, except that either Bach or the duke was determined that the instrument include a Glockenspiel; great trouble was taken over obtaining bells from dealers in Nuremberg and Leipzig, and it seems that the original set of 29 (a number hard to account for) had to be replaced because of difficulties over blend and pitch. In 1737 the organ had a Glockenspiel on the Oberwerk.

In December 1709 and February 1710 Bach was paid for repairing harpsichords in the household of the junior duke, Ernst August and Prince Johann Ernst. On January 17, 1711 he was godfather to a daughter of J.C. Becker, a local burgher. From February 21, 1713 Bach was lodged in the castle at Weissenfels. Duke Christian’s birthday fell on February 23, and Cantata bwv208 was written for this occasion, and re-performed later at the Weimar court (see below). In October 1713 Bach composed the strophic aria bwv1127 on a German poem based on Duke Wilhelm Ernst’s motto ‘Omnia cum Deo and nil sine eo’ and written for his 51st birthday. About May 1713 the young prince Johann Ernst, an active composer who had studied with both Walther and Bach, returned from a grand tour, apparently with a good deal of music, for in the year from June 1 there were bills for binding, copying, and shelving (some of the music came from Halle). In February 1713 he had been in Amsterdam and may have met the blind organist J.J. de Graff who was in the habit of playing recent Italian concertos as keyboard solos. This may have given rise to the numerous concerto arrangements made by Walther and Bach, but they both may have been exposed to Italian (and French) music prior to 1713 via the Halle and Leipzig bookseller A. C. Sellius, an agent for E. Roger in Amsterdam.

On September 7, 1713 Bach was probably at Ohrdruf, standing godfather to a nephew; and on November 6 he took part in the dedication of the new Jakobskirche at Weimar (there is no evidence that he composed any of the music). On November 27 he was at Weimar, as godfather to Trebs’s son. At about this time he seems to have gone to Halle, perhaps to buy music, and to have become accidentally involved with the authorities of the Liebfrauenkirche. The organist there (Zachow, Handel’s teacher) had died in 1712, and the organ was being enlarged to a three-manual of 65 stops. The story has to be pieced together from hints in an incomplete correspondence; but it looks as if the pastor, J.M. Heineccius, pressed Bach to apply for the vacant post. Bach may have been involved in planning the enlargement of the organ, when Zachow became incapacitated; at all events, he stayed in Halle from November 28 to December 15 at the church authorities’ expense. He also composed and performed a cantata (probably bwv21), attended a meeting on December 13, 1713, was offered the post, and let the committee suppose that he had accepted it, although he had not had time to find out what his casual fees would amount to. On December 14 they sent him a formal contract. Bach replied on January 14, 1714, saying cautiously that he had not been released from Weimar, was uneasy about his salary and duties, and would write again within the week. Whether he did so is not known; but in February the committee resolved to tell him that his salary was not likely to be increased. Thus, at Halle he could expect a slightly smaller salary than he was already getting; the attraction was the organ, more than twice as large. Bach must then have approached the duke, for on March 2, ‘at his most humble request’, he became Konzertmeister (ranking after the vice-Kapellmeister), with a basic salary of 250 florins from February 25. In finally refusing the Halle post, he probably mentioned that figure, for the committee accused him of having used their offer as a lever to extract more money from the duke. This he denied on March 19, in a letter so reasonable and so obviously honest that he remained on good terms with Halle and was employed there as an organ examiner in 1716. Gottfried Kirchhoff had meanwhile been appointed organist on July 30, 1714.

Few cantatas (apart from the secular bwv208) can be ascribed to the early Weimar years. bwv18, 21, 54, and 199 appear to date from about 1713 and clearly have no specific connection with the cantatas composed with an eye to the church calendar from March 1714 onwards. On March 23, 1714 it was ordered, at the request of the new Konzertmeister, that cantatas should in future be rehearsed in the court chapel, not at the Kapellmeister’s home or in other lodgings; and on Palm Sunday, March 25, Bach performed bwv182. This was the fourth Sunday after his appointment as Konzertmeister, when he had become responsible for writing a cantata every four weeks. As he evidently hoped to complete an annual cycle in four years, he did not keep strictly to this rule; having written a cantata for Advent Sunday in 1714 (bwv61), he wrote for the 23rd Sunday after Trinity in 1715 (bwv163) and the second Sunday in Advent in 1716 (bwv70.1); in 1717 he no longer wrote canatas for the court chapel. Apart from such intentional irregularities, there are gaps in the series, and the strange thing is that these gaps became more numerous after the end of 1715. One of the gaps is accounted for by the death at Frankfurt on August 1, 1715 of the musically gifted Prince Johann Ernst, plunging the duchy into mourning from August 11 to November 9, 1715, when not a note of music might be played. From 1717 there are no cantatas at all. A tentative explanation will be suggested for this; but it is hard to see why Bach’s usual allowance of paper was paid for on May 16, 1716 when he is known to have performed church cantatas not monthly, but only sporadically, after January 19. Most likely, however, he wrote the funeral music bwv1142 for Prince Johann Ernst on April 2, 1716. On the high feast days, the ducal family attended the main service at Sts Peter and Paul Church in the city of Weimar. There was more space for a larger performing ensemble such as required for bwv172 (Pentecost 1714), bwv63 (Christmas 1714), and bwv31 (Easter 1715).

On April 4, 1716 Bach, like the librettist Salomo Franck and ‘the book-printer’, was paid for ‘Carmina’, bound in green taffeta, that had been ‘presented’ on some unspecified occasion – perhaps on January 24 when Duke Ernst August had married Eleonore, sister of the Prince of Cöthen. Ernst’s birthday was celebrated in April; two horn players from Weissenfels came to Weimar, probably brought over for a repeat performance of bwv208. Meanwhile, the new organ at Halle had been making progress, and on April 17 the council resolved that Bach, Kuhnau of Leipzig, and Rolle of Quedlinburg should be invited to examine it on April 29. They all accepted; each was to receive 16 thaler, plus food and travelling expenses. The examination began at 7 a.m. and lasted three days – until sometime on May 1, when the experts wrote their report, a sermon was preached, and fine music performed; there followed a tremendous banquet given by the town council. On May 2 the organist and the three examiners met the builder to discuss details.

On July 31, 1716 Bach and an Arnstadt organ builder signed a testimonial for Johann Georg Schröter, who had built an organ at Erfurt. In 1717 Bach was mentioned in print for the first time: in the preface to Mattheson’s Das beschützte Orchestre, dated February 21, Mattheson referred to Bach as ‘the famous Weimar organist’ saying that his works, both for the church and for keyboard, led one to rate him highly, and asked for biographical information that Bach never provided.

It is against this background that Bach’s departure from Weimar has to be considered. In 1703 he had been employed by Duke Johann Ernst; since his return in 1708, by Duke Wilhelm Ernst, Johann’s elder brother. The brothers had been on bad terms, and when Johann Ernst died in 1707 and his son Ernst August came of age in 1709, things became no better. For some time, the ducal disagreements do not seem to have affected Bach; perhaps they were kept within bounds by Superintendent Lairitz, and Ernst August’s younger half-brother (Johann Ernst, the composer) may have had some influence. But the latter died in 1715, Lairitz on April 4, 1716, and the new superintendent certainly failed to cope with the ‘court difficulties’; like the rest of Wilhelm’s household, he was forbidden to associate with August. The musicians, though paid by both households, were threatened with fines of 10 thaler if they served August in any way.

It seems unlikely that the cantata gaps in 1716 and the total lack of cantatas for 1717 were due to casual losses. It is tempting to suppose, however, that Bach found his position increasingly embarrassing (owing to his early connection with the junior court) and expressed disapproval of Duke Wilhelm’s behaviour by evading his own responsibilities. In fact, Bach does not seem to have disapproved of the duke’s behaviour until he discovered that a new Kapellmeister was being sought elsewhere. Drese senior died on December 1, 1716; his son, the vice-Kapellmeister, was by all accounts a nonentity. Bach produced bwv70.1 for December 6 and completed only the first movement of bwv186 for December 20, but there were no more cantatas to follow, as far as is known. By Christmas 1716, Bach may have found out that the duke was angling for Telemann. Negotiations with Telemann came to nothing; but apparently Bach now set about looking for a post as Kapellmeister. He was offered one by Prince Leopold of Cöthen, brother-in-law to Duke Ernst August (Bach and the prince had probably met at August’s wedding in January 1716) and the appointment was confirmed on August 5, 1717. No doubt Bach then asked Duke Wilhelm’s permission to leave, and no doubt he was refused – the duke being annoyed because his nephew had obviously had a hand in finding Bach a job that carried more prestige and, at 400 thaler, was better paid.

The duke and Bach must nevertheless have remained on reasonable terms for the time being, for at some date hardly earlier than the end of September Bach was in Dresden and free to challenge the resident French keyboard virtuoso Louis Marchand. Versions of this affair differ, but according to Birnbaum (who wrote in 1739 under Bach’s eye), Bach ‘found himself’ at Dresden, and was not sent for by ‘special coach’. Once there, some court official persuaded him to challenge Marchand to a contest at the harpsichord; the idea that they were to compete at the organ seems to have crept in later. Whatever may be the truth about these and other details, it is universally agreed that Marchand ran away.

On his birthday, October 30, 1717, Duke Wilhelm set up an endowment for his court musicians; and the second centenary of the Reformation was celebrated from October 31 to November 2. Presumably Bach took part in these ceremonies, though there is no evidence that he set any of the librettos that Franck had provided. Emboldened, perhaps, by the Marchand affair, he then demanded his release in such terms that the duke had him imprisoned from November 6 until his dismissal in disgrace on December 2. The Cöthen court had paid Bach an advance of 50 thaler on August 7 and ‘the newly arrived Kapellmeister’ received his first salary on December 29, suggesting that Bach and his family left Weimar for Cöthen only after Christmas and just in time for the traditional New Year’s festivities at the Cöthen court. His duties in Cöthen began on January 1, 1718 with the traditional congratulatory music for the princely family. However, he was in Leipzig on December 16–18 examining the new organ at the university church (Paulinerkirche). The three-manual instrument with 48 stops was the largest in Saxony at the time and was built by Johann Scheibe, with whom Bach later collaborated and whose son he was later in dispute with. He may have returned from Leipzig to Weimar for Christmas, performing his final service for the ducal court, but his post of Konzertmeister disappeared after he left. The younger Drese had become Kapellmeister in his father’s place and Bach’s pupil J.M. Schubart took the post of court organist.

6. Cöthen.

Except during the few last months of his Weimar period, Bach had been on good terms with Duke Wilhelm; but his relations with that martinet must always have been official. At the princely court of Anhalt-Cöthen, until the end of 1721, things were different; Prince Leopold was a young man who, as Bach himself said, loved and understood music. He was born in 1694, of a Calvinist father and a Lutheran mother. The father died in 1704, the mother ruled until Leopold came of age on December 10, 1715. There was no court orchestra until October 1707, when Leopold persuaded his mother to take on three musicians. While studying in Berlin in 1708, he met the bass singer and composer Augustin Reinhard Stricker (1675–1720); from the end of 1710 to 1713 he was on the usual grand tour, during which he studied with Johann David Heinichen at Rome. He returned capable of singing bass, and of playing the violin, viola da gamba, and harpsichord. The Berlin court orchestra had broken up in 1713, and from July 1714 he employed Stricker as Kapellmeister and his wife as soprano and lutenist; by 1716 he had 18 musicians. However, by August 1717 Stricker and his wife seem to have resigned, leaving the prince free to appoint Bach.

The late-medieval St Jacob Cathedral in Cöthen was Calvinist and had an organist, but elaborate music was rarely performed there. The two-manual organ built in 1674–6 by Z. and A. Thayssner had 25 stops. The court chapel in the castle had a small organ of unknown specifications. The newly built Lutheran St Agnus Church (completed in 1699) had a two-manual organ of 27 stops by J.H. Müller (1707–8). Bach probably used the St Agnus and St Jacob organs for teaching and private practice. He communicated at St Agnus and took part in the baptisms at the court chapel but had no official duties in either. He may, however, have been involved in the affair of May 1719, when a cantata was put on for the dedication festival of St Agnus, and 150 copies of (presumably) the libretto were printed. The printer’s bill for one thaler and eight groschen was endorsed by the pastor: ‘The churchwardens can give him 16 groschen; if he wants more, he must go to those who gave the order’.

Bach’s basic salary, 400 thaler, was twice Stricker’s, and extra allowances made it up to about 450. Only one court official was paid more, and there is other evidence that Bach was held in high esteem. On November 17, 1718 the last of his children by his first wife (a short-lived son) was named after the prince, who himself was a godfather. Bach’s residence in Cöthen is not definitely known, but it seems likely that he began as a tenant in Stiftstrasse 11; in 1721, when that house was bought by the prince’s mother for the use of the Lutheran pastor, he moved to Holzmarkt 10. The orchestra needed a room for their weekly rehearsals; the prince supplied it by paying rent to Bach (12 thaler a year) as well as a regular sum for the maintenance of a harpsichord.

Bach’s official functions at the court were comparatively light since they involved no weekly church music. The regular major duties were congratulatory cantata or serenata performances for New Year’s Day and the prince’s birthday on December 10, and occasionally for other special occasions in honor of the princely family. His double title of “Kapellmeister and Director of Chamber Music” indicates his responsibility for chamber music performances throughout the year. Some extant printed texts help to identify the occasion for the works, but the music has survived for only a few of the relevant dates: bwv66.1 (December 10, 1718), bwv134.1 (January 1, 1719), and bwv173.1 (December 10, 1722). bwv184.1 was performed on New Year’s Day in either 1718 or in 1721–3, and bwv194.1 on an unknown occasion sometime between 1718 and 1723. Only texts have survived for bwv1150–53. In 1721 there may have been no birthday celebrations, for the prince was married, at Bernburg, the next day. There is, however, no evidence for the performance of a wedding piece. After Bach had moved to Leipzig, bwv36.2 was performed on November 30, 1726 for the birthday of the prince’s second wife.

On May 9, 1718 the prince went to drink the waters at Carlsbad for about five weeks, taking with him Bach, five other musicians, and a harpsichord. The upper nobility from all over Europe used to spend their summer vacation at this famous spa in Bohemia, but it is not known whom the Cöthen musicians entertained. Early in 1719 Bach was in Berlin, negotiating for a new harpsichord. About this time he seems to have been busy composing or buying music, for between July 1719 and May 1720 some 26 thaler were spent on binding. During 1719 Handel visited his mother at Halle, only some 30 km away; Bach tried, but failed, to make contact with him. He also disregarded a renewed request from Mattheson for biographical material.

In January 1720 Bach stood godfather to the daughter of the Cöthen viola da gamba player C.F. Abel. W.F. Bach was nine in December 1719; the title page of his Clavier-Büchlein is dated January 22, 1720; it contained among other works early versions of 11 preludes from Das wohltemperiete Klavier and the composing scores of the two-part preambles (inventions) and three-part fantasias (sinfonias). In May Bach again went to Carlsbad with the prince for another extended stay. He used these leisurely weeks for making a fair copy of his sonatas and partitas for solo violin, by far the most calligraphic of all his surviving musical autographs. The date of his return to Cöthen does not seem to have been recorded; but apparently it was after July 7, for that was the date of Maria Barbara’s funeral, and there is no reason to doubt Emanuel’s story that his father returned to find her dead and already buried. The death of his nearly 36-year old wife may well have unsettled Bach, and even led him to think of returning to the service of the church. But there was a more practical reason for his taking an interest in St Jacobi at Hamburg. The organist there, Heinrich Friese, died on September 12, 1720; Bach had known Hamburg in his youth, and must have been attracted by the organ, a four-manual Schnitger with 60 stops. There is no evidence that he was actually invited to apply for the post; but he may well have made inquiries of his own.

At all events, his name was one of eight being considered on November 21, and he was in Hamburg at that time but had to leave on November 23. A competition was arranged for November 28, five days after Bach had had to leave for Cöthen; three candidates did not appear, and the judges were not satisfied with the other four. An approach was made to Bach, and the committee met on December 12; as Bach’s reply had not arrived, they met again a week later, when they found that Bach had refused. Perhaps he was unable, or unwilling, to contribute 4000 marks to the church funds, as the successful candidate actually did.

From the way in which the committee kept the post open for Bach, one may suppose that they had heard his recital at St Katharinen. Exactly how this performance was arranged, no-one knows; but in the obituary Emanuel stated that Bach played before Jan Adam Reincken, the magistracy, and other notables; that he played for more than two hours in all; and that he extemporized in different styles on the chorale An Wasserflüssen Babylon for almost half an hour, just as the better Hamburg organists had been accustomed to doing at Saturday Vespers. As a fantasia on this chorale was one of Reincken’s major works, this may seem a tactless choice; but the obituary makes it clear that the chorale was chosen by ‘those present’ and not by Bach himself. Reincken is reported to have said, ‘I thought this art was dead, but I see it still lives in you’, and showed Bach much courtesy. A later remark of Mattheson’s has been taken to imply that Bach also played the G minor Fugue bwv542. He also appears to have presented his Cantata bwv21, perhaps in a second concert.

In early 1721 Bach must have been preparing the Brandenburg Concertos, whose autograph full score was dedicated on March 24, 1721 to the Margrave Christian Ludwig of Brandenburg-Schwedt, brother of the first Prussian king, before whom Bach had played in Berlin while acquiring the new Cöthen harpsichord in February 1719. What he played is not known; but he was invited to send in some compositions. The margrave’s reaction to the receipt of the six works assembled for him by Bach is also not known.

One of Bach’s friends at Cöthen was the goldsmith C.H. Bähr; Bach stood godfather to one of Bähr’s sons in 1721, and deputized for a godfather to another in 1723. About the beginning of August 1721 he gave a performance of some unspecified kind for Count Heinrich XI Reuss of Schleiz; this may have been arranged by J.S. Koch, the Kantor there, who had held a post at Mühlhausen, though possibly not in Bach’s time there. On June 15, 1721 Bach was the 65th communicant at St Agnus; one ‘Mar. Magd. Wilken’ was the 14th. This may well have been Bach’s future wife – the mistake in the first name is an easy one – but Anna Magdalena makes no formal appearance until September 25, when Bach and she were the first two among the five godparents of the son of the princely wine stewart Christian Hahn. This baptism is recorded in three registers. In two of them Anna is described as ‘princely singer’, in the third, simply as ‘musician’ (Musicantin). In September Anna was again a godmother, to a child called Palmarius; again the registers differ in describing her occupation. Her name does not appear in court accounts until summer 1722, when she is referred to as the Kapellmeister’s wife; her salary (half Bach’s) is noted as paid for May and June 1722.

Practically nothing is known of her early years. She was born on September 22, 1701 at Zeitz. Her father, Johann Caspar Wilcke, was a court trumpeter; he worked at Zeitz until about February 1718, when he moved to Weissenfels where he died on November 30, 1731. The surname was variously spelt. Anna’s mother (Margaretha Elisabeth Liebe, d March 7, 1746) was daughter of an organist and sister of J.S. Liebe who, besides being a trumpeter, was organist of two churches at Zeitz from 1694 until his death in 1742. As a trumpeter’s daughter, Anna may well have met the Bachs socially. Since the famous cantatrice Pauline Kellner (1664–1745) was active at the Weissenfels court from 1716, it is likely that Anna Magdalena was trained by her. She was paid for singing, with her father, in the chapel at Zerbst on some occasion between Easter and midsummer 1721. At about June 15 and not quite aged 20, she joined the Cöthen capelle at the rank of chamber musician. Her wedding with Bach (aged 36) took place on December 3. The prince saved Bach 10 thaler by giving him permission to be married in his own lodgings. At about this time Bach paid two visits to the city cellars, where he bought first one firkin of Rhine wine, and later two firkins, all at a cut price, 27 instead of 32 groschen per gallon. Probably early in 1722, Bach gave his wife an album, the Clavier-Büchlein for Anna Magdalena Bach. 25 leaves are extant, about a third of the original manuscript. The album contains fragmentary composing scores of the French Suites and has a later replacement title-page, on which Anna Magdalena wrote the title and the date and Bach noted the titles of three theological books. It seems to have been filled by 1725 when Bach started a second album for his wife.

On December 11, 1721 the prince married his cousin Friderica Henrietta, Princess of Anhalt-Bernburg. The marriage was followed by five weeks of illuminations and other entertainments at Cöthen. This was not however an auspicious event for Bach because the prince’s second wife was ‘eine Amusa’ (someone not interested in the Muses) and probably had a somewhat negative influence on the generally happy relationship between Bach and her husband. But what certainly played a more important role in the court’s attitude toward musical affairs were the budgetary reductions beginning in 1720 and the resulting personnel reductions in the capelle. Family pressures involving the powerful princess mother Gisela Agnes and her younger son August Ludwig as well as external events forced Prince Leopold into economizing. That may have already been on the horizon when Bach explored the Hamburg situation in late 1720.

A legacy from Tobias Lämmerhirt (Bach’s maternal uncle) had facilitated Bach’s first marriage; Tobias’s widow was buried at Erfurt on September 12, 1721, and Bach received something under her will too, though not in time for his second marriage. On January 24, 1722 Bach’s sister Maria, together with one of the Lämmerhirts, challenged the will, saying that Bach and his brothers Jacob (at Stockholm, Sweden) and Christoph (at Ohrdruf) agreed with them (Christoph had died in 1721). Bach heard of this only by accident; and on March 15 he wrote to the Erfurt council on behalf of Jacob as well as himself. He objected to his sister’s action and said that he and his absent brother desired no more than was due to them under the will. On April 16 Jacob died; and the matter seems to have been settled on these lines towards the end of the year. Bach’s legacy of 224 thaler amounted to more than a year’s pay.

In summer 1722 there was no Kapellmeister at the neighbouring court of Anhalt-Zerbst, and Bach was commissioned to write a birthday cantata for the prince; for this he was paid 10 thaler in April and May. The birthday was in August, and payments made during that month presumably refer to the performance of bwv1154, a lost work which was scored for two oboes d’amore and ‘other instruments’.

Several didactic works for keyboard belong to the Cöthen period. The autograph of Das wohltemperirte Clavier (book 1 of the ‘48’) is dated 1722 on the title-page. The handsome fair copy was preceded by drafts, like those in W.F. Bach’s Clavier-Büchlein (begun in 1720); and some of the movements look earlier than that. Presumably Bach brought them together for convenience, partly to serve as the last step in his keyboard course, focusing on contrasting free-improvisatory and strict-imitative styles, partly to exhibit the advantages of not-quite-equal temperament (the term ‘wohltemperirt’ being borrowed from Werckmeister).

Open in new tab

1. Autograph MS of Bach’s organ chorale Der Tag, der ist so freudenreich bwv605 (completed in new German organ tablature) from the Orgel-Büchlein, composed mostly c1710–15 (D-B Mus.ms.Bach P 283, f.9r)

Staatsbibliothek zu Berlin

The title page was almost certainly the only part of the Orgel-Büchlein (Fig. 1) that Bach wrote at the turn of 1722 while still at Cöthen, but as another educational work it is best mentioned here. It was meant to be a collection of chorale preludes, not only for the ordinary church seasons but also for occasions when such subjects as the Lord’s Prayer, or Penitence, were being emphasized. The paper is of a kind that Bach used only in Weimar. Very few items (revisions of bwv620 and 631, composition of bwv1169) date from the Leipzig period; in the rest, the writing resembles that of the cantatas of 1715–16. Of the 164 preludes Bach allowed for, he completed fewer than 50. Last in this group of works come the Auffrichtige Anleitung (‘upright instruction’) with 15 Inventions and 15 Sinfonias, whose autograph fair copy is dated ‘Cöthen, 1723’. Its contents had already appeared, in earlier versions and under different titles, in W.F. Bach’s Clavier-Büchlein of 1720.

The story of Bach’s move to Leipzig begins with the death of Kuhnau, Kantor of the Thomasschule there, on June 5, 1722. Six men (Bach not among them) initially applied for the post, among them Telemann, who was still remembered for the good work he had done at Leipzig 20 years before. He had been doing a similar job at Hamburg for about a year and was probably the most famous of German musicians actually living in Germany. One of the Kantor’s duties was to teach Latin. Telemann refused to do that; nevertheless, he was appointed on August 13. But the Hamburg authorities would not release him and offered to increase his pay; in November he declined the Leipzig post. At a meeting on November 23 Councillor Platz said that Telemann was no loss; what they needed was a Kantor to teach other subjects besides music. Of the remaining five candidates, three were invited to give trial performances; two dropped out, one because he would not teach Latin. By December 21 two Kapellmeisters newly applied, Bach and Graupner, and both were immediately invited by the town council to audition for the position. The remaining other candidates were Kauffmann of Merseburg, Schott of the Leipzig Neukirche, and Rolle of Magdeburg. Of the five candidates, Graupner was preferred; he was a reputable musician, and had studied at the Thomasschule. He successfully performed his test on January 17, 1723 with two cantatas, one before and one after the sermon. But on March 22 he too withdrew, having been offered more pay at Darmstadt. Meanwhile, Bach had auditioned on February 7, 1723 with the two cantatas bwv22 and 23. He was not an unknown in Leipzig, particularly since Kuhnau had invited him to perform bwv172 on Whitsunday 1721 at the two Leipzig main churches. Rolle and Schott were also heard, and possibly Kauffmann too, but each of them with only one cantata. On April 9, after Graupner had withdrawn, the council considered Bach, Kauffmann, and Schott. Like Telemann, none of them wished to teach Latin. Out of fear that the shortlisted candidate Bach might also be prevented by his prince from accepting the position, Councillor Platz said that as the best men could not be got, they must make do with the mediocre. But the council resolved to approach Bach, for on April 13 he obtained written permission to leave Cöthen. On April 19 he signed a curious document that reads as if he were not yet free from Cöthen but could be free within a month; like Telemann before him, he also said he was willing to pay a deputy to teach Latin. On April 22 the councillors agreed on Bach, one of them hoping that his music would not be theatrical. On May 5 he came in person to sign an agreement; on May 8 and 13 he was interviewed and sworn in by the ecclesiastical authority; on May 15 the first instalment of his salary was paid. With family and furniture, he moved to Leipzig on May 22, and performed his inaugural cantata, bwv75, at the Nikolaikirche at the opening of the school year, on May 30, the first Sunday after Trinity. On June 1, at 8.30 a.m., he was formally presented to the school.

This story has been told in some detail because it throws light on the circumstances in which Bach worked at Leipzig. To him, the Kantorate was a step downwards in the social scale, and he had little respect for his employers. The stage was set for trouble, and in due course trouble came. Councillor Platz on Telemann is curiously echoed by Councillor Stieglitz, ten days after Bach’s death: ‘The school needs a Kantor, not a Kapellmeister; though certainly he ought to understand music’. In Cöthen, Bach was not replaced and remained as titular Kapellmeister with minimal obligations at the court about 40 miles away from Leipzig, until the death of Prince Leopold in 1728. Performances there are documented for July 1724 (with A.M. Bach), December 1725, November 1726 (bwv36.2), January 1728, and on the occasion of the Prince’s funeral ceremonies in March 1729 (bwv1143).

7. Leipzig, 1723–9.

The position of Kantor at the Thomasschule, held conjointly with that of civic director of music, had been associated with a wealth of tradition since the 16th century. Ever since the Kantorates of Seth Calvisius (1594–1615) and Johann Hermann Schein (1615–30) the institution had become the preeminent training ground for young musicians in Protestant Germany. Up to a quarter of the graduating students, who came from all over the country, ended up in musical professions. It was one of the most notable positions in German musical life both in this and in the esteem it commanded; and there can be little doubt that the general attractiveness of the position in itself played a part – very likely the decisive part – in Bach’s decision to move from Cöthen to Leipzig. His subsequent remark about the social step down from Kapellmeister to Kantor must be seen in the context of his later disagreements with the Leipzig authorities, as indeed the letter in question (to Erdmann, a friend of his youth, on October 28, 1730) makes unequivocally clear. In any event, Bach was not the only Kapellmeister to apply for the post. The duties were incomparably more varied and demanding than those in Cöthen or Weimar (to say nothing of Mühlhausen or Arnstadt) and more or less corresponded to those undertaken by Telemann in Hamburg. It cannot have been mere chance that Bach wanted to tackle a range of duties comparable with those of his friend. Above all he must have preferred the greater economic and political stability of a commercial metropolis governed democratically to the uncertainties of the court of an absolute prince, where personal whim often held sway. The university – the foremost in the German-speaking world at the time – must have been another special attraction in the eyes of a father of growing-up sons.

Musical aptitude was a decisive factor in the selection of pupils for the Thomasschule, and it was the Kantor’s responsibility to assess and train them – altogether 55 ‘alumns’, i.e., scholarship recipients to be housed in the school’s dormitory. The alumns were joined by a substantial number of ‘external’ students from families residing in Leipzig. Their musical education was furthered by the daily singing lessons, mostly given by the Kantor. There was also instrumental instruction for the ablest pupils, which Bach had to provide free of charge but was thus enabled to make good any shortage of instrumentalists for his performances. Indeed, the number of professional musicians employed by the town (four Stadtpfeifer, three fiddlers, and one apprentice) was held throughout his period of office at the same level as had obtained during the 17th century. For further instrumentalists Bach drew on the apprentices and students of the Stadtpfeifer as well as university students. In general, the age of the Thomasschule pupils ranged between 12 and 23. Remembering that voices then broke at the age of 17 or 18, it is clear that Bach could count on solo trebles and altos who already had some ten years’ practical experience – an ideal situation, impossible in boys’ choirs today.

As far as church music was concerned, Bach’s duties centred on the principal services on Sundays and church feasts, as well as some of the more important subsidiary services, especially Vespers. In addition, he could be asked for music for weddings and funerals, for which he would receive a special fee. Such additional income was important to Bach, as his salary as Kantor of the Thomaskirche and director of music came to only 87 thaler and 12 groschen (besides allowances for wood and candles, and payments in kind, such as corn and wine). In fact, including payments from endowments and bequests as well as additional income, Bach received annually more than 700 thaler. Further, he had the use of a spacious official residence in the south wing of the Thomasschule, which had been renovated at a cost of more than 100 thaler before he moved in in 1723. Inside the Kantor’s residence was the so-called ‘Komponirstube’ (‘composing room’), his professional office containing his personal music library and the school’s. The buildings of the old Thomasschule were, scandalously, demolished in 1903 to make room for senior minister’s quarters (today church offices). Already in the 1880s, the west façade of the Thomaskirche (originally attached to the Thomasschule) was rebuilt in the neo-Gothic style, with a significant increase of the space for the organ and choir gallery.

During his early Leipzig years, Bach involved himself in church music with particular thoroughness and extreme energy. This activity centred on the ‘Hauptmusic’ composed for Sundays and church feasts. The performance of a polyphonic cantata, with a text related as a rule to the Gospel and/or Epistle for the day, was a tradition inherited from previous Kantors. Even so, Bach engaged on a musical enterprise without parallel in Leipzig’s musical history: in a relatively short time he composed at least three complete or nearly complete cycles of cantatas for the Church year, with about 60 cantatas in each, making a repertory of roughly 200 sacred cantatas extant today. The first two cycles were prepared immediately, for 1723–4 and 1724–5; the third took rather longer, being composed between 1725 and 1727. The obituary lists a total of five annual cycles, yet it remains unclear how exactly the large body of cantatas was shelved in Bach’s office and counted after the composer’s death. The fourth cycle, to published texts by Picander, presumably dates from 1728–9, but the texts were apparently only selectively set to music; the formation and makeup of the fifth cycle proves totally elusive. The established chronology of Bach’s vocal works demonstrates that the main body of the cantatas was in existence by 1729, and that Bach’s development of the cantata was effectively complete by 1735. The existence of the fourth and fifth cycles has been questioned, because of their fragmentary survival compared with the almost complete survival of the first, second, and third; but until a positive argument for their non-existence can be put forward the number of five cycles, laid down in the obituary of 1754, must stand. Compared with the high proportion of Bach’s works of other kinds that are lost (orchestral and chamber music, for instance), the disappearance of about 100 cantatas would not be exceptional. (The preservation of Bach’s works is discussed below, §11; see §III, (7), 15 for the correspondence of excess chorales in the Breitkopf collection of 1784–7 to the number of lost cantatas.)

The first cycle begins with the beginning of the academic school year on the first Sunday after Trinity 1723 with Cantata bwv75. It was performed ‘mit gutem applausu’ at the Nikolaikirche, followed by bwv76, for the second Sunday after Trinity, performed at the Thomaskirche. The two largest churches in Leipzig are both Gothic in style, and in Bach’s time they contained stone and wooden galleries. The choir lofts were on the west wall of the nave above the council gallery. The organs too were in the choir lofts (the ‘Schüler-Chor’): the Nikolaikirche and the Thomaskirche each had a three-manual organ with 36 and 35 stops respectively (Oberwerk, Brustwerk, Rückpositiv, Pedal). The Thomaskirche had a second organ, fitted to the east wall as a ‘swallow’s nest’, with 21 stops (Oberwerk, Brustwerk, Rückpositiv, Pedal); this fell into dilapidation and was demolished in 1740. The organs were always played before cantata performances, during which they would provide continuo accompaniment; they were played by the respective organists at each church; during Bach’s term of office these were Christian Heinrich Gräbner (at the Thomaskirche until 1729), Johann Gottlieb Görner (at the Nikolaikirche until 1729, then at the Thomaskirche), and Johann Schneider (at the Nikolaikirche from 1729). Bach himself, who had not held a regular appointment as an organist since his time in Weimar, directed the choir and the orchestra, and would not normally be playing the organ. However, he frequently must have directed his church ensemble from the harpsichord, as is documented for the performance of bwv198 in 1727. At any rate, the harpsichord was often, if not regularly, employed as a continuo instrument in addition to the organ.

The cantata was an integral part of the Leipzig Lutheran liturgy. It followed immediately on the reading from the Gospel, preceding the Creed and the sermon (the second part of a two-part cantata would follow the sermon, ‘sub communione’). Apart from organ playing and the congregational singing of hymns, selected by the Kantor, the other musical constituent of the liturgy was the introit motet, which would be taken from the Florilegium Portense (1618) by Erhard Bodenschatz, a collection mainly drawn from the 16th century (Lassus, Handl etc.), and was performed a cappella with harpsichord continuo. Services began at 7 a.m. and lasted three hours; this allowed a mere half-hour for the cantata, and Bach rarely overstepped this duration. The normal performing forces consisted of about 12–16 singers and 18 instrumentalists; the precise number varied according to the work, but it was rare for the total number of singers and players to fall below 25 or to exceed 40 (the figure required on exceptional occasions, like the St Matthew Passion, which demanded two Kantoreien and double the normal number of instrumentalists). Ordinarily the performing forces under Bach’s direction consisted of four groups: pupils from the Thomasschule (the first ‘Kantorei’); the eight salaried town musicians, until 1734 headed by J.G. Reiche and thereafter by J.C. Gentzmer; University students (principally Bach’s private pupils); and additional assistants (probably regularly including one or two paid soloists) and guests. The first Kantorei alternated weekly between the two main churches, St Thomas and St Nicholas, while the other three under the direction of choir prefects provided a cappella music at the three other churches under Bach’s purview.

Bach took up his additional duties as music director to the university, a post traditionally held by the Thomaskantor, by August 9, 1723 when he performed the Latin Ode bwv1155 (lost) at the university’s festivities marking the birthday of Duke Friedrich II of Saxe-Gotha. The major part of his duties for the university comprised the musical provisions for the so-called quarter-day orations and the ‘old’ services in the Paulinerkirche, the University church, employing pupils from the Thomasschule and town musicians on the four major festivals of Christmas Day, Easter Day, Whit Sunday, and Reformation Day; Bach was paid 2 thaler and 6 groschen on each occasion. He carried out the most important of his civic duties for the first time on August 30, 1723, when he introduced bwv119 as part of the annual celebration of the change of town council at the Nikolaikirche. The enormous scope of Bach’s new responsibilities, as well as his vast workload, may be gauged from the fact that the day before (14th Sunday after Trinity) bwv25 was heard for the first time, and the first performance of bwv138 (for the 15th Sunday) was soon to follow.

September 1723 saw the start of Bach’s protracted wrangle with the university. In a written request for payment, he laid claim to the traditional right of the Thomaskantor to be responsible for the ‘old’ services and the quarter-day orations. The university, however, wanted to combine these duties with responsibility for the ‘new’ services (normal Sundays and holy days), which it had in April 1723 entrusted to J.G. Görner, organist of the Nikolaikirche, together with the title of ‘Musik Direktor’. On September 28 Bach’s request was turned down, and he was paid only half the fee. He would not give in and turned to the Elector of Saxony in Dresden with three petitions. Following the intervention of the Dresden court, the university decided to put Görner in charge of the ‘new’ services only, and awarded Bach his traditional rights with payment as before. Thereafter, as the regular fee payments prove, Bach retained responsibility for the ‘old’ services and quarter-day orations until 1750.

On November 2, 1723 Bach inaugurated a new organ (which he had previously appraised) in Störmthal, outside Leipzig, with bwv194. Then, from the second Sunday in Advent to the fourth, came his first break in the weekly routine of composing and performing cantatas; in Leipzig, unlike Weimar, this period was a ‘tempus clausum’, as was Lent up to and including Palm Sunday. On Christmas Day figural music returned, in a particularly splendid manner, with bwv63 and the D major Sanctus bwv238 at the main service and the Magnificat bwv243.1 at Vespers; these were Bach’s first large-scale compositions on Latin texts such as were customary in Leipzig on major feast days. At this point in the calendar his duties were unimaginably heavy, yet he carried them out with incomparable creative vigour, producing bwv40 and 64 for the feasts of St Stephen and St John the Evangelist, bwv190 for New Year, bwv153 for the Sunday after New Year (January 2, 1724), bwv65 for Epiphany (January 6) and bwv154 for the following Sunday (January 9); after that, normal weekly services were resumed.

During the next ‘tempus clausum’ Bach composed his first large-scale choral work for Leipzig, the St John Passion bwv245.1, first performed at Vespers in the Nikolaikirche on Good Friday (April 7). This Vespers service had been introduced specially for the performance of a Passion only in 1721; in that year Kuhnau’s St Mark Passion (now lost) had been performed. Performances alternated annually between the Thomaskirche and the Nikolaikirche, an arrangement to which Bach strictly adhered. bwv245 had several further performances, each time in a greatly altered version (see §III, (7), 14): on March 30, 1725 (in a second version adapted to the annual cycle of cantatas), probably on April 7, 1730 (in a third version), and on April 4, 1749 (fourth version); in about 1738 Bach undertook a revision of the work which remained unfinished.

With the first Sunday after Trinity 1724 (June 11) Bach began his second cycle; these were chorale cantatas. Not least because it included works composed at Weimar, the first cycle had been thoroughly heterogeneous in character, both musically and textually, but Bach gave the new cycle a unifying concept, with all the works based on texts, and their melodies, from the hymnbook. Unfortunately, this series of chorale cantatas, beginning with bwv20, O Ewigkeit, du Donnerwort, and its programmatic overture, was interrupted early in 1725 after 44 weeks; Bach continued the weekly production of new works, but with non-chorale cantatas such as settings of texts by the young poet Christiane Mariane von Ziegler. In subsequent years he added to the cycle but never completed it. On June 25 he was in Gera for the dedication of the organ at the Salvatorkirche. In July he went to Cöthen with Anna Magdalena for a guest appearance as a performer; he had retained the title of Court Kapellmeister there, and it lapsed only on the death of Prince Leopold in 1728. There is evidence of further visits to Cöthen, with Bach performing alongside his wife (who sang as a soprano), in December 1725 and January 1728. During 1725 Bach started to prepare a second Clavier-Büchlein for Anna Magdalena. On February 23, 1725 he performed the Tafel-Music bwv249.1 at the Weissenfels court for the birthday of Duke Christian; this was the original version of the Easter Oratorio bwv249.3, first given at Leipzig the following April 1. bwv249 represents the beginning of a long-standing collaboration with the fluent Leipzig poet Christian Friedrich Henrici (Picander), the chief supplier of texts for Bach’s later Leipzig vocal works.

Bach produced congratulatory cantatas for two Leipzig University professors in May and August (bwv36.1 and 205). On September 19–20 he played the Silbermann organ at the Dresden Sophienkirche before the local court musicians (on one of the two days performing concertos, with orchestral accompaniment provided by Dresden court musicians), thus continuing his practice of giving virtuoso organ performances on concert tours – and undoubtedly in Leipzig, too, although he no longer held a post as organist. His favourite instrument in Leipzig was evidently the great organ of the Paulinerkirche built by Johann Scheibe in 1716, with 53 stops, three manuals (Hauptwerk, Seitenwerk, and Brustwerk) and pedals; Bach had been one of its examiners in 1717. Early in 1726 – during the third cycle, which had started in June 1725 – there was an interruption of Bach’s production of cantatas, for reasons that remain obscure: between February and September 1726 he performed 18 cantatas by his cousin Johann Ludwig Bach (3/72). In particular, between Purification and the fourth Sunday after Easter, he performed none of his own music at the main Sunday services; even on Good Friday he used a work by another composer, Reinhard Keiser’s St Mark Passion, which he had performed once before, in Weimar. Difficulties with performers may have been partly responsible; the instrumental forces required in J.L. Bach’s cantatas are more modest than those Bach himself normally used. Even apart from this, however, the pattern of Bach’s cantata production – as far as can be judged from the available material – changed during the third cycle; there are considerable gaps as early as the period after Trinity Sunday 1725, and it seems that the third cycle, unlike the first two, extended over two years. In the gaps, cantatas by other composers and further performances of Bach’s own works were given.

The extended interruption of cantata production may have been – in part, if not primarily – a scheme in order to turn to other creative activities such as the composition of the St Matthew Passion and the preparation of the keyboard partitas. In fact, the fall of 1726 saw the appearance in print of Partita bwv1, under the general title of Clavier-Übung: with this Bach began his activity, later to increase in scope, as a publisher of keyboard music. Partita bwv1, published singly, was followed by nos. 2 and 3 (1727), bwv4 (1728), bwv5 (1730), and bwv6 (1730 or 1731; no copy is known). Evidently the series was originally planned to comprise seven partitas. There are early versions of nos. 3 and 6 in the second album for Anna Magdalena of 1725. Bach sent bwv1, with a dedicatory poem, to the Cöthen court as a form of congratulation on the birth of an heir, Prince Emanuel Ludwig (born September 12, 1726). In December 1726, on the installation of Dr Gottlieb Kortte as university professor, Bach produced a more sizable occasional work, the dramma per musica bwv207.1.

In 1727 Bach produced two extremely important works. The St Matthew Passion, for double choir to a libretto by Picander, was performed on Good Friday (April 11; there is evidence that it was repeated in the Thomaskirche in 1729, 1736, and 1742; see §III, (7), 14). The other work was the Trauer Ode bwv198, performed in October at a memorial ceremony, planned by the university, on the death of the Electress Christiane Eberhardine, who had remained a Protestant when her husband, August the Strong of Saxony, converted to Roman Catholicism. For this Bach was commissioned to set a text by the Leipzig professor of poetry, Johann Christoph Gottsched. This became a somewhat controversial affair, as the university director of music, Görner, felt he had been slighted. Bach however retained the commission and performed the two parts of his work, ‘composed in the Italian manner’, directing it from the harpsichord, in the university church, on October 17. Between September 7, 1727 and January 6, 1728 there was a period of national mourning, with no other musical performances.

In September 1728 a brief dispute with the church authorities flared up. The sub-deacon, Gaudlitz, demanded that he himself should choose the hymns to be sung before and after the sermon at Vespers; as it was usual for the Kantor to select these hymns, Bach felt that his rights had been encroached upon. As the dispute was settled in the sub-deacon’s favour Bach must have seen this as a setback, for once again his grievances had not been met; but his relations with the ecclesiastical authorities were on the whole good throughout his time at Leipzig. His relations with the town council, in particular the supervisor of the Thomasschule, Councillor Stieglitz, went less smoothly and were to become even more difficult in the 1730s. Documents dealing with the various disputes show Bach to have been a stubborn defender of the prerogatives of his office who frequently reacted with excessive violence and was often to blame if there was a negative outcome. It would be wrong, however, to draw hasty inferences about Bach’s personality and his relations with the world about him. It is an unfortunate coincidence that about a half of Bach’s surviving correspondence is concerned with generally trivial but often protracted disputes over rights. This material is extant in public archives, while utterances of kinds not appropriate to archival preservation, which might have complemented this rather austere view of his personality, have survived in only small quantity. From Bach’s behaviour during these disputes it can be seen that, under pressure, he would defy bureaucratic regulations in order to preserve his independence and to clear himself an artistic breathing-space. His taking over of the collegium musicum in 1729, to be directed under his own management, must be seen in this context, as it represents something more than an incidental biographical fact.

Early in 1729 Bach spent some time at the Weissenfels court in connection with the birthday celebrations in February of Duke Christian, with whom he had long been associated. In January the title of court Kapellmeister of Saxe-Weissenfels was conferred on him (his Cöthen title had expired with the death of Prince Leopold); he retained the title until the duke’s death in 1736. At the end of March, he went to Cöthen to perform the funeral music for his former princely employer; only the text survives of this large-scale work in four parts (bwv1143), but much of its music can be reconstructed as it consists of parodies of bwv198 and 244. On April 15 (Good Friday) the St Matthew Passion was performed again at the Thomaskirche. On the second day of Whit week (June 6) and very near the end of the school year, the Picander cycle cantata bwv174 was performed. The manuscript, uniquely for Bach, is dated ‘fine d: 5 Junij 1729. Lipsiae’; perhaps this represents some sort of final gesture after a heavy, six-year involvement in cantata composition.

Beside the production of cantatas, Passions, and other vocal occasional works, both sacred and secular, instrumental music retreated to the background during Bach’s first years in Leipzig. Apart from some keyboard and chamber works (including the sonatas for harpsichord and violin bwv1014–19) there appear to have been only a relatively small number of organ works (preludes and fugues, trio sonatas) which are hard to date individually but will have been primarily connected with Bach’s activities as a recitalist.

In June 1729 an invitation to visit Leipzig was delivered to Handel, then in Halle, by Wilhelm Friedemann, in place of his father who was ill at the time; but nothing came of it. Thus, Bach’s second and last attempt to establish contact with his highly esteemed London colleague met with failure. Significantly, in both cases the initiative was taken by Bach.

8. Leipzig, 1729–39.

On his appointment as director of the collegium musicum, decisive changes came about in Bach’s activities in Leipzig; and at the same time new possibilities were opened up. The collegium had been founded by Telemann in 1702 and had most recently been directed by Georg Balthasar Schott (who left to become Kantor at Gotha in March 1729); it was a voluntary association of professional musicians and university students that gave regular weekly (and during the fair season even more frequent) public concerts. Such societies played an important part in the flowering of bourgeois musical culture in the 18th century, and with his highly reputed ensemble, in such an important commercial centre as Leipzig, Bach made his own contribution to this. After having previously been loosely connected with the collegium (e.g., performing bwv193.1 in 1727), he took over its direction before the third Sunday after Easter – in other words, by April 1729 – and retained it until 1741.

Bach must have had strong reasons for wanting to take on this fresh area of work in addition to his other duties. To some extent it is possible to guess those reasons. For six years he had immersed himself primarily in the production of sacred music, and he had created a stock of works sufficient to supply the requirements of his remaining time in office. In his efforts to provide sacred music that was at once fastidious and comprehensive he had met with little appreciation from the authorities, and no additional facilities (for example, much needed professional instrumentalists) had been placed at his disposal: it would be understandable if he now felt resigned to the situation. Further, as a former Kapellmeister, he must have been attracted by the prospect of working with a good instrumental ensemble, and another important incentive must have been the thought that, as director of the collegium, he would be able to establish a wholly independent musical praxis, in accordance with his own ideas and, in fact, be practically in control of musical life in Leipzig. Whether the new position brought him some additional income remains unknown.

Nothing, unfortunately, is known about the programmes of the ‘ordinaire’ weekly concerts. But the surviving performing parts for such works as the orchestral suites bwv1066–8, the violin concertos bwv1041–3, and the flute sonatas bwv1030 and 1039 demonstrate that Bach performed many of his Cöthen instrumental works (some in revised form) as well as new compositions. The seven harpsichord concertos bwv1052–8, collected together in a Leipzig manuscript, also belong in this context. Bach often performed works by other composers as well, including five orchestral suites by his cousin Johann Ludwig, secular cantatas by Handel and Porpora, and the flute quartets that Telemann wrote for Paris. Further, Bach’s many musical acquaintances from other places must have made frequent appearances, including his colleagues in the Dresden court orchestra (there is evidence of visits from Johann Adolf Hasse and his wife Faustina, Georg Benda, Silvius Leopold Weiss, Carl Heinrich Graun, and Jan Dismas Zelenka). C.P.E. Bach’s remark that ‘it was seldom that a musical master passed through [Leipzig] without getting to know my father and playing for him’ must refer to performances of the collegium musicum, which took place on Wednesdays between 4 and 6 p.m. in the coffee-garden ‘before the Grimmisches Thor’ during the summer months and on Fridays between 8 and 10 p.m. in Zimmermann’s coffee-house during the winter. In addition, there were ‘extraordinaire’ concerts, to mark special events; on these occasions, during the 1730s, Bach performed his large-scale secular cantatas. The number of text booklets (300) printed for those special events provide an idea about the size of the audience. His activities with the collegium must have made heavy demands on him, and the reduction in his production of sacred music is easy to understand.

This does not, however, mean that his interest in sacred music was diminished. Such a view is contradicted not only by the major ecclesiastical works written after 1730 but also by the simple fact that, throughout his period of office, Bach provided performances of his cantatas, a repertory largely completed before 1729, every Sunday at the two main Leipzig churches. His reference to the ‘onus’ of such undertakings, in connection with the performance of a Passion planned for 1739, might just as well have been made in the 1720s. Admittedly, his difficulties became particularly acute about 1730, as his important memorandum of August 23, 1730, dealing with the state of church music in Leipzig and outlining his remedies, testifies. His letter of October 28 that year, to his old friend Erdmann in Danzig, may be read in the same sense; sheer frustration that the memorandum had proved ineffectual drove him to consider leaving Leipzig. It would seem that his work with the collegium musicum had not yet brought about the intended equilibrium in his activities.

The situation had been aggravated by other, external factors. The old headmaster Johann Heinrich Ernesti had died in 1729 (Bach had performed the motet bwv226 at his funeral in October). During the subsequent interim in the Thomasschule’s direction the organization of school life was disturbed. Problems of space appear to have arisen too. It was in this context that complaints were made about Bach’s neglect of his school duties (the dropping of singing lessons, absence on journeys without leave); in August 1730 there was even a question of reducing his salary ‘because the Kantor is incorrigible’. It would appear that things were put right by Johann Matthias Gesner, an old acquaintance from the Weimar years who took over the headship of the school in the summer, and who seems soon to have established friendly and familiar relations with Bach.

On Good Friday 1730 Bach apparently performed the St John Passion in its third version bwv245.3, and in the two following years the St Mark Passion bwv247 and an anonymous St Luke Passion bwv246. From June 25 to 27, 1730 the bicentenary of the Augsburg Confession was celebrated across Lutheran Germany, and Bach wrote three cantatas for the event (bwv190.2, 120.3, and 1139.2). They are not untypical of his church compositions of this period, most of which were put together as reworkings of earlier compositions and parodies; that is also true also of the major vocal works like the St Mark Passion, the B minor Mass, the Kyrie-Gloria masses, and the oratorio trilogy for Christmas, Easter, and Ascension Day. Among the relatively few new cantatas composed between 1729 and 1735, the chorale-text-only works bwv9, 14, 97, 100, 112, 117, 140, 177, and 192 stand out.

In 1731 a collected edition of the six partitas appeared as op.1, under the title I.Teil der Clavier-Übung. From this formulation it is clear that Bach planned further ‘parts’ in a series of ‘keyboard exercises’, and these he now proceeded to produce. His new and continuing interest in publishing his own compositions is a clear sign of a new determination with regard to independent and freely creative activity. At the end of June 1731 Bach and his family had to move to temporary quarters while rebuilding and extension work were being carried out on the Thomasschule. His residence must have become increasingly cramped, for his family was growing. In the early years in Leipzig Anna Magdalena had borne a child almost every year, but few of them survived infancy: Christiana Sophia Henrietta (b spring 1723; d June 29, 1726), Gottfried Heinrich (48), Christian Gottlieb (bap. April 14, 1725; d Sept 21, 1728), Elisabeth Juliane Friederica (bap. April 5, 1726; d Leipzig, Aug 24, 1781), Ernestus Andreas (bap. Oct 30, 1727; d Nov 1, 1727), Regina Johanna (bap. Oct 10, 1728; d April 25, 1733), Christiana Benedicta (bap. Jan 1, 1730; d Jan 4, 1730), Christiana Dorothea (bap. March 18, 1731; d Aug 31, 1732), Johann Christoph Friedrich (49), Johann August Abraham (bap. Nov 5, 1733; d Nov 6, 1733), Johann Christian (50), Johanna Carolina (bap. Oct 30, 1737; d Leipzig, Aug 18, 1781), Regina Susanna (bap. Feb 22, 1742; d Leipzig, Dec 14, 1809).

Joy and sorrow were everyday matters. But Bach’s family life must have been harmonious in more than one sense; in 1730 he reported, as a proud paterfamilias, that with his family he could form a vocal and instrumental concert ensemble. The family moved back into their refurbished apartment the next April. The school was reconsecrated on June 5, 1732 with a cantata, bwv1162. In September 1731 Bach had been to Dresden for the first performance of Hasse’s opera Cleofide and to give concerts at the Sophienkirche and at court (there were enthusiastic reports in the newspapers). In September 1732 he went with his wife to Kassel for the examination and inauguration of the organ of the Martinskirche, where he reportedly played the ‘Dorian’ Toccata and Fugue in D minor bwv538.

With the death of Elector Friedrich August I of Saxony on February 1, 1733 a six-month period of national mourning began.During the mourning period and as soon as the date was set for the fealty visit in Leipzig by Elector Friedrich August II, Bach appears to have composed the Missa (Kyrie and Gloria) of what would later become the B minor Mass bwv232 for the festive service in the Nikolaikirche on April 21. However, whether or not the performance actually took place on that day – after it was decided that the protestant court without the Roman-catholic Elector would attend the Lutheran service – remains unclear. The collegium musicum obtained special permission to restart its performances in the middle of June, when a new harpsichord, or possibly a fortepiano, was introduced. During the mourning period Bach had completed the D major version of the Magnificat bwv243.2, which was probably first heard in Leipzig when the church music restarted on July 2 (Visitation). Later in July, Bach travelled to Dresden where he presented the new Elector with a set of parts of the Missa, along with a note dated July 27, 1733 that expressed his hope of obtaining a title at the court Kapelle. There is evidence to suggest that the Missa was also performed at this time, perhaps at the Sophienkirche in Dresden, where W.F. Bach had been working as an organist since June 1733. Not until November 1736, however, was the title of ‘Royal Polish and Electoral Saxon Court Compositeur’ conferred on Bach because it conflicted with the Weissenfels court title he held until the duke’s death in June 1736. But after the intervention of his patron Count Keyserlingk in September of that year, the court appointment became effective on November 19. As a gesture of thanks, Bach paid his respects to the Dresden royal household and an enthusiastic public with a two-hour recital on the new Silbermann organ at the Frauenkirche on December 1. Henceforth, his name and title as 'church composer’ was kept on the rolls of the Dresden Royal Kapelle and Chamber Music through 1750.

After the dedication of the Missa in July 1733, Bach kept the Saxon royal family’s interests in mind with his ‘extraordinaire’ concerts of the collegium musicum. On August 3, the name day of the new elector, Bach began his remarkable series of secular cantatas of congratulation and homage with bwv1158 (music lost), followed by bwv213 (September 5, for the heir to the electorate), bwv214 (December 8, for the electress), bwv205.2 (February 19, 1734, for the coronation of the elector as King of Poland; music lost), an unknown work (August 3, again for the elector), and bwv215 (October 5, also for the elector, who was at the performance in Leipzig). Much of the festive music was performed in the open air with splendid illuminations, and according to newspaper reports the music benefited from a resounding echo. (On the day after the performance of bwv215 Bach’s virtuoso trumpeter and the leader of the Leipzig Stadtpfeifer, Gottfried Reiche, died as a result of the exertions of his office.) During the following Christmas season Bach gave the people of Leipzig a chance to hear much of the music from his secular festive cantatas in modified form, as the Christmas Oratorio, which was heard in six sections between Christmas Day 1734 and Epiphany 1735 (and consisted predominantly of parodies of bwv213–15).

On November 21, 1734 the new headmaster of the Thomasschule, Johann August Ernesti, was officially welcomed; the distinguished classicist Gesner had moved to the newly founded University of Göttingen as its first dean. Bach’s dealings with the school leadership had been untroubled for four years, thanks to his friendly relations with Gesner; but with Ernesti he experienced the most violent controversies of his entire period as Thomaskantor. A dispute flared up in August 1736 over the authority to nominate the choral prefect, in which the interests of the Kantor and the headmaster were diametrically opposed. With his neo-humanist educational ideals, which placed priority on high academic standards, Ernesti showed little appreciation of the musical traditions. The tendency at the Thomasschule, at least from the start of Bach’s period of office, had been to restrict musical activities, or at any rate to reduce their proportions; Bach, on the other hand, demanded the best-qualified pupils to assist him, and certainly he must often have overburdened them (with music copying, rehearsals, and so on). Against what were to some extent unfair arguments on the headmaster’s part, his struggles were doomed to failure. The grievances arising from the nomination of the choir prefect were taken before the courts in Dresden; the affair, which led to Bach’s having disciplinary difficulties with his pupils, was settled early in 1738 (the precise outcome is not recorded). The prefect in question, Johann Gottlob Krause, whom Bach refused to acknowledge, had already left the Thomasschule in 1737. Bach’s Dresden court title made him not quite unassailable by the Leipzig authorities, but it definitely provided him with considerable protection.

Open in new tab

2. Bach’s autograph letter of recommendation for his pupil Johann Ludwig Krebs, August 24, 1735 (D-Zsa III Z.40.7, f.36r)

Stadtarchiv, Zwickau

Among the more important events of 1735 was the appearance of the second part of the Clavier-Übung at Easter. In the context of Bach’s activities as a publisher it should also be mentioned that by 1729 he was also involved in the distribution of musical publications by other authors and kept a stock, including Heinichen’s book on figured bass, Walther’s Lexicon, and keyboard works by Hurlebusch, Krebs (Fig. 2), and his own sons. In June he travelled to Mühlhausen, where he had spent part of his early career, to appraise the rebuilt organ in the Marienkirche, where his son Johann Gottfried Bernhard (47) had just been appointed organist. During Advent 1735, when no music was performed, and Lent 1736 Bach was probably engaged on the revision of the St Matthew Passion and in making a carefully laid-out fair copy of the new expanded version. In this form, characterized by its writing for double chorus and orchestra (now with two continuo parts), the work was performed in the Thomaskirche on March 30, 1736, with the cantus firmus parts in the opening and closing choruses of part 1 played on the ‘swallow’s nest’ organ. Also at Easter the Schemelli Hymnbook, on whose tunes and figured basses Bach had collaborated, was published.

In summer 1737 Bach temporarily resigned the direction of the collegium musicum and turned it over to his trusted pupil C.G. Gerlach, music director of the Neukirche. For the last ‘extraordinaire’ concert on October 7, 1736 he had written the congratulatory Cantata bwv206 on the birthday of the elector. Only two further works of homage are known from 1737–8, bwv30.1 for Baron J.C. von Hennicke (1737) and bwv1161 for the Elector and his wife (1738), which indicates that Bach was occupied primarily with the other things for which he had time after his release from the work associated with the collegium. For Easter and Ascension Day 1738 he returned to what he had worked on already in 1735: the revision of the Easter Cantata bwv249.3, now renamed ‘oratorio’ bwv249.4, and the composition of the Ascension Oratorio bwv11 – both complementing the Christmas Oratorio and completing the oratorio cycle on the major festivals celebrating the life of Christ. Moreover, at about the same time he composed the four Kyrie-Gloria Masses bwv233–6 and worked on a substantial revision of the St John Passion and prepared a new fair copy of the piece but broke off after only 20 pages for unknown reasons. He never completed the project even though he continued to perform this Passion at least one more time, in 1749.

Bach also turned to keyboard music, working on the second part of Das wohltemperirte Clavier, the revision of large organ chorales from the Weimar period (‘the 18’), and the third part of the Clavier-Übung, the largest of his keyboard works. This collection of organ pieces, some freely composed, some based on chorales, with large-scale works for a church organ and small-scale ones for a domestic instrument, appeared at Michaelmas 1739.

In the late 1730s, Bach apparently devoted himself more than previously to private teaching. Between 1738 and 1741, for example, Johann Philipp Kirnberger and Johann Friedrich Agricola were studying with him in Leipzig – probably the most important and influential of all his pupils except for his own sons. Some of the surviving teaching materials focussing on contrapuntal techniques (bwv1129–32) originate from this period. Over the Leipzig years Bach taught more than 80 private pupils; the more notable among them include Carl Ffriedrich Abel (c1743), Johann Christoph Altnickol (1744–8), Johann Friedrich Doles (1739–44), Georg Friedrich Einicke (1732–7), Heinrich Nikolaus Gerber (1724–7), Carl Gotthelf Gerlach (1723–9), Johann Gottlieb Goldberg (c1740), Gottfried August Homilius (1735–42), Johann Christian Kittel (1748–50), Lorenz Christoph Mizler (1731–6), Johann Gottfried Müthel (1750), Christoph Nichelmann (1730–33), Johann Georg Schübler (after 1740), Georg Gottfried Wagner (1723–6), and Christoph Gottlob Wecker (1723–8).

In October 1737 Bach’s nephew Johann Elias (39) came to live with the family, as private secretary and tutor for the younger children; he remained until 1742. The surviving drafts of letters he prepared give a lively picture of Bach’s correspondence in these few years – and cause for regret that no other period is similarly documented. At this period Bach gave especially close attention to the study of works by other composers. He was a subscriber to Telemann’s Parisian flute quartets of May 1738; but more typical is his preoccupation with Latin polyphonic liturgical compositions. The stile antico tradition seems to have held a particular fascination for him. In the first place he owed his knowledge of this repertory, to which he marginally contributed by making transcriptions (works by Palestrina, Gasparini, Lotti, Caldara, and others), to his connections at Dresden. His knowledge of Pergolesi’s Stabat mater of 1736, which he reworked during the later 1740s as a setting of Psalm 51, Tilge, Höchster, meine Sünden bwv1083 is also surprising; the earliest trace of Pergolesi’s work north of the Alps thus leads to Bach – a sign of the latter’s remarkable knowledge of the repertory.

On May 14, 1737 Johann Adolf Scheibe, in his journal Der critische Musikus, published a weighty criticism of Bach’s manner of composition. This seems to have come as a severe blow to Bach. Evidently at his urging, the Leipzig lecturer in rhetoric Johann Abraham Birnbaum responded with a defence, printed in January 1738, which Bach distributed among his friends and acquaintances. The affair developed into a public controversy, the literary conduct of which, at least, was suspended only in 1739 after further polemical writings by Scheibe and Birnbaum. Scheibe acknowledged Bach’s extraordinary skill as a performer on the organ and the harpsichord, but sharply criticized his compositions, claiming that Bach ‘by his bombastic and intricate procedures deprived them of naturalness and obscured their beauty by an excess of art’. Birnbaum’s elaborate and verbose replies fail to recognize the true problem, which lies in a clash of irreconcilable stylistic ideals. Nevertheless, his discussion of naturalness and artificiality in Bach’s style, and his definition of harmony as an accumulation of counterpoint, make some important statements about the premisses and unique character of Bach’s compositional art, and Bach himself must have been involved in their formulation. This is clear above all in the way in which ‘the nature of music’ is represented, with references to biographical details (such as the challenge to Marchand) and express mention of composers and works in Bach’s library (Palestrina, Lotti, and de Grigny). The controversy smouldered on for several more years. Mizler, too, shook a lance, pointing to ‘the latest taste’ in Bach’s cantata style with reference to bwv1161 of 1738 (‘so well does our Kapellmeister know how to suit himself to his listeners’). In the end Scheibe climbed down, with a conciliatory review (1745) of the Italian Concerto in which he apologized handsomely (‘I did this great man an injustice’).

9. Leipzig, 1739–50.

Assessed by his traceable musical output, Bach had given the office of Thomaskantor in the later 1730s not the kind of emphasis as before. He set his priorities differently by turning his attention to compositional projects totally independent of the office and of primary interest to him. His two-year ‘sabbatical’ from the collegium may also have been prompted by the need for time to focus on his own projects. Bach did not neglect his duties at the school and the church,but he increasingly administered the office by relying on his assistants (the choir prefects) and trusted colleagues. The headmaster complained in 1736 that ‘when he was away, he regularly had the organist of the Neukirche, Mr. Schott or Mr. Gerlach, conduct’. And in an application for a Kantor’s position, Bach’s pupil G.B. Fleckeisen boasted in 1746 – most likely with some exaggeration – that he had ‘to direct and conduct … for two entire years’ musical performances at Leipzig’s two main churches and ‘completed that service with honor’. Most of Bach’s outside travels, such as a five-week absence in the spring of 1744, are not specified but they probably involved organ exams, guest appearances as virtuoso performer, and the like.

By October 1739 – Clavier-Übung part III had just been published – Bach resumed the direction of the collegium musicum, which had temporarily been in the charge of C.G. Gerlach. A composition for the birthday of the elector (October 7; music lost) and a re-performance of bwv206 on August 3, 1740, the elector’s nameday, date from this time, but it would seem that Bach’s ambitions and activities in connection with the ‘ordinaire’ and ‘extraordinaire’ concerts were considerably diminished. There are no signs, however, that Bach’s interest in instrumental ensemble music slackened; if anything, it underwent a certain revival and he continued to produce chamber music steadily throughout the 1730s.

Bach withdrew from the collegium musicum in 1741. With the death of the coffee-house owner Gottfried Zimmermann (May 30, 1741) the collegium had lost its landlord and organizer, and without him it could not long continue. The collegium had made an important contribution to musical life in Leipzig for 40 years, both with and without Bach’s leadership, and even its demise was not without consequences for the future. In both its function and its membership, it served to prepare the ground for a new focal point in civic musical life, the Grosses Concert, founded on the lines of the Parisian Concert Spirituel in 1743 by a group of younger and wealthy merchants, almost all of them members of the newly founded first Masonic Lodge in Leipzig. The Grosses Concert was destined to be the immediate predecessor of the Gewandhaus concerts that began in 1781. As the successor organization to Bach’s collegium, it moved to a new venue – the inn ‘Zu den drei Schwanen’ – and started under the well-tried direction of C.G. Gerlach. Quite a number of the musicians from the collegium were carried over and it is likely that from time to time Bach presided over performances.

In August 1741 Bach went to Berlin to visit Carl Philipp Emanuel who in 1738 had been appointed court harpsichord player to Crown Prince Frederick of Prussia (later King Frederick II 'The Great’) and to reconnect with former students and friends. In the two previous years Bach had made brief journeys to Halle (early 1740) and Altenburg (September 1739; he gave a recital on the new Trost organ in the castle church). In November 1741 there was a further journey, this time to Dresden, where he visited Count von Keyserlingk. In the same year, probably in the autumn, the ‘Aria with 30 Variations’, the so-called Goldberg Variations, appeared in print. Bach’s visit to Dresden may lie behind the anecdote related by Forkel, according to which the variations were commissioned by the count as a means of ameliorating sleepless nights, but the lack of any formal dedication in the original edition suggests that the work was not composed to a commission. It is conceivable, on the other hand, that after publication the count received a copy of the work for the use of his young resident harpsichord player Johann Gottlieb Goldberg, who was a pupil of both J.S. and W.F. Bach. In his own copy (which came to light only in 1975) Bach added a series of 14 enigmatically notated canons on the bass of the Aria (bwv1087) in about 1747–8. They place a special and individual accent on the canonic writing that occupied him so intensively at that period.

On August 30, 1742, on the Kleinzschocher estate near Leipzig, a ‘Cantata burlesque’ (known as the Peasant Cantata bwv212) was performed in homage to the new lord of the manor, Carl Heinrich von Dieskau; this work is unique in Bach’s output for its folklike manner (except perhaps for the quodlibet in the Goldberg Variations). The thoroughly up-to-date characteristics of parts of the work show that Bach was not only intimately acquainted with the musical fashions of the times but also knew how to adapt elements of the younger generation’s style for his own purposes (as he also did in the third movement of the trio sonata from the Musical Offering).

Alongside this work, apparently his last secular cantata, Bach’s only vocal compositions of the 1740s were isolated sacred works (including bwv118.2 and 195.2), some new, some refashioned. There is evidence, on the other hand, that he gave numerous performances of works by other composers, some newly arranged or revised. These included a German parody of Pergolesi’s Stabat mater (Tilge, Höchster, meine Sünden bwv1083, c1746–7), a Latin parody after the Sanctus and ‘Osanna’ from J.C. Kerll’s Missa superba (Sanctus in D bwv241, c1747–8), Handel’s Brockes Passion (c1746–7), and a pasticcio Passion after C.H. Graun bwv1167 (with inserted movements bwv1088 and ‘Der Gerechte kömmt um’ bwv1149). Bach also often repeated his own earlier sacred works. Evidence does not exist to form a complete picture, but they included revised versions of the St Matthew and St John Passions; the latter was performed for the last time during Bach’s lifetime on Good Friday 1749.

The only new vocal composition of any size was the Credo and following sections of the Mass, which, when added to the Missa of 1733 (bwv232.2), produced the B minor Mass bwv232.4 – a continuation of Bach’s preoccupation with Latin figural music during the late 1730s. No specific reason for the composition of the B minor Mass, and no evidence of a projected or actual performance, has so far come to light. One of the more plausible hypotheses is that the composition of the work (which is described in C.P.E. Bach’s Nachlass as ‘the large Catholic Mass’) was connected with the consecration of the Catholic Hofkirche in Dresden, planned for the late 1740s and then postponed until after 1750 (building started in 1739). All that is known for certain is that the expansion of the 1733 Missa by the addition of a Credo, a Sanctus (1724), and the movements from ‘Osanna’ to ‘Dona nobis pacem’ and the fusing of the various sections to create a unified score (see also §14) were done in the last years of Bach’s life – more precisely, between August 1748 and October 1749.

Instrumental music, however, once again came to the fore during the 1740s. Bach had begun to sift through his older organ chorales about 1739–42, probably following completion of Clavier-Übung III. Some of the Weimar pieces were extensively reworked and gathered into a new manuscript collection (the ‘18’, bwv651–68). These revisions may have been undertaken with a view to the subsequent appearance of the chorales in print, as happened with the six chorales on movements from cantatas (the ‘Schübler Chorales’) about 1748. Apparently Bach was still engaged in work on the chorales in the last months of his life. The copying from dictation of the chorale Vor deinen Thron bwv668, later the subject of legend, was in fact probably confined to an improvement of an existing work (the chorale bwv641 from the Weimar Orgel-Büchlein).

Bach retained his interest in organ building to the last. In 1746 alone there were two important examinations and inaugurations of organs: on August 7, in Zschortau and on September 26–9 in Naumburg. Bach’s appraisal of the large Hildebrandt organ in the Wenzelskirche, Naumburg, was one of his most important. He customarily subjected instruments to the most searching examinations, both of their technical reliability and of their tone quality. He had also taken a critical interest in the pianos that Gottfried Silbermann was building during the 1730s, proposing alterations in the mechanism which Silbermann evidently adopted. At all events, Bach praised Silbermann’s later pianos and promoted their sale (a receipt for one sold to Poland, dated May 6, 1749, survives). On his visit to Potsdam in 1747 he played on a range of Silbermann pianos of the newer type which had been purchased by the Prussian court.

The visit to the court of Frederick the Great in May 1747 is one of the most notable biographical events in Bach’s otherwise unspectacular life. The invitation probably came about through Count Keyserlingk, who was then in Berlin. Bach’s encounter with Frederick began on May 7 at the palace of Potsdam during the chamber music which was a feature of every evening of court life there. Bach’s execution on the piano of a remarkable improvisation on a theme supplied by the king met with general applause. The next day Bach gave an organ recital in the Heiliggeistkirche in Potsdam, and during chamber music that evening he improvised a six-part fugue on a theme of his own. He also visited the new Berlin opera house, and possibly went to look at organs in Potsdam and Berlin. On his return to Leipzig, probably in the middle of May, he worked industriously on an ‘elaboration of the royal Prussian fugue theme’, beginning with writing down the fugue he had improvised (a three-part ricercare), which, while in Potsdam, he had announced that he would print. But he now decided on a larger project and under the title Musikalisches Opfer (‘Musical Offering’) he prepared a work in several movements dedicated to Frederick the Great; this work was printed in its entirety by the end of September (Michaelmas) 1747. The royal theme serves as the basis for all the movements (two ricercares, in three and six parts, for keyboard; a trio sonata for flute, violin, and continuo; and ten canons for various instruments).

In June 1747, after some hesitation, Bach joined the Correspondirende Societät der Musicalischen Wissenschaften founded by Lorenz Mizler. It was probably still in 1747 that he submitted, as a ‘scientific’ piece of work, his canonic composition on Vom Himmel hoch bwv769. At the same time he sent the members an offprint of the six-part canon bwv1076 from the series on the fundamental bass of the Goldberg Variations. He seems, however, to have taken no further interest in the society’s affairs as (according to C.P.E. Bach) he thought nothing of the ‘dry, mathematical stuff’ that Mizler wanted to discuss. Besides his long acquaintance with his pupil Mizler, Bach’s most likely reason for joining the society was that prominent colleagues such as Handel, Telemann, and Graun were fellow members.

The beginnings of his work on the Art of Fugue date from about 1740, or before. It is impossible to give an exact date as the original composing score is now lost, but early traces of work on the Art of Fugue can be seen in the contrapuntal studies bwv1132 from about 1738. However, what must be a first version of the work survives in an autograph fair copy containing 14 movements (12 fugues and two canons) and dating from 1742 at the latest. Thereafter Bach expanded and revised the work in readiness for printing. He himself supervised the printing to a large extent, and the process was probably largely complete by about the end of 1749 (in other words, before his son Johann Christoph Friedrich, who had helped to correct the proofs, left to join the court at Bückeburg in January 1750). But Bach was not to see the entire work (eventually comprising 14 fugues and four canons) in print; C.P.E. Bach took charge of the publication and the work appeared posthumously in spring 1751. Bach had been unable to complete the fair copy of the last movement, a quadruple fugue, and so the fugal cycle ends with an unfinished movement. The editor decided to mitigate the effect of that by adding the organ chorale Wenn wir in höchsten Nöten sein bwv668.1 at the end. The revision of this chorale, under the new heading of Vor deinen Thron tret ich hiermit bwv668.2 as copied by Bach’s student Christoph Transchel, had been the last piece of work to occupy the composer.

In his final years Bach suffered from increasingly severe trouble with his eyes, seriously restricting his ability to work and leading eventually to total blindness. He definitely composed nothing major after autumn 1749. The last known examples of his handwriting, which give an impression of increasing irregularity, clumsiness, and cramping, go up to October 1749 (parts of the score of the B minor Mass). Other documents to which he put his signature date from as late as spring 1750. The cause of the eye disease seems to have lain in untreated (and untreatable) diabetes, which may also have caused neuropathy and degenerative brain disease, evidence of which is found in the dramatic change in his handwriting in manuscripts of 1748–9. He gave a performance of the St John Passion on Good Friday 1749 without completing the revision of the work begun in about 1740. His health must have been very poor by spring 1749; otherwise, the Leipzig town council would surely not have been so tactless as to submit Gottlob Harrer, a protégé of the Dresden prime minister Count Brühl, to examination for the post of Kantor on June 8, 1749. Out of consideration for Bach the cantata performance was in a concert hall rather than one of the churches. The town chronicle reported that the authorities expected Bach’s death. Bach, however, recovered at least temporarily and was able to demonstrate his fitness by a performance of bwv29 on August 25 (town council election service). At another undated occasion in 1749, he presented the satirical dramma per musica bwv201 and could not resist mocking the Saxon prime minister by making some subtle text changes. When his grandson Johann Sebastian Altnickol (his pupil Johann Christoph Altnickol had married Elisabeth Juliane Friederica Bach) was baptized on October 6, 1749 in Naumburg Bach was unable to make the short journey to stand godfather in person.

Bach’s state of health and ability to work must have fluctuated during his last year. He appointed Johann Nathanael Bammler, a former choir prefect at the Thomasschule for whom he provided two excellent references in 1749, to deputize for him as occasion warranted. But in spite of everything Bach was not entirely inactive. In spring 1749 he is known to have corresponded with Count Johann Adam vom Questenberg, apparently about a commission or some other project. Although no details are known, this reaffirms Bach’s obviously well-established connections with some major noble patrons from the area of Bohemia (Count Sporck of Lissa and Kukus), Moravia (Count Questenberg of Jaroměřice), and Silesia (the Haugwitz family). From May 1749 to June 1750 he was engaged in a controversial correspondence about the Freiberg headmaster Biedermann. In May 1749 Biedermann had violently attacked the cultivation of music schools; Bach immediately felt himself called into battle and made some openly critical remarks. His involvement is understandable, for he must have seen parallels with the state of affairs at the Thomasschule, where the same tendency fuelled Ernesti’s reforms. Bach solicited a rejoinder on the part of Christoph Gottlieb Schröter, a member of Mizler’s society, and even Mattheson joined in, from Hamburg. Once again, the affair throws light on the situation in German schools during the early Enlightenment and Bach’s last years as Thomaskantor. The integration of academic and musical traditions, which had been an institution for centuries, was in the process of turning into an irreconcilable confrontation.

At the end of March Bach underwent an eye operation, performed by the English eye specialist John Taylor who was on a lecture tour in Germany (and was later to perform a similar operation on Handel). It was only partly successful, however, and had to be repeated during the second week of April. The second operation too was ultimately unsuccessful, and indeed Bach’s physique was considerably weakened. Yet as late as the beginning of May 1750 Johann Gottfried Müthel could go to Leipzig, stay at Bach’s house and become his last pupil. To what extent regular instruction was possible under these circumstances remains uncertain. In the next two months Bach’s health had so deteriorated that, on July 22, he had to take his last Communion at home. He died only six days later, on the evening of July 28, after a stroke. He was buried two days later at the cemetery of the Johanniskirche. It is not known what form the funeral ceremony took or what music was performed.

Bach’s wife, Anna Magdalena, who in addition to her domestic tasks was a loyal and industrious collaborator, participating in performances and copying out music, survived him by ten years. After vacating the Kantor’s apartment at the Thomasschule she moved to a new home at nearby Hainstrasse. She received a modest pension that she supplemented by income earned from a small sewing business which she ran with her unmarried daughters; she died in 1760. On his death Bach had left a modest estate consisting of securities, cash, silver vessels, instruments – including eight harpsichords, two lute-harpsichords, ten string instruments (among them a valuable Stainer violin), a lute and spinet – and other goods, officially valued at 1122 thaler and 22 groschen; this had to be divided between the widow and the nine surviving children of both marriages. Bach himself had evidently given instructions for the disposition of his musical Nachlass, which is ignored in the official valuation. According to Forkel, the eldest son Wilhelm Friedemann ‘got most of it’ (see §III, (7), 11).

10. Iconography.

The oak coffin containing Bach’s remains was exhumed in 1894: the detailed anatomical investigation by Professor Wilhelm His confirmed their identity and showed that Bach was of medium build. From a skull impression Carl Seffner, in 1898, modelled a bust, which shows an undoubted similarity with the only likeness of Bach that can be guaranteed as authentic, that of the Leipzig portraitist Elias Gottlob Haussmann. That portrait exists in two versions, one dating from 1746 (Museum für Geschichte der Stadt Leipzig; property of the Thomasschule) and one of 1748 (bequested in 2015 by W.H. Scheide to the Bach-Archiv Leipzig; see below, Fig.3). The earlier, signed ‘E.G. Haussmann pinxit 1746’, was presented to the Thomasschule in 1809 by the then Thomaskantor, August Eberhard Müller. Müller had obtained the painting from the descendants of W.F. Bach in Berlin. It is often supposed that the Thomasschule portrait is one that members of Mizler’s society were required by statute to donate to that institution, but that is highly unlikely: Bach probably did not present a portrait, at least not in the form of a painting, to the society. With the passage of time the Thomasschule picture was severely damaged and repeatedly painted over. Thorough restoration in 1912–13 and a more recent one in 2018–19 returned it more or less to its original condition, but it remains inferior to the excellently preserved replica of 1748. The latter has a reasonably secure provenance, out of C.P.E. Bach’s estate; it was owned privately for many years by the Jenke family in Silesia and then from 1938 in England, before being exhibited in public by Hans Raupach in 1950 at the Bachfest Göttingen, and subsequently acquired by the American collector William H. Scheide of Princeton, NJ.

The authenticity of an unsigned pastel portrait, painted in 1773 by Johann Philipp Bach while visiting his godfather C.P.E. Bach, is beyond doubt; it is painted in imitation of the 1748 Haussmann portrait then in Hamburg. During the 18th and 19th centuries many copies were made of the Haussmann portrait in either version, both in oils and in various types of print; an engraving (1794) by Samuel Gottlieb Kütner, an art student at the Zeichenakademie, Leipzig, along with C.P.E. Bach’s son Johann Sebastian (1748–78), was said by Emanuel himself to be ‘a fair likeness’. The nearest we can nowadays get to the composer’s true physiognomy is probably in the 1748 version of Haussmann’s portrait, wherein, as a man in his early 60s, Bach is represented as a learned musician, with a copy of the enigmatic six-part canon bwv1076 in his hand to demonstrate his status (Fig.3).

According to GerberL, probably authentic portraits that no longer survive were once owned by J.C. Kittel (from the estate of the Countess of Weissenfels) and by J.N. Forkel. A pastel from C.P.E. Bach’s collection (not the one referred to above) has not survived. A portrait of the supposedly young Bach surfaced in 2016 and was acquired by the Bach-Archiv Leipzig; its authenticity has not yet been established. The painting (in Bach-Dokumente, ix, p.39) originates from the early 18th century and shows a young man in front of a façade of organ pipes and holding a scrolled-up musical score. On its back, an old slip of paper with an inscription in an unknown hand (around 1800) reads somewhat mysteriously ‘The famous organ player and organ builder, died July 28, 1750’. All the other better-known and much reproduced portraits turn out to be misattributions. The oil by Johann Jacob Ihle, dating from about 1720 and purporting to show Bach as Kapellmeister in Cöthen, comes from the palace at Bayreuth and was identified as a ‘picture of Bach’ only in 1897. But there is no concrete support for that identification, and the portrait’s earlier provenance is obscure; it now hangs in the Bachhaus in Eisenach. The portrait by Johann Ernst Rentsch the elder (Städtisches Museum, Erfurt), allegedly representing Bach at the age of about 30, came to light only in 1907 and has no credible documentation. Many other apocryphal portraits, including the ‘portrait in old age’ discovered by Fritz Volbach in Mainz in 1903 (private collection in the USA), are of the ‘old man with a wig’ type and have nothing to do with Bach. Questionable, too, is the unsigned and undated group portrait of musicians, supposedly executed by Johann Balthasar Denner and showing Johann Sebastian Bach (with violoncello piccolo) and three of his sons (collection Ralph Kohn, London; a less well preserved replica at the Internationale Bachakademie, Stuttgart).

Open in new tab

3. Johann Sebastian Bach: portrait by Elias Gottlob Haussmann, 1748, replica of a 1746 version (Bach-Archiv Leipzig); the composer holds a copy of his six-part canon bwv1076

W. H. Scheide Library, Princeton, New Jersey / photo Willard Starks

11. Sources, repertory.

The earliest catalogue of Bach’s compositions – admittedly a very rough one – was included in the obituary that C.P.E. Bach and J.F. Agricola wrote immediately after Bach’s death but did not publish until 1754. It scarcely provides an adequate idea of the extent of Bach’s works, but it lists in the first section all the works published during Bach’s lifetime: the four parts of the Clavier-Übung; the Musical Offering; the Canonic Variations bwv769; the Schübler chorales; the Art of Fugue (not included are the Mühlhausen town council cantata bwv71, its two successor works, the Schemelli Hymnbook, and the single print of bwv1076). The second, much larger section presents a summary listing of the works in manuscript as the great majority of Bach’s compositions remained unprinted. Vocal compositions are followed by keyboard and other instrumental solo and ensemble works.

On the assumption that Bach managed to keep his music together as far as possible during his lifetime, it seems that major losses occurred only by way of the division of his legacy in 1750, when the manuscripts, especially of the vocal works, were divided between the eldest sons and Bach’s widow. Most of them reportedly went to Wilhelm Friedemann, but he, unfortunately, was the least successful at managing his inheritance; he was compelled for financial reasons to sell them off item by item, and the material is not simply scattered but for the most part lost. Carl Philipp Emanuel Bach’s and Anna Magdalena’s shares were better preserved. Bach’s widow gave her portion (the parts of the cycle of chorale cantatas) to the Thomasschule (today in the Leipzig Bach Archive) while most of C.P.E. Bach’s inheritance (published in his estate catalog of 1790) passed through Georg Poelchau’s collection into the Berlin Königliche Bibliothek (later the Preussische Staatsbibliothek and today the Staatsbibliothek zu Berlin). Only very few of the items inherited by Johann Christoph Friedrich and Johann Christian, including a printed copy of the Musical Offering and the autograph of the organ Prelude and Fugue in B minor bwv544 (signed with Johann Christian’s nickname ‘Christel’), can be traced.

The most serious losses occurred among the cantatas: altogether perhaps more than 100, including church cantatas and secular occasional works. The funeral music for Prince Leopold of Cöthen (1729) and the St Mark Passion (1731) are among large-scale vocal works of which only the texts survive. A greater proportion of the music for organ and other keyboard instruments has probably survived than that in any other category. Losses among the orchestral and chamber works are impossible to estimate but may be regarded (on the evidence of existing transcriptions, for example) as substantial.

Today more than 90% of the surviving original Bach sources are held by two German institutions, the Staatsbibliothek zu Berlin (D-B) and the Bach-Archiv Leipzig (D-LEb); the Staatsbibliothek has the much larger holdings. During the 19th century this library acquired further Bach collections, notably those from the Singakademie and the estates of Forkel, Franz Hauser, and Count Voss-Buch. The core of the Bach Archiv holdings consists of the chorale cantatas (A.M. Bach’s inheritance), supplemented by manuscripts from the Leipzig publishing firms Breitkopf and C.F. Peters and other acquisitions.

Besides the original manuscripts – the autograph scores, and parts prepared for performances under Bach’s direction – which, in their essentials, Bach kept in his library, many copies were made in the circle of his pupils, particularly of organ and harpsichord music. As many autographs of the keyboard works are lost, this strand is particularly significant for the preservation of Bach’s works. Important copies have come down through members of Bach’s family (including the Möllersche Handschrift and the Andreas-Bach-Buch, both compiled by Sebastian’s elder brother Johann Christoph), through J.G. Walther, and through Bach’s pupils Krebs and Kittel. After Bach’s death Breitkopf in Leipzig became a centre for the dissemination of his music. In Berlin a notable Bach collection was made for Princess Anna Amalia of Prussia, under the direction of Kirnberger, in which all facets of Bach’s creative output were represented (now D-B Amalien-Bibliothek). These secondary sources have to serve when autograph material is not available – relatively often with the instrumental works (e.g. a large percentage of the organ pieces; the English and French Suites, toccatas, fantasias, and fugues for harpsichord; duo and trio sonatas; concertos and orchestral works), more rarely with the vocal ones (e.g. cantatas bwv106 and 159; motets bwv227–30; and the masses bwv233 and 235).

Research into source materials, notably in conjunction with the Neue Bach-Ausgabe, has proved fruitful. The use of diplomatic research methods has allowed most of the copyists who worked for Bach – and all the important ones – to be identified: ‘Hauptkopist A’ was J. Andreas Kuhnau (b 1703); ‘Hauptkopist B’ was C.G. Meissner (1707–60); ‘Hauptkopist C’ was J. Heinrich Bach (1707–83); ‘Hauptkopist D’ was S.G. Heder (b 1713); ‘Hauptkopist E’ was J.G. Haupt (b 1714); ‘Hauptkopist F’ was J.L. Dietel (1715–73); ‘Hauptkopist G’ was R. Straube (b 1717); ‘Hauptkopist H’ was J.N. Bammler (1722–84), and Hauptkopist K (Anon N1, Anon. L107) was F.A. Kuhl (1719–1779). The majority of the other anonymous Bach scribes has also been identified by name, primarily by H.-J. Schulze, Y. Kobayashi, and P. Wollny (see bwv³, p. xxxix–xliv). Papers, inks, and binding have been evaluated for the purposes of identification and dating; but above all Bach’s own handwriting, in its various stages of development, has served as the criterion for dating. A far-reaching revision of the chronology of Bach’s works (only some 40 of the originals are dated) has been made possible, leading to a substantial revision of previous conceptions, which were based for the most part on Spitta’s work. The new chronology was established in its important details by Dürr and Dadelsen during the late 1950s. Since then it has been variously added to, modified, or confirmed. For the vocal works it is now essentially complete; sometimes it is precise to the actual day, notably for the Leipzig church cantatas. With the instrumental works the situation is more complicated, because the original manuscripts are often lost; in consequence, results have been less precise since the history of the secondary sources permits of only vague conclusions about composition dates (for example, copies originating from the circle around Krebs and J.G. Walther point to a date in the Weimar period); this makes it unlikely that any complete and exact chronology will be established for the instrumental works, though a relative one is now largely achieved.

Investigations of source material have also led to the solution of crucial questions of authenticity, particularly in connection with the early works but also affecting some of the later ones. For example, Cantata bwv15, hitherto regarded as Bach’s earliest cantata, has been identified as by Johann Ludwig Bach; similarly, cantatas bwv53, 189, and 142 have been excised from the list of his works. Some instrumental works, such as bwv835–8, 969–70, 1024, and 1036–7, have been assigned to other composers. On the other hand, an important early organ work, bwv739, has now been authenticated and its manuscript ranks as probably Bach’s earliest extant musical autograph. Completely new finds have been made (bwv1081–1128) and numerous copies by Bach of other composers’ works have come to light or have been identified stemming from his library (e.g. works by Caldara, Durante, and Gasparini); these provide additional information about his repertory and its context.

12. Background, style, influences.

Bach’s output, unparalleled in its encyclopedic character, embraces practically every musical form of his time except opera. The accepted genres were significantly added to by Bach (notably with the harpsichord concerto and chamber music with obbligato keyboard); further, he opened up new dimensions in virtually every department of creative work to which he turned, in format, density, and musical quality, and also in technical demands (works such as the St Matthew Passion and the B minor Mass were to remain unique in the history of music for a long time to come). At the same time Bach’s creative production was inextricably bound up with the external factors of his places of work and his employers, as was normal in his time. The composition dates of the various repertories thus reflect Bach’s priorities in his various professional appointments; for instance, most of the organ works were composed while he was active as an organist at Arnstadt, Mühlhausen, and Weimar, whereas most of the vocal works belong to the period of his Kantorate at Leipzig. But Bach’s production was by no means wholly dependent on the duties attaching to his office at the time. Thus during his Leipzig period he found time to produce a body of keyboard and chamber music to meet his requirements for concerts, for advertisement, for teaching, and for other purposes. And his career may be seen as a steady and logical process of development: from organist to Konzertmeister, then to Kapellmeister, and finally to Kantor and director of music – a continual expansion of the scope of his work and responsibilities. This is no matter of chance. Bach chose his appointments and chose the moment to make each move. If he was unable to accomplish what he required (as was often the case in Leipzig), he was capable of turning his attention elsewhere in pursuit of his creative aims. Bach was a surprisingly emancipated and self-confident artist for his time.

The uncertainty about the dating of Bach’s early works, with so little help in the form of source materials, makes it difficult to reconstruct and assess the beginnings of his work as a performer and composer. The Weimar Tablature brought to light two significant manuscripts in the hand of the 13- to 15-year-old Bach: copies of D. Buxtehude’s chorale fantasia Nun freut euch, lieben Christen gmein buxwv210 and J.A. Reinken’s fantasia An Wasserflüssen Babylon – two of the technically and musically most demanding organ works of the 17th century. They provide evidence of the astonishing early accomplishments of the young virtuoso who then intended to benefit from further instruction by Georg Böhm in Lüneburg. It is to be supposed that he started to compose while still under the tutelage of his elder brother in Ohrdruf, but although he took no formal lessons with an established composer, as Handel did with Zachow, it would be mistaken to call him self-taught as a composer, for the significance of his belonging to a long-standing family of professional musicians should not be underestimated. Composing was probably overshadowed by instrumental playing in Ambrosius Bach’s family; this must to some extent have applied to the young Johann Sebastian, and probably he devoted more attention to developing his skills as an instrumentalist, especially as an organist, than to composition studies. But the art of improvisation – in those days inseparably bound up with practice on the instrument – would at the very least prepare the ground for his work as a composer. This reciprocity between performing and composing is reflected in the unruly virtuoso and improvisatory elements in Bach’s early works as exemplified e.g. by the Toccata and Fugue bwv565.

As composers who influenced the young Bach, C.P.E. Bach cited (in 1775, in letters to Forkel) Froberger, Kerll, Pachelbel, Frescobaldi, Fischer, Strungk, certain French composers, Bruhns, Buxtehude, Reincken, and Böhm – almost exclusively keyboard composers; C.P.E. Bach also said that Bach formed his style through his own efforts and developed his fugal technique basically through private study and reflection. In his letter of resignation from Mühlhausen Bach himself wrote of having procured a good supply of the very best vocal compositions, suggesting that in vocal music too he was decisively stimulated by the study of other composers’ music. Bach came into personal contact with the last three of the composers named by C.P.E.; there was no question of any teacher–pupil relationship. No record survives of what works he collected at Mühlhausen, but they might have included Keiser’s St Mark Passion, a six-part mass by Peranda, and an italianate chamber concerto by Biffi, for his early autograph copies of all these survive, demonstrating the breadth of his knowledge of the repertory. As later influences, C.P.E. Bach named Fux, Caldara, Handel, Keiser, Hasse, the two Grauns, Telemann, Zelenka, and Benda. This list, though certainly less representative than the earlier one, suggests that Bach’s main interests still lay in his great contemporaries, whose music he not only heard but also studied in transcripts. With them he abandoned his one-sided attention to the organists among older composers, but his interest in the retrospective style represented by Fux and Caldara, complemented by his enthusiasm (mentioned by Birnbaum, 1737) for Palestrina and Lotti, is notable, and is borne out by tendencies in his music from the mid-1730s. Clearly he also became interested in, and ready to follow, more recent stylistic trends, particularly in respect of the music of Hasse, the Graun brothers, and Benda (for example in the ‘Christe eleison’ of what was to become the B minor Mass) and in such works as the Peasant Cantata, the Goldberg Variations, and the Musical Offering). Mizler, in an article of 1739 on Bach’s cantata style, referring to the Scheibe–Birnbaum controversy, mentioned a work (bwv1161, lost) composed ‘perfectly in accordance with the newest taste’ (‘vollkommen nach dem neuesten Geschmack eingerichtet’).

Curiously, C.P.E. Bach’s list of the masters his father had ‘loved and studied’ contains no mention of Vivaldi and the two Marcellos, or of Corelli, Torelli, and other late Baroque Italian composers. Forkel compensated for this by his emphasis on the importance of Vivaldi’s concertos, without citing any particular source to support his claim. Indeed, it was Vivaldi who exercised what was probably the most lasting and distinctive influence on Bach from about 1712–13, when a wide range of the Italian repertory became available to the Weimar court orchestra. Bach drew from Vivaldi his clear melodic contours, the sharp outlines of his outer parts, his motoric and rhythmic conciseness, his unified motivic treatment, and his clearly articulated modulation schemes. His confrontation with Vivaldi’s music in 1713–14 provoked what was certainly the strongest single development towards his own personal style. In Forkel's words, Vivaldi ‘taught him to think musically’; his musical language acquired its enduring quality and unmistakable identity through his coupling of italianisms with complex counterpoint, marked by busy interweavings of the inner voices as well as harmonic refinement. It is impossible to describe Bach’s personal style by means of simple formulae; but the process of adaptation and mutation that can be felt throughout his output seems to have taken a particularly characteristic turn at that point in 1713–14 whose principal landmarks are the Orgel-Büchlein and the first Weimar series of cantatas. His adaptation and integration of various contemporary and retrospective styles represent his systematic attempt at shaping and perfecting his personal musical language (‘unlike that of any other composer’, according to J.F. Agricola in the obituary) and expanding its structural possibilities and its expressive powers.

An essential component of Bach’s style can be seen in his combination of solid compositional craftsmanship with instrumental and vocal virtuosity. The technical demands made by his music reflect his own prowess as an instrumentalist. Bach’s own versatility – his early involvement in singing (it is not known whether he was later active as a singer), and his experience as a keyboard player, violinist, and viola player – was partly responsible for the fact that demanding technical standards became the norm for every type of composition he wrote. This led to Scheibe’s famous criticism: ‘Since he judges according to his own fingers, his pieces are extremely difficult to play; for he demands that singers and instrumentalists should be able to do with their throats and instruments whatever he can play on the keyboard. But this is impossible’. It makes no essential difference at what level these demands are made (for instance between the Inventions and the Goldberg Variations, the four-part chorale and the choral fugue); everywhere Bach’s requirements are the antithesis of conventional simplicity. Yet technical virtuosity never predominates; it becomes a functional element within the composition as a whole. Bach’s impulse towards integration is also manifested in the typically instrumental idiom in which he cast his vocal parts. He thus produced in his music for voices and instruments a homogeneous language of considerable density. Even so, he differentiated between instrumentally and (less often) vocally dominated types of writing; but even in such vocally dominated pieces as the Credo of the B minor Mass he maintained both the density and the uncompromising, yet appropriate technical standard. It is of course significant, as regards both matters of technique and the quality of his music in general, that, as far as we know, he wrote almost exclusively for himself, his own ensembles, and his own pupils, and never for a broader public (let alone a non-professional one). This partly explains why his music – unlike, say, Telemann’s or Handel’s – was disseminated within unusually narrow confines.

13. Cantatas.

Up to about two-fifths of Bach’s sacred cantatas must be considered lost; of the secular cantatas, more are lost than survive. Thus it is difficult to draw firm conclusions about the evolution of the cantata in Bach’s hands, even though the surviving repertory is considerable and roughly proportional to the number of cantatas composed at each place where he worked.

The earliest surviving cantatas, and probably Bach’s first, date from the Mühlhausen and perhaps even the Arnstadt period; they include – as the earliest of all – bwv150, 131, 106, and 196 (c1707). The best, in both form and content, are bwv106 and 71. The latter is especially sumptuous, and its appearance in print bore the young composer’s reputation far beyond the boundaries of Mühlhausen. The early vocal works belong almost without exception to the category of ‘organist’s music’, that is they are pieces composed for particular occasions, not regular cantatas for the Sundays and feast days in the church calendar. Nor do they conform to the type established as modern by Neumeister in 1701, but they rely closely on central German tradition. Their texts are mostly taken from the Bible or the chorale repertory; freely conceived poetry is rare (found only in bwv71, 106, and 150). Musically they consist of a succession of different formal types – concerto, motet, (strophic) aria, and chorale – adapted and combined to suit the composer’s purpose. Bach did not call them cantatas: as a rule he reserved that term for the solo cantata of the Italian type (like bwv211 and 212), calling his sacred cantatas ‘Concerto’, and in earlier works ‘Motetto’, sometimes ‘Dialogus’ (depending on the text) or simply ‘Music’.

Bach’s early cantatas are distinguished from their central German precursors, which must have been familiar to him from his upbringing, by his tendency to give each movement a unified structure and his development of a broad formal scheme. He found the means to unify movements that for the most part do not function as closed numbers by reducing motivic material (in the solo movements). Reacting against haphazard sequential form, with its danger of formal dissolution, he began to use strictly symmetrical sequences of movements to underpin the overall cyclic structure: for example, chorus–solos–chorus–solos–chorus (bwv106).

During Bach’s early Weimar years, organ music dominated his output; on the other hand, the letters written in 1712–13 by his pupil at Weimar, Johann Philipp Kräuter, show that Bach encouraged him to write cantatas. 1713 is the date, too, of what seems to be Bach’s first secular cantata, the Jagd-Kantate bwv208, written to a commission from the Weissenfels court. The piece shows Bach, obviously newly acquainted with the Italian style, taking up the recitative and the modern kind of aria (for preference the da capo aria), a step which had a decisive effect on the next sacred cantatas, bwv199, 21, and 63 (bwv21 was probably written in connection with his application to succeed Zachow in Halle in December 1713). With his nomination as court Konzertmeister on March 2, 1714, he started to produce cantatas on the whole regularly from the end of March onwards, in accordance with an agreement ‘to perform a piece of his own composition under his own direction, in the chapel of the royal castle, on every fourth Sunday at all seasons’. Beginning with bwv182, performed on March 25, 1714 (Feast of the Annunciation/Palm Sunday), this was an opportunity to compose a cantata cycle for each Sunday and feast day of the ecclesiastical year, albeit over a time-span of four years or more. However, as things turned out, the number of cantatas he wrote in Weimar 1714–16 amounted to little more than 20, instead of about 30 (he composed no cantatas in 1717).

The principle of the annual cycle is closely bound up with the history of the cantata from Neumeister on; the texts were mostly published in cycles, one for each Sunday and feast day in the church year. Bach, admittedly, never adhered strictly to a single poet (except in the incomplete Picander cycle of 1728–9), preferring to pick and choose. In Weimar he turned for the first time to librettos by Neumeister (bwv18 and 61) and used texts by Georg Christian Lehms (1684–1717; bwv199 and 54), but evidently preferred texts by the Weimar court poet Salomo Franck (1659–1725), the author of extremely original and profoundly felt sacred and secular poetic texts, among the best Bach set. bwv54, 21, and 199 are among cantatas dating from before 1714; regular production began with bwv182 on March 25, 1714. There followed, usually at four-week intervals, in 1714 bwv12, 172, 61, and 152; in 1715 bwv18, 54, 31, 165, 185, 163, and 132; in 1716 bwv155, 80.1, 161, 162, and 70.1. Repeat performances of bwv21, 199, 31, 165, and 185 were slotted into the cycle. Gaps are accounted for by the loss of certain cantatas and in one case by the period of a state mourning from August 11 to November 9, 1715.

Musically the works are of particular importance for the development they show in Bach’s personal style of writing for voices and instruments. The recitatives contain extensive arioso sections to begin with, but these gradually disappear (although the combinatorial element was to remain typical of Bach throughout his life); the arias become longer, in free or (more usually) strict da capo form and occasionally using more complex structures. The choruses embrace a multiplicity of formal principles, among them fugue and canon (bwv182), chaconne (12), concerto (172), motet (21), and French overture (61). Also notable are the overlapping of instrumental and vocal formal schemes (the use of Chor- and Vokaleinbau) and instrumental quotations of chorale melodies. The extraordinarily colourful instrumentation is especially characteristic: within the smallest of performing ensembles Bach tried out a great variety of combinations, for example recorder, oboe, viola d'amore, and viola da gamba in Cantata bwv152. Following the Italian ideal, his orchestral writing moved away from the French practice of five-part writing (with two violas, which predominates in the early cantatas) towards a more flexible four-part style. Instead of the harmonic weight of the middle voices in five-part writing Bach provided a rhythmically and melodically active viola part that is particularly characteristic.

In Cöthen, corresponding to Bach’s official responsibilities, only secular cantatas were composed (with the single exception of bwv1147) and those were mostly written for New Year celebrations and the prince’s birthday. Bach’s principal librettist was Christian Friedrich Hunold (‘Menantes’, 1681–1721). Among the Cöthen cantatas, many survive only as verbal texts (bwv1151–3) or are lost altogether; a substantial part of the music survives only for bwv66.1, 134.1, 173.1, 184.1, and 194.1. These pieces mostly exemplify the ‘serenata’ type of work, with succinct operatic treatment in dialogues between allegorical figures. It is not surprising that they reflect Bach’s study of the instrumental concerto of the period (in part in the solo–tutti differentiation) or that dance characteristics appear, notably in the solo movements. Bach used transverse flutes in Cantata bwv173.1, evidently for the first time (Fig.4).

Open in new tab

4. Autograph MS of Bach’s Serenata ‘Durchlauchtster Leopold’ bwv173.1, composed 1722, with the text of the sacred cantata ‘Erhöhtes Fleisch und Blut’ bwv173.2 added 1724 (D-B Mus.ms.Bach P 42, f.1r)

Staatsbibliothek zu Berlin

At Leipzig the performance of sacred cantatas on Sundays and feast days (some 60 a year) was one of Bach’s chief tasks, and during the first several years he produced a large number of new works. His vast workload meant that within the first cycle, beginning on the first Sunday after Trinity (May 30), he not only had to rely on repeat performances of earlier sacred cantatas but also had to resort to parodies of secular cantatas written at Cöthen. Nevertheless, his first cycle (1723–4) contains the following new compositions: bwv75, 76, 24, 167, 136, 105, 46, 179, 69.1, 77, 25, 119, 138, 95, 148, 48, 109, 89, 60, 90, 40, 64, 190, 153, 65, 154, 155, 73, 81, 83, 144, 181, 67, 104, 166, 86, 37, and 44; to these must be added his test works (bwv22 and 23, for Quinquagesima 1723) and bwv194.2, composed for the consecration of the new organ in Störmthal. Apart from bwv24 (Neumeister) and bwv64, 69.1, and 77, the poet or poets of this first cycle remain for the most part unknown. The use of Knauer’s Gotha cycle of 1720, which provides two texts for each Sunday and feast day, together with the fact that cantatas in two parts, or two separate cantatas, were sometimes performed (before and after the sermon) – such as bwv75, 76, 21, 24+185, 147, 186, 179+199, 70.2, 181+18, 22+23, 172+59, and 194.3+165 – indicates that Bach designed his first Leipzig cycle, in part at least, as a double cycle.

Thus in his first year at Leipzig Bach furnished himself with an astonishingly concentrated repertory, and his emphasis on the cantata genre also gave him mastery over an incomparable variety of forms, free from any schematicism. Three favourite groundplans are: biblical text–recitative–aria–recitative–aria–chorale (bwv46, 105, 136 etc.); biblical text–recitative–chorale–aria–recitative–aria–chorale (bwv40, 48, 64 etc.); and biblical text–aria–chorale–recitative–aria–chorale (bwv86, 144, 166 etc.). A constant feature, characteristic of the Leipzig cantatas as a whole, is the framework, comprising an introductory choral movement in the grand style (solo pieces appear rarely at the start) and closing four-part chorale, simple but expressive. Compared with the Weimar cantatas, the orchestral forces are larger. From bwv75 onwards the brass (mainly trumpets and horns) are more strongly deployed, the flute is brought into play increasingly after 1724, and the oboe d’amore (from bwv75) and oboe da caccia (from bwv167) are introduced as new instruments, as are the violino piccolo and violoncello piccolo at a later date. Instrumental virtuosity is heightened, and the melismatic quality of the vocal writing is further developed. The ‘prelude and fugue’ type of movement is frequently used for the introductory chorus (as in bwv46).

The second cycle, dating from 1724–5, consists mainly of a series of freshly composed chorale cantatas (i.e. cantatas of which both text and music are based on hymns): bwv20, 2, 7, 135, 10, 93, 107, 178, 94, 101, 113, 33, 78, 99, 8, 130, 114, 96, 5, 180, 38, 115, 139, 26, 116, 62, 91, 121, 133, 122, 41, 123, 3, 111, 92, 125, 126, 127, and 1. From Easter 1725 this series was continued at first with cantatas of the traditional kind, that is with texts related to the prescribed scriptural readings for the day (bwv249.3, 6, 42, and 85), and then with nine cantatas to texts by Mariane von Ziegler (1695–1760): bwv103, 108, 87, 128, 183, 74, 68, 175, and 176, in all of which there is a tendency to use forms closer to those of the first cycle. 1724–5 was not only the most productive year for cantatas, as far as is known from the surviving works at least; it also, with the chorale cantata, saw the beginnings of a type that perhaps represents Bach’s most important contribution to the history of the genre. What is particularly striking is his endeavour to lay out the introductory movements as large-scale cantus firmus compositions, each adhering to a different structural principle. Cantata bwv20, and with it the second cycle, opens with a chorale movement for chorus in the form of a French overture which it is possible to regard as a kind of programmatic statement, whereas the opening chorus of bwv2 takes the retrospective form of a chorale motet. By this means Bach marked out a broad framework, in terms of both musical style and compositional technique, to indicate the conceptual range of the cycle he was starting. Cohesion between the movements within each cantata is guaranteed, at least from the textual point of view, by their relationship to the fundamental chorale (with chorale paraphrases for the solo pieces, as opposed to the chorale-text-only procedure in bwv4); often it is further emphasized by references to the cantus firmus and by the use of various ways of intermingling cantus firmus and free material. The author of the texts for the chorale cantatas is not known – Pastor Christian Weiss of the Thomaskirche, who used to preach chorale sermons, is a possibility.

With the third cycle, from 1725–7, the continuous, weekly production of cantatas ends, or so the sources indicate. It appears, however, from a surviving printed textbook of 1725 covering the third to the sixth Sunday after Trinity, that this cycle must have suffered substantial losses. When his production was actually interrupted Bach usually filled the gaps with works by other composers, including no fewer than 18 cantatas by his cousin Johann Ludwig Bach of Meiningen. The cantatas of the third cycle offer no major innovations in the way of musical structure, but they notably include solo (bwv52, 84, 35 etc.) and dialogue cantatas (bwv58, 32, 49 etc.), as well as large-scale works in two parts. There is an absence of overall formal integrity in the planning of this cycle, but Bach reveals a wide variety of ambitions and intentions, among them completing the cycle of chorale cantatas with further works of that type (bwv137), using new texts by Christoph Birkmann (bwv49, 52, 55, 56, 58, 82, 84, 98, and 169), reverting to older texts by Neumeister (bwv28), Franck (bwv72), Lehms (bwv110, 57, 151, 16, 32, 13, 170, and 35), or from a Rudolstadt textbook (bwv17, 39, 43, 45, 88, 102, and 187), and experimenting with the use of complementary texts from the Old and New Testaments (the former in the opening movement, the latter in a central one: Rudolstadt texts). One remarkable trait of the cycle is the frequent introduction of older instrumental movements, pre-eminently as sinfonias but sometimes also with choral participation (the reconstruction of the first movement of the Orchestral Suite bwv1069 to open bwv110 is an example of this). A remarkable innovation in summer 1727 was the appearance of obbligato organ parts (bwv34, 146, 169, 49, and 188), found in both sinfonias (recycling instrumental concertos) and arias (Fig. 5).

The third cycle was followed by the 1728–9 cycle on texts by Picander, which has been transmitted in a fragmentary state (1728: bwv149, 188, and 197.1; 1729: bwv171, 156, 159, 145, and 174). Yet, that Bach really did set the whole of Picander’s Cantaten auf die Sonn- und Fest-Tage durch das gantze Jahr (Leipzig, 1728) as his fourth cycle cannot be accepted without reservation. The poet must have been expressing something more than a pious hope when he wrote in the preface ‘that any lack of poetic charm may perhaps be compensated for by the gracefulness of the incomparable Herr Kapellmeister Bach and these songs [Lieder] may be performed in the principal churches of prayerful Leipzig’. Bach apparently assigned some texts from this cycle to his students; e.g. in 1732 C.P.E. Bach set the text for Septuagesimae Sunday ‘Ich bin vergnügt mit meinem Stande’ (br-cpeb: F30). One of the characteristics of Picander texts is the frequent interpolation of chorale verses in the free poetry, creating attractive opportunities for mingling choruses and arias, which were not wasted on Bach (see bwv156 and 159, or the first movement of the St Matthew Passion). The cantatas written after 1729 offer nothing essentially new in formal terms, as far as can be determined from those that survive, but they show signs of a later style, manifested (in bwv30.2, 36.5, and 195.3 for example) above all in a more refined shaping of the accompanied recitative and a more integral, polyphonic treatment of the final chorale (entailing some modification of the cantus firmus). Some of the later cantatas (bwv117, 192, 112, 177, 97, and 100) show an interesting modification of the chorale type: they relinquish freely composed texts but (unlike the older per omnes versus type represented by bwv4) set the central movements as recitatives and arias.

It is impossible to reconstruct a fifth cycle worthy of the name from the surviving works (not even given the large number of unattributed four-part chorales: see §III, (7), 15). The summary worklist of the obituary was somehow based on a shelf-reading in Bach’s study, but it is impossible to rebuild the order in which he kept his music shelves. Hence it remains unknown whether he kept the Weimar cantatas separate from the Leipzig cantata scores or if the 1723–4 cycle was counted as a double cycle (see above) – in short, the fifth cycle remains a mystery. Yet the mention in the obituary of ‘five cycles of church pieces, for every Sunday and holy day’ is just a tantalizing hint of how much has been lost.

Besides the cantatas composed in connection with the church year, Bach also wrote sacred cantatas for specific occasions, like changes of town council, weddings, funerals, the bicentenary of the Augsburg Confession (1730), and inaugurations of organs; in style these are essentially indistinguishable from the other works. The body of cantatas, for all its variety, has an unusually self-contained character, maintained above all by its consistently high musical quality and its unfailing expressive profundity. The distinctive expressive power of Bach’s musical language did not merely evolve in the cantatas, in many essential respects, but also finds its most characteristic representation in them. His expressive urge, as seen in individual arias and choruses, was not confined to single words as the primary bearers of expression but was geared to movements and formal sections as a whole, in keeping with Baroque formal models (like the ABA of the da capo aria). Only within the context of a movement’s structural and expressive unity did he regard the special treatment of single words as possible or meaningful. Among the tools of Bach’s craft, the traditions of musica poetica and musical rhetoric (the theory of musical figures) must certainly be reckoned. They were deeply rooted in him. Yet to reduce Bach’s intentions to their rhetorical and figural components, or even to emphasize those components, would be to diminish their true breadth. Over and above this objective of expressive unity, Bach was always primarily concerned with the contrapuntal organization of melodic-rhythmic and harmonic textures to establish coherence. That is a principal reason why his cantata movements lend themselves so readily to parody. The technical prerequisites for producing a parody work – which Bach did so often – are metrical similarity and expressive affinity; the most essential requirement, however, is self-sufficiency of the musical substance, and its flexibility leaves considerable scope for the musical interpretation of a new text.

During his early Leipzig years Bach wrote only isolated secular cantatas, but these became more frequent as time passed. They were produced for various occasions: university ceremonies (bwv36.1, 198, 205, and 207), celebrations at the Thomasschule (bwv36.1 and 1162), festivities in the houses of noblemen and prominent citizens (bwv202, 216, 210, 249.1, 30.1, 210.1, and 212), and commissions from court (bwv249.1 and 36.1). Most of his large-scale congratulatory and homage cantatas written for the electoral house of Saxony were produced at the collegium musicum. A favourite format was the operatic dramma per musica, with a simple plot suited to the specific nature of the occasion being celebrated (bwv213, 206, 214, 207.1, and 215). The more lyrical cantatas such as bwv204, or the two Italian works, bwv203 and 209, would certainly have been performed at the collegium musicum. The Coffee and Peasant Cantatas (bwv211 and 212), to some extent tinged with folk style, are distinguished by their lifelike and humorous characterization. The librettist of most of the works of 1725–42 was the versatile Picander, the only other important poet for Bach’s cantatas during this period being Johann Christoph Gottsched (1700–66), the influential Leipzig professor of rhetoric (bwv198, 1161, and 1163). There is concrete evidence of about ten cantatas for the Cöthen court and some 50 secular cantatas composed during the Leipzig years, but in most cases only the texts survive. Their occasional nature is the main reason why so many have been lost: few could have been given a second performance, and then only after alterations to the text. Bach was of course aware that their best chance of survival lay in parody, and he took such opportunities as occurred to save the music, as in the case of the Christmas Oratorio (see §III, (7), 14).

14. Oratorios, Passions, Latin works.

The three works that Bach called ‘oratorios’ fall within a relatively short period: the Christmas Oratorio of 1734–5 as well as the Easter Oratorio and Ascension Oratorio planned for 1735 but completed in 1738. They were planned as a festive cycle for the major jubilant feasts of the ecclesiasticval year. The librettists are not known for certain, but Picander (librettist of the Easter Oratorio) is the principal candidate. The place for Bach’s oratorios in the Lutheran liturgy was the same as that for the cantata; the only difference between the oratorio and cantata texts is that the former have a self-contained ‘plot’ or take the form of narration with dialogue. This conforms with the history of the genre, although Bach held the tendency to formal expansiveness firmly in check, in comparison with standard Italian practice. In the Christmas Oratorio, especially, the normal character of a single self-contained work is contradicted by its being split into sections for six different services between Christmas Day and Epiphany, and this is further emphasized by Bach in his use of different performing forces for the sections (although these are based on an underlying general scheme, and are grouped round six scenes from the Bible, with certain divergences from the allocation of lessons to be read at the various services). The unusual conception of an oratorio performed over several days is reminiscent of the Lübeck Abendmusiken, and the Christmas Oratorio obviously belongs to the oratorio tradition established by Buxtehude. All three of Bach’s oratorios are essentially based on parodies of secular cantatas whose music, initially associated with a particular occasion, could reasonably be re-used in this way (the Christmas Oratorio from bwv213, 214, and 215 among other works; the Easter Oratorio by a reworking of parts of bwv249.1; the Ascension Oratorio above all from bwv1162). However, there is so much that is new and individual in the Christmas Oratorio, especially in the biblical choruses and the chorales, and in the Ascension Oratorio, that the works are in no sense subordinate to their originals. The pervasive use of texts from the Gospels, moreover, gives the works a special status, linking them to the Protestant historia and thus ultimately to the Passion.

Of the five Passions mentioned in the necrology two survive (St Matthew and St John), for one the text survives (St Mark), and the other two are lost. Judging from the source it seems probable that the anonymous St Luke Passion – which is certainly not by Bach – was included among his works in error because the score, dating from about 1730, was copied largely in his hand and contained additions by him. This means that only one Passion remains to be accounted for, but a reasonable solution to the problem might be that the second version of the St. John Passion bwv245.2 was kept in its own score. Curiously enough, Hilgenfeldt (1850) mentioned a Passion by Bach dating from 1717, giving no indication of the source of his information, and Bach gave a guest recital at the Gotha court during the Passion period in 1717, making it conceivable that he put on a Passion performance while the post of Kapellmeister was vacant. However, it is more likely that he performed a work by another composer. Also, he performed Keiser’s St Mark Passion in Weimar in about 1713, so his interest in the genre is established for the period.

The three known works represent the same type of oratorio Passion, in the tradition of the historia, in which the biblical text is retained as a whole (with ‘parts’ for soloists – Evangelist, Jesus, Pilate etc. – and the turba choruses for disciples, high priests etc.), and is interrupted by contemplative, so-called ‘madrigal’ pieces set to freely composed verse, as well as by chorales. A special feature of Bach’s Passions is the unusual frequency of the chorales, which are set in simple yet extremely expressive four-part writing. The text of the St John Passion of 1724, Bach’s first large-scale vocal work for Leipzig, is not a unified piece of work. The freely composed parts rely heavily on the famous Passion poem by Barhold Heinrich Brockes (Der für die Sünde der Welt gemarterte und sterbende Jesus, 1712) and on texts by Christian Heinrich Postel (c1700) and Christian Weise (1675); besides this, the Evangelist’s part contains interpolations from St Matthew’s Gospel. Unlike any other of Bach’s large-scale works, the St John Passion underwent substantial changes of every kind in the course of its various performances. For the second performance, in 1725, Bach produced a much altered version adapted conceptually to the cycle of chorale cantatas (see §III, (7), 13) and by the incorporation of movements based on texts by C. Birkmann, some also using a cantus firmus. In a third version (probably of 1730) the interpolations from St Matthew were cut and a new aria and sinfonia added (both lost). Finally a fourth version of 1749 saw the work restored to something much closer to its original form; besides some changes to the text, for his last performance of a Passion in Leipzig Bach greatly enlarged the performing forces (by a part for bassono grosso among other things). It seems that Bach began a thorough-going revision of the work in about 1738, but for some reason abandoned the process halfway through movement 10 and did not resume it; furthermore the alterations he made at that time were not adopted in the 1749 performance. For all the modifications made over the 25-year period, the setting of the biblical Passion text remained the work’s constant centre, around which the madrigalian movements in particular were fitted in various ways like different settings for a gemstone. Bach skilfully exploited the network of internal textual correspondences which is unique to St John’s Gospel, and convincingly translated it into an ‘architectural’ structure.

The history of the St Matthew Passion, with its double chorus, is less complicated, though not entirely straightforward. In this case the date of the first performance seems to be established (the Thomaskirche, Good Friday 1727), but some details of that occasion and of the repeat performance two years later remain unclear because of lacunae in the source material (version bwv244.1). Furthermore, some ten movements from the St Matthew Passion were incorporated into the Cöthen funeral music of 1729 (bwv1143), and the consequences of that for the repeat of the Passion in the same year are not known. On the whole the St Matthew Passion is a considerably more unified piece than the St John, for which the primary reason is its use of Picander’s text. Its greater textual and musical scale allows more space for the arias and ‘madrigal’ pieces in which the coupling of arioso with aria is an especially characteristic feature. Another special feature is the way the strings provide an accompanying ‘halo’ in Jesus’s recitatives. The pervading cyclical formation of the work (from the interrelating of the chorales, tonal organization, and paired movements) is in some respects even more pronounced than in the St John Passion, while it lacks the earlier work’s ‘architectural’ centre. After 1729 the St Matthew Passion had at least two more performances under Bach’s direction. In 1736 he made some important changes, chief among them emphasizing the separation of the two choruses and orchestras by division of the continuo, exchanging the simple chorale at the end of part i for ‘O Mensch bewein’ and replacing the lute in ‘Komm süsses Kreuz’ with bass viol. The additional alterations of about 1742 and 1746 were mainly a matter of meeting practical performing conditions.

In its main sections, that is in the ‘madrigal’ pieces, the St Mark Passion of 1731 was a parody work whose main sources are the Trauer Ode (bwv198) and the Cöthen funeral music (bwv1143). While only the text survives, the musical design can in part be deduced from these models, although they scarcely permit it to be reconstructed satisfactorily. The Bach literature includes discussion of parody relationships which go further than this, but they seem to raise more questions than they answer. The most plausible suggestion, made by Smend (1940–48), is that some of the exceptionally large number of chorales in the St Mark Passion may have survived in the collections of Bach’s four-part chorales, and indeed the chorale collection by Bach’s student Johann Ludwig Dietel from about 1735 (D-LEb: Ms. R 18) contains a group of chorales from the St Mark Passion. A more recently surfaced text booklet for a performance of the St. Mark in 1744 in the Thomaskirche indicates that this Passion, too, was expanded and revised at some point after its first performance.

In Bach’s time Latin polyphonic music was still often used in ordinary Lutheran Sunday worship, particularly, in Leipzig, at important church feasts. Further, the concerted Magnificat continued to hold its place in Vespers on high holidays. Bach had been interested in Latin polyphonic music at least since his Weimar period, as his copies of pieces by other composers demonstrate (Peranda, Durante, Pez, Wilderer, Bassani, Caldara, Lotti, Gasparini, Palestrina etc.; catalogue in Wolff, 1968 and Beisswenger, 1992). He also wrote insertions in this style for other composers’ works and made some arrangements (Sanctus bwv241; Credo intonation for a mass by Bassani bwv1081; ‘Suscepit Israel’ for a Magnificat by Caldara bwv1082). His earliest surviving work of this type is probably the Kyrie bwv233.1 on the cantus firmus ‘Christe, du Lamm Gottes’. Then in his first year at Leipzig came the five-part Magnificat, first the E♭ version with four inserted Christmas pieces (bwv243.1), revised in D major in 1733, without the Christmas pieces, for use on any major feast day (bwv243.2). Among the various Sanctus settings attributed to Bach, apart from bwv232.1, only bwv237 and 238 (both 1723) are original compositions. The four short masses (bwv233–6), mostly parody works based on cantata movements, date from about 1738. In the careful selection of models and the subsequent reworking of the musical material, these works, together with the B minor Mass, amount to a valuable anthology of Bach’s vocal writing in music of outstandingly high quality. The transposition of German cantata movements into mass settings did more than replace German words, contingent on the time and occasion of their writing, with the timelessness of the Latin (and Greek) texts; it also removed the limitations imposed on the cantatas by their place in the annual church cycle and gave them a more general validity. The longer-term outcome of this was seen soon after 1750, when specifically the Latin sacred music was hailed by connoisseurs like Marpurg, Kirnberger, Hiller, and even the south German Prince-Abbot Gerbert as a particularly important sector of Bach’s music.

Open in new tab

5. Autograph MS of the Sinfonia (arranged from the Prelude of the Partita for violin bwv1006) from Bach’s cantata ‘Wir danken dir, Gott’ bwv29, 1731 (D-B Mus.ms.Bach P 166, f.1r)

Staatsbibliothek zu Berlin

Bach’s masterpiece in this genre is of course the work known – though not conceived as a unity – as the B minor Mass (bwv232.4). Its genesis stretched over more than two decades. Bach’s aim seems originally to have been to bring together a collection of exemplary large-scale mass movements rather than to create a single, cyclical work on an unprecedented scale. In assembling the whole score in 1748–9, however, the composer undoubtedly had the intention of making it a comprehensive work of consistent quality. The oldest section is the Sanctus of 1724 (bwv232.1). The Kyrie and Gloria come from the 1733 Missa (bwv232.2) dedicated to the Dresden court, while the Credo or ‘Symbolum Nicenum’ was composed only during Bach’s last years. In many respects the two main sections, Kyrie-Gloria and Symbolum Nicenum, represent Bach’s ideals not of Latin polyphonic music alone but of vocal music altogether: in their stylistic multiplicity (the contrast of deliberately archaic and modern styles; the experimentation with the widest variety of instrumental and vocal techniques); their abandonment of the da capo aria and the recitative; and in their formal perfection. The 1733 Missa (reminiscent of the Magnificat in its five-part writing) emerges as a completely integrated, unified whole, typified by the inner logic of the tonal organization (B minor–D–F♯ minor–D–A–D–G–B minor–D) and the disposition of the vocal and instrumental solos. The Credo in its final form is a particularly good example of Bach’s many-layered and symmetrical layout (Table 1). The Missa and the Credo have a series of parody originals (including movements from Cantatas bwv29, 46, 171, 12, and 120); in the latter the ‘Credo’, ‘Et incarnatus’, and ‘Confiteor’ seem to be the only original compositions.

Open in new tab

Table 1

An earlier version of ‘Credo in unum Deum’ in G (bwv232.3) exists, dating from the early 1740s, while the ‘Et incarnatus’ chorus may be the last vocal composition that Bach completed. However, Bach’s reworking of earlier material went much further than usual. In ‘Agnus Dei’, in particular, nearly half the movement was completely revised, using new thematic material. When the entire work was nearly finished Bach revised it once more, probably in 1749, adding ‘Et incarnatus’ (the words of which he had originally set as part of the aria ‘Et in unum Dominum’). The music of the new ‘Et incarnatus’ chorus is reminiscent of a movement in Pergolesi’s Stabat mater, and in its combination of unorthodox polyphony and musically expressive gesture points the way forward to a new stylistic sensibility. It is all the more astonishing that Bach successfully followed it with the earliest music in the mass, the ‘Crucifixus’ (from the second movement of Cantata bwv12) – though he did bring this up to date with a more empfindsam style of continuo and more subtle instrumentation of the upper parts.

It was obviously not by chance that Bach turned in his old age to the mass genre. With its centuries-old tradition, by comparison with such modern genres as the cantata and oratorio, the setting of the mass had a natural affinity to the historical and theoretical dimensions of Bach's musical thinking, which also bore fruit in the monothematic instrumental works of his last years.

15. Motets, chorales, songs.

In Bach’s time motets were sung as introits for services and on certain special occasions. The tradition established at Leipzig was to select introit motets from the Florilegium Portense (1603), a classical repertory from the 16th century compiled by Erhard Bodenschatz. For this reason, Bach wrote motets only for special occasions, probably only for burial services, although in only one case, Der Geist hilft (for the funeral of the Thomasschule headmaster Ernesti in 1729), is there documentary evidence of this. Bach’s motet texts, following the tradition, are based on biblical quotations and chorales; freely composed poetry is used in only one case, and even this is hymnbook poetry (Komm, Jesu, komm, Paul Thymich). On the occasions for which the motets were composed, Bach normally had more than the school choristers at his disposal; he was thus able to use between five- and eight-part writing, as he did in six pieces (bwv225–9 and 1165). In line with normal central German practice since the 17th century, it was a rule in the performance of motets at Leipzig, including those from Florilegium Portense, that a continuo part should be included – to be precise, organ, harpsichord (in Leipzig the so-called motet harpsichord), and lute, with violone, cello, and bassoon. In this way the bass of a vocal (choral or polychoral) movement was supported by a larger or smaller continuo depending on the circumstances, in the manner of a basso seguente. Colla parte accompaniment was required only occasionally. The performing parts that have survived for Der Geist hilft, with strings (first chorus) and reed instruments (second chorus) doubling the voices, must be connected with the exceptional nature of the occasion and cannot necessarily be taken as applicable to the other motets; similar special cases, with partly obbligato instruments, are bwv118, O Jesu Christ (both versions) and Der Gerechte kömmt um (bwv1149).

Bach’s use of double chorus and his exposition of forms of chorale treatment link the motets with the central German tradition in which he had grown up. That it was part of his direct family inheritance is illustrated by the fact, which can scarcely be coincidental, that motets are particularly well represented in the Alt-Bachisches Archiv. Bach’s earliest motet, Ich lasse dich nicht bwv1165, long attributed to Johann Christoph Bach of Eisenach, adheres extremely closely to Thuringian models. Composed by 1712 at the latest, the work’s foundations in the tradition are typified by the highlighting of upper parts and the largely homophonic conception of the first section, and by the interweaving of a chorale tune in large note values in the second; by contrast, the harmonic intensity of the work (in F minor) and the unified, almost rondo-like, thematic construction of its first section are innovatory. Among later works, Bach’s debt to the tradition is best illustrated by the closing section of Fürchte dich nicht, in its combination of cantus firmus (‘Warum sollt ich mich denn grämen’) and freely imitative writing, and the opening section of Komm, Jesu, komm, with its chordal writing for double chorus. As a whole, the style of bwv118 too is retrospective, with its archaic instrumentation and its homophonic choral writing.

By contrast, most movements in the motets have a markedly polyphonic vocal manner, dominated by instrumental style and showing unifying motivic work. Another characteristic is the clear formal articulation, with multi-movement works demonstrating different kinds of treatment. Thus Jesu, meine Freude, the longest work of this kind, in 11 movements, is the most strictly (that is, symmetrically) conceived: the opening and closing movements are identical, the second to fifth correspond to the seventh and eighth, and the central sixth movement is a fugue. Der Geist hilft begins with a concerto-like movement, followed by a double fugue and a simple chorale setting. The form of the instrumental concerto (fast–slow–fast) is used in Singet dem Herrn. Precise dating is possible only in the case of Der Geist hilft (October 24, 1729). While Fürchte dich nicht may – like bwv1165 – have originated from the Weimar period the other motets appear to date from the Leipzig years. This is certain in the case of O Jesu Christ bwv118.1 (c1736–7): its instrumentation was revised for a repeat performance in about 1746–7, with strings, oboes, bassoons, and horns (bwv118.2); the original had only two litui, cornets, and three trombones. The authenticity of Lobet den Herrn has been questioned, probably groundlessly, but the paucity of material that would permit comparisons weakens the arguments on either side.

Bach’s composition of chorales is most closely associated with his production of cantatas. Four-part chorale style, or stylus simplex, was normal for his closing movements, particularly in the Leipzig cantatas; it also often occurred at the ends of subsections in the Passions and oratorios. Bach’s chorale writing is characterized by the ‘speaking’ quality of the part-writing and the harmonies – meaning that they aim to be a direct interpretation of the text. In its pervasive counterpoint and its expressiveness, Bach’s harmonic style stands out from that of his contemporaries, who preferred plain homophonic textures in their chorales. This simpler approach, found in the chorales of such as Graupner or Telemann, with movement mostly in minims, was well suited to congregational singing, but Bach took no account of that in his chorales, which are deliberately more artistic, rhythmically often more lively (written in crotchets), and frequently bolder in their harmonies. The first four-part chorale settings are in the Weimar cantatas (the last movement of bwv12, performed on April 22, 1714, is among the earliest examples), and Bach’s stylistic development in this type of composition reached a final stage 30 years later in the chorales of the Christmas Oratorio, with their elegantly mobile bass lines and their polyphonic refinement of the inner voices. His training as an organist probably contributed to the personal stamp of his style; organ settings such as bwv706 display similar stylistic traits. Chorales such as bwv371, conceived with accompanying orchestral forces in mind, act furthermore as reminders that chorales were Bach’s favourite medium of instruction. C.P.E. Bach wrote in 1775: ‘His pupils had to begin by learning four-part thoroughbass. After that he went on with them to chorales; first he used to write the bass himself, then they had to invent the alto and tenor for themselves … this way of leading up to chorales is indisputably the best way of learning composition, including harmony’.

The posthumously published collections (Birnstiel, 2 vols., 1765, 1769; Breitkopf, 4 vols., 1784–7) contain almost all the chorales known from Bach’s vocal works, some under different titles. The Breitkopf edition, prepared by C.P.E. Bach and Kirnberger, contains 371 chorales, among them more than 100 not found in the extant vocal works. This possibly provides an important pointer to lost vocal music, and though extremely difficult to follow up it has borne some fruits, as in the reconstruction of the St Mark Passion or the Picander cycle. It is worth remarking that the number of excess chorales, that is those that cannot be assigned to extant works, more or less corresponds to the number thought to exist in the lost cantatas and Passions.

Under the generic heading of ‘sacred songs’ come the 69 melodies with figured bass in G.C. Schemelli’s Musicalisches Gesang-Buch (1736). According to the foreword, Bach edited the figured bass for some of the melodies, while others were entirely new compositions by him. Three are demonstrably his (bwv452, 500, and 505); of the rest at least seven pieces for two voices and ten ‘improved’ continuo parts can be associated with him. He seems to have been only peripherally occupied with the composition of songs and strophic arias, for which he took texts from religious poetry of the 17th and 18th centuries: that, at least, is the inference to be drawn from the limited surviving repertory, for which the only source is the second Clavier-Büchlein for Anna Magdalena Bach (1725) containing bwv511–14 and 516 – works which probably have a direct association with the Schemelli Gesangbuch. Comparison of bwv512 with 315, and of bwv452 with 299, draws attention to the conceptual association between the composition of chorales for two and for four voices. The collection of four-part chorales which Bach’s pupil J.L. Dietel extracted from his teacher’s works (Leipzig, c1735), like the Schemelli Gesangbuch (1736), indicates that Bach was working on chorales rather intensively and systematically at the time, perhaps with a view to a more compendious publication.

Only exceptionally did Bach compose secular songs. A quodlibet for four voices and continuo (bwv542), surviving only in fragmentary form, is unique among his vocal works. It was probably composed for a wedding in Erfurt, at the latest by mid-1708. With its admixture of various melodies and humorous words, the piece forms a link with the musical games played, so tradition relates, when the Bach family got together (see §III, (7), 1 above). A particular rarity, from a later period when he was settled in the university town of Leipzig, is the song addressing a pipe of tobacco (bwv515).

16. Organ music.

The obituary written immediately after Bach’s death and published in 1754 contains the following statement: ‘For as long as there is nought to confute us other than the mere possibility of the existence of better organists and keyboard players, we cannot be reproached if we are bold enough to persist in the claim that our Bach was the most prodigious organist and keyboard player that there has ever been. It may be that this or that famous man has accomplished much in polyphony on these instruments but was he for that reason as expert – with hands and feet together – as Bach was? Whosoever had the pleasure of hearing him and others, being not otherwise disposed by prejudice, will agree that this doubt is not unfounded. And whosoever looks at Bach’s pieces for the organ and the keyboard, which he himself, as is universally known, performed with the greatest perfection, will likewise have nothing to say in contradiction of the above statement’. The claim illustrates the well-nigh legendary reputation that Bach enjoyed in his lifetime. His fame had already spread beyond the confines of central Germany by 1717, when he challenged the French virtuoso Louis Marchand to a competition at the court of Dresden and won by default when the Frenchman took flight. ‘It would be wrong to conclude from this defeat of Marchand in Dresden that he must have been a poor musician. Did not as great a one as Handel avoid every opportunity of confronting the late Bach … or of getting involved with him?’ (Marpurg).

Keyboard music as a whole occupies a crucial position in Bach’s life in many respects, but this is even more true of the works for harpsichord than of those for organ. No other genre occupied Bach so consistently and intensively from the beginning of his career to the end. His life as a professional musician began with learning to play on a keyboard, above all in Ohrdruf in 1695–1700 under the tuition of his elder brother Johann Christoph, and his study of keyboard music by the best composers of the 17th century laid the most important foundations of his training as a composer. The compositions for harpsichord, in particular, provide the opportunity to assess Bach’s development at each stage of his creative life.

Bach was bolder than any of his contemporaries: from the first he set no limits to his keyboard skills, and accepted no restrictions to his horizons – from the breadth of the foundations of his style to the comprehensive range of genres in which he composed. The stylistic basis was laid in his youth, and it was undoubtedly important that growing up in the central German environment of his time gave him the opportunity to learn about different stylistic tendencies side by side, without any bias towards one rather than another. As a result his models came from a highly diverse repertory. The north German school, including such masters as Buxtehude, Reincken, Bruhns, Lübeck, and Böhm, were ranged alongside central German composers such as Pachelbel’s circle and older pupils (J.H. Buttstedt, for example, or A.N. Vetter) and Witt, Krieger, Kuhnau, and Zachow, as well as their southern German colleagues J.J. Froberger, J.C. Kerll, and J.C.F. Fischer. Italians such as Frescobaldi and Battiferri confronted Frenchmen such as Lully, Marais, Grigny, and Raison. Many of these names are to be found in the large manuscript collections (the so-called Andreas-Bach-Buch and Möllersche Handschrift) copied by the Ohrdruf J.C. Bach. They give a clear picture of the repertory that his younger brother grew up with, and which showed him – like the young Handel, learning his craft in a similar environment – ‘the manifold ways of writing and composing of various races, together with each single composer’s strengths and weaknesses’. No comparable sphere of influence served to challenge this broadly based group of musicians and exemplars later in Bach’s life. There were, of course, individuals who had an effect on him, such as Vivaldi after 1710, or probably Couperin, or his exact contemporary Handel, but no group of musicians of a comparable range or variety.

Bach’s dedication to every keyboard genre and form appears equally boundless. The range remains constant throughout his career, from the earliest to the last compositions. All the major types are represented: the freely improvisatory (prelude, toccata, fantasia), the imitative and strict (fugue, fantasia, ricercar, canzona, capriccio, invention), the combinatory (multi-part preludes, prelude and fugue), and multi-movement forms (sonata, suite or partita, overture or sinfonia, chaconne or passacaglia, pastorale, concerto, and variations); and then there are the various types and forms of chorale arrangement.

Unlike the vocal music and the chamber and orchestral works, Bach’s keyboard output covers his entire creative life. There are quite lengthy periods of heightened activity – organ music before 1717, harpsichord music after that date. As a whole, however, Bach seems to have cultivated the two genres alongside each other. It is thus the more surprising that, right from the beginning, consistently, and in defiance of inherited 17th-century tradition, he abandoned the conventional community of repertory between organ and harpsichord, choosing to write specifically for the one or the other. The uncompromising use of obbligato pedals, in particular, is a distinguishing mark of Bach’s organ style. Only exceptionally (for example in the chorale partitas and the small chorale arrangements from the third part of the Clavier-Übung) do the performing possibilities coincide so that organ and harpsichord become truly interchangeable.

Since most of Bach’s keyboard works from the pre-Leipzig years survive in copies (generally made in the circle of Bach’s pupils) rather than in autograph scores, it is not possible to establish a precise chronology. Even a relative one is possible only in general terms, with considerations of style and authenticity holding the balance. In the earliest works the influence of Bach’s models is pronounced. Pachelbel had taught Johann Christoph Bach, and the master’s influence extended to the younger brother, most visibly and prevalently in the earliest of his extant compositions. The chorales in the so-called Neumeister collection, which came to light in 1984 (bwv1090–1120, and bwv714, 719, 737, 742, and 756), are taken to be among his earliest works, from about 1700. Although the Neumeister manuscript represents neither an integrated body of work nor a unified collection, in its dazzling variety it embodies some contradictory and simultaneously essential traits of Bach’s early organ music: imperfect technique alongside daring innovation; reliance on models such as Pachelbel, Johann Michael and Johann Christoph Bach, and masters from north, south, and central Germany, together with a determination to surpass and dispense with such models; and an entirely unorthodox mixture of free composition and strict polyphony, unconventional harmony, and pronounced virtuosity.

A subsequent stage in Bach’s development is found in the chorale partitas bwv766–8, mostly wrought in the manner of Böhm (bwv768 was revised and expanded during Bach’s Weimar period). The Canzona bwv588 shows south German and Italian characteristics, while the Fantasia in G bwv572 looks to the French style. With their sectional layout, the preludes in E and G minor, bwv566 and 535.1, must have been written under Buxtehude’s immediate influence.

The extraordinary harmonic boldness and the richness of fermata embellishment in the pieces bwv715, 722, and 732, intended to accompany chorales, imply that they belong to the Arnstadt period when Bach’s treatment of chorales caused confusion among the congregation. The fugues after Bonocini (formerly Legrenzi) and Corelli, bwv574 and 579, can also be placed among the early works. Admittedly, the scarcity of autographs, combined with the complicated situation surrounding the other sources, makes it difficult to establish a reliable chronology. It is scarcely possible even to draw definite conclusions about which of the early keyboard works belong within the period of Bach’s youth, if that is set at about 1700–07.

The models recede in importance from the short Mühlhausen period, at the latest, and Bach’s individuality begins to pervade every note of his compositions. This applies particularly to the many extended organ chorale settings probably dating from between 1709 and 1712–13 and already so much in accordance with Bach’s later ideals that he found this group of 18 chorales (bwv651–8) worthy of revising in and after about 1740. In his freely composed organ works (toccatas, preludes, fantasias, and fugues) Bach tightened up the formal scheme, preparing the way for the two-movement prelude and fugue through an intermediate type in which the fugue was a long, self-contained complex but the prelude was not yet a unified section (such as the first movement of bwv532). Here is an early manifestation of one of the peculiarities of Bach’s working methods, encountered later in the ‘48’: fugues attain their final form almost instantaneously, preludes often go through several stages of development. Probably the most important work of these years, and datable to before 1709, is the Passacaglia in C minor bwv582.

In about 1713–14 a decisive stylistic change came about, stimulated by Vivaldi’s concerto form. Bach’s encounter with Vivaldi’s music found immediate expression in the concertos after Vivaldi’s opp.3 and 7 (bwv593 etc.). Features adapted from Vivaldi include the unifying use of motivic work, the motoric rhythmic character, the modulation schemes and the principle of solo–tutti contrast as means of formal articulation; the influence may be seen in the Toccatas in F and C bwv540 and 564. Apparently Bach experimented for a short while with a free, concerto-like organ form in three movements (fast–slow–fast), as in the Prelude, Trio, and Fugues bwv545.2 and bwv541.2, but finally turned again to the two-movement form, as in bwv534 and 536. Of comparable importance to the introduction of the concerto element is his tendency towards condensed motivic work, as in the Orgel-Büchlein. Bach’s conception of this new type of miniature organ chorale, combining rhetorical and expressive musical language with refined counterpoint, probably dates back to a relatively early point, possibly the beginning of the Weimar period, but he cannot have started to collect them systematically in the autograph before 1713–14. Among the earliest entered in the manuscript are, among new compositions, bwv608, 627, and 630, and about 1715–16 Bach added bwv615, 623, 640, and 644 (to cite some typical examples). Some of the pieces, such as bwv601 and 639, are of earlier date. By the end of the Weimar period the Orgel-Büchlein was complete in all essentials, although a few isolated pieces were added later, such as bwv620 and 631 (c1730), the fragment O Traurigkeit bwv1169 and bwv613 (c1740). The final total of 45 pieces falls considerably short of the 164 originally projected, but Bach had already ceased to work consistently at this major undertaking as early as 1716. The reason for this is unknown; when he took it up again in Leipzig it was only sporadically and apparently in connection with teaching, or so a copy made about 1727–30 suggests.

Bach composed few organ pieces at Cöthen, but among them is undoubtedly the C major Fantasia bwv573 which he added to Anna Magdalena Bach’s first Clavier-Büchlein (1722). In Leipzig, in about 1727, he composed the trio sonatas, a new genre for the organ, which he wrote, according to Forkel, for his eldest son Wilhelm Friedemann. It was probably in conjunction with renewed activity as a recitalist – he is known to have performed in Dresden (1725, 1731, and 1736), Kassel (1732), Altenburg (1739), and Potsdam (1747) – that he returned to the prelude and fugue genre. Now, surely as a consequence of the ‘48’, he always wrote them in two sections, with the preludes as important as the fugues. There was a final flourish of virtuosity (especially in the writing for obbligato pedal) in works such as bwv544, 546, and 548 (all early 1730s), but always in the context of a clearcut structure (there is a da capo fugue in bwv548).

In 1739, as the third part of the Clavier-Übung, Bach published a comprehensive and varied group of organ works. Framed by a Prelude and Fugue in E♭ (bwv552), there are nine chorale arrangements for Mass and 12 for the catechism, followed by four duets. Bach’s encyclopedic intentions can be seen in the form of the work – that of a collection of specimen organ pieces for large church instruments and smaller domestic ones (including the harpsichord), symbolized in his invariable coupling of a large piece with a small; they can equally be seen in the variety of his contrapuntal methods, whereby he constantly produced fresh kinds of cantus firmus treatment. At the very end of Bach’s output for the organ are such disparate works as the C minor Fantasia and Fugue bwv562 (1747–8), the ‘Schübler’ chorales (arrangements after solo movements from cantatas) and the canonic variations on Vom Himmel hoch bwv769. The variations, written for Mizler’s society in 1747, survive in two original versions, printed and autograph, whose different sequence of movements shows Bach experimenting with symmetrical form and the placing of climaxes.

17. Music for harpsichord and lute/lute harpsichord.

Just as Bach learnt most about the craft of composition from keyboard music, so too did he use it for preference in teaching others. He was obviously already a sought-after teacher when still in Weimar, but the move to Leipzig brought a decisive expansion of his teaching activities. Heinrich Nikolaus Gerber, who studied with him in the early Leipzig years, left an account of Bach’s method of introducing the widest variety of composition by gradual stages, along with the technical premisses of their performance. According to Gerber he used to begin with the Inventions and the French and English suites, and conclude the course with the ‘48’. This canon of characteristic works from the decade 1715–25 constitutes, so to speak, the stylistic core of Bach’s music for keyboard and for that reason served later as the yardstick by which to settle questions of authenticity. Nowadays, however, the yardstick’s usefulness has become somewhat problematic, since it does not take fully into account either the stylistic breadth of Bach’s early output or the unorthodox musical language of the late works.

One of the essential elements of Bach’s art as a keyboard composer is the attention he gave, from the first, to the idiomatic qualities of the individual instruments, respecting not only the differences between organ and harpsichord but also those within the family of string keyboard instruments, of which he used at least four types: harpsichord, clavichord, lute-harpsichord, and fortepiano. He is specific about the main kinds of harpsichord in the Clavier-Übung (the first part is for one-manual harpsichord, the second and fourth for a two-manual instrument). One of the earliest manuscript sources refers to the suitability of the E minor suite bwv996 for the lute-clavier (‘aufs Lauten Werk’). Bach took an active interest in Gottfried Silbermann’s experiments in developing the fortepiano during the 1730s and 40s. There is reliable testimony that he improvised on several new Silbermann fortepianos of different types in the presence of Frederick the Great in Potsdam in 1747, which makes it possible to regard the three-part ricercar of the Musical Offering as conceived primarily for this new kind of keyboard instrument.

There is an obvious association between Bach’s renown as a keyboard virtuoso, together with his work as a teacher, and the fact that his keyboard music is among the most accessible of his entire output, and also that it was the most widely available. Its dissemination shows a marked rising curve during the 18th century, internationally as well as within Germany. Bach’s harpsichord works were available in Italy, France, Austria, and England by 1750, and in view of this it is not surprising that the young Beethoven was schooled in the ‘48’. The growing recognition of the significance of this part of his output was reflected in the first complete edition of the works for harpsichord (begun in Leipzig in 1800 by Hoffmeister & Kühnel and continued by C.F. Peters), in which Forkel, among others, was involved.

Bach’s early harpsichord compositions are in a similar situation to the early organ works as regards dating and evaluation. None of the very earliest can be dated precisely. The Capriccio bwv992 has been assigned to 1704; there are no biographical data to support this (it is extremely doubtful that it was written for Bach’s brother Johann Jakob), but it certainly belongs to the period immediately after 1700. Before 1710 there were countless individual pieces like toccatas, preludes, and fugues (these last mainly using a ‘repercussive’ thematic technique like the early organ fugues); variation form is represented by the Aria variata bwv989. In the toccatas (bwv910 etc.) Italian, north German, and French influences conjoin in equal importance (bwv912 is an interesting counterpart to the organ work bwv532); Bach’s penchant for the French style is evident in his abundant use of the style brisé. After 1712 the particular influence of concertos by Vivaldi, Marcello, and others can be seen in Bach’s numerous concerto arrangements (bwv972–87).

To the last years in Weimar and the early years in Cöthen belong works such as the so-called English Suites and the Fantasia chromatica bwv903, and also the Clavier-Büchlein for Wilhelm Friedemann of 1720, which is predominantly didactic in layout. It is however less important for its instruction in playing technique (the Applicatio bwv994 gives fingering and tables of ornaments after D’Anglebert) than as a book of instruction in composition. For Bach himself, the two could not be dissociated: the Clavier-Büchlein contains eleven early versions of preludes from the ‘48’ as well as early versions of the Inventions and Sinfonias, under such titles as ‘preambulum’ and ‘fantasia’.

Then followed, also in 1722, Das wohltemperirte Clavier (book 1 of the ‘48’), with its 24 preludes and fugues in all the major and minor keys, surpassing, in logic, in format, and in musical quality, all earlier endeavours of the same kind by other masters, such as J.C.F. Fischer’s Ariadne musica. The work shows a perfectly balanced contrast between free and strict styles, each represented by several different types of prelude and fugue. Bach’s writing in book 1 of the ‘48’ in the most varied fugues – from two- to five-part, in a wide range of styles – represents the culmination of a 20-year process of maturation and stands unparalleled in the history of music. The final version of the two- and three-part Inventions and Sinfonias, in a collection entitled Auffrichtige Anleitung (‘upright instruction’) also arranged by key but representing a different method of composition whose object (according to Bach’s foreword) was ‘to teach clear playing in two and three obbligato parts, good inventions [i.e. compositional ideas], and a cantabile manner of playing’, dates from 1723.

The first traces of the subsequent great works of the Leipzig period are to be found in the two Clavier-Büchlein for Anna Magdalena of 1722 and 1725, which in fact anticipate the so-called French Suites bwv812–17 and the Partitas bwv825–30 – both initially dedicated to his wife. The Partitas in particular (appearing in print singly from 1726) represent a further culmination in Bach’s keyboard output; whereas the ‘48’ shows the prelude and fugue type developed to its most consummate maturity, these present similarly matured specimens of the most popular harpsichord genre of the time, the partita, comprising a suite of dance movements and ‘galanteries’. These – the burlesca, capriccio, and the like – do not appear in the English or French Suites; as in the English Suites, each partita begins with a large-scale movement, each differently titled and each in a different style. Later, with the collected publication of all six in 1731, Bach inaugurated his series of published works under the general title Clavier-Übung (the title was borrowed from a publication by Kuhnau, his predecessor in office).

In 1735 appeared the second part, whose contents were intended to be representative of the most prominent and fashionable styles: the Concerto in the Italian Style bwv971 embodies the ultimate stage in the process of transcribing instrumental concertos for keyboard, and stands in contrast to an Overture in the French Manner bwv831 which, more markedly than the partitas, represents what was specifically French in harmony, rhythm, ornamentation, and melodic invention. 1741–2 eventually saw the end of the Clavier-Übung series with the aria and 30 variations known as the Goldberg Variations. Apparently Bach had not cultivated the variation form since his youth, so that the contrast between the Goldberg Variations and the early works (chorale partitas and the Aria variata) is the more marked. This work outshines all others as far as performing technique is concerned (Domenico Scarlatti’s influence is unmistakable in places). The large-scale cyclical layout (based on a sequence of 10 x 3 movements, incorporating a series of nine canons, one at every third variation, arranged in order of ascending intervals to move towards a climax, with a final quodlibet) is without precedent. The basis of the composition is a ground bass of 32 bars – identical with, and alluding to the fundamental bass of Handel’s Chaconne with 62 Variations hwv 442 – first presented in the aria and then subjected to free and canonic elaboration in a wide variety of ways. In their monothematic and emphatically contrapuntal conception, the Goldberg Variations set the scene for Bach’s last keyboard works – the Musical Offering and Art of Fugue.

Besides the harpsichord works published in the 1730s, the only other major work is the second part of Das wohltemperirte Clavier (not so titled – the complete autograph does not survive). This companion-piece is less unified than book 1 and was partly assembled from existing preludes and fugues, some of them transposed. The freshly composed pieces probably date chiefly from the late 1730s; the work was complete by 1744 at the latest. Apart from this one major undertaking, Bach appears to have composed very few keyboard works at this period: perhaps the Fantasia ‘sur un rondeau’ bwv918, certainly the Fantasia in C minor with fragmentary fugue bwv906 from about 1738.

The dates of composition of the seven surviving works for lute – apparently his total output for the instrument – cover at least 30 years; four of the seven works are alternatively designated for lute clavier (bwv 996–8 and 1006.2). The earliest work is the Suite in E minor bwv996, which dates from the Weimar period, c1715; it already shows a surprisingly balanced construction. The Prelude in C minor bwv999 shows an affinity with the ‘48’, and likely belongs to the Leipzig years. All the other lute works were composed in Leipzig, starting with the Fugue in G minor bwv1000, an expanded polyphonic development from the violin fugue (in bwv1001), which (like bwv997) is in a tablature copied by Bach’s friend, the Leipzig lawyer and lutenist Christian Weyrauch. The Suite in G minor bwv995 (after 1011, for cello) dates from the period 1727–31 and is dedicated in Bach’s autograph to the Leipzig publisher and lutenist Jakob Schuster. The Suite in E (bwv1006a, after 1006 for violin) also survives in autograph form and is a much less demanding arrangement of its model as compared with bwv1000 and 995; it dates from the second half of the 1730s. Bach must have composed the Suite in C minor bwv997 before 1741; this is an original lute composition and is laid out in a similar virtuoso fashion to the Prelude, Fugue, and Allegro in E♭ bwv998 which can be ascribed to the early 1740s. The late works may have been written for the Dresden lutenists Silvius Leopold Weiss and Johann Kropffgans, and in any case were probably played by them. There is evidence that Weiss and Kropffgans performed at Bach’s house at least once, in 1739. Bach’s arrangement for violin and harpsichord of Weiss’s lute suite in A major (bwv1025) may also have been made in connection with this occasion. His contributions to the repertory of the lute, long past its heyday but enjoying a final flowering in the German-speaking countries, represent, along with the works of Weiss, the culmination of the instrument’s 18th-century repertory. They require an instrument with 10 to 14 strings, but in Bach’s day were alternatively played on the lute-harpsichord, an instrument in whose construction Bach had assisted. The indistinct line between lute and harpsichord music is illustrated by the autograph of bwv998, marked ‘pour La Luth ò Cembal’.

18. Orchestral music.

Many of Bach’s orchestral compositions must be presumed lost. The surviving repertory can in any case give only an incomplete idea of his output for larger instrumental ensembles, for he must have written many further works during his years at Cöthen and while he was working with the collegium musicum in Leipzig. Traces of lost concerto movements may be found in numerous cantatas, such as bwv42 (first movement), and other large-scale vocal works, such as the Easter Oratorio (first two movements); and various of the surviving harpsichord concertos, in particular, invite inferences about lost originals.

In the score bearing the dedication to Margrave Christian Ludwig of Brandenburg, the so-called Brandenburg Concertos are dated March 24, 1721. This is merely a terminus ante quem, for the concertos themselves must have been written over a considerable period before being assembled in 1721 as a collection of ‘Concerts avec plusieurs instruments’ (not as a single work in several parts). It cannot be proved that Bach composed instrumental music in his capacity as Konzertmeister in Weimar; but his position there and his preoccupation with the Italian concerto style during those years make it seem probable that he did. Of the Brandenburg Concertos, the sinfonia version of no. 1 was performed in 1713 as part of bwv208; and 6 points to the Weimar period, partly because of its indebtedness to the Italian type of concerto (above all in the middle movement) and also because of its unusual instrumentation (the particular combination of low strings is otherwise found only in Weimar cantatas). But it is not possible to draw any firmer conclusion about a Weimar orchestral repertory.

The special significance of the Brandenburg Concertos resides in the fact that, like Vivaldi’s, they abandon the standard type of concerto grosso and use a variety of solo combinations. The originality of Bach’s ideas extends far beyond Vivaldi’s, as do the density of the compositional texture and the level of professional virtuosity. The devising of concise head-motifs, particularly in the first movements, shows a strong Italian influence. All of Bach’s instrumentations are unprecedented. They feature all kinds of combinations, from homogeneous string sound (nos.3 and 6) to the heterogeneous mixing of brass, woodwind, string, and keyboard instruments. Just as unusual is Bach’s conflation of the group concerto with the solo concerto in nos. 2 and 5. No. 5 probably represents the latest stage in the set: it may have been written for the inauguration of the new harpsichord he brought back from Berlin early in 1719 (an earlier version of the concerto survives from about this date). At the same time, it marks the beginnings of the keyboard concerto as a form.

For a long time, Bach scholars schematically assigned most of his chamber and ensemble music to the Cöthen years. Recent studies based on original sources and style criticism have led to a thorough revision of the chronology affecting this part of his output. It now seems that only the smaller part of the instrumental ensemble music (or at least of what survives of it) belongs to the Cöthen period, while the greater part was composed at Leipzig, and principally for the collegium musicum which Bach was associated with from 1723 and which he directed from 1729 to the early 1740s. Thus the four Orchestral Suites, with their leaning towards French style, were written in Leipzig: no. 1 perhaps as early as 1725, nos. 3 and 4 in about 1725 and after 1730 respectively, and no. 2 about 1738–9. The B minor Suite (bwv1062), with its hybrid mixture of concerto elements and suite form and the extraordinary virtuosity of its flute writing, is most likely Bach’s very last orchestral work. The only solo concertos to survive in their original form from this time are the violin concertos in A minor and in E and the double violin concerto in D minor, which again obviously relate to the collegium musicum. Pointers to lost works that may be supposed to have been composed in Cöthen can be obtained from Leipzig pieces showing clear signs of reworking, above all cantata sinfonias with obbligato organ and the harpsichord concertos. Despite the many attempted ‘reconstructions’ of lost originals, however, there is not a single case in which a putative original is clearly discernible in a later version. The intended instrumentation of the original cannot be conclusively determined from the later version. Moreover, allowance must also be made for substantial differences between the two versions, so that a faithful reconstruction of a supposed but lost original is not really possible. Bach never proceeded in a mechanical way; rather, he strove to give the arrangement an identity of its own by subjecting the model to further development and exhausting its potential. This often involved the addition of fresh contrapuntal parts, the alteration of detail, and structural modification. Of special interest are Bach’s adaptations of instrumental works into vocal ones, such as the derivation of the opening chorus of bwv110 from bwv1069; also of note is the wresting of the outer movements of an ensemble concerto (bwv1044) out of the Prelude and Fugue in A minor for harpsichord (bwv894).

The most noteworthy of the later concertos composed in the 1730s, with substantial changes to the originals on which they draw, are the Triple Concerto in A minor bwv1044 (sharing several features with Brandenburg Concerto bwv5), the seven harpsichord concertos bwv1052–8, and the concertos for two or more harpsichords bwv1060–65, all but one of them reworkings of earlier works by Bach himself (except for bwv1064, an arrangement of Vivaldi’s Concerto in B minor for four violins, op.3 no. 10). In fact, Bach’s alterations and restructurings are sufficiently important – especially the deployment of the left hand of the harpsichord part and the invention of idiomatic harpsichord figuration – for works of this rank to be considered compositions in their own right. They owe their special historical importance to their occurrence at the beginning of the history of the keyboard concerto, a form which was to be taken up above all by Bach’s sons so that in Germany, until about 1750, it remained the exclusive preserve of the Bach family. A stimulus for the composition of the harpsichord concertos may have been the new instrument introduced on June 17, 1733 (‘a new harpsichord, the like of which no-one here has ever yet heard’), according to the announcement advertising the collegium musicum concert.

19. Chamber music.

As with the orchestral music, a great many chamber compositions are thought to be lost. Once again the greatest losses affect the Cöthen period, but the Weimar years also suffer. When the summary worklist at the very end of the obituary mentions ‘a mass of other instrumental pieces of all sorts and for all kinds of instruments’, it probably refers first and foremost to works for various chamber ensembles.

The unusual flexibility with which Bach manipulated the conventional genres of sonata and suite is comparable to his orchestral output, as regards formal and compositional aspects as much as textures. Particularly important is his emancipation of the harpsichord from its role as continuo instrument and its deployment as a true partner in the sonatas for harpsichord with violin (bwv1014–19), flute (1030–33), and viola da gamba (1027–9). The cycle of six harpsichord and violin sonatas (c1725–6) were the first in a series of works with obbligato keyboard and paved the way for a new musical genre. The traditional trio sonata with continuo still cast its shadow (for example, in the opening movements of bwv1015 and 1019), but it yielded by stages to a more integrated three-part style (for example, the opening movements of bwv1014 and 1018). The only genuine trio sonatas to survive, apart from the one in the Musical Offering, are bwv1038 and 1039, both dating from Leipzig. Bach’s arrangement of the gamba sonata bwv1027, after bwv1039 for two flutes and continuo is an illustration of the development of the new type of trio writing from the trio sonata. A similar procedure stood behind his earlier development of the organ sonata. Most movements of the organ sonatas are based on instrumental trios, as the arrangement of the first movement of bwv528 from a trio sonata movement for oboe d’amore, viola da gamba, and continuo in Cantata bwv76 illustrates. This same movement preserves a trace of the many lost trio sonatas of the Weimar and Cöthen years. Yet the trio sonatas of the Leipzig period, too, may represent only a small fraction of their original numbers, if the way the genre lingers on in the Musical Offering is any guide.

The list of surviving duo sonatas with continuo is also relatively short, and again dominated by works of the Leipzig period: the violin sonatas bwv1021 and 1023 and the flute sonatas bwv1034–5. The Fugue in G minor for violin and continuo bwv1026, from before 1712, is not only Bach’s earliest surviving piece of ensemble music, it is also the only chamber-music piece of the pre-Cöthen years to have survived as an independent entity. The only other sources we have for an idea of what kind of chamber music Bach wrote in his early years are the instrumental sonatas and sinfonias of the Weimar cantatas.

Bach’s creative powers in the Cöthen years appear in a special light in the sonatas and partitas for solo violin, dating from 1720, and the suites for solo cello from about the same time. The two were apparently planned as a double ‘opus’ since the autograph fair copy of the violin solos is specifically designated ‘libro primo’; no autograph of the cello suites has survived. The single partita for solo flute (bwv1013) is also likely to have been composed in Cöthen, but the playing technique is much more advanced than, for example, the flute writing in Brandenburg Concerto no. 5. Yet all the works senza basso not only demonstrate Bach’s intimate knowledge of the typical idioms and performing techniques of each instrument, but also show his ability, even without an accompanying bass part, to bring into effective play dense if illusive counterpoint and refined harmony coupled with distinctive rhythms. The special qualities of Bach’s chamber music were recognized fairly early. Johann Friedrich Reichardt wrote in 1805, reviewing the first edition of the solo violin music, that the pieces represent ‘perhaps the greatest example in any art of the freedom and certainty with which a great master can move even when he is in chains’.

20. Canons, ‘Musical Offering’, ‘Art of Fugue’, theoretical writings.

Bach’s preoccupation with the canon as the strictest form of counterpoint can be traced back to the Weimar period. In his organ chorales and particularly in the Orgel-Büchlein the canonic principle plays a major role. Canonic elements are present also in several of the early vocal works. Here however it is a matter of canonic technique cropping up in a context of complex contrapuntal construction; as a genre in its own right, the canon, in Bach’s day, would appear almost exclusively as a theoretical example in composition teaching. It was in this sense that it was often favoured – generally in the form of a circular canon – by musicians for entries in students’ albums: such entries were normally notated in enigmatic fashion, setting the would-be solver an intellectual exercise. Bach wrote such canons in albums more than once; for the most part they are probably lost. Except for bwv1076–7, all the surviving individual canons (1072–5, 1078, 1086) were probably dedicatory works of this kind; the dedicatee of bwv1074 is the Hamburg jurist L.F. Hudemann. bwv1077 was re-used for this purpose and entered in the album of the Leipzig theology student J.G. Fulde. What is probably the earliest of them is dated August 2, 1713 (bwv1073) and dedicated to the Weimar town scribe J.C. Gebhardi; the latest is dated March 1, 1749 (bwv1078); the dedicatee is the Leipzig medical student Benjamin Faber.

A new kind of theoretical canon came into being in connection with the Goldberg Variations, in which the canonic principle played a special part. In his personal copy of the Goldberg Variations Bach wrote in 1747–8 a series of 14 perpetual canons on the first eight bass notes of the aria ground (bwv1087), exploring the most varied canonic possibilities of the subject, subsequently arranging the individual perpetual canons in a progressive order, organized according to their increasing contrapuntal complexity. The types included range from simple, double, and triple canons, retrograde canons, and stretto canons to a quadruple proportion canon by augmentation and diminution. Nos. 11 and 13 of this series are identical with bwv1077 (see above) and 1076 (depicted on Haussmann’s Bach portrait of 1746).

Closely related to these (and likewise probably dating from the later 1740s) are the Vom Himmel hoch variations, in which Bach first used a strictly canonic scheme for a monothematic work in several movements of progressive difficulty. The Musical Offering (1747) is also plainly influenced by this mode of musical thinking. Here, for a theme incomparably more complex than that of bwv1087, he devised ten canons of differing structural types, notated as puzzle canons in the original printed edition of 1747. The series of canons on the ‘royal theme’ includes a canonic fugue, providing a bridge between the canons, which are primarily theoretical in conception though also intended for performance, and the two keyboard fugues or ricercares in three and six parts. A further constituent part of the Musical Offering is a trio sonata for flute, violin, and continuo, also based on the royal theme. In its second slow movement Bach introduced echoes of the fashionable style practised at the Prussian court. The Musical Offering, in effect a compendium in three sections, shows Bach elaborating on the theme supplied to him by Frederick the Great in every imaginable way for an ensemble of up to three instruments.

The Art of Fugue constitutes the final contribution to this group of monothematically conceived works intended as representative examples of a specific principle. As a didactic keyboard work, the Art of Fugue in some ways forms a counterpart to the two books of the ‘48’, with the difference that here it is exclusively the fugue that is in question, and, what is more, the fugues are developed from a single theme. Bach’s work on the Art of Fugue was accomplished in two stages – from before 1740 to about 1742, resulting in an autograph fair copy that contains a total of 12 fugues and two canons. And then (in connection with preparing the work for publication) from the mid-1740s to 1749–50. The extant autograph score represents the conclusion of the first stage, in which the conception of the work already appears clearly: beginning with simple fugues (Bach avoided this term, speaking of ‘contrapunctus’), progressing through ‘counter-fugues’, double fugues, and triple fugues, with interpolated canons, and culminating in a mirror fugue. For the printed version the number of movements was not only increased by four (two canons, a fourth simple fugue, and most notably a closing quadruple fugue) but their order was rearranged so as to expound more logically the ‘chapter of instruction on fugues’. When Bach died the work may have been more ‘complete’ than it is in the form in which it has survived. In particular the quadruple fugue had surely been completed in all essentials, since the composition of its combinatorial section must necessarily be an early stage in the composition of a quadruple fugue. Only the three opening sections of the exposition, however, are extant, and these – further abbreviated by the editors, give the Art of Fugue the appearance of being a mighty torso. The editor of the posthumous first edition, C.P.E. Bach, added the organ chorale bwv668.1 at the very end in order to mitigate the work’s ending with a fragmentary fugue.

The Musical Offering and the Art of Fugue mark both the end and the culmination of Bach’s activity as a keyboard composer in the broadest sense. While the two ricercars on the ‘royal theme’ of the Musical Offering represent different fugal styles (forward- and backward-looking) and different textures (three- and six-part polyphony), the Art of Fugue explores a notably more intensive monothematic conception. As a didactic keyboard composition in some sense it counterbalances the two parts of the ‘48’, yet with the difference that it concerns itself with fugue alone, in a series of compositions developed out of a single ‘principal composition’ (theme) – and does so using a technique in which forward- and backward-looking styles operate alongside each other, synoptically as it were. It was probably unintentional, and yet it is hardly by chance, that the initial premise and the goal of Bach’s keyboard art and his musical thinking come together in the Art of Fugue.

Throughout his Leipzig years Bach was engaged in teaching composition, primarily to some of his private students. His teaching always involved practical components such as the performance of thorough bass. For this purpose, he apparently formulated some basic sets of rules: Einige höchst nöthige Regeln vom General Bass bwv1133, transmitted in the early 1740s by his son Johann Christoph Friedrich, as well as Vorschriften und Grundsätze zum vierstimmigen Spielen des General-Bass oder Accompagenment bwv1134, copied about 1738 by Carl August Thieme, a former St. Thomas School student. Apparently in the later 1730s, Bach became increasingly interested in dealing with contrapuntal issues as some of his surviving teaching notes indicate. They include rules in German and Latin on five-part polyphonic writing from about 1740 (Regula Joh. Seb, Bachii bwv 1129), autograph excerpts from Zarlino (Canones aliquot per Josephum Zarlinum bwv1130) dating from 1742–3, and autograph rules for the treatment of syncopations in double counterpoint (bwv1131) from about 1743–6, as well as joint contrapuntal exercises with his oldest son Wilhelm Friedemann (bwv1132) going back to 1736–9.

21. Methods of composition.

Bach’s methods of composition can be outlined only roughly: the sources, musical and literary, present no more than a fragmentary picture. ‘Methods’ here refers to Bach’s general procedures of composition, as far as these can be described objectively (without venturing into conjecture about creative psychology) and can be related to certain essential impulses and particularly characteristic approaches.

Bach’s vast knowledge of the musical repertory was a decisive factor behind his art. He had an intimate knowledge of the types and styles of composition of his time and in particular of the work of his most important contemporaries; moreover, he had a sound idea of the music of the past, extending back as far as Frescobaldi and Palestrina. The study of works by other masters went hand in hand with experimentation in his own. It is thus characteristic that his acquaintance with the works of Buxtehude and Böhm, with Vivaldi’s concertos, with the Passions of Keiser and Handel, and with the masses of Lotti, Gasparini, and Palestrina should have left an immediate imprint on his compositions in the same genres. It was less a matter of imitation of a model than of an awareness of the possibilities, an expansion of his own manner of writing and a stimulation of his musical ideas. This is confirmed in a contemporary report by T.L. Pitschel on his manner of improvisation, according to which, before beginning his own fantasia, Bach as a rule played from music a work by another master (or perhaps one of his own) which would ignite his imagination. Further, C.P.E. Bach wrote that, in accompanying a trio, his father liked to extemporize a fourth part. This tendency to take compositions by others as a starting-point is paralleled in his late adaptations: in his arrangement of Pergolesi’s Stabat mater an obbligato viola part is added, replacing the one following the continuo in the original; and his version of the ‘Suscepit Israel’ from Caldara’s Magnificat in C expands it from a five-part into a seven-part piece. An important aspect of Bach’s procedure of composition is its systematic and encyclopedic nature. He habitually wrote works of one particular type within a relatively limited period: for example the Orgel-Büchlein, the ‘48’, the solo violin sonatas and partitas, the canons, the chorale cantatas etc. He was concerned to try out, to develop, and to exhaust specific principles of composition. There are practically no completely isolated compositions. Relationships, correspondences, and connections with other works can constantly be found. This approach to the procedure of composition is at once deep and yet of great natural simplicity; and it never results in mere repetition. Certainly there is repetition, of a kind, in the case of parodies or transcriptions of existing works. Yet even here it is inappropriate to speak of repetition, since in the process of parodying and transcribing, Bach always modified so that the end-product represents a fresh stage in the development of the original composition.

C.P.E. Bach related that his father did not actually compose at the keyboard – apart from some keyboard works whose material originated in improvisations – but that he often tried out his music on the keyboard afterwards. This procedure may be seen in the few instrumental works of which Bach’s autograph draft survives, for example the early versions of the Inventions in the Clavier-Büchlein for Wilhelm Friedemann, where an abundance of inserted corrections are to be found. In the vocal music, where a wealth of source material is available, the main stages of composition can often be reconstructed. In thematically and motivically self-contained movements, like arias and choruses, Bach normally began with the development and formulation of a motif, a phrase, or a theme, which would be guided by the prosody of the text; he then added the contrapuntal voices, and continued in the same way, sometimes using ‘continuation sketches’ to plan the music’s progress in advance (see the critical edition of the sketches, Marshall, J(i) 1972). In choral fugues he usually began by outlining the thematic entries and wrote in the accompanying parts afterwards. The decisive step was the embarkation on the writing of a movement, for progress was in its essentials determined by established models (harmonic-tonal groundplans, modulation patterns, aria schemes) and governed by the principle of unified continuation (‘style d’une teneur’ and ‘Affekteinheitlichkeit’ – ensured by a unified motivic organization and interchange, permutation, and transposition of component sections). The invention of the central idea was for Bach the critical moment in the process of composition, as the title page of the Inventions specifies: ‘gute inventiones zu bekommen’ (‘how to achieve good inventions’); and this is borne out by C.P.E. Bach’s report that his first requirement of his composition pupils was the invention of ideas. With this the die was cast, down to a work’s emotional content. Outlines and sketches relating to this operation can sometimes be found in the original manuscripts; typically, however, Bach hardly required more than one or two attempts before arriving at the definitive form of his principal idea. The further elaboration of the idea – the dispositio, elaboratio, and decoratio – required mastery of his craft rather than inspiration.

In composing multi-movement vocal works Bach, understandably, began as a rule with the self-contained movements and only afterwards worked at the recitatives and chorales. In the recitatives he normally first wrote out the text and then added the melody and bass, section by section. In the chorales the bass was added to the melody and the middle parts were inserted later. Then all the movements were revised in detail, and sometimes corrections were made. The appearance of Bach’s working drafts is thus unusually clear and neat as a whole although it is mainly in his fair copies that the particular quality of his handwriting, a quality comparable to that of his music, is expressed. The physical state of the fair copy had to reflect the degree of artistic perfection to which the composer aspired, and the pains taken to achieve neatness and clarity in the copy are not evidence of pedantry. Rather, Bach was aware of the dichotomy between the perfection of the musical idea and that of its representation in performance. For this reason and no other he made the following statement in 1738, through the mouth of his spokesman J.A. Birnbaum: ‘One does not judge a composition first and foremost by the impression of its performance. Yet if such judgment, which can be deceptive, is not to be taken into consideration, then I see no other way of forming an opinion about it except by looking at the work as it is set down in notation.’

Ultimately, for Bach, the process of composition was an unending one. Dynamic markings and indications of articulation would be inserted as he looked through the parts; he would revise and improve a work when he was copying it out, and when giving further performances would make fresh alterations and improvements. He also inserted corrections in works already in print. Throughout his life Bach was his own severest critic. Even in works which went through two or three different versions, like the chorale prelude An Wasserflüssen Babylon bwv653, the ‘final’ version does not represent a definitive one but merely a further stage in the search for perfection – the central and ultimate concern of Bach’s method of composition.

Works