The second version of the Latin treatise known now as the Leges Edwardi Confessoris, which was not issued by King Edward the Confessor (1042–1066), but is an anonymous composition, perhaps originating in Lincoln, and expanded by the same author as the first version soon after its creation in the 1130s. Like the first version, it concerns different aspects of the peace of the Church and king, as well as enforcement of the peace, but adds chapters on the history of the laws of Edgar (king of England, 959–975), unjust execution, usury, and pledges and witnesses for sales to produce a 39-chapter treatise.

Digital edition

Edited by Bruce O'Brien


Other versions of this law

Text filiations


by Bruce O'Brien

The treatise known as the Leges Edwardi Confessoris survives in several versions, all from the twelfth and early thirteenth centuries. The treatise likely represents a regional understanding of the customs an episcopal see in the Danelaw would have to understand in the management of its estates and people. The most recent study of this treatise, my God’s Peace and King’s Peace: The Laws of Edward the Confessor (Philadelphia, 1999), provides a detailed examination of the contents of the treatise and the development of its different versions. The following introduction is derived from that study and concerns the structure of the treatise and the relationship of the witnesses to the text.

Development of the Text

Despite the claims of the treatise, the laws of this treatise (hereafter ECf1) are not in any sense the laws of the Danes, Norwegians, English, Welsh, and Picts living under William I, let alone a record of all the English customs that had force under the Confessor. Instead, the ‘laws and customs’ in the Leges Edwardi concern different aspects of the peace or, better, the types of legally established peace and security that the Church and king had power to create and administer. This choice means that many subjects of law do not appear, as they rarely occasioned any breach of peace—land law is the most visible omission. The author has not signaled his focus on peace in any obvious way; no title or rubrics announce it. The earliest versions of the Leges Edwardi occur as running text without rubrics marking the different aspects of the peace. Nevertheless, the first version, consisting of thirty-four chapters (ending at 34.1a), falls rather neatly into five reasonably distinct parts, including prologue and epilogue. The three that form the substantive body of the treatise are all devoted to some aspect of the peace, royal, or ecclesiastical. The divisions are as follows:

[Prologue (Prol.)]

1. Peace of God and Holy Church (1–11)

2. Peace of the king (12–19)

3. Enforcing the peace (20–33)

[Epilogue (34)]

The first section, on ‘the peace of God and Holy Church’, is a collage of a continental text of the peace of God and regulations governing sanctuary from a Carolingian capitulary, pasted onto what appear to be original accounts of English customs, which, incidentally, differ little in thrust from the imposed foreign texts. The author construed this ‘peace of God and Holy Church’ broadly enough to cover matters from the inviolability of churches to the conduct of ordeals. He has bound these chapters together by emphasizing the responsibility of royal justices to support the Church's power to punish those who broke its peace. Thus, the author concluded, ‘will the sword justly help the sword’.

The treatment of the royal peace that follows the peace of God mirrors the Church's peace in outline — a basic definition followed by special conditions or aspects. The parallel images supply an internal symmetry. For example, as those traveling to churches for prayer are protected by the peace of God and the Church, so those journeying on the king's roads and waterways are under his peace. Noticeably absent, though entirely consistent with how the Church's pax had been handled, is any discussion of the abstract legal meaning of grið 'peace' or any mention of its double or cousin, mund 'protection, peace'. The author did not state principles; he did not imagine that the applications he described would have had greater clarity had he done so. For the author to have abstracted such principles would have made him stand out from almost all of his peers. Instead of principles, the author described some of the circumstances in which the peace of the king can be broken. He provided only a selection from the many possibilities — for instance, breach of a peace received from the king's hand (the kinges hand salde grid) is mentioned, but not the king's peace given by an earl or sheriff.

Finally, the author turned in the third section to enforcement of the peace and addressed the question, Who is looking after whom? Here the weight of enforcement is placed on various forms of suretyship: the permanent liability of friborg or frankpledge, the temporary liability of hospitality, and the special responsibility of the king to protect the Jews. There is also some discussion of the officials who governed the peace and judged offenders, as well as of the fine levels in the Danelaw and in areas under English law. The author's mention of the large fines levied in the Danelaw provides the transition to the series of events in the epilogue, namely, William I's wish to impose that fine and all other Norwegian laws, the immediate outcry of his English nobles, and his accession to their request that he confirm the laws of Edward that they had just recited.

Surprisingly, the author made no visible use of texts of Anglo-Saxon law, as had most of the other treatise writers. Felix Liebermann, who edited the studied the text in the alte nineteenth century, thought that perhaps a few chapters were derived from contemporary treatises or acts. For example, he thought that chapter 2.9 might have been derived from William I's writ separating ecclesiastical and royal pleas. More obviously, the Leges' chapter 10 appears to have come from the same written source as Leis Willelme's chapter 17. However, these are exceptions. The author seems, as Liebermann put it, not to have troubled himself in the least to discover what had already been written about preconquest law. Except for these few resemblances and an infrequent appearance of a continental source, the treatise constitutes a description in the author's own words of the laws and customs of England.

The Leges Edwardi began to circulate in this thirty-four-chapter form. Six manuscripts attest to this. Whether this circulation was planned or not is difficult to say. It was the habit of classical and medieval authors to circulate their work privately as or as if a draft. Geoffrey of Monmouth and John of Salisbury both did so. Ralph Niger, the twelfth-century English chronicler, urged William, archbishop of Rheims, ‘to have his book De re militari examined to see if it contained any error’. In like manner or with similar intent the author of the Leges Edwardi might have passed around a draft of his work. But it may also be the case that it was released without any plan and before its completion. Guernes of Pont-Sainte-Maxence claimed that the first version of his Life of Thomas Becket ‘was stolen from me by scribes before I had corrected and finished it, before I had balanced the bitter and the sweet or shortened the passages that were too long. It is sometimes incorrect and is not complete’. The earliest version of the Leges may represent a similarly unauthorized release.

From Version 1 to Version 2

Subsequent revisions suggest that the treatise was not viewed as finished in its first version. But here the trail becomes complicated. At issue is who revised and enlarged the treatise — the author of version 1 or one or several revisers? Did he or they work in collaboration with or independent of the author? What is at stake is small but in fact important. Answering the question is made more difficult by the possibility that before the next major revision was undertaken, resulting in the thirty-nine-chapter text, an intermediate thirty-seven-chapter text was completed and copied. The evidence is inconclusive but still deserves consideration. This intermediate version may have constituted a general but slight revision of the style and an enlargement of the treatise by the addition of the history of Edgar's laws (c. 34. 1b–3), the story of Edward the Confessor's search for an heir among the grandchildren of Edmund Ironside and the subsequent rejection of Edgar Aetheling and nomination of William of Normandy (c. 35), a chapter on procedure to be followed when a relative has been executed unjustly as a thief (c. 36), and a rhetorical set piece to explain the Confessor's attitude toward usury (c. 37). While one manuscript witnesses this thirty-seven-chapter text, its variants may equally suggest that a thirty-nine-chapter text was shortened and then collated against a text of the first verison. This would explain why this text shares so many readings with the archetype of the first version (alpha).

Perhaps not. If this manuscript does witness an intermediate version, it may be possible to say something about when this version was finished. Chapter 36 may have been added in response to a royal campaign against theft, which was reported by the Peterborough Chronicle to be on the rise at the beginning of Stephen's reign. Chapter 37 could be a byproduct of the prohibition of usury by the legatine council at Westminster in December 1138. While knowledge of this prohibition was likely widespread in the English Church, it is nevertheless striking that a council's prohibition of usury in 1138 would be backdated in the Leges to Edward the Confessor's reign. The period immediately prior to the council included the first major campaign to have Edward canonized, led by Osbert de Clare, prior of Westminster, who was not only a biographer of the king but also a notorious forger of charters in favor of the abbey that the Confessor had rebuilt and dedicated in 1065. It is possible that Osbert had an influence on the composition of chapter 37, and perhaps even on its addition to the Leges.

This intermediate version may be a figment born of truncation, verbal contamination, and careless copying. However, either it or the first version was revised stylistically and received further chapters, bringing the total to thirty-nine: one on selling goods without sureties (c. 38) and another on slaughtering and selling animals without witnesses (c. 39). The fact that the last chapter is partly in the first person plural suggests that it may be from a written source, perhaps from a royal edict, since the first person plural is usual in royal charters and edicts. Overall, changes in style throughout the treatise were slight. It is not always possible to distinguish errors in the archetype of the first version (alpha) from the author's own words, and thus it is not always possible to prove how the first version was revised. But in general, ignoring the obvious errors, the revision is visible on a number of levels. In some places where the first version is overly general, the second version, as if responding to a demand to eliminate any ambiguity, is specific. For example, the revision's "atrium" clarifies version i's "illam" as a more discrete goal for the fugitive seeking sanctuary. In other places, this revision cluttered what was already clear. The revision adds "omnes" to "eorum possessiones" (c. 1.1) and "semper" to "ad hora nona" (c. 2.5), with little resulting improvement of sense. In these cases, it appears that the changes are merely the result of the author's changing taste.

Or, others have suggested, the changes are the result of a reviser's taste. Some have questioned whether the original author was responsible for the added chapters and general revision. Liebermann entertained the possibility that the last five chapters (cc. 34.1b–39.1) were added by someone other than the author, but had decided, for a mixture of reasons (both good and bad), that all thirty-nine were the work of one person, and that the thirty-four-chapter text was in fact probably the later product of some capricious scribe who had cut off the last five chapters and modernized the language throughout. For Liebermann, it was unlikely that there had first been a thirty-four-chapter form, a view that informed his interpretation of textual variants in the manuscripts of the long version. His conclusions went unchallenged (and largely unexplored) until the 1960s, when H. G. Richardson and G. O. Sayles argued that all that followed chapter 34. 1a constituted an interpolation that "marred" the symmetry of the original version. For this reason, the additions could not have been done by the author, but had to represent the work of a less gifted reviser. Although there is some sense to their argument, it seems probable that these additions were made by the author of the original because they betray, as Liebermann noticed, the same interest in etymology and historical explanation found scattered throughout the earlier chapters.

The Leges Edwardi in its first and second versions resembles the other Anglo-Norman legal treatises as much as can be expected from sharing their subject. The treatise's plan is as visible as the explicit program of the Quadripartitus (which includes its second book, the Leges Henrici Primi), but it lacks Quadripartitus's ambitious coverage. The Leges Edwardi's treatment of its subject is in particulars reminiscent of the Articuli of William I, but, except possibly for chapter 39, it is not in any obvious way a pastiche of current legislative enactments and custom as that short tract appears to be. It is most comparable to the earliest recension of the Leis Willelme, a version witnessed only by what is also the oldest manuscript copy of any version of the Leges Edwardi, British Library, Additional MS 49366. This early version of the Leis is for the most part free from dependence on any textual source and covers many of the same points as the Leges Edwardi— the peace of the Church, sanctuary, tithes, royal peace, and murder fines. The difference in language between the two makes comparison more complicated but not inappropriate; the Leis's adoption of Old English terms for most rights and crimes suggests that its author felt the same overriding practical concern to render English law accurately that is evinced by the author of the Leges Edwardi and all but one of the other Latin legal treatise writers and translators. The Leges, however, has much greater clarity of purpose and organization than the Leis Willelme and, in fact, has arguably more structure than any of its peers.

The Edition

There are more than forty manuscripts of the Leges Edwardi, fourteen of which probably reflect the work of the author (versions 1 and 2) rather than of subsequent redactors (versions 3 and 4). All fourteen belong to two versions of the treatise (with one possible exception): to version 1, ending at c. 34.1, or to version 2, ending at c. 39.2. All manuscript descriptions have been arranged by sigla, which in most cases are those used by Thorpe, Schmid, and Liebermann in their editions. The abbreviations and designations are those used by N. R. Ker. Manuscripts are vellum or parchment unless otherwise stated. Discussion of contents and paleographical and codicological aspects is meant to provide only what is relevant to an understanding of the manuscript context in which the Leges Edwardi is found. Descriptions of witnesses to alpha, the archetype of Version 1, can be found in the introduction to its edition on this website.


Ar London, College of Arms, Vincent 98, [fols. 36–41v, unnumbered]. 315 × 210 mm. Written space: 260 × 163 mm. 41–48 long lines. When Liebermann completed the first volume of his Gesetze der Angelsachsen in 1903, he confessed that he could not locate the manuscript holding a text of the third version of the Leges Edwardi that was supposed to be in the library of the College of Arms. This manuscript does indeed exist, but was rebound with two other manuscripts in the eighteenth century and thereafter proved difficult to locate. It does not hold a text of the third but of the second version, transcribed in 1564 from a "parchment exemplar" provided to the copyist, according to a marginal note, by a person identified only as "Lovelace." This was likely Sir William Lovelace (d. 1577), who was admitted at Gray's Inn in 1548 and so might have merited the title Iurisperitus given to the owner in that marginal note. The College of Arms acquired Vincent MS 98 as a bequest from Ralph Sheldon in 1684, though it takes its name from its previous owners, John Vincent (d. 1671) and his father Augustine Vincent (d. 1626). The manuscript appears in the 1690 catalogue (p. 29) as a Vincent/Sheldon MS, no. 98, and again in the 1788 catalogue, which includes the note that no. 98 had been rebound with Vincent MSS 65 and 75 in August of that year. The Leges is preceded by extracts and notes relating to the time of Henry VIII and earlier, written in a hand contemporary to that which copied the Leges. The copyist of the treatise was also responsible for the two texts that follow it: the Libertas civium Londoniensis (Liebermann's Lib Lond; fols. 41v–42) and the Instituta Cnuti (fols. 42v–49v). A treatise entitled "For the Mariage [sic] of A Prince's Daughter" (fols. 50–60) follows in another late-sixteenth-century or early-seventeenth-century hand, and it is followed in turn by notes on accounts (fols. 60v–63) and a list of "the Nobility of England" (fols. 64v–65).

Cl London, British Library, Cotton Cleopatra A. XVI, fols. 56–61v. 173–76 × 125 mm. Written space: 125–30 × 95–100 mm. 31 long lines. Collation (part 1, fols. 1–68, only): two singletons (fols. 1–2) 1–78 86 94. This manuscript consists of two formerly independent manuscripts that were not joined until after the fifteenth century. One manuscript, now folios 69–195, holds the only surviving copy of the Westminster Chronicle that includes the section covering the late fourteenth century. The other (fols. 1–68) is a collection of texts relating principally to law and the exchequer. This manuscript contains a thirteenth-or early-fourteenth-century copy of the Dialogue of the Exchequer (fols. 3–34), a text of the Leges Edwardi ending at c. 37.1, and several writs and short treatises concerning coinage and the exchequer. There are also a number of extracts taken from memoranda and plea rolls. All of the quires have been rearranged since the late sixteenth century, when their order, in so far as it can be reconstructed from a sixteenth-century table of contents (fol. 65v), was quires 7–9 (fols. 51–68) followed by quires 1–6 (fols. 3–50). Quire 9 may in fact have originally followed quire 6; the table of contents does not provide sufficiently detailed titles for reconstructing with certainty the original order of all the texts. Quires 1–9 were set in their present order sometime after the late sixteenth century and before the late-seventeenth-or early-eighteenth-century table of contents was added on a loose leaf at the front of the manuscript (now fol. 2). The script of the Leges Edwardi dates approximately from the first half of the fifteenth century; a number of the extracts from official rolls concerning appointments to the office of comptroller up to 14 Henry VI (1435/36) are in a hand contemporary to the one responsible for the Leges. The Leges Edwardi was probably copied before these extracts were added. It is also possible that, given the manuscript's contents, it originated with someone interested in the work of the exchequer at Westminster.

Ha London, British Library, Harley MS 785, fols. 6–iiv. 318–20 × 207–10 mm. Written space: 257–73 × 150–55 mm. Double cols. 37–41 lines/ page. Most of the folios of this paper manuscript have been separated into leaves and mounted on strips, making collation impossible. This late-sixteenth-or early-seventeenth-century collection of legal and historical texts includes mostly extracts from Lambarde's Archaionomia, Polydore Vergil's Historia Anglica, and Raphael Holinshed's Chronicles, but also, inter alia, copies of the Instituta Cnuti, version 2 of the Leges Edwardi, and the Libertas ciuium Londoniensis as an appendix. The last three texts (In Cn, ECf, Lib Lond) were transcribed from a manuscript that is no longer extant. The hand that copied the extracts from the published sources as well as most of the texts in the manuscript up to fol. 103 also copied the Leges. The rest of the manuscript, which may have once traveled on its own, is principally a commonplace book of John Burche, one of the barons of the exchequer under Elizabeth I.

Hk London, British Library, Additional MS 49366, fols. 144v–154. 152–56 × 108–10 mm. Written space: 122–28 × 73–78 mm. 25 long lines, though fols. 17–24, 108v, 109v–16, and 141–57 have 26. Ruling: fols. 17–107r are ruled with pencil; fols. 107v–16 are ruled with a hard point, with occasional penciling; and fols. 141–57v are ruled with hard point with a fair amount of penciling, though in a pattern distinct from that of earlier folios. The impression rather than the pattern of ruling changes midway through 105, when a scribe with a lighter touch took over. Collation: 1–26 34 (early modern vellum quires with central bifolia of paper) 4–128 13–1410 158 16–214 (these quires are early modern vellum additions) 2210 wants 5 and 7, probably blank (fols. 141–148) 238 a singleton (fol. 157) 24–256 26 six (fols. 169–174, early modern additions). The medieval quires of this important manuscript hold a copy of Quadripartitus (fols. 17–105), Ulpianus de edendo (fols. 105–116v), the only text of the earliest version of the Leis Willelme (fols. 141–44v), the earliest copy of the second version of the Leges Edwardi, to which have been joined selections from the Consiliatio Cnuti, consisting of some or all of thirty-nine chapters from Cnut, all of Be Blaserum and Forfang, and the prologue and six chapters from the Hundred Ordinance. The rubricator of Ulpianus de edendo also added the rubrics or rubricated initials and capitals to the texts of the Leis Willelme, Leges Edwardi, and Consiliatio Cnuti, which suggests that despite the intervening early modern quires, the texts listed above were all part of the same manuscript. The medieval portions are the work of four hands: Quadripartitus is the work of a single scribe; two scribes worked on Ulpianus de edendo (fols. 105–108v; 109–116); and one scribe wrote Leis Willelme, Leges Edwardi, and Consiliatio Cnuti. The dates for the script of the Leges Edwardi have ranged widely: from the first quarter of the twelfth century to ca. 1230. The hand responsible for Quadripartitus is mid-twelfth-century. The hands of Ulpianus de edendo, Leis Willelme, and Leges Edwardi are from the third quarter of the twelfth century.

Hr London, British Library, Harley MS 1704, fols. 7v–12v. 255–65 × 177–80 mm. Written space: 193–200 × 132–40 mm. Double cols. 42–44 lines/page. Collation: the manuscript consists of a single quire, 112 (fols. 1–12), now bound at the front of a later medieval paper manuscript (fols. 13–106). These added quires contain no legal material. The vellum section includes two legal treatises, the Consiliatio Cnuti (fols. 1–7v) and a text of version 2 of the Leges, that have been joined together, without break, under the title "Incipiunt leges que uocantur leges Edwardi quia cum diu essent dimisse ipsas fecit reparare et de legibus regis Willelmi" (fol. 1). The vellum manuscript appears to have included, or to have been intended to include, a text of Magna Carta 1225 but probably not a text of Glanvill. The scribe responsible for both the Consiliatio Cnuti and the Leges Edwardi wrote a note on the bottom half of the last folio (12v) that described Glanvill's compilation of a book of "multe leges" and claimed that much of this book had fallen into desuetude after being superseded by laws of John and Henry III, specifically Magna Carta and the Charter of the Forest, "which the king made . . in these words." What follows this note on the next line is the rubric "Magna Carta de libertatibus Anglie." The legal texts are in a script of the first quarter of the fourteenth century; the hand itself is similar to the hand that wrote fols. 74–86 in MS Ad.

Pl London, British Library, Additional MS 35179, fols. 4–8v. 230–32 × 168–72 mm. Written space: 177–82 × ca. 90 mm. Double cols. 44 lines/ page. Collation: (foliation begins on the third flyleaf) 12 (fols. 2–3) 2–312 44 wants 4, probably blank (fols. 28–30) 5–812 96 wants 1–3, probably blank (fols. 79–81) 106 wants 3, probably blank (fols. 82–86), 116 wants 1,2, probably blank (fols. 87–90). This manuscript holds a copy of version 2 of the Leges Edwardi, the Libertas ciuium Londoniensis, the Instituta Cnuti (fols. 9–16), Henry of Huntingdon's description of Cnut (fol. 16), the Articuli Willelmi (fols. 16v–17), the Norman genealogy from Rollo to John (fols. 17–18), a treatise on the legal and ecclesiastical divisions of England, a veterinary treatise, a register of writs (fols. 31–36), pleas from 36–40 Henry III (fols. 36–39), an alpha text of Glanvill (fols. 39v–73), a second register of writs (fols. 74–83v), a manorial text on stewards' accounts, the Provisions of Merton (1236) (fols. 84v–85), a text relating to an eyre of 1248 (fols. 85v–86), and a collection of thirteenth-century assizes and nonlegal treatises in the final quire (fols. 87–90). Many of the same texts are in S, but not in the same order or of the same version (e.g., Glanvill), although the sequence from the Instituta Cnuti to the genealogy is preserved. In Pl, all of the texts in fols. 4–86 (excluding occasional additions to blank folios) were the work of one scribe. The oldest text he copied dates from 1254 (39 Henry III); a different hand has added (on fol. 83v) to the manuscript's register of writs a note dated 1 Edward I (1272/73). It is likely, though not certain, that the work of the original scribe was completed between those dates. The script is consistent with such a date. A number of texts in the final quire relate to Chester in the second half of the thirteenth century, suggesting that this manuscript, if not made in Chester, was located there soon after its creation.

Rl Oxford, Bodleian Library, Rawlinson C. 641, fols. 3–10. 185–88 × 135–40 mm. Written space: 142–45 × 90–95 mm. 27–31 long lines, except double cols. on 13v–15. Ruling is by pencil and changes with fol. 19 (Assize of Clarendon) from light guides to heavy double lines marking the left and right edges of the writing space. Collation: 18 28 32 48 (or a combined quires 3 and 4 for a quire of ten) a singleton (fol. 27), attached either to the fourth or fifth quires 52 68 78 According to its fifteenth-century table of contents, the manuscript began not with philosophical treatises (fols. 1–3), but with church canons, followed by the Instituta Cnuti (now fols. 30–43), Henry I's coronation charter (Hn com, fols. 43–44), the Articuli of William I (fols. 44–45), a work entitled Tractatus Henrici regis secundi de legibus et consuetudinibus Anglie compositus per Ranulphum Glanvillano, the Assize of Clarendon (fol. 19), and Magna Carta 1215 (fols. 21v–29). The synodal decrees have disappeared, but the rest remain, albeit with what were the first two quires placed at the end of the manuscript. While the fifteenth-century table of contents identifies the Tractatus Henrici regis secundi as a copy of Glanvill, the original order of texts shows that it was referring to the Leges Edwardi and not to Glanvill. N. R. Ker saw first that the original order of the folios was 30–45, 1–29, and that the text of Magna Carta, which would in the original order have come last, was added later. Other than the early thirteenth-century hand responsible for Magna Carta, the other seven hands are late twelfth-century and bridge all the gaps between quires but one (between quires 5 and 6, originally the manuscript's seventh and first quires). The version 2 text of the Leges Edwardi was written by two hands (3–7r, 7v–10), the latter of which is a very pointed charter hand, earlier rather than later in the last quarter of the twelfth century. Ker rejected Liebermann's claim that the hands were French and instead concluded that at least the prickly script on fols. 7v–10 and the links of the contents to Textus Roffensis "suggest Kentish provenance." Although an origin in Kent for so late a manuscript may not be supportable on paleographical grounds, its Old English variants do link the scribe, or his exemplar, to Kent. A further clue to its origin may be provided on fol. 5v, where, in an exceptional departure from 31 lines of writing per folio, the scribe has given the words "ad thesaurum regis" on line 9 ascenders that take up line 8. This raises the possibility that the manuscript or its scribe was connected to the treasury, at Winchester or in the Tower of London, or to the lower exchequer, the treasury's receipt office at the exchequer, which followed the exchequer itself and which, from the reign of Henry II, usually met at Westminster.

S London, British Library, Harley 746, fols. 49–54v. 255 × 157–62 mm. Written space: 189–93 × 103–9 mm. Double cols. 42 lines/page. Collation: 1 three, a bifolium and a singleton (fols. 1–3) 2–612 712 wants 12, probably blank (fols. 63–74) 8–912 1014 wants 1, 9–11, 13 (fols. 99–107). This collection of legal texts is related to Pl, including many of the same texts. Here they follow a different order, beginning with a beta (rather than Pl's alpha) text of Glanvill (fols. 4–49), version 2 of the Leges Edwardi, the Libertas ciuium Londoniensis (fol. 55), the only surviving Latin version of the Leis Willelme (fols. 55v–59), Henry I's coronation charter (fol. 59), Magna Carta 1215 (fol. 59v–64), an emended version of the Charter of the Forest 1217 (fols. 64–65v), Magna Carta 1225 (fols. 65v–68), a writ of Henry III concerning hundreds (fols. 68v–69), the Provisions of Merton (fol. 69), a register of writs (fols. 70–74), four pieces on the ecclesiastical, legal, and political divisions of England (fols. 75–76), a sequence of texts — the Instituta Cnuti (fols. 77–84), Henry of Huntingdon on Cnut (HA 6.17, pp. 366–69), the Articuli of William I (fol. 84), and the Norman genealogy (fols. 85–86) — in the same order as in Pl, several statutes (fols. 87–99), a French treatise on royal writs (fols. 100–103), some songs, and a number of miscellaneous documents. All of the texts were copied by one or more scribes at the same time; they all share the same layout and were rubricated by the same hand. The terminus post quem is 1275, the date of the Statute of Westminster (I) (fols. 91–98v). The hands date from the late thirteenth or early fourteenth century.

Relationship of the witnesses

>The relationship of the witnesses to Version 1 has been described in the introduction to the edition of that text on the Early English Laws website. Version 1 was revised and enlarged to produce the thirty-nine-chapter text, version 2. However, one manuscript, Cl (s.xiv1), stands in a relationship with both versions. Its pattern of readings may be a sign that a version 2 text was contaminated by readings from a version 1 text, or vice versa. It is also possible that its readings are evidence for an intermediate version of the treatise. Clues are few and problems of interpretation many. The text in Cl ends at c. 37.1 and shows textual affinities with both versions. Cl, however, is late and its text was carelessly copied. The length of the text may not be significant; although Cl's scribe clearly intended the Leges to end at c. 37.1, it cannot be known whether or not he was following his exemplar, nor whether or not he was ultimately reflecting any of the manuscripts standing between his text and the author's. The text's tie to either version is at best murky, owing in large part to the many unique readings in Cl. Nevertheless, the simplest explanation for all of these observations is that Cl is a version 2 text contaminated by readings from version 1; as such, when it does not agree with alpha it can at times assist in the reconstruction of version 2.

The manuscripts of version 2 probably descend from a common archetype, beta (β), and fall into two families. One, consisting of Ar (1564), Ha (s.xvi/xvii), Pl (s.xiii2), and S (s.xiii/xiv), descends from a common hyparchetype, zeta (ζ). These four texts share many errors in comparison to all other texts. The following is a partial list of these errors:

Prol. 1

coram eo om. zeta

c. 5.2

si alpha Cl epsilon; quod zeta

c. 5.3

ad ecclesiam om. zeta

c. 7.2

uitulis om. zeta

c. 9

Die (De Cl) illo quo iudicium fiet alpha Cl Rl; De illo de quo aque uel ferri iudicium fiet Hk; Quando de aliquo iudicium fiet diuinum Hr; om. zeta

c. 11a

suo om. zeta reddi (om. Ad) alpha Cl epsilon; reddendum zeta

c. 11.1

habebant in orationibus (oratione Hk) sancte ecclesie alpha Cl epsilon; in orationibus sancte ecclesie habebat zeta

c. 20.1a

terminus alpha Cl epsilon; ad minus zeta spacium add. zeta

c. 20.6

eum om. zeta

c. 21

milites dittograph in zeta

c. 22.5

etiam add. zeta

c. 34.2b

eorum add. zeta

c. 34.2e

Suuue Cl; Emme Hk; etiam Hr; Emma Rl; Iunia zeta

c. 34.3

dimissam Cl epsilon; omissam zeta

c. 35.1b

longo Cl epsilon; multo zeta

c. 35.1c

uocant add. zeta

c. 36.1

ei om. zeta

c. 38.1

ipsi si epsilon; si ipsi zeta

Further support for the reconstruction of zeta comes from manuscript contents. The Leges Edwardi in S and Pl is accompanied by many of the same texts, though they are not arranged in the same order. The other two manuscripts of this family, Ar and Ha, both share three of their legal texts, the Instituta Cnuti, Leges Edwardi, and Libertas ciuium, with S and Pl.

Of the four witnesses to zeta, Ar, Ha, and Pl are closely related. They share a number of common errors and therefore descend from a common hyparchetype, iota (ι). For example:

c. 2.9b

Et sic iuste gladius gladium iuuabit om. iota

c. 12.8

ad disturbationem eorum preparantur, ipsa opera om. iota

c. 22.2

emendi om. iota

c. 23

aliquem om. iota

c. 23.2

ad om. iota

c. 30.2

non om. iota aliquis om. iota

c. 34

Pictorum alpha Cl epsilon S; Pictanorum iota

c. 34.2d

putatus om. iota

c. 35.1

accepit om. iota

c. 38.1a

eius om. iota

c. 38.3a

scit epsilon; sciat S; faciat iota

That Ar, Ha, and Pl are all independent from one another is proved by omission by homoeoteleuton in each where the other two texts preserve the passage. The witnesses, therefore, to zeta are S and iota.

The situation of zeta is complicated, however, by its relationship with version 3. In addition to taking most of its rubrics from the text of version 3 found in Henry of Huntingdon's Historia, zeta also shows evidence of collation against version 3.

Prol. 1

nil addentes add. zeta version 3

Prol. 1

commutantes Hk Hr Cl; committentes Rl; mutantes zeta version 3

c. 15.1

autem alpha Cl epsilon; uero zeta version 3

c. 24.1

prefectos alpha Cl epsilon; prepositos zeta version 3 (abl. pl.)

c. 31a

uocabantur om. zeta version 3

c. 35.1b

eius Cl epsilon; eorum zeta version 3

Since version 3 is descended not from zeta but from the other beta family, epsilon (as explained below), it is more likely that the influence runs from version 3 to zeta than from zeta to version 3. This contamination, however, means that zeta can only be used confidently to reconstruct version 2 when it disagrees with the other beta witnesses but does not agree with version 3.

The other family of beta consists of Hk (s.xii 3/4), Hr (s.xiv in.), and Rl (s.xii 4/4). All three manuscripts descend from a hyparchetype, epsilon (ε), as is shown by their common errors.

c. 8.3

populis alpha Cl zeta; populo epsilon

c. 8.3a

multum curiosi erant alpha Cl zeta; erant multum curiosi epsilon

c. 12.6

eam om. epsilon

c. 14.1

si argentum, dimidium regis est alpha Cl zeta; dimidium argenti epsilon

c. 19

Et add. epsilon

c. 20.6

aut add. epsilon

c. 27

primum add. epsilon

c. 28.1

ordinationes alpha Cl zeta; concordationes epsilon

c. 31.1

uocabant alpha Cl; uocant zeta; om. epsilon

c. 33

emendationem forisfacture, ut alpha Cl zeta; emendatione forisfacture ubi epsilon

c. 34.2e

matre Cl zeta; matris epsilon

c. 39.1

suis om. epsilon

c. 7.4

de apibus add. eta

c. 11a

eius om. eta

c. 15.5

marcas suas alpha Cl zeta Rl; suas marcas eta

c. 16.2

Itaque alpha Cl zeta Rl; Ita quod eta

c. 20.1a

dierum om. eta

c. 26.1

seruicium alpha Cl zeta Rl; seruicia eta

c. 31.2

Et om. eta

c. 34.2a

regnum suscepit Cl zeta Rl; suscepit regnum eta

c. 34.2b

ut Cl zeta Rl; quod eta hac add. eta

c. 34.2e

sorore Cl zeta Rl; sororis eta

c. 35

ista terra Cl zeta Rl; terra ista eta

c. 36.5

forisfacturas tres Cl zeta Rl; tres forisfacturas eta

c. 37.1

sepe se Cl zeta Rl; se sepe eta

c. 38.2

emptor ipse zeta Rl; ipse emptor eta

c. 39

emerent zeta Rl; emerentur eta

Texts in two of these manuscripts, Hk and Hr, descend from a common hyparchetype, eta (η).

All three individual witnesses to epsilon show signs of interpolation, correction, and revision. Hk on a number of levels shows adjustments being made to the Leges. Those chapters taken from the Consiliatio Cnuti and joined without break to the end of c. 39 have also been adjusted in style, length, and content. This combined witness has been, in fact, edited separately for Early English Laws, where it is called the Holkham Leges Edwardi (ECf2 Hk). Hr appears to have created the majority of its rubrics. Rl occasionally misunderstood the text or exemplar (e.g., c. 33) or revised it (c. 22.4). Thus, while two of epsilon's manuscript witnesses are the earliest extant for version 2, each must be used with caution when its reading is unique.

Using epsilon and zeta to reconstruct beta is further complicated by the relationship each family has with version 3 of the Leges. At various points, each hyparchetype, or occasionally only an individual witness to a hyparchetype, shares readings with version 3 against the other hyparchetype of version 2 as well as against version 1. The links between zeta and version 3, examined above, suggested that zeta was contaminated by readings from a text of version 3 close to that found in Henry of Huntingdon's Historia. The evidence for a relationship between epsilon and version 3 is somewhat different and can be laid out as follows:

c. 2

pacem (object of habere in c. 1.1) alpha Cl zeta Rl; pax (sit add. Hr) eta version 3

cc. 7.4, 8.1

de apibus add. Rl as rubric at c. 7.4 and et apibus in text at c. 8.1; et apibus add. in text at c. 7.4 eta; de apibus in rubric and text at c. 8.1 version 3

c. 8.3

populis alpha Cl zeta; populo epsilon version 3

c. 10.1

et ultra alpha Cl zeta Rl; ut ultra eta version 3

c. 11a

eius om. eta but in Rl version 3

c. 12.1

sue om. Rl Hr (ergo epsilon) version 3

c. 12.3

suorum were alpha Cl iota Rl; suum were (were suum version 3) eta S version 3

c. 12.6

fer eam alpha zeta; fer eas Cl; feras eta; ferre Rl; fer version 3

c. 12.7

quod alpha Cl Hr zeta; quam Rl Hk (ergo epsilon) version 3

c. 14.1

si argentum, dimidium regis est alpha Cl zeta; dimidium argenti epsilon; medietas argenti version 3

c. 15

infra alpha Cl Rl zeta; post Hk; ipsos Hr; infra ipsos version 3

c. 18.3a

scienter eos alpha Cl zeta Rl; eos scienter eta; eos gratis version 3 (some manuscripts)

c. 20.1a

dierum alpha Cl zeta Rl; om. eta version 3

c. 24

antequam (et Cl) introducat illud in domum suam uel alterius et dicat (dixerit Cl Hr zeta) se inuenisse alpha Cl zeta eta; et dixerit se inuenisse antequam (priusquam version 3) introducat illud (om. version 3) in domum suam uel (etiam add. version 3) alterius Rl version 3

c. 27

viiitodiebus quibus coronatus alpha Cl zeta; viii dierum (diebus Hk) quibus primum coronatus epsilon version 3

c. 30

Notinghamsyre eta S version 3; om. alpha Cl iota Rl

c. 31.1

uocabant (uocant zeta) alpha Cl zeta; om. epsilon version 3

c. 32.1

ue Latine alpha Cl zeta Hk; ue Latino Hr Rl version 3

c. 32.2

etiam add. Hk Rl version 3 (some manuscripts)

c. 33

ut alpha Cl zeta; ubi epsilon version 3

c. 34.2e

sorore Cl zeta Rl; sororis eta version 3

c. 35.1b

Edgarum (autem add. Cl; uero add. Hr) filium eius (eorum zeta), secum retinuit Cl eta zeta; Edgarum uero filium eius rex


E secum retinuit Rl; Rex uero Eadwardus Eadgarum, filium eorum, secum retinuit version 3


ibi Cl zeta; sibi eta; interfectori Rl version 3

For all of these readings to be construed as evidence that epsilon was contaminated by version 3 would require the belief that in almost every case the borrowing was slight and not substantive, was limited to odd words here and there, and included no rubrics or major additions (e.g., version 3's c. 38). This scenario is very improbable. Given that the readings common to epsilon and version 3 are spread throughout the text of the Leges and are of all levels of significance, while those common to zeta and version 3 are few, the simplest explanation is that version 3 is descended from a text of version 2 related to epsilon. But to which specific text? In a number of places, version 3 shares unique readings with either Rl or eta; some of these cannot have been in epsilon. Version 3, then, is not derived directly from epsilon, but must have drawn from both branches of epsilon and so preserves readings otherwise unique to Rl or to eta.

Second, there is the issue, raised above, of the rubrics in Rl, Hr, Cl, and zeta. Version 1 had no chapter list at the beginning of the treatise or rubrics within the text. The earliest text of version 2, Hk, likewise has no chapter list or rubrics, except apparently at c. 36. Since version 1 lacks them, as does the earliest witness to version 2, it seems reasonable to suppose that version 2 originally lacked a chapter list or rubrics. Where then did the rubrics in the manuscripts other than alpha and Hk come from?

The explanation is complicated and tentative because, as is commonly the case, the rubrics charted their own textual courses. No one set of rubrics is derived principally from any other single set, except in one case, and that must be the starting point. Zeta's rubrics are the same as or very close to those in the version 3 text inserted into Henry of Huntingdon's Historia. As already pointed out, the text of zeta shows evidence of contamination by readings from a text of version 3 that was related to a copy of the Historia's text of version 3. Zeta probably took its rubrics from the same source. This transfer of readings and rubrics from version 3 to zeta likely occured after 1161, Liebermann's terminus post quem for the rubrics. However, the earliest witness to zeta, Pl, is from the third quarter of the thirteenth century and so the question of when zeta borrowed these rubrics cannot be resolved.

Rl and Hr, the two remaining epsilon texts, also have rubrics, though they differ significantly in many instances. Only a small portion of the chapters of Rl are rubricated, most of which rubrics it shares with other witnesses, while the majority of the chapters of Hr are rubricated, and many of these rubrics are unique. Yet Rl's and Hr's rubrics are related in some way. Although it is possible that epsilon itself (labeled ε1 in Figure 2) had some rubrics, perhaps a chapter list that the scribe of Hk chose to omit, it seems more likely that epsilon, after eta was copied, received a few rubrics (ε2), and that these rubrics appear in Rl. Later, eta (after the copying of Hk) or Hr was corrected against epsilon (ε2) by the addition of epsilon's rubrics. A text descended from ε2 became one of the bases for version 3. All witnesses to version 3 have rubrics that can be divided roughly into two types. The first type appears in the earliest witness, Cb (see Figure 1), and is simple and related to those in Rl and Hr. This rubrication was subsequently passed on to version 4 from a text related to Cb. The second type, found in the Historia Anglorum's text, influenced zeta. So far so good.

From where, then, did Cl acquire its rubrics? Cl's rubrics are related to both zeta and version 3 in the Historia Anglorum. The majority are closer to zeta (cc. 1, 5, 17, 19, 20, 23–27, and 37), but others are closer to the Historia's wording (cc. 2, 7, and 14). If this were all, then it might appear that Cl had taken its rubrics from some third-version text between that in the Historia and zeta, or more likely had compiled its rubrics using zeta and the Historia. However, the water is muddied by evidence of a relationship between Cl and version 4. One of Cl's rubrics (c. 9.1: "De baronibus qui curiam habent") is shared only with version 4 ("De baronibus qui suam habent curiam et consuetudines").66 While neither the Historia Anglorum's version 3 text nor zeta has a rubric for this chapter, nevertheless a number of other rubrics (cc. 12, 15, 20, and 27) are slightly closer to version 4's, which makes a borrowing from version 4 plausible. The simplest explanation is that the rubrics in Cl represent the work of a redactor between the early thirteenth and early fourteenth centuries who selected rubrics from available copies of the Leges, a pool consisting of witnesses to zeta, one branch of version 3, and version 4. Cl, which itself may represent the work of this redactor, was at Westminster, and so was close enough to London collections (the Guildhall's, for example) to borrow without difficulty from manuscript witnesses to these different versions.

Given contamination of one hyparchetype of beta by version 3, the reconstruction of beta is not as straightforward as was the reconstruction of alpha. The earliest text of version 2 reconstructed from these witnesses is based, then, on two criteria: (1) agreement of zeta and epsilon will usually be taken to represent the beta archetype, and (2) when zeta and epsilon disagree (e.g., 11.1, 14.1, 15.7) but offer two possible readings, alpha, Cl, and version 3 will be employed to establish which witnesses to accept. Here version 3 is used only to identify those passages that zeta may have borrowed, and this identification allows zeta's readings to be eliminated from the reconstructed text. When it disagrees with alpha, Cl's agreement with either epsilon or zeta will be used cautiously to reconstruct beta.

While beta has been emended in only one place (c. 25), nevertheless the text presented in this edition differs in many instances from Liebermann's text and is arguably somewhat closer to the author's own words.

If the author titled his work, this is probably lost to us. All of version I's titles attribute the laws to "the time of King William the Great"; the fact that this title is found in all witnesses to the earliest version creates a prima facie case for accepting it as the author's title. However, its resemblance to the title of Glanvill raises the concern that, as all version 1 texts travel behind texts of Glanvill, it may be the case that a later editor has modeled a title for version 1 on the title of Glanvill. It would then be no guide to the author's intentions. Zeta and Cl both title the laws as King Edward's, but both take their rubrics, and probably also their titles, from copies of later versions of the Leges Edwardi. While the latest witness to epsilon (Hr) says that these laws are the "Laws of King William," the other two epsilon manuscripts, Hk and Rl, which also constitute the oldest witnesses to version 2, have no title whatsoever. Given the confusion and contamination of witnesses, it appears that the original title of version 2, if there ever was one, is unrecoverable. For this reason, I have chosen to retain the now conventional title, Leges Edwardi Confessoris, which was first used only in the seventeenth century.

Editorial Procedures

The numbering of chapters in Liebermann's edition has been maintained, except in the case of his chapter 8, the first clause of which enjoyed no textual support in versions 1, Cl, and zeta and appears only inconsistently in epsilon. This chapter seems to have been added by epsilon to the text of the treatise. Consequently, Liebermann's chapter 8 has been re-numbered. The apparatus is organized by archetype, hyparchetype, and then alphabetically by manuscript sigla.

The orthography of the earliest witnesses, usually Hk, but also Rl, has in most cases been used. The apparatus includes, for the most part, neither purely orthographic variations in Latin nor trivial errors that do not illustrate the relationships of the manuscripts. Agreement by all of the witnesses to an archetype or hyparchetype is represented in the apparatus by the appropriate Greek letter; one witness is used to supply the orthography: Ad for alpha, Do for theta, S for zeta, Pl for iota, and Hk for eta and epsilon. All Old English words in the text, whether Latinized or not, represent forms found in Hk. The text uses the Old English of this single witness not simply to show the instability of twelfth-century English orthography (see, e.g., c. 21.1, infangeþef, and c. 22.4, infangetheof), but also because it is impossible to choose the best reading when no clear standard can be employed to discriminate. Furthermore, the orthography of the oldest manuscript is the closest representation of the practices of the author's day. It is likely, for example, that Hk's greue for Old English gerefa at c. 32 represents a northern dialect, which is the form that would be expected of an author working in the northern Danelaw in the twelfth century. All variant readings of the Old English, including the merely orthographic, and of proper nouns, have been included in the apparatus because they may preserve dialectal clues to the origins of the manuscripts. Scribal use of the special letters of the Old English alphabet (e.g., æ, ð, and þ) is preserved in text and apparatus. Numerals within the text appear as they do in Hk or in the earliest manuscript from which the numeral has been adopted. The text has been capitalized and punctuated according to modern English conventions. All Latin abbreviations are silently expanded. In the apparatus, abbreviations from the witnesses are silently expanded when the full form is clear.

The Translation

The Latin of version 2 has been rendered into an English prose style that makes legal sense in twelfth-century terms. The Leges Edwardi was meant to convey some knowledge about the law as it stood after the Conquest. Judging by its popularity in the twelfth century, it was thought to do a pretty good job. The belief in the sensibleness of the treatise has guided the translation. When the translation is not immediately accessible, clarification has been provided in the notes. In addition, Old English sayings and more unusual terms have been left intact in the translation. In the originals, of course, these passages stood out as distinct to contemporaries. Most of them were translated into Latin by the author immediately after they appeared in the text. Translating the Old English into modern English, then, would introduce tautologies not present in the treatise's original text and would fail to convey its bilingual feel. In all cases, the literal meaning of the Old English is given in notes to the translation.

Previous Editions

The editio princeps was William Lambarde's Archaionomia (1568), which was followed in the next century by Henry Spelman's initially unpublished Codex legum veterum statutorum regni Angliae (1626), an edition later incorporated into David Wilkins's Leges Anglo-Saxonicae (1721). Other than the translation-paraphrases of James Tyrrell (1700) and W. Guthrie (1744), very little work was done after Lambarde and Spelman on the text of the Leges Edwardi, which may explain why Thorpe's edition (1840) was overly faithful to Lambarde's text. Reinhold Schmid's Gesetze der Angelsachsen (1832; 2d ed. 1858) —based on some, but not all, manuscripts of the text—was followed soon after by the work of Felix Liebermann, whose own Gesetze der Angelsachsen appeared between 1903, with the publication of the texts and apparatus, and 1916, with the publication of what for the most part was a textual introduction. William Stubbs had edited from two manuscripts the copy of the version 3 text in Roger of Howden's Chronica (1868–71). Since Liebermann, there has been no work on the text of the Leges Edwardi or on its manuscript witnesses, except for Ker's study (1954) of the Guildhall manuscripts, which hold copies of version 4.