The Text of the Old Testament in Anglo-Saxon England 0521464773, 9780521464772

Citation preview

C A M B R ID G E S T U D IE S IN A N G L O -S A X O N E N G L A N D G E N E R A L E D IT O R S

S IM O N K E Y N E S M IC H A E L L A P ID G E

THE TEXT OF THE OLD TESTAMENT IN ANGLO-SAXON ENGLAND

A S S IS T A N T E D IT O R : A N D Y O R C H A R D Cambridge Studies in Anglo-Saxon England is a series of scholarly texts and monographs intended to advance our knowledge of all aspects of the field of Anglo-Saxon studies. The scope of the series, like that of Anglo-Saxon England, its periodical counterpart, embraces original scholarship in various disciplines: literary, historical, archaeological, philological, art historical, palaeographical, architectural, liturgical and numismatic. It is the intention of the editors to encourage the publication of original scholarship which advances our under­ standing of the field through interdisciplinary approaches.

R IC H A R D M A R S D E N Girton College, Cambridge

Volumes published

I

Anglo-Saxon Crucifixion Iconography and the A rt of the Monastic Revival b y It A It it A it A ■ &TtlBllS?X>NflS'r fl.X-iL.ICR ■: ?s C'.r«CRHÍ>ri’(p«NON?l>AfJC'I >*rOíKn e-'t'öe»pL«ijcrcixs íicsic.ch-ncpm OBSÍIÍ íliXfjtlf\H' xptAsistKX'iijo (DtN» -vnpictis : NON?

: cn

\ V R U N '^ T p Ö j|:x3Í.Ti#C*

i ’.

■

|:.\CKlí^CX>MT|?eC«IX?'rC*B ftTfTRCtirt)ORS(’.X’SXrX « SK»-. \smHX‘.i--i \*mc-i«iv

II London, British Library, Loan 81, verso ( X 0.33). From one of the sister pandects of the Codex Amiatinus. Sir. XXXV. 10-XXXVII.2. The chapter division in col. 2 was added in the thirteenth or fourteenth century, probably at Worcester (see pp. 93 and 129). I Florence, Biblioteca Medicea Laurenziana, Amiatino 1,485r ( X 0.33). Sir. XXXVI.19—XXXVII. 13. The corrections in col. 1 were made in Northumbria but the chapter numbers added much later, in Italy.

" lífí |fU *yW ífof 4*ÆT t

vW Y ^ I W fobb ntOíáz qln 2*if oy ítttiy’jn

oyWt* jSj|fö|g ^ < ; í p k t ^ ^iwnjð

n a m moywT

rfí 4 r ' v m

É

tjwí ^jvrw^wr, • -, ^ u ip ö

v .

.

>' i l N ’ ~M ti*

•: .4 , ‘ 4 |t[ ^ w K U ' ^w nsm ii*

mfM

* y , \ -

o^ptwc^^mW^aMw^ *tatm íb i í *

W ife

'.*••

f

1)

Y^vm^wij' ' i g j f p . d t f ð j t t t

«j«*W tf lf C if

1

%

^4w «m ídtwníí

rirtrpjtfrr- cfar**l

trn^Vqp '•n^nwy» Ol^titr> Acptif&m MomtiMTíJtTTTV

«i*vptfpúÚ/wr. J 'o f’irfl/ue *

c T d c q iT ^ c e - x l l i i m

c o iitu ru tm

*

|i i e « r a i c } ð i f l : jO T T c e - œ rtc e ^ p o t m i u . /

.$

itfaxf&im Itrmr

m a w i V Cambridge, Magdalene College, Pepys 2981 (4), verso (X 0.85). Dan. VIII.3—VIII.7. Pepys pasted the fragment firmly down on page 3 of his calligraphic scrapbook, making the recto invisible (see p. 254).

IJlrgTB-ltistrncan cjm ludt 1 c a n s c en n am S e n a re -c le d n ö Infiomttxcea-lnwttipíiaroxe-condis qtiotfnrcs-iUanv C|íb Innefircan ctfc Ins cjui non TtBiOTiKrtUimi j^ppcmetrh* « 9 Cjwpdein lioti finr

\ mitium pen ueuscce- önitnco^itœ I nom^sepanaritr xdo • p n o b a w c tr aittm conniprtrlnstpi ' eK-rcesQ m iiitnaliifola tmtmcc non Ijtctum I bnrsœ ptenacs nec InhabracuioInam pono-suborco peccocns

qiliArnrttPTvvrtivrr' wt.if.ua,w

rt«}«?me t>»T«^»rr4nt(f «wi^rrot«im nait^Mwn ^ ^ i r a^titvTrTner^

r ' Í C Í T T ’«tM.imr-tí-p^ixJnfw^ nm;Wtr- „ ..* ■ ’’^pmen-vM*!;^. ItiCuyfev. / ojrto{;isj u c| clnornfti

V**>*r>rtúi «vmatnffliur «vmuuv tfrt>bt> n fujquem alcendtfti ncmdtiCer^ d ef.fa lm o itr m ccu m f (^ u d tx tta f Cui fiqursc eiiKabttui^uir q m occum r uqbif ðdocunifr- tu rb a b o a f A iítU rr'^(Jae cja)edt|ii;fc^ppintr u rté inm i cp2crr;jQúcjdríi^ dnoft-orcrirmxúT LaLopttr enTpopti m m ulroi^ c^gmrrT Itiuactiu £rdeftcierrr-Vp repLektr"qrtUirc nofca^gtamdnK^ac^uacc^icrrrrTmare>Uáribe tu cýcj-rr xrca.Lix aex€ir^ ^uörtrtt ignomm iae flip gtam ru^ícp imcpafLiL 'm » n t o p t c tfre* < y u A & r a f a t i i m a l t u d t f r ^ t r e c r p Q f f a

I B

VII London, British Library, Royal 1. E. VII, 113r (X 0.3). The opening of IV Kings. The emendations were made in the post-Conquest period (see pp. 333—4).

VIII Columbia, Missouri, University of Missouri-Columbia Libraries, Fragmenta Manuscripta 4, verso ( X 0.8). Hab. II.5—17. There is a correction in line 14, and emendations in lines 1, 4, 9, 10 and 15 (see pp. 383—4).

1

Introduction W hat page, what word in that divine authority, the Old and the New Testament, is not a most proper standard of human life? Rule of Benedict, ch. lxxiii Bibles laid open, millions of surprises George Herbert, Sitine In letters sent from the spiritual battleground of northern Germany to his friends back home, the Anglo-Saxon missionary and saint, Boniface, made two regular demands: for prayers and for books. He could be quite specific about the latter. W riting at some time between 742 and 746 to Daniel, bishop of Winchester, he asked for a liber prophetarum, which he said had belonged to his late master, Abbot Winbert of Nursling, Hampshire. It was the one, he explained, ‘in which six prophets are to be found together in a single volume, written out in full with clear letters’.1 Nothing in Boniface’s surviving correspondence proves conclusively that he received this part-Bible, but the fact that he was able to cite from Isaiah, Ezekiel and two of the Minor Prophets in a letter to Archbishop Cuthbert of Canterbury, in 747, may be significant.2 We know that other requested books reached him safely.3 Whatever the case, Boniface’s letter highlights the part played by codices of the Old Testament in the life of the Anglo-Saxon monastery and mission. This study will confirm that the IX London, British Library, Sloane 1086, no. 109, verso (X 0.9). Num. 1.52—II. 10. Oddities in the Style-IV minuscule include (in line 11) capital S and G in septuaginta (see p. 392).

1 Ep. lxiii: ‘ubi sex prophete in uno corpore claris et absolutis litteris scripti repperientur' (MGH, Epp. select. I, 131). On the identity of the six prophets, see below, pp. 45—6 and, on Boniface and his letters, pp. 68—72. 2 Ep. lxxviii (MGH, Epp. select. I, 161—70). In the case of one of the Minor Prophets, however (Habakkuk), Boniface cites from a canticle version, which he will have known by heart; see below, pp. 69—70. 3 In a letter (Ep. lxxv) written in 746-7 to Archbishop Egbert of York (r. 732-66), Boniface acknowledges receipt of ‘gifts and books' (‘muneribus et libris susceptis’) and asks for more of the latter, namely some of Bede's treatises’ ( de opusculis Bedan lectoris aliquos tractatus’) (MGH, Epp. select. I, 156—8). 1

1

Introduction W hat page, what word in that divine authority, the Old and the New Testament, is not a most proper standard of human life? Rule of Benedict, ch. lxxiii Bibles laid open, millions of surprises George Herbert, Sitine In letters sent from the spiritual battleground of northern Germany to his friends back home, the Anglo-Saxon missionary and saint, Boniface, made two regular demands: for prayers and for books. He could be quite specific about the latter. W riting at some time between 742 and 746 to Daniel, bishop of Winchester, he asked for a liber prophetarum, which he said had belonged to his late master, Abbot Winbert of Nursling, Hampshire. It was the one, he explained, ‘in which six prophets are to be found together in a single volume, written out in full with clear letters’.1 Nothing in Boniface’s surviving correspondence proves conclusively that he received this part-Bible, but the fact that he was able to cite from Isaiah, Ezekiel and two of the Minor Prophets in a letter to Archbishop Cuthbert of Canterbury, in 747, may be significant.2 We know that other requested books reached him safely.3 Whatever the case, Boniface’s letter highlights the part played by codices of the Old Testament in the life of the Anglo-Saxon monastery and mission. This study will confirm that the IX London, British Library, Sloane 1086, no. 109, verso (X 0.9). Num. 1.52—II. 10. Oddities in the Style-IV minuscule include (in line 11) capital S and G in septuaginta (see p. 392).

1 Ep. lxiii: ‘ubi sex prophete in uno corpore claris et absolutis litteris scripti repperientur' (MGH, Epp. select. I, 131). On the identity of the six prophets, see below, pp. 45—6 and, on Boniface and his letters, pp. 68—72. 2 Ep. lxxviii (MGH, Epp. select. I, 161—70). In the case of one of the Minor Prophets, however (Habakkuk), Boniface cites from a canticle version, which he will have known by heart; see below, pp. 69—70. 3 In a letter (Ep. lxxv) written in 746-7 to Archbishop Egbert of York (r. 732-66), Boniface acknowledges receipt of ‘gifts and books' (‘muneribus et libris susceptis’) and asks for more of the latter, namely some of Bede's treatises’ ( de opusculis Bedan lectoris aliquos tractatus’) (MGH, Epp. select. I, 156—8). 1

The text of the Old Testament in Anglo-Saxon England

Introduction

books of the Old Testament circulated most commonly throughout the Anglo-Saxon period, alongside gospelbooks, epistles and psalters, in just the sort of handy part-Bible requested by Boniface. Complete Bibles were known at least as early as the end of the seventh century but were doubtless always something of a rarity. Only at the end of the period, according to the manuscript evidence, may their numbers have increased significantly. The legibility of W inbert’s book of prophets was important to Boniface, whose eyesight was failing,4 and a carefully produced manuscript in half-uncial or hybrid minuscule is indicated. There are several surviving analogous examples from English scriptoria of the eighth or early ninth centuries, although it is conceivable that the part-Bible in question was an Italian import, written in uncials. The English Vulgate tradition, as we shall see, was based largely on such exemplars. No investigation of the text or texts of the Vulgate Old Testament known to the Anglo-Saxons - circulating in their complete Bibles, part-Bibles and other manuscripts, and turned into Old English by the earliest translators —has ever been made. It is true that the manuscript evidence for the Old Testament is sparse in comparison with the number of gospelbooks and psalters which survives, but it is not negligible. Seventeen manuscripts, dating from the second half of the sixth to the middle of the eleventh centuries, comprise the raw material for this study.5 The seventeen are defined by their library shelf-marks. In fact, two trios of manuscripts consist of the dispersed fragments of single codices (or, in one case, possibly two codices) and one pair represents the com­ ponents of a two-volume Bible. Conversely, one manuscript is a conflation of the remains of two originally separate codices. Thus the evidence for the Old Testament in Anglo-Saxon England encompasses at least thirteen original books. Two are complete, or almost complete, large-format Bibles, and one is a small-format part-Bible. The other manuscripts consist of isolated leaves or fragments, but their codicological origins can be established by the criterion of page size, which can be estimated with some confidence. This indicates that four more of the books, with com­ paratively large pages, were complete Bibles, and that the six others, with smaller pages, were further part-Bibles, such as Heptateuchs and collec-

tions of the ‘wisdom’ books or, as in the case of Abbot W inbert’s volume, the prophets.6 The count, then, is at least eight small Old Testament volumes and six complete Bibles. Chance will have played an important part in producing the detail of the survival pattern of these biblical manuscripts, but its outline reflects the familiar Anglo-Saxon history of intellectual endeavour, with its two apparent peaks of achievement - the first in Bede’s Northumbria, at the turn of the seventh century, and the second in the wake of Benedictine monastic revival of the later tenth. From the former comes the complete 'Codex Amiatinus’, from the latter the almost-complete Royal 1. E. VII + VIII - both of them books of high textual and codicological quality. Associated with Amiatinus are leaves from a second, similar pandect; more or less contemporary with Royal 1. E. VII + VIII are fragments Irom three other complete Bibles. In between, the evidence is of partBibles, of variable - and sometimes very poor —quality. Vulgate history is inseparable from monastic and church history. In the Anglo-Saxon context, this encompasses the arrival of Christianity from Italy, directly or via Ireland, in the sixth and seventh centuries; the dispatch of AngloSaxon missions back to the Continent in the eighth century; and later a return movement of men, ideas and manuscripts from the Continent in t he post-Carolingian period —a movement that was sporadic in the time of King Alfred and his immediate successors, but continuous from the ntiil-tenth century onwards, when the spirit of continental reform trans­ formed Anglo-Saxon monasticism, at least in the south. The Anglo-Saxon texts of the Old Testament cannot therefore be studied in isolation from the wider history of the Vulgate Bible. Their idiosyncrasies can only be identified and assessed in relation to the textual forms transmitted in other manuscripts. In view of the number of such manuscripts extant and the great length of the Old Testament, the task is an immensely complex one. Adequate tools, in the form of comprehensive collations of at least a representative selection of the manuscripts, are a prerequisite. Historians ol the text of the New Testament text have for many years had the guidance of the complete critical edition by Wordsworth and W hite.7

4 Ep. lxiii: ‘caligantibus oculis’ (MGH, Epp. select. I, 131). 5 I list them below, pp. 40-41. Bogaert, 'La Bible latine’, p. 304, notes a general neglect of the history of the Latin text of the Old Testament, a history of great complexity, in comparison with that of the New Testament.

f' I «li.se uss (lie problems involved in making such assessments below, p. 43Nn//uft/ (IKK1) 193d). For the history of the text, see esp. Metzger, Early VnMon\ and l'is< her, 'Der Vulgaic-Tcxi des Neuen Tcstamentes’ and ‘Das Neue Testa­ men! ni l..i(cmis< her Npi.it he', in lic/írá^c, pp. 31 7 S and I 36 274, respectively.

The text of the Old Testament in Anglo-Saxon England

Introduction

books of the Old Testament circulated most commonly throughout the Anglo-Saxon period, alongside gospelbooks, epistles and psalters, in just the sort of handy part-Bible requested by Boniface. Complete Bibles were known at least as early as the end of the seventh century but were doubtless always something of a rarity. Only at the end of the period, according to the manuscript evidence, may their numbers have increased significantly. The legibility of W inbert’s book of prophets was important to Boniface, whose eyesight was failing,4 and a carefully produced manuscript in half-uncial or hybrid minuscule is indicated. There are several surviving analogous examples from English scriptoria of the eighth or early ninth centuries, although it is conceivable that the part-Bible in question was an Italian import, written in uncials. The English Vulgate tradition, as we shall see, was based largely on such exemplars. No investigation of the text or texts of the Vulgate Old Testament known to the Anglo-Saxons - circulating in their complete Bibles, part-Bibles and other manuscripts, and turned into Old English by the earliest translators —has ever been made. It is true that the manuscript evidence for the Old Testament is sparse in comparison with the number of gospelbooks and psalters which survives, but it is not negligible. Seventeen manuscripts, dating from the second half of the sixth to the middle of the eleventh centuries, comprise the raw material for this study.5 The seventeen are defined by their library shelf-marks. In fact, two trios of manuscripts consist of the dispersed fragments of single codices (or, in one case, possibly two codices) and one pair represents the com­ ponents of a two-volume Bible. Conversely, one manuscript is a conflation of the remains of two originally separate codices. Thus the evidence for the Old Testament in Anglo-Saxon England encompasses at least thirteen original books. Two are complete, or almost complete, large-format Bibles, and one is a small-format part-Bible. The other manuscripts consist of isolated leaves or fragments, but their codicological origins can be established by the criterion of page size, which can be estimated with some confidence. This indicates that four more of the books, with com­ paratively large pages, were complete Bibles, and that the six others, with smaller pages, were further part-Bibles, such as Heptateuchs and collec-

tions of the ‘wisdom’ books or, as in the case of Abbot W inbert’s volume, the prophets.6 The count, then, is at least eight small Old Testament volumes and six complete Bibles. Chance will have played an important part in producing the detail of the survival pattern of these biblical manuscripts, but its outline reflects the familiar Anglo-Saxon history of intellectual endeavour, with its two apparent peaks of achievement - the first in Bede’s Northumbria, at the turn of the seventh century, and the second in the wake of Benedictine monastic revival of the later tenth. From the former comes the complete 'Codex Amiatinus’, from the latter the almost-complete Royal 1. E. VII + VIII - both of them books of high textual and codicological quality. Associated with Amiatinus are leaves from a second, similar pandect; more or less contemporary with Royal 1. E. VII + VIII are fragments Irom three other complete Bibles. In between, the evidence is of partBibles, of variable - and sometimes very poor —quality. Vulgate history is inseparable from monastic and church history. In the Anglo-Saxon context, this encompasses the arrival of Christianity from Italy, directly or via Ireland, in the sixth and seventh centuries; the dispatch of AngloSaxon missions back to the Continent in the eighth century; and later a return movement of men, ideas and manuscripts from the Continent in t he post-Carolingian period —a movement that was sporadic in the time of King Alfred and his immediate successors, but continuous from the ntiil-tenth century onwards, when the spirit of continental reform trans­ formed Anglo-Saxon monasticism, at least in the south. The Anglo-Saxon texts of the Old Testament cannot therefore be studied in isolation from the wider history of the Vulgate Bible. Their idiosyncrasies can only be identified and assessed in relation to the textual forms transmitted in other manuscripts. In view of the number of such manuscripts extant and the great length of the Old Testament, the task is an immensely complex one. Adequate tools, in the form of comprehensive collations of at least a representative selection of the manuscripts, are a prerequisite. Historians ol the text of the New Testament text have for many years had the guidance of the complete critical edition by Wordsworth and W hite.7

4 Ep. lxiii: ‘caligantibus oculis’ (MGH, Epp. select. I, 131). 5 I list them below, pp. 40-41. Bogaert, 'La Bible latine’, p. 304, notes a general neglect of the history of the Latin text of the Old Testament, a history of great complexity, in comparison with that of the New Testament.

f' I «li.se uss (lie problems involved in making such assessments below, p. 43Nn//uft/ (IKK1) 193d). For the history of the text, see esp. Metzger, Early VnMon\ and l'is< her, 'Der Vulgaic-Tcxi des Neuen Tcstamentes’ and ‘Das Neue Testa­ men! ni l..i(cmis< her Npi.it he', in lic/írá^c, pp. 31 7 S and I 36 274, respectively.

The text of the Old Testament in Anglo-Saxon England

Introduction

Those concerned with the Old Testament have been at a disadvantage. Only in 1994 did the large critical edition, initiated by Pope Pius X in 1907 and executed by the scholars of the Benedictine Order in Rome, reach completion with its eighteenth volume, the Liber MaccabaeorumA The first, the Liber Genesis, had appeared as long ago as 1926, under the editorship of Henri Quentin. The handy two-volume edition of the whole Vulgate by Robert Weber, with minimal critical apparatus, largely reproduces the Benedictine text of the Old Testament (and Wordsworth’s and W hite’s text of the New) but introduces some modifications, resulting from further study of the manuscripts.89 An invaluable by-product of the Rome enterprise has been a series of monographs on the transmission and textual history of biblical and liturgical manuscripts, along with numerous papers by the Benedictine scholars associated with the enterprise.101The more recent volumes of the Rome edition have their own extensive critical introductions. Meanwhile, however, an even more ambitious Benedictine critical project is underway at the Vetus Latina Institut at Beuron, directed by H. J. Frede, as successor to Bonifatius Fischer. Some thirty volumes, covering both Testaments, will eventually provide a collation of all known sources of the Old Latin, pre-Hieronymian texts and will thus replace, after two and a half centuries, the three-volume pioneering work of Pierre Sabatier, Bibliorum Sacrorum latinae uersiones antiquae seu uetus Italica. n The Beuron volumes also include information on the Vulgate manuscripts and their coverage in this respect is more comprehensive than that of the Rome Biblia Sacra. Fischer’s edition of Genesis was published in 1949, but the other Old Testament books have been slower to appear, and only Sapientia

Salomonis, edited by Walter Thiele, was complete by 1994.12 A parallel series of volumes from Beuron provides historical surveys and patristic investigations and prints important texts.13 In Madrid, T. Ayuso Marauela has directed a separate project to produce a critical edition of the Vetus Latina Hispanica, based on the conviction that there existed a separate Spanish Old Latin translation. An introductory volume, including useful general information on the transmission of the Latin Bible, has appeared, along with editions of the Octateuch and Psalms.14 The textual history of the Latin Bible, including the Vulgate Old Testament, is still being written, and the survey of the Anglo-Saxon manuscripts which this study presents is a contribution to it. Before listing the manuscripts and explaining the method of my approach to their contents, I give a brief account of the history of the Vulgate text and the forms in which it circulated, and of the continental background against which the Anglo-Saxon witnesses must be considered.

8 On the history of the edition (directed initially by Cardinal Gasquet and housed since 1933 at the Abbazia San Girolamo), see esp. Quentin, Essais, pp. 22-7; Gasquet, ‘Revision’; Gribomont, ‘L’édition vaticane’, pp. 473—9, and ‘Les editions critiques'; and below, pp. 15-18. A useful bibliographical survey of the Latin Bible is in MacNamara, 'Latin Bible’, pp. 10-15. Also helpful, despite gaps, is A. Vernet, La Bible au moyen age: Bibliographic (Paris, 1989). 9 Biblia Sacra iuxta Vulgatam Verstonem, 3rd ed. (Stuttgart, 1983); see p. xxii, on the ‘new’ text. Fischer’s Nouae amcordantiae (1977) are based on an earlier edition of this work. 10 See the series Collectanea Biblica Latina (Rome, 1912—), and the publications of Henri Quentin, Donatien De Bruyne, John Chapman, Pierre Salmon, Henri de Sainte-Marie, Robert Weber and Jean Gribomont. 11 LVA (1743-9).

2

THE VULGATE

Beginnings and early transmission The vernacular of early Christianity, and consequently the Vulgate’ lan­ guage of both the Old Testament (the Septuagint translation from the Hebrew) and the New Testament, was Greek,15 but as the church prospered and spread during the second and third centuries, and the influence of Greek declined, a demand for Latin translations had to be m et.16 The first were probably made in north Africa and may have been

11

" 15 1,1

VL X I/1. Sirach (VI. XI/2), Song of Songs (VL X/3) and Isaiah (Ví. XII) are near completion; see Bibliography. Several volumes of New Testament epistles are already available from Beuron, to complement the earlier edition of the gospels, Itala. Das Neue Testament in altlateinischer Uberlieferung nach den Handschriften herausgegeben, ed. A. Julicher, 4 vols., (Berlin, 1938-63). Vetus Latina: Aus der Geschichte der lateinischen Bibel (Freiburg, 1957-). Fifteen volumes have appeared. Important work on old biblical and liturgical texts was presented in an earlier series, Texte und Arbeite 1-54 (Beuron, 1917-64). An annual Bericht from Beuron describes the Institut’s work. VLH (I 953-). Kenyon, Creek: Bible: h. Tov, ‘The Septuagint’, in Mikra, ed. Mulder, pp. 161—88; W . F. Howard, 'The Greek Bible', in Versions, ed. Robinson, pp. 39—82. die earliesi Latin versions, see esp. Kedar, ‘Larin Translations’, pp. 299-313; Gribomonl, Les plus am iemies’, pp. 1 1 65; and Bogaert, ‘La Bible latine’, pp. 143—56.

The text of the Old Testament in Anglo-Saxon England

Introduction

Those concerned with the Old Testament have been at a disadvantage. Only in 1994 did the large critical edition, initiated by Pope Pius X in 1907 and executed by the scholars of the Benedictine Order in Rome, reach completion with its eighteenth volume, the Liber MaccabaeorumA The first, the Liber Genesis, had appeared as long ago as 1926, under the editorship of Henri Quentin. The handy two-volume edition of the whole Vulgate by Robert Weber, with minimal critical apparatus, largely reproduces the Benedictine text of the Old Testament (and Wordsworth’s and W hite’s text of the New) but introduces some modifications, resulting from further study of the manuscripts.89 An invaluable by-product of the Rome enterprise has been a series of monographs on the transmission and textual history of biblical and liturgical manuscripts, along with numerous papers by the Benedictine scholars associated with the enterprise.101The more recent volumes of the Rome edition have their own extensive critical introductions. Meanwhile, however, an even more ambitious Benedictine critical project is underway at the Vetus Latina Institut at Beuron, directed by H. J. Frede, as successor to Bonifatius Fischer. Some thirty volumes, covering both Testaments, will eventually provide a collation of all known sources of the Old Latin, pre-Hieronymian texts and will thus replace, after two and a half centuries, the three-volume pioneering work of Pierre Sabatier, Bibliorum Sacrorum latinae uersiones antiquae seu uetus Italica. n The Beuron volumes also include information on the Vulgate manuscripts and their coverage in this respect is more comprehensive than that of the Rome Biblia Sacra. Fischer’s edition of Genesis was published in 1949, but the other Old Testament books have been slower to appear, and only Sapientia

Salomonis, edited by Walter Thiele, was complete by 1994.12 A parallel series of volumes from Beuron provides historical surveys and patristic investigations and prints important texts.13 In Madrid, T. Ayuso Marauela has directed a separate project to produce a critical edition of the Vetus Latina Hispanica, based on the conviction that there existed a separate Spanish Old Latin translation. An introductory volume, including useful general information on the transmission of the Latin Bible, has appeared, along with editions of the Octateuch and Psalms.14 The textual history of the Latin Bible, including the Vulgate Old Testament, is still being written, and the survey of the Anglo-Saxon manuscripts which this study presents is a contribution to it. Before listing the manuscripts and explaining the method of my approach to their contents, I give a brief account of the history of the Vulgate text and the forms in which it circulated, and of the continental background against which the Anglo-Saxon witnesses must be considered.

8 On the history of the edition (directed initially by Cardinal Gasquet and housed since 1933 at the Abbazia San Girolamo), see esp. Quentin, Essais, pp. 22-7; Gasquet, ‘Revision’; Gribomont, ‘L’édition vaticane’, pp. 473—9, and ‘Les editions critiques'; and below, pp. 15-18. A useful bibliographical survey of the Latin Bible is in MacNamara, 'Latin Bible’, pp. 10-15. Also helpful, despite gaps, is A. Vernet, La Bible au moyen age: Bibliographic (Paris, 1989). 9 Biblia Sacra iuxta Vulgatam Verstonem, 3rd ed. (Stuttgart, 1983); see p. xxii, on the ‘new’ text. Fischer’s Nouae amcordantiae (1977) are based on an earlier edition of this work. 10 See the series Collectanea Biblica Latina (Rome, 1912—), and the publications of Henri Quentin, Donatien De Bruyne, John Chapman, Pierre Salmon, Henri de Sainte-Marie, Robert Weber and Jean Gribomont. 11 LVA (1743-9).

2

THE VULGATE

Beginnings and early transmission The vernacular of early Christianity, and consequently the Vulgate’ lan­ guage of both the Old Testament (the Septuagint translation from the Hebrew) and the New Testament, was Greek,15 but as the church prospered and spread during the second and third centuries, and the influence of Greek declined, a demand for Latin translations had to be m et.16 The first were probably made in north Africa and may have been

11

" 15 1,1

VL X I/1. Sirach (VI. XI/2), Song of Songs (VL X/3) and Isaiah (Ví. XII) are near completion; see Bibliography. Several volumes of New Testament epistles are already available from Beuron, to complement the earlier edition of the gospels, Itala. Das Neue Testament in altlateinischer Uberlieferung nach den Handschriften herausgegeben, ed. A. Julicher, 4 vols., (Berlin, 1938-63). Vetus Latina: Aus der Geschichte der lateinischen Bibel (Freiburg, 1957-). Fifteen volumes have appeared. Important work on old biblical and liturgical texts was presented in an earlier series, Texte und Arbeite 1-54 (Beuron, 1917-64). An annual Bericht from Beuron describes the Institut’s work. VLH (I 953-). Kenyon, Creek: Bible: h. Tov, ‘The Septuagint’, in Mikra, ed. Mulder, pp. 161—88; W . F. Howard, 'The Greek Bible', in Versions, ed. Robinson, pp. 39—82. die earliesi Latin versions, see esp. Kedar, ‘Larin Translations’, pp. 299-313; Gribomonl, Les plus am iemies’, pp. 1 1 65; and Bogaert, ‘La Bible latine’, pp. 143—56.

The text o f the Old Testament in Anglo-Saxon England

Introduction

known to Tertullian (c. 130-230), but the earliest firm evidence comes in long passages cited by Cyprian of Carthage in the mid-third century, which narrowly predate the first European evidence.17 We should prob­ ably envisage the translations being made locally and piecemeal, as demand dictated, and it is unlikely that there was ever a single, received ‘Old Latin’ version.18 The picture is one of variation and confusion and Augustine was driven to complain that Latin translators were ‘out of all number’, compared with those who translated from Hebrew to Greek.19 Evidence for the preparation of complete Bibles is scanty, although Bonifatius Fischer has drawn attention to the intriguing report of the seizure of biblical codices, including one described as ‘very large’, during a raid by the imperial authorities on a north African sect in 3 03.20 Our most reliable record is the late-sixth-century account in the Institutiones of Cassiodorus of his own monastery of Vivarium, which he says possessed a nine-volume Old Latin Bible and also a large pandect, which he called the Codex grandior?1 This was probably the volume reported by Bede as having reached Wearmouth-Jarrow in 686,22 but it has not survived. The seventh-century Lyon Heptateuch’ is the only substantial extant Bible manuscript with the old text, although it has lost much of Genesis and Exodus.23 There is, however, a large number of fragments.24 Conscious of the uncontrolled proliferation of varying and often inac-

curate Latin texts and the need for a single, authoritative one, Pope Damasus commissioned the scholar Jerome, in 382, to make a revision of the gospels from the Greek. These were finished by 383 and Jerome seems to have followed them with a more cursory revision of some, at least, of the rest of the New Testament. Thereafter, between about 390 and 405, he worked on a revision of the books of the Old Testament.25 He had settled by now in Bethlehem and was able to make use of Origen’s compendious Hexapla Bible (available at Caesarea), and in particular the heavily anno­ tated Septuagint version in the fifth column of that work.26 Before long, however, Jerome became dissatisfied with the Greek and decided to begin the task anew, this time going direct to the Hebrew for his authority, although it is clear that he continued to refer also to both Greek and Old Latin texts. How far his initial ‘hexaplaric’ revision had extended is a matter of dispute.27 Certainly the book of Psalms was completed, for it became the received ‘Gallican’ psalter, and also Job and Song of Songs, which survive in large part, and Proverbs and Ecclesiastes, whose texts are partially recoverable, and a surviving prologue to Chronicles shows that this book should be included also.28 Whatever the case, Jerome thoroughly reworked all the Old Testament books,29 except those rejected from the Hebrew canon, the six so-called deuterocanonical books; he did, reluctantly and rapidly, make fairly free versions of Tobit and Judith,

Also useful are Billon, Old Latin Texts and Sparks, Latin Bible , pp. 100—10. On the classification of Old Latin texts, see Fischer in VL II, l4 * -2 2 * ; and Marsden, Interven­ tion’, p. 235 17 Bogaert, ‘La Bible latine’, pp. 143-4; Gribomont, 'Les plus anciennes’, pp. 47-51; Metzger, Early Versions, pp. 285—93; Kedar, ‘Latin Translations , pp. 299—300. 18 Ibid., pp. 300 - 1 . 19 Doctr. donst. II.xi: ‘Qui enim Scripturas ex hebraea in graecam uerterunt, numerari possunt, latini autem interpretes nullo modo’ (CCSL 32, 42). It is to Augustine that we owe the identification of one Old Latin version as the Ttala’, but the term is confusing and, although used by Jiilicher (above, n. 12), is usually avoided today; see Metzger, Early Versions, pp. 290—3Fischer, LB, p. 38; McGurk, ‘Oldest Manuscripts’, p. 1. Instit. I.v.2 and xiv.2. On the Bibles of Vivarium, see below, pp. 130-9. HA, ch. 15; below, p. 86 and n. 52. Lyon, Bibliothéque de la Ville, 403 + 1964 (VL I, no. 100; CLA VI, no. 771). See VL II 5*_6*; and Fischer, LB, p. 121. The codex is ed. Robert, Pentateuchi Versto (1881) and Heptateuchi Partis Posterioris Versio (1900), with extensive critical material. 24 The list in VL I, 11-34 includes about two hundred. For a comprehensive catalogue of manuscripts, presented book by book for both Testaments, see VLH I, 205-27.

2H 21 22 27

6

The literature on Jerome’s translations and revisions is extensive. Useful are Kedar, 'Latin Translations’, pp. 313-34; Metzger, Early Versions, pp. 330-4; Semple, ‘Biblical Translator ; Sparks, CHB 1, 517—26; and Condamin, ‘Les caractéres’. On Origen’s Hexapla, see Kenyon, Greek Bible, pp. 22-4, and Roberts, Old Testament, pp. 18-20. Origen used a system of asterisks and obeloi to indicate, respectively, elements of the Hebrew missing from the Septuagint version and additions in the latter not authorized by the Hebrew. Sparks, CHB 1, 515 and 531; A. Vaccari, ‘Recupero d’un lavoro critico di S. Girolamo’, in Scritti di erudizione e di filologia II: Per la storia del testo e dell’ esegesi biblica, Storia e letteratura raccolta di studi e testi 67 (Rome, 1958), 83-146, at 94-9. Much of Job, and extracts from Proverbs, Ecclesiastes and Song, survive in St Gallen, Stiftsbibliothek, 11 (before 781; CLA VII, no. 896); Fischer, LB, pp. 180-1. De Bruyne, ‘lin e nouvelle preface de la traduction hexaplaire de saint Jérðme’, RB 31 ( I 914-19), 229-36, identified a frequently occurring addition to a preface to Esther as ‘hexaplaric’. In translating from (he Hebrew, Jerome probably made use of the literal Greek translations »1 Aquila and Symmachus in Origen‘s Hexapla: Bogaert, ‘La Bible latine’, p. I ^8 unii n, 76.

7

The text o f the Old Testament in Anglo-Saxon England

Introduction

known to Tertullian (c. 130-230), but the earliest firm evidence comes in long passages cited by Cyprian of Carthage in the mid-third century, which narrowly predate the first European evidence.17 We should prob­ ably envisage the translations being made locally and piecemeal, as demand dictated, and it is unlikely that there was ever a single, received ‘Old Latin’ version.18 The picture is one of variation and confusion and Augustine was driven to complain that Latin translators were ‘out of all number’, compared with those who translated from Hebrew to Greek.19 Evidence for the preparation of complete Bibles is scanty, although Bonifatius Fischer has drawn attention to the intriguing report of the seizure of biblical codices, including one described as ‘very large’, during a raid by the imperial authorities on a north African sect in 3 03.20 Our most reliable record is the late-sixth-century account in the Institutiones of Cassiodorus of his own monastery of Vivarium, which he says possessed a nine-volume Old Latin Bible and also a large pandect, which he called the Codex grandior?1 This was probably the volume reported by Bede as having reached Wearmouth-Jarrow in 686,22 but it has not survived. The seventh-century Lyon Heptateuch’ is the only substantial extant Bible manuscript with the old text, although it has lost much of Genesis and Exodus.23 There is, however, a large number of fragments.24 Conscious of the uncontrolled proliferation of varying and often inac-

curate Latin texts and the need for a single, authoritative one, Pope Damasus commissioned the scholar Jerome, in 382, to make a revision of the gospels from the Greek. These were finished by 383 and Jerome seems to have followed them with a more cursory revision of some, at least, of the rest of the New Testament. Thereafter, between about 390 and 405, he worked on a revision of the books of the Old Testament.25 He had settled by now in Bethlehem and was able to make use of Origen’s compendious Hexapla Bible (available at Caesarea), and in particular the heavily anno­ tated Septuagint version in the fifth column of that work.26 Before long, however, Jerome became dissatisfied with the Greek and decided to begin the task anew, this time going direct to the Hebrew for his authority, although it is clear that he continued to refer also to both Greek and Old Latin texts. How far his initial ‘hexaplaric’ revision had extended is a matter of dispute.27 Certainly the book of Psalms was completed, for it became the received ‘Gallican’ psalter, and also Job and Song of Songs, which survive in large part, and Proverbs and Ecclesiastes, whose texts are partially recoverable, and a surviving prologue to Chronicles shows that this book should be included also.28 Whatever the case, Jerome thoroughly reworked all the Old Testament books,29 except those rejected from the Hebrew canon, the six so-called deuterocanonical books; he did, reluctantly and rapidly, make fairly free versions of Tobit and Judith,

Also useful are Billon, Old Latin Texts and Sparks, Latin Bible , pp. 100—10. On the classification of Old Latin texts, see Fischer in VL II, l4 * -2 2 * ; and Marsden, Interven­ tion’, p. 235 17 Bogaert, ‘La Bible latine’, pp. 143-4; Gribomont, 'Les plus anciennes’, pp. 47-51; Metzger, Early Versions, pp. 285—93; Kedar, ‘Latin Translations , pp. 299—300. 18 Ibid., pp. 300 - 1 . 19 Doctr. donst. II.xi: ‘Qui enim Scripturas ex hebraea in graecam uerterunt, numerari possunt, latini autem interpretes nullo modo’ (CCSL 32, 42). It is to Augustine that we owe the identification of one Old Latin version as the Ttala’, but the term is confusing and, although used by Jiilicher (above, n. 12), is usually avoided today; see Metzger, Early Versions, pp. 290—3Fischer, LB, p. 38; McGurk, ‘Oldest Manuscripts’, p. 1. Instit. I.v.2 and xiv.2. On the Bibles of Vivarium, see below, pp. 130-9. HA, ch. 15; below, p. 86 and n. 52. Lyon, Bibliothéque de la Ville, 403 + 1964 (VL I, no. 100; CLA VI, no. 771). See VL II 5*_6*; and Fischer, LB, p. 121. The codex is ed. Robert, Pentateuchi Versto (1881) and Heptateuchi Partis Posterioris Versio (1900), with extensive critical material. 24 The list in VL I, 11-34 includes about two hundred. For a comprehensive catalogue of manuscripts, presented book by book for both Testaments, see VLH I, 205-27.

2H 21 22 27

6

The literature on Jerome’s translations and revisions is extensive. Useful are Kedar, 'Latin Translations’, pp. 313-34; Metzger, Early Versions, pp. 330-4; Semple, ‘Biblical Translator ; Sparks, CHB 1, 517—26; and Condamin, ‘Les caractéres’. On Origen’s Hexapla, see Kenyon, Greek Bible, pp. 22-4, and Roberts, Old Testament, pp. 18-20. Origen used a system of asterisks and obeloi to indicate, respectively, elements of the Hebrew missing from the Septuagint version and additions in the latter not authorized by the Hebrew. Sparks, CHB 1, 515 and 531; A. Vaccari, ‘Recupero d’un lavoro critico di S. Girolamo’, in Scritti di erudizione e di filologia II: Per la storia del testo e dell’ esegesi biblica, Storia e letteratura raccolta di studi e testi 67 (Rome, 1958), 83-146, at 94-9. Much of Job, and extracts from Proverbs, Ecclesiastes and Song, survive in St Gallen, Stiftsbibliothek, 11 (before 781; CLA VII, no. 896); Fischer, LB, pp. 180-1. De Bruyne, ‘lin e nouvelle preface de la traduction hexaplaire de saint Jérðme’, RB 31 ( I 914-19), 229-36, identified a frequently occurring addition to a preface to Esther as ‘hexaplaric’. In translating from (he Hebrew, Jerome probably made use of the literal Greek translations »1 Aquila and Symmachus in Origen‘s Hexapla: Bogaert, ‘La Bible latine’, p. I ^8 unii n, 76.

7

Introduction

The text of the Old Testament in Anglo-Saxon England

but left untouched Wisdom, Sirach, Baruch and Maccabees.30 These were thus to reach what would eventually be known as ‘the Vulgate in the old textual forms.31 The habit of using ‘Vulgate’ and ‘Hieronymian’ as synonymous terms is therefore not historically accurate, although con­ venient.32 In today’s Vulgate, the Old Testament books revised from the Hebrew, along with Tobit, Judith and the gospels are Hieronymian; the other deuterocanonical books and the rest of the New Testament are not, but are based on revisions of old texts by others. Nevertheless, in this study I use the term ‘Hieronymian’ (abbreviated to Hier.) to designate all the Old Testament texts established in the volumes of the Rome Bihlia Sacra. In retrospect, the eventual success of the new texts seems to have been inevitable, not least because they were demonstrably more reliable than the old. The Council of Trent made them official in 1546 and the Bible used today by the Roman church is thus substantially Jerome’s. Yet the general acceptance of the Vulgate and its replacement of the old texts were by no means immediate.33 Jerome’s translations had appeared as indi­ vidual books or as small groups of books over a period of many years, and after the death of Damasus (in 384) they enjoyed no official endorsement. The old versions continued to be used and copied alongside the new, and inevitably there was interaction and mutual contamination. ‘Mixed’ texts became common, some involving merely the sporadic use of readings from one version in the other, others the interpolation of long passages. Sometimes whole books in the Old Latin tradition were included in Bibles which otherwise transmitted a Vulgate text.34 The Old Latin traditions were to assert their influence on Vulgate transmission throughout the Middle Ages and as late as the thirteenth century, not only on the

Continent but in the Insular area, too.35 Small-scale contamination of Vulgate texts occurred in a number of ways. At the simplest level, during the period before which the new version had become ‘canonized’, copyists probably retained readings which sounded familiar and clearer to them, especially in frequently cited texts,36 but Old Latin readings were some­ times added deliberately as marginal variants or glosses, which were then a prime source for subsequent variation in copying. The earliest surviving manuscripts of the gospels, which may date from Jerome's time, have such marginal variants. 37 A small group of Spanish manuscripts, including the tenth-century ‘Codex Gothicus’, preserves almost three thousand Old Latin readings from the Old Testament, 778 of them in the Heptateuch, in the form of marginal notes.38 The accessibility of good Vulgate texts for copying purposes could be an important factor in determining the quality of transmission from a particular scriptorium and would itself depend on the wealth of the monastery in question. The more impoverished houses might have to make do with fewer and poorer exemplars. In connection with gospelbooks in Ireland, it has been shown that the purer texts tend to be found at, and in the vicinity of, the larger houses.39 We cannot always be certain whether mixing was deliberate or acci­ dental. An interesting example from sixth-century Britain is the version of Malachi used by Gildas for citations in his De excidio Britanniae. It seems to have been copied from an Old Latin exemplar, in which a leaf had been lost and subsequently replaced by one which contained the appropriate text in the new version, but we can only guess whether this was the only text by now available to the supplier or whether it was chosen because it was considered a better text.40 A damaged Vulgate text, and the avail­ ability only of an old text for repairs, may explain the case of the northern Italian ‘Codex Ottobonianus’, an Octateuch copied around 700, in which

30 In his prologus to the three 'books of Solomon’, Jerome wrote that Wisdom and Sirach, like Judith, Tobit and Maccabees, might be read in the churches ad aedificationem plebis, although they were not canonical (BS XI, 3—5, at 5). 31 The use of the term can be traced back to the thirteenth century, but not before, according to Sparks, CUB 1, 518; cf. the sixteenth-century date given by Sutcliffe, CHB 2, 99, and see his ‘The Name “V ulgate” , Biblica 29 (1948), 345-52. 32 Cf. Bogaert, ‘La Bible latine', pp. 139 40 and Biblia Sacra, ed. Weber, p. xx. 33 Gribomont, ‘Leglise’, pp. 51-60, gives a useful account of the early spread of the Vulgate, and Fischer, LB, pp. 405-7, summarizes the processes of Old Latin survival. 34 For example, Salzburg, St Peter, Stiftsbibliothek, a. IX. 16 (Salzburg, before 798; siglum T]) has Song of Songs in an Old Latin form, and Paris, BN, 1 15()4 + 1 1505 (N. France, 822; siglum P) has Old Latin texts of Tobit and Judith. On P, sec below, p. 27.

35 Fischer, LB, pp. 418-20. I discuss evidence for the influence of Old Latin texts in England below, pp. 49-54. On Irish mixed texts of the seventh and eighth centuries, see Cordoliani, Trlande', pp. 13-15. 36 Gribomont, ‘La transmission’, p. 740; Chapman, ‘Families’, p. 29. 37 Fischer, Beitrdge, pp. 57—63. ix Ayuso Marazuela, ‘La Biblia Visigótica’, pp. 18-32; VLH I, 409-36. The ‘Codex Gothicus’ is León, Real Colegiata de San Isidoro, 2 (Castille, 960); see VLH I, 368 and Fischer, LB, pp. 72-3. The other manuscripts are of later date. Old Latin addition or substitution was especially prevalent among Spanish manuscripts, including the ‘Codex Toletanus'; see VLH I, 402—6, and below, p. 436.

8

')

D o y le , 'B ib le in Ire la n d ', p. C l.

111 B u r k i t t , ‘B ib le o f G ild a s ’, p p . 2 0 8 - 9 .

Introduction

The text of the Old Testament in Anglo-Saxon England

but left untouched Wisdom, Sirach, Baruch and Maccabees.30 These were thus to reach what would eventually be known as ‘the Vulgate in the old textual forms.31 The habit of using ‘Vulgate’ and ‘Hieronymian’ as synonymous terms is therefore not historically accurate, although con­ venient.32 In today’s Vulgate, the Old Testament books revised from the Hebrew, along with Tobit, Judith and the gospels are Hieronymian; the other deuterocanonical books and the rest of the New Testament are not, but are based on revisions of old texts by others. Nevertheless, in this study I use the term ‘Hieronymian’ (abbreviated to Hier.) to designate all the Old Testament texts established in the volumes of the Rome Bihlia Sacra. In retrospect, the eventual success of the new texts seems to have been inevitable, not least because they were demonstrably more reliable than the old. The Council of Trent made them official in 1546 and the Bible used today by the Roman church is thus substantially Jerome’s. Yet the general acceptance of the Vulgate and its replacement of the old texts were by no means immediate.33 Jerome’s translations had appeared as indi­ vidual books or as small groups of books over a period of many years, and after the death of Damasus (in 384) they enjoyed no official endorsement. The old versions continued to be used and copied alongside the new, and inevitably there was interaction and mutual contamination. ‘Mixed’ texts became common, some involving merely the sporadic use of readings from one version in the other, others the interpolation of long passages. Sometimes whole books in the Old Latin tradition were included in Bibles which otherwise transmitted a Vulgate text.34 The Old Latin traditions were to assert their influence on Vulgate transmission throughout the Middle Ages and as late as the thirteenth century, not only on the

Continent but in the Insular area, too.35 Small-scale contamination of Vulgate texts occurred in a number of ways. At the simplest level, during the period before which the new version had become ‘canonized’, copyists probably retained readings which sounded familiar and clearer to them, especially in frequently cited texts,36 but Old Latin readings were some­ times added deliberately as marginal variants or glosses, which were then a prime source for subsequent variation in copying. The earliest surviving manuscripts of the gospels, which may date from Jerome's time, have such marginal variants. 37 A small group of Spanish manuscripts, including the tenth-century ‘Codex Gothicus’, preserves almost three thousand Old Latin readings from the Old Testament, 778 of them in the Heptateuch, in the form of marginal notes.38 The accessibility of good Vulgate texts for copying purposes could be an important factor in determining the quality of transmission from a particular scriptorium and would itself depend on the wealth of the monastery in question. The more impoverished houses might have to make do with fewer and poorer exemplars. In connection with gospelbooks in Ireland, it has been shown that the purer texts tend to be found at, and in the vicinity of, the larger houses.39 We cannot always be certain whether mixing was deliberate or acci­ dental. An interesting example from sixth-century Britain is the version of Malachi used by Gildas for citations in his De excidio Britanniae. It seems to have been copied from an Old Latin exemplar, in which a leaf had been lost and subsequently replaced by one which contained the appropriate text in the new version, but we can only guess whether this was the only text by now available to the supplier or whether it was chosen because it was considered a better text.40 A damaged Vulgate text, and the avail­ ability only of an old text for repairs, may explain the case of the northern Italian ‘Codex Ottobonianus’, an Octateuch copied around 700, in which

30 In his prologus to the three 'books of Solomon’, Jerome wrote that Wisdom and Sirach, like Judith, Tobit and Maccabees, might be read in the churches ad aedificationem plebis, although they were not canonical (BS XI, 3—5, at 5). 31 The use of the term can be traced back to the thirteenth century, but not before, according to Sparks, CUB 1, 518; cf. the sixteenth-century date given by Sutcliffe, CHB 2, 99, and see his ‘The Name “V ulgate” , Biblica 29 (1948), 345-52. 32 Cf. Bogaert, ‘La Bible latine', pp. 139 40 and Biblia Sacra, ed. Weber, p. xx. 33 Gribomont, ‘Leglise’, pp. 51-60, gives a useful account of the early spread of the Vulgate, and Fischer, LB, pp. 405-7, summarizes the processes of Old Latin survival. 34 For example, Salzburg, St Peter, Stiftsbibliothek, a. IX. 16 (Salzburg, before 798; siglum T]) has Song of Songs in an Old Latin form, and Paris, BN, 1 15()4 + 1 1505 (N. France, 822; siglum P) has Old Latin texts of Tobit and Judith. On P, sec below, p. 27.

35 Fischer, LB, pp. 418-20. I discuss evidence for the influence of Old Latin texts in England below, pp. 49-54. On Irish mixed texts of the seventh and eighth centuries, see Cordoliani, Trlande', pp. 13-15. 36 Gribomont, ‘La transmission’, p. 740; Chapman, ‘Families’, p. 29. 37 Fischer, Beitrdge, pp. 57—63. ix Ayuso Marazuela, ‘La Biblia Visigótica’, pp. 18-32; VLH I, 409-36. The ‘Codex Gothicus’ is León, Real Colegiata de San Isidoro, 2 (Castille, 960); see VLH I, 368 and Fischer, LB, pp. 72-3. The other manuscripts are of later date. Old Latin addition or substitution was especially prevalent among Spanish manuscripts, including the ‘Codex Toletanus'; see VLH I, 402—6, and below, p. 436.

8

')

D o y le , 'B ib le in Ire la n d ', p. C l.

111 B u r k i t t , ‘B ib le o f G ild a s ’, p p . 2 0 8 - 9 .

*'r ^ e

text of the Old Testament in Anglo-Saxon England

Introduction

Old Latin material has replaced the Vulgate in over two hundred verses in Genesis and ExodUs.41 Even when there were no practical obstacles to the efficient spread of the Vulgate, the wi|fu[ conservatism of those in authority, or humble copyists, must not be underestimated. At one extreme there was Augustine, who never fully accepted Jerome’s new version, although he was himself quoting from the revised gospels by 404. He acknowledged the need for revision of the old texts but was scandalized by Jerome’s by-passing of both these and the Septuagint in favour of the Hebrew for his Old Testament translations.42*A t ^ otjier^ were the scribes in the monasteries, who reacted unfavourably against the new version and clung to long-cherished readings. T he very idea of the monastic life was born out of scripture, an its regulation and practice were closely bound up with specific citations from it.44 The liturgy, above all, suffused as it was with biblical ections and canticles and a natural conservator of their forms, was an important and often inadvertent source of old readings.45 This was a problem of w h ic h Bede was acutely aware in his work at Jarrow.46 The canticles, circulating as they did appended to the psalter, were particularly influential. A series of nine was common but their number, and their arrangement, Varied widely. They were known in both Vulgate and Old Latin forms, a n d the latter featured most notably in the series associated wit the R o m a n ’ psalter, which was dominant in England during most of the Anglo-Saxon period.47 Familiarity with the writings of the Latin

fathers will have been another important influence on some copyists and could lead easily to the incorporation, either from memory or by deliberate copying, of patristic versions of scriptural passages which were themselves based on the old texts.48 The overwhelming majority of the tens of thousands of biblical citations in Cyprian, Hilary, Rufinus (in his versions of Origen), Jerome, Ambrose, Gregory the Great and Augustine - to name only some of the most influential - are Old Latin in form.49 Sometimes, however, a writer would use the old alongside the new, with specific comments on the differences between the versions.50 To these works of commentary must be added florilegia, such as the Liber de diuinis scripturis (or Speculum), which, like the Speculum of Augustine (to whom it used to be attributed), consists of hundreds of biblical citations, some of which provide rare parallels for specific Vulgate variants.51 The example of Gregory the Great shows that the church fathers themselves could give respectability to the interpolation of old readings into Jerome’s new version. In a prefatory epistola to his Moralia in lob, he explains that he will expound the new version but will not hesitate to introduce the old, to justify his exposition {per testimonia)}2 At the humble level of the monastic scriptorium, in an era long before the scholastic ascendency, the urge to edify the faithful seems to have taken precedence over fidelity of copying, and good sense over precise details of expression.53

^ ^ lty " ^ 'k b o te c a Apostolica Vaticana, Ottob. lat. 66 (CLA I, no. 66; VL, no. Old L ee ^ 401-2; and Gribomont, ‘Les plus anciennes’, p. 55. The atm read i n g s from the codex are printed in Vanae lectiones, ed. Vercellone I, 183-4 10 ( s i g l Um py Ottobonianus is an important Vulgate witness in BS I-IV (siglum O). 42 Loewe, CHB 2 i l n c . Qc ,, _ .. > t 10; Sutcliffe, Jero m e, p. 96. Gribomont, ‘A u x origines’, p. 19 44 Biarne, 'La v i e ^ O n a stiq u e ’, p. 416 and passim. On the B ible a n d the liturgy, see Saxer, ‘Bible et liturgie’, esp. pp. 170-5; and on the iturgy as p r e s e r v e r of old texts, Salmon, 'Le texte biblique’, pp. 501 and 504-5; and Gribom ont, T ■’é ’ -g lise ’. For the manuscript evidence relating to Spain, see VLH I, 437-60.

The earliest surviving witnesses to the Vulgate are Italian. There is a gospelbook which dates from the first half of the fifth century, possibly

42 O nT hert; rf>e d e t ^le Scholar’, p. 49. e o n g i n anc| complex transmission of the canticles, Schneider, Cantica, is ispensable. g;C(;i a[SQ £)A(j L II, 2; Mearns, Canticles (useful, but unreliable in manuscript d a t i n g ) . Fischer, LB, pp. 414-15; and below, pp. 51-2. On the Roman psa ter in E n g l a n c j , see also pp. 28 and 69-70. 10

The pre-Carolingian Vulgate

tH Gildas again provides an early Insular example, for he seems to have quoted Old Latin versions from Lucifer in otherwise pure Vulgate passages; Burkitt, ‘Bible of Gildas’, p. 210. 4I' See VL I/1B for a list of the writers used by the scholars at Beuron; cf. the exhaustive list in VLH I, 229—312. McNally, Bible, pp. 95—104, gives a useful index of commentaries on each book of the Old Testament for the period 650-1000. On the Latin texts used by the fathers, see Gribomont, ‘Leglise', esp. pp. 51-3; Bogaert, ‘La Bible latine’, pp. 152-4; and VLH I, 147-51. , t 10; Sutcliffe, Jero m e, p. 96. Gribomont, ‘A u x origines’, p. 19 44 Biarne, 'La v i e ^ O n a stiq u e ’, p. 416 and passim. On the B ible a n d the liturgy, see Saxer, ‘Bible et liturgie’, esp. pp. 170-5; and on the iturgy as p r e s e r v e r of old texts, Salmon, 'Le texte biblique’, pp. 501 and 504-5; and Gribom ont, T ■’é ’ -g lise ’. For the manuscript evidence relating to Spain, see VLH I, 437-60.

The earliest surviving witnesses to the Vulgate are Italian. There is a gospelbook which dates from the first half of the fifth century, possibly

42 O nT hert; rf>e d e t ^le Scholar’, p. 49. e o n g i n anc| complex transmission of the canticles, Schneider, Cantica, is ispensable. g;C(;i a[SQ £)A(j L II, 2; Mearns, Canticles (useful, but unreliable in manuscript d a t i n g ) . Fischer, LB, pp. 414-15; and below, pp. 51-2. On the Roman psa ter in E n g l a n c j , see also pp. 28 and 69-70. 10

The pre-Carolingian Vulgate

tH Gildas again provides an early Insular example, for he seems to have quoted Old Latin versions from Lucifer in otherwise pure Vulgate passages; Burkitt, ‘Bible of Gildas’, p. 210. 4I' See VL I/1B for a list of the writers used by the scholars at Beuron; cf. the exhaustive list in VLH I, 229—312. McNally, Bible, pp. 95—104, gives a useful index of commentaries on each book of the Old Testament for the period 650-1000. On the Latin texts used by the fathers, see Gribomont, ‘Leglise', esp. pp. 51-3; Bogaert, ‘La Bible latine’, pp. 152-4; and VLH I, 147-51. ,v and A>p, both associated with Rheims, are represented among the eight Alcuinian manuscripts whose text is collated for at least some volumes of the Rome Biblia Sacra} 16 W hat may be the oldest surviving Alcuinian Bible (cÞ1), dating from the turn of the eighth century, contrasts starkly with the later examples, for it is an ill-ordered and calligraphically unimpressive book.117 Lowe associ­ ated it with an experimental stage of production, long before Saint-Martin achieved its special status. The Bible which Alcuin himself finished some time during 800 (as we know from his correspondence), for presentation at Charlemagne’s coronation, was probably a similar, inferior production.118 Only after Alcuin’s death, during the abbacies of Frithugils (807-34) and Adalhard (834—43), were the scribal standards and decorative distinction which we now associate with Alcuinian Bibles reached.119 Production continued under Abbot Vivian (843-51) and, despite disruptions, for another forty years or so.120 As a group, the Bibles shared a large format and a common style of presentation and were clearly designed, unlike the products of Orleans, for public use. In the complete examples, the number of pages actually varies between 420 and 450, according to whether or not

no pjere i follow Quentin, Memoire, pp. 290—3; cf. Berger, Histoire, p. 165. See also Fischer, LB, pp. 140—1; and Dahlhaus-Berg, Nona Antiquitas, p. 44. Where I cite marginal readings from the Theodulfian Bibles in this study, I expand al to alia. 111 Ibid., pp. 52 and 53, respectively. 112 See esp. Pseudo-Jerome, ed. Saltman, pp. 5-11; also E. Power, ‘Corrections from the Hebrew in the Theodulfian Manuscripts of the Vulgate’, Bihlia 5 (1924), 233—58; Berger, Histoire, pp. 179 and 181; Fischer, LB, pp. 9 5 -6 and 141, with n. 37; Psalterium, ed. Sainte-Marie, pp. xxxi—xxxii. M3 ‘p raeterea Ebrei . . . modernis temporibus in legis scientia non ignobiliter eruditi’ (MGH, Epp. select. V, 403). On Hebrew scholars in Carolingian France, esp. in rela­ tion to psalm glosses, see S. L. Keefer and D. R. Burrows, ‘Hebrew and the Hebraicum in Late Anglo-Saxon England’, ASE 19 (1990), 67-80, at 68-70. 114 Fischer, ‘Die Alkuin-Bibeln’, LB, pp. 203-403 and pis. 1-8, is the most complete textual study of the Tours Bibles, expanding his introduction to Die Bibel von Moutier-Grandval, British Museum Add. Ms. 10546 {facsimile}, ed. Verein Schweizerischer Lithographiebesitzer (Bern, 1971), pp. 49—98. His Alkuin-Bibel adds some detail. See also Berger, Histoire, pp. 185—258; E. K. Rand, Studies in the Script of Tours I. A Survey of the Manuscripts of Tours (Cambridge, MA, 1929) and II. The Earliest Book of Tours with Supplementary Descriptions of Other Manuscripts of Tours (Cambridge, MA, 1934). Quentin discusses the Bibles, as part of the ‘Amiatinus family’, in Mémoire, pp. 406—13. Ganz, ‘Mass Production’, surveys the practical aspects of Bible production at Tours and lists (pp. 61—2) surviving Bibles and fragments in assumed chronological order, with manuscript dimensions.

1n Fischer, LB, p. 2691u' h Propagators of rhe ‘textual standard’ view have included Saltman, Pseudo-Jerome, p. 4; Loewe, CHB 2, 139 and 146; Ganshof, ‘La révision’, p. 11 and ‘Charlemagne’, p. 280. The importance of the Alcuinian text is, in my view, over-emphasized, too, in the preface to Biblta Sacra, ed. Weber, p. xx. Most recently, Margaret Gibson, although accepting that the Alcuinian text was not ‘official’, yet insists that it was ‘both stable and widely diffused. In practice - and by default - it represents the standard text of the Carolingian Bible, outside Italy’ (Bible in Latin West, p. 6). Fischer’s statement, in an early essay, that Alcuin’s text became, at least in France, the ‘Normaitext’ (AlkuinBibel, p. 19) may have influenced some of these accounts. For his later, very different, view, see below, n. 14). , 'The Alcuinian (cxi in pariieular gained an easy entrance into the English convents’ (H is to r y , p. 0 8 ). Ml) See Ins d is i ussum ul t he 'W in i lle s le l ( io s p e ls ', /b i l l . , p p . 1 ) 0 —8.

The text of the Old Testament in Anglo-Saxon England

Introduction

cannot be discussed here.131 Whatever the case, it should now be accepted that the idea of the triumph of the Alcuinian text is a chimaera, even though post-Carolingian Bibles may have been influenced to a consider­ able extent by the usage at Tours in their ancillary features (such as the choice and ordering of books and systems of chapter division). For the Vulgate in England, this independence from Alcuin’s text is confirmed in chs. 9—11 of this study, in which I analyse the texts of manuscripts copied in the tenth and eleventh centuries —following, that is, the great influx of continental ideas and manuscripts which accompanied the monastic reform movement. It is instructive to examine the influence of the other prominent Carolingian textual tradition, the Theodulfian. The orthodox view has been that it was small, owing to overwhelming competition from Alcuin’s text. Berger plausibly sought a further explanation in the complexity of Theodulf’s recension, which was too ‘riche de science’, an idea echoed in Chapman’s perception of its ‘artificial’ nature.132 This reminds us of one of the difficulties involved in establishing ‘Theodulfian’ traces in later manu­ scripts —that is, the scholarly and evolving nature of the recension. Yet the Theodulfian text does have, in most books, a distinctive overall character and the evidence for its influence beyond the diocese of Orleans is in fact impressive. Not only did manuscripts originating unmistakably in the tradition of Orleans spread widely, but their influence persisted for a considerable time, even as late as the thirteenth century, when they affected in part the text of the wisdom books in the Paris Bibles.133 Among earlier examples, a Pentateuch from St Gallen, written partly by the scribe W inithar, has an eighth-century text of mixed origin, with corrections made in the early part of the ninth century, partly from one of

the younger Theodulfian Bibles.134 As late as the eleventh century, revisions were being made to a three-volume Bible at Saint-Hubert directly from the text of the Theodulfian Bible then in that monastery’s possession (@H).135 W hat is particularly significant is that the monastic centres at which Theodulfian influence is evident are frequently those where the earlier, or even contemporary, use of Alcuinian texts can be seen also. Two Bibles associated with Corbie in the first half of the ninth century provide important evidence in this context. The one (P) was written in 822 and has an original text showing strong Alcuinian influence (the first such influ­ ence recorded outside Tours), but there are frequent alterations by a contemporary corrector, made without doubt from a Theodulfian exemp­ lar.136 The other (E) is a two-volume Bible written between 830 and 850, which shows a strong Alcuinian element with an overlay of Theodulfian readings in the text of the Octateuch, Kings and Psalms.137 Significantly, the text of the latter book is in Jerome’s ‘Hebrew’ version, the choice of Theodulf for his Bibles, not the Gallican, which was used at Tours. This shows that even the common identification of the Gallican version as a specifically ‘Alcuinian’ legacy to the later Vulgate must be treated with caution. Its dominant position was in fact not assured before the early thirteenth century, and even then was not universal. In Italy, Jerome’s earliest version of Psalms, the ‘Roman’, which is in all the Cassinian (II) and other manuscripts of the tenth to the twelfth centuries, continued to circulate widely until the sixteenth century and is still in use at St Peter’s in Rome.138 Jerome’s last version, iuxta Hebraeos, which had been chosen for the Codex Amiatinus at the turn of the seventh century and was used in all the Theodulfian Bibles, was adopted (presumably under the

151 See the comments of McKitterick, ‘Tours Anomaly’, p. 65; also Light, ‘French Bibles’; H. Deniflé, Die Handschriften der Bibel-Correctorien des 13. Jahrhunderts, Archiv fiir Literatur- und Kirchengeschichte des Mittelalters 4 (Freiburg, 1888), 263-311 and 471—601; and J. P. P. Martin, ‘Le texte parisien de la Vulgate latine’, /VIuston 8 (1889), 444-66, and 9 (1890), 55-70 and 301-16. 1,2 Berger, Histoire, p. 146; Chapman, ‘Families’, p. 398. Berger clearly regarded Theodulfian influence as very limited, compared with Alcuinian (Histoire, p. 184). 133 Fischer, LB, p. 51. Other Theodulfian influences are noted ibid., pp. 146-50; Berger, Histoire, pp. 177-8; and McKitterick, ‘Tours Anomaly, p. 68. On the formal influence of Theodulfian Bibles, see the New Testament example cited by Berger, Histoire, pp. 181-4; and Fischer, LB, p. 48 and n. 50.

IB

1.1 St Gallen, Stiftsbibliothek, 2, pp. 3-300, used for Numbers and Deuteronomy only (siglum S); CLA VII, no. 893a-b; Fischer, LB, p. 181; BS III, x; and Berger, Histoire, p. 120. 1,3 Fischer, LB, p. 146. One of the three volumes is lost. The others are Brussels, Bibliothéque Royale, II. 1639, and Namur, Musée Archéologique, Fonds de la Ville 4. 11.1 Paris, BN, lat. 11504 + 11505 (originally a single volume). See Fischer, LB, pp. 45 and 149—50; Berger, Histoire, pp. 93—6; and Quentin, Memoire, pp. 395—9 (siglum Geo). P is of great interest in relation to the textual history of the late Anglo-Saxon Bible, London, BL, Royal 1. E. VII + VIII, the subject of ch. 10. Iw Paris, BN, lat. I 1532 i 1 1533, the so-called ‘Corbie Bible’; Fischer, LB, pp. 154—5; Berger, llnloire, pp. 1()4 8; and Quentin, Memoire, pp. 402-3 (siglum Corb). 1' H W e b e r, t.e joa n U e r roiihim, p. tx.

27

The text of the Old Testament in Anglo-Saxon England

Introduction

cannot be discussed here.131 Whatever the case, it should now be accepted that the idea of the triumph of the Alcuinian text is a chimaera, even though post-Carolingian Bibles may have been influenced to a consider­ able extent by the usage at Tours in their ancillary features (such as the choice and ordering of books and systems of chapter division). For the Vulgate in England, this independence from Alcuin’s text is confirmed in chs. 9—11 of this study, in which I analyse the texts of manuscripts copied in the tenth and eleventh centuries —following, that is, the great influx of continental ideas and manuscripts which accompanied the monastic reform movement. It is instructive to examine the influence of the other prominent Carolingian textual tradition, the Theodulfian. The orthodox view has been that it was small, owing to overwhelming competition from Alcuin’s text. Berger plausibly sought a further explanation in the complexity of Theodulf’s recension, which was too ‘riche de science’, an idea echoed in Chapman’s perception of its ‘artificial’ nature.132 This reminds us of one of the difficulties involved in establishing ‘Theodulfian’ traces in later manu­ scripts —that is, the scholarly and evolving nature of the recension. Yet the Theodulfian text does have, in most books, a distinctive overall character and the evidence for its influence beyond the diocese of Orleans is in fact impressive. Not only did manuscripts originating unmistakably in the tradition of Orleans spread widely, but their influence persisted for a considerable time, even as late as the thirteenth century, when they affected in part the text of the wisdom books in the Paris Bibles.133 Among earlier examples, a Pentateuch from St Gallen, written partly by the scribe W inithar, has an eighth-century text of mixed origin, with corrections made in the early part of the ninth century, partly from one of

the younger Theodulfian Bibles.134 As late as the eleventh century, revisions were being made to a three-volume Bible at Saint-Hubert directly from the text of the Theodulfian Bible then in that monastery’s possession (@H).135 W hat is particularly significant is that the monastic centres at which Theodulfian influence is evident are frequently those where the earlier, or even contemporary, use of Alcuinian texts can be seen also. Two Bibles associated with Corbie in the first half of the ninth century provide important evidence in this context. The one (P) was written in 822 and has an original text showing strong Alcuinian influence (the first such influ­ ence recorded outside Tours), but there are frequent alterations by a contemporary corrector, made without doubt from a Theodulfian exemp­ lar.136 The other (E) is a two-volume Bible written between 830 and 850, which shows a strong Alcuinian element with an overlay of Theodulfian readings in the text of the Octateuch, Kings and Psalms.137 Significantly, the text of the latter book is in Jerome’s ‘Hebrew’ version, the choice of Theodulf for his Bibles, not the Gallican, which was used at Tours. This shows that even the common identification of the Gallican version as a specifically ‘Alcuinian’ legacy to the later Vulgate must be treated with caution. Its dominant position was in fact not assured before the early thirteenth century, and even then was not universal. In Italy, Jerome’s earliest version of Psalms, the ‘Roman’, which is in all the Cassinian (II) and other manuscripts of the tenth to the twelfth centuries, continued to circulate widely until the sixteenth century and is still in use at St Peter’s in Rome.138 Jerome’s last version, iuxta Hebraeos, which had been chosen for the Codex Amiatinus at the turn of the seventh century and was used in all the Theodulfian Bibles, was adopted (presumably under the

151 See the comments of McKitterick, ‘Tours Anomaly’, p. 65; also Light, ‘French Bibles’; H. Deniflé, Die Handschriften der Bibel-Correctorien des 13. Jahrhunderts, Archiv fiir Literatur- und Kirchengeschichte des Mittelalters 4 (Freiburg, 1888), 263-311 and 471—601; and J. P. P. Martin, ‘Le texte parisien de la Vulgate latine’, /VIuston 8 (1889), 444-66, and 9 (1890), 55-70 and 301-16. 1,2 Berger, Histoire, p. 146; Chapman, ‘Families’, p. 398. Berger clearly regarded Theodulfian influence as very limited, compared with Alcuinian (Histoire, p. 184). 133 Fischer, LB, p. 51. Other Theodulfian influences are noted ibid., pp. 146-50; Berger, Histoire, pp. 177-8; and McKitterick, ‘Tours Anomaly, p. 68. On the formal influence of Theodulfian Bibles, see the New Testament example cited by Berger, Histoire, pp. 181-4; and Fischer, LB, p. 48 and n. 50.

IB

1.1 St Gallen, Stiftsbibliothek, 2, pp. 3-300, used for Numbers and Deuteronomy only (siglum S); CLA VII, no. 893a-b; Fischer, LB, p. 181; BS III, x; and Berger, Histoire, p. 120. 1,3 Fischer, LB, p. 146. One of the three volumes is lost. The others are Brussels, Bibliothéque Royale, II. 1639, and Namur, Musée Archéologique, Fonds de la Ville 4. 11.1 Paris, BN, lat. 11504 + 11505 (originally a single volume). See Fischer, LB, pp. 45 and 149—50; Berger, Histoire, pp. 93—6; and Quentin, Memoire, pp. 395—9 (siglum Geo). P is of great interest in relation to the textual history of the late Anglo-Saxon Bible, London, BL, Royal 1. E. VII + VIII, the subject of ch. 10. Iw Paris, BN, lat. I 1532 i 1 1533, the so-called ‘Corbie Bible’; Fischer, LB, pp. 154—5; Berger, llnloire, pp. 1()4 8; and Quentin, Memoire, pp. 402-3 (siglum Corb). 1' H W e b e r, t.e joa n U e r roiihim, p. tx.

27

The text o f the Old Testament in Anglo-Saxon England

Introduction

influence of the latter) in several ninth-century Carolingian Bibles, including, as we have seen, E, as well as in at least two Spanish Bibles (AL and XM). A Mozarabic version was used in Madrid, Biblioteca de la Universidad Central, 31 (X), and the Codex Cavensis (C) carried both this and the Hebrew version. One of the Paris Bibles had both the Hebrew and the Gallican Psalms, presented in parallel columns.139 In England, the Gallican version was not adopted until late in the tenth century, and the earliest surviving Anglo-Saxon witness is the ‘Salisbury Psalter’ (c. 975), but the Roman version was still being copied in the middle of the follow­ ing century.140 The most telling evidence of Theodulfian influence comes from the two late Alcuinian Bibles associated with Rheims, v and . xxxv: ‘Sic et adhuc deprecor, ut augeas quod cepisti, id est, ut mihi cum auro ( onscribas epistolas domini mei sancti Petri apostoli . . . ’ (MGH, Epp. select. I, 60) ' ' lift. Ixiii. See above, p. 1. 1,1 (if. die translation by Talbot, Missionaries, p. 1 18. ‘ ( )n lianu h, see below, p. ^ 0 . See above, pp. 37—8.

The text o f the Old Testament in Anglo-Saxon England

Introduction

surviving volume from St Gallen, which contains just such a mixture, although there are only four books in all: Isaiah, Hosea, Zachariah and Daniel.227 It is certainly not unreasonable, therefore, to speculate that, when Boniface cited from Isaiah, Ezekiel, Zephaniah and Habakkuk in his letter to Cuthbert of Canterbury, he was indeed using the book he had requested from England.228 However, Habakkuk may not necessarily have been in this codex, for Boniface appears to have taken his citations of this prophet from a canticle version.229 The lack of Anglo-Saxon biblical witnesses written in the second half of the ninth century is not surprising, for there is a dearth of manuscripts of any kind from English scriptoria during this period, and the abysmal palaeographic and linguistic standards revealed in contemporary Latin charters suggest that book production had effectively ceased.230 By now, anyway, a reasonable number of part-Bibles may already have been in circulation. More curious is the lack of Old Testament evidence for so much of the tenth century, in the form either of imported manuscripts or of newly written ones, for it is difficult to believe that the period of the monastic reform and expansion from mid-century onwards did not create a significant new demand for biblical manuscripts of all kinds and that this was not met, by whatever means and however inadequately. For the earlier part of the century, we do know that King Athelstan, who succeeded in 924, presented gospelbooks to Christ Church and St Augustine’s, Canter­ bury, and to the community of St Cuthbert’s, Chester-le-Street.231 These, and most of the other books associated with the king, were brought from the Continent or Ireland. There is evidence that further gospelbooks were imported during the succeeding decades of monastic reform, when contact with continental houses continued, and at least one was written in England.232 The first trace of the Old Testament is the fragment of Minor Prophets (Columbia, Missouri, University of Missouri-Columbia Libraries, Fragmenta Manuscripta 4), which was apparently copied during the second half of the century. It was part of a complete Bible, and could be more or less contemporary with a second such Bible (London, BL, Royal 1. E. VII

+ VIII), which survives almost complete and was copied towards the end of the century. Booklists, however, clarify our understanding of Old Testament use in some measure, although they involve only books extant at the very end of the Anglo-Saxon period. One list, made between 1069 and 1072, probably relates to the books procured for the church of Exeter by Bishop Leofric, after his arrival there in 1050. It includes, among some fifty books, a volume of four prophets (presumably the Major Prophets) and separate volumes of Isaiah, Ezekiel, Song of Songs and Maccabees.233 Another, made in about 1070, which may describe the personal library of Sæwold, abbot of Bath, at the time of the Conquest, includes a Heptateuch among its thirty-three items.234 Such a book, or at least a Hexateuch, was considered an appropriate form in which to circulate vernacular versions of scripture in the late Anglo-Saxon period.235 W hat we cannot know is how many of this total of six listed part-Bibles (four of them consisting of single books) were imported, and how many had been produced in English scriptoria. There remains, nevertheless, the curious imbalance in the record of biblical manuscripts shown in the above tables. They suggest that nothing but large-format manuscripts (presumably complete Bibles) was produced during the period of the Benedictine reform and the closing decades of the Anglo-Saxon period. This is probably an aberration, which I consider further in my conclusion to ch. 11. The survival of a fragment of Lamentations in a liturgical version in Columbia, Missouri, University of Missouri-Columbia Libraries, Fragmenta Manuscripta 1, may be a clue to how at least some of the demand for the Old Testament in late Anglo-Saxon England was met, other than in part-Bibles. The prescribed scriptural readings for the Night Office were mostly from the Old Testament and the fragment, probably copied in the mid-tenth century, may have belonged in the sort of lectionary or breviary for which we have

227 229 2,11 2.1 2.2

St Gallen, Stiftsbibliothek, 41 (s. ix/x). 228 See above, p. 1. See below, pp. 69—70. Brooks, Early History, pp. 164—74, esp. 171—4. Cf. Morrish, ‘Manuscripts’. Keynes, ‘Athelstan's Books’; and below, pp. 321-3. Boulogne, Bibliothéque Municipale, 10 (Gneuss, no. 79S; below, p. 322). 4 6

Lapidge, ‘Booklists’, pp. 64—9. Cf. P. W. Conner, Anglo-Saxon Exeter: A Tenth-Century Cultural History (Woodbridge, 1993), pp. 226—9, who argues against the assumption i hat the books were all additions made by Leofric. M| Lapidgc, ‘Booklists’, p. 39; ptd with commentary, P. Grierson, ‘Les livres de l’Abbé Sciwold tic Bath’, RE 32 (1940), 96-116. See below, pp. 102-6. I have noticed a dearth of independent Heptateuchs or (k iaieu ih s from the ninth and tenth centuries in the catalogues of the continental libraries. 4 7

The text o f the Old Testament in Anglo-Saxon England

Introduction

surviving volume from St Gallen, which contains just such a mixture, although there are only four books in all: Isaiah, Hosea, Zachariah and Daniel.227 It is certainly not unreasonable, therefore, to speculate that, when Boniface cited from Isaiah, Ezekiel, Zephaniah and Habakkuk in his letter to Cuthbert of Canterbury, he was indeed using the book he had requested from England.228 However, Habakkuk may not necessarily have been in this codex, for Boniface appears to have taken his citations of this prophet from a canticle version.229 The lack of Anglo-Saxon biblical witnesses written in the second half of the ninth century is not surprising, for there is a dearth of manuscripts of any kind from English scriptoria during this period, and the abysmal palaeographic and linguistic standards revealed in contemporary Latin charters suggest that book production had effectively ceased.230 By now, anyway, a reasonable number of part-Bibles may already have been in circulation. More curious is the lack of Old Testament evidence for so much of the tenth century, in the form either of imported manuscripts or of newly written ones, for it is difficult to believe that the period of the monastic reform and expansion from mid-century onwards did not create a significant new demand for biblical manuscripts of all kinds and that this was not met, by whatever means and however inadequately. For the earlier part of the century, we do know that King Athelstan, who succeeded in 924, presented gospelbooks to Christ Church and St Augustine’s, Canter­ bury, and to the community of St Cuthbert’s, Chester-le-Street.231 These, and most of the other books associated with the king, were brought from the Continent or Ireland. There is evidence that further gospelbooks were imported during the succeeding decades of monastic reform, when contact with continental houses continued, and at least one was written in England.232 The first trace of the Old Testament is the fragment of Minor Prophets (Columbia, Missouri, University of Missouri-Columbia Libraries, Fragmenta Manuscripta 4), which was apparently copied during the second half of the century. It was part of a complete Bible, and could be more or less contemporary with a second such Bible (London, BL, Royal 1. E. VII

+ VIII), which survives almost complete and was copied towards the end of the century. Booklists, however, clarify our understanding of Old Testament use in some measure, although they involve only books extant at the very end of the Anglo-Saxon period. One list, made between 1069 and 1072, probably relates to the books procured for the church of Exeter by Bishop Leofric, after his arrival there in 1050. It includes, among some fifty books, a volume of four prophets (presumably the Major Prophets) and separate volumes of Isaiah, Ezekiel, Song of Songs and Maccabees.233 Another, made in about 1070, which may describe the personal library of Sæwold, abbot of Bath, at the time of the Conquest, includes a Heptateuch among its thirty-three items.234 Such a book, or at least a Hexateuch, was considered an appropriate form in which to circulate vernacular versions of scripture in the late Anglo-Saxon period.235 W hat we cannot know is how many of this total of six listed part-Bibles (four of them consisting of single books) were imported, and how many had been produced in English scriptoria. There remains, nevertheless, the curious imbalance in the record of biblical manuscripts shown in the above tables. They suggest that nothing but large-format manuscripts (presumably complete Bibles) was produced during the period of the Benedictine reform and the closing decades of the Anglo-Saxon period. This is probably an aberration, which I consider further in my conclusion to ch. 11. The survival of a fragment of Lamentations in a liturgical version in Columbia, Missouri, University of Missouri-Columbia Libraries, Fragmenta Manuscripta 1, may be a clue to how at least some of the demand for the Old Testament in late Anglo-Saxon England was met, other than in part-Bibles. The prescribed scriptural readings for the Night Office were mostly from the Old Testament and the fragment, probably copied in the mid-tenth century, may have belonged in the sort of lectionary or breviary for which we have

227 229 2,11 2.1 2.2

St Gallen, Stiftsbibliothek, 41 (s. ix/x). 228 See above, p. 1. See below, pp. 69—70. Brooks, Early History, pp. 164—74, esp. 171—4. Cf. Morrish, ‘Manuscripts’. Keynes, ‘Athelstan's Books’; and below, pp. 321-3. Boulogne, Bibliothéque Municipale, 10 (Gneuss, no. 79S; below, p. 322). 4 6

Lapidge, ‘Booklists’, pp. 64—9. Cf. P. W. Conner, Anglo-Saxon Exeter: A Tenth-Century Cultural History (Woodbridge, 1993), pp. 226—9, who argues against the assumption i hat the books were all additions made by Leofric. M| Lapidgc, ‘Booklists’, p. 39; ptd with commentary, P. Grierson, ‘Les livres de l’Abbé Sciwold tic Bath’, RE 32 (1940), 96-116. See below, pp. 102-6. I have noticed a dearth of independent Heptateuchs or (k iaieu ih s from the ninth and tenth centuries in the catalogues of the continental libraries. 4 7

The text o f the Old Testament in Anglo-Saxon England

Introduction

no other evidence but which has been postulated by Helmut Gneuss.236 The Old Testament circulated also in manuscripts which were not Bibles or part-Bibles, but compilations of miscellaneous items for exegetical, devotional or pedagogical use. There are two examples of complete Old Testament books in such contexts, one from the late Anglo-Saxon period (Proverbs), the other from eighth-century Northumbria (Job). Another manuscript, from the early eleventh century, has a large extract from Sirach. I give their texts equal status with those of biblical manu­ scripts in this study, and list them here:

Staatsbibliothek, Misc. Patr. 17 (B. II. 10), fols. 133—61, is of particular importance, because of the amount of citation which it contains, most of which has not previously been analysed.238 A similar, but anonymous, work transmitted large parts of the Old Testament in Anglo-Saxon England in the form of moral and legal precepts taken from Exodus, Leviticus, Numbers and Deuteronomy. It was known as the Liber ex lege Moysis and was of Irish origin.239 Although I have ignored its text in this study, because the two main manuscripts were written on the Continent, it should be noted as the kind of source which might have affected the copying of Old Testament texts by those familiar with it. The manuscripts are Cambridge, Corpus Christi College 279, 55v-80r (Tours?, s. ix2), which was at Worcester,240 and London, BL, Cotton Otho E. xiii (s. xln), which was at St Augustine’s, Canterbury, but was later badly damaged in the Cotton fire.241 I have examined the text of the Cambridge manuscript and found it to be eccentric and full of errors.

St Petersburg, Public Library, F. v. I. 3, fols. 1-38 s. viii2, uncial; complete Job, with contemporary interlinear commentary in set minuscule (CLA VI, no. 1599; D) London, BL, Cotton Vespasian D. vi, fols. 2—37 s. xmed, Anglo-Saxon minuscule; complete Proverbs, accompanying texts on vice and virtue, with interlinear Old English gloss (Gneuss, no. 189) London, BL, Royal 7. C. IV, 100v-102r s. xi1, Caroline minuscule; Sir. XXV. 17-XXVI.28, appended to Defensor’s Liber scintillarum, with interlinear Old English gloss (Gneuss, no. 470) A further example of a complete Old Testament book in a non-biblical manuscript, which is of immense importance in the history of the Insular Vulgate text, is Cornish, not Anglo-Saxon. It is a copy of Tobit, included in a manuscript which is now part of Oxford, Bodleian Library, Bodley 572, and probably written in the early years of the tenth century. W ith remarkable fidelity, it transmits the idiosyncratic text of Tobit used at Wearmouth-Jarrow two hundred years earlier for the Codex Amiatinus and Ceolfrith’s other two pandects.237 I devote two chapters of this study to a further source of information about the text of the Old Testament, scriptural citation by the authors of works of exegesis, history, hagiography or devotion. These scholars and writers include Boniface and Felix among the Southumbrians (ch. 2) and Bede and Alcuin among the Northumbrians (ch. 6). Alcuin’s work, a florilegium entitled De laude Dei, which I have collated from Bamberg, 236 Voigts, ‘Fragment’, esp. pp. 87-9; Gneuss, ‘Liturgical Books’, pp. 110-12 and 122. Voigts’s identification of the script of the fragment as Anglo-Caroline is disputed by Dumville, 'On the Dating’, pp. 44 -4 , who believes it is continental, and probably Breton, and that the book involved was imported into Fngland in the tenth century. 237 Gneuss, no. 583- See my ‘Survival’, and the summary below, pp. 179—81. 4 8

Old Latin influence The pre-Hieronymian, Old Latin, textual traditions played little part in the known history of the Old Testament in Anglo-Saxon England. Insular evidence for the existence of Old Latin or mixed texts is most apparent for Ireland, which was an important agent for the transmission of old texts from the fifth to the seventh centuries.242 After the arrival of the Vulgate in Ireland in the late sixth or early seventh century, mixed texts were established alongside the old and were not displaced, at least in the gospels (to which the bulk of the surviving evidence relates), until the ninth century.243 Apart from citations in church writers, which show that an old version of Genesis was in use until the twelfth century, evidence for the Old Testament in Ireland comes from only three biblical fragments.244 One of these is a palimpsest with an Old Latin text of Daniel, proving that 23K See below, pp. 222-35. 239 On the use and transmission of the Liber, and on two further manuscripts, see Kottje, ‘Dcr Liber ; also Fournier, ‘Le liber . 210 James, Corpus Christi II, 42—4; Ker, Libraries, p. 206; Kottje, ‘Der Liber, p. 62. 211 Thomas Smith, Catalogus librorum manuscriptorum bibliothecae Cottonianae (Oxford, 1696), p. 79; Ker, Libraries, p. 43; Kottje, ‘Der Liber, p. 62. 212 Kenney, Sonnes, p. 625. 211 Doyle, ‘Bible in Ireland’, pp. 34—7. 2 11 Mc Na ma r a , l.ai in I hhl e ’, pp

s i 9.

Iollil. Prm. mrg. xxiv (ibid., p. 256). /IOil II, 313.

6 6 6 7

The text of the Old Testament in Anglo-Saxon England

Early Southumhrian scholars and writers

excellent Vulgate texts, too. In the citations from these in his works, there is no sign of any textual link with the Vulgate in use at Theodore’s school in Canterbury.

letter written at some time between 742 and 746, to send him.41 The letter to Cuthbert contains several short citations from Isaiah, longer extracts from Ezekiel and two from the Minor Prophets, Habakkuk and Zephaniah. Whatever its origins, Boniface uses in general an excellent, accurately transmitted Vulgate text. From Ezekiel, for example, he cites most of XXXIV.2—5 and 9—10, with only one or two minor omissions (such as eis from XXXIV.4 and pastores from XXXIV.10).42 Ezek. III.18, however, receives a different treatment:

TH E LETTERS OF BO NIFAC E

From 716 until his death at the hands of pagans in Frisia at the age of nearly eighty, Wynfrith, a native of Wessex who received the name Bonifatius from Pope Gregory II in 719, worked tirelessly to promote Christianity among the tribes of Germany.37 A body of correspondence relating to this period, collected soon after Boniface’s death and containing a hundred and fifty items, is extant in three manuscripts from the eighth or ninth centuries.38 Of greatest interest in the context of Vulgate history are nearly forty letters which Boniface himself wrote from the Continent, either to friends and acquaintances back in England or to the pope at Rome, and another thirty written to Boniface (or, after his death, about him) from England. Many of the letters contain scriptural allusions and some have verbatim citations from Old Testament books. Although Boniface repeatedly asked for books of both scripture and commentary to be sent to him in Germany from England, we cannot be certain that all the texts which he used, and to which his letters give intermittent access, were of English provenance. He made three visits to Rome and may have brought volumes of scripture back from there.39 It is, however, tempting to believe that the book of prophets from which he cited in a long letter (no. lxxviii), written in 747 to Archbishop Cuthbert of Canterbury,40 was the one which had belonged to Abbot Winbert at Nursling and which Boniface had asked the bishop of Winchester, in a

37 The Vita Bonifatii by Willibald, written soon after Boniface's death, is ed. Levison in Vitae Sancti Bonifatii, MGH, SS rer. Ger. LVII, 1-57, and trans. Talbot, Missionaries, pp. 25—62. See also Stenton, Anglo-Saxon England, pp. 167—74; Levison, England and Continent, pp. 70—93; and Greatest Englishman, ed. Reuter. 38 The letters are ed. Michael Tangl in MGH, Epp. select. I, with a discussion of the manuscripts, pp. vi—xiv, and a scriptural index to the letters, pp. 315—17. 39 In 719, he was made a missionary by Gregory II in Rome and, in 722, was consecrated bishop; and during his third and last visit, in 737—8, Gregory III made him papal legate for Germany. 40 MGH, Epp. select. I, 161-70. AL) declinauit] inclinauit (MZ20 m his XXII.7* supputur > supputetur XXII.8 elegit > et legit This list may be compared with the following twelve corrections in the equivalent text in A, among which it will be noted that four (curruum, silete, silete, Moyses) are shared with b:

III Kgs XI.40 XXII. 10 XXII. 24 XXII. 32 IV Kgs II. 3 II. 5 11.25 XVII. 7 XXI. 8 XXII.4

fugit > fuit prohetebant > prophetebant me ergo > mene ergo currum > curruum Dominus hodie > hodie Dominus silite > silete silite > silete in montem inde > inde in montem factum enim est > factum est enim uerbum facerent hoc > facerent uerbum hoc Moysi > Moyses pecuniam > pecunia

In four of these cases (IV Kgs II.3, 11.25, X V II.7 and XVII. 12 ) A originally had a unique word order, presumably as a result of scribal error, which was subsequently corrected.240 There are also four cases of erasure (in III Kgs X X II.50 and IV Kgs 1.13, 11.23 and IX.17) where it may be assumed that there was copying error, but the original letters are not now visible.241 In IV Kgs IX.17, for example, ego was originally followed by several letters which were subsequently erased and are unrecoverable. Several different hands were involved in the corrections in b, though there is no need to doubt that all were made at W earmouth-Jarrow, where the pandect probably remained until the last quarter of the eighth century. Two corrections w ithin a few lines of each other, of elegit (III Kgs XXII. 8) 240 In II.3 and XVII.7, where there was transposition of two words, the correction is

T h e C e o lfr ith ia n te x t

and supputur (X X II.7), are in two different hands, to judge by the forms of t used. Some of the correcting was probably done by the copyist himself. An example is the alteration of Moysi to Moyses, where the original uncial i was made into a capital E w ith added horizontal strokes, the topmost sharing exactly the angled fork characteristic of the main script. The added s is w ithout a fork and is written smaller, but similar versions of s are found at line-ends elsewhere in the manuscript. The two corrections of silite to silete may also be by the copyist. The fact that the corrections of both silite and Moysi had to be made also in A indicates that the errors were in the exemplar. There is no evidence that the two pandects shared a corrector, although it is in the nature of most corrections to offer very little material for comparison and most correctors show great freedom in their letter forms. I do not agree with the Benedictine editors that the correction of fecerunt (IV Kgs XVII. 11) by means of q', added above, was made by the copyist.242 The letter form is quite unlike the scribe’s usual q, and the comma-shaped sign for om itted ue differs radically from the three-part zigzag sign he normally uses. It is possible that this correction was made by the corrector who added u to hie (IX.17), u and a suspension mark to corrumeorum (XIX.23) and h to is (X X II.5). The corrector who added p to prohetas (III. 13) and i to videntum (XVII. 13) may have been the same.

Emendations in b There are six further emendations in b’s text of Kings which alter, or add to, a text which was not obviously incorrect as it stood. None is paralleled in A:

Kgs habitu > habitus turbinem + quasi turbinem + quasi usque aquas + et non sunt diuisae 11.14 XVII. 13 uidentum > uidentium XIX.27 introitum om.

1.7

II.l 11.11

indicated by short lines and dots above each word. The other two cases involve three words. In 11.25, in montem inde made poor sense, and has been corrected by means of two single lines with dots (in montem being apparently treated as one word). In X VII.12, the correct order is indicated by a single line above what should be the first word (facerent), two lines above the second and three above the third. 241 In IV Kgs 1.13, the erasure after genua may be of sua, an addition which occurs in E.

In the case of the first alteration, habitu was the established older reading, occurring in RD, as well as A, and has been selected by the Benedictine

196

197

MJ /IS VI, 291.

T h e t e x t o f th e O l d T e s ta m e n t i n A n g l o - S a x o n E n g l a n d

editors as Hieronymian, but it gave way to habitus in almost all ninthcentury and later manuscripts.243 In X V II.23, the form uickntum, also judged to be Hieronymian, is recorded only in ACIIFS T . In X IX .27, introitum is used by the manuscripts of all periods, except RXA and, before correction, one Paris Bible, but has not been judged to be Hieronymian. The manuscripts which include it show much variation in the wording of the colon in question. Only bA have the rather awkward version et egressum tuum et introitum uiam tuum (cf. Dp XT*IIL.54 The variant prope is used in all other manuscripts, including A and f2, the fifth-century earliest witness to the Vulgate Daniel.55 It is unsafe to try to deduce the textual history of the part-Bible from which the Pepys fragment comes on such limited evi­ dence, but the indications are that we could expect a text originating in a good exemplar in the Italo-Insular tradition, uninfluenced by the new Carolingian revisions.

siglum Dj ,57 The pages of the codex measure 325 X 220 mm, with a text area of 250 X 175 mm, and the script is a Northumbrian uncial with a squarish aspect and distinctive knobbed finials on h and m. The long lines were copied with extra space between them to accommodate an interlinear gloss, and this was written, intermittently, in a neat set minuscule script with prominent ligatures. It shares with the main script the Insular abbreviations for autem, enim and est but has also a large number of other abbreviations and suspensions. The interlinear gloss consists mainly of extracts taken from the Commentarii in librum lob by Philippus presbyter,58 with some also from Gregory’s Moralia in lob.59 The text and gloss of the St Petersburg manuscript were reproduced by J. Martianay as an appendix to an edition of the works of Jerome made at the end of the seventeenth century, and this is the main source for the collation in the Rome Biblia Sacra.60 The St Petersburg text of Job is remarkable for its use of a large number of variants paralleled neither among the collated Vulgate manuscripts nor in known patristic or liturgical sources. Characteristic are:

4

T H E ST P E T E R S B U R G JO B

In 1805, the former Imperial Library in St Petersburg acquired a codex of one hundred and eight folios containing a complete book of Job (fols. 1-38) and an abbreviation of Jerome’s commentary on Isaiah (fols. 39-108). Both were copied in the second half of the eighth century, probably in the same scriptorium, the first in uncial, the second in Anglo-Saxon set minuscule. The early history of the codex, which is now St Petersburg, Public Library, F. v. I. 3, is obscure, but its origin was probably Northumbrian, although Lowe left open the possibility that it had been written by Northumbrian scribes in a continental centre.56 The codex belonged at one time to Corbie and was taken to Saint-Germain-desPrés in the seventeenth century. Job was clearly part of the composite codex from the start, accompanying the commentary on Isaiah, and thus cannot strictly speaking be classified as a part-Bible, but this does not lessen its value as a rare witness to the text of Job circulating in early Anglo-Saxon England. It was collated for the Rome Biblia Sacra with the 54 On propter, see BS XVI, xl. 55 For f2, see above, p. 12 and n. 55. 56 CLA IX, nos. 1599 (fols. 1—38) and 1600 (fols. 39—108); A. Staerk, Les manuscrits latins du V‘ au X lir siecle conserves a la Bibliotheque Imperiale ck Saint-Pe'tersbourg, 2 vols. (St Petersburg, 1910) I, 34-5 and pi. IX and II, pi. XXVIII; Lowe, English Uncial, p. 23 and pi. XXXVII.

256

II.

7

III. 13 VI. 30 IX.24 X IV .18 XVI. 12 XVIII. 15 XX. 3 X X II. 9 X XIII. 6 XXIV. 10 X X X II. 7 X XX III. 19 X XX VII. 20 X LI.l

,7 ” 60 61

uerticem] ceruicem enim + lob lingua mea] ore meo est quis ergo est] quis ergo esset cadens defluet] cadet defluens impiorum] iniquorum tabernaculo] habitatione61 mihi] tibi comminuisti] commouisti fortitudine] formidine62 incedentibus + aliis prolixior] proueetior marcescere] marcere quis] subito quis non] numquid non

BS IX, viii-ix; Fischer, LB, pp. 151-2. 58 PL 26, 619—802. CCSL 143, 143A and 143B. Sancti Hieronymi opera II, Appendix, cois. 69—114; and see BS IX, ix. Cf. ©AKlS12 habitaculo. 'W ith much fear’ is not appropriate in the context; cf. Hier, nolo fortitudine contendat mecum.

257

The text of the Old Testament in Anglo-Saxon England

A few of D j’s otherwise unique variants, including a notable concentration of them in ch. 1, appear to originate in Septuagint readings. Such are: 1.5 1.6 1.8 1.15 XIX. 21

enim + lob filii} angeli seruum] puerum gladio} in gladio amici} o amici

The first three occur also in the hexaplaric versions of Job transmitted in St Gallen, Stiftsbibliothek, 11 (s. viii2), Oxford, Bodleian Library, Auct. E. inf. 1 (s. xii) and Tours, Bibliothéque Municipale, 18 (s. xi).63 D : shares many other variants with these manuscripts, a few of them also exclus­ ively, such as iratus for indignatus in XXXII.3 and dico for dicam in XXXII. 10. Most, however, occur in Vulgate manuscripts, including stultis + mulieribus (II. 10; A2D,)

Other correspondences between D[ and the lemmata in the commentary involve minor variation, such as the addition or omission of the conjunc­ tion et, and they are often found in other manuscripts. Furthermore, there are many distinctive readings throughout the commentary, which are not in D, and not often in any known Vulgate manuscript. The most likely explanation for the apparent, but comparatively small, connection between D, and Philippus is that the latter used a text which had some relation with an ancestor of the former. It is possible, however, that the commentary itself influenced the text of D lf in respect of one or two unique amplifications. In XX.2, D! adds eo quod haec mala poenarum innocentem asseris te sustinere to et mens in diuersa rapitur. The relevant part of the commentary, in the printed version, has: ideo iniquit aestuo et cogitatione conturbor quia docis te haec mala sine causa a Deo judice sustinere, cum de Deo nihil sinistrum debeat suspicari.66

This is somewhat altered in its extracted version in Dt , written above the text of XX.2: ideo iniquit cogitationibus conturbor quia dicis te haec mala sine causa a Deo justo judice sustinere.67

The commentary could certainly have provided the substance of the addition in D1; although little of the actual wording, except mala and sustinere. The editors of the Rome Biblia Sacra draw attention also to D j s uerumtamen congnoscite quia mihi est cor sicut for et mihi est cor sicut, in XII.3.68 The version of the commentary printed, however, has nothing specific to say on the colon in question. The interlinear gloss in Dj has ita ego noui sicut et uos, which bears no relation to the textual amplification. The exact nature of the relationship of D fs text to the commentary of Philippus remains to be established. There is no textual connection between Dj and the Codex Amiatinus, whose text of Job, though containing errors, is nearer the mainstream M' PL 26, 668. 67 Sancti Hieronymi opera II, Appendix, col. 88. 6H Thus Martianay, ibid., col. 80. BS IX prints cognoscitis.

259

The text o f the Old Testament in Anglo-Saxon England

Part-Bibles of the eighth and ninth centuries

Hieronymian tradition.69 Our only other source for the text of Job in Northumbria in the early Anglo-Saxon period is Alcuin’s florilegium, De laude Dei, which has a column of extracts from chs. I, II, VII, IX, X and XIX.70 The most interesting reading is in 1.21, where the florilegium shares with D, the amplification sit nomen Domini benedictum + in saecula. This parallels the Greek but has been noted in its Latin version only in a commentary on Psalms by Hilary.71 There are two other correspondences between Di and Alcuin’s text. One, also in 1.21, is the absence of a six-word addition found in all the Vulgate manuscripts, except for a small group which includes not only D, but also A. The other is a minor variant, aufers for auferes, which is shared with several manuscripts (but not A). Against these, however, must be set a number of differences between the texts of Di and the florilegium, as in II. 10, where Alcuin has the rare sustineamus for suscipiamus. There are other readings in which either Alcuin or, more often, D t is unique. The addition of in saecula in 1.21 is not enough to confirm the direct effect of D, on the florilegium. It may reflect some currently well-known liturgical form of the text. A clue to the history of the St Petersburg text of Job is provided by a fragment of that book, perhaps from a part-Bible, which was written in an Irish majuscule script in the eighth century and survives as the back fly-leaf in a tenth-century commentary on the Minor Prophets. This belonged formerly to the monastery of Saint-Martin, on the Moselle. The fragment, written presumably in Ireland, is now numbered fol. 172 in Ghent, Universiteitsbibliotheek, 2 54 (445).72 It consists of one leaf and a narrow strip of the conjoint leaf, which carries only a few letters. The written area of the single leaf is 240 X 160 mm, and its full width about 190 mm. The text is Job XXXIII.24-XXXIV.22, which is given the siglum f in the Benedictine edition, and it shares three variants with D 1; two of them exclusively:

These are positive correspondences, but there are also many differences between f and D1; even in this short section of text. Most are due to the poor copying standards in f. Clear errors include si non om. (XXXIII.33), uirolenta for uiolenta (XXXIV.6), concordati for cordati (XXXIV. 10) and spiritus for spiritum (XXXIV. 14). Among other unique f readings, which may also be errors but which offer more or less acceptable alternatives, are addet for reddet (XXXIII.26), loquaris + tu (XXXIII.32), iniquis for impiis (XXXIV.8), cucurrit for cucurrerit (XXXIV.9) and redigetur for reuertetur (XXXIV. 15). In a few cases, it is D which appears to be the deviant manuscript, as in sapientiam + meam (XXXIII.33) and aduersus for contra (XXXIV. 19). On the evidence of the three shared variants, a textual relationship between f and D] may be accepted as very likely. If it was constant throughout the lost bulk of f, a plausible explanation for it would be that the exemplar for Dh or its immediate precursor, had reached Northumbria from Ireland, where it had shared ancestry with f. Such an origin, suggesting an association with the earliest biblical manuscripts to reach the Insular area, would help to explain the strong traces of Old Latin influence and the many variants which, in relation to known texts, are unique.74 74 On the early history of the Latin Bible in Ireland, see above, pp. 14 and 49-50.

XXXIII. 27 homines! homo ()73 XXXIV. 18 duces] iudices XXXIV. 19 eius] suarum See above, p. 170. 70 See above, pp. 228—9. 71 Tract, psalm. CXVIII (CSEL 22, 384). 72 CLA X, no. 1557; BS IX, ix; Fischer, LB, p. 152. 73 Hier, respiciet homines et dicet peccaui (where ‘men’ is the object of respiciet)', cf. LXX sita tots áitopÉ(U|/ETai avðpam oc; cidtck; tames: ‘even then a man shall blame himself’.

260

261

The Egerton codex

8 The Egerton codex

DESCRIPTION

The codex, preserved now in an unattractive Victorian binding, consists of forty-eight trim m ed and gilt-edged leaves, measuring c. 310X225 mm, with a tex t area throughout of c. 260 X 200 mm.4 The present first folio is greatly discoloured and the text on the recto is hardly legible in places, but most of th e other leaves are well preserved. The codex comprises the following m aterial, which falls into three distinct sections: I

In 1834, the London bookseller William Pickering offered for sale fifty manuscripts and prints, along with over four thousand printed books. Among the manuscripts, as item no. 2 in the catalogue, was a Latin codex containing the five Old Testament wisdom books. This was, declared Pickering, ‘of great importance to the Biblical Scholar from its high antiquity and . . . a desideratum in any public or private library’. The ‘fragmenta biblica’ had been written in the eighth century in characters ‘very similar to the Saxon, and . . . usually designated Semi-Uncial’.1 Such was the obvious attraction of the codex that excellent facsimiles of two small portions of it, complete with yellow and pink shading of some letters, were used as a frontispiece to the catalogue.2 It was the only item in the sale without a list price but was offered to the British Museum, in June 1834, for £105.3 Not until 11 November 1843, however, according to a note at the beginning of the codex, did the British Museum buy it from Pickering. The money came from the funds given to the Museum, along with his own collection of manuscripts, by Francis Henry Egerton, ninth Earl of Bridgewater, and the codex became Egerton 1046. Pickering did not exaggerate its importance, but it presents considerable problems of interpretation, in respect both of its calligraphy and codicology and of its eccentric text. 1 Catalogue of Biblical Classical and Historical Manuscripts and of Rare and Curious Books .. . on Sale by William Pickering . . . (London, 1834), p. 1. 2 The facsimile, by J. Netherclift, reproduces the opening four lines of both Sirach (30v) and Ecclesiastes (lOv). 3 According to a pencilled note added to the item in the copy of the catalogue owned by A. N. L. Munby, and now in the Cambridge University Library (Munby d. 45), it was offered ‘to J. P .’. Item 1 in the catalogue was a ‘Biblia Sacra Hebraea’, in four manuscript volumes, on sale at £120.

2 62

II III

lr - 1 0 v lOv—l4 v I4 v —I6v 17r—30v 30v—31v 3 2 r-4 8 v

Prov. IV.9 gratiarum —end Ecclesiastes (complete) Song of Songs (complete) Wisdom (complete) Sir. prologue and 1.1—35 mansuetudo Sir. IV. 10 in iudicando —XLIV. 13 semen eorum

The present fols. 7 and 8 became transposed at some time before their num bering, so that two large sections of the Proverbs text (XXVI.2 quia rapinas—X X V I.26 in concilio and XXI.10 anima impii—XXIV. I cum eis) are now out o f order. Because of the continuity of material and the complemen­ tary nature of the components of the codex, Lowe noted that they ‘apparently were kept together from the beginning’.5 If he meant to imply that those components had never been used independently, he was cer­ tainly w rong. Two manuscripts of very different character, intended originally, or at least used, for different purposes, were brought together to make the present codex. The one (section II in the above scheme) was p u t between two parts of the other (I and III). They may have been w ritten at different times, although both during the eighth-century. The m anuscript which comprises sections I and III —that is, fols. 1—16, with the three books of Solomon, and fols. 32-48 with Sir. IV. 10— XLIV. 13 — is written in a variable Insular minuscule. One scribe was probably responsible, although the variation makes assessment difficult. At its m ost careful, the writing has a certain formality but sometimes (as 4 CLA II, nos. 194a and 194b; Gneuss, no. 410; Catalogue of Additions [1843 section], p. 103; Ancient Manuscripts, ed. Thompson II, 13-14 and pi. 26 (and see p. 41 for partial co llatio n o f th e te x t of Proverbs and Ecclesiastes with Add. 10546 and 24142); Brown, 'N ew F ra g m e n t’, pp. 41—2. See pi. VI. ’ CLA II, nos. 194a and 194b. Cf. the implicitly opposite view of the British Museum catalo g u er, T hom pson (Ancient Manuscripts II, 13).

263

The text o f the Old Testament in Anglo-Saxon England

The Egerton codex

at the bottom of the first column on 7v) it degenerates into a distinctly untidy form; fewer pen-lifts give it an almost cursive aspect, and there is much irrational spacing and word-division. Yet the script remains a set minuscule, for even where the words on a line are squeezed, joining of letters does not occur. An ‘underslung’ / is much used and there is some evidence of looping, rather than wedging, on ascenders. Uncial n is more or less consistently used, but minuscule forms of s and r are usually preferred. Abbreviations are few and are mainly restricted to the nomina sacra, -bus, -que and final -m, but quo for quoniam may also be noted. The irregularity of the script and the inconsistency of letter forms identify it as Phase-I set minuscule and indicate that it was written before the middle of the eighth century.6 Lowe saw a calligraphic resemblance to some lines on 62v of Cassiodorus’s commentary on Psalms in Durham, Cathedral Library, B. II. 30.7 The Egerton text is written per cola et commata, and there are in fact more divisions than in any of the other manuscripts of the wisdom books for which information is available, including Amiatinus. The extent of the division is not immediately apparent, however, for an effort has been made to fit each period into one line, so that the empty part­ lines which are so characteristic a feature of the more formal Bible manu­ scripts divided per cola et commata are largely absent. Sometimes this has resulted in the squeezing of words and it accounts for the uneven look of the right-hand side of each column. Where a colon does extend onto a second line, however, the text here is indented. Only occasionally does the scribe err in not thus indicating a follow-on; an example occurs on 9r, where a break seems to be indicated between tisanas and feriente in quasi tisanas feriente desuper (Prov. XXVII.22), owing to the scribe’s (or his exemplar’s) failure to indent feriente. The set-minuscule leaves are apparently in gatherings of eight, but the present binding prevents close examination of their codicological relation­ ships. Two quires thus make up the first section, and the quire signature II at the bottom of 9r confirms this. However, it should be noted that the missing head of Proverbs (prologue and I—IV.9) would have required another bifolium before our ‘quire I’ (the first page of which, lr, carries no signature). The leaves of bifolia within the surviving quires are no longer always conjoint, as the transposition of fols. 7 and 8 proves. There is no

decoration in the manuscript, but simple ornamental initials in black are used intermittently, some of them quite large, and red ink is used for some colophons. Capitulum numbers occur at divisions in Proverbs and Ecclesi­ astes, and Song and Sirach have within their texts the tituli which are characteristic of their respective Vulgate versions. An immediately obvious aspect of the set minuscule portion of Egerton 1046 is the extent to which it has been corrected or otherwise altered. No page is without such alterations, sometimes as many as twelve or fifteen of them. They are in a confusing variety of hands and they seem to testify to the manuscript’s use over an extended period of time; I discuss them further below. At the bottom of the second column on lOr, the beginning of a lection (Prov. XXXI. 1, Verba Lamuel) is marked in the left margin by a cross, made up of five dots, as found in a number of other manuscripts associated with Northumbria.8 There is no prefatory matter before Ecclesiastes or Song. A Hieronymian preface to Proverbs is characteristic of Vulgate manuscripts and may have been in the missing head of this codex.9 The manuscript which forms section II of Egerton 1046 has fifteen leaves (fols. 17-31) and contains the whole text of Wisdom and the opening of Sirach, as far as 1.35, where mansuetudo completes a colon but does not form a natural break in the text. With its larger script, cleaner presentation and more generous use of space, it contrasts dramatically with the companion manuscript. The script is a rather compressed hybrid minuscule, probably of Phase II, for there is a general consistency in the letter-forms, despite frequent odd changes of aspect. The date is therefore likely to be the mid-eighth century or later. Lowe (and Pickering) described the script as ‘half-uncial’, which pays due regard to the elevated character of the writing but does not take account of the dominance of minuscule forms and the use of a slanting pen. The key letters d, n and r are almost invariably minuscule, and a usually has the distinctive Insular majuscule ‘oc’ form, though at the beginning of words an uncial form with a notably pointed bow may be written. Usually, s is uncial but m very rarely so. Descenders, especially those of q and g, are often very long, and this emphasizes the tall, compressed aspect of the script. Lowe again

6 If the manuscript was in fact produced at a centre in the south of England, the same calligraphic features would permit a slightly later date. 7 CLA II, no. 152; Gneuss, no. 237.

K Lowe, CLA II, no. 194a, noted five-point crosses in the gospel manuscripts Durham, Cathedral Library, A. II. 17, London, BL, Royal 1. B. VII and Wurzburg, Universitatsbibliothek, M. p. th. f. 68. 9 A prologue to Wisdom appears in a few Vulgate manuscripts. The Codex Amiatinus has capitula before Proverbs, Ecclesiastes, Wisdom and Sirach.

264

265

The text o f the Old Testament in Anglo-Saxon England

The Egerton codex

invoked Cassiodorus’s commentary on Psalms, in Durham B. II. 30, noting ‘some kinship’ with the distinctive type of hand mainly used there. This in turn had ‘a distinct similarity’ to one of the hands used in the eighth-century Northumbrian gospelbook, Cambridge, University Library, Kk. 1. 24.10 Lowe believed that the Egerton leaves and the Durham Cassiodorus could have been produced in the same Northumbrian centre. The differences, however, cannot be ignored. Unlike the scribe of Egerton, for instance, the scribes of the Durham book nearly always used minuscule r in final position (it occurs seldom in Egerton); they often used uncial R, and when they used the minuscule form they kept it on the line. Ascenders and descenders are shorter and compression is not so pro­ nounced as in Egerton. In this latter respect, the Cambridge gospelbook’s script is in fact rather closer to Egerton than to that of the Cassiodorus, although again there are differences of detail. The text has no decoration, except for yellow shading to most of the large letters, which are used regularly to start new lines at divisions of the text. This division is carefully and spaciously made, per cola et commata. There is no consistent punctuation, but occasionally a medial punctus, apparently made by the original scribe, ends a colon. The pages are otherwise almost without additions or alterations of any sort, but on 19v a small cross in the inner margin marks the start of Wisd. V.16, lusti autem in perpetuum. The lack of alterations, including corrections, produces one of the most obvious contrasts with the Egerton codex’s set minuscule manu­ script. It is not a consequence of there being no errors. On the contrary, the hybrid minuscule section has almost as many, proportionally, as the other sections; the greater care taken with the script and presentation has not extended to producing a more accurate text. The lack of corrections indicates that the higher-grade manuscript was never read critically, perhaps never read much at all, and this seems to be good evidence that it led, at least at first, a life independent of the other sections of the codex. To some extent, colon division in the manuscript follows a pattern characteristic of the Italo-Insular tradition, as witnessed in G2A and also CM, but often it divides uniquely, and sometimes eccentrically, especially in the later chapters. Sirach is preceded by a short prologue, which is common to almost all Vulgate manuscripts and here is followed without a

break by the main text. This is interspersed with tituli, as in most Vulgate versions of Sirach. The two manuscripts which comprise Egerton 1046 have been prepared similarly, with ruling made after folding and guided by prickings in both inner and outer margins. Slight differences reflect the needs of the two styles of presentation. In the set minuscule pages, there are double bounding lines on the outer sides of each column but single on the insides, with an extra line down the middle of the intercolumnar space.11 In the hybrid minuscule pages, the extra line is further to the right, providing the right-hand column with a double inner boundary. The space between the lines is used to guide the writing of the capital letters, often coloured, with which the cola begin; run-on lines start at the inner boundary. The double boundary lines to the left of the left-hand column are used in a similar way.12 The intercolumnar space is a little wider in the later manuscript and the maximum ruled width of the columns narrower. This results in rather longer lines in the set minuscule portions, but the average overall text area in both cases is similar, at about 260 X 200 mm. Although there is much variation in the quality of the parchment throughout the codex, that of the hybrid minuscule manuscript tends to be thicker and better, and is little discoloured. The continuity of the material in the composite codex is more apparent than real. If it ever contained the complete text of the five wisdom books, and this must be seriously doubted, several losses have occurred. At least two leaves are missing from the head of the codex. The text of Prov. I-IV.10, written in set minuscule, would not have filled both of these but would have needed more than one, and no doubt prefatory matter, including the Hieronymian prologue to Proverbs, took up the rest of the space. Four leaves would have been required to complete Sirach at the end of the codex (XLIV.13—LII.13). In set minuscule, the text would have filled probably seven of these eight pages. The third lacuna, between Sir. 1.35 on 31v and IV. 10 on 32r, is less easy to evaluate. If the codex originally contained the complete text of the book, and the hybrid minuscule continued throughout the missing portion, four leaves would

10

CLA II, no. 138; Gneuss, no. 21. Lowe was referring to 186v—187v, containing the prologue and capitula for John, where the hand differs considerably from the hands in the rest of the manuscript.

266

11

12

At the outer edges of the columns, the innermost of the two lines has apparently not always been drawn. On the preparation of the manuscripts, see the detailed descriptions in CLA, no. 194a—b. I disagree with Lowe’s statement that double boundary lines are provided on the inside l both columns of the hybrid section.

267

The text o f the Old Testament in Anglo-Saxon England

The Egerton codex

have been needed. If it were all in set minuscule, however, the missing text could probably have been accommodated on two. A change in script at some intermediate point is possible. Whatever the case, a complete codex would have needed at least fifty-six leaves, and perhaps fifty-eight. The text of Song finishes two-thirds of the way down 16v and is followed by a blank space; the text of Wisdom (in hybrid minuscule) then starts on a fresh leaf. No such gap has been left in the case of either Proverbs or Ecclesiastes, although neither finishes at the end of a page. Ecclesiastes starts half way down the second column of lOv, following immediately the end of Proverbs, and Song begins after the end of Ecclesiastes with an incipit at the very bottom of the first column of I4v. However, the gap after Song does not prove that the codex was planned as a mixed-script part-Bible from the start. It may be explained more simply by the fact that I6v appears to complete a quire. Whatever script was to be used then for the next book, practical considerations could have led to the starting of a new quire for Wisdom. It is most unlikely that a composite codex was planned from the start. If we were to assign the set minuscule manuscript to the very end of Phase I, and the hybrid minuscule manuscript to the beginning of Phase II, the two might be contemporary, dating from around the middle of the eighth century, but a gap of a generation or so looks probable. The set-minuscule manuscript is likely to have been part of a complete wisdom codex, written during the first half of the eighth century, much used and much corrected. At some stage, its middle section may have become lost or irreparably damaged and the gap may have been filled, at least in part, by the supply of text, to order, in a different script. Perhaps the process was not completed, or several leaves (containing Sir. 1.35 and IY.10) were later lost.13 It is also plausible, however, to suggest that, at some time after the writing of both manuscripts, Egerton 1046 was put together from such incomplete or damaged materials as were available. These may have included parts of a hybrid-minuscule manuscript of the two deuterocanonical books, Wisdom and Sirach, which had perhaps been begun with some special purpose in mind but had never been completed. Perhaps it was the set-minuscule manuscript which was deliberately cannibalized, in order to

expand the higher-grade manuscript into a wisdom codex. Such cannabalization is likely to have occurred at a period when a Northumbrian scriptorium was reaching the end of its active life, with production of new manuscripts at a standstill, perhaps during the ninth century. The process may have taken place, however, much later in the Anglo-Saxon period; indeed, I see no reason why we should insist that it took place within this period at all. Perhaps we owe the present form of the codex to the activities of a later antiquary. The textual complexities of Egerton 1046 (hereafter L3) echo those of its codicological history and remain to a great extent unresolved. Below, I deal with each book in turn, giving first an assessment of the character of the underlying text and then a brief account of corrections and emendations, although I make no attempt to deal with these comprehen­ sively. As far as the base-text is concerned, I pay particular attention to L3’s relationship, or lack of it, with A (the Codex Amiatinus), for this witnesses a text already established in Northumbria before L3 was written. In theory, it or manuscripts derived from it were a potential source for L3’s copyists. In examining wider textual links, I make persistent reference to the manuscripts discussed in ch. 5 in connection with Amiatinus and loosely grouped under the heading ‘Italo-Insular’.14 Two of them (k and D2) are older than both L3 and A but are incomplete. All of the others, each containing all five wisdom books (sometimes with other Old Testa­ ment books also), probably post-date L3, although none was copied later than the ninth century.

13

The loss of the opening and closing leaves of the codex, and of several within it, is likely to have occurred during its long (and unknown) post-Conquest history. Pickering’s description of the manuscript as ‘fragmenta biblica’ may indicate that it consisted of loose leaves at the time of its sale.

268

The correction process Most of the many corrections and emendations which characterize the Egerton codex occur in the set minuscule manuscript. The immense problems of classifying and dating the hands involved, and establishing the sequence in which the various alterations were made, demand separate and expert study. The difficulties arise from the confused mixture of correction, recorrection and emendation, the various scribes who seem to have been responsible (although this impression may be deceptive) and the sheer quantity of the alterations of all sorts made. The copyist himself certainly made a number of the more obvious M They are (in approximate chronological order) kD 2 T 1 WS 2 YQ. See above, p. 153, n. for details, including a note on Z (Metz, Bibliothéque Municipale, 7; s. viii“ ).

269

6 8

,

The text o f the Old Testament in Anglo-Saxon England

The Egerton codex

corrections throughout the codex.15 There may then have been a phase of correction from the exemplar soon after the copying was finished, but the overall impression given is that many of the alterations were made haphazardly over an extended period. It is impossible to separate clearly a ‘correcting’ phase from an ‘emending’ phase, and to relate each to specific scribes. Some basic corrections were made in a small hand somewhat reminiscent of the sort o f‘capitular’ uncial used to correct Amiatinus, with a sloping tendency and a characteristic open a.16 Corrections in a similar style, but using an ink which has become red-brown in colour, occur occasionally.17 A few changes were made in a hand with majuscule characteristics, using a very fine pen and an ink which appears (now, at least) much blacker than that of the surrounding text.18 Many corrections and additions were made in a large, often untidy minuscule with cursive tendencies, using an ink which is now pale and a pen which was defective, producing broad strokes imperfectly formed in the middle.19 Most of the longer additions are in this hand and we might assume them to be part of a later reworking of the text from a different exemplar, and yet the hand is

hardly distinguishable from the hand of the main text, where this is at its more informal. There are further corrections and emendations in hands different from those already noted, such as the addition of uiro (Prov. XXVI.5), in a very small and wiry script, and of uero (Prov. XXII.1), in an equally wiry but larger script, apparently written by an untrained scribe. The progressive nature of the alteration process in the codex is well illustrated in Eccl. 1.8, from which two consecutive words had been omitted by the copyist: cunctae res difficiles non potest {eas homo} explicare sermone. One corrector added homo above the line (in a small, neat set minuscule) and a second supplied eas in the left-hand margin (in a larger, untidy script). Probably homo came first. Both omissions are signalled by an oblique stroke with a dot above and the signes de renvoi sit one above the other between potest and est. The stroke with dot, used both at the point of omission and alongside the added word or phrase, is the usual way of signalling omissions throughout the manuscript. One interesting excep­ tion occurs in Proverbs, where the omission and subsequent restoration (at the bottom of the column) of the whole of XXIX.20 is signalled by vertical s-shaped signs with dots or wedges either side. The most common way of marking letters for deletion is by means of dots, placed beneath. In at least one instance, in Sir. IV. 11, the process appears to have been modified, for the three dots used to correct magister to magis have not only been placed above the word but highlight gis, rather than the offending ter. Perhaps the intention in this case was to indicate the syllable which was to be retained.20 In Sir. V.4, the phrase in corde et ui has been deleted by means of a point on the line before in, three dots above both corde and tui and a medial point after ui. The reason for the use of three dots in the codex is not always apparent, as in the cases of Sir. VIII.3—4, where they are written before 3b and 4b, and Sir. XI.3, where they are written before et initium. It may be that, in Sir. VIII.3-4, transposition of the cola was intended, although this would be quite inappropriate. Transpositions are usually marked in the codex with the diagonal line with dot above. This sign was also used consistently by the scribe in Sirach, when providing the text of section headings, ready for copying in red ink into gaps in the text.21

15

16

17

18

19

In Song, for instance, he seems to have inserted e above langu{e}o (II. 5) and / above f{l}ores (11.12) and, in Sirach, he restored omitted splendidae (XXIX.29) and simul (X XX II.12) and added ten above auer{ten}tem VIII.6 ). The correction in 11.12 is not noted in BS XI, nor that in VIII . 6 in BS XII. The critical apparatus in the latter volume also contains some errors: in Sir. XVIII.25, it is indicated that die was supplied after in as an emendation, but in fact the scribe wrote the word originally and the emendation is its erasure; in XXVII.30, it is recorded that the scribe wrote, and then deleted, unde, but in fact he first wrote ad unde, and it is ad which he (or a subsequent corrector) marked with dots for deletion. Some other errors in BS are noted below, but I have made no compre­ hensive check. These include the addition of quod in Prov. XII. 17 and ti above intimi{ti}s in Sir. XIII. 15 (for inmitis; the erroneous ti before m has been marked for omission with dots beneath each letter). Probably the same scribe made good the omission of the first colon of Sir. XXI.25 (six words) and two cola in Sir. XXII. 1-2 (twelve words; in this case, lost by homoeoteleuton). Examples are the change of nequitiae in to nequitia et (Sir. X IX.20), made by means of dots beneath e and in and the addition of et above in and, in the same verse, the adding of the second colon, the whole of which had originally been omitted. I judge that all six words were omitted, not only the last three, as the editors of BS XI record. Examples include qui abscondit frumenta (Prov. XI.26, on an erasure) and doctrinam (XIX.27, above disciplinam). Examples in Proverbs are the additions of occasionem (IX.9; signalled by an oblique stroke with a single dot), eleven words (XI. 16) and spem (XXIV. 14) and, in Ecclesiastes, dolorem (1.18), arbitratus (II.10) and et adpropinqua ut audias (IV.17).

2 70

20

21

I have assumed, charitably, that the corrector did not intend to delete gis and produce mater. However, in a similar case in IX. 13, where alteris ceris was written for alterceris, it is quite clearly the letters with dots above them (that is, the first is) which are to be deleted. See below, p. 293.

271

The Egerton codex

The text o f the Old Testament in Anglo-Saxon England PROVERBS

Proverbs begins at IV.9, opening the first set-minuscule section of the codex. It is the most intensely corrected or emended book. As I noted above, the text is divided per cola et commata more frequently than in the the other Vulgate Bibles for which information is available. Capitulum division is indicated throughout the text with a series of numerals written in red in the left margins. The first in the surviving text is cap. VII, at VI. 1 (lv).22 The first one or two words are also written in red, usually in minuscule, although fugit in XXVIII. 1 (which, coincidentally, is also the start of cap. XXVIII) is in rustic capitals. The scribe of the main text in each case left a space at the beginning of the relevant lines for the later insertion of these words. At XXVI. 27 the number XXVII was written but the gap for the first two words, qui fodit, was never filled. In its capitula, L3 exactly follows the divisions classified as ‘series B’ by the Benedictine editors.23 Amiatinus is the only early witness but the series is also in 0 mkQ and two later Theodulf-influenced Bibles.24 At the start of X. 1, the extra heading Parabolae Solomitis is inserted, and a further main division of the text is indicated at XXV. 1, with a large h in haec quoque and also a large g in gloria at the start of the next colon. These features are also in A, except for the large initial letter in gloria. Text In about 30 per cent of the readings where there is variation in surviving manuscripts of Proverbs, A and L3 concur,25 but this is almost certainly due to their common origin in an Italian tradition, not to any direct connection. Convincing evidence of such a connection is lacking, and only two variants exclusive to AL3 are of note.26 They are non for numquam

(XXX. 15)27 and senioribus for senatoribus (XXXI.23).28 Against these must be set the absence from L3, not only of A’s fifty or so unique ‘Ceolfrithian’ variants,29 but also many of the other distinctive variants which A shares with other manuscripts, readings such as expugnator est for expugnatore (XVI.32; AQAZR GVPp*) putamine] uel staminis* (4>AZRG, 0 Mmarg.)116 ad bella} ad bellandum (TZRGVPP) debetis] debeatis (MT, one of the earliest Alcuinian manuscripts (c. 802). In II Maccabees, the character of the text, at least in respect of

A persistent difficulty in the analysis of the text of the two books of Maccabees (as in Chronicles and the prophets) is a general lack of positive variation, and especially of clear-cut differences between the two main Carolingian traditions. No other useful textual landmarks are apparent. In the opening few chapters of I Maccabees, although the character of Royal’s text can scarcely be described as unequivocally Alcuinian, nevertheless

Four Theodulfian manuscripts have been collated in BS XVIII (0 SHAM). In the earlier chapters of I Maccabees, 0 s often concurs with the Alcuinian manuscripts, or most of them, against ©AM. 0 H usually agrees with 0 AM. The latter two are very often joined also by corrected