UNDER THE IMPRESSION: MULTISPECTRAL IMAGING OF LORD FREDERICK CAMPBELL CHARTER XXI 5

Abstract

Lord Frederick Campbell Charter xxi 5 is the only surviving English document that still has an authentic, legible, pre-Conquest seal attached to it. The text purports to be a writ of Edward the Confessor (1003x5–1066) granting a slew of rights to Christ Church Cathedral, Canterbury. We examined the writ using multispectral imaging to recover layers of erased text.

Many scholars have noted that the text of the writ was altered on at least one occasion. Now, multispectral imaging confirms that there were multiple layers of erasures, even more than previously anticipated. The original writ may have been inscribed on reused parchment. This can be used as evidence for the conditions (and even the immense quantity) in which writs were produced during Edward’s reign. Alternatively – or additionally – the writ’s multiple alterations could suggest that it was rewritten repeatedly after the Conquest, during various phases of Canterbury’s post-Conquest property disputes. The results confirm Nicholas Brooks’ hypothesis that at one stage the text was altered from referring to the rights to the archbishop alone (in the singular), to instead refer to the whole community at Christ Church (in the plural). Taken together, these results reveal shifts in legal thinking in Canterbury between 1066 and 1100, while demonstrating the enduring authority of Edward the Confessor’s seal. These results also show the potential for using multispectral imaging to illuminate – literally – the history of manuscript production.

British Library manuscript Lord Frederick Campbell Charter xxi 5 (BL, LFC xxi 5) is a relatively small rectangle of parchment attached to a hefty wax seal (fig 1). One of these parts is a rare, pre-Conquest survival; the other has visibly been subject to post-Conquest alterations. To hold it is literally to have pre-Conquest evidence on one hand and post-Conquest confections on the other.

Fig 1. LFC xxi 5 photographed in normal light. Photograph: © The British Library Board.

LFC xxi 5 has the distinction of being the only surviving English document that has an identifiable, pre-Conquest seal still physically attached to it. Other documents made in England before 1066 contain fragments of what appear to be seal tags; some even have a few scraps of wax attached. ¹ A now-detached seal of Edward the Confessor also survived into the modern period, along with a variety of earlier seal matrices, ² but only LFC xxi 5 has an identifiable, authentic seal still attached, albeit with a few repairs made to both the wax seal and the seal tags in the intervening 900 years. However, while the seal appears to be authentic, most of the visible text of the writ was rewritten after the Norman Conquest, as scholars have long acknowledged. ³ This paper will discuss recent multispectral imaging (MSI) of the writ, which has shown that the alterations to the text were more complex than previously thought. These results show multiple layers of erasure, under all parts of the visible text. As Nicholas Brooks predicted, at least one layer contains erased text granting rights to the archbishop alone (in the singular). This was altered around the turn of the twelfth century, so that Edward appeared to grant the rights to the whole community at Christ Church Cathedral, Canterbury (in the plural). ⁴ MSI also revealed at least one other layer of erasures that previous scholars did not anticipate. These results provide significant evidence for the production of writs, which were probably the most numerous type of royal document in eleventh-century England, but which have not survived in large numbers. Additionally, this analysis illuminates how legal thinking and strategies changed in the decades following the Norman Conquest.

DESCRIPTION OF THE WRIT

LFC xxi 5 consists of a piece of parchment that is 95mm long on its left side and about 200mm wide at its widest point. Two seal tags were cut from the bottom of the parchment. Each of these measures about 5mm wide, as seems to have been typical of eleventh-century English seal tags. ⁵ These have been repaired in later centuries with small pieces of thin paper and thread. At some stage, the writ was folded, with at least three horizontal folds and three vertical folds. The back contains endorsements in a twelfth-century hand (Carta edwardi regis de saca et socne et libertatibus ecclesie cristi), in a thirteenth- or fourteenth-century hand (· xlviii ·) and in a sixteenth-century hand (Ed. confess.).

The wax seal impression is attached to one of the tags. The seal is two sided. One side shows a bearded, seated figure holding a sceptre and an orb. The other side shows a seated figure holding a sword and a staff, although the part of the impression with the top of the staff has now crumbled away. At some stage, the inscription and the edges of the seal disintegrated and the wax was repaired to prevent further damage. This can be seen under infra-red light, where the repairs appear to be a darker colour, as shown in fig 4. Even though the inscription does not survive, this seal can be safely attributed to Edward the Confessor through comparisons with the detached seal of Edward that survived in Paris with most of its inscription intact. ⁶ The current diameter of the wax disc, with repairs, is about 80mm. The maximum diameter of the surviving original, impressed wax is about 60mm.

The parchment contains ten lines of visible text. This visible text purports to be a writ issued by Edward the Confessor to all his officials – bishops, reeves and thegns – in the shires where Archbishop Stigand of Canterbury and the monks of Christ Church had property. The current text confirms that Stigand and the monks have a variety of rights and revenues: to sake and soke, toll and team and over breach of the peace or royal protection (grithbrice), assault on a person in a house (hamsocn), assault on royal roads or a royal official (foresteal) and the apprehension of thieves (infangentheof) and harbouring fugitives (flymenafyrmth) ‘over their own men within boroughs and without, as fully and to the same extent as my own officials would expect it, and over as many thegns as I have assigned them to have’. ⁷ These rights echo a series of writs attributed to different kings confirming the rights of Christ Church Canterbury, starting with the freols confirmed by Cnut c 1018. ⁸ However, it has long been acknowledged that after the first three lines of LFC xxi 5, the text was erased and rewritten in a slightly lighter ink in the late eleventh or early twelfth century. ⁹ This rewritten section of the writ begins with the list of rights in the fourth line (heora saca ⁊ socne wurþe. on strande ⁊ on streame. on wudan ⁊ on feldan. tolnes ⁊ teames).

The first three lines were written in a hand that could date from the mid to late eleventh century. This hand used a ‘teardrop’-shaped a throughout, with a ligature between the g and the a in Stigand’s name (line 2). The ascenders on þ, b, h and l terminate in a thick end with a slightly wavy top. The letters m, n and h end with serifs extending to the right. The ascenders on lowercase d extend almost horizontally from the top of the bowl of the d, curving upwards at the end. S was written in three ways. This scribe wrote a rustic capital S for Archbishop Stigand’s name. In the second line, for words beginning with s such as sciran and se, the scribe used a tall, crook-shaped letter that curves over the letter to the right. Elsewhere, the scribe used an s consisting of a down stroke, which sometimes curved slightly to the left at the base, plus a curve situated slightly below the start of the highest point on the down stroke. Ys are dotted. Ts are short and curved. Gs are Insular. The tail of g curves around and sometimes attaches to the first curve under the crossbar to form a circle. The scribe used both þ and ð. Unlike the short, almost horizontal ascender on d, the scribe wrote ð with a long, diagonal ascender stretching to the left. The downstroke on the scribe’s Tironian nota et curves to the left.

The scribe who rewrote the last seven lines of the writ has been identified as the main scribe of the F manuscript of the Anglo-Saxon Chronicle. ¹⁰ This scribe, hereafter referred to as scribe F, worked at Christ Church, Canterbury, in the late eleventh and early twelfth centuries. Peter Baker has argued that scribe F was possibly the precentor of Christ Church. ¹¹ In addition to the F manuscript of the Anglo-Saxon Chronicle, scribe F also made additions to Anglo-Saxon Chronicle A ¹² and created other charters and forgeries. ¹³ The compiler of Anglo-Saxon Chronicle F – who was possibly scribe F himself – also altered or embellished some of the diplomas he inserted into his recension of the Anglo-Saxon Chronicle. ¹⁴ The last seven lines of LFC xxi 5 were copied with an ink and script that were very similar to those used in scribe F’s other works. ¹⁵ The last seven lines have many characteristics of his hand, which has been described in detail elsewhere. ¹⁶ These features include an irregular script, mixed ascenders – some forked, some thickened – and various forms of a, some two-storied, some more triangular in shape.

Traditionally, scholars assumed that the first three lines represented an original writ of Edward the Confessor and only the last seven lines were altered. ¹⁷ However, the first three lines may have been altered as well. Even in normal lighting conditions, the shadow of a large h can be seen between the visible words on and Cantƿara[byrig] at the end of line 2. In 2012 Elaine Treharne suggested that the whole document had been erased and rewritten. Treharne argued that the first three lines were also the work of scribe F, who had tried to ‘emulate the palaeographical forms of the original scribe’; ¹⁸ however, it is not clear why scribe F would abandon both an imitative script and a darker formula of ink after the first three lines. Other parts of the writ may have been altered by scribes other than scribe F as well. In 2013 Susan Kelly and Nicholas Brooks argued that a different hand may have added the n to ‘beon’, at the end of the third line. That n is more compressed than the ns in the first visible hand, its right leg curving rather than coming straight down and ending with a wedge. ¹⁹ Clearly, further analysis of the writ was needed. Such analysis became possible thanks to the advent of new technologies that enable the examination of erased documents written with iron gall ink.

MULTISPECTRAL IMAGING

In August 2017 Christina Duffy conducted MSI of LFC xxi 5. MSI is a non-invasive imaging technique that uses lights of different wavelength to recover faded, erased, damaged and often lost text. The British Library captures multispectral images using a MegaVision Cultural Heritage EV Imaging System. The extracellular vescicle (EV) camera includes MegaVision’s monochrome E7 50-megapixel back, computer controlled shutter and aperture, and custom hyperspectral parfocal lens, which is responsive over the entire range of silicon sensitivity.

The MegaVision system uses narrow-band light-emitting diode (LED) illumination, which subjected the writ to only the required light energy to expose the sensitive unfiltered monochrome sensor. Images were captured over fourteen spectral bands from the near ultra-violet (UV, 365nm) to the near infra-red (IR, 1050nm). The chemical composition of the inks, waxes and parchments in the writ is varied, and so these components absorb light differently. It is this difference in spectral reflectance that allows us to filter through noise and extract legible passages of text. Images can be further processed to clarify areas of illegibility. In this case a method called principal component analysis (PCA) was used. Mathematically, PCA is a multivariate statistical analysis variant, and through iterations enables us to reduce spectral noise and draw out features from a stack of images.

Fig 1 shows the writ as seen in normal light, as if viewed with the naked eye. Fig 2 is a pseudocolour image highlighting the most legible PCA component drawn from the multispectral image stack. Fig 3 shows further PCA iterations demonstrating an improvement in legibility. In fig 3, you can see traces of letters of the erased text in white. The letters of the visible text appear in a slightly darker grey. Fig 4 is an image of the seal under near infra-red light, where later wax repairs to the seal appear darker.

Fig 2. A pseudocolour image of LFC xxi 5 highlighting the most legible PCA component drawn from the multispectral image stack. Photograph: © The British Library Board.

Fig 3. Further PCA iterations. Photograph: © The British Library Board.

Fig 4. LFC xxi 5 with its seal photographed under near infra-red light. Photograph: © The British Library Board.

ERASED TEXTS

Figs 2 and 3 reveal the existence of at least eleven lines of erased text under all parts of the visible text, including the first three lines. Ascenders, descenders, curves and letterforms – which could not have been created by damage or folds – appear behind and slightly above the visible text, including the visible text of the first three lines. The scribes who erased and rewrote this writ often wrote directly over previous lines of text, thereby obscuring much of the text underneath. The parchment also bears traces of damage and deterioration, which may further obscure some of the earlier layers of text. Darker patches are especially noticeable on the left half of the parchment, possibly the results of damp or slight water damage. However, a few letters and words can be recovered, as detailed in the list below. All these suggestions are provisional.

The letterforms that can be recovered are listed here by the lines of visible text that are nearest to them. ‘Below’ refers to erased letters that appear lower down on the page than the nearest visible letters, while ‘above’ refers to letters that appear higher on the page than the nearest visible letters. ‘Behind’ refers to erased letters that were written over by the visible text. ‘After’ refers to erased letters that appear to the right of the nearest visible text, ‘before’ to letters that appear to the left.

Line 1:

‘lo’ or ‘le’ or ‘leo’ appear after the visible ‘d’ in ‘Eadƿeard’ ²⁰

Three down strokes and a line that curves to the right at its end appear above and behind the second ‘mine’

‘u’ (?) appears above the ‘e’ of the second ‘mine’

A variety of curved lines appear above ‘eorlas’, including an ascender attached to the left of a curve – possibly part of a ‘b’, ‘h’, or ‘þ’ – above the ‘a’ in ‘eorlas’

A further ‘l’, or letter or shape with a descender that curved to the right, can be seen behind the second ‘⁊’

‘þ’ appears after last ‘e’ in ‘ealle’.

Line 2:

A capital ‘N’ or a letter with two lines and a diagonal crossbar appears before ‘æt’

A line that curves to the left slightly, possibly an ‘i’ or ‘l’, appears before ‘on’

A large ‘H’, parts of which are visible to the naked eye, appears behind the ‘n’ of ‘on’ and the ‘c’ of ‘Cantƿara’.

Line 3:

A descender curving to the right appears under the n in ‘land’

A descender curving to the right appears under the e in ‘eoƿ’

Part of a ‘d’, with the curved top of the bowl and an ascender on the right, appears above the m in ‘him’.

Line 4:

A tall ‘s’ may be seen in the space between ‘⁊ socne’

The tall ascender of a ‘ð’ appears over the ‘u’ in ‘ƿurþe’

A letter with an ascender (possibly a tall ‘s’) appears in the space between ‘on streame’. The remains of the ascender can still be seen in normal light.

Between lines 4 and 5:

‘h’ or ‘n’ below the first ‘⁊’

‘⁊’ below the ‘n’ in ‘on streame’

‘g’ below the ‘t’ in ‘tolnes’.

Line 5:

‘n’ or ‘h’ appears after the ‘s’ in ‘brices’.

Between lines 5 and 6:

‘…y þ(ƿ?) r(n?) i(l?) c(t?) e’ (possibly ‘[gr]yþbrice’ or ‘write’?) appears below the visible word ‘hamsocne’

A line that curves to the left below the ‘s’ and the ‘t’ in ‘forestealles’

‘h(u?n?a?o?)is’ below the ‘te’ in ‘forestealles’

‘ag[en?]e menn binnan’ appears below the visible words ‘…es ⁊ infangenes þeoues’.

Between lines 6 and 7:

‘þ’ or ‘ƿ’ appears below the ‘g’ and also the ‘n’ in the visible word ‘burgan’

‘g’ appears below the n in ‘butan’

‘þa ic’ appears below the visible word ‘full’

‘a(o?e?) g’ appears below the visible word ‘forþ’

‘þ’ or ‘ƿ’ appears below the ‘n’ in the visible word ‘agene’.

Line 7:

‘g’ between the ‘o’ and ‘u’ in ‘ouer’

A letter with a long descender or ascender (possibly a long ‘s’) between the ‘e’ and ‘r’ in ‘ouer’

‘e’ between ‘ouer’ and ‘sƿa’

Line with a crossbar (part of a long ‘s’ or ‘f’?) between ‘þegena’ and ‘sƿa’

‘e’ between ‘s’ and ‘ƿ’ of ‘sƿa’

‘7 his þam þ(ƿ?)’ (or ƿamb?), starting behind and slightly below ‘sƿa ic…’

‘ea’ in the space between ‘hæbbe’ and ‘⁊’.

Between lines 7 and 8:

‘m’(?) or ‘us’ below the ‘s’ in ‘swa’

‘ǷER’ or possibly ‘þer’ or ‘þeƿ’ (possibly the Old English word for man or part of a place name?).

Line 8:

Long ‘s’ behind the ‘t’ in ‘teo’

Capital ‘I’ between ‘teo’ and ‘butan’

‘g’ below the ‘t’ in ‘butan’

‘r’ or ‘s’ below the ‘n’ in ‘butan’

Short down stroke (part of an ‘i’) before ‘hy’

‘a’ after ‘y’ in ‘hy’

‘forgyf’ below and behind ‘7 heora…’.

Between lines 9 and 10:

‘Fr’ above ‘sƿa’ in line 10

‘þy’ above ‘cnut’ and ‘cyng’; the y appears to have been dotted

‘e’ with a very long tongue above ‘ng’ in ‘cyng’

‘ba(?) cl(d?) u(y?) d þi riw [p]?’, beginning below the visible word ‘forgiuen’.

Line 10:

‘b’, ‘h’, or ‘þ’ between ‘cyng’ and ‘ær’ (ascender runs over upper e’s baseline).

Below line 10:

A group of strokes, blurred, very close to where the parchment has been cut, below the visible word ‘geþauian’.

The results suggest that there are not one but several layers of erased text in each section of the parchment. Some of the lines of erased text seem to be too close together to have been part of the same original document. For example, between lines 9 and 10, near the word cyng there is an ‘e’ with a very long tongue. However, one of the erased ascenders nearby – under cyng in what is now line 10 – runs over the baseline on which the ‘e’ sits. Either a scribe wrote on a steep diagonal, or these letters represent two erased lines that intersect. Moreover, those underlying layers of text were probably also in Old English: runic-derived characters thorn (þ) and wynn (ƿ) are visible in the erased text across the page, including in the first three lines.

While it is harder to tell what any one of those erased layers contains, at least one layer seems to have been a writ or a forged writ. The words that can be recovered from this layer echo phrases found in other purported writs from Canterbury. In fact, these phrases often echo the wording of the visible text in the last seven lines of LFC xxi 5. Between lines 5 and 6, the term gryþbrices may appear, referring to the right to fines levied over breaking the peace. Forgyf appears under line 8. Variants of the word are found in other (purported) writs associated with Canterbury: forgyfen appears in a writ of Cnut for Canterbury copied into the MacDurnan Gospels, ²¹ forgefen in a writ granting the community of St Augustine rights over their own men, ²² and forgeuen in a writ purporting to grant Stigand and the community at Christ Church all the property that they had during the reigns of Edward’s predecessors. ²³ Forgiuen appears in the visible text of LFC xxi 5. ²⁴ The word does not appear in any other surviving pre-Conquest writs; it only appears in writs or forgeries from Canterbury. Also between visible lines 5 and 6, the words [h]is ag[en?]e menn binnan (‘his own men within …’) appear under MSI. Again, variations on that exact phrase appear in three aforementioned Canterbury writs and the visible text of LFC xxi 5, to refer to Christ Church’s rights over its own men, within burhs and without. ²⁵ Similar sentiments were expressed in writs that were not associated with Canterbury. A writ of Edward confirming the rights of Bury St Edmunds – which survives in a single-sheet in a mid- to late eleventh-century hand – contains the phrase ofer ealle heora agene menn ægðer ge binnan burh ge butan (‘over all their own men, both within burhs and without’). ²⁶ The phrase his hagan land her binnan ⁊ ofer his agene man appears in a potentially authentic writ for Wulfwold, abbot of Chertsey, giving him sake and soke over ‘his land enclosed within [London] and his own men’. ²⁷ Meanwhile, various spellings of ofer his menn binnan appear in other writs of Edward the Confessor that are generally considered to be largely authentic, for the monk Ælfstan and for Abbot Leofwine of Coventry. ²⁸ Thus, some of the phrases that can be recovered from the erasures in LFC xxi 5 suggest that at least one layer was a writ or a forged writ, and some of the language seems to be closely connected to the wording found in writs associated with Canterbury.

However, another layer of erased text may not have been a writ of Edward the Confessor, or at least it does not use the layout of surviving writs from eleventh-century England. Some of the erased text – including the H under line 2 and some erased text between lines 7 and 8 – appears to have been written in capitals or majuscule letters. None of the five surviving, contemporary, single-sheet writs of Edward the Confessor (as opposed to diplomas) include words written in capitals. Capitals do appear in some mid-eleventh-century Canterbury copies of earlier eleventh-century writs; ²⁹ however, it is unclear if these copies reflected the format of original documents, or even if these copies were based on original documents and not forgeries. That being said, these examples do show that Canterbury scribes were willing to add capitalisation to writs, at least. Names were written in capitals in some charters of Edward the Confessor, including in a diploma for Saint-Denis that was accompanied by a writ and seal; however, the accompanying writ does not use capitals. ³⁰

Additionally, erased letterforms under the first three lines could potentially be earlier than the mid-eleventh century: the long, curved ‘l’ shape visible in the first line is intriguing in this respect, since it is not found in many mid-eleventh-century examples. Moreover, none of the texts of Edward’s surviving writs allow for an ‘l’ to appear so soon in the text. Some version of the formula Eadweard cyng gret (‘King Edward greets …’) begins ninety of the ninety-two surviving texts that purport to be writs of Edward and that were written in some form of English, whether they were authentic or forged. Both of the two writs that have a different second or third word are dubious. ³¹ One of these is, admittedly, a single-sheet forgery from Christ Church, with spellings suggesting that it was created in the late eleventh or early twelfth century, so later scribes at Canterbury were not averse to diverging from that phrase that usually opened writs. ³² However, neither of these two exceptions has an ‘l’ that appears sufficiently early in the text.

POSSIBLE EXPLANATIONS FOR THE ERASURES IN LFC XXI 5

There are two main explanations for the erasures revealed by multispectral imaging:

1. 1. The first three lines of LFC xxi 5 represent a genuine writ of Edward the Confessor, copied on reused parchment that originally contained something other than a writ, hence the capital letters. The writ was then partially erased and recopied in the first decade of the twelfth century.

2. 2. None of the visible text is part of the genuine document issued in Edward’s name. The genuine document was erased and written over several times.

At least one layer of the writ seems to have been a contemporary document of Edward the Confessor, since the seal is genuine and there is no evidence that it was added later. As noted above, the design of the seal matches images of the seal of Edward that was held in Paris. ³³ The measurements of the seal tags are consistent with the dimensions of the remnants of other possible seal tags on writs of Edward the Confessor, noted above. Nor is there evidence that the seal was cut from another document and attached to this parchment: the seal tags were clearly cut from the bottom of this parchment, although they have been reinforced in modern times with sewing and thin paper supports.

That leaves the question of whether the first three lines were part of that genuine document of Edward the Confessor. The erasures under the first three lines could just show that it was written on reused parchment. Certainly, the erased letterforms under the first three lines could potentially be earlier than the mid-eleventh century: the long, curved ‘l’ shape visible in the first line might be compared to manuscripts from Cnut’s reign. ³⁴ Meanwhile, the script of the opening lines is not implausible for Edward’s reign. ³⁵ If this were the case, it could mean that a layer of unrelated text was erased in order for a scribe to write a writ of Edward. The first three visible lines and the erased text that echoes other writs – such as agene menn – could have been part of a genuine writ of Edward, copied on reused parchment.

To support this view, there may be at least one other possible example of a writ of Edward copied on reused parchment. Richard Mortimer has drawn our attention to Westminster Abbey, WAM xii, a writ of Edward the Confessor granting land at Perton, Staffordshire, to Westminster Abbey. ³⁶ The drypoint ruling on the parchment does not match up with the visible lines of text, suggesting that the parchment was at least prepared for another text, if not erased and reused. WAM xii also contains a fragment of a seal, making it even more relevant to the case of LFC xxi 5 and questions about the production of sealed writs. The seal is badly damaged and very little can still be seen of the original impression. A ‘V’ shape on one side might once have been the folds in the robe of a seated figure of Edward. The seal’s authenticity cannot be confirmed; however, the seal tags measure 6mm. This is very close to the measurements of LFC xxi 5 (each tag is about 5mm wide) and the fragments of authentic tags that still survive, whereas seal tags on forged documents were generally larger, as noted above. However, the reused parchment hypothesis assumes the visible text of WAM xii is itself genuine and the drypoint ruling was used for a non-documentary text; but the visible text of WAM xii is not above suspicion. That text refers to Perton, Staffordshire, a controversial estate that was one of Westminster Abbey’s most distant holdings. Westminster’s control of Perton had to be reasserted and confirmed with writs as late as the reigns of Henry i and Richard i. ³⁷ Members of the community at Westminster therefore had motivation to preserve – or perhaps even alter – writs relating to Perton. As far as we are aware, no studies have been conducted to see if there is any erased text underneath the visible text and, if so, if that text was unrelated or the text of a genuine writ that was altered. More work also remains to be done on WAM xii and the possible erasures there.

If a genuine writ was copied on reused parchment, there are many possible implications for our understanding of the production of writs in mid-eleventh-century England. The reuse of parchment might indicate economic factors that made accessing parchment difficult or expensive for kings’ administrators. ³⁸ Alternatively, reused parchment might suggest that writs and other royal documents were being produced quickly and/or on a relatively large scale, to the extent that production outstripped the supply of new parchment. This reinforces James Campbell’s point that writs may have been far more numerous than charters, even though fewer writs than charters survive to this day. ³⁹ Writs’ very frequency, along with their small size, might have meant few were saved or survived. ⁴⁰ Moreover, many writs contained announcements of new bishops, instructions or other information that quickly became redundant and therefore did not need to be saved. It is notable that the writs that do survive were often related to property or rights and, in the case of LFC xxi 5, might have been altered to make sure their terms applied beyond the lifetime of one individual. (For Richard Sharpe’s arguments that writs were reissued with each new king, please see below.) MSI of LFC xxi 5 offers some new physical evidence from this important class of documents that could corroborate the frequency and speed at which they were produced. As such, these MSI results perhaps even hint at the importance and schedule of Regenbald, the priest who was described as Edward’s sigillarius, perhaps meaning his seal keeper. ⁴¹

However, even if LFC xxi 5 was written on reused parchment, the possibility remains that the original text of the writ or other document was erased and all of the visible text of LFC xxi 5 is the result of later alterations. The first three lines, written in mid-eleventh-century vernacular minuscule script, could have been erased and rewritten during or soon after the end of Edward’s reign: hands that are plausible for the 1060s are also plausible for the 1070s. The writ could then have been partially erased and rewritten again by Scribe F in the late eleventh or early twelfth century.

In this scenario, the original document might have pertained not to rights, but to property that was eventually secured or some other information that did not need to be saved in its original form. Farther down, the capital letters – ǷER… – could conceivably refer to one of Christ Church’s properties, at Warehorne. Warehorne was spelled variously as Ƿerhornas, Uuerahornas, Werhorna and Werehorne in eleventh-century manuscripts. ⁴² Warehorne was an estate with a complex eleventh-century history. In the early eleventh century, the monks of Christ Church made a single-sheet copy (or forgery) of a ninth-century charter which claims that King Egbert of Wessex granted land at Warehorne and Flotham to Ætheric, minister. ⁴³ In 1010, Archbishop Ælfheah granted land at Warehorne to Christ Church, according to a note preserved in one of Christ Church’s luxurious gospel-books. ⁴⁴ Then, in the 1030s, Eadsige bought Warehorne from Christ Church, with a reversion to Christ Church after his death and Eadwine’s death, as recorded in a thirteenth-century register from Canterbury. ⁴⁵ Warehorne was also one of the estates that was listed in the de luxe gospel-book as having been confirmed by Edward the Confessor, in an entry that seems to have been copied down in the 1070s. ⁴⁶ It is possible that an earlier stage of text in LFC xxi 5 referred to this estate’s reversion or confirmation to Christ Church. Once the estate was secured – and especially after Domesday Book recorded Warehorne as Christ Church’s property – any other documents related to Warehorne could safely be altered to refer to more general rights. However, this must remain a hypothesis. The letters following ǷER in the document are now too blurred to recover, having been effectively obliterated by an earlier eraser. And the interpretation of these letters as Warehorne also assumes that some underlying text was part of an earlier document pertaining to Christ Church.

REMAINING QUESTIONS

Due to these uncertainties, even with multispectral imaging we still cannot answer a key question about this writ: did it fit into a series of genuine writs or were all the ‘writ’ elements forged later? This question has implications for the use of writs in eleventh-century England as a whole, as outlined in Richard Sharpe’s magisterial examination of the subject. ⁴⁷ Sharpe suggested that a series of writs from Bury showed that each new king reconfirmed rights and privileges to individual bishops/abbots via writs, if these writs could be shown to be genuine. ⁴⁸ However, he noted that this hypothesis was unproven in the case of Canterbury, because several Canterbury writs, including LFC xxi 5, had clearly been altered later. If some of the erased text is genuine, it could prove that a genuine writ of Edward confirmed that the archbishop, at least, had some of those rights, such as grithbrice.

Unfortunately, MSI still cannot prove definitely whether the partially erased writ in LFC xxi 5 was the original document issued by Edward or not, since there are other layers of erasure underneath it. These layers have been too obscured by erasures and other texts to recover enough data to confirm their contents definitively, and to prove that the original writ was part of the series. Nor does the text recovered from the erased writ layer prove that the writ was genuine. The erased text echoes language that is found primarily or exclusively in writs and forged writs from both Christ Church and St Augustine’s Canterbury, including the rewritten text of LFC xxi 5, as discussed above. Susan Kelly has shown the close interrelationship between forgeries for Canterbury and St Augustine’s in the period around the Conquest, particularly during the abbacy of Æthelsige (r 1059–70). ⁴⁹ It is possible that one of the layers of erasures in LFC xxi 5 contained a forged writ that was modelled on the writ of Cnut copied into the MacDurnan Gospels in the early eleventh century. Of course, such shared phrases do not mean all these documents are open to suspicion. Kelly argues that the writ for St Augustine’s might be authentic, even though it only survives copied into later records. ⁵⁰ That being said, those later records also mention that the writ for St Augustine’s had a seal, and, without the original single sheet surviving, there is no way to rule out the possibility that that the St Augustine writ was altered, as LFC xxi 5 was, to take advantage of its seal. ⁵¹ While writs were not entirely formulaic (except in their opening words), it is nevertheless suspicious that some phrases occur exclusively in Canterbury documents and not in writs promising similar rights to Bury and other houses. We cannot prove the status of the erased writ layer in LFC xxi 5 either way. The erased text in LFC xxi 5 could have been genuine and theoretically could have been a model for some forgeries at Canterbury; equally, it could have been a later concoction based on documents already held at Canterbury.

And even if the erased writ layer was forged or altered, we cannot prove when it was altered. There were plenty of occasions for a writ to be re-rewritten again and again at Christ Church Canterbury between c 1050 and c 1120, as recent scholarship has emphasised. ⁵² Even before the Norman Conquest, Christ Church was ‘in disgrace’, as Kelly puts it, under Archbishop Stigand, who had not gone to Rome to receive his pallium and who had jeopardised Canterbury’s primacy. ⁵³ The monks of Christ Church had further motivations to secure their property after the Norman Conquest of 1066 and after a fire at Christ Church in 1067. The monks needed to pay for rebuilding, and the fire also offered an opportunity to ‘recreate’ or ‘improve’ some allegedly lost documents. The community also needed to protect its holdings after the deposition of Stigand in 1070 and during the ascendancy of Odo of Bayeux, who, as earl of Kent, adopted a predatory land acquisition policy. Odo and his followers came to control substantial estates in the southeast, including some claimed by Christ Church. In the 1070s, a number of those estates were adjudged to belong to Christ Church at an event known as the Penenden Heath trial. At the same gathering, the archbishop’s rights were reconfirmed, at least in the accounts of the trial that survive from around 1100. ⁵⁴ While the archbishop’s rights were confirmed, these surviving accounts do not suggest that the estate at Warehorne was contested at the Penenden Heath trial, so – if these accounts can be trusted – this might have been a moment when the monks of Christ Church could have found it useful to convert a writ about Warehorne into a writ about archbishops’ rights. Indeed, the rights listed in accounts of the trial are the same rights that appear in the writ. ⁵⁵ Accounts of the Penenden Heath trial also suggest that the participants were concerned with pre-Conquest precedents: the king commanded the elderly Bishop Æthelric of Selsey be fetched in a cart to comment on English law. ⁵⁶ During or just before these events in the 1070s, there seem to have been several campaigns to preserve and forge documents relating to property and papal letters at Canterbury. Copies of documents and possibly some forgeries were added to an ornate gospel-book in the 1060s and 1070s. ⁵⁷ The script of the writ is similar to, if not precisely the same as, writs copied in the ornate gospel-book during Christ Church’s attempt to corral or create records of their property and privileges during the late eleventh century. ⁵⁸ Robin Fleming argued that, additionally, a cartulary was compiled in the 1070s, and Brooks and Kelly have accepted that a Christ Church cartulary was at least compiled sometime in the early Anglo-Norman period. ⁵⁹ Such a context would also explain the links between some of the erased text and documents from St Augustine’s and Christ Church.

That being said, Christ Church’s legal wranglings – and possible contexts for alterations to the erased writ – did not end with the Penenden Heath trial. In the 1080s, the compilation of Domesday Book again caused the monks of Christ Church to assess their property. Around 1100 and after, some of these property assessments and details were copied into a manuscript now known as the Domesday Monachorum. ⁶⁰

HIS AND HERA

But while the MSI results cannot confirm whether Edward the Confessor confirmed this series of rights to Stigand or whether that was a later forgery – or even when such a forgery might have been made – the MSI results do reveal important evidence for post-Conquest legal strategies at Canterbury. The writ was altered one last time around 1100, when Scribe F subtly but importantly changed the text of the writ in LFC xxi 5 to link Canterbury’s rights to the whole community, and not just to an individual archbishop.

At first, Scribe F’s interventions in the writ seem puzzling, since MSI reveals that he echoed much of the same phrasing and claimed the same rights as the writ he partially erased. He did, however, change singular pronouns to plural pronouns. Although the erased text between lines 5 and 6 is too close to the visible text of line 5 to determine whether there are any ascenders, the erased word preceding agene menn looks like the singular [h]is, rather than the plural hera. The erased text thus seems to corroborate Nicholas Brooks’ hypothesis that Scribe F altered the text of LFC xxi 5 in order to change it from confirming rights just to Archbishop Stigand (in the singular) to the whole community (in the plural). ⁶¹

This switch – from individual archbishops’ rights to communal rights – relates directly to Archbishop Anselm’s struggles with William Rufus and Henry i in the 1090s. When Anselm was initially offered the archbishopric, he insisted that William Rufus restore all of Christ Church’s property and give him the archbishopric as Archbishop Lanfranc had held it. However, according to Anselm, Rufus later argued that nobles had held some lands from Canterbury in the pre-Conquest period and that he was able to appoint heirs to those nobles – and those Canterbury estates – during his reign. George Garnett has suggested that Anselm’s insistence that Rufus review Canterbury’s property might have backfired in some respects. ⁶² One wonders if the references to Canterbury’s agene menn in the erased writ might have influenced some of these debates. Writing to his mentor Hugh, archbishop of Lyons, Anselm characterised the situation thus:

The king gave me the archbishopric just as Archbishop Lanfranc held it until the end of his life, and now he is stealing from the church and from me what it and the archbishop freely held for so long, that he himself gave to me. I am certain that the archbishopric will be given to no one after me except in the condition I hold it on the day of my death. Nor, if another king should accede during my lifetime, will he bestow anything on me unless he finds that I am already holding it. ⁶³

Whether or not Anselm’s letter accurately depicts the reasoning of Rufus and his agents, it does highlight how Anselm’s thinking about property was developing. Anselm seems to have been particularly worried that Canterbury would permanently lose any lands or rights that he was not able to reclaim. Concerns about kings having the power to grant and withhold Canterbury’s land were also amplified by Anselm’s interest in reform and his connections to the reforming pope Urban ii and the ongoing Investiture Contest. As George Garnett has detailed, Anselm would eventually refuse to do homage to Henry i over these issues. ⁶⁴ These concerns seem to have changed the strategies Anselm and the monks of Canterbury used to try to secure their property. At some stage, Anselm and his associates switched from trying to claim land based on previous archbishops’ rights to claiming land as belonging to the community at Canterbury. They seem to have realised that land given to a church could be held more securely than land given to an individual archbishop. The community had continuity, whereas the ascension of new archbishops – or kings – could provide opportunities to challenge property that was held primarily by archbishops. ⁶⁵ The concept that churches’ property and rights could not be challenged, even if the heads of those churches changed, had a long history in England, ⁶⁶ and by the turn of the twelfth century, the members of Christ Church Canterbury seem to have been developing a variation on this theme, to cope with Anglo-Norman kings’ claims to rights and forms of land tenure. Anselm’s and his associates’ shifting strategies are demonstrated in Scribe F’s changes to the visible text of the writ, changing it from his rights to hera rights and from his men to hera men. These changes would have a long impact and shaped the wording and the claims of later writs, as Alger Doane has shown. ⁶⁷

Scribe F’s use of parchment attached to Edward’s seal was a significant riposte to some royal officials’ arguments, too. Anselm fretted that Rufus was able to justify his actions with reference to pre-Conquest precedent. Although Garnett argues that, in the 1090s, ‘Anselm appears not to have grasped the significance of Rufus’s appeal to the status quo TRE’, the multiple alterations of LFC xxi 5 suggest that the community at Christ Church appreciated the power of Edwardian precedents. Scribe F, and perhaps others, chose to reuse this particular piece of parchment and its attached seal, even though there were other options at their disposal. For example, in other cases Scribe F – or someone with a very similar hand – forged documents on reused pieces of parchment from a different era from the purported document he was creating: he forged a charter from the 1020s on erased parchment that possibly contains the palimpsest of a ninth-century diploma. ⁶⁸ By contrast, the LFC xxi 5 parchment seems to have been valued for its seal and was used for forged writs of Edward. This matches other evidence for the value of Edward’s seal in the post-Conquest period. Sealed writs were accepted as evidence of ownership during Domesday inquests, although Fleming has noted that juries’ testimony often relied on more than just documents. ⁶⁹ Edward’s seal itself was even forged in the late eleventh century, such was its usefulness in establishing documents’ significance. ⁷⁰

Scribe F’s alterations to the writ of Edward were of a piece with the rest of his programme of history writing and document alteration, even down to the partial erasure and rewriting of the text. ⁷¹ The same process is evident in a writ purporting to be of William i. The first line remains in a mid-eleventh century hand, but the rest was rewritten, possibly by Scribe F. ⁷² (Even then, the hand of the first line has been detected in another Canterbury document, pertaining to an early eleventh-century dispute between Christ Church and St Augustine’s, suggesting Canterbury’s interference – if not forgery – of some existing visible text. ⁷³ ) David Bates has argued that John Joscelyn copied a document that contained a copy of the previous, erased version of this writ’s text, since the opening is very similar. If so, Bates notes that ‘the improvements made in the twelfth-century rewriting were fairly small’, just as the erased text of LFC xxi 5 is very similar to Scribe F’s altered, visible text. ⁷⁴ The second version of the writ omits one of the references to French and English thegns and a phrase about Christ Church’s lands in King Edward’s day – possibly, Bates suggests, in reaction to the Penenden Heath trial. ⁷⁵ It also changes a general reference to mention Christ Church specifically. But while those alterations may have involved only a few words, Scribe F’s changes to writs signal a larger transformation in legal and social thinking in Canterbury around 1100, as individual and communal responsibilities began to be defined more sharply.

The shift in attitudes towards communal versus individual responsibilities was not limited to thinking about Christ Church’s property and rights, either. Recent scholarship has noted a post-Conquest shift in attitudes to collective versus individual responsibility in the various manuscripts of the Anglo-Saxon Chronicle. As Alice Jorgensen and Emily Winkler have noted, the F recension of the Chronicle focuses on individual, royal responsibility for invasions, unlike earlier recensions which blame invasions on collective sin. ⁷⁶ This complements and mirrors the shift from the individual to the collective in Scribe F’s alterations of LFC xxi 5: communities could not be blamed for individual leaders’ failures. To be secure, rights had to be given to the whole community, not just to archbishops, and especially not just to the controversial Stigand. Given the worldview found in Scribe F’s historical writings, as well as his forged documents, it was worth altering a whole writ in order to change key singulars to plurals. ⁷⁷ As such, LFC xxi 5 is a witness to a wider cultural shift.

CONCLUSIONS

While the extent of the erasures makes it difficult to provide definitive answers about all phases of LFC xxi 5’s production and alteration, MSI analysis has raised some possibilities that are worth considering. First, the MSI analysis seems to confirm Brooks’ suggestion that scribe F changed singular pronouns to plural pronouns when he altered the writ. This is significant because it shows how the community at Christ Church literally rewrote their approach to protecting their property in the decades following the Norman Conquest. This change can be seen both in this individual piece of parchment and in the series of writs that Christ Church possessed – and at least partially fabricated – by the end of the eleventh century. ⁷⁸ Whereas writs were received or forged on behalf of the archbishop in the early eleventh century, by the 1090s and the early twelfth century the monks of Christ Church switched to emphasising the rights of the institution as a whole. LFC xxi 5 was altered to show Edward giving rights to the whole community at Christ Church at a crucial moment, shortly after Anselm’s attempts to protect Christ Church’s property in the 1090s, in reaction to claims made by kings like William Rufus and Henry i. Scribe F’s activities may have been part of a wider change in strategies to protect the church’s property. These alterations reveal conceptual shift around the roles, rights and responsibilities of individual leaders versus communities.

Additionally, MSI analysis revealed that all parts of the writ – including the first three visible lines of text – had erasures underneath them. This suggests either that the writ was erased and rewritten more times than had previously been thought, or that the original writ was copied on reused parchment. If the first possibility is correct, then the writ might originally have pertained to lands that were secured by the 1070s or at least by the Domesday survey. After their land was secure, the monks of Canterbury might have refocused on their rights, and reused a document that was no longer necessary. Alternatively, if the original writ was composed on reused parchment, then it might clarify the production of royal writs in pre-Conquest England. Such evidence would have implications for the volume of work, economics or working conditions of the scribes attached to Edward the Confessor’s administration. Whether the writ was copied onto reused parchment or its contents quickly became obsolete and could be reused, LFC xxi 5 highlights the ephemerality of writs and their extensive production.

While the MSI results show how some thinking changed in the decades after the Conquest, LFC xxi 5 also highlights continuity over the Conquest in at least one respect: the continued value that the monks of Christ Church – and possibly others – attributed to seals of Edward the Confessor. LFC xxi 5’s alterations reinforce the sense that precedents from Edward’s reign were perceived as a powerful legal force for decades after the Conquest.

This small piece of parchment provides a rare insight into a type of document that may have been the most numerous type of royal document produced in eleventh-century England. Happily, the advent of technology such as MSI analysis is creating new opportunities for making the most of such fragmentary survivals and for illuminating – literally – a dramatic and formative period in legal, intellectual and governmental history.