The Staffordshire Hoard inscription

Elisabeth Okasha (University College Cork)

Revised version - October 2011

The inner face of the inscribed strip has now been cleaned by conservators from Birmingham Museum and Art Gallery, and microscopically examined by Elisabeth Okasha, who has provided this updated version of her Symposium paper.

Description

The inscribed strip (StH 550) is made of gold alloy and is now folded over in half on itself. When straight it would have measured 179 mm in length, 15.8 mm in width and 2.1 mm in thickness. On the outer or front side of the strip is Text 1, which is set in two lines. The decorated gem-setting at the beginning of the text, and the incised animal head at the end, indicate that the text is complete. Text 1 is likely to be primary, that is, the object seems to have been designed from the outset to be inscribed. The letters are formed by incisions which were then filled with niello.

On the inner side or back of the strip is Text 2, also in two lines, which is set upside down with respect to Text 1. The letters are incised but are not filled with niello. The inside of the strip contains an identical animal head to the outside, but it too is again incised only. Text 2 contains substantially the same text as Text 1.

There are holes and a pin on the strip, which were presumably used for fastening it on to some larger object. This larger object seems likely to have been flat, not curved, and fairly substantial. Since all the other objects from the hoard are battle-related, it is tempting to see the inscribed strip as having formed part of a shield or weapon. However parallels are scarce. A book-cover has been suggested, but it is hard to see what that has to do with a battle-field. More promising suggestions are that the strip might have come from some ecclesiastical object, perhaps a shrine, a reliquary or a cross, any of which might have been carried into battle.

The only other flat inscribed strip known to me which is in any way similar is one from North Petherton in Somerset (Okasha 2004, no. 230, pp. 243-4). However, it is not particularly helpful as a parallel: it is much later, from the 11th or 12th century, is made from copper alloy, not gold, has different texts on front and back, and the language used is Old English not Latin. Moreover, its function is equally unclear.

Texts

Transliteration

The letters are transliterated as capitals with deliberate spaces in the texts retained. An underlined letter indicates that the letter, though damaged, is legible. An underlined letter inside square brackets indicates that the letter is damaged but the reading is probable. A dot inside square brackets indicates that one letter is lost, and three dots indicate the loss of several letters. In addition, ':' indicates a deliberate single dot in the text and '/' indicates a ligature.

Text 1, the text on the outer side, reads:

In line 1, the first letter seems to have contained a top curve and is most probably S. In line 2, the first letter is probably an F and the penultimate letter appears to be U but of an unusual shape, resembling U ligatured to I.

There are two deliberate dots in the text, one each around the letters DNE, the abbreviation of domine. These dots could have been used to indicate word separation: inconsistent use of word separation is indeed common in Anglo-Saxon inscribed texts. Alternatively, or perhaps as well, the dots could have been used to highlight the nomen sacrum. The deliberate space in line 2 of the text, preceding the letter Q, may indicate word separation, or may have been an attempt to better fit the remaining letters into the space available. The ligature E/T is in the form of the usual manuscript abbreviation for et.

Divided into words, with abbreviations expanded and likely letters assumed, the text reads:

[s]urge domine disepentur inimici tui et [f]ugent qui oderunt te a facie tua

Text 2, the text on the inner side, reads:

In line 1, the first lacuna originally contained 2 or 3 letters consisting of several vertical lines, so a reading of INI is quite possible. In line 2, the lacuna probably contained two letters A, one following the other. The penultimate lost letter could be A or D or T while the final letter is totally lost.

There are two abbreviation marks, one on the N of DNI and one on the U of FUGIU[NT]. The latter seems unnecessary, unless the N is not really to be read there. There are seven deliberate spaces: on either side of DNE; between the O and D of ODE[R]UNT; on either side of AF and at the end of AF ACIE (for a facie); preceding NOS. The ligature E/T is again in the form of the usual manuscript abbreviation for et.

Divided into words, with abbreviations expanded and likely letters assumed, the text reads:

surge domine disepintur [...mic]i tui et fugiu[nt] qui ode[r]unt te a facie tu[a] [a]diuie nos[.r.]

It seems clear that this is substantially the same text as Text 1, but a little longer, and less carefully executed. There are two specific differences in spellings. The first is that Text 1 has DISEPENTUR, while Text 2 has DISEPINTUR; these correspond to classical Latin dissipentur. In so far as this is relevant, the spelling of the word in Text 1 is closer to the classical spelling, but non-classical spellings are common in Anglo-Saxon inscriptional texts. The second difference occurs in the verb form corresponding to classical Latin fugiant. Text 1 has [F]UGENT while Text 2 has FUGIU[NT]. In this case, and again in so far as this is relevant, Text 2 has a form closer to the classical spelling.

Other differences between the two texts, and the relationship between them, are discussed below, where it is argued that Text 2 is likely to be a first, practice attempt, Text 1 the finished version of the text, ready for public show.

Source of the text

The text to be inscribed would have been chosen by an educated person, presumably after consultation with the person commissioning the inscribed object. The commissioner must have been well-to-do and, presumably, someone with a large army, perhaps a king or local leader. The educated person is likely to have been a cleric or religious, that is a priest, monk or nun. He/she would then have written the text on a piece of vellum or on a wax tablet, and this exemplum would have been copied by the goldsmith who was, we may presume, illiterate.

The inscribed text is a rendering of a well-known Vulgate text, which appears in the following form in Numbers 10, 35:

cumque elevaretur arca dicebat Moses surge Domine et dissipentur inimici tui et fugiant qui oderunt te a facie tua, 'When he had lifted up the ark, Moses said "Rise up, Lord, and may your enemies be dispersed and those who hate you flee from your face"'.

As noted above, there are some minor spelling differences between the Vulgate text and Texts 1 and 2 on the strip. There is also the omission of et following domine. This latter difference could have been an error made by the goldsmith in copying the exemplum or could alternatively have appeared in the exemplum. The spelling differences could also have arisen due to errors in copying. However, since there are frequent examples of non-classical spellings in medieval Latin, they might have originated in the exemplum. The exemplum might have been written from memory rather than copied from a manuscript.

Although it is likely that the person who wrote the exemplum had the text from Numbers in mind, it is also possible that he/she was thinking of a similar text from the book of Psalms. Indeed, the Psalms might have been better known to him/her since the Psalms were chanted daily throughout the year in the monastic liturgy. The following text appears in Psalm 67, 2 in the Vulgate:

exsurgat Deus et dissipentur inimici eius et fugiant qui oderunt eum a facie eius, 'Let God arise and his enemies be dispersed and those who hate him flee from his face'.

The differences occur in the first two words and also in the use of the third person pronouns ('his', 'him'). The version in the Old Roman Psalter is similar but has a slightly different word order, suggesting that it is less likely to be the source of the text.

This text, in one form or another, was quoted fairly often in medieval documents. A relevant instance occurs in Felix's Vita Sancti Guthlaci, which is of interest both because of its date and because of its geographical location. Guthlac was born around 674, probably of Mercian royal lineage, and his abode as a hermit in the Fenlands is likely to have been in the area bordering Mercia and East Anglia (Colgrave 1956, 1-3). Felix probably composed his Vita Sancti Guthlaci 'somewhere between 730 and 740' (Colgrave 1956, 19).

Felix quoted both versions of the text in the Vita. The version from Numbers occurs when Guthlac meets Æthelbald, who subsequently became king of the Mercians, ruling from 716 to 757. Guthlac foretells the future to Æthelbald and in the process says et fugient a facie tua qui te oderunt, 'those who hate you shall flee from your face', with one manuscript using the spelling fugiunt (Colgrave 1956, 150-1). This alternative spelling is the same as that used in Text 2.

The version from Psalm 67 occurs earlier in the Vita, when Guthlac is vanquishing devils who have appeared to him in a vision. Felix says: Tum vero vir Dei tandem ... velut prophetico ore sexagesimi septimi psalmi primum versum psallebat: Exsurgat Deus, et reliqua, 'Then at length the man of God ... sang the first verse of the sixty-seventh psalm as if prophetically, "Let God arise", etc' (Colgrave 1956, 110-11). Colgrave notes that instead of et reliqua, 'etc', some manuscripts have the completed text et dissipentur inimici eius (Colgrave 1956, 111, note to lines 37-8).

Script

The script used in both the inscribed Texts 1 and 2 is insular majuscule. While the epigraphic use of this script obviously bears a relationship to its manuscript origin and use, it is hard to argue directly from manuscripts to inscriptions. The tools used, the conventions of use, and the personnel involved, would all have been quite different in the two mediums. A literate scribe, wielding a pen, might well not have produced the same letter-forms as an illiterate goldsmith copying an exemplum.

As is common with insular majuscule when used in inscriptions, the texts include the occasional capital form. For example, in Text 1, three out of the five examples of N, and all three instances of R, are capital. The G used in both texts is the insular minuscule form and the A is the uncial form. This inconsistency in letter form is entirely usual in Anglo-Saxon inscriptions, whatever script is used, and may also be seen in Text 1 in some letters E being high, some not. There is also apparent inconsistency in the use of word separation in Text 1, which again is entirely usual in inscribed texts of all scripts and all dates within the Anglo-Saxon period.

There are some specific differences in the letter-forms used in Text 1 and Text 2. For example, although both texts use insular G, the second example in Text 2 (in FUGIU[NT]) is rather an odd shape. Again, while both texts have capital N in DNE and ODERUNT, in DISEPENTUR, Text 1 has an insular N, whereas Text 2 has a capital. Text 2 has a capital N in NOS, where there is no corresponding word in Text 1. Text 2 has a capital S in SURGE and an insular S in DISEPINTUR; Text 1 has a capital S in DISEPENTUR and, possibly, an insular S in SURGE.

What is quite unusual is the considerable use of large open serifs on some letters in Text 1, used alongside ordinary serifs and half serifs. For example, in the word qui in Text 1, the lower leg of the Q has a large open serif while the curved end has a small ordinary serif. Nevertheless, the use is not consistent. For example, of the eight instances of the letter I (capital i), seven contain a large open serif at the top of the stem while one (in tui) does not.

Text 2 does contain one or two serifs, but there are many fewer of them and they are generally less ornate, although one open serif does occur on the I of FUGIU[NT].

The fact that these open serifs are not used consistently in Text 1, and probably not at all in Text 2, suggests that they are likely to be an artistic flourish on the part of the engraver, rather than being due to the difficulties of engraving on gold. These open serifs may resemble manuscript serifs more than epigraphical serifs. However, large open serifs of this sort do occur sporadically on inscribed texts from elsewhere, for example on the stone from Coldingham, Berwickshire, probably dating from the 8th or the 9th century (Okasha 1992, no. 188, p. 43) and one of the stones from Carlisle, which is probably 8th century in date (Okasha 1992, no. 24, p. 60). Moreover, it is the frequency of their use in Text 1 that is remarkable.

There are some other differences in script between Text 1 and Text 2. The letters in Text 2 are less carefully executed than those in Text 1, are not as carefully spaced and show a greater variation in letter height. There are more deliberate spaces between letters in Text 2: Text 1 has one such space while Text 2 has seven, including some in the middle of words. Text 2 has no dots in the text while Text 1 has two, around DNE, but Text 2 contains two abbreviation marks, in DNE and FUGIU[NT]; the latter is otiose, if the reading as given is accurate.

Another clear difference between the two texts is that Text 2 contains extra letters at the end. These appear to read [a]diuie nos[.r.] unless the first letter is to be read as part of the final A of tu[a]. These extra letters do not seem to be a continuation of the Vulgate text of Numbers 10, 35, nor indeed of Psalm 67, 2.

There are two general possibilities. The first is that these are practice letters, in line with the more casual nature of the whole of Text 2. Might the goldsmith have been trying again to engrave letters that he/she had found hard the first time? In this case these letters were not on the exemplum but were copied from the rest of Text 2. Alternatively, the letters are not random practice letters but did appear on the exemplum. In this case a well-known phrase might be expected. However, the reading of the letters is not certain and it is therefore hard to be sure what text, if any, the scribe had in mind. Some possibilities are: dei nostri; dme nostri; dne nostri; ad ...nostri, as in a phrase like ad patrem domini nostri; ad ... nos dei, as in a phrase like ad nos arcam dei.

If the extra letters did appear on the exemplum, we then have to enquire why they were omitted from the finished text, Text 1. It is possible that when the final text came to be put on the outer side of the strip, there was found to be insufficient room for further letters at the end of Text 1. However, since Text 1 appears to be complete in itself, and indeed finishes at the end of verse 35 of Numbers 10, it seems more likely that the extra letters are practice letters. The use of texts consisting of practice letters can be paralleled from a number of Anglo-Saxon inscriptions on portable objects, such as Waltham Abbey (Okasha 1983, no. 178, p. 100), Dublin V (Okasha 1992, no. 190, pp. 44-5) and Nassington (Okasha 2004, no. 229, pp. 242-3).

Suggested relationship between the inner and outer sides of the strip

The two texts, Text 1 on the outer side and Text 2 on the inner side, are clearly versions of the same text, yet they are not identical in either spelling or letter-forms. The two animal heads are also very similar, so similar that they clearly had one exemplum, yet they too are not quite identical.

I start with two assumptions. The first is that the goldsmith was illiterate and therefore, as assumed above, would have needed an exemplum from which to copy. The second is that it would not have been at all easy for a scribe or the goldsmith to copy one text directly from the other, due to the need to keep turning the strip over. The letter-forms seem much too similar to make this likely.

Being illiterate, the goldsmith needed an exemplum to copy, for both the text and the animal head. I suggest that the goldsmith wanted to practise engraving the letter forms and the animal head and so copied them roughly from the exemplum on to the inner side of the strip, knowing that the inner side would be invisible when the strip and its artefact were assembled.

The exemplum is likely to have been written and drawn by a scribe on a wax tablet or a scrap of vellum. I suggest that the exemplum contained the spellings and letter-forms of Text 2 but did not contain the extra letters at the end of Text 2. The exemplum might have had the text written straight out or divided into two lines as it was to appear in Text 1. It may have been that on the exemplum the scribe did not space the text carefully, nor write the letters with care, nor add serifs to most of the letters and the goldsmith reproduced this in Text 2. Alternatively, the exemplum may have contained a more finished text which the goldsmith copied roughly in order to gain practice. The extra letters at the end of Text 2 may represent a second attempt at particular letter-forms, perhaps those that had proved troublesome to him/her on the first attempt.

It is possible that, after making his/her first attempt in Text 2, the goldsmith felt able to proceed straight to engraving the final version on the outer side of the strip. Alternatively, and perhaps more likely, the scribe could have carefully copied the text and the animal head from the exemplum on to the outer side of the strip, using pigment (a possibility proposed by George Henderson). In either case some slight changes were made from what was on the exemplum: two small spelling changes, some slight changes in letter-form, the omission of the abbreviation marks, the addition of serifs, and the addition of two dots. The spelling changes and letter-form changes may have been deliberate or accidental. The text was divided into two lines so as to fill the available space, either copying the two lines as on the exemplum, or spacing them carefully with pigment on the strip. Having engraved the text and the animal head on to the outer side of the strip, the goldsmith washed off the pigment (had it been used), and filled the incisions with niello.

Parallels

Anglo-Saxon inscriptions using insular majuscule script are not very frequent in occurrence: I know of only 18, comprising about 8% of the total number of inscribed texts. All but three are dated 8th to 9th century, the three being either much later or, in one case, undatable. Of these 18, five texts can be isolated where the script used resembles fairly closely that of the inscribed strip. They are the stones from Ardwall, Kirkcudbrightshire, (Okasha 1971, no.3, p. 48), Dewsbury, stone no. I (Okasha 1971, no.30, pp. 65-6), and Yarm, North Yorkshire (Okasha 1971, no.145, p. 130). There are also the lead plates from Kirkdale (Okasha 2004, no. 225, pp. 238-9, Kirkdale II) and from Flixborough (Okasha 1992, no.193, pp. 46-7, Flixborough II).

The Yarm, Kirkdale and Flixborough letter-forms all contain at least some large serifs, although not of the same open shape as those on the Staffordshire strip. All also contain uncial A, capital R, and have inconsistency of letter-form. In general terms, the closest parallel in terms of script is the lead plate from Flixborough. In addition to the features already noted, the Flixborough text uses insular minuscule G as well as both high and low forms of E. The Flixborough lead plate is dated on linguistic, epigraphic and archaeological grounds to the 8th to the 9th century.

(The Flixborough lead plate)

Anglo-Saxon inscriptions that contain two copies of the same text are extremely unusual; indeed I know of only three examples. One can be disregarded, a lost stone from Marcham, Berkshire, of which no illustration exists (Okasha 2004, no. 228, p. 242). The other two are both late, 10th or 11th century, and are both made of bone. One is a trial piece from the London Guildhall site which contains the personal name Aelburh incised twice, once on each side (Okasha 1992, no. 198, pp 50-51). The likelihood is that both names represent practice lettering. The other is probably a weaver's sword, from Wallingford, Berkshire (Okasha 1971, no. 118, p. 119). The text, the second version of which is repeated immediately below the first, reads + Eadburh mec ah ..., '+ Eadburh owns me ...'. In this case, the lower version of the text appears likely to be a copy of the upper version. Neither of these seems to furnish a very useful comparison.

But there is one parallel of interest. The Flixborough lead plate which, as mentioned above, is similar in script to the Staffordshire strip inscription, contains traces of what are likely to be letters beneath the finished text. These letters are not now legible, but it has been suggested that they could have formed an initial stylus sketch for the finished inscription (Brown and Okasha 2009, 138-41).

It seems likely that Text 1, the finished and visible text, set on the outer side of the strip, was addressed to God and perhaps also to the saints in heaven. It is most unlikely at this date that many people, unless they were in the church, were literate. Text 2, on the inner side, would have been invisible and, if it consisted merely of practice letters, then this is sufficient explanation for it. If Text 2 was meant to be read, we have to enquire who its intended readers were. It may be, however, that to discuss readership of either text is to miss the important point that, in Anglo-Saxon England, the written word was itself considered to be a symbol of power and authority, whether or not what was written could be understood.

Date

It is not easy to date any Anglo-Saxon inscribed text on the basis of the script alone. However, the following features of the script used suggest a date in the 8th century:

• the use of insular majuscule
• the considerable use of large open serifs
• the similarity of the script to the five inscribed texts mentioned, all of which are of a similar date
• the particular similarity with the script of the Flixborough lead plate.

To this may be added the fact that niello work in general, and in particular in inscriptions, is typical of the 8th to the 9th century.

If it were to turn out that all the other objects in the hoard dated from the 7th century or earlier, then the strip would represent one of the latest of the objects in the hoard. In my view, it is hard to argue that the strip dates from much before AD 700.

(X-radiograph of the Flixborough plaque, showing traces of possible letters beneath the finished text)

It seems likely that Text 1, the finished and visible text, set on the outer side of the strip, was addressed to God and perhaps also to the saints in heaven. It is most unlikely at this date that many people, unless they were in the church, were literate. Text 2, on the inner side, would have been invisible and, if it consisted merely of practice letters, then this is sufficient explanation for it. If Text 2 was meant to be read, we have to enquire who its intended readers were. It may be, however, that to discuss readership of either text is to miss the important point that, in Anglo-Saxon England, the written word was itself considered to be a symbol of power and authority, whether or not what was written could be understood.

Date

It is not easy to date any Anglo-Saxon inscribed text on the basis of the script alone. However, the following features of the script used suggest a date in the 8th century:

• the use of insular majuscule
• the considerable use of large open serifs
• the similarity of the script to the five inscribed texts mentioned, all of which are of a similar date
• the particular similarity with the script of the Flixborough lead plate.

To this may be added the fact that niello work in general, and in particular in inscriptions, is typical of the 8th to the 9th century.

If it were to turn out that all the other objects in the hoard dated from the 7th century or earlier, then the strip would represent one of the latest of the objects in the hoard. In my view, it is hard to argue that the strip dates from much before AD 700.

References

• Brown, M. P. and Okasha, E., 2009. 'The Inscribed Objects', in D. H. Evans and C. Loveluck (eds), Life and Economy at Early Medieval Flixborough, c. AD 600-1000: The Artefact Evidence (Excavations at Flixborough 2, Oxbow Books, Oxford), 138-141
• Colgrave, B.,1956. Felix's Life of Saint Guthlac (Cambridge University Press, Cambridge)
• Okasha, E., 1971. Hand-List of Anglo-Saxon Non-Runic Inscriptions (Cambridge University Press, Cambridge)
• Okasha, E., 1983. 'A supplement to Hand-List of Anglo-Saxon Non-Runic Inscriptions', Anglo-Saxon England 11, 83-118
  
• Okasha, E., 1992. 'A second supplement to Hand-List of Anglo-Saxon Non-Runic Inscriptions', Anglo-Saxon England 21, 37-85
• Okasha, E., 2004. 'A third supplement to Hand-List of Anglo-Saxon Non-Runic Inscriptions', Anglo-Saxon England 33 (2004), 225-81