Eadgifu, Governor of Kent, in a Lost Charter of Edgar

Abstract

The witness-list to a charter of Edgar the Peaceable, now lost but quoted by William Somner, designated Eadgifu, Edgar’s grandmother, as in Cantia etiam gubernator. Although doubts about the authenticity of Somner’s source are now unlikely to be resolved, a study of Eadgifu’s life and comparison with contemporary cases of shire administration and female secular power show that, at least in theory, the claim that she was governor of Kent is more credible than it may seem.

The use that historians should make of early medieval English charters is not always a straightforward matter. Much depends on the context of the charter’s creation or its preservation and the degree of authenticity attributable to it. Those charters that survive as a full text, whether as original copies or later transcripts, can be difficult enough to assess but special problems are posed by those charters that no longer exist at all but survive only as allusions or paraphrases in other documents or historical works.

One such ghostly charter was discussed by William Somner in the first edition of his Antiquities of Canterbury, published in 1640.¹ In a section devoted to the history of municipal government, Somner had occasion to discuss the portgerefa, literally ‘town-administrator’, the office that would be re-named bailiff after the Norman Conquest. He cited several examples of port-reeves of Canterbury derived from primary sources, including a lengthy quotation from the witness-list of a charter that, since the publication of his book, has gone missing. The excerpt that Somner quoted is now the only portion of the text still available for study:

… in the yeare 956. to a Deed of the sale of a parcell of Land in Canterbury to one Ethelstane, by two Knights Ethelsi and Wlfsi, is the subscription (amongst other witnesses) of one Hlothewig Portgerefa, on this wise: viz. After King Edgar, Queene Eadgive (who writes herselfe, in Cantia etiam Gubernator), and some others, it followes. + Ego hloðwig portgerefa. [and] se hiored to xpescircean. [and] se hiored to sce Agustine. [and] ða ðreo geferscipas innan burhwara utan burhwara miccle gemittan.²

This ghostly charter raises several interesting points. Some uncertainty must attend its date. 956 is impossible for a charter of King Edgar. His brother Eadwig had become king in 955. Edgar would not have been referred to as a king until 957, when a schism in the kingdom made him king of Mercia and Northumbria, and he did not become king of all England until 959, so would not have witnessed a charter dealing with land in Canterbury before that date. It is unlikely that Somner’s Edgar might be an error for Eadwig, since Eadgifu appeared in only one charter of that king (and that of disputed authenticity) but did appear in several charters of Edgar.³

The likeliest script for Somner’s charter would have been Phase IV Square minuscule and the type used by Somner’s printer attempts to imitate this, from the chrismon to gemittan, even using wyns.⁴ This suggests that Somner was working off either an original or an early copy. The date of 956, however, is given in Arabic numerals before the transcription proper, so the error need not belong to the original charter. Nicholas Brooks and Susan Kelly suggested three possible solutions: 956 might simply be a printer’s error for 965 or 966, either of which would be an acceptable date for a charter featuring both Edgar and Eadgifu; the original charter might have been drawn up in 956 and later confirmed by Edgar, a nuance obscured by Somner’s selective quotation of the text; or the charter might originally have been undated but was endorsed erroneously by a later medieval scribe.⁵

The first suggestion is attractively simple but unfortunately is the only one of the three that can actually be ruled out by comparison with Somner’s own copy of The Antiquities, which contains annotations that he made in preparation for a revised edition.⁶ The relevant page contains several alterations but the date of 956 is untouched, so Somner must have been satisfied with how it had been printed. The second suggestion seems unlikely, since a careful reading of Somner’s description shows that he assigned not the original property transaction but the subscription to 956. A fourth possibility is that the Roman numerals of the original date were simply misread, either by a very early copyist or by Somner himself, perhaps by mistaking xi for vi. In that case, Somner might have seen the numerals dcccclxi (giving the perfectly credible date of 961) but misread them as dcccclvi (956) The problem with this explanation is that x and v appear quite distinct in other original charters of the period, with the left lower trait of the x reaching below the line such that it would be difficult to mistake it for v. This leaves Brooks’ and Kelly’s third suggestion, that of an erroneous date endorsed on the charter by a later scribe. In the absence of the original copy, this explanation is unverifiable and feels rather unsatisfying but it is credible and seems to be the only possibility left.

The second point to consider is the resolutely local, Kentish context in which this charter was produced. Hlothwig the port-reeve makes only one other appearance, in a charter that survives as an original single-sheet diploma.⁷ This charter, dated to 968, concerns an exchange of land in Heronden, near Tenterden. Like Somner’s charter, it was confirmed by King Edgar. Its witness-list is a most extraordinary creature, boasting the king, Archbishop Dunstan, Abbot Sigefrith, Ælfstan and Birhtsige praepositi, Hlothwig portgerefa, seven ministri, nine rustici, ‘se hioræd to Cristes ciricean. [and] se hioræd to sancte Agustine [and] ða þreo geferscipas innan burhwara [and] utan burhwara [and] micle gemettan’, Ælfsige burþen, the hioræd of Appledore and five witnesses who lack a title.

Somner’s charter appears to have been of a similar species. For one thing, Hlothwig attested both and in both he used his vernacular title, rather than a Latin translation. Furthermore, both charters are also distinguished by corporate attestations from the convents of Christ Church and Saint Augustine’s and from the three fellowships: that within the borough, that outside the borough and the ‘many guests’. The three fellowships are also found in two earlier Canterbury charters.⁸

The Heronden charter’s idiosyncrasies are explicable from its local provenance. It was produced not by a royal scribe attending a national assembly but by a Kentish scribe attending a Kentish assembly (albeit one graced by the unusual presence of the king) and interested in Kentish people and institutions.⁹ It would be naive to assume that it was the only local charter produced in Canterbury in this period, or the only charter that its scribe wrote. It is perfectly credible that other charters produced at Kentish assemblies, especially those produced by the same scribe, would have followed a similar style. That Somner’s charter is another local production is confirmed by his dialect: the spellings Ethelstane and Ethelsi show the Kentish preference for E- over Æ-.

That the scribe of Somner’s missing charter was a local adds to the significance of the title that he accorded to Eadgifu. Brooks and Kelly confessed themselves at a loss to explain why Eadgifu was called gubernator of Kent.¹⁰ The remainder of this paper will be devoted to testing the credibility and examining the implications of this, the most intriguing point raised by Somner’s charter.

The simplest explanation is that Somner misquoted his source, that he was tricked by eye-skip into assigning to Eadgifu a title, in Cantia etiam gubernator, which really belonged to a different testator, presumably one of the ‘some others’ to whom he alluded between Eadgive and Hlothewig. Two observations support this hypothesis: first, the gender of gubernator is wrong for Eadgifu; secondly, in Somner’s personal copy of The Antiquities the whole section in parenthesis after Eadgifu’s name has been crossed out and accordingly it does not appear in the second edition.¹¹ The deletion is the only significant alteration that Somner made to this page, the others being minor corrections of spelling or syntax. Frustratingly, he included no marginal note to explain the edit. These two points will be addressed in reverse order.

For an academic to make a clumsy error in print is, unfortunately, not unknown but the corollary of this explanation, namely that someone else was in Cantia etiam gubernator, loses credibility on examination. If the title attributed to Eadgifu were a common one, such as episcopus, dux or minister, and were without a territorial designation, then this would be an obvious case of eye-skip, but casually assuming that there just must have been some other gubernator is dangerous when nothing like the phrase quoted appears in any other known charter. That might be partly explained by the scarcity of local productions in this period, making it difficult to know what was normal in a charter of this nature but it may be significant that the only other charter comparable to this one in date, context and style, the Heronden charter, lacks both the gubernator in Cantia and Eadgifu.

The eye-skip explanation also fails on a priori grounds: first, Somner could simply have inserted the correct gubernator’s name in the revised edition. The amendment would have taken only a few words, yet instead he chose to delete the whole entry, which implies that the problem with it was something more complicated than misattribution. Secondly, the title’s attachment to a queen was apparently what originally attracted Somner’s attention, in which case it is unlikely that he would have made an error transcribing it. Perhaps the simplest explanation for Somner’s change of mind is that, since his focus was on the port-reeve, he had decided on reflection that Eadgifu the gubernator, a puzzle that, however intriguing, he was unable to explain or explore further, was an unwelcome distraction.

The masculine gender of gubernator may also not be as serious a problem as it seems. It was not uncommon for women to use masculine titles if they were performing a traditionally masculine role. In 990, Empress Theophanu of Germany (serving as regent for her son Otto III), who usually styled herself imperatrix, issued a charter in her own name calling herself ‘Theophanius gratia divina imperator augustus’.¹²

This usage also appeared in literature. Henry of Huntingdon rhapsodised about Æthelflæd, lady of the Mercians, that ‘Hec igitur domina tanta potentie fertur fuisse, ut a quibusdam non solum domina uel regina, sed etiam rex uocaretur, ad laudem et excellentiam mirificationis sue.’¹³ To hammer home his point, he wrote a brief poem that is deliberately ambiguous about her sex. One line in particular sums up his point: ‘Tu regina potens rexque trophea parans.’¹⁴

It was not unusual for a woman performing what was perceived to be a man’s role to use a grammatically masculine title. What was unusual was for a woman to perform what was perceived to be a man’s role at all, so the alleged case of Eadgifu gubernator deserves to be closely scrutinized.

Eadgifu herself provided us with valuable information about her early life in a written statement that she made c. 959 for the (unnamed) archbishop and convent of Canterbury.¹⁵ According to this, her father was Sigehelm, who left her some land just before he and the men of Kent went to battle at the Holme, where he was killed. This information identifies her father with the Kentish ealdorman of that name, whose death leading his shire against the Danes at the Holme in 904 is vividly described in the A, B, C and D manuscripts of the Anglo-Saxon Chronicle,¹⁶ by the Annals of St Neots¹⁷ and by Æthelweard the Chronicler, ealdorman of the Western Provinces, under the year 902.¹⁸

The date of Eadgifu’s birth is unknown but Æthelweard recorded that Sigehelm was a young man (lanugine tenera – ‘soft of beard’) when he was killed, so his daughter must have been a child at the time. This seems to be confirmed by her tenurial arrangements. Putting his affairs in order before the battle, Sigehelm paid off a debt to one Goda, reclaimed an estate at Cooling that he had pawned pending payment of the debt and left the estate to Eadgifu. After Sigehelm’s death, however, Goda claimed that the debt had not been repaid and continued to hold Cooling, while one Byrhtsige Dyring ceaselessly claimed the estate from him, presumably on Eadgifu’s behalf. Finally, after six years of vicarious litigation, Eadgifu appeared in court herself, presumably now an adult (though she still need only have been in her mid-teens), and swore that the debt had been repaid. Goda still, however, would not give up, until some of Eadgifu’s friends interceded for her with King Edward and he deprived Goda not only of Cooling but of all his estates as well (most of which Eadgifu generously returned to him).

Surprisingly, Eadgifu’s marriage to King Edward is not mentioned in her statement but the fact that she needed friends to intercede with him for her c. 910 implies that they were not yet personally acquainted. It is also unknown why Edward married her. He already had sons by two previous marriages, so the need to produce an heir cannot have been a factor. The marriage is likely to have been political, an attempt to curry favour with the powers in the south-east whilst Edward was absent campaigning in the Midlands.¹⁹

The marriage seems to have occurred c. 920 (so about ten years after they first met over the Cooling affair): Edward’s previous wife, Ælfflæd, had been consigned to a nunnery by 918,²⁰ and Eadgifu’s first son Edmund was probably born in 921.²¹ Eadgifu would give Edward another son, Eadred, before his death in 924. Her movements during Æthelstan’s reign are unknown but the accession of her own son Edmund in 939 brought her to prominence. The earliest charters that she attested date from 940, where she appears immediately after Edmund as eiusdem regis mater. ²² She remained prominent throughout the reign of her second son Eadred (946–955) but she was one of the many who fell from favour under her grandson Eadwig (955–959), when, as she complained to the archbishop, she was deprived of all her estates.

The accession of her second grandson Edgar restored her to her estates and her former prominence.²³ She attested three of his surviving charters (though not all are of unquestioned authenticity)²⁴ and was the beneficiary of another.²⁵ It has even been suggested that she may have retired to a nunnery.²⁶ Her last attestations were in 966, by which time she must have been aged around seventy.²⁷ She had died by 967, when Edgar granted part of Meon, an estate that he had formerly given to Eadgifu, to his maternal grandmother Winflæd.²⁸ A later medieval Christ Church obituary list gives the date of her death (sans year) as 25 August.²⁹

This was the woman who, at least in her twilight years, apparently served as gubernator of Kent. If this is true, what might being gubernator have involved? The natural assumption is that the office was comparable to that of ealdorman. The last two ealdormen of Kent as such were Eadgifu’s father Sigehelm and his co-ealdorman Sigewulf (the common prototheme of their names suggests that they were related), who was killed alongside him at the Holme. It was conventional for sons to be appointed to their father’s ealdormanries but Sigewulf’s son Sigeberht was also killed in the battle. In the absence of sons, did Edward resort to appointing a daughter?

The idea should not be dismissed out of hand, for one other woman is known to have served as an ealdorman when male candidates for the post had run out. His sister Æthelflæd was of Mercian royal descent through her mother and their father King Alfred married her off to Æthelred, ealdorman of English Mercia. After Æthelred fell ill in the 900s, it was Æthelflæd who took over his duties and after his death in 911 she assumed his role fully.³⁰ After her own death in 918, she was apparently succeeded by her daughter Ælfwynn, only for her to be ælces onwealdes on Myrcna benumen by Edward and transported to Wessex, never to be heard from again.³¹ This tale of marriage, power and female sovereignty may provide an analogy for the case of Eadgifu.

Eadgifu was a minor at the time of her father’s death, so the office of ealdorman apparently fell vacant. Although the possibility of unrecorded or unidentified successors cannot be ruled out, such a vacancy would make Kent fit into the general shift in how shires were administered that is seen in this period. The expansion of West Saxon power into Mercia and Northumbria, rather than increasing the number of ealdormen, actually reduced it, shires (both the historic shires of Wessex and the newly created shires of the Midlands) being bundled together into groups assigned to a diminishing number of ealdormen, whose regular succession was not always maintained.³² This would remain the standard policy for ealdormanries until the Norman Conquest.³³

Increasingly, other officials are seen to shoulder the burden formerly borne by ealdormen.³⁴ One such official was the bishop, who presided jointly with the ealdorman over the shire court.³⁵ When the ealdormanry was left vacant or merged with others, bishops would have been left on their own at the pinnacle of the shire hierarchy and there is no reason why they should not have officiated on their own. Archbishop Wulfstan (himself a prolific legislator) encouraged bishops to take an active interest in judicial affairs and local administration,³⁶ and there is even one known example of a bishop (in the first half of the tenth century) who sentenced thieves to death.³⁷ In the Danelaw, the jarls seemed to have ruled small districts and their continued appearance in charters suggests that their services were retained after their territories were conquered.³⁸ A good example is Earl Thurferth, who led the submission of ‘þa holdas, [and] eal se here þe to Hamtune hierde norþ oþ Weolud’ to King Edward at Passenham in 917.³⁹ These geographical points suggest that his jurisdiction covered Northamptonshire. He went on to appear among the duces in four of Æthelstan’s charters.⁴⁰

The main official seen to undertake responsibility for shires was the reeve, a development that appears in Kent around the time of Eadgifu’s alleged governorship. The Heronden charter was attested, just before Hlothwig, by two praepositi, Ælfstan and Birhtsige. This term can mean two things. It might be used to refer to religious dignitaries, specifically the prior of a religious house or the reeve of its estates.⁴¹ Three religious houses were corporately present at the assembly, to any of which Ælfstan and Birhtsige might have been attached. The appearance of monastic praepositi in charter attestations was, however, always rare and is last seen in Kent in the early ninth century.⁴² Praepositus was also one of several Latin words used for vernacular gerefa, so these two men may have been joint-reeves of Kent (which had formerly had two joint-ealdormen).

The first unambiguous example of a reeve who administered Kent was Wulfstan, who is reported to have acted as King Edgar’s representative in the memorandum of a land dispute.⁴³ A royal assembly at London agreed that certain charters (and therefore their lands), which had been stolen from Rochester Cathedral and were now held by a widow, were forfeit to the king. In execution of this judgement, Wulfstan se gerefa attempted to take possession of the estates to ðæs cinges handa (on the king’s behalf). The widow, however, outmanoeuvred the king by making a private agreement with the bishop that he would buy the estates from the king but allow the widow their usufruct for the rest of her life.

A second assembly witnessed the bishop’s purchase of the estates from the king and the memorandum ends with what purports to be a list of those present, including (though not immediately sequential) Wulfstan of Dalham and ‘seo duguð folces on Westan Cænt’ (the nobility of the people in West Kent). This makes it tempting to identify Wulfstan the reeve as Wulfstan of Dalham. There is a potential connection with Eadgifu here, since Wulfstan of Dalham was also ‘seneschal of Queen Eadgifu’s East Anglian properties’.⁴⁴ There is beautiful neatness in imagining that it was Eadgifu who appointed him reeve of Kent and that he continued in office after her death.

There are, however, several problems with this interpretation. First, Wulfstan was a very common name, so two Wulfstans should be identified only on strong evidence. Secondly, the witness-list is not strong evidence, for it contains several anachronisms. It is attested by King Edgar yet also by an ealdorman who was not appointed until Edward the Martyr’s reign, by two bishops who were not appointed until Æthelred II’s reign and by ðæs cinges modor Ælfðryð, a designation possible only in Æthelred II’s reign.⁴⁵ Ironically, a more credible witness-list appears in a near-contemporary forgery that purports to be the charter recording the purchase.⁴⁶ Its supposed date of 955 is too early for Edgar and its indiction of 966 is wrong, for the witness-list would fit only the year 973, thus giving an approximate (though authoritatively uncertain) date for these events. The charter describes the activities of Wulfstan regis prefectus, who is described handing over the charters in exchange for the bishop’s cash and he may be intended as the same man as the Wulfstan minister who appears in the witness-list but Wulfstan of Dalham as such does not appear.

There is rather more certainty about the identity of the next reeve of Kent. Another memorandum records a shire court held at Crayford between 984 and 988. It was attended by a certain Wulfsige, described variously as Wulfsige the priest and Wulfsige se scirigman, who accepted an oath to ðæs cinges handa. ⁴⁷ Scirigman (or scirman) does not necessarily mean a man who administered a shire but rather was, like gerefa, a generic term for any kind of administrator.⁴⁸ The context in which Wulfsige appears, however, makes it quite clear that he administered the shire of Kent. Leofric se scyresman is mentioned as present at a shire court convened by Archbishop Sigeric (995–1005).⁴⁹ An Ælfwine praefectus meus in Kent is claimed in a Westminster forgery purporting to date from 998,⁵⁰ but he may be a confused memory of Æthelwine, who attested charters as variously scirman, sciregerefa and prefectus early in Cnut’s reign.⁵¹ From Æthelwine onwards, Kent has been administered by sheriffs as such.

In sum, Kent was administered by ealdormen until the death of Sigewulf and Sigehelm. There then ensues a gap in our knowledge of about seventy years, after which the shire is seen to be administered by officials called variously gerefa, scirman or finally scirgerefa. Eadgifu’s life fell almost entirely within that gap and, just before Wulfstan’s emergence (and perhaps overlapping with it), she herself was apparently gubernator of Kent. With the exception of Somner’s charter, this title was never used in witness-lists and only ever appeared in charters as part of the royal title. King Edgar used it at least twenty-one times, often in apposition to rex, rector or basileus or in an extravagant combination, such as ‘industrius Anglorum rex ceterarumque gentium in circuitu persistentium gubernator et rector’.⁵²

Gubernator, it seems, had regal connotations and there is further evidence that it was used as a euphemism by people who were not monarchs officially but did regard themselves as such. In 901, Æthelred and Æthelflæd, officially the ealdorman of the Mercians and his wife, granted a cyrograph which describes them as monarchiam Merceorum tenentes honorificeque gubernantes et defendentes. ⁵³ Æthelflæd alone used a similar phrase in a charter (apparently authentic but surviving only as an adjusted copy), in which she described herself as gubernacula regens Merciorum. ⁵⁴ Alfred and Edward may have denied Æthelred and Æthelflæd the regal title but they were not afraid to flirt with it and gubernare was one of the terms they used to imply it.⁵⁵

This provides some clues to how gubernator was meant when applied to Eadgifu. For one thing, although she is not called regina in any authentic contemporary source,⁵⁶ she had been a king’s wife, so it was not inappropriate for a quasi-regal title to be applied to her. For another, like Æthelred and Æthelflæd, she governed no ordinary shire but a former kingdom. To call her Queen of Kent as such would have been provocative but, at least in the eyes of the Kentish scribe, she could justly be called something close to that.

The example of Æthelfæd also warns against assuming that Eadgifu’s sex disqualified her from occupying an administrative office. Her case is certainly unusual when set against Anglo-Saxon history in general but it becomes more credible when set in its proper context, in the mid-tenth century, when kings were experimenting with new ways of administering shires.

Indeed, this paper is not the first to argue that Eadgifu had a role in administering Kent. Pauline Stafford, who was unaware of Somner’s charter, pointed out that Eadgifu was listed first, before Archbishop Oda (941–958), among the witnesses to a will made in Kent.⁵⁷ Stafford argued that this pointed to some administrative role in the shire.⁵⁸ Such a reading may draw too much from too little: as a king’s widow and an ealdorman’s daughter, Eadgifu was naturally the kind of high-status individual whose attestation was useful in a will and her precedence over the archbishop follows the pattern established for her in royal charters. The title that she used in the will was ðære hlæfdian, the generic title for women of distinction and reveals nothing for or against Stafford’s interpretation.

More suggestive of an administrative role are Eadgifu’s landholdings. She held considerable property in Kent, some of it inherited from her father and some granted to her by her sons. The gift most relevant to this paper’s inquiry was made in 955, when King Eadred willed to Eadgifu all his booklands in Kent, Sussex and Surrey.⁵⁹ That her son was simply celebrating his mother’s Kentish origins is a possible explanation but rather trite. The only previous document to treat the king’s booklands in Kent as a parcel was King Alfred’s will, under which he bequeathed them to his son and successor Edward.⁶⁰ This implies that the Kentish lands had a special relationship with the monarchy. They have been construed as Crown lands, what a later generation would call ‘ancient demesne’ (it may have been for this reason that King Eadwig confiscated them).⁶¹ Lands were also assigned to ealdormen and sheriffs ex officio. ⁶² Eadred’s gift thus gives the impression that he was establishing his mother in an administrative or even quasi-monarchical position in Kent.

Ealdormen and sheriffs were assisted in the performance of their functions by a staff of under-reeves.⁶³ Reference has already been made to the praepositi Ælfstan and Birhtsige, who appeared in a Kentish charter shortly after Eadgifu’s disappearance. The evidence is circumstantial but it is possible that, if ‘reeve’ is the correct translation of their title, they had been Eadgifu’s own subordinates and continued in office even after she had died. Under-reeves also serve to make Eadgifu’s position as governor of Kent seem more credible by creating the possibility that it was only an honorary position, that the under-reeves performed her duties on her behalf while she gloried in being nominal governor, the shire an ornament for her and she an ornament for the shire.

There is one last doubt that might be raised against Eadgifu’s governorship and that is the question of whether or not Somner’s charter was authentic at all. The most serious difficulty with the charter’s authenticity is that, apparently, it called Eadgifu regina, or at least that is what Somner implied when he called her ‘Queene Eadgive (who writes herselfe, in Cantia etiam Gubernator)’. The only other charters in which Eadgifu used the title regina are either known or suspected to be forgeries and all hailed from the same archive as Somner’s charter: Canterbury. Was Somner’s charter just another Christ Church forgery?

One must remember that this charter was a local production, so that to judge it by rules established for central productions may be unfair. Eadgifu’s attestation as implied by Somner would have been something like Eadgifu regina in Cantia etiam gubernator, which resembles several of Edgar’s titles. Surely, a forger cannibalizing Edgar’s charters for inspiration would have applied such a title to him, not to his grandmother? Conversely, a local scribe with an inflated view of Kent’s importance (and one can never have too inflated a view of Kent’s importance) might indeed have modelled an ad hoc title for his shire’s governor on the king’s verbose titulature. The attestation is also made in a flamboyant way, as though to attract attention. The scribe could simply have written regina et in Cantia gubernator but instead he used the adverb etiam, giving her attestation a literary flourish and making it sound like a public announcement.⁶⁴

There is some evidence that scribes acting in an unofficial capacity did have a habit of being less reserved about Eadgifu’s title. Wulfstan of Winchester, writing in 996 or shortly after, and Adelard of Ghent, writing during the archiepiscopate of Ælfheah (1006–1012), both called her regina in reference to events in Eadred’s reign.⁶⁵ Eadgifu applied no title to herself in her aforementioned statement to the archbishop but the document is endorsed in an eleventh-century hand Eadgyua Regina. It was later used as the basis of a post-Conquest forgery in which Eadgifu is called regina et mater Eadmundi et Eadredi, which in itself is only a statement of the truth.⁶⁶ Similarly, Eadgifu’s title ðære hlæfdian, used in the will that she witnessed with Archbishop Oda, was rendered regine in a later Latin translation.⁶⁷ Finally, she was referred to as regina Eadgiua in a memorandum, apparently based on a lost charter or charters, in the Anglo-Norman Christ Church cartulary.⁶⁸ Historian, endorser, forger, translator and cartularist may have inserted information not present in the original texts but it was not factually wrong. If the text that Somner was paraphrasing was itself a translation of a vernacular document or a later transcript of the original, then a similar interpolation may have been made with equal sincerity and authority.

It is even possible that ‘Queene’ was an interpolation by Somner himself. His transcription imitates contemporary minuscule only from the chrismon onwards, so that it cannot be certain how much of the text before then was quoted directly from the charter in front of him. This might also explain why he gave Eadgifu’s name in the later medieval form Eadgive (cf. Godiva from Godgiefu), although that form did occasionally appear in contemporary charters.⁶⁹

Finally, the charter’s most outlandish feature, its claim that Eadgifu was governor of Kent, far from arousing suspicion, might actually support the case for its authenticity. A forgery that makes blatantly false claims that do not assist its litigious purpose defeats itself, so, even if the land grant that this charter concerned were fictitious, or even if this charter were manufactured by some Christ Church scribe cribbing off the Heronden charter, that would still not explain why it called Eadgifu in Cantia etiam gubernator. That claim had no bearing on its case and cannot have been copied from any other known charter, even the Heronden charter. Unless the forger simply had a flight of sheer fancy, this at least must be something that he believed had been true and expected his audience to believe as well.

CONCLUSION

Eadgifu was born in the 890s, the daughter and sole known child of Sigehelm, one of the last two ealdormen of Kent, who fell in battle with the Danes in 904. She was introduced to King Edward the Elder around 910, shortly after she came of age. Ten years later, she married him, becoming the mother of successive kings Edmund and Eadred. During her sons’ reigns, she was a prominent member of the court and Eadred may have put her into some administrative position in her father’s shire of Kent. The brief reign of Eadwig saw her banished from her former position of influence and her lands were confiscated but she was restored by her younger grandson Edgar. She was fondly remembered in her ancestral shire: Somner recorded, in another of his works, that a picture of Queen Eadgifu was, until recently, preserved in the treasury of Canterbury Cathedral, in gratitude for her gifts.⁷⁰ She was even styled, in a charter written locally and surviving only in a brief quotation, gubernator of Kent.

In the absence of the original, or even a cartulary copy, of that charter, its evidence will always be attended by a quantum of doubt. Its date is impossible and there is even the nagging suspicion that Somner transcribed it incorrectly, but its claim about Eadgifu, however superficially preposterous, proves on examination to be at least theoretically acceptable. Though examples of female secular power are rare in this period, Eadgifu’s case is not incongruous amongst those that do exist and it also makes sense in the context of contemporary developments in local administration. In her lifetime ealdormen, who formerly administered shires, were being reduced in number, while other officials increasingly assumed the burden of their functions, culminating in the emergence of the sheriff in the half-century after her death. Eadgifu’s governorship of Kent, however brief or nominal, would fit into this overall pattern and she, herself daughter of the last ealdorman, was beautifully qualified for the job.⁷¹