Copyrighted Material

Preface

This book is the first full-scale study of libraries in the ancient

world. It presents whatever is known about them from their debut
in the ancient Near East in the third millennium B.C. down to the
early Byzantine period, the fourth and fifth centuries A.D., when
the spread of Christianity and of monasticism fundamentally
changed the course of library history.
Ancient writings contain only haphazard bits of information
about libraries. To fill out the picture we must turn to a variety of
sources. One of the most helpful is archaeology: at a number of
sites excavation has laid bare library remains, and these give us an
idea, in some cases a fairly good idea, of the physical facilities. Inscriptions relating in one way or another to libraries, from decrees
honoring wealthy benefactors to epitaphs on the gravestones of
humble employees, provide welcome detail on important aspects.
But there are certain matters, such as the nature of the holdings in
the various libraries, the readership they served, the way they
made their acquisitions, their methods for shelving and cataloguing, and so on, for which we are reduced to inference based on
faint clues, or even, at times, to pure speculation.
The book is addressed to both general readers and scholars. For
the latter I have supplied, in a section at the back, complete documentation; it is keyed to the statements concerned by means of
page references and brief clues rather than footnote numbers to
avoid the distraction these can cause.
It was Harry Haskell of Yale University Press who suggested the
idea of a book on ancient libraries, and I am deeply grateful for his
ix

Copyrighted Material

Preface

stalwart support during all the time it took to bring it to fruition.

My good friend N. Lewis, author of the fundamental study of the
manufacture of papyrus paper, was kind enough to check my
statements on that subject.
The book is dedicated to my wife whose encouragement, as always, was unfailing and who, as always, passed a keen and critical
eye over the manuscript.

xi
A F

I C

Thamugadi

The Greco-Roman world: the west.

S IC ILY

Naples
Pompeii

Carthage

SEA

UM

Rome

MEDITERRANEAN

Comum

Seville

Bilbilis

F R A N C E

R
ine .
Rh
IA
BR

S P A I N

A T L A N T I C
O C E A N

200

Syracuse

Tarentum

Narona

Danub
eR
.

100

GREECE

300 Miles

Copyrighted Material

Pergamum

Chios
Athens

BOEOTLA

xii

Cyrene

EGYPT

Jerusalem

Nineveh
Aleppo
Ashur
Euphra
tes
R.
AK

The Greco-Roman world: the east.

Alexandria

CYPRUS

Antioch

MINOR

Tigris R.

100

200

CASPIAN
SEA

300 Miles

Uruk

Babylon S U M
ER
Nippur

KA

IA

SEA

Smyrna
ASIA
Ephesus
Miletus
Halicarnassus
Cos
Astypalaea
Rhodes
CRETE

AEGEAN
SEA

Ancyra (Ankara)

YR

MEDITERRANEAN

Patras
Argos

Delphi

Pharsalus

Constantinople

S
AS

PHOENICI
A

ON
ED
AC
M

Sinope

BLACK SEA

PERSIAN
GULF

Copyrighted Material

Copyrighted Material

1
The Beginnings
The Ancient Near East

It was in Egypt and Mesopotamia, lands abundantly watered by

great rivers, that civilization arose. And it is there that we find the
earliest examples of that key feature of civilization, writing: inscribed clay tablets that date shortly before 3000 B.C. have been discovered among the archaeological remains of the Sumerians, a
gifted people settled in southern Mesopotamia.
The Egyptians were not far behind, but we cannot follow the
history of their writings nearly as well because they used a perishable writing material. In ancient times the banks of the Nile were
lined with papyrus plantsthe bulrushes of the ark of bulrushes
in which the infant Moses was placedand from the reeds the
Egyptians made a form of paper; it was excellent in quality but, like
any paper, fragile. Mesopotamias rivers boasted no such useful
reeds, but its land did provide good clay, and as a consequence the
clay tablet became the standard material. Though clumsy and
bulky it has a virtue dear to archaeologists: it is durable. Fire, for
example, which is death to papyrus paper or other writing materials such as leather and wood, simply bakes it hard, thereby making
it even more durable. So when a conqueror set a Mesopotamian
palace ablaze, he helped ensure the survival of any clay tablets in it.
Clay, moreover, is cheap, and fashioning it into tablets is easy, factors that helped the clay tablet become the preferred writing material not only throughout Mesopotamia but far outside it as well, in
Syria, Asia Minor, Persia, even for a while in Crete and Greece. Excavators have unearthed clay tablets in all these lands. In the Near
East they remained in use for more than two and a half millennia,
1

Copyrighted Material

The Beginnings: The Ancient Near East

and in certain areas they lasted down to the beginning of the Christian Era until finally yielding, once and for all, to more convenient
alternatives.
The Sumerians perfected a style of writing suited to such a material. It consists basically of simple shapes, just wedges and lines,
which could easily be incised in soft clay with a reed or wooden
stylus; scholars have dubbed it cuneiform from the wedges (cunei
in Latin) that are its hallmark. Although the ingredients are merely
wedges and lines, the combinations of these to form signs standing
for sounds or words number in the hundreds. Learning them
required long training and much practice; inevitably, literacy was
by and large limited to a small professional class, the scribes.
The Akkadians conquered the Sumerians around the middle of
the third millennium B.C., and they took over the various cuneiform signs used for writing Sumerian and gave them sound and
word values that fit their own language. The Babylonians and Assyrians did the same, and so did peoples in Syria and Asia Minor.
For the scribes of these non-Sumerian users of cuneiform, training
was doubly demanding. The literature of the Sumerians was treasured throughout the Near East, and long after Sumerian ceased to
be spoken, the Babylonians and Assyrians and others kept it alive,
the way Europeans kept Latin alive after the fall of Rome. Their
scribes, as a consequence, had to know the values of the various
cuneiform signs for Sumerian as well as for their own language.
The contents of the earliest clay tablets are simple notations of
numbers of commoditiesanimals, jars, baskets, etc. Writing, it
would appear, started as a primitive form of bookkeeping. Its use
soon widened to document the multitudinous things and acts that
are involved in daily life, from simple inventories of commodities
to complicated governmental red tape.
Archaeologists frequently find clay tablets in batches, sometimes batches big enough to number in the thousands. The batches
consist for the most part of documents of the types just mentioned:
bills, deliveries, receipts, inventories, loans, marriage contracts, divorce settlements, court judgments, and so on. These records of
2

Copyrighted Material

The Beginnings: The Ancient Near East

factual matters were kept in storage to be available for reference

they were, in effect, files, or, to use the term preferred by specialists
in the ancient Near East, archives. Now and then these files include
pieces of writing that are of a distinctly different order, writings
that do not merely record some matter of fact but involve mental
activity. They range from simple textbook material to creative literatureand they make an appearance very early. Near Nippur in
southern Mesopotamia, for example, excavation brought to light a
group of tablets, dating to the middle of the third millennium B.C.,
on which were inscribed lists of geographical names, lists of gods,
lists of professions, writing exercises, a number of hymns. They
could well be from a collection belonging to a school for scribes,
perhaps one maintained by a temple, a collection of works that
were kept handy for consultationin other words, its library.
Around 1980 archaeologists working at the site of ancient Ebla
in Syria, thirty or so miles southwest of Aleppo, while digging out
the remains of Eblas royal palace, had a stroke of phenomenal
good luck: they came upon its main archive room filled with some
two thousand clay tablets that were on file there when, around
2300 or 2250 B.C., invaders set the palace on fire. The room was a
rectangle roughly 3.5  4 m, and the tablets were found in heaps
on the floor. Apparently they were kept on wooden shelves along
the walls and, when the shelves burned through, the tablets fell
down. The great majority were administrative records: almost a
thousand pieces listed distributions of textiles and metals by the
palace authorities, while another hundred or so had to do with cereals, olive oil, agricultural land, breeding of animals. But, along
with these was a group of tablets whose contents were totally different. There were some sixty on which lists of words in Sumerian
were inscribednames of professions, of geographical locations,
of birds, of fish. There were twenty-eight with bilingual lists, in
Sumerian accompanied by translation into Eblaite. There were
more than a dozen incantations, some in Sumerian and some in
Eblaite. And there were two tabletsduplicateswith the text of
a Sumerian myth. The excavators have determined that they must
3

Copyrighted Material

The Beginnings: The Ancient Near East

have come from a top shelf along one wall, the north wall of the
room. The clear implication is that, shelved amid the palace
records, was the working library of the palace scribes.
Eblas collection was small enough for users to consult by simply
browsing through the shelf of tablets. As collections grew larger,
this would at some point cease to be feasible. Modern libraries
meet the problem by drawing up catalogues of their holdingsand
so did the scribes of the ancient Near East. Among the tablets found
at Nippur were two, dating to around 2000 B.C., both inscribed
with a list of Sumerian works of literaturevarious myths,
hymns, laments. One, slightly longer, has sixty-eight titles, the
other sixty-two. They clearly have to do with one and the same
collection because forty-three titles appear on both. The longer has
twenty-five titles that are not on the shorter, which, in turn, has
nineteen that the longer lacks. It could well be that the two lists
arose because the scribes catalogued the collection involvedperhaps a batch of tablets in a given areaat two different times, the
second time after a re-shuffling that removed a sizable number of
works and replaced them with others.
A catalogue, even so primitive a type as the one embodied in the
Nippur tablets, a mere listing of titles in no consistent order, was a
notable step toward systematizing a collection. Two further steps
were improved cataloguing and the adding of identifying notes to
tablets. Both of these steps had been taken by at least the thirteenth century B.C., as can be seen in finds from Hattusas.
Hattusas is an ancient site, located two hundred miles or so
southeast of Ankara, which was the capital of the empire of the
Hittites from the seventeenth to the thirteenth century B.C. Here
archaeologists uncovered a huge mass of tablets that came from
the royal palace there. Inevitably most are of documentary type
connected with governmental activity, but there are a good number that are not, ranging from prosaic handbooks to Hittite renderings of Sumerian and Babylonian epics. Some of these have, added
after the end of the text on the back surface, several lines of writing
that identify the work more or less the way a title page does today,
the colophon, as it is called; the term is derived from the Greek
4

Copyrighted Material

The Beginnings: The Ancient Near East

word kolophon, finishing touch, reflecting the ancient practice of

providing as a finishing touch at the end what we provide
through a title page at the beginning. Here are a few examples:
Eighth tablet of the Dupaduparsa Festival, words of Silalluhi and
Kuwatalla, the temple-priestess. Written by the hand of Lu, son
of Nugissar, in the presence of Anuwanza, the overseer.
Third tablet of Kuwatalla, the temple-priestess. Not the end.
When I treat a man according to the great ritual [i.e., the first
line used as title].
Second tablet of Tudhaliyas, the great king, on the oath. End.
This tablet was damaged; in the presence of Mahhuzi and
Halwalu, I, Duda, restored it.

Each colophon begins with the number of the tablet it is on. This
was of vital importance, for, though the scribes wrote on both faces
of a tablet, and often wrote very small, many works required more
than one, even as works today require more than one page. But
tablets, unlike pages, could not be bound; the best that could be
done was to keep them together, either in stacks on top of each
other or on edge alongside each other, both of which arrangements
left ample opportunity for individual tablets to get misplaced or
lost. The colophon often includes, as in the first example, the name
of the local scribe who copied the text. The third example involves
a tablet that had come from elsewhere and had suffered damage;
it records the scribe who took care of making it usable.
Not all tablets had colophons. Where they were present, they
unquestionably were of great help to users of a collection: a glance
at a colophon immediately revealed a tablets contents and the part
of the work it represented.
A discovery at Hattusas reveals that catalogues had come into
existence that were a great advance over the bare listings in those
found at Nippur. The discovery was of a series of tablets, probably
belonging to the thirteeth century B.C., that contained detailed bibliographical entries. Each entry begins by giving the number of
tablets that made up the work being recorded, just as modern cata5

Copyrighted Material

The Beginnings: The Ancient Near East

logues give the number of volumes in a multi-volume publication.

Then the entry identifies the work itself by giving the title, which
may take the form of citing its first line, or by giving a capsule description of the contents. Then it tells whether the tablet marked
the end of the work or not. At times the entry includes the name of
the author or authors, or adds other useful information. Here are
some samples that illustrate the nature of the entries and their
variety:
Three tablets on the spring festival of the city of Hurma. How the
presiding official celebrates the festival. First and second tablets
missing.
Single chapter. When the singer in the temple of the deity Inar
breaks the offering bread he then recites as follows in Hattic.
The end.
One tablet. Words of Annana, the old woman. When one supplicates the Storm-God. Not the end.
Single chapter. Words of Ammihatna, Tulpija, and Mati,
priests. . . . When in a cultic place in a temple that is pure one
finds fraud, here is the ritual therefor. The end.
One tablet on the fine oil of Azzari, the Hurrian woman doctor.
When one leads the troops in battle against an enemy city and a
charm using fine oil is put on the general who commands the
army, one anoints the general, his horse together with his chariot, and all his battle gear. The end.
One tablet, the end, on the purification of a murder. When the
exorcist-priest treats a city for a murder. Words of Erija.
Two tablets. When the king, queen, and princes give substitutefigures to the Sun-Goddess of the Earth. The end. However we
have not found the first tablet belonging to it.
Words of Ehal-Tessub. . . . When for a man his male and female
slaves do not get along, or a man and a woman do not get along,
or a man and a woman see bad dreams; how one carries out to
Istar (?) the magic ceremony of the evil hatred.

Copyrighted Material

The Beginnings: The Ancient Near East

In addition to noting missing tablets, the entries now and then

provide information about shelving. There is an entry, for example, which in listing a work that happens to be in two tablets notes
that they do not stand upright; presumably, in the part of the
palace holdings represented by this catalogue, most tablets were
stored on edge while these two, exceptionally, lay flat.
A good number of the works recorded in the entries have
turned up in the unearthed mass of tablets. Some of these have
colophons that are similar in wording to the entriesindeed, one
colophon corresponds, word for word. Quite possibly the compilers of the catalogue, as they passed along the shelves, saved time by
simply copying or paraphrasing colophons whenever they came
across them.
Practically all of the works listed in the catalogue have to do with
religion: ceremonies for public occasions, rituals for a variety of circumstances from putting a protective charm on a general headed
for combat to solving a domestic quarrel, interpretation of omens,
and so on. Indeed, the subject matter is so consistently of this nature that the sudden appearance, right in the middle of a
series of entries about rituals, of one listing a work on a treaty between the king of the Hittites and some local ruler may well reflect
an ancient instance of mis-shelving! The holdings of the palace at
Hattusas were, to be sure, far from being limited to such materials;
they included myths, legends, historical annals. The catalogue, it
would seem, was of one particular collection that, to judge from the
contents, was for use by the palace clergy. It would have been an invaluable tool: any priest who needed a ritual for a given problem,
instead of picking up tablet after tablet to read the colophon if there
was one, or some lines of text if there was not, had only to run an
eye over the entries in the catalogue. It was a limited tool: the order
of the entries is more or less haphazard (alphabetization, for example, lay over a millennium and a half in the future) and they give no
indication of location. But it was, no question about it, a significant
step beyond the simple listing of titles of the Nippur tablets.
The finds at Hattusas, in short, reveal the development of procedures for organizing a collection of writings. The palace holdings
7

Copyrighted Material

The Beginnings: The Ancient Near East

were certainly extensive enough to require them: the catalogue

alone, representing, as we have seen, just the clergys working
library, lists well over one hundred titles. Now, modest collections
of tablets were to be found throughout the Near East, for they were
a natural result of the way scribes were trained. The key element
was repeated copying of well-known works, so every master had
on hand a collection of these to serve as models, and his students
could build up collections of their own by simply keeping some of
the copies they ceaselessly turned out. Scribes associated with temples by the same assiduous copying built up collections for their
temples; the scholarly material served the scribes themselves and
the religious, the temple priests. It was in this fashion and not by
purchase that collections were put together, for a book trade does
not seem to have existed in the Near East.
On the other hand, a library the size of that at Hattusas was
strictly a prerogative of kings, all-powerful figures who could create holdings en masse by plundering the collections in places they
conquered or, when that recourse was not available, by sending
out the palace scribes to copy whatever was desired. Which particular monarch was responsible for Hattusas library or for any of the
previous royal libraries, such as Eblas, we do not know. It is not
until the end of the twelfth century B.C. that we can finally cite the
name of a founder of a libraryTiglath-Pileser I, one of Assyrias
greatest rulers, whose reign lasted almost forty years, from 1115 to
1077 B.C.
The proof was unearthed in the ruins of the temple of Assur, Assyrias chief god, at Ashur, the nations religious capital. Excavators
found in a certain area there numerous tablets that seemed to belong together: the writing is similar, the nature of the clay is similar, and a number of sure indications of date point to their having
been assembled during the reign of Tiglath-Pileser I. It has been
convincingly argued that they were part of a library that he created, possibly while he was still crown prince. Some one hundred
different works can be ascribed with a fair degree of certainty to it,
a number that surely represents only a part of what the holdings
were in his day. As at Ebla and Hattusas, only a handful of these are
8

Copyrighted Material

The Beginnings: The Ancient Near East

purely literary; most are the professional writings that were the indispensable tools of scribes and priests. The biggest component
were works dealing with omens, the determination from stars and
other heavenly bodies, from sacrificial animals, natural events, and
the like, of what lay in the future. The next biggest were standard
handbooksvocabulary lists, list of plants, trees, animals, gods,
place names, a multiplication table, an astronomical text. There
were some hymnsand even a catalogue of musical compositions,
whose entries went so far as to include the instruments they were
to be sung to, for example:
5 Sumerian psalms comprising one liturgy, for the adapa [probably a tambourine].
Song to the reed flute in Sumerian.
Three recitations to the pipe in Semitic.

Thus, if scholarly inference is right, to Tiglath-Pileser I goes the

credit for being the first founder of a library whom we know by
name. It is, however, a later king of Assyria, Ashurbanipal, who, no
question about it, was founder of a library. Moreover, what he
founded has every right to be called the first systematically collected library in the ancient Near East. In size it was many times
greater than any before and was not to be surpassed for the next
three and a half centuries.
Ashurbanipal, Assyrias last important ruler and one of its most
noted, held the throne for well-nigh half a century, from 668 to
627 B.C. He was himself literate: he boasted not only that he
achieved the highest level in the scribal art but that among the
kings, my forerunners, none had learned such an art. During the
latter half of the nineteenth century, British archaeologists working at Nineveh, the royal seat of Ashurbanipal and his predecessors, came upon a vast quantity of tablets in the ruins of two of his
palaces. The find turned out to be a treasure trove: it has provided,
over and above hundreds of examples of the various types of professional writings, the texts of the Epic of Gilgamesh, the Epic of
Creation, and most of the other noted works of ancient Near East9

Copyrighted Material

The Beginnings: The Ancient Near East

ern literature that we possess today. Colophons provide the link to

Ashurbanipal. Some read baldly: Palace of Ashurbanipal, King of
the World, king of Assyria. Others are long and detailed:
Palace of Ashurbanipal, King of the World, King of Assyria, who
trusts in Ashur and Ninlil, whom Nabu and Tashmetu gave wideopen ears and who was given profound insight. . . . The wisdom
of Nabu, the signs of writing, as many as have been devised, I
wrote on tablets, I arranged [the tablets] in series, I collated
[them], and for my royal contemplation and recital I placed them
in my palace.

1.1 Clay cuneiform tablet, 14.6 cm  13.3 cm, with the Assyrian version
of the Gilgamesh epic. From the palace of Ashurbanipal at Nineveh.

10