Sweelinck: Erbarm dich mein o Herre Gott

Erbarm dich mein | ô Herre Gott. | Joann Pieters. ||
[vv. 3–4] Manualiter | unndt | Pedaliter. ||
[vv. 5–6] Uff 2. Clavir. ||
Jan.P..S.

  Lübbenau MS LyA1

[v.1] Versus | Quartus ||
[v.2] 2.da variatio ||
[v.3] Versus | Quintus ||
[v.4] Versus Sextus ||
[v.5] Versus | Septimus ||
[v.6] Versus | Octavus ||

  Turin Tablature TorG5
(South Germany, Austria, or Northern Italy, 1637–40)

[vv.1–2] Erbarm dich mein | O Herre | Gott | H.S.M ||
[vv.5–6] Erbarm | dich meiner | O Herre Gott. | H.S.M ||

  Zellerfeld Tablature Ze2
(Caspar Calvör, Braunschweig, 1668?)

‘Erbarm dich mein, o Herre Gott’, one of two popular early Reformation versions of Psalm 51 (the other being ‘O Herre Gott, begnade mich’, both with tunes in the plaintive phrygian mode), has spawned many fine polyphonic settings. Among them is a cycle of variations by Sweelinck, one of his finest, most comprehensive such works. (Scheidemann apparently thought so too; Sweelinck’s work clearly inspired (at least) two different works by his student: a pair of verses on the same song which quote two passages from the Sweelinck work, and a much later set of two pairs of chained verses on ‘Mensch, willst du leben seliglich’, which shares the first phrase of its melody with ‘Erbarm dich mein’).

The most reliable manuscript of ‘Erbarm dich mein’ and many other works by Sweelinck, the staff-notation Lynar A1, includes six variations, the middle pair of which have the cantus firmus notated in tablature below the staves for pedaliter realization (the same kind of notation used in Van Noordt’s 1659 Psalm settings) – one of only two explicit obbligato pedal parts in Sweelinck’s œuvre. A later tablature source, Zellerfeld 2, however, includes only vars. 1–2 and 5–6 (attributed to Scheidemann, whose own two ‘Erbarm dich mein’ verses are also found in this source). A third source, Torino G5, includes all six variations, also in tablature, but in a form that, according to Dirksen and Vogel (the MS does not seem to be available online), obscures the part-writing and is intended for performance on one manual only. For the sake of curiosity I have chosen to record a manualiter performance of all six variations, though I have mostly followed the readings of LyA1 rather than TorG5 in the middle pair. The listener may thus choose to hear any combination of pairs of variations.

Why are there six variations for a song that has only five stanzas, when usually Sweelinck provides a number of variations or verses equal to or fewer than the number of stanzas? A couple of possibilities present themselves:

1   the (or a) four-variation version was the original (as Dirksen suggests), and should be prefaced with a relatively simple harmonization representing the first stanza, as the composer himself did with ‘Allein Gott’ and should likely be done with ‘Ich ruf zu dir’ and perhaps others (see Julia Dokter’s important article on this subject, referred to in a recent entry on this page); in this case the extra pair of verses might be an optional alternative. (Dokter’s argument that the two ‘inserted’ verses are alternatives for stanzas 2 and 4 cannot stand, however, since these two variations are ‘chained’ or dovetailed even more intricately than some other variation pairs of Sweelinck’s. Her statement that TorG5 contains only this middle pair of verses – entirely understandable from the somewhat unclear presentation of this information in the new Sweelinck edition – also needs correcting; as noted above, the whole set of six verses is present there among the sequence of untitled ‘Versus’ by Sweelinck that also includes Psalm 36 and ‘Wij geloven in eenen God alleen’. This does not detract from the importance of her article.)

2  the six-variation version stands for the text as augmented by a sixth, doxological, stanza, as in, e.g., the 1616 Straßburg Kirchen-Gesangbuch. I am unfortunately not qualified to undertake the kind of study of text-painting that Dokter has furnished, so as to offer a plausible alternative to her analysis.

As it is, the form of this work remains something of a mystery and perhaps remains somewhat flexible.

The distribution of the music on the manual and pedal keyboards of the organ is also perhaps flexible, even beyond the question of pedaliter (clearly preferable, I think) or manualiter realization of the cantus firmus in variations 3 and 4.

The first verse can certainly be played on one manual, provided the active left-hand voice does not disrupt the right-hand cantus firmus with excessive wind disturbance. Two-manual performance is perhaps preferable, however, as directed by Scheidt: right hand (assuming it is playing the chorale) on the upper manual, the active left-hand voice on the Rückpositiv, which is not only closer to the listener but also usually more pentrating in sound. One must also pay attention to balance, however, and, in Sweelinck variations, the dovetailing. In the last bar of variation 1 of ‘Erbarm dich mein’, for example, I think the left-hand e' must either be played either on the same manual as the immediately preceding e' in the right-hand voice (i.e., the bicinium should be played on one manual, with the right hand moving to a new manual in this final bar), or be of similar volume and relatively similar timbre. The present recording takes the second approach.

The second variation can, technically speaking, be played in three ways. One-manual performance – I think probably the most authentic solution – was my choice here; in this case, the left hand joins whatever manual the right hand uses for the last bar of the first variation – which might have been the same manual it used in the bicinium, or a third manual). The piece can also be played on two manuals, with the two hands perhaps remaining on the manuals used in the bicinium; it may be significant that in LyA1, the upper staff contains only one voice throughout save for the last four eighth-notes in the middle voice in bar 64, and the widest span for the left hand there would only be a tenth, which is not uncommon in this literature. (The keyboard staff notation of the period is supposed to denote hand division, but one occasionally encounters an interval wider than a tenth written on a single staff, which suggests that the hand division may not have been absolute.) A two-manual performance of the second variation would, perhaps even more than in the first, require careful attention to balance, since the two upper voices are roughly equal in importance. Finally, since the cantus firmus is in the same tenor octave as in variations 3–4, it could be played pedaliter in variation 2 as well, with one or two manual keyboards employed (I am not sure there is any evidence in this literature for trios à 3 claviers – and given the almost entirely manualiter concept of Sweelinck’s keyboard works, a trio realization would be truly exceptional – but the possibility technically remains).

Whether the pedal is used in variations 3–4 or not, the manual voices must be played on one keyboard. The 4-voice writing and mostly slow-moving texture suggest the use of a plenum registration (the greater activity in variation 4 occurs in a higher tessitura, where the wind is less likely to be disturbed), but I think a strong reed-based combination is also a possibility, given that the Oude Kerk organ contained five manual reed stops and the two manuals controlling them could even be coupled together.

Variations 5–6 are explicitly designated in LyA1 for two manuals (I do not know whether this is the case in Ze2). In fact variation 5 can be played successfully on one manual; the two unisons between the upper voices are by no means disturbing or technically difficult. Variation 6, however, is clearly a two-manual work, as the several voice-crossings would create a number of musical and technical difficulties if it were played on one manual. It might be a betrayal of the composer’s intentions to play variation 5 on one manual, but the sudden entrance of the colored cantus in variation 6 would be all the more stark if it were carried out on a new registration. Dirksen rightly points out that this voice, like Sweelinck’s other coloratura settings, exactly fits the range of the Zink* stop found on many organs of the period and may well have been intended for it (possibly in a combination with the Trompet, Nasat, and flutes like that famously favored by his pupil Jacob Praetorius) – but other options might well be considered, as in the present recording.

*  Though the Zink was one of the Renaissance voices eschewed by Schnitger, it was still thought important enough to retain in the 1662 rebuild of the organ of the Severikirche in Ottendorf, which Scheidemann tested and dedicated and probably advised on; and at Altenbruch, the organbuilder Joachim Richborn replaced a Schalmei 4 with one in 1668 upon advice from none other than Matthias Weckman, star student of Jacob Praetorius. (Klapmeyer replaced this with a Vox Humana in the 1727 rebuild in which the Brustwerk was also added.) Buxtehude’s instrument in Lübeck also still had one, since he never got the Schnitger rebuild he tried for twice.