Professional Documents
Culture Documents
Sacred Bilingualism Code Switching in Me
Sacred Bilingualism Code Switching in Me
By
2002
The Dissertation Committee for Christopher Charles LeCluyse
Certifies that this is the approved version of the following dissertation:
Committee:
___________________________
Mary Blockley, Supervisor
___________________________
Thomas M. Cable
___________________________
Mark R. V. Southern
___________________________
Almeida Jacqueline Toribio
___________________________
Marjorie Curry Woods
Sacred Bilingualism: Code Switching in Medieval English Verse
by
Dissertation
in Partial Fulfillment
of the Requirements
Doctor of Philosophy
May 2002
UMI Number: 3114768
Copyright 2002 by
LeCluyse, Christopher Charles
________________________________________________________
Throughout the Middle Ages, English poets produced works alternating English
and Latin, a practice known as code switching. Studying how these macaronic poems
integrate languages reveals the practices and values of medieval bilinguals in ways
that monolingual texts cannot. This study examines bilingual verse from linguistic,
Starting with three Anglo-Saxon poems that alternate Old English and Latin,
this study characterizes code switches within a generative syntactic framework. Only
three switches in the Anglo-Saxon poems violate the Functional Head Constraint
(FHC), which stipulates that functional elements and their complements should share
the same language and that code-switched constituents must occupy positions
poetic factors displace the complement of a functional head. Code switching allowed
Anglo-Saxon poets to evoke the origins of neighboring texts and to emulate liturgical
iv
After the Conquest, macaronic poetry became especially popular during the
fifteenth- and sixteenth centuries in the form of bilingual carols. More than two-
hundred such carols can be classified into twelve types according to the degree to
which they integrate English and Latin. In most cases, poetic conventions actually
Anglo-Saxon poems, the sixteen ungrammatical switches in the carols typically occur
when poetic constraints override linguistic ones. Comparing code switches in carols from
the Middle- and Early Modern English periods demonstrates writers’ increasing
preference for overtly connecting ideas on the model of Latin and French texts.
community that persisted even as the Reformation radically altered its membership.
Despite these sweeping changes, the inherently public and social nature of carol
performance remained an ideal means for bilinguals to reify their identity, in part by
exhibited here thus lay the groundwork for future studies of code switching in medieval
v
Table of Contents
Introduction 1
Bibliography 180
Vita 194
vii
Introduction
The canon of English literature has long been a monolingual affair, bounded and
historians of the language have from the beginning admitted the influence of other
languages on English (particularly Latin, Norman and Franconian French, and Old
Norse), literary scholars have been slow to study works that bring together the various
languages used on the isle of Britain. Those works that combine languages broadly
labeled macaronic have often been rejected as neither fish nor fowl, unsuitable for
This trend seems to be gradually changing for prose works of the later Middle
Ages. Medievalists have paid increasing attention to texts from the thirteenth through
sixteenth centuries that combine English and Latin: medical texts in Linda Voigts s
work, legal and business documents in Laura Wright s, and sermons in Sigfried
1
In this study I use macaronic neutrally to denote works that include words from more than one
language, with none of the disparagement connoted in the word s early usage. The term macaronic
was coined by Teofilo Folengo (1491—1544) to describe Latin poetry that incorporated Italian words
with Latin inflections (Wehrle vii). According to the Oxford English Dictionary, Folengo was most
likely citing the title of Tifi Odasi s Macaroneae (c. 1490). Both of these authors equated the mixture
of languages in their works with doughy foodstuffs: Odasi called his poetry gnoccolosa (gnocchi-like),
and Folengo explained in the second edition of his Liber macaronices (1521) that the name of the style
came from macaroni, quoddam pulmentum farina, caseo, botiro compaginatum, grossum, rude, et
rusticanum (a certain pottage made from flour, cheese, and butter thick, crude, and rustic)
( macaronic ). In keeping with the qualities he ascribed to this dish, Folengo stipulated that the
sentiments [of such poems] must be gross, vulgar, and obscene (Wehrle xi). Although in time the
term macaronic came to be applied to works of all sentiments, involving any juxtaposition of
languages, the negative connotations of the word persisted in the usage of some critics. Wenzel
summarizes the derivation of the term macaronic and briefly summarizes examples of macaronic
English poetry before contrasting them with the macaronic sermons of his own study (1—12). Wehrle
recounts writers and critics application of the term from the sixteenth through the early twentieth
centuries (xi—xxxvii).
1
Wenzel s. 2 Each of these studies has turned up an impressive number of multilingual
documents enough, according to Wright, to challenge the traditional notion that the
of Latin and French. Most of these studies introduce the linguistic phenomenon of
utterance. They generally use that term only in its broadest sociolinguistic sense,
poems from Anglo-Saxon England, the last lengthy study to address the combination
of languages in works from the Old, Middle, and Early Modern English periods was
of Rossell Hope Robbins, Richard Leighton Greene, and David Jeffrey has treated
macaronic poems from the Middle English period but only in the context of larger
discussions of medieval English lyric poetry.4 None of these scholars has continued
Wehrle s juxtaposition of early and late medieval poems, and none Wehrle
2
These scholar s relevant studies are included in the bibliography.
3
Two emerging scholars who have approached code switching in medieval texts in a linguistically
sophisticated fashion are Elaine Rebecca Miller, who studies multilingualism in documents produced
by fifteenth-century Spanish Jews, and Mary Catherine Davidson, who approaches code switching in
medieval English legal and poetic texts from a discourse perspective. Titles of their studies are
included in the bibliography.
4
See pp. 70—81 and p. 82, n. 9, for a discussion of Jeffrey s work.
2
In this study, I will try to fill the seventy-year void in scholarship by bringing
English history. The following chapters will serve as snapshots of linguistic, literary,
and social practice at two crucial junctures in the English Middle Ages: at one end,
the late-tenth- and early-eleventh centuries, at the other, the fifteenth and early-
here share much: the regular alternation of English and Latin in highly regulated
verse forms, the incorporation of liturgical material into extraliturgical poetic settings,
and the use of code switching to signal membership in bilingual speech communities.
English poetic works, I will not only characterize written code switching in each era
but also compare code-switching practice among eras, revealing how developments in
English syntax, poetic conventions, and written style play out in the poems. Changes
in English syntax altered how the language could interact with Latin within bilingual
As this study will show, the writers of macaronic poems, by operating within
formal poetic restrictions, produced much more regular and artificial code switches
than those occurring in spoken discourse but nevertheless maintained the grammatical
integrity of those switches with the ease expected of fluent bilinguals. The poems
written forms and structural features in fact amplify rather than distort the grammar of
3
code switching, drawing attention to the process itself by alternating languages in a
regular and highly predicable fashion. In the few instances of ungrammatical switches
between English and Latin, poetic constraints often supercede linguistic ones, as the
The interplay of syntactic and poetic structure in these bilingual works reminds
us that linguistic and literary concerns are not mutually exclusive. Nor must a more
from their social function or their place in the development of the English language.
problem of projecting from the present to the past and developing an interdisciplinary
framework for socio-historical linguistics (11). Far from being limited to literary
texts or representing only deviations from spoken usage, style, they argue, is in fact a
exactly how Latin and English interact in these texts therefore gives us a more
detailed understanding of the bilingual communities in which they circulated and the
linguistic habits of those communities. To account for the literary and the social as
well as the linguistic, I will alternate chapters detailing the what of code switching in
the poems with chapters explaining the why. That is, detailed analyses of where
5
Susan M. Wright applies such considerations of style to syntactic analysis in her article On the
Stylistic Basis of Syntactic Change.
4
languages alternate will be followed by discussion of how the macaronic poems in
these influences will in turn demonstrate how macaronic poetry allowed writers and
the Benedictine monasteries reformed before the Norman Conquest to the much more
diffuse constellation of clerics, religious, and lay people who enjoyed reading and
with that in better-known works by Chaucer and Langland (a topic treated at the end
scholars may be quite familiar of Chaucer s use of Latin to satirize pretentious layfolk
confusio in the Nun s Priest s Tale , ll. 4353—56) and Langland s incorporation of
Latin scriptural phrases for more edifying purposes. These canonized depictions of
bilingual phenomena reinforce the notion of Latin and the vernacular as separate and
literary studies , I hope to show how these oft-neglected works embody a variety of
influences at work in medieval England, they do not deserve to slip through the
disciplinary cracks. They are in fact English literature in the fullest sense.
6
Chapter 1: Code Switching in Anglo-Saxon Poetry
wucan lengtenes, / on †am mon†e †e man Martius / geond Romwara rice nemne∂”
(during the first week of Lent, in the month that one calls March throughout the
kingdom of the Romans) (ll. 48–50). Still other works, most notably the freely
adapted liturgical texts grouped as the Benedictine Office, alternate a line of Latin
with several lines in Old English. None of the aforementioned poems entail much
linguistic integration, however. Those poems containing isolated Latin words use the
foreign terms as momentary borrowings and may even draw attention to their foreign
origins, as the example above shows. Likewise, poems translating liturgical texts
generally restate each Latin element before substantially amplifying it in the
vernacular. This and similar linguistic juxtapositions may more properly be
considered translation phenomena—in effect Latin glosses that have been
incorporated into vernacular texts.
In contrast to those poems that bring Old English and Latin together only
superficially, three surviving works—“The Phoenix” (in its closing ten lines), “A
1
For more on the kinds of Latin words borrowed in the poems and how those words are inflected,
see the studies by Otto Funke and Peter Baker. Helmut Gneuss relates Old English words borrowed or
loan-translated from Latin to manuscript glosses in “Anglicae linguae interpretatio: Language Contact,
Lexical Borrowing and Glossing in Anglo-Saxon England.”
7
Summons to Prayer,” and “Aldhelm”—alternate strings of words in either language
within sentences. Previous scholars seem to have been caught short by this linguistic
alternation and have not proceeded to address substantively how the poems’ creators
went about switching languages and why they would choose to do so. Those scholars
who have not dismissed the multilingual character of these texts outright have
focused primarily on how the Old English vernacular is peculiarly elevated by
juxtaposing it with Latin, whether for spiritual effect or as a virtuosic flourish. A
typical comment is Leslie Whitbread’s pronouncement that “all three specimens may
be called mere exercises in ingenuity and in the literature of display” (“The Old
English Poem Aldhelm,” 197–98).
Such dismissals overlook the fascinating interplay between languages in the
poems, which goes far beyond the borrowings and glosses found in other Anglo-
Saxon texts. Unlike the translations on which most scholarly accounts of bilingualism
in Anglo-Saxon England focus, these works portray bilingualism in its most overt
manifestation: code switching.2 Studying these poems, then, illuminates how writers
went about switching languages in texts written for monastic audiences and demonstrates
how bilingualism itself could be a topic of vital interest within such communities.
Before treating the language alternation in these poems in a more substantive
and linguistically informed manner, I will explain their manuscript contexts,
2
Studies of linguistic interaction between Old English and Latin by and large focus on the processes
Anglo-Saxon translators used to render and modify Latin material; see, for example, Anne Savage’s
and Brian Shaw’s articles on “The Phoenix,” discussed below. As a translator and grammarian, Ælfric
has been a frequent focus of discussion of bilingualism (for example, in Connie Eble’s brief sketch).
Ælfric did not display his bilingualism by code switching in the manner of the Anglo-Saxon macaronic
poems, however. Switches between Old English and Latin in his Grammar, for example, are instead
made to cite Latin texts or to translate Latin words and phrases into Old English. Ælfric also borrows
Latin grammatical terms and literally translates other terms as Old English calques (for example, Old
English foresetnyss for Latin praepositio [Zupitza 10.15]). My thanks to Glenn Davis for providing
this last example.
8
reproduce and translate the texts themselves, and summarize previous treatments of
the Exeter Book. This late tenth century manuscript contains nearly a fourth of the
surviving corpus of Old English poetry (Ker 152). The origins of the Exeter Book are
unknown, but textual scholars have proposed several places of origin. The leading
theory, reviewed by Richard Gameson, posits this famous manuscript as the product
(135–85). Patrick Conner argues, however, that the codex may have originated at
Exeter, based on its similarities to five other manuscripts with Exeter connections.
Although Gameson counters Conner’s analysis by arguing that the features the Exeter
Book shares with the other manuscripts do not necessarily indicate a common Exeter
origin, he concedes that given the obscure early history of the Exeter Book, even
Glastonbury and Christ Church, Canterbury, “are little more than educated guesses” (179).
In its first section, “The Phoenix” describes the appearance, travels, and life
cycle of the mythical bird—a description largely derived from the Carmen de ave
phoenice by the second-century poet Lactantius. In the second section, the Anglo-
Saxon poet added Christian elements to Lactantius’ pagan poem and further expanded
9
upon it by allegorizing the phoenix as a type of the good Christian. This moral lesson
[The Author of Light has delivered us so that we might here deserve to acquire through good
deeds joys in heaven, where we might seek out the greatest kingdoms and sit in high seats,
live in the joy of light and peace, possess the pleasant dwelling of happiness, enjoy the
harvest days, see the mild and gentle Lord of Victories without end, and sing him praise in
eternal exultation, blessed among the angels. Alleluia.]
Although “The Phoenix” is the most thoroughly studied of those poems that
alternate Old English and Latin, its macaronic ending has received relatively little
attention. Most scholars focus on the ways in which the poem translates and expands
upon Lactantius’s Carmen de ave phoenice (Krapp and Dobbie xxxiv–xxxv). Anne
Savage, for example, contextualizes the poet’s translation in the bilingual atmosphere
3
The text for “The Phoenix” is from Blake 63 (cf. the edition by Krapp and Dobbie 112–13). I have
omitted Blake’s macrons and changed his punctuation as indicated. Texts for “The Summons to
Prayer” and “Aldhelm” are from Dobbie 69–70 and 97–98. I have altered some punctuation and
italicized Latin words. All translations are my own.
4
Here I follow Krapp and Dobbie’s punctuation on the grounds that the most straightforward
translation of gesittan is its modern derivation, “sit,” and that the ablative case of sedibus altis
indicates the place where the sitting is done. Blake places a comma after gesittan, translating the word
as “occupy” and reading maxima regna as its object (87).
10
process of translating themselves to themselves, from a Latin embodying Christianity
(123). Savage explains that by adding new material to the source text and by
departing from the original story to introduce the themes of exile and solitude so
popular in Old English poetry, the poet recast the Latin poem culturally, allowing
readers to relate to its subject more immediately (131; 133–34). Savage does not
mention the poem’s macaronic coda, however—a curious omission since the bilingual
passage overtly demonstrates how the bilingual poet “translated himself to himself.”
first half of “The Phoenix” corresponds to commentary in the second half, Shaw
shows how the poem recasts Lactantius’s material in a more Christian light and how
it relates the various aspects of the phoenix to Christ and his resurrection. Unlike
Savage, though, Shaw does draw attention to the macaronic ending, noting that it
“relies for a degree of heightened feeling on the use of Latin … and the liturgical
echoes” (178). Shaw does not explain, however, what this “heightened feeling”
entails or what these liturgical echoes are. In Chapter 2, I will make the
11
1.1.2 “A Summons to Prayer”
Another macaronic work similar to the end of “The Phoenix” in its execution is
other poetry, a partial Old English translation of the Regularis concordia, numerous
homilies, and an Old English prose version of Appolonius of Tyre. Ker dates the
section that contains the Regularis concordia and “A Summons to Prayer” to the
beginning of the eleventh century (82–90). Although the manuscript’s place of origin
later articles prefers a Christ Church, Canterbury, origin (Dobbie lxix–lxx; Whitbread,
translation of Bede’s De die judicii) and another poetic section commonly called “An
6
Dobbie sees the manuscript form rogo as incorrect and suggests the imperative roga instead. On
the variant readings of this line and their impact on our interpretation of the underlying syntax, see
below, p. 37, n. 26.
13
“A Summons to Prayer” has been discussed especially in relation to the
material preceding it in CCCC 201: “Judgment Day II” and “An Exhortation to
studying the manuscript and its affinities with other contemporary homily collections,
noted the textual and thematic similarities among the three poems. Whitbread read the
poem as a final absolution following the call to confession in “Judgment Day II” and
Summons to Prayer,” Whitbread follows the traditional editorial assumption that the
two are separate texts. Fred C. Robinson, however, shows that these supposedly
separate texts are in fact presented as a single work in the manuscript (“The Rewards
1.1.3 “Aldhelm”
two only in its multilingual character. The poem appears in Cambridge, Corpus
Christi College 326 between the Latin text of Aldhelm’s Prosa de virginitate and the
table of chapters for that work. The manuscript is one of five surviving copies of
14
Aldhelm’s treatise produced between the late tenth and early eleventh centuries at
Christ Church, Canterbury, but is the only copy to feature the poem (Ker 107–08;
Brooks 267–70). As the title modern editors have given the poem indicates, the poem
preeminent writer of Anglo-Latin verse and prose. Unlike the end of “The Phoenix”
and “A Summons to Prayer,” the poem “Aldhelm” includes isolated Greek borrowings
(underlined in the text below) and switches languages throughout most of the lines.
(3) “Aldhelm”
†us me gesette sanctus et iustus
beorn boca gleaw, bonus auctor,
Ealdelm, æ†ele sceop, etiam fuit
ipselos on æ†ele Angolsexna,
5 byscop on Bretene. Biblos ic nu sceal
ponus et pondus pleno cum sensu,
geonges geano†e geomres iamiamque,
secgan so†, nalles leas, †æt him symle wæs
euthenia oftor on fylste,
10 æne on e†le ec †on †e se is
yfel on gesæd. Etiam nusquam
ne sceal ladigan labor quem tenet
encratea, ac he ealneg sceal
boethia biddan georne
15 †urh his modes gemind micro in cosmo,
†æt him drihten gyfe dinams on eor†an,
fortis factor, †æt he for† simle
[Thus a holy and just man accomplished in books, a good author, Aldhelm, the noble
poet set me. He was also eminent in the homeland of the English Saxons, a bishop in
Britain. I, a book, shall relate truly, not at all falsely, a toil and a burden with full
meaning and moreover in the mourning [or birth] of fresh affliction, that abundance
was always there for him in aid, alone [or fame] likewise in his homeland, more
often than for one about whom evil is spoken. He whom self-control possesses must
certainly never excuse himself from hardship, but he must always eagerly pray for
support through his spiritual consciousness in the little world, that the Lord might
give him power on earth, the strong Creator, so that he might always henceforth …]
15
“Aldhelm” is especially challenging to translate because of two difficult
passages containing words that could be either transliterated Greek, like the
unambiguously Greek terms that dot the rest of the poem, or Old English.7 The word
geano∂e, in the seventh line, is unique to this poem; Arthur S. Napier, the poem’s first
modern editor, believed that it was an Old English word akin to Gothic gaunò†us,
transliteration of Greek genethle or genete, “birth” (Dobbie 194; Whitbread, “The Old
English Poem Aldhelm” 196). Likewise, æne, in line 10, could be either an Old English
prompting critics to assume that the text is incomplete (cf. Dobbie xci, 194;
Whitbread, “The Old English Poem Aldhelm” 193). Fred C. Robinson has tentatively
suggested, however, that the last sentence of the poem could in fact connect the
metrical work to the following treatise, which begins with Aldhelm’s formulaic
meaning “send” was traditionally omitted in such opening formulae. This assumed verb
of sending could be predicated by he at the end of the poem, making for a seamless fit
(and a final code switch) between it and the prose work: “… so that he might always
7
Anglo-Saxon poets writing in Latin likewise showed a predilection for borrowing Greek terms, and
this so-called hermeneutic style was practiced extensively at Canterbury, where “Aldhelm” was
written. See Section 2.1 for a discussion of how the poem’s inclusion of Greek borrowings emulated
contemporary Canterbury style and characterized the writing of Aldhelm himself.
16
henceforth [send] to the most reverend virgins of Christ … his best wishes for perpetual
have generally held it in low esteem. Whitbread makes his comment regarding the
compose a poem in three languages at once exceeded his talents” and denigrates “the
poet’s clumsy flourishes and often uninflected Latin and Greek” (197). Lapidge is
Style” 121). The subsequent analysis of where code switches occur in the poem,
however, will show that for all of the poem’s supposed flaws, Latin and Old English in
“Aldhelm” alternate grammatically. As Section 1.4 will explain, the smooth switches
between Latin and Old English in “Aldhelm” suggest that the poet knew Latin quite
poems illustrates an aspect of code switching linguists have noted for some time:
governed behavior. The Anglo-Saxon poems are, of course, shaped by poetic as well
17
as linguistic principles and are written texts rather than spoken utterances.8 The
fundamental linguistic patterns of code switching. These poems are some of the most
Determining exactly what rules the code switches in the poems follow is a more
difficult matter, however. Particular Latin and Greek elements favor a variety of
Prayer” and “Aldhelm” (for example, salus mundi and fortis factor) could be
interpreted as Latin insertions into an Old English matrix. In many cases, however,
sequences of words in Old English alternate with sequences in Latin, with neither
Prayer,” “ˇænne gemiltsa∂ †e, N., mundum qui regit,” languages switch between a
Latin noun clause acting as the subject of the sentence (mundum qui regit) and the
Old English verb gemiltsa∂. Still, the Greek words in “Aldhelm” appear to be
borrowings; as Dobbie points out, the poet need not have been fluent in Greek to
8
These written texts nevertheless retained aspects of oral composition. See Katherine O’Brien
O’Keeffe’s Visible Song for a thorough discussion of transitional literacy in Anglo-Saxon literature.
While most reviewers have wholeheartedly accepted her argument, John D. Niles faults her for
considering only poetic texts and Peter J. Lucas questions, among other things, her assertion that
manuscript punctuation before the eleventh century did not mark meter.
9
Anglo-Saxon poets apparently obtained such Greek terms by imitating the works of Aldhelm and
Abbo of Germain-des-Prés and by consulting Latin-Greek glossaries (Lapidge, “The Hermeneutic
Style” 114–15). See pp. 50–51, below.
18
Many linguists subscribe to one of these various interpretations in their studies
of modern bilingual data. Some studies, such as Miwa Nishimura’s study of Japanese-
English code switching and the David Sankoff and his colleagues’ article on so-called
nonce loans in Tamil, favor an insertional model, according to which words from one
language are interpolated into a base or matrix language. Others, such as Carol Pfaff’s
groundbreaking study and that by Hedi Belazi and his colleagues, described in
alternational model posits that bilinguals switch between two discrete grammars.10
Pieter Muysken has negotiated the differences between these two models of
code switching by treating both as viable alternatives. In his study of the multilingual
different groups of speakers seem to favor different code switching strategies.11 For
example, people who speak Chinese as well as Dutch are more likely to insert short
strings of words in one language into extended sequences in the other. By contrast,
10
Or, in minimalist approaches to syntax, two manifestations of one underlying grammar.
11
Muysken identifies a third model of code-switching, congruent lexicalization, according to which
particular words find a place in the vocabularies of both languages. As he explains, however, this
notion “more often underlies the study of style shifting and dialect/standard variation, rather than
bilingual language use proper” (“Patterns of Language Mixing” 400). The practice is particularly
common in “cases where word forms are either already identical in the two codes involved or
pronounced in an ambiguous way” (“Patterns of Language Mixing” 401).
19
sizeable sequences in both languages (“Patterns of Language Mixing” 401–02). As
Muysken explains, a variety of factors accounts for such differences, including the
similarities between the languages involved, the speaker’s degree of fluency in the
borrowing. In addition to the Latin names cited at the beginning of this chapter, we
find a number of isolated Latin words in otherwise Old English poems: for example,
the poem “Christ” in the Exeter Book contains the line “Eala sibbe gesih∂, sancta
Hierusalem” (O peaceful sight, holy Jerusalem, l. 50).12 Although both code switching
and borrowing are superficially similar in that they involve the use of more than one
from code switching is therefore important because, as Carol Pfaff explains, the two
phenomena “are usually construed as making vastly different claims about the
competence of the individual speaker [in that] ‘borrowing’ may occur in the speech of
12
Fred C. Robinson also notes instances where Anglo-Saxon scribes apparently used Latin words in
place of Old English ones in cases where the meter and alliteration of poetic texts would require the
vernacular (“Latin for Old English”).
20
those with only monolingual competence, while ‘code-switching’ implies some
the former generally involves single lexical units, the latter involves phrases
(Muysken, “Patterns of Language Mixing” 399). Borrowings usually fall outside the
words from strings of words in another language) is therefore a necessary first step in
seems best suited to the data at hand.In studying poems split equally into Old English
and Latin phrases, one would be hard pressed to determine what the base language is,
13
Romaine thoroughly reviews prevailing explanations of borrowing (51–67) and distinctions
between borrowing and code switching (142–61).
14
Although borrowed words that are used regularly throughout a community, called established
borrowings, may conform to Pfaff’s generalization, the use of borrowings is not limited to
monolingual speakers. Borrowings that occur momentarily in an individual’s speech, called nonce
borrowings, may in fact be employed primarily by speakers who are moderately to extremely fluent in
two languages (Poplack and Sankoff, “Code Switching” 1176; Romaine 67). Moreover, the initiation
of an established borrowing does of course require a bilingual speaker to introduce the word to other
speakers (see Romaine 59–60).
15
Elaine R. Miller reaches a similar impasse in trying to apply Aravind K. Joshi’s insertional model
to Hebrew-Spanish code switching in fifteenth-century Sephardic texts (“Medieval Spanish-Hebrew
Language Switching: A Comparison with Joshi's Constraint”).
21
“residue of indeterminate cases” in studies that seek to determine a base language for
tells us nothing of the communicative competence of the speaker” (148). That is,
determining that the base language of the first line of “A Summons to Prayer” is in fact
Old English because the verb of that sentence is in Old English does not indicate how
well the writer knew Latin. To avoid such problems and enable further consideration of
how the poems reflect the poets’ Latin competence, I will employ an alternational
In their 1994 study of code switching, Hedi M. Belazi, Edward J. Rubin, and
code switching generally fail to describe actual code switching behavior, either
because these constraints are too restrictive or not restrictive enough (222–28). The
Functional Head Constraint (FHC), on the other hand, describes code switching more
words that have real semantic content (called lexical heads, the latter term denoting
that they characterize the phrases in which they occur) and those that carry primarily
complementizers such as that in “She said that she studies code switching,”
22
According to the FHC, switching between a lexical head and its complement
(the word that “completes” its meaning, such as the direct object of a verb) is
unrestricted as long as the words occupy positions both languages allow (Belazi et al.
221).16 Such switches include those between an adjective or adverb and the word it
modifies. For instance, since the grammar of French allows certain kinds of adjectives
to follow the nouns they modify, unlike English, the code switch “Voici l’auto rouge
that she might buy tomorrow” would be grammatical but “Voici le red auto … ”
would not. Switching between a functional head and its complement, on the other
l’auto rouge qu’elle might buy tomorrow” would be grammatical because the
languages switch before the functional head might (a modal), the sentence would be
ungrammatical if it contained the switch “that she might acheter demain.” The
Subsequent studies have confirmed that speakers observe the FHC as they gain
16
Belazi et al. derive this last condition from the Equivalence Constraint proposed by Sankoff and
Poplack and incorporate it as a corollary to the FHC.
23
elements and their complements, bilingual speakers would unconsciously replace
Other studies involving children raised in a bilingual environment and adults learning
a second language likewise show the constraint at work. Bilingual children begin to
switch languages in accordance with the FHC once they have acquired a full
the FHC as they gain mastery in the second language (Toribio, “Emergence of Code-
switching Competence”).17
The operation of the Functional Head Constraint will become clearer as we see
demonstrate, most of the switches in the poems are grammatical because they occur
either between sentences or between lexical heads. For example, in “The Phoenix,”
“agan eardinga alma letitiæ” (possess the pleasant dwelling of happiness, l. 673) is
grammatical because the languages switch between two lexical heads: the Old
English noun eardinga and the Latin adjective which modifies it, alma. Both
17
Some linguists have attempted to refute the FHC by presenting evidence that supposedly
contradicts it. Shahrzad Mahootian and Beatrice Santorini, for example, fail to see the constraint at
work in their recordings of spontaneous speech. As Toribio explains in “Emergence of Code-Switching
Competence,” however, their study is methodologically flawed because such data is “incompatible
with linguistic-theoretical modes of inquiry.” Only responses elicited to determine a speaker’s
evaluation of ungrammatical utterances can support or disprove linguistic constraints. Jeffrey
MacSwan does use elicited responses to support his conclusion that switches between a
complementizer and its complement are permissible, but, Toribio contends, overlooks data obtained
from his own subjects indicating that a short pause between the functional head and its complement
makes such sentences more acceptable. Toribio also demonstrates that the counterexamples Miwa
Nishimura offers in Japanese/English Code-switching are borrowings accommodated by Japanese
morphology rather than true code switches.
24
constituents occupy positions sanctioned by the grammars of Old English and Latin,
since both languages allow modifying phrases to come after the words they modify.
The very few switches that violate the Functional Head Constraint do so because they
occur between a functional element and its complement—the case in “se so∂fæsta
summi filius” from “A Summons to Prayer,” in which the switch occurs structurally
between the Old English determiner se (a functional head) and its Latin complement
filius. After characterizing the grammatical switches in the poems, the subsequent
Proper evaluation of the code switches in the poems can occur only after we
have an overall idea of how they integrate Latin and Old English syntax. Most of the
poems’ code switches occur within rather than between sentences (that is, most
switches are intrasentential rather than intersentential). Only six of nearly sixty
sentences in the poems. The macaronic coda to “The Phoenix,” for example, is a single
ten-line long sentence. Intrasentential switches between Old English and Latin are
classified in Tables 1 and 2, below. Greek borrowings in “Aldhelm” have not been
included in the tally because they do not constitute genuine code switches, as explained
in Section 1.2.1.
The descriptors used in the first column of each table follow the general format
Sankoff and Poplack use in their formulation of code-switching grammar. The first or
25
only constituent listed in the leftmost column, “Switch Site,” is in the language that
appears first in the title of the table; for example, in Table 1.1, “Verb and
complement” indicates a switch between an Old English verb and its Latin
complement. Switch sites are grouped according to the kinds of constituents they
involve: lexical heads that are necessary for the sentence to be complete (obligatory
constituents), lexical heads that are not obligatory (adjuncts), and functional heads.
The label under “Switch Site” denotes where code switches occur on the most
fundamental syntactic level—not necessarily as they appear in the poem texts. For
instance, in the clause “†æt hi ealle †e unica uoce †ingian to †eodne,” lines 24–25a of “A
Summons to Prayer,” the ablative phrase unica voce adjoins the verb †ingian, whose
object is †e. Even though the languages switch twice in surface linear order, from the
Old English object to the adjunct phrase and back, structurally this is a single switch
between the adjunct and the verb. Here I differ from linguists who regard code
switching as a surface process. Muysken, for example, states that “the evidence
(Bilingual Speech 118). He supports this generalization, however, with only two
18
Both examples include English phrasal verbs in otherwise Spanish sentences, the first from
Poplack’s initial study, the second from Carol Pfaff’s article: “Me iban a lay off” (“They were going to
lay me off”) and “Bueno, por qué te hicieron beat up ese?” (“Well, why did they make you beat up that
[person]?”) (Bilingual Speech 118–19). Muysken argues that lay off in the first example would be
blocked if me had not already been moved from its original position to before the verb iban; likewise,
he reads the movement of the subject of the infinitive te in the second example as enabling the
26
the surface linear order of words, Muysken would be hard pressed to explain cases of
express the same material in both languages (cf. Romaine 125–26). My analysis of
motivated by the assumptions expressed through the Functional Head Constraint: that
code switching involves the alternation between two discrete linguistic systems and
Constituent types that are not involved in any code switches are not listed in the
tables. Table 1.2, for instance, does not list switches involving functional heads
because no such switches involving Latin functional heads occur in the poems. The
like that between ll. 669 and 670 of “The Phoenix,” “… gaudia in celo, / †ær we
motum maxima regna secan,” are said to occur “preceding” that relative pronoun or
relative adverb (†ær in this case) to distinguish them from switches that may occur
between the relative and the subject of the following clause. In the case of switches
following Old English functional heads, an asterisk indicates that the switch comes
between the functional element and its complement and is therefore ungrammatical.
2: “beorn boca gleaw, bonus auctor,” from Old English to a Latin appositive) may
inclusion of beat up. This seems to be a chicken-or-egg problem, however, since the movement could
have been triggered by the input of these phrasal verbs from the lexicon.
27
occur either preceding or following the constituent named. Since such modifiers are
neither obligatory parts of the sentence nor functional heads, noting the exact
The second column in each table indicates the number of switches occurring at
that location in all three poems. The third column expresses this number as a
percentage of all the switches recorded in the table—that is, all switches from Old
English to Latin in Table 1.1, and all switches from Latin to Old English in Table 1.2.
The fourth column expresses the number of switches as a percentage of all the
switches tend to occur in the poems. Because of the relatively small number of
28
Table 1.2: Intrasentential Switches Between Latin and Old English
macaronic poems occur either between Latin adjuncts and the Old English words they
modify or between Latin obligatory constituents and subsequent Old English sentence
elements. Those switches involving adjuncts run no risk of violating the Functional
Head Constraint and are therefore grammatical. The two kinds of Latin adjuncts
featured most frequently in the poems are ablative noun phrases and appositives.
Latin ablative phrases appear in “The Phoenix” (sedibus altis, l. 671, and laude
partu, l. 10; clara uoce, l. 16; bona uoluntate, l. 17; casta uiuendo, l. 19; unica uoce,
l. 24; and the mixed-case luce perhennem,19 l. 29). Latin appositives can be seen both
19
If the scribe of this portion of the manuscript omitted a nasal abbreviation mark over the e of luce,
the phrase may in fact be in the accusative case. Dobbie emends the phrase to “in lucem perhennem (or
perennem),” most likely under the same assumption (185).
29
in “A Summons” (auctor pacis, l. 5; salus mundi, l. 6; factor cosmi, l. 9; Christus in
orbem, l. 11; mundi redemptor, l. 12; clemens deus, l. 14; summus iudex, l. 27; factor
aeternus, l. 28; and regna caelorum, l. 31) and “Aldhelm” (bonus auctor, l. 2 and
Summons” displays a double convergence of language and poetic form. While the
modified to include a distinctly Latin phrase type, it also uses Latin appositives to
serve a distinctly Anglo-Saxon poetic purpose.20 The Latin appositives included in the
a verb or between a verb and its complement, make up the next largest group of
switches. These switches likewise run no risk of violating the Functional Head
Constraint because they occur after lexical, not functional, heads.21 Here the Latin
and Old English data is almost in complementary distribution: whereas all of the
switches between subjects and verbs occur between a Latin subject and an Old
English verb, all but two of the switches between verbs and their complements occur
20
On the use of apposition for thematic emphasis in Beowulf, see Fred C. Robinson, Beowulf and the
Appositive Style, esp. 68–80. Mary Blockley further explains the syntactic ramifications of apposition
in the poems in two articles, “Apposition and the Subjects of Verb-Initial Clauses” and “Old English
Coordination, Apposition, and the Syntax of English Poetry,” a study incorporated into Chapter 2 of
her book Aspects of Old English Poetic Syntax 55–57. Christopher M. Cain discusses the phonological
adaptation of Latin words in the macaronic poems to serve Old English meter.
30
between and Old English verb and its Latin complement. Specifically, we find the
subject-verb switch “Hafa† us alyfed lucis auctor” in l. 677 of “The Phoenix”; three
others in “A Summons to Prayer,” “ˇænne gemiltsa∂ †e, N., mundum qui regit” (l. 1),
“Geunne †e on life auctor pacis” (l. 5), and “summi filius fo on fultum” (ll. 8b–9a); and
“ne sceal ladigan labor quem tenet encratea” in “Aldhelm” (ll. 12–13a). Of the two
switches between Latin verbs and Old English complements, one occurs in “The
Phoenix”: “†æt we motun her merueri goddædum begietan” (ll. 678–79), in which the
Latin verb merueri takes an Old English infinitive, begietan, as its complement. The
other occurs in “A Summons to Prayer”: “†u †a so∂fæstan supplex rogo fultumes bidde
friclo uirginem almum …” (ll. 20–21), in which the Latin verb rogo takes the Old
English complement †u.22 Switches between Old English verbs and Latin
complements include the following: in “The Phoenix,” “begietan gaudia in celo” (l.
679) and “motun maxima regna secan” (ll. 4–5a); in “A Summons to Prayer,” “scolde
cennan Christum regem” (l. 18), “fultumes bidde friclo uirginem almum” (l. 21); and in
“Aldhelm,” “ne sceal ladigan labor” (l. 12). The one example of a switch between a
preposition and its complement is “to omnes sancti” in “A Summons to Prayer” (l. 22).
The three ungrammatical code switches in the poems follow Old English
21
In her typology of code switching, based on data from modern Spanish-English bilinguals, Poplack
notes that after intersentential switches, those between obligatory constituents are the most common.
31
and evaluating these switches is somewhat difficult since each switch exposes a
switches is relatively straightforward, however, since all three show the effects of
poetic usage on syntax. In these particular cases, the poets’ efforts to maintain
determiner: “se so∂fæsta summi filius” in “A Summons to Prayer” (l. 8). As explained
previously, here the languages switch at the structural level between the Old English
determiner se and Latin filius, the functional element’s complement. Code switches
following determiners are difficult to distinguish from borrowings because nouns are
frequently involved in code switches and are also the most frequently borrowed
lexical items (cf. Belazi et al. 226–27; Romaine 124–25). In this particular example,
one cannot establish once and for all whether summi filius is simply a Latin
commonplace borrowed for the occasion or a bona fide transfer from Old English to
Latin grammar. Nor, for that matter, can one rule out that the complement of se is
22
On the switch between the modal motun and the anomalous infinitive merueri, see Section 1.3.2.2.
23
In this and subsequent examples, I have identified code switches by (1) bracketing the larger
constituent in which they occur, (2) labeling the functional head, (3) enclosing its complement in an
additional set of brackets, and (4) underscoring the head of that complement.
32
so∂fæsta, used as a substantive (“the truth-fast one”), rather than filius. If summi filius
is not a borrowing and filius is the complement of se, however, then the switch
meter, alliteration, and language alternation in the poem.24 The half-line summi filius
conforms to the metrical and alliterative pattern seen throughout the macaronic
coincident with the onset of the off-verse, as if delaying the primary governing
principle of the verse would present complications” (281).25 To make the job easier,
Latin and in Old English by choosing words and formulae in which there is no
contradiction between Latin stress assignment and Old English rules of alliteration,”
24
According to the traditional description of Old English meter developed by Eduard Sievers, each
line is composed of two half-lines, or verses, the first of which is sometimes referred to as the on-verse
and the second as the off-verse. Each half-line must contain two stressed syllables, and either or both
of the stressed syllables in the on-verse must alliterate with the first stressed syllable of the off-verse.
Additional aspects of Sievers’ metrical theory concern the number of unstressed syllables in each half-
line and where they can occur; characteristic alternations of stressed and unstressed syllables produce
five different types of half-lines, lettered A through E. See Mitchell, A Guide to Old English 161–67
for a succinct summary of this traditional system.
Mary Blockley and Thomas Cable have since presented a “more adequate theory of meter than that
described by Sievers’ Five Types,” which systematically accounts for the number and location of
unstressed syllables in the half-line:
The meter of Old English poetry is alliterative-syllabic, each verse containing four positions,
which are realized as four syllables or resolved equivalents. There is one optional expansion
of unstressed syllables in either of the first two positions of the verse (273).
Although Cain’s analysis, cited below, operates under the traditional model, my subsequent
discussions of how particular words occupy metrical positions within the half-line follow Blockley and
Cable’s theory.
33
as is the case with the disyllabic summi (Cain 282). Because the second half-line
determiner like se could not occupy a position at the beginning of the off-verse even
if Old English were allowed there. The determiner therefore was relegated to the first
half-line, delaying the arrival of the complement until after the switch to Latin.
1.3.2.2 Switch Following a Modal: [ModP motun [her merueri] (“The Phoenix,” l. 2)
As with the previous example, analyzing the switch in “The Phoenix” between
Old English motun and its Latin complement merueri leads to a conundrum of
hence a functional head. Whether Old English possessed true modals is a contentious
issue, however. David Lightfoot argues that although words like motun developed
into true modal auxiliaries in Early Modern English (might, should, and would), they
still possessed a primarily verbal character in Old English (98). Such “pre-modals,”
Lightfoot argues, were fully inflected, took objects, and could occupy the same
positions as other verbs throughout the Old and Middle English periods. Not until the
formulating the syntactic category “modal” for Old English would be unmotivated
25
By Cain’s count, thirty-seven of the fifty-two alliterating Latin half-lines in the three poems place
the alliterating syllable “in the onset of a disyllabic word, which, of course, must always bear word
stress” (284). On alliteration of Latin and English words in Piers Plowman, see Sullivan 91.
34
English modal formation, therefore, Old English auxiliaries like motun would be
the development of English modals. Bernd Heine explains, for example, that
particular auxiliaries may have a range of functions along a continuum he calls the
“Verb-to-TAM [Tense, Aspect, and Modality] chain.” At one end of the continuum
are the full-fledged lexical properties of the verb; at the other, the purely functional
properties of tense and aspect markers like the English past-tense suffix -ed. The use
of particular auxiliaries varies not only among languages and across historical periods
but also among speakers and groups (53–71). Contrary to Lightfoot’s account of the
development of modal auxiliaries, then, modals did not change as a class, nor did they
all go through the same set of processes in the development from fully substantive
verbs to the functional elements they are today (Heine 75). As a result, at any given
period of the language’s history we cannot rule out the functional roles of modals any
more than we can rule out their semantic content. Motun cannot therefore be fully
Acknowledging motun as a functional head that does not share the language of
its complement, merueri, we are left to puzzle out the idiosyncratic form of that
complement. As Blake explains, the word “contains the ending of the passive
infinitive attached to the stem of the preterite” (86). The editors of the Anglo-Saxon
meritare, and mereri, but conclude that retaining the manuscript form “seems best in
35
the absence of any convincing emendation” (Dobbie 279-80). Considered from a
infinitive for a Latin word that was historically inflected both in the second
conjugation as mereo (in which case the infinitive would be merere) and as the
deponent mereor (for which one would expect the infinitive mereri) (Stelten 161).
The strange morphology of merueri may show the strains of fitting a Latin word
into a doubly foreign context: first, the linguistic context of a modal phrase, and
second, the poetic context of macaronic alliterative verse. Since Latin indicates verbal
mood through inflection, it has no need for modals. Therefore, the Anglo-Saxon poet
trying to make a Latin verb function as the complement of an Old English modal
would be required to innovate an appropriate form for the verb. The infinitive is a
natural choice, since Latin verbs indicating ability or obligation, such as possum, take
infinitival complements and Old English modals, like their Modern English
successors, do likewise.
This hybrid verb form also indicates the degree to which the poetic structure of
this line overrides a variety of linguistic principles, phonological and syntactic as well
as morphological. To begin with, neither “proper” infinitive form, mereo or mereri, can
occupy an entire half-line—incorporating the -u- of the preterite allows the word to fill
the four mandatory syllables of the verse. To do so, however, the word must be stressed
differently, in what Cain calls “an apparent conflict between prosodic stress and
metrical ictus” (284 n. 30). Although merueri would normally be stressed on the
36
penultimate syllable, Cain suggests that the first syllable might have been lengthened,
as other Latin open syllables in the poems seem to have been, to maintain the proper
verse form (285). Since merueri occupies the second half-line by itself, any head taking
it as a complement must perforce occur in the first half-line, on the other side of the
linguistic divide. Poetic unity seems to have compensated for the linguistic disunity,
though. Not only does the modal motun set up alliteration on m; the inclusion of the adverb
her also creates assonance on u and e and further consonance on r: motun her merueri.
Prayer,” the languages switch between the conjunction and and the head of its
head, this switch violates the Functional Head Constraint.27 Classifying coordinating
26
Editors have cast the inflection of rogo in “A Summons” into doubt, depending on how they
interpret the line and its textual transmission. Dobbie holds that the “correct” form of the Latin verb is
the imperative roga, leading to a translation along the lines of “and you, suppliant, ask the truth-fast
one” (185) Such a reading would make the imperative roga parallel to the imperative bidde in the
following line. Since the subject of the imperative in unexpressed, the pronoun †u according to Dobbie’s
reading would be a vocative—a direct address that is not a part of the fundamental structure of the
sentence.
27
In current theories of generative grammar, the head of a clause is the category I, for “Inflection,”
comprised of features such as a tense, aspect, and number. Most of these features are mapped onto the
verb and indicated morphologically. In the case of two conjoined clauses, the second clause, headed by
37
may not contradict received wisdom, calling them heads is another matter. According
to X-bar theory, heads characterize the phrases in which they occur, projecting their
unit of linguistic structure that encodes the features of the conjunction itself, requiring
Traditional descriptions, on the other hand, treat coordinators as connecting words but
functional heads is put forth by Janne Bondi Johannessen in her study Coordination.
By applying the criteria of Arnold Zwicky, Richard Hudson, and Stephen Abney, she
makes a convincing case for the coordinating conjunction not only as a head (in
accordance with Zwicky and Hudson’s criteria) but also as a functional head (in
resemble functional heads in the following ways. First, coordinating conjunctions are
conjunctions besides and, or, but, for, and so can be added to the lexicon. Second,
coordinating conjunctions usually are small words that do not receive stress and in
languages other than English can even be added to other words, as in the Latin
conjunctions –que (and) and –ve (or). Third, coordinating conjunctions do not bear
I, is the complement of the conjunction; therefore, a switch between the conjunction and I is
ungrammatical.
38
“descriptive contents”; they indicate linguistic relationships between and among
words that carry meaning rather than express their own semantic value. Finally,
coordinating conjunctions cannot be separated from their complements (in this case,
the second word or clause that they conjoin) (Johannessen, Coordination 74–103).
Citing languages that allow for the verb to agree with the second of two conjoined
demonstrates that conjunctions are functional heads because they encode the features
conjunction and the second word or clause that it conjoins violates the Functional
Head Constraint. The switch between Old English and and Latin rogo in the present
cause this linguistic division. The complement of and does not occur in the on-verse
because the object of the verb, †a so∂fæstan, fills out that portion of the line
metrically.28 Consequently, the primary constituent of the complement does not occur
until the second half-line, requiring the verbal inflection to be in Latin. The
ungrammatical switch may seem less abrupt, however, because the conjunction and
the verb appear at opposite ends of the line, concealing the mismatch between
28
While verb-final order is generally considered to be characteristic of Latin clauses of all types
(with significant stylistic variation), in Old English it is found particularly in clauses introduced by
coordinating conjunctions and other connecting words (Pinkster 168–69; Mitchell, Old English Syntax
39
1.4 Code Switching as an Indicator of Latin Competence
switches in the poems gives us some idea of how well the poets knew Latin. As
Shana Poplack argues, the degree of integration of the two languages and the
grammaticality of code switches increase with the competence of the speaker; the
more fluent the speaker is in the languages at his or her disposal, the more likely he or
she will switch languages within rather than between sentences and the more
grammatical those switches are (“Sometimes I’ll Start” 615–16). The high proportion
switches therefore suggest that the poems were written by people with considerable
Summons to Prayer” only because it is the longest poem of the three and therefore
Although the poems display a variety of different kinds of code switches, the
poets’ preferences for switches involving nominals and adjuncts reduce the risk of
their violating the Functional Head Constraint. Such classes of words come with few
syntactic strings attached, making them more likely candidates for switches (cf.
Poplack, “Sometimes I’ll Start” 602–03). Switches between adjuncts and the words
they modify also run little risk of being ungrammatical because they steer clear of
§§905, 3091). The placement of the object †a so∂fæstan between the conjunction and and the verb rogo
is therefore appropriate for both languages.
40
combines the optimal features of both kinds of constituents. Thus, while poetic
conventions seem largely responsible for the few ungrammatical switches in the poems,
us to reevaluate their status within the corpus. Rather than being overly ambitious
scraps of bilingual doggerel, they are in fact the products of poets well versed in the
dual modes of literacy in Anglo-Saxon England. As such, they embody the language
contact that had already produced changes in the Old English lexicon and would
considered how the poems brought Latin and Old English together at the most
fundamental syntactic level, we may now relate these poems to other texts circulated
within the same monastic communities during the same linguistic convergence.
41
Chapter 2: Textual and Social Motivations for
Code Switching in Anglo-Saxon Poetry
Exclusively maintaining such a syntactic focus, however, does not allow us to explain
why the poets switched languages in the first place. To explore the motivations for
poetic code switching, we must turn from the syntactic study of the phenomenon to
cohesion, and as the work of William Labov demonstrates, “a very important factor in
group identification, group solidarity and the signalling of difference” (Trudgill 12).
Moreover, as John Gumperz argues, in many cases “the motivation for code
switching languages, speakers group utterances into “contrastable units” and provide
purposes (Gumperz 89–90; Romaine 161). Although attitudes toward code switching
vary from community to community,1 the practice can be a means of positive group
1
For example, in a study of Spanish-English bilinguals from southern California, Toribio finds that
even people who regularly switch languages in conversation may dislike the behavior. On one extreme
was a respondent who felt that code switching “reflects a deterioration or inadequate mastery of both
languages” (“Spanglish” 127). Other respondents felt more comfortable code switching but noted that
they did not find the practice to be aesthetically pleasing and that it bothered them when people switch
languages (“Spanglish” 118, 133). By contrast, Spanish-English bilinguals in other parts of the country
express more positive attitudes toward code switching, as Shana Poplack shows in her study of Puerto
Rican bilinguals living in New York (“Contrasting Patterns of Code-switching”).
42
characterizations of code switching on the speech of modern speakers, but their ideas
also clarify the motivations of Anglo-Saxon bilinguals for alternating between Old
As the ensuing discussion will demonstrate, the bilingual poets who switched
between Old English and Latin in “The Phoenix” and “A Summons to Prayer” and
the poems’ relationships to other texts and to signal their own membership in a
prestigious community. The switches between Old English and Latin and the isolated
adjoining monolingual works and connect the poems with the liturgies whose
celebration constituted the primary focus of Anglo-Saxon monastic life. Making these
liturgical connections enabled the poets to position their own works within the textual
to Prayer,” and “Aldhelm” are similar in that they are placed near texts that either
translate Latin or present it untranslated. 2 The coda to “The Phoenix” follows that
Summons to Prayer” occurs one page after “Judgment Day II,” an Old English
translation of Bede’s De die judicii. And “Aldhelm” comes just before its namesake’s
2
Robinson makes a similar observation in “The Rewards of Piety” 196–98.
43
Prosa de virginitate. The linguistic alternation in all three poems points toward the
origins, whether textual or authorial, of adjoining texts. In the first two poems, the
inclusion of Latin in an otherwise Old English context evokes the Latin sources
and Greek in an otherwise Latin context characterizes the author of the treatise that
The coda to “The Phoenix” and “A Summons to Prayer” are similar in their
by new material explaining the allegorical importance of the creature and its
relevance to the Christian life. The ending of the poem enjoins the reader to apply the
preceding material to his or her own life, moving beyond commentary to a first-person
recognition of God’s gifts and a prayer for continued blessings in the hereafter.
Prayer.”3 These three works are closely related, recorded as they were by the same
scribe and placed on successive pages in Cambridge, Corpus Christi College 201
Summons” are in fact one work, which he calls the Rewards of Piety. As he explains,
3
Whitbread also observed the similarity between the sections of “The Phoenix” and the poems in
CCCC 201, but considered “An Exhortation to Christian Living” and “A Summons to Prayer” to be
separate “postscripts” or “supplements” to “Judgment Day II” rather than parts of a single,
thematically unified work (“Notes on Two Minor Old English Poems” 127–28; “The Old English
44
although editors have traditionally presented “An Exhortation” and “A Summons” as
separate poems, the two supposedly distinct works are grouped together in the
manuscript. As Robinson observes, the macaronic section begins with a capital not
much bigger than the one marking the final section of “An Exhortation” (“The
isolated from the preceding translation of Bede’s De die judicii. Rather, these works
function as a group with its own distinct program. “Judgment Day II” is separated
from “An Exhortation to Christian Living” by a prose rubric announcing “Her enda†
rice!fara† …” (Here ends this book, which is called Inter florigeras, that is in English
“among the flowering [ones]” who journey to God’s kingdom) (Harlow pl. 23.5). The
out, echoes the words blowende and rice in its second line (“The Rewards of Piety”
199 n.7). Moreover, in the opening proposition of “An Exhortation,” “Gif †u wille †æt
blowende rice gestigan,” the demonstrative †æt refers deictically to the heavenly
kingdom described in “Judgment Day II” and mentioned in the rubric. “An
Day II” by shifting from depiction of the blessed life to instruction on how to obtain it.
Poems of the Benedictine Office” 48–49; “The Old English Poem Judgment Day II and Its Latin
Source” 646).
45
The “Judgment Day II”-Rewards of Piety grouping closes with the macaronic
“A Summons to Prayer,” which asks God to bless the penitent so that he or she may
attain the eternal reward promised and explained in the previous sections. Here again,
manuscript layout reflects the thematic unity of the texts. The final words of “A
Summons,” regna caelorum, are spaced more widely than any of the lines of
“Judgment Day II” or “An Exhortation,” so that regna takes up about a third of a line
and caelorum occupies an entire line by itself, followed by a blank row.4 The end of
the macaronic section is therefore set off more noticeably from the following text
(“The Lord’s Prayer II”) than any of the previous, supposedly separate texts are from
we see that just as in “The Phoenix,” an Old English translation of a Latin work is
followed by an explanation and application of that work’s theme to the moral life and
a macaronic closing that relates the preceding material directly to the reader or
listener. Whereas “The Phoenix” brings the point home by shifting to the first-person
plural, “Hafa† us alyfed lucis auctor,” “A Summons to Prayer” relates the preceding
lessons to its audience even more directly: “Îænne gemiltsa† †e,” followed by a
By switching between Old English and Latin at the ends of these programmatic
groupings, the poets manifested the Latinity that inspired their works in the first
4
Robinson notes blank space between “Judgment Day II” and “An Exhortation,” but none of the lines
between these texts are entirely blank (“The Rewards of Piety” 194; Harlow pl. 23.5).
46
place—the Latin poems of Lactantius and Bede. I do not mean to suggest, however,
that the Latin elements in “The Phoenix” are direct borrowings from the Carmen de
ave phoenice, nor that “A Summons to Prayer” quotes De die judicii.5 Rather, the
relationship between the Latin in the macaronic poems and that of the source texts is
iconic, representing the fact of Latinity. Had Latin not been used in the poems, the
translations they follow would appear no different on the surface than texts conceived
and written entirely in the vernacular. The macaronic endings thus function as
linguistic colophons: by switching from Old English to Latin, they announce the
Latin origins of the preceding sections.6 I will explain further how the macaronic
passages interpret these Latin sources in Section 2.2.2, which demonstrates how the
different way. It introduces the Prosa de virginitate, proceeding into Aldhelm’s own
preamble to his Latin work. Unlike “The Phoenix” and “A Summons to Prayer,”
which introduce Latin into an otherwise Old English context, “Aldhelm” brings
Old English and Greek into Latin surroundings. Since the Prosa de virginitate is not
translated but in the original language, “Aldhelm” does not foreground the Latinity behind
5
Whitbread does note that some of the Latin in “A Summons to Prayer” translates expressions found
in “An Exhortation to Christian Living” (“Notes” 129). These similarities may arise, however, because
“An Exhortation” adopts the same Latin liturgical vocabulary as the macaronic ending. In Sections
2.2.2 and 2.2.3, below, I examine how “A Summons to Prayer” employs this liturgical vocabulary; see
p. 65, n. 22.
6
Whitbread states more generally that each macaronic ending “might by its use of Latin round off
the structure with some show of symmetry” (“Notes” 128).
47
Prayer” do. Instead, “Aldhelm” evokes the author of the treatise and adopts salient elements
come to his aid more often “æne on e†le”—either “alone in his homeland” or, if one
reads æne with Napier as Greek ainh, “fame in his homeland” (Dobbie 194).
presents the nation as a modern phenomenon, Kathleen Davis demonstrates that the
same nation-making processes Anderson and others ascribe to the modern world were
also at work in England at least as early as the ninth century.8 The double
conflated with a real location as the name of a people, Angolsexna, is juxtaposed with
the name of a place, Bretene.9 Consequently, as Davis explains, “place and people
7
On the denotation of Angolsexna, see n. 8, below.
8
Davis explains the construction of national identity in Alfred’s famous preface to Gregory’s
Pastoral Care. Nicholas Howe, writing before the publication of Anderson’s Imagined Communities,
examines similar processes at work in Exodus and Beowulf.
9
The poem is unique in its use of the term Angolsexna—a search through the Dictionary of Old
English Corpus reveals that no other vernacular poem from the period does so. Originally formulated
48
become indissociable” (620–21).
past in which one figure, Aldhelm, distinguishes himself as both poet and ecclesiast
throughout the æ†el Angolsexna. Such a remembrance of a past ideal “not only posits
the nation as a preexisting, homogenous entity, but also authorizes the contemporary
emphasis). As a result, the poem’s depiction of Aldhelm tells us less about the
political reality of the poet-bishop’s own time than it does about the political
a unified English kingdom was enabled in part by the monastic reform movement that
made possible the production of the manuscript containing “Aldhelm.” Through the
culture” sanctioned and in many cases enabled by the monarchy (Davis 627).
Although the use of Old English is not in itself sufficient to establish Aldhelm’s
product. As such, the incorporation of Old English points to both the national and
by continental Latin writers to distinguish the Saxons of England from those living on the continent,
the term had “become rare by the later tenth century” and did not reenter common usage until the
sixteenth. The use of Anglo-Saxon to refer to all of the Germanic inhabitants of early medieval
England, regardless of their continental origins, is a modern innovation (Reynolds 398).
49
If switching between Latin and Old English helps indicate Aldhelm’s
nationality, the inclusion of Greek emulates the style Aldhelm was known for—one
Reform. As Michael Lapidge explains, Aldhelm was one of the original exemplars in
England of the so-called hermeneutic style of Latin writing, which featured the
(“Hermeneutic Style” 105). Aldhelm was one of the principal authors studied in
Anglo-Saxon schools and attracted especial interest for his novel use of language
advanced this style in their own works and made frequent use of Greek loanwords;
every known tenth-century Latin author or work which has survived, [with] only one
notable exception: Ælfric,” who purposefully rejected the style for the sake of clarity
The poem “Aldhelm” was produced at Canterbury, an active center not only for
the reform movement but also for the promulgation of the hermeneutic style (Ker
writings were studied all over England, that tenth-century writers at Canterbury
should emulate his works was especially apt because he himself studied there almost
three centuries earlier. Adherents to the hermeneutic style at Canterbury include Oda
and Dunstan (both archbishops of Canterbury and leaders of the reform), and Oda’s
student Frithegod, who in turn taught the reformer Oswald (Lapidge, “Hermeneutic
50
Style” 111). Like Aldhelm himself, later practitioners of the style need not have had a
Rather, they derived Greek words from the writings of Aldhelm as well as from
Greek-Latin glossaries, which provided the Greek words used in the poem “Aldhelm”
appears in the same manuscript as “Aldhelm” and the Prosa de virginitate, along with
the first seventeen lines of another work studied as a model of the hermeneutic style,
the “Aldhelm” poet matched the style of the manuscript’s principal text—a style
Aldhelm himself.
Since, as I explained at the beginning of this chapter, code switching can serve
well as texts. The bilingual poets who produced “The Phoenix” and “A Summons to
Prayer” did so in part to signal the Latin origins of the preceding vernacular
translations; however, they did so in a way that promoted their identity as bilingual
speakers. By switching languages, these poets could be seen, like Aldhelm, as both
10
Lapidge locates all of the Greek terms used in “Aldhelm” in Georg Goetz’s Corpus Glossariorum
Latinorum (Leipsig: B.G. Teubner, 1888). He acknowledges that Goetz’s compilation “gives only an
approximate notion of what words might have been found in a [tenth-century English] glossary”
51
“æ†ele sceop” and “bonus auctor,” expressing a confluence of languages that their
monastic audiences could admire and with which they could identify.
macaronic poems with neighboring texts, it also connects them on a larger scale with
the textual community that produced the manuscripts themselves, one brought
As part of the same movement, liturgical texts previously unknown to England were
introduced from the continent: the chants of the Mass were expanded by adding new
lyrics and melodies called tropes, while a new repertoire of hymns was incorporated
into the daily cycle of liturgies known as the Divine Office.12 Comparison of the
Latin elements in the poems with these liturgical texts reveals a number of striking
because the Corpus brings together dozens of glossaries, many of which were produced after the tenth
century (“Hermeneutic Style” 115).
11
The beginning of the Benedictine Reform is conventionally dated to Dunstan’s appointment as
abbot of Glastonbury in 940 or 943. Dunstan became archbishop of Canterbury in 959 or 960, the first
of a series of archbishops from reformed monasteries (Foot 179; Brooks 258). Dunstan’s disciple
Æthelwold in turn spread the reform to Winchester during his tenure as bishop there (c. 963–984).
Seminal surveys of the reform include Knowles, The Monastic Order in England 18–62; Stenton,
Anglo-Saxon England 433–69; and Symons’ introduction to his edition of the Regularis concordia. For
a discussion of the proliferation of Latin and vernacular works during the reform, see Lapidge,
“Schools, Learning and Literature” 31–48.
12
Although the practice of troping is believed to have originated on the continent in the late-eighth-
and early-ninth centuries, the earliest surviving tropes date from a century later. The trope continued to
be a vital liturgical genre until around 1150, and particular kinds of tropes persisted well into the
thirteenth century. Planchart summarizes the early development of the trope and thoroughly traces the
relationships among tropes from English and continental manuscripts (1: 131–291). On the impetus for
troping, see van Deusen 165–200. The subsequent development of the trope genre, particularly in
France, is traced by Fassler in Gothic Song. On the introduction of the new hymn repertoire to
England, see Section 2.2.3.
52
correspondences on a functional as well as a lexical level; that is, the poems not only
Although several scholars have noticed a similarity between the Latin in the
poems and Latin liturgical texts, they have not explained such correspondences in
detail. For example, as mentioned in Chapter 1, Brian Shaw notes that the macaronic
ending to “The Phoenix” “relies for a degree of heightened feeling on the use of Latin
… and the liturgical echoes” but explains neither what this “heightened feeling”
entails nor what these liturgical echoes are (178). Leslie Whitbread similarly notes
that “For Latin (and Greek) phrases the author of ‘Aldhelm’ seems to have drawn on
liturgical texts” and that “the same may be true of some of the Latin in the ‘Summons.’”
Whitbread cites the Canon of the Mass and Matthew 8:2 and 10:7 as analogues for “A
Summons” and suggests that “Aldhelm” may have drawn on contemporary hymns, but
he does not specify what hymns inform the poems’ Latin elements (“Notes” 129).
In the subsequent discussion, I will explain in much greater detail how the
poems borrow from liturgical texts not only by adopting the vocabulary of tropes and
hymns but also by interpreting neighboring texts in the manner of tropes. The poems’
incorporation of the liturgical vocabulary is most obvious in the various epithets for
God they contain. But on a more fundamental level, the poems act as tropes by
developing the meaning of adjacent authoritative texts through the various modes of
scriptural exegesis. Thus, just as the bilingual poets who wrote the macaronic
passages cited the language of both eucharistic and official liturgies, they employed
the hermeneutic strategies of liturgical texts to extract figurative meaning from literal
53
accounts—in the case of “The Phoenix,” of the mythical bird’s life cycle; in the case
writer’s creation of his Prosa de virginitate. Moreover, by highlighting the lexical and
stylistic similarities of their own works to tropes and hymns, these poets
demonstrated that their works were part of the prestigious textual community created
As explained in Chapter 1, the three poems that interweave Old English and
Latin are all found in manuscripts produced between the late-tenth- and early-
eleventh centuries, the period of the monastic reform initiated by Dunstan in the 940’s
and promulgated by his disciples Æthelwold and Oswald. The places of origin
identified or proposed for the manuscripts containing the poems were all reformed
Summons to Prayer,” bears the strongest link to the reform since it appears along with
the partial Old English translation of the Regularis concordia, the English redaction
CCCC 201 may have originated, was reformed in 961 by Oswald (Lapidge, “Schools,
Learning, and Literature” 31). The various places of origins proposed for the Exeter
13
In addition to Æthelwold’s implementation of the reform at Winchester, two of Dunstan’s
successors as archbishop also had direct Winchester connections: Æthelgar, elevated to the
archbishopric in 988, was a disciple of Æthelwold and the first abbot of the reformed New Minster at
Winchester; Ælfheah, succeeded Æthelwold as bishop of Winchester in 984 and was consecrated as
archbishop in 1008 (Brooks 278–79).
54
Book—Exeter, Glastonbury, and Christ Church, Canterbury—were likewise reformed
houses. Exeter itself was reformed in 968, when King Edgar commissioned a group
of monks from Glastonbury, the starting point for Dunstan’s reform efforts, to
reestablish Benedictine practices there (Conner 29–30). Exactly when Christ Church,
Canterbury (where “Aldhelm” was produced and “A Summons to Prayer” may have
community is uncertain, but the earliest documentary evidence indicates that monks
occupied the priory by 1020 (Brooks 256). Although the poems, particularly “The
Phoenix,” may have been composed before the manuscripts were compiled, their
strong affinity with liturgical texts introduced to England during the second half of
the tenth century and recorded in contemporaneous manuscripts suggests that they
Not long before the manuscripts containing these poems were compiled,
liturgy, most likely according to the liturgical practices of the reformed monastery of
14
Based on his codicological study of the Exeter Book, Patrick Conner argues that the part of the
manuscript containing “The Phoenix” was composed before the English reform and derived its
inspiration from continental models (148–59). Since the reform movement on the continent began
almost a half century earlier, however, with the founding of Cluny in 910, English monastic poets
writing before the reform of their own communities could very well have based their poems on those
of poets in reformed continental houses. Even if the poet of “The Phoenix” modeled his work on a
continental poem composed before the Cluniac reform, this pre-reform model could have still drawn
on the language of tropes and hymns, since the development of tropes and the expansion of the hymn
corpus into the New Hymnal were well underway by the ninth century. On the beginnings of the
reform at Cluny, see Knowles, Monastic Order 28–30.
55
Fleury.15 His efforts are attested in three collections of tropes, the two so-called
Winchester Tropers (Oxford, Bodleian Library, Bodley 775 and Cambridge, Corpus
Christi College 473) and a later manuscript partially included in London, British
Library, Cotton Caligula A xiv. The Oxford manuscript dates from around 1050, but a
rubric and other textual references indicate that it is a copy of a manuscript composed
between 978 and 986, very close to the period when the Regularis concordia was
dated to around the middle of the eleventh century. Alejandro Enrique Planchart
convincingly argues, however, that the order of tropes in the manuscript and its
connections with Wulfstan the Cantor’s Vita Æthelwoldi suggest a date between 996
and 1006 (1: 26–33). The manuscripts, Planchart concludes, therefore “probably
reflect directly the Winchester liturgical customs at the time of St. Æthelwold” (1:
11). These dates also make the Winchester Tropers contemporaneous with those
The British Library troper postdates both the Winchester Tropers and the
poems, having been compiled around 1050 (Planchart 1: 50). Although the troper is
missing numerous folios, Planchart deduces that it “apparently contained virtually the
whole of the Winchester repertory, conflated with other continental tropes and a large
number of unica” (1: 49, 55–60). The two fragments with which the troper is
15
Fleury was a regular point of contact between reformers in England and those on the continent.
Æthelwold, unable to visit Fleury himself, sent representatives there to study the community’s
observance of the Benedictine rule and its performance of plainchant. Æthelwold’s fellow reformer
Oswald spent eight years at Fleury before becoming bishop of Worcester, and advisors from Fleury
56
currently bound—a twelfth-century troper and a collection of Old English saints’
for the eleventh-century troper based on this contextual evidence and its inclusion of
commerce between Canterbury and Winchester (1: 47–50). Following his usage, I
The tropes in these manuscripts embellish parts of the Mass by setting new texts
preexisting ones, and by interpolating new texts and melodies within established
chants. For example, whereas the basic text of the Agnus Dei is a threefold invocation,
[Lamb of God, who takes away the sins of the world, have mercy on us. Lamb of
God, who takes away the sins of the world, have mercy on us. Lamb of God, who
takes away the sins of the world, grant us peace],
a trope found in both Winchester Tropers varies each petition by adding new text (in
keeping with Planchart’s catalog, italicized words are those of the original chant):
attended the Council of Winchester to aid in the drafting of the Regularis concordia. See Knowles,
Monastic Order 39–61ff. for an account of how continental reforms were emulated in England.
57
[You whom John preached in the desert, rejoicing and saying, “Behold the Lamb of
God, who takes away the sins of the world,” have mercy on us. You who sit at the
right hand of the father, the one unseen king, have mercy on us. King of kings, joy of
angels, God, have mercy on us. Unfailing light, eternal peace, and redemption of all,
eia, have mercy on us.]
These kinds of additions are labeled differently in the tropers depending on which
parts of the Mass they modify: whereas tropes to proper portions (the Introit,
Offertory, and Communion, whose texts change depending on the occasion) are
generally labeled tropi, additions to the Gloria, Sanctus, and Agnus Dei are called
some, for example, call new texts set to newly composed melodies tropes but those
set to existing melodies prosulae (Bjork 2–3). Tropes to the Alleluia form a special
the melody of such an addition (the sequence proper) and its text (called a prosa or
prose) (cf. Fassler). For the sake of simplicity, however, I will refer to all
modifications to Mass chants apart from the Alleluia as tropes and to modifications of
Comparing the Latin phrases in the macaronic poems to texts in the Winchester
and Canterbury Tropers reveals how much the poems resemble tropes in their style
and function as well as their vocabulary. Like many tropes, the poems alternate
16
For example, in the Oxford Troper, tropes to the propers of feasts celebrating virgin saints are
rubricated “TROPI IN NATALE DE VIRGINIBUS” (“Tropes on the birthday of virgins”), those for
the Sanctus begin “INCIPIUNT LAUDES AD DULCIA CANTICA SANCTUS” (“Here begin laudes
to the sweet songs ‘Sanctus’”), and those for the Agnus Dei begin “INCIPIUNT LAUDES
RESONANT QUAE DULCITER AGNUM QUI VENIENS PECCATA PIUS TULIT IMPIA
MUNDI” (“Here begin the laudes which sweetly express the Lamb who, in coming, mercifully took
away the wicked sins of the world”) (Planchart 2: 18, 19, 20). Interestingly, like the texts it introduces,
this last rubric tropes the Agnus Dei.
58
languages. Like the tropes, the poems introduce and conclude preexistent
texts. Most apparently and perhaps most significantly, the poems and the tropes also
employ a remarkably similar Latin vocabulary. After analyzing the textual, stylistic,
“Aldhelm” share with the tropes, I will demonstrate how the poems employ the Latin
lexicon that tropes inherited from hymns, the ultimate source of this vocabulary.
Prayer,” and “Aldhelm” was a rare phenomenon particular to marginal texts. Like
these poems, a number of the tropes are themselves macaronic, alternating between
Latin and Greek. For example, the following Kyrie trope, which appears three times
in the Winchester Tropers, alternates Greek, grecized Hebrew, and Latin epithets for
the three persons of the Trinity. In the manner of a glossary, many of the Greek and
Hebrew terms are followed by Latin synonyms. A rubric in the Oxford Troper
Other tropes borrow isolated Greek epithets such as theos ymon, “our God,” into an
inclusion of Greek loanwords in this trope manifests the hermeneutic style popular
throughout the history of Anglo-Latin writing but especially cultivated during the
reform. In addition to the tropes, other Latin writings from Winchester display what
Michael Lapidge terms a “predilection … for the grecism” (“Hermeneutic Style” 139,
123–29). By alternating languages, the poets who wrote the macaronic poems
therefore adopted not only the style of the tropes but also the larger stylistic
movement of which the tropes themselves were a part. Rather than writing outside the
their affinity with tropes. As explained previously, all three macaronic passages either
conclude larger works or introduce them; the core of each of these works is a
preexistent Latin text, either translated or in the original language. Particular kinds of
the Introit, the introduction to the Mass, generally precede the established chant, acting
as an introduction to an introduction. On the other end, Alleluia tropes set new lyrics to
60
the closing melodic run of the praise song, allowing that conclusion to be sung to a
complete text rather than on the final -a. In time, such tropes became sequences, stand-
textual junctures calls attention to the sources of neighboring material. To explain the
significance of these adjoining texts, the poems resort to all four patristic modes of
grouping runs the entire gamut of medieval exegesis, here applied to the Latin source
texts that underlie both works. Both poems begin with the literal (the presentation and
readings (explanation of “The Phoenix” as a type for Christ and the exhortation to
moral behavior inspired by Bede’s vision of the Last Judgment, respectively) before
closing in the anagogical mode with a final look forward to the joys of a blessed
afterlife. Both the end of “The Phoenix” and “A Summons” signal this shift from
previously.
18
Fassler traces the development of the sequence from a trope of the Alleluia to an independent
liturgical form (38–43, 47–57).
19
The classification of scriptural signification into four “senses” is most closely associated with
Augustine and Gregory, but can be traced to Clement of Alexandria and Origen. These four modes
include the historical or literal, what the actual words of the text mean; the allegorical, how the text
metaphorically relates the tenets of Christianity; the tropological, how the text reflects precepts of
moral behavior; and the anagogical, how the text reveals the ultimate purpose and outcome of creation
De Lubac comprehensively treats the origins and development of the four senses of scripture. See also
Evans 67–71, 114–22.
61
“Aldhelm” likewise runs through a variety of interpretational modes, in this
case applied both to Aldhelm’s life and to the Prosa de virginitate: the literal account
of the treatise’s creation (“Thus a holy and just man accomplished in books, a good
author, Aldhelm, the noble poet set me,” ll. 1–3) is followed by a tropological
possesses must certainly never excuse himself from hardship, but he must always
eagerly pray for support through his spiritual consciousness in the little world,” ll.
11b–15) and an anagogical look forward to the blessings the fortis factor will bestow
on the righteous (ll. 16–17a). The treatise proper then shifts back into the tropological
methods of tropes and sequences. Scholars have long acknowledged that these
scriptural gloss interprets a particular lemma (cf. Fassler 20). Developing this idea,
William T. Flynn argues that medieval clerics equated liturgical troping with the
classical and medieval rhetoric involve embellishing an argument through the use of
ornamental language, liturgical tropes involve embellishing set chants through the use
of newly composed words and music. Liturgical troping, therefore, allowed for “an
interplay between the liturgy and scripture and ultimately provided a means of
scriptural interpretation and commentary within the liturgy” (Flynn 15–16, 48–51).
Drawing on the work of Nancy van Deusen, Flynn observes that sequences interpret
62
scripture by combining the various senses, facilitating the liturgical and exegetical
transition they make from an anagogical praise song (the Alleluia) to a literal Gospel
reading (53–54).20 Such tropes therefore offered the poets who wrote “The Phoenix,”
The comparisons I have made so far focus on how the modus operandi of the
poems resembles that of the tropes. The poems connect with these liturgical texts
most directly and most literally, however, in their numerous Latin epithets for God. In
“The Phoenix,” God is called lucis auctor (creator of light, l. 667) while “Aldhelm”
refers to the Divinity as fortis factor (strong creator, l. 17). Such epithets are
additionally referred to as auctor pacis (creator of peace, l. 5), salus mundi (Salvation
of the World, l. 6), summi filius (Son of the Most High, l. 8), factor cosmi (Creator of
the Universe, l. 9), mundi redemptor (Redeemer of the World, l. 12), clemens deus
(merciful God, l. 14), Christum regem (Christ the King, l. 18), summus iudex (highest
Judge, l. 27), and factor aeternus (eternal Maker, l. 28). Mary is renamed in the same
These frequent sacred invocations recall a similar usage in the tropes. Since two
parts of the Mass ordinary, the Kyrie21 and the Agnus Dei, already included repeated
20
Relating the four senses of scripture to liturgical tropes, van Deusen observes that the similarity
between the terms trope and tropological is by no means accidental, since “tropes which most closely
conformed to the analogical-tropological modus of interpretation were most widely circulated” (166).
21
For a list of extant manuscripts containing Kyrie tropes and a discussion of the tropes’ textual and
musical forms, see Bjork.
63
petitions to God, troping them often involved magnifying these petitions by renaming
the Deity with various epithets not in the original texts, as seen in the Agnus trope
given above and in the following Kyrie trope from the Oxford Troper:
The epithets for God in this trope—among them rex genitor, luminis fons, rerum
conditor, lux oriens, and vitae vis—superficially resemble the Latin epithets in the
macaronic poems. A more detailed survey of the tropes, however, reveals that many
of the Latin phrases in the poems echo these liturgical texts verbatim.
The following table lists Latin elements from the macaronic poems that also
appear in the tropes, sometimes with variations of case or word order—a difference I
will address later. The first column lists a phrase from a particular poem, indicated in
“Aldhelm.” Subsequent columns give the corresponding phrase in the Winchester and
64
Canterbury Tropers, the kind of trope in which that phrase occurs, and the incipit of
that trope. As can be seen, although some of the poems’ Latin elements do in fact
vocabularies of the poems and the tropes are not limited to epithets; nor are the
liturgical texts bearing such similarities limited to the epithet-rich tropes of the Kyrie
22
Whitbread cites Old English phrases from other poems in CCCC 201 that correspond to some of
the Latin phrases from “A Summons to Prayer”: clara uoce, clænre stefne (“Gloria I” l. 37); clemens
deus, rumheort hlaford (“Lords Prayer II” l. 63); luce perhennem, ece leoht (“An Exhortation to
Christian Living” ll. 35, 73). He argues that the Latin in these cases “is an attempt to translate phrases
current in contemporary O.E. religious verse, especially those in the poems surrounding the ‘Summons’
…” (“Notes” 129). Given the evidence from the liturgical texts, however, I think it much more likely that
these Old English phrases themselves adopt and translate the liturgical vocabulary that “A Summons”
presents in its original language; all three expressions appear in the tropes, as shown in Table 2.1.
23
Bo indicates Bodley 775; CC indicates CCCC 473; and Ca indicates Cotton Caligula A xiv.
Numerals following the sigla indicate the page of the trope text in Planchart, vol. 2. An asterisk
designates a work as unique to that particular Anglo-Saxon troper.
65
sine fine Communion Virginis antique sceleri CC, Bo,
Ca 224
sine fine Introit Postquam factus homo CC, Bo,
Ca 151
sine fine Introit Omnibus ecce piis CC, Bo,
Ca 186
sine fine Introit Quos mea perpetuo CC, Bo,
Ca 187
sine fine Kyrie Pater creator omnium Bo 257
sine fine Kyrie Kyrie rex sempiterne Bo 251*
salus mundi (S) salus mundi Introit Hodie natus est nobis CC, Bo
142
virginem almum (S) alma virgo Introit Adest alma virgo parens CC, Bo,
Ca 176
As Table 2.1 shows, the poem that most frequently echoes the trope vocabulary is “A
Summons to Prayer”—not surprising since this poem, the longest of the three,
Still more Latin elements in the poems do not echo the tropes directly but come
close to doing so. For example, lucis auctor in “The Phoenix” resembles origo lucis
in a Kyrie trope. Whereas “A Summons to Prayer” has mundi redemptor, the Introit
trope “Ecclesiae sponsus illuminator” and the Sanctus trope “Pater ingenitus” have
orbis redemptor (Planchart 2: 70, 323). Likewise, unica uoce in “A Summons” recalls
una uoce, which appears in the Offertory trope “Ab indignatione et ira furoris,” the
Communion trope “Laus, honor, virtus,” and the Kyrie trope “Christe redemptor,
miserere nobis” (Planchart 2: 215, 228, and 254). The poems glancingly cite the trope
I have listed these correspondences not to suggest that the poets directly quoted
specific tropes in their macaronic works but rather to demonstrate the degree to which
66
both the poems and tropes share a lexicon marked by a relatively small set of terms:
significations and attributions for God and Christ (deus, rex, iudex, redemptor and
clemens, summus), abstractions (lux, pax), and stock expressions denoting the eternity
and unity of heavenly praise (sine fine, laude perhenni, una uoce). The stock
expressions and designations for God ultimately derive from scripture and the
traditional (non-troped) chants of the Mass. The abstractions, however, are a feature
the macaronic poems and the tropes share with texts from outside the eucharistic
liturgy, namely the hymns sung as part of the Divine Office. As the following
subsection will explain, while the writers of the macaronic poems incorporated the
same sorts of terms into their own works, they varied these abstractions in telling ways.
Tropers share much in common with trope collections from throughout northwestern
Europe and Norman Sicily. Gunilla Iversen has identified certain thematic elements
Magloire, and Saint Evroult. Among these shared elements is the frequent naming of
God with abstract epithets like pax, lux, and sapientia—the same sorts of abstractions
that appear not only in the troped Agnus Dei and Kyries cited previously but also in
the Anglo-Saxon macaronic poems (compare these terms with lucis auctor and lucis
liturgical sources, including texts from the Gallican liturgy. Naming God with
67
abstract epithets is not original to these liturgical sources, however, as they
themselves reflect a centuries-old tradition first found in patristic hymns and poems,
such as those of Ambrose and Ennodius, and promulgated in the religious poetry later
composed by Rhabanus Maurus, Florus of Lyons, and Sedulius Scotus (Iversen 24–31).
In the next subsection I will explain in greater detail how the macaronic poems,
like the tropes, adopt the vocabulary of hymns recorded in hymnals produced in
England during the Benedictine Reform. Before doing so, however, I will consider
two final points of difference between the poems and the tropes. First is the difference
clemens Deus, the Gloria trope “Quem glorificant sancti angeli” has Deus clemens.
Such syntactic inversion makes the Latin elements conform to Old English word
order. Whereas in Latin an attributive adjective like clemens would typically follow
the noun it modifies, in Old English the adjective would come before the noun
(Mitchell, A Guide to Old English §159). Joseph Crowley has noted a similar
tendency in London, British Library, Royal 2. A. xx, a glossed prayer book compiled
at Worcester two centuries before CCCC 201. While Old English glosses of the same
texts in other manuscripts preserve the Latin word order, the Royal glosses
education and not regular monastic background” presenting the glossed prayers to
24
Note, however, that the abstractions in the poems do not name God directly. See below, p. 70.
68
In “A Summons to Prayer,” however, such inversion helps the switches between
languages conform to both poetic and linguistic constraints. Following Old English
word order facilitates the code switch by making the syntaxes of the two languages
more equivalent. As David Sankoff and Shana Poplack first observed, code switches
are more likely to occur where the syntaxes of the languages involved are congruent
(“A Formal Grammar for Code Switching”). Pieter Muysken subsequently notes that
alternational code switching is more likely to occur between languages that are
401–04). Ordering Latin adjectives within noun phrases according to Old English
conventions—while still placing Latin noun phrases where Old English grammar
allows—therefore creates a local equivalence between the languages that helps make
A second difference between the poems and the liturgical texts is that although
the poems incorporate some of the same abstract nouns used to refer to God in the
tropes and hymns, they generally avoid naming God solely with an abstraction. The
following trope of the Agnus Dei, for example, includes much of the same imagery of
69
The third line of this trope includes a string of abstract epithets: pax, pietas,
called auctor pacis rather than pax—“creator of peace” versus peace itself—and
kindness.” Likewise, numerous tropes and hymns refer to God as lux, but “The
Phoenix” has lucis auctor. In fact, of all the Latin nouns used to designate God in the
poems, only one, salus mundi, is an abstraction. In other cases, the abstract terms
abstract epithet with a genitive or adjectival modifier (for instance, lux lucis [“light of
light”] or pax vera [“true peace”]) (26). Prosodic factors no doubt influenced this
point of difference, since inflecting single-syllable nouns like lux and pax in the
genitive allowed them to fit Old English meter, as discussed in Section 1.3.2.1.
Benedictine life, and almost every liturgical hour included the singing of a hymn. The
singing of hymns and other sacred music was so prevalent in Anglo-Saxon monastic
life that during the summer, as Mary Berry calculates, a Winchester monk would
have sung for eleven hours each day on average (151). The prevalence of hymns in
70
monastic life, moreover, makes it all the more likely that monastic authors would
Those hymns sung as part of the Office in Anglo-Saxon England can be found
in eight surviving manuscripts from the tenth and eleventh centuries, items K1
Old English Liturgical Terminology.” These manuscripts incorporate hymns from the
New Hymnal repertory, developed on the continent during the ninth century and most
Municipale, 231 [A 44] and Vatican City, Biblioteca Apostolica Vaticana, Reg. lat.
338) are neither available in print nor collated in Gneuss’s Hymnar und Hymnen im
Englischen Mittelalter, I will exclude them from the present study. The six remaining
hymnals are listed below according to the order in which they will be discussed;
preceding each manuscript’s shelf number is its siglum and, in parentheses, the item
25
See Gatch 341–62 on the sources used to reconstruct the Office as it was practiced during the
reform period and the relevance of such liturgical study to the interpretation of Ælfric’s homilies and
of Old English literature in general.
71
B and D date from the same period as the Winchester Tropers and the
manuscripts containing the Anglo-Saxon macaronic poems. The former, called the
Canterbury Hymnal, shares the same manuscript as the Bosworth Psalter and dates
from the second half of the tenth century. The latter, called the Durham Hymnal,
dates from the first half of the eleventh century (Gneuss, “Liturgical Books” 119).26
Both manuscripts were written in Canterbury and therefore reflect practices particular
to the community that produced “Aldhelm” and possibly the Exeter Book.
H and V, both compiled in the middle of the eleventh century, assign hymns to
the same hours and feasts as B and D. The former, in the manuscript known as the
Leofric Collectar, originated at Exeter while the latter possibly originated at Christ
Church, Canterbury (Gneuss, “Liturgical Books” 119). Gneuss proposes in his study
of the hymns that the order shared by B, D, H, and V was formulated at Christ
J may have originated at Christ Church, Canterbury, and dates from the middle
of the eleventh century; C can be traced to Worcester and is possibly the latest of the
six, dating from around 1065 (Gneuss, “Liturgical Books” 119). J and C differ from
the other four in the hours to which they assign particular hymns. According to
Gneuss, since many of the hymns in J and C are assigned to the same hours in the
Regularis concordia and Ælfric’s Letter to the Monks of Eynsham, their order was most
likely formulated at Winchester on the model of a hymnal from Fleury (Hymnar 69–74).
26
B and D have been edited by Wieland and Stevenson, respectively. Korhammer edits the Old
Testament (known as “monastic”) canticles in B, C, D, J, and V but not the hymns proper.
72
The provenances of these hymnals trace the same network of reformed houses
that produced not only the tropers but also the macaronic poems, a network
them to communities at Exeter and Worcester. It therefore comes as little surprise that
the poems echo the hymns as much as they do the tropers, and in certain
circumstances echo only the hymns. The following table shows the direct
correspondences between the poems and the hymns and indicates the manuscripts in
which those hymns can be found. I have cited hymns from B and D by their page
numbers in Wieland’s and Stevenson’s editions, and separated the Canterbury group
J, and V is from Table II in Gneuss’s Hymnar, to which I have added those hymns
found in H according to Dewick and Frere’s index (Gneuss, Hymnar 60–68; Dewick
and Frere 650–52). Gneuss also compares the contents of the hymnals with the hymns
27
Wieland in turn cites the hymns in B by their number in Dreves and Blume’s Analecta hymnica.
73
mundi redemptor (S) mundi sator et redemptor Alleluia dulce carmen B 69, D 55,
H, V; Ae, C, J
mundi redemptor Auctor salutis unicus B 86, D 79,
H, V; Ae, C, J
salus mundi (S) mundi salus Christe redemptor omnium B 60, D 39
sine fine (P) sine fine Iam rutilat sacrata dies et splendida B 64, D 47,
valde H, V; C, J
sine fine Christe caelorum habitator alme D 143
sine fine Christe sanctorum decus atque virtus B 80, D 69,
H, V; C, J
sine fine O pater sancte mitis atque pie D 145; J
virginis partu (S) partus virginis Audi redemptor gentium B 62, D 38,
H, V; Ae, C, J
partum virginis Veni redemptor gentium D 43, H, V;
Ae, C, J
As was the case with the tropes, several of the Latin elements in the poems
partially resemble phrases from the hymns without matching them word for word.
Table 2.3: Near Lexical Correspondences Between Macaronic Poems and Hymns
Phrase in Macaronic Poem Phrase in Hymn First Line of Hymn MSS
alta polorum (S), regna regna polorum Ecce iam noctis tenuatur umbra B 28, D 8, H,
caelorum (S) V; Ae, C, J
blandem et mitem (P) mitis atque pie O pater sancte mitis atque pie D 145; J
lucis auctor (P) lucis creator Lucis creator optime B 32, D 13,
H, V; Ae, C,
J
mundi redemptor (S) redemptor saeculi Iesu redemptor saeculi B 33, H
redemptor gentium Veni redemptor gentium D 43, H, V;
Ae, C, J
redemptor gentium Iesus refulsit omnium D 48, H, V;
Ae, C, J
salus mundi (S) omnium salus Audi redemptor gentium B 62, D 38,
H, V; Ae, C,
J
summus iudex (S) iuste iudex Nox atra rerum contegit B 46, D 23,
H, V; C, J
iudex omnium Hymnum canamus domino B 91, D 87,
(gloriae) V; Ae, C, J
unica uoce (S) voce simul una O pater sancte mitis atque pie D 145; J
74
A comparison of Tables 2.2 and 2.3 with Table 2.1 shows that the poems share
some Latin phrases with both the tropes and the hymns: Christum regem, clara uoce,
laude perenne, mundi redemptor, salus mundi, and sine fine. Moreover, most of the
hymns containing these phrases appear in all six of the hymnals, demonstrating that
the poems draw upon a hymn repertory equally broad as that referenced in the tropes.
The wide distribution of the hymns recalled in the poems and the correlation between
Winchester and Canterbury as loci of both hymn and trope production make it all the
more likely that monastic audiences would appreciate the liturgical resonance of the
As was the case with the tropes, the poems in several instances invert the word
order prevalent in the liturgical vocabulary by placing Latin words in a more typically
Old English sequence: clara voce for vox clara and virginis partu for partus virginis.
The latter example exhibits another tendency Crowley observes in the glosses from
London, British Library, Royal 2. A. xx. Where Latin would usually place the
genitive after the noun it modifies, the glosses place the genitive in its common Old
English position, before the noun (126 ff.). The poetic reasons for such inversions are
even clearer here; following Old English word order allows laude to alliterate with lof
in the first half-line (“Phoenix,” l. 676) and salus to alliterate with sibbe and gesæl∂a
(“A Summons,” l. 6). Moreover, bringing Latin syntax in line with that of Old
English once again facilitates alternating the two languages by creating a local
75
Some expressions in the poems, however, do not show such parallels with the
tropes and instead hearken only to the hymn tradition. Specifically, all three
macaronic poems use the word auctor—“The Phoenix” and “A Summons to Prayer”
to refer to God (lucis auctor, auctor pacis), and “Aldhelm” to refer to the man
himself (bonus auctor). Whereas none of the tropes refer to God as auctor, the
appellation is found frequently in hymns, in which God is called not only lux et
auctor, as seen in Table 2.2, but also auctor humani, auctor omnium, auctor saeculi,
auctor salutis unicus, and salutis auctor.28 Although God is not referred to as auctor
in the Vulgate Bible, Sedulius used the term in his alphabetic hymn “A solis ortus
cardine,” found in both the Canterbury and Durham Hymnals. The macaronic poets
auctor. Designating the Deity not only as a creator but as an author befits poems so
overtly about the business of connecting texts from both within and without the
liturgy. The “Aldhelm” poet’s use of bonus auctor to name his subject conveys all the
more praise given the sacred associations of the designation auctor in the hymns.
These multiple points of contact and between the macaronic poems and
liturgical texts situate the bilingual identity of the poets squarely within the
Benedictine Reform. In this context, code switching was a means of drawing attention
28
These terms are found in the hymns “Christe sanctorum decus angelorum” (B 108, D 116, H, V;
C, J), “Ad cenam agni providi” (B 87, D 82, V; Ae, C, J), “A solis ortus cardine” (B 65, D 50, H, V),
“Auctor salutis unicus” (B 86, D 79, H, V; C, J), and “Christe redemptor omnium, ex patre” (B 60, D
39, H, V; Ae, C, J), respectively.
76
community. Had the poets not switched languages and instead used only Old English,
the liturgical elements would neither stand out linguistically nor reproduce the words
the language and practice of both the tropes to the Mass and the hymns of the Divine
Office, these poets bridged the two types of liturgies that defined their monastic lives.
Using code switching to connect texts within manuscripts and within communities
77
Chapter 3: Macaronic English Carols: Sources and Structure
both in the languages of ecclesiastical and secular administration and in the new
orders founded by subsequent continental reforms. The English church came under
the control of French-speaking abbots and bishops, and, in the two centuries
following the Conquest, members of the old monastic order were joined by their
and expressed the changing linguistic and ecclesiastical scene. In the process, such
1
Although William the Conqueror was not able to appoint Normans to the various English abbacies
as quickly as he was to English bishoprics, a year after his death only one Anglo-Saxon abbot
remained (Knowles, Monastic Order 106, 111). Among the new monastic orders to arrive in Britain
were the Cistercians, whose first settlement on the island was made around 1114 in western Wales;
their first establishment in England occurred July 1124 in Lancashire (Knowles, Monastic Order 227).
The first community of Carthusians was established at Witham in 1180 (Knowles, Monastic Order
383). Operating alongside these monastic orders but in a much more public role were the Augustinian
canons, communities of which began to proliferate during the first half of the twelfth century
(Knowles, Monastic Order 140–42). The first mendicants in England were the Franciscans, who arrived
in 1224, fourteen years after their rule was officially approved (Knowles, Religious Orders 1: 130, 128).
78
Although multilingual poets writing in the thirteenth and fourteenth centuries
continued to compose works alternating English and Latin, they also incorporated
Anglo-Norman French into the mix.2 Their efforts are attested in manuscripts that
contain works alternating French with Latin or English, as well as combinations of all
three languages. One of the earliest English-Latin works from the period is found in
London, British Library, Egerton 613, written in a thirteenth-century hand different from
that of the surrounding Anglo-Norman material (Greene 299).3 The first stanza reads
Nearly contemporary with this poem are others that switch between French and
Latin. The following example is the beginning of a Marian hymn found in London,
British Library, Harley 978, a manuscript dating from the mid-thirteenth century and
2
David Jeffrey and Brian Levy note that few Anglo-Norman lyrics (macaronic or otherwise) have
survived from the twelfth century because Anglo-Norman rulers of the time employed continental
French and Occitanian troubadours almost exclusively (2).
3
Carol Harvey seems to have been unaware of this lyric, much less the Anglo-Saxon macaronic
poems, when she asserted that “it is in the later poetry of the Anglo-Norman period that the English
language is used for the first time in macaronic compositions” (75).
4
With the exception of the translation to Example 2, all translations are my own.
79
(2) Ave gloriosa
Duce creature,
Mater salvatoris
Virgine Marie,
Ave speciosa
Chaste, nette e pure
Virgo, flos pudoris,
E sanz vilenie (Jeffrey and Levy 106). 5
[Hail, full of glory: sweet creature, mother of the Savior, Virgin Mary. Hail, lovely
one: chaste, clean, and pure virgin, flower of modesty and without fault (Jeffrey and
Levy 107).]
Rossell Hope Robbins, David Jeffrey argues that most of the manuscripts that contain
Anglo-Norman lyrics (Harley 978 included) were compiled for use by mendicant
such collections (Jeffrey and Levy 3–6). As a result, Jeffrey argues, starting in the
forms of lyric poetry to England after continental Franciscan practice (The Early
5
Jeffrey and Levy label this poem as a “Macaronic Ave Maria,” but the terms gloriosa and speciosa
are found in the hymn “Ave regina caelorum” (Paroissien Romain 278). Other Latin elements in the
poem are either newly composed or derived from a variety of other Marian hymns and antiphons.
6
The major exception Jeffrey and Levy note is Harley 2253, which contains the Harley Lyrics (4).
This collection, compiled between 1314 and 1325, includes three macaronic works: the French and
Latin “Song Against the King’s Taxes” (Coss 182–87), the French and English “Mayden moder milde”
(Brook 66–8; Jeffrey and Levy 41), and “Dum ludis floribus” (Brook 55; Jeffrey and Levy 248), a
poem that switches between Latin and French in each line for the first four stanzas and then ends by
switching between Latin, French, and English.
80
Subsequent scholars have taken issue with the Franciscan hypothesis, however,
observing that no manuscript dated earlier than the 1300’s shows definite signs of
Franciscan ownership (Coss lviii–lix n. 94; Frankis 179). What indications thirteenth-
century manuscripts do contain point to other kinds of users. Harley 978, to return to
the previous example, lists the dates of death for abbots at the Benedictine monastery
of Reading and seems to have been kept there (Coss lvi; Frankis 176). Another
manuscript that contains only Anglo-Norman texts (London, Lambeth Palace 522)
originated at the Benedictine abbey of St. Augustine’s, Canterbury, but pays tribute to
the new mendicant orders in its illustrations. Preceding each major work is a
depiction of either the contemplative or the active life. Whereas the figures shown in
are Dominican and Franciscan friars (Frankis 177). Without denying the role of the
Friars Minor in popularizing and spreading lyric works in the fourteenth century,
then, we may acknowledge the involvement of other religious and indeed of lay
macaronic works,8 the most substantive body of Latin-English macaronic verse did
not emerge until the fifteenth century, when bilingual poetry became increasingly
7
See Frankis on the lack of evidence for mendicant involvement prior to the fourteenth century
(179–183) and indications of lay ownership (183–84).
8
Other manuscripts from the period that contain macaronic poems include Oxford, Bodleian Douce
137 (Jeffrey and Levy 233), Harley 913 (Coss 210–12), Reg. 12. C. xiii (Coss 251–253), and
Edinburgh, National Library of Scotland, Advocates art. 21 (the Auchinleck Manuscript) (Coss
253–58).
81
popular (Jeffrey, “Early English Carols” 213; Wehrle 64). By that point, the poetic
form known as the carol had caught hold in England and quickly lent itself to
bilingual expression.
The bilingual carols written in the fifteenth and sixteenth centuries continue
the general linguistic patterns of earlier macaronic works (for example, switching
languages between, rather than within, lines) but fit those patterns into a distinct
stanzaic form. The defining characteristic of the carol is its burden, a chorus sung
before the first stanza and between the stanzas thereafter (Greene xxix, xxx–xxxi;
Bukofzer 121).9 Works in other languages written in a similar refrain form include the
Latin cantilena and the Italian ballata (Greene cix–cxvii; Jeffrey, The Early English
Lyric 131–33).10 The carol was originally sung to accompany a ring dance: the burden
by the dancers themselves, and stanzas by a soloist or small group (Bukofzer 119). In
time, however, carols became too musically complex for dancing, so that by the
9
The carol’s straightforward textual form does not always indicate an equally straightforward
musical one. As Manfred Bukofzer explains, “in spite of its stereotyped literary structure the carol
could be set to music in various ways” (122). Although some of the carols recorded with music exhibit
the syllabic text setting and simple refrains we associate with modern Christmas carols, many were set
to complex polyphonic arrangements. For example, “The best song as hit semeth me,” a bilingual carol
in London, British Library, Addit. 5665 sets the burden and stanzas to intricate melismatic lines in
close counterpoint. The burden itself is sung twice each time it occurs, the first time by two voices and
the second time by three voices (Stevens 103). The disparity between the meter of the poetic text and
the rhythm of the music to which it is set makes guessing the musical character of carols recorded
without musical notation extremely difficult, if not impossible. Jeffrey ignores this disparity when he
concludes that the meter and poetic form of James Ryman’s carols, written around the same time as
those in Addit. 5665 and recorded without music, suggest that they were set to the melodies of familiar
Latin hymns (based on, for example, the “jolly, almost boisterous” rhythm of one of Ryman’s carols
[“Early English Carols” 219]).
82
fifteenth century the genre included multi-voice part songs and poems intended to be
to 1550, without which the present study would not be possible, contains almost five
hundred lyrics conforming to this pattern. More than two hundred of these works
Cambridge, St. John’s College S. 54, is typical of those that alternate languages to a
significant degree:11
10
For other similarities between the form of the English carol and that of continental dance songs,
see Greene liii–lxiii.
11
All carol texts are from Greene’s collection. I have omitted his editorial indications for
emendations and have set Latin words in italics.
83
Mete, drynke, and cloth—we haue no more
Pro nostri graui opere (Greene no. 274).
[Now Jesus, Ruler of the Soul, support us lest we fail. [1] God, who has made all this
world and has bought it with his precious blood, have thought of us sinful men;
support us lest we fail. [2] You are the Lord who made all things, for all grace is in
this gift. Save us from confounding by the fiend, our defender, O Lord. [3] We have
three enemies wherever we go: the word, the flesh, and the fiend. Save us from them,
that we may not disgrace ourselves; keep us from meeting them. [4] We are here out
of our entire free will. We have only woe, travail, and care, food, drink, and
clothing—we have no more because of our grievous deeds.]
As the above text demonstrates, medieval carols were sung on occasions other than
Christmas. Although the Christmas season and the feast days that fell within it were
indeed popular occasions for carol singing, carols also celebrated other feast days
throughout the year, offered ethical and catechetical instruction, and accompanied
Greene assigns these compositions illustrates this point: alongside “Carols of the
The final Latin line wrapping up each quatrain of this carol, called a cauda,
accentuates the stanzaic form. Such Latin caudae were typical of macaronic poems
developed later in this chapter will make clear, Example 3 resides at the more
macaronic end of the language alternation spectrum because it not only switches
languages in the burden but also includes new Latin material at the end of each
stanza. Although some carols contain even more Latin and switch more frequently
within sentences, most contain far less, anything from a single Latin word borrowed
84
into an otherwise English carol to an unvarying Latin refrain repeated at the end of
each stanza.
religious and lay people. Macaronic carols from medieval England are recorded in
forty extant manuscripts, fragments, and early print books from a remarkably broad
range of sources throughout England and Wales.12 Twenty-one of these sources can
lay users.
Ee. 1. 12) was written by the Franciscan James Ryman and compiled by him in 1492,
but his is the only macaronic carol manuscript with definite Franciscan origins
particular religious communities were the work of monastic (that is, non-mendicant)
British Library, Sloane 2593; Oxford, Bodleian Library, Ashmole 1393; and
12
By comparison, Greene’s compilation draws on 103 sources, indicating that nearly two-thirds of
those manuscripts, printed books, and fragments contain only monolingual carols. (Many monolingual
carols, however, are found in sources that contain other kinds of texts in Latin and French.)
13
Ryman, a member of the Franciscan community at Canterbury, was ordained as an acolyte and a
subdeacon in 1476. His impressive corpus of 166 carols constitutes almost one fourth of those extant,
and 77 of his works are macaronic (Little 4; Jeffrey, “Early English Carols” 210–211).
85
houses (Greene 306–07, 316, 332).14 A manuscript and a two-leaf fragment in the
British Library (Egerton 3307 and Additional 40166) most likely come from
Bodleian Douce 302, including five macaronic carols, was written by the Augustinian
canon John Audelay (Greene 317).16 Even the hermits of the Carthusian order seem to
have been involved in macaronic caroling: London, British Library, Lansdowne 379
contains a single macaronic carol as well as the prayers of a member of the London
generalization that “the mendicant friars, and particularly the Minorites, were
churches also went enthusiastically about the business of recording bilingual carols.
One British Library manuscript, Harley 3300, contains a carol largely in English but
with the refrain Sanctus alongside documents from the ecclesiastical court at Exeter
(Camargo 104–08). The Ritson Manuscript, also in the British Library (Addit. 5665),
includes nineteen macaronic carols and may also be an Exeter product (Greene
307–08).17 Worcester Cathedral and the college church of Beverley Minster have
14
Facsimiles of several pages from the Bodleian manuscript can be found in Stainer 1: xxvi–xxviii,
transcribed in 2: 61–65.
15
The carols in Egerton 3007 are recorded with music, edited in Stevens, Mediæval Carols.
16
The contents of this manuscript are available in Whiting’s edition. According to references within
the manuscript, completed in 1426, Audelay resided at Haghmond Abbey, near Shrewsbury. By the
time the manuscript was compiled, Audelay was both blind and deaf (Whiting xiv–xv).
17
Stevens, Early Tudor Songs and Carols, includes the music from the Ritson Manuscript pieces.
Exeter converted from a Benedictine community to a college of secular canons in 1050 (Knowles and
86
been identified as places of origin for two important carol manuscripts in the
two macaronic carols in his schoolbook, now Cambridge, Gonville and Caius College
383 (Greene 324–25). Finally, evidence of interest in macaronic carols among parish
clergy outside of England can be found on the back of an indenture dated August 8,
1471, for the rental of a church in Llanguollo, Wales (Bridgwater, Somerset, Town
attested by sources created by and for lay users. Two manuscripts known to have
been owned by lay aristocrats—Yale University Beinecke 365 and London, British
Library, Addit. 31042—are normally studied for other texts they contain, but each
also includes a macaronic carol.19 Another source contains texts strongly suggesting
aristocratic ownership: the macaronic carol in London, Lambeth Palace, Lambeth 306
is included alongside the Brut, courtly poems, hunting texts, and “a list of the retinue
of Edward III at Calais” (Greene 312–13). Merchants also collected carols that
alternate languages: two manuscripts in the British Library’s Harley collection (nos.
Hadcock 330). In addition to its Exeter connections, the Ritson Manuscript is also linked to Henry
VIII, whose own composition, “Passetyme with good cumpanye,” appears twice in the manuscript
(Stevens, Early Tudor Songs and Carols xvii, 338).
18
Music for the carols in Arch. Selden B. 26 is available in Stevens, Mediæval Carols.
19
The Yale manuscript, conventionally known as the Book of Brome, contains the mystery play
Abraham and Isaac and was owned by the Cornwallis family, who resided at Brome Manor in Suffolk
(Greene 335). Like many manuscripts in which macaronic carols are found, it is a commonplace book,
containing everything from saints’ lives to business records. The British Library manuscript, known as
87
4294 and 5396) include merchants’ accounts (the latter in the same hand as that of the
carols), and Oxford, Balliol College 354 contains thirty-four bilingual carols as well
as a staggering variety of texts collected by the London grocer Richard Hill starting
around 1508 and continuing until 1536 (Greene 303, 320–21; Parker, Commonplace
Book 48–49).20
Suggesting a still broader audience for macaronic carols are those that appear in
printed collections published in the sixteenth century. Three of the print sources for
the carols in Greene’s compendium contain bilingual works: a single leaf printed by
Wynkyn de Worde in 1521 (Bodleian Rawlinson 4to. 598 [10]), a four-leaf fragment
from a book possibly printed by William Copland around 1550 (Bodleian Douce
Fragments f. 48), and a pamphlet collection now at the Huntington Library in San
Marino, California (Greene 339–41). The one complete pamphlet in the Huntington
1542 and 1546 (Greene 340; Reed liv).21 Three of the carols in the Huntington
(James Ryman’s compilation), Egerton 3307, and Balliol 354 (Greene 43, 54, 220).
Fragmentary though they are, these printed collections indicate that macaronic carols
Section 5.3.2.
the London Thornton Manuscript, has been studied for the romances and alliterative poetry it contains
and was owned by a variety of northern English recusant families (Thompson 5–7).
20
Most of the contents of Balliol 354 appear in Dyboski’s edition. Parker includes a detailed
inventory of the contents of the manuscript and speculates on the interests and life of its compiler.
88
3.3 Categorizing Carols According to Degree of Language Alternation
Having described the variety of sources in which macaronic carols are found,
we may now take a more detailed look at how these lyrics alternate languages. To lay
the groundwork for further linguistic study, I will first generate a more specific
system of classification to describe bilingual carols and distinguish them from each
other. This system will separate those carols with the highest degree of language
alternation from those that do not include much Latin or that simply repeat the same
switches between Latin and English. In Chapter 4, I will analyze where languages
alternate in those carols, evaluate those switches according to the Functional Head
Constraint, and explain how changes in the locations of code switches over time
For the purposes of this study, macaronic is a variable term. While many
scholars use macaronic to describe any texts that include words in more than one
language, some works are more macaronic than others. The degree of language
considered in sequence:
21
Reed includes a facsimile of all three of these print sources. Greene observes, however, that
Reed’s bibliographical description of the Huntington collection is incomplete and inaccurate (340–41).
89
These criteria address the nature, frequency, and variety of language alternation in
the carols and allow us to compare the language alternation in a particular carol with
that of others. By evaluating carols accordingly, we may place them on the continuum
of language mixing first considered in Section 1.2: from borrowing and insertion to
less macaronic than one that contains code switches. A carol that switches languages
frequently and incorporates a greater variety of Latin phrases is more macaronic than a
carol that does not switch with as much frequency or variety. According to these
criteria, carols can be classified into twelve descriptive types, ranging from least
macaronic (Type A) to most macaronic (Type L).22 The following list describes each of
these types and indicates how many carols of that type can be found in Greene’s
edition. The Appendix specifies which carols conform to each of these types.
22
William Otto Wehrle, interested in matters of poetic form rather than linguistics, classified the
carols according to where Latin appeared in each stanza. His system, for example, distinguishes a
quatrain with a Latin fourth line (his Type A, “The Macaronic Cauda”) from one with a Latin second
line (his Type M, “Latin Second Line in Quatrain Stanzas” (65–89, 163–64). For my purposes, however,
both forms are identical because both involve the same degree of language alternation. Consequently,
several of my categories cut across his classificatory lines.
90
3.3.1 Type A: Isolated Latin Borrowings
otherwise English context. Such carols, hereafter designated Type A, are the least
macaronic according to criterion (a), above. Often the Latin words are borrowed
directly from liturgical sources and designate either abstractions or sacred epithets (an
interesting parallel with the Anglo-Saxon macaronic poems discussed in Chapter 2),
This single borrowing, Rex Pacificus, constitutes the only Latin in the six-stanza carol.
Still other borrowings in Type A carols occur within reported speech, again
usually borrowed from scriptural or liturgical sources.24 For example, one stanza from
(5) The ritche and the pore ther title did reherse:
The pore clamed heauen throughe his pacient havour;
He saide, “Beati pauperes,” and further the verse;
The riche man by ritches thought hym in favour,
For who was so ritche as was our Saviour? (Greene no. 95)
[The rich and poor uttered their titles: the poor claimed heaven through his (Christ’s)
patient behavior (or possession); he said “Blessed are the poor” and the rest of the verse.
The rich man thought that he was in favor because of his ritches, but who was as rich as
our Savior was?]
23
On the similar use of Latin epithets and abstractions in Piers Plowman, see Sullivan 93 and 97–98.
24
Linguistic changes after verbs of speech have also been noted among modern bilinguals (cf.
Gumperz 75–84; Romaine 162).
91
Numerous carols similarly quote Gabriel’s “Ave” to Mary at the Annunciation and her
response, “Ecce ancilla Domini,” returning to English after those memorable words.25
Although such carols can be called macaronic in that they involve a mixture of
languages, their paucity of Latin material and the fact that they generally include
simple borrowings rather than code switches make them the least macaronic carols in
the corpus. Many other carols contain borrowings, especially after verbs of speaking
such as saide in Example 5, but can be classified as more macaronic types by virtue
include a single line of Latin or several code switches between English and Latin
within a particular stanza. For example, one of James Ryman’s Christmas carols ends
This closing line involves a code switch between the prepositional phrase In ortu
Regis omnium and the preceding verb phrase, wille be—not simply a single borrowed
word. Similarly, a fifteenth-century carol by John Audelay ends after eight English
25
Examples include Greene nos. 172a, 248, and 253.
92
(7) Mater, ora Filium
That he wyl affter this outlere
Nobis donet gaudium
Sine fyne for his merce (Greene no. 311).
[Mother, pray to the Son that he will after this exile may (sic) give us joy without end
out of his mercy.]26
In Type B carols, stanzas containing Latin generally come at the end of the work—a
Prayer.”27 Such carols entail a much more substantive integration of Latin and
English than the single-word borrowings of Type A carols; according to criteria (a)
and (d), the carols contain code switches rather than borrowings and may alternate
languages several times within a particular stanza. However, although English and
Latin are more thoroughly integrated in one particular stanza, code switches do not
occur elsewhere, making these works less macaronic than carols that alternate
languages throughout.
Some carols switch between Latin and English in the burden but nowhere else.
Burdens from these carols, those of Type C, typically consist of two lines, the first in
College O. 3. 58:
26
See Section 4.3.3 for an explanation of this ungrammatical switch.
27
Examples of such carols in Greene’s collection include nos. 93, 148 B, and 239.
93
A few carols, however, reverse this arrangement, as in the burden to another fifteenth-
Still others have four-line burdens that are mostly Latin with a few words or a line of
interpolated English, as in this carol from a pamphlet grouped with Richard Kele’s
Christmas carolles newely Inprynted. As in Example 5, the code switch here involves
multiple code switches before and/or after each stanza, depending on where the
switches occur in the burden. These carols keep syntactic integration to a minimum,
however. Although the burden may itself contain an intrasentential switch, the
switches between the end of the burden and each stanza would by and large occur
between sentences (a distinction observed in criterion (b)). The Latin elements also
do not vary, making carols of Type C more macaronic than carols that do not switch
as often (according to criterion (d)) but less macaronic than carols with a greater
94
3.3.4 Type D: Latin Burden, One Line Repeated as a Cauda
By far the largest group of macaronic carols are those that start with a Latin
burden and repeat one line or a word of that burden as a cauda to each stanza, an
arrangement I will term Type D. Notable examples (still performed today even
outside of early music circles) include “Gabriell of hye degree” (Greene no. 238 C),
which in one of its earliest versions carries the burden “Noua, noua, ‘Aue’ fit ex
‘Eua’” (News! News! ‘Ave’ is made from ‘Eve’) and repeats Noua at the close of each
stanza, and the Agincourt Carol, the burden and first stanza of which are given below.
Dozens of Type D carols are preserved in the collection of James Ryman’s works,
and several can be found in manuscripts and printed books from the sixteenth century.
Interestingly, almost all surviving carols that include a burden entirely in Latin
repeat at least some of the burden in subsequent caudae.28 Modern editions of the
carol texts such as that in Example 11 obscure the fact that in performance, the
repetition of the burden would entail multiple code switches: a switch after the Latin
28
Only two carols from Greene’s collection, nos. 1 and 2, feature Latin burdens without Latin caudae.
Carols can include Latin caudae without having a Latin burden, however. See Type G, below.
95
burden, when the language changes to English for the bulk of each stanza, and again
before the Latin cauda to each stanza. As in carols of Type C, however, code
switches generally occur between sentences and do not vary, features that make such
3.3.5 Type E: Latin and English Burden, Latin Line Repeated as a Cauda
Carols of Type E switch between Latin and English in their burdens and then
repeat at least part of the Latin from the burden at the end of each stanza. This pattern
is illustrated by the following burden and stanza from one of James Ryman’s carols:
As in carols of Type C, the burden may also reverse the arrangement seen in Example
12, beginning with Latin rather than English. Although some Type E carols repeat the
burden text twice or follow the two-line burden with a single Latin tag line, none have
the four-line burdens seen in a few Type C carols, such as Example 10. The Latin line
in the burden of Type E carols augments the number of switches and often occurs
96
macaronic than those carols that do not switch within the burden but less than those
With the five surviving carols that have English burdens but end each stanza
to see the variety of code switches that makes these and subsequent types of carols
especially appropriate for linguistic scrutiny. Perhaps one of the best known examples
new music by Benjamin Britten in his Ceremony of Carols. This carol’s Latin
Looking back on the previous three carol types, we note an interesting distinction.
Unlike carols with bilingual burdens (Types C and E), almost all carols with entirely
Latin burdens (Type D) include Latin caudae. Carols of Type F, however, include
Latin caudae without featuring Latin in their burdens. Thus, whereas invariable
caudae, as in Types D and E, generally repeat material from the burden, variable
caudae, as in Type F, need not hearken back to previous Latin material. These changing
Latin elements therefore become a means of developing the carol text, contributing new
meaning to each stanza rather than repeating previously stated ideas. Moreover, because
Type F carols do not repeat the same Latin elements at the end of every stanza, the
Type G carols include Latin burdens and end each stanza with new Latin
material. The “Boar’s Head Carol,” the singing of which is still a tradition at Queen’s
College, Oxford, is a prime example of Type G. The following version, from a quarto
leaf printed in 1521 by Wynkyn de Worde, conforms to the Type G pattern in all but
the last stanza. Modern versions of the carol (from 1718, 1811, and 1921) end the
98
third stanza with a Latin cauda that makes explicit reference to Queen’s College
(Greene 379–381).
The variability of the Latin caudae in carols of Type G contributes to the earlier
characterization of carol types in Section 3.3.6. Carols with Latin burdens (Types D
and G) almost always have Latin caudae, and invariable caudae such as those in carols
of Types D and E generally repeat Latin material from the burden. In some carols with
entirely Latin burdens, however, these Latin lines do vary, as evidenced in Example 14.
Carols of Type H switch languages within the burden and incorporate new Latin
99
material at the end of each stanza. The relationship between carols of Types G and H
is therefore much like that between those of Types D and E, with the salient
difference that the caudae in carols of Types G and H vary from stanza to stanza. One
different feast day of the Christmas season in each stanza. The Latin in the burden
quotes the hymn for Christmas matins, Christe redemptor omnium, and the Latin
cauda to the second stanza quotes an antiphon for the feast of St. Stephen (December
26) (Greene lxxxvi, civ).29 Such Latin phrases therefore cite the texts of the Divine Office
even more directly than those in the Anglo-Saxon macaronic poems do, linking the carol to
Because languages switch within the burden and before the new material that ends
each stanza, carols of Type H are the most macaronic of those that whose stanzas
Unlike the previous kinds of carols, carols of Type I incorporate two lines of
30
Among the textual problems in this version of the carol are the repetition of convivio in stanzas
four and five and the repetition of knyghtes in stanzas five and six. In both cases, the second usage
101
Latin per stanza and thus switch between English and Latin more frequently. This
pattern is more common in works from manuscripts produced in the second half of
the fifteenth century and afterwards but is also seen in some carols recorded earlier in
the fifteenth-century.31 While the second Latin does not vary in earlier carols of this
form, later carols show more variety. Below, for example, is a carol from Bodleian
Arch. Selden B. 26, dated to the middle of the fifteenth century (Greene 314). Each
stanza contains a varying Latin second line and an unvarying Latin fourth line that
does not make sense. Greene notes that “the substitution [of knyghtes] for ‘kings’ may be merely a
copyist’s error” but offers no explanation for the repetition of convivio (344).
31
For example, British Library, Sloane 2593, which contains one carol with two lines of Latin per
stanza, has been dated to the first half of the fifteenth century (Greene 306). British Library, Egerton
3307, which contains two such texts, dates from around 1450 (Greene 299).
102
Bytwene mankynd and the Chylde;
Aue domina.
[5]
Lady, quene of paradyse,
Mater Dei electa,
Thou bare oure Lorde, that hye Justyse;
Aue domina.
[6]
With merthe and alle solempnite
Nato canunt omnia;
Thou berde of ble, welcome thou be;
Aue domina (Greene no. 179)
[Hail Lady, queen of heaven. [1] May your (Christ’s) birth be worshiped, whom you
carried, Mary, both in the home and in the city; hail, Lady. [2] For we were lost
through our sins by hellish punishment, but now he whom you have borne will save
us; hail, Lady. [3] It was almighty God’s will, happy fertile one, that his grace should
light upon you; hail, Lady. [4] May you be blessed, mild maiden, who are always a
friend between mankind and the Child; hail, Lady. [5] Lady, queen of paradise, mother
chosen by God, you bore our Lord, that high justice; hail, Lady. [6] With joy and all
solemnity, all sing to your son; you bride of bliss, may you be welcome; hail, Lady.]
In later Type I carols, new material in the second Latin line of each stanza adds to the
variety of the code switches. Other Type I carols may switch to Latin in the first and
3.3.10 Type J: Latin and English Burden, Two Latin Lines per Stanza
Type J carols switch to and from Latin not only twice in each stanza but also in
the burden. The majority of extant Type J carols date from the second half of the
fifteenth century (none are attested, however, in Ryman’s work). At least one Type J
carol was recorded around 1450, in Egerton 3307. Variants of this carol appear in two
strips in Bodleian Bodley 77 (Greene 12–14; 348–49). Greene notes that with the
103
exception of the line “Consors paterni luminis,” from an Epiphany hymn, and Mary’s
32
Incensum in the last stanza, as Greene notes, “should be ‘myrrham’” (349).
104
[Almighty Jesus, King of Bliss, takes on the flesh of a virgin; he was always and
always is a sharer of the Paternal light. [1] The holy Church brings him to mind: he
entered the bridal chamber of the womb; from heaven to earth to save mankind the
Father commissioned the Son. [2] A messenger came to Mary bearing the salvation
of humanity, and she answered with a quiet voice, “Behold the handmaiden of the
Lord.” [3] Through the power of the Holy Spirit entering the palace of the womb,
meekness is best above all things in the sight of the Most High. [4] Three kings upon
the twelfth day, a shining star leading, they took the way to seek our Lord, bearing
gifts. [5] A star (was) always before the kings: the first king offered gold. He is God
and true Lord: the second king presented incense. [6] He was named the third king who
gave over beautiful incense. May He bring us all to bliss who chose to die on a cross.]
Carols of Type J are the most macaronic of those carols operating within the
traditional confines of the carol form: quatrain-long stanzas with code switches
typically between, but not within, lines. Some carol writers produced even more code
switches, however, by going beyond these restrictions, either by writing stanzas more
than four lines long or by switching languages within the lines themselves.
3.3.11 Type K: Latin Burden, Three or More Latin Lines per Stanza
Carols of Type K, with Latin burdens and at least three Latin lines in each
stanza, are late and rare developments in the carol form: two were written by James
Ryman in the late fifteenth century and the other is recorded in the Ritson
Manuscript. One of Ryman’s carols, which tells of the Three Kings’ journey to
Bethlehem and Joseph and Mary’s flight into Egypt, has three Latin lines per stanza,
but each stanza ends with an English and Latin refrain that repeats a Latin line from
[The Sun of Justice has risen from an unblemished virgin. [5] These three kings were
warned in dreams by the Most High that they should by no means go back to the
most wicked Herod. Now we worship him born so nobly from an unblemished virgin.
[6] Not by Herod, that wicked night, but by another way they have gone home again
most truly by the providence of God. Now … virgin. [7] Joseph then fled, as well as
Mary, with the child into Egypt, where they stayed until King Herod departed from
this world. Now … virgin.]
The Ritson Manuscript carol, in which King Herod again plays a starring role, has
three lines of Latin per stanza but does not repeat Latin from the burden:
106
[2]
Membra figi tenera
Thou gauest thy comowndement,
Matrem tenens viscera.
Thy hope thou losste and thyn entent,
Sternit dum milicia.
[3]
Reus nunc extinguere
Infynyte and most of pyte,
Verens regnum perdere,
In sorwe and woo thy see ys dyghhte;
Vixit Dei milicia (Greene no. 109).
[Let praise resound throughout the world, the glory of the Innocents.
[1] Say, impious Herod, what do your cruelties avail, in chains for
bloodshed? Put in pain with great distress, does your army help you?
[2] That the young limbs be pierced, you gave your commandment, that the
mother clutch her innards. You lost your hope and your desire, as your
army scattered. [3] Now to expire guilty, exceedingly and most piteous,
fearing to loose your kingdom, your realm is given to sorrow and woe; the
army of God lived.]
Both of these examples demonstrate that although carols of Type K depart from the
quatrain stanza typical of so many other macaronic carols, they observe a similar
linguistic protocol by switching between lines but not within them and by alternating
lines of Latin and English. Type K stanzas may begin in either English or Latin,
however, and Latin lines need not always be followed by English ones: each stanza in
Example 19, for instance, proceeds directly from the fifth Latin line to the Latin burden.
3.3.12 Type L: Latin and English Burden, Three or More Latin Lines per Stanza
Two carols can be considered the most macaronic in the corpus because they
switch languages not only in the burden but also between or within more than two
lines per stanza. Even this small sample shows more stylistic variety than those of
(Greene civ–cv). Ryman doubled the usual quatrain carol stanza, producing eight-line
108
crucified and nailed upon a tree. He conquered death and lived again and took with
him those whom bitter death afflicted to be with him in bliss. [3] O Mary, flower of
virgins, truly most fair and sweet, like a rose or a lily whose blossom shall not fail,
secure our prayers to the Son, who is both God and man, for the salvation of the
faithful, that he may grant us bliss.]
This carol clearly illustrates how code switches can be shaped by prosodic
conventions. Within each stanza, lines in the same language rhyme, English lines
with English and Latin lines with Latin—a particular challenge for the poet to sustain
over eight lines. Other writers were not able to manage linguistic and prosodic
The other Type L carol is a lament on mortality found in Balliol 354. It contains
the same number of code switches per stanza as the previous carol but arranges those
switches quite differently, switching within rather than between lines in a fashion
states, “Terribilis mors conturbat me” (Terrifying death disturbs me), a reworking of
“Timor mortis conturbat me” (The fear of death disturbs me). The latter line from the
Office for the Dead appears in several other carols as well as in John Lydgate’s poem
of the same name and in William Dunbar’s “Lament on the Makaris” (Greene 441).
[Alas, my heart will break in three. Terrifying death troubles me. [1] That time of
youth, which is so foolish, lead me away into vain contrivance; I am ill, I may not
rise. Terrifying death troubles me. [2] When I was young, I feared little, but always in
sin I ate my bread. Now I have been led into my bed; Terrifying death troubles me.
[3] The body departs; in my soul a demon looks at his list; he himself desires to have
his toll. Terrifying death troubles me. [4] Christ himself when he had to die, his
humanity cried to his Father, “Look down on me, Father, who is so high; terrifying
death troubles me.” [5] Now I ask the Trinity, “Lead me from this vanity into heaven,
where there is joy with you; terrifying death troubles me.”]
Although no other surviving carols switch within lines as frequently as this lament,
macaronic songs and poems of other forms do, from the Anglo-Saxon poems studied
in Chapter 1 to sixteenth-century works like the part-song “Up I Arose,” in the Ritson
Manuscript.33
33
This part-song, a pastourelle in which the narrator tells of meeting a woman who laments her
pregnancy, begins “Up Y arose in verno tempore and found a maydyn sub quadam arbore that dyd
complayn in suo pectore, saying ‘Y fele puerum movere’” (“Up I arose in springtime and found under a
certain tree a maiden who complained in her breast, saying, “I feel the child move’”) (C. Miller 160, 244;
Stevens, Tudor Songs and Carols 19).
110
This survey of the various types of macaronic carols not only illustrates the
diversity of their forms but also makes it possible to distinguish those carols that
that either limit Latin material to a single word, line, or stanza or repeat the same
mention unrewarding—results: since Types C through E bring back the same Latin
again and again, in many carols the same code switch occurs multiple times. Only in
carols of Types F through L do switches vary from stanza to stanza. I have therefore
chosen in the subsequent chapter to limit my analysis to carols of these types. As the
chart in the Appendix shows, however, the corpus of bilingual carols is sufficiently
large that even focusing on the most macaronic yields sixty-five works for closer
study. After describing where Latin and English alternate within sentences in these
works, I will place code switching in the carols in its linguistic, historical, and social
context. The following chapters will illustrate a threefold diversity, as the variety of
forms macaronic carols take parallels the variety of code switches within them and
111
Chapter 4: Code Switching in Medieval English Carols
The period during which extant macaronic carols were recorded, extending
from the early fifteenth century through the middle of the sixteenth century,
fortuitously coincides with the transition from Middle to Early Modern English.1
Analyzing where code switches occur in the most macaronic carols—those of Types
alternation in the carols changed along with the English language and its usage. After
addressing whether the Latin phrases cited in the carols can in fact be considered code
switches, I will analyze how often and where language alternation occurs in the
carols. Comparison of carols written during the Middle and Early Modern English
periods will expose differences in code switching practice, marking the development
English contexts bear all the hallmarks of borrowings. For example, the following
verse from London, British Library, Lansdowne 379, written around 1500, includes
1
Historians of the language date the beginning of Early Modern English to around 1500 in
recognition of a variety of convergent technological, political, and cultural as well as linguistic
developments. The years leading up to and immediately following the turn of the sixteenth century saw
the introduction of the printing press to England (1476), the coronation of Henry VIII (1509), and the
beginning of the Renaissance and the Reformation (Fennell 2, 135; Lass 1, 6). During the same period,
the English language underwent the phonological change known as the Great Vowel Shift, in addition
to lexical and syntactic changes such as those mentioned on p. 138, n. 22 (Lass 11–12).
2
Compare this usage with the similar incorporation of Latin abstractions in Piers Plowman,
identified by Sullivan (97–98).
112
(1) A mervelus thyng I hafe musyd in my mynde:
Howe that Veritas spronge owghte of the grounde,
And Justicia for all mankynde,
From heuen to erthe he cam adowne (Greene no. 94).
[I have considered a marvelous thing in my mind, how Truth sprang out of the ground
and Justice for all mankind. From heaven to earth he came down.]
The carol writer borrowing veritas and justicia need not have been fluent in Latin any
more than a modern art critic noting that a painting possesses “a certain élan” need be
fluent in French. Most borrowings are indeed nouns such as those in the carol—a fact
Although these isolated Latin words are easy to label as borrowings, entire Latin
phrases quoted in numerous other carols are harder to classify. For example, the
phrase “a solis ortu cardine” (at the foremost rising of the sun), from a Christmas
hymn sung at Lauds, is found in five macaronic carols. Greene lists numerous carol
lines similarly quoted from Latin hymns, prosae, and antiphons in tables that, in his
that because the Latin phrases derive from outside sources, these elements are simple
borrowings? Carol writers who were familiar with the phrases could have borrowed
them without understanding them word for word, much as Christmas carolers today
can sing “Gloria in excelsis Deo” and even understand what it means without
knowing that Deo is in the dative case. If we do consider the Latin phrases in the
who knew Latin well enough to speak or write it fluently. Under this assumption,
113
macaronic carols become a collection of commonplaces assembled by writers who
could appreciate the significance of the Latin they were borrowing but could not fully
understand it. How we classify the Latin elements in the carols therefore determines
Consideration of the pragmatic function of these Latin elements and the way in
which they were integrated with English suggests that they are in fact code switches.
these elements, which she calls high-reference phrases, are linguistically marked, but
this markedness does not make them ungrammatical (“Code-Switching and Switching
Code”).3 Instead, the linguistic marking afforded by code switching either signals a
shift in interpretational strategy (as in the case in some medieval legal records) or, in
Switching Code”).
Moreover, although the high-reference phrases in the carols did not originate
with the writer, he or she would require advance knowledge of Latin to integrate them
grammatically with English, as is so frequently the case. For example, the following
3
The inclusion of such high-reference phrases in many cases reverses the process of troping
discussed in Chapter 2. Instead of new text being incorporated into an existing composition, phrases
citing various preexisting texts are incorporated into a new work. Some carols, however, are like tropes
in that they augment liturgical texts with newly composed material. Many of James Ryman’s carols,
114
stanza from a sixteenth-century carol in Balliol 354 contains two Latin lines from
liturgical sources:
[The holy breast of chastity conceived the Son through the Word, brought so before
the Trinity as a lily of chastity.]
According to Greene, “verbo consepit Filium” comes from the same hymn for
Christmas Lauds as does “a solis ortu cardine,” mentioned above, and “castytatis
lyllyum” comes from the prosa for the Mass of the Blessed Virgin Mary (lxxxviii,
xcvii). Consepit in the first Latin phrase is predicated by the English subject brest. In
the second Latin phrase, “castytatis lyllyum” is the object of a preposition (ut) not in
the prosa. The sense of the sentence is not complete without the Latin. To integrate
the Latin so thoroughly with the English, the writer would therefore have to have
brest could take the third-person singular consepit as a verb and that the Latin
preposition ut could take the dative noun phrase found in the prosa. The thorough
integration of the two languages and the writer’s demonstrable knowledge of Latin
grammar precludes these phrases’ being simple loan words.4 Many other carols that
for example, translate and develop traditional hymns and antiphons such as the Magnificat and the Te
Deum and include phrases of untranslated Latin from the sources.
4
Those carols that in part translate preexisting Latin hymns likewise suggest that the writer would
have had to have been fluent in Latin to switch grammatically from the original Latin to the English
translation (or to combine an existing vernacular translation with the Latin text).
115
English and Latin syntax and switch languages in the wide variety of ways detailed in
the next section. Such evidence supports interpreting shifts to and from these high-
Occurring with predictable regularity every second or fourth line, the switches
in the most macaronic carols may seem overdetermined when compared to the much
evaluated according to the Functional Head Constraint (FHC), the code switches
carols, only the sixteen analyzed below violate the FHC. Although this trend bears
witness to carol writers’ bilingual abilities, these bilinguals were also helped along by
the form of the carol itself, whose prosodic conventions actually favor grammatical
code switches.
code switches in that carol lines tend to comprise complete phrasal or clausal
5
Wenzel identifies forty-three medieval English sermons that fully integrate English and Latin
within sentences. The spontaneity of code switches in some of the most macaronic sermons (relative to
the more regular switching in the carols) can be seen in the following passage from a sermon in
Oxford, Bodleian Library, Bodley 649, written in the early fifteenth century (Wenzel 160):
Domini gouernoris most eciam be merciful in punchyng. Oportet ipsos attendere quod of stakis
and stodis qui deberent stare in ista vinea quedam sunt smo†e and li∑tlich wul boo, quedam sunt
so stif and ful of warris quod homo schal to-cleue hom cicius quam planare (Wenzel 274).
116
constituents and languages generally do not switch within lines. A given carol line
Example 16 from Chapter 3, almost every Latin line is a complete adverb clause
Domina”). Since phrases begin and end at the line boundaries and languages do not
generally switch until a new line begins, an ungrammatical code switch within the
phrase is highly unlikely. This principle can be seen at work in the following stanza from
James Ryman’s paraphrase of “Regina celi laetare,” quoted in full at Example 20 in the
previous chapter.
Two lines begin with the coordinating conjunction and functional head and, which in
both cases receives the first weak stress of the iambic line and takes a complement
that fills out the rest of the line (“naylde vppon a tre” in the first case and “them with
[The lord governors must also be merciful in punishing. They should take notice that of the stakes and
supports that should stand in this vineyard, some are smooth and will easily bend, others are so stiff
and so full of obstinacy that a man will split them sooner than straighten them out (Wenzel 275).]
117
him toke he” in the second).6 A code switch between and and its complement would
languages do not change until that important functional boundary is safely past.
phrasal structures that traverse the line break, resulting in a functional head in one
line and a switch before that head’s complement in the next. The third line in this
stanza from a fifteenth-century carol to St. John the Evangelist in Bodleian Eng. poet.
e. 1 begins with and, just as the two lines in the previous example do, but in this case
the head of the conjunction’s complement (the inflection of the second conjoined
clause) does not arrive until after the line has ended and the languages have switched:
[O most noble of all evangelists, request grace from our Maker for us, and serve us
goblets of celestial sweetness.]
The phrase “off sweteness celstyall” obscures where the conjunction’s complement
begins on the most fundamental structural level. Whereas the phrase would normally
occur after the word it modifies, pocula, here has been displaced before the verb,
prebe. As was the case in the Anglo-Saxon macaronic poems, poetic aims here
counter grammaticality; the same phrase that allows the line to fit into the carol’s
meter and rhyme scheme also delays the arrival of the complement’s head until the
final Latin line. In such instances (detailed in the following section), the form of the
6
For a discussion of coordinating conjunctions as functional heads, see Section 1.3.2.3.
118
carol contradicts the grammar of code switching rather than reinforcing it.7
switches more or less likely. A carol type that incorporates less Latin or does not vary
what Latin it has runs less risk of violating the FHC. Carols of Types I, J, K, and L,
on the other hand, increase the risk of ungrammatical code switches because they
include two or more lines of Latin in each stanza, increasing the likelihood that
languages will switch after a functional head. Most of the ungrammatical switches in
A survey of the code switches in over two hundred macaronic carols uncovers
sixteen switches that violate the FHC: one switch after a determiner, one after a
switches occur after other kinds of functional heads, such as quantifiers or negators.
The following subsections describe switches after each kind of functional head.
first addressed in Section 1.3. As the code switch in Example 4 (“And off sweteness
celestyall / Prebe nobis pocula”) demonstrates, where a code switch can be said to
7
The grammatical effects of poetic conventions on code switching are by no means limited to
medieval English carols. Pieter Muysken has noted similar processes at work in modern bilingual
songs that switch between Quechua and Spanish. Some of these songs depart from normal Quechua
word order by placing Spanish direct objects after Quechua verbs. Muysken suggests that these
counter-examples can be explained “through appealing to requirements of rhythmic structure specific
to the genre of bilingual songs” (Bilingual Speech 118).
119
occur depends on the focus one takes. According to the surface linear order, the code
switch occurs between English celestyall and Latin prebe. Since these words are
is a verb—the switch cannot violate the FHC (applied at surface level). Seen in light
conjunction and and the head of its complement, prebe, and is therefore
artifacts of my linguistic assumptions. They would also find among the switches I
treat as grammatical a number that would no doubt violate the constraints they favor.
and evaluate switches according to their overarching linguistic structure. As the above
The carols contain only one switch after a determiner, found in a fifteenth-
century Type I carol from London, British Library, Sloane 2593. The switch occurs in
120
the carol’s first stanza and is identified below.8
As explained in Section 1.3.2.1, code switches after determiners like on (ModE an)
determiner, ave, is linguistically marked not only because it is reported speech, one
step removed from the speaker, but also because it is an interjection used as a noun
phrase. If these marked features lead us to read ave as a borrowing, the true code
switch does not begin until the following line with the relative clause “Que vocatur
Maria,” which although grammatical seems displaced since it modifies mayde rather
than ave. Classified as a code switch, the change of language at ave is indeed
directly as possible and to maintain the rhyme scheme. In either case, this example
contradicts the pattern established in most other carols because the change in
8
As in Chapter 1, I have identified code switches by (1) bracketing the larger constituent in which
they occur, (2) labeling the functional head, (3) enclosing its complement in an additional set of
brackets, and (4) underscoring the head of that complement.
121
4.3.2 After a Complementizer
Although every stanza in that carol ends with the same cauda, “Saluator mundi natus
est,” which repeats the second line of the burden almost verbatim, this final line is
introduced differently each time and thus varies in its grammatical function. In the
second stanza, given below, that repeated clause occurs after the complementizer that,
[To the shepherds keeping their flock on Christmas night, an angel said that in
Bethlehem among brave beasts the Savior of the World was born.]
Unlike the previous example, this one includes adjunct phrases between the
complementizer and the head of its clausal complement, the inflection of the verb natus
est. As explained previously, this pattern also holds for most of the subsequent
phrases that delay the arrival of the complement head until after the code switch.
Two ungrammatical switches in the carols occur after modal auxiliaries and
modal—a situation previously considered for Old English in Section 1.3.2.2. One
switch, from a sixteenth-century Type H carol to St. John the Evangelist in Balliol
122
354, pairs a English modal with a Latin infinitive—the same inflection one would
The other switch occurs in the fifteenth-century Type B carol by John Audelay cited
in the previous chapter at Example 7; in the last stanza, an English modal takes a
One would expect the Latin verb following a modal to be in the infinitive form, as it
is in Example 8; such a choice would fit the meter just as well since dare, like donet,
is a trochee. Instead, the tense and mood of the verb are contradictorily indicated both
by the English future tense indicator wyl and by the present subjunctive inflection of
syntactic grounds. Even in this case, however, the interposition of the adjunct
prepositional phrase affter this outlere and the Latin dative nobis obscures the FHC
123
4.3.4 After Coordinating Conjunctions
Most of the code switches that violate the FHC occur after coordinating
Rather than presenting every such code switch, the following is a representative
Selden B. 26:
[To slay our sin, he took the way into the world from the joy of the Kingdom of
Heaven, and therefore both night and day the earth echoes with joys.]
The same carol contains a similar switch in the previous and the following stanzas;
likewise, a Type K carol by James Ryman contains two switches after and in
neighboring stanzas (Greene no. 75). Only one carol, quoted in Example 4, contains
an isolated switch after and. All of the switches after and interpolate adjuncts
between the conjunction and the verb of its complement clause, filling out the line as
124
A different pattern predominates among the four switches that occur after for,
The burdens of two fifteenth-century Type H carols in Bodleian Eng. poet. e. 1 contain
Here the use of for allows the memorable quote from the Gospel of John to fill out the
iambic tetrameter line but makes the switch ungrammatical. The disjunction does
draw attention to the quote, however, raising the possibility that this violation of the
Unlike switches after for, those after English but and its Latin equivalent sed
material. Switches after both contrast words occur only once apiece in the corpus of
medieval English carols, and both appear in carols from the Early Modern period.
The first switch of this kind occurs in one of James Ryman’s many adaptations of the
closing Vespers hymn Te Deum laudamus. The carol incorporates new Latin material
in the second line of each stanza but repeats “Te Deum verum laudamus” as a cauda,
combining features of Types D and J. In the fourth stanza, the English introduction to
9
Toribio notes a similar practice in modern Spanish-English poetry, in which some ungrammatical
code switches seem to be made intentionally to draw attention to the act of code switching (“Speaking
in Tongues”).
125
(12) By day and by nyght, as it is ryght,
A laude tua non cessamus,
[ConjP Butte [with the tyght of alle oure myght
Te Deum verum laudamus]] (Greene no. 294).
[By day and by night, as it is right, we do not cease from your praise, but with the
guidance of all our strength, we praise you, true God.]
Even though the conjunction butte and the verb laudamus occupy opposite ends of
the clause in linear surface order, the switch is ungrammatical since it occurs between
reproduced below.
This is the only switch after a Latin conjunction in the carols, and it conforms to the
pattern seen throughout the ungrammatical switches by including two modifiers, the
Latin adverb semper and the English adverbial prepositional phrase in sinne, after the
conjunction. The ungrammatical switch in this case is the result of a linguistic and
poetic domino effect: each line comprises a complete clause, but languages switch
within each line; by conjoining the first two clauses, the carol writer practically
126
Since, contrary to this last example, most of the ungrammatical switches in the
carols occur after line breaks, prosodic factors could have made these switches
acceptable to the poets writing them. Modern evidence suggests that speakers do not
note as great a disjunction when a pause occurs after the functional head. For
brief silence between a complementizer in one language and its IP head in another
made the sentence more acceptable (241).10 The aural (and visual) disjunction
between the end of one poetic line and the beginning of another could have marked
such a pause. In carols set to music, such a disjunction would have been even greater
since each poetic line corresponds to a separate musical phrase.11 Such prosodic
medieval English carols, we may now analyze the grammatical switches occurring in
the most macaronic works. As mentioned in the description of various carol types,
code switches in carols of Types F through L are particularly suitable for study
because these carols do not generally contain repeated material. Although some carols
10
MacSwan goes on to conclude that switching between a complementizer and its complement is
therefore not prohibited by any code switching constraint. As Toribio explains, however, MacSwan
must overlook the speakers’ responses in order to reach this conclusion (“Emergence of Code-
switching Competence”).
11
All of the songs in Stevens’ Mediæval Carols, derived from those manuscripts that record carols
with music, follow this pattern.
127
of Types D and E do alter the syntactic function of a repeated cauda by introducing it
in different ways, in most carols the function of a repeated cauda generally does not
vary from stanza to stanza, constituting in essence a single code switch made several
that the switches themselves vary, to avoid artificially inflating the number of
switches under consideration. Since even the most macaronic carols repeat Latin
material, I have tallied code switches before such repetitions only once apiece,
sample into two groups: one comprising carols dated in Greene’s collection up to
1492, the date given in the colophon to James Ryman’s collection, and another
comprising carols dated after 1492 through the sixteenth century. The two groups
thus roughly divide, however arbitrarily, into carols that contain Middle English and
those than contain Early Modern English, conventionally understood to begin around
1500. Since the dates for most of the carol manuscripts have not been pinpointed to a
particular year in the fifteenth century, the first group conservatively comprises all
carols found in manuscripts dated generally to the fifteenth century, including those
dated to the second half, and the second group comprises only those from manuscripts
Tables 4.1 through 4.4 below summarize where code switches occur in carols of
Types F through L. The first two identify switches from English to Latin, the other two
from Latin to English. As with the tables in Chapter 1, the constituent indicated in
128
Column 1 is in the language named first in the title (that is, Middle English in Table
4.1, Early Modern English in Table 4.3, and Latin in Tables 4.2 and 4.3). The number
a percentage both of the total switches treated in that table (Column 3) and of all
intrasentential switches in carols of Types F through L from that period (Column 4).
12
Eighty-two switches in Middle English carols occur between sentences.
13
Fourteen of these switches occur between a verb of speech and its complement.
129
Table 4.2: Intrasentential Switches between Latin and Middle English
Switch Site Number % Latin-ME % All
LEXICAL HEADS
Obligatory Constituents
Subject NP and VP 4 2.2% 1.6%
V and complement 2 1.1% 0.8%
Adjuncts
ADJ 7 3.8% 2.7%
Genitive NP 3 1.6% 1.2%
Dative NP 2 1.1% 0.8%
Ablative NP14 24 10.4% 9.4%
Appositive 19 10.4% 7.5%
Direct Address 20 11.0% 7.8%
Participial phrase 18 10.0% 7.1%
Prepositional phrase 47 25.8% 18.4%
Interjection 20 11.0% 7.8%
Relative pro. and preceding 14 7.7% 5.5%
Subord. conj. and preceding 2 1.1% 0.8%
Total switches 182 100% 71.4%
As the data in Tables 4.1 and 4.2 show, Latin elements are the structural loci of
carols. Most of these Latin elements (69 percent of all intrasentential code switches)
and nouns of direct address; and interjections.15 In other words, most of the Latin
elements describe or otherwise limit Middle English material. The next largest group
of code switches (12.1 percent) occur after obligatory Middle English constituents,
especially between Middle English verbs and their Latin complements; in these
instances, Latin and Middle English are much more closely intertwined, since
elements in both languages are necessary for the completeness of each sentence.
14
Five of these switches occur before an ablative absolute.
15
A very fine line separates switches before or after interjections from those involving nouns of
direct address since many Latin direct addresses are preceded by an exclamation of greeting or by an
invocation such as O or Ave.
130
Switches before Middle English adjuncts are a fraction of their Latin counterparts:
only 9.1 percent of switches involve Middle English elements that modify Latin
material. A still smaller group (7.4 percent) of switches entail Middle English
functional heads. Unlike the ungrammatical switches discussed in Section 4.3, almost
half of these switches are grammatical since they occur before the functional heads in
question. The smallest group of code switches (2.4 percent) occur after obligatory
Latin constituents and tend to fall between subjects and verbs rather than between
verbs and their complements— opposite the trend for switches after obligatory
Table 4.3: Intrasentential Switches between Early Modern English and Latin16
Switch Site Number % EModE-L % All
LEXICAL HEADS
Obligatory Constituents
Subject NP and VP 4 3.5% 1.6%
V and complement17 26 23.0% 10.5%
Preposition and complement 2 1.8% 0.8%
Adjuncts
ADJ 5 4.4% 2.0%
Appositive 1 0.9% 0.4%
Direct Address 10 8.8% 4.0%
Participial phrase 1 0.9% 0.4%
Prepositional phrase 26 23.0% 10.5%
ADV 2 1.8% 0.8%
Adjunct infinitive phrase 3 2.7% 1.2%
Nominal absolute 1 0.9% 0.4%
Interjection 1 0.9% 0.4%
Subord. conj. and preceding 15 13.3% 6.1%
Relative pro. and preceding 5 4.4% 2.0%
FUNCTIONAL HEADS
*Mod and V 1 0.9% 0.4%
Coord. conj and preceding 7 6.2% 2.8%
*Coord. conj and complement 3 2.7% 1.2%
Total switches 113 100% 45.5%
16
Thirty-six switches in Early Modern English carols occur between sentences.
17
Twenty-two of these switches occur between a verb of speech and its complement.
131
Table 4.4: Intrasentential Switches between Latin and Early Modern English
Switch Site Number % L-EModE % All
LEXICAL HEADS
Obligatory Constituents
Subject NP and VP 2 1.5% 0.8%
V and complement 2 1.5% 0.8%
Preposition and complement 2 1.5% 0.8%
Adjuncts
ADJ 3 2.2% 1.2%
Genitive NP 1 0.7% 0.4%
Dative NP 1 0.7% 0.4%
Ablative NP18 9 6.7% 3.6%
Appositive 15 11.2% 6.1%
Direct Address 14 10.4% 5.7%
Participial phrase 7 5.2% 2.8%
Prepositional phrase 24 17.9% 9.7%
Adjunct infinitive phrase 3 2.2% 1.2%
Nominative absolute 1 0.7% 0.4%
Interjection 5 3.7% 2.0%
Relative pro. and preceding 20 14.9% 8.1%
Subord. conj. and preceding 14 10.4% 5.7%
FUNCTIONAL HEADS
Coordinating conj. and preceding 10 7.5% 4.0%
*Coord. conj. and complement 1 0.7% 0.4%
Total Latin switches 134 100% 54.1%
Tables 4.3 and 4.4 reveal that code switching in Early Modern English carols
differs markedly from that in carols from the first nine decades of the fifteenth
century. Perhaps most noticeably, code switches entailing English elements and those
before Latin adjuncts are still the largest group of switches (47.3 percent), they are no
longer a majority. The next largest group of switches now occur before English
adjuncts—at 28.2 percent a threefold increase over the earlier fifteenth-century data.
As was the case in those carols from the Middle English period, switches after
obligatory English constituents still comprise about twelve percent of the total
132
intrasentential switches (12.9 percent) and far outnumber switches after obligatory
Latin constituents (2.4 percent). Unlike the earlier carols, however, in which no
switches involved Latin functional heads, the Early Modern carols include as many
involving Latin functional heads as they do English ones (4.4 percent in both cases).
Coordinating conjunctions are the sites of all Latin switches either before or after
functional heads. While almost all of these Latin switches occur before the
The marked differences between code switching in the early and late carols is
all the more remarkable given the arbitrary grouping of both samples. Since the data
for Tables 4.1 and 4.2 include any fifteenth century carol that could not be pinpointed
to the last decade of the century, the “Middle English” sample could in fact contain a
number of carols that incorporate Early Modern English. By the same token, Late
Middle English did not become Early Modern overnight; conservative speakers
writing in the early sixteenth century could have retained linguistic features most
individuals no longer employed. Despite the ambiguities created both by the inexact
science of manuscript dating and the fluid nature of linguistic change, however, the
two groups of carols display key differences. Tables 4.5 and 4.6 juxtapose the data
from the previous two tables, bringing these differences further to light.
18
Six of these switches occur before an ablative absolute.
133
Table 4.5: Comparison of Switches Between English and Latin as a Percentage
of Total Intrasentential Switches
A survey of the largest differences between switch sites in the two groups of carols
reveals an important trend. Fewer switches in carols containing Early Modern English
occur before Latin prepositional, participial, and ablative phrases. At the same time,
more switches occur before English prepositions, Latin and English subordinating
conjunctions, and Latin coordinating conjunctions. Carols written during the Early
Modern English period thus depend less on Latin adjuncts than their Middle English
predecessors do. Rather than limiting and relating ideas in English and Latin by using
Latin modifying phrases, the Early Modern carols rely increasingly on coordination
and subordination, which bring together entire clauses. As the next section will
is the result of linguistic changes that were at once stylistic and syntactic. During the
fifteenth and sixteenth centuries, English writers emulating Latin and French models
developed a style that made more extensive use of subordination and coordination
than the spoken language did. These stylistic developments affected written English
135
syntax as they encouraged the use of conjunctions, classes of words instrumental in
creating complex syntactic structures. Code switching in the carols was directly
affected in the process: not only did these linguistic developments alter the patterns of
code switching; the alternation between languages also foregrounded the newly
developed style.
Throughout the Late Middle English Period, writers made increasing use of
Fisher, this stylistic movement developed between 1420 and 1440 at Chancery, the
center of the English royal bureaucracy throughout the fifteenth century. Fisher
argues that writers at Chancery not only worked to standardize English orthography
style, under the influence of the Latin and French documents they continually copied
(880–87). By 1460, the style had been adopted by professional scribes throughout
England (Fisher 896).19 By the beginning of the sixteenth century, this hypotactic
lawyers, and later educators began to emulate it in their official and private
circles, the later macaronic carols therefore register more code switches before
English subordinating conjunctions like as, because, and that (used to express
19
Here and throughout I cite Fisher’s original 1976 article, which is included with only minor
alterations in his 1997 essay collection, The Emergence of Standard English (36–64).
136
purpose).20 Since the clerical writers of macaronic carols would have read (and very
well may have produced) such prose, it is not unreasonable to suggest that they
adopted the same style in their poetic works as well.21 Even if bilingual carol writers
did not consciously mimic Chancery in their poetic works, they could have easily
made the same choice as Chancery writers by adopting the hypotactic style of the
Latin prose works with which they were conversant. Indeed, the carols themselves
illustrate the combination of Latinate and vernacular influences that shaped English
The discussion so far has been limited to the increased use of subordination
from the late fifteenth century onward. The use of explicit coordination, however,
was also a hallmark of the written standard developed just before and during the Early
Modern period. As Laura Wright notes in her brief study of the syntax of sixteenth-
century witness narratives, Early Modern English speech was like spoken registers of
Present Day English in that it did not organize utterances according to sentences,
made little or no distinction between coordination and subordination, and in fact did
20
Examples of such switches include “As he ys Lord both day and nyght, Venter puelle baiulat”
(Greene no. 52), “Because he sayde he saw a syghte, Lapidaverunt Stephanum” (Greene no. 9), and
“Of Criste that we may haue a sight Funde preces ad Filium” (Greene no. 201). These switches are
from Balliol 354 (1508–c.1536), Huntington HM 147 (c. 1500), and Cambridge Ee.1.12 (c. 1492),
respectively. Switches before adverbial that occur only in James Ryman’s works.
21
Many records similar to those kept at Chancery were in fact macaronic, for example the Latin-
English accounting records Laura Wright studies in “Macaronic Writing in a London Archive.”
22
In time the English language would acquire a more diversified vocabulary of subordinating
conjunctions. Although the carols themselves do not demonstrate this trend, by the seventeenth century
English subordinators became more specialized in their functions. Causal subordinators diversified as
Old English forms incorporating the conjunction for gave way to more specialized markers (see
Traugott and Rissanen). Temporal subordinators such as after likewise became distinguished from
non-temporal “prepositions” such as unless in the kinds of complements they could take, to the point
that, as Dubinsky and Williams argue, temporal subordinators became complementizers.
137
not significantly distinguish between coordinated and noncoordinated sentences (the
1589, divided utterances into sentences and clearly coordinated and subordinated
ideas (Wright, “Syntactic Structure” 104).23 Such evidence upholds the notion that the
enabled the further standardization of written English but also facilitated switching
between English and Latin. Patterning English written usage on that of Latin made
the syntaxes of the two languages more alike on a local level. Doing so in turn
enabled bilinguals to code switch at the syntactic sites English and Latin now shared
Such newly licensed sites included the explicitly marked subordinating and
23
Wright comes to these conclusions by applying David Crystal’s observations of present-day
conversational speech to sixteenth-century court transcripts. Among his findings, Crystal notes that
modern speech (1) is not organized into discrete sentences, (2) does not meaningfully distinguish
between the conjunctions and and but, and (3) is comprised either of multiple simple utterances
beginning with connecting expressions or of long, complex utterances held together by loose
coordination. Wright finds the same features in a number of witness accounts from sixteenth-century
courts, which though written approximate the spoken register. The doctor’s testimony, however, bears
all the hallmarks of modern written style (Wright, “Syntactic Structure” 104).
24
See p. 23 for a discussion of the Equivalence Constraint, which Belazi et al. incorporate as a
corollary to the Functional Head Constraint.
138
The use of code switching to highlight aspects of written style reminds us that
code switching itself can be employed stylistically—a fact underscored by the various
patterns identified in the previous chapter. What better way to mark distinguishing
features of the written idiom than by switching languages before those features (cf.
Traugott and Romaine 26–28)? Comparing code switching in Middle English and
Early Modern carols thus not only exposes a series of important developments in
English writing, but also underscores the stylistic utility of code switching among
139
Chapter 5: Macaronic Carols in a Community of Practice
The previous two chapters have discussed bilingual carols from medieval
England primarily as linguistic and literary artifacts. Carols were not performed and
read in isolation, however, and thus merit study both as expressions of social groups
and as a means of constituting them. Although these works have received little
scholarly attention in recent years, their broad social, geographical, and chronological
distribution make them more representative of medieval culture than many of the
I will now return to the remarkably varied array of religious and lay users
identified in Chapter 3 to detail how they could form bilingual identities in part by
writing, performing, and disseminating macaronic carols. I will hereafter refer to this
set of interrelated activities as “caroling” and the people who engaged in these
activities as “carolers.” To describe this network of carolers and the ways in which
macaronic carols allowed bilinguals to give their linguistic identities concrete form, I
will employ the model set forth by Etienne Wenger in Communities of Practice:
those of anthropologists Jean Lave and Julian Orr, sociologists Pierre Bourdieu and
140
Jürgen Habermas, and psychologist Lev Vygotsky, as well as poststructuralist critics
processors, and illustrates concepts throughout his book with examples from that
recreational. And although Wenger generally applies his model to modern situations,
theories of social structure, situated experience, social practice, and identity (11–15).
Within this theory, he defines practice as “doing in a historical and social context that
gives structure and meaning to what we do” (47). Negotiating meaning through
always involves a social component), they generate artifacts that reify the practice
itself as well as the values and assumptions that motivate it (54–62). The medieval
communities that produced the texts treated throughout this study offer numerous
examples of these dual dynamics. Monastic orders, for example, were founded quite
141
explicitly to ensure the continuance of various practices—not only of prayer but of
lifestyle, work, and education. To reify such practices, monastic founders produced
rules stipulating how each practice should be carried out, and in the process expressed
transmit the practice to others; in fact, we as modern scholars can infer past practices
only through such reifications: the manuscripts and other artifacts left by medieval
people. Additional features of Wenger’s theory will become relevant as I address how
Macaronic and monolingual carols were performed in similar social contexts and
often recorded in the same manuscripts. Although three of every five sources Greene
cites contain only monolingual carols, very few of the remaining two-fifths contain
only macaronic ones, and the largest collections include monolingual works alongside
those that alternate languages.1 Moreover, other than in their linguistic alternation,
macaronic carols did not differ from monolingual works in their poetic forms or in their
The two can be distinguished from each other, however, on the level of
reification. That is, although macaronic carols were recorded and no doubt performed
identity. After fleshing out caroling in general as a practice, I will identify the aspects
1
See the characterization of carol manuscripts on p. 85, n. 12.
142
of that practice that enabled the formation of a community, drawing examples
whenever possible from macaronic texts. I will then focus on the community involved
depended on fluent bilinguals to generate carol texts, I will argue that even those who
Because this community was constituted by practice rather than predetermined social
markers, it could continue even after the Reformation radically altered its
membership. I will then explain how macaronic carols reified a positive conception of
Throughout its development from a dance song with a single melody line to a
complex multi-voice composition—and, I will argue, even in cases where carol texts
were written to be read, not sung—the carol was a quintessentially public form,
requiring or at least implying the engagement of a group. The social nature of the
carol developed from its origins as a ring dance sung alternately by the dancers
themselves (on the burden) and a soloist or small group outside the circle (on the
stanzas) (Bukofzer 119). Besides the carol texts themselves, scenes of caroling in
contemporary historical accounts and chivalric romances and entries for carol-related
expenses in monastic business records demonstrate that carols were very much a
From the texts of both macaronic and monolingual carols we see that the carol
was often part of group celebrations and feasts. That carols were sung or recited in
143
groups is clear from the first-person plural viewpoint of many carols, expressed in
both Latin and English—for example, the burden from a Christmas carol in the Ritson
Manuscript reads:
One of James Ryman’s carols likewise comes from the assembly, who bid farewell to
Advent both in the burden and at the end of each stanza with “Farewele fro vs both
alle and sume” (Greene no. 3). The celebratory nature of occasions for caroling is
evident not only from the feast days for which carols were composed, but also from
references to conviviality in the carols themselves. For instance, the Boar’s Head
Carol, presented in Chapter 3 at Example 14, is clearly sung from the standpoint of a
lone singer bringing the dish into a hall, inviting the participation of those “qui estis
in conuiuio” (who are at the banquet). Some carols even set the rules for such group
Descriptions of public ceremonies held over a century before most carols were
recorded also indicate the social character of carol performance. For example, the
144
Annales Londoniensis describe townspeople attending the coronation of Edward II in
1308 as “coram rege et regina karolantes” (caroling [that is, both singing and
dancing] before the king and queen) (Withington 1: 125). The same annals report that
five years later London fishmongers staged a pageant to celebrate the birth of Edward
III and were likewise “coram regina karolantes” (caroling before the queen)
(Withington 1: 126). These early examples show that the events at which carols were
sung and disseminated brought together people of various occupations and social
Even after carols became too musically complex for dancing, they remained
122).2 The Tudor scholar and playwright Nicholas Udall, in his 1542 translation of
Erasmus’s Apophthegmes, mentions “syngyng merie songes and rymes for makyng
laughter and sporte at marryages, euen like as is nowe vsed to syng songes of the
Frere and the Nunne, with other sembleable merie iestes, at weddynges, and other
feastynges” (qtd. in Reed 78). Two surviving “songes of the Frere and the Nunne,”
which portray a Franciscan in flagrante delicto with a sister religious, are carols. Both
songs come from sixteenth-century sources and have the same burden: “Inducas,
2
Bukofzer cites formal correspondences between the carol and the processional hymn to support the
use of carols in processions. The fact that he makes this observation right after citing rubrics in the
Ritson Manuscript (London, British Library, Addit. 5665) suggests that his conjecture might also be
based on that manuscript’s inclusion of Latin pieces derived from the Sarum processional. Greene
145
Lord’s Prayer (Greene nos. 461 and 461.1). Udall’s description presents carols as
Venus: “Festes, instrumentz, caroles, daunces, / Lust and array, and alle the
circumstaunces / Of love” (Robinson ll. 1931–33). In Sir Gawain and the Green
Knight, the Christmastide feasting at Bercilack’s court also includes carols, here
[Much splendor and joy shone forth therein, around the fire upon the floor, and in many
ways at the supper and afterwards, many noble songs such as motets for Christmas and
new carols, with all the befitting happiness of which one may tell.]
Such depictions of caroling were still stock features of the romance repertoire a
century later, to the extent that in the prose romance Merlin (recorded in the middle of
the fifteenth century) sung and danced carols are mentioned off-handedly as part of
notes, however, that “the MS. as a whole is not made up of processional music,” undermining
Bukofzer’s assertion somewhat (308).
3
Greene likewise notes this passage (xxviii).
146
various celebrations and feasts. For example, when entertaining Kings Ban and Bors,
Arthur and the kings “yede to here euesonge, and than thei yede to soper, and after
begonne the carolles and to speke of the turnemente” (went to hear evensong, and
then they went to supper, and afterwards began the carols and to speak of the
tournament) (Wheatley 138). These examples from the poetry show royalty and
Although the convivial settings evoked by the carols themselves could in many
cases apply to celebrations hosted by either religious or lay persons, the historical and
literary descriptions cited so far portray predominantly lay carolers. Records from
relates that the accounts kept by William More, prior of the Benedictine chapter at
Worcester cathedral until its dissolution, list “among the most frequent items of
expense in the years 1518 to 1532 … malmsey and other wines, minstrels and other
entertainment, and singers of carols” (xl). More’s accounts also suggest that the priory
subsequently recorded such carols in manuscripts like Bodleian Arch. Selden B. 26,
carols. For Christmas 1518, More notes not only a sum paid to “syngars of carralls at
cristmas day at ny∑th” but successive payments to “Richard skryvenar for wryting …
147
The fact that a religious house like the Worcester priory hosted banquets at
which carols were sung does not mean that such celebrations were limited to those in
clerical orders. Even celebrations hosted by monasteries and cathedral chapters would
have included lay guests and servants. Major religious houses, after all, were
house may have contained no more than a few dozen core members (Prior More’s
Worcester, for example, held between forty and forty-five monks), that core group
notaries, unattached and would-be chantry priests, scribes, school- and choir-masters
as well as other clerks and literati who formed the clerical underworld of the late
medieval cathedral” (Greatrex 120; Dobson 27). Maintaining those clerics in turn was
these supporters in turn obtained goods and services for the community from an array
of tradespeople and provisioners.4 Add to their numbers those who rented lands held
by the religious house, and those in turn who benefited from the house’s spiritual and
religious foundation’s daily life becomes even larger and more diverse. The members
of such large and disparate communities came closest together during holiday
4
Here I follow R. B. Dobson’s characterization of northern English cathedral chapters. To take one
of the largest examples, Dobson determines that the Benedictine priory at Durham housed
approximately seventy monks (23). Secular clergy there were numerous enough that in the early
fifteenth century Prior John Wessington “could regularly assemble a dozen or more … clerks to
witness his convent’s official acta and judicial appeals” (Dobson 27). As for its lay constituency, the
priory “normally employed a labour-force well over twice its own size,” and “their absolute reliance
148
celebrations and the numerous transactions necessary to prepare for them—the public
occasions during which carols were sung, heard, and passed along.5
Caroling practice as described so far involves only texts set to music, sung at
first to accompany dancing and later by themselves. Before we consider how such
works enabled the formation of a community of practice, there remains the question
of whether poems in carol form were likewise intended for public reception. Even if
one accepts the compelling arguments advanced by M. T. Clanchy that reading in the
Middle Ages was primarily an aural and performative act, poems read alone need not
have been construed as public (267–71, 284–85). The carol was unlike poetic forms
developed entirely in the written medium, however, because of the social nature
imparted in its early days as a dance and retained in its most distinguishing feature:
the burden. This repeating chorus continued to serve a useful purpose for two
centuries after the carol ceased to be a dance form, namely as the vehicle for a public
voice that placed the rest of the poem in a larger textual and social context.
upon an elaborate network of local and not so local provisioning facilities is manifest in every
surviving … account roll” (Dobson 29).
5
The complex nature of such religious communities makes it unwise to apply too broadly Roger
Bowers’s characterization of the carol repertoire in Cambridge, Trinity College, O.3.58:
It seems most likely that [the carols] were conceived and sung as home-made entertainment
… round the firesides of the lodgings and refectories of the vicars choral and singing-men of
the greater secular churches and aristocratic household chapels. Despite their vernacular texts
(albeit commonly interspersed with phrases in Latin), it must be emphasized that … these
carols were by no means the music of the people, but exclusively of the educated and
sophisticated elite (88).
In the popular music history text Antiquity and the Middle Ages, Peter M. Lefferts presents Bowers’s
conjecture out of context as a characterization of all carols (187)—an inaccurate portrayal not only
because it excludes most lay carolers but also because most carol manuscripts come from monastic
communities, not secular churches or aristocratic houses (cf. Greene’s response to a similar statement
by Frank Ll. Harrison regarding London, British Library, Egerton 3307 [300]).
149
As Lori Ann Garner argues in her comparison of the English carol with
continental refrain forms, burdens served a much more important function than
simply signaling when the group of dancers could join in the action (469–70).
Because the same burdens could appear in more than one carol and could state
general themes outside the particular focus of the stanzas, they linked the carol to
other texts and genres. In the process, the burden set up a larger interpretive context.
For example, a burden could present a traditional proverb explaining the moral of the
stanzas, or a burden normally associated with festive songs could be used to predict
the happy outcome of a story even before it had run its course (as is the case with an
amorous pastourelle Greene numbers 463.1 in his collection) (Garner 478–80). This
contextualizing function proved useful enough to warrant the persistence of the carol
form outside its original musical setting so that “the burdens index[ed] the way in
which the songs as a whole [were] to be received not only by the immediate
audiences in the context of the dance, but also by reading audiences who continued to
are designed to impart—is inherently public, requiring the reader to relate the text at
hand to other texts and thus to place him- or herself within a greater textual
sphere for the genre even when they were read in private as poems. Just as the first-
person plural pronouns of so many carol texts refer deictically to a larger group of
participants beyond the individual singer, the burden likewise indicates a plural
150
perspective. In this sense, it helps achieve what Benedict Anderson terms unisonance,
suggests—above all in the form of poetry and songs” (145).6 Even if solitary, the
medieval reader of the burden could experience the unanimity and solidarity of sung
and danced burdens, those moments of the carol requiring group participation.
carols could be passed along orally after people had memorized what they had heard
and how carols, once performed, could be written down for later performance. Since
both books and brains are portable, the transmission of carols was limited only by the
Richard Hill’s commonplace book (Balliol 354), contains the texts of nineteen carols
found more or less verbatim in other manuscripts and another sixteen that appear with
more significant variation elsewhere (based on a tally of the contents listed in Greene
320). One macaronic text in this collection, a Type D carol to St. Thomas of
Canterbury with the burden “A, a, a, a, / Nunc gaudet ecclesia” (Ah! Now the Church
rejoices) appears in three other manuscripts: Sloane 2593; Gonville and Caius
College 383; and the two-leaf fragment London, Public Record Office, Chancery
Miscellanea, Bundle 34, File 1, No. 12 (Greene 60–61, no. 114). Taken together,
6
I use Anderson’s term here only for its usefulness in explaining how a group of people, however
separate, can feel united through “imagined sound” (Anderson 145). I do not mean to suggest,
however, that carols were a means of nation-building, in keeping with Anderson’s thesis, nor do I
share his suspicion of those who participate in unisonance.
151
these four sources demonstrate the transmission of this carol among both clerics and
lay people of both middle and upper classes, including a London grocer (Balliol 354),
383), and perhaps an aristocrat (the Public Records Office fragment) (Greene 320,
Merchants like Hill who supplied the religious houses that produced carol
Library, Harley 4294 and 5396) likewise bear indications that they were owned and
used by merchants (Greene 303, 320–21). Although Hill’s book, unlike the Harley
manuscripts, does not include business accounts, it does contain ample evidence of
his business travels, including a list of dates and locations of English fairs,
and taxes for French as well as English locales (Dyboski xxxi–xxxii; Parker,
Commonplace Book 50, 67). Fairs would have been prime opportunities for
7
This last attribution is admittedly based on indirect evidence. In addition to the carol, the Public
Records Office fragment also contains a twelve-line macaronic poem beginning “Pryde pryde wo thou
be mater viciorum” (Pride, pride, woe, you are the mother of vices). The same poem also appears on a
velum strip from the late fifteenth century once kept with the papers of the earls (later dukes and
marquesses) of Ormonde, who resided at Kilkenny Castle, Ireland, from the twelfth until the twentieth
century (Greene 313; Seymour 207). Assuming that the strip was added to the Ormonde collection
around the time it was written, the Public Records Office fragment may have likewise come from a
manuscript owned by aristocrats. When Greene prepared the 1977 revision of his edition, he was
apparently unaware that the Ormonde manuscripts had been donated to the National Library of Ireland
and the Bodleian in the 1940’s. He instead listed the strip’s location as Kilkenny Castle based on
Carleton Brown and Rossell Hope Robbins’s 1943 Index of Middle English Verse (313). Repeated
searches through the National Library of Ireland’s Manuscript Sources for the History of Irish
Civilisation have failed to locate the strip’s present location. My thanks to Dolores Gaffney and Izabel
152
merchants and customers alike to hear and exchange carols, since fairs were often
held around the religious feastdays on which carols were sung (for example, Hill
notes a fair “in the wedynsday in esterwek [Easter week] [at] Huchyn” [Dyboski
xxiv; Parker, Commonplace Book 50]). Moreover, even grocers were in the business
of selling books, and books of carols may have well been part of their wares (Thrupp
The shape and size of many carol manuscripts would have also facilitated their
being carried from place to place, especially by people on business. Hill’s manuscript
commonly used for ledgers and attested in other commonplace books (for example,
Cambridge, Trinity College O.9.38) (Greene 320; Parker, Commonplace Book 38;
Rigg 1). Harley 4396 is even smaller at eight-and-a-half by five inches, and
by four-and-one-eighth inches that “looks as if it had been folded into half its size”
(Greene 303; James and Macaulay 68). This last manuscript also sports an original
vellum cover long enough to wrap around it entirely, a feature that protected the book
The various aspects of caroling practice, from the public nature of carol
varied group of people to participate in it. The next section will further explain this
Pennec at Kilkenny Castle for answering my queries concerning the current whereabouts of the
Ormonde manuscripts.
153
practice in Wenger’s terms by demonstrating how it helped form a community. In the
identity based upon linguistic competence. For bilinguals and those who identified
(72–85). To take each of these dimensions in turn, practice first requires participants
not only to interact with each other but also to actively negotiate the meaning of what
they are doing. Since the community is defined primarily by the practice in which it
engages, its members need not have anything in common besides the practice itself;
among participants, practices rely on the environmental and situational factors that
The various factors that made caroling a public group activity ensured mutual
other people in the moment of performance, even in cases where carols were read and
not sung. The social heterogeneity of the individuals who engaged in caroling
154
occupation, kept the practice alive. This heterogeneity was manifested both locally
and globally: locally, in that the feasts and celebrations at which carols were sung
communities and individuals known to have utilized carols differed from each other
in marked ways but shared the practice of composing, recording, reading, and
performing and transmitting the same works, as the previous example of Richard
negotiation and subsequently hold each other accountable to the goals they have
the community, whatever outside pressures may be operating upon it, making the
goals and accountability may seem too strict when applied to the festive setting of
carol performance, but we have already considered carols that set the terms for the
the rest of the assembly. The practice of singing carols on religious feast days
church. Although the church established the dates for religious feasts and defined the
155
accompanying liturgical celebrations, carols were employed in less decorous and
more personal convivial settings. The Latin liturgical content of many macaronic
“reflects a history of mutual engagement” and enables the creation of new meanings
(Wenger 83). The liturgical tropes discussed in Chapter 2 were quite clearly the
and performed over time eventually became the kernels of new tropes. So too with
the later carols: not only were carol repertoires shared by various participants, as
of liturgical texts were adapted to inform both monolingual and macaronic carols. In
both cases, existing bodies of texts recorded current practice but also led to the
8
In this fashion the carols resemble the Anglo-Saxon macaronic poems, which likewise brought
liturgical language into extraliturgical settings. By “indigenous” here I mean to indicate only that the
impetus for these developments originated within the community of carolers, not that the carol was
inherently a popular folk genre. Although Greene argues for “the ultimately popular character” even of
Latin cantilenae (cxvii) and points to supposed pre-Christian origins of the carol genre (cxxxix), the
performance and certainly the recording of English carols and their continental counterparts was
continuously facilitated by Christian clerics (cf. the performance of the ballata as related by Jeffrey,
The Early English Lyric 131–33).
156
notion of textual communities Brian Stock sets forth in The Implications of Literacy
(Wenger himself does not cite Stock, however). A brief consideration of the
similarities between the two kinds of communities will help make Wenger’s theory
more relevant to the Middle Ages, since Stock derives his ideas from the example of
eleventh-century heretical and reformist groups. These groups, Stock argues, arose
with the advent of literate culture in medieval Europe, a development with profound
consequences for social organization (88). Both reformers and heretics constituted
minded founder (Stock 90). This textual focus informed the community’s practice,
In Wenger’s terms, these groups developed a shared repertoire of written texts and
orally transmitted norms, which reified the practices of the dissenters, held participants
accountable, informed further practice, and thus enabled the community to persist.
Moreover, in keeping with Wenger’s notion that practice can unite individuals that hold
little else in common, Stock notes that “eleventh-century dissenters may not have
shared profound doctrinal similarities or common social origins, but they demonstrated
a parallel use of texts, both to structure the internal behavior of the groups’ members
reformers Stock discusses, but they formed a textual community in a similar fashion.
In their case, carols, not revelatory texts, formed the core repertoire of the
repertoire provided a means of solidarity. With Stock’s model in mind, then, I will
forge a shared identity through the performance and transmission of macaronic carols.
could express and construct their linguistic identity through macaronic carols, we
must first consider how many people could participate meaningfully in that
Latin. As with any community of practice, members varied in their competence and
“trajectories” (153–5). Fluent and literate bilinguals charted what Wenger terms
“insider trajectories” within the community by virtue of their ability to generate and
record as well as to understand macaronic texts. Among these insiders we may count
the scribes who created carol manuscripts as well as poets like James Ryman and
trajectories” in the community, but could nevertheless find the carols meaningful by
158
virtue of their festive and liturgical associations (Wenger 154). Attending Mass and
the Divine Office regularly would have familiarized monolingual English speakers
with a wide array of Latin phrases associated with particular feasts and prayers. One
need not have “known” Latin to understand that Christmas was ushered in with the
phrase from John’s gospel “Verbum caro factum est” or that proper thanks to God
was rendered with the phrase “Te Deum laudamus.” And like many choral singers
today, monolinguals could learn to pronounce Latin and even understand the general
meaning of the Latin phrases they sang without a formal education in the language.
5.3.1 Carolers Inside and on the Periphery: The Problem of Lay Literacy
ownership of these manuscripts gives us an idea of that community’s breadth but does
not indicate who besides the compilers of these manuscripts could be considered
competence in Latin, we are forced to conflate the abilities to speak and understand
the language with the abilities to read and write it. In picking through this thorny
issue, I will attempt as much as possible to distinguish these various skills, so often
9
Clanchy, following H. J. Graff, notes that “the automatic coupling of reading with writing and the
close association of literacy with the language one speaks are not universal norms, but products of
159
Since acquisition of Latin literacy was a fundamental part of medieval clerical
education, the problem of determining the size of the late medieval bilingual
community is more acute with respect to the laity. Anecdotal reports of medieval
English lay people learning and speaking Latin are rare because, as M. T. Clanchy,
explains, to medieval writers the categories of literate and laity were mutually
exclusive. The ability to read and write Latin qualified one as clericus, just as the
inability to do so made one laicus, regardless of one’s tonsured state. Before the
were noted for their erudition, not their ability to read or write (Clanchy 226–34).
term litteratus changed with the spread of benefit of clergy in England (233). Before
the fourteenth century, one could establish right to trial in ecclesiastical courts (and
thus escape the death penalty) only by presenting proof of membership in holy orders,
but by the later Middle Ages passing a literacy test was sufficient to establish clerkly
status (Gabel 64–65). The test itself generally involved reading a brief passage from
the Psalter, particularly the first verse of Psalm 51 (Gabel 72–73). 10 By the mid-
fifteenth century, Clanchy argues, increasing numbers of lay people were classified in
court records as literate not because they had necessarily acquired greater proficiency
modern European culture.” By contrast, medieval learners learned to form letters, to sound them out, to
read entire words, and to write them at separate stages in their education (Clanchy 232).
10
Increased use of the Psalter and the Primer for private devotional reading likewise would therefore
have enabled more people—women as well as men—to prove their “clerical” status; see pp. 163–64,
below.
160
in Latin but because the standard for classifying one as clericus had lowered (Clanchy
literacy can lead to an overrepresentation of Latin literacy rates among medieval lay
people. In her influential study The Merchant Class of Medieval London, for
example, Sylvia Thrupp takes the records of medieval ecclesiastical courts at their
ecclesiastical court between 1467 and 1476 were registered as literate and by
comparing that figure with tallies from “two much smaller groups of witnesses in
1373 and 1466,” Thrupp estimates a forty-percent Latin literacy rate for male London
however, that “the standards [of Latin literacy] for witnesses may not have been so
high as those set for allowance of benefit of clergy” and presumes “that in most cases
[being registered as literate in ecclesiastical court] meant the ability to read a little
Latin” (156–57). The limited sample size and the specialized nature of the documents,
Even a lower Latin literacy rate among London merchants is of interest, however, since
it indicates a broader knowledge of Latin among the lay population than modern
11
This fact does not seem to hinder recent scholars from unquestioningly adopting Thrupp’s figures,
however. David R. Parker, for example, quotes Thrupp in his 1998 study, The Commonplace Book in
Tudor London.
161
scholars conventionally admit and suggests that the merchants who kept manuscripts
containing macaronic carols were not isolated in their familiarity with Latin.
Thrupp’s figures make no account for Latin literacy rates among women, which,
if they followed the trends seen in vernacular literacy during the same period, were
English women were illiterate during the reign of Henry VII—a figure that dropped
to ninety percent by the time of the English civil war (176). Cressy’s findings,
however, are based on the numbers of signatories in wills and other legal documents
who could write their names. As Clanchy observes, the ability of medieval people to
write or to sign their names did not necessarily correlate with their ability to read
since students acquired these skills at different times (232). The continuation of this
Spufford’s work on literacy in the early modern period.12 Since English schools
became open to girls and women in 1400, fifteenth- and sixteenth-century women
could have followed the trend of many of their male counterparts, attending school
long enough to learn to read the vernacular but leaving before learning to write
Those women who did not attend school could have nevertheless gained the
ability to read the vernacular and to pronounce Latin through their use of devotional
12
Spufford explains that throughout the seventeenth century, children were taught to read English
before the age of seven, but did not learn to write until later. Since many working-class children
entered the labor force when they were seven, their ability to read is therefore not reflected in their
162
texts, namely the Primer and the Psalter. As Marjorie Curry Woods explains, while
the first kind of prayer book was commonly owned by lay women, the latter was
often used by women in holy orders (182–84, 186). Latin versions of these texts were
the most common in fifteenth-century England, and pious laywomen and religious alike
recited or sang from these books regularly. This constant repetition and reinforcement,
Woods argues, made it possible for these women to read the vernacular:
The same held true, of course, for men who had learned to pronounce letters in either
Latin or the vernacular: even if they did not understand Latin, they could pronounce
With the foregoing evidence in mind, then, we may assume that the core of the
bilingual caroling community was not limited to clergy and that it could have
fewer than Thrupp’s forty percent. While few women could have occupied insider
trajectories within the community, more could have taken peripheral trajectories than
ability to write (408–15). See Houston 116–29 for a discussion of the problems of determining literacy
rates throughout early modern Europe.
13
Transferal of pronunciation skills gained in one language to the other would also have been aided
by the phonological similarities between medieval English Latin and Middle or Early Modern English.
As was the case throughout Europe, Latin was generally pronounced after the fashion of the
vernacular—a point modern scholars tend to overlook. See McGee et al. for a detailed consideration of
regional varieties of Latin throughout Europe and a reconstruction of medieval English pronunciation
of Latin (46–64).
163
Cressy’s dismally low vernacular literacy rates suggest. Moreover, people who were
able to read the vernacular could also pronounce Latin, allowing them to participate
in bilingual carol performance. Even those who could not understand Latin, therefore,
present themselves as bilingual. Although the identity thus constructed did not
personal meaning from bilingual carols and to consider themselves part of the
community that valued such works. I will hereafter use bilingual to include members
part of a community of practice helps us explain why these bilingual works continued
to be popular even after the English Reformation radically altered its core membership.
As explained in Chapter 3, the latest carol sources contained in Greene’s collection are
from printed collections published in the first half of the sixteenth century. Two of the
three print sources that contain macaronic carols were printed during and after some of
the most radical reforms, Richard Kele’s Christmas carolles newely Inprynted (in the
Huntington Library) between 1542 and 1546, and the similarly titled Christmas
carolles newely Imprinted (Oxford, Bodleian Library, Douce Fragments f. 48) around
164
1550 (Reed liv; Greene 339–41).14 By the time these collections were published,
religious houses and monasteries had been dissolved (throughout the 1530’s), and
rituals and entire feast days had been abolished (starting in 1541). Around the time the
latter collection was printed, chantries and religious guilds were prohibited (in 1547),
and the Latin Mass was replaced with a eucharistic liturgy entirely in English (in
1549).15 The printed collections thus made macaronic carols available after the feasts
on which carols were sung had been curtailed, Latin had begun to fall out of favor as a
The printed medium of these collections suggest that a lay audience who valued
bilingual carols persisted amid these sweeping changes.17 The potential size of that
lay audience is increased by virtue of the black-letter type employed for carol
collections, a matter that returns us to the earlier consideration of lay literacy rates.
argues, sixteenth-century lay people could have read black-letter print without being
14
See Section 3.2.1 for a summary of print sources that contain macaronic carols. Despite their titles,
these publications also contain carols on themes other than the Christmas season. Kele’s collection and
the fragments bound with it also feature carols on Christ’s passion and humorous songs, including one
of the “songes of the Frere and the Nunne” discussed on pp. 146–47 (Greene 340). The Douce
fragment also contains a carol on the Passion amid other texts for Advent (Greene 339).
15
Marsh 18–21 offers a succinct summary of early reform measures. A more thorough-going
account can be found in Duffy, esp. 377–565. Skeeters offers a detailed explanation of how Bristol
responded to the loss of clergy.
16
Marsh relates that during the same period, Latin fell from grace not only within the liturgy but also
in printed texts, so that eventually “the use of Latin all-but disappeared” in “the more accessible forms
of print” (143).
165
able to read roman type or handwritten script. The first texts children read in early
letter. Readers used to this “type of the common people” therefore had difficulty
deciphering the roman type we find so legible today. “Black-letter literacy,” Thomas
explains, “in short, was a more basic skill than roman-type literacy; and it did not
follow that the reader fluent in one was equally at home in the other” (99). As in the
medieval period, students learned to write after they had learned to read black-letter
texts, and the hands in which they wrote required advanced technical facility gained
could have been able to read printed sixteenth-century carol collections without being
able to read other types and scripts, to write, or even to sign their names.
The establishment of black-print literacy among lay people and the coincidental
bilingual caroling community. The clergy who had formerly occupied the core of the
responsibility for maintaining the practice, and lay printers to an extent took on the
This transition was no doubt eased by a substantial appreciation for bilingual carols
among the laity before reformist measures were enacted. Equally vital to the
17
See Marsh 138–146 and Watt (throughout) on the other kinds of printed texts read by lay people
during the Reformation.
166
continuation of the practice were the qualities that had enabled a community to
coalesce around it in the first place: the mutual engagement, joint enterprise, and
shared repertoire that had maintained the community for two centuries and that
carolers could not continue amid such drastic alterations without renegotiating the
values their practice embodied. Continuing to sing songs celebrating feasts and
doctrines that had become officially circumspect and to incorporate liturgical material
in a language no longer approved for public worship brought the practice up against
argued regarding the Corpus Christi carol in Balliol 354 (Commonplace Book 64;
“The Act of Supremacy”).18 But the appearance of bilingual carols in print suggests
much more public promotion of Catholic doctrine and practice. Whether we ascribe
bilingual carols in print remind us that the Reformation was by no means a quick or a
monolithic process.19
18
The carol (Greene no. 322 A) begins with the burden “Lulley, lulley; lully, lulley; / The fawcon
hath born my mak away” and in successive stanzas reveals a wounded knight bleeding upon a bed,
next to which an inscribed stone reads “Corpus Christi.” Parkes explains how the imagery of the carol
very specifically evokes the decorations of a Lenten Mass, arguing that the carol is a lament over the
revocation of the Mass and Henry’s Act of Supremacy (“The Act of Supremacy” 9). Parkes reasons
that Richard Hill’s “Catholic sympathies,” expressed elsewhere in Balliol 354, motivated his inclusion
of the carol in his commonplace book (The Commonplace Book 64).
167
Although bilingual caroling persisted during the initial decades of the
expressed fell out of favor. Among English Catholics the carol form persisted into the
Bodleian Library, Eng. poet. b. 5, compiled by the recusant Thomas Fairfax between
1651 and 1657 (37). Ultimately, however, reformers brought about the end of the
veneration of Mary and the saints and the “mystery of Christ’s birth” (McKay 46). As
the devotional outlook of the songs increasingly raised suspicions that carolers were
To complete our view of the practice according to Wenger’s model, I will now turn to
the concurrent dynamic of reification. Written texts and melodies are the most
obvious reification of caroling practice: recording songs and poems in concrete form
allowed for their continued performance and the maintenance of the practice as a
the what and the who of code switching. In the carols’ regular alternation between
English and Latin, we see a formal realization of code switching practice. Because of
that formal alternation, individuals depicted in the carols are often presented as
19
For more on the development of the English Reformation, particularly among the laity, see Marsh.
168
bilingual themselves. The sacred status of those thus depicted suggests that the people
We have already seen in Chapter 3 how regular the alternation between Latin
and English is in most macaronic carols. In the pattern typical of Types C through H,
four-line stanzas alternate three lines of English with one line of Latin. This poetically
recorded today—or, for that matter, code switches in medieval English prose (see p.
117, n. 5). Precisely in that regularity, however, lies the potential for reification. The
carols reduce what could be a much less predictable and more fluid process to a more
most concrete form, presenting moments of bilingualism neatly ranged one after the
other. Bilingual carolers on insider trajectories could find and express in these tidy
code switches the fact of their dual linguistic competence. Once given concrete form,
the community to participate as well, thereby investing the entire community with a
bilingual identity.
regularly, macaronic carols that report a particular figure’s speech may switch
speaker. Since most carols commemorate sacred feast days, the speakers portrayed
169
code switching occupy the highest echelons of the medieval sacred hierarchy: angels
announcing Christ’s birth and praising God in heaven, the archangel Gabriel, the
Virgin Mary, and even Christ himself. Such portrayals occur in dozens of carols,
announcement of Christ’s birth to the shepherds. Each stanza ends with a Latin line
[The shepherds heard an angel cry; he then sang a merry song: “Why are you so
sorely aghast? Now (is the) the foremost rising of the sun.”]
Likewise, in one of James Ryman’s carols the same angel switches from Latin to
[The angel sang then with many more, “Glory in the highest! May there be peace on
earth as well as to man, and may there be joy among the angels.]
We are thus led to assume that the angel not only can speak English as well as Latin
but also alternates between the two in conversation. That the Latin elements
paraphrase a Christmas hymn and the scriptural account does not detract from its
material, they are still presented as original to the angel; uttering these memorable
170
lines motivates his own code switch between sentences. Similar instances of angels
code switching between English and Latin high-reference phrases can be found in
other carols on the Nativity as well as several works by James Ryman that paraphrase
Moving further up the angelic hierarchy, we find eleven carols in which the
nos. 38, 232 A–B, 234 A–D, 240, 243, 244, 245). Some of these carols contain
bilingual renditions of the Ave Maria’s opening lines, as in this example from a
fifteenth-century carol on the Five Joys of Mary in Oxford, Bodleian Library, Eng.
poet. e. I:
[The first joy that came to you was when the angel greeted you and said, “Mary, full
of charity, hail, full of grace.”]
Still others incorporate the Marian antiphon Alma Redemptoris mater into Gabriel’s
[An angel came to her with great light and said, “Be well, blessed creature! You are
correctly to be called kind mother of the Redeemer.”]
The same stanza occurs in variant forms in at least three other carols from the
171
Medieval English carol writers also depicted the Virgin Mary as a fluent
bilingual speaker in her response to Gabriel at the Annunciation. In at least six carols,
Mary switches from English to lines from the Vulgate (Greene nos. 238 B, 239, 240,
244, 245, 248). The following carol, for instance, also from Balliol 354, shows Mary
Other carols have Mary switch from English to her memorable proclamation, “Ecce
ancilla Domini” (Behold the handmaiden of the Lord) (Greene nos. 238 B, 248)
Finally, the most sacred bilingual speaker depicted in the carols is Christ
himself, shown code-switching as he suffers on the cross and presides over the Last
Judgment (Greene nos. 266, 277, 364, 370, 372). Such is the case in the carol on
mortality recorded with variation in the Bodleian and Balliol manuscripts cited above
since the narrator of the carol hears this account of the Passion from a talking bird:
[Jesus Christ, when he had to die, spoke to his Father; “Father,” he said, “in Trinity,
the fear of death troubles me.”]
172
Christ also code switches in a carol on the Last Judgement from Christmas carolles
newely Inprynted, in which he repeats Latin imperatives to the blessed and cursed in
In these carols, then, code-switching not only distinguishes the most sacred
Christian figures but also punctuates the most important moments of Christian
salvation history: the Annunciation, the Nativity, the Crucifixion, the Last Judgement,
and the final vision of God exalted in heaven. The writers of these carols obviously
saw nothing amiss in having the holiest of personages code switch. Otherwise, they
surely could have written the stanzas so that the switches between languages did not
occur in mid-quotation. Nor did they, like other medieval authors and even some
itself,” as Peter Nolan argues in his discussion of Latin phrases in Piers Plowman
173
(544–45).20 Rather, they presented the vernacular alongside Latin as equally fitting
forms of sacred speech and demonstrated the equality of the two by switching
between them.
By incorporating code switches into the speech of sacred figures, bilingual carol
writers lent prestige to that linguistic behavior and reified the practice in especially
positive terms. If the Latin featured in many of the carols may be considered “high-
reference” because of its liturgical origins, here the speakers themselves are
especially marked. These heavenly figures were thus presented as sharing a linguistic
practice with the much more mundane folk who produced and performed the carols.
This offered those who esteemed bilingualism and valued code switching a special
point of contact with the subjects of the carols, more so than if the sacred figures
This is not to deny that a few carols satirize bilingual clerics by reporting code
switching in speech, as Chaucer does throughout the Canterbury Tales and Langland
does in parts of Piers Plowman (Nolan 541–42). One convivial carol from the fifteenth
century (Greene no. 418.3) portrays a clerkly fox who greets the geese he is about to
terrorize with “Pax vobis,” and the “songes of the Frere and the Nunne” mentioned above
describe clerics who seduce women at prayer times, quoting Latin from the Mass all the
while. Simply satirizing the clergy, however, does not make a work anticlerical or
20
Eamon Duffy succinctly expresses this commonplace: “It was part of the power of the [Latin]
words of consecration that they were hidden, too sacred to be communicated to the ‘lewed’, and this
very element of mystery gave legitimacy to the sacred character of Latin itself, as higher and holier
than the vernacular” (218).
174
suggest that it was written by a lay person: the same bilinguals who validated code
switching in sacred carols may very well have poked fun at it in more raucous works.
Like the macaronic carol form itself, these portrayals promoted the alternation of
Latin and the vernacular as an appropriate vehicle for the divine Word. By presenting
angels, archangels, and the Mother and Son of God as fellow bilinguals, writers of
macaronic carols advanced their own suitability as messengers of that Word. The
sociable nature of the carol and its widespread popularity ensured that this laudatory
175
Implications for Further Study
Since this study is organized chronologically, its ending coincides with that of
both a popular form of macaronic verse and the medieval period in England. Its
Anglo-Saxon poems and the later carols allows one to study them via the same
modern linguistic descriptions of code switching. In the same way, their mutual
extending the methodologies of Chapters 2 and 5 to both early and late texts.
Examining how code switches in the carols enact particular discourse functions as
Catherine Davison and Elaine Rebecca Miller do for other medieval texts would
interpretive methods outside the liturgy, long after the liturgical trope had ceased to
be a productive genre. And since, as Brian Stock and Etienne Wenger demonstrate,
the circulation of texts is key to the expression of communal identity, their models
could be applied not only to the disparate carolers of late-medieval England, but also
carols to other kinds of macaronic texts would identify code switching strategies
parallel to the trends Pieter Muysken sees at work among modern immigrants to the
increase in code switches before conjunctions in later carols, for example, and the
ablative phrases could signal a general shift from an insertional mode to a more
alternational one. Linking clauses in either language rather than embedding Latin
phrases was perhaps facilitated by the development of Early Modern carol types
explained largely on stylistic grounds, macaronic texts also provide a useful means
for studying language change. We have already seen how bringing English and Latin
together within sentences shaped the syntax of either language locally, as when Latin
elements in the Anglo-Saxon macaronic poems follow a more typically Old English
word order or when English sentences in later medieval works adopt the more
hypotactic structure of French or Latin writing. Additional inquiry of this sort would
and syntax of some Latin elements in the Anglo-Saxon macaronic poems, for
example, could have resulted from writers adapting Latin linguistic structure, not out
177
of ignorance but as the result of language-contact phenomena still at work in the
While the previous examples demonstrate how code switching influences the
structure of the languages involved, changes in the languages themselves can in turn
affect code switching. Comparison of the switches in bilingual carols with those in
other macaronic texts from the Middle and Early Modern English periods may reveal
way that it could switch with Latin also changed: comparing code switches in
multilingual medieval texts across time therefore can reveal historical changes in
Such comparison can extend to the present day. This study would not be
possible unless we presumed similarities in code switching practice that obtain across
languages and periods. More overt juxtaposition of medieval and modern bilingual
texts and modern bilingual speech would foreground these similarities and display
additional points of difference. The code switches in the verse considered here may
often violate linguistic constraints as a way to draw attention to code switching itself,
following a written rather than a spoken norm.1 These modern writers may in fact
find useful models in the work of their medieval forebears as they likewise reify their
1
This observation was first suggested to me by Almeida Jacqueline Toribio.
178
Appendix: Macaronic Carols by Type
(Designated by Greene no.; those of Types F—L dated on or after 1492 are underscored.)
A B C D E F
Isolated L 1 L Line / L L & E Burden L Burden, L & E Burden, E Burden,
Borrowings Throughout 1 Repeating Repeating Varying Caudae
Stanza Caudae Caudae
95, 130, 148A, 3, 51, 93, 118, 19, 20, 70, 76, 35A, 49, 60.1, 34, 35B, 45, 14, 26, 33, 36,
172a, 241, 148B, 239, 92, 94, 101 B, 69, 80, 91, 96, 46, 56, 59, 62, 68, 82, 110,
248, 253, 258, 311 126, 140, 125.1, 178, 64, 72, 77, 81, 114, 122C,
322A 152A, 152˚B, 179.1, 194, 103, 125.2, 131, 132, 187,
157A, 168, 196, 197, 198, 138, 185, 186, 188, 189, 201,
169, 265, 312, 199, 200, 195, 215, 204, 232, 461,
319, 333, 335, 200.1, 202, 234D, 245, 461.1
399, 418.3 206, 207, 208, 266, 276, 289,
209, 210, 211, 293, 295, 296,
212, 216, 220, 297, 298, 299,
221, 222, 223, 334, 353, 358,
224, 225, 226, 367, 370, 378
227, 228,
234A, 234B,
234C, 237,
238A, 238B,
238C, 243,
244, 262, 267,
277, 278, 279,
279.1, 284,
285, 286, 287,
288, 289.1,
290, 300, 301,
302, 303, 305,
313, 330, 351,
352, 410, 426,
429
G H I J K L
L Burden, L & E Burden, L Burden, L & E Burden, L Burden, L & E Burden,
Varying Caudae Varying Caudae 2 L Lines per 2 L Lines per 3+ L Lines per 3+ L Lines per
Stanza Stanza Stanza Stanza
122A, 122B, 9, 21, 24, 29, 39.1, 75, 86A, 23A, 23B, 109, 127 218, 372
173, 242, 366 38, 39, 99, 86B, 179, 23C, 31, 52,
100, 104, 105, 191A, 229, 98, 273
119, 176, 230, 275 , 291, 292,
240, 274, 304, 294, 474
307, 364, 369,
375, 457
179
Bibliography
Berry, Mary. What the Saxon Monks Sang: Music in Winchester in the Late Tenth
Century. Bishop fithelwold: His Career and Influence. Ed. Barbara Yorke.
Woodbridge, Suffolk: Boydell, 1997. 149—160.
Baker, Peter S. The Inflection of Latin Nouns in Old English Texts. Baker and
Howe 187—206.
Baker, Peter S. and Nicholas Howe, eds. Words and Works: Studies in Medieval
English Language and Literature in honour of Fred C. Robinson. Toronto: U of
Toronto, 1998.
Belazi, Hedi M., Edward J. Rubin, and Almeida Jacqueline Toribio. Code Switching
and X-Bar Theory: The Functional Head Constraint. Linguistic Inquiry 25
(1994): 221—237.
Blockley, Mary. Apposition and the Subjects of Verb-Initial Clauses. Baker and
Howe 173—86.
____________. Aspects of Old English Poetic Syntax: Where Clauses Begin. Urbana:
U of Illinois, 2001.
Blockley, Mary and Thomas Cable. Kuhn’s Laws, Old English Poetry, and the New
Philology. Beowulf: Basic Readings. Ed. Peter S. Baker. New York: Garland,
1995. 261—79.
Brook, G. L., ed. The Harley Lyrics: The Middle English Lyrics of MS. Harley 2253.
3rd ed. Manchester: Manchester UP, 1964.
Brooks, Nicholas. The Early History of the Church of Canterbury: Christ Church
from 597 to 1066. Leicester: Leicester UP, 1984.
180
Brown, Carelton and Rossell Hope Robbins. The Index of Middle English Verse. New
York: Columbia UP, 1943.
Bukofzer, Manfred F. Popular and Secular Music in England (to c. 1470). The New
Oxford History of Music 3: Ars Nova and Renaissance 1300-1540. Ed. Anselm
Hughes and Gerald Abraham. London: Oxford UP, 1977. 107—33.
Cain, Christopher M. Phonology and Meter in the Old English Macaronic Verses.
Studies in Philology 98 (2001): 273—91.
Camargo, Martin. Two Middle English Carols from an Exeter Manuscript. Medium
fivum 67 (1998): 104—11.
Coss, Peter, ed. Thomas Wright s Political Songs of England. London: Camden
Society, 1839. Rpt. Cambridge, U.K.: Cambridge UP, 1996.
Cressy, David. Literacy and the Social Order: Reading and Writing in Tudor and
Stuart England. Cambridge, U.K. : Cambridge UP, 1980.
Crowley, Joseph. Anglicized Word Order in the Old English Continuous Interlinear
Glosses in London, British Library, Royal 2. A. XX. Anglo Saxon England 29
(2000): 123—152.
181
Davis, Kathleen. National Writing in the Ninth Century: A Reminder for
Postcolonial Thinking about the Nation. Journal of Medieval and Early
Modern Studies 28(3): 611-37.
De Lubac, Henri. Exegese medieval: Les quatres sens de l’ecriture. 4 vols. Paris,:
Aubier, 1959. Trans. Mark Sebanc. Medieval Exegesis. Grand Rapids: W.B.
Eerdmans, 1998.
Dewick, Edward Samuel and William Howard Frere. The Leofric Collectar. 2 vols.
Henry Bradshaw Society 45 and 56. London: Henry Bradshaw Society, 1914
and 1921.
Dobbie, Elliot Van Kirk, ed. The Anglo-Saxon Minor Poems. Anglo Saxon Poetic
Records 6. New York: Columbia UP, 1942.
Dobson, R. B. Cathedral Chapters and Cathedral Cities: York, Durham and Carlisle
in the Fifteenth Century. Northern History: A Review of the History of the
North of England and the Borders 19 (1983): 15—44.
Dreves, Guido Maria and Clemens Blume, eds. Analecta hymnica medii aevi. 55 vols.
Leipzig: Fues’s Verlag, 1886—1922. New York: Johnson Reprint Corp., 1961.
Dyboski, Roman, ed. Songs, Carols, and Miscellaneous Poems from the Baliol MS.
354, Richard Hill s Commonplace Book. Early English Text Society e. s. 101.
London: Kegan Paul, Trench, Tr bner, and Co., 1907.
Eble, Connie. filfric and Bilingualism in Anglo-Saxon England. The Fourth Lacus
Forum 1977. Columbia, S. C.: Hornbeam, 1978. 423—32.
Evans, Gillian Rosemary. The Language and Logic of the Bible: The Earlier Middle
Ages. Cambridge, U.K.: Cambridge UP, 1984.
182
Fennell, Barbara A. A History of English: A Sociolinguistic Approach. Oxford:
Blackwell, 2001.
Fisher, John H. Chancery and the Emergence of Standard English in the Fifteenth
Century. Speculum 52 (1977): 870—899.
Funke, Otto. Die gelehrten lateinischen Lehn- und Fremdw rter in der altenglischen
Literatur. Halle: Max Niemeyer, 1914.
Gabel, Leona C. Benefit of Clergy in England in the Later Middle Ages. Diss. Bryn
Mawr, 1928. Smith College Studies in History 14. Northampton, Mass.: Dept.
of History of Smith College, 1929.
Gameson, Richard. The Origin of the Exeter Book of Old English Poetry. Anglo-
Saxon England 25 (1996):135—85.
Gatch, Milton McC. The Office in Late Anglo-Saxon Monasticism. Lapidge and
Gneuss 341—62.
183
Gneuss, Helmut. Anglicae linguae interpretatio: Language Contact, Lexical
Borrowing and Glossing in Anglo-Saxon England. Proceedings of the British
Academy 82. Oxford: Oxford UP, 1993. 107—148. Rpt. inLanguage and History
in Early England. Aldershot: Variorum, 1996. Item V.
Greatrex, Joan. The Cathedral Monasteries in the Later Middle Ages. Monks of
England: The Benedictines in England from Augustine to the Present Day. Ed.
Daniel Rees. London: Society for Promoting Christian Knowledge, 1997.
118—34.
Greene, Richard Leighton. The Early English Carols. 2nd ed. Oxford: Clarendon,
1977.
Harlow, Geoffrey, ed. Old English Verse Texts from Many Sources: A Comprehensive
Edition. Early English Manuscripts in Facsimile 23. Copenhagen: Rosenkilde
and Bagger, 1991.
184
Universitatis Stockholmiensis: Studia Latina Stockholmiensia 29. Stockholm:
Almqvist and Wiksell International, 1985. 23—58.
Jeffrey, David L. Early English Carols and the Macaronic Hymn. Florilegium 4
(1982): 210—27.
Jeffrey, David L. and Brian J. Levy, eds. The Anglo-Norman Lyric: An Anthology.
Studies and Texts 93. Toronto: Pontifical Institute of Medieval Studies, 1990.
Knowles, David. The Monastic Order in England: A History of its Development from
the Times of St. Dunstan to the Fourth Lateran Council, 940—1216.2nd ed.
Cambridge, U.K.: Cambridge UP, 1963.
Knowles, David and R. Neville Hadcock. Medieval Religious Houses: England and
Wales. London: Longmans, Green, and Co., 1953.
Krapp, George Phillip and Elliott Van Kirk Dobbie, eds. The Exeter Book. Anglo-
Saxon Poetic Records 3. New York: Columbia UP, 1936.
185
______________. The Hermeneutic Style in Tenth-Century Anglo-Latin Literature.
From Athens to Chartres: Neoplatonism and Medieval Thought. Ed. Haijo Jan
Westra. Leiden: E.J. Brill, 1992. 97—114.Rpt. in Lapidge, Anglo-Latin
Literature 105—49.
Lapidge, Michael and Helmut Gneuss, eds. Learning and Literature in Anglo-Saxon
England: Studies Presented to Peter Clemoes on the Occasion of His Sixty-Fifth
Birthday. Cambridge, U.K.: Cambridge UP, 1985.
Lefferts, Peter M. Medieval England, 950—1450. Antiquity and the Middle Ages:
From Ancient Greece to the 15th Century. Ed. James McKinnon. Englewood
Cliffs, N. J.: Prentice Hall, 1991. 170—196.
Lucas, Peter J. Rev. of Visible Song: Transitional Literacy in Old English Verse, by
Katherine O Brien O Keeffe. Review of English Studies n. s. 44 (1993), 401—03.
"macaronic, a. and n." Oxford English Dictionary. Ed. John Simpson. 3rd ed. OED
Online. Draft Mar. 2001. Oxford UP. 30 Oct. 2001.
<http://dictionary.oed.com/cgi/entry/00298768>.
186
McGee, Timothy J. with A. G. Rigg and David N. Klausner. Singing Early Music:
The Pronunciation of European Languages in the Late Middle Ages and
Renaissance. Bloomington, Ind.: Indiana UP, 1996.
McKay, Frank. The Survival of the Carol in the Seventeenth Century. Anglia 100
(1982): 36—48.
_____________. Linguistic Identity in the Middle Ages: The Case of the Spanish
Jews. Crossing Boundaries: Issues of Cultural and Individual Identity in the
Middle Ages. Ed. Sally McKee. Arizona Studies in the Middle Ages and
Renaissance 3. Turnhout, Belgium: Brepols, 1999. 57—77.
Mitchell, Bruce. A Guide to Old English. 5th ed. Oxford: Blackwell, 1992.
Niles, John D. Rev. of Visible Song: Transitional Literacy in Old English Verse, by
Katherine O Brien O Keeffe. Speculum 68 (1993): 851—53.
Nolan, E. Peter. Beyond Macaronic: Embedded Latin in Dante and Langland. Acta
Conventus Neo-Latini Bononiensis. Ed. R. J. Schoek. Medieval and Renaissance
Texts and Studies 37. Binghamton, N.Y.: Center for Medieval & Early
Renaissance Studies, 1985. 539-548.
Parker, David R. The Act of Supremacy and the Corpus Christi Carol. English
Language Notes 30 (1992): 5—10.
Paroissien Romain (Liber usualis). Paris: Soci t de S. Jean L vang liste, Descl e et
Cie, 1935.
Pinkster, Harm. Latin Syntax and Semantics. Trans. Hotze Mulder. London:
Routledge, 1990.
188
Poplack, Shana. Contrasting Patterns of Code-switching in Two Communities. The
Sociolinguistics Reader 1: Multilingualism and Variation. Ed. Peter Trudgill
and Jenny Cheshire. London: Arnold, 1998. 44—65.
Poplack, Shana, David Sankoff, and Christopher Miller. The Social Correlates and
Linguistic Processes of Lexical Borrowing and Assimilation. Linguistics 26
(1988): 47—104.
Reed, Edward Bliss, ed. Christmas Carols Printed in the Sixteenth Century.
Cambridge, Mass.: Harvard UP, 1932.
Robinson, F. N., ed. The Works of Geoffrey Chaucer. 2nd ed. Boston: Houghton
Mifflin, 1957.
189
Romaine, Suzanne. Bilingualism. Oxford: Blackwell, 1995.
Sankoff, David, Shana Poplack, and Swathi Vanniarajan. The Case of the Nonce
Loan in Tamil. Language Variation and Change 2: 71—101.
Schofield, B. The Provenance and Date of Sumer is Icumen In. Music Review 9
(1948): 81—86.
Seymour, St. John D. Three Medieval Poems from Kilkenny. Proceedings of the
Royal Irish Academy 41 (1932—34): B 205—09.
Shaw, Brian A. The Old English Phoenix . Medieval Translators and Their Craft
Ed. Jeanette Beer. Studies in Medieval Culture 25. Kalamazoo: Medieval
Institute Publications, 1989. 155—83.
Stelton, Leo F., ed. Dictionary of Ecclesiastical Latin. Peabody, Mass.: Hendrickson,
1995.
Stenton, Frank Merry. Anglo-Saxon England, 3rd ed. Oxford: Clarendon, 1971.
190
Stevens, John, ed. Early Tudor Songs and Carols. Musica Brittanica 18. London:
Stainer and Bell, 1962.
______________. Medi val Carols. Musica Brittanica 4. London: Stainer and Bell,
1952.
Stevenson, Joseph, ed.. The Latin Hymns of the Anglo-Saxon Church, with an
Interlinear Anglo-Saxon Gloss. Surtees Society 23. Durham, U.K.: G. Andrews,
1851.
Sullivan, Carmeline. The Latin Insertions and the Macaronic Verse in Piers
Plowman. Diss. Washington, D.C.: Catholic U of America, 1932.
Thomas, Keith. The Meaning of Literacy in Early Modern England. The Written
Word: Literacy in Transition. Ed. Gerd Baumann. Oxford: Clarendon, 1986.
97—132.
191
_______________________. Speaking in Tongues: Spanish, English, and the
Grammars in Between. U of Texas at Austin. 27 March 1998.
Traugott, Elizabeth Closs and Suzanne Romaine. Some Questions for the Definition
of Style in Socio-Historical Linguistics. Folia Linguistica Historica 6 (1985):
7—39.
Vantuono, William, ed. Sir Gawain and the Green Knight. Notre Dame, Ind.: U of
Notre Dame, 1999.
Watt, Tessa. Cheap Print and Popular Piety, 1550-1640. Cambridge, U.K.:
Cambridge UP, 1991.
Wheatley, Henry B., ed. Merlin, or The Early History of King Arthur: A Prose
Romance. Early English Text Society o. s. 10, 112. London: Kegan Paul,
Trench, Tr bner, & Co., 1899.
Whitbread, Leslie G. Notes on Two Minor Old English Poems. English Language
Notes 4 (1967): 241—43.
192
________________. The Old English Poem Aldhelm. English Studies 57 (1976):
193—97.
________________. The Old English Poem Judgment Day II and Its Latin Source.
Philological Quarterly 45 (1966): 635—56.
_______________. The Old English Poems of the Benedictine Office and Some
Related Questions. Anglia 80 (1962): 37—49.
Whiting, Ella Keats, ed. The Poems of John Audelay. Early English Text Society o. s.
184. London: Oxford UP, 1931.
Wieland, Gernot R., ed. The Canterbury Hymnal: Edited from British Library Ms.
Additional 37517. Toronto Medieval Latin Texts 12. Toronto: Pontifical
Institute of Mediaeval Studies, 1982.
Woods, Marjorie Curry. Shared Books: Primers, Psalters, and the Adult Acquisition
of Literacy among Devout Laywomen and Women in Orders in Late Medieval
England. New Trends in Feminine Spirituality: The Holy Women of Li ge and
Their Impact. Ed. Juliette Dor et al. Medieval Women: Texts and Contexts 2.
Turnhout, Belgium: Brepols, 1999.
11, 1973, the son of Robert Clifford LeCluyse and Josephine Marie Dragna LeCluyse
of Lenexa, Kansas. After graduating in 1991 from Rockhurst High School in Kansas
City, Missouri, he entered Oberlin College and the Oberlin Conservatory of Music in
Oberlin, Ohio. In May 1995, he received a Bachelor of Arts with Highest Honors in
English from Oberlin College, majoring in English and Music Performance (voice).
He entered the Graduate School of the University of Texas in August 1995 and
received a Masters of Arts in English in May 1997. While preparing the prospectus
for this dissertation, he worked as an associate editor at Holt, Rinehart, and Winston,
from June 1998 to August 1999. During his candidacy for the Ph.D., he taught
introductory courses in the Division of Rhetoric and Composition and the Department
of English and worked as a writing consultant and assistant director in the University
194