University of Virginia Library

MEMORIAL TRANSMISSION, SHORTHAND,
AND JOHN OF BORDEAUX

by
Gerald E. Downs

The hypothetical process known as memorial reconstruction was developed
in the twentieth century by Sir Walter W. Greg and the "New Bibliogra-
phers" to explain printed Renaissance playtexts identified as "bad quartos." Evi-
dence cited to support the concept (if not the concept itself) has been questioned,
principally by Paul Werstine and Laurie E. Maguire.[1] Arden Shakespeare editor
Giorgio Melchiori replies that critics "intended to nullify all narratives by their
predecessors. Theirs was not an alternative narrative…."[2] However, Werstine
and Maguire do specify other possible origins of these problematic publications
relevant to the renewal in this essay of a hypothesis that New Bibliographers
themselves discount. My analysis of the manuscript playtext John of Bordeaux,[3]
aptly described by Harry R. Hoppe as "a bad quarto that never reached print,"[4]
indicates that it is transcribed from the stenographic recording of a stage perfor-
mance. I compare this inference to prior scholarly opinion that the text is either
a memorial reconstruction or a descendent transcription of authorized text and
I discuss the implications of my findings.

Before shorthand can be fairly argued as a cause of any playtext, a history of
the current opinion devaluing that method of transmission must be reviewed, be-
ginning with a clarification of the terminology. Early in Suspect Texts Maguire as-
serts that "memorial reconstruction" is synonymous to "bad quarto" (15), and to
"reporting" (18), apparently in agreement with Fredson Bowers, whose thirteenth
and final listed class of Elizabethan printer's copy is "memorial reconstruction
of the text without direct transcriptional link with any manuscript derived from
author's autograph, in other words, the copy for a so-called 'bad quarto.'"[5] But
earlier in the century Greg had carefully defined "reporting," listing six of its

forms, including "shorthand reporting" and "memorial reconstruction."[6] These
and other terms may be redefined not only to adhere to their historical usages,
but to maintain distinctions necessary to discussion.

When in 1938 Leo Kirschbaum held that a bad quarto "cannot possibly rep-
resent a written transcript of the author's text,"[7] he referred to a physical set of
fewer than fifty editions printed between 1591 and 1620 that have been suspected
historically to be so corrupt that a significant part of each of these dramas re-
sults from memorial transmission—a process recovering text without access to an
original or descendent manuscript from which the agent's knowledge of the text
derives, or to any other transcribed descendant. This term may be usefully ap-
plied to oral, written, or printed text without quantification or identification of
any method of transmission.

A memorial report (or report) is a relatively full text derived at least in part from
any originating form of memorial transmission. Hypothetical cases of reporting
that may result in bad quartos include "memorial reconstructions" and "theatri-
cal reports," where a memorial reconstruction is the written report of a playtext, not
taken from performance, but recalled by an actor or actors (or someone pos-
sessed of a like knowledge of the text). A theatrical report is a text derived directly
from stage performance, when the oral report is recorded and the performance
is otherwise described. A shorthand report is a category of theatrical report com-
prising stenographic notes, their longhand transcription and other descendants,
including revisions, printed texts, and performances.[8] Similarly, others of these
terms are applied hypothetically by convention to particular bad quartos. To
imply descent from a category or a subcategory, for example, the 1597 edition
of Shakespeare's Romeo and Juliet has been called a "report" and a "memorial
reconstruction."

Maguire suggests that memorial reconstruction is conceptually flawed be-
cause it is "capacious, being able to explain almost any textual problem" (Texts
7). Werstine more accurately remarks that memorial reconstruction "by an actor
or actors identified with specific parts has never proved an adequate explanation
for the genesis of any 'bad' quarto; the case…has needed to be supplemented
by secondary hypotheses" (82). When John Jowett responds to Maguire "that for
some texts a capacious theory seems to be needed, and that few are available,"[9]
he apparently accepts the weakness of multiple hypotheses in order to retain
memorial reconstruction as the best practicable explanation of some bad quartos.
But the "alternative narrative" of a shorthand report, as suggested by Werstine

in his criticism of the New Bibliography, has been "available" since Thomas
Heywood's introduction to The Rape of Lucrece in 1608:

Maguire's examination of memorial reconstruction conforms to her obser-
vation that a "connection between suspect texts and shorthand reports has not
been a component of studies … for many years, having been cogently disproved
by G. I. Duthie in 1949" (Texts 18).[10] However, Adele Davidson ably questions
the consensus, concluding in part that in "analyzing Elizabethan shorthand,
Duthie tackled an abstruse subject, and students of Shakespeare have referred
to Duthie's work without examining fully either Willis's system or Duthie's own
methods and conclusions."[11] Further criticism of his analysis is warranted, but
Heywood's authority and the related capacity of a shorthand theory should first
be addressed. According to Werstine,

Heywood referred to this circumstance again in 1637 by publishing the pro-
logue (but not the playtext) to a then-recent stage revival of his If You Know Not
Me, You Know Nobody, originally printed in 1605 (and probably one of Heywood's
"mangled" subjects in 1608). He laments that his old play is:

Except for those knowingly assisting a memorial report, no judge could be
better placed than this prolific dramatist who twice condemns the transmission of

his own work. Yet he is often cited to deny his assertion. For instance, after sug-
gesting that Heywood's memory in 1637 was faulty, Maguire observes of G. N.
Giordano-Orsini's conclusion that If You Know Not Me is a memorial reconstruc-
tion that "Heywood was mistaken in his suggestion of shorthand, or, if shorthand
were used, it was supplemented by other means. If Giordano-Orsini's conclusion
is correct, it suggests that Heywood had no idea how his text had been con-
structed …" (Texts 103).[13] But for thirty years Heywood had a definite idea that
his text was reproduced by stenography from a performance by players whose
characterizations he must have known intimately.

B. A. P. van Dam observes of a bad quarto derived from shorthand that its
errors will be those "of the scribe who has copied the acting parts, the mistakes
made by the actors … those of the stenographer, made in taking it down and
afterwards when transcribing his notes into longhand, and … the mistakes of the
printer also. It is a highly interesting question whether it is possible to distinguish
these various mistakes."[14] Many modern scholars have not thought this matter
through. For example, Arden editor T. W. Craik suggests that "shorthand would
not produce all these [invariable] characteristics" of bad quartos.[15] Yet nothing
theoretically prevents a shorthand report from incorporating these features, ex-
cept insistence that other agents, including actors, did not exhibit them. The only
evidence of performance standards is the bad quarto itself, and then only if it is
a theatrical report. Van Dam asserts that some actor-errors confirm theatrical
reporting, while he acknowledges that few other kinds of error are peculiar to
the stenographer. Shorthand reporting dependent on multiple agents inherently
accommodates evidence that need not be explained by stenography alone. The
hypothesis therefore enjoys some immunity to negative evidence that is fatal to
theories of unitary origin. In other words, shorthand reporting is capacious. How
was it discredited?

The first pages of Duthie's Shorthand and Lear examine Timothy Bright's Char-
acterie, first printed in 1588.[16] Joseph Quincy Adams wrote in 1934 that "Wil-
lis' system is entirely based on phonetics; Bright's system, on 'characters' that
stood for individual words without respect to sound. Only Bright's system can
explain the errors in the [1608 first quarto of King Lear]."[17] For Adams the Q1
errors are determined by what "Shakespeare wrote, according to the Folio"
(149). This adheres to Greg's opinion that "where the [1623] folio differs from
the quarto its readings … must be derived from the authoritative playhouse

manuscript."[18] Yet Adams's case for characterie was quickly confuted by Mad-
eline Doran and by William Matthews.[19] Doran notes that differences in the two
early Lear texts were possibly caused by multiple-agent "alteration of unknown
proportions; in that case argument in terms of mere quantity is of little value"
(142). Nevertheless, Duthie adopts an argument similar to Adams's, though he
means not to support shorthand, but to deny it by suggesting that errors in Q1
Lear, again determined by Folio variance, are unlikely to derive from Willis's
system. This strategy dominates Duthie's argument, as one may infer from his
title.

When words variant in Q1 and F Lear begin with the same letter, Duthie
accepts, strictly for the sake of argument, that the Q1 word could be a shorthand-
caused error, which he accounts for in one of two ways. First, the reported word
is assumed to have been on the stenographer's memorized list of common words
and their representations: e.g. "m" = "mother", but was miswritten ("M") or
mistranscribed ("moon"). Or he creates a phonetic stenograph to explain the er-
ror. The "initial-letter" rationale borrows anti-Bright argument, but in every case
(the symbolic or the phonetic) Duthie discovers alternative causes of the variants
that reduce the probability of the shorthand errors actually having occurred. His
discussion of line 3.2.58 (by Greg's line numbering) is typical.

Assuming that a mix-up of "c" symbols caused the quarto error "centers" for
the Folio reading "Continents", Duthie suggests instead that the "copy for Q may
have had 'continents'. The compositor may have managed to deal only with the
'contin' and the final 's'. He may have read the 'o' as 'e', the 'i' as 'e', and the
'n' as 'r'…. Since 'centers' is a word he may have decided to be content …"
(Shorthand 67). This response to a supposed shorthand error "may have" no ra-
tionale other than a determination to list alternatives. Stone offers a case for
"cincture" to emend Q1 and questions "Continents" as a failed correction (196).
Yet even such a plausible error is not exclusive to stenography; the variants are
no more than suggestive.

A second example of questionable analysis of possible shorthand error occurs
at 2.2.151; variants Q1 "ont" and F "out". Duthie argues that if "a stenographer,
intending to write [Duthie's stenographic "out"], made the great character too
long at the foot and also straightened the foot out a little … he might be under-
stood to be conveying 'ont'…. But 'ont' is a very likely misreading of a longhand
'out'" (Shorthand 44). Supposition that a stenographer would be unable to write or
read "out" in his notes seems artificial when all forms of error are likely to occur
in bad quartos, regardless of their sources. As Davidson observes, the listing of
alternatives does not logically disprove an explanation ("Stenography Reconsid-
ered" 80), and Doran's objection to "mere quantity" also applies.

Duthie did not disturb prevailing assumptions when he accepted that Q1 Lear
can be judged by the F text, but modern opinion must reckon with later think-
ing, as expressed by P. W. M. Blayney: "The differences between Q and F are
not—most emphatically not—primary evidence for Q's origin. They show only
what was done to Q at some unknown date after its publication. The only source
of primary evidence for the origin of Q is Q itself."[20] This statement invalidates
Duthie's argument. For example, recent debate over Lear's possible revision from
a Q1-like text to an F-like text by Shakespeare himself has seen many scholars
commit to Blayney's hypothesis that Q1 derives from an authorial draft.[21] Any-
one holding Blayney's opinion must reject Duthie's acceptance of the 1608 text as
a memorial reconstruction; they cannot have Q1 Lear printer's copy both ways.

Although Q1 Lear is seldom categorized now as a bad quarto, Duthie be-
lieved the printer's copy was a memorial reconstruction, and his Folio-based
case against shorthand depends on it. His explanation for twenty variants is
"certainly or probably, a memorial corruption"; for another fifty-six, "substitu-
tion by the actor for the correct (F) reading" (Shorthand 69). These kinds of error
will occur in a report of performance. But for scholars—such as Maguire (Texts
270)—contending both that Q1 is not a memorial report and that Duthie's argu-
ment against Willis's system is sound, alternatives must be found for the variants
consistent with their own conclusions.

However, Duthie's argument resonated with the mid-century New Bib-
liography: in 1955 Greg stated that "all speculation on [shorthand] has been
scotched by Duthie's recent investigations and is not likely to be revived."[22]
Yet Greg had earlier supported shorthand as explaining Q1 King Lear, despite
his claim to know "all the objections to the theory…."[23] Therefore, of his
later acceptance of Duthie's case without published analysis, we may infer that
Greg was influenced after all by the failed reconciliation of shorthand to the
Lear Q1/F variants.

Duthie's stenographs were developed to show how Willis's system could ac-
count for Q1, but they lead to a contrary conclusion. In a review of Duthie,
Matthews observes that in "many of the [variant] pairs, the [requisite] symbols
are so different that the quarto forms could not possibly have arisen from a con-
fused shorthand form."[24] But Q1 was a candidate for shorthand only because it

seemed to report the "authorized" text better than a reconstruction. For Q1 Lear
to be a shorthand report of F, these irreconcilable variants must be F revisions or
Q1 errors, some attributable to the players. Yet neither possibility was, a priori,
admitted by the New Bibliography;[25] the textual status quo welcomed both nul-
lification of shorthand and acceptance of Q1 as a memorial reconstruction. But
when recent opinion that Q1 does not report F reduces Duthie's title-argument
to irrelevancies, his general objections to shorthand make an inadequate case.

Duthie addresses one pertinent topic only in his separate analysis of Q1
Hamlet, where he repeats others' prior argument.[26] Scholarship acknowledges
shorthand piracy of sermons preached extemporaneously during the era of the
bad quartos. Victimized clergy sometimes responded by authorizing minimally
altered reprints, where little opportunity arises for comparison. However, in 1922
H. T. Price reprinted Henry Smith's A Fruitful Sermon in parallel texts (both from
1591) and in 1933 he analyzed two editions of a 1589 Stephen Egerton sermon.[27]
Price's case for Bright's system was dismantled by William Matthews and when
Adams's argument for characterie reporting King Lear was confuted at the same
time on the same grounds, the study of sermon reporting was abandoned with
some issues yet to be resolved. Matthews addresses one of these:

Alan Herr accepts this inference as "entirely reasonable,"[28] though it is also
likely that the names were adopted by writers who were not shorthand practitio-
ners, which may apply to opinion matching Heywood's in authority. Sir George
Buc, nephew and chief assistant to Master of the Revels Edmund Tilney, and
later himself Master of the Revels from 1610–22, assesses one of these terms:

Of this positive statement by one in a position to know, Maguire suggests that Buc
"actually claims more for Brachygraphy than Bales ever did" (Texts 99), though
Buc probably meant the word to refer to shorthand systems generally—and
expertly—in use.

Although I agree to some extent with Herr that "accuracy of a shorthand
transcript of an Elizabethan sermon is almost certainly incapable of verification"
(82), my reading of many of these sermons leads me to conclude that the quality
of reporting is quite high. For instance, the style of Henry Smith's remarkable
oratory is consistent through piracies of long, involved scriptural themes. The
fact that Smith augmented many sermons before his death persuades me that he
accepted them as his own preaching, as demonstrated by A Fruitful Sermon:

The second edition makes minor changes to this passage and restores the brack-
eted line. A marginal reference "Iob. 1." is retained in the first edition, but
reoccurrence of "Blessed be the name of the Lord." probably caused a scribe's or
a compositor's eyeskip omission of the reference to Job in the body of the text.
Smith accepts the report to the extent that his restoration of an omission also
restores the cause of the omission. But he does reject first-edition additions to
the sermon:
Brackets in this instance indicate omissions in the second edition; their style identi-
fies a revising hand whose first-printing augmentations (concentrated in the final
pages) Smith cuts on sight. Apparently the preacher accepts the straightforward
exhortation as his own.

While the quality of these reports should be properly recognized, it is unnec-
essary to argue their absolute accuracy. Matthews and Doran effectively elimi-
nate characterie from consideration to force the inference that a different system
was used to report the sermons, when "not in print," in a manuscript culture,
cannot mean "nonexistent." Though phonetic, the system in use in 1590 was also
probably not Willis's.

Argument against specific features in Willis may not then apply to an al-
ternative technique. For example, though Duthie aptly notes that "all short-
hand should be cursive" (Shorthand 29), as Willis's characters are not, another
reporter could avoid this failing. Generalizations based on Willis's practice are
to be trusted even less when they are inapt. For instance, Duthie attempts to
show from Willis's prefatory use of the word treatably that the system was too
slow for verbatim reporting. But the sense of that word depends on who does the
treating. Speaking of both Elizabethan and modern stenographers, E. H. Butler
observes that "the ability of the writer is the thing that tells, even more than the

construction of the system."[30] Nevertheless, Duthie envisions a reporter able only
to "note down the gist of the speaker's argument, using the drastic methods of
abbreviating clauses and sentences.… It might be used to summarize a speech
briefly" (Shorthand 24–25). These limitations do not account for the quality of
sermons reported from 1589 on.

Conversely, evidence may show features of another shorthand system that are
consistent wi6th Willis's authority and experience. For instance, a stenographer
may act on his dictum that "not the orthographie, but the sound is respected"
(ch. 8, note a). We are left with Duthie's inadequate a priori arguments. For exam-
ple, he remarks of Willis's advice to use special stenographs (e.g. "ao" = "also")
that "The necessity for learning lists of words is a demerit in a shorthand system"
(Shorthand 36). But for the expert such lists aid speedy and exact writing. Not
surprisingly, Matthews detects Duthie's inexperience, whose objections to Willis

Many Duthie assumptions may then be met with counter-assertion. For ex-
ample, though he suggests that "In all probability the stenographer could not
escape detection" (Shorthand 75), Doran had suggested in 1935 that to "speculate
on whether or not he would be detected in the theater, or allowed to continue
if detected, is more or less useless, since we know so little about conditions in
the playhouse and the attitude of the companies toward such practices" (154).
Likewise, while Duthie asserts (mimicking Willis on the difficulties presented by
a rapid speaker) that "a stenographer present at a performance would find his
imagination transported beyond the endeavour of his hands" (25), Stone contends
that "We should have in mind…the deliberate pace of a speaker addressing a
large audience: actors…cannot speak more rapidly than the physical conditions
allow" (35). The shorthand hypothesis must be decided on other issues.

If bad quartos are bad and if memorial reconstruction is doubted as their
cause, then evidence relevant to reconsideration of an older theory of their ori-
gins should be welcome. John of Bordeaux has been studied by Laurie Maguire in
Suspect Texts and separately in reply to Hoppe's essay.[31] Further analysis of the text
as a potential bad quarto offers to shed new light on the genre.

William L. Renwick transcribed the untitled play for the Malone Society, as-
signed its name, and designated its scribe S,[32] who cannot have been the author
of the legible but partially damaged manuscript. Maguire lists several reasons
for scholarly interest in the text: it is slightly annotated by known but unidenti-
fied theatrical hands; the name of the actor John Holland is added marginally
as three entries; and a speech is added in the hand of playwright Henry Chettle.

Additionally, "it appears to have been written to dictation" ("Misdiagnosing"
114). However, Maguire neither accepts Renwick's suggestion to that effect, nor
does she agree with Hoppe that the play is a memorial reconstruction. Instead,
she posits scribal transcription, with simultaneous revision, as the necessary ori-
gins to explain the lone early copy of the play ("Misdiagnosing" 124–125). By
Maguire's criteria, a bad quarto is a memorial reconstruction or it is neither;
her "alternative narrative" can only be that corruption arose coincident with
recourse to an authorial manuscript or its descendant. I will examine John of
Bordeaux not only for evidence of transcription and memorial reconstruction, but
also for shorthand reporting.

The play is a sequel to Friar Bacon and Friar Bungay with no cause to doubt
that Robert Greene is its original author.[33] Because my investigation does not
rely on the plot, the briefest synopsis may suffice: John of Bordeaux is a military
commander opposing the Turks warring against the German emperor Frederick,
whose son Ferdinand attempts to win John's wife Rossalin by causing her husband's
exile. Prince Ferdinand is aided by the German magician Vandermast, whose ri-
valry with Friar Bacon in the earlier play is renewed when Bacon takes the side
of John and Rossalin. Throughout, Bacon's assistant Perce provides comic relief.

The text of John of Bordeaux fills most of its folio pages and only occasionally
limits line length to pentameter. Hoppe lists these facts as evidence of reporting.
Maguire reasonably argues that verse displayed end-to-end is still verse, and
she is surely correct to suggest the scribe's purpose was to save paper because,
as Hoppe also notes, the last six pages are more normally lined.[34] Yet Maguire
mistakenly asserts that the manuscript "simply lines verse as prose" and she con-
tends that "frequent coincidence of punctuation with metrical line ends suggests
that the scribe had an ear for, or knew he was copying, verse. Three out of four
metrical units are signaled by punctuation in the following quotation" ("Misdi-
agnosing" 117). The unrelined passage reads,

"Punctuation" connotes frequency and variation, but in twenty-six crowded
pages S uses only eight full stops (four on folio 1a) and 109 commas. This aver-
ages one mark for every eleven lines, or fewer than five per page. Even at this low
rate, the punctuation tends to be bunched: folio 6a has eight commas in ten lines,
none in the other forty-seven. Folio 7b opens with the three commas in Maguire's
citation but has only one in the next forty-five lines. Hoppe accurately describes

the greater part of the text as written with "total disregard of verse-lining or
punctuation" (123), which may well be a meaningful characteristic. For example,
Greg defended his opinion that the first quarto of King Lear could have derived
from shorthand, where the longhand transcription was "undivided metrically
and practically unpointed."[36] Though relatively few lines in John of Bordeaux are
irreparable, editorial care would have been needed to put it in order for printing
and "bad quarto" corruptions could have resulted, as where Bacon addresses the
chained Vandermast: "so famed for magick and for excorsismes here let lose /
yor selfe no willt not be way with him I saye" (1170–71).

Maguire suggests that from "folio 11 onward there are few errors of lineation"
("Misdiagnosing" 117). Though later lines approach metrical length, many er-
rors remain and end-to-end verse continues. For example, on 12a Bacon enters,
saying

The first isolated regular line is 1132. At 1153, Bacon instructs the devil Astrow
(verse line divisions added):
Despite a superficial regularity, the verse is mislined, beginning with the exten-
sion "say Bacon charged". The last line can be considered regular only by ac-
cepting the prior two as corrupt. Earlier on the same page, Bacon converses with
Astrow, who replies:
Is this Prose, corruption, or stannic verse? Similarly, at 1195 (folio 13b) Bacon
has stricken Vandermast mad:
Consecutive verse lines of twelve, nine, nine, fifteen, and eleven syllables would
not have been written by Robert Greene. If this is not corruption but prose, the
printer would also probably have lined the passage as verse. Although so much
error would probably not occur in circumstances other than memorial transmis-
sion, the corrupt metrical line itself is more telling than simple mislineation.
Consider these consecutive passages from 11b:
The iambic pentameter lines suggest no authorial reason to deviate from form;
lines 1078, 1082, and 1084 are faulty with twelve syllables. In contrast, John's
speech beginning at 1090 is error-free, punctuated (though not merely to mark
line-ends), well-spelt, metrical, and properly lined. Written in Henry Chettle's
hand, this addition probably mirrors the fair copy of John of Bordeaux initially
delivered to the players who purchased it from Greene. But the rest of the play's
divergences from metrical regularity are evidence of corruption from a variety
of causes. For example, Maguire lists for John of Bordeaux "extra-metrical phrases
(e.g. 115, 218, 225, 497–8, 598–9, 1045)" (Texts 266). The list is incomplete,
perhaps because Maguire believes that a "feature which can arise 'in almost any
mode of transmission' should have no place in a diagnostic study of memory"
(Texts 185). But evidence should not be excluded without contextual analysis.

Maguire dismisses the internal repetitions adduced by Hoppe as evidence of
memorial transmission. For example, Hoppe suggests that Perce's "parroting of
the previous speaker's closing words is … a reporter's repetition" (127). But the

recurring trait is obviously a clown's affectation, as Maguire shows by compari-
son with Friar Bacon ("Misdiagnosing" 120). Whether the clown speaks more than
is set down for him—as Renwick suggests (viii)—is uncertain.

Hoppe cites another sort of repetition that fails to support memorial recon-
struction at line 774: "the prayes that vertue [waits] guids and pittie waights
vpon". Maguire rightly objects that anticipation of "waits" involves a common
error that occurs simply because the anticipated word intrudes on a scribe's mind
from the nearby context. However, the well-known phenomenon of scribal eye-
skip is caused by nearby identical (or similar) words and phrases:

Hoppe correctly observes that "gates" was omitted and then added to line 459
to replace "turks". But he attributes the error to Kirschbaum's "mnemonic tele-
scoping" by a reporter (128), which draws on the eyeskip concept without estab-
lishing any analogous criteria. Hoppe discounts eyeskip by noting the prima facie
improbability of dictation taken from copy read aloud that could be transcribed
directly. But eyeskip will occur in transcribing shorthand into longhand. Hoppe
also cites a number of quickly amended scribal errors, as when Ferdinand's line
at 301 contains a repetition:
A marginal correction such as "holds", despite Hoppe's suggestion (127), need
not have been immediate. Still, the eyeskip repetition (of "scornes", caused by
reoccurrence of "she") was easily seen and corrected. More significantly, eyeskip
anticipation may cause omission that can escape notice. One of Hoppe's examples
is the lament of the Turkish emperor Amurath:
After writing "now", the scribe's eye falls on the second "now" and he writes
"selimus". The omission resulting in "how can I live now Selemus my only son
is slain" is avoided because S catches the error in time to make correction currente
calamo. This kind of evidence indicates not only transcription from written copy,
but transcription of a few words at a time, which is compatible with convert-
ing shorthand notes into longhand. Frequent transcription errors are unlikely
to occur in a manuscript dictated by reporters; the tedious exercise ("beyonnd
the commane reatch") would probably have engendered more errors, but of dif-
ferent kinds.

Hoppe cites Greg (Abridgements 320) on more distant anticipations as evidence
of memorial transmission that "can only occur where the whole of the text and
not merely the immediate context is present to the mind" (121). Maguire com-
ments on a potentially meaningful anticipation:

Maguire's "usable evidence" is two printed plays that "owe their textual dis-
order to other causes," and two manuscripts, each with a distant repetition. She
observes "that we have only one instance of anticipation in John of Bordeaux.
Isolated examples of repetition are insufficient indication of memorial recon-
struction." Yet Maguire does not explain why her citations are analogous to
Perce's anticipation. For example, she suggests that the scribe of The Fatal Mar-
riage "makes a number of bizarre errors, not least of which is his repetition of lines
1956–66 … at lines 2056–57 (sic); he realized his error while writing the second
set of lines, and then broke off midsentence …" ("Misdiagnosing" 122).

The repetition in The Fatal Marriage is not bizarre.[39] The scribe numbered a
sheet at the top left of its fold to transcribe on four folio pages. On completing
the first page he wrongly continued on the verso of the second leaf. Crossing out
his error, he transcribed nearly identically on the verso of the first leaf. The era-
sure (written earlier but appearing later than its correction) results from a known
cause that cannot form part of an argument about Perce's anticipation, which
must be evaluated on its own characteristics.

To begin, Hoppe does mention the passage (131, example 3, where both line
sets are quoted), though his treatment is inadequate: "It is almost impossible to
suppose that a reader-dictator would commit such a haplography … on the other
hand, a reporter-dictator's memory would be only too apt to skip and jump in this
way …" (131). This fails to explain how the error occurred, or how it was corrected.

Although the anticipation is nearly identical to the correctly placed lines,
there is no other evidence of scribal error. Perce takes two "honggerie hosborge
scollers" under his wing and offers to show them how to "geet meat with oute
monie" (504–505):

After another twelve lines the first scholar asks:
The deleted lines (508ff) are more appropriate later, when the ale-house is com-
pared to the scrivener's shop. The deletion is then correct, though part of line 507
and more of line 509 should also have been erased, when Perce's (modernized)
continuation would have probably been:
If this is correct, we may infer that associating "can not you come in to an ale-
house" with "is there any ale-house" led the actor to the wrong speech, when
"be wise and I'll teach you to get meat without money" was his redundant, ad
lib recovery. This significant repair would be unnecessary when dictating to a
scribe, but in performance an actor must extemporize. A theatrical report ac-
commodates the sequence; memorial reconstruction and transcription of autho-
rized copy do not.

This evidence of anticipation is admissible and singularly convincing of me-
morial transmission. But Maguire argues further from a mistaken assumption:

There is no evidence that S changed his mind at 508–509. Reviser A (not S)
deleted these lines, including a part of Perce's transition back to the correct se-
quence. Renwick prints the revisers' additions in bold type but differentiates their
strikeouts only rarely in discussion. Hoppe and Maguire mistakenly assume that
S is responsible for substantive deletions. For example, at line 35, even though
Renwick identifies the interlined "Jade" as reviser A's, Hoppe takes the deletion
as evidence that S corrects an anticipated word:

Maguire discounts Hoppe's inference, though she does not note that the correc-
tion was made by a reviser who probably was unaware of "steed" in the much
later passage. On the same assumption, Hoppe analyzes a deletion at line 423:
the "scribe inserted a speech for Vandermast … which was canceled, probably
immediately after being written" (128). But the deletion is again a later revision.

However, another instance may signify memorial repetition, as Hoppe sug-
gests (126). At lines 308–309 Ferdinand asks of Vandermast to "let magick be
amenes to get me grace of Lovlie Rossaline / and I will mak the partener of
my wellth". The same request is exactly repeated more than 300 lines later (my
verse-lining is indicated):

Maguire suggests that the request at 647 is consistent with the author's charac-
terization of Ferdinand, when "in his insincerity, he resorts to repetition" ("Mis-
diagnosing" 121). Maguire relines the dialogue as if to restore the playwright's
original, but she takes no account of the verse: Ferdinand's "I will what will I"
is left short; adding a second short line ("tut tut my Lord" or "too soon forgot")
is hypermetrical.[40] Intentional repetition does not account for this deficiency or
the insensible reference to a forgotten oath in the text. A better explanation is
that the players failed to follow their script. In the preceding lines, Vandermast
interpolates a vocative and reassures Ferdinand:
Vandermast had earlier (310ff) denied Ferdinand's request "to get me grace of
Rossalin" as impossible; he now reminds the prince that he has not forgotten the
one promise that he did subsequently make to help force her submission:
Vandermast's later reference to this promise at line 643 is probably meant for
Ferdinand to confirm his reward for a scheme; but the actor mistakenly repeats
earlier lines that no longer have any point. Aware of his error, he fails to recover
and turns to his fellow actor for help, who wittily regains momentum by alerting
the audience to the error, marking the irony that an oath was literally forgotten.
Modernized, the lines become:

Renwick suggests "break" for "<t>ak" at line 650, but even on correction
the awkward line appears to be a theatrical modification to fit the circumstance.
If these suggestions are credited, the lines derive from performance.[42] Memorial

reconstruction allows on-the-spot correction, but a theatrical report holds a mir-
ror up to nature. Suggestion that Greene drew Ferdinand's character by these
lines denies evidence of a sixteen-word repetition.

Another Ferdinand repetition may be related to his character: "Daphni tut
wer she but as coy as she …" (91); "ha vandermast Daphnie was ner so [q]
coye …" (293). For Hoppe, it is "hardly conceivable that a competent Eliza-
bethan dramatist would perpetrate … such repetitions" (126); Maguire holds
that to "Ferdinand, love means 'the Daphne simile' as automatically as war to
Hotspur means 'honour'" ("Misdiagnosing" 121). However, the evidence is con-
sistent with an actor's imperfect performance, whose errors paradoxically take
Ferdinand's character beyond the author's or the player's conception. We may
then suggest that Greene was not responsible for all nuances discernable in the
characters of the manuscript. This possibility will be convincing if a totality of ev-
idence leads to the conclusion that John of Bordeaux accurately reports an inexact
performance.

Hoppe correctly argues that speech ascription confusions would have re-
sulted in printed error (124). Maguire observes that S "regularly writes 'Ros-
sacler' for 'Ferdinand' (this mistake also occurs in dialogue). At 697 the [speech
prefix] 'fredrick' should be 'ferdinand'" (Texts 267). Maguire does not address
this evidence further,[43] but contrary to her assertion, the Ferdinand/Rossacler
error does not occur in dialogue, though it is found in mid-line speech ascriptions
(marked by S with brackets): "before [be falce] I falce my fayth vnto my lord)
Rossacler my …" 275.

Hoppe suggests without elaboration that the "best-reported" roles (those of
Frederick, Rossacler, and Ferdinand) help to identify the actor-reporters, "which
may explain why at certain points there is confusion in character-names" (132).
Renwick implausibly suggests that the confusions result from cuts. A careful
analysis of the evidence yields other inferences.

The scribe's Frederick/Ferdinand mistake occurs once in dialogue and twice
in marginal speech prefixes. The origin of this error in the similarity of names is
indicated by reviser A's own added entry and correction at line 1011, "[ffred] ffer-
dynand". There are, nevertheless, other anomalous headings. For example, on
folio 2a, line 110, S has an entry for "Basshaws", and at 115 "Bassha" as a speech
prefix. Renwick's "reviser B" adds a majuscule "B" to subsequent "Bassha(w)"
prefixes (at 124 and 129) to differentiate a pasha who enters unannounced by the
scribe. The letter "B" is also added to the speech prefix "souldier" at line 140,
but not at line 142, though it is not clear how many speaking characters are on

stage. On folio 3a the attendants to the Turkish emperor become "noble" and
"lord" (twice, with late-added "B"s). I believe that all these confusions may result
not from careless transcription or dictation, but from the necessities of speedy
stenography.

An instructive revision appears on folio 2b, where Bacon uses magic to extort
from the Turkish Emperor his sword, robe, and crown, by conjuring a scene
where the Turk's son Selimus is threatened:

The marginal "sound" and "Sellimus" are directions added by a reviser, who
would not have intended to assign to Selimus the threat to Bacon, though the
heading could be confusing to a compositor. However, the lesson in this instance
is that a reviser may identify characters from several manuscript sources, but
a stenographer's identifications depend on the dialogue. In this sequence, the
reporter would have learnt of the "soulder" at line 188. Only at line 219 is he
identified, again in dialogue, as the devil "Astrough". The playwright would not
have withheld this information for more than thirty lines.

In other cases a stenographer will have a choice of names. For example, S
refers to "John" or "Bordeaux" interchangeably. A similarly perceived scribal
option may help to undo the confusion surrounding the speech prefix for "Ferdi-
nand." The play opens, "Enter Emperor, Iohn of Burdiox his wif Rossalin his son
Rossacler Vandermast frier Bacon with attendants". A stenographer unfamiliar
with the play may account for the extant manuscript; his initial stage direc-
tion will have been made up from his notes or from what he has independently
learned of the play. When Ferdinand enters at line 71, the scribe knows that
he is Frederick's son, but identifies him as "Rossaclere". The error may have
seemed consistent with the list reproduced in the first stage direction if S mistook
"Emperor" as antecedent to "his son Rossacler". However, the dialogue names
Ferdinand at line 75, and his lines in the scene are all properly ascribed before
his exit at line 103.

Ferdinand enters at line 255 again as "rossaclere", which name remains the
prefix throughout the scene—even when "Ferdinand" is finally identified by
name at line 310. However, when "yong rossaclere" enters with Ferdinand at
lines 408–09, neither character is misnamed again. Rossacler is identified in the
dialogue at line 487, and he first speaks at line 491. While Ferdinand is silent
until line 450, one must ask how a stenographer could have differentiated the
names in the massed entry:

Renwick suggests (perhaps by inference from excessive room in the paren-
theses enclosing speech ascriptions) that speakers' names "may often have been
added after the text was written" (xiii). When the scribe discovered whom the
"yong" actor personated, he would then have retroactively differentiated the
characters in stage directions and speech headings beginning with their entry
at 409. This sequencing could hardly be authorial and normal transcription is
obviously not to blame. A scribe creating his own headings and learning as he
goes explains this crux. The stenographer, whose speech ascription was a speedy
symbol indicating a change of speaker (represented by the scribe's ubiquitous
right-hand parenthesis), would add names while transcribing, and he would re-
duce the possibility of error by noting the context of the dialogue.[44]

This practice accounts for the play's midline "headings." Further, the same
symbol near the end of a transcribed line would signal a line-break to accom-
modate an ascription in the left margin, as is especially indicated by numerous
down-turnings:

The expedient is also indicated at line 167, where the Turkish emperor Amurath
continues to speak after a parenthesis:[45]
A pause in the Turk's speech occasioned the sign for a change of speaker. How-
ever, the dialogue resolves the error, and the transcribed parenthesis is superflu-
ous. The alternative possibility, that "Amewrothe)" was in the authorial copy-text,
is denied by the fact that no other instance of a lone parenthesis appears in the
text after a name. Yet a number of such marks occur when extended pauses are
plausible.[46] The inference is that the stenographer wrote first and asked ques-
tions later.

After rejecting evidence of erroneous memorial repetition by actors in John
of Bordeaux, Maguire discounts the evidence Renwick and Hoppe cite for dicta-
tion in "numerous examples of unusual, phonetic spellings," which, she suggests,
"can be explained by a scribe 'sounding' words in his head" ("Misdiagnosing"

123). To accept this argument is not merely to concur that S was free to spell
as he pleased—which was then true to some extent for all English-writers—but
also to accept that a scribe's spelling flagrantly disregarded his copy-text. The
inefficiency can be obvious in combination:

Had the play been printed, confusions such as the ambiguities of "loks"
would have led to corruption. However, the phonetic spelling is generally read-
able; "grattewlat", "meadesen", and "anenstrewment" are clear enough. For this
reason, the common-sense assertion that no one would enlist so unlikely a scribe
might be denied, despite the large number of truly strange spellings.

The case is different for the Latin orthography; anyone familiar with its rules
would have no need or desire to spell phonetically. An unlearned scribe would
rely on his copy, not on the "sound" of written words he could not in any case
pronounce. Renwick does assert that the scribe was uneducated (xii), and it is
impossible to disagree. Still, an English-only penman could apply his skills to
Latin in dictation, though the quality of his transcription would suffer without
grammatical, syntactical, and semantic understanding. For example, Ferdinand's
citation at lines 81–82 is erased by reviser A, perhaps in recognition of its cor-
ruption: "… [o well sayd Ovid in his wanton] / [wright venus it vinis egnes et
egna fuet]". Maguire must assume that the scribe read "Venus in uinis ignis in
igne fuit" but "sounded" the words to himself before transcribing unacceptably.
Lines copied so badly in normal circumstances would force subsequent transcrip-
tion to proceed by one or two words at a time, at most. Yet even when evidence
of currente calamo correction of short scribal errors suggests that S transcribed at a
slow pace, the Latin passages in John of Bordeaux are spelled only by sound:

At folio 12a, the magician Bacon addresses a recalcitrant spirit:

Renwick identifies "Seleno" as (possibly) Apollo, brother of Hecate-Silene; "eri-
nis" as Erynnis; and "Cleno" as Celaeno (xiv); he then emends lines 1149–51:

Of this "hopelessly corrupt" invocation of underworld spirits, Renwick suggests
that a "scribe so ignorant of Latin as to be capable of these errors might be ex-
pected to transcribe more carefully" (xiii). Cellini attests to the difficulty of the lines
by taking "Seleno" to be Celaeno, but denies that identity to "Cleno" (170–171).[47]

The Latin of lines 1147–51 is reminiscent of the corrupt Italian in the bad
quarto of another Robert Greene play, where the source in Ariosto corresponds
well to Alleyn's Orlando manuscript player's part:[48]

Although the first line is revised, the transcription by the theatrical scribe is ac-
curate in comparison to the quarto of 1594:

Questioning the assumption that Q copy was itself transcribed, Greg reacts
to Orlando as Renwick does to the Latin in John of Bordeaux:

Anyone may prove by personal experience the necessity of transcribing for-
eign words singly (or partially) to minimize error. Greg's inference of a phonetic
transmission must be right, even though he envisioned memorial reconstruction
instead of shorthand, and evidence in John of Bordeaux leads Renwick to the same
conclusion. Further, Greg describes a ten-line Latin passage in Orlando (1275–84)
comprising "a superficial layer of evident errors" above a "comparatively sound"
text with "a few deep-seated errors, some evidently corruptions." He also infers
"that the passage was first corrupted and then subjected to editorial correction"
before printing (Abridgements, 239). So might have been described the Latin of a
printed John of Bordeaux.

Renwick also observes the scribe's "persistent errors which point to a per-
sonal idiosyncrasy: 'quoine', 'quill', and 'quoy'; and, still more, the voicing of
initial 'p' to 'b' nine or ten times, of which seven were corrected currente calamo"
(xii–xiii). Why does S confuse b and p? By "voicing", Renwick seems to suggest
that dictation is a more likely explanation of the failure to distinguish these let-
ters than misreading or "head-sounding." The supralinear minuscule "b" would
not often be confused with an infralinear "p." Although printing is not in the
equation, the suggestion of substitution by "sounding" must take into account
the immediacy of scribal transmission, which affords less time than typesetting
for such mental errors to germinate. Though similarly voiced letters could be
misheard in dictation, shorthand reporting allows another explanation.

One symbol used for two like-sounding letters aids rapid writing. In the p/b
case, the time used to decide upon which is correct would always be left to the
context taking shape in the longhand transcription. In most instances where S
miswrites these letters, at least momentary confusion is possible. On writing "an
aged man porlie atiered in robes of penurie", "ropes" may have been the first
word to come to mind (when "robes" might not be suggested by "poor attire").
Reading further, S will have seen by "penurie" that "p" must be overwritten by
"b". This instance shows the error was not confined to the initial letter or to a
mistaken "b", despite the predominance of those mistakes. At 796, "bade" would
have been supposed before correction to "payd". Similarly, several confusions of
"but" and "put" may be in part caused by temporary semantic ambiguity:

In each case, "put" was required, yet "p" was written over "b" only at line 389;
the mistake can stem from confusing "I put" with "aye, but." At line 347, "p" is
written over "b" twice:
Neither "burse" for "purse" nor "boot" for "pot" seems a likely semantic confu-
sion, but each is consistent with phonetic transmission. Transcription from writ-
ten copy cannot credibly account for these numerous cases.

Using identical symbols for "k", "c", and "q" is also a stenographic likeli-
hood. The scribe usually let stand his use of "qu" for initial "c"; but at line 293
S wrote "q" before erasing it and continuing "coye". S spelt effectively by sound
with little to indicate the influence of a formal education. His habits in John of
Bordeaux include other possibly vestigial indications of rapid writing.

S interchanged the letters u and v somewhat in accord with a familiar system.
When a word called for either letter initially, v was used. S followed this conven-
tion with few exceptions: "vpon"; "velian"; "veri". When either letter was medial,
u was the norm, as in Chettle's added lines: "couert"; "releeues"; "pouertie";
"thou"; "disguise". Medial u for v was not universal, yet most writers complied
even though readers had to determine from context the letter intended. Despite
the good sense of allowing each letter to stand for itself, those who did so were
ahead of their time. S was among this group, though not because he had consid-
ered the matter thoroughly. His medial u is occasionally "v", when the context
is then necessary to know that "u" is meant. Reader familiarity will have been
in judging the other way round: "Coniver" (conjure); "inivrius" (injurious). The
reversals derive perhaps from the scribe's determination to write medial v as "v",
a practice for which there are very few exceptions. For example, 11 a has "dis-
cover"; "prove"; "beleve"; "resoulve"; and "hevines". If the scribe transcribed
notes utilizing necessarily different stenographs for the sounds of u and v, the func-
tionality of "v" for v in longhand would have been apparent. Such advancement
will derive from practicalities for the self-taught stenographer, rather than from
overcoming the inertia of an illogical custom.

A phonetic speller has perhaps two advantages as a note-taker; he will not
be tied to habits unrelated to representations of sound, and he will more quickly
realize and learn shortcuts. The play has many spellings consistent with these
possibilities. For example, the sounded but non-essential end-letter t is often
missing: "waigh" = wait; "kep" = kept; "suplian" = suppliant; "temp" = tempt;
and "spigh" = spight. Most of these words S spelt correctly elsewhere. A scribe
working from good copy has no reason to omit a letter that belongs, when a
stenographer will at times transcribe a word as he has recorded it.

S indulged a common habit in separating an initial syllable from the rest of
a word. But the extent of his usage indicates a time-saving utility: "in treted";
"in chaunt"; "in cappable"; "in senced"; "in swes". If a separate symbol were
used for the word "in" (most likely a sign for the letter n), the expert could more
quickly move to other syllables. The same holds for "a" and "b": "a gaynst";
"a monste"; "be case"; "be wrayes". The manuscript word "be" is mistakenly
used for "been", which could also be represented by the single letter. The scribe
may have used a symbol for a quickly to represent the sound, and later to influ-
ence the spelling of a phrase: "a nasse"; "a nanggree". The usage was extended
to other initial elements: "my narme"; "thy nei"; "my nier". For each of these
cases a letter or common word is used for which there could be a ready symbol,
while leaving the "n" sound for the following syllable. Reduction of the number
of symbols is desirable; but unambiguous use of one symbol for added purposes
can only be helpful.

The definite article will rate its own sign in such a system, but S misuses "the"
for "thee", "then", and "they"; "thy" may have been represented by the same

sign, as it is used in error for "thee". Similarly, "a" stands mistakenly for "as"
and "at", and colloquially for "he". These words are suited to instant writing and
contextually assisted transcription that on occasion will be imperfect. The majus-
cule "I" is in one instance used mistakenly for "it"; in another telling instance,
"I" is stricken and followed by "it". If each of these words was represented by a
symbol for i, they would be differentiated by context. Such errors are not likely
to occur in transcription of written copy or in dictation to longhand, when the
rest of the text exhibits many spellings longer than necessary.

The foregoing analyses lead to a number of conclusions. First, scribal errors
indicate that the text of John of Bordeaux is a transcript that cannot result from
dictation only. Second, phonetic orthography throughout the text (most nota-
bly in Latin passages) nevertheless indicates a stage of oral transmission. Third,
mislineation and metrical corruption argue against transcription of authorized
or descendent copy. Fourth, distant anticipation and repetition (with following
dialogue serving to recover from these errors) point to memorial transmission
without the capability of retraction. Fifth, confusion of speech headings and
character identifications in early scenes are explained by a reporter's unfamil-
iarity with the text, but not by transcription of authorized copy. Sixth, sparse
set directions, including mid-line speech headings, indicate a rapidity of writing
inconsistent with transcription of authorized text. Seventh, scribal habits point
to a phonetically based, highly idiosyncratic speed-writing. Taking the inferences
together leads me to conclude that John of Bordeaux is the initial transcription of
shorthand notes of a performance.

Numerous anomalies in the playtext do not rise to the level of corroboration
for a shorthand hypothesis, but this proposal does accommodate evidence that
is difficult to fit to alternative theories. For instance, when Vandermast tries to
coax his own wife (supposing she is Rossalin) to accept Ferdinand's advances,
the addition at 688 of "young and"—anticipating line 690—disturbs both the
meter and the balance of the argument. The mistake, very easy in performance,
is probably not scribal:

Recognition of the likely source of the manuscript also calls into question
the assumptions and rationalizations characterizing prior scholarly opinion. For
example, Renwick asserts that the manuscript "was made for a company already
familiar with the play" (viii). However, there is no way to know whom S ad-
dresses at 447 by "Exent Bacon to bringe in the showes as you knowe". Renwick
assumes an awareness in the recipient players, but a reminder of the placement
of a dumb show may have been addressed to an intermediary (Chettle?). The
evidence suggests that reviser A was not familiar with the play, whose corrections
are often misguided:

The reviser interlined "and" to replace the blotted word before "exiled", but
the sense and meter suggest "For Rossalin her husband is exiled." Ultimately,
many features of the text, if inessential to the case for shorthand reporting, are
invaluable to our understanding of the theater as evidence of a play captured in
its performance and in its purloining.

Maguire ends "by examining the alternative explanations for the manu-
script … which show that New Bibliographical eagerness for memorial recon-
struction conceals potentially more interesting and more important scenarios
for the history of Elizabethan theater" ("Misdiagnosing" 115). Her own alter-
native—that the play is transcribed from authorized copy—is argued on the
mistaken assumption of scribal correction and an assertion that the scribe spelled
phonetically by repeatedly sounding the words of his copy to himself. I maintain
instead, for the convergent reasons listed above, that the much "more interesting
and important scenario"—and the reality—is a shorthand report.

The essential proposition is an astounding mental and physical activity of
two hours' duration. Achieved (and preserved) once, we may presume it was
accomplished before and after by the same practitioner and probably by others
systematically adept and motivated. "Suspect texts" inevitably come to mind. For
example, in one of the notices taken of John of Bordeaux, Sidney Thomas cites
the speech added at line 1090 in his argument that Henry Chettle contributed
to a bad quarto:

Addressing the same Romeo and Juliet quarto, John Jowett remarks:

The last-mentioned publication is the same Fruitful Sermon discussed earlier
that exhibits literary additions written or approved by the printer and author
Henry Chettle.[51] Thomas and Jowett approach the hypothesis that texts related

by common features and publishers are theatrical reports. New study is required
of the player's part and the printed text of Orlando Furioso—and of all other bad
quartos, including Q1 Romeo and Juliet. But the hypothesis is supported by the
text of John of Bordeaux, which is explained by the accurate verse, the close pho-
netic rendering, and the recording of on-stage errors, as a remarkably well done
stenographic report.

These observations deny the historical intuitions that early English short-
hand was unworkable and that Renaissance dramatic performances were ideal
reproductions of authorial works. They also suggest that persons such as Henry
Chettle were employers—and self-employed—in the innovative acquisition of
plays for the use of rival companies and for publication. Acceptance of John of
Bordeaux as the record of a performance as it occurred—just as sermons were
recorded live, and not as literature—is an inference that (multiplied by large
numbers of corrupt texts) has profound implications for the study of Renaissance
Theater, and especially of bad quartos.

However, performances closely adhering to authored dialogue are not pre-
cluded by the existence of bad quartos. Rather, extended effort is necessary to
determine whether texts not so categorized are good performances well reported.
As shown by revisions and corrections begun on John of Bordeaux, successive
transcriptions and printing of reports will more or less carefully remove evidence
of memorial transmission and stenography. Playtexts may survive in forms not
easily betraying extraordinary histories. Because John of Bordeaux suggests that
bad quartos originated from shorthand, textual scholarship should be alert to
evidence of reporting that survives in other texts. Theatrical reporting may ac-
count for a wide range of printed plays, but memorial reconstruction should be
recognized as a viable explanation for only a few special cases.