¹ As Lendinara, P. has observed: ‘The Abbo Glossary in London, British Library, Cotton Domitian i’, ASE 19 (1990), 133–47, at 133.Google Scholar

² The only general study of the Excerptiones is by Law, V., ‘Anglo-Saxon England: Ælfric's “Excerptiones de Arte Grammatica Anglice”’, Histoire Épistémologie Langage 9 (1987), 47–71.CrossRefGoogle Scholar The same author's Insular Latin Grammarians (Woodbridge, 1982)Google Scholar is the standard point of reference for Anglo-Saxon grammatical studies. For a description of Ælfric's use of the Excerptiones, see Bender-Davis, J., ‘Ælfric's Techniques of Translation and Adaptation as Seen in the Composition of his Old English Latin Grammar’ (unpubl. PhD dissertation, Pennsylvania State Univ., 1985).Google Scholar The Grammar is edited in Zupitza, J., Ælfrics Grammatik und Glossar (Berlin, 1880) [henceforth cited as Grammatik].Google Scholar

³ See Ker, N. R., Catalogue of Manuscripts Containing Anglo-Saxon (Oxford, 1957), pp. 1–3 (no. 2) and 442–3Google Scholar (no. 371). Ladd, C. A., ‘The “Rubens” Manuscript and Archbishop Ælfric's Vocabulary’, RES 44 (1960), 353–64, discusses the modern history and dismemberment of the Antwerp–London manuscript.Google Scholar

⁴ Ker, , Catalogue, no. 2Google Scholar, inventories the contents and the order in which they were written. The additions, beside the glossarial material described below, include Remigius's commentary on Donatus's Ars minor, Ælfric's Colloquy, several Latin poems, and a letter in Latin prose. Förster, M., ‘Die altenglische Glossenhandschrift Plantinus 32 (Antwerpen) und Additional 32246 (London)’, Anglia 41 (1917), 94–161Google Scholar, describes the manuscript and establishes the original arrangement of the leaves (pp. 97–8). He also edits the additions of the Antwerp segment, with the exception of the Remigius commentary and the poem by the Frenchman Herbert, which is discussed below.

⁵ Ker, , Catalogue, no. 2Google Scholar, describes the four lists containing English (one on flyleaves, three in margins). The three marginal lists (two alphabetical and one bilingual class list) are in Kindschi, L., ‘The Latin-Old English Glossaries in Plantin-Moretus MS 32 and British Museum Add. 32, 246’ (unpubl. PhD dissertation, Stanford Univ., 1955)Google Scholar [henceforth cited as Kindschi]. All citations below are by page and line from this edition. The glossarial material from the Antwerp segment only is printed in Förster, , ‘Die altenglische Glossenhandschrift’, pp. 101–46.Google Scholar The glossaries are best known from Wright, T. and Wülcker, R., Anglo-Saxon and Old English Vocabularies, 2nd ed. (London, 1884), cols. 104–91Google Scholar, edited from the Junius transcript (Oxford, Bodleian Library, Junius 71). The relationship of this transcription to the original is elucidated by Ladd, , ‘The “Rubens” Manuscript’.Google Scholar I discuss the five lists (a Latin-Latin list of architectural terms on Antwerp 43v, in addition to those mentioned by Ker, , in ‘On the Antwerp-London Glossaries’, JEGP 98 (1999), 170–92Google Scholar). Other studies touching on this glossarial material include that of Gillingham, R., ‘An Edition of Abbot Ælfric's Old English—Latin Glossary with Commentary’ (unpubl. PhD dissertation, Ohio State Univ., 1981)Google Scholar, which analyses the relationship of the class list to Ælfric's Glossary, and that of Lazzari, L., ‘II canto liturgico nel glossario in latino-inglese antico del ms Antwerpen, Plantin Morenas M. 16.2 (47) + London, BL, Add. 32246’, Linguistica e filologia 2 (1996), 193–221Google Scholar, which examines a discrete batch of the class list. Tania Styles of the University of Nottingham is preparing a thesis on the vocabulary of family relationships in the class list.

⁶ Ker, Catalogue, no. 2, art. b. It is in the first glossing hand, the same elegant hand that wrote text glosses and marginal scholia but none of the main text aside from two supply sheets. The list is mostly Latin; there are only a half dozen vernacular interpretamenta among the 1000 items.

⁷ BNF, nouv. acq. lat. 586 is thus to be added to T. A. M. Bishop's list of nine Anglo-Saxon manuscripts that use this abbreviation (‘Lincoln Cathedral MS 182’, Lincolnshire Hist. and Arch. 32 (1967), 73–6, + 2 plates, at p. 73, n. 3Google Scholar): Lincoln 182 (Bede); Antwerp/London (Priscian); Antwerp, Plantin-Moretus Museum 16.8 (Boethius); Brussels, Bibliothèque Royale, 1650 (Aldhelm); Brussels, Bibliothèque Royale, 1828–30 (glossary); Copenhagen, Kongelige Bibliotek, 1595 (sermons); London, BL, Cotton Cleopatra D.i (Vitruvius); London, BL, Egerton 267, fol. 37 (Boethius); and Vatican City, Biblioteca Apostolica Vaticana, Reg. lat. 123, fol. 114 (misc.). Bishop puts the Lincoln Bede, the Brussels Aldhelm and both the Antwerp and the London Boethius at Abingdon along with the Antwerp—London manuscript. For other evidence linking the Paris manuscript to the same scriptorium, see below.

⁸ A stain covers several pages in the first original quire, a probable indication of the mishap that occasioned the twelfth-century repairs. Trimming at the outer top margins of the first four original leaves, fols. 16–20, was probably to remove areas disfigured by the same stain. Small spots of the stain are scattered throughout the manuscript, without impairing legibility.

⁹ Beneath the back pastedown, which has separated from the board, is written ‘Demigieu 1754, 12^ct’. The reference is to the Marquis de Migieu, who acquired the manuscript in 1752 (Ker, , Catalogue, p. 443Google Scholar). Probably it was during this rebinding that the manuscript was trimmed to its present size.

¹⁰ The lower left margin of 113v does, however, hold the sketch of a well-dressed man clasping his hands in front of him. Executed in dry-point and only faintly visible, the drawing has no connection with the text.

¹¹ This ancestor was related to a manuscript used in the composition of Ælfric's Grammar. Both Paris and Antwerp—London originally had deneger for degener (corrected on Paris 34r, uncorrected on Antwerp 6v), a mistake shared with four copies of the Grammar, including the oldest, Oxford, St John's College 154 (Grammatik, p. 45Google Scholar). And Ælfric's copy of the Excerptiones shared the same tradition of scholia with the extant copies, judging from the items of the paratext absorbed by the Grammar, e.g., ‘quaternio, cine oððe feower manna ealdor’ (Grammatik, p. 35Google Scholar); ‘hie quaternio, qui preest iiii militibus uel quattuor dyplomata’ (Paris 31r); ‘glabrio, calu oððe hnot’ (Grammatik, p. 35Google Scholar); ‘glabrio, clauus uel sine pilis’ (Paris 31r). ‘hie lar, ðis fyr/ hi lares, ðas hus. lardum, spic, forðan ðe hit on husum hangað lange’ (Grammatik, p. 42Google Scholar); ‘Lar numero tantum singulari ignem significat. In plurali domus significat. Unde lardum nomen accepit, quod in laribus pendet diu’ (Paris 33r).

¹² E.g., the omission of homuncio from a list of diminutives on London 9r (cf. Paris 16v and Grammatik 17.6); the omission on London 9r of several lines of text (beginning ‘ut tignum…’ and ending ‘…ante um’ on Paris 17v); qui for quia and sonor for soror on the same page (cf. Paris 18r); and there are many other examples.

¹³ The Paris scribe first misinterpreted an apostrophe as the almost identical abbreviation of -us. The -us ending has been pointed for deletion but no apostrophe has been inserted. This is the context: ‘Apostrophus dexter pars est circuli, sed ad summam litteram apponitur. Qua nota deesse ostendimus parti orationis ultimam vocalem, ut Tanton' (Paris, Tantonus) pro Tantóne, et similia’ (Paris 129rv, Antwerp 46v).

¹⁴ Grammatik, p. 168.10.Google Scholar It must be said that there is no evidence in either of the manuscripts for textual contamination from the Grammar, copies of which were numerous. Any such editing would become immediately evident, since anomalous words would appear among the lists of illustrations. The diripio paradigm, however, is the only example where Antwerp—London and the Grammar share material omitted by the Paris manuscript. Some vernacular glosses may have originated with Ælfric's Grammar, or perhaps even with Ælfric's own copy of the Excerptiones (see Ker, , Catalogue, p. 442Google Scholar), but there is no reason to suspect that the monolingual Latin text of the Excerptiones differs much from its author's autograph copy.

¹⁵ Catalogue, p. 1.Google Scholar

¹⁶ Vocabularies II, 130.Google Scholar These glossaries have recently been re-edited by Rusche, P., ‘The Cleopatra Glossaries: an Edition with Commentary on the Glosses and their Sources’ (unpubl. PhD dissertation, Yale Univ., 1996).Google Scholar The Antwerp—London class list has three lemmas beginning with k: Karchesia, melas; Kalende. i. uocationes, gehealddagas uel halige dagas; Kalo g[rece], uoco Latine (Kindschi, , pp. 82.14 and 223.8–9Google Scholar). The compiler of the a-order list could have found four k words in the Excerptiones. karibdis on London 13r, kalipso on Antwerp 5v, kalends on Antwerp 15r and Kartago on Antwerp 33r.

¹⁷ The block, beginning ‘ut si falsum…’, ending ‘…proderam. proderas. proderat’, is on Paris 121v–123v.

¹⁸ The loss of the first folio of quire nine produced an anomalous quire of seven leaves. It is impossible to say whether this loss occurred during medieval times or during the modern dismemberment of the manuscript. All the r words from Wright and Wülcker's edition of the Junius transcript can still be found in the manuscript, but Junius copied only items with English, of which the r batch probably had none, because the a-order glossary is overwhelmingly Latin—Latin.

¹⁹ ‘Xisma. tis. i. nouaculum’, omitted by Kindschi.

²⁰ Each alphabetical batch begins with a capital in the middle of the top margin. Some of these are in red, though now, because of fading, the colour is distinguished only with difficulty. Capitals l, m, n and o (Antwerp 20r, 24r, 28r and 32r) are certainly red.

²¹ See the discussion and examples in Porter, ‘Glossaries’.

²² Gneuss, H., ‘Anglicae linguae interpretatio: Language Contact, Lexical Borrowing and Glossing in Anglo-Saxon England’, PBA 82 (1992), 107–48, at 134–7.Google Scholar

²³ E.g., Dictionnaire historique de la langue française (Paris, 1992), s.v. mentir.Google Scholar

²⁴ Oxford Latin Dictionary, ed. Glare, P. G. W. (Oxford, 1982), s.v. detestor.Google Scholar

²⁵ London 7r: ‘Alia que primitiuorum similem possunt habere significationem’; London 10v: ‘Denominatiuum apcllatur a uoce primitiui sui non ab aliqua speciali significatione’; London 24v: ‘Species uerborum duae sunt, primitiua et diriuatiua, quae inueniuntur fere in omnibus partibus orationis. Est igitur primitiua quae primam positionem ab ipsa natura accepit’; Antwerp 20r: ‘Et seruant significationes primitiuorum, quamuis uideantur quedam ex his in alium sensum transire … Sed si quis attentius inspiciat, non penitus absistunt haec a primitiuorum significatione’; Antwerp 45v: ‘Et sciendum quod poete sepe diriuatiuis utuntur pro primitiuis.’

²⁶ Antestor, ‘to call as a witness’, probably derives from a contraction of *ante-testor (Oxford Latin Dictionary, s.v. antestor). The a-order list makes a logical but wrong connection with antistes, derived from the prefix ante- and the verb sto, ‘to stand’.

²⁷ Of the more than 1000 items in the a-order list, about 150 contain repeated material. The verb asciscere, for instance, appears twice as a headword on London 2r (asscisco, Kindschi, , p. 42.17;Google Scholaradscisco, ibid. p. 43.3). Aqualiculus appears on Antwerp 2r and again on 2v (ibid. pp. 39.2 and 41.18), as does aplustra (ibid. pp. 40.13 and 42.4), etc.

²⁸ Examples include contiguous glosses on mutuo and mentio, first on Antwerp 24r and again on 24v (Kindschi, , pp. 256.7–8 and 257.15–16Google Scholar), on oria and obsonor, twice on Antwerp 32r (ibid. pp. 259.14–15 and 260.11–12), and on nebulo and nequito, twice on Antwerp 28r (ibid. pp. 258.8–9 and 258.13–14).

²⁹ Cf., for example, Kindschi, , pp. 255.16–256.15 and 257.1–257.20, or 259.7–259.15 and 260.1–260.12.Google Scholar

³⁰ The variations show that the scribe was not copying slavishly but was editing and interpreting as he transferred glosses from the working copy to the margins, as in the following example (Kindschi, , pp. 258.8–9 and 258.13–14Google Scholar):

nebulo. i. mendax

nequito et nequitor. i. nequiter ago, uel inutilem rem ago

nebulo. i. mendax qui mendaciis quibusdam ueritatem obscurare nititur, uel qui suos

fallit auditores

nequito, as. i. nequiter ago. Necatus ferro dicimus. Nectus uero. Alia re peremptus dicitur.

In this instance the compiler expands the original definitions, but in other instances he abbreviates, sometimes severely (e.g., the two glosses on incubo, Kindschi, , pp. 115.13 and 202.7Google Scholar). At a later date I plan to examine the details of the working methods.

³¹ Kindschi, , p. 116.4–5:Google Scholar ‘Cedes. i. interemptio. Strages uero est ubi multi occiduntur’; ‘Glades. i. morbus’. Kindschi, , p. 254.1–2: ‘Lues. i. illuuies. sordiditas’; ‘Labes. i. interims’.Google Scholar

³² See above, pp. 90–4.

³³ Another distinctive scribal habit is forming the bowl of the g as a circle.

³⁴ There are, for example, mistakes in Paris that are not repeated in Antwerp—London: incorrect ma for ama (Paris 80r, Antwerp 23r), incorrect frigero for frigeo (Paris 87v, Antwerp 27r; cf. Grammatik, p. 156.6), and incorrect indutabilem for indubitabilem (Paris 104v, Antwerp 35r), as well as the omission of the diripio paradigm (Paris 90v, Antwerp 28v). Other evidence against direct copying is the relatively large number of text glosses, Latin-Latin or Latin-English, unique to one or other of the manuscripts (especially evident in the section on adverb inflexion, Paris 99v-106r, Antwerp 32v-35v), and the very different pattern of word accents on, for example, Paris 83v and Antwerp 25r. It also seems implausible that an eyeskip error on Antwerp 41r (the omission of the phrase ‘…aduerbia poni debent. Sunt autem pleraque huiuscemodi aduerbia…’) could have arisen from the Paris manuscript, where the two instances of the word aduerbia appear on different pages, 116r and 116v.

³⁵ Both lemma and interpretamentum mean ‘to surpass’ or ‘to excel’. The French verb in the form surmunter is first attested c. 1119, in the form surmonter c. 1155 (Dictionnaire historique de la langue française, s.v.).

³⁶ Coniueo means ‘to wink’ or ‘to leave uncensured’, ergo ‘to connive at’. Old French ceners means ‘to give a sign’. Latin nuo (‘to nod’) is a back formation from adnuo. Ceners is the ancestor of modern signer (‘to sign’, ‘to make a sign’). For the environment of all these words in the glossary, see p. 104. The earliest attestation of ceners is again the Chanson de Roland, with the meaning ‘to make the sign of the cross’, according to the Dictionnaire historique de la langue française. The same source gives the gloss meaning as common from the twelfth and into the seventeenth century.

³⁷ Sufulcirs is a hapax legomenon, cited in neither Godefroy's Dictionnaire de l'ancien française, 10 vols. (Paris, 1880–1902)Google Scholar nor in Altfranzösisches Wörterbuch, ed. Tobler, A. and Lommatzsch, E. (Berlin and Wiesbaden, 1925–; in progress).Google Scholar The word is apparently a loan formation from Latin suffulcio, ‘to prop from below’, ‘to keep from falling’. The form of the word in the Paris manuscript is certain: the final s is tall and quite clear. Old French foucir is the well-attested reflex of the Latin Lemma., fulcio (Altfranzösisches Wörterbuch, p. 2178Google Scholar).

³⁸ Because it is attested only very late (fourteenth century), it is problematic to link perfumers with the eleventh-century scribe. Nevertheless, the lemma suffio occurs twice in the a-order list (below, p. 104). Ancestor of modern parfumer ‘to perfume’, porfumers is of undetermined Romance origin (Dictionnaire historique de la langue française, s.v.).

³⁹ Merchier, ancestor of modem marquer ‘to mark’, is attested in this form from c. 1120 (Dictionnaire historique de la langue française, s.v.). The OE cognate mearcian is found in the works of Alfred and Ælfric (Oxford English Dictionary, s.v. mark). It is interesting that Latin insignire is interpreted in Ælfric's Grammar by OE mœrsian (Grammatik, p. 192Google Scholar), which is phonetically similar to the French interpretation here.

⁴⁰ Catolliers, ancestor of modern chatouiller ‘to tickle’ or ‘to titillate’, is attested c. 1220. The primitive spelling ca- indicates that the word was in currency at an early date (Dictionnaire historique de la langue française, s.v.).

⁴¹ The gloss, damaged by trimming, is illegible at beginning and end. The meaning may be ‘[?to remain] all day is to sojourn’. Whatever the Latin verb, this is evidendy a learned definition arrived at by translating the lemma by a vernacular phrase to which it is linked etymologically. Soiorners, ancestor of modern French séjourner, has been attested as a late-eleventh-, early-twelfth-century word (Dictionnaire historique de la langue française, s.v.); the shift of vowels (/o/ to /e/) is owing to the medieval school-pronunciation of Latin words (Ewert, A., The French Language (London, 1933), p. 55).Google Scholar

⁴² Past participle of suffulcio ‘to prop below’.

⁴³ In Kindschi, , pp. 105.10, 105.12 and 272.13.Google Scholar The first two instances are in the Latin-English class list, the third in the mosdy Latin a-order list.

⁴⁴ Förster, , ‘Glossenhandschrift’, p. 121.Google Scholar

⁴⁵ A. Campbell represents the most extreme view in minimizing the French influence: ‘No loanwords which can certainly be regarded as French occur in manuscripts older than 1066, except prūd, prīct proud, whence are derived pryte, pryt pride. There are … a few words like capun, castel, which might be derived from Latin or French’ (Old English Grammar (Oxford, 1959), p. 221).Google Scholar

⁴⁶ Ker, , Catalogue, p. 3.Google Scholar The poem (edited by Förster, , ‘Glossenhandschrift’, p. 154Google Scholar) and its Abingdon associations are discussed by Lapidge, M., ‘Æthelwold and the Vita S. Eustachii’, in his Anglo-Latin Literature 900–1066 (London, 1993), pp. 213–23, at 218.Google Scholar

⁴⁷ Porter, D., ‘Æthelwold's Bowl and the Chronicle of Abingdon’, NM 97 (1996), 163–7.Google Scholar I omit from discussion here the elegy on the death of Ælfric, archbishop of Canterbury (Ker, , Catalogue, p. 3Google Scholar), as that reflects as much a Canterbury as an Abingdon connection.

⁴⁸ Ker, , Catalogue, p. 382.Google Scholar

⁴⁹ The folio is an originally blank flyleaf. The poem is ed. Dümmler, E., ‘Lateinische Gedichte des neunten bis elften Jahrhunderts’, Neues Archiv der Gesellschaft für ältere deutsche Geschichtskunde 10 (1885), 333–57, at 351–3.Google Scholar

⁵⁰ Michael Lapidge draws my attention to the tenth-century Horace manuscript of Fleury provenance, Paris, BNF, lat. 7971. On 3r a metrical ex-libris in elegiacs names Herbertus as the donor: ‘Hic liber est, Benedicte, tuus, uenerande, per [euum]; / Obtulit Herbertus seruus et ipse tuus’. Another ex-libris on 1v names Constantius, who had close ties with the Ramsey monk Oswald when Oswald was studying at Fleury. See Vidier, A., L'historiographie à Saint-Benoît-sur-Loire et les miracles de Saint Benoît (Paris, 1965), p. 53Google Scholar, and Mostert, M., The Library of Fleury (Hilversum, 1989), no. BF1140 (p. 222).Google Scholar

⁵¹ Relations between Canterbury and Fleury were particularly strong, for example. V. Ortenberg documents communications over several decades of the late tenth and early eleventh centuries: The English Church and the Continent in the Tenth and Eleventh Centuries (Oxford, 1992), p. 10Google Scholar, n. 28. There was likewise considerable communication between Canterbury and Abingdon. Ælfric, archbishop of Canterbury from 995 to 1005, had been an Abingdon monk, and Siward, abbot of Abingdon from 1030 to 1044, travelled to Canterbury in 1044 to become executive assistant to the enfeebled archbishop. He returned to Abingdon in 1048, dying there on 23 October. There must have been many others who travelled between Canterbury and Abingdon, either alone or in the train of great men like Ælfric and Siward. See Graham, T., ‘CCCC 57 and its Anglo-Saxon Users’, Anglo-Saxon Manuscripts and their Heritage, ed. Pulsiano, P. and Traherne, E. (Aldershot, 1998), pp. 21–69Google Scholar, for an analysis of the historical and codicological links between Canterbury and Abingdon. As for the presence of a Frenchman at Abingdon, it may be relevant that an anonymous redactor of the Abingdon version of the Anglo-Saxon Chronicle exhibited both interest and expertise in French: see The Anglo-Saxon Chronicle, MS B, ed. Taylor, S., The AS Chronicle: a Collaborative Edition, ed. D. Dumville and S. Keynes 4 (Cambridge, 1983), p. lxi.Google Scholar

⁵² For an example, see Byrhtferth's Enchiridion, ed. Baker, P. and Lapidge, M., EETS ss 15 (Oxford, 1995)Google Scholar for an assessment of the career and influence of Abbo of Fleury, one of the luminaries of monastic culture near the turn of the millennium.

⁵³ ‘Perlege tarnen et, que peto, perfice clemens et mihimet misero miseriis miserere misello…’ (Förster, , ‘Glossenhandschrift’, p. 153Google Scholar). Cf. Dümmler, , ‘Lateinische’, p. 351.8: ‘Perlege tu pastor perfice quodque precor’; and 353.89: ‘Tu mihimet misero miserans miserere misello’.Google Scholar

⁵⁴ An entry from the class glossary, for example, holds the solution to the Latin riddle (my ‘A Double Solution to the Latin Riddle in MS Antwerp, Plantin-Moretus Museum 16.2’, ANQ 9.2 (1996), 3–9, at 3Google Scholar). Some lexical rarities from Herbert's elegy, amphibalis, birrus, cauma, also appear in the class list (Kindschi, , pp. 88.3, 155.13 and 234.10Google Scholar). And it is worth noting the elegy's adjectival forms lupinus and ferinus, which recall the discussion in the Excerptions of adjectives derived from animal names (London 7v): ‘Omnia quae a nominibus mutorum animalium cuiuscumque sint declinationis diriuantur in formam possessiuum i penultimam longam seruant, ut aper aprinus, caper caprinus, ceruus ceruinus, porcinus, taurinus, ferus uel fera ferinus, caninus, leporinus, lupinus.…’

⁵⁵ Kindschi's comment on this topic is misleading: ‘The Latin—Old English class glossary … is crowded into such marginal space as remained [after the alphabetical list was written]’ (p. 2).Google Scholar Antwerp—London in fact holds much empty marginal space which the scribe of the class glossary chose to ignore. Instead he squeezed his text onto a few leaves, between earlier strata in the forward-sloping hand.

⁵⁶ This seems a strong possibility, and if it is so, one with relevance for the word porfumers, which as we noted above is unattested before the fourteenth century. The earliest attestation of mouton is some 150 years after the date of the glossary, and even then it is considered a dialectal form (Dictionnaire historique de la langue française, s.v.). Mutinus must be related to mutina, which in the Paris Priscian (20v) is glossed ‘…bestia sine cornibus’, an animal without horns, i.e. a sheep.

⁵⁷ One may speculate whether the hapax sufulcirs was a loan formation in imitation of Anglo-Saxon scholastic practice.

⁵⁸ In the course of this work I incurred many debts. I thank the institutions that allowed me to examine manuscripts in their keeping: the British Library, the Bibliothèque Nationale de France and the Planrin-Moretus Museum. Thanks also to the National Endowment for the Humanities, which supported this work with a research fellowship, and to the English Department of Louisiana State University, which welcomed me as a visiting scholar in 1997.