Square Minuscule in the Age of Cnut the Great

Abstract

Beowulf scholars wishing to eliminate a late date for the poem have embraced David Dumville’s stringent position that, although the scribes almost certainly copied the manuscript in the early eleventh century, it is “in the highest degree unlikely” they copied it after 1016. There is no obvious nor expressed justification for limiting a scribe’s life or career to the year Æthelred died and Cnut the Great became king. Dumville’s argument was that he could find no examples of Square minuscule that anyone could prove was written after 1013. In his exhaustive Catalogue of Manuscripts containing Anglo-Saxon, Neil Ker cautiously avoids such narrow dating ranges for undated manuscripts, relying instead on thirty-year or (more often) fifty-year dating ranges. By his system, Ker classified more than two dozen scribal hands as late types of “square Anglo-Saxon minuscule.” Analyses of Ker’s examples of early eleventh-century Square minuscule prove that scribes maintained the script type throughout the reign of Cnut. Some examples suggest that the scribes most likely copied Beowulf in the first half of his reign.

The dating of the Beowulf manuscript, as distinct from Beowulf, became controversial in 1988, when David Dumville published an article he called “Beowulf Come Lately: Some Notes on the Palaeography of the Nowell Codex.”¹ Until then, Beowulf scholars accepted the paleographical dating of Neil R. Ker at s. X/XI, interpreted following Ker’s sensibly cautious fifty-year range as circa 975 x 1025.² I discussed in Beowulf and the Beowulf Manuscript, and presented at the Dating of Beowulf symposium in Toronto in 1980, the obvious but previously unremarked consequence that this dating range included nearly ten years of the reign of Cnut the Great, 1016–35.³ I also noted at the meeting that, had the first scribe copied all of Beowulf, paleographers would likely agree that the manuscript was early eleventh century, not excluding Cnut’s reign.⁴ Although the second hand displayed an earlier-looking script, I reasoned that this “late type of square Anglo-Saxon minuscule,” as Ker classified it, must also be early eleventh century. In “Come Lately,” Dumville seemed to agree. “We may say that the Beowulf-manuscript is datable, in Ker’s terms, to ‘s. XI in.’ [early eleventh century]: it is (in the highest degree) unlikely to have been written before A.D. 1000.”⁵ However, he also concluded that it was “in the highest degree unlikely that the Beowulf-manuscript was written later than the death of Æthelred the Unready” in 1016, the year that Cnut the Great ascended the throne of England.⁶ To maintain his precise sixteen-year calculation, Dumville had to narrow the dating ranges—s. X/XI, s. XI in., s. XI¹—of many other manuscripts containing Old English that Ker similarly classified as late versions of Square minuscule. This study challenges Dumville’s arguments by presenting substantial paleographical and text-historical evidence that scribes continued to write versions of Square minuscule in manuscripts containing Old English after 1016, into and perhaps beyond Cnut’s reign.

Dumville’s groundbreaking investigations of the earliest phases through the mid-tenth-century phases of Square minuscule have only tenuous bearing on the early eleventh-century manuscripts he seeks to re-date in “Come Lately.”⁷ In the 1987 article, Dumville discusses Phases I and II, up to circa 940, but briefly mentions the “mature and concluding phases,” III through VII, which he says will appear in a sequel. “It will consider,” he predicted, “the script of the reigns of Edmund, Eadred and Eadwig (Square minuscule of Phase III), the interaction of Caroline and Square minuscule (Phase IV), informal Square minuscule, centres of production of the mature script (Phase V), scribal treatment of Latin and Old English, the latest phases (VI, VII) of Square minuscule, and the transmutation of the script into the forms characteristic of eleventh-century Anglo-Saxon minuscule.”⁸ However, after reprising Phase II, the 1994 sequel considers only Phases III–IV up to circa 960. Phases V–VII and beyond have yet to appear.⁹ While acknowledging Dumville’s contribution to the study of tenth-century Square minuscule, some paleographers have critiqued his approach in ways that also apply to the continued use of the script in the early eleventh century. Christine Voth, for example, discusses “different styles . . . some regional and others located to a specific house” in early tenth-century Square minuscule. Rather than phases emanating from royal writing centers, as Dumville held, Voth agrees with Malcolm Parkes that “the presence of different styles of scripts within Square minuscule would suggest that different writing centers and scribes had different exemplar scripts from which to model their own.”¹⁰ By focusing on places and individual scribes, who might use competing styles in the same manuscript, Voth’s approach uncovers important features the theory of phases overlooks.¹¹ David Ganz, who also emphasizes scribes over scripts, confesses he “cannot find in the manuscripts assigned to [Dumville’s] phases enough shared conventions in the shaping of individual letters to link them in a common group, or enough such differences to separate them from manuscripts assigned to other groups.”¹² Peter Stokes throws any orderly progression of phases from 960 to 1035 into disarray when he observes that “characteristics of the earlier phases of Square minuscule reappeared in later phases, after having vanished from the extant manuscripts which were produced in between.”¹³

For manuscripts written in the early eleventh century, it is necessary to make a distinction between Latin and Old English. By 1000, scribes were almost exclusively using Caroline script developed on the Continent for Latin texts, and English Vernacular minuscule, including late versions of Square minuscule, for English texts, even in the same manuscripts. As Julia Crick puts it, “By the eleventh century, Latin had parted company from Old English in visual terms.”¹⁴ She argues that “most scribes probably never learned to write Caroline Minuscule at all, as it was a foreign script difficult to master,” and raises the valid question of whether unlearned readers were able to cope with it.¹⁵ “By the eleventh century,” Crick observes, “written English had begun to colonise areas of communication which, in other times and in other places, might normally have been reserved for Latin. It serves as the vehicle,” she adds, “for prescriptive and salvatory texts (law, penitentials, prayers), much translated from Latin, and its use for rendering the Bible . . . attests the extraordinary authority attached to it.”¹⁶ One can see this cultural shift in the Nowell Codex, which begins with an Old English translation in English Vernacular minuscule of the Latin Life of St. Christopher and ends with a poetic Old English rendering of the biblical book of Judith in late Square minuscule.¹⁷ Peter Stokes acknowledges the longevity of Square minuscule, or elements of it, in his chapter “Scribal Continuity in Bookhands and Charters: The ‘Square-Influenced’ Hands.”¹⁸ His terminology, Style-II English Vernacular minuscule, is the equivalent of Ker’s late square Anglo-Saxon minuscule, as Stokes makes clear at the end of this chapter.¹⁹

Dumville’s paleographical notes alluded to in his subtitle seek to eliminate the many additional hands Ker identified as late examples of Square minuscule. Because he includes no illustrations in “Come Lately,” it is often difficult to evaluate his statements. In some cases, he simply dismisses Ker’s description of a late example without offering alternative descriptions, much less illustrations. In others, notably specimens in Ælfric manuscripts of the Catholic Homilies, Dumville asserts that manuscripts belonging to untraceable transmissions with non-Ælfrician accretions date to the late tenth, rather than, as Ker thought, the early eleventh century. Ælfric did not even begin composing the First Series of Catholic Homilies until 990. For Ælfric’s earliest manuscript, Ker’s “s. X EX.” does not mean the last third or the last quarter of the tenth century, but only the last decade. Some period of time would have had to elapse for the later manuscripts to abandon Ælfric’s earliest exemplar, preceding British Library, Royal MS 7 C. xii, which scholars agree is a copy of a copy.²⁰ Unlike those wanting to date Beowulf early, Ælfric scholars have yet to embrace Dumville’s challenges to Ker’s dating ranges. The following analyses scrutinize the many manuscripts Ker specifically identified as “late” examples of “square Anglo-Saxon minuscule” that include and sometimes encompass Cnut’s reign. To evaluate Ker’s conclusions fairly, it is essential to keep in mind that scholars of Ker’s generation and earlier often used the term “Anglo-Saxon” as we use “Old English” today for the language. When he uses “Anglo-Saxon” in the title of his Catalogue of Manuscripts Containing Anglo-Saxon and in referencing eleventh-century examples of “square Anglo-Saxon minuscule” in this Catalogue, Ker is referring to texts written in Old English, not Latin. All of the examples of late Square minuscule discussed in this study are in Old English, descendants of the tenth-century tradition of Square minuscule used for both Latin and Old English.

As is apparent in the different terms favored by Ker and Stokes, paleo-graphical terminology can change from source to source. In his articles for Anglo-Saxon England, Dumville uses the recognized but ambiguous phrase “English square minuscule” for both Latin and Old English texts. In “Come Lately,” he sometimes uses “Insular script,” higher in the taxonomy, in reference to Ker’s examples of Square minuscule in Old English. Dumville insists that Ker “often spoke of ‘square’ aspects of this phase of Insular script, but he consistently called it ‘Anglo-Saxon minuscule,’ ” suggesting that Square is not integral to Ker’s paleographical classification.²¹ In fact, Ker consistently classified late examples of this script as “square Anglo-Saxon minuscule.” Dumville repeatedly asserts that his mentor, T. A. M. Bishop, was the first to name the script “Square minuscule.”²² He goes further in “The Beowulf-Manuscript and How Not to Date It,” claiming that when Ker’s Catalogue was published in 1957, “and for a decade thereafter, Square minuscule had not been recognised as a separate and canonical form of Insular script. That was the contribution,” he maintains, “of T. A. M. Bishop.”²³ Bishop himself makes no such claim in “An Early Example of the Square Minuscule,” whose title acknowledges the known script, the Square Minuscule. Bishop even mentions that an early form of “the Square Minuscule” was recognized in the Parker Chronicle in the early twentieth century.²⁴ In 1931 Kenneth Sisam, Ker’s mentor, described the main hand of an Ælfric manuscript as “a fine square English hand which, if normal conditions are assumed, would fall in the first quarter of the eleventh century.”²⁵ In 1957, describing the Old English glosses in Bodleian Library, MS Junius 27, Ker classified the early tenth-century hand that wrote the Latin psalter “a rather stiff square Anglo-Saxon minuscule, resembling [the Square minuscule in Old English] in the Parker Chronicle, ff. 16^v–25^v,” lines 891–924, CCCC MS 173.²⁶ In 1966, nine years later, Bishop closely echoed Ker’s words when he described the same hand in the same folios (16–25v) as “a rather primitive Square minuscule resembling that of the Parker Chronicle, 891–924.”²⁷

While he would agree with Dumville that Square minuscule (in both Old English and Latin) was a quintessentially tenth-century script, Ker in his Catalogue shows that he was keenly interested in tracing the efforts to retain this native script in eleventh-century manuscripts containing Old English. Agreeing with Ker, Peter Stokes argues that, “rather than being a deliberate script style which was practised throughout Anglo-Saxon England, this second style is characterised more by the principle of retaining much of the aspect and many letter-forms from the tenth century and includes the hands which Neil Ker described as ‘late’ forms of, ‘manifest descendant[s]’ of, or ‘influenced by’ Square minuscule.”²⁸ Ker put the second hand of Beowulf in this group. In “Come Lately,” Dumville steadfastly refused to recognize that Ker in all of these manuscripts was purposely tracing the continued use of Square minuscule in early eleventh-century manuscripts.

Ker specifically classifies more than twenty late tenth- and early eleventh-century manuscripts written in Old English as late examples of Square minuscule, and in most cases adds descriptions of specific letterforms in them to underpin this overarching classification. The sine qua non that gives the script its name is the square or flat-topped a, usual also in the digraph æ. This a is often almost literally square. The bodies of letters within the minim lines also tend to be as wide as they are high. Ker also notes as occurring in the script e ligatures, particularly high e ligatures (also usual in æ) with a and o and a number of consonants (g, t, r, low s, etc.), and the incidence of low, round, and long s, and of curved or straight y, both dotted and undotted. Only taking note of Caroline incursions, Ker does not normally single out other Insular letterforms, which he describes in detail in his introduction; their forms remained fairly stable (the bottom of Insular g is more often closed in the early eleventh century than in the tenth).²⁹

Dumville also maintains that “no specimen of Square minuscule is datable later than [1001 x 1013]; in fact,” he insists, “on the strictest interpretation of dating criteria, no example need be dated [his emphasis] after A.D. 1000.”³⁰ Ker gives dating ranges for scores of specimens of Square minuscule after 1000. Among the manuscripts he broadly dates by their script, along with Beowulf and Judith, circa 975–1025 (s. X/XI) are a wide range of texts. They include copies of comparatively minor works, such as a short inscription, medical recipes and a herbal, a booklist, glossaries, a homily based on the Gospel of Luke, and an exhortation to confession.³¹ But they also encompass such major Old English collections as Ælfric’s homilies, the Blickling Homilies, and copies of the Pastoral Care, the Martyrology, the Anglo-Saxon Laws, and of course Beowulf and Judith.³²

Ker specifically classifies the hands of all of these late Old English manuscripts as Square minuscule, often appending such descriptive epithets as “showing the influence of,” “influenced by,” “the same kind of script as,” “in the tradition of,” “a late type of,” “a graceful late type of,” “a rather rough and debased,” “a fine,” and “retaining features of ” Square minuscule. Ker likewise classifies six additional manuscripts, including three of Ælfric’s Homilies and one of his Lives of Saints, Ælfric’s Grammar, Gregory’s Dialogues, and a folio from the Old English version of Bede’s Ecclesiastical History, from the first third of the eleventh century (s. XI in.).³³ Dumville argues that by s. XI in., ineunte, Ker meant only the first twenty-five years of the century, while s. XI med., medio, meant the next fifty years, and s. XI ex., exeunte, meant only the last twenty-five years.³⁴ Even if Ker meant such non-intuitive three-part divisions, a terminus of 1025 for s. XI in. still includes nearly ten years of Cnut’s reign, the same as Ker’s s. X/XI. Ker qualifies the examples he dates s. XI in. as “handsome,” “late,” “stiff,” “fine,” and “influenced by” Square minuscule. Finally, Ker classifies “all the writing” of four different hands copying the West-Saxon Gospels as “a late uncalligraphic type of square Anglo-Saxon minuscule,” which he dates s. XI¹, the first half of the eleventh century.³⁵ He dates additional manuscripts this late using the classification “Anglo-Saxon minuscule” with descriptive (square) epithets, as, for example, with Bodleian Library, MS Bodley 865, discussed below. Using 1025 as the midpoint, the dating range even for twenty years, let alone for twenty-five or fifty years, includes all of Cnut’s reign.

The Old English Gospels in Cambridge, Corpus Christi College, MS 140 refute Dumville’s claim that eleventh-century forms of Square minuscule in Old English totally disappeared before Cnut became king of England. Dumville dismisses Ker’s identification of these four hands without providing analysis or illustration. “In Ker’s dating scheme,” he writes, “the latest specimen of Square minuscule is provided by the copy of the Old English Gospels now Cambridge, Corpus Christi College, MS. 140, written by four scribes all practising ‘a late uncalligraphic type’ of that script which he dated ‘s. XI¹.’ This broad extension of the dating range of Square minuscule must be rejected, however; examination of the manuscript shows that Ker’s dating is correct but that his description of the scribes’ handwriting is very questionable indeed. Their Insular script is of eleventh-century type.”³⁶ In quoting Ker, Dumville substitutes “that script” for “square Anglo-Saxon minuscule,” the classificatory phrase he says Ker never uses.³⁷ Peter Stokes agrees with Ker’s assessment. “All four retain features from Square minuscule,” he says, and concludes, “These scribes at Bath, then, seem to have retained the weight and many of the letter-forms from Square minuscule.”³⁸

Figures 1–4 illustrate what Ker meant by late examples of Square minuscule in these four hands.³⁹ Taken together, four different scribes writing Old English in the second quarter of the eleventh century at Bath each practiced different versions of Square minuscule. With slight variations, all four scribes form their horned, square a in two strokes: the first stroke starts with a tiny horn at the top of a minim that finishes by curving right, across the bottom minim line; the second stroke completes the square from the horn across the top minim line, then descending to meet the end of the first stroke. All four scribes regularly use horned e; g with closed bottom; low, long, and round s; wedged ascenders; mostly interchangeable þ and ð; and dotted y with straight or curved descender.

Identifying himself as Ælfric at Bath on the last page (fol. 45v), the scribe of Matthew (fig. 1) has rounded a in digraph æ, which has a slightly elevated e-head ligaturing on this page with f, r, and low s and with most other consonants elsewhere. The scribe uses the three forms of s, low, round, and long (notably in the st ligature). Horned e does not usually have a high head, but its tongue ligatures with o and many consonants, and sometimes extends at line ends. The bottom of Insular g is usually closed and occasionally open (e.g., fol. 27v). The relatively short ascenders and descenders add to the square look. At times the scribe cannot resist using Caroline a and, on the first two pages, a playful, flamboyant Caroline r, sadly abandoned thereafter. The Latin colophon, strictly observing linguistic barriers, is in a neat Caroline, except for the English names Ælfric and Brihtwold, which are in Square minuscule. Other calligraphic flourishes include a larger graceful ampersand, enlarged ð with elaborate crosses, and atypically oversized letters, notably high e with extended tongue, none of which overpower the square look. On most folios the aspect is not vertically compressed

Figure 1.

First page of the Gospel of Matthew. Cambridge, Corpus Christi College, MS 140, fol. 2r.

With permission of the Masters and Fellows of Corpus Christi College, Cambridge.

and elongated, contrary to Dumville’s characterization of eleventh-century examples.

Figure 2.

First page of the Gospel of Mark. Cambridge, Corpus Christi College, MS 140, fol. 46r.

With permission of the Masters and Fellows of Corpus Christi College, Cambridge.

Mark’s horned, flat-topped a is slightly narrower at the top than the previous hand, but the scribe maintains a square aspect, neither vertically narrow nor elongated (fig. 2). The ascender of d stays close to the top minim line. There are short descenders; wedged ascenders for b, h, l, and þ; long s, with top curl added after the post in the st ligature; curving cross-stroke for ð and abbreviation for þæt; high e-head alone and with æ ligatures with low s and t, but low e-tongue alone and in æ ligatures with a, o, and many consonants (e.g., f, g, n, r, s); and three forms of s. As if anticipating the Luke hand, the scribe of Mark, who normally shuns terminal þ, uses it twice in four lines on the last page of the stint (fol. 72v1–4).

Figure 3.

First page of the Gospel of Luke. Cambridge, Corpus Christi College, MS 140, fol. 73r.

With permission of the Masters and Fellows of Corpus Christi College, Cambridge.

Figure 4.

First page of the Gospel of John. Cambridge, Corpus Christi College, MS 140, fol. 116r.

With permission of the Masters and Fellows of Corpus Christi College, Cambridge.

The Luke hand has other minor variations (fig. 3). Sometimes the top stroke of his or her square a is slightly rounded in both a and æ, but still maintains equidistant proportions. The horned e and æ both ligature with following letters—e with g, n, r, low s, t, w, and æ with c, f, r, s, and t; all forms of s; and straight-limbed, dotted y. Characteristic features of this hand are terminal long s and þ; unligatured long s and t; ð crossed with a simple line or a quirky upstroke; and prominent ð, once appearing here as a large capital resembling recumbent orb and cross (also used by the Matthew scribe), but normally with a long ascender often starting with a downward hook to the left.

Although the first three hands differ from one another, the overall appearance of their versions of Square minuscule are reasonably complementary. The fourth hand for the Gospel of John (fig. 4) is noticeably different, seemingly unpracticed, displaying a rather cockeyed version of Square minuscule letters, one by one. Although sharing most of the main features of the others, there is at the beginning little concept of ligaturing. Here the scribe seems to be a novice whose work lacks finesse. On the opening page, for example, this scribe writes du-n-e-ard for duneard, we-ald for weald, wæ-ron for wæron, be-arn for bearn (fol. 116r13–15). Tasked with providing an easily accessible text for readers of the Bible in Old English, the scribe seems to struggle to achieve a legible, squarish minuscule. However, after a few pages, the Luke scribe suddenly shows off a practiced Square minuscule, but with the narrow and elongated proportions that Dumville used to characterize all eleventh-century specimens. In other places—for example, on folio 138r—the script has returned to normal proportions, more like the previous three hands, particularly Luke’s. These gospels copied at Bath attest to a working writing center, perhaps a nascent scriptorium, willing to tolerate a variety of hands producing Square minuscule.

It is not an early eleventh-century foible that multiple hands, not always compatible, practiced different scripts in a single manuscript. Richard Gameson discusses the same phenomenon from a century earlier in the early tenth-century manuscript of the Old English Bede, and other examples from the tenth century are not hard to find.⁴⁰ There are, in any case, important paleographical lessons one should not hastily dismiss in CCCC MS 140. Chronological phases cannot apply to the four different hands, as they all occur in the same manuscript. Here chronological, and to some extent scriptorial, arguments must surrender to the work of individual scribes. The scribe who copied Luke gives the scribe who copied Mark a break by copying folio 64r, covering most of Mark 12:26–38. The change of hands is subtle and easy to miss, showing their essential compatibility. Ker no doubt noticed one of the Luke scribe’s idiosyncrasies, the use of terminal long s on the first line. Other telltale features on the page are the terminal þ, the simple cross-strokes of ð, and the unligatured long s and t. Even with the differences among the scribes, their work, taken individually or together, illustrates that there was undoubtedly a continuity, even a demand, in the attempts to write Square, or Square-influenced, minuscule well into the eleventh century. Above all, these sometimes successful, sometimes clumsy, efforts did not end when Æthelred died in 1016. Ker was obviously intrigued by these late manifestations in the early eleventh century of Square minuscule, and elements of it, in Old English manuscripts. His Catalogue documents the work of other scribes in other places who contributed to this continuity from the end of the tenth century at least until the scribes at Bath copied CCCC MS 140. In his chapter on “Scribal Continuity in Bookhands and Charters: The ‘Square-Influenced’ Hands,” Stokes validates Ker’s brief, classificatory insights.⁴¹

Dumville implies that CCCC MS 140 was the “latest specimen” of Square minuscule that Ker found. There are many other specimens. Ker does not always explicitly define a script at all in his briefer descriptions. A good example is the eleventh-century translation of the Capitula of Theodulf in Oxford, Bodleian Library, MS Bodley 865. Although Ker does not say so, the Latin parts are in Caroline script, while the Old English is written by three different scribes in late Square minuscule. Stokes agrees that they “show much variation, but they all retain many forms from Square minuscule.”⁴² Ker dates this manuscript s. XI¹, the same as the hands of the gospels in CCCC MS 140. He also provides a rare facsimile of Bodley 865, folio 107v, in Plate III, one of only eight plates in his entire Catalogue (fig. 5). Ker briefly describes the Old English script, saying only that “the square flat-topped

Figure 5.

Oxford, The Bodleian Library, University of Oxford, MS Bodley 865, fol. 107v.

a is used regularly, high e ligatures occasionally, and round s often.”⁴³ There are distinct differences between this hand and the second hand of Beowulf, such as the horned a, the long r descenders, and the comparatively infrequent use of high e ligatures. But the script shares many of the letterforms and resembles the weight, aspect, and proportions of the second hand of Beowulf.⁴⁴

In discussing his methodology for dating in the introduction to his Catalogue, Ker states that his “chief guides” were the “dated charters.” Unfortunately, these charters “are poorly represented” for the “important years [for script] about and soon after 1000.”⁴⁵ The paucity of dated charters for the controversial transitional period Ker dated s. X/XI is particularly lamentable. Dumville asserts that “no book (or charter) certainly datable by its contents to after A.D. 1000 is a specimen of Square minuscule.” For him, Square minuscule is an exclusively tenth-century script. He here implies that tenth-century Square minuscule was a uniform, canonical script, which is belied by his four phases, ending circa 960.⁴⁶ Ker nowhere suggests that his many specimens of eleventh-century Square minuscule conform to one style established in the tenth century. Three charters help illustrate why Ker was confident in dating specimens of late Square minuscule from the beginning to the end of Cnut’s reign, s. X/XI, s. XI in., and s. XI¹. The first (S.1220), The Godwine Charter, dates from 1013 to 1020.⁴⁷ One of the witnesses is Lyfing (d. 1020), who as Archbishop of Canterbury crowned Cnut in 1017. King Cnut and Queen Ælfgifu/Emma witness the second (S.963) and third (S.971) charters, royal diplomas written by the same scribe and dated internally in the same year, 1031 (“millessimo xxxi”).

The hand of the Godwine Charter, preserved in the top half of a chirograph, practiced a late type of Square minuscule somewhat resembling the second hand of the Beowulf manuscript. The similarities are illustrated by a comparison of the right side of the charter with the first eight lines copied by the second Beowulf scribe in British Library, Cotton MS Vitellius A. xv, fol. 172v4–11 (fol. 175 in the BL foliation) (figs. 6 and 7). Both hands have quadrangular, flat-topped a alone, with æ, and with ea ligature. Both scribes use high e ligatures with a range of Insular letterforms (e.g., f, g, n, r, s, t); both make the back of d short and round, and of ð high and concave-down; and both close the bottom of g. The two scribes use two forms of y (straight-limbed and rounded, both dotted), three forms of s (round, long, and low), and ligature long s and t. The hands are similar but not identical (the onset of cross-strokes differ for ð, but not for þ in the abbreviation for þæt; the eo ligatures are formed differently; the charter’s r descender is longer than the second scribe’s, but both are shorter in relation to other descenders). Both specimens clearly have the overall aspect, weight, and proportions of late Square minuscule. There is no convincing reason why two scribes in different places could not have written both manuscripts early in Cnut’s reign.

Figure 6.

Canterbury, Canterbury Cathedral Archives, DCc-ChAnt/S/458 (Godwine Charter), right side. Reproduced courtesy of the Dean and Chapter of Canterbury.

Figure 7.

London, British Library, Cotton MS Vitellius A. xv, MS fol. 172v (BL 175), lines 4–11. Reproduced from Electronic Beowulf 4.0.

With permission of The British Library Board.

Two royal diplomas, prepared by the same scribe and dated 1031, show that royal scribes late in Cnut’s reign used a form of Square minuscule in Old English. These large, handsomely executed royal charters grant lands to his thanes, Ætheric and Hunewine, a moneyer under both Æthelred and Cnut. Hunewine’s name appears on, among other mints, Cnut’s Quatrefoil circa 1017–23, the same period as the Godwine Charter.⁴⁸ The royal charters use Caroline for the Latin sections and a neat and distinct Square minuscule for everything in English. In both documents, the scribe distinguishes between Latin and Old English, including the boundaries (in smaller lettering), and at normal size the names of Ætheric and Hunewine (the first line of fig. 8, but deep within the otherwise Latin Caroline text of the full charter). At normal size as well, in Old English, the scribe lists in Square minuscule the names of all the witnesses, beginning with the king and queen, and including the archbishop, bishops, dukes, abbots, and ministers. These features are illustrated in figure 8, from the left side, lines 8–25, of this large document. That this form of Square minuscule was the scribe’s normal hand is manifest in two ways in the later of the two royal charters, S.971. In the line listing Æthelnoth, Archbishop of Canterbury, the scribe inadvertently continued using Square minuscule for doro-in Latin

Figure 8.

Exeter, Exeter Cathedral Library, Dean and Chapter MS 2525 (Exeter Charter), lines 8–22, Sawyer 971.

With permission of Ann Barwood, Canon Librarian, on behalf of the Dean and Chapter.

Figure 9.

Exeter, Exeter Cathedral Library, Dean and Chapter MS 2525 (Exeter Charter), dorse.

With permission of Ann Barwood, Canon Librarian, on behalf of the Dean and Chapter.

Dorobernensis (the old name for Canterbury) before finishing the Latin name in Caroline. On the dorse (fig. 9), the scribe paraphrased the document in Old English using full-sized Square minuscule, not the thin, compressed form of the charter boundaries: “þis is þære anre hyde land boc to stoke” (“This is the landbook for one hide at Stoke”). Although faded and obscured by shine-through, there is a longer endorsement mentioning Cnut in the scribe’s uncompressed late Square minuscule in Cotton MS Augustus ii 69 (S.963): “þis is þære healfan hide landboc to Mæwi þe Cnut het gebocian Æþerice his þegene on ece yrfe” (“This is the landbook for a half-hide at Mæwi, which Cnut ordered to be granted forever to his thegn Ætheric”).⁴⁹

These dated royal charters must have been among Ker’s “chief guides” in dating other late examples of Square minuscule. Dumville’s neglect of them in relation to late Square minuscule is strange. For the tenth century, his Phases I–IV supposedly derive from the styles of Square minuscule issued from the royal writing offices, his argument being that scribes copied their script because of their prestige. Here we have evidence of two dated royal diplomas witnessed by Cnut, Ælfgifu/Emma, and Cnut’s Witan. Dumville dismisses them in a book on Caroline script, saying only that they “were both written by one and the same scribe, and both concern land in Devon: their script can be characterized only as regressive.”⁵⁰ If they are regressive, they look back to Dumville’s so-called Phase IV of Square minuscule, five charters written between 960 and 963 by one and the same scribe, called Edgar A, in the early years of the reign of King Edgar. The only other instance Dumville found of this supposed phase occurs for the annals 963–64 on less than a page (fol. 28r8–25) of the Parker Chronicle.⁵¹ His description of these few specimens is almost the same as his description of eleventh-century English Vernacular minuscule. They are “Insular minuscule, but with changed proportions—narrower and somewhat elongated,” whereas the proportions of late tenth-century boundary clauses also “underwent change, the resulting letter-forms being narrower and more elongated than those of Square minuscule. Such a laterally compressed form of writing was, at full size, to find favour in the next century.”⁵²

Without explicit internal dates, as with the two Cnut royal charters of 1031, Ker knew that paleographical dating was inexact and that one had to take into account historical, linguistic, and contextual factors, as well, in efforts to date hands. This is nowhere more apparent than in his dating of Ælfric manuscripts. For example, as a leading Ælfric scholar, Ker was well aware that the earliest manuscript of Ælfric’s Sermones catholici, London, British Library, Royal MS 7 C. xii, could not be earlier than when Ælfric began composing the homilies in or after the year 990. With this starting date, one can express Ker’s “s. X EX.” as 990 x 1000.⁵³ The two hands, which are far from identical, are polished examples of Square miniscule.⁵⁴ Peter Clemoes at first wanted to date the Royal manuscript as early as 989.⁵⁵ Although later persuaded that 989 was too early, Clemoes still maintained that the composition of the Royal manuscript was the first half of 990.⁵⁶ Malcolm Godden has persuasively argued that Ælfric sent the Second Series to Sigeric in 995, and the First Series a year or two earlier.⁵⁷ If Godden is right, Ælfric must have sent his Latin Preface to Sigeric at the same time. Several factors have bearing on the debated dating range over this important manuscript. No one has doubted that someone in medieval times cut the first folio from the first quire of Royal 7 C xii. Clemoes realized that the Latin Preface, but not the longer English one, would have fit on this missing folio.⁵⁸ He doubtless realized as well that if the excised leaf had been the Latin letter to Sigeric, a 990 dating of the Royal manuscript would be untenable.⁵⁹ Offering no alternative explanation for its absence, Clemoes rejected this possibility, saying, “We have no evidence at all to suggest that Ælfric ever thought of these prefaces as anything but complementary.”⁶⁰ His comment is unconvincing, because we do have evidence that Ælfric first requested corrections from Sigeric, before sending the first series with its Old English preface to the “unlearned” secular clergy (“ungelærede men”).⁶¹ “Now I earnestly beseech your kindness, most mild father, Sigeric, that, if you find any blemishes of malign heresy or dark fallacy in our translation, you deign to correct them through your industry” (“Precor modo obnixe almitatem tuam, mitissime pater Sigerice, ut digneris corrigere per tuam industriam, si aliquos nevos malignę heresis aut nebulosę fallacię in nostra interpretatione repperies”).⁶² The evidence is even clearer in his Latin preface to the second series: “I beseech your benignity to read through our translation, as with the previous one, and to judge if it is to be received by the catholic faithful or to be thrown away” (“Perlegat queso benignitas vestra hanc nostram interpretationem, quemadmodum et priorem, et diiudicet si fidelibus catholicis habenda est, an abicienda”).⁶³ If Sigeric was already dead when Ælfric was correcting the Royal manuscript, Ælfric would have likely removed the Latin letter to him on this first leaf, just as he removed a (by then) repetitious text on folio 64. In this case, “s. X EX.” signifies 995 x 1000, at the earliest.

The late type of Square minuscule in Ælfric’s marginal note on folio 64, written in his own hand, is particularly interesting by comparison to the very late tenth-century examples of Square minuscule used by the scribes of the main text. After deleting a long passage from folio 64r4–25 through 64v1–4, Ælfric himself wrote in the margin on the recto, “ðeos racu [is] fullicor on ð[ære] oðre bec, 7 w[e hi] forbudon on ð[ys]sere þy læs þe h[it æ] þryt þince gif [heo] on ægðre bec b[eo]” (“This account appears more fully in the second book, and we have forbidden it in this, lest it should seem tedious if it appeared in both books”).⁶⁴ His complaint makes no obvious sense unless this note post-dates the second series—that is, after 995. Following David Dumville’s description of eleventh-century vernacular script, Ælfric’s hand is “laterally compressed” with “the resulting letter-forms being narrower and more elongated than those of Square minuscule.”⁶⁵ Of particular interest is the long descender of the narrow r, which usually has a squarer look with a short descender in tenth-century Square minuscule (figs. 10 and 11). Sisam and Ker believed that this copy “was written certainly in Ælfric’s lifetime and represents the earliest extant state of the first series,” but both assumed Ælfric died circa 1020–25.⁶⁶ In fact, Ælfric’s marginal

Figure 10.

London, British Library, Royal MS 7 C xii, fol. 64r.

With permission of The British Library Board.

Figure 11.

London, British Library, Royal MS 7 C xii, fol. 64r, detail of Ælfric’s hand in right margin.

With permission of The British Library Board.

note is one part of plausible evidence that in Royal 7 C. xii Ælfric might have been correcting in the early eleventh century the manuscript of the first series that he sent to Sigeric in the late tenth century.⁶⁷

Ker was and remains an authority on manuscripts containing any of Ælfric’s homilies. The modern editors of the Catholic Homilies, Peter Clemoes, John Pope, Malcolm Godden, and (of the Prefaces) Jonathan Wilcox, all acknowledge Ker and accept his dates in their accounts of the manuscripts in their respective editions. Ker studied manuscripts at Oxford under the direction of Kenneth Sisam and wrote his dissertation on “A Study of the Additions and Alterations in MSS Bodley 340 and 342.” His Catalogue meticulously describes all major manuscripts containing any of Ælfric’s homilies item by item, allowing scholars and editors to compare the vastly different collections. To facilitate this kind of research, Ker provides at the end of his Catalogue comprehensive tables of manuscripts containing any of Ælfric’s Sermones catholici.⁶⁸ The tables are arranged in columns, beginning with the only complete collection of the first and second series, to show to what extent the homilies appear in later collections. This information provides indispensable guides to the extremely complicated textual history of Ælfric’s Sermones catholici.

Steeped in knowledge of this history, Ker was well aware that the appearance of a script might be deceptive. When he noted in his introduction that “sometimes there are special reasons for a different form of dating” a script, Ker was alluding to the complex textual history of Ælfric’s Sermones catholici. He specified three Ælfric manuscripts—Oxford, Bodleian Library, Bodley MSS 340+342 (Ker 309, MS D, the topic of his dissertation), Cambridge, Corpus Christi College, MS 162 (Ker 38, MS F), and London, British Library, Cotton MS Vitellius C. v (Ker 220, MS H)—which “contain texts by Ælfric and are in a type of script which seems to have gone out of use soon after 1000.” If one went by their script alone, Ker acknowledged, “a date for these manuscripts a little before 1000 would be palaeographically possible . . . but historically not very likely.”⁶⁹ In this understated observation, Ker is pointing out that Square minuscule continued after 1000 in Ælfric manuscripts. Paleographers in general recognize the importance of history, including the history of a text, to paleographical judgments. Julian Brown, for example, stressed the importance of paleographers, historians, and philologists working together to derive meaning from paleographical evidence.⁷⁰ With these additional considerations in mind, Ker placed these three manuscripts in the first third or, if Dumville is right, the first quarter of the eleventh century, s. XI in.⁷¹

As Elaine Treharne observes, “Extant copies and adaptations of Ælfric’s homiletic texts do not . . . reveal a knowledge of Ælfric as the actual author, and thus while it appears his work had prestige and authority within its circulating institutions, it is unclear whether or not those producing his material had any idea that it was his.”⁷² The disappearance of Ælfric’s Prefaces from these early eleventh-century manuscripts may suggest that they post-date both Archbishop Sigeric, circa 994–95, and the unlearned readers, ungelærede men, who could not read Latin and so first received them, as Ælfric says, with an Old English preface. Without Ælfric’s name in the prefaces, the unauthorized manuscripts in effect became anonymous collections of homilies. They could not have come from Ælfric’s workshop, because, in addition to omitting his name, all three ignore his express wishes. In his first Latin Preface written for Sigeric, Ælfric tells the Archbishop that he is preparing a second series of forty homilies intended to be read on alternating years. He tells Sigeric, “we give license, if it pleases anyone better, to arrange both into one book” (“tamen damus licentiam, si alicui melius placet, ad unum librum ambos ordinare”), but it does not appear that he ever contemplated the kind of transformations that show up in these three manuscripts.⁷³ In the Latin Preface that he wrote to Sigeric for the second series, he drops all mention of combining the two series.⁷⁴ There is no hint of this license in either of the Old English Prefaces he subsequently wrote for the secular clergy. In them, he uses the same strict admonition emphasizing the importance of copying the exemplars accurately.⁷⁵ In a final prayer at the end of the second series, Ælfric explicitly says: “Gif hwa ma awendan wille, ðonne bidde ic hine for godes lufon þæt he gesette his boc onsundron fram þam twam bocum ðe we awend hæbbað” (“If anyone wants to translate more, then I ask him for God’s love that he set his book apart from these two books that we have translated”).⁷⁶

Figure 12.

Oxford, The Bodleian Library, University of Oxford, MS Bodley 340, fol. 1r.

As Ker detailed in his dissertation, Bodley 340+342 (Ker 309, MS D) creates its own first and second series, combining non-Ælfrician homilies with a mixture of Ælfric’s homilies from his two series.⁷⁷ Bodley 340 begins with a non-Ælfrician homily, followed by a conglomeration of Ælfric’s firstand second-series homilies and other non-Ælfrician ones (fig. 12). Similarly, Bodley 342 begins its own “second series” with Ælfric’s first homily of his First Series, followed by another commingling of homilies (fig. 13).⁷⁸

Figure 13.

Oxford, The Bodleian Library, University of Oxford, MS Bodley 342, fol. 1r.

Corpus Christi College Cambridge MS 162 (Ker 38, MS F) is a shorter, mixed collection of fifty-five homilies, drawing heavily on Ælfric’s first and second series, with some non-Ælfrician material.⁷⁹ In addition to dissociating Ælfric from his homilies and reordering them against his express wishes, both of these collections ignore his instructions, which he added in his own hand in the earliest manuscript, “Ciriclice þeawas forbeodað to secgenne ænig spell on ðam ðrim swigdagum” (“Church customs forbid saying any homily on the three silent days,” that is, on Holy Thursday, Good Friday, and Holy Saturday).⁸⁰ Ælfric, if he was still alive, would be displeased to see sermons added for these days.⁸¹

The complexity of the early eleventh-century transmission, involving a deliberate reimagining of Ælfric’s homilies in Bodley 340+342 (MS D) and CCCC MS 162 (MS F), is well illustrated by Malcolm Godden’s stemma showing the relationships of the DEF group of manuscripts for the second series (fig. 14).⁸² The stemma shows that these D and F manuscripts grew from different branches of a lost common archetype, X⁽¹⁾, while four of the homilies in F derived from yet another non-extant archetype, X⁽²⁾. Although Dumville, seeking to place these manuscripts in the 990s, argues that “Ælfrician textual history . . . is in fact chronologically unquantifiable,” one cannot reasonably explain the evolution of these manuscripts as a simple, rapid circulation of the original versions sent to Archbishop Sigeric by 995.⁸³ Dumville attributes to the extant manuscripts what only partly applies to their lost archetypes X⁽¹⁾ and X⁽²⁾.

Figure 14.

Godden, Ælfric’s Catholic Homilies 2, p. lxi.

Reproduced by permission of Malcolm Godden and the Council of the Early English Text Society.

Adding more complexity to what requires a different stemma from another lost archetype, Cotton Vitellius C. v (MS H) remarkably preserves the evolution of a newly conceived collection at different times through the first half of the eleventh century. Ker says that “the older leaves are in a graceful late type of square Anglo-Saxon minuscule, s. X/XI.”⁸⁴ However, in his introduction he explains that “sometimes there are special reasons for a different form of dating.” Four manuscripts, he says, including Bodley 340+342 (MS D), CCCC MS 162 (MS F), and Vitellius C. v (H), “contain texts by Ælfric and are in a type of script which seems to have gone out of use soon after 1000. A date for these manuscripts a little before 1000 would be palaeographically possible . . . but historically not very likely. ‘s. XI in.’ seems the best date for them.”⁸⁵ Clemoes accepted that the first stage of Cotton Vitellius C. v came later than Bodley 340+342 and CCCC MS 162.⁸⁶ The position of the Vitellius text in Ælfrician textual history is thus chronologically quantifiable. Godden demonstrates that it was a “second-recension” manuscript, “reflecting a uniformly late stage of his work.”⁸⁷ Dumville is accordingly misguided to ignore textual history in attempting to place these three manuscripts in the 990s on the basis of handwriting alone.⁸⁸

Ker would not have disagreed with Dumville’s assessment that s. X/XI for the Beowulf manuscript was a compromise that could be more precisely dated s. XI in., early eleventh century, as in the case of Cotton Vitellius C. v. Ker assigned the same dating range to Cambridge, University Library MS Gg.3.28, the only complete set of the first and second series of Ælfric’s Sermones catholici that contains the Latin and English prefaces to both series. As Ælfric began composing the Catholic Homilies in 990 and submitted the second series to Archbishop Sigeric in 995, this manuscript was written post-995, not 993, as Dumville felt free to argue by neglecting textual history.⁸⁹ In his 1968 paleographical description of the Will of Æthelgifu, Ker seems to have broadened his dating range for “s. XI in.” and “s. XI¹” to 990 x 1040 for some key examples of Square minuscule.⁹⁰ There he says that “the failure of Anglo-Saxon minuscule at the end of the tenth century led to a period of some fifty years, approximately 990 to 1040, during which there was great variety in the writing of books and charters in England, with some good writing and, especially in the vernacular, much rather poor imitative writing with no character of its own.” “In this period,” Ker explains, “great differences are to be seen between the hands of scribes writing at the same time and in the same place, between, for example, the first and the second hand of Beowulf and the major and minor hand of the early Ælfric manuscript Gg.3.28.”⁹¹ Peter Stokes says much the same thing in arriving at the same conclusion: “Although a coherent scribal practice did ultimately emerge, the intervening period [ca. 990 to 1035] was one of great variation not only between practices at different scriptoria but even between scribes writing at the same time, at the same place, and in the same book.”⁹²

By “imitative writing,” Ker is referring to examples of late Square minuscule in Old English, as these three manuscripts attest. Ker tersely distinguishes the two hands of CUL Gg.3.28 by describing the copy as “mainly in an unusual forward-sloping hand” with “seven short passages . . . in an earlier-looking and more ordinary kind of script.”⁹³ He explicitly says that the main scribe of this Ælfric manuscript and the first scribe of Beowulf “have gone some way toward Caroline minuscule.”⁹⁴ For Beowulf, Dumville argues that “it seems logical to deduce that the transitional phase (represented by those manuscripts containing different Insular script-types contemporaneously written) should have been placed by Ker in the first part of the eleventh century” (fig. 15). However, Dumville fails to apply the same logic to CUL Gg.3.28. Both scripts of CUL Gg.3.28 are compressed and elongated with longer descenders, even for r, than tenth-century Square minuscule. The script also exhibits “the lateral compression so characteristic of eleventh-century English vernacular script,” yet in this case Dumville describes it as “an early example of the move toward new proportions, being perhaps datable as early as 993.”⁹⁵

Figure 15.

Cambridge, Cambridge University Library, MS Gg.3.28 fol. 225v, lines 19–30.

Reproduced by kind permission of the Syndics of Cambridge University Library.

Although aware of its eleventh-century appearance, Stokes does not contradict Dumville. “The manuscript was written in the early 990s by two scribes,” he says, “one of whom wrote a late form of Square minuscule, while the other wrote a very early form of English Vernacular minuscule.”⁹⁶ Stokes had misgivings, however. On the next page he acknowledges, “if the dating of CUL Gg.3.28 is correct, then this presupposes that Style I was written at Winchester in the 990s even though we have no surviving examples which can be dated so early.”⁹⁷ The sensible conclusion is not to push Ker’s dating range of 990 x 1040 to its earliest outer limits, but to accept that CUL Gg.3.28 should be dated, by the same reasoning Dumville used for the Beowulf manuscript, s. XI in., 1000 x 1030, the early eleventh century, possibly in the reign of Cnut. Ker also placed the Will of Æthelgifu in the same broad, amorphous period of 990 x 1040 as the Ælfric manuscript CUL Gg.3.28 and the Beowulf manuscript. While all three exhibit late forms of Square minuscule, they are all quite different from one another. Together they offer graphic evidence that there is no uniform Square minuscule during this period. Indeed, Ker described the hand of the Will of Æthelgifu as “dully imitative,” “a collection of letter-forms stiffly independent of one another.”⁹⁸

Figure 16.

Princeton, Princeton University Library, William H. Scheide Collection, M 140 (Will of Æthelgifu), detail of line 51. Courtesy of Special Collections, Princeton University Library.

Despite Ker’s broadest paleographical dating to 990 x 1040, Lord Ren-nell of Rodd claims that “On examination the document, written in Old English, can apparently be dated on internal evidence to 980–990 A.D.”⁹⁹ Neither Ker nor Dorothy Whitelock, whom Lord Rennell had persuaded to undertake, respectively, the paleographical analysis and the translation and commentary, give such an early dating range. Ker’s extreme, fifty-year outer limits were 990 x 1040. The thirty-year range for Ker would be 1000 x 1030, early eleventh century. In his analysis of the script, Ker highlights several eleventh-century features. “The square flat-topped a is an imitation of the square a common in late-tenth-century” Square minuscule, he begins, “but it is essentially different, being in fact no more than a u closed at the top.” Most examples of this form, he concludes, are “early eleventh century,” citing examples from Bodley 342 and Bodley 865, both discussed above. The form of d “is commoner in the eleventh century than in the tenth,” and the low-e combinations with g and t “occur frequently in the first hand of Beowulf.” Perhaps the most striking feature is the inadvertent use at line 51 of Caroline r in dohtur (fig. 16). While he draws attention to these eleventh-century aspects and gives a dating range that covers most of the early eleventh century, Ker says, “The evidence, such as it is, is not inconsistent with a date in the last decade of the tenth century.”¹⁰⁰ In this concession, Ker was accommodating Dorothy Whitelock, who argued on historical grounds for a date “probably not long before 990,” even earlier than Ker’s earliest outer limit.¹⁰¹ In her recent edition of the will, Julia Crick more realistically decides that “the script of the extant single leaf . . . belongs most comfortably in the first half of the eleventh century.”¹⁰²

Figure 17.

Will of Æthelgifu. Princeton, Princeton University Library, William H. Scheide Collection, M 140 (Will of Æthelgifu), detail of line 39. Courtesy of Special Collections, Princeton University Library.

The basis for Whitelock’s historical argument, the “internal evidence” Lord Rennell references, is Æthelgifu’s request of “hire hlæfdian,” her lady the queen (who is not named), to allow her beneficiary, Leofsige, to serve the ætheling, or prince (who is not named) (fig. 17).¹⁰³ Whitelock favored Æthelstan, the first son of Æthelred II and Ælfgifu of York, as the unidentified ætheling. Because Æthelgifu makes the request of the queen, this ætheling must have been a minor at the time. Whitelock considered it “unlikely that Æthelstan was born before 985” and concluded that, if he was the ætheling mentioned in the Will, “the outside limits of date would be c. 985 to 1002, probably not long before 990.”¹⁰⁴ She believed that Ælfgifu of York was already dead and that the person Æthelgifu referred to as “her Lady” was Æthelred’s mother, Ælfthryth, “the Queen Dowager, who appears prominently in records of the period.”¹⁰⁵ Throwing this theory into serious doubt, Simon Keynes has noted in another connection that Ælfthryth is “wholly absent [in royal diplomas] in the period 984–93 (when she seems to have been out of favour).”¹⁰⁶ Moreover, it is unlikely that anyone would refer to the Queen Dowager as hlæfdige, “lady.” Pauline Stafford points out that “mater regis” (“King’s Mother”) was a common title for Ælfthryth, with only four exceptions, where she is called “regina” (“queen”).¹⁰⁷ The most serious problem with the identification of Æthelstan as the unnamed ætheling, as Whitelock recognized, is that by 990 Æthelred and Ælfgifu (and possibly another unknown wife or wives before Emma of Normandy in 1002) had four sons—four æthelingas—and six æthelingas by 994. All of them were apparently brought up by Ælfthryth on her estate called Æthelingadene, the valley of the princes, a name accentuating the inevitable confusion.¹⁰⁸ All six sons witness a royal charter in 1001.¹⁰⁹ Similar, but less extensive, confusion would arise for Alfred and Edward, the two æthelingas of Æthelred and Emma of Normandy (also known officially as Ælfgifu).

It is surprising in view of her thorough commentary that Whitelock did not consider Harthacnut, the only son of Cnut and Ælfgifu/Emma of Normandy, as the ætheling Æthelgifu meant. Cnut heightened his legitimacy by marrying Æthelred’s widow Emma, accepted and recodified Anglo-Saxon law, and surely would have followed English tradition by using ætheling, the Old English word for prince, for their son. Harthacnut, born in 1018 near the midpoint of Ker’s and Crick’s dating ranges for the Will, is in fact the only unambiguous ætheling of them all. Æthelgifu’s address to “hire cynchlaforde ond hire hlæfdian” (“her King and her Lady”) is particularly appropriate for Cnut and Ælfgifu/Emma. Keynes has drawn attention to the close relationship between them, and has highlighted how her subjects consistently referred to her as “seo hlæfdige” from the year she came to England from Normandy in 1002 until the year she died in 1052.¹¹⁰ When she dies in 1052, the C version refers to her as “seo ealde hlæfdige” (the Queen Dowager or Queen Mother, literally but honorifically “the old lady”), while the D version simply calls her seo hlæfdige.¹¹¹ Keynes gives a few pertinent examples from her lifetime: “In 1020 Archbishop Wulfstan directed a writ “to his lord King Cnut and to the Lady Ælfgifu” (“Ælfgyfe þa hlæfdian”). In another document, King Cnut “and the Lady Ælfgifu” (“Ælfgifu seo hlædige”) granted permission. At lower levels of the social hierarchy, Mantat the Anchorite announced his will in the form of a document addressed to King Cnut and Queen Emma (“Emma hlæfdie”).”¹¹² If Æthelgifu was addressing them, too, it was natural for her to refer twice to the queen as her Lady (“hire hlæfdian”) and to Harthacnut simply and unambiguously as the ætheling. Harthacnut was a minor circa 1020–25, providing credible limits for the date of Æthelgifu’s will.

Likewise, the first half of Cnut’s reign, 1016–25, is a reasonable time period for the late form of Square minuscule practiced by the second scribe of the Beowulf manuscript. At least some communities during Cnut’s reign, it is now apparent, preferred to use this kind of script for texts in Old English, perhaps because they were meant for readers who did not know Latin, or perhaps because some scribes were proud of an indigenous script. This enduring preference is manifest in its use in late charter boundaries in Old English, as well as in the homiletic tradition, poetic and prose translations of the Bible, and prose and poetry about “geardagum,” “former days, days gone by,” including stories in the Nowell Codex about Danish Scyldings, the ancestors of Cnut the Great. The first scribe who copied stories about Alexander the Great and the first part of Beowulf was, as Ker said, “influenced by Caroline minuscule,” and indeed had “gone some way toward Caroline minuscule.”¹¹³ There is evidence that this first scribe tried to suppress some of this foreign influence and conform to a more normal English Vernacular minuscule. The scribe sometimes inadvertently used Caroline a on the opening pages of the codex, but then got rid of all vestiges of it.¹¹⁴ The same scribe uses Caroline s, however, throughout. The third Beowulf scribe, first identified as “a later hand” by Julius Zupitza in 1882, rewrote all of folio 179 (fol. 182 in the BL foliation) in a more normal Style I Early Vernacular minuscule.¹¹⁵ Dumville acknowledged “evidence for scribal development and adaptation to more modern forms” on this folio.¹¹⁶ But in the end Dumville is uncompromising: “The Beowulf-manuscript provides for the composition of the poem a terminus ante quem which excludes the reigns of Cnut and his sons.”¹¹⁷

Dumville’s focused efforts to move each and every late example of Square minuscule between 990 and 1016 have as much to do with the dating of Beowulf as the dating of the Beowulf manuscript. Yet, maintaining Ker’s broader dating of these manuscripts allows scholars to consider early eleventh-century influences in and on all these texts, including a credible transmission for Ælfric manuscripts and an unambiguous ætheling for Æthelgifu’s hlæfdige. Instead, a recent collection reassessing the dating of Beowulf from an exclusively early perspective is free to ignore manuscript evidence. Michael Drout, in “ ‘Give the People What They Want,’ ” discourages anyone thinking of analyzing the manuscript evidence that the Beowulf we have was composed during Cnut’s reign. “Detailed analysis of Kiernan’s specific claims is beyond the scope of this paper,” he says, “but it is sufficient to note that if his argument about the eleventh-century origin of Beowulf is correct, then arguments for a ninth-or tenth-century poem are just as wrong as those for a seventh-or eighth-century poem.”¹¹⁸ No one wants that dismal outcome, which perhaps explains why the latest possible date for Beowulf is currently a lonely place. Rather than simply dismissing the hypothesis because so few people have accepted it, serious scholars should begin to weigh all of the evidence (not just pieces here and there) that supports it. Anyone accepting the extensive, corroborating manuscript evidence after analyzing it must still reconcile many powerful arguments espousing earlier dates, from the seventh through the tenth centuries. They cannot reconcile circa 750 with circa 1020, but they must explain, for example, how a seemingly ancient linguistic form, or metrical type, or oral formula, or topical allusion might still be used by poets and understood by readers and listeners in the early eleventh century. Above all, they should not begin by changing the previously accepted paleographical dating of the Beowulf manuscript. One cannot test a hypothesis by removing its premise. Neil Ker has provided considerable proof that scribes continued to write late versions of Square minuscule in manuscripts containing Old English in the time of Cnut the Great.