3. Handel with care?

The blurb on the back of the recent Cambridge Companion to Textual Scholarship (2013) begins

Absolutely.  And although the focus of this blog will generally be sources written down and printed before the digital age, increasingly our access to such sources is through some digital means—an image on the screen, whether or not it ever makes it on to paper in our hands.  Inevitably this blog will often deal with the IMSLP, the International Music Score Library Project—a massive database of musical sources, mostly printed editions, but some manuscript copies or and even “born digital” files that need not necessarily ever have made it to paper.  There are full scores, vocal scores, parts, arrangements and transcriptions, and more recently even recordings.  The idea behind it is to make public domain material accessible globally—although of course what is in the public domain varies across the globe according to the copyright laws of any given jurisdiction.

There is much good to be said about the IMSLP, but I think many of its users have little sense of the problems inherent in this sort of resource.  Principally, as a wiki there is tremendous inconsistency in the quality, quantity, and reliability of the information it makes so readily accessible—even when the contributors to the site include major research libraries.  As an example, here is what the IMSLP currently displays at the head of the section marked “Full Scores” for Handel’s oratorio Messiah (which Handel was in the middle of writing exactly 275 years ago):

SOURCE:  cropped screenshot of http://imslp.org/wiki/Messiah,_HWV_56_(Handel,_George_Frideric) (accessed 30 Aug. 2016)

These first items seem to present two different digital images of the autograph manuscript (the “composing score,” as it is generally called).  The second of these items (it currently appears in two different scans—IMSLP #18920 and #296169) is a scan of a facsimile published in 1892 as Vol. 45a of the German Handel Society’s complete edition.  The bulk of the project was the work of an individual, Friedrich Chrysander (1826-1901).  For his Messiah facsimile, Chrysander sought to draw together all known autograph material for the work.  This volume thus contains not only the whole of the composing score (British Library R.M.20.f.2), but also a few settings in Handel’s hand in the “conducting score” (held in the collections of the Bodleian Library, Oxford), as well as two settings of the text “How beautiful are the feet” which Chrysander mistakenly thought related to Messiah (amounting to some 25 pages from yet another manuscript in the BL—whose Handel holdings have all been digitized and made available), and a few sketch leaves at the Fitzwilliam Museum, Cambridge.  This facsimile is therefore not intended to represent a single source:  it is an anthology of sources.

The first item listed on the IMSLP (divided into four files—#414200-414203) has a very misleading description.  Despite the claim that this is the “holograph manuscript,” these files together comprise a color scan of the same (black-and-white) 1892 facsimile that appears in black-and-white scans immediately below.  It troubles me that the scan omits the title page and front matter of the 1892 facsimile, and thus presents itself to be a scan of the autograph itself.  There are plenty of tidbits to reveal its true identity.  The source is indicated as being “State and University Library Carl von Ossietzky, Hamburg (D-Hs):  M/B/1722.” A user familiar with Handel sources would know that D-Hs has a very important collection of Handel’s conducting scores, the bulk of it coming to Hamburg through Chrysander himself.  The composing score of Messiah, however, is not in that collection, but in the British Library, as noted above.  D-Hs has made available a digital scan of M/B/1722, and it is much more honest than what is in the IMSLP:  there is the front matter for all to see—and indeed we see from the book plate on the inside front cover that this copy was originally the property of its editor.

Source:  cropped screenshot of http://gcs.sub.uni-hamburg.de/PPN818310642/2000/0/00000002.tif

IMSLP #414200-414203 does not include that image.  But Chrysander’s printed page numbers are there, which should arouse some suspicions.  There is no signal to the user, however, that these files have pages from several different documents—anthologized in print in 1892.  Anyone turning to these files to do their Handel research will get a false sense of the document(s).

That would be bad enough, but Chrysander’s facsimile is false in another important respect—no matter which scan is consulted.  In his efforts to produce as clean a facsimile as possible, Chrysander doctored the images.  In his preface, he complains of an earlier facsimile issued by the Sacred Harmonic Society in 1868:

“Handel’s music paper is always the best of his age, but yet the ink often shows through.  In the printed photograph, the ink of these passages appears on the wrong side of the page with the same thickness and blackness as real notes, if it has not been previously carefully removed.  In the London [1868] facsimile the photographic plate is generally printed off rudely without any such cleansing.  The result is that the notes that show through seem to have equal value with the written ones, and make the page not only ugly, but in numerous places illegible, and even give rise to false readings.  I have removed more than ten thousand such blots from the London edition by comparison with the autograph.”  (Preface, p. v)

In his removal of “ten thousand such blots,” Chrysander sometimes went too far.  Here is a glimpse of Chrysander’s facsimile:  the tenor staff of bb. 111-122 of the chorus “And the Glory of the Lord”:

Source:  cropped screenshot of IMSLP #18920

http://imslp.nl/imglnks/usimg/7/7f/IMSLP18920-PMLP22568-HG_Band_45a.pdf p. 21 (pdf p. 41)

Here is the same excerpt as it appears on the scan available on the British Library website:

Source:  cropped screenshot of http://www.bl.uk/manuscripts/Viewer.aspx?ref=r.m.20.f.2_f012r

 And, for good measure, the same selection as it appears in the new facsimile issued by Bärenreiter in 2008:

Source:  cropped from a digital scan (600 dpi) JPEG of the printed facsimile.

It appears to me that these last two are identical—and I would guess that the scans that appear on the BL website are the same digital files used in the production of the Bärenreiter volume.  (I gladly acknowledge that it was Donald Burrows’s commentary in the new Bärenreiter facsimile that drew my attention to the doctored Chrysander facsimile, and to these details particularly.)  The text in the new BL scan is replete with the blotches and blains that Chrysander so painstakingly removed.  But in his clean-up, Chrysander changed the text:  his fourth note is clearly on the fourth line (a C# in the tenor clef), where the blotted version reproduced in the later scan extends well into the fourth space—and that D is confirmed by the doubling in the viola line (which had no blot and so did not need to be cleaned):

Source:  viola line of Chrysander facsimile, as above; this is in alto clef

This D is the reading Chysander ultimately adopted in his edition of the work, with no comment about the apparent discrepancy in his facsimile.

Speaking of the blotchiest bar in the tenor line here (b. 119), Max Seiffert (who, after Chrysander’s death in 1901, brought the edition of Messiah through the press) comments

Source:  screenshot of http://petrucci.mus.auth.gr/imglnks/usimg/0/0b/IMSLP17693-Handel_Messiah.pdf, p. x (pdf p.8).  Seiffert’s A = Handel’s composing score; O, G and H are copies by John Christopher Smith (O is the copy in the Bodleian, G is at the Morgan Library in New York, and H a subsequent copy that was acquired by Chrysander for the Hamburg library.  Why doesn’t Hamburg scan THAT for the IMSLP?).

Compare the two images of the composing score here (bb. 119-120 from “And the glory of the Lord”—Chrysander from IMSLP #414203 on the left; BL scan on the right):

John Tobin (who edited Messiah for the new Handel complete edition, the Hallische Händel-Ausgabe—a project to which I will have cause to return in later posts) reads Handel’s procedure here differently than Seiffert—but then again Seiffert may have only seen the score via the already doctored facsimile.  According to Tobin (who translates this into the treble clef):

Source:  scan of John Tobin, Handel’s Messiah:  A Critical Account of the Manuscript Sources and Printed Editions (1969), p. 190

“Obviously in error,” and it is easy enough to see how Handel’s copyist could make such an error.  Tobin’s putative original reading fits the blots which the BL scan shows, but would be less clear from the 1892 facsimile’s reading (even blotty as it remains in this instance).  Seiffert complete misses out the E.

Just as another example of Chrysander’s tidying up, here is the 1892 facsimile’s presentation of something out of the Bodleian conducting score (which he included because it was in Handel’s hand); note particularly the annotations at the top of the page.

Source:  screenshot of IMSLP #18920 http://imslp.nl/imglnks/usimg/7/7f/IMSLP18920-PMLP22568-HG_Band_45a.pdf p. 281 (pdf p. 301)

Here is the page as reproduced in the 1974 facsimile of the conducting score:

Source:  a digital scan (600) JPEG of the printed facsimile (f. 57 of part II).

Not only has Chrysander eliminated a lot of the marginalia, as Donald Burrows points out, he has misread “Miss Young” and converted it into “Mißion.”  Note also his tempo marking:  allegro Larghetto.  (Whatever that means.)  In the original, is it allegro, that is, struck through—or is that just a stray smudge?  (In Chysander the smudge is eliminated.)  Its placement to the far left suggests to my eye that it was added after Larghetto, and thus maybe less likely to be a cancelled (earlier) marking.  But the sources are inconsistent for the tempo marking of this movement:  the version (for bass) preserved in the composing score is marked NBallegro; another version reads Andante.  Yet another lacks any instruction.

And so on.  With a work as textually well-documented as Messiah, the problems posed by these IMSLP items (and what they claim to be) scarcely do any real harm.  Indeed, by the time you read this, it may have been fixed.  (Check here.)  Even so, there have been tens of thousands of downloads of these files—if the IMSLP figures are to be believed—so somebody is using them.  This is surely the tip of the iceberg, and caution is advised.