Introduction | Bede’s Latin | Cases | Genitive | Accusative | Verbs | Shifted pluperfect | Participles | Syntax of Subordinate Clauses | Relative Clauses | Cum Temporal Clauses | Indirect Discourse | Vocabulary | Bede’s Style | Hyperbaton | Connective Relative | Bibliography
At the beginning of the Historia Ecclesiastica Bede names the five languages then spoken in Britain, Anglo-Saxon, British, Irish, Pictish, and Latin:
Haec in praesentī iuxtā numerum librōrum quibus lēx dīvīna scrīpta est, quīnque gentium linguīs, ūnam eandemque summae vēritātis et vērae sublīmitātis scientiam scrūtātur et cōnfitētur, Anglōrum vidēlicet Brettōnum Scottōrum Pictōrum et Latīnōrum, quae meditātiōne scrīptūrārum cēterīs omnibus est facta commūnis. (1.1.13)
Bede connects these languages to the unity of the Church. There are five languages, but they are all devoted to exploring and confessing one and the same truth. And it is Latin—the language of the Church and the language in which Bede himself writes—that unifies Christian Britain through the study of the Scriptures.
Language is central to the story of the evangelization of Britain, and to Bede’s conception of the overall unity of the Church. In his Biblical commentaries, particularly On Genesis and On the Temple, Bede develops a contrast between the linguistic confusion of the tower of Babel and the mutual understanding of Pentecost. Bede recognized that a knowledge of languages, an ability to make oneself understood, was essential to the spread of Christianity and the unity of the Church.
One of the significant figures in the Historia Ecclesiastica is the interpres. The root meaning of interpres is “go-between” or “middleman”—the word seems originally to have been associated with negotiating business transactions (Brown 1993, 43–44)—but for Bede an interpres is a translator. The word first appears in the Historia Ecclesiastica in Book 1, when St. Augustine’s mission picks up translators (interpretes, 1.25.5) in Gaul before embarking for England. The last appearance is in Book 5, when King Nechtan of the Picts receives a letter from the English church instructing him in the Christian faith. The letter has to be translated into the King’s own language by an interpres (5.21).
In Book 3, King Oswald acts as an interpres for Aidan, translating the Irish bishop’s teachings into the language of the Northumbrian people (3.3.9). Later in the same book, Bishop Cedd acts as an interpreter for the parties at the synod of Whitby (3.25). Finally, in Book 4, interpretes teach Caedmon scriptural lessons, and he translates them into vernacular song (4.24.1). Bede sees himself as this kind of interpres: through his writings, in particular his Biblical commentaries, Bede is engaged in interpretatio (5.24), interpreting the spiritual meaning of Biblical texts for his readers.
In all of these instances, the figure of the interpres is crucial in spreading Christianity from the Continent to England, and from England to the furthest reaches of the British Isles. The interpres facilitates the unity of the Church by providing a bridge between Christians who speak different languages. Through the interpres, the Word becomes domesticated to different languages, places, and cultures.
The verb from which we derive the word “translate,” transferre, has a somewhat more complicated story. It can mean “to translate” from one language to another, as it does in the story of Caedmon. Holy men and women, at their deaths, are often “translated” from earth to heaven. And the bones of saints are “translated” from the monastic cemetery to a special place inside the church, where they become the locus of pilgrimages and miracles. Translation is as much as spiritual as a linguistic process connecting the ordinary with the sacred, this life with the next.
Transferre can also mean “to convert,” as when the monks of Iona are converted (translati, 3.4) to the canonical observance of Easter by Ecgbert. Ecgbert’s original plan was to travel to Germany to attempt to convert the pagan Germani—to rescue them from Satan and convert them to Christ (ereptos Satanae ad Christum transferre, 5.9).
It’s interesting to note how Ecgbert is prevented from carrying out his plan of traveling to Germany. A vision of Boisil, the late prior of Melrose, appears to one of his former students, now a brother in Ecgbert’s abbey. In the vision, Boisil tells the brother to convey a message to Ecgbert: Ecgbert is needed not in Germany, but in the monastery at Iona. The unnamed brother acts as an intermediary, a kind of interpres, between Boisil and Ecgbert. In this case, the brother is an unsuccessful go-between. The message is repeated on a subsequent night, but Ecgbert still chooses to set out for Germany. It takes a storm at sea to convince Ecgbert to turn back and direct his attention to Iona. Direct signs from God are a back-up plan when the words of intermediaries fail.
Another way of looking at it is to see the sign as visible proof of the validity of the words. Miracles are often proofs of the sanctity of a life lived according to the Gospel. The Word is translated into the actions of a human life, the piety of which is in turn demonstrated through miracles. Miracles are visible signs (signa) of a spiritual truth: a kind of translation.
Simple, modest, unpretentious, pure: these were some of the adjectives used to describe Bede’s prose during the Middle Ages (Sharpe 2005, 340). This assessment of Bede’s style is echoed by modern scholars, who have called it “pure, simple, and efficient” (Wetherbee 1978, 23) and “clear and limpid” (Plummer 1896, I:liii), and have remarked on its “remarkable naturalness and simplicity,” its clarity, and its “great purity of language” (Druhan 1938, xx–xxii).
Latin was not Bede’s native language. The language of his Northumbrian family was Old English (Anglo-Saxon), but Bede would have been introduced to the study of Latin when he was sent to the monastery of Wearmouth-Jarrow for his schooling at the age of seven. The monastery was founded in 673 (at about the same time as Bede’s birth) by Benedict Biscop, who built up an extensive library at Wearmouth-Jarrow from his own book buying trips to the Continent. Estimates of the number of books available in the library at Wearmouth-Jarrow range from 130 (Plummer 1896, I:l–li) to 150 (Laistner 1935, 263–266) to 250 (Lapidge 2006, 36)—in any case, it would have been the most extensive library in Britain at the time of Bede.
Most of the books available to Bede would have been Christian texts, including scriptural commentaries, theological treatises, histories (like Gregory of Tours’ Historia Francorum), and poetry. But among the few pagan texts in the library at Wearmouth-Jarrow would have been the grammatical treatises of Donatus, Servius, Consentius, and others. Bede seems to have studied those grammars carefully. Druhan concludes from “the general regularity of [Bede’s] syntax that he followed as his guide to syntax … a definite body of precepts laid down by grammarians whose works he had before him” (1938, xxiii). Shanzer (2007) argues that Bede’s developed the elegant, periodic style of historical writing seen in HE not from grammar books but through close reading of Christian historians, especially Orosius and Rufinus of Aquileia.
Terence Tunberg cautions that “it is actually not easy to isolate features that are unequivocably and exclusively peculiar to … Medieval Latin.” In Latin textbooks, he argues, “the syntactical and grammatical norms … reflect the prose usage of only two canonical authors, Cicero and Caesar, while the full range of ancient Latin, from Terence to St. Augustine, demonstrates a wide range of variation from the Ciceronian norms” (2004, 157–158). Most features of Bede’s Latin that appear to be deviations from classical usage are, in fact, attested elsewhere in classical Latin outside the works of Cicero and Caesar. As Druhan concludes, “Bede’s Latinity is almost classical” (212).
Below are a few of the more common deviations in Bede from the “normative” Latin grammar and syntax found in introductory Latin textbooks and in Cicero and Caesar.
Druhan notes that “in the use of the genitive case, extensions of the classical usages are considerable in Bede” (1938, 197). Most of these are not difficult, and do not require special attention here. Two are worth noting for their frequency or peculiarity:
1.1 “Chorographic Genitive”
Bede regularly uses the genitive of a place name, or the genitive of the name of the inhabitants of a place, after nouns such as urbs, provincia, oppidum, flumen, etc. In the Praefatio, for example: in provinciā Oriēntālium Anglōrum, “in the province of the East Angles” (i.e., East Anglia), in provinciā Lindissi, “in the province of Lindsey,” in ecclēsiā Cantuariōrum, “in the church of the Kentish people” (i.e., Canterbury).
Bede occasionally uses the genitive of an abstract noun instead of an attributive adjective. In the Praefatio, for example, he writes studium tuae sinceritātis (“the zeal of your sincerity”) where studium sincerum (“sincere zeal”) would have been more straightforward. This adds a certain poetic impressiveness to the style in expression of key ideas.
2.1 Accusative of motion without preposition
The use of the accusative of motion towards, without a preposition (ad, in), is expanded in Bede. In addition to the expected nouns (e.g., domum, humum, cities, small islands, AG 427), Bede employs the construction with patriam and the names of countries and provinces (e.g., Galliam, Britanniam). For example, in 4.23.2: desiderāns ... Galliam pervenīre, “desiring to get to Gaul.” Normally the proposition in or ad would be used (Logeion, s.v. pervenio).
3.1 Ablative of extent of time
Bede often uses the ablative to express extent of time and space, rather than the accusative (AG 423.2). For example, in 4.20.35: bis sex rēgnāverat annīs, “had ruled for twelve years.” The accusative is also used.
3.2 Connected ablative absolute
Contrary to common usage, in which the noun in an ablative absolute very seldom denotes a person or thing elsewhere mentioned in the same clause (AG 419), Bede often employs the ablative absolute where the ablative noun is identical with the subject of the sentence. For example, in 4.23.30: nūntiāvit mātrem illārum omnium Hild abbātissam iam migrāsse dē saeculō, et sē aspectante ..., “She announced that the Abbess Hild, the mother of them all, had passed away, and while she herself was watching ...”
Bede frequently employs the shifted form of the pluperfect, using fuisse, fuisset, or fuerat instead of esse, esset, or erat. For example: 4.23.28: cum quis eōrum dē saeclō fuisset ēvocātus, “when anyone of them had been summoned from earthly life.”
5.1 High frequency and extended use of present participles
“Bede exemplifies a distinctive trait of Late Latin in the great abundance of participles which he employs and in the extended uses he makes of them” (Druhan 1938, 138). Particularly noticeable is the abundance of first conjugation present participles in the nominative case, which occur more than four times as often in HE than they do in samples of Caesar and Cicero (In Caesar’s Gallic War and Cicero’s De Amicitia such words account for 0.04% of total word forms, compared to 0.18% in HE Book IV).
Bede frequently employs a present participle where one might expect, for example, a cum-clause or ablative absolute. In 4.8.2, for example: sīc termināns temporālem vītam intrāvit aeternam, “and thus ending his earthly life, he entered eternal life.”
“The frequency of the subjunctive in relative clauses in the Historia Ecclesiastica is very high ....This shows a preference for the subjunctive mood characteristic of later writers” (Druhan 1938,156). Druhan argues that all of these cases fall within common usage, and can be construed as relative clauses of characteristic (AG 534), relative clauses of purpose (AG 531), or subjunctive “by attraction,” that is, when the relative clause is part of an indirect statement or ut-clause of purpose (AG 591).
As in classical and later Latin, in Bede the conjunction cum can introduce clauses that are either temporal, causal, or concessive (AG 544–549). Cum is the most frequent subordinating conjunction in Bede. But its purely temporal use (with the indicative) is fairly rare in Bede. In these cases the clause is purely temporal, and cum indicates contemporaneous action in the subordinate and main clauses (e.g. 2.1.12 cumque … sparserit [fut. pf.], etiam cum … appetit, “when he dissipates …, even when he reaches for …” See note ad loc.). For temporal clauses Bede prefers dum. See Druhan 1938, 174–176.
Bede regularly introduces indirect statements with quod (or quia), followed by a subordinate verb in the subjunctive, though he also employs the more expected accusative and infinitive construction The use of quod + subjunctive after verbs of saying and thinking occurs in some classical authors (first in the Bellum Hispanicum), but becomes common in later Latin. The use of quia + subjunctive in indirect statement is first observed in Petronius, and in Christian Latin it gains predominance over quod. In Bede such quia clauses are very frequent. On occasion, the verb following quod or quia will be in the indicative (see 4.19.18: crēdō quod ideō mē superna pietās dolōre collī voluit gravārī, “I think that heaven has wished me to weighed down by neck pain”). Druhan notes: “Bede seems to use the subjunctive consistently whenever the statement is advanced as that of another, without any implication as to the truth or falsity of the statement expressed in the quod- or quia-clause. The indicative is found only when the substantive clause contains the thought or statement of the author” (211).
Bede’s vocabulary is “fundamentally that of the classical and Silver Age, supplemented by the traditional Christian vocabulary of writers like Augustine, Jerome, Ambrose, Prosper, and Gregory the Great” (Druhan 1938, xxii).
The following graph shows the ratios of DCC Core and Non-Core vocabulary in the first section (156 words) of Bede’s Praefatio and in a 156-word selection from Cicero’s Dē Amicitiā (1st c. BCE), Gildas’s Dē Excidiō et Conquestū Britanniae (6th c. CE) and the Hisperica Famina (7th c. CE). In the case of the Hisperica, the comparison is somewhat artificial because the Hisperica is in verse, and likely aims at a parodic effect with its deliberately obscure vocabulary (Ó Cróinín 1995, 179).
In the passage from Cicero, 84% of the vocabulary is in the Core; in Bede, 71%; in Gildas, 60%; and in the Hisperica Famina, 22%. Of the Non-Core vocabulary in Bede, several of the words (such as transcribō and transmittō) are compounded forms of Core vocabulary words.
The largest class of non-Core vocabulary words in Bede are Christian Latin vocabulary words like abbas (abbot), episcopus (bishop), monasterium (monastery), and rēgulāris (governed by a monastic rule). Most words in Bede’s Historia Ecclesiastica can be found in standard Latin dictionaries such as the Oxford Latin Dictionary and Lewis and Short. But certain words that appear in these dictionaries will have different, specifically Christian connotations in Bede. Most of these can be found in the Dictionary of Medieval Latin from British Sources (“DMLBS”) and the Glossarium mediae et infimae latinitatis (“DuCange”), both of which are included on Logeion.
Finally, most Latin dictionaries (e.g., Lewis and Short, Oxford Latin Dictionary) give the assimilated form of verbs that begin with a prepositional prefix, such as compono (for conpono) or afflictus (for adflictus). Bede, however, generally uses the unassimilated forms of the verb (i.e., conpono or afflictus). One exception is ammoneo, which Bede uses instead of the more standard unassimilated admoneo. On consonantal assimilation, see AG 16.
Bede’s style, while generally described simply as “clear” and “pure,” is in fact remarkably varied. His prose can be fast-moving and dramatic (as when he recounts the assassination attempt against Edwin, 2.9.13–19), polished and periodic (as when he reports the preaching of Paulinus, 2.16.1–3, or of Wilfrid), brief and simple (as in the story of Caedmon, 4.24), or highly ornate (as in the simile of the sparrow in 2.13.9–13, see Shanzer 2007, 333–336). He includes poetry in the work, as in the acrostic hymn in praise of virginity (4.20). Bede is adept at varying his style while maintaining a sense of overall stylistic unity. In the words of Gregory Hays: “Medieval Latin works are not always stylistically homogenous; even a text by a single author may vary in register from section to section and even from one section to the next. Not all authors are capable of this kind of versatility, but for those who are it is an extraordinarily effective tool .... Bede’s mastery of multiple styles is one of his particular gifts as a writer” (2012, 227).
Let us look at a passage from the Historia Ecclesiastica to identify some of the more common stylistic features of Bede’s prose. This passage is from Book 4, Chapter 6, where Bede talks about Eorcenwold’s foundation of two monasteries: Chertsey and Barking.
 Hic sānē, priusquam episcopus factus esset, duo praeclāra monastēria, ūnum sibi alterum sorōrī suae Aedilburgae, cōnstrūxerat, quod utrumque rēgulāribus disciplīnīs optimē īnstituerat; sibi quidem in regiōne Sudergeonā iuxtā fluvium Tamēnsem in locō quī vocātur Cerotaes Eī, id est Cērōtī īnsula, sorōrī autem in Orientālium Saxonum prōvinciā in locō quī nuncupātur In Berecingum, in quō ipsa Deō dēvōtārum māter ac nūtrīx posset existere fēminārum.  Quae susceptō monastēriī regimine, condignam sē in omnibus episcopō frātre et ipsa rēctē vīvendō et subiectīs rēgulāriter ac piē cōnsulendō praebuit, ut etiam caelestia indiciō fuēre mīrācula.
Even before he had been made bishop, he founded two famous monasteries, one for himself and the other for his sister Æthelburh, both of which he established the best rule of monastic discipline: for himself in the territory of Surrey, beside the River Thames, in a place called Chertsey, or the island of Ceorot; and for his sister in the kingdom of Essex, in a place called Barking, where she was the mother and nurse of women devoted to God. Having undertaken the rule of the monastery, she showed herself equal to her brother in worth, both in her own righteous life and in the pious and disciplined care she took for her subjects—as, indeed, a series of heavenly miracles demonstrated.
Bede likes to set the scene with a temporal or circumstantial clause (cum-clauses with the subjunctive are the most common) or an ablative absolute. You’ll be seeing a lot of those constructions as you read Bede. What follows is a remarkably controlled and balanced construction, as Bede alternates between Chertsey Abbey, which Eorcenwold established for himself (ūnum sibi ... sibi quidem ...), and Barking Abbey, which he established for his sister (alterum sorōrī ... sorōrī autem ...). The result is almost architectural. Here, as elsewhere, Bede exercises “unobtrusive but complete linguistic control” over his material (Wetherbee 1978, 26; see also Shanzer 2007, 335).
At the same time, Bede demonstrates his mastery of variation, both in his choice of words (substituting nuncupātur for vocātur, for example) and in the length of his clauses and sentences. This section comprises two sentences—one of 62 words, the other of 27 words—of varying syntactical complexity. The second sentence, though shorter, has a more complex periodic structure: the sense of the main clause is not complete until the verb praebuit. As Hays remarks: “An author’s ability to mix long and short, paratactic and hypotactic, is a measure of his compositional skill” (2012, 221).
In his chapters on Barking Abbey (4.7 ff.), Bede makes use of a lost book, written at the abbey itself, as his principal source. Although it is impossible to determine the extent of his borrowings from that lost book, in other places in the Historia Ecclesiastica it is clear that Bede incorporates direct quotations from his sources. In Book I, for example, he lifts long passages from Orosius without attribution (Plummer’s edition prints the borrowed words in italics, which makes his debt clear). With other sources, such as Gildas, Bede carefully rewrote and adapted the original to his own stylistic preferences (see Shanzer 2007, 331–333, for an example). Bede was as skilled an editor as he was a writer.
Bede often separates words that belong together, such as nouns and their modifying adjectives. For example, five words come between the adjective dēvōtārum and the noun fēminārum at the end of the first sentence; two words separate caelestia and mīrācula at the end of the second sentence. This separation of associated words is a type of hyperbaton, a word derived from Greek hyper (“over”) and bainein (“to step”). It is a very common, indeed integral, feature of Latin poetry, and a regular feature of more artistic varieties of Latin prose. Bede himself discusses types of hyperbaton in his rhetorical treatise De schematibus et tropis (On Figures and Tropes), defining it as “a kind of transposition of words which upsets their natural order” (112). According to Kendall, hyperbaton is the rhetorical figure that “more than any other, gives Bede’s prose its distinctive flavor” (1978, 153). Kendall notes that Bede is especially fond of separating adjective-noun phrases (as in our example) and of separating a preposition from its object.
Hays points out that this device “is common in verse, for metrical reasons,” and that “its use in prose texts can confer a poetic flavor" (2012, 221). Bede may have reserved fēminārum for the end of the sentence because it created a more pleasing rhythm than existere posset. Medieval Latin prose authors prefer sentences that end in words of three or four syllables, and in certain rhythms (Tunberg 1996, 114–118), but these are preferences, not hard-and-fast rules. In this chapter, for example, six sentences end with words of 3 or 4 syllables, but one ends with a monosyllable. As always, Bede strives for variation.
This instance of hyperbaton (dēvōtārum māter ac nūtrīx posset existere fēminārum) is an example of what Kendall calls compound hyperbaton, in which “a phrase ... is interrupted by two or more words not in themselves forming a single integral phrase” (154). Kendall observes that hyperbaton is “an indicator of stylistic level,” and that more complex forms of hyperbaton “are found, almost without exception, in moments of high religious seriousness ... to secure a pious tone or assert matters of faith” (157). As Bede comes to the end of the sentence, he shifts from a plain historical narrative (explaining which monastery was established where) to a characterization of Barking Abbey’s pious community. The use of hyperbaton elevates the stylistic register.
Bede likes to connect sentences with a relative pronoun, as he does with quae in the second sentence (AG 308.f). Bede is, of course, concerned not only with the structure of individual sentences, but also with the larger structures of chapters, books, and the overarching structure of the Historia as a whole. Notice how the first sentence begins with hic, referring to Eorcenwald, but ends with a new subject, māter ac nūtrīx, referring to Æthelburh, which becomes the antecedent of quae at the beginning of the next sentence. Through his careful use of syntactic parallelism and subordination, Bede shifts his focus from Eorcenwald to his sister, and illustrates his assertion that Eorcenwold and Æthelburh are equally worthy (condignam). The passage as a whole moves from Eorcenwold (hic) to Æthelburh (quae) to the inhabitants of her abbey (subiectīs): the syntactic dependency mirrors the flow of authority from the founding bishop to the abbess to her subjects.
In the end, the piety of Æthelburh and the community of Barking Abbey is demonstrated through miracles. Bede moves from a straightforward narrative of events—the brick and mortar founding of Barking—to an exploration of the spiritual significance of those events. He acts as both narrator and interpreter.
For stylistic discussions of other passages in Bede’s Historia Ecclesiastica, see Hays 2012, 224–227, Shanzer 2007, and Wetherbee 1978. For breakdowns of some of Bede's particularly impressive periodic sentences, see the articulated texts in this edition.
Brown, George H. 1993. “The Meaning of Interpres in Aldhelm and Bede.” In Interpretation: Medieval and Modern, edited by Piero Boitano and Anna Torti, 43–65. Cambridge: D.S. Brewer.
Druhan, D.R. 1938. The Syntax of Bede’s Historia Ecclesiastica. Catholic University Studies in Medieval and Renaissance Latin, vol. 8. Washington, DC: Catholic University.
Grocock, C.W. 1999. “Bede and the Golden age of Latin Prose in Northumbria.” In Northumbria’s Golden Age, edited by J. Hawkes and S. Mills, 371–382. Stroud: Sutton Publishing.
Hays, Gregory. 2012. “Prose Style.” In The Oxford Handbook of Medieval Latin Literature, edited by Ralph J. Hexter and David Townsend, 217–238. Oxford: Oxford University Press, 2012.
Kendall, Calvin B. 1978. “Bede’s Historia Ecclesiastica: The Rhetoric of Faith.” In Medieval Eloquence: Studies in the Theory and Practice of Medieval Rhetoric, edited by James J. Murphy, 145–172. Berkeley: University of California Press.
Laistner, M.L.W. 1935. “The Library of the Venerable Bede.” In Bede: His Life, Times, and Writings, edited by A.H. Thompson, 237–266. Oxford: Clarendon Press.
Lapidge, Michael. 2006. The Anglo-Saxon Library. Oxford: Oxford University Press.
Ó Cróinín, Dáibh. 1995. Early Medieval Ireland, 400–1200. Longman History of Ireland. New York: Routledge.
Plummer, Charles. 1896. Venerabilis Baedae Historia Ecclesiastica Gentis Anglorum Historia. 2 vols. Oxford: Clarendon Press. Extensive introduction, Latin text with apparatus criticus, notes, and valuable indices. Scans on Archive.org: vol. 1. vol. 2.
Shanzer, Danuta. 2007. “Bede’s Style: A Neglected Historiographical Model for the Style of the Historia Ecclesiastica?” In Source of Wisdom: Old English and Early Medieval Studies in Honour of Thomas D. Hill, edited by Frederick M. Hall et al., 329–352. Toronto: University of Toronto Press.
Sharpe, Richard. 2005. “The Varieties of Bede’s Prose.” In Aspects of the Language of Latin Prose, edited by Tobias Reinhardt, Michael Lapidge, and J.N. Adams, 339–356. Oxford: Oxford University Press.
Tannenhaus, Gussie Hecht. 1973. Bede: Concerning Figures and Tropes. In Readings in Medieval Rhetoric, edited by Joseph M. Miller, Michael H. Prosser, and Thomas W. Benson, 96–122. Bloomington, IN: Indiana University Press.
Tunberg, Terence. 1996. “Prose Styles and Cursus.” In Medieval Latin: An Introduction and Bibliographical Guide, edited by Frank Mantello and Arthur G. Rigg, 111–121. Washington, DC: Catholic University of America.
Tunberg, Terence. 2004. “The Latinity of Erasmus and Medieval Latin: Continuities and Discontinuities.” Journal of Medieval Latin 14: 145–168.
Wetherbee, Winthrop. 1978. “Some Implications of Bede’s Latin Style.” In Bede and Anglo-Saxon England: Papers in honor of the 1300th anniversary of the birth of Bede, given at Cornell University in 1973 and 1974, edited by Robert T. Farrell, 23–31. Oxford: British Archaeological Reports.