University of Virginia Library

SCRIBAL INTENTIONS IN MEDIEVAL ROMANCE:
A CASE STUDY OF ROBERT THORNTON

by
John Ivor Carlson

I. A Rationale of "Scribal Corruption"?

Although codicologists have long been familiar with the tendency among
medieval copyists to shape their texts through amateur editing, the recent popu-
larity of critical approaches that treat books as social constructs has renewed and
redirected interest in this fact.[1] Complexities introduced by independent-minded
scribes, once approached solely as impediments to establishing archetypal read-
ings, have now become a subject for analysis under broader interpretations of
textual criticism. As Derek Pearsall notes, "scholars are learning the value of bad
manuscripts: how in the work of interfering and meddling scribes, for instance,
can be seen the activities of our first literary critics. The methods of compilers
and manuscript editors of all kinds, whether professional or amateur, need to
be studied, if we are to understand the reception and readership assumed for
the literary works contained in their collections."[2] Sometimes, as in the case of
Pearsall himself, this interest in scribal agendas has been presented as part of a
wider critique of eclectic editing and its pursuit of authority. It stands to reason,
though, that more precise knowledge about copying habits would be as much
a boon to those seeking authorial readings as to their antagonists.[3] Whatever a

scholar's attitude toward the issue of intention, it is hard to take exception to
calls for increased understanding of those who shaped the evidence with which
every textual critic, whether a disciple of George Kane or Jerome McGann, must
inevitably contend.[4]

If a nearly universal desire for more thorough analysis of scribal intention can
be taken as a given, the practicality of such study is far more problematic than
many recent advocates of this methodology have been willing to admit.[5] Identify-
ing the contributions of individuals in any text represents a daunting task under
ideal circumstances, but when collaborators are anonymous, indeterminate in
number, and of unknown purpose, it requires overwhelming evidence to defend
anything more than the most banal assertions. The collation of independent wit-
nesses for at least one work in a manuscript is essential, of course, when isolating
unique copying features. Still, even then it would assume much to attribute ap-
parent tendencies to a single scribe, let alone extrapolate a coherent rationale of
intervention, without first repeating this comparison across several other texts in
the same hand.[6] It is difficult to imagine how scholars lacking fortuitous textual
circumstances could improve much on the standard description of usus scribendi
already familiar to those even passingly acquainted with the works of Kane, Eu-
gene Vinaver, or their predecessors.[7] The most prohibitive requirement of study-
ing unique scribal agendas is therefore the reliable identification of individual
copyists with sizable oeuvres that encompass multiple texts extant in at least one
other independent witness.

Among the select group meeting this minimal demand in full, Robert Thorn-
ton stands out as particularly well-suited for a case study in scribal intention
owing to both his status as an amateur copyist and the well-deserved critical at-
tention his writings have previously attracted. Thornton, otherwise an unremark-
able member of the fifteenth-century English gentry, created two miscellanies
intended for his family's private edification: Lincoln Cathedral MS 91 (olim Lin-
coln Cathedral A.5.2) and British Library MS Add. 31042.[8] These compendiums
of devotional literature and courtly poetry, while doubtless shaped in part by the
availability of suitable exemplars, also appear indebted to Thornton's own tastes
as an independent scribe.[9] One example of these interests is his large collection
of alliterative romances, which raises his work from curiosity to marvel with the
inclusion of Wynnere and Wastour and the Morte Arthure among other notewor-
thy witnesses. Clearly, as the sole source for documentary evidence regarding
these highly regarded poems, Thornton's manuscripts possess enough historical
import to justify reanalysis of his scribal habits. Yet while this selection of texts
adequately explains his work's literary interest, Thornton's unusual freedom has
direct bearing on our ability as textual critics to evaluate his intent as well: sub-
ject to external pressures less demanding than those facing professional scribes,
Thornton had greater opportunity to develop idiosyncratic copying habits that
suggest personal motives more nuanced than either expediency or accuracy. This
providential combination of scribal independence with textual merit makes ten-
able an investigation into one copyist's compilation methods that could reshape
widely-accepted notions about an important source of medieval literature.

II. Past Assessments of Robert Thornton

Before we move forward with an investigation of Thornton's idiosyncrasies
focused on his scribal agenda, however, it is instructive to examine those conclu-
sions reached by critics whose primary concern has been with reconstructing
authorial texts. These analyses, well-argued and reflecting sincere conviction,
highlight in their range and contradiction the impressionistic character of many

such considerations of scribal intent. Consider, for example, the amusing contrast
between the appraisals of Thornton offered by L. F. Casson and Mary Hamel:
It is difficult to imagine any reconciliation between these conclusions, but gener-
ous readers might perhaps imagine that inherent differences between the texts
under consideration led to these incompatible assessments. Unfortunately, the
arguments of Sir Israel Gollancz and Stephanie Trigg regarding Wynnere and
Wastour undermine any such glossing interpretation:
Less generous critics, in light of these disagreements and their chronology, might
note here that Casson and Gollancz represent an earlier generation of scholar-
ship and contend that it should be expected that their work lacks the critical
nuance of more enlightened days. Frances McSparran's recent and persuasive
work, though, complicates that condescending narrative by taking a more skepti-
cal view of Thornton's reliability:
What is one to make of these conflicting interpretations? Rather than differences
in textual evidence or critical acumen, it would seem that the obstacles faced
by editors in identifying authorial readings have most influenced their findings
about scribal intent: those casting doubt on Thornton's faithfulness arrive at their
readings through collations of independent witnesses (Casson and McSparran) or
appeals to an aesthetic ideal (Gollancz) ultimately posited on the imperfection of
all surviving documents; those defending his work (Hamel and Trigg) invariably
lack the resources for the former approach and doubt the validity of the latter.
Bluntly speaking, explications of Thornton's habits to date have told us much
more about how certain editorial methods encourage assumptions about scribes
generally than how one fifteenth-century copyist interacted with his exemplars.
The elimination of such bias is the most important reason to stop discussing
scribal habit only in the context of authorial intent and instead begin treating it
as a subject worthy of investigation in its own right.

III. The Methodology of Reassessment

Allowing that any true reassessment of Thornton must approach his scribal
contributions as legitimate objects of study, the present inquiry employs a two-
step process that first contrasts his manuscripts against other witnesses to isolate
unique readings and then seeks explanations for those variants consistent with
what is known of his background. When this strategy is followed, patterns of
change emerge that commonly involve the insertion of extraneous dialogue at-
tributions, conjunctions, negations, and intensifiers as well as more generalized
stylistic smoothing through the rearrangement of word order and explicit iden-
tification of a sentence's subjects or objects. On occasion these changes seem
indifferent in effect, but more often than not the copyist sacrifices poetic fea-
tures like metricality to make basic syntactic structure and narrative content
more explicit even when these elements should be perfectly clear to any reader.
Evaluating these elements of his scribal profile in light of what we can ascertain
about Thornton, the cumulative evidence suggests that he engaged in predictable
interventions motivated by a desire to present these tales orally for the moral
education of his family.

Providing a factual basis about Thornton's characteristic habits on which a
broader analysis and interpretation of his scribal agenda can rest presents its own
challenges. Foremost among these difficulties is the selection of specific works
from which reliable evidence regarding scribal habit may be inferred. Given that
the purpose of this study is to differentiate one scribe's emendations from those
originating with either the author or other copyists, the most fundamental criteria
in choosing texts must be the existence of independent witnesses to contrast with
the Thornton redactions. This basic principle rests on the assumption that shared
readings between related manuscripts often represent inherited text, whether ar-
chetypal or not, whereas unique passages imply an original scribal intervention.[15]
Among those writings that meet this requirement, it would also seem reasonable
to limit the study further to the same poems that have focused scholarly attention
on Robert Thornton to ensure some homogeneity within the evidence.[16] These
prerequisites yield a basic corpus of six Middle English romances: Sir Degrevant
(which survives in one other manuscript), Octovian (existing in two manuscript
besides Thornton's), The Siege of Jerusalem (for which eight other witnesses are
known), The Parlement of the Thre Ages (partially preserved in another manuscript),
The Awntyrs off Arthure (recorded in three other manuscripts), and Richard Coer de
Lyon (existing in six other witnesses).[17]

With the issue of which works to include resolved, the next step in defining
the scope of this investigation involves choosing specific passages in each work
for collation. Since an acceptable policy must minimize the chance that unchar-
acteristic copying stints or problematic stretches of text will skew the study's
conclusions, selecting lengthy passages far removed from one another seems the
most conservative approach. With these criteria in mind, one finds that groups
of one hundred lines near the beginning and end of the manuscript witnesses for
each text will serve admirably in almost every case. An exception to this deci-
sion only appears justified when looking at Richard Coer de Lyon, where accidental
duplication of twenty lines by Thornton generates unique evidence outside these
boundaries equally deserving of inclusion in the study.[18] The total body of text
under investigation according to these standards will therefore be more than
four thousand lines in six texts witnessed across twenty-seven manuscripts, an

extensive collection of data that should provide enough material to accurately
characterize Thornton's scribal habits.

Having established this study's textual boundaries, we can now address the
crucial matter of adopting guidelines to differentiate Thornton's scribal interven-
tions from variants introduced by other copyists. Considering the wide-spread
confusion evident in previous scholarship on this topic, it is only sensible that
these principles set a standard for acceptable evidence that is both concise in its
definition and equally clear about its limitations.[19] Consequently, no variant in
Thornton's manuscripts will be definitively ascribed to his intervention in this
study unless analogous readings appear several times across multiple texts in
his copies while similar passages not found in his texts occur rarely in other wit-
nesses. This simple procedure will allow for the swift identification of features
deserving closer evaluation and eliminate from consideration those of ambiguous
provenance: for example, if comparisons between the manuscript witnesses re-
veal that both Thornton and other scribes regularly omit pronouns, this tendency
could not be credited with certainty to any one copyist even if each instance were
singular; evidence that Thornton alone inserts disproportionately large numbers
of extraneous conjunctions in his romances, on the other hand, would signal
that this idiosyncrasy represents an aspect of his individual copying habits. Such
enumerative analysis is admittedly incapable of establishing the scribal nature of
particular readings, no doubt explaining its dismissal in the past by editors con-
cerned with authority, but it will allow for the description of those broad patterns
in variation relevant to the issue of scribal intent. While this strict methodology
might disqualify some lightly attested features contributed by Thornton to his
works, guidelines less restrictive would certainly lead to uncertain if not errone-
ous attributions that would undermine the conclusions extrapolated from this
investigation.

IV. Constructing Thornton's Scribal Profile

Applying these criteria for determining scribal intervention to the corpus
selected earlier, one finds that the most striking examples of idiosyncratic copy-
ing involve dialogue attributions. The collations suggest that Thornton, rather
than allowing readers to determine the presence of direct speech and a speaker's
identity using contextual clues, commonly inserted phrases like "he said" or
"she spoke" to remove all potential for syntactic ambiguity.[20] Fifteen instances
of this type of intervention occur in Thornton's manuscripts, thirteen of which
are unique, compared to only nine instances found elsewhere that he does not
share.[21] These numbers might not seem particularly remarkable, but it is im-
portant to remember that in context such a distribution means that close to

two thirds of such dialogue attributions occur only in one copyist's manuscripts
whereas the remaining third is divided across the other twenty-seven indepen-
dent witnesses.[22] Octovian, as copied in Lincoln Cathedral MS 91, demonstrates
Thornton's characteristic treatment of dialogue attributions in its very first in-
stance of direct speech:
This stands in contrast to the same passage in Cambridge Ff.2.38, which at the
same point has a syntactically unmarked transition from narration to dialogue:
There is no way to determine the relative authority of these readings in the ab-
sence of clear metrical evidence, but the first iteration's attempt to smooth away
ambiguity through explicitness suggests the scribal nature of the Lincoln redac-
tion. This failure in poetic economy, while not conclusive, leads to the stylistic
flatness often ascribed by textual scholars to copyists and thereby supports the
assertion that such expanded readings are non-authorial.[25]

While typifying the changes to dialogue, though, that example cited above
from Octovian only becomes significant in light of supporting evidence found in
other romances proving it reflects Thornton's usus scribendi. These examples are
cited by line number in this article's second appendix, but special note should be
taken of those instances within the portion of Richard Coer de Lyon copied twice by
our scribe. One can find in that short passage startlingly clear corroboration of
the scribal nature of these attributions spread across three lines grouped closely
together:

Each of the dialogue attributions above is left out of all other witnesses to Rich-
ard and, in two instances, this also includes the accidental second copying of the
passage later in the same manuscript by Thornton himself. This contrast within
his own work appears very suggestive in regards to the handling of dialogue and
firmly establishes that the scribe modified his exemplars freely:
Given that six independent witnesses along with this second copy by Thornton
indicate no explicit dialogue attributions, it is extremely probable that these read-
ings originate with our scribe and not some earlier archetype. This pattern of
intervention implies that Thornton consciously marked the move from narration
to dialogue, but, despite the conspicuous nature of these changes, they are not
the most pervasive modifications he makes to syntactic transitions.

Another element of Thornton's scribal practice involving the connections
between syntactic elements manifests itself in an abundance of unnecessary con-
junctions and conjunctive adverbs. Scanning the collation sample for evidence
of these paratactic features uncovers two hundred and twenty-one instances, of
which one hundred and fifty-eight are unique, showing clearly the copyist's re-
markable dependence on this grammatical crutch.[28] Other manuscript witnesses,
in contrast to such profusion, exhibit only ninety-seven examples not found in
either the Lincoln Cathedral or British Library manuscripts. Typifying the con-
structions favored by Thornton, the following passage appears in his copy of The
Awntyrs off Arthure:

Although the other three witnesses all differ slightly from one another, the follow-
ing reading found in MS Douce 324 of the Bodleian Library illustrates the aver-
sion to repetitive connective words common to these other versions of the text:
Despite the brevity of this passage, the thoroughness with which Thornton fills
out his text's transitional elements seems difficult to mistake. If one ignores the
medial conjunction in the second line because of its presence in another witness,
there remain four unique conjunctions and adverbs within these lines of the
Lincoln redaction not found elsewhere. Two of these, furthermore, introduce
independent clauses that grammatically require no explicit linkage to the preced-
ing text. This lack of function, along with the combined authority of the other
manuscripts, implies the scribal nature of these elements and suggests their origin
lies with Thornton.

As with dialogue attributions, though, the firm association of paratactic in-
sertions with one scribe relies upon the distribution of the feature across several
independent texts. This seems especially true in regard to these transitional ele-
ments because Thornton's approach differs from his scribal peers not so much in
nature as in degree. Along with the numerous other examples listed in the second
appendix, The Parlement of the Thre Ages provides support for this observation
while also demonstrating that Thornton's extraneous conjunctions and adverbs
are notably more widespread even when compared against what has been identi-
fied as a more corrupt witness:[31]

Thornton begins each of these four lines with an extraneous conjunction, whereas
the Ware scribe, typically more prone to copying errors and sophistications than
Thornton, records only one of those verses in a similar manner:
Although at this point the specific rationale for such interventions remains un-
clear, the impression that these changes amount to something more thorough-
going than the usual scribal tendency to round out lines is strengthened greatly
by comparison to the Ware text. Thornton's indulgent use of these simple
function words ultimately complements the dialogue attributions and confirms
that an interest in syntactic transitions is a central component of his scribal
profile.

Differing slightly in emphasis from previous examples of his unique contribu-
tions, Thornton's preference for doubling negative modifiers introduces a new
wrinkle into his tendency to use certain common grammatical constructions with
unusual frequency. This particular rhetorical ploy, which emphasizes negation
without changing or adding nuance to meaning, appears uniquely eleven times in
Thornton's texts and only six times when he does not record it.[34] A fairly typical
example of this change's effect on the text can be seen in Sir Degrevant's descrip-
tion of the festivities surrounding the titular knight's marriage:

This passage shows how the copyist's use of doubling emphasizes what could be
thought of as a semantic transition from positive to negative, indicating some
continuity with but also extension of the scribal features previously noted. The
Cambridge Manuscript version of Sir Degrevant, in contrast, does not demonstrate
any such concern for highlighting negative statements:
The difference between the two readings here may have been inspired by minim
confusion, but, even assuming mechanical error, the change suggests a predis-
position toward doubled negative constructions. Indeed, this preference pres-
ents itself even more forcibly in passages where there is no easy paleographic
explanation:
Although the additions in these lines do not undermine the meter, this reading
compares unfavorably in its redundancy with the less convoluted statement of
the other witness:
Straining to accommodate the doubled negatives, the Thornton passage lacks
the jauntiness of expression evident in the other manuscript's reading and shows
the copyist's willingness to sacrifice style in order to accentuate sense. Thornton's
emphasizing of such semantic markers, while not overwhelmingly attested given
the limited environments that lend themselves to such constructions, clearly jus-
tifies their inclusion on a list of his scribal habits alongside more pronounced
intrusions.

The unique stress Thornton places on some semantic markers becomes easier
to discern when one considers his handling of adverbial intensifiers. These much-
maligned "filler" words appear seventy-five times in Thornton's work, sixty-one
of which are unique, as opposed to the twenty-seven times he does not share in
all the other manuscripts combined.[39] Although such intensifiers sometimes prove
original in Middle English verse, the disparity between the number found in
Thornton's copies and other redactions suggests that once again the multiplica-
tion of examples is the responsibility of this particular copyist. Octovian illustrates
this increased density quite dramatically in an early passage involving several
unique instances of scribal filler:

Altogether there are four words in this passage that qualify as intensifiers, three
instances of "full" along with one of "fele," and three of these have no parallel
in the reading of the other manuscript witness:
It should be immediately apparent that, since their very nature ensures that such
words are extraneous to sense, the simple violation of poetic economy cannot
be held as proof of their scribal nature. Additionally, readings such as these sel-
dom lead to metrically unacceptable lines since the differences they create rarely
exceed a single unstressed syllable. Instead, the best evidence identifying them
as a copyist's intrusion is provided by the large number of unique examples in
well-attested texts like The Siege of Jerusalem:
Given that six other witnesses exist for these lines, including a group represented
by the second passage cited above that has been identified as closely related to
the Thornton copy quoted first, the unique nature of the examples in his redac-
tion strongly suggests the scribal nature of such additions. With similar situa-
tions existing in Awntyrs and Richard, the insertion of intensifiers can therefore
be accepted as one last example of the copyist's attempt to highlight important
semantic markers through his additions.

However, the addition of new material is not the only means by which
Thornton clarifies syntactic meaning and this is demonstrated most clearly by
his regular tinkering with both word order and the degree of explicitness in his
identification of sentence elements. These instances of syntactic smoothing oc-
cur ninety-five times in Thornton's work, sixty-nine of which are unique, and
only thirty-eight times in other manuscripts where they are not repeated in his
work.[44] Sir Degrevant provides a cluster of changes typifying these tendencies early
in its text, as shown here in excerpts from Thornton's copy and the Cambridge
University Library manuscript respectively:

In this passage, Thornton adds pronouns to explicitly identify what are treated as
implied subjects in the other manuscript, thereby incidentally altering the syntax
to create two independent sentences as reflected in the modern punctuation cited
here. Again, the lines as written in the other witness are not difficult to interpret
and this action appears to be hyper-correction, just like the following sentence
inversion in the Thornton and Cambridge copies of Octovian:
This example shows Thornton moving the object of "wedde" from an anterior to
posterior position, creating the standard subject-verb-object sentence order most
common in English prose but less rigidly adhered to in poetry. Wholesale recast-
ing of sentence order is a bit rarer than the explicit identification of subjects and
objects, but such interventions do convey admirably the overall essence of the
changes Thornton introduces into his texts. Correcting what is not essentially in
error, such inversion shows our scribe once again replacing poetic features with
more prosaic expressions whenever the latter are achieved through grammatical
constructions that rely on an audience's close attention to detail.

V. From Profile to Rationale

Although the firm identification of five distinct types of scribal intervention
constitutes in itself a worthwhile discovery, especially for editors attempting to
ascertain Thornton's role in his unique texts, the results also encourage con-
sideration of what motivation might lie behind these changes. Two possibilities
immediately suggest themselves, random chance and scribal laxity, but these
simple explanations prove inadequate. Attributing the copyist's unique readings
to chance, for instance, overlooks both the pervasiveness of the changes made
and the striking regularity in both their method and function. It clearly violates
the principle of Occam's razor to suppose each of these individually minor altera-
tions could occur by accident in such predictable and patterned ways. However,
while chance may be quickly ruled out as unlikely, the tendency among scribes
to "simplif[y] their task by reproducing content and sense rather than exact de-
tails of wording and language" that Frances McSparran credits for Thornton's
variations in Octovian requires more deliberation.[49] McSparran's belief that such
changes reflect the translation encouraged by relaxed copying procedures, espe-
cially if one defines this process in terms of idiolects rather than dialects, offers a
tempting rationale for the choices Thornton makes. Still, this approach fails to
account for the somewhat obsessive spelling changes identified by Mary Hamel

in her study of this copyist's self-corrections.[50] It defies reason to propose that
the same scribe who commonly altered his spellings in an attempt to mimic his
exemplar would add several hundred words to each text without considering
their purpose. One must therefore conclude that the emphases placed on syn-
tactic structure and semantic markers by the scribe's intervention originate in a
conscious attempt by Thornton to shape his work.

As previously indicated, determining the specific agenda behind those pur-
poseful changes Thornton made to his exemplars requires an examination of
his scribal habits in light of what can be established about his background. This
involves, in effect, an appeal to the "cultural milieu" in which the Lincoln and
British Library texts were compiled.[51] Thornton's particular social environment,
as reflected in the choice and arrangement of his manuscripts, appears to value
especially the didactic potential of the written word. As George Keiser points
out, "Robert Thornton [along with another of his contemporaries, Elizabeth
Sywardby] … clearly found in the book an important, perhaps even an essential
means to spiritual fulfillment and … their interest in the book for this purpose
does not make them by any means unusual members of their society."[52] By divid-
ing his miscellanies evenly between religious tracts and moralistic secular content,
Thornton signals to his audience that these texts should function as exemplars
for good-living. Even the vernacular romances largely conform to this principle,
on occasion stating explicitly in their incipits that the portrayals of "owre eldurs
olde" were meant to teach "soothe sawys."[53] Although perhaps not self-evident,
a connection between Thornton's interest in presenting chivalric tales as if they
were secular saints' lives and his copying habits can be made. Considered as a
system of scribal intervention, the changes revealed in the preceding consider-
ation of the Thornton manuscripts seem complementary to each other in their
efforts to simplify, smooth, and emphasize different grammatical elements essen-
tial to sense. All of these interventions, then, reveal the common scribal desire to
alert readers to potentially confusing narrative elements and ensure that no one
misinterpret the author's perceived intent.[54] This inclination can be reconciled
with a counter-tendency to preserve minutiae like the exemplar's spellings if one
allows that, given Thornton's concern that each tale's didactic message reach his
audience, an intermittent willingness to sacrifice literal fidelity to preserve his
texts' underlying meaning might represent another form of conservatism. In a
sense, the pattern of intervention revealed here suggests that Thornton elected

to follow a rational tertium quid between the extremes of mechanical and creative
copying ascribed to him by previous scholars to ensure the moralistic themes he
saw as important would be accessible to his audience.

While concern for his readers' understanding provides a unifying motivation
for the copyist's interventions, it falls short of explaining why Thornton em-
phasized one specific set of syntactic transitions and semantic markers so much
more than his contemporaries. The final explanation of Thornton's grammati-
cal tinkering instead depends upon the overlap between the ways in which the
scribe emended his exemplars and the textual features some critics have used to
argue that his romances were copied down from performances by minstrels. The
proliferation of linking words, poetic filler, and other redundancies created by his
changes were all considered symptomatic of oral composition and the demands
it placed on the poet's creative resources. Representing a standard interpretation
of Thornton's romances grounded in oral composition theory, Casson claims that
"X [Sir Degrevant's archetypal form] was committed to memory by a minstrel.
He probably transmitted the poem to another of his kind, and the process may
have continued for several generations (in the textual sense). This is the most
satisfactory way of accounting for most of the numerous periphrases, anticipa-
tions, recollections, upset rhyme schemes, inconsistencies, insertions of redun-
dant words and general literary dilution to be found on almost every page of
[Thornton's Lincoln manuscript]."[55] However, disenchantment with theories of
oral composition and dissemination has led to a rejection of such generalizations
as demonstrated in McSparran's recent introduction to Octovian: "The rather free
variation of wording.… seems inconsistent with this preservation of phonologi-
cally non-significant details of spelling from the archetype. The first would be
compatible with oral transmission, the second could not survive it, and must de-
pend on a written tradition."[56] McSparran's explanation of variation between the
two copies of Octovian, supported by Hamel's work on self-correction, seems ef-
fectively to refute Casson's argument proposing oral composition. However, her
subsequent admonition to "consider the [presumably written] cultural context in
which popular romances were produced and circulated" must be combined with
recognition that those oral elements identified by Casson aptly summarize the
grammatical features privileged by Robert Thornton's scribal habits.

While the widespread oral composition of late medieval literature has seemed
increasingly unlikely to critics in recent years, study of Thornton's cultural con-
text suggests that the oral recitation of romances may represent the key to our
copyist's patterns of intervention. The most plentiful evidence of such oral reci-
tation comes from the works themselves, wherein poets repeatedly exhort their
audience to listen attentively, although Thorlac Turville-Petre is certainly right to
warn that such evidence "may be suspected of being an idealization."[57] Despite
such cautious doubt, though, even Turville-Petre concedes that "[i]t is probable
enough that recitation, whether by a professional.… or member of the family,
was one method of dissemination [for romances]" and provides intriguing evi-

dence of such oral delivery in his edition of the fragmentary "Lament for Sir John
of Berkeley."[58] This rhymed alliterative poem gives a short account of another
unremarkable fifteenth-century member of the English gentry whose rank and
interests seem suggestively similar to those of Thornton:
The audience is told that John of Berkeley was "a generous householder who
entertained in the grandest fashion … with food and jollity, song and … the
reading of romances."[60] If one remains open to the possibility that Thornton, like
Sir John, intended either to read his texts aloud for friends and family or have
them read by another, his copying style begins to make better sense. This slightly
unorthodox explanation of Thornton's alterations would mean that emphasis on
syntactic structure and semantic meaning stems from the scribe's belief that his
audience would misunderstand more subtle grammatical constructions better
suited to the eye than the ear when listening to the tales presented aloud.[61]

The conclusion, albeit tentative, that Thornton consciously crafted the ver-
nacular romances in his miscellanies for spoken delivery suggests an approach
to the relationship between scribe and exemplar that would significantly change
our view of this copyist's work. While previous characterizations of Thornton
as alternately conservative or innovative seem to capture in part the complex
interaction he engaged in with his texts, these arguments have never been able
to account for what seem to be his contradictory tendencies. These failures most
likely reflect the assumption that the value of a scribe's work can only be mea-
sured against the yardstick of authorial intention and that the need to recover
an original meaning precludes in-depth study of individual copyists. Concentrat-
ing on scribal intention, though, encourages one to see consistent patterns of
intervention like those Thornton engages in as rational attempts to reconcile a
received text with an anticipated use. These patterns and the statistical evidence
illustrating them can be used then to establish with greater accuracy archetypal
readings. In fact, combined with other types of textual evidence, knowledge of

Thornton's scribal agenda could have a profound effect on the editing of those
texts he copied by allowing for the recovery of authorial readings even in unique
witnesses where collation is not possible.

VI. Conclusions and Application

While the restoration of authorial readings has not been the primary focus
of this article, a brief look at two passages in Thornton's unique copies of Wyn-
nere and Wastour and the Morte Arthure will begin to suggest practical applications
for the scribal profile outlined above. Since information about copying habits
can only be used to anticipate tendencies and not identify specific errors, lines
have been selected that also exhibit obvious deficiencies when judged according
to metrical and alliterative norms. The first of these exemplary passages, taken
from the Morte Arthure, describes the arrival of Emperor Lucius' envoys at King
Arthur's court and demonstrates how the scribal inversion of sentence structure
could upset a poem's delicate structure:

The first three lines maintain the widely attested A-A-A-X pattern, in which
the first three of four stressed words carry the expected alliteration. This is the
standard format of the Middle English alliterative long-line and therefore, at
least in terms of poetic structure, all three can be taken without further comment
as authorial. Unfortunately, in the fourth line the pattern switches to the non-
standard A-A-X-A, which is generally not considered an acceptable variation
since without alliteration the third stressed syllable cannot fulfill its traditional
role as the link between the two half-lines.[63] In light of Thornton's penchant for
syntactic smoothing, however, it is probable that the authorial version of this
line had a less prosaic ordering of verb and object: "Ilke A kyng Aftyre kyng and
his enclines mad."[64] A slightly different situation arises, on the other hand, dur-
ing Wynner's speech about responsible forestry in Wynnere and Wastour where it
is Thornton's weakness for paratactic expression that suggests a correction to a
flawed b-verse metrical pattern:
Hoyt Duggan's widely-accepted rules of b-verse metricality stipulate that any
b-verse with two or more dips of multiple unstressed syllables are categorically
non-authorial, and that restriction would clearly apply to the third line quoted
here.[66] While the only logical correction for this error is the elimination of the
extraneous conjunction, Thornton's consistent interference with syntactic transi-
tions offers yet another means to justify this emendation. Given how glaringly
obvious the errors are in both this and the previous example, however, some
might question whether the ability to confirm their existence through knowl-
edge of Thornton's copying habits is any great advantage. Judging the scribal
evidence solely on its ability to unearth new error misrepresents the challenges
inherent to the editorial process, though, which is more often stymied by the lack
of corroborative evidence for corrections than any shortage of errors needing
emendation. An understanding of Thornton's unique tendencies provides the
confirmation an editor would need to justify adopting specific changes in both
these and many other lines, representing an improvement over past invocations
of a generic usus scribendi too easily accommodated to preconceived notions about
scribal fidelity.

While we are fortunate in the identification of Thornton and his manuscripts,
especially given his importance in preserving alliterative verse, the same prin-
ciples demonstrated here should inform other new studies of scribal habit as well.
That is to say, textual scholars should seek whenever possible to explain scribal
agendas in the same detail previously reserved only for the authorial intent. This
does not mean abandoning interest in traditional forms of authority, but rather
realizing that without such complementary efforts we will never understand fully
either aspect of medieval literature. Most editorial appeals to scribal habit carry
little weight, unfortunately, representing impressionistic judgments indebted to
preconceived notions about scribal fidelity and the best means of recovering
authorial readings. The previous scholarship on Robert Thornton demonstrates
this susceptibility toward bias, and the case study offered here is intended both as
a correction of that specific problem and as a general template for new research.
Especially now, as the wider availability and functionality of electronic editions
makes such descriptive work less burdensome, textual scholars and editors have
little excuse for treating the study of scribal agendas as a mere afterthought.

Increased understanding of a copyist's motives may not be possible in every cir-
cumstance, but the statistical method offered here at least provides one practical
method of answering recent calls to extrapolate from scribal interventions some
idea of how early audiences received and interacted with medieval texts.

APPENDIX 1
Manuscripts and Lines Collated and Critical Editions Consulted

As indicated in the article's main body, the default choices for each text were
one-hundred-line groups shared by Thornton's copy with other witnesses at the
beginning and end of each text. Those instances where the number of lines ex-
ceeds two hundred, excepting the case of Richard Coer de Lyon, reflect the need to
compensate for lines omitted in the Thornton texts.

• I. The Awntyrs off Arthure at the Terne Wathelyne, ll. 1–100; 605–709.
• Manuscripts. Lincoln Cathedral MS 91; MS Douce 324; Bibliotheca Bodmeriana, Coligny,
  Geneva, f.5a; Lambeth Palace Library MS 491.
• Critical Edition. Gates, Robert J., ed. The Awntyrs off Arthure at the Terne Wathelyne, A Critical
  Edition. Philadelphia: Univ. of Pennsylvania Press, 1969.[67]
• II. Octovian, ll. 1–100; 1365–1464.[68]
• Manuscripts. Lincoln Cathedral MS 91; Cambridge University Library MS Ff.2.38.
• Critical Edition. McSparran, Frances, ed. Octovian, Edited from Lincoln Dean and Chapter
  Library MS 91 and Cambridge University Library MS Ff.2. 38. EETS 289. London:
  Oxford Univ. Press, 1986.
• III. The Parlement of the Thre Ages, ll. 226–325; 559–664.
• Manuscripts. British Library MS Add. 31042; British Library MS Add. 33994 (olim Ware
  MS).
• Critical Edition. Offord, M. Y., ed. The Parlement of the Thre Ages. EETS O.S. 246. 1959;
  repr. London: Oxford Univ. Press, 1997.
• IV. Richard Coer de Lyon, ll. 1–100; 983–1002; 7105–7209.
• Manuscripts. British Library MS Add. 31042; Auchinleck MS; MS Egerton 2862; Gonville
  and Caius College, Cambridge MS 175/96; MS College of Arms HDN 58; MS
  Douce 228; MS Harley 4690.
• Critical Edition. Brunner, Karl, ed. Der mittelenglische Versroman über Richard Löwenherz,
  kritische Ausgabe nach allen Handschriften mit Einleitung, Anmerkungen und deutscher
  Űbersetzung. Wiener Beiträge zur englischen Philologie, XLII. Wien und Leipzig:
  Wilhelm Braumüller, 1913.
• V. The Siege of Jerusalem, ll. 1–104; 1223–1326.
• Manuscripts. British Library MS Add. 31042; Bodleian 1059 (Laud Misc. 656); British
  Library Cotton Caligula A.ii; Cambridge University Library Mm.5.14; Lambeth
  Palace 491; Ashburnham 130.

---

69




• Critical Edition. Hanna, Ralph, III, and David Lawton, eds. The Siege of Jerusalem. EETS
  320. London: Oxford Univ. Press, 2003.
• VI. Sir Degrevant, ll. 1–100; 1820–1919.
• Manuscripts. Lincoln Cathedral MS 91; Cambridge University Library MS Ff.1.6.
• Critical Edition. Casson, L. F., ed. The Romance of Sir Degrevant: A Parallel-Text Edition from
  MSS. Lincoln Cathedral A.5.2 and Cambridge University Ff.I.6. EETS O.S. 221. 1949; repr.
  London: Oxford Univ. Press, 1970.

APPENDIX 2
Examples of Scribal Habits

The sections below provide a comprehensive list of examples within the sam-
ple corpus for each scribal habit considered in this article. These citations are
organized both according to the work in which they appear and their distribu-
tion across the manuscript witnesses: unique examples appear only in Thornton's
manuscript; shared examples appear in both Thornton and at least one other
witness; counter-examples appear only in manuscripts other than Thornton's.

• I. Dialogue attributions
• Unique examples
• Awntyrs: l. 621
• Octovian: ll. 55, 62, 1422, 1461
• Parlement:—
• Richard: ll. 986, 995, 998, 7180
• Siege: ll. 82, 97, 101, 1231
• Degrevant:—
• Shared examples
• Awntyrs: l. 664
• Richard:—
• Siege: l. 15
• Counter-examples
• Awntyrs: ll. 96, 627, 640, 646
• Octovian: l. 1371
• Parlement:—
• Richard:—
• Siege: ll. 1234, 1236, 1304, 1305
• Degrevant:—
• Total (and percentage of all examples)

       

Unique examples:	13	(54%)
Shared examples:	2	(8%)
Counter-examples:	9	(38%)
All examples:	24	(100%)

• II. Paratactic expressions (conjunctions and linking adverbs)
• Unique examples
• Awntyrs: ll. 14, 22, 32, 42, 52, 55, 57, 62, 80, 82, 608, 621, 663, 668, 677, 696, 698,
  699, 701, 705, 709.
• Octovian: ll. 7, 10, 17, 20, 27, 36, 37, 43, 46, 49, 50, 57, 61, 67, 73, 79, 83, 85, 98, 1378,
  1385, 1386, 1400, 1405, 1407, 1410, 1416, 1419, 1435, 1443, 1452, 1455, 1463, 1464
• Parlement: ll. 232, 236, 238, 248, 249, 251, 268, 276, 286, 297, 299, 301, 321, 568,
  570, 571, 578, 588, 603, 607, 617, 649

---

70




• Richard: ll. 4, 7, 10, 12, 15, 17, 19, 20, 27, 29, 33, 39, 45, 49, 56, 69, 60, 67, 73, 75, 77,
  94, 95, 97, 983, 985, 986, 991, 993, 994, 997, 998, 999, 1001, 7110, 7111, 7115, 7117,
  7132, 7155, 7156, 7157, 7167, 7191, 7195, 7203, 7209
• Siege: ll. 8, 11, 17, 33, 59, 74, 77, 81, 88, 93, 95, 102, 1231, 1232, 1233, 1251, 1318
• Degrevant: ll. 6, 11, 12, 26, 72, 79, 85, 99, 1841, 1849, 1864, 1867, 1873, 1874, 1879, 1901,
  1905
• Shared examples
• Awntyrs: ll. 3, 6, 9, 14, 19, 27, 53, 54, 61, 63, 66, 613, 618, 627, 638, 641, 653, 686, 696,
  699, 707
• Richard: ll. 57, 987, 7135, 7141, 7143, 7145, 7146, 7149, 7151, 7175, 7187
• Siege: ll. 5, 13, 22, 23, 47, 51, 61, 62, 63, 67, 71, 78, 90, 92, 104, 1228, 1236, 1245, 1258,
  1261, 1264, 1267, 1275, 1276, 1283, 1284, 1288, 1295, 1299, 1322, 1324
• Counter-examples
• Awntyrs: ll. 3, 26, 27, 29, 33, 40, 45, 53, 81, 86, 606, 612, 614, 615, 619, 620, 624, 631,
  640, 645, 646, 662, 691, 693, 695, 703
• Octovian: ll. 41, 45, 66, 1374, 1421, 1445
• Parlement: ll. 237, 253, 256, 293, 588, 619, 620, 624, 644, 650, 658
• Richard: ll. 25, 55, 7114, 7163, 7172
• Siege: ll. 2, 14, 18, 19, 29, 30, 31, 34, 35, 48, 56, 65, 66, 68, 69, 89, 1225, 1229, 1230, 1241,
  1242, 1248, 1249, 1251, 1253, 1255, 1259, 1263, 1264, 1286, 1293, 1296, 1297, 1298,
  1301, 1304, 1317, 1321
• Degrevant: ll. 9, 86, 1823, 1839, 1842, 1853, 1859, 1885, 1907, 1908, 1909
• Total (and percentage of all examples)

       

Unique examples:	158	(49%)
Shared examples:	63	(20%)
Counter-examples:	97	(31%)
All examples:	318	(100%)

• III. Doubled negative constructions
• Unique examples
• Awntyrs:—
• Octovian: ll. 19, 1380
• Parlement: ll. 288, 583, 631
• Richard: ll. 22, 7118, 7166
• Siege:—
• Degrevant: ll. 62, 78, 1873
• Shared examples
• Awntyrs:—
• Richard:—
• Siege:—
• Counter-examples
• Awntyrs: l. 59
• Octovian: l. 1404
• Parlement:—
• Richard: l. 7139
• Siege: ll. 1246, 1250, 1289
• Degrevant:—
• Total (and percentage of all examples)

       

Unique examples:	11	(65%)
Shared examples:	0	(0%)
Counter-examples:	6	(35%)
All examples:	17	(100%)

• IV. Adverbial intensifiers
• Unique examples
• Awntyrs: ll. 21, 33, 55, 85, 88, 607, 614, 619, 633, 654

---

71




• Octovian: ll. 7, 8, 12, 74, 79, 86, 87, 93, 1365, 1421, 1429, 1430, 1439, 1463
• Parlement: ll. 227, 281, 310, 321, 565, 568, 573, 595
• Richard: ll. 7, 55, 59, 70, 74, 81, 88, 92, 97, 989, 990, 997, 999, 7112, 7134, 7152
• Siege: ll. 54, 56, 57, 68, 78, 1254, 1268, 1269, 1276
• Degrevant: ll. 1831, 1859, 1899, 1902
• Shared examples
• Awntyrs: ll. 67, 70, 87, 620
• Richard: ll. 5, 8
• Siege: ll. 37, 47 (x2), 69, 1226, 1257, 1269, 1275
• Counter-examples
• Awntyrs: ll. 8, 26, 27, 61, 69, 82, 86, 652, 658, 674, 682, 691, 705, 708
• Octovian: ll. 23, 97, 1433
• Parlement: ll. 249, 263
• Richard: ll. 44, 62, 7137
• Siege: ll. 29, 91, 1258
• Degrevant: ll. 81, 99
• Total (and percentage of all examples)

       

Unique examples:	61	(60%)
Shared examples:	14	(14%)
Counter-examples:	27	(26%)
All examples:	102	(100%)

• V. Syntactic smoothing
• Unique examples
• Awntyrs: ll. 1, 45, 49, 57, 693
• Octovian: ll. 10, 15, 20, 22, 44, 65, 67, 74, 94, 1369, 1374, 1378, 1384, 1387, 1421
• Parlement: ll. 234, 287, 298, 302, 636, 640, 653
• Richard: ll. 8, 13, 18, 30, 31, 52, 66, 79, 84, 86, 983, 987, 996, 7113, 7114, 7115, 7116, 7118,
  7177
• Siege: ll. 4, 19, 48, 56, 1297, 1315
• Degrevant: ll. 31, 38, 85, 93, 98, 1822, 1823, 1826, 1838, 1842, 1853, 1857, 1865, 1878,
  1886, 1905, 1913
• Shared examples
• Awntyrs: ll. 2, 31, 34, 645
• Richard: ll. 32, 984, 986, 992, 7136, 7144
• Siege: ll. 3, 17 (x2), 19, 21, 32, 39, 41, 47, 55, 82, 1229, 1241, 1243, 1287, 1317
• Counter-examples
• Awntyrs: l. 33
• Octovian: ll. 11, 23, 80, 1367, 1380
• Parlement: ll. 231, 235, 246, 281, 562, 570
• Richard: ll. 21, 995, 7119, 7132, 7173, 7180
• Siege: ll. 18 (x3), 21, 29, 64, 80, 101, 1260, 1264, 1274, 1300, 1321, 1325
• Degrevant: ll. 57, 77, 83, 1850, 1873, 1903
• Total (and percentage of all examples)

       

Unique examples:	69	(51%)
Shared examples:	26	(20%)
Counter-examples:	38	(29%)
All examples:	133	(100%)