Ancrene Wisse

Ancrene Wisse, a Middle English rule or guide for female recluses, was composed in the West Midlands in the early thirteenth century. There is still no scholarly consensus on its exact date, localization, authorship, or audience, but recent research has helped to narrow down the possibilities, and to suggest at least what its broader cultural and institutional context might have been.

It is a carefully-structured work, divided into the Preface edited here and eight parts (called by the author distinctiones). The first, which deals with the anchoresses' devotional routine, and the eighth, which covers their remaining outward observances, constitute the Outer Rule. The remaining distinctiones, which deal with the custody of the senses, the nature of the solitary life, the Seven Deadly Sins and their remedies, confession, penance, and the love of God, constitute the Inner Rule, and offer general spiritual guidance as well as specific advice on the anchoritic life. The work draws on a wide variety of sources: the Church Fathers, particularly Augustine of Hippo and Gregory the Great; later monastic writers, particularly Bernard of Clairvaux and Aelred of Rievaulx (whose Latin rule for anchoresses, De Institutione inclusarum, c. 1160, is an important source); and the pastoral manuals and preaching aids developed in the Paris schools during the late twelfth and early thirteenth centuries (see Millett 1999).

Ancrene Wisse is clearly the work of a highly-educated author, but it does not assume an equally highly-educated audience; although both French and Latin versions survive, it seems to have been composed originally in Middle English, and most of the Latin it contains, apart from the prayers and hymns which would have formed part of the anchoresses' daily routine, is translated or glossed. It appears to have been composed originally for three well-born (gentile) sisters who gave up the world in the flower of their youth to become anchoresses (Nero A. xiv, f. 50, ed. Day 1952, 80/23-37), and a probably authorial addition to the later form of the work in Corpus 402 addresses a larger and more scattered group of anchoresses, twenty now or more (4. 69r/14); but the internal evidence of the work suggests that from the beginning the author had other possible audiences in mind, not only anchoritic but secular (in the section on confession, Part 5, he tells the anchoresses that most of its material is relevant to everyone, addressing their specific concerns only briefly at the end; and much of the material on the sins in Part 4 is similarly generalized). The accessibility of his work, its vivid, colloquial, and elegant style, and the relevance and interest of its content to a variety of audiences, meant that from a very early stage it began to be revised, adapted, and translated for other kinds of audience, both religious and lay. It survives in seventeen medieval manuscripts and manuscript fragments, and borrowings from it also appear in a number of late-medieval sermons (see Spencer 1993) and devotional works (Millett 1996a; see also Diekstra 1998).

It now seems increasingly likely that the origins of Ancrene Wisse are Dominican; both its structure and some of its content appear to have been influenced by the early Dominican constitutions, and there are some textual indications that its author itself was a friar (see Millett 1992, Millett 2000). If so, it must have been written after 1216, when the Dominican order was founded. The revised form of the work in Corpus 402, which refers approvingly to regular visits to the anchoresses by both Franciscan and Dominican friars, must postdate the arrival of both orders in England (by 1224), and probably their settlement in the West Midlands (by the early 1230s); but since the original version does not mention such visits, and indeed warns the anchoresses not to trust clerical visitors of any kind, it seems likely that it was composed earlier than the 1230s.

This places Ancrene Wisse in the period following the Fourth Lateran Council of 1215, which initiated a major programme of pastoral reform and religious education of the laity across Western Europe. From an early stage, the newly-founded orders of friars were recommended by the Papacy to support local bishops in the implementation of this programme by preaching and hearing confessions. In this period, the West Midlands dioceses of Worcester, Hereford, and Coventry and Lichfield were occupied by a series of reforming bishops (one of whom, Alexander Stavensby, bishop of Coventry and Lichfield 1224-38, had taught Dominic himself at Toulouse, and founded Dominican priories in Shrewsbury and Chester in the 1230s); it is possible that Ancrene Wisse (which includes a substantial proportion of general material on confession), should be seen in this broader context of pastoral reform, as part of a drive to provide a literature of pastoral instruction in the vernacular (see Millett 2003b).

The Preface to Ancrene Wisse

The Preface to Ancrene Wisse is a more complex piece of work than it appears to be at first sight, simultaneously drawing on and reacting against an earlier tradition of monastic legislation.

Its overall model (see Dobson 1976, Ch. 1), supplying its framework of argument and some of its content, is the prologue first added to the statutes of the Premonstratensian canons when they were revised c. 1174 (ed. Lefèvre 1978). The Premonstratensians were one of the independent congregations of Augustinian canons, following the Rule of St. Augustine but supplementing its relatively brief prescriptions with their own more detailed regulations; like the Cistercians, the reformed order of Benedictine monks on whose legislation they initially drew, they were noted for the strict discipline and austerity of their way of life. When St Dominic—who had been an Augustinian canon himself—founded his new order of friars in 1216, and found himself obliged by Canon 13 of the Fourth Lateran Council to accept the rule and regulations of an existing order, he adopted the Premonstratensian regulations (though in a modified form, to fit the very different mission of the Dominicans); it is likely that this is the route by which their prologue, which remained largely unaltered in the early Dominican constitutions (ed. Thomas 1965), reached the author of Ancrene Wisse. The influence of the Premonstratensian/Dominican legislative tradition is also reflected in some of the content of Parts 1 and 8, the Outer Rule, and in the structural subdivision of Ancrene Wisse into distinctiones and chapters, a feature whose explicit mention here goes back to the original Premonstratensian prologue (which drew attention to a formal innovation, since twelfth-century monastic regulations were normally subdivided only by chapter); see Millett 2000, Millett 2003a. The Preface goes beyond this model, however, to a critical reassessment of what constitutes the religious life, based on a discussion of some key terms associated with it, religio (ME religiun), ordo (ME ordre), and regula (ME riwle). The author tells us that the anchoresses have repeatedly asked him for a rule, and that they complain that people keep asking them what religious order they belong to. Taken together, these points suggest that they felt a certain unease about their institutional status; and they had some reason, in the early thirteenth century, for feeling marginalized. Lay-anchorites (that is, anchorites who, like these anchoresses, had entered the solitary life from the world rather than from a religious community) counted technically as religious (religiosi): they had taken formal vows of chastity, obedience, and stability of abode, and led an enclosed life under the supervision of the local bishop. However, by the early thirteenth century, the term religio had come to be used in an increasingly narrow sense; it tended to be applied primarily (as it is in Canon 13 of the Fourth Lateran Council) to the way of life of those belonging to a religious order (ordo or, sometimes, religio), renouncing personal property, living in an enclosed community, wearing a distinctive habit, and following a written rule (regula). We do not know whether the anchoresses addressed in Ancrene Wisse had turned to the solitary life as their first choice, or whether they had chosen it as an alternative to communal religious life because of the excess of demand over supply in this period for places in women's religious houses; in either case, a central theme of the Preface is the value of forms of religious life other than the traditional monastic model (see Millett 2002). The terminology of religious life is reclaimed for extra-monastic forms of vocation, and for the inner, rather than the outer, life: the religious life (religio / ordo) is redefined, using an unusual allegorization of James 1: 27, in terms of pastoral care and the solitary life rather than monasticism, and the rule (regula) that the author offers does not have the legislative status of a traditional monastic rule, but is made up instead of a combination of optional external recommendations (the Outer Rule) and general moral guidance, which the author emphasizes is of divine rather than human origin (the Inner Rule). The Preface to Ancrene Wisse, like the work as a whole, reflects an ambivalent attitude to the monastic tradition on which it draws. Although the anchoresses are offered a Rule which owes much to an existing tradition of monastic legislation, the Preface also represents a reaction against traditional monasticism, looking back beyond it to the solitaries of the early Church, and outside it to other kinds of religious life, both in the world and in solitude. Similar attitudes can be found in other writers from the late eleventh to the early thirteenth centuries, reflecting the development of new approaches to the religious life during what has been called the Medieval Reformation (see Bolton 1983). Although Giles Constable saw the essential feature of this movement as the application of the monastic life to all people, and the interiorization of monastic values and spirituality (Constable 1996, p. 7), the democratization of monastic ideals seems, in some cases at least, to have been less a process of peaceful dissemination than of active reaction against the identification of religio with traditional monasticism. A similar reclamation of the terminology of religious life, in this case for all devout Christians, can be found in the Historia Occidentalis composed in the early 1220s by the reforming preacher James of Vitry: We do not consider regulars [regulares, i.e. those living according to a rule] only those who renounce the world and go over to a religious life [religionem], but we can also describe as regulars all the faithful of Christ who serve the Lord under the rule [regula] of the Gospels and who live in an ordered way [ordinate: i.e. each following their own ordo or vocation] under the one highest and supreme Abbot (Ch. 34, ed. Hinnebusch 1972). James's concluding image is paralleled in the later addition to Ancrene Wisse Part 4 mentioned above, where the larger group of anchoresses is described as as if in a cloister over which Jesus is the high Prior (4. 69v); the author of Ancrene Wisse offers his readers not a real but a virtual cloister, a religio which is a matter of inner disposition rather than outward observance.

Principles of editing

The existing EETS editions of Ancrene Wisse provide a detailed record of the readings of its individual MSS, but are difficult for non-specialists to use. The first aim of this edition is to provide a corrected and fully-annotated working text, based on a good manuscript, for general use; the second is to establish a context for the edited text within the broader historical development of the work.

Ian Doyle noted the exceptionally dynamic character of the textual transmission of Ancrene Wisse (Doyle 1954, i. 234). Between the early thirteenth century and the end of the Middle Ages, the work was repeatedly revised and adapted for different audiences and for different functions; it was also translated once into Latin and twice into French. Some major early revisions were probably authorial. Dobson made a strong case for identifying an early annotator of Cleopatra (Scribe B, referred to in this edition as C2), who both corrected and revised its carelessly-copied text, with the author himself (Dobson 1972, pp. xciii-cxl). The more extensively revised text in Corpus 402 (A), adapted for a larger group of anchoresses, incorporates both some of the C2 revisions (in some cases further revised) and additional revisions, some shared with other manuscripts, a few unique to A; some at least of these additional revisions are likely to be authorial. However, it is clear that from its earliest stages the work was seen by other users as open to alteration and adaptation, and we cannot assume that all the additions in A were the responsibility of the original author. The internal evidence of a revision (e.g. marked similarities of style and approach to the original, or the use of the first person) may sometimes suggest authorial origin; but these clues are not always present, particularly in shorter additions, and in a social and institutional context where works of popular instruction might be composed, annotated, revised, and sometimes copied, by clerics with relatively similar educational backgrounds, it is not always easy (or even useful) to distinguish between authorial and non-authorial contributions.

Most recent editors of Ancrene Wisse have based their editions on A, and it has generally become the standard reference text. There are good reasons for this. Corpus 402 is a carefully-copied, consistently-spelt manuscript, which offers a particularly good text with relatively few errors; it also brings together most (though not all) of the early revisions to the work.

Dobson argued further for its textual authority as a close copy of the author's own final and definitive revision of his work (Dobson 1962, p. 163). However, there are problems with this view. Although much of the added material in A is likely to be authorial, it is less clear that its text as a whole is the product of direct authorial supervision, or that it constitutes (as has often been assumed) a distinct version of the work. The only part of the work which has been thoroughly overhauled is Part 8; elsewhere, there is less evidence of systematic revision. The significant gap left in the argument by the cutting of an address to the original three anchoresses in Part 4 has not been repaired, and a briefer reference to them in Part 2 remains undeleted; the attribution of more offspring to Envy and Wrath in Part 4 has not been accompanied by a corresponding adjustment to their numbers in the surrounding text; the A text of the long Quomodo obscuratum addition in Part 2 is markedly inferior to those in F and V; and some originally marginal annotations have been entered at unsuitable points in the text. The A text seems to be less a coherent and harmonious recasting of the original (to borrow a phrase from Bédier) than the multi-layered product of an ongoing process of textual change; its revisions and deletions reflect a sequence of changing—sometimes even conflicting—intentions, not all of them necessarily authorial.

It can be argued, then, that A is not a definitive authorial version of Ancrene Wisse, or (because of the extensive revisions it incorporates) even a representative form of the work; there is also the problem that its relatively unsystematic incorporation of these revisions does not always make sense in isolation, leaving apparent inconsistencies, discontinuities, or even contradictions in its text. To be understood fully, it needs to be contextualized within the broader textual history of the work, and presented in a way which will make the multi-layered nature of its textual development clearer to the reader.

The editorial approach I have taken does not fit easily into the traditional classifications. This edition is not critical in the traditional sense, although it uses the tools of the critical editor (the analysis of manuscript relationships and of individual variants) to help establish the direction of textual change; it does not attempt to recreate a hypothetical author's original (either of the original work or of a later version), but to provide an edition of a specific historical form of the work, the text in Corpus 402. It is not a best-text edition, since it does not give privileged status to the A text, but uses it rather as a vantage-point from which the early textual development of the work can be retrospectively surveyed. It is not a parallel-text edition (a function already carried out in practice by the EETS series of diplomatic or semi-diplomatic editions of the work); although it prints alongside the text material from other early manuscripts not found in A, it offers a convenient overview of their most significant variations rather than providing full evidence. It could be described as an attempt to realize in practice the concept of social editing proposed by Jerome J. McGann (see McGann 1992), which sees the work not simply in terms of authorial intention, but as a communal product which may be significantly influenced by the broader social context of its production and transmission. My aim has been not, as with traditional textual criticism, to reverse the process of transmission, but to trace its historical development, using the evidence of the manuscript tradition as a whole to remove mechanical errors and misunderstandings from the A text, and to identify and account for other differences between A and the original form of the work.

The main influences on my editorial conventions have been editions of Latin works showing analogous patterns of continuing textual development: A. H. Thomas's edition of the early Dominican constitutions (Thomas 1965), and David d'Avray's edition of thirteenth-century model sermons on marriage (d'Avray 2001), both of which use a combination of typographical mark-up and extended citation of variants to indicate the relationship of their base text to other forms of the work.

Methods of editing

The edited text is based on Corpus 402 (A). Alterations by the main scribe have been silently incorporated in the text. Where the text has been emended, the MS reading is given in the Apparatus criticus. Square brackets are used to enclose emendations, and material omitted (for whatever reason) in A which has been been supplied from other MSS.

Material found in A but not in the original version of Ancrene Wisse is indicated by bold type. My use of bold type to indicate later additions does not distinguish between earlier and later stages of revision (which may sometimes coexist in a single passage), but is simply intended to alert the reader to the presence of changes from the original form of the work; fuller information is given in the Apparatus criticus and the Textual Commentary.

Rubrics in A are indicated by red type. Rubrics in MSS other than A are not recorded in the footnotes or Apparatus criticus unless there is a corresponding rubric in A, though they may be commented on in the Textual Commentary (for a discussion of possibly authorial rubrics in the early MSS and in F not found in A, see Dobson 1972, pp. lvi-lxii).

Abbreviations are silently expanded, and capitalization, punctuation, and paragraphing modernized. Word-division has been standardized, and in most cases modernized. The accent-like hairlines sometimes used by the A scribe above vowels (see Ker in Tolkien 1962, p. xiii) are not reproduced.

For the Middle English, I have followed the spelling of A, unless it is clearly erroneous by the scribe's normal standards. Words or short passages incorporated into the text from other MSS are normally silently adapted to A's spelling-conventions; where longer passages from other MSS are used to supply gaps in A, the spelling of the alternative base MS is retained. The spelling of the Latin in A is retained.

On the line-numbering and paragraph-numbering used in the text, see the Editor's Preface, under Problems. Also incorporated in the text are folio-references to the base MS, Corpus 402.


The footnotes are used to present separately, with a translation, material from the original version of Ancrene Wisse, and from the earlier stages of its textual development, either not found, or found in altered form, in A. Normally only the more substantial variants (i.e., more than a few words long) are included; shorter ones are recorded only in the apparatus criticus, unless they change the meaning of a larger unit of text.

Apparatus criticus

The Apparatus criticus is selective. Variations in spelling, and variations in word-form which do not affect the sense, are not normally cited, unless the variant is ambiguous or helps to cast light on other MS readings. Variations between þe and þet (both as demonstrative adjective and as relative pronoun) are not cited. As a rule, only those variants are cited which are shared by more than one MS (though single-MS variants may be cited where they are related to a preceding or following variant shared by more than one MS, or where they are of interest in themselves).

Variants are recorded more fully from CFGNTV than from the rewritten versions LPSR (for the MSS and their sigla, see the following section of the Introduction, About the manuscripts). The revisions made by C2 to C are recorded (except where they do no more than correct errors in C, or modify its spelling and punctuation); also the alterations to adjust the work to a different audience found in two of the early MSS, G and T. Modernizations of vocabulary and word-order, and minor verbal expansions which add nothing to the sense, are not normally recorded.

Lemmata are from the A text unless otherwise indicated; where a different base MS is being used, the MS sigil is given before the line-number. If the word in a lemma occurs more than once in the line, the particular instance is identified by number.

The form of a variant given in the Apparatus criticus is that of the first MS listed; the readings of the French and Latin versions are not given separately where it is clear that they are translating one particular variant in the English. Variants are normally cited in alphabetical order of sigil; but this order may be departed from to group variants according to their distance from the lemma, to put the English form of a variant before the French and Latin, and (occasionally) to avoid citing a variant in an anomalous form. It should not be assumed that all MSS not cited support the reading of the lemma; fuller variants, however, are cited where the reading in A is not shared by the other MSS, has only minority support, or is otherwise problematic. Where supporting evidence for the A reading is given, any MS not cited is either temporarily not running or casts no light on the variants in question (e.g. where the French and Latin versions could be translating either of two English synonyms).

Abbreviations are silently expanded, and capitalization, punctuation, word-division, and paragraphing are modernized; where any of these are relevant to the reading, however, they are discussed in the Textual Commentary.

Typographical conventions Bold type indicates material in the A text which was not in the original version. Red type indicates rubrication. In the text, square brackets indicate emendations, and material from other MSS supplying omissions; in the translation, they enclose words supplied for the sense (e.g. where a Latin quotation in the text has been cited in truncated form), and occasionally brief editorial explanations. In the Apparatus criticus, a single square bracket follows the lemma. In the Apparatus criticus, diamond brackets enclose illegible text, indicated by three dots regardless of its length. In the Apparatus criticus, round brackets enclosing three dots indicate intervening words not relevant to the reading in question; they are also used to enclose surrounding or intervening words which are not part of the variant but are supplied to explain it, and editorial comments (in italics).


added altered conjecturally emended (by) continued (citation of Scriptural texts or prayers) corrected (to reading of lemma) deleted (by any method) expuncted interlined omitted (referring back to reading of lemma) replaced (by) similarly (referring back to immediately previous reading) supplied transposed variant reading(s)