Full text of "Latin manuscripts : an elementary introduction to the use of critical editions for high school and college classes"

This book belongs to
THE CAMPBELL COLLECTION

purchased with the aid of
The MacDonald-Stewart Foundatior

and
The Canada Council

it. II, BLACK WELL LTD.

HOOKSKLLEBS
, r ," and ,.i, HHOAD 8TBKBT
OXFORD

mi 0501 f miQVmiAAl [VGHAIU U5tf 01CUS
flJMf M W5GVmiAmiUUJlMUUNUJUsl DO" .

JlC!UAI-5AlVOAlQVUr :
I
If Jk5VN teOM AT 1-1 MEON f M U U.0055AAV

iSMMt i 10&COM I HAJll MVjLHi

AVICUAIS.QHNOVO.1 1 iXBLXSi N 10A
NOa U IVtS AV!U\55I liVUtNOCU

i IdVlMAlM Jl05t(i M^NUJMVAl 1 N 1-5LGM 15'
t aVi Gi UKI I It U00VII\tt5i N5HCAI ACVIO

;AiaifcrimittA5i

rAVI OUICISA\V51 1 VU1CAN.OJMCOQV1 1 VAIOMM

srVAUIAfNll:

^AVACOION o.

I. VKRGII.: Palatinu

INTER-COLLEGIATE

LATIN SERIES

UNDER THE EDITORIAL SUPERVISION OF

HAROLD W. JOHNSTON, PH. D.

PROFESSOR OF LATIN IN THE UNIVERSITY OF INDIANA

LATIN MANUSCRIPTS

HAROLD W. JOHNSTON

CHICAGO

SCOTT, FORESMAN & COMPANY
1897

LATIN MANUSCRIPTS

AN ELEMENTARY INTRODUCTION TO THE
USE OF CRITICAL EDITIONS FOR

HIGH SCHOOL AND COLLEGE CLASSES

HAROLD W. JOHNSTON, PH. D.

PROFESSOR OF LATIN IN THE UNIVERSITY OF INDIANA

CHICAGO

SCOTT, FORESMAN & COMPANY
1897

SCOTT, FORESMAN & Co.,

CHICAGO, ILL.

PRESS OF ROGERS .M SMITH CO., CHICAGO.

EDWARD B. CLAPP,

PROFESSOR OF GREEK

IN THE UNIVERSITY OF CALIFORNIA.

PREFACE.

"pXURING the last session of the Summer School of Indiana University
"^"^ I gave a course of lectures to the Teachers' Class on Paleography,
Hermeneutics and Criticism. My attention was then called to the fact
that even in secondary schools many questions relating to Paleography
and Criticism are asked by pupils who find different texts of the same
author used in the same class. Some of their text books, too, go so far
as to give and discuss various readings of difficult passages, as does Green-
ough's Caesar, for example. A wish was therefore expressed by several
teachers of Latin that a manual might be published for the use of High
Schools, answering the more common questions of this sort. In response
to their wish I have prepared this volume. It gives a mere outline of
the subjects of which it treats in broad strokes, but contains, I hope, all
that students in High Schools and in the lower classes of Colleges will
need in order to understand the critical notes found in the text books
commonly used by these classes. For University use it should be supple-
mented by lectures upon the several authors of the sort admirably illus-
trated by Mr. W. M. Lindsay's Introduction to Latin Textual Emendation,
Based on the Text of Plant us (New York, 1896).

The elementary nature of this manual excludes references to authorities,
but I must mention some of the most important which were used in the
preparation of the lectures from which these chapters are condensed. On
ancient books the standard work is Birt's Das antike Buchwcsen (Berlin,
1882). On the book trade in antiquity there are Haenny's Schriftstellcr
und Buchhandler im altcn Rom (Leipzig, 1885), and (to be used cautiously)
Putnam's Authors and their Public in Ancient Times (New York, 1894). On
Paleography Thompson's Handbook of Greek and Latin Palaeography (New

PREFACE.

York, 1893) is the best modern work; to supplement it the best collection
of fac-similes of Latin manuscripts is, perhaps, Chatelain's Paragraphic des
Classiqucs Latins (Paris, 1884, fol.). On Criticism there is a valuable article
by Friedrich Blass in I wan Mullers Handbuch (Vol. I, Munich, 1892).. For
the use of young students teachers will find good material for parallel
reading in Gow's Companion to School Classics (New York, 1888), from which
I have drawn several -paragraphs, and in the Dictionaries of Antiquities,
under the words c/iarta, codex, liber, papyrus, volumen, etc.

The illustrations are from the works mentioned above, and from
Schreiber's Atlas and Baumeister's Denkmaler.

The plates are from Chatelain, except that of the Codex Romanus of
Catullus, which was furnished by its discoverer, Professor William Gardner
Hale, of the University of Chicago.

Besides owing to Professor Hale the privilege of first publishing a
fac-simile of a page of the most important Latin manuscript discovered
in many years, I am under obligations to Professor Edouard Baillot and
Mr. Charles H. Beeson, of this University, and to Dr. Edward Capps, of
the University of Chicago, for assistance generously given me.

H. W. JOHNSTON.
INDIANA UNIVERSITY,
Feb. 5, 1897.

TABLE OF CONTENTS.

I. THE HISTORY OF THE MANUSCRIPTS, 1-89.

THE MAKING OF THE MANUSCRIPTS, 1-24 .

Writing Materials, 2. Paper and Vellum, 3. Papyrus, 4-13. Pens
and Ink, g 5. Rolls, 6-8. Reading the Rolls, 9. Size of the Rolls,
lo-n. Preservation of the Rolls, 12. Parchment, 13-24. Instru-
ments for Writing, 15. Books (Codices), 16, 17. Odd Forms, 18.
Size of the Books, 19. Parchment vs. Papyrus, 20, 21. Tardy Use of
Parchment, 22. Age of Parchment Books, 24.

THE PUBLICATION AND DISTRIBUTION OF BOOKS, 25-41 . 27

The Authors, g 25. Copyright, 26, 27. Plays, 28. Uncommercial
Publications, 29, 30. Commercial Publications, 31. Process of Publi-
cation, 32. Dictation, 33- Rapidity of Publication, 34. Cost of the
Books, 35. Stichometry, 36. Correctors, 37-39. Titles, 40, 41.

THE TRANSMISSION OF THE BOOKS, 42-67 35

Period Covered, S 42. Period of the Decline, 43. Public Libraries, 44.
Schools and Universities, 45. The Classics, 46. Scholia, 47, 48.
Glosses, 49. The Grammarians, 50. Opposition to Christianity, 51.
Subscriptions, 52. Their Value, 53. Summary, 54. Lost Works,
55. The Dark Ages, 56. Indifference to Learning, 57. The Church,
58, 59. The Revival of Learning, 60, 61. Invention of Printing, 62.
Summary, 63. Editiones Principes, 64. Ancient Manuscripts, 65-67.

THE KEEPING OF THE MANUSCRIPTS, 68-89 48

Care of the Manuscripts, 68. Naming of the Manuscripts, 69, 70.
Descriptions, 71. Important Libraries, 72-80. Index to Collections,
SS 78-80. Symbols for the Manuscripts, 81-83. First and Second
Hands, 84, 85. Collation of the Manuscripts, 86-87. Uncollated
Manuscripts, 88. Critical Editions, 89.

II. THE SCIENCE OF PALEOGRAPHY, 90-149.

STYLES OF WRITING, 91-115 61

Scope of the Science, 91, 92. Uses of Paleography, 93, 94. Ancient
Forms of Letters, 95, 96. National Hands, 97. The Majuscules,

10 TABLE OF CONTENTS.

SS 98-107. Capitals, 100. Specimens : Square, 102. Rustic, 103.
Uncials and Half-uncials, 104, 105. Specimens : Uncials, 106. Half-
uncials, 107. The Minuscules, 108-114. Specimens, 113. Abbre-
viations, 114. Summary, 115.

THE ERRORS OF THE SCRIBES, 116-149 79

The Codex, 116-118. Faulty Copies, 119. The Classification of
Errors, 120. Unavoidable Errors, 121-124. Intentional Errors,
125-127. Accidental Errors, g 128-146. Errors of the Eye, 129-
136. Dittography, 133. Lipography (Haplography), 134. Skipping,
135-136. Errors of the Memory, 137-141. Transposition, 138,
139. Substitution, 140. Omissions and Additions, 141. Errors of
the Judgment, 142-146. Wrong Division of Words, 143, 144. Wrong
Punctuation, 145. Interpolation, 146. Uncertain Sources of Errors,
I47-I49-

III. THE SCIENCE OF CRITICISM, 150-208.

METHODS AND TERMINOLOGY OF CRITICISM, 150-160 95

Subdivisions of the Science, 151. The Critical Doubt, 152, 153.
Causes of Doubt, 154-158. Kinds of Criticism, 159. Criterion, 160.

TEXTUAL CRITICISM, 161-193 . . . 100

Apparatus Criticus, 162. The Manuscripts, 163. Examination of the
Manuscripts, 164, 165. Possible Results, 166-169. Stemmata,
170-172. Uses of the Stemmata, 172 Ancient Translations, 173-
174. Ancient Commentaries, 175, 176. Citations, 177. Imitations,
178. Use of the Apparatus, 179-187. Relative Worth of Manuscripts,
183-185. Test of Worth, 186, 187. Conjectural Emendation, 188-
193. Criticism and Conjecture, 190, 191. Limits of Emendation, 192.
Opposing Views, 194.

INDIVIDUAL CRITICISM, 194-208 .... "4

Purpose of Individual Criticism, 195. External Evidence : Manuscripts,
196. Ancient Writers, 197, 198. Internal Evidence: Historical,
199. Individuality, 200. Language and Style, 201, 202. For-
geries, 203. Tests of Proposed Authors, 204. Illustration of Proof,
205, 206. Summary, 207, 208.

DESCRIPTION OF PLATES, 209-224 125

INDEX . ! 3 2

THE HISTORY OF THE MANUSCRIPTS.

THE MAKING OF THE MANUSCRIPTS.

THE PUBLICATION AND DISTRIBUTION OF BOOKS.

THE TRANSMISSION OF THE BOOKS.

THE KEEPING OF THE MANUSCRIPTS.

THE HISTORY OF THE MANUSCRIPTS.

THE MAKING OF THE MANUSCRIPTS.

1V/TANUSCRIPTS and books were formerly studied as a part of
Paleography, and were so treated by scholars until very
recent times. At the present time separate treatment is given to
this subject, although even now it may scarcely be regarded as a
distinct branch, or discipline, of Philology. Under this head we
have to consider the materials for writing, so far as these have to
do with works of formal literature, the manufacture, distribution
and sale of books, their destruction and preservation in the dark
ages, and their present condition and keeping.

WRITING MATERIALS. We are concerned now with those mate-
rials only, by the aid of which the literature of classical antiquity,
chiefly Roman, was published to the world and afterwards trans-
mitted to us. Almost all the substances for receiving writing
known to the ancients were used at one time or another, for one
purpose or another, by the Romans. Some of these were merely
the makeshifts of rude antiquity and antedated all real literature,
as, e. g., bark and leaves of trees, skins or tanned hides of animals,
and pieces of linen cloth: all these are mentioned in works of lit-
erature, but none were used to receive them. Others, such as stone,
metal tablets, coins, etc., have preserved inscriptions of great im-
portance to the study of antiquity and therefore of great interest
to philologists, but belong rather to Epigraphy and Numismatics
than to our present subject. Of more general use than any of
these were the tablets covered with wax, which are mentioned so
frequently by Cicero, and were used as late as the fourteenth cen-
tury ; even these are excluded, however, by our definition, as they
were at best used for merely the rough drafts of literary composi-
tions. For the publication of works of literature in classical times

14 LATIN MANUSCRIPTS.

the one recognized material was Papyrus, and for their further
transmission to our times Parchment alone need be considered.

3 PAPER AND VELLUM. While parchment (vellum) was known
to the classical writers, and perhaps used to a limited extent

instead of the bulky tablets, and while papyrus (paper) was occa-
sionally used for works of literature until the seventh century and
for correspondence until the thirteenth century, their general rela-
tion to each other is correctly given above : papyrus was the stand-
ard commercial material at the time when the classics were written,
and the tough parchment, upon which these works were copied
centuries after their authors had passed away, has preserved these
works to us, and is the material of the manuscripts with which
modern scholars work. To Caesar and Cicero, for example, a
parchment book would have been as great a curiosity as are to us
the papyrus rolls that have lived through the centuries. This
distinction is of great importance to the further study of this
subject.

4 PAPYRUS. The manufacture of papyrus from the reed of the
same name, which was known to the Egyptians from very ancient
times, reached its height in that country under the earlier Ptole-
mies (third century B. C.), and was improved and perfected in
Rome. Ennius (239-170 B. C.) is the earliest Roman writer to
mention the material and is supposed to have been the first to use
it for literary purposes. The papyrus reed has a jointed stem of
triangular shape, five or six inches in diameter, and grows to a
height of six or eight feet. The paper (char to) was made of the
pith by a process substantially as follows: Strips of the pith as
long as the joints would permit were cut as thin as possible and
arranged side by side, as closely as possible, upon a board. Across
these at right angles other strips were laid in the same manner,
with perhaps, a coating of paste or gum between the two layers.
The strips were then thoroughly soaked in water, and pressed or

THE MAKING OF THE MANUSCRIPTS. 15

hammered into a substance not unlike our paper. After this sub-
stance had been dried, and bleached in the sun to a yellowish
color, the sheets were rid by scraping of any irregular or rough
places that remained, and were trimmed into uniform sizes depend-
ing, of course, upon the length of the strips of pith which com-

."'. : -^ ~~~.*Z.'i'.''f*Ss'

FIG. I. PAPYRUS PLANTS.

posed them. According to Pliny (23-79, A. D.) the quality of the
sheets, which were sold under eight or nine special names, varied
with their width. Sheets of the best quality were about ten inches
wide, while the inferior sorts decreased to a width of six inches
or less. The height of the sheets varied from seven and a half
inches to twelve or thirteen.

PENS AND INK. Only the upper surface of the sheet was com-
monly written upon, the surface, that is, formed by the horizontal
layer of strips, and these, showing even after the process of manu-

LATIN MANUSCRIPTS.

PENS, PEN-CASE AND CRAYON HOLDERS

VARIOUS WRITING MATERIALS FROM WALL
FIG. 2. INSTRUMENTS USED IN WRITING

THE MAKING OF THE MANUSCRIPTS. 17

facture, served to guide the pen of the writer. The pen (cala-
mus or calamus scriptorius) was made of a reed, and was shaped
to a coarse point and cleft with a knife much as our quill-pens
used to be. Quill-pens are first mentioned by Isidorus (f 636 A.
D.), a bishop of Seville, and cannot have been known to the
classic writers. Metal pens, of one piece with the holders, were
also used in ancient times, but cannot be accurately dated. The
ink (atramentuni) for papyrus was made of soot mixed with glue
and thinned with water or vinegar. It was more like paint than
ink, and was easily removed when fresh with a damp sponge
which the writer kept by him for the correction of mistakes.
Even when the ink had become dry and hard it could be washed
(not scraped) away sufficiently to fit the sheet for use a second
time. A sheet thus used a second time was called a palimpsest
(cf. liber palimpsestus below) , but its use was a mark of poverty or
niggardliness (Cic. Fam. VII, 18). Of course the reverse side of
charttz, which had served their purpose, was often used for scratch
paper, as old letters and envelopes are used to-day, and rare
instances are known of the original writing covering both sides
of the sheet.

BOOKS. A single sheet of papyrus might serve for a very 6
brief document, such as a short letter, but for literary purposes
many such sheets would be necessary. These were not fastened
side by side into a book, as are the separate sheets in our books,
or numbered and placed loosely together, as we arrange them in
our letters or manuscripts. The papyrus book was really a roll as
its Latin name (volumen] implies, made up of the necessary num-
ber of sheets glued together at the sides (not at the tops), with
the lines upon each sheet running parallel with the length of the
roll, and with each sheet forming a column perpendicular to the
length of the roll. It was necessary, therefore, to leave on the
side of the sheet as it was written a broad margin, and these

LATIN MANUSCRIPTS.

margins overlapping each other and glued together made a thick
blank space (i. <?., a double thickness of papyrus) between the
7 columns. When the sheets had been securely glued together in
their proper order, a thin slip of wood was glued to the left edge,
or margin, of the first sheet, and a second like slip (umbilicus)
was attached in the same way to the right edge of the last sheet,
much as a wall map is mounted at the present time. The volu-
men was then rolled tightly around the wood attached to the
last sheet, the top and bottom (frontes} of the roll were trimmed
smoothly and polished with pumice stone, and the roll was rubbed
with cedar oil to protect it from worms and moths. For purposes
of ornament the frontes were sometimes painted black, and knobs,
often painted or gilded, were added to the umbilicus upon which
the volume was rolled, or the umbilicus itself was made long
enough to project beyond the frontes and was carved at its
extremities into horns (cornud). Even illustrations were not
unknown; at least a portrait of the author sometimes graced the
first page of the roll, and it is barely possible that the portraits
found in late manuscripts may be copies of these and entitled to
8 more respect than is usually paid them. To the top of the roll,
that is, to the top of one of the sheets (probably the last), was
attached a slip of parchment (titulus} upon which was written the
title of the work with the name of the author. For
each roll a parchment case was made, cylindrical in
form, into which the roll was slipped from the top,
and above which the titulus was visible. If a work
FIG. 3 . CASK FOR was divided into several volumes (see below) the rolls

ROLLS (CAPSA). .. i_ n / /- \

were put together in bundles (fasces) in a cylindrical
wooden box (capsa or scrinium] with a cover, like a modern hat
box, in such a way that the tituli were visible when the cover
was removed.

FIG. 4. READING A ROLL.

THE MAKING OF THE MANUSCRIPTS.

READING THE ROLLS. When a volume was consulted the roll 9

was held in both hands and unrolled column by column with the

right hand, while the left rolled up

upon the other slip of wood the part

that was already read. When the reader

had finished, it was customary to roll the

volume tightly upon the umbilicus by

holding the roll beneath the chin and

turning with both hands. In the case

of a long roll this turning backwards

and forwards must have required much

time and patience, and at the same time

must have sadly worn the roll itself.

These considerations bring us naturally to the size of the rolls.

SIZE OF THE ROLLS. Theoretically there was no necessary 10
limit to the number of sheets that could be glued together, and
consequently none to the size, or length, of the roll: all depended
upon the taste or caprice of the writer. We should suppose that
the author would naturally take as many sheets as were necessary
to contain his work and make them into one roll, and this was
undoubtedly the early custom. So we find that in ancient Egypt
rolls were put together of more than one hundred and fifty feet
in length, that in Greece the complete works of Homer and Thu-
cydides were written upon single rolls (that for Thucydides accord-
ing to careful calculation must have been fully two hundred and
forty feet long), and that in Rome the Odyssey of L/ivius Andro-
nicus (third century B. C.) was originally contained in one roll.
Such rolls were found in the course of time to be inconvenient
to read and liable to break and tear from their own bulk. The H
Alexandrian scholars (about the third century B. C.) were the first
to devise a better plan, and introduced the fashion of dividing
literary works of considerable length into two or more parts, or

LATIN MANUSCRIPTS.

'books," each of which was written upon a separate roll. So
sensible a plan was sure to be followed in time by authors gen-
erally, but its adoption was compelled, or at least hastened, by an
innovation on the part of the manufacturers of papyrus, who
began to sell their product not in single sheets, but in ready-
made rolls of convenient lengths. These rolls varied in length
according to the style of compositions for which they were
intended: rolls intended, e. g., for poems and collections of
letters were shorter than those intended to receive historical and
scientific works. Of the former the roll would receive about one
thousand lines, of the latter about twice as much. Authors had
now to adapt their works more or less to an arbitrary standard,
sometimes perhaps to the detriment of the quality of their writings
(Martial I, 16), and some ancient works were divided for republi-
cation into "books" which had not been so divided by their
authors, e. g., Herodotus, Thucydides and Xenophon among the
Greeks and Naevius (Suet. De Gram. 2) among the Romans.
12 PRESERVATION OF THE ROLLS. The number of papyrus rolls
preserved to us is quite considerable, although none of them con-
tain any complete Latin work of importance and most of them
are in a badly damaged and fragmentary condition. There are
large collections, owned by the state, in London, Paris, Berlin,
Naples and Vienna. Most of them came from Egypt, but many
were found in 1752 in the ruins of Herculaneum so badly burned
that they were taken at first for charcoal and have not yet been
fully deciphered. Of all that are preserved to us the oldest is at
Paris, and was written fully twenty- five hundred years before
Christ, while the most important perhaps is one containing a copy
of Aristotle's Constitution of Athens, a work which had been totally
lost for over a thousand years. This roll came into the posses-
sion of the British Museum in 1890, and contained the accounts
of a farm bailiff, or steward, in Egypt, rendered in the reign of

THE MAKING OF THE MANUSCRIPTS. 21

Vespasian, 78-79 A. D. On the back of this worthless document
some unknown scholar had written, or caused to be written, a
copy of this work of Aristotle for his own use. This recovery
of a lost classic of such traditional fame is one of the most
notable events of the sort of the nineteenth century, and gives
new hope of regaining from the tombs of Egypt other works
of Greek and Roman writers, which scholars have given up as
lost forever. Should this hope be realized parchment may have
to yield to papyrus its claim to the honor of preserving to us the
literature of classical antiquity ( 3).

PARCHMENT OR VELLUM. It has been remarked above (2) 13
that the use of skins or hides to receive writing was not unknown
to the Romans before the dawn of literature: we are told by
Dionysius (f 7 B. C.) that the treaty between Tarquinius Super-
bus and the people of Gabii (Livy I, 54) was written upon a
leather covered shield. The revival of this ancient material after
papyrus had been introduced was due to an improvement in the
treatment of the skins which made it possible to write on both
sides of them. Pliny (23-79 A - D.) asserts upon the authority of
Varro (116-28 B. C.) that this improvement was made in the
reign of Eumenes II (197-159 B. C.), King of Pergamum in Asia
Minor, and was due to the rivalry between the libraries at Alex-
andria and Pergamum. The King of Egypt, he says, tried to 14
embarrass the rival library by forbidding the exportation of papy-
rus, and the scholars of Pergamum were driven to invent a sub-
stitute. The story is untrue, but shows that in Varro's time
Pergamum was noted for its parchment (membrand] and explains
the name by which the material came to be known in much later
times, pergamena, from which our own word parchment (see Web-
ster) is derived. Parchment was known to the Romans at an
earlier date even than Varro's story would imply, but was used
merely for temporary purposes side by side with the wax-tablets,

22 LATIN MANUSCRIPTS.

because the form (see below) was more convenient than the papy-
rus roll, and the writing could be easily and repeatedly erased.
15 INSTRUMENTS FOR WRITING. - - The parchment, unlike the
papyrus ( 5), had to be ruled to insure straight lines. For this
purpose the position of the lines was marked
with a pair of dividers by means of punctures
on both sides of the page, and the lines were
drawn with the aid of a ruler and a bodkin
(stilus}. Sufficient pressure was put upon the
DIVIDERS FROM stihis to cause the line to show through upon
the reverse side (where it would be raised above
the surface), and to save the trouble of repeated measurements
and rulings several sheets were often laid one upon another and
all ruled at once. The pen was the same as for papyrus, but
the smoother surface of the parchment made it possible to use
a sharper point, and as a result to make finer strokes and get
more letters into a line. The ink for papyrus was not suitable
for parchment, and recourse was had to gallnuts, which contain
tannin, and are still used for making inks and dyes. Vitriol was
added in later times and heat applied (encaustum, whence the
Italian inchiostro, French enque, encre, and English ink}. Various
colors were manufactured, of which black was used for ordinary
purposes, and for ornament red and gold. The parchment tituli
( 8) for papyrus rolls were in red.

16 PARCHMENT BOOKS. As the parchment could be written upon

on both sides, the sheets were put together as are the sheets of
paper in modern books. This form resembled that of the wooden
tablets covered with wax, and hence the parchment book received
the same name, codex (originally, "a block of wood"). The
sheets were of various sizes, but the most common dimensions
were such as to give a page of what we now call quarto size,
being about as wide as long. As the flesh side of the parchment

b*>.?ceco}

- " -<4

i COMlO Ci JUAOn AlAr^lp.

*. *v r * s,<fct

.\coeALjetnscet|(\y -1'

II. ClCERO: Schedae Vaticanac, Sacc. // -/".

THE MAKING OF THE MANUSCRIPTS. 23

was almost white and the hair side a light yellow, care was used
in arranging the sheets. Ordinarily the book was made up of
quires of eight leaves (sixteen pages), composed of four folded
sheets. The first of the four sheets was laid with the flesh side

c, o a B. . o; o? o ' * a ' a' o

FiG. 4. BOOKCASE AND WRITING MATERIALS

down, upon it the second with the hair side down, the third as
the first and the fourth as the second. These were then folded
down the middle, and the quire was ready for ruling as ex-
plained above. When a quire was arranged in this way the
colors of every two adjacent pages would be the same, no matter

24 LATIN MANUSCRIPTS.

where the book was opened, and the loss of a sheet would
17 be at once detected by a difference in the color. The sheets
were sometimes arranged in quires of three, five and even
sheets The quires composing a book were lettered consecutively
to assist their arrangement in the proper order, and sometime.
the pages of the several quires were numbered.
was done after the quire was put together, vertical lines being
ruled upon the page to keep the horizontal lines of the same
length and to insure a uniform margin. The writing sometimes
ran across the full page, exclusive of these margins, but was moi
frequently arranged in narrow columns, usually two to the pa
but sometimes three or even four. When the work was finished
the quires were stitched or glued together, and the book
formed if intended to be preserved, was protected by a covering
of the same material, not unlike our own flexible bindings.

considered even a passing fashion.

19 SIZE OF THE CoDEX.-The parchment was so thin and
that a single codex could contain the complete works of an author
or even of several authors, that in papyrus form had to be divide
into several rolls: all of Vergil, ,. & made a codex of very con
venient size, and Catullus is commonly joined with some other

author or authors.
,0 PARCHMENT vs. PAPYRUS.-The superiority of parchment

20 papyrus is obvious: it was more durable and d,d uot become
frayed at the edges; both sides were available; more words

be written in a line of the same length; works of large compass

could be comprised within a codex of moderate size;

could be read more easily and consulted more couvemently, ,th

THE MAKING OF THE MANUSCRIPTS. 25

no time to be lost in rolling it up and restoring it to its cover;
besides, as it would lie open of itself, the hands of the reader
were free to copy from one codex to another, if he pleased, with-
out assistance.

Despite these numerous and manifest advantages, parchment 21
was slow to supersede papyrus. In classical times it was used
merely for accounts, notes, letters, etc. Martial (40-102 A. D.) is
the first to mention parchment copies of works of literature, and
even his words (XIV, 184, 186, 188, 190, 192) are not decisive
in the opinion of certain scholars. The fates seem to have
decreed that papyrus should be the perishable material for pagan
literature, and parchment reserved for the Christian world. We
find, as a matter of fact, that Bibles were early written in the codex
form, and that the works of bishops and saints were soon spread
upon the same material. The great law books, following upon
the compilations of Theodosius and Justinian, demanded a more
convenient form than the volumcti, and seem to have been pub-
lished from the first as codices. The law and the gospel! Next
came what we call the great classics, that is, the choicest works
of Greek and Roman literature, and from the third century of our
era parchment was the favorite for current publications. By the
seventh century papyrus had practically retired from the field ( 3).

TARDY USE. The slowness of parchment to supplant papy- 22
rus is not satisfactorily accounted for by the natural conser-
vatism of the Romans. It can be explained, perhaps, by sup-
posing that parchment was much more expensive than papyrus,
but no proof can be adduced to support this supposition. In
fact, what little we know of the relative price of the two sub-
stances seems to indicate that papyrus was more expensive
than parchment. The real reason is yet to be discovered.
PALIMPSESTS. The word palimpsest has been explained already 23
( 5). It has also been remarked ( 14) that parchment was used
at first for note books and memoranda because the sheet could be
cleaned easily and used repeatedly by washing off the writing

2 6 LATIN MANUSCRIPTS.

when it had served its purpose. This statement is true only of
the inferior ink employed in earlier times. As the ink was
gradually improved in course of time ( 15), it became almost
indelible, especially when fixed by age, and even rubbing and
scraping, to say nothing of washing, failed to remove all traces
of the earlier writing. In such cases the second copy was some-
times written between the lines of the older copy, and both writ-
ings may now be read under favorable circumstances. This fact
is of great importance to scholars, as will be explained hereafter.
A book thus rewritten is called liber palimpsestus or codex

rescript**.

24 AGE OF PARCHMENT BOOKS. From the history of the intro-
duction of parchment ( 13) it will be understood that the oldest
parchment books (codices] which we possess are of a very late date
as compared with the papyrus rolls (volumina) which are still
extant ( 12). Our very oldest codices do not go back beyond
the third or fourth century of our era, and very few are older
than the ninth. This will be considered more fully hereafter.

Ut *

** v -. -;> ^iU^IS.-^*~.,Oc2"i:

S-o c s ~ 3-g &S3-&JG^ I ?

ne '? S3 : s**i > *

^^^ 5H-V -^Ci^<J r

^?f

.ZJ <-, r

>J -5

~1 ^ *- .*

y < ** 1 *5 * f** ^^ ^T

^ _: 3- ~ '*> ^-* **

\J ^^ "*-- *<* ^^ * * ~ ^*f

3'.<: ?2. 5 JS. t.-.^ :*

^ < u-2 ^ ^ ^

^sS^-sw^ -^^.*

?5^J : Ssf!

,-> - "S r *^^ 5 ^.S- j?*

^ t . ^^f%SS
s- gSs^f>

*$ %~3#S

- - 5>f^-5^

<z o

if-Mj- 5 ft
?---> ~ ^"Sca^-^^^**

- -: < *== ;? 5 ^ *-i ^ ^'

*-*? * 1 ''vS ^* ri ^
M^Se

w
o
x
w

THE PUBLICATION AND DISTRIBUTION OF BOOKS.

AUTHORS. The men whose names are famous in the 25
history of Roman literature may be divided into two classes.
Some were men of high position in society and in the state, to
whom literature was but one form of a many sided activity : such
men are Caesar and Cicero and Sallust. Others are persons of
distinctly inferior station, freedmen perhaps, or sons of freedmen,
who won their bread by their pens: such persons are Terence
and Vergil and Horace. One fact in regard to the authors of the
second class forces itself at once upon our attention: Each is
attached to some powerful friend, to whom he seems to owe all
his material prosperity. This fact is the more striking, because
the works of many authors of this class, of all of those whom we
have directly mentioned, were widely read during their lifetime,
and must have had a ready sale and considerable market value
even then. We should expect such poets as Horace and Vergil
to have had a generous income independent of the bounty of their
patrons. It seems to have been otherwise.

COPYRIGHT. The natural inference is that the author had, 26
little pecuniary interest in the sale of his works. There is no
direct evidence, i. e., no statement in the works of such authors,
to support this assertion, but there is none to controvert it. As
each copy of a book was made by itself, page by page, with pen
and ink, as no costly plant was necessary to multiply these copies,
and no special skill, it is hard to see how the author could retain
any control over the reproduction of a work when it had once got
into circulation. Even in our day any one may make a manu-
script copy, or any number of them, of any book which he is
unable to buy, whether the author likes it or not. This seems

28 LATIN MANUSCRIPTS.

to have been the case in Rome, and this state of helplessness
fully accounts for the dependence of the poet upon the patron,
and the absence of any feeling of shame or degradation, on the

27 part of the dependent. The first copy of his book he could sell,
or as many copies as he could make, or have made, before any
left his possession, but these would at best be very few. That
even this chance, poor as it was, was precarious is shown by the
theft of Cicero's De Finibus (Att. XIII, 21, 4 and 5) in advance
of publication. Worse than this, the hapless author had not even
the privilege of deciding whether a book that he had written
should be published or not: at least Ovid declares (Trist. I, 7)
that he had intended to destroy his Metamorphoses, but the work
was published from copies taken by his friends without his con-
sent or knowledge. Cicero let the first draft of his Academica get
out of his possession while he was considering a different form for
the treatise, and the consequence was that two very different ver-
sions were circulated at the same time.

28 PLAYS. The fact that a dramatist received pay when one
of his plays was presented at the public games has nothing to
do with the question of property rights in works of general
literature. As a matter of fact the attacks made upon Terence
by rival dramatists show that they were acquainted with his
plays before they were put upon the stage, and justify the
suggestion that they may have been in more or less general
circulation for the purpose of private reading.

29 UNCOMMERCIAL PUBLICATION. Every Roman of position kept
in his employ several trained scribes (librarii'], usually slaves or
freedmen and often highly educated and accomplished, who served
him as amanuenses, secretaries, etc. Under the Republic the
author must have had his book copied by these librarii, either
his own or his patron's. Many of these copies would be intended
for dedication or presentation purposes, but some would find their
way into the market. These were sold in book shops (taberntz
librarifc. Cic. Phil. II, 9, 21), which were set up in Rome long

THE PUBLICATION AND DISTRIBUTION OF BOOKS. 29

before there was any organized publishing business. The first
impulse toward such an enterprise may have been given by the
bringing to Rome by Sulla and Lucullus of whole libraries from
Greece and Asia Minor. It at once became the fashion to make
large collections of books, and in Cicero's time no house was com-
plete without a spacious library fully stocked with books, although
the owner was often wholly ignorant of their contents. Cicero
had great numbers of books not only in his house at Rome, but
also at each of his half dozen country-seats. He was assisted in
collecting them by his friend T. Pomponius Atticus, a man noted as
much for his love of literature and learning as for his vast wealth 30
and far reaching business enterprises. He seems to have had a
commission from Cicero to buy for him every book that could be
bought, and to make copies of those that were valuable or rare.
Atticus had numerous librarii (Nepos XXV, 13, 3), and these he
employed also in making copies of Cicero's works and of such
others as Cicero recommended to him. All these he sold to good
advantage (Att. XIII, 12, 2), but the gain was merely incidental
and by no means the object he had in view. His success, how-
ever, added to the constantly increasing demand for books, seems
to have led to the establishment of the business upon a commer-
cial basis, and in so far as this is true it is permissible, perhaps,
to speak of Atticus as the first of Roman publishers.

COMMERCIAL PUBLICATION. Under the Empire the business 31
seems to have reached large proportions almost at a stride. The
publishers were at the same time wholesale and retail dealers in
books. Their establishments were found in the most popular and
generally frequented parts of Rome, were distinguished by the
lists hanging by the door of books kept for sale, and soon became
the resort of men of culture as well as of those who sought merely
after the novel and the entertaining. Even under Augustus (29
B. C.-I4 A. D.) the works of Roman authors were read not only

30 LATIN MANUSCRIPTS.

in Italy but also in the provinces, and even crossed the sea.
Public libraries were established in many places, and in the schools
the antiquated works that had been the text books for generations
(e. g., the Twelve Tables and the translation of Homer by
Andronicns) began to give place to those of contemporary authors.

32 PROCESS OF PUBLICATION. It is evident that the publisher had
no more control over works once in circulation than the author
had ( 26), and he must therefore have relied upon the elegance,
correctness and cheapness of his editions of the classics to insure
their sale, and in the case of a new work upon the quickness
with which he could supply the demand. The general process
was something like this: The book to be copied, furnished by
the author if a new work, bought or borrowed or hired (see below)
if an old one, was read to the scribes, some of which were the
slaves of the publisher and others perhaps hired for the occasion,
but all trained copyists. Other slaves arranged the sheets in the
proper order as fast as they were written, pasted them together
(Cic. Att. IV, 4 b.), mounted them and supplied them with their
parchment tituli and cases (see 7). Errors were then corrected
and the book was ready for sale.

33 DICTATION. No ancient authority can be quoted in sup-
port of the statement that the books were copied from dicta-
tion, but this must have been the case in all large establish-
ments. To say nothing of the fact that even private letters
were usually dictated, and of the difficulty of managing the
roll, which served for copy, while writing ( 20), the slowness
of the other method, if but few slaves were employed, and the
impracticability of furnishing copy to a large number without
great loss of time, seem enough to justify the statement. In
later times, especially during the middle ages, the scribes
worked independently.

34 RAPIDITY OF PUBLICATION. Cicero tells us (Pro. Sulla, XIV,
42) that Roman senators could write fast enough to take down
evidence verbatim, and the trained scribes must have far surpassed
them in speed, even if the system of shorthand often mentioned

THE PUBLICATION AND DISTRIBUTION OF BOOKS. 31

by ancient authorities was not used for books intended for gen-
eral circulation. Martial tells us (IT, i, 5) that his second book
could be copied in an hour. It contains ninety-three epigrams
amounting to five hundred and forty verses, which would make
the scribe equal to nine verses to the minute. It is evident that
a small edition, one, that is, not many times larger than the num-
ber of scribes employed, could be put upon the market much
more quickly than it could be furnished now. When the demand
was great and the edition large (Pliny, Ep. IV, 7, 2, mentions
one of a thousand copies) the publisher would put none on sale
until all were ready, thus preventing rival houses from using one
of his books as copy. If he overestimated the demand, unsold
copies could still be sent to the provinces (Hor. Ep. I, 20, 13) or
as a last resort be used for wrapping paper (Mart. Ill, 2).

COST OF THE BOOKS. The cost of the books varied, of course, 35
with their size and with the style in which they were issued.
Martial's first book, containing eight hundred and twenty lines
and covering twenty-nine pages in Teubner's text, was sold
(Mart. I, 66; 117, 17) at thirty cents, fifty cents, and one dollar;
his Xenia, containing two hundred and seventy-four verses and
covering fourteen pages in Teubner's text, was sold (XIII, 3) at
twenty cents, but cost the publisher less than ten. Such prices
are hardly more than we pay now. Much would depend of course
upon the demand, and very high prices were put upon particular
copies. Gellius (II, 3, 5) mentions a copy of Vergil, supposed to
be by his own hand, which had cost the owner over one hundred
dollars, and copies whose correctness (see below) was attested by
some good authority were also highly valued. The same circum-
stances would increase the price of modern books materially.

STICHOMETRY. The ancients did not measure their books, as 36
we do, by the pages, but by the verse in poetry and the line in
prose, and the number contained in the work was written at the

32 LATIN MANUSCRIPTS.

end of the book. The Alexandrian librarians seem to have entered
the number along with the title of the work in their catalogues,
and to have marked the number of lines, at every fiftieth or hun-
dredth line, in their copy of the book. This system of measure-
ment was carefully employed by the publishers, and furnished an
accurate standard by which to fix the price of the book and the
wages of those scribes who were not slaves. For this purpose
they selected the hexameter verse as the unit for poetry, and as
its equivalent in prose a line of sixteen syllables or thirty-five
letters. This standard line, were it actually written, would require
one of the broader sheets mentioned above ( 4), but such a sheet
was not necessary and perhaps not usual. It was merely neces-
sary to find the ratio of the line actually written to the stand-
ard line, for the scribes were careful to keep their lines of the
same length, and the number of lines on the page constant,
throughout the work upon which they were engaged. Frequently
we find the number of lines written very much greater than the
number registered at the end of the roll, because the page was
too narrow to contain the standard line from which the registered
number was calculated. We do not know the price paid for
ordinary works of literature.

37 CORRECTORS. The very rapidity with which the scribes worked
would lead us to look for many mistakes in their copies, and from
the earliest times authors and scholars have complained of their
blunders. Cicero says (Q. Fr. Ill, 5, 6) that he knows not where
to turn for books: they are written so badly and put upon the
market with so many imperfections. He took every precaution to
have his own books as free from errors as possible. His famous
freedman, Tiro, read the copy carefully before it was sent to Atti-
cus, and Atticus had each book examined and corrected before it
passed out of his keeping. Even after the earlier copies were sent
out he introduced improvements in the later editions at Cicero's
suggestion.

THE PUBLICATION AND DISTRIBUTION OF BOOKS. 33

Similar precautions were taken by at least the best commercial 38
houses. They had competent correctors in their employ, but as
each copy had to be examined independently, the labor was far
greater than that of the modern proof-reader, and the results
much less satisfactory. Martial (II, 8) warns his readers that the
errors which they may detect in his books are to be ascribed to
the publisher, not to him, and elsewhere (VII, u) he gives us to
understand that authors corrected with their own hands the copies
which they presented to their friends (cf. Cell. II, 3, 5). Quin-
tilian prefaces his Institutions with a letter to his publisher, beg-
ging him to issue the work as free from blunders as he can, and
Irenaeus, bishop of Lyons, 177 A. D., urges that each copy of his
work be compared with the original.

Persons buying books sometimes had them examined first by 39
a competent critic (Gell. V, 4, 2), or corrected by comparison with
a copy known to be accurate. Such standard copies were not
always to be had, but were consulted if possible to decide disputed
readings (Gell. I, 7), and were sometimes hired for this purpose
(Gell. XVIII, 5, n) at large expense. It is beyond question that
errors in the codices of later times, which have descended to us,
are in some cases derived from blunders made at the time when
the books were first published.

TITLES. As in the papyrus roll the title was no part of the 40
work itself, but rather of the mounting ( 8), so in the later
parchment codex it was the ancient custom to write the title,
together with the number of the lines ( 36), at the end, instead
of at the beginning where we should look for it. This must be
explained, of course, from the standpoint of the scribe, who was
concerned only with what he had written and how much, and left
the purchaser to mark the volume or leave it unmarked at his
pleasure. The manuscripts of the middle ages usually have the
title both at the beginning and at the end of the book, frequently

34 LATIN MANUSCRIPTS.

adding a word of good omen (fe/ia'ter), or an expression of grati-
fication at the conclusion of the task (see Plate VIII). These
titles vary greatly in different manuscripts of the same work,
sometimes even in the same manuscript, and suggest that the
classic writers were far less anxious about getting good titles for
their works than modern authors are, and may perhaps have pub-
41 lished them without any formal titles at all. Cicero refers to his
essay on Old Age indifferently as the Cato Maior (Off. I, 42, 151)
and De Senectute (Div. 2, 3). If Macrobius (Sat. I, 24, n) is to
be trusted, Vergil seems to have spoken of the Aeneid by its
hero's name Aeneas (cf. Hamlet, Ivanhoe, etc.). Sallust's mono-
graph on the Conspiracy of Catiline is called in the best manu-
script Bellum Catulinarium at the beginning and Bellum Catilinae
at the end. Quintilian (35-100 A. D.) calls it Bellum Catilinae,
and so does Nonius (beginning of fourth cent.); Servius (end of
fourth cent.) has the shorter title Catilina (cf. Aeneas and Cato
Maior above), Priscian (sixth cent.) has Bellum Catilinarium, and
in other ancient authorities we find Historia Catilinae. The best
form nowadays is Bellum Catilinae, which is rapidly driving out
the De Coniuratione Catilinae Liber of our school books, just as
Belli Gallici Liber I. (//., ///., etc.) is displacing the Commentarius
De Bella Gallico Primus (Sccundus, Tertius, etc.) familiar to us
all. No title is absolutely certain.

rf^^

, ' *

'<^^

optnriio' pert&r2ce< t?n *k\y\ jfWtnomk; du^

ran
o

^Li!

fup
tLy TuUu& cet^us coi

tjersJt^^fyuri^cjt^

I\' C.USAR : /'arisiiiH.f

THE TRANSMISSION OF THE BOOKS.

PERIOD COVERED. The creative genius of the Ro- 42
mans ends, so far as literature is concerned, with the reign
of Trajan (97-117 A. D.). From this time until the invention of
printing the preparing and publication of books did not vary from
the methods described above, except so far as the parchment codex
differed in form from the papyrus roll. During this period of
about thirteen centuries we have now to consider the fates of the
published works, or in other words of the manuscripts that con-
tained them: the means that were taken to preserve them, how
they were lost, and then after nearly a thousand years partially
recovered. This period may be naturally divided into three very
unequal portions: i. The Period of the Decline, extending roughly
to the Germanic invasions of about the fifth century; 2. The
Dark Ages, extending to about the thirteenth century; 3. The
Revival of Learning. It must be remembered that we are con-
cerned with the social, political and literary history of these times
so far only as it relates to the Transmission of the Manuscripts.

THE PERIOD OF THE DECLINE. It is a fact well known to all 43
students of literature that at the time when genius is least pro-
ductive and originality most torpid the masterpieces of an earlier
day will be most carefully studied and appreciated. This is emi-
nently true of Roman literature: its darkest period saw the estab-
lishment of public libraries, the growth of schools and universities
on humanistic lines, the rise of the grammarians, and the classics
made the last defense of paganism against Christianity. All these
agencies made for the preservation of literature, so far as it was
preserved at all, and must be examined therefore in some detail.

36 LATIN MANUSCRIPTS.

44 PUBLIC LIBRARIES. The growth of private libraries ( 29)
steadily increased during the empire, for we read that the gram-
marian Serenus Sammonicus (f 212 A. D.) left 62,000 volumes to
his son, but the largest of these collections are of little impor-
tance compared with the public libraries that were founded during
the same period. The first of these to be opened in Rome was
established by Asinius Pollio (f 4 A - D -) during the reign of
Augustus in the atrium of Libertas. Augustus himself opened
two, and by his successors the number was gradually increased to
twenty-eight. Of these the most magnificent was the Bibliotheca
Ulpia, founded by Trajan. Smaller cities had their libraries too.
Pliny, Trajan's governor of Bithynia, tells us (Ep. I, 8) of having
given one himself to his native town, Comum, supported by an
endowment yielding annually thirty thousand sesterces. The im-
portance, from our standpoint, of these public libraries lies in the
fact that such collections were universal in their character, while
private libraries are usually gathered in a less catholic spirit. The
former would tend to preserve, therefore, the less popular and
attractive works that might otherwise have disappeared.

45 SCHOOLS AND UNIVERSITIES. A still more important part in
the preservation of the literature of the past was taken by the
schools and universities. These had been established on Greek
lines in the city of Rome at least as early as the time of Cicero
and Varro, and had spread throughout the empire until in the
centuries just preceding the Germanic invasions all the intel-
lectual life of the Roman was connected with education. The
branches taught were grammar, rhetoric, logic, arithmetic, geome-
try, astronomy, and music, but the central thing was the study
of the older and greater writers of Greece and Rome. Original
creation had virtually come to an end, and it seemed to all educated
persons that the study of the works of the past was the most
profitable of intellectual pursuits. Two facts in relation to the
schools affect the transmission of the manuscripts.

THE TRANSMISSION OF THE BOOKS. 37

THE "CLASSICS. "--The choosing of materials for pupils to 46
study and imitate would lead gradually to the fixing with more
or less precision of the canon of the classics, those writers, that
is, whose works were regarded as the best of their kind in the
various lines of literature. Of some of these authors the complete
works were used in the schools; of others certain parts complete
in themselves (e. g., the first and third decades of Livy) were
carefully studied, while of other parts epitomes were made for
reference purposes ; of others still selections were made for specific
objects, as when, for example, the letters and speeches scattered
through the various works of Sallust were brought together in
one volume for rhetorical purposes. The result, so far as it affects
the transmission of the manuscripts is apparent: of some authors
the whole works would be in constant demand and copies would
be multiplied almost beyond numbering; of others parts only
would be so treated; still others would be wholly neglected. It is
evident, also, that these school editions would be especially liable
to errors, and even to arbitrary changes for the purposes of
instruction.

SCHOLIA. The needs of the pupils would lead, in the second 47
place, to the preparation of notes and commentaries upon those
authors whose language or matter was found to require such helps.
Such notes are added to the works of English authors in our own
schools now, and must have been even more needed by Roman
school boys because no books were then written especially for the
young. These school commentaries, to distinguish them from the
works of modern scholars, are called scholia, and their authors, or
(more usually) their compilers, are called scholiasts. Some of
these notes were published separately, and have come down to us
with the name of the author attached, as, e. g., the commentaries
of Asconius (first century) on some of Cicero's Orations, of Por-
phyrio (second century) on Horace, of Tiberius Claudius Donatus

38 LATIN MANUSCRIPTS.

(fourth century) on Vergil, and of Aelius Donatus (fourth century)

48 and Eugraphius (sixth century) on Terence. Other Scholiasts,
and by far the larger number, wrote their notes on the margins
and between the lines of their manuscripts of the authors they
explained, and of these as a rule inferior scholars we seldom know
the names. When it is necessary to distinguish them, they are
called by the name of the author ("Scholiast on Juvenal," etc.) or
even of the manuscript ( 69) on which their scholia are found.
These scholia are chiefly valuable for the subject matter of the
author, but they give some help also in the text. In the first
place, those scholiasts whose commentaries were published sepa-
rately, frequently quote the passage of the text which they
explain, and thus give us the reading of the manuscripts they
used, in .most cases older and therefore better than our own. In
the second place, they sometimes help us to fix the date of a
manuscript or its relations to others even when the scholia are of
little value and the name of their author is not known.

49 GLOSSES. One sort of scholia is often mentioned in editions
of the classics. An unusual word was called glossa, and in the
course of time the definition or explanation of such a word was
called by the same name. Collections of these words and ex-
planations were made, called glossac, whence our words "gloss"
and ''glossary." Now when the scholiast found in his text such
a word, for example a foreign or obsolete Latin word, he often
wrote the word of the same meaning which was current in his
time (Latin also, of course) directly above it in the text or
close to it in the margin. A later copyist was very apt to
take such a gloss for either a correction or an omitted word,
and accordingly to omit the original word from his copy, or to
write both words together.

50 THE GRAMMARIANS. Close upon the writing of commentaries
to explain the subject matter of the classics followed the composi-
tion of scholarly works, dealing directly with the language itself,
the sounds, inflections, syntax, prosody, lexicography and so on.
The writers upon these subjects, differing widely in their learn-

THE TRANSMISSION OF THE BOOKS. 39

ing and ability, are grouped together under the name of Gramma-
rians, as opposed to the Scholiasts, although many belong to the
one class as much as to the other. For the preservation of the
classics they are valuable, entirely apart from their scholarship,
in proportion to the number of quotations which they make in
illustration of the matters of which they treat. Among those help-
ful in this way may be mentioned Charisius (fourth century),
Diomedes (sixth century), Macrobius (fifth century), Nonius
(fourth century), Priscianus (sixth century), Scaurus (second cen-
tury), and Victorinus (fourth century).

OPPOSITION TO CHRISTIANITY. It is well known that the 51
higher classes in Rome were the last to embrace Christianity.
For resisting the spread of the new faith they found the most
effective weapon to be the literature in which were embodied all
the beauty and power of pagan morality, culture and refinement.
Men of the highest social standing, senators, statesmen, consuls,
devoted their energy and talent to fostering the ancient classics.
They succeeded in maintaining the old system of education, pre-
vented the establishment of separate schools for the benefit of
their opponents, and even endeavored to put the texts of the great
Roman writers upon a sounder basis. For this purpose they had
made or made with their own bands copies of manuscripts of
known excellence (see 39), or in default of these used their own
knowledge of the language to remove the more obvious errors due
to the carelessness or ignorance of successive copyists. Some of
these editions they attested by their own names, and these names
have occasionally come down to us in later copies.

SUBSCRIPTIONS. These signatures, technically called subscrip- 52
tiones, date mostly from the fourth to the sixth century, although
a few are earlier, and are known to us in copies hundreds of
years later, accompanied perhaps by the subscription of some later
reviser. For example, many manuscripts of Terence, dating from

4O LATIN MANUSCRIPTS.

the ninth to the twelfth century, have preserved an ancient sub-
scription in two forms :

CALLIOPIUS RECENSUI CALLIOPIUS RECENSUIT.
This shows that much as these manuscripts may differ from each
other, all are derived ultimately from a revision of the text of
Terence made by Calliopius, who is otherwise unknown, but is
believed for certain reasons to have lived in the third or fourth
century. Again, several manuscripts of Caesar, dating from the
ninth and eleventh centuries, have the subscription :
JULIUS CELSUS CONSTANTINUS vc LEGI.

We do not know anything more about this man of high position
(vc= vir darissimns, see Harper's Dictionary, s. v. clams), but the
name seems to show that he lived no earlier than the fourth

century.

53 VALUE. We are able to test the value of these revisions,
because we have other manuscripts of Terence and Caesar that
are independently derived. Of Terence we have but one manu-
script (Codex Bembinns, see Plate III) that has escaped the
corrections of Calliopius, but this shows us that he used_ either
inferior manuscripts as his guide, or else relied upon his own
insufficient knowledge in correcting the text current in his
time. With Csesar the case is different. The manuscripts
derived from the revision of Celsus have been, until very re-
cently, regarded almost the only reliable authorities, and even
now Celsus is credited (see Kiibler, Teubuer's text, p. ix) with
having used good copies in making his text, even if he did
rely sometimes too much upon his own guesses.

54 SUMMARY. From the preceding paragraphs it ought to be
evident that in the period of the decline all conditions were favor-
able for the preservation in some form of the manuscripts. The
influence of the schools, however, and the well meant, but not
always successful, efforts of the revisers would lead us to expect
variations in the texts of the more popular authors, and the disap-
pearance of those thought less useful for instruction and less
admirable in style.

C ATI

IX.

- - t

51 B AN

DA5V5I
tN VIslLRIISVSCtN

ici X.NI t\iEi45N T .ira.v.G.i

LtM

VB* 1 SI D I AH O XB ASCES S A1VIJSJ O V At I S

tc.S: E \i r ATI 'it MS n v D v RISC E RI e AM p v \ v

VI IB 1 1 LAVAS I RIS.\A' r XAlO.S IDE RElARR A

"V PUT

INI

RJ1

ITS LV&O R \m.t O AT >
P V-DY ATS:O LAN

V. VERGII, : Schedae Vaticanae, Sacc. // '.

THE TRANSMISSION OF THE BOOKS. 41

LOST WORKS. It is well known that the works of certain 55
Roman authors have been entirely lost, that of others we possess
parts only, that there are few whose writings are wholly preserved.
We should not regret this, if the works of inferior authors only
had been lost but among the missing are some of the most
famous in the lines of history, oratory, philosophy, and poetry.
We should expect it, if the works of early writers only had per-
ished but whole volumes of Cicero, two-thirds of Tacitus, three-
fourths of Livy are gone, to mention those names only that are
as familiar to us as our own. No imperial library could have
lacked complete editions of their works, they must have been
included in hundreds of private collections, school boys must have
studied them, and teachers commented upon them, but they are
no more to be found. We have therefore to explain how so much
has disappeared, and how so much has been preserved.

THE DARK AGES. It was at the very time when Roman lit- 56
erature was the center of all intellectual activity ( 43) that the
catastrophe came that was to overwhelm learning, literature and
even Rome itself. In the fourth century the Roman empire was
divided ; Valentinian took the eastern half with Constantinople for
his capital, leaving Rome and the west to his brother Valens.
The fifth century had only just begun when the hordes of the
north fell upon the western half and made havoc of it. First the
Vandals, turned from Italy, established themselves in Gaul. Then
the Visigoths sacked Rome, passed into Gaul, and drove the Van-
dals into Spain. The Vandals, again, crossed over into Africa,
ravaged that province, and returned to Italy by the south. The
Tartar Huns came next and disappeared leaving desolation behind
them. The Franks attacked Gaul, the Saxons Britain. The Os-
trogoths disputed Italy with the Vandals, and both were dispos-
sessed by the eastern Emperor, Justinian (527-565). He died and
the Lombards appeared. Then the Saracens came from the south

42 LATIN MANUSCRIPTS.

and the Danes from the north. It was not until the time of
Charles the Great (Charlemagne), in the last part of the eighth
century, that order was restored in Western Europe. Cities had
been pillaged, provinces laid waste, empires overturned, a great
civilization overwhelmed, and a literature that antedated the cities,
provinces and empires, and had inspired the civilization, had prac-
tically disappeared.

57 INDIFFERENCE TO LEARNING. The worst, perhaps, was yet to
come. These three centuries of destruction were followed by five
centuries of indifference to learning. It is impossible to give
within our limits an adequate idea of the ignorance of the period :
the ninth chapter of Hallam's Middle Ages cannot be condensed
into a paragraph. During this time Latin ceased to be a spoken
language; inflections were neglected, syntax ignored, sounds modi-
fied, and Spanish, French and Italian began to be. There was not
even an educated class. The nobles could not sign their names:
until seals were brought into use they subscribed to their charters
with the sign of the cross. The ignorance of the church was the
subject of reproach in every council; in one held in 992 it was
asserted that not a single person in Rome knew the first elements
of letters. In the time of Charlemagne not one priest of the thou-
sand in Spain could address a common letter to another. In Eng-
land King Alfred said that he could not remember a single priest
south of the Thames, the most civilized part of his realm, that
knew the meaning of the common prayers. Alfred himself had
difficulty in translating a pastoral letter of Saint Gregory on
account of his ignorance of Latin, the one written language of the
time. Charlemagne could not write at all. If the ignorance of
nobles, priests and kings was so appalling, that of the commons
must have been sublime, and we are ready to find the loss of
Roman literature less surprising than its partial recovery.

THE TRANSMISSION OF THE BOOKS. 43

THE CHURCH. The one preservative agency was the church. 58
In spite of the gross ignorance, the narrow-mindedness, the world-
liness of the priesthood, there were three influences in connection
with the church that made for the preservation of classical litera-
ture. These were the papal supremacy, the liturgy and the mo-
nastic establishments. For our present purpose we may pass over
the first two with the short statement that the liturgy was in
Latin, and that the need of the church of some one language as a
means of communication with its branches everywhere served to
keep alive some faint knowledge of the Latin tongue, corrupted as
it became. The third must be more fully considered. Of the re- 59
ligious orders of Western Europe one of the most ancient was that
founded in 529 on Monte Cassino, near Naples, by Saint Benedict.
Its rule was less severe than that of the others, but it enjoined
upon its members frugality, soberness and above all industry.
From various kinds of manual labor the copying of manuscripts
was finally selected as the most likely to keep the mind from car-
nal thoughts, and so all over Italy, Switzerland, France, England
and Ireland the pious monks laboriously copied and recopied the
manuscripts of Latin authors amid all the destruction of barbaric
invasions, and the poverty of learning that followed. It must be
clearly understood that these manuscripts were not copied for pub-
lication. The work was purely mechanical, a treadmill process.
The completed codices were stored away in the vaults of the abbeys
to molder and decay, until, in later times, when the very knowl-
edge of their meaning was lost, they were brought out to be
washed and scraped and made fit to receive other copies by other
generations of monks. It was from no love of learning, therefore,
that the Benedictines and the allied brethren saved the literature
of Rome, so much of it, that is, as did not rot in cellars and dun-
geons, or was not remorselessly rubbed away to make room for
hymns and homilies and lives of the saints and martyrs. For

44 LATIN MANUSCRIPTS.

such precious compositions as these were the parchments used that
a king's ransom would not now purchase.

60 THE REVIVAL OF LEARNING. It is impossible to give here an
intelligible account of the gradual revival of learning during the
period which we have described above as the Dark Ages. The
history of the five hundred years from 800 to 1300 comprises the
growth of schools, the planting of universities, the cultivation espe-
cially of the more useful sciences of medicine, law and theology.
It was not until the fourteenth century that literature felt the new
movement, and that in Italy. Petrarch (1304-1374) and Boccaccio
(1313-1375) were the first to turn for better models to the almost
forgotten classics of their countrymen of an earlier day, and the
finest minds of the next generations followed their guidance. The
last quarter of the fourteenth century saw all Italy permeated with
the new enthusiasm, and a positive fever was inspired for recov-
ering the lost literature of Rome. Then it was that the stores of
manuscripts buried in the monasteries were eagerly brought to
light. Vast quantities were found at Monte Cassino (see 59),
and at Bobbio in Italy, at St. Gallen and Einsiedeln in Switzer-
land, at Fulda and Mainz in Germany, and in far distant England
even, wherever the copying of manuscripts had been the employ-

61 ment of the monks. Petrarch was especially active in searching
for new treasures and protecting those that were discovered -for
the danger of losing them again was not over in the fourteenth
century. A treatise of Cicero De Gloria had been in his posses-
sion, but was afterwards irretrievably lost. He declares that in his
youth he had seen the works of Varro, but all his efforts to
recover these and the second decade of Livy were fruitless. He
did find in 1350 a copy of Quintilian, the only one known until
sixty-four years later another copy was found in a dungeon under
the monastery of St. Gallen. By this time the awakening had
touched all classes. Princes and popes gathered scholars at their

THE TRANSMISSION OF THE BOOKS. 45

courts as the surest means of obtaining fame for themselves. The
representatives of the popes in other countries sent to Italy all
the classical manuscripts of which they could possess themselves
by fair means or foul. Almost all the Latin manuscripts which
we now have were thus discovered between 1350 and 1450. Many
very ancient manuscripts known at that time have since been lost,
but so many copies were made that, so far as we know, but one
entire work has disappeared, the Vidularia of Plautus.

INVENTION OF PRINTING. The fortunate invention of printing 62
about 1450 made secure what had been recovered. The first Latin
author to be sent abroad in the new form was Cicero, whose De
Omciis was printed in 1465. In less than twenty years from this
time the Venetian printer, Aldus Manutius, had begun his great
work of giving to the world almost the whole body of ancient lit-
erature in the form that has made his name a synonym for taste-
ful and convenient volumes.

SUMMARY. This sketch, short and colorless as it is, helps to 63
explain several important facts, often referred to in critical editions.

1. The largest collections of valuable manuscripts are in Italy.

2. The very oldest manuscripts are likely to be palimpsests.

3. The large majority of our manuscripts were written in the
fourteenth and fifteenth centuries.

4. Many extant manuscripts are copies of an older manuscript,
also extant.

5. Some manuscripts were written by persons with little or no
knowledge of Latin.

6. The printed cditio princcps of certain authors is valuable,
because it may have been derived from good manuscripts since
lost to us.

EDITIONES PRINCIPES. The following list includes the prin- 64
cipal Latin authors: Apuleius, Rome, 1469; Caesar, Rome,
1469; Catullus, Venice, 1472; Cicero, De Omciis, Rome, 1465,
Opera Omnia, 1498; Gellius, Rome, 1469; Horace, Venice, 1470;

46 LATIN MANUSCRIPTS.

Juvenal, Rome and Venice, 1470; Lactantius, Rome, 1465; Livy,
Rome, 1469 ; Lucan, Rome, 1469 ; Lucretius, Brescia, 1473 ;
Martial, Rome, 1470; Nepos, Venice, 1471; Ovid, Rome and
Bonn, 1471; Persius, Rome, 1470; Plautus, Venice, 1472; Pliny
the Younger, Venice, 1485; Propertius, Venice, 1472; Quintilian,
Rome, 1470; Sallust, Venice, 1470; Seneca's Prose Works,
1475, Tragedies, Ferrara, 1484; Statius, Venice, 1472; Sueto-
nius, Rome, 1470; Tacitus, Venice, 1470; Terence, Strassburg,
1470; Tibullus, Venice, 1472; Valerius Flaccus, Bonn, 1474;
Velleius Paterculus, Basle, 1520; Vergil, Rome. 1469.

Arranged chronologically: 1465 Cicero's De Officiis, Lac-
tantius ; 1469 Apuleius, Caesar, Gellius, Livy, Lucan, Vergil ;
1470 Horace, Juvenal, Martial, Persius, Quintilian, Sallust,
Suetonius, Tacitus, Terence; 1471 Nepos, Ovid; 1472 Catul-
lus, Plautus, Propertius, Statius, Tibullus ; 1473 Lucretius ;
1474 Valerius Flaccus ; 1475 Seneca's Prose Works ; 1484
Seneca's Tragedies; 1485 Pliny the Younger; 1498 Cicero's
Opera Omnia; 1520 Velleius Paterculus.

65 ANCIENT MANUSCRIPTS. The following list gives the dates of
all extant Latin manuscripts which are thought to be no later
than the sixth century. As will be explained hereafter (115),
the dates are merely approximate, and any of the older parch-
ments may be later by a century than the date here assigned to
it. It is also possible that some may have been written at an
earlier time. FIRST CENTURY : Two papyrus fragments from Her-
culaneum containing selections from prose writers. A papyrus roll
from Herculaneum containing the Carmen De Bello Actiaco, a
specimen is given in 103. THIRD or FOURTH CENTURY: The
seven oldest manuscripts of Vergil, specimens of three, Plates I,
V and X. Three fragments of Sallust's Histories, at Berlin, Rome
and Naples, a specimen is given in 103. Palimpsest fragment
of Juvenal and Persius at Rome. Palimpsest of Livy at Verona.

66 Fragment of Livy, Book XCI, at Rome. FOURTH or FIFTH CEN-
TURY : Fragments of a palimpsest of Lucan at Vienna, Naples and
Rome. The Codex Bembinus ( 53) of Terence at Rome, for speci-
men see Plate III. The palimpsest of Cicero De Re Publica at

THE TRANSMISSION OF THE BOOKS- 47

Rome, for specimen see 106 and Plate II. Palimpsest of Cicero's
Orations at Turin. Milan and Rome (from Bobbio. see 60 above).
Palimpsest of Cicero's Orations against Verres at Rome. A few
leaves of a palimpsest of Livy at Turin. Palimpsest of Gaius at
Verona. Palimpsest of Merobaudes (first half of the fifth century)
at St. Gallen. Fasti at Verona. FIFTH or SIXTH CENTURY:
Palimpsest of Ulpiaii at Vienna. Palimpsest of Lactantius at St.
Gallen. Vatican fragments of the Jurists, Rome. Palimpsest of
Plautus at Milan (from Bobbio). Fragment, De Jure fisa\ at Ve-
rona. A few leaves of a palimpsest of Hyginus at Rome. Palimp-
sest of Gellius and fragments of Seneca at Rome. Manuscript of
the Grammarian Cledonius (fifth century) at Berne.

It will be noticed that of these twenty-four manuscripts, many 67
of which are badly mutilated, no less than fourteen are palimpsests,
but it must also be noticed that, valuable as these palimpsests are.
none has furnished us with the complete text of any work of any
author. Their testimony is usually decisive for such portions of a
given text as they contain, and, more than this, they often enable
us to select from later, more legible, and complete codices, the one
which is truest to the original.

THE KEEPING OF THE MANUSCRIPTS.

68 QARE OF THE MANUSCRIPTS. The manuscripts recov-

ered as described above remained sometimes the property of
the abbeys in which they were found, but more often passed by
purchase, gift or theft into the possession of individual owners, and
were at all times liable, as articles of ordinary commerce, to be
mutilated, lost, or destroyed. Those that have come down to mod-
ern times receive better treatment. All of any value are kept in
the great libraries of Europe, the property of the universities or
* even of the various states. The rules governing their use vary
with their value and the spirit of the libraries where they are
kept. Some may be taken from the libraries for the purpose of
study, others may be examined freely within the library itself, but
may not be removed from it, others still must be handled only by
an officer of the library, who finds the passage which the student
desires to examine, and reads or shows it to him. In general it
may be said that, when scholars are properly introduced to the
authorities, all reasonable facilities are given them for examining
and comparing even the most valuable manuscripts. The greatest
obstacle is the lack of complete descriptive catalogues to some of
the most interesting and important collections.

69 NAMING OF THE MANUSCRIPTS. Every library has its own sys-
tem of identifying its books and manuscripts by letters or num-
bers, and by these letters or numbers added to the Latin name of
the library or city where they are kept the manuscripts are now
known and described. A manuscript that has passed from library
to library, as almost all have done, has borne of course the special
name and mark of each, and so has been known and described
differently at different times. Besides, many manuscripts were

THE KEEPING OF THE MANUSCRIPTS. 49

used by scholars when they were the property of individuals, and
were then called merely by the names of their owners. It follows,
therefore, that in using editions of an author separated by many
years we may find the name of a given manuscript varying with
the dates of the several editions. Owing to these changes in the 70
name it has sometimes happened that a manuscript has been sup-
posed to be lost which really existed but was disguised by a differ-
ent name, and also that readings from the same manuscript have
been quoted under its several names so as to lead to the belief
that the one manuscript was two or more. Such errors are sure
to be detected in the course of time by the identity of the quoted
readings, but they show how necessary it is to have a full history
and an accurate description of every valuable manuscript.

DESCRIPTIONS. As an example of the brief descriptions given 71
in modern critical editions the following is taken from Kiibler's
edition of Caesar's Gallic War (1893) in Teubner's series: "Codex
Amstelodamensis Si saec. VIIII-X, olini Floriacensis, postea inter
libros Petri Danielis Aurelianensis, deinde Jacobi Bongarsii, inde
Bongarsianus primus dictus." The manuscript is number 81 in
the library of Amsterdam, was written in the ninth or tenth cen-
tury, was previously in the abbey of Fleury-sur-Loire (in France),
afterwards in the private library of Pierre Daniel of Orleans
(born 1530, died 1603), then in that of Jacques Bongars (born
1554, died 1612), and was consequently called Bongarsianus primus.
Critical editions usually add particulars as to the condition of the
manuscript, the size and number of pages, its style of writing, the
errors that occur most frequently, etc., etc. Examples are given
in connection with the plates. These descriptions are often hard
reading, because names of modern places and even persons are
Latinized, and these names are not given in our dictionaries. Some
help in interpreting these names is given in the following para-
graphs, but completeness is not attempted.

5 LATIN MANUSCRIPTS.

72 IMPORTANT LIBRARIES. The libraries with collections of clas-
sical manuscripts are too numerous to be described here, but the
most important are named in the following list in alphabetical
order by countries. For further information see the article Li-
braries in the Encyclopaedia Britannica, from which ' this is con-
densed. There are no Latin manuscripts of any value in the
United States.

AUSTRIA: The Imperial Library at Vienna (Vindobona),
founded in the fifteenth century, contains 500,000 volumes and
20,000 manuscripts (codices Vindobonemes) . There are besides good
manuscripts in some of the monastic establishments, e. g., at
Saltzburg (Salisburgum, codices Salisburgenses). The University
Library of Prague contains 200,000 volumes with 3,800 manu-
scripts (c. Pragenses}.

BELGIUM: The libraries of the universities at Ghent (Ganda-
vjini) and at Liege (Leodicuni) have together over 3,000 manu-
scripts (c. Gandavenses and Leodicenses). The Royal Library at
Brussels (Bruxellae] contains 30,000 manuscripts (c. Bruxellenses).
73 DENMARK: The Royal Library at Copenhagen (Haimia),
founded in the sixteenth century, has 500,000 volumes and nu-
merous manuscripts (c. Haunienses).

ENGLAND: At Cambridge (Cantabrigid) the University Library
has 6,000 manuscripts (c. Cantabrigienscs], with many others of
great value in the library of Trinity College. At Oxford (Oxonia]
the Bodleian Library, founded in 1602 by Sir Thomas Bodley,
contains 30,000 manuscripts (c. Bodleiani, or Oxonienses) and a
valuable collection of first editions (see 64) of Greek and Latin
authors. At London (Londinium] is the library of the British
Museum, one of the largest and most important in the world,
which was founded in 1753 and contains 1,600,000 volumes,
including more than 50,000 manuscripts (c. Britannia' or Londi-
nenses). These manuscripts are often further described by the

m petfmrr-tfr offta^i^uL faekmn^ SsfreuaiC- uliiucb*

a*ferf ^r-ftAr^man LufttfefluLmd*

f tfiquecumpAiictrr-eLiStm tucLtf
tncpnf^rarfimof ^AtTincui^ntr 1 l?

'T' ' p<iu la turn

If " t '

. P fL - J^brmaunncriien -cam _
- Sj 4.m fcr-enuTrfimyT cuii f Autpcct deratf __

amtcualti - VAT-fboCmm

erctcum . ltCLA^m ^-iyr-. lomifa^ ucf t

Til CILIS H BLL\TAV

ix,PLlCiT -" 1 K ClTlT

VI. SALLVST: Parisintis 1602^, Sure. /A" A'.

hf Am fTprftjiv^n cjut fq -, muufj?

wCcoh&^r vr-fier-i adifiecerz

ion

f <ft

wiucn,*, . m OT rT inwn-T.* i ^

/* f * >*J /^^ "t* wL*frt^'* H-X/r* ^*.LA is \^\j * 1^147

dceffe- - S ertd-u^oircj', im pA-zttprri ttr^mfn-rv^UtT Ammufe^t'* auiu-Jit^to^iM -

i -^

v >3 (
iiKa

THE KEEPING OF THE MANUSCRIPTS. 5!

names of previous owners, e. g., codices Townlciani, from the col-
lection of Charles Townley (1737-1805) and codices Harlciani, col-
lected by Robert Harley (1661-1724), Earl of Oxford, and his son. 74

FRANCE: At Paris (Lutetia Parisioruni] the National Library
is the largest library in the world, founded in the fourteenth cen-
tury, containing 100,000 manuscripts (c. Parisini, or Parisiaci}.
Many of these were formerly in the ancient Royal Library (c.
Regii} or less important collections e. g., at St. Germain (c. San-
germanenses), and at Fontainebleau (c. Bliaudifontani}. Some few
good manuscripts still remain in provincial towns, e. g., at Mont-
pellier (c. Montepessulani}.

GERMANY : Almost all the universities have large libraries
containing manuscripts of value. The University of Heidelberg
(Heidelberga), situated in the Palatinate, has over 400,000 vol-
umes and many manuscripts (c. Palatini], and the University of
Strassburg (Argentoratum] has 500,000 volumes and some good
manuscripts (c. Argentoratenses}. At Berlin (Berolinum} the Royal
Library contains 15,000 manuscripts (c. Berolinenses]. At Dresden
(Dresdena] the Royal Library has about 500,000 volumes with
4,000 manuscripts (c. Dresdenses}. At Gotha the Ducal Library has
more than 6,000 manuscripts (c. Gotham}. At Munich (Mona-
chiuni) the Royal Library, founded in the sixteenth century, is the
largest in the empire and contains 30,000 manuscripts (c. Mona-
censes], while the University Library has 1,800 more. The Royal
Public Library at Stuttgart (Stuttgardia) has 3,800 manuscripts.

HOLLAND: At The Hague the Royal Library has 4,000 manu- 75
scripts. At Leyden (Lugduniim Batavorum] are 5,000 manuscripts
(c. Leidcnses or Lugduncnses Batavi}. At Amsterdam (Amsteloda-
mnm] are some very valuable manuscripts (c. Amstelodamenses) in
the library of the university.

ITALY: Of the numerous collections of manuscripts in Italy
(63) only the most noteworthy can be mentioned here. At

52 LATIN MANUSCRIPTS.

Florence is the Laurentian library attached to the church of St.
L/orenzo; it contains some 10,000 manuscripts, chiefly from the
library of San Marco, the collections of the Medici and Leopoldo
families, and the library of John Ashburuham, of England, pur-
chased by the Italian government in 1884 (c. Florentini, Lauren-
tiani, Medicei, S. Mara, Lcopoldini, Ashburnhamii, etc.). The

FIG. 7. VATICAN LIBRARY.

Biblioteca Riccardiana, founded by the Riccardi family and pur-
chased by the government in 1812, contains 3,800 manuscripts (c.
76 Rtccardiani}. At Milan (Mediolanum} the Ambrosian library has
8,000 manuscripts (c. Mediolanenses or Ambrosiani), including some
famous palimpsests. At Naples there are 4,000 manuscripts in the
National library and museum (c. Neapolitani), some from the old

THE KEEPING OF THE MANUSCRIPTS. 53

Bourbon library (c. Borbonici). At Rome is the Vatican Library,
the most famous and magnificent but not the largest in the world,
containing some 23,000 manuscripts (c. Vaticani or Romam).
Among these are most of the manuscripts brought from Bobbio
( 60), 3,500 taken from Heidelberg in 1623 (c. Palatini, see
above), many bequeathed to the library in 1600 by Fulvius Orsini
(Ursinus, c. Ursiniani}, others purchased from Duke Federigo of
Urbino in 1655 (c. Urbinates], from Queen Christina of Sweden by
Pope Alexander VIII (c. Reginenses or Alexandrini}, and from
Cardinal Mai, and many other only less famous collections. The
library is not fully catalogued and its management is far from lib-
eral. Two other libraries, the Biblioteca Cosanatense and the Bib-
liotcca Vittorio Emanuelo, have recently been united and contain 77
more than 6,000 manuscripts, most of them from the important
collections of the Jesuits of the old Collegia Romano. At Turin
(Augusta Taurnwrum] are some good manuscripts (c. Taurinenses)
in the University Library. At Venice (Venetiae) the Marcian
Library, founded in the fifteenth century, contains many valuable
manuscripts (c. Veneti, Marciani, or Veneti Marciani], and there
are others at Verona in the Cathedral Library (c. Veronenses].

SWITZERLAND: There are good libraries with valuable manu-
scripts at Basle (c. Basilienses], at Berne (c. Bernenses}, at Ein-
siedeln (c. Einsidlenses}, at St. Gallen (c. Scmgallenses), and at
Zurich (c. Turicenses).

INDEX TO COLLECTIONS. In the following list are arranged 78
alphabetically the names of manuscripts mentioned above, with
a few others occurring in critical editions of school classics :
Alexandrini (Rome), Ambrosiani (Milan), Amstelodamenses
(Amsterdam), Argentoratenses (Strassburg), Ashbumhamiani
(Florence), Basilienses (Basle), Bembinus (of Cardinal Pietro
Bembo, 1470-1547), Bernenses (Berne), Berolinenses (Berlin),
Blandiniani (Blankenberg, Belgium), B liaudifontani (Fontaine-
bleau), Bodleiani (Oxford), Bongarsianus (\ 71), Borbonici (Na-
ples), Britannia (London), Bruxellenses (Brussels), Budenses

54 LATIN MANUSCRIPTS.

79 (Buda, Hungary), Cantabrigienses (Cambridge), Caroliruhenses
(Carlsruhe), Colbertini (of Jean Baptist Colbert, 1619-1683,
statesman, France), Colonienses (Cologne), Corbciensis (of Cor-
vey, town with monastery, in Germany), Cuiacianus (of Jacques
Cujas, 1522-1590, France), Einsidlenses (Einsiedeln), Florentini
(Florence), Floriacensis ( 71), Fuldenses (Fulda, Germany),
Gudiani (of Marquard Glide, 1619-1700, Germany), Graevianns
(of J. G. Greffe, 1632-1703, Netherlands), Gnelferbytani (Wol-
feiibu'ttel, Germany), Haunienses (Copenhagen), Laurentiani
(Florence), Leidenses (Leyden), Lcopoldini (Florence), Lipsienses
(Leipzig), Londinenses (London), Marciani (Venice), Matritenscs
(Madrid), Medicei (Florence), Mcdiolanenses (Milan), Minorau-

gienses (of Augia Minor, an ancient abbey in Austria), Mona-
censes (Munich), Montepessulani (Montpellier), Moysiacenses (of
the abbey of Moissac, France), Neapolitani (Naples), Ottoboniani
(of Cardinal Pietro Ottoboni, 1668-1740, nephew of Pope Alex-
ander VIII, Vatican, Rome), Oxonienses (Oxford), Palatini
(Heidelberg, Rome), Parisiaci, or better Parisini (Paris), Petro-
politani (St. Petersburg), Pragenses (Prague), Reginenses (Rome),

80 Regii (Paris), Regionwntani (Konigsberg), Riccardiani (Flor-
ence), S. Marti (Florence; to be carefully distinguished from
Marciani above), Sangallenses (St. Gallen), Sangermanenses (St.
Germain), Scaligcranus (of J. C. Scaliger, 1484-1558, or J. J.
Scaliger, 1540-1609), Sorboniani (of the Sorbonne, a depart-
ment of the University of France), Taur incuses (Turin), Thua-
neus (of Jacobus Augustus Thuaneus (De Thou), a statesman
and historian of France, 1553-1617), Toletani (at, Toledo, Spain),
Turicensis (Zurich), Urbinatcs. Ursiniani, Vaticani (Rome),
Veneti and Veneti Marciani (Venice), Veronenses (Verona), Vin-
dobonenses (Vienna), Vossianus (of Isaac Voss, 1618-1689),
Vralislavienses ( Breslau ) .

81 SYMBOLS FOR THE MANUSCRIPTS. In editions of the classics
in which the manuscripts are frequently mentioned, it is custom-
ary for the editors to use in place of the name or names of each
manuscript, often long and unwieldy ( 71), an arbitrary symbol,
usually a letter of the alphabet or a numeral. These symbols
are prefixed to the descriptions of the manuscripts where they are
first given, usually in the introduction or the critical appendix.
For example, to the description quoted above ( 71) Kiibler has

THE KEEPING OF THE MANUSCRIPTS. 55

prefixed the letter A, and by this symbol the given manuscript,
codex Amstelodamensis <?/, is known throughout his edition of the
Gallic War. Scholars may therefore call this manuscript briefly
" Kiibler's A." It happens unfortunately that there is no gener-
ally accepted system in accordance with which these symbols are
selected and used by scholars. Some editors arrange their manu-
scripts in the order of their supposed importance and letter them
A, B, C, etc. Others use for each manuscript the first letter of its
name or of some one of its several names. Others still, using 82
manuscripts quoted in some earlier standard edition, retain the
symbols adopted by the earlier editor, as Kiibler seems to have
done in the case just cited, adding new symbols, of course, for such
manuscripts used by them as the earlier editor did not quote. In
using at the same time different editions of the same author, the
student has, therefore, to be constantly on his guard against con-
founding these symbols. For example, in the three editions of
Caesar's Gallic War by Holder (1882), Kiibler (1893) and Meusel
(1894), the symbols for the six most important manuscripts are
shown in the following table :

NAME OF MANUSCRIPT. Holder. Kiibler. Meusel.

Codex Amstelodamensis, 81 . . . . A A A
Codex Parisinus Latinus, 505*5 . . . M M Q
Codex Parisinus Latinus, 5763 B B B

Codex Romanus Vaticanus^ 3864 . . R R M
Codex Parisinus Latinus, 5764 T T a

Codex Vaticanus, 3324 U U h

It will be seen at once that the three editors agree in the sym- 33
bols of but two manuscripts out of the six, and that, while Holder
and Kiibler may be used together without confusion, great care
must be taken when the readings of Meusel are compared with
those of either of the others. Such changes of the symbols are,
of course, even more confusing when they are made in the same

56 LATIN MANUSCRIPTS.

work. Thus, in the fourth edition of Orelli's Horace (1886-1889)
the codex Ambrosianus is marked O in the first volume (Odes and
Epodes, by Hirschfelder) , while in the second volume (Satires and
Epistles, by Mewes) it is marked a. No reason is given for the
change, except that Mewes adopted the symbols of Keller and
Holder (1864-70).

84 FIRST AND SECOND HANDS. Mention has already been made
( 37) of the correction of errors in ancient manuscripts, and it is
hardly necessary to say that such errors and corrections are no
less frequent in those of later date. Sometimes the scribe himself
discovered his mistake and erased the wrong letters, inserting the
right ones in their place, or wrote the correct reading between the
lines above the blunder, or in the margin, in the last case indi-
cating by dots or other simple marks the place where the correc-
tion was to be made. Sometimes a later reader introduced in the
same way corrections, or at least variations, derived from other
manuscripts or from his own sense of the appropriate. Now, it is
often important to distinguish these corrections from the original
reading and from each other, if they were made by different per-

85 sons. If some of the corrections are seen to be in the same writ-
ing as the text they are said to be by "the first hand;" others
are said to be by the second or third hand, according to their age.
These hands are indicated by the editors in several ways: some-
times by small figures written as indices after the symbol used
for the manuscript, e. g., A 1 , A\ etc.; sometimes when the manu-
script is denoted by a capital letter, e. g., P, the correctors will
be marked p, p\ etc. Here, too, there is a lack of agreement
among editors.

86 COLLATIONS OP THE MANUSCRIPTS. It is no longer necessary,
as it once was, for a scholar engaged upon a given work to travel
all over Europe, from library to library, to examine the scattered
manuscripts of his author. Of all important manuscripts of the

f-f'* '&' J 'si <r

iS3 iTiJttty ^fetfC-

rctutw

r\^ ^ffv*

vii. c / /<&< xi J xi n.

THE KEEPING OF THE MANUSCRIPTS. 57

classics copies have been made, called collations, and published to
the world. These collations are not complete copies of the manu-
script, word by word, much less fac-similes, but give merely the
variations of the given manuscript from some other manuscript,
or better from some printed edition, of the author, which the col-
lator has taken as his standard. For example, take the third sen-
tence in Caesar's Gallic War with I/owe and Swing's text as the
standard: Gallos ab Aqnitanis Garumna fltimcn, a Belgis Matrona
et Sequana dividit. Now, if we wished to publish the reading of
the codex Vaticanus 3324, marked U by Kiibler, a complete copy
would require eleven words. As it happens, however, U differs
but once from the text we have taken as our standard, and U's
reading of the whole sentence is sufficiently indicated by printing
this one word, preceded, of course, by the number of the line in
the standard text, in which the variation is found: 5, gamnna, U.

The saving of time, labor, and expense by these collations, to 87
say nothing of the wear and tear of the manuscripts, is very great,
but over against this advantage must be set the danger of errors,
owing in the first place to slips of the collator, and in the second
place to slips of the printers who reproduce his work. These
errors are being gradually removed by new collations made with
far greater care and skill than were formerly employed. Of some
very valuable manuscripts, however, which have been destroyed or
lost, or which by mutilation or decay have become illegible, editors
have still to depend upon early collations which are known to be
inaccurate and untrustworthy.

Of a very few manuscripts exact reproductions have been made,
either from type or by photography. The latter process may be
depended upon accurately to reproduce the original (see the Plates
in this book) when the ink is not too dim; the former (e. g.,
Merkel's Aeschylus, Studemund's Plautus A) is exposed to the
same risks of error as the less elaborate collation.

58 LATIN MANUSCRIPTS.

88 UNCOLLATED MANUSCRIPTS. It is not to be supposed that
all Latin manuscripts have been collated. The vast majority
have been found after cursory inspection to be copies of older
manuscripts in our possession and are therefore of no value
except as curiosities. It may be that some of them have been
unfairly judged and are deserving of closer study, but not
much is to be hoped for from them. Besides these, some few
manuscripts may yet be discovered in the large collections
which have not been fully or accurately arranged and catalogued
( 76). Thus, Professor Hale discovered in 1896 a manuscript
of Catullus hidden behind a false number in the Vatican library.
For specimen page see Plate XI.

89 CRITICAL EDITIONS. A critical edition gives in the form indi-
cated above the readings of all the manuscripts of the given work
which the editor deems valuable, together with certain other evi-
dences for the original text which will be considered hereafter
( lyS- 1 ? 8 )- S uch editions are usually very elaborate and costly
but of some of the authors read in schools there are critical edi-
tions to be had which represent sound scholarship and yet are
inexpensive. Among these are Kiibler's Csesar, Meusel's Caesar
(the best), Jordan's Sallust, Kloucek's Vergil, and Ribbeck's (1894)
Vergil. There are unfortunately no equally convenient and satis-
factory editions of Nepos, Cicero and Livy.

II.
THE SCIENCE OF PALEOGRAPHY.

STYLES OF WRITING.

THE ERRORS OF THE SCRIBES.

THE SCIENCE OF PALEOGRAPHY.

STYLES OF WRITING.

pALEOGRAPHY treats of ancient methods of writing. It inves- 90

tigates the history of the characters used to represent speech,
traces the changes from age to age in those of the same language,
teaches the art or science of deciphering documents, and deter-
mines their date and place of origin from the style of writing.
Paleography was not recognized as a science until the publication
in 1 68 1 of the De Re Diplomatica of Jean Mabillon (1632-1707),
who gave directions in this work for distinguishing by the writing
itself between the genuine documents of the middle ages and the
forgeries that were current in his time.

SCOPE OF THE SCIENCE. By the definition given above Pale- 91
ography should include the study of writings of every sort, of all
times and all peoples, regardless of the material (2) which re-
ceived them. As a branch of classical philology, however, its scope
has been greatly restricted. In the first place it is limited to the
Greek and Latin languages, and in the second place Epigraphy
and Diplomatics, once mere branches of Paleography, have won for
themselves the rank of independent sciences. The former treats
of the very oldest written records of Greece and Rome, those, that
is, cut in stone and metal, or scratched and painted upon wood or
other hard substances; the latter deals with the charters, wills,
deeds, grants, etc., of the centuries following the breaking up of
the Roman empire. To Paleography is left, therefore, merely the 92
study of the writing, or various styles of writing, found in the
manuscripts of the works of literature that have descended to us in
the manner just described. Limiting our study of Paleography to
Latin manuscripts as we do, the period covered extends from the
fourth century A. D., the time when the oldest codices now exist-

62 LATIN MANUSCRIPTS.

S 65) were written, to the fifteenth century, when the scribe
was succeeded by the printer.

93 USES OF PALEOGRAPHY. From the history already given of
the manuscripts it is evident that, other things being equal, the
older of two manuscripts is the better: the nearer, that is, its date
is to the time of the author, the less the number of transcriptions,
in all probability, between it and the original copy, and the less
the chances for errors by successive copyists. It is therefore of
great importance to scholars to be able to fix the time, even the
century, of the writing of a given manuscript, and this is the first
thing that paleography undertakes to do. Again, if the study of
ancient characters shows that certain letters, for instance, were
very much alike and were often mistaken for each other, it may
be possible for us, when we find in a manuscript a combination of
letters that makes no sense, to reverse the process and discover
the letter or letters that the last copyist ought to have written.
This is the second use of paleography, and the one that is of

94 greatest importance to the ordinary student. As will be shown
below, the dating of manuscripts is largely a matter of practice
and experience, not of rules and regulations, and requires direct
and long continued contact with the manuscripts themselves. It
is a science for experts, and of these there are very few whose
opinions have the force of authority. The correction of errors, on
the other hand, is of great interest in itself, and may be under-
stood and practiced by persons who have never so much as seen
a genuine manuscript. For these reasons the chief stress will be
laid here upon the errors of the scribes, and only so much atten-
tion will be given to the styles of writing as is necessary to
understand the causes of these errors and the methods of correct-
ing them.

95 ANCIENT FORMS OF LETTERS. Two styles of lettering were
known to the Romans at the time when the classics were written.

!HL^KJV01\TlMNAMlNCiTlTLlBC'lV I'MWvS-

tfd VSOve -TANU6 'bpVl&&

\'TTT

/i?/

STYLES OF WRITING. 63

One style, called the Majuscule (litterae maiusculae), is used in
inscriptions as much older than the classics as our oldest manu-
scripts, written with letters of the same form, are younger. These
majuscules were the only style used for the formal publication of
works of literature until the eighth century, and were used even

T^^^ NFF1

DVONORO-O TYMO'fVlgt-

SOfCEAfSOR -AtonU "

r T''rv

^^X^ 4 - ' '<- ' * Mf u. > ' /T> > L.M '> -

ll., B EPITAPH OF L. CORNF.UUS SCIPIO.

to the invention of printing for certain works held in extraordi-
nary esteem (the Scriptures, Vergil) and in other works for the
titles and the headings of chapters. From this last use was de-
rived the name Capital (capnt, chapter) which is still used for one
style of these majuscules, the oldest known to the Romans. It 96
may be studied in the copy (Fig. 8) of the inscription upon the

FIG. 9. POMPEIAN WALL INSCRIPTION.

Surda sit oranti tua |janua laxa ferenti] | audiat oxclusi verba [receptus amans] | janitor
ad dantis vigilet [si pulsat inanis] | surdus in obductam sofmniet usque scram]

64 LATIN MANUSCRIPTS.

tomb of Lucius Cornelius Scipio, dating from the third century
B. C. The second style is known as the Cursive. It was used for
less formal purposes than the publication of books, c. g., for mem-
oranda, accounts and correspondence. It is known to us from
words scratched or written upon all sorts of objects found in the
ruins of Pompeii, and also from a number of wax tablets, dating
from the second century A. D., which were found between 1786
and 1855 in the mines of Dacia. A specimen from Pompeii is
given here (Fig. 9) and a comparison of these cursives with the
majuscules above will discover differences not unlike those exist-
ing between our small script letters and printed capitals.

97 NATIONAL HANDS. This old Roman cursive had nothing
directly to do with the transmission of the classics, and is
therefore of less interest to the student of Paleography than
of Diplomatics. Employed for almost all purposes except for
the publication of books, but characteristically for legal and
administrative documents, it gradually developed, under local
influences and modified by the prevailing book hands, into
three strongly marked National Hands, the so-called Visigothic,
Merovingian and Lombard, very much as the Latin language
at the same time was' becoming the vernacular languages of
Spain, France and Italy. The most important of these is the
Lombard, which reached its fullest development at Monte Cas-
sino ( 59) during the ninth century. The Irish hand has a
different origin ( 105), and is far above the continental hands
in firmness and beauty. None of the National Hands were
destined to endure long, all being superseded after the time of
Charlemagne by the Minuscule type, which is discussed below
( 108).

98 THE MAJUSCULES. The Latin Majuscules divide into two types,
the Capital and the Uncial. The Capital is the more ancient, is
derived directly from the pattern used for carving upon hard mate-
rials, and therefore prefers the straight line to the curve, because
curves are hard to manage upon stone, wood and metal. So far
as formal literary works are concerned the Capital is the character-
istic type for the papyrus roll. It was so stiff and slow to write

ROLLS FROM

CAPITALS
FROM CODICES.

ANCIENT
CURSIVES.

UNCIALS.

N(A,A)

N X A

\. ( A) aeAA

& " '

U b *

c r

<X c) ^

oi (a )

^ (F)

/! U

ees.e,

f(F)

/' ff

f Ff

G qc

05 Y

q ( j ,

i_l U(

yb &

\1 M AA.

fuv /^ AV

ro<X30D

M M r?

^ ^

P P

r-?

PPP

Q 9

s <

T7 g

V(X'M)

V U U

U V

U U L|

FIG. 10 MAJUSCULE AND CURSIVE LETTERS.

66 LATIN MANUSCRIPTS.

that even in books it begins to be less common in the fifth cen-
tury, and then disappears altogether, except in artificial reproduc-
99 tions of the ancient style, in titles and headings ( 95). The Uncial
is a modification of the Capital type due to the effort to combine
rapidity with dignity. It is essentially a round hand, making a
single curve do duty for several of the straight lines found in a
letter of the older type, and changing the forms slightly so as to
allow the pen to remain upon the parchment while the whole char-
acter was traced. The name Uncial, first found in the writings of
St. Jerome (f42O A. D.), is not descriptive, because it means simply
"inch high" letters (litterae nnciales), a name which fits the Capi-
tals just as well. We do not know when the Uncial type was first
introduced. It is a common book hand in the fourth century, and
is the prevailing book hand from the fourth century to the eighth.
It was perhaps the parchment hand, as papyrus was not adapted
to fine strokes ( 15) or hasty pens.

100 CAPITALS. Two styles of capitals are recognized by experts,
called respectively the square and the rustic. In square capital
writing the letters are disproportionately wide, and commonly of
the same height except that F and L sometimes extend above the
others. The angles are right angles, and the bases, tops and ex-
tremities are usually finished off with the fine strokes and pend-
ants which are familiar to us in our modern printed copies of the
same letters. Rustic capitals, on the other hand, are of a more neg-
ligent type, but as a style of writing for choice books were no less
carefully formed than the square capitals. The strokes, however,
are more slender, cross strokes are short and are more or less
oblique and waved, and finials are not added to them. There are

101 also more letters of superior height. Of the two styles the square
is certainly the older, although the rustic happens to be found in
our oldest manuscript. Capital manuscripts have few contractions
and no punctuation marks. Originally words were not separated

STYLES OF WRITING. 67

but ill some that have come down to us the words are marked off
by dots placed about midway of the vertical space occupied by the
letter (see Fig. n); in some of these the dots may have been
added by a later hand.

SPECIMENS. Of Square Capitals very few specimens are pre- 102
served, a few leaves of a manuscript of Vergil divided between
Rome and Berlin (Plate V), and a few leaves of another manu-
script, also of Vergil, in the library of St. Gallen (Plate X). Both
are assigned to the fourth century, and the use at so late a period
of a style so laborious and wasteful shows the esteem in which
Vergil was then held (95). It has been remarked that Homer
in the Greek world, Vergil in the Roman world and the Bible
in the early Church were published with a sumptuous elegance
to which no other works could aspire. It is a manuscript of
Vergil which is singled out by Martial (XIV, 186) to be adorned
with the author's portrait.

Of Rustic Capitals specimens are more numerous, although in 103
this style, too, Vergil is reproduced more frequently than any other
author. In addition to the fac-similes of the famous Codex Pala-
tiiins of Vergil (see Plate I), and the even more valuable Codex
Bembinus of Terence (see Plate III), we give two specimens. The
first represents the oldest Latin manuscript extant, a poem by an
unknown author upon the Battle of Actium, found in the ruins
of Herculaueum. It is written upon papyrus ( 65) and from it
and others from the same place is taken the first column in
Fig. 10. The second is a fragment of Sallust's Histories (Fig. 12),
from a badly mutilated manuscript, of which there are a few
scattered leaves at Berlin, Rome and Orleans ( 65.)

UNCIALS AND HALF-UNCIALS. The two characteristic features 104
that serve to distinguish the Uncial type of the majuscule from
the capital are these: the letters A, D, E, H, and M are curved,
and the main vertical strokes tend to rise above or fall below the

X srf*

/ v^, w -^

; s< r~t <*->
< ^ V -7

rf I? **

1 1 5 =

2 I I

.2 a s o

S .2

^ U O QJ

* * I

JS S "3 "

3 3 P

<-> O .5

4J t/i *J

U U CJ

.' -'3

(/) TO yi Q

V -^ ffi ^

3 C JS

K c c

U 9 fi

aj O C

X3 .2 rt ^2 3

HI i-* ^" -

3 >

a .5" w

<2 C S
D O

COC,K)lt<S

Kio. 12. HRAGMtNT Ol ; SAU.USTS HIM'()Rli:s

70 LATIN MANUSCRIPTS.

line. Otherwise the letters are like capitals, and like these the
words are not separated and the abbreviations are few. In the
fourth century it wae the prevailing book hand, except for the
sumptuous editions already mentioned; in the fifth and sixth the
letters were still formed with much beauty and precision; in the
eighth it was rapidly degenerating. As a test of age the letter M
has been selected: in its earliest form the first limb is straight, or
at least is not curved inward at the bottom. The letter E is only
less characteristic: in the earlier centuries the cross stroke is con-
sistently placed high, but when the hand begins to break the posi-
tion of this stroke is variable, sometimes high, sometimes low, in
the same manuscript.

105 Half-Uncials are derived from the uncials and represent the
last efforts of the book hand to differentiate itself from the im-
proved business hand of the time. It is first found, but not as a
book hand, as early as the close of the fifth century, and is charac-
terized by exaggerated vertical strokes, by the close approach to
our small print of the letters e, m, n, and r, and by frequent liga-
tures, contractions and abbreviations. It is also called the Roman
Uncial and Pre-Caroline Minuscule. It was from the half-uncial
style that the independent Irish hand ( 97) was derived. For
specimens of the latter see Plates XII and XIII.

106 SPECIMENS. Of Uncial writing there are extant a considerable
number of specimens, of which two examples are here given. The
first is from the most ancient manuscript of this type which has
come down to us, the fourth century palimpsest of Cicero De Re
Publica, now in the Vatican library, with the superimposed writing
omitted. The letters are massive and regular, and the columns
are very narrow. An idea of the amount of material required for
the whole work may be gained from the five lines here given,
there being but fifteen such lines to the column and two columns
to the page. The second specimen presents a decided contrast in

STYLES OF WRITING.

FIG. 13. CICERO DE RE PUBLICA. (See Plate II.)
qui bona nee | putare nee ap | pellare soleat | quod earum | rerum vide [atur]

the size of the characters. It is from the famous fifth century
codex of Livy at Vienna.

Km. 14. LIVY FIFTH CENTCRY.

ri oppido posset ante ipsam Tempe in fau | cibus situm Macaedoniae claustra |
tutissima praebet et in Tessaliam | opportunum Macedonibus decur | sum cum
et loco et praesidio valido in

Of the Half-Uncials we give one specimen, taken from the 107
sixth century manuscript of St. Hilary ( f 368 A. D.) on the
Trinity now in the Archives of St. Peter at Rome. It will serve
at the same time as an example of the compositions which effaced
so many classical manuscripts (59)- The specimen presents
almost the entire alphabet, and it will be noticed that, while the
round style of uncial writing is retained, very few of the letters
are real uncials

72 LATIN MANUSCRIPTS.

u
lie
q mco 6

FIG. 15. ST. HILARY

damnationem fidei esse | te aboletur par aiteram | rursus abolenda est cufius] |
cpiscopi manuin innocente[m] | [lin]guam non ad falsiloquium coe [gisti]

108 THE MINUSCULES. The next improvement in the direction of
a rapid hand is the Minuscule writing. This is in all essentials
the same as the small (lower case) letters in our present Roman
alphabet. It is historically the product of all the factors already
described, the uncial and half-uncial book hands and the con-
temporary business hands acting upon each other. Chronologically
the Minuscule follows directly upon the half-uncial, the cursive
and national hands being merely subordinate local forms of no
importance in the case of classical manuscripts.

109 Minuscule writing, as are all the styles already mentioned, is
at its best in its earlier stages. Of the better forms the Caroline
(Carolingian) may be regarded as the type, as it finally became
the literary hand of all Western Europe, although in the differ-
ent countries certain peculiarities of the national hands survived
sufficiently for experts to tell with a good deal of accuracy the
place of origin of manuscripts of this age. In general it may be
said that the Caroline minuscule is round, heavy, almost sprawl-
ing. The letter o, for example, is not slender or egg-shaped, but
either a true circle or else shaped like an apple, broader than it
is long. On the same pattern are formed the left part of d and

110 the right part of b. The up strokes of certain letters, b and 1 for
example, are club like (thicker toward the top), owing to the run-

HALF
UNCIAL

VIST- MERO-

LOMBARD. GOTHIC. VINGIAN

OLDER

MINUSCULES.

LATER
MINUSCULES

a a

CC(ae If)

(aj^d

d d

b 6

b d

a-e

f ( f fi)

g
h

k 9

ll di \\ )

Klc

N ptt

f ( ri r> >

r ir

f(*fl

f S

iST'T*' final

X;*. d

CT (ti 9

C (ti as

*- Jlerov.)

? ?* P

vr y

* 1

FIG lf>. HAIJ- CNCIAI. AND MINUSCULE LETTERS.

74 LATIN MANUSCRIPTS.

ning together of the up and down strokes which we keep sepa-
rated in our script (). The letter r is not perpendicular to the
line of writing, as are the other letters, but inclines to the right
and has the side stroke broad and sweeping. There is almost no
distinction between f and s, as in our own books a century ago,
and the i has no dot. The Minuscule introduces the separation
of words, and a feeble attempt at punctuation. Abbreviations are
not especially numerous at first.

HI In the eleventh century the club-like vertical strokes disappear,
the writing becomes noticeably more slender, and the o and rounded
parts of b and d become egg-shaped. From this time abbrevia-
tions become more and more numerous and arbitrary.

In the thirteenth century the rounded character, which has
increased with every improvement in the book hand, begins to
disappear. The o, for example, is made with two strokes o, and
so the other letters with rounded parts, and finally all the curved
lines become straight. This is the Gothic type, forced and arti-
ficial, requiring two or three times the care and time to write: cf.
the four stroked o (0). For the reader it is especially trying. It
' is almost impossible to distinguish the letters i, n, u and m, espe-
cially when several occur in combination (e. g., minimum}: this led
to the writing of double i with accented letters (11), and finally to
the accent over a single i (i), whence our dotted form. It is from
this Gothic Minuscule that the German lower case letters are
derived.

112 In the fifteenth century came a reaction. The Humanists
( 60, 61) with a finer taste turned back to the Caroline Minuscules
as the characters for their copies of the precious manuscripts they
were searching for so eagerly and copying as fast as found. Here,
too, they made improvements. From the majuscules they borrowed
initial letters for sentences and proper names, and used them, as
has been remarked already, for titles and chapter headings.

' l
6

i 1 f 11 -I

-3 2 v fa !r "2 S s.

X w > -T*

e MIL*

~- V r C r i
fi ? I 1' g

76 LATIN MANUSCRIPTS.

113 SPECIMENS. The vast majority of our classical manuscripts are
written in minuscule letters, and specimens are therefore easy to
obtain, even if hard to select. The fac-similes of manuscripts of
Caesar, Sallust and Cicero (Plates IV, VI, VII, VIII, etc.) are
excellent examples of their several dates. In addition to these is
given an example of a fifteenth century manuscript, a Munich
codex of Livy (Fig. 17), to show the improved forms of the Hu-
manists. It is fortunate for us that the invention of printing
came during this period of simple good taste, for it fixed the
Caroline character forever as the type for modern books.

114 ABBREVIATIONS. In the later styles of the minuscules the
number of contractions, abbreviations and ligatures increases to an
enormous extent. The object was to save not merely time and
labor, but also parchment which was exceedingly costly. The use
of these abbreviations has greatly increased the labor of the pale-
ographer, because there was no general system in accordance with
which they were used, and a scribe's misinterpretation of a prede-
cessor's symbols might introduce, and has introduced, endless con-
fusion into our texts. It is impossible to give any connected
treatment of the subject. A table of the most frequent contrac-
tions is given (Fig. 18), with the warning that practice only will
enable one to read with accuracy, not to say facility, the manu-
scripts of the later centuries.

115 SUMMARY. From what has been said it will be understood
that the age of the ordinary manuscript can be fixed only within
very wide limits. The various styles of writing shade so grad-
ually into each other, that it is hard to tell where the earlier ends
and the later begins. In general it may be said that a codex
wholly in capitals is earlier than the eighth century, and if the
words are not divided earlier than the seventh; that an uncial
manuscript was written between the fourth and the eighth; the
minuscule prevails after the ninth, and if marked by many abbre-

apud

quae.qui,

quod: fj quae, d & a qui, o> &

autem

t-f, }/ (An.Sax.)A~U A u^-

dct & CJ^ quod.

ber

quando

tin , also <]<ta

-bis

V (urbis Ufv )

que

O (majuscule also), CjQ

-bus

B (Capit.), b. (s. also -us).

(XII. Cent.); fl;

con-

^ (shorthand; in XIII. Cent.)

quia

Ofl (in Gaius CK ).

9).

quid

ji ' i

T ^P' i

quidem

*4 ' ^>

deus.i etc.

D5 Dl etc., etc.,

quoniam

C| <V) . Q.VC^. (majuscule

enim

N ; lator tt . ft .
< <: -

too), c] uo

esse

:more commonly C^C u. C C

1 (in XV. Century also er,

est

V, -r, -7 ; An. Sax. ^o ; in

ir, re, r, e after r).

XV. Cent. 6. "^ etc.

runt

- -. 1.x. 2,? ? V <&,<&.

ser /

haec hie hoc

: "h haec 6c hoc, h hie, H fl

-sis \

hoc.

sunt

f s, ft

-. ~ ~

tera, ten,

ter f

mem, men

fri

-ud

o e. g. ilrS illud.

,~ { WO non).

-um

/ , for ex. R. rum, O$.

-nt

W/, (at end of word); N, U, U

orum.

-or

Ct , later also C? , Q?

, e. g.T T tor; later

per

Jp (Visigoth), ^

tQ tur.

pra

a *
P- P

^ <) ,e. g. l?T?iusrus; al-

prae

P IP'' P )

so: U -e. g.conftxTlz

pri

p (An. Sax. also /f )

Constantinus.

pro

f, ^. |

V Vt

proprio

vel

propter

ff'^

versus

pur

I'll', iS ABBREVIATIONS AND CONTRACTIONS.

yS LATIN MANl'SCRIPTS.

viations is not earlier than the eleventh ; manuscripts of Vergil are
later than the style of writing would indicate in other authors; of
two manuscripts written in the same style the more carefully
written is, other things being equal, the older; in any type stiff-
ness and constraint indicate a later date. As a last test we may
appeal to the spelling. We know from inscriptions the spellings
in general use at various periods, and if a manuscript varies in
spelling from the use o'f the author's day we may thus fix its
date. If, on the other hand, it retains the spelling current in the
author's time, it may be taken to be a careful copy of an earlier
manuscript, whatever its own date is (see, however, 128).

Experts go much further than this, but their results are
reached by practice and experience.

THE ERRORS OF THE SCRIBES.

CODEX. We may turn now from theoretical, or histor- 116
ical, Paleography to its practical side, the discovery of the
errors which a scribe would be apt to make in copying such a
manuscript as we have described. It will be convenient at this
point to review briefly the essential features of the manuscript
which would serve as his copy, and which he would pass on in
turn to his successors. So few of the papyrus rolls have come
down to us ( 12) that we may henceforth disregard them alto-
gether and concern ourselves with the parchment codices only.
The manuscript, then, or codex, will mean to us a parchment book
with the leaves stitched or glued into a cover, or binding, more or
less like our own books ( 16). These leaves are usually of folio 117
or quarto size ( 16), with writing on both sides ( 13), sometimes
arranged in narrow columns (Plates II, XII), sometimes running
clear across the page (Plate I) exclusive of the margins. These
margins are often covered with notes ( 48), written perhaps by
several different hands ( 84), but all as a rule later than the
codex itself. Some codices were written between the fourth and
tenth centuries, more from the tenth to the thirteenth, but most
from the thirteenth to the fifteenth ( 63). Of these the oldest
( 98) are written in capitals or uncials, without separation of
words and without punctuation marks ( 101), but with few
contractions; the later are written in minuscules, with a few stops
and numerous contractions ( in). The evidence, however, goes 118
to show that all our manuscripts, no matter how written, are de-
rived from originals written in capitals. To the earlier codex from
which the later are derived the name archetype is given. We
know, of course, that all our manuscripts are later by many hun-

80 LATIN MANUSCRIPTS.

dred years than the authors' copies, but we caunot tell in any
case how many reproductions may have been made between the
originals and our copies. There is absolutely no foundation for
the idea once fondly cherished, that among our manuscripts of
some author, may be one written under his own eye.

119 FAULTY COPIES. Now there is no manuscript extant which
can be depended upon to reproduce accurately the original copy.
Such manuscripts were almost unknown in the times of the au-
thors themselves ( 37), and all have suffered from successive
transcriptions. The best manuscript, therefore, will be found to
contain many blunders : lines that will not scan, words and sen-
tences that have no meaning. These may be corrected perhaps by
comparison with other manuscripts, but it sometimes happens that
none of our manuscripts gives the passage correctly, or that, while
several or all give good scanning and good sense, they do not give
the same words. In all such cases, i. e., where the manuscripts
support each other in obvious blunders or contradict each other,
it is the business of the critical editor to restore the text as nearly
as he can to its original form, either by determining for one manu-
script against the others, or by emending all. In doing so he
proceeds according to certain rules of paleography derived from a
study of the errors which the scribes were most apt to make.
These errors are of three kinds: Unavoidable, Intentional and
Accidental.

120 CLASSIFICATION OF ERRORS. This classification is practi-
cally exhaustive but it is not the common one. Scholars
usually say that the faulty copy varies from the original by
giving more words, fewer words, or different words. This
differs from our classification simply by looking at the result
rather than at the process. The usual treatment is given in
Freund's Triennium Philologiciim, Vol. I, p. 250 f., another
in Lindsay's Introduction, p. 10. A less formal treatment is
that in Gow's Companion to School Classics, p. 60 f., from
which are taken several of the examples used below.

C D~o

"lull

I & .a* * i ~ i 1:? H r e s **J 5 .1- a. d rf ?

j*f
ill

li'l'r

'?i !

, -a ^ 4 J;

' 8 f- tf fc TJ*

J?.

THE ERRORS OF THE SCRIBES. 8l

UNAVOIDABLE ERRORS. These are due to injuries suffered by 121
the manuscript used as copy after it was itself completed. The
most conscientious and painstaking scribe could only copy what was
before him: if that was defective his own work could be no better.
Now, time has dealt hardly with all our codices, and all good
manuscripts, exposed to the ravages of fire, damp, mildew, moths
and mice, are more or less defective or illegible. Sometimes in
the course of years or centuries, one or more leaves would be lost
from the codex that served as copy: all our manuscripts of Cor-
nelius Nepos have the same gap (lactina, it is called) in the life
of Lysander, showing that all are derived from the same arche-
type, itself defective here. We have no means of telling whether
much or little has been lost.

Again, leaves sometimes became loose, and were replaced in the 122
wrong order, or were carelessly put together at the end of the
codex, in either case sure to make the next copy wrong. Thus,
in the second book of the Annals of Tacitus chapters 62-67, as
our one manuscript and the older editions give them, belong after
chapter 58 and before chapter 59. On the other hand, of the two
manuscripts of Lucretius at Leyden one (Munro's B : Vossianus Q.
gq] has at the end 206 verses that cannot be read consecutively;
fortunately the other manuscript (Munro's A : Vossianus F. jo) has
these same verses in the proper order, 52 in Book I, 104 in Book
II, 50 in Book V, showing that A is an older manuscript than B,
having been copied from their common archetype before the leaves
were disarranged.

Again, a part of a leaf might be torn off. Thus a manuscript 123
of Horace (Keller and Holder's e : Einsidlensis 361} gives only the
latter parts of lines 1-18 of Epist. I, x, showing that the upper
left hand portion of one page of the archetype was gone. In such
a case if the loss were small and the sense clear the scribe might
be tempted to supply the missing words from his own inner con-

LATIN MANUSCRIPTS.

sciousness. So we account for the strange variation at the end of
line 126 of Horace Sat. I, vi, where the manuscripts are divided
between two readings, each making good sense:
Ast ubi me fessum sol acrior ire lavatum

Admonuit, fugio I rabiosi tem P or a signi,

{ campum lusumque trigonem.

And so in Tibullus I, ii, 25, a missing pentameter was filled in in
four ways:

En ego cum tenebris tota vagor anxius urbe.

{Securum in tenebris me facit esse Venus.
Praesidium noctis sentio adesse deurn.
Ille deus certae dat mihi signa viae.
Usque meum custos ad latus haeret amor.

124 Lastly, when a scribe discovered that he had omitted some-
thing he was accustomed to write it in the margin or at the top
or bottom of the page ( 84). His successors, copying his work
in turn, are scarcely to be blamed if they have inserted the
omitted words in the wrong place. So in Horace Ep. I, xv, verses
43 and 44 are omitted in some manuscripts, put after 38 in sev-
eral and after 39 in others. So also in Caesar B. G. I, xiii, 6, all
the manuscripts and many school editions still give the senseless

. ut magis virtute quani dolo contenderent aut insidiis
niterentur.

The words quam dolo undoubtedly should follow contenderent.

125 INTENTIONAL ERRORS. These are of two kinds, due either to
the bad faith of the scribe when he tried to pass off as authentic
what he knew was not written by the author, or to his ignorance
when he tried recklessly to make sense of what he did not under-
stand. Errors of the first kind are not the source of much trouble ;
they would be dangerous to make and not likely to be perpetu-
ated. Plenty of passages there are the genuineness of which is
doubted by some editor or editors, but most of these are due to
another mistake to be discussed hereafter the incorporation into
the text of marginal notes and in others the jury of scholars dis-

THE ERRORS OF THE SCRIBES. 83

agree so widely that no verdict can be brought against the scribe
Thus Horace Carm. IV, viii can hardly be right: it contains 34 126
verses, while all the others have a number divisible by four. So
the editors proceed to reduce the lines to a multiple of four by
rejecting the spurious verses. Knightly and Martin reject verses
7 and 8, leaving 32; Schiitz 14, 15, 16, 17, one-half of 24, 25,
and one-half of 26, leaving 28; Kiessling 17 and 33, leaving 32;
Nauck 17 and 28, leaving 32. Hardly any two agree upon the
inserted lines. One other passage in Horace has a suspicious
look. In Carm. Ill, xviii, n and 12,

Festus in pratis, vacat otioso

Cum bove pagus

some manuscripts give for the last word pardus, evidently a de-
vout scribe's reminiscence of Isaiah xi, 6: ''The leopard shall lie
down with the kid."

Errors due to ignorance are more frequent though confined in 127
the main to points of grammar. Thus in Horace Carmen I, iv, 12,

Nuuc et in umbrosis Fauno decet immolare lucis

Seu poscat agna sive malit haedo

some manuscripts have agnam and haedum because the copyist
unable to govern the ablatives as they stood made them objects
of poscat and malit. Precisely the same change has been made in
Vergil EC. iii, 77:

Cum faciam vitula pro fugibus, ipse venito

where the variant vitulam is found. So in Horace Carm. I, viii,
two indirect questions are followed by four direct. In most of the
manuscripts the direct questions were deliberately changed to indi-
rect until the process was stopped in line 8 by the impossible
scanning of timeat for timet. Again in certain manuscripts of
Cicero all subjunctives after quod were quietly changed to indic-
atives because in the middle ages it was believed and taught that
quod was always followed by the latter mood. Foreign words suf-

84 LATIN MANUSCRIPTS.

fered most from the ignorance of the copyists. Often transliterated
correctly enough by one scribe they were ruthlessly "corrected"
by his unscrupulous successors into such Latin words as they
were thought to resemble. The Gallic names in Caesar are cases
in point and many Greek proper names in the Latin poets, and
common nouns and other parts of speech in Cicero's letters, have
come down to us distorted out of all resemblance to their original
forms.

128 ACCIDENTAL ERRORS. These are of four kinds: of the Ear,
the Eye, the Memory, and the Judgment. It has been asserted
by good authorities that there are no errors in Latin manuscripts
due to dictation imperfectly heard. It is of course true that most
of our codices were written at a time when dictation was no
longer usual, but they are none the less derived from archetypes
that were almost certainly written from dictation ( 33). The
question is complicated moreover by our ignorance of many points
relating to the pronunciation of Latin, especially the changes that
must have taken place in the later periods of the language. We
do know that many words were spelled in more than one way and
that various spellings of the same word were current at the same
time. Hence some of the cases where manuscripts differ in spell-
ing may be explained by supposing that the copy was read to a
number of scribes who spelled the words according to their indi-
vidual preferences, without regard to the spelling of the author.
This same variation in spelling shows us that the sounds of -ae,
-oe and -e, of b and v, of t and d, of ci and ti must have been iden-
tical or very similar, and to the ear there could not have been
much difference between -et, -at and -it. So we may account for
the readings exitium and excidium, Horace Carm. I, xv, 21, volun-
tas and voluptas, I, xxvii, 13, citrea and cyprea, IV, i, 20. So
in Vergil, Aen. I, 48 and 49:

THE ERRORS OF THE SCRIBES. 85

. . . Et quisquam numen Itmonis adorat
Praeterea, aut supplex aris inponet honorem

we find adorat . . . inponit, adorat . . . inponet, adorat . . . tnponat,
etc., to say nothing of the varying prefixes im- and in-.

ERRORS OF THE EYE. Of these we make four kinds: Ablepsy 129
or Blindness, Dittography, Lipograpy or Haplography, and Skip-
ping. By ablepsy is meant the failure to distinguish between
words that look somewhat alike. A study of the tables of letters
and of the fac-siiniles already given will show that in the manu-
scripts many letters and combinations so closely resemble each
other as to be scarcely distinguishable when read rapidly. Errors
arising from this source are more common in the later codices,
written in minuscules, than in the earlier majuscule writing. Still
they are sometimes found even in capital writing, as the taking of
F for E, C for G in the case of single letters, and in combinations
the taking of N T (often ligatured, KT) for V T or V I. The letter
S was often turned the wrong way 8, and if written too closely
after I was liable to be taken for R (18). An example of the 130
former confusion may be found in Velleius Paterculus II, xxix, 3,
where it is said of Pompeius that he was

"... potentiae, quae honoris causa ad eum deferretur, non
VT ab eo occuparetur, cupidissinius."

Madvig suggested VI for VT (vi for ut) which removes all diffi-
culty. An example of the latter is found in Horace Carm. I, iv, 8:

. . . dum gravis Cyclopum
Voleanus ardens VISIT officinas

where most of the manuscripts have urit, i. e., VISIT. So we
shall never know when Terence was born. He died in 160 B. C.
when, according to the archetype, he was passing his Illcesimum
(=triccsi)>ium) qitintum annum: this might be carelessly written
however for (///=I/I) vicesimum quintum, a difference of ten years.

In minuscules, on the other hand, the single letters / and /, 131
b and v, c and e are very similar, and so too are the combinations

86 LATIN MANUSCRIPTS.

in, m, ;// ( in), ut and lu, iec, lee, and tec, cl and d (Fig. 15), lu
and /zz'. Our texts still hesitate between Verucloetius and Verudoc-
tius. Then there are whole words that closely resemble each other :
omnia and omina, fulmen and flwnen, dearum and dierun\, numen
and nomen, and very many others. In Horace Carm. II, vi, 19, some
manuscripts have minium for minimum; in Ep. II, i, 198, they read
nimio and mimo where Bentley thought Horace wrote nummo.

132 Lastly under this head we may speak of marks of abbreviation
(see Fig. 17, p. 77), some of which were conventional, others
arbitrary, all liable to be overlooked by one scribe or wrongly
interpreted by another. This last source of error may be more
properly classified below under errors of Judgment.

133 DlTTOGRAPHY. This is the writing twice of what ought to be
written but once, and is an error only less frequent than ablepsy.
Turning over a few pages of Reid's Cicero's Academica I found
mallent for malent, antiqui qni for antiqui, and (an excellent ex-
ample) materiam iam and materiam etiam for materiam. In Sal-
lust Catiline, xix, 3, the inferior manuscripts have quos sine exercitu
for quos in exercitu. Tyrrell's foot notes to Plautus Miles Glo-
riosus show autem milia for autem ilia (ablepsy, too), vim me cogis
for vi me cogis. This error is especially exasperating in the case
of numbers expressed by capital letters: thus in Caesar B. G. II,
iv, 9, one manuscript says the Menapii promised VII thousand men,
another vin and another vim. Fortunately in this case they could
go no higher by dittography, but in such matters as these little
reliance can be placed upon the text. As a last example take
Livy XXVII, xi, n:

. . . cui di sortem legendi dedissent, et ius liberum eosdem

dedisse deos.

Here the writer of the best manuscript has expanded the six words
dedissent . . . dedisse into the sixteen dedissent et ius liberum eosdem
dedissent et ius liberum eosdem dedissent et ius liberum eosdem dedisse.

THE ERRORS OF THE SCRIBES. 87

LIPOGRAPHY OR HAPLOGRAPHY. This is the converse of dit- 134
tography, that is, the writing once of what ought to have been
written twice, and is equally common. What has been said above
about dittography in the case of numbers applies also to lipography.
In the Acadernica (I, 4 and 32) are found abhorrent for abhorrerent,
probatur for probabatur, and in the Miles mortem ale for mortem
male, simile sciat for similes sciat. So we find such errors as dicit
for didicit, decus for dedecus, etc. In Cicero Laelius xiv, 48, si
qua significatio should be, in Reid's opinion, si quasi significatio.

SKIPPING. The last and worst error of the eye is skipping, 135
the worst, because where we depend upon one manuscript only we
frequently find passages where something must have been omitted,
and where there is no clue but the context to what has been lost.
The error is due of course to the scribe losing his place as his
eye traveled from copy to parchment, and is caused almost always
by the occurrence of similar words, or at least similar syllables,
in the same relative position in different lines. Thus in the best
manuscript of Sallust (codex Parisinus Sorb. 500, saec. X] Cat. xx, 1 1 :

. . . nobis rem familiarem etiani ad necessaria deesse? Illos
binas aut amplius domos continuare, nobis larem familiarem
nusquam ullum esse?

the eye of the scribe passed from the first familiarem to the second,
omitting all the words between them. Here the omitted words
are supplied in the margin and by other manuscripts. Such mis- 136
takes are rare in poetry, where the prosody, and sometimes the
stanza, may be a safeguard. Still in Horace Carm. I, xii, 25-27:

Dicam et Alciden puerosque Ledae,
Hunc equis, ilium superare pugnis
Nobilem; quorum simul alba nautis
Stella refulsit,

where lines 26 and 27 end in the same syllable -is, several scribes
have omitted line 26, although the omission spoils the sense and
stanza, and one even changed nobilem in 27 to nobiles to make

88 LATIN MANUSCRIPTS.

it agree with pueros in 25. Of course this error may sometimes
be charged against the scribe when the fault is due to a defect in
the copy ( 121), but the two sources may usually be distinguished
by the fact that a torn (but not a stained) archetype causes loss
on two pages at least.

137 ERRORS OF .THE MEMORY. Every one who has had occasion to
do any copying knows how wearisome it is to copy word by word.
It is much faster and less irksome to copy by groups of words,
taking each time as much as one can retain in memory. There
are three dangers : that getting the correct sense of the passage we
may arrange the words of the copy wrongly, or substitute for
them synonymous expressions, or omit or add words where the
omission or addition will not affect the general sense. We have
therefore these three errors to consider.

138 TRANSPOSITION. The changing of the order of words in a
Latin prose sentence is more frequent than in English on account
of the greater freedom of arrangement in the inflected language.
It is hardly possible that there is a single chapter in Caesar, or
Cicero, for example, where all the manuscripts give the same
order. The following examples from Cicero Pro Cluentio, vii, will
show how common the error is: dnbitatio ulla . . . ulla dubitatio,
scire volui vos . . . scire vos volui, causa accusandi . . . accusandi
causa, rebus suis diffidentes . . . diffidentes suis rebus, et fuit apud
eum . , . et apud eum fuit, mortmis est . . . est mortuus, is heredem
fecit . . . is fecit heredem, sororis filium suae . . . sororis suae filium.

139 All the manuscripts give the first words of Livy's (59 B. C.-iy
A. D.) preface: factumsne sim operae pretium, whereas Quintilian
(40-118 A. D.) expressly quotes them (IX, iv, 74) as the begin-
ning of an hexameter: facturusne operae pretium sim. So even
syllables or single letters were transposed, as in Vergil Georg. IV,
71, where one manuscript gives aries for aeris, and in II, 356,
where the best manuscript (Ribbeck's M : codex Mediceus 39, /,

THE ERRORS OF THE SCRIBES. 89

saec. V.} ends the verse with submoveret ipsa, for sub vomere et ipsa,
adding the sin of Lipography to that of Transposition (v and m
are transposed, and the e is written once). For transpositions due
to marginal corrections see 124.

SUBSTITUTION. The substitution of synonymous expressions 140
for the exact words of the author is nearly as common. Hence
the various readings of et, -que, ac, and atqne, turn and tune, ut
and uli, ni and nisi, ob and propter, quod and quia. Sometimes a
singular may be substituted for a plural, as hoc postulo for haec
postulo, or one tense for another practically the same, as dico for
dicam, debetis for debebitis, or words of nearly the same signifi-
cance, as putaretur for videretur, defendenda for depellenda, adiun-
gerer for adhibcrer. All these examples are taken from a few
pages of Cicero Pro Cluentio.

OMISSIONS AND ADDITIONS. Unimportant words, that is, words 141
whose presence or absence will not essentially affect the meaning,
are frequently left out or .put in by the copyist. Such words are
especially the personal and possessive pronouns (for reflexive pro-
nouns see Reid on Cicero Acad., p. 115, 1. 12), certain prepositions,
especially in and cum, the verb esse, most frequently when it would
be part of a compound tense, words repeated for emphasis in the
same clause, and vocatives, especially such as indices, patres con-
scripti, and the like. Less frequently subordinate clauses are
omitted, but this is more apt to be due to Skipping than to
imperfect memory. To this fault belongs also the confusion in
proper names when two or more stand near each other: frequently
the scribe would repeat the first and leave us with no clue to the
second.

ERRORS OF JUDGMENT. Of these three classes may be made: 142
the wrong division of words, faulty expansion of abbreviations, and
the insertion in the text of scholia written between the lines or in
the margin.

90 LATIN MANUSCRIPTS.

143 WRONG DIVISION OF WORDS. It has been remarked that in
majuscule writing ( 101) words were not separated and sentences
were not divided by punctuation marks. When later copyists,
writing in minuscules, undertook to divide the words ( no) and
to punctuate the sentences mistakes were numerous, and these
were the cause afterwards of still further corruptions of the text.
In Caesar B. G. I, xxxi, 12: quod proelium factiim sit ad Mage to-
brigam, some manuscripts divide differently, Admagetobrigae, chang-
ing the ending to fit the construction. In Vergil Aen. III. 150:

. . . visi ante oculos adstare iacentis

In soninis, multo manifest! lumine, qua se . . .

we have a dream, but it becomes a vision if we write insomnis
(genitive).

144 This error is usually complicated (see above) by one or more
of those already mentioned. Thus the manuscripts read in Plautus
Mil. Glo. 309: hocine simile sciat . . . aedis tollat, which makes no
sense. The error arose by the scribe first taking nc, the conven-
tional abbreviation for nunc, for -ne ( 132), then si miles was
written similes, and finally, by Lipography ( 134) an ^ was lost
before sciat. The original text must have been: hoc mine si miles
sciat . . . acdis tollat. There is another good example in line 1262,
where the manuscripts read:

Non video: ubi est?

Videre spolia mares.

These words give a meaning, but one foreign to the context. The
last words should be: videres, pol, si amarcs. The words were
first divided wrongly, and then the meaningless spolsia was turned,
perhaps by a later copyist, into the nearest Latin word, spolia.

145 FAULTY PUNCTUATION, or incorrect division of clauses and sen-
tences, scarcely requires illustration, but one example may be given
to show how it may be complicated with the error just described.
In Seneca Ep. LXXXI, 4, the codices have :

THE ERRORS OF THE SCRIBES. 91

Philosophia unde dicta sit apparet : ipso enirn nomine fatetur.
Quidam et sapientiani ita quidam definierunt ut dicerent, etc.

This was corrected by Madvig: . . .fatetur quid amet. Sapient 'iam, etc.

Of the trouble caused by the misinterpretation of abbreviations
something has been said under another head ( 132), an example
is given above ( 144) in the Miles, 1. 309.

INTERPOLATIONS. It has been remarked ( 47-49) that on the 146
margins of their manuscripts the ancient grammarians and critics
often wrote their glossae and scholia, just as nowadays the mar-
gins of too many school books are used for the same purpose.
These notes might consist of a single word or several lines, all
written in Latin, of course, and often in close imitation of the
author's style. On these same margins ( 84) the scribes wrote
the words they had accidentally omitted from the text. Now when
in the middle ages a scribe undertook to copy an ancient manu-
script covered with notes and corrections, he could not always tell
them apart, could not, that is, in all cases distinguish the several
hands ( 85), and so the comments of the scholiast were often
incorporated in the text. Some of these are readily detected by
comparison with other manuscripts and some less easily by the dif-
ference in style, and the [ ] in our text books mark off such words
as scholars generally believe to be spurious. Some, of course, may
never be detected. The supplanting of a genuine word by its gloss
( 49) is a more serious matter, for only rare words needed
glosses, and as students of language we cannot afford to lose a
single word that the Romans ever used.

UNCERTAIN SOURCES OF ERROR. While the most common 147
errors of the scribes have been pretty fully illustrated above, it
must not be supposed that the cause of a given blunder can
always be confidently named as soon as the blunder itself is de-
tected. In some cases the corruption existing in our manuscript
may be due to error superimposed upon error in successive copies.
Sometimes the given fault may be due to any one of several pos-

92 LATIN MANUSCRIPTS.

sible causes, the particular one being of no possible consequence.
As an illustration we may take the omission of a word : it may be
due to haplography, or to skipping, or to the deliberate act of the
scribe. For instance, in Plautus Amph. 723:

Enimvero praegnati oportet et malum et malum dari
some manuscripts give the words et malum but once. This may
be explained either as a case of haplography, or an attempt by the
copyist to correct what he conceived to be a blunder of dittography

148 on the part of his predecessor. There are still other reasons for
the omission of a letter or word. In some elaborate manuscripts
(see Fig. 17) initial letters of verses or chapters were made very
ornamental, and would be left by the scribe for the "rubricator",
or "miniator" to fill in. In many such manuscripts, e. g., the
codex Ursinianus (D) of Plautus, these letters were not supplied,
and later copies are either defective or variously emended. This
lack of initial letters may account for the variation in Horace,
Carm. I, xix, 1 1 :

Et versis animosum equis,
where for the first two words (et perhaps abbreviated) we find the

149 manuscripts give us variously versis, aversis and et versis. Again,
in the scriptoria of the monastic establishments there were correct-
ors whose duty it was to revise the manuscripts as fast as they
were written, comparing them perhaps with the originals or with
other standard copies. If a scribe found in his copy a word which
he could not make out, or did not understand, or which he took to
be a corruption, he might, leave it for this corrector to fill in either
from his superior knowledge or after comparison with a copy more
plainly written. The corrector might fail, however, to notice the
omission, and so the copy would be even more faulty than its prede-
cessor. It is evident that the certain correction of such errors will
very rarely turn upon the particular explanation which we adopt for
them, and this is true of many errors other than those of omission.

W <-*!

^ Q
!5 oj

h t

Q i

2?'
H -5

: M

*-,-"

', W

h u)

S 2 < '^

UJ m
, * < ^^^

feg--hi

. *^*^* l*-^ 3 * ^

\ * ^

L ' ' ^**

*v *^ f
^ 4 n ^

'rf-r> ;ii -5.
^ H % p

dx ;s

u w 5

r^ i p" 1 ^ ^-=*

5 o

c/J ^

tu H

wr-SpSlS

fer^

O s s H .tu a;^w ; . *g ^
w cj J? J cj ^ H /- ^w

< w > 2 < 5 ?> ^
< ^ < -

/ J

III.
THE SCIENCE OF CRITICISM.

METHODS AND TERMINOLOGY OF CRITICISM.
TEXTUAL CRITICISM.
INDIVIDUAL CRITICISM.

THE SCIENCE OF CRITICISM.

METHODS AND TERMINOLOGY OF CRITICISM.

/ ~pHE general functions of Philological Criticism may be said to 150

be to examine the works of antiquity that have come down to
us, to determine and restore so far as possible their original forms,
and to assign to them their proper place among other works of
the same sort. By a gradual limitation, which we have already
seen ( 91) in the matter of Paleography, the field of Criticism
has been narrowed to the literatures of Greece and Rome, works
of art being considered the province of the younger science of
Archseology. By common consent, too, the last function named
above has been left to general literary criticism. In this book we
confine ourselves to Roman materials, and may therefore take for
our last subject of discussion the restoration to their original forms
(see 119) of the works of Latin Literature so far as they have
been transmitted to us.

SUBDIVISIONS OF THE SCIENCE. While the general definition of 151
Philological Criticism given above would probably be accepted by
writers upon the science, even the most superficial reading of their
works discloses a wide difference in their methods of treatment,
due largely to their various subdivisions of the science into what
may be called its branches. We read to our confusion of Higher
and Lower Criticism, of Diplomatic and Conjectural Criticism, of
Grammatical, Historical, Aesthetic, and Individual Criticism, to use
only the names employed by English authorities. These names
are not arbitrary and meaningless, but are more or less naturally
derived from some work which the critic does, or some result at
which he aims. We must therefore try to understand them clearly,
whether all prove to be of practical value or not; and we can un-
derstand them best by considering the nature of the critic's work.

96 LATIN MANUSCRIPTS.

152 THE CRITICAL DOUBT. We are reading the text of some classic
and find something that we do not understand, or we find some-
thing peculiar in the expression of what we do understand. The
standard by which we judge is our knowledge of the language
which we are reading and our familiarity with the facts about which
we read. If we do not distrust our own knowledge, the perplexity
occasioned by the unintelligible or unusual expression will cause
a doubt about the integrity of the text we are reading, because
we know from experience with books in our own language that
the author's words are not always correctly reproduced upon the
printed page. This doubt may be thus formulated: Was this
word or this phrase or this sentence which offends us written in
this way by the author? Of course if we had in our hands the
original manuscript of the author the doubt would not be pre-
153 cisely the same. In this case, if a similar offense is given, as it
is given in almost every letter we receive, we ask : Did the au-
thor intend to write this which offends us, or have we here a slip
of the pen? In the case of inscriptions we have what approaches
very nearly to an original manuscript; in the case of Caesar, Sal-
lust and Nepos, however, we do not have the original manuscript
but a copy, one out of a long series of copies, no one knows how
long, and the doubt as to the integrity of the text is all the
more natural. This same doubt can arise and does arise in
another way. We find in the text which we are reading, be it in
a printed book or a manuscript, in a passage which of itself
causes no perplexity and gives no offense, one or more variants.
Even if all these, taken one by one, are intelligible and satisfac-
tory, still a doubt is started, for the author can not have used all
the variants, if indeed any one of them is his.

154 CAUSES OF DOUBT. Now it is evident that these doubts, or
the causes of these doubts, that is, the offenses against the stand-
ard by which we judge, are very numerous, and of many kinds.

tnci m>ttu ccmiiitncti ^hiit-
1 4<uUtni^ tnlnuf oil ^uioi r
r rw&t-tmix petrmintic; jtcr
iu? Uor tn^tnjo ut^hit; _
cUufl wpini^ cut 1 .

Time Uf ant; TICUUTTI UUIluj
mow ttmiior cvtc
. tum^-fti

Am tarn mm auUt^ c>T~ unuf
iatx

tuvc

Ixr lOxUt

Ixiix

(X lulcmqi t|' pAtrcru
u^ unc Tttjhot IT

pctu^ polTerwi Icfbic-
i-j "Utter ttltttc mtf bxitlle

Cliu an Utrett qAcm wi {Inii tcnttt
CX in inimir, DMvitiim <utt a
t acrid fclce matait>. riuttfuc .
^ V. um ^clmfrio mco Tjitx^nft .
^ Arum nciac quio hlxi- tccin
t PoUnoluni (in I
"t

" nil

-j- T1 fpwcut puctlc

auttolum -ftuflc -trulm

fottu^r

7iu

XI. CATru.uS: Romanus, San. A'//".

METHODS AND TERMINOLOGY OF CRITICISM. 97

and it is the attempt to classify these offenses that has given rise
to so many divisions and subdivisions of the science of criticism
with distinctive names. To understand these names let us carry
the process a little further. We find in the text which we are
reading an expression which violates the formal laws of the lan-
guage as they are already known to us, or which is at least con-
trary to the usage of the writer or of his time so far as these
usages are already known to us ; or else we find something wrong
with the thought: it either gives no sense or a sense which con-
tradicts what has gone before. In other words the language con-
sidered either in itself or as an expression of thought offends us,
and the effort to remove this offense is called Grammatical Crit-
icism, or Literal Criticism. Again, we find in the statement of 155
facts already known to us, something that contradicts what we
have learned from other sources and have believed to be true.
This contradiction may raise a doubt as to the facts themselves,
a doubt which leads to what is called Historical Criticism. Or, if
we do not distrust the facts, we may distrust the good faith of the
author, and this is correctly enough though not so obviously
referred to the same branch of criticism. Or, if our confidence in
the fides of the writer cannot be shaken we have no recourse but
to doubt the integrity of the text, which carries us back to Gram-
matical Criticism, or as a last resort to wonder whether the pas-
sage may not be an interpolation or the whole work perhaps
assigned to the wrong author. To settle these last questions we
appeal to what is usually called the Higher Criticism: a better,
because more suggestive, name is Individual Criticism. Finally, 156
we may find something that offends our taste, something correct
enough in itself but out of place in the poem, as unpoetical, in
the oration, as unoratorical, iu the history, as undignified or what
not, and this brings vis to Technical or Aesthetic Criticism. But
before we inflict upon the poet or orator or historian the penal-

98 LATIN MANUSCRIPTS.

ties for violating the canons of taste we may find reason as before
to distrust rather the integrity of the text or the identity of the
reputed author.

157 Given now one of these three causes of doubt, and the three
classes will be found to be practically exhaustive, we have next to
find the appropriate word or phrase or sentence to remove the
offense, and lastly to establish such appropriate term as the orig-
inal term used by the author. The appropriate term may be easy
or hard to find. Sometimes it is our consciousness of the appro-
priate that takes offense at the reading: stella darus, Romulus
seen/idles fuit rex Romanus, although even here we might hesitate
between sol and clara for the first, and Numa and primus for the
second. It may be that among the variants recorded iu the edi-
tions which we consult, or actually existing in the manuscripts
which we examine, we discover an appropriate term, and this the
only one that is appropriate. It is evident that this may be pro-
nounced with confidence to be the original term, the one used by
the author. This process of finding what is appropriate and estab-
lishing it by comparison of manuscripts is called Diplomatic Crit-
icism, and depends for its success and certainty upon our knowl-

158 edge of Paleography and the making and fate of manuscripts. But
it may well be that nothing appropriate is found in the variants,
and the critic is left to find it in his own sense of what is suit-
able. If he can invent such an appropriate term, which removes
the original offense and gives rise to no other, he may put it for-
ward as probably the original word or phrase, etc., and if other
scholars accept his suggestion (technically termed a "conjecture")
as certain, it is called an "emendation" and becomes part of the
text. Such a process is very inappropriately called Conjectural
Criticism; inappropriately, because it has in it no element at all
of criticism. It is to be hoped that a better name may be sug-
gested for it.

METHODS AND TERMINOLOGY OF CRITICISM. 99

KINDS OF CRITICISM. These illustrations will serve to explain 159
the technical terms used by writers upon criticism, and will also
show that, however convenient these names may be to describe
more or less distinct processes, they are of very little practical
importance, because of the close relation of these processes to each
other and their mutual dependence. No hard and fast line can be
drawn between any two of them even theoretically, and in practice
no such line ever is drawn. Still, for convenience of treatment we
may make two divisions: the two that are commonly but very
inappropriately called the Lower and the Higher Criticism. The
former undertakes to determine and restore so far as possible the
original form of the composition, while the latter characterizes the
style of the work and identifies its author. We shall use the
more suggestive and therefore more appropriate names Textual
and Individual Criticism.

CRITERION. It must be remembered that the standard by 160
which we determine the appropriateness of a given term is
simply our own knowledge. When therefore we take offense
at a certain reading it may well be that our knowledge of gen-
eral usage or of the author's peculiar usage is at fault and not
the traditional text. What seems to the critic at one time
faulty may in the light of fuller knowledge seem correct and
appropriate (see Munro's Lucretius, third edition, p. x). This
change of attitude is more likely to occur when the period be-
tween successive revisions is longer than the span allotted to
a single scholar. It is not merely the advances made in crit-
ical science during the intervening years, but far more the
wider range of modern scholarship in all fields of investigation
together with its microscopic minuteness, that puts our recent
texts above those of the past, and gives them promise of more
of permanence as well as of authority. And this is our chief
hope (see 12 and 88) for further improvement.

TEXTUAL CRITICISM.

161 JT has been remarked above ( 152, 153) that the impulse to a
critical examination of the text we are reading may come from
two sources. We find something inappropriate, which offends us,
or we find an appropriate but unfamiliar term in a familiar pas-
sage, which excites our surprise. Our first thought naturally is
to inquire whether or not the perplexing term rests upon any
authority, for it may be a misprint or the invention of the editor
whose text we are using. If we find authority for the inappro-
priate term, or for both of the appropriate terms, our perplexity
can be removed only by determining the value of the authority in
the first case, or of the opposing authorities in the second. This
process is called Textual Criticism and can best be understood if
we suppose the case of an editor undertaking the independent
determination of the text of some classic.

162 APPARATUS CRITICUS. Such an editor has first to collect all
the testimony available for the original work. The amount and
importance of this testimony, called collectively the Critical Appa-
ratus, will vary widely with different authors and with different
works of the same author, but will consist of some or all of the
following: manuscripts of various dates, early printed editions,
translations into other languages (available for Greek authors and
the Scriptures only), ancient commentaries, citations and imita-
tions by later writers, technically termed " testimonial These we
shall consider in the order given.

163 THE MANUSCRIPTS. These are the most important witnesses
and the editor will make his collection as complete as possible and
will study them with peculiar care. He will seldom or never have
the manuscripts themselves before him as he works but merely

TEXTUAL CRITICISM. IOI

collations ( 86, but cf. 87 at the end), that is, witnesses to wit-
nesses. These collations are not so trustworthy as the manu-
scripts, because liable themselves to the same defects as the manu-
scripts, the errors of the copyist, and in addition in most cases to
the errors made by printers and undetected by proof readers. Be-
sides these the collator of several manuscripts of the same work is
liable to confuse the readings and to assign them to the wrong
manuscripts. Sometimes we have several collations of the same
manuscript made by different scholars, and where these vary the
editor will append the name of the collator (see Tyrrell's Plautus
Mgl., p. 7). The testimony of any particular manuscript is not
therefore absolutely reliable, and never can be made so, but modern
collations are more trustworthy than earlier ones, not only on ac-
count of the increased attention given to Paleography, but also on
account of the more general recognition of the fact that in de-
scribing manuscripts the least and apparently most insignificant
detail may prove to be of great importance.

EXAMINATION OF THE MANUSCRIPTS. Having collected as 164
many manuscripts, i. <?., collations of manuscripts as possible, the
editor next examines them all methodically to determine the rela-
tive weight that is to be attached to their individual testimony.
This does not mean that all will be scrutinized with the same
minuteness, for a very cursory examination of one manuscript may
show the editor that it is merely the copy of an earlier one in his
possession and may therefore be entirely neglected. In the same
way he will disregard the printed editions that are founded upon
manuscripts which he still has, for these editions have no inde-
pendent value and are one step further from the original. On the 165
other hand, if the foundation manuscripts are no longer extant
(as in the case of Velleius Paterculus, see Rockwood p. xvii f.)
the editio princeps ( 64) represents them and testifies for them.
In this case \ht fides of the editor of the book must be considered:

102 LATIN MANUSCRIPTS.

the readings of the manuscripts X and Y of Cicero's letters, upon
which Orelli largely based his edition and which were known only
from the edition of Simeon Du Bos (born 1635), ^ lave been proved
to be forgeries with which the unscrupulous editor undertook to
bolster up his often convincing conjectures (see Tyrrell Cicero in his
Letters, p. c). So the great problem of Horatian criticism turns
upon the confidence to be reposed in the readings from certain
lost manuscripts preserved in the edition of Jacobus Cruquius,
issued in 1578 (see Wilkins Horace Ep., p. xxvii, and Palmer
Horace Sat. 3d ed., p. xxix f).

166 POSSIBLE RESULTS. This examination of the manuscripts and
the editions which represent lost manuscripts will result in one of
four possibilities, as follows: i. The editor may find but one
manuscript of his author upon which to base his text. This con-
dition obtains in the case of Hyperides and Babrius among Greek
writers, and the first six books of the Annals of Tacitus among
the Latin writers. The condition is so rare because it means that
the works in question were entirely lost for a time and were not
recovered until after the invention of printing (about 1450). In
such a case the editor's task is comparatively simple: he has but
to take the necessary pains to read the manuscript correctly and
reproduce it accurately. All that he does beyond this is, strictly

167 speaking, not textual criticism. 2. The editor may be able to
trace all existing manuscripts back to a single manuscript also ex-
isting. This is 'true of certain Orations of Lysias, true of Athe-
naeus, true of books XI-XVI of the Annals and I-V of the His-
tories of Tacitus, and is believed by certain scholars to be true of
some few other works. The course of the editor in this case is as
clear and simple as in the first case, when once the derivation of
the manuscripts is demonstrated. This demonstration, however, is
a matter of exceeding nicety and corresponding difficulty. The
mere agreement in all or almost all readings with an important

wipfttieqty-Affi'tfG

^^^ft^W^. ^^w? 2 !

0<frtS^/64, r . ff^Jl. >=*' 4>1\ tn^Lrrn

i WH^-^^^jS^ fl^^

y \ '

7U#

XII. HORACE: Bcrncnsi*

IX,

TEXTUAL CRITICISM. 103

difference in age is not enough to prove the descent. The most
convincing proof is a lacuna ( 121) in all the younger manu-
scripts with no evidence of mutilation, while such mutilation is
found in the oldest manuscript at the place where the lacuna oc-
curs. 3. The editor may find that all existing manuscripts may 168
be traced to one manuscript no longer extant, which can, however,
be more or less completely and accurately reconstructed from copies
in his possession. Such a manuscript is the Henoch's codex of
the Dialogue of Tacitus (see Gudemann's edition, p. cxv) and the
Verona manuscript of Catullus. This last was used in the tenth
century and disappeared, was found again and copied in the four-
teenth century and has again disappeared (see Merrill's Catullus,
p. xxxvi). The proof of the descent is, of course, even more diffi-
cult here than in the second case, although essentially the same
in kind. The task of the editor, moreover, will be simplified only
so far as as he is able to reconstruct the archetype. When this
cannot be accomplished we have the fourth and last case. 4. The 169
editor may find his manuscripts hopelessly confused, or divided
into several families whose connection cannot accurately be deter-
mined, and to which the several manuscripts can be assigned only
doubtfully or provisionally. Here the difficulty increases in propor-
tion to the extent of the work and the number of the manuscripts.
Sometimes, when the manuscripts are very numerous, the problem
may be solved by some favorable, almost lucky, circumstance, as
e. g., the superiority of P. (see Plate VI) over all the other manu-
scripts of Sallust, even those of the same class. On the other
hand the problem may baffle generation after generation of schol-
ars, as has been the case, and seems likely to be the case forever,
with Horace. The consideration of these four cases will show how
the discovery of a single manuscript, although of no great value
in itself, may completely overthrow the accepted text of a given
author.

104 LATIN MANUSCRIPTS.

170 STEMMATA. The derivation and relation of the manuscripts
of certain authors may be represented by diagrams, called
stemmata (genealogical tables), varying in complexity with the
number and character of the manuscripts. The most interest-
ing are those which illustrate the descent of our existing manu-
scripts from a supposed original no longer extant (case three
above). As an illustration Meusel's stemma of the manuscripts
of Caesar's Gallic War is here given, taken from his edition,
Berlin, 1894. Of the twenty manuscripts which he names on
p. xi, eleven are disregarded because they are believed to be
copies of some of the remaining nine ( 164). These nine
manuscripts, distinguished by the letters A, Q, B, M, S, a,/, //,
/ ( 82), Meusel arranges in four groups, A Q, BMS, af, ///,
because the members of each group agree so closely in their
readings as to warrant a belief in a common origin for each

set. This origin for A and Q, i. e., the lost
x manuscript from which both A and Q were

copied, he calls x ' wherever, therefore, A and
Q have common readings, we have the read-
ing of x ; whenever they disagree we must
^__ ,-~^ ~^~ ^ . assume an error on the part of one copyist

171 A Q BMS a. fh i or both. In like manner the common source

of B, M, S is called $, of a and /TT, of h and
/ p. These four supposed originals \, <f>, ir, p, each reconstructed
as explained in the case of X, are now carefully compared and
are found to divide into two groups, x resembling < very closely,
and TT resembling p. The archetype of x a "d </> is called a, that
of TT and p is called ft, and these two, reconstructed as were the
other supposed originals, are found to have so many common
readings, and so many variations that can be explained paleo-
graphically as coming from a common source, as to point at a
common origin for both. This source, called X, is therefore
the common archetype of all our manuscripts of the Gallic War.

172 USE OF THE STEMMATA. The use of the stemmata has two
advantages. First, the relationship of the manuscripts is shown
at a glance, and also the relative importance of their readings
singly and in combination. Secondly, much less space is needed
for recording their readings ( 86 at end), X denoting the read-
ing of all the manuscripts (A, Q, B, M, S, a, f, h, /), a the
reading of five (A, Q, B, M, S), and so on.

A more complicated stemma may be studied on p. cxxxiv of
Gudemann's Dialogue of Tacitus:

TEXTUAL CRITICISM.

[Archetypon Fuldense (?), Corbeiense (?) ]
VIII.-IX. Cent.

[Codex Hersfeldensis (?)]
XIII. Cent.(?)

[Apographon ignoti inventoris]
XV. Cent. (c. 1457).

[X]

[Y]

V D C X A

ANCIENT TRANSLATIONS. As the testimony of the manuscripts 173
is indirect to an extraordinary degree, and as between the most
ancient of these and their originals hundreds of years, perhaps
even a thousand years, intervene, the editor looks eagerly for tes-
timony more nearly or quite contemporary with the original. The
most ancient testimony is furnished by translations into other lan-
guages, and while, as has been said ( 162), the Latin classics
derive no benefit from this source, they do throw some light upon
the earlier writings of the Greeks. It is well known that even in
the earliest times the Romans had translations from the Greek.
Fragments of the translation of the Odyssey by Livius Andronicus
have come down to us, and Cicero not only made set translations
of whole works, but filled his philosophical writings especially with
translations either made by him or taken from earlier Roman poets.
Such translations are very free, but those made by writers of the
early church and even in the middle ages are painfully literal,
almost word for word. It goes without saying that the freedom of 174
the most ancient translations detracts from their value for critical

IO6 LATIN MANUSCRIPTS.

purposes, and yet some assistance has been derived from this source,
as e.g., by Spengel and Rauchenstein on Phaedrus 279 A from
Cicero Orator 41. It is also evident that the help thus gained
will avail more in making a selection between two or more vari-
ants than in restoring a passage hopelessly corrupt. Another
fact that depreciates the value of this testimony is that it is trans-
mitted to us through manuscripts no older or better perhaps than
those which we have of the original, for instance all the manu-
scripts of the passage of Cicero referred to above are later by half
a century than the best manuscript of the Phaedrus. Worse than
this, it has been shown that some ancient translations have been
"corrected" in later times, that is, modified so as to bring them
into harmony with the corrector's text of the original.

175 .ANCIENT COMMENTARIES. Of far more importance are the
ancient commentaries or scholia ( 47) upon the masterpieces of
antiquity, many of which have been preserved to us. We have
complete commentaries, for example, upon the works of Hippo-
crates (fl. 400 B. C.) by Galen (f 200 A. D.) in eighteen books,
upon Aratus (fl. 270 B. C.) by Hipparchus (f 125 B. C.), and sev-
eral upon the writings of Aristotle and Plato. Besides these there
are extracts more or less valuable from commentaries upon Homer,
Aristophanes and the tragedians. Less assistance is given by the
commentaries upon the Latin writers but some of these are very
valuable, as, e.g., the scholia of Aelius Donatus on Terence (about
the middle of the fourth century), of Servius and Tiberius Donatus
on Vergil (also fourth century), of Porphyrio and the Pseudo-
Aero on Horace (date uncertain), all in a very unsatisfactory form.
All these commentaries are of more value from the standpoint of
interpretation than of textual criticism, but it was customary then
as it is now to prefix to the note one or more words of the text,
and of course many notes are concerned with the words them-

176 selves. Unfortunately, the lemmata, as the words from the original

TEXTUAL CRITICISM. 107

prefixed to the notes are called, have often been altered by later
students to fit the texts current in their times, and the precise
form of even a word which is discussed is seldom of moment to a
commentator concerned chiefly with the meaning. Besides, the
critic has here, as sometimes in the case of translations, to deal
with texts even more corrupt than those of the original upon
which he is engaged, for it is only very recently that an effort
has been made to settle on scientific principles the texts of these
commentators.

CITATIONS. To the apparatus criticus must be added the cita- 177
tions by later writers. So large a body of the post- classical lit-
erature has come down to us that the occasional quotations found
therein from the works of the classical writers are collectively very
numerous and of considerable importance. The works of the
Church Fathers are filled with quotations from the heathen
writers, e. g., St. Augustine with maxims from Sallust. Still
richer sources are the fragments of the grammarians and lexicog-
raphers ( 50). But here, too, the editor is apt to be disappointed.
Too often he finds numerous citations from precisely those pas-
sages of his author that present no critical difficulties, and none
that bear upon the term in doubt. Then conies the usual diffi-
culty in regard to the text of the grammarians and lexicographers
and saints whom he consults, and the additional vexation that
they often quote from memory only and sometimes give in differ-
ent passages different forms of the same quotation. Their value
for these reasons is, therefore, very differently estimated by differ-
ent editors.

IMITATIONS. Lastly, attention is now being given to the imita- 178
tions in late writers of favorite predecessors in the same style of
composition. It is evident that when such an imitation is estab-
lished the evidence may be made to point in either direction, i. e.,
from the better established text fonvard or backward to the text in

108 LATIN MANUSCRIPTS.

doubt. The weight to be attached to evidence of this sort is at
best not very great, but may serve to turn a nicely balanced scale.
In general it will be found to count more for the interpretation of
the author than for his text. Friedlander's Martial gives a good
idea of the use of these imitations.

179 USE OF THE APPARATUS. The apparatus having been collected
as completely and as accurately as possible, its use has next to be
considered. Here we must bear in mind the fact that by far the
larger part of the text of the average classic presents no critical
difficulty at all. For nine words out of ten, to put the case mildly,
the evidence against a given term or for it will be so over-
whelming, that no occasion for perplexity will be given. For the
tenth term no rules can be given that will have in practice any
general application or binding force, but these hints may be sug-
gested: First, the authority of the apparatus should be heeded
so far as it can be determined ; secondly, the fitness and appro-
priateness of every suggested term in itself must be considered;
thirdly, the possibility of deriving paleographically ( 93) the re-
jected term from the one received into the text has great weight.
It is evident that the second suggestion will depend largely upon
the editor's taste and his familiarity with the usage of his author.

180 Take an example: In two manuscripts of equal authority we
find a different number of words; the additional term in the one
is appropriate, its absence from the other causes no perplexity.
The editor will first ask how the term got into one copy if it was
not in the original. This will be easily explained if the same
word occurs near by or if it is a word often supplied or likely to
be added as an explanation. He will next inquire how the term
was lost from one copy if it stood in the original. This loss is
always possible, but it is especially easy and therefore probable if
it is a word not needed for the sense and likely to be lost by a

181 failure of the memory ( 141). Again, suppose that two terms

TEXTUAL CRITICISM. 109

occur ill two manuscripts of equal authority but in reverse order,
res publica and publica res. Here the editor will ask which is the
usual and natural order, which the unusual and artificial, and will
assume that the latter is the order in the original as being more
likely to be changed by the scribe. This brings us face to face
with Griesbach's famous canon for New Testament criticism: That
the more difficult reading is to be preferred to the easier, be-
cause the latter is more apt to be an alteration than the former.
This is true so far as intentional variations from the original are
concerned, but it is not true of unintentional errors. Uninten-
tional errors, however, are far more numerous than intentional
errors ( 125), and the canon is therefore of very limited applica-
tion. So it will be found to be with every rule that may be laid 182
down, there will always be exceptions and exceptions. Every editor
will consciously or unconsciously adopt rules for his own pro-
cedure, based upon his familiarity with the apparatus he is using
and varying with the author whom he considers. And even should
different editors of a given author agree upon the rules to be ap-
plied, a rare thing for editors to do, their decisions upon specific
applications of the rules would be sure to vary in many cases. It
was the failure to recognize the reasonableness of many of these
differences of opinion that caused the bitter feeling of the older
critics toward each other personally, a feeling that still finds vent
occasionally in our philological journals and reviews.

RELATIVE WORTH OF MANUSCRIPTS. We have now to con- 183
sider the meaning of such expressions as " greater or less manu-
script authority," " a better or poorer manuscript," etc. The
first editions were based upon such manuscripts as their pub-
lishers could procure, sometimes upon the first manuscripts they
chanced upon, and presented texts of little critical value. When
scholars began to turn from these editions to the manuscripts,
their first impulse was to count the manuscripts for or against
a given term, and give to the greater number the respect due
to superior authority. A little thought will show how utterly

HO LATIN MANUSCRIPTS.

unscientific such a procedure is : Suppose that in the tenth

century two copies, A and B, were made of the

*i manuscript X, which was then lost, and that in the

A~ 13 fifteenth century four copies, c, d, e, and /, were

< e j, made of B, which also was lost. Now when the

scholar of the sixteenth century found A and c for

a certain reading, and </, e and / for the only variant, he would

decide against the true authority of the manuscripts if he gave

his decision in favor of the majority.

184 The next step was to select some one manuscript, usually
on account of its age or the care showed in its writing, and to
vary from it only when its readings could not be made to yield
a satisfactory sense. This was a step forward, for age is pre-
sumptive evidence of worth ( 93), and so is careful writing
( 115), and so also is freedom from interpolation. But pre-
sumptive evidence is not enough: Suppose that in the tenth
century two copies, A and B, are made of the manuscript X,

x which immediately disappears. Suppose that A is

L carelessly made, while B is a good copy, and that

] in the fifteenth century a good copy, c, is made of

B, which then disappears. Now when the_ scholar

of the seventeenth century finds A for a certain reading and

c, later by five hundred years, for the only variant, he will

naturally side with A, but his decision may be wrong. Mere

age is evidently not enough.

185 Take these same manuscripts, A and B, and suppose that
from A is made a careful copy, d, and from B a careless copy,
c, and that these last copies only are preserved : who can tell
which is the better manuscript with no further information
than these data give? Take a last example: Suppose that

from A is made the careful copy ft, and from B the
careless copy k, and that the owner of k corrects

his copy by comparing it a few years later with B.
A and B are now lost, and in the seventeenth cen-
tury an editor finds that the interpolated manu-
script // has one reading and the uninterpolated manuscript k
another : where lies now the balance of authority, if he cannot
distinguish the second hand ( 84) in // ? If he can distin-
guish it?

186 TEST OF WORTH. From these illustrations it ought to be

evident that no rule of general application can be laid down
that will determine from external considerations only the worth

TEXTUAL CRITICISM. Ill

of any manuscript. Its own readings are the only index to its
worth, i. e., the manuscript which varies least frequently from
the correct text is the best. But, it is urged, we get our text
from the manuscripts, and to get the correct text we must
know what manuscripts have superior authority. Two consid-
erations will help to explain this apparent paradox. The first
has been mentioned already ( 179) : a very large part of the
text of every author is now and has been for five hundred
years, perhaps, critically certain. This of itself gives opportu-
nity to test the value of any manuscript, a process shown with
admirable clearness by Professor Pease in his treatment of the
manuscripts of Terence (Tr. A. Ph. Ass'n, 1887). In the second 187
place scholars make themselves so familiar with an author's
way of thinking and with his style of expression, worked out
from passages critically certain, that when they come to an
uncertain passage they are able to test the opposing manu-
scripts by their fidelity to the known usage of the author.
Both these considerations call for a considerable period of
study, extending over generations perhaps, and it is this long
and careful study that really tests the manuscripts. It is true
that when the best manuscript is found by some such process
as this, it will usually prove to be an old manuscript (as com-
pared with its fellows), and carefully written, and free from
interpolations, but no one of these qualities, no two or three of
them, is a certain indication of excellence.

CONJECTURAL EMENDATION. No matter how excellent and 188
numerous the manuscripts of a given author are, no matter how
complete the other materials ( 162) of the critical apparatus,
there will still remain occasional passages where all the help which
the apparatus renders cannot furnish a satisfactory text. At this
point textual criticism has reached the limits of its obligation,
beyond this it does not go. Scholars, however, are not content to
stop even here; they undertake by a process of divination, not of
criticism, to give to us the words written by the author, although
lost or distorted beyond recognition in the course of time. The
process is one which we all almost unconsciously employ to a
limited extent at least: the sentence we are reading does not end
at the bottom of the page, but we can guess a word or two more

112 LATIN MANUSCRIPTS.

before we turn the leaf; a word in a friend's letter is carelessly
written or blotted, but we divine the meaning from those that pre-
189 cede or follow it. The process, when we come to analyze it, is
this: the reader puts himself in the position of the writer and by
a purely intellectual effort tries to realize what the writer under
the given circumstances must have thought and therefore written.
No rules can be given for such a process. It depends essentially
upon the reader's ability to identify himself with the writer; it
calls not merely for the fullest intellectual sympathy and appre-
ciation, but also for that fullness of knowledge which is the goal
of modern scholarship, for the reader who understands the writer
best, will be most successful in his conjectures.

190 CRITICISM AND CONJECTURE. But while the act of conjecture
has nothing to do with criticism, criticism resumes its functions
so soon as the conjecture is made. All the critical tests ( 154-
156) grammatical, historical, individual, and technical, must be
applied. If these are satisfied by the conjectured term, it is ' pos-
sible.' It becomes ' probable ' if it will satisfy the one diplomatic
test that can be applied to it: can all the variants, or at least the
best attested variant, be derived from the conjectured term by the
processes (any or all) of corruption known to us from our study
of paleography? Further than the 'probable' conjecture cannot
go, although in the fullness of our enthusiasm over some unusually

191 brilliant suggestion we pronounce it 'certain.' Certain it becomes
only when it is confirmed by the discovery of new manuscript
authority, as, e. g., Reuter's conjecture on Plautus Trin. 297 and
Bishop Hare's on 313 afterwards found in A. This confirmation
is very rare, but it is the hope of such rewards which has made
conjectural emendation with some scholars a passion, almost a
madness. It ought to be remembered that it is no less a triumph
of scholarship to vindicate the soundness of a manuscript reading
impugned by others, than to emend a passage that has been
despaired of for centuries.

^^tf ki^ W-r^wfrinA, PimuHrWA.^.deLtro^

^&3&8&& fe^fere^&r,

l inefcefo .*** 9^*' c ^* ^skujijw ^rpz

- n e*,it*r*r l/v '- m

"*r.

rl&tnu

ttw

tint

lib,

*3^g3

feS^Sfe?**-r

,^^7n?- t

XIII. HORACE: lierncu. fX.

TEXTUAL CRITICISM. 113

LIMITS OF EMENDATION. As the most conservative defender of 192
the manuscripts must admit the necessity of some conjectural
emendation, so the boldest emendator admits the necessity of
restraint. To draw a line, however, and say to the followers of
Bentley and Madvig and Cobet and Dobree "Thus far and no far-
ther" is impossible. Professor Chase puts the case in a nutshell ,
when he says of Madvig's emendations of Livy that he feels less
sure "that the words Madvig gives are those which Livy actually
wrote, than that they are the best possible expression in Latin of
what Livy wished to convey."

OPPOSING VIEWS. A less temperate expression of the same 193
thought will be found on p. xxxvii of Professor Tyrrell's preface
to his edition of the Miles Gloriosus of Plautus, over against
which, for the sake of fairness, should be placed Professor Palmer's
outburst on pp. xxxviii-xliv of his preface to the Satires of Horace
in the same series, or Munro's defense of the emendation of
Horace on pp. xix-xxxii of the edition by King and Munro.
Bentley, the most subjective of all critics, has put his view of the
matter into the often quoted phrase: Nobis et ratio et res ipsa
centum codicibus potiores stint. Against this stands the fact that
modern editors do not accept one in a hundred of the changes
introduced by Bentley into his text of Horace.

INDIVIDUAL CRITICISM.

194 ' I S HE name we have chosen for the second division ( 168) of

criticism is not the common one. Higher criticism is the
name most frequently heard, because so frequently used by theolog-
ical writers. It might, perhaps, be well to leave this name to
denote the more restricted application of the science to the Scrip-
tures only. Philologists object to the term not so much on account
of its arrogant and invidious sound, as because it fails to describe
the function of this branch of criticism. As a descriptive name
the German Kritik des Echteu und Unechteu, criticism of the
genuine and the spurious, is especially apt and appropriate, but
long and unwieldy, and even among German scholars the term
Individual-Kritik is perhaps as common.

195 PURPOSES OF INDIVIDUAL CRITICISM. Individual Criticism is
concerned chiefly with the matter of authorship, and undertakes to
answer such questions as these: If a given composition is ascribed
by tradition to a particular author, did he produce it? If to more
than one, which (if any) of the number produced it? If to none,
who did produce it? Such questions are not new, they have en-
gaged the attention of scholars as far back as the history of phi-
lology goes. The very earliest students of classical literature were
in doubt as to what poems were to be assigned to Homer out of
the mass of what we now call cyclic literature. To early times
also goes back the custom of ascribing for the sake of profit infe-
rior works to writers of established reputation : only twenty-one out
of one hundred and thirty plays called Plautine in the time of
Varro were pronounced genuine by that great critic (Gell. Ill, 3).
Soon after the death of Horace spurious poems were circulated under
his name, and similar examples might be multiplied. Some such

INDIVIDUAL CRITICISM. 115

questions are still busily discussed that have engaged the attention
of scholars for generations (e. g., the Dialogue of Tacitus [?] ),
and new ones constantly are propounded. To understand the
process followed in investigating these problems it will be best to
put ourselves in the position of an editor examining the title of
his author to a given composition, as we did ( 161) in the case of
textual criticism. The evidence which he has to consider is partly
external and partly internal and may be examined conveniently
under these heads.

EXTERNAL EVIDENCE MANUSCRIPTS. The editor will first 196
examine the manuscripts to see how far the traditional authorship
is supported by their authority. The value of their testimony is
in no case very great, but varies with the age of the manuscript,
deliberate falsification being more likely in those of later date. In
any case the manuscript will hardly do more than show where the
"burden of proof" lies. Sometimes they will agree in assigning
the work to an author even when his title to it is known from
other sources to be weak, e. g., in the case of the rhetorical treatise
Ad Herennium, which all incorrectly assign to Cicero. Here, of
course, the burden of proof lies upon those who claim the work
for another author. Sometimes the manuscripts will be found to
disagree, or perhaps the best manuscript will offer a choice of
authors. The latter is the case with the valuable literary treatise
riepi ui/ous, once ascribed to Cassius L/onginus (third century) but
now generally attributed to Dionysius of Halicarnassus (f 7 B. C.),
the best manuscript of which has the title Ato^ucriou r\ AoyyiVou.
Finally the manuscript may suggest no author at all, as, e. g., in
the case of the Lives of Eminent Commanders universally pub-
lished now under the name of Cornelius Nepos.

ANCIENT WRITERS. The editor will turn next to the ancient 197
writers, scholiasts, grammarians, lexicographers, etc., for evidence
(^testimonia" see 162) older and therefore more valuable than

Il6 LATIN MANUSCRIPTS.

that of the manuscripts (177). He may find the title of the
given work among those credited to a certain author, in which
case he must consider the probability of several works on the same
subject having the same title, or better he may find among the
citations from a certain author passages which have come down to
us in the given work only. The value of such testimony turns
upon two points. First, the competence of the witness must be
considered, for it is evident that a writer will speak with greater
authority upon the works of a friend than upon those of a
stranger, upon works in his own peculiar province than upon
those in a less familiar field. So we will instinctively heed the tes-
timony of Aristotle the pupil in regard to the works of Plato the
teacher, of the younger Pliny in regard to the works of his uncle,
of a professional rhetorician in regard to works of rhetoric. Again,
some writers have a better reputation than others for care and
198 painstaking. In the second place, the character of the testimony,
affirmative or negative, must be taken into account. Except in
the case of the expert testimony, just mentioned, affirmative evi-
dence, i. e., evidence supporting the current tradition, is not of
great value: it simply carries the tradition back a step further
than the manuscripts. Negative testimony is much more impor-
tant, especially that of experts, for it shows either that the tradi-
tion is later than the time of the witness, or that he had reasons
for disbelieving it. A good illustration is furnished by the Ad
Herennium mentioned above. All the manuscripts (and the oldest,
Parisinus 7714, dates from the ninth century) ascribe the work to
Cicero, and so do Saint Jerome (331-420 A. D.) and the famous
grammarian of the sixth century Priscian; but Quintilian, III, i,
21, mentions Cornificius (fl. 90 B. C.) as the author, and to him all
scholars now assign the work. In this case the negative testi-
mony is aided by the date of the witness (about 35-100 A. D.), his
right to speak upon rhetorical questions, and his reputation for
carefulness and truthfulness in his statements.

INDIVIDUAL CRITICISM. 117

INTERNAL EVIDENCE 'HISTORICAL. From the external testi- 199
mony the editor will turn to that furnished by the work itself, and
will first apply the tests of historical criticism ( 155): Does the
work, in the first place, betray an ignorance of facts which must
have been known to its supposed author? Does it, in the second
place, show an acquaintance with events that occurred only after
the time of the supposed author? This is perhaps the safest test
that can be applied to a work internally, for it leaves least to
depend upon the partiality or prejudice of the editor. The results
will depend upon the character of the work, which may or may
not contain many allusions capable of historical identification, and
also upon the greater or less precision with which we can deter-
mine the time and the place of the reputed author. One thing
the editor must consider with great care, the possibility of inter-
polation. Diplomatic tests ( 155) will reduce this danger to a
minimum, but for this minimum due allowance must be made.
This allowance will vary of course with the editor.

INDIVIDUALITY. Next comes the agreement of the views ex- 200
pressed in the given work with the reputed author's views upon
the same subjects as known to us from other sources. An editor
would justly scout the idea of Sallustian authorship for a pam-
phlet in support of the nobility, or of defending as the work of
Horace a newly discovered epic in praise of Augustus. The
results in this case will depend, of course, upon the fullness and
accuracy of our knowledge of the views of the traditional author,
and due allowance must be made for alteration in these views in
the case of a man whose literary career covers a long period of
time, and also for the freedom or reserve with which he expresses
his views under different circumstances. Cicero's changed political
views between the impeachment of Verres (70 B. C.) and the de-
fense of Sulla (62 B. C.) furnish a case in point, and so does the
different tone with which he treats of matters of state in his con-

Il8 LATIN MANUSCRIPTS.

fidential letters to Atticus and in liis public deliverances upon
these same themes from the rostra or in the senate.

201 LANGUAGE AND STYLE. Next comes the linguistic test, fur-
nished by grammatical criticism ( 154). Sometimes very general
considerations will be sufficient for a negative decision. No fact
is more familiar than the growth and development of a language,
the change from age to age in forms, syntax, prosody, vocabulary.
If a work written at one time is assigned to an author of a widely
different time, the study of the language will surely detect the
error. The same thing is true of variations due to differences in
place of birth or education, but in less degree: does not Asinius
Pollio charge Livy with Palavinity (provincialism)? In most cases
for negative testimony, however, and in all cases for affirmative
testimony, these general considerations will not suffice. The solu-
tion of the problem will require that a thorough comparison be
made between the language of the work under consideration and
the language of the author known to us from other sources. How
minute and thorough must be the editor's knowledge of both only
the highest scholarship can realize, and besides this we must have
enough writings of the reputed author, unquestionably genuine, to
make the test extensive as well as minute. Under these condi-
tions it is probable that this test is the safest and surest of all,
but we must be very sure that the conditions are fulfilled: the
supposed genuine works may be in part spurious, the comparison
may be partial, the editor's knowledge may be defective ( 160).

202 The early application of the test is shown in Varro's judgment
upon the plays of Plautus (Gell. Ill, 3, i and 2), and is ascribed to
Caesar by Cicero (Fam. IX, 16, 4) ; the danger is shown by Cicero
no less clearly in the famous letter (Att. Ill, 12, 2) in which he
suggests that the authenticity of one of his genuine compositions
might be successfully denied because it was written less carefully
than usual. Modern scholars, famous for critical talents, have been

JB-

tto.cftfccc/it " Xcpoflpcrulo- cognofarmonurn (X

faac"'fl-m
tluf

circum uerj iAnc-

fi ccrn fef borccm? r-- tma>lio7-ref

ctcouecuctaerc-oppzcto pp

Arf* ficcenflurnifiaem

<rerurtir'
<nenc-

Titto

1 dof>n du

ubt ^f

' iw

f . t

aein.-imduj; ftT77ulcLxmor- fn

r c-

aro.'t<9?<xut- erocrrr ' ~&wn cuyadfamccrn iiccer-axzur-

' AT rr n?f en

r r

' n VA

a< ^ r

XIV. SALI.UST: /^n-isiniix i ()(>>$. Sari. iX X.

INDIVIDUAL CRITICISM. 119

deceived again and again. Muretus managed to insert a few
trimeters of his own among some verses of Accius without detec-
tion by Scaliger, who published them without question in his first
edition of Varro, and Wolf, failing to find in the printed editions
of Cicero a letter ascribed to him in a manuscript in the library
at Berlin, pronounced the letter spurious from internal considera-
tions, but was forced to retract when a pupil showed him that he
had merely looked for the letter in the wrong place in the book.

FORGERIES. In cases where one author has deliberately 203
imitated the style of another, no matter whether he intended
to pass the spurious work off as genuine, or whether in after
times the mistake was made by others, the problem is much
harder to solve. It may be taken for granted that the imitator
would conform as closely as possible to the historical condi-
tions, would reproduce only the best known and safest views
of his model, and would follow closely his peculiarities in style
and language so far as these would be known to him. Detec-
tion from internal considerations will depend entirely, therefore,
upon the ratio between his knowledge and that of modern
scholarship, a ratio that is constantly changing in favor of de-
tection. External tests are the main reliance, however, in cases
of this sort. Excellent material for practice of a sort not too
difficult may be found in the letters and speeches which Sal-
lust has inserted in his account of the Conspiracy of Catiline,
especially the letter of Lentulus (Cat. xliv, 4) and the speech
of Caesar (Cat. li).

TESTS OF PROPOSED AUTHORS. Now if from all the evidence 204
thus obtained the editor regards the case as proven against the
traditional author, he next proceeds to find the real author, and
the problem is not unlike those proposed for solution in textual
criticism. If the manuscript or the ancient authorities suggest
another author than the one usually received (as in the case of
the Ad Herennium mentioned above, 198), or other authors, the
editor will assume such author, or such authors one by one, as
real and will then apply all the tests which he has used in the case
of the traditional author. If none stands the test, the problem
now resembles that of conjectural emendation: the editor endeavors

120 LATIN MANUSCRIPTS.

to find some author who could have written the given work, a
process of divination pure and simple. If he hit's upon one his-
torically possible, the tests are again applied, and if the results
are satisfactory the probable ( 190) author is found. So it stands
with Cornelius Nepos, the argument for whose authorship of the
Lives of Eminent Commanders is given, as an instructive illustra-
tion, in the next section. Further than the probable this process
cannot go, and hence the energy with which it is pursued, and the
envy and hatred and malice and all uiicharitableness which such
an investigation discovers as that of Professor Gudemann into
the history of the discussion over the authorship of the Dialogue
of Tacitus.

205 ILLUSTRATION OF PROOF. There have come down to us
and are now published under the name of Cornelius Nepos
twenty-three short lives of Eminent Commanders (Vitae or
Liber de Excellentibus Ditcibns), not Romans. It is to be
noticed that these are not only not ascribed by any ancient
authority to Nepos, but are apparently assigned by the manu-
scripts to Aemilius Probus (of the time of Theodosius). The
authority of the manuscripts is easily disposed of by considera-
tions that need not be discussed here, but the claim of Nepos
to the authorship of the Lives rests upon the following argu-
ments only :

206 Several passages in the Lives (17, 4, 2; 18, 8, 2; i, 6, 2)
seem to have been written during the transition between the
republic and the monarchy, and the last two of these seem to
refer to the year 36 B. C. Now, at this time lived T. Pompo-
nius Atticus, and the Lives are dedicated to an Atticus, whose
nomen and praenomen are unfortunately not given. About this
time moreover lived also Cornelius Nepos, a close friend of T.
Pomponius Atticus, who is known from other sources (Cell, n,
8, 5, and see the other refs. in Teuffel 198, 5) to have written
"Lives" of eminent men. The inference that the extant Lives
without an author are by this author without other extant
"Lives" is natural, and becomes almost irresistible (i. e., prob-
able) when we find (i) that these Lives are followed in all
extant manuscripts by lives of Cato and T. Pomponius Atticus,
which are known to be by Cornelius Nepos, and (2) that these

INDIVIDUAL CRITICISM. 121

lives of Cato and Atticus show the same characteristics in sub-
stance and diction, the same kind of generalization, and the
same tendency to exculpate and exalt their subjects as do the
Lives of Eminent Commanders. All modern scholars, therefore,
of reputation agree in assigning the authorship of the Lives of
Eminent Commanders to Cornelius Nepos.

SUMMARY. It is greatly to be regretted that there is no his- 207
tory of Philological Criticism in English, no biographical dictionary
even, to which the student can turn for information about all the
scholars whose names he finds in philological publications. A
sketch of the development of Textual Criticism may be found in
Harper's Dictionary of Classical Literature, with a brief bibliog-
raphj 7 . In Gow's Companion, p. 66 f, is given a list of famous
critics and scholars, and some account of the greatest of these may
usually be found in the biographical dictionaries and encylopedias,
if they are not mentioned in Harper's Dictionary. Besides, the
introductions to the more elaborate editions commonly describe the
contributions made to the criticism and elucidation of the several
authors by the scholars who have worked upon them. Students
who can read German will find abundant helps at their disposal, 208
the most important of which are named in the bibliographies ap-
pended to the several essays in Miiller's Handbuck, Vol. I.

In general it may be said that the greatest advances made in
criticism, with few and rare exceptions, have been made in very
recent times. This is largely due to the increased attention given
to Paleography, a study made possible by the cheapening of fac-
similes on account of improved methods of pictorial reproduction,
but more largely perhaps to the general spread of scientific methods
through all branches of study: fas est et ab hoste doceri! It can
hardly be said that there are now any "opposing schools" of crit-
icism, so far as classical philology is concerned, however individ-
uals may differ in their methods and results. Differences there are
in the texts of even those authors that have been most carefully

122 LATIN MANUSCRIPTS.

studied, but these differences are on the whole very few and insig-
nificant, and it may be affirmed that the texts of all the great
classics of Latin literature, with the possible exception of Plautus,
are to-day more trustworthy and nearer to the original than is, for
example, the accepted text of our English Shakespeare.

a **>-ttUu:

umm^ ^^rcu^^^'^rnpo P

^ V>annt,,

tiarurCiuir

? dafLt&dtuirw

r ttyc^rf^ftul&~ ft & w euiquiabrum

i egp gte(tu>td<Sb ^4u<dtC^fmagnd,d curartrpar-

nn~\ N

cJ/t>imtrrifn r iu^C5C s t> w<?

-n

^^t VJI

X\". CICERO:

7v>/.

DESCRIPTION OF PLATES.

LIST OF PLATES.

OPP.
PAGE

I. VERGIL : Palatinus, Saec. IV- V . Frontispiece

II. CICERO: Schedae Vaticanae, Saec. IV-V . .22

III. TERENCE: Bembinus, Saec. IV-V . . 26

IV. C/ESAR: Parisians 5763, Saec. IX-X . . 34
V. VERGIL : Schedae Vaticanae, Saec. IV . . 40

VI. SALLUST: Parisinus 16024, Saec. IX-X . . 50

VII. CAESAR: Vindobonensis 95, Saec. XII-XIII . 56

VIII. CICERO: Ambrosianus C. 29, p. inf., Saec. X . 62

IX. CICERO : Rhenaitgiensis 127, Saec. XI . . 80

X. VERGIL : Sangallensts 1394, Saec. IV . . 92

XI. CATULLUS : Romanus, Saec. XIV . . 96

XII. HORACE: Bernensis 363, Saec. IX . . . 102

XIII. HORACE: Bernensis 363, Saec. IX . . 112

XIV. SALLUST: Parisinus 16025, Saec. IX-X . .118
XV. CICERO: Vossianus Fol. 86, Saec. X . 122

XVI. CICERO: Parisinus 18525, Saec. XII . .126

DESCRIPTION OF PLATES.

T^AC-SIMILES are given of one or more manuscripts of each of 209

the following authors: Caesar (Plates IV and VII), Catullus
(XI), Cicero (II, VIII, IX, XV, and XVI), Horace (XII and
XIII), Sallust (VI and XIV), Terence (III), and Vergil (I, V,
and X). Of equal value for purposes of study are the fac-siniile
of a fragment from Sallust's Histories (Fig. 12, p. 69) and the
reduced specimen of the Munich Livy (Fig. 17, p. 76). The speci-
mens are taken from the authors read early in our courses of
study and have been selected to illustrate the styles of writing
described in 95-115 rather than to represent the apparatus
criticus of the several authors.

Plate I, Frontispiece. The Codex Palatinus, or Codex Vaticanm 210
/<5j/, of VERGIL, Ribbeck's P. Written in rustic capitals ( 103)
of a style as early as the third century, and variously dated ( 65,
102, 115) from the fourth to the fifth. The volume has 571
leaves, counting the sheets of blank paper inserted by the binder
between every two leaves of parchment; 33 leaves have been lost.
It belonged originally to the library in Heidelberg ( 74) but was
taken to Rome in 1623 ( 76), thence to Paris in 1797 and
returned to the Vatican in 1815. The page given contains lines
277-299 of the first book of the Georgics. The letters barely dis-
cernible at the ends of the lines are from the opposite side of
the leaf.

Plate II, p. 22. The Schedae Vaticanae, or Palimpsestus Vati- 211
canus 5757, of CICERO'S De Re Publica. Uncial writing of the
fourth or fifth century ( 66, 106) covered by half-uncials of the
eighth. This is the famous codex rcscriptus originally at Bobbio
( 60), which was discovered at Rome by Cardinal Angelo Mai

126 LATIN MANUSCRIPTS.

(1782-1854) and published by him in 1822. The manuscript is
badly mutilated ( 67), whole quires as well as single leaves being
missing. The specimen page gives De Re Publica I, xvii, 26-27.

212 Plate III, p. 26. The Codex Bembinus, or Vaticanus 3226,
of TERENCE, Umpfenbach's A. It is written in rustic capitals
(Thompson's Palaeography, p. 189) of the fourth or fifth century,
called uncials by some editors. It had originally 114 leaves, but
16 whole leaves and parts of others are wanting. It was owned
by Cardinal Pietro Bembo (1470-1547), passed into the hands of
Fulvius Orsini ( 76) and is now ( 214) in the Vatican. It is the
oldest and best of the manuscripts of Terence, and one of the best
manuscripts of Latin authors preserved to us. The specimen
gives Phormio 179-223. The codex has valuable marginal notes.

213 Plate IV, p. 34. The Codex Floriacensis, or Codex Parisinus
5763, of C^SAR, Meusel's B. Minuscule writing of the ninth cen-
tury. The volume has 180 leaves, of which leaves 1-112 contain
Caesar's Gallic War, and leaves 113-180 books xiii-xvi of Jose-
phus. The manuscript was anciently in the monastery of St.
Benedict at Fleury-sur-Loire, and is variously known as Colberti-
nus 897 and Regius 4 ^-. Besides the subscription of Constantinus
( 52) it has at the close of book II:

FLAVIUS LICERIUS FIRMINUS LUPICINUS LEGI.
This subscription is supposed to date from the sixth century.
The codex has marginal readings from other manuscripts, is
incomplete, and belongs to the "first class" of the manuscripts of
Csesar, marked a in the stemma on p. 104. The specimen gives
the close of book II and the beginning of book III.

214 Plate V, p. 40. The Codex Vaticanus 3256, or Schedae Vati-
canae or Puteanae, of VERGIL, Ribbeck's A. Written in square
capitals ( 100, 102) and now generally assigned to the fourth
century ( 65) although at one time believed to be much older
( 115), even of the time of Augustus, and hence called the

t irot ; fWifclVTl^wiAiw -C^r (i }<&:^ i C I f- 4 ""

T- S; nvfcv>q?*&e tmrnafcfifrvA-ufrt f$u ivnfejudjmf martaSm nrtcluttf'

^in

l *nd>;

j
.

X\'I. CICKRO: /-tins/nil^ V//.

DESCRIPTION OF PLATES. 127

"Augustean Fragment." Of this manuscript, possibly the oldest
extant of any Latin classic, seven leaves only are preserved, four
in the Vatican and three in the Royal Library in Berlin (Schedae
Berolinenses), containing Georgics I, 41-280, and II, 181-220.
The manuscript was once in the library of St. Denis in France,
but the time of its mutilation is unknown. The Vatican leaves
have the memorandum: Claudius Puteanus Fulvio Ursino d. d.,
and it is known that Du Puy gave them to Orsini in 1574-75.
When Orsini died in 1600 they passed with his other books into
the possession of the Vatican. The history of the leaves at Berlin
cannot be traced so far back. A fragment of the same manuscript
is known to have been in the library of Pierre Pithou (scholar
and jurist, 1539-1596), but is now lost. From it, however, four
verses, Aen. IV, 302-305, are preserved in a fac-simile made for
the second edition of Mabillon's De Re Diplomatica ( 90). The
specimen page gives Georgics I, 61-80.

Plate VI, p. 50. The Codex Parisinus 16024, or Sorbonianus 215
500, of SALLUST, Jordan's P. Minuscule writing of the ninth or
tenth century. The manuscript had originally at least 190 leaves,
of which 144 are lost from the beginning. The remaining 46
leaves contain the Catiline and Jugurtha with the lacuna in the
latter (J. ciii, 2-cxii, 3) which characterizes all the manuscripts of
the first class. The last page cannot be read, having been pasted
upon a piece of blank paper, perhaps by a binder, and the pre-
ceding page ends abruptly with the words fuit ante diem ( J. cxiii,
3). This is the best manuscript of Sallust, the only one of the
first class whose readings are quoted separately by Jordan. The
specimen page gives the close of the Catiline and the beginning
of the Jugurtha.

Plate VII, p. 56. The Codex Vindobonensis 95 of CAESAR, 216
Meusel's /. Minuscule writing of the twelfth or thirteenth cen-
tury. The manuscript has 182 leaves containing besides the Gallic

128 LATIN MANUSCRIPTS.

and Civil Wars the Alexandrian, African and Spanish Wars. It is
a manuscript of the second (ft) class, see 53 and the stemma p.
104. The specimen page gives Bell. Civ. I, xxv, 6 p0sset-x.-x.vii,
2 quod ab. The scribe omitted a few words in line 18 of the sec-
ond column, indicating the omission by the letter d (i. e., desunt],
and added them at the bottom of the page, preceded by the letter
h (i. e., haec}.

217 Plate VIII, p. 62. The Codex Ambrosianus C. 29, part. inf.
of CiCERO, Baiter and Kayser's A. Minuscule writing of the tenth
century. The codex has 158 leaves, containing the De Omciis, the
orations against Catiline, and those for Marcellus, Ligarius and
Deiotarus. It was once the property of Cardinal Federigo Borro-
meo (1564-1631), the founder of the Ambrosian library. The
specimen gives the beginning of the first oration against Catiline.

218 Plate IX, p. 80. The Codex Rkenaugiensis 127 of CICERO,
Baiter and Kayser's R. Minuscule writing of the eleventh cen-
tury. The manuscript has 62 leaves, containing the Cato Major
and the orations against Catiline. The Laelius originally preceded
these but has been lost. The specimen gives the close of the Cato
Major and the beginning of the first oration against Catiline.

219 Plate X, p. 92. Codex Sangallensis 1394, or Schedae Sangall-
enses Rescriptae, of VERGIL, Ribbeck's G. Square capital writing
( 65) of the fourth century. The volume is composed of a num-
ber of fragments gathered by Ildefonse d'Arx in 1822. Of Vergil
eleven leaves only remain (and of these three or four were written
over in the twelfth or thirteenth century) containing Georg. IV,
345-566; Aen. I, 381-418 and 685-722; III, 191-228 and 457-532;
IV, 1-38; VI, 688-724. The specimen gives Aen. VI, 688-705,
with verse 678 inserted between 695 and 696. Notice the correc-
tion of inter to iter in the first line.

220 Plate XI, p. 96. The Codex Romanus, or Ottobonianus 1829,
of CATULLUS, Male's R. Minuscule writing (North Italian Gothic)

DESCRIPTION OF PLATES. 129

of the close of the fourteenth century. The manuscript has 73
leaves, containing Catullus alone (19). It was discovered ( 88)
by Professor Wm. Gardner Hale in 1896 in the Vatican library,
and is considered by him to be of the same rank as O and G,
standing in the same relation as G to the "lost Verona" manu-
script ( 168). It is the most carefully and beautifully executed
manuscript of the three and the richest in variant readings. A
full collation will shortly be published by Professor Hale, and the
Vatican will give out at the same time a complete fac-simile. The
specimen page is the first of the codex.

Plates XII and XIII, pp. 102 and 112. The Codex Bernensis 221
j<5j of HORACE, Keller and Holder's B. Minuscule writing (Irish
Hand, see 105) of the ninth century, the oldest manuscript of
Horace extant. It has 197 leaves, of which leaves 167-186 con-
tain parts of Horace arranged irregularly (see Orelli's preface).
It bears the name of Bongars ( 71) and was once in the library
of Fleury-sur-Loire, to which it was supposed to have been brought
by Alcuin (735-804) or one of his fellow-workers. Plate XII
reproduces the obverse of leaf 167, containing a short life of
Horace, of no particular value, and the first of his odes. Plate
XIII shows the reverse of leaf 168 containing Carm. I, xxii, 9-24;
xxxii, 1-16; xxxviii, 1-8; II, ii, 1-24; iv, 1-4.

Plate XIV, p. 1 1 8. The Codex Pansinus 16025, or Sorboni- 222
anus 7576, of SALLUST, Dietsch's P'. Minuscule writing of the
ninth or tenth century. The codex has 47 leaves containing the
Catiline and Jugurtha with the lacuna ( 215) marking the manu-
scripts of the first class. It is little inferior to P, although not
quoted separately by Jordan. The page reproduced contains
Jugurtha Ivi, 2 praeterea-\\\\\, 3 editiorem. The glosses are as
follows :

Marginal In tempore, subaudi necessitates \ id est oportune
Id est querere frumcntum Nota jnobilitatem Numidarum

130 LATIN MANUSCRIPTS.

glans Remissis, id est ociosis, cessantibus Ediciorem, id est
apciorem excelsiorem.

Interlinear de signo dato perticas bant in several places
to show the force of the historical infinitive over which it is
written

223 Plate XV, p. 122. The Codex Vossianus Fol. 86 of CICERO.
In the library of the University of Leyden, minuscule writing of the
tenth century. The manuscript has 192 leaves containing parts of
Cicero's philosophical works. The page here reproduced gives De
Natura Deorum II, Ixvi, 165 Ergo et to the end. A later hand
has added the marginal note, Finitur disputatttm Balbi and the
caricatures.

224 Plate XVI, p. 126. The Codex Parisinus 18525 of CICERO.
This codex, written in minuscules of the twelfth century, has but
six leaves left, containing fragments of the first and second ora-
tions against Catiline. The specimen shows Catiline I, xii, 29
vocibus to the end and the beginning of II. The last four lines
give a summary* of the second oration, as follows:

Superiore libro Catihna circumventus eloquio Ciceronis spon-
taneum elegit exilium, unde oratori maxima venisse videbatur
invidia; sed postero die timore dissimnlato processit ad popuhim
fingens se timere quod emiserit Catilinam ut minus sit invidio-
sum quod in exilium expulerit. Prohemium (i. e., prooemium)
sumptiim ab exultatione diccndis verbis pcne triumphantibus qui
sine damno populi Romani belhim superare potuerit.

.NDEX.

INDEX.

References are to the sections.

Abbreviations : for names of Manu-
scripts, 81; in Manuscripts, no, 114,

132, '45-
Accidental Errors: of the Eye, 129-136;

of the Ear, 128; of the Memory, 137-

141; of the Judgment, 142-146.
Addition of unimportant words, 141.
Age of Manuscripts, 12, 24,65; as test

of value, 93, 184; tests of age, 115.
Aldus Manutius, 62.
Alexandrian Scholars : divide the Rolls,

11; measure them, 36; as critics,

195-

Ancient Commentaries, 175. 177; Crit-
icism, 195; Manuscripts, 65; Transla-
tions, 173, 174.

Apparatus Criticus, 162-178; Manuscripts,
i63f; Translations, 173; Commen-
taries, 175; Citations, 177; Imitations,
178; Use of Apparatus, 179.

Archetype, 118, 170, 171.

Aristotle's Constitution of Athens, 12.

Atramentum, 5. See Ink.

Atticus, as publisher, 30, 37.

Authors: position of, 25; relation to
patrons, 26; to publishers, 27; prop-
erty right in books, 26; complaints
f 37> 38. See also Publisher, Li-
braries.

Authorship : test of, 196-203; proof of,
205, 206.

Barbarian Invasions, 56.

Bookhands, 98, 105.

Books : papyrus, 6; parchment, 16; manu-
facture, 32; cost, 35; measured, 36;
titles, 40; copyright in, 26; market
for, 29,34; published without author's
consent, 27; rare books used as copy,
32; corrected, 37.

Bookstores, 29, 31. See Tabernae.

Coesar : leading Manuscripts, 82; sub-
scriptions, 52,53; stemma, 170.

Calamus scriptorius, 5.

Calliopius, 52, 53.

Capital letters, 95, 96, 98, 100, 102, 103,
115; similarities between, 129.

Capsa, 8.

Caroline minuscules, 109, 112.

Causes of critical doubt, 152-158.

Chartae, 4.

Christianity, opposition to, 51.

Church, influence of, 58.

Cicero : his publisher, 30; libraries, 29;
complains of errors, 37; books pub-
lished without his consent, 27.

Citations in later authors, 177.

Classics, canon of, 46.

Clergy, ignorance of, 57, 58.

Codex: meaning of word, 16; form, 16;
size, 19; convenience, 20, 33; Codex
rescriptus, 23, 67, see Palimpsest; de-
scription of Codex, 116, 117. See also
Books, Parchment.

Collations of Manuscripts, 86, 87, 88,
163.

Conjectural Criticism, 151, 188-193; ob-
jection to the name, 158. See Emen-
dations.

Conjectures confirmed, 191; "probable"
and "certain," 190.

Constantinus, 52.

Cornua, 7.

Correction of Errors, 37. See Errors.

Correctors: in authors' times, 37; later,

52, 53-

Cost of Books, 35.
Criterion of the Appropriate, 160.
Critical Editions, 71, 89.
Criticism : Science of, see Contents;

Higher and Lower, 151,155, 194-206;

132

INDEX.

133

References are to the sections.

Diplomatic, 151, 157, and see Text-
ual; Conjectural, 151, 188-193; Gram-
matical, 151, 154; Historical, 151,
155; Aesthetic, 151, 156; Individual,
194-206; applied to Emendations,
190.

Cursive Writing, 96, 97; little to do with
Classics, 97; influence on bookhand,
108.

Dark Ages, 56.
Decline of Literature, 43.
Descriptions of Manuscripts, 71.
Dictation, 33; errors from, 128.
Diplomatic Criticism, 151, 157, 190, and

see Textual.
Diplomatics, 91.
Dittography, 133.
Divisionof Words, see Separation; wrong

division, 143, 144; of sentences, etc.,

i45-

Doubt, critical, 152, 153; causes of, 154-
156; removal of, 157, 158.

Ear, errors of, 128.

Editions, 34; by subscriptors, 51, 52;
first, 64, 165.

Education, 45; disappearance of, 57;
revival of, 60.

Emendation defined, 158; conjectural,
190-193; possible, probable and cer-
tain, 190; limits of, 192; opposite
views of, 193.

Epigraphy, 2, 91, 95.

Errors : in the authors' time, 32, 37; in
our Manuscripts, 119, 149; classifica-
tion of, 120.

Examination, critical, of Manuscripts,
164, 165; possible results, 166-169.

Families of Manuscripts. See Stemmata.

Fasces, 8.

Faulty Manuscripts, 119.

First Editions, 64, 165.

Forgeries, detection of, 203.

f routes, 7.

Glosses, 49, 146.
Gothic Letters, in.
Grammarians, 50, 175.
Grammatical Criticism, 151, 154.

Griesbach's Canon, 181.

Half- Uncials, 105, 107.

Hands ; first and second, 84, 85, 146,

185; National, 97.
Haplography, 134.
Higher Criticism, 151, 155, 194-206, see

Individual; objections to the name,

. '55. 194-
Historical Criticism, 151, 155, 199.

Ignorance of all classes, 57, 58; of copy-
ists, 63, 125.

Illustrations in Manuscripts, 7, 102.

Imitations, 178.

Inconsistencies in Names and Symbols
of Manuscripts, 69, 70, 71, 81-83.

Individual Criticism, 194-206. See
Contents.

Individuality as test of Authorship, 200.

Ink, for papyrus, 5; for parchment, 15;
colors, 15; washed away, 5; im-
proved, 15; scraped off, 23.

Instruments for writing, 5, 15. See Ink,
Pens, Stilus.

Intentional Errors, 125-127.

Internal Evidence of Authorship, 199.

Interpolations, 146.

Irish hand, 97, 105.

Judgment, Errors of, 142-146.

Keeping of the Manuscripts, 68-89.
Kinds of Criticism, 159.

Lacunae, 121.

Learning, Indifference to, 57; revival
of, 60.

Libraries first brought to Rome, 29; pri-
vate, 29; public, ancient, 44; modern,
72-80; use of Manuscripts in, 68;
Latin names of, 78-80.

Librarii, 29, 30.

Limits of Emendation, 192; of Textual
Criticism, 188.

Linguistic test of Authorship, 201.

Lipography, 134.

Literal Criticism, 154.

Lombard Hand, 97.

Lost Manuscripts, 61, 168; works, 55.

Lower Criticism, 151. See Textual.

INDEX.

References are to the sections.

Majuscules, g8f; errors in, 129.

Manuscripts: lost, 61; names changed,
70; in apparatus criticus, 163; rela-
tive worth, 183-187. See Contents,
and also Codex, Book, Papyrus,
Parchment.

Marginal Notes, 47, 49, 84, 124.

Membrana, 14. See Parchment.

Memory, errors of, 137.

Merovingian Hand, 97.

Minuscules, 108-115; 131.

Monastic establishments, 59, 60.

Mutilated Manuscripts, 67, 121.

Names : various, for same Manuscripts,
69; of Manuscripts commonly used,
78, 79; of places, Latin, 72-80.

National Hands, 97.

Numismatics, 2.

Offenses, critical, 152-158.

Oldest Manuscripts, 65; printed classics,
6 4 .

Omission of unimportant words, 141.

Opposition to Christianity, 51.

Ovid : Metamorphoses published with-
out his consent, 27.

Paleography, 90-149. See Contents;
practical, 116-149; theoretical (his-
torical), 90-115

Palimpsests, 5, 23, 59; preserved, 65, 66;
value of, 67.

Papyrus: relation to parchment, 3; manu-
facture, 4; size of sheets, 4; instru-
ments for writing on, 5; palimpsest,
5; rolls, 6; odd forms of, 18; pre-
served, 12, 65.

Parchment : relation to papyrus, 3; his-
tory of, 13, 14; instruments for writ-
ing on, 15; books, 16; odd forms, 18;
superior to papyrus, 20; size of
sheets, 16; of books, 19; tardy use,
22; age of books, 24; oldest pre-
served, 65, 66; first books written
on parchment, 21.

Patron and Author, 25.

Pens for papyrus, 5; finer for parch-
ment, 15; metal, quill, 5;

Pergamena, 14. See Parchment.

Pergamos, 14.

Period of Decline,43 ; of Transmission, 42.

Petrarch, 60, 61.

Plays circulated for reading, 28.

Pre-Caroline Cursive, 105.

Preservation of Rolls, 12; of codices, 65;
of Literature, 43 f.

Printing, 62; source of our letters, 100,
113; of German letters, in; first
editions, 64.

Proof of Authorship, 205, 206.

Publication of Books, 25-41; uncom-
mercial, 29; commercial, 31.

Publishers, 31; relation to Authors, 26.

Punctuation in majuscule writing, 101;
minuscule, no; errors in, 145.

Purpose of Individual Criticism, 195.

Quill pens, 5.
Quires, 16.

Rapidity of Publication, 34.

Reading the Rolls, 9.

Recensions of Early Scholars, 52, 53.

Relative worth of Manuscripts, 183-187.

Revival of Learning, 60.

Rolls: how made, 6; read, 9; size, 10;

ready made, n; preserved, 12. See

Papyrus.

Roman Uncial, 105.
Ruling Parchment, 15.
Rustic Capitals, 100, 103.

Scholia, 47, 146, 175.

Schools and contemporary Authors, 31;

preserve literature, 45; corrupt it, 54;

revival of, 60; school books, 47;

against Christianity, 51.
Scrinium, 8.

Selections from complete works, 46.
Separation of Words: in Capitals, 10 1 ;

uncials, 104; minuscules, no; errors

in, H3-

Sheets : sizes of Papyrus, 4; parchment,
16; ruled, 15; quires, 16; arrange-
ment of, 16; lost or transposed, 122.

Similar sounds, 128; letters, capital, 129,
130; minuscule, 131.

Size of Rolls, 10; codices, 19; papyrus
sheets, 4; for poems, etc., n; parch-
ment sheets, 16; editions, 34; meas-
urement of, 36.

INDEX.

'35

References are to the sections

Skipping, 135, 136.

Slaves as copyists, 29, 30, 32, 33, 36.

Sounds confused, 128.

Specimens, of Capitals, 102; rustic capi-
tals, 103; uncials, 106; half-uncials,
107; minuscules, 113.

Spelling as a test of Manuscripts, 115;
concurrent spellings, 128.

Square Capitals, 100, 102.

Stemmata, 170-172; of Caesar, 170; Taci-
tus' Dialogue, 172.

Stichometry, 36.

Stilus, 15.

Styles of Writing, 91-1 15. See Contents.

Subdivisions of Criticism, 151; in this
book, 159.

Subscriptions, 52, 53.

Substitution, 140.

Symbols for Manuscripts, 81.

Tabernae librariae, 29.
Tardy use of Parchment, 22.
Testimonia, defined, 162; explained, 173-

'77, '97, !9 8 -

Tests of Authorship, 196-202; of Manu-
scripts, age, 115; worth, 186, 187.

Texts, modern, soundness of, 208.

Textual Criticism, 161-193. See Con-
tents.

Titles, doubtful forms of, 40, 41.

Tituli, 8.

Transmission of Manuscripts, 42-67, See

Contents.
Transposition, 138, 139.

Umbilicus, 7.

Unavoidable Errors, 121-124.

Uncials, 98, 99, 104, 106, 115.

Uncollated Manuscripts, 88.

Unimportant words added or omitted,
141.

Uses of Paleography, 93, 94; of Pa-
limpsests, 67; of Stemmata, 172; of
Subscriptions, 53.

Vellum, 3. See Parchment.

Vergil : a supposed original Manuscript,

35; sumptuous editions, 102; portrait,

102; date of codices, 115.
Visigothic hand, 97.
Volumen, 6, 7.

Wax Tablets, 2, 96.

Writing as manual labor, 59; rapidity of,

34; materials for, 2, 5, 15.
Wrong Division of Words, 143, 144; of

clauses and sentences, 145.

St. Michael's College
Library

PHONE RENEWALS

Z 114 .J73 1897 SMC
Johnston, Harold Whetstone
Latin manuscripts 47076251