Medieval multilingualism is a subject of recent strong interest in historical linguistics. In medieval literary studies it naturally challenges distinctions maintained by disciplinary boundaries with origins in the nineteenth century’s “conception of nation-ness as linked to a private-property language.”¹ But the subject of medieval multilingualism also complicates some intra-disciplinary orthodoxies. For example, much of what has passed for the study of multilingualism in the Anglo-Saxon period in the history of English tends to focus the subject almost exclusively on loanwords and structural correspondences due to language contact, while ignoring other realities of medieval multilingualism.² One reason why scholarship on multilingualism in Anglo-Saxon England tends to be so reductive is an apparently pervasive rigid understanding of “multilingualism,”³ in which evidence of lexical and structural borrowing is the only kind of evidence of multilingualism that matters. But one of the lessons of scholarship on later medieval multilingualism is that broad linguistic competencies have realities different in kind – not just in degree – from those which produce borrowing. The study of borrowing values clear delineation through etymological and structural descent. But the study of medieval multilingualism often suggests that awareness of and use of multiple languages in a variety of texts in the Middle Ages occupied a space not easily explained by the standard descriptive narratives of modern scholarship, because multilingualism’s realities are more than the sum of borrowed parts.

In addition to the indirect evidence of multilingualism through linguistic borrowing, the direct evidence of multilingual encounters in Anglo-Saxon England has drawn frequent commentary in scholarship. But records of interaction between English speakers and speakers of other vernaculars in use in Anglo-Saxon England are exceedingly thin on the ground. Only very rarely do we encounter direct commentary on what must have been common speaker-to-speaker interactions of the sort that Bede describes when he tells us in III.3 of his Historia ecclesiastica that King Oswald of Northumbria (d. 642) mediated the limited English that Bishop Aidan, an Irish speaker “qui Anglorum linguam perfecte non nouerat” (who did not know perfectly the language of the English), used in his preaching, since Oswald had acquired his own excellent knowledge of Irish during his long exile among Irish speakers.⁴ And although Bede names Pictish as one of the five languages spoken in his day, images of the language in the records of Old English are absent except for Bede’s acknowledgment. The records of Old English preserve some eighty occurrences of forms (e.g., Peohta, Pyhtas, Pyhtish, etc.) referring to the Picts, but none – even with the derivational suffix -isc – refers to the language, other than the three forms found in versions of the topographical preface to Chronicle D, E, and F (Pihttisc, Pyhtisc, and Pihtisc, respectively), which, of course, was taken from Bede. Likewise, instances of various forms referring to Wales/Welsh and Britain/British (made distinct in the topographical preface to some versions of the Chronicle, where Bede only makes mention of British) abound in Old English records. But exceedingly rarely do they refer to language, such as when Felix in his Vita S. Guthlaci notes that the devils attacking the saint spoke British, which Guthlac understood since he had lived among British speakers.⁵

Similarly, various forms of Frankish and Anglo-Saxon interaction are known to have taken place from even very early in the period. Frankish people resided in Kent at least as early as the late sixth century, many of them, like Æthelberht’s queen Bertha, of relatively high social status; Bede tells us that the daughter of the Kentish king Eadbald went to a Frankish monastery, a practice of many Kentish Anglo-Saxons at the time; similarities between Frankish and Kentish law suggest further ties; cross-Channel trade remained important even after kinship alliances seem to peter out with the eclipsing of Kent by other Anglo-Saxon kingdoms; and, of course, the export of Anglo-Saxon churchmen like Alcuin and Boniface to the Continent can only have bolstered cross-cultural interactions. But, once again, the records provide little direct commentary on the linguistic realities of such interactions. In fact, most of the Old English records of frencisc – which only once refers to a language – are quite late, so that we are dealing with a linguistic and cultural identity, Norman, quite different from the cross-Channel influence on south-eastern England in the sixth and seventh centuries. In his Enchiridion, Byrhtferth says: “Se ðe his agene spræce awyrt, he wyrcð barbarismum, swylce he cweðe þu sot þær he sceolde cweðan þu sott. Se ðe sprycð on Frencisc and þæt ne can ariht gecweðan, se wyrcð barbarolexin, swylce he cweðe, inter duos setles cadet homo þonne he sceolde cweðan, inter duos sæles” (Whoever corrupts his own language, he commits barbarism, as if he said þu sot [you soot] where he should say þu sott [you fool]. Whoever speaks in French but cannot speak it correctly, he commits barbarolexis, as if he said inter duos setles cadet homo [the man falls between two seats] when he should say inter duos sæles).⁶ But, of course, Byrhtferth is talking about frencisc as the late tenth-century language of his teacher, Abbo of Fleury. Even Anglo-Saxons’ impressions of interchange with Norse speakers seem to be noticeably absent, as Matthew Townend notes that only two references to Norse in England survive: in Æthelweard’s Latin Chronicle and in Ælfric’s De Falsis Diis.⁷

Whatever linguistic borrowing resulted from language contact situations, the plain fact is that they left little impression on the Anglo-Saxons that we can discern from their record-keeping. We can reason any number of explanations for this, including the incomplete and contingent survival of the records. But surely the most fundamental explanation is a philosophical difference, not a material vacuum: present-day English language scholars and the Anglo-Saxon clerics who were responsible for our records do not share idea-making on language. Modern idea-making on language has bred for a vision focused on matters like the negotiation of language differences as actuated on structural and social levels. But, of course, these are ideas well outside the field of vision of medieval ideologies of language difference, which for cleric and layperson alike were mostly invested in a religious grammar for understanding language difference as a form of God’s judgment for mankind’s pride and presumption at Babel. Learned interest in the sacred languages of Hebrew, Greek, and Latin (and even in remote and esoteric languages like Egyptian and what the Anglo-Saxons called indisc – the language of India) registers at least as much of a presence in the records as does interest in vernaculars that could have been heard at market or at court or in other ordinary interactions in many places and times throughout Anglo-Saxon England. Interest in language differences in the records is shaped mostly by clerical ideas instead of by languages on the ground and in use, such as would draw the gaze of modern scholarship. Accordingly, the sacred languages of Hebrew, Latin, and Greek form a much larger piece of direct evidence of Anglo-Saxons’ commentary on language in general. So if we wish to use our records for the direct evidence that they offer on languages in contact in Anglo-Saxon England, we have to re-conceptualize the text as a location where languages meet, where differences between languages are negotiated, and where language relationships are established. And in this imagining of medieval English, one in which clerical interests shape the dialectic of linguistic thought, Latin, naturally, forms the most important multilingual context with English.

I want to suggest that, quite apart from the indirect evidence of borrowing and from the almost entirely absent documentation of cross-linguistic speaker interactions, the records of Old English nevertheless suggest the possibility for a broadly based multilingual linguistic event: preaching. There is only a single statement on vernacular preaching that survives from Anglo-Saxon England, a letter composed by Ælfric at the request of and under the name of Bishop Wulsige and instructing the priests of his diocese to relate the meaning of the Gospels to the people in English on Sundays and Mass days.⁸ However, the majority of the extant corpus is made up of texts that were used for or relate to preaching, and while vernacular homilies, sermons, and other pastoral texts continue to be of intense interest to scholars studying sources, themes, rhetoric, audiences, uses, and stemmatic relationships, linguistic scholarship has mostly overlooked the homily and sermon text as a record of what could be considered the best-documented linguistic event. And since these texts very often mix in elements of the sacred languages (with a strong emphasis on Biblical Latin, naturally) with translation and explanation for audiences that could have ranged from monastic clergy to the laity, they offer us some opportunity for a direct view of a certain kind of multilingual dynamic in Anglo-Saxon England that goes beyond the standard narratives of linguistic borrowing.

Macaronic sermons in later medieval English have been the subject of intense scrutiny for the questions they raise about multilingualism, code-switching, and the linguistic multi-competence of sermon writers and of audiences whose exposure to mixed English–Latin preaching challenges simplistic views of the Latin Middle Ages as rigidly diglossic. In his seminal work on the subject, Siegfried Wenzel argues that the later medieval sermons that he studies, along with some direct commentary on mixed-language preaching, indeed offer evidence of actual macaronic preaching at least to clerical audiences but perhaps, too, to clerical and lay audiences together.⁹ More recently, though, Alan J. Fletcher¹⁰ and Herbert Schendl¹¹ emphasize the mixed-language nature of macaronic sermons as a primary feature of written composition. Still, attention to mixed-language texts has been far more extensively applied to later medieval English than to mixed-language texts from the Anglo-Saxon period.¹² Surely, one reason for the absence of attention to mixed-language texts from Anglo-Saxon England is that there are far fewer of them than can be found in later periods, and this fact has boosted some orthodox thinking: Latin and the vernaculars were much more thoroughly segregated by use in the early Middle Ages than they were in later centuries.¹³ Even if the relative difference in Latinities between the two eras is significant – and there is every reason to believe that it was – the possibility that contexts for Latin and English use in the early Middle Ages are greater than has been suggested is not automatically precluded, even though “orthodox thinking” has a way of shutting down thinking altogether.

Although code-switching has been a subject of study as a feature of Old English charter materials¹⁴ and of a small amount of Old English verse,¹⁵ mixed-language contexts in Old English texts remain largely unexplored from the viewpoints of multilingualism and sociolinguistics. The term “code-switching” itself has so many different applications that its meaning can be too imprecise to describe the mixed-language contexts of the Old English preaching materials,¹⁶ which are decidedly not of the sort of intrasentential switching most often associated with “code-switching” (and with later macaronic sermons). So I wish to avoid further use of the term, which, as Ardis Butterfield says, suggests “a professionalism of approach” that might not be in tune with the medieval texts under discussion here.¹⁷ Nor do I wish simply to register an opinion in the debate on whether mixed-language sermons were composed to be delivered from the pulpit or to be read only by the learned (although it is an issue of importance addressed below). Crucially, mixed-language contexts in the Old English preaching materials do not by themselves say anything about the multilingual competence of the audience of preaching (although they may suggest something of it). Rather, the mixed-language contexts of the Old English preaching materials demonstrate one way in which Latin could have had a much broader reception in Anglo-Saxon England than is usually allowed for under the standard narratives of early medieval Latin’s and vernacular languages’ isolation from one another.

Go a step more: preaching provided a context for transmitting direct knowledge of Latin, and this knowledge – limited though it was – ipso facto could have formed part of ordinary Anglo-Saxons’ metalinguistic knowledge and could even have formed a small part of the linguistic repertoire of Anglo-Saxons beyond the formal education in Latin afforded only to the clergy. In other words, the records provide some evidence for thinking that the artificial linguistic stratum of Latin formed a kind of “contact situation” of its own in the form of mixed-language preaching. Obviously, this is a proposal that strains against the standard view of the uses of Latin and even against the meanings of medieval literacy, since they carry with them the presumption of the inaccessibility of any Latin to all but the clerical class. But it is not implausible that preaching in Anglo-Saxon England established an oral context for laypersons (and semi-literate or poorly educated clergy) to acquire some exposure to Latin beyond liturgical Latin. And, more often than not in the Old English preaching materials, this exposure came with translation and explanation of a sort that would have made preaching a linguistic interchange reminiscent of speaker-to-speaker interactions (where comprehensibility is at issue) that we imagine to have been common but for which we lack extensive evidence. þæt is on englisc is such a frequently encountered locution used to explain Latin in late Old English preaching texts that it suggests that the reception of preaching could well have been a multilingual space for expansion of the linguistic repertoire and for reinforcing an awareness of linguistic differences.

It is widely acknowledged that the Old English preaching materials had three main uses: (1) for preaching to the laity; (2) for use in the Night Office in monastic houses; and (3) for private reading (presumably for clerics but with the rare exception of literate laypersons, too).¹⁸ Where earlier scholarship cast doubt on the idea that homilies proper (i.e., exegetical presentations on Gospel lections) were intended for the laity,¹⁹ more recent scholarship has focused on some of the ways that Old English homilies and sermons were used in the satisfaction of the church’s pastoral obligations.²⁰ But locating Old English homilies and sermons within the sphere of preaching to the laity has often seemed of minor importance in scholarship precisely because of the monastic contexts of our records. There is a somewhat deceptive pseudo-syllogism implicit in the logic of a good deal of scholarship on Old English, in general, that forms something of a cognitive bias: monks were responsible for making and collecting manuscripts; manuscripts preserve our texts; therefore, our texts must be, first and last, of monkish concern. Leaving aside the expanding role of written language in Anglo-Saxon England to preserve matters beyond religious concern and, of course, the question of what constitutes monkish interest in the early Middle Ages, in our modern preoccupation with categorization and boundaries we tend to underestimate the ways that even monastic learning could trickle down to the laity.

Thacker reminds us that through much of Anglo-Saxon history, monasteries were very loosely organized establishments with wide-ranging pastoral responsibilities in many places, the monastic ideal of enclosure notwithstanding.²¹ Blair points out that “Bede’s practice of handing out English versions of the Creed and Lord’s Prayer to ‘many monolingual priests’ rather startlingly reveals a direct channel from the greatest scholar of the age to Northumbrian peasants.”²² Furthermore, lay sponsorship of monastic houses established one channel by which transmission of monastic learning to lay audiences could have taken place – gifts of lands and goods stitched together monastic and secular communities in ways that fostered exchange. And the imitation of monastic discipline in lay households explains Ælfric’s vernacular collection of saints’ lives for his lay patron Æthelweard, which Wilcox theorizes had a broad audience extending to the “non-social elite,” including “[s]ervants and children” and possibly the “full cross-section of village life.”²³

But a wider channel of transmission from monastic learning to broad reception is preaching, for many vernacular preaching texts have their origins in monkish concern for the proper instruction of the unlearned by secular clergy with limited Latin literacy. Townsend cautions us that images of Latin in the Middle Ages as a monolithic proprietary right of a tiny elite minority are illusory and that it is more useful to think of “plural Latinities articulated in concrete relation to the variable circumstances of vernacular culture.”²⁴ There is evidence that Old English homilies’ audiences could be quite mixed, so that “[a] bishop, for example, might compose model sermons ad clerum with the expectation that they would subsequently be preached in some form ad populum.”²⁵ While some homilies seem aimed at monastic audiences and others at lay audiences on the basis of content appropriate to each group, the mixed nature of many texts suggests no obstacle to delivery in some form to lay audiences. To further blur the division between clerical and lay preaching, which is a harder division in the modern imagination than may have actually existed, Hall finds convincing evidence of Latin sermons from post-reform Canterbury intended for lay audiences.²⁶ In short, just as recent scholarship has increasingly viewed preaching and texts used for preaching in the early Middle Ages as evidence of a permeable boundary between clerical and lay pieties, so homilies and sermons therefore may provide opportunities to question the linguistic divisions between clerical Latinity and lay vernacularity.

Also germane to any discussion of preaching as a linguistic event are questions of homilies’ and sermons’ performative aspects. Hall asks,

These are all questions which bear on our interpretation of preaching texts, not as records of speech per se, but as records of texts composed, in part, to be delivered as speeches. Clayton says that the evidence suggests that “Ælfric intended his homilies to be read at Mass by bishops and priests, both secular and monastic,”²⁸ and the language of homilies and sermons often deploys the rhetoric of address, indicating less base texts for preparation than full texts for reading aloud to an audience. Though practices were fluid and how homilies and sermons as written texts related to oral performance no doubt varied considerably, there seem not to be any insurmountable obstacles to considering the homiletic records as in part scripted speech. Swan points out some of the ways that the performativity of homiletic texts constructs speaker and audience identities and ideologies within the context of a “public, oral delivery and reception.”²⁹ And though she is most concerned with ideologies of authority and control, her observations on the performative aspect of preaching surely extend to linguistic idea-making, especially where the rhetoric of comparison of Latin with English pervades the homilies and sermons themselves. Kienzle suggests that modern performance theory can contribute to our understanding of how medieval homilies and sermons might have been delivered, particularly as the conceptual framework for preaching as part ritual and part theater illuminates matters of messaging and preacher–audience identities.³⁰ Most pertinent to the conceptualization of preaching as a multilingual linguistic event, though, is “the relationship between medieval preaching texts – by far our most plentiful surviving evidence for preaching – and actual medieval preaching events, or performances, of which we have only relatively few, indirect, descriptions, and, of course, no direct experience.”³¹ These are the questions to which I will turn now: how and to what extent do Old English preaching texts perform multilingualism?

Old English preaching texts – homilies and sermons (though scholarship does not always observe a strict separation of terminology for exegetical and non-exegetical preaching texts) – form the largest coherent group of extant Old English texts. They occur as compositions attributed to Ælfric (c. 950 – c. 1010) and to Wulfstan (d. 1023) and as some 140 anonymous texts of mostly indeterminate date, although the Vercelli Homilies along with the Blickling Homilies are major collections that generally point to a date of composition somewhat earlier than those of Ælfric and Wulfstan.³² Across these four groups of preaching texts – leaving aside the precise uses and audience of any individual text but allowing that homilies and sermons existed in large part for oral delivery, as the arguments above demonstrate – there is an extensive pattern of Latin and English immediacy, adjacency, and juxtaposition in which relationships between Latin and English are spelled out for listeners, unlike the ordinary of the Mass or other liturgical ceremonies exclusively in Latin. The positional rhetoric of English and Latin is inherent to the form, of course, because homilies (properly) expanded on the pericopes found in Gospel lectionaries and other service books.³³ So a typical formula is to recite the opening of the pericope in Latin, then translate the passage in English and launch an extended explanation of its meaning and importance.

But beyond the requirements of the form, preaching texts in Anglo-Saxon England often demonstrate a marked and natural English–Latin bilingualism. Part of this bilingualism is surely aimed at the many poorly educated Mass priests who required translation and explanation to be able to deliver accurate expositions of scripture to lay audiences, as Ælfric says in the Preface to his translation of Genesis in admonition of “ungelæredan preostas” (unlearned priests), who, understanding a little Latin, miss larger and more critical senses of meaning.³⁴ But this bilingualism fell, too, on the ears of homilies’ audiences when the texts were used for public preaching, and the pattern of quotation in Latin followed by translation in English can only have worked to promote the kind of common multilingual environment in the early English Middle Ages that recent scholarship has labored to bring to the fore. As Tyler says, “part of the symbolic value of multilingualism is that it reacts against dominant paradigms” of, in this case, language history in England, which reflexively emphasizes monolingualism and an insuperable divide between elite users of Latin and the broader linguistic repertoire of ordinary Anglo-Saxons.³⁵

Let us consider the Latin–English bilingualism of the homiletic collection in the Vercelli Book, a manuscript from the second half of the tenth century. Scragg says that “the material’s practical use for preaching was not significant in the compilation and ordering” of the manuscript, since it is a mixture of poetry and prose and since many of the prose items mix elements of the sermon, homily, and saint’s life. These are the features which distinguish the manuscript as different from continental or English books compiled solely for preaching.³⁶ Even though the manuscript appears not to have been compiled for the use of a preacher, rubrication indicates that many of the homiletic texts that it preserves were copied from proper homiliaries.³⁷ Therefore, the manuscript may well offer versions (produced by a single scribe copying rather mechanically³⁸) of preaching texts that were in circulation through oral delivery. In any case, the standard pattern of Latin–English bilingualism in the Vercelli prose texts is the presentation of partial Latin and an immediately following English translation when the Latin text quoted appears in direct speech in the Bible. Take the three following examples of the predominant type of Latin–English bilingualism in the Vercelli texts:

In these and in some four dozen other similar occurrences of Latin in the Vercelli prose texts, the speech of a biblical figure – very often Jesus, as one would expect – is partially rendered in Latin and then treated in English. To be sure, other types of Latin–English bilingualism occur in the texts, such as in Homily XI when the homilist quotes and translates Psalm 125:5 in a passage on ultimate rewards after a life of struggle,⁴² or in Homily II when the homilist quotes the words of Jesus from the Sermon on the Mount in Luke 6:37, “<Dimittite> & <dimittetur> uobis,” but leaves them untranslated (in what might be considered a case of intersentential code-switching).⁴³ But the standard pattern of partial duplication of Biblical Latin followed by relatively close English translation does suggest something of the possibility of the performance of multilingualism in preaching: the preacher performs the “roles” of speakers from Jerome’s Vulgate by quoting them in Latin and translating (and expanding) their speech for the audience. Furthermore, the arrangement has the effect of linguistic framing, in which the relationship between English and Latin is not declared but performed: Latin is the “authentic” voice from the Bible, and English is the substitutive medium; the prestige of Latin is mediated by the utility of English; and the preacher deploys “expected patterns of language”⁴⁴ to establish ritualistic efficacy in which the judicious use of Latin formalizes and “theurgizes” the vernacular context of the English homily. Scragg notes that errors in the Latin quotations in the Vercelli Homilies are “often so gross as to produce sheer nonsense,” which he attributes to the single scribe’s inability “to recognize or correct mistakes in his Latin.”⁴⁵ But this is only an observation of scribal transmission and does not disrupt the interpretation of the pattern of Latin–English bilingualism in the Vercelli Homilies (or, more properly, in the sources of the Vercelli Homilies) as evidence of a kind of staged multilingualism in preaching.

The Blickling Homilies, another important group of anonymous homilies, in a manuscript from the end of the tenth century, also demonstrate a pattern of quoted direct speech from the Bible and from other Latin sources (like apocrypha and saints’ lives) rendered in Latin and then in English. But, on the whole, they deploy Latin quotations much less often than the Vercelli Homilies. Still, the pattern occurs more than twenty times across four of the eighteen homilies in the collection. For example:

Again, even though the use of Latin quotations is a less pervasive feature of the Blickling Homilies, the strategy the texts adopt is very similar to that observed in the Vercelli Homilies – the Latin–English bilingualism largely (but not exclusively) supports the performance of the speech of biblical or saintly persons. In the context of preaching, of course, the reality of performing multilingualism takes on meanings for delivery and reception quite different from encounters with these texts through reading. The preacher “enacts” the bilingual experience between English and Latin for the aural reception of his audience; for an individual reader of these homiletic texts, the experience of their bilingualism lacks the performative aspects framed by preaching in a ritual setting by an authority using expected patterns of language.

Gatch observes that “it is not necessarily the case that authors will tailor their materials to the special needs and conditions of those who will read their writings or hear them read. The nature and purpose of the Anglo-Saxon sermon books are very difficult and complex matters.”⁴⁹ But a pattern as clear as the Latin–English bilingualism of preaching texts certainly suggests design; it is even possible that the type of Latin–English bilingualism that I highlight here reflects a part of Anglo-Saxon homilists’ rhetorical strategy for preaching specifically because it is laden with performance value in oral delivery. There, the preacher “plays” both the speaker of Latin and the speaker of English and negotiates comprehension for the audience (and it is not hard to imagine comprehension aided by the preacher’s intonation and gesturing). Uncertainty about whether or not “authors will tailor their materials to the special needs and conditions” of their audience does not interdict texts’ reception dynamics. Through vernacular preaching in which there is a widely embedded Latin–English bilingualism, a clerical or lay Anglo-Saxon audience receives messaging on the relationship of Latin to English, both in its prestige hierarchy and in its structural correlations.

Composition of homilies (in the term’s strict sense of exegesis) in the vernacular was extremely rare in the early Middle Ages except, obviously, in Anglo-Saxon England.⁵⁰ If the preaching texts in the Vercelli and Blickling collections are as late as the mid tenth century and last quarter of the tenth century, respectively, it is possible that they reflect something of the earliest efforts of the Benedictine reform movement. After the Benedictine reforms, one of the more robust circumstances of vernacular culture was an increased emphasis on instructing the laity. Ælfric is the model in Anglo-Saxon England for monastic involvement in activities related to the instruction of the laity through preaching; his letters, prefaces, and homilies demonstrate acute sensitivity to the unlearned audience. The pattern of Latin–English bilingualism observed in the Vercelli and Blickling homilies expands and deepens in the fabric of Ælfric’s preaching texts to such a degree that we cannot but imagine that the performance of multilingualism through preaching formed a context for the development of ordinary Anglo-Saxons’ linguistic consciousness, i.e., their ideas about language alternatives, language differentiation, and language variety. Metalinguistic knowledge underwrites speech communities’ broad linguistic mentalities, as early modern lexicography and orthoepy demonstrate in the history of the English language. If preaching represents a fairly wide channel of transmission of religious thought, its Latin–English bilingualism could represent a channel of (passive) transmission of metalinguistic thought through the presentation, juxtaposition, and negotiation of the vernacular and the prestige codes.

Ælfric very often translates and explains entire Latin phrases from the Bible (and from other sources) instead of simply rendering them in English, a strategy that provides side-by-side comparison for listeners. And, unlike the Vercelli and Blickling Homilies, he frequently employs rather pointed discourse markers, like þæt is on englisc or þæt is on urum geþeode, before rendering his preceding Latin quotation in the vernacular. What the pattern of Latin–English bilingualism in Ælfric suggests is rhetoric for oral delivery – readers of the texts would have the benefit of adjacency and backtracking to aid their comprehension without such distinctive discourse marking. But constructions like þæt is on englisc, þæt is on urum gereorde, or simply þæt is signal to an audience that a translation of the Latin just spoken follows; as metadiscourse, such phrases would make less sense otherwise. To be sure, the use of phrases like þæt is on englisc is not an Ælfrician innovation; Alfred uses the phrase twice in his translation of the Cura Pastoralis to explain “sacerdas” as “clænseres” (priests) and “Galað” (Gilead) as “gewitnesse heap” (heap of testimony). But, certainly, these phrases occur far more often in Ælfric’s homilies than anywhere else in the corpus of Old English texts. These usages, again, indicate genre-specific formulations, which I would suggest is due in part to their use as discourse markers in preaching to help the audience bridge the divide between Latin and English.

Consider the following examples from Ælfric’s Catholic Homilies I and II collections:

There are sixty-seven constructions similar to these examples in Ælfric’s Catholic Homilies I and II, in addition to another twenty in which the English translation follows the Latin quotation without an introductory marker (in the manner of what we observe in the Vercelli and Blickling Homilies). There are also a number of occasions when Ælfric explains individual words from the sacred languages of Hebrew, Greek, and Latin or when he explains names, as in his homily on the nativity of St. Paul: “Annanias is gereht on ebreiscum gereorde scep” (Annanias signifies “sheep” in the Hebrew language).⁵⁵ And while quotation-translation of direct speech from the Bible often occurs in Ælfric, as in the third example above from John 8:46, it is far less rigid a pattern of Latin–English bilingualism than that found in the Vercelli and Blickling Homilies. Indeed, the Latin–English bilingualism in Ælfric applies not necessarily to direct speech but also to quotations found in saints’ lives, exegetical tracts, and the Bible. On the whole, there is a much wider use of quoted Latin in Ælfric’s two series of homilies, and a greater emphasis on Latin–English bilingualism through the use of þæt is … constructions than can be found in two important sets of homilies that scholarship generally judges to be somewhat earlier than Ælfric’s work. The quotation of a wider range of Latin sources and the systematic linkage of quotation with translation with the repeated þæt is … construction surely reflect the fact that Ælfric’s homilies are more exegetical than those found in the Vercelli and Blickling collections. But it also may suggest one of the ways that Ælfric sought to improve the quality of available preaching texts, since the effect of his method is a rhetorical shift in which the Latin–English juncture is spelled out in the text through discourse markers instead of through the possibly uneven and ineffective performance of a preacher. Ælfric’s pattern of Latin–English bilingualism may represent something of a post-reform modification of the bilingualism standard in early or pre-reform preaching texts like those found in the Vercelli and Blickling Homilies. In any case, there seems to be a clear shift to the use of þæt is … constructions for the expression of bilingualism in preaching texts after Ælfric.

This pattern of quotation-translation extends to Ælfric’s Lives of Saints collection, which was composed very probably so that Ælfric’s lay patron, Æthelweard, and his family could imitate at home devotions to saints observed in Benedictine houses. As stated in note 23, the collection also could have found a very wide audience beyond his patron’s immediate household; þæt is … constructions in fact occur some thirty times in Ælfric’s Lives of Saints. The pattern is also (rather less extensively) found in the preaching works of Ælfric’s contemporary Wulfstan (especially in his eschatological works), for example in his sermon on “The Last Days”:

In other words, though our records from Anglo-Saxon England say so very little about what we increasingly imagine to have been a multilingual environment between groups of speakers on the ground that was adequate to promote interaction and, thus, various language-contact phenomena, they nonetheless inscribe quite a full picture of the exchange between Latin and English. This is true in the context of reading, of course (consider glossed texts like psalters, for example), but, by extension, in a context of spoken language, too, since so many texts designed for preaching, as we have seen, emphasize the Latin–English interface in direct ways that cannot be found elsewhere in the corpus, excepting grammatical texts composed for direct instructional use in Latin in the cloistered environment of religious houses. Texts used for preaching in the Middle Ages “spoke” to audiences through the oral delivery of clergy, so homilies and sermons that mix Latin and English – and emphasize their relationship through formulaic phrases like þæt is on englisc – form a multilingual space for Anglo-Saxons that is, in fact, more substantially documented than the other presumed encounters of English with other languages in Anglo-Saxon England. An understanding of preaching as a multilingual space problematizes the usual conception of Latin as the exclusive provenance of learned clergy, for the context of preaching as an oral medium bridges the gulf between the unlearned and illiterate in the Middle Ages and the Latin language, especially since preaching often occurred before a mixed clerical and lay audience. It also complicates the notion that only lexical and structural borrowing provides evidence of multilingualism. A broader conception of multilingualism, one which includes the scripted and ritualistic language of preaching, brings into view a wider range of multilingual encounters than a rigid focus on borrowing allows.

The Latin–English bilingualism of the preaching texts briefly surveyed here suggests: (1) that preaching formed not simply a textual but an oral context for the delivery of English in contact with another language; (2) that vernacular homiletic composition in its earlier stages focused on quotation and translation of direct speech from the Bible; (3) that quotation and translation of direct speech in homilies boost the performative aspects of preaching as a ritualized setting and in which the preacher “enacts” the bilingual experience for the audience; (4) that the linguistic framing of Latin and English in comparison worked to help establish the hierarchical relationship between the two, in which Latin is naturally the higher-prestige, authentic voice; and (5) that expansion of Latin quotation and use of þæt is … constructions in later homiletic compositions (especially as found in Ælfric’s works) express a shift in the rhetorical strategies of performing multilingualism in preaching. “Imagining” English in the Middle Ages invokes an inventiveness about how we frame the language in the distant past as well as a resistance to static thinking and standard narratives. Multilingualism, by itself, complicates our modern instinct for categorical reification precisely because it represents a blurring of distinctions, an attrition of the structural logic that conditions our images of language in the Middle Ages. In our enthusiasm to value the exceptional vernacularity of Anglo-Saxon England, we may sometimes overlook the ways that English and Latin together cast a kind of language contact and a kind of multilingualism that are not in evidence in the standard accounts of English language history.

Languages and literatures do not develop in isolation from each other; and that is all the more true when speakers of different languages live in close proximity, or when two or more languages are spoken by the same person. The linguistic environment of Middle English is especially interesting and complicated, not only because of the many languages it was in contact with but also because of the many variables involved in language choice. In medieval England the language(s) you spoke depended on social status and mobility, profession and business, location and education. All these variables were in turn historically contingent and variable across time, for by the end of the fifteenth century the linguistic landscape had changed almost beyond recognition from that of the early twelfth century. This chapter provides a history of these variables by looking across the wide spectrum of languages spoken and written in England from 1100 to 1500. These include not only the Big Three (English, French, and Latin) but also the Celtic languages and early Scandinavian.

A convenient starting point is the Norman Conquest, which in many histories of England features misleadingly as the moment when the national mother tongue became subservient to the language of French colonizers.¹ Like many other myths, this one has its origin in the medieval period. Writing in the first half of the fourteenth century, the Dominican friar Robert Holcot looked back at the Norman Conquest in terms that have become very familiar:

What Holcot says about his own times is confirmed by other sources. Before the Conquest, English and Latin had been the languages of the classroom and the law courts,³ but after 1066 French became the language of pleading at court (see below), and anyone “put to letters,” that is, educated in Latin, also had to cope with French, since that was the language in which Latin was normally construed. This situation was not to change until after Holcot’s death in 1349.⁴ However, the myth that the Norman Conquest involved the imposition of a foreign language on a nation previously united by its linguam saxonicam is an anachronism based on an association between nation and national language that only became thinkable around the fourteenth century,⁵ when English alone was left as the dominant vernacular of medieval England.

The generations of English people who lived during and just after the Conquest had different ideas about language. For instance, instead of the conspiracy theory that William the Conqueror wanted to destroy English, Orderic Vitalis (c. 1075-1143) says that the king intended to learn English, so that he could understand the people he was to govern, but was just too old and busy to keep his resolution.⁶ These generations would also have understood that the Norman Conquest merely accelerated a “process of ‘normanization’” that predated the Conquest.⁷ King Edward the Confessor, the “English” king from 1042 to 1066, was the son of a Norman queen (Emma), having grown up in Normandy, whence he returned to England with a large French-speaking entourage. Alongside the French presence at court, English co-existed with various other spoken vernaculars. Only a few decades before William the Conqueror, England had been ruled by a dynasty of Danish monarchs founded by King Cnut (d. 1135), and at the time of the Conquest Old Norse was still spoken in northern parts of Britain.

The oldest indigenous languages at the time of the Conquest, however, were the Celtic ones. Before the fifth-century diaspora of Germanic-speaking tribes, three of these Celtic languages – Welsh, Cornish, and Cumbric – had formed a unified language, Brittonic (or Brythonic). As place-names indicate, Brittonic was spoken as far north as Dumbarton (< Dun Breatann, “fort of the Britons”) in Scotland, and down from northern England (Cumbria, related to Cymru) to the southernmost tip of Cornwall.⁸ However, by the eighth century, when eastern and central parts of Britain had been thoroughly anglicized, what had once been a unified language had been broken up and receded into the west – Welsh into Wales and Cornish into Devon and Cornwall – and the north, where Cumbric lingered on in areas of Cumbria and the Scottish lowlands – notably Strathclyde – until the early twelfth century. Further north in Scotland, Gaelic was spoken, this language descended from Goidelic, a different branch of the Celtic language family. According to a geographical treatise, De situ Albanie, the dividing line separating the Gaelic-speaking area from the rest of Scotland was the Forth (north of Edinburgh). The relevant passage is worth citing in full because, in the course of naming the Forth in three languages, the author reveals the complexities of multilingualism:

Forth and Werid are the Gaelic and Cumbric names respectively; and, although Scotwater is immediately recognizable as an English name rather than a French one, the writer was not confused. He was evidently a French speaker,¹⁰ who made use of an English name when no French one was available. By the time this was written, c. 1170, Cumbric had almost certainly died out (though some old names were still being remembered), and the boundary separating Gaelic and English speakers lay along the Forth; in the centuries to come it was to shift ever further north, until at the end of the medieval period it more or less coincided with the natural border separating the highlands from the lowlands.¹¹

Unlike Cumbric, Cornish survived beyond the Middle Ages, but little literature was produced in Cornish after this period, and it had died out as a community language by the end of the eighteenth century.¹² However, in the eleventh and twelfth centuries it was spoken both in Cornwall and in areas of Devon. For instance, a royal charter datable to 1156 or 1157 is addressed to the king’s “fidelibus suis francis et anglis et wallensibus Cornubie et Deuonie” (sworn men, French and English and the Welsh of Cornwall and Devon): “wallensibus” here does not here mean “natives of Wales” but rather “Welsh speakers,”¹³ for Cornish and Welsh were not always perceived as different languages, even by those proficient in the language.¹⁴ By 1500, English had ousted Cornish from all but the western end of Cornwall.¹⁵

Welsh was spoken not just in Wales but also in areas bordering Wales.¹⁶ For much of the period, Wales was only half conquered, and even the conquered parts were only nominally under the direct control of the English crown. Until the reign of Edward I, the business of conquering and colonizing Wales was left to the private enterprise of French-speaking marcher lords, and wherever they went English-speaking settlers followed. A century or so after the Conquest, French and English had penetrated into all areas of Wales conquered by the Normans, as we learn, for example, from the works of Gerald of Wales, born in Manorbier (Pembrokeshire), but himself a French speaker, Paris-educated, with excellent Latin and some rudimentary Welsh and English.¹⁷ The second redaction of Gerald’s Journey Through Wales (c. 1197) tells the story of Henry II’s encounter with an old man in Cardiff, on the king’s return from an expedition to Ireland.

The prophecy (relating to Henry II’s betrayal by his sons) may well be Gerald’s invention but the circumstantial detail is plausible. Philip de Marcross (a small town in the Vale of Glamorgan) appears in local charters,¹⁹ and it is not unlikely that a landowning knight would be bilingual in French and English and so be able to act as an interpreter. King Henry, as another witness, Walter Map, confirms, only spoke French and Latin, but according to Map he could understand various other languages, including English – which is why the peasant is able to speak to him directly while the King’s answers to him (in French) are related back in Philip de Marcross’s English translation.²⁰ That the peasant speaks English is another realistic detail: this part of Wales had been subdued, and English was now “overwhelmingly the language of the small boroughs which came to pepper the coastal lowlands of south Wales.”²¹ In higher circles, however, French remained de rigueur, and it is revealing that here and throughout his writings Gerald thought it necessary to translate English quotations into Latin. By contrast, French quotations in his works are not translated. In the circles where Gerald moved, knowledge of French and Latin could plainly be taken for granted, but knowledge of English could not.

The impact of Welsh on the vocabulary of Middle English was minimal (and this is true for all the Celtic languages in general), but some significant cultural and linguistic transfer from Welsh to English took place in the Welsh marches. A number of Welsh loanwords appear in poetry and prose from this area. Examples are baban (Welsh for “baby”) and cader (“cradle”) in Ancrene Wisse, Hali Meiðhad, and the Lincoln’s Inn manuscript (from Shropshire) of The Siege of Troy; genow (Welsh genue “mouth”) in Seinte Margarete; keis (“retainers”) in Sawles Warde;²² while the Harley lyrics, from Herefordshire, contain the Welsh borrowings miles (< mil “wild animal”), wolc (< gwalch “hawk”), and crouþ (< crwth).²³ Only the last of these, first recorded in the Harley lyrics, gained general currency (OED, s.v. croud), no doubt because it referred to a stringed instrument that was peculiar to Wales and for which no word existed in the Germanic or Romance lexicon.²⁴

Whether the writers and original readers of these West Midland texts would also have regarded these words as Welsh “borrowings” is another question. Probably, some of these words had become naturalized in the dialect of a region where native Welsh and English speakers had long mingled and where English–Welsh bilingualism must have been common. Charter evidence from Hereford indicates the presence of Welsh speakers amongst all classes of society, including the landowners who attended the shire court;²⁵ and even after the whole of Wales had been annexed by the English crown, Welsh-speaking gentlemen from the area continued to commission poetry in Welsh.²⁶ The fifteenth-century poet Guto’r Glyn addressed a number of his poems to Henry Griffith, who lived at Newcourt in the parish of Bacton.²⁷ Since it was the duty of clergy to catechize and preach to the laity in their mother tongue, some proficiency in Welsh amongst clerics active in the Welsh borders was to be expected. We have some interesting proof of this from the records pertaining to the canonization of St. Thomas of Cantilupe (Bishop of Hereford) in 1320.²⁸ In preparing the case for his canonization as a saint, a papal delegation, headed by the Bishop of Mende, was dispatched to England and Wales to interview locals who had witnessed the miracles following his death. Unusually, the documents specify the language used in the depositions. In Hereford and Conway, the delegation – speaking French and Latin but equipped with English-speaking interpreters – found communication easy. The local clergy spoke either Latin and French or both. About half of the laity in Hereford and Conway could also answer in French (a few even used some Latin), while the other half replied in English – though competence in French was significantly lower in rural areas. But in Swansea (where again half the laity replied in French and half in English) the delegates encountered a monolingual Welsh speaker who could neither understand them nor make himself understood. Their solution was to look for Welsh interpreters, and these were eventually recruited from the Franciscan monastery in Hereford, where two brothers were found who had been born in Wales and were fluent in Welsh.

It is not surprising in this context that some Middle English poetry from the Welsh marches seems to show influence of the Welsh poetic tradition. Of particular interest is the use of consonance in a Harley lyric, originating from Herefordshire, known as The Poet’s Repentance,²⁹ which begins:

In addition to linking stanzas by means of concatenation, as illustrated by the last line of the first stanza and the first line of the second (e.g., “Ywis hit is al wrong. / Al wrong Y wrohte for a wyf,” lines 12-13), the poet has linked consecutive line ends by means of pararhyme (wet/wyt, bet/byt, etc.). This use of pararhyme is rare in Middle English. To my knowledge it occurs in only two other poems, both also from the Welsh marches: Three Dead Kings (which also employs concatenation) and Pater Noster.³¹ However, it may owe something to Welsh verse.³² Thus the requirement in englynion proest that rhyme words must end in the same final consonant, but in a different vowel, not uncommonly results in pararhyme, as in the following stanza by the above-mentioned Guto’r Glyn:

As in The Poet’s Repentance and Three Dead Kings, so in this Welsh poem the play of consonance is accompanied by stanza concatenation (known as cyrch-gymeriad in Welsh).

Before leaving the topic of Welsh, we should not forget that Welsh speakers could, and did, contribute very directly to Middle English literature by writing in English. The works of Chaucer’s contemporary John Clanvowe, a Welsh expatriate,³⁴ are a case in point. But the most remarkable Welsh contribution to Middle English poetry is surely the fifteenth-century Welsh–English “Hymn to the Virgin.”³⁵ According to a prologue to the poem, the poet Ieuan ap Hywel Swrdwal composed it in Oxford as a rebuke to a couple of fellow students (both English) who had dismissed the Welsh tradition as far inferior to English or Latin verse. To prove the contrary, Ieuan wrote an English poem to the metrical rules of Welsh verse. It begins as follows:

The reason why the poem looks so peculiar is that it was written according to Welsh rules of phonology and orthography. Thus <dd> represents English <th>; <w> represents [u]; single <f> is [v], and [j] is lost before front vowels (hence i for ye). The scribe of the prologue (from London, British Library, MS Add. 14866) alerts his readers to this issue: “Ond am fy mod i’n scrivennu’r lyffr hwn oll ag orthographi kymbraeg e gaiff hyn o saesneg ganlyn yn llwybr ni: darllenwch hi val kymbraeg” (But because I am writing this book in Welsh orthography, this much of English can follow our manner; read it as Welsh). The scribe, this implies, could have written this in “normal” English, and was writing for people accustomed to reading such English. However, in Ieuan’s poem and this scribe’s copy, English was for once made subservient to the laws of Welsh.

Having dealt briefly with the Celtic languages, I now come to the languages that had a far greater impact on English: Old Norse, French, and Latin. The vast majority of Old Norse words entered the English language in the period 1150-1299. In any total tally of loanwords, Old Norse certainly lags well behind French and Latin, but, as Philip Durkin points out, if we narrow our focus to a shortlist of the hundred most common English words, Old Norse edges ahead of French and Latin.³⁶ High-frequency loans include take, they, hit, as well as words such as give and egg for which Old English had variant forms with /j/.

Whether any Old Norse would still have been spoken in the Middle English period depends on what arbitrary date is selected as its beginning. If 1066, the answer is certainly yes; if 1150 the answer is probably not. Some historical context will be useful.³⁷ Viking raids on England (and Wales and Scotland) began in the eighth century and were soon followed by conquest and large-scale settlement. The resulting presence of a community of people who spoke a different Germanic vernacular, Old Norse, was officially recognized in 886 by the establishment of the Danelaw. The border of this territory ran roughly along a diagonal line from London in the east to Chester in the west. North of this border Vikings were free to settle and live by their own laws. The area where the Old Norse element penetrated most deeply into English is known as “the Great Scandinavian belt,” from Cumberland and Westmoreland in the west to the North and East Ridings of Yorkshire and Lincolnshire in the east, where Scandinavian affected not just lexis but also grammar (as reflected in, for instance, the construction at + infinitive).³⁸ Although the defeat of a Norwegian army by Harold Godwinson in 1066 marked the end of the Viking Age, evidence from charters, place-names, and inscriptions suggests that Old Norse survived in parts of the Danelaw until the early twelfth century.

How is the demise of Old Norse in the first half of the twelfth century to be reconciled with the fact that it began to exert its greatest influence on the English lexicon around 1150? Bearing in mind the inevitable time-lag between spoken language and the written record, the paradox shows that the agents of linguistic change were not English speakers borrowing words from an influential language but rather Scandinavian speakers abandoning Old Norse in favour of English and carrying over Old Norse words (and to a lesser extent grammatical structures). The situation is thus not really one of borrowing but rather one of “language-shift-induced imposition,” that is, the importation of linguistic elements by speakers shifting from one language to another.³⁹ In the case of Old Norse speakers, this language shift was facilitated by the fact that English and Old Norse were closely related: indeed, speakers of these two languages were probably mutually intelligible, as no records of encounters between Vikings and English make mention of communication problems.⁴⁰

The perceived standing of English in relation to French differed crucially from that which it enjoyed with respect to the other languages discussed above: vis-à-vis the Celtic languages, English stood on a superior footing, and Old Norse was at best an equal, but French was a language of prestige. The reasons for this were historical. While the numbers of invading Frenchmen were insignificant compared with the size of the population they colonized, the invaders took control over secular and ecclesiastical government, and their language, as that of a ruling elite, became a badge of cultural superiority. The symbolic significance of French also explains why the writing and speaking of it were kept up in England long after it had ceased to be anyone’s mother tongue, and why the manner in which one spoke French (“vulgarly” or “properly”) carried considerable social significance.

The question of when French ceased to be a mother tongue in medieval England has been much debated. According to Douglas Kibbee and William Rothwell, it had ceased to be a living vernacular by around 1200.⁴¹ From then on it was, so they argue, an “artificial language,” one which, like Latin, had to be learned from books and teachers. However, there is good evidence that, well into the fourteenth century, children of the higher echelons of society acquired French at an age young enough to speak and write it with native-like competence.⁴² Higden’s Polychronicon (1350) provides some revealing evidence about the contemporary linguistic situation. I cite John Trevisa’s translation:

There is mention of “teaching” here, but there is no question of book learning: “gentilmen children” have learned to speak French before they are old enough to go to school. For them, it is a cradle-tongue, and according to Higden it is only “vplondisshe men” who strive “wiþ greet besynesse” to acquire it.

There were various ways in which “gentilmen children” could acquire French. Exposing children to French in the family household was one way. Walter de Bibbesworth’s Tretiz (c. 1280) provides some insight into the process. Walter’s aim was to introduce his readers to the specialized language of estate management, not to basic French, which he says is known to everyone:

To avoid misunderstanding, when Walter says that “everyone” has basic French he is restricting himself to the landowning classes and those who could read (those who worked the land were illiterate and ignorant of French). Walter advises ladies to speak to their children in French from the cradle. To ensure that children acquired proper grammar as well as vocabulary, he advises ladies to use the noun with the correct article: the system of grammatical gender in French could thus become second nature. Sons of the well-to-do were also sent abroad for their formation or placed as pages in aristocratic courts, which were predominantly Francophone until the late fourteenth century.⁴⁵ The preferred option for daughters was to board them at a nunnery, where they received an education in good manners, which included the accomplishment of speaking and writing in French. Chaucer’s courtly Prioress learned her French in this fashion (Canterbury Tales I.124-5).

In these conditions, French continued to thrive as a second vernacular, remaining until the mid fourteenth century not only the preferred language of literary entertainment for the upper classes but also sufficiently dominant and vigorous to take on a range of other functions previously monopolized by Latin: the writing of letters, the keeping of records and accounts, and last but not least the law. The hegemony of French as the viva voce language in legal courts was challenged by the Statute of Pleading (1362), which permitted English to be the spoken language in royal and baronial law courts. In practice, however, knowledge of French remained a necessity for anyone going to court, not only because the legal profession, trained in law French, jealously guarded its turf,⁴⁶ but also because English did not have the lexical resources to take over.⁴⁷

The continued vitality of French in England also explains why the language evolved, undergoing some of the same syntactic changes as did continental French while also developing characteristics of its own that marked it out as a peculiarly insular variety.⁴⁸ There is much comment, both from insular writers and continental ones, about the “bastardized” varieties of French that flourished in England. Gower self-consciously apologizes for the French he acquired in England, though in fact it is excellent and very close to continental norms.⁴⁹ A number of fabliaux poke fun at Englishmen speaking French;⁵⁰ Walter Map ridiculed the spoken French of Henry II’s bastard son Geoffrey as “Marlborough French,”⁵¹ people from Marlborough being famed for their ineptitude.⁵² Chaucer’s dig at the Prioress’s French “After the scole of Stratford-atte-Bowe / For Frensch of Paris was to hire unknowe” (Canterbury Tales I.125-6) also belongs to this tradition.

Many English people were well qualified to hear the differences between the “French of Paris” and Anglo-French, either because they had travelled abroad or because they had studied in France, where the University of Paris, in particular, attracted many intellectuals. John Barton, author of a fifteenth-century French grammar, mentioned his Paris education to advertise his credentials as someone well qualified to impart “proper French” (droit françois), that is, la droit language de Paris.⁵³ Yet whatever the real dialectal differences between the “French of Paris” and insular French, the snobberies and insecurities about the “purity” of one’s French also bear witness to an imagined “perfect” language which really existed nowhere except in the superegos of those who aspired to it.⁵⁴

This attitude to “pure” French is relevant not only to our understanding of Anglo-French writing but also to our appreciation of monolingual English works, as Chaucer’s works show. The Tale of Sir Thopas is a parody of the Middle English popular romances and their “drasty rymyng” (VII.930), but the humor works in part by appealing to our shared sense of social superiority, our ability to discriminate between what is, culturally speaking, “proper” and what is not. One sensitive area in which Chaucer exercises this acquired knowledge is pronunciation. For instance, in Anglo-Norman the French vowel [y] (as in tu) had fallen together with Old English [u], still pronounced [u] in Middle English, as in OE hus, ME hous.⁵⁵ The pronunciation was ridiculed in French texts (some of insular origin) that show Englishmen making all kinds of mistakes, including confusing fut (was) with fout (fucks).⁵⁶ Chaucer, who had traveled widely and had learned high-class French as a page at the household of Elizabeth of Ulster, avoids such howlers and never rhymes French-derived [y] with native [u] except in Sir Thopas, which blithely rhymes armoure with sowre (VII.819-22) and cote armour with lily flour (VII.866-7). We are, of course, meant to snigger at these rhymes, because no one with “proper” French would pronounce armure as armour.

Another characteristic of Anglo-Norman was the loss of final-e (e.g., in French belle), still pronounced as schwa in continental French.⁵⁷ Chaucer, too, pronounced his final-es in French-derived words – except, again, in Sir Thopas, where, for example, plas (French place) is made to rhyme with was (VII.778) and entent (< French entente) with verayment (VII.2). Again we are meant to laugh: though such rhymes are widespread in Middle English and reflect a perfectly normal pronunciation of French words by an English tongue, some knew that a “proper” Frenchman would not have used them.

The insistent use of Sir before proper names (as in “Sir Thopas”) is similarly non-U. It had been normal in Anglo-French and Old French carried on the practice,⁵⁸ but in refined French writing of the fourteenth century it had become déclassé. Jean Froissart, for instance, only uses appellative sir with names of people from the lower orders.⁵⁹ Chaucer also knew better than to call knights sir (Arcite, for instance, is never “Sir Arcite”). And so when Thopas is repeatedly referred to as “Sir Thopas,” a discerning audience au courant with “proper” French usage concludes that he is not a proper knight but a knight as a burgher might imagine him (as indeed Chaucer insinuates in various other ways). In matters of pronunciation and taste more generally, Chaucer closely identified himself with his image of a refined French gentleman, and winced at anything such a gentleman might have associated with insular usage.

According to recent research by Richard Ingham, the demise of French as a second vernacular in medieval England took place after the Black Death in the mid fourteenth century. His findings show, for example, that after 1350 Anglo-French scribes lost mastery of the French system of grammatical gender which scribes of an earlier generation controlled with native-like competence.⁶⁰ Ingham’s linguistic evidence is corroborated by other types of evidence. William Langland, for example, complains about a new generation of clerics (“thise newe clerkes”) who cannot construe “but [except] in Laytyn or Englissh” (Piers Plowman, B.15.372). The context for this remark is the adoption of English as the medium for Latin instruction. (Langland was not in a position to know that this “new” development was merely a return to pre-Conquest practice.) John Trevisa credited the change to the innovative methods of the schoolmaster John Cornwall, author of a Latin Speculum grammaticale, which does indeed include more English glosses and translations than any grammatical work since Ælfric’s Latin–English grammar,⁶¹ but the shift to English in schools must have been gradual and organic. The only reason it was ever practical to teach Latin in the medium of French is that most children privileged enough to receive an education had some French already. When that was no longer the case, schoolmasters naturally had to change their methods. In the same period, King Edward III is said by Froissart to have urged the English nobility to help with the colonization of newly conquered lands by continuing to teach their children French.⁶² The implication would be that the custom of teaching French to young children in a household setting was falling into disuse, and this impression is confirmed by John Trevisa (1342-1402), who in his translation of Higden’s Polychronicon added that “gentilmen haueþ now moche i-left for to teche here children Frensche.”⁶³

The consequences of the demise of French as England’s second vernacular were many. One of them was that English could now be regarded as the language of the nation, and French as a foreign tongue. Holcot’s myth of the Norman Conquest as a French plot against the national language is symptomatic of this change of linguistic climate, as is the demand for English translations of texts which had earlier been read in Anglo-French.

The textual history of Robert Grosseteste’s Château d’Amour (c. 1240) offers an interesting illustration. Like many intellectuals of his day, Grosseteste had studied in Paris, before returning to become Chancellor of the University of Oxford (1215-21) and finally Bishop of Lincoln (1235-53).⁶⁴ His serious theological writings were in Latin, but he chose French for a treatise on husbandry⁶⁵ and for the Château d’Amour, an allegorical verse treatise which he aimed at a wider audience: “En romanz comenz ma reson / Pur ceus ki ne sevent mie / Ne letrreüre ne clergie” (I begin my discourse in French for those who have neither education nor learning).⁶⁶ The topos of “accessibility to the layman” was a common one in prologues. It is striking that until 1300 the language normally used in this context is French.⁶⁷ By the fourteenth century, however, it became normal to say this in (and of) English, and the fact that French could no longer be regarded as the “comun langage pur amis / Ke de clergie ne ount apris”⁶⁸ (the common language for friends who have not had a formal education) in turn explains why Grosseteste’s Château had to be translated into English.⁶⁹ One translation, The King and Four Daughters (c. 1400), adapts Grosseteste’s prologue as follows:

Here a “romance,” that is, a book in French, is automatically associated with the “lond of France.” Whether the fifteenth-century translator assumed that “Grostyd” was a foreigner or that he wrote his “romance” while in France is unclear, but evidently the translator could no longer entertain the possibility that a French text could be home-produced.

With Anglo-French in recession, English gradually replaced it in the many discursive domains and professional functions where French had earlier been used: in court poetry, in letter writing, record keeping, wills and deeds, petitions to Parliament, charters, and so on. However, it could only do so by usurping the words of the language which it was ousting – which explains why the period of heaviest borrowing from French words is 1350-1500.⁷⁰ The French words that ceased to be used in Anglo-French simply moved across into English. As in the case of Old Norse, the end of Anglo-Norman was the beginning of the Norman conquest of the English language.

The second language of prestige, Latin, derived its high status from the fact that it was the lingua sacra of the medieval West, the language of the Bible (which was normally read in St. Jerome’s Latin Vulgate edition) and the language of church and university. Its acquisition was simply a matter of going to school, because all education began with the learning of Latin. The teaching of Latin literature and versification was the business of all grammar schools, and even before that, at elementary school, boys learned Latin from basic chants and prayers. The seven-year-old “clergeon” in Chaucer’s Prioress’s Tale had learned the hymn Alma redemptoris mater by heart (without knowing what the words mean). At advanced level, at universities all across Europe, Latin was also used, and well-educated men felt thoroughly at home in it. Indeed, Roger Bacon (d. 1294) remarked that he had three mother tongues: English, French, and Latin.⁷¹

The Norman Conquest had had the paradoxical effect of enlarging the scope of Latin. The new ruling elite was French-speaking, of course, but writing documents was no job for the elite, and the clerical business of government administration and record keeping was initially left to English scribes, who were used to writing in English and Latin. Since English was incomprehensible to almost all Frenchmen and French to many Englishmen, it was Latin, incomprehensible only to the uneducated, that prevailed as the written language of government administration, until in the thirteenth century French became sufficiently established in England to be used alongside Latin in this sphere.⁷² Latin had the further advantage of being the language least subject to linguistic variation. This was less true of Anglo-French,⁷³ which had high- and low-prestige varieties (with prestige accruing from conformity with la droit language de Paris⁷⁴) and not at all true of English, which differed enormously from region to region.

And yet even Latin was subject to change and variation, because it was not just a “written” language, used for “formal-official” purposes, as is often assumed,⁷⁵ but a language spoken by students, churchmen, and even by some lay people, as we have seen already. Children heard and spoke Latin as soon as they went to school and, because it was a normal medium of communication for schoolboys and university students, the language constantly evolved to retain full functionality in the environment which its speakers operated in. Glossaries such as Alexander Neckam’s De nominibus utensilium (c. 1180) provide lists of words for everyday objects, and the reference in the titles of such glossaries to utensilia (useful objects) here “suggests that the pupils would be expected to use the vocabulary dealing with daily needs.”⁷⁶ Many medieval coinages for utensilia did not survive into the modern beyond the medieval period. The Renaissance heralded a return to the original “purity” of Roman Latin, and with it came a retreat from the Latin of “daily needs” and a normalization of the language based on classical norms that put an end to local divergences in Latin – divergences so great, as Erasmus lamented in 1528, that Latin “was no longer mutually intelligible among nations.”⁷⁷

Returning to the linguistic situation of the present, characterized, at least in England and the United States, by the hegemony of a single “national” language, we should now be in a better position to appreciate both our differences from and our connections with medieval England. As we have seen, the idea of a country united by a single language, English, first took hold in the fourteenth century, when other vernacular languages of equal or superior status (first Old Norse and then French) had died out as birth tongues of medieval England. Of the Celtic languages, Cumbric was extinct by c. 1100, and while Cornish and Welsh remained alive they were of low prestige (in the opinion of English speakers) and had little impact on Middle English, except on the borders with Wales.⁷⁸ As I suggested at the beginning of this chapter, Holcot’s myth of a pre-Conquest nation united by the English language is a back-projection of his (and now our) reality. His myth has come true, as has the humanists’ dream of Latin as a lingua eterna far removed from the contingencies of the spoken vernacular. Certainly, Latin enjoyed a different status in the period we have surveyed – it was not a mother tongue in the modern sense – but, as we have seen, people like Roger Bacon, who spoke and wrote it from boyhood on, had good reason to describe it as one of his linguae maternae. In C. S. Lewis’s words,⁷⁹ it was humanists like Erasmus who began the “process of classicization which was finally to kill Latin” and to separate it off entirely from the vernacular languages we still speak today.