Portal de Periódicos da CAPES

New Light on the Sir Gawain and the Green Knight Manuscript: Multispectral Imaging and the Cotton Nero A.x. Illustrations

2017; University of Chicago Press; Volume: 92; Issue: S1 Linguagem: Inglês

10.1086/693361

2040-8072

Murray McGillivray, Christina Duffy,

Eurasian Exchange Networks

Previous articleNext article FreeNew Light on the Sir Gawain and the Green Knight Manuscript: Multispectral Imaging and the Cotton Nero A.x. IllustrationsMurray McGillivray and Christina DuffyMurray McGillivray and Christina DuffyMurray McGillivray, University of Calgary ([email protected])Christina Duffy, British Library ([email protected])PDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreAmong the striking features of the modest manuscript, London, British Library, MS Cotton Nero A.x. (art. 3), that contains the only known copies of the Middle English poems Sir Gawain and the Green Knight, Pearl, Cleanness, and Patience, are ten full-page illustrations of the poems (on fols. 37/41r, 37/41v, 38/42r, 38/42v, 56/60r, 56/60v, 82/86v, 90/94v, 125/129v, and 126/130r1) and a further two taking up most of their pages (on fols. 82/86r and 125/129r). Formerly these illustrations were treated with some disdain by critics (a typical dismissal is Israel Gollancz’s blunt “They are certainly of crude workmanship”2); lately they have attracted more measured attention,3 but some mysteries have remained, including the relation between the ink drawing and the coloration of the illustrations, the pictorial content of the original images in places where these are damaged or faded beyond legibility, and the possibly iconographically significant use of particular pigments by the colorist. Scientific study using multispectral imaging has allowed us to resolve some of these mysteries and contribute some knowledge to the ongoing discussion of others. Because this manuscript is in many ways a paradox and mystery in itself—containing as it does the unique versions of four of the finest works of Middle English literature, recognized in scholarly consensus as courtly productions of the highest literary artistry, but in a manuscript not representing at all a correspondingly elevated standard of craftsmanship—study of the illustrations emerges as one of the ways in which that paradox can be evaluated. The digital methods we have applied in our research have allowed us to see them in ways impossible with conventional methods.Multispectral imaging is a photographic technique in which data from both the visible spectrum of reflected light and the spectral ranges outside the visible spectrum, such as ultraviolet and infrared light, are recorded using digital technology for further analysis. As one of the technologies comprising the broad range of spectral imaging, multispectral imaging has been used for a variety of kinds of scientific study in fields such as geology and astronomy, but from our point of view its use for investigation of cultural heritage objects, such as medieval manuscripts, is most interesting.4 Some recent equipment for multispectral imaging of heritage objects uses tuned light-emitting diodes (LEDs) to expose an object extremely briefly to a succession of narrowly delimited wavelengths of light, recording the reflected light from each wavelength in a grayscale photograph. Filters can also be used in conjunction with particular wavelengths: for example, an ultraviolet-blocking filter used with ultraviolet illumination blocks the reflected ultraviolet light itself, but lets through to the camera any fluorescence of pigments caused by the UV illumination. Applications of multispectral imaging that are important for medieval manuscript study include the possibility of revealing erased writing in the case of palimpsests5 or of accidentally damaged pages where the ink is no longer visible due to abrasion or other damage;6 of documenting manuscripts in connection with conservation efforts or to preserve the data they contain against the possibility of further degradation; and of exposing manuscript materials to study and analysis of kinds that would not be possible without this kind of digital technology.We used Megavision’s Cultural Heritage EV Imaging System over a period of several weeks in January and February 20147 to investigate the illustration pages in the manuscript. All the illustration pages were imaged using the equipment. The special capabilities of this system are founded on its hyperspectral lens, which maintains a high level of focus across the spectral range from ultraviolet to near infrared, and its high-resolution charge-coupled-device array camera back, also highly responsive across the same range. The computer-controlled setup of camera and LED arrays was arranged on a table top, with screens diffusing the direct illumination of the LEDs; two arrays of LEDs fired simultaneously for smoother illumination. The manuscript was photographed in a cradle below the camera in what was essentially a copy-stand setup. The LED arrays flash seven visible-spectrum wavelengths, namely royal blue, light blue, cyan, green, amber, bright red, and dark red, and also one ultraviolet and four infrared wavelengths at the manuscript without filtration. Then a succession of filters is combined with particular wavelengths to create additional images.8 In total, the system produces a “stack” of twenty-three different images of each page in rapid succession, with a total light exposure minimal enough not to cause anxiety for conservation personnel. The final image, a color image created by adding together the seven visible-spectrum grayscale images, can then be compared with the individual spectral images.Our post-imaging exploration of results over the ensuing months focused on three areas: inspection of what we initially considered to be the underdrawings of the colored images using the infrared images; examination of illustration details now missing due to abrasion or fading that were revealed by ultraviolet fluorescence; and study of the varying behavior under different wavelengths of apparently similar pigments. Examination of the ink drawings underneath the current colored pictures is possible because the iron-gall ink that was used to draw them has its own unique response to the various spectra produced by the imaging setup. In particular, the ink absorbs infrared radiation in the 780-nm range quite well, whereas almost all the other pigments that are on the illustration pages simply reflect that wavelength of light up towards the camera, with the result that the iron-gall ink lines show as dark gray, while the other pigments show as shades of white or light gray. The effect is as if we were able to look right through the pigment on top and see the original ink drawing, which in a way is a correct description of what is going on. Multispectral imaging also recovered and virtually “restored” portions of the images that have long been lost to personal visual inspection because fluorescence under ultraviolet light can reveal damaged or missing artwork. The reflected ultraviolet light itself can be blocked from the high-resolution camera lens with an ultraviolet-blocking filter so that only the visible-spectrum fluorescence reaches the sensitive camera back. Although examination with handheld ultraviolet lights has long been used by scholars to reveal such missing detail, that method does not allow the results of examination to be shared as evidence and is moreover somewhat dangerous both for the documents and the eyes of the scholars because of the length of time required. Use of the ultraviolet-blocking filter permits fluorescence to be isolated from other results of illumination with ultraviolet light, including ultraviolet absorption. Finally, similar-colored paints may have different reactions to portions of the visible or invisible parts of the spectrum used in the multispectral imaging because of their chemical composition, and for this reason deployment of pigments by the colorist can be examined in a more detailed and comprehensive way by digital analysis than personal inspection allows.The “Underdrawings”It is clear that the illustrations of Cotton Nero A.x. were added some time after the text was written, and probably after the quires had been loosely sewn or even after the book had been made up.9 W. W. Greg first noted the main evidence for this in his review of the Gollancz facsimile: “[T]here are two holes (now repaired) in the inner margin of fol. 82 and the painting and design on the recto of this leaf are carried through the holes onto fol. 83r, showing that the picture was painted after the sheets had been sewn and probably after the volume was bound.”10In addition, as Greg does not note but others do,11 several of the illustrations are on unused, already ruled pages, again suggesting that provision of illustrations was an afterthought. The final three illustrations, to Sir Gawain and the Green Knight, are on unused leaves after this final text in the manuscript, at the end of a quire (fols. 125/129 and 126/130). Finally, fols. 37/41 and 38/42, consisting of four illustration pages to Pearl, constitute a bifolium that was probably added to the beginning of the manuscript at the time the illustrations were planned; it would stand to reason that the manuscript would be disbound but remain in its gatherings for illustration if the bifolium needed to be added.The illustrations were carried out in two steps: first a line drawing was made in ink, then it was colored by applying paint with a brush. Jennifer A. Lee first raised the possibility that the colors were applied by a different person than the artist who made the ink drawings; this was rejected by Kathleen L. Scott but reaffirmed by Maidie Hilmo, who even suggests that a third person added further color.12 Comparison of the ink drawing itself with the total impression created by an illustration when the overlaying pigment is added is possible because multispectral imaging can virtually strip away the paint pigment. Image 11 of each stack of multispectral images, made with infrared light with a wavelength of 780 nm, best captured the ink drawings underneath the colored paint that was applied over them (see Figs. 1 and 213 for an example comparing a full-color image with the infrared view of fol. 37/41r, the Dreamer reclining by the grave mound of the Pearl).14Fig. 1. MS Cotton Nero A.x. fol. 37/41r.View Large ImageDownload PowerPointFig. 2. MS Cotton Nero A.x. fol. 37/41r (infrared).View Large ImageDownload PowerPointIn general, these infrared images strongly support the conclusion of previous scholars, even those who presume that the same artist was responsible, that the drawings are considerably more skillfully done than the paint application, which does not in itself conclusively show that different artists were responsible but would tend to suggest that. More apposite, however, is what seems to be evidence that the draftsperson and the colorist did not always share illustrative goals and approaches, a situation that could arise if the same person colored the drawings after a period of time but seems more likely to be associated with two separate artists approaching the task seriatim.In Fig. 3, a detail of fol. 90/94v, the Green Knight, having had his head chopped off by Gawain at Arthur’s court at the beginning of the story, holds it up and prepares to ride out of the banqueting hall. The line drawing (Fig. 4) shows details hard to see in the painted page, such as the ruffing of the Green Knight’s doublet at the bottom, the way it blouses in folds over its tight-fitting lower portion, which follows the poem in showing that “his wombe and his wast were worthily smale” (Sir Gawain and the Green Knight 14415), or what may be genitalia just in front of the strap that comes down from the horse’s rump. The hand applying color to the manuscript pages was not only not as skilled as the hand that did the line drawings, but was also not intent on following them with any precision: all those details are obscured by a heavy-handed and almost slapdash application of pigment.Fig. 3. Detail of MS Cotton Nero A.x. fol. 90/94r.View Large ImageDownload PowerPointFig. 4. Detail of MS Cotton Nero A.x. fol. 90/94r (infrared).View Large ImageDownload PowerPointAnother example of a mismatch between what seem to be different illustrative goals is the handling of drapery in the illustration of one of the lady’s visits to Sir Gawain’s bedroom (fol. 125/129r). Here the lady of the castle leans over Gawain in his bed and (in the illustration only, not the poem), chucks him suggestively16 under the chin. The bed is heavily decked out, first in a red patterned cover, then in a blanket or coverlet striped in three different shades of green (Fig. 5). In the line drawing (Fig. 6), however, there is no preparation for all those green stripes: what stripes there are in the drawing in the material covering Gawain represent folds in the material and, though tucked under Gawain’s body, are meant to be seen as continuous with the vertical folds that hang down the side of the bed and end near the bottom of the illustration in swooping volutes. Drooping wrinkles sketched between these folds, now almost entirely obscured with a stripe of the medium green but revealed by the multispectral imaging, contributed in the original drawing (in a graphic way more associated with line drawing than painting) to the impression of massing and limp hanging of the cloth down the viewer’s side of the bed. The very same method of portraying drapery using the vertical hanging folds, the swooping volutes at the bottom, and the drooping wrinkles between the folds can be seen as pure uncolored line drawings on fol. 56/60v (Fig. 7, the tablecloth at Belshazzar’s feast in Patience) and on fol. 90/94v (Fig. 8, an image enhanced by capturing ultraviolet fluorescence, as explained in the next section of this article, showing the tablecloth in Arthur’s hall).Fig. 5. Detail of MS Cotton Nero A.x. fol. 125/129r.View Large ImageDownload PowerPointFig. 6. Detail of MS Cotton Nero A.x. fol. 125/129r (infrared).View Large ImageDownload PowerPointFig. 7. Detail of MS Cotton Nero A.x. fol. 56v/60v.View Large ImageDownload PowerPointFig. 8. Detail of MS Cotton Nero A.x. fol. 90/94v (UV light with UV-blocking filter).View Large ImageDownload PowerPointIt seems very unlikely, having made these comparisons, that even the vertical folds drawn by the draftsperson on the fol. 125/129r illustration and their continuations in the part of the blanket encircling Gawain were intended by the draftsperson to be filled in with dark green as the colorist has done: rather, they seem in the drawing intended to be a continuous part of the same surface of cloth as the rest of the blanket. The conclusion that the colorist had a different pictorial intention here than the draftsperson seems inescapable, and we would on the basis of comparisons of this kind argue that Lee’s suggestion that they were two different people17 is correct. It would also seem that the draftsperson’s intention was not to draw scenes for later coloring but to create line drawings as black-and-white illustrations complete in themselves, using the illustrative vocabulary of line art rather then preparing a drawing expressly for coloring.18 One could reasonably also suggest that the manuscript remained as illustrated with uncolored line drawings until an owner became dissatisfied and employed another artist to color them.The areas on fol. 83/87r (a text page, so not separately imaged by us using multispectral images—but the areas can be inspected in the facsimile at http://gawain.ucalgary.ca) that have been painted through the holes at the gutter that Greg discusses19 are captured in our images of fol. 82/86r, which shows Jonah being fed to the whale. The colored image (Fig. 9) shows the blue-and-white paint shading of the waves continuing through the lower hole, which is a half circle to the far left of the image and towards the bottom of the page, and the infrared image (Fig. 10) shows part of the ink line encircling the whale’s head also continuing through the hole (and part of the drawn eye of the whale can just be made out, though mostly obscured with a stain). Both the ink drawing and the coloring must then have been done after the sheets had been at least assembled into quires, and the most likely scenario is that Cotton Nero A.x. had already existed as a bound volume for some time before the line drawings were added to it, and then existed as a volume with black-and-white line-drawing illustrations for some time before the color was added.Fig. 9. Detail of MS Cotton Nero A.x. fol. 82/86r.View Large ImageDownload PowerPointFig. 10. Detail of MS Cotton Nero A.x. fol. 82/86r (infrared).View Large ImageDownload PowerPointRestoration of Missing Parts of IllustrationsAnother use for multispectral imaging that we have employed is in recovering and virtually “restoring” portions of the images that have been lost. The principle here is one that has been long known: fluorescence under ultraviolet light can reveal damaged or missing manuscript text, and the same is true of the artwork here. Ultraviolet light excites atoms and molecules of various substances and causes them to emit photons of visible-spectrum light, which is called fluorescence. Iron-gall ink, even when no longer visible on the surface of the parchment, strongly absorbs the incident ultraviolet light. Because the reflected ultraviolet light itself can be blocked from the high-resolution camera lens with an ultraviolet-blocking filter so that only the visible-spectrum fluorescence reaches the sensitive camera back, and because the exposure though very brief is uniform across the whole page, this can be a very powerful tool, though it cannot restore every part of every image, since some damage is too severe. Image 14 of each stack exposed the page to ultraviolet light of 365 nm; fluorescence was captured without any reflected UV light by interposing a UV-blocking filter. The advantage of this filtering is that the resulting image is not complicated by reflected ultraviolet light, so that fluorescence in the visible range is captured in an isolated way.There are several very sad areas of damage in the Cotton Nero A.x. illustrations. In an illustration to Pearl (fol. 38/42v; Fig. 11), the kneeling Dreamer raises joined palms to the Pearl Maiden, who extends a hand in his direction over the wall of her city across the stream from him. His face is almost completely obliterated, apparently by purposeful abrasion; hers is fairly unclear, too, and both the wall of the city and the bush in the top left corner seem faded or damaged or both. In the UV view (Fig. 12), many of these details return. Both faces are now clear, the wall structure is much clearer, and even the bush is slightly clearer. The Dreamer’s face in the naked-eye view has only vague slits where eyes and mouth should be and no nose; in the UV view he is wearing the full, somewhat placid, expression the artist originally gave him. Unfortunately, a similarly obliterated face, also that of the Dreamer, on fol. 37/41v is only partially recoverable with this technology (Fig. 13), likely because that face was the object of a more concerted attempt at obliteration.Fig. 11. Detail of MS Cotton Nero A.x. fol. 38/42v (color).View Large ImageDownload PowerPointFig. 12. Detail of MS Cotton Nero A.x. fol. 38/42v (UV light with UV-blocking filter).View Large ImageDownload PowerPointFig. 13. Detail of MS Cotton Nero A.x. fol. 37/41v (UV light with UV-blocking filter).View Large ImageDownload PowerPointIn another illustration from Pearl, fol. 38/42r (Fig. 14), the Dreamer stands across the same stream from the Pearl Maiden, gesturing with a pointing finger with his right hand and making some kind of motion with his left hand that is no longer readable with the naked eye because of damage to the place. (The fact that the pointing finger of the right hand has been painted over with the blue of the stream is often mentioned as an indication that the colorist was less skilled than the draftsperson or that the coloring was done with less care than the drawing.)20 Here UV light (Fig. 15) reveals the obliterated hand fairly clearly, showing no particular gesture, just an open hand, but also that the draftsperson has accidentally given the Dreamer a second right hand rather than a left one.21Fig. 14. Detail of MS Cotton Nero A.x. fol. 38/42r.View Large ImageDownload PowerPointFig. 15. Detail of MS Cotton Nero A.x. fol. 38/42r (UV with UV-blocking filter).View Large ImageDownload PowerPointIn both of these cases, an individual scholar in the past, when UV handheld lamps were still permitted for this manuscript, might have been able to see the missing details personally during the few seconds of exposure permitted but would have been unable, except through verbal description or a rapid pencil sketch, to communicate them to others. This new digital technology allows their entry into evidence: not only we but anyone who gets access to these high-resolution digital photographs can now evaluate these new details and their contribution to the iconography of the manuscript’s illustrations. The publication of these images of previously invisible portions of the illustrations, in other words, not only constitutes a contribution in itself to knowledge akin to the recent publication of photographs of the newly discovered underwater wreck of Sir John Franklin’s HMS Terror, the previously unknown and invisible being made publicly available as an object of knowledge, but also creates a resource for further study.This combination of UV light and a UV-blocking filter can sometimes be added to the previous method—in which infrared light was used to isolate the iron-gall ink of the underdrawing—to further clarify an image, since ink that has disappeared from an image will not be seen under infrared examination, whereas the regions of parchment that formerly carried the ink may still show it under UV illumination. In another view of the illustration from Sir Gawain and the Green Knight on fol. 90/94v previously discussed (Fig. 16), the faces of several of the participants have lost some ink, as have other parts of the drawing. In particular, the faces of Sir Gawain wearing red at the top left and of Sir Gawain (again) wearing red at the bottom right are much damaged. Infrared illumination does not help much at all in this case (Fig. 17), but UV light causes fluorescence that helpfully fills in some of the detail (Fig. 18). Here the top-left Gawain’s face has reappeared, and the bottom-right Gawain’s face is, while not completely visible, much easier to make out. Such details as the item of table furniture on the left below Gawain’s ax and the modeling of the tablecloth (Fig. 7) are much more prominent and easier to see.Fig. 16. Detail of MS Cotton Nero A.x. fol. 90/94v.View Large ImageDownload PowerPointFig. 17. Detail of MS Cotton Nero A.x. fol. 90/94v (infrared).View Large ImageDownload PowerPointFig. 18. Detail of MS Cotton Nero A.x. fol. 90/94v (UV light with UV-blocking filter).View Large ImageDownload PowerPointA more complex case is that of the casting of Jonah to the whale on fol. 82/86r (Fig. 19), an illustration to Patience. The contributions of the two artists seem again in conflict, as Hilmo points out: “Parts of Jonah’s head and the handle of the two-pronged fork seem to have been accidentally painted over.”22 Jonah’s face has also largely disappeared, apparently through abrasion, though the evidence of purposeful attack is not as clear as in the case of the two Pearl images just discussed. Infrared illumination (Fig. 20) does not restore the damaged face but makes clearer aspects of the ink drawing that are partially obscured by the paint covering them. In particular, the prophet’s right arm and hand are revealed to be behind two (or three?) whale teeth, the whole painted with ocean blue by the colorist. In addition, it is clear that the original drawing did not extend the left “prong” of the “two-pronged fork” over the nose of the beast, where in the colored drawing it seems to either overlap or enter its nostril. Instead, the drawing of that portion of the tool ends at the edge of the nose (the slight appearance of continuation over the nose in the infrared view is due to absorption of some infrared by the red paint applied by the colorist), so that the tool was behind the head of the whale (with an error in drafting making it unclear whether the right “prong” is in front of or behind the uppermost tooth). Probably, rather than the “two-pronged fork” the colorist has made it into, this was intended by the draftsperson as an oar, disappearing under the water and behind the head of the beast.23 It would seem that the colorist intended the entirety of the whale to be below the water, but that this caused him or her conceptual difficulties with the oar (perhaps misinterpreted as some kind of fork) and with the prophet, whose head is in the whale’s mouth but who is not covered with the blue of the waves, except for the right arm. The draftsperson’s intention was clearly different: in the ink drawing the billows are indicated by lighter pen lines, which have been a partial inspiration to the colorist for the contours of his waves; but unlike the waves of the colorist, the billows of the draftsperson do not appear to cover the whale’s head or tail. Probably the draftsperson’s intention was to show the whale’s head, outlined with a large circle, as reaching up out of the water to catch the prophet; and the whale’s tail, the drawing of which ends abruptly, just shy of the edge of the page, as disappearing into billows to its left and right but also sticking out of the water. The UV plus UV-blocking filter image of the same scene (Fig. 21) achieves only a partial restoration of the prophet’s face but does at least suggest the possibility that the draftsperson has given it an expression of shock or horror with an open mouth, a departure from his or her usual placid visages. However, the UV image considerably clarifies the drafting of the billows by the draftsperson and makes it certain that they did not, unlike the colorist’s waves, extend over either the tail or the head and maw of the beast.Fig. 19. MS Cotton Nero A.x. fol. 82/86r.View Large ImageDownload PowerPointFig. 20. MS Cotton Nero A.x. fol. 82/86r (infrared).View Large ImageDownload PowerPointFig. 21. MS Cotton Nero A.x. fol. 82/86r (UV light with UV-blocking filter).View Large ImageDownload PowerPointIn this kind of case, prolonged study of individual images from the stack produced in a multispectral imaging run is combined with prolonged comparison of them to reveal results that no one image alone could give. Both the infrared image and the UV plus UV-blocking image show the power of multispectral imaging by revealing details that are entirely invisible to conventional imaging and to naked-eye study using visible light. And because these are high-resolution grayscale digital photographs, the images can be blown up, placed side by side on a monitor, massaged with common or specialized graphics programs to increase clarity and legibility, and otherwise manipulated; and they, and the results of such manipulation, can be placed into the public record as evidence of, and new knowledge about, an earlier state of the illustrations to the manuscript than we would otherwise be able to access.Differentiating PigmentsThe principle used to differentiate pigments using multispectral imaging is that similar paints may have different reactions to portions of the visible or invisible parts of the spectrum used in multispectral imaging because of their chemical composition.24 We know, for example, from multispectral imaging as well as from X-ray fluorescence, that the blue used in the illustrations is chemically different from that used in the red-and-blue decorated initials. It behaves differently at various wavelengths, and it also fluoresces differently under the X-rays. X-ray fluorescence can show us the chemical reasons for the different behavior of the two blues under the portions of the spectrum used in multispectral imaging (in this case, Paul Garside has shown that the blue used in the initials is from azurite, whereas the blues used in the illustrations are organic blues from indigo woad).25In the illustrations themselves there are clearly two reds used, although they look very similar or identical now on the manuscript page. In Fig. 22 (fol. 38/42r) the Dreamer gestures to the Pearl Maiden across the stream. His gown is red, of a chemical composition that Paul Garside has identified, using X-ray fluorescence because of the presence of mercury, as vermillion.26 There are also numerous red or red-and-white flowers depicted in the locus amoenus where both figures stand, and it would seem likely from even prolonged direct visual inspection that the artist, whose palette was not large, would have used the same pigment for the red flowers as for the red gown. However, illumination with the farthest infrared emitted by an LED in our setup, at a wavelength of 940 nm (Fig. 23), shows that most of the flowers were painted using a completely different pigment. The vermillion almost disappears at this wavelength, but many of the flowers now appear as a dark black: they must be chemically different, even though the hues now seem so similar. Further research would be needed to establish what pigment was used for them, since we had not noted this difference during our manuscript study week when X-ray fluorescence was used; but a possible culprit is minium, or red lead. It seems possible to speculate that the appearance of this pigment when the artist used it was visually quite different from the red of the gown, perhaps brighter, livelier, more suggestive of the delight of the heavenly surroundings of the poem’s locus amoenus.Fig. 22. MS Cotton Nero A.x. fol. 38/42r.View Large ImageDownload PowerPointFig. 23. MS Cotton Nero A.x. fol. 38/42r (far infra