Finding the Female Voice in Anselm's Orationes sive Meditationes: New Manuscript Evidence

Abstract

This article presents new manuscript evidence of one of the most widely read devotional texts of the Middle Ages – the Prayers and meditations (Orationes sive meditationes) of St Anselm, abbot of Le Bec and archbishop of Canterbury – to argue that some of these prayers were originally written and disseminated in the voice of a woman. Composed and copied in grammatically feminine forms, these prayers were, as I show here, deliberately designed to be used and recited by women. This new and hitherto completely overlooked evidence is transformative for our understanding of women's active voices in medieval cultures of devotion.

In 1104, Anselm, exiled archbishop of Canterbury (1093–1109), sent a very special book to Matilda, margravine of Tuscany (1055–1115), ‘the greatest woman of her time’,¹ with whom he shared a close political relationship and personal friendship.² What made this book so special was less its material preparation than its contents. Unlike the lavish presentation copies for which medieval authors often spared no effort or expense to help convey their debt and gratitude to powerful patrons and benefactors, the book for Matilda was a much more modest affair. It had been prepared in haste and with limited resources during Anselm's second exile abroad (1103–6), far from the well-oiled machine that was the domestic scriptorium and manuscript workshop available to him back at Christ Church Cathedral Priory in Canterbury.³ In a letter dispatched shortly before or together with the now-lost manuscript, Anselm expresses surprise (and perhaps embarrassment) upon discovering that Matilda was not yet in possession of his prayers and meditations (‘orationes sive meditationes, quas ego dictavi et putabam vos habere, non habebatis’), prompting him to furnish her with a personal(ised) copy (‘ideo mitto eas vobis’) as a matter of urgency.⁴ After all, some of these prayers were circulating widely at this point,⁵ so Anselm's astonishment was likely genuine. Indeed, few (if any) of those he sent to Matilda were new compositions written specifically for her, but most (perhaps all) were copies of prayers Anselm had published previously for his monastic brethren and/or for the friends (‘iuxta desiderium et petitionem amicorum suorum’) he had made both at Canterbury and during his earlier career as prior (1063–78) and then abbot (1078–93) of Le Bec-Hellouin in Normandy.⁶ Anselm admits as much in a second (and much shorter) dedication letter to Matilda that serves as an authorial preface for the prayer collection sent to her. He even asks the countess – who had a reputation for sponsoring authors attached to her court⁷ – to help disseminate his prayers further by granting interested third parties access to her personal copy:

It pleased your highness [Matilda] that I send her the prayers I had produced for different brothers at their individual requests. Though there are some amongst them that are not appropriate to your person (‘In quibus quamvis quaedam sint quae ad vestram personam non pertinent’), I wanted to send them all so that anyone to whom they are pleasing may take them from this exemplar (‘ut, si cui placuerint, de hoc exemplari eas possit accipere’).⁸

A question that has been asked repeatedly by scholars but which has yet to receive a definitive answer is why Anselm felt compelled to point out to Matilda that certain prayers he had sent were not suitable for her. What did he mean when writing that these ‘are not appropriate to your person’ (‘ad vestram personam non pertinent’)? To which prayers in particular was he referring, and what made them unbefitting? Some have suggested that the entire prayer collection received by Matilda had been assembled previously for an altogether different purpose, possibly to serve as the author's personal reference or working copy, and that Anselm simply ‘recycled’ it wholesale for the countess.⁹ Building on this hypothesis, it has been proposed that Anselm might have felt self-conscious and perhaps uncomfortable about presenting his venerated patron and protector with an off-the-peg compilation, and that ‘his [Anselm's] apology that some of the prayers he did send were “not appropriate to [her]” may reflect knowledge that another Anselm, Anselm of Lucca (d. 1086), had composed a series of five prayers especially for the countess’.¹⁰ It has even been conjectured that Matilda would have found the prayers disheartening – and that Anselm must have known full well she would – since they preached virginity and withdrawal from the world to a secular lord and twice-married woman who had experienced the trauma of child loss, but whose innermost desire it was to take monastic vows and live the life of a cloistered bride of Christ, as she herself had confessed to Anselm on multiple occasions.¹¹

Intriguing though they may be, such explanations find little concrete support in the sources, and none of them is corroborated by the extant manuscripts. I will thus propose a new explanation borne out directly by new manuscript evidence, one that revisits some fundamental assumptions about the original design, dedication and dissemination of Anselm's Orationes sive meditationes. I will argue that the most plausible reason why some – indeed, the majority – of the twenty-two prayers and meditations Anselm sent to Matilda in 1104 were flagged as not appropriate to her persona was that they had been written from a male perspective and, crucially, in a male first-person voice (i.e. using grammatically masculine forms of nouns such as servus, peccator, debitor and adjectives like dignus/indignus when speaking in the first person or referring to oneself in the third person), making them difficult to use and recite even for a woman as educated and versed in Latin as Matilda, of whom Anselm ‘expect[ed] and assum[ed] an accomplished level of literacy’.¹² Excellent research has been done on the use(s) and reception of prayers amongst medieval women both generally and with specific regard to Anselm's writings, including from art-historical perspectives,¹³ but so far little consideration has been given as to how these female voices may be reflected in the texts’ grammar as it is presented and preserved in the surviving manuscripts themselves. Consequently, it has gone all but unnoticed that there is one particularly early manuscript of the so-called ‘Matildan recension’ (i.e. the arrangement of prayers and meditations that derives, directly or indirectly, from Matilda's lost personal manuscript) in which several of Anselm's prayers are written wholly in the female voice (i.e. using grammatically feminine forms such as ancilla, peccatrix, debitrix and digna/indigna). Unknown to the texts’ editors and thus completely overlooked in scholarship to date, this new manuscript evidence is – as I will demonstrate – uniquely significant and transformative for our understanding of women's active voices in one of Anselm's most widely read and transmitted works.

New manuscript evidence

Copies of Anselm's prayers and meditations survive in over a hundred medieval manuscripts,¹⁴ thirteen of which have so far been identified as belonging to the ‘Matildan recension’.¹⁵ To these we can add a fourteenth witness in the shape of Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (hereinafter B). To appreciate why this little-known manuscript is of particular significance, more so perhaps than any of the better-known copies of the same recension, we must first investigate its contents and materiality, including its scribal preparation, which will then allow us to locate and contextualise it within the existing manuscript tradition (see Table 1). The fullest – and, to my knowledge, the only – published description that provides the basic point of reference and the starting point for any more detailed investigation of B is provided by Irene Stahl in her catalogue of the University of Bremen's medieval manuscript collection.¹⁶ The first matter to note is that B is a composite codex consisting of two previously independent codicological units, both from the twelfth century, which were combined at an unknown point but very likely prior to the insertion of the later medieval (saec. xv?) ownership mark (fo. i r) that names the book as the property of the Cistercian monastery of Schönau. The second and shorter unit (fos 49–56) contains a selection of seven texts (in excerpts) that mostly deal with eschatological and apocalyptic matters.¹⁷ B's first and longer codicological unit (fos i–xlvii), which is the one of interest here, contains eleven prayers and one meditation written by Anselm (see Table 2), arranged in a way that is unusual and indeed unique within the work's manuscript tradition. I return to this arrangement below. This unit is a fragment both textually and in terms of its quire structure. Several sheets have been cut severely, often leaving little more than stubs (fos viii, xv–xvi, xliv–xlv), one sheet is missing entirely (fo. xvi), and the final quire (vi) is missing the last sheet (fo. xlviii) that would have completed the text of Oratio 12 addressed to St John the Evangelist.¹⁸ We can no longer know how many (if any) additional sheets or quires would have followed after that, but probably just about enough to accommodate the remaining eight prayers and two meditations so as to complete the set – the ‘recueil complet …, définitif et cohérent’,¹⁹ to borrow André Wilmart's terminology – in keeping with the thirteen other known manuscripts.²⁰

Table 1. Extant manuscripts of the ‘Matildan recension’ of Anselm's Orationes sive meditationes.

* Sigla established in Anselmi Cantuariensis opera omnia have been adopted to facilitate cross-reference; all other sigla are mine.

** Added by Schmitt, ‘Zur neuen Ausgabe’.

Table 2. Contents of Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B).

* Rubric lost due to cropping or loss of parchment.

Or. = Oratio.

Med. = Meditatio.

PL = Patrologia Latina.

Another striking feature of B's materiality and scribal preparation is just how small and compact it is in size. The sheets containing Anselm's prayers and meditations measure slightly less than c.17 × 11 cm with no apparent cropping, or at least not substantially so, considering that the margins surrounding the ruled writing area (c. 13.5 × 7.5 cm) are sizeable and the pin-prick holes from the ruling (20 lines/page) are fully visible along the vertical edges. The foliation with Roman numerals written centrally above the first line of text on the top of every page is also intact and contemporaneous with the hand that copied the prayers, and in the opening quire (i) this was done by the main scribe him-/herself (fos i–v) before a contemporary, possibly a collaborator or assistant/amanuensis, completed the remainder (fos vi–xlvii). The entire text – including the rubrication and, possibly, the large pen-flourished red initials characteristic of many Cistercian (and some Benedictine) manuscripts from the twelfth century – is the work of a single scribe.²¹ The visual impression (aspect) of the scribe's straight, angular handwriting with upright letter forms and a high degree of horizontal compression (see Figure 1) is so unmistakably characteristic of twelfth-century scriptoria and scribal culture in the region north of the Alps that today comprises southern Germany, Austria and Switzerland as to eradicate all doubts that B's fragment of Anselm's Orationes sive meditationes originates there, even if the exact provenance eludes us. (We only have to look to centres of book production such as Admont, Göttweig, Melk, Engelberg, Lambach etc. to detect similarities.²²) This localisation is significant, not least since all of the previously identified twelfth-century manuscripts of the ‘Matildan recension’ come from this region (N from Traunkirchen or Nonnkirchen; Z from Zwettl; S from Zwiefalten; K1 from somewhere in Austria) (see Figures 2–4).²³ Samu Niskanen, in his recent research on the authorial publication and early manuscript dissemination of Anselm's works (including Cur Deus homo and De incarnatione verbi), likewise stresses the importance of Germany (or rather the German-speaking territories) as a hotbed of transmission in the twelfth century.²⁴

Figure 1. Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B), fo. i v. Reproduced under Public Domain licence (Mark 1.0 Universal) from the Digital Collections of the Staats- und Universitätsbibliothek Bremen.

Figure 2. Benediktinerstift–Stiftsbibliothek, Admont, ms Cod. 289 (N), fo. 3r. Reproduced with permission.

Figure 3. Augustiner-Chorherrenstift–Stiftsbibliothek, Klosterneuburg, ms CCl 798 (K1), fo. 3r. © Stiftsbibliothek Klosterneuburg. Reproduced with permission.

Figure 4. Württembergische Landesbibliothek, Stuttgart, ms Cod. theol. et phil. qt. 234 (S), fo. 14r. Reproduced with permission.

If B's origin within this recognised hotbed positions it as a prime witness for the earliest stages in the manuscript tradition of the Orationes sive meditationes, specifically the ‘Matildan recension’, then this position is underscored by the added fact that the palaeographically significant features in the hand of its scribe point to a date of production at least contemporaneous with – if not slightly earlier than – that of the codex usually given pride of place in scholarship (N). Whilst it is impossible to establish an absolute date of production for B based on internal evidence, we can determine a relative date by situating the scribe's penmanship within the wider European development of what Erik Kwakkel has dubbed the ‘transitional script of the long twelfth century’.²⁵ Generated from a corpus of hundreds of dated manuscripts made c.1075–1225 and distinguished by geographical areas – amongst them the shared Kulturraum of Germany, Austria and Switzerland – the chronological progression of twelfth-century script(s) across the Latin West traced by Kwakkel provides a helpful benchmark for dating otherwise undated manuscripts based on quantifiable palaeographic features like angularity, fusion between adjacent letters and the execution of minims.²⁶ Judging from these kinds of features, B was likely produced in the second or third quarter of the twelfth century (c.1125–75) around the same time as or perhaps even before N (held to have been produced c.1150 × 60), which scholars studying the manuscript tradition of the Orationes sive meditationes have considered the manuscript closest to (or even a copy of) the lost archetype that was Matilda's personal copy prepared by Anselm himself or at his behest in 1104.

Another reason (besides its date) why N has been given preferential treatment amongst the earliest extant manuscripts of the ‘Matildan recension’ is its extensive programme of illustration. One of only three surviving illustrated twelfth-century copies of Anselm's prayers and meditations (the other two being Bodleian Library, Oxford, ms Auct. D.2.6, most probably from Littlemore or Harrold, and Bibliothèque d’étude du Grand Verdun, Verdun, ms 70, from St Albans, both belonging to different recensions), N features eleven multi-coloured miniatures, several of them full-page images, including the famous frontispiece (fo. 1v) that depicts Matilda receiving her copy of the Orationes sive meditationes from Anselm (see Figure 5).²⁷ The scene shown on this frontispiece should not be taken as evidence to suggest that Matilda's copy was itself an illustrated one, however. Indeed, even the depicted volume's intricate (bejewelled?) binding is likely a product of the miniaturist's own imagination and artistic licence. As noted above, Anselm had produced the book he sent to Matilda in haste and with limited resources whilst abroad (at Lyon), and a deluxe presentation copy boasting a bespoke cycle of nearly a dozen illustrations designed from scratch would almost certainly have been beyond his means at that point.²⁸ As others have argued compellingly, it is much more plausible that these miniatures and the codices in which they survive were made for – and perhaps by – twelfth-century communities of religious women. Dorothy Shepard's perceptive discussion of the arrangement of the miniatures in N and those in the so-called ‘Littlemore/Harrold Anselm’, which also takes into consideration the manuscripts’ visual accessibility and punctuation, thus plausibly suggests that they would have found a locus in both private and communal settings.²⁹ As Linda Olson observes, however, illustrated codices such as N in which ‘Matilda's literate and devotional presence is grafted onto – even inscribed into – Anselm's Orationes sive meditationes’ are really ‘only the tip of the iceberg when it comes to women's involvement in the circuits of textual exchange associated with Anselm's popular devotional writings’.³⁰ Focusing too narrowly on their vivid visual testimony to the neglect or exclusion of non-illustrated manuscript witnesses may be running the risk of unduly prioritising image over text in ways that hinder rather than help our understanding of twelfth-century reading practice.

Figure 5. Benediktinerstift–Stiftsbibliothek, Admont, ms Cod. 289 (N), fo. 1v. Reproduced with permission.

Whilst considerable attention has been paid to medieval women's engagement with Anselm's prayers and meditations through manuscript imagery and other forms of non-verbal mediation, much less thought has been given as to how women engaged directly with the texts themselves. Though it is surely correct to suggest that women (and men, for that matter) would have read these images as texts,³¹ we should not assume that they did not at the same time read and recite the texts proper. And yet, the practical challenges of taking prayers originally composed by/for men in a male voice and then giving them to women to read and perform – and, ultimately, to embody – have never been considered with regard to the Orationes sive meditationes.³² The tacit assumption often seems to be that women simply adopted Anselm's prayers wholesale, male voice and all, and spontaneously adapted them to suit their female personae, which in practice would have meant instinctively and systematically changing various grammatically masculine forms to their feminine equivalents when reading out these texts straight from the written exemplars or committing them to memory, be that during private devotional practice or in communal settings. As Mary Edsall reminds us, however, even memorised prayer informed by the lectio divina ultimately ‘needs a textual base … from which the ruminating mind can take off into its own prayer, meditation, and contemplation’.³³ Moreover, the difficulty of making impromptu grammatical adjustments to an orally delivered text with any level of consistency and accuracy should not be underestimated, even for experienced readers with high levels of literacy/Latinity such as Matilda or the skilled female scribes who produced the earliest copies of Anselm's prayers in twelfth-century Germany;³⁴ unless, of course, some of the prayers had in fact been written down in a female voice from the very beginning to help enable and facilitate their delivery by women, which – as I will demonstrate now – is precisely what we can see in B.

Anselm's female voice

Texts – especially prayers – written in a woman's voice are a rare and precious source for scholars of the Middle Ages.³⁵ Besides later medieval women's books of hours, relatively few specimens survive compared to the numerous examples that preserve the voices of male authors and their male audiences, so that whenever a new one resurfaces it is almost guaranteed to make the news both inside and outside the academy.³⁶ It is surprising, therefore, that no notice seems to have been taken by scholars of Stahl's important observation, made in passing in her 2004 catalogue description of B, according to which ‘the praying person of the final text [Oratio 12] is grammatically feminine, having been corrected to the masculine form by a second hand’.³⁷ What escaped Stahl's attention, though, is that this is true not just of Oratio 12 (fos xlv r–xlvii v), but equally of four of the remaining eleven prayers also copied in B (Oratio 1, fo. ii r–v; Orationes 8–10, fos xxx v–xlv r), all of which are written – grammatically speaking – from the perspective of a woman (see Figures 6–7). As a result, these prayers channel the first-person voice of not a male but a female speaker; they embody the persona not of God's faithful son (filius) and servant (servus) but of his daughter (filia) and handmaiden (ancilla); they ask mercy and forgiveness of sins not for the peccator but for the peccatrix; etc. All in all, there are one hundred and twenty-three instances in which B has a feminine form in lieu of the equivalent masculine form found in other copies of the Orationes sive meditationes, and in all but a handful of instances a later medieval corrector (judging from the letter forms and ink, we are probably dealing with the same scribe who wrote the ownership mark on fo. i r) – has added the corresponding masculine endings (usually superscript -us) in the space between the lines. Pronouns (ista, ea, illa, ipsa, haec etc.) apart, the most frequently used/corrected nomina feminina are peccatrix (for peccator) and filia (for filius), followed by ancilla (for servus) and famula (for famulus), with the most frequent adjectives being mortua (for mortuus), misera (for miser) and viva (for vivus) (see Table 3).³⁸ The choice of feminine forms whenever the grammatical subject is the reader herself is consistent – virtually without exception – in Orationes 1, 8–10 and 12, textual lacunae caused by (partial) page loss notwithstanding, whereas the other prayers (including the one meditation) extant in the manuscript use the masculine forms (e.g. fo. xvii v: ‘ego peccator’; fo. xxiiii r: ‘ego ipse’; fo. xxix r: ‘ego non sum dignus’; etc.).

Figure 6. Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B), fo. xxxviii r. Reproduced under Public Domain licence (Mark 1.0 Universal) from the Digital Collections of the Staats- und Universitätsbibliothek Bremen.

Figure 7. Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B), fo. xlii r. Reproduced under Public Domain licence (Mark 1.0 Universal) from the Digital Collections of the Staats- und Universitätsbibliothek Bremen.

Table 3. Feminine forms of Latin terms in Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B).

* All pronouns, nouns and adjectives are rendered in the nominative case, though the texts preserved in B also feature accusative, genitive, dative and ablative cases.

** Only in one instance (Oratio 8) does the late medieval corrector gloss ancilla with servus.

*** Not glossed by the corrector.

Or. = Oratio/Orationes.

How can we explain this intriguing use of grammatically feminine forms, then, especially given that none of the other extant manuscripts of the Orationes sive meditationes exhibits this phenomenon? The most obvious and straightforward explanation would seem to be that B – like N and the ‘Littlemore/Harrold Anselm’ – was made for, and possibly by, a female monastic community, whose female(?) scribes made the conscious effort of ‘gendering’ the prayers in the process of copying them from a ‘male exemplar’ (i.e. one that used the standard masculine forms) so they and their fellow sisters could use them more easily and effectively in their own devotional routine. Though the first part of this explanation (i.e. the manuscript's possible origin in a female community and/or scriptorium) seems to stack up against the evidence due to palaeographic similarities between B and other twelfth-century German manuscripts known to have been produced by women,³⁹ the second part (i.e. crediting the introduction of the female voice to the initiative of these nuns) is implausible for several reasons. First, had it been the community of religious women who first introduced the feminine forms into the manuscript tradition of Anselm's Orationes sive meditationes in an attempt to enhance their usability in a female context of reception, then why would they have limited their intervention to just five of the twelve (and perhaps once twenty-two) prayers/meditations copied in B, rather than adjusting the entire corpus? And why those particular prayers, but none of the others? Second, the notion that B was produced from a ‘male exemplar’ and had its grammar and voice adjusted in the process is rendered, if not altogether impossible, then at least improbable by the fact that its internal arrangement of the prayers and meditations is unlike that of any other surviving manuscript of the ‘Matildan recension’ (see Table 4). The majority (>70 per cent) follow the standard arrangement identified by Wilmart,⁴⁰ which is held to be ‘the nearest we have to Anselm's final intention [in which he] placed his last words (prologue, Oratio 1, Meditatio 3, Oratio 3) at the front of the sequence already established for the rest (Orationes 2, 4–19, Meditationes 1–2), when arranging the book for the countess’.⁴¹ B differs from this supposedly definitive arrangement in two significant respects: not only does it organise the first five texts differently by allocating Oratio 3 a later place within the sequence (e.g. Oratio 1, Meditatio 3, Oratio 2, Oratio 4, Oratio 3… vs. Oratio 1, Meditatio 3, Oratio 3, Oratio 2, Oratio 4…), but it also omits Oratio 11 altogether. Whilst the first of these notable idiosyncrasies is shared by three other manuscripts (E, L, S) that are of later dates than B, the second is completely unique within the known corpus.⁴² Finally, and perhaps most significantly, there is important evidence in Anselm's extant personal correspondence that helps resolve these conundra and allows us to identify with confidence the source of the female voice uniquely preserved and promoted in B: Anselm himself.

Table 4. Arrangement of the Orationes sive meditationes in the extant manuscripts of the ‘Matildan recension’

* Cf. Wilmart, ‘Les Prières’, 41–2; Sharpe, ‘Anselm as author’, 57–8 n. 155.

Or. = Oratio.

Med. = Meditatio.

a.r. = alia recensio.

As has often been noted, Matilda was neither the only nor the first recipient of Anselm's prayers and mediations in the form of a personalised copy prepared or overseen by the author himself, though she was probably the first to receive the complete set.⁴³ Over half (64 per cent) of the twenty-two prayers accepted as Anselm's authentic authorial compositions were certainly written prior to his archiepiscopal appointment,⁴⁴ with three of uncertain date (Orationes 10, 12, 17) and four known as post-1093 creations (Orationes 1, 3–4; Meditatio 3). Three of the prayers from the pre-1093 corpus (Orationes 5–7) were composed at the request of an unknown monk of Le Bec and sent (c.1063 × 77) to Gundulf, a fellow monk who had left Le Bec for the recently founded (1063) abbey of Saint-Étienne de Caen with Lanfranc and who would subsequently become bishop of Rochester. As Anselm explains to Gundulf in an enclosed letter, the three texts represent the author's successive and increasingly desperate attempts at crafting a single prayer to the Virgin Mary.⁴⁵ All three are included in B, but none of them has been rendered in the female voice. Around the same time, or shortly afterwards, between 1075 and 1078, Anselm once again sent a collection of prayers to a personal acquaintance alongside a personal letter. Unlike Gundulf, however, the recipient of this letter (and of the enclosed prayers) was not a man, but – like Matilda three decades later – a well-connected aristocratic woman: Adeliza, domina regia nobilitate, eldest daughter of King William i ‘the Conqueror’ (1066–87).⁴⁶ Anselm informs Adeliza that he has sent her seven prayers (‘orationes septem’) for her own devotional practice, one of which should be thought of as a meditation (‘[q]uarum prima non tantum Oratio quantum Meditatio dicenda est’), with two others being dedicated to SS Stephen and Mary Magdalene, respectively (‘oration[es] sancti Stephani et sanctae Mariae Magdalenae’).⁴⁷ The latter have been identified as Oratio 13 and Oratio 16, and the meditation as Meditatio 1, all three of which might originally have featured in the now lost final quire(s) of B.⁴⁸ Scholars have failed to reach a consensus as to the identification of the other four prayers Anselm sent to Adeliza. The most recent editor of Anselm's letters, Niskanen, concludes that they ‘cannot be identified’,⁴⁹ echoing Richard Sharpe's earlier assessment, according to which ‘there is no way of knowing’.⁵⁰ Previous scholars were less pessimistic, with Wilmart tentatively suggesting an identification with Orationes 8, 9, 10 and 14,⁵¹ though Richard Southern argued that the last must postdate Anselm's letter to Adeliza and thus championed Oratio 11 instead whilst maintaining the validity of Wilmart's other identifications.⁵² Southern's identification (Orationes 8, 9, 10, 11) was adopted and reinforced by Jean-François Cottier.⁵³

The general scepticism of recent scholarship notwithstanding, the new manuscript evidence provided by B may serve to rehabilitate both Wilmart's and Southern's suggestions, though Southern was, as we shall see, also mistaken concerning the identity of one of the four unspecified prayers. Besides Matilda, Adeliza is the only known female recipient of a personal copy of Anselm's prayers.⁵⁴ As already noted above, three of the seven prayers sent to her (Orationes 13 and 16; Meditatio 1) are no longer present in B, though given the fragmentary state of preservation they might well have been at the point of B's creation in 1104. The same is true of Oratio 14, though, as Southern rightly noted, we may safely discard this prayer from the collection received by Adeliza in the later 1070s. The alternative offered by Southern (Oratio 11) is, as we saw, the very one that is absent from B but preserved in all other witnesses of the ‘Matildan recension’. If this were indeed one of the prayers sent to Adeliza because its author deemed it particularly appropriate for a woman, then its unique omission from B would seem extremely difficult to explain. Unlike Orationes 12–19 and Meditationes 1–2, the absence of Oratio 11 cannot be ascribed to a loss of parchment, with the text of Oratio 10 leading directly and seamlessly into that of Oratio 12 on what little remains of B's mutilated fo. xlv r today (see Figure 8). Let us assume, therefore, that whoever designed B or the unknown/lost exemplar from which it was made with a female readership in mind would have had a vested interest in preserving not just some but all the prayers Anselm had composed for and/or sent to women, namely Matilda and Adeliza, and we can now try – by means of elimination – to identify amongst its extant contents (Orationes 1–10 and 12; Meditatio 3) those prayers that might originally have been intended for the Conqueror's daughter.

Figure 8. Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B), fo. xlv r. Reproduced under Public Domain licence (Mark 1.0 Universal) from the Digital Collections of the Staats- und Universitätsbibliothek Bremen.

Four of the twelve prayers still present in B (Orationes 1, 3–4; Meditatio 3) we can discount immediately from being part of the ‘Adelizan recension’ (to introduce a shorthand) since they are amongst those which have been identified as written after Anselm's archiepiscopal appointment in 1093. We should probably also reject the three prayers that Anselm wrote for Gundulf (Orationes 5–7) as we cannot be certain which commission (Gundulf's or Adeliza's) came first, and neither of their accompanying letters bears a date of composition. Even if Gundulf's were slightly earlier, it would seem peculiar for Anselm to ‘rebrand’ prayers dedicated explicitly to his male friend and former brethren by sending them to a secular woman shortly afterwards with detailed guidance as to how Adeliza should read them for her spiritual wellbeing.⁵⁵ Had that been the case, we could reasonably expect Anselm's letter to Adeliza to feature an apology rather like that found in his letter to Matilda and cited at the beginning of this article, warning Adeliza that the prayers sent to her were actually written for someone else (i.e. Gundulf) and therefore not appropriate to her person, but it does not.⁵⁶ This leaves five prayers (Orationes 2, 8–10 and 12) as possible candidates. The first of these to appear in B's unconventional arrangement (Oratio 2) is the least likely. Described by its modern translator as ‘the prayer that belongs most completely to the “new style” of devotion of the eleventh and twelfth centuries’,⁵⁷ this prayer to Christ is generally considered amongst the works composed during ‘Anselm's last phase’, to adopt Southern's terminology, which began c.1075 at the earliest and extended to the very end of the eleventh century.⁵⁸ The prayer's passionate reflections on the Lord's presence/absence in the world and on Christ's becoming flesh and shedding blood to redeem mankind puts it in close semantic proximity with Anselm's meditation on human redemption (Meditatio 3) by which it is directly preceded in B, as well as in E, L and S, and which we know from Eadmer's first-hand testimony was composed c.1099/1100, during Anselm's exile at Lyon, after the completion of Cur Deus homo, thus suggesting that these two prayers were written around the same time.⁵⁹ In fact, Oratio 2 is not attested in any extant version of the Orationes sive meditationes before the ‘Matildan recension’, meaning we actually have no manuscript evidence at all of its existence pre-1104.⁶⁰

Having thus narrowed down the list of candidates for the four unidentified prayers Anselm sent to Adeliza to precisely four prayers not discarded on chronological or logical grounds or otherwise accounted for in B, we can see something remarkable: these four prayers (Orationes 8–10 and 12) are the very ones which in B are written in a female voice. (The only other text in B using the female voice is Oratio 1, on which see below.) Not only does this identification serve to confirm, at least in part, the suspicions concerning the identities of Adeliza's prayers expressed by Wilmart and Southern but dismissed in subsequent scholarship, but it also provides us with an eminently credible – in fact, the single most plausible – explanation as to why these prayers are written in the female voice: namely because they had been written for a woman from the very beginning (Adeliza) and then, many years later, were sent to another woman (Matilda) who would equally have benefited from their feminine grammar when reciting them as part of her own devotional practice. What had started out as Adeliza's voice – channelled by the prayers’ author, Anselm – thus became Matilda's voice and, ultimately, the voice(s) of the religious women who copied the prayers preserved in B for their own communal usage somewhere in the German-speaking world of the mid-twelfth century.

There are three matters that remain before it is possible to conclude this study. The first is to explain why there is one more prayer (Oratio 1) in B written in a woman's voice even though this one postdates those for Adeliza by some decades. Again, the easiest and most compelling explanation is that, contrary to previous scholarly assessments according to which ‘there is nothing to suggest that Anselm composed any of them [Orationes sive meditationes] expressly for Matilda, nor is there anything to suggest that he tailored them to suit her sex’,⁶¹ he in fact did exactly that with Oratio 1, which is widely accepted as one of the very last prayers Anselm wrote and, to quote Sharpe, ‘may date from no earlier than the prologue [to Matilda]’.⁶² When opening the manuscript sent to her by Anselm in 1104, the first text Matilda would have encountered after the dedication that doubled as a new bespoke prologue was, of course, none other than Oratio 1, which takes pride of place in all known copies of the ‘Matildan recension’, B included. What better way for the author to please his patroness than by opening with a prayer he had either composed specifically for her or customised to suit her person (and gender) by rendering it in a feminine voice. Given that Oratio 1 contains but a single yet prominently placed self-referential term (peccatrix, fo. ii r, line 4; see Figure 9), this would have meant minimum effort for maximum effect, and one which was almost guaranteed to be noticed and appreciated by the recipient. It was also noticed – but, judging from the erasure and correction, less appreciated – by the manuscript's later medieval corrector, who manually amended peccatrici to peccatori alongside the more than one hundred such amendments he made later in the book to the four prayers written for Adeliza (Orationes 8–10, 12).

Figure 9. Staats- und Universitätsbibliothek, Bremen, ms msc 0037 (B), fo. ii r. Reproduced under Public Domain licence (Mark 1.0 Universal) from the Digital Collections of the Staats- und Universitätsbibliothek Bremen.

Next to consider is why this corrector (presumably a man) went through the entire manuscript to ‘re-gender’ its grammatically female forms where he found them by adding the masculine equivalents as interlinear glosses.⁶³ A plausible answer is provided by the same corrector's other contribution in the shape of the ownership mark he inscribed manu propria (fo. i r) to place B in the possession of the – presumably his – monastery of Schönau, a community of male Cistercians founded 1142 in the Odenwald, southern Germany. The likely date of the corrector's hand indicates that the codex only came to Schönau during the fifteenth century, not long before the monastery's dissolution in 1558. When it did, Schönau's monks probably found it challenging (or at least cumbersome) to identify with and recite the five prayers written in a female voice,⁶⁴ thus one of them annotated these prayers to facilitate their delivery – a highly instructive example of what was outlined earlier in this article concerning the difficulty of spontaneously adjusting a text's grammar during oral delivery, even amongst medieval literati accustomed to praying both privately and collectively on a daily basis. The fact that no prior attempt was made to insert the masculine forms further corroborates my hypothesis that B had previously been owned – and was likely produced – by a community of religious women, perhaps one located not far from Schönau, who would have had no such difficulty in using the prayers’ female voice.

It is naturally tempting to imagine that the lost exemplar from which these female monastic scribes copied B during the second or third quarter of the twelfth century was none other than Matilda's personal volume made by Anselm. In favour of this possibility is not only the fact that no other extant manuscript shares B's unique arrangement of the prayers and/or its gendered versions of Orationes 1, 8–10 and 12, but also, as noted above, B's remarkably compact size, which may reflect the size and shape of the book Anselm had prepared for Matilda at Lyon with limited time and resources. Another possibility worth entertaining in principle is that Anselm might have kept an authorial draft or duplicate of the manuscript he sent to Matilda for himself, and that B was copied directly from it, similar to the case Niskanen has made for the copies of Anselm's letters preserved in Stadtbibliothek, Trier, ms 728/282, according to which they descended from a now-lost copy kept in Anselm's personal archive (Niskanen's ζ) without any known intermediaries.⁶⁵ Like B, the manuscript containing these letters is of conspicuously small format, and its likely origin at the abbey of St Eucharius/Matthew in Trier in the mid-/late twelfth century again testifies to the importance of Germany and its two southern neighbours as a hotbed for the early transmission of Anselm's works, possibly based on his autographs or, given Anselm's rank as abbot and later archbishop, texts he himself had dictated to a secretary/amanuensis.⁶⁶ Whilst, on balance, B's descent from Matilda's personal manuscript seems the more plausible, especially since she is known to have ‘contributed actively to the dissemination of the books gifted to her’,⁶⁷ we should remain open to both possibilities.

The third and final matter, which brings our investigation full circle, is what Anselm meant when he told Matilda that some of the prayers he sent to her were ‘not appropriate for your person’ (‘ad vestram personam non pertinent’). The most probable answer – as I hope to have shown here – has neither to do with the prayers’ actual content, nor is it suggestive of a self-conscious author selling old wine in new bottles by ‘re-packaging’ an existing collection wholesale for a new recipient. Nor, indeed, does it indicate a rivalry between Anselm and his namesake from Lucca. Rather, Anselm's allusion in his letter to Matilda was probably related, quite simply, to the fact that most of the twenty-two prayers of the so-called ‘Matildan recension’ had actually not been written from the perspective and/or in the voice of a woman. This is surely why he deemed them inappropriate for Matilda, or at least not as appropriate as the four prayers once designed for Adeliza (Orationes 8–10, 12) and the one he had more recently composed (or customised) for the countess herself (Oratio 1). Had B not lost several sheets/quires and become a fragment whose text terminates prematurely, there is a good chance that it might also have provided us with female versions of Anselm's three remaining prayers for Adeliza (Orationes 13, 16; Meditatio 1), which would have made the collection as a whole a more balanced and – in Anselm's own words – more appropriate one for its designated female recipient and other medieval female users.

The significance of the arguments presented in this article extends far beyond the context from which they were generated. Enabled by new manuscript evidence of Anselm's Orationes sive meditationes in the ‘Matildan recension’, the findings are transformative not only for our knowledge of these important and widely read texts and the history of their composition and earliest transmission in twelfth-century Europe, but also, and perhaps more significantly, for our understanding of women's participation in high medieval cultures of devotion, in particular the power of the female voice(s) in literary conversations that, though often (mis-)read as male monologues, were really dialogues involving both men and women.⁶⁸ The voices of Princess Adeliza, Countess Matilda and the (anonymous) members of a twelfth-century community of religious women resonating from a hitherto overlooked manuscript – and amplified here using forensic palaeographical and codicological study – can be appreciated both in their own right and in concert with the voices of others, including Anselm's own authorial voice. As author, Anselm wanted the voices of Adeliza and Matilda to be heard clearly and distinctly in the prayers he sent to them, so in anticipation of their primary (i.e. oral) mode of delivery he deliberately composed some (though not all) of them from the first-person perspective of a grammatically feminine subject. The appreciation of this empathetic strategy amongst the prayers’ medieval audiences is evidenced by the existence of B, which – as I have argued – might have been derived from Matilda's personal copy, though the fact that this important manuscript has escaped our attention until now means there has been little appreciation amongst modern audiences. Various interpretations have been presented in scholarship as to how Anselm's Orationes sive meditationes and similar devotional texts could be tailored to accommodate the needs and identities – indeed, the personae – of medieval female readers, but so far they have relied primarily on visual (e.g. manuscript illustrations) or quasi-textual (‘image-as-text’) kinds of evidence to prove that woman ‘were not only actively included in the learned Anselmian devotional network, but even prioritized as receivers and readers, disseminators and users of Anselm's devotional writings’.⁶⁹ Being able to draw on B's unique textual evidence opens entirely new doors for such lines of argumentation by furnishing us with concrete testimony of female praying voices that did not require alternative or auxiliary modes of engagement such as illumination, but which were inscribed directly in the Latin texts themselves. The fact that B is the sole surviving witness of Anselm's prayers and meditations in this female or gendered version does not have to mean that they were not in wider circulation during the twelfth century, but it certainly reinforces the likelihood that this manuscript was indeed produced straight from the book sent to Matilda in 1104, which Anselm himself explicitly presented to her as an exemplar (exemplar) suitable for copying.⁷⁰ A supremely powerful woman and patroness who readily granted others access to her personal library, Matilda was not just a primary recipient and active user of Anselm's prayers in the same way that Adeliza had been several decades earlier, but she was also a conduit for the prayers’ dissemination to other medieval women (and men) who adopted (and adapted) her voice, which had been conscientiously channelled by Anselm as author, and lent their own voices in turn. The research presented in this article now allows us to hear and appreciate these long-silenced voices anew.