literacy in pre-islamic arabia 255

Literacy in Pre-Islamic Arabia: An Analysis of the

Epigraphic Evidence*

Peter Stein

Introduction

When an extensive literary work, and especially a holy scripture like

the Qurʾan, is committed to writing, this appears to presuppose the
existence of established writing practices and a minimum of literacy
among the community addressed by that scripture. The holy scrip-
tures of Judaism and Christianity, for example, emerged in the cultural
centers of Mesopotamia and the Mediterranean, that is, in an area
and at a time where the use of writing had been a familiar element
of everyday life for centuries. By contrast, the region in which the
Qurʾan came into being—namely, western Arabia—does not seem to
have been characterized by extensive literary production. But even
though archeological evidence for this region, and especially for the
sixth century, is sparse, there are indicators of a limited spread of
literacy even in this area.
There can be no doubt that the Qurʾan was put into writing under
the influence of the Christian (i.e., Syriac, and probably also Ethiopian)
and Jewish literary traditions. The Christians and Jews who were pres-
ent in the urban centers of the Arabian Peninsula must have brought
their holy scriptures with them into the region, and liturgical practices
in these communities must have involved some use of written docu-
ments. A minimum of scribal production in Aramaic (or even Hebrew)
among these communities can thus safely be assumed, even though
this supposition still awaits archeological confirmation. Yet since lit-
erary production of this kind was imported from Syria-Palestine and
did not have a properly Arabian origin, it will not be pursued any
further within the scope of this contribution.1 Rather, the present

* This paper is an abbreviated version of a more extensive paper published in

German (Stein, “Stein vs. Holz”).—I am very indebted to Michael C. A. Macdonald
(Oxford) for having read a draft of the manuscript.
1
  These communities were dependent on the literary culture of the Hellenistic
world, which is largely different from the genuinely Arabian one, and in particular
256 peter stein

article will focus on the question of a genuinely Arabian culture of

writing, that is, one developed within the Arabian Peninsula itself
and existing prior to, or contemporary with, those writing traditions
imported from outside. It should be noted that “Arabian” in this
context is used as a purely geographical term rather than an ethnic
or linguistic one, comprising all the cultures of the Peninsula (except
for the regions in the far north and northwest, which were subject to
strong Hellenistic influences).2 Bearing in mind the explicit self-clas-
sification of the Qurʾan as “Arabic,” I will therefore ask to what extent
the inhabitants of pre-Islamic Arabia were able to express and receive
their mother tongues by means of script.
For this purpose, the available material sources documenting the
use of writing will be surveyed and classified with respect to their
distribution and their importance within public and private social
life. Subsequently, the question of literacy in the particular societies
will be discussed. This procedure fits nicely under three questions: 1.
Which materials were written upon? 2. What was written? 3. Who
did the writing? Due to the fact that the archeological exploration of
the Arabian Peninsula is as yet rather scanty, the following analysis
will naturally be restricted to conclusions that will have to be critically
reviewed once new sources become available. Nevertheless, an exhaus-
tive examination of the epigraphic material presently known promises
to yield important insights for the study of the genesis of the
Qurʾan.3
The following statements will concentrate exclusively on firsthand
epigraphic material, while later Islamic accounts of the topic will be
referred to only if they overlap with the epigraphic evidence (for a
discussion of the Islamic sources in connection with the Qurʾan, see
Gregor Schoeler’s contribution to this volume). Due to my concentra-

from the South Arabian one (as will be demonstrated further below). Yet even for the
Mediterranean world, the degrees of literacy maintained by different scholars differ
widely (cf., e.g., Blanck, Buch, 29, against Harris, Literacy, 140–141, 306–322).
2
  As the examples of the Nabateans and the Amīr (in Qaryat al-Fāw, Najrān and
the Yemenite Jawf) show, members of ethnically “Arab” tribes could assimilate to a
neighboring sedentary culture and take over its way of life, including the use of writ-
ing. While the Nabatean script was borrowed from older non-Arabic (i.e., Aramaic)
traditions, the script used by the Amīr was a genuinely South Arabian development.
3
  The only recent attempt at providing an overview synthesizing both pre-Islamic
epigraphy and traditional “Arabic studies” (études arabes) was undertaken by Robin,
“Inscriptions.” However, his focus is mainly on linguistics and research history, while
the material aspects of writing are only marginally touched upon.
 literacy in pre-islamic arabia 257

tion on epigraphic sources, South Arabia, which is comparatively well

documented, will receive far more attention than other parts of the
Arabian Peninsula. Yet this imbalance may not be purely accidental,
since South Arabia gave birth to a highly developed culture unparal-
leled in the rest of the peninsula already in pre-Islamic times. From
the situation prevailing in South Arabia, extrapolations concerning
other, and less extensively documented, regions can then be made,
provided of course that a comparable level of social and economic
organization can be assumed. Chronologically, the following inquiry
will cover the period stretching from the eighth century bce until the
mid-sixth century ce.

Which Materials were Written Upon?

First of all, let us turn to the material basis of writing in pre-Islamic

Arabia, i.e., the available writing materials. The considerable progress
made in the field of South Arabian epigraphy during the last decades
makes it possible to sketch a fairly detailed picture of the materials
used for writing down various kinds of texts. The order in which these
materials will now be treated is based on their respective importance,
as reflected in the number of preserved inscriptions. Let it be noted
right away that the following account, based as it is on epigraphic
evidence, coincides only partially with the information given by later,
and mostly Islamic, sources.4 In fact, this should come as no surprise,
since the fact that a given writing material is mentioned in a later
source does not necessarily mean that it had actually been in wide-
spread use.
The writing material most extensively used in the Arabian Peninsula,
at least according to the present state of research, were rocks and
stones.5 The graffiti sc attered over the entire peninsula count several
ten thousands.6 It must be admitted that due to their firmness and

4
  For the latter, cf. the overview by Grohmann, Paläographie, and Maraqten,
“Writing Materials.”
5
  However, it is significant in this context that ca. 7,000 of the presently known
Ancient South Arabian minuscule inscriptions on wood (see below) stem from only
one place in Yemen. If we extrapolate on this basis with respect to all important cities
of South Arabia alone, their number would by far exceed that of the monumental
inscriptions.
6
  Cf. n. 44. By contrast, inscriptions carved on hewn stones or other artificial
objects number only a few thousands.
258 peter stein

immobility, more inscriptions on stone and rock may have survived

than those executed in other materials, which renders any immediate
quantitative comparison impossible. On the other hand, as will be
shown further below (cf. n. 45 below), the background of those numer-
ous and widely scattered rock graffiti is probably different from that
of all other kinds of inscription. That is, rock as a writing material
does not enter into competition with other materials, because it is
restricted to a very specific kind of literacy, namely, writings by shep-
herds passing by (and also by campaigning generals). Even though
rock, because of its nearly unlimited size, can be considered the most
representative writing surface, it cannot be moved and is thus not
everywhere available. Inscriptions of public relevance, like royal
decrees, are thus only rarely executed in rock but rather on hewn
blocks of stone, which can be erected wherever needed, for example,
in the center of a town. Hence, the distribution of stone and rock as
writing materials mainly depends on the distribution of representative
rock surfaces in the immediate surroundings of a sanctuary or settle-
ment: while in the rocky region of the Radmanite area, for example,
even building or votive texts are written on rock (e.g., in Qāniya and
the Jabal al-Miʿsāl), in oasis settlements like Mārib or the Jawf settle-
ments, we find merely inscriptions on stone blocks. With respect to
ancient South Arabia, then, stone and rock occupy an analogous rep-
resentative status, yet are exchangeable only to a limited degree.
Outside South Arabia, inscriptions on rock or (more rarely) hewn
stone are distributed more unevenly.7 While the short inscriptions
traditionally called Safaitic and Thamudic are spread over almost the
entire north and west of the peninsula, more elaborated inscriptions
concentrate on merely two centers of settlement along the interna-
tional trade routes: the numerous Dedanic (or Lihyanic) and Minaic
inscriptions from the oasis of al-ʿUlā in the north-west, and the homo-
geneous group of the Hasaitic tomb inscriptions from the north-
eastern coast. Since the Aramaic and Nabatean inscriptions from
Taymāʾ and Madāʾin Ṣāliḥ, respectively, must be considered part of
the Aramaic cultural area, they are not dealt with in this paper. In

7
 On the following, cf. Macdonald, “Reflections,” with some new suggestions con-
cerning the terminological definition of single languages and dialects. For the most
recent editions of some of these corpora, see Sima, Lihyanischen Inschriften, and Sima,
“Hasaitischen Inschriften”; for pre-Nabatean Aramaic texts, see Schwiderski (ed.),
Inschriften.
 literacy in pre-islamic arabia 259

addition, there are a small number of isolated inscriptions in Old

Arabic from different places of the peninsula.8
The second most important writing material is wood (cf., however,
n. 5). Thousands of inscribed palm-leaf stalks and other wooden sticks
from South Arabia have come to light during the last decades.9 These
sticks, which on average are 10 to 15 cm long and only occasionally
measure up to half a meter, were scratched with a pointed style.
During the course of centuries, this practice supported the develop-
ment of a specialized minuscule script used almost exclusively for
inscribing sticks, whereas contemporary inscriptions on stone or
metal were executed in a different monumental script which had
evolved separately. In South Arabia, such wooden sticks clearly con-
stituted the most practical writing material, since they could be used
for transferring as well as storing information, while the requisite raw
material is renewable, cheap, and readily available, provided there
exists a minimum of agricultural activity. Against this background,
other wooden surfaces like tablets, which require more elaborate pro-
cessing, play only a marginal role.10 Sticks made from palm-leaf stalks
can be identified with what the Arabic-Islamic tradition refers to as
ʿaṣīb (plural ʿuṣub),11 on which parts of the Qurʾan were supposedly
recorded. Even though Arabic literature mentions other kinds of
wood, these have not yet been botanically identified with specific
kinds of sticks.
The third rank in the list of writing materials is occupied by pieces
of terracotta and pottery, which generally contain short inscriptions,
often no more than a single word. In this context, however, a some-
what narrow definition of the term “writing material” must be adopted:
objects of daily use like jars which are marked with names, indications
of their content and the like do not count as writing media, since
they were not produced or prepared primarily in order to be written
upon. Nevertheless, even though the majority of inscribed containers
do not therefore enter into the present discussion, there still remains

8
  The Old Arabic texts which have so far been discovered are listed in Macdonald,
“Reflections,” 61; cf. also Robin, “Inscriptions,” 545–550, and Macdonald, “Old Ara-
bic.”
9
  The most comprehensive publication on the subject is Ryckmans et al., Textes
du Yémen. A supplementary bibliography is given in Stein, “Minuscule Inscriptions,”
183, with n. 9.
10
  There is only a handful of inscribed boards and tablets, mostly in the shape of
“eye stela,” which resemble monumental inscriptions on stone.
11
  Cf. the references given by Maraqten, “Writing Materials,” 292–293.
260 peter stein

a considerable amount of ostraca used as writing surfaces. In contrast

with the widespread use of ostraca in the Mediterranean, however,
their use in South Arabia seems to have been limited to one specific
field, namely, the identification of persons buried in large tomb houses
like the Awām cemetery in Mārib.12 More elaborate texts on terracotta
objects are almost entirely absent.13
The next material in terms of archeological attestation is metal,
which is almost exclusively bronze. In some sanctuaries (e.g., the
temple of Almaqah in the city of ʿAmrān), inscriptions executed in
bronze tablets are found instead of the stone inscriptions which were
common elsewhere. Other metals like silver and gold, by contrast,
play almost no role.14 Given the high costs of metal objects as com-
pared with other materials, this perhaps comes as no surprise. Besides
these sanctuary inscriptions, the vast majority of inscribed objects of
metal do not qualify as writing media in the sense defined above.
Other materials which could be used for writing—namely, papyrus,
leather, and parchment, all of which were widespread throughout the
Mediterranean, and also linen, cotton, silk, and animal bone—have
not been documented archeologically, apart from very few excep-
tions.15 That these materials were at least in limited use seems probable
given the later Arabic-Islamic sources.16 According to them, parts of
the Qurʾan were first written down on a variety of disparate materials,
among them leather (adīm), parchment (raqq), papyrus (qirṭās), as
well as the shoulder bones and ribs of camels (aktāf and aḍlāʿ). In
addition, mention is made of splinters of limestone, called likhāf,
which may parallel the ostraca found in other regions of Arabia.
Together with the above-mentioned palm-leaf stalks (ʿuṣub) and
bones, these splinters can be characterized as the most easily acces-

12
  Cf. the overview by Gerlach, “Friedhof,” 57. Obviously, to every corpse an ostra-
con bearing his name was allotted.
13
 Two exceptions to this, which consist in tomb inscriptions from the Gulf region
(ash-Shāriqa), are presented by Robin, “Documents,” 80–81. Some few votive inscrip-
tions on pottery vessels belong to the category of objects of daily use not specifically
produced for writing. They are similar in kind to the more numerous votive texts on
metal vessels.
14
  There is merely half a dozen amulets made of gold, silver and bronze (cf. recently
Nebes, “Goldanhänger”).
15
 Some camel bones painted with Ancient South Arabian characters have been
found at Qaryat al-Fāw, cf. Ansary, Qaryat al-Fau, 99, figs. 3–4.
16
 See Grohmann, Paläographie, 66–117; Maraqten, “Writing Materials,” 288–300
and 305–310; with respect to the Qurʾan, see GdQ, vol. 2, 13, and Schoeler, “Schrei-
ben,” 21–22 and 24.
 literacy in pre-islamic arabia 261

sible among the genuinely “Arabic” writing materials. By contrast,

leather and parchment require an extended production process, while
papyrus had to be imported entirely. In all likelihood, materials such
as these were brought to the area only in the course of the establish-
ment of Jewish and Christian communities.
Finally, the exact meaning of the terms “tablets” (alwāḥ) and
“leaves” (ṣuḥuf), which are also referred to in the context of the writ-
ing down of the Qurʾan, is difficult to determine. As we have seen
(cf. n. 10 above), wooden tablets were not used as writing materials
for daily use. It is also questionable whether the natural environment
of the Arabian Peninsula would have made an extended production
of wooden objects larger than simple sticks possible. Since all the
other materials enumerated in connection with the Qurʾan were either
imported, produced specifically in order to be inscribed, or simply
picked up from the ground, the alwāḥ were probably wooden tablets
prepared for writing purposes which, like papyrus and parchment,
originated from the Mediterranean. An identification of the alwāḥ
with the bronze tablets known from South Arabia seems far less plau-
sible: this hard and expensive material is simply not suited to being
inscribed spontaneously with matters of daily relevance.17
The term ṣaḥīfa (plural ṣuḥuf), the basic Arabic meaning of which
is “smooth surface suitable for writing upon” and which is commonly
translated as “leaf, page (of a book),” has a graphical, and probably
also morphological, equivalent in Ancient South Arabian ṣḥft.18 Since
the root ṢḤF occurs in Ancient South Arabian much earlier than a
possible Ethiopian influence in the region can be assumed,19 a bor-
rowing of ṣḥft from Ethiopic can be ruled out, which possibly holds
true for the Arabic ṣaḥīfa as well.20 Ancient South Arabian ṣḥft means
primarily “written account, document,”21 without being restricted to

17
  Though it may be possible that the heavenly tablet (lawḥ) mentioned in the
Qurʾan (Q 85:22) was thought of in terms of these representative metal tablets hung
in South Arabian temples, which may have been known also to visitors from other
regions of the peninsula.
18
  Cf. al-Selwi, Jemenitische Wörter, 128.
19
  This is demonstrated, for instance, by the fragmentary boustrophedon inscrip-
tion YM 546/2 (ḏt ṣḥft[n], “this (present) document”) from the Yemenite Jawf, which
belongs to the oldest known monumental inscriptions in South Arabia (Wissmann,
Geschiche , 69f.).
20
  In contrast to this, the Arabic term muṣḥaf (“book”) is derived from Ethiopic
maṣḥaf (“written document, book”) (cf. Leslau, Dictionary, 552; Burton, “Muṣḥaf”).
21
  Cf. Beeston et al., Sabaic Dictionary, 142; Ricks, Lexicon, 135.
262 peter stein

a certain type of writing material. This is evident especially from

inscriptions in which ḏt ṣḥftn, “this document,” explicitly denotes the
respective text, carved in stone or rock.22 Yet the term cannot simply
be synonymous with s1ṭr, “written account,”23 as is suggested by con-
structions like ʾs1ṭr ḏt ṣḥftn, “the written (lines) of this document” (J
2361/14) or ḥg ḏt ṣḥftn w-ʾs1ṭr-s1, “according to this document and its
written (lines)” (R 3688/10; both Qatabanic), and by the genre of the
respective documents, which are of a legal character. Thus, the seman-
tic horizon of ṣḥft is best located in the vicinity of “written document
of legal relevance,” and the respective verb ṣḥf hence denotes the
writing of such a document. This fits well with the information that
the ṣuḥuf of pre-Islamic and early Islamic times, albeit not sealed,
were of a legal character and could be hung in representative public
places.24 The specific meaning of the Classical Arabic term should
therefore be considered a later derivation.25
With regard to the writing techniques employed, the writing was
generally carved on the surface of the respective medium, i.e., it was
chiseled in rock and stone, or scratched in wood or terracotta. The
only exceptions are the above-mentioned bronze inscriptions, which,
as a rule, were cast in models made by lost-wax technique, thus pro-
ducing an inscription in relief. As already mentioned, two different
types of script were in use in ancient South Arabia: a monumental
script (in Arabic, musnad),26 which in principle can be used for all
writing materials, and a minuscule script (zabūr),27 limited to writing
on wooden sticks and palm-leaf stalks.28 The former is characterized

22
 E.g., M 316/1 and YM 546/2 (fragments of stone blocks), J 2361/14 and
R 3688/10 (rock surface).
23
  Cf., for example, the rock inscriptions J 539/1, Naqīl-Kuḥl 1/3=Gr 157/3, Ry
506/2, and the minuscule inscription on a palm-leaf stalk YM 11749/1 (all Sabaic), in
which ḏn s1ṭrn (“this document”) refers to the respective text; examples from the
Qatabanian area are J 2361/7, R 3688/11.14, and R 3689/8. – By the way, also the
Ancient South Arabian verb s1ṭr (“write”) was eventually transferred to Arabic as a
loan word (saṭara, vgl. al-Selwi, Jemenitische Wörter, 109–110.).
24
  Cf. Ghédira, “Ṣaḥīfa.”
25
  Thus already GdQ, vol. 2, 24, n. 4, and recently Maraqten, “Writing Materials,”
309.
26
  For terminology, cf. al-Selwi, Jemenitische Wörter, 114; Beeston, “Musnad,”
704–705.
27
  Cf. Müller, “Zabūr”; Maraqten, “Writing Materials,” 301–306. The Arabic term
is, like the root ZBR in general, of South Arabian origin (cf. already al-Selwi, Jeme-
nitische Wörter, 103).
28
  The few examples of minuscule characters on stone or bronze objects are hardly
of any importance. The relevant evidence is given in Ryckmans, “Écriture minuscule,”
 literacy in pre-islamic arabia 263

by a high degree of symmetry, based on simple geometrical forms,

while the latter has during the course of centuries developed from its
original monumental style to a completely independent cursive script
(cf. fig. 1).29
Writing with ink has not yet been attested archeologically, and the
use of paint, too, has been documented in only a few instances.30 Both
techniques thus do not even come close to rivaling the ubiquity of
engraved writing. According to the Arabic-Islamic sources, however,
paint- and ink-based writing was practiced in western Arabia in the
sixth century ce. These techniques are of course required for materials
such as leather, parchment, and papyrus, which were reportedly used
for writing down the Qurʾan. In any case, the early Islamic tradition
takes the use of the reed pen (qalam) dipped in ink for granted.31 It
is perhaps not surprising that for the Qurʾan, a writing technique
modeled on the holy scriptures of Christians and Jews was used; it
does not necessarily mean that other techniques, like the scratching
of wooden sticks, were unknown in those areas. As will be shown
further below, however, there are reasons to assume that the inhabit-
ants of western Arabia did not develop a writing culture of their own
but rather, if they did write at all, should have adopted the established
systems of their neighbors.

What was Written? The Content of the Epigraphic Sources

After having examined the materials that were written on, I will now
turn to the content of the surviving firsthand documents. In order to
be able to determine whether there existed “literary” writings in the
narrow sense of the term, I will distinguish between writing linked

188ff.; cf. also Ryckmans, “Pétioles,” 257, with nn. 8 and 9, and Ryckmans, “Origin
and Evolution,” 235, n. 14.
29
  The palaeographical sequence of the minuscule script has been established by
Ryckmans (cf. most recently “Origin and Evolution”).
30
 Some examples of painted rock inscriptions are shown by Jamme, Miscellanées,
82–83 (J 2477–2481, personal names), Pirenne, “Les témoins écrits,” 23ff., with pl. 13,
15–21, and al-ShaÎri, “Epigraphic Discoveries,” 179, with pl. Ib–IIa and IVb. Inscrip-
tions on building walls have been uncovered in Qaryat al-Fāw (mostly proper names
in combination with wall paintings on plaster, cf. Ansary, Qaryat al-Fau, 130–133,
and 136–137.). For samples of painted pottery, see the specimina published by Sedov,
“Denkmäler,” 94–95, fig. 38d and 39d), and some of the ostraca mentioned in n. 12
above.
31
  Cf. Grohmann, Paläographie, 117–123.
264 peter stein

to the needs of ordinary life, on the one hand, and writing that does
not arise from everyday concerns, on the other. In this connection,
the question of which kind of material was used for which genre of
texts is of some importance, since there are clear correlations between
certain sorts of texts and certain media. From this, one can draw a
number of conclusions concerning the possibility that there may have
existed hitherto unknown literary genres.

Documents Connected with Everyday Life

The by far most extensive genre of everyday writings consists in docu-
ments concerning legal and economic matters. Above all there are
numerous orders, cover letters, and delivery invoices, which make
up the bulk of correspondence by letter, and which are recorded on
wooden sticks and palm-leaf stalks. Letters containing exclusively
private information, by contrast, are very rare and for the most part
also concern matters of everyday relevance, for example, health prob-
lems of relatives. Confidential or even intimate correspondence has
not yet been attested, something which is perhaps hardly surprising
given the very open and public nature of the ancient South Arabian
postal system (see n. 54 below).
Deeds of legal and economic affairs were also written mainly on
wooden sticks; if they were of public relevance, they were additionally
carved on a representative block of stone. While these latter served
the aim of publication, the handy wooden sticks themselves, which
bore the signatures of the parties involved or of witnesses, were kept
in an archive. If any doubt was cast on the stone version of the docu-
ment, its wooden original could be consulted for confirmation.32
Lists of personal or family names recorded on wooden sticks prob-
ably also fulfilled an economic function, since they frequently include
numerals and units of measurement. Finally, isolated personal names
on funerary monuments, building walls, or ostraca, must be assigned
to a similar context, provided that they serve to identify the legal
affiliation of the inscribed object.

32
  Cf. the so-called Öffentlichkeitsklausel in Sabaic: ʾ-hn-n ʿkr l-yyfʿn, “whenever
there is an objection, it (sc. the original document) ought to be produced.” See Stein,
“Inscribed Wooden Sticks,” 269–272.
 literacy in pre-islamic arabia 265

A group of legal texts hitherto exclusively found on stone or rock33

consists in decrees and prohibitions, issued mainly by the ruler, by a
deity, or by some other high-ranking official of the community. Since
these documents were relevant for a wider section of the public, they
were generally set up at a prominent place immediately near the object
they dealt with (e.g., a sanctuary or a piece of property).
Also inscribed in stone are the numerous building inscriptions that
document the construction not only of dwelling-houses and temples,
but also of fortifications, tombs, and agricultural constructions like
irrigation devices, cisterns, and wells. On the one hand, these texts
exhibit a profoundly religious component, since all building activities
are placed under the protection of the gods. On the other hand, they
are also of legal relevance, since they were prominently placed on the
building’s façade and thus document the builder’s claim of owner-
ship.
A special subgenre of legal texts, which is unique not only in terms
of its content but also its writing medium, is constituted by confes-
sional, or expiatory, inscriptions, which were normally cast in bronze
tablets and were probably hung in temples. These texts contain public
confessions of cultic offenses committed by certain individuals or
groups, and also record the penances imposed on them.34
The remaining inscriptions of a religious character, above all the
numerous votive inscriptions carved on stone, are to be counted as
everyday documents as well. Even though commissioning a stone-
mason to execute such an inscription must have constituted an excep-
tional, if not unique, event in the life of the person ordering it, within
a society characterized by a predominance of ritual this practice
reflects essential needs of everyday life, such as the birth of a child,
recovery from a disease, or safe and successful return from a military
campaign. In fact, the production of votive inscriptions appears to
have been almost “industrialized,” as can be inferred not only from
the large amount of such texts but also from the use of stereotyped
formulas, which are composed of exchangeable elements occurring

33
  There are, however, some indirect indicators of the existence of such documents
in the form of wooden sticks as well (cf. R 3566/21, Qatabanic: w-l yftḫ ḏn ftḥn w-mḥrtn
b-ʿḍm ʾw ʾbnm kn-m byḥrg mlkn, “and this law and decree is to be written down on
wood or stone according to the command of the king”). Perhaps such wooden ver-
sions of royal decrees were kept in archives in the royal palaces that have not been
discovered yet.
34
  For the background of this kind of texts, see Sima, “Parallelen.”
266 peter stein

in more or less all of these inscriptions.35 In view of such formulaic

constraints, there is thus no room to speak of a “literary” composition
of these texts. A votive inscription is to be considered an accessory
to the votive object itself, such as a bronze statuette.36 The value of
the offering made is determined not by the literary value of the text,
but by the material value of the votive object itself.
In addition there are some—as yet still isolated—instances of reli-
gious texts that can be considered “literary” in a more strict sense of
the word, such as oracles, omens, and proverbs. These do not, how-
ever, appear as representative monumental inscriptions, but in a
rather modest form on the wooden sticks that were used for ordinary
writing purposes. The content of these texts makes it likely that they
were not recorded in order to preserve the society’s religious or cul-
tural memory for future generations, but rather served practical func-
tions within certain cultic ceremonies (e.g., oracular requests and
replies, such as A 40-3 and Mon.script.sab. 85).37
As I have already pointed out, all of the above-mentioned texts are
attested archeologically almost exclusively in South Arabia.
Nevertheless, some of them also appear in the Arabic-Islamic litera-
ture on pre-Islamic and early Islamic history. According to these
accounts, at least in the urban centers of the Ḥijāz legal documents
and contracts, decrees regulating social life, and even letters were
recorded in writing as early as the sixth century ce.38 Yet the marked
difference of the writing culture of western Arabia from that of South
Arabia is evident if one considers the almost complete lack of monu-
mental inscriptions on rock or stone from the centuries immediately
preceding Islam.39
An important, yet not easily answerable question that arises in this
context is in which script business correspondence in pre-Islamic

35
  Cf. Stein, “Schreibfehler,” 453–460.—Nevertheless, many texts do include pas-
sages that contain background information other than these stereotyped formulas.
These do not, however, exert any influence on the content and syntax of these votive
formulas and are therefore of no relevance for a general evaluation of the genre.
36
  After the Early Sabaic period, in particular from the first century ce onwards,
these votive inscriptions were usually recorded on a stone block serving as the statu-
ette’s pedestal. In other cases (such as metal vessels), the inscription is carved on the
object itself. Even when the writing medium and the votive object are identical, as it
is the case for votive inscriptions on bronze tablets, the actual value of the gift was
determined by the material of the writing medium rather than by the text itself.
37
  Cf. Stein, “Minuscule Inscriptions,” 187.
38
  Cf. Schoeler, “Schreiben,” 1–6.
39
  Cf. Robin, “Inscriptions,” 557–558.
 literacy in pre-islamic arabia 267

Ḥijāz, if assumed to have existed at all, might have been written. If

one goes by the few Old Arabic monumental inscriptions, the Arabic
script came into use only about 500 ce, whereas before this people
resorted to neighboring scripts like the Nabatean one in Syria-
Palestine, and the Ancient South Arabian one in Qaryat al-Fāw.40
Consequently, if there was any writing activity in western Arabia
before the sixth century ce, it was probably followed either Nabatean
or South Arabian writing practices. The former alternative is sup-
ported by the fact that the Arabic script is generally believed to derive
from the Nabatean one, while the second hypothesis derives its plau-
sibility from the coincidence of some specifics of the particular writing
materials (above all the ʿuṣub).

Inscriptions without Direct Connections to Everyday Life

All the text genres mentioned above have emerged from the needs
of everyday communication. We will now turn to texts that did not
arise from the economic or social requirements of ordinary life, but
rather were composed without any specific communicative intention.
First of all, there are numerous rock graffiti left by passing nomads
all over the Arabian Peninsula. Their coarse style suffices to show that
these texts are far from any kind of standardization, and thus can be
considered original creations of their authors. But since these inscrip-
tions hardly ever contain anything beyond their author’s name and
ancestors, they are of rather limited literary expressiveness.
Far more expressive are commemorative inscriptions, usually exe-
cuted on rock surfaces as well, which were commissioned by persons
of high social rank (and thus able to employ a professional scribe41)
on the occasion of some outstanding event. Most of these texts contain
accounts of military campaigns, but also commemorate construction
works and political activities, and sometimes even recount an entire
curriculum vitae (e.g., ʿAbadān 1). In contrast with the stereotyped
texts of ordinary building or votive inscriptions, which are character-
ized by their use of standardized patterns, commemorative texts are
freely composed due to the singular nature of the particular events
40
  Ibid., 545–550, and Macdonald, “Reflections,” 36–37 (cf. also the compilation
of all the textual evidence ibid., 61).
41
  These scribes sometimes recorded their names next to the inscription proper,
such as in ʿAbadān 2: w-ktb ḏn m2s3ndn ʾlzʾd 3 grbyn bn yḥḍb, “And this inscription
(sc. the huge commemorative text ʿAbadān 1 immediately above) has written Ilzaʾad
the stonemason, the son of Yaḥḍib.”
268 peter stein

and, hence, the lack of models that could have been imitated.
Inscriptions of this kind, even though they belong to the most exten-
sive epigraphic texts, are extremely rare. Stylistic elements of these
texts, however, also occur in the historical passages of votive inscrip-
tions.
There is one unique composition which stands out against all the
above-mentioned kinds of texts and deserves to be classified as “lit-
erature” in the strict sense of the word. This is the so-called Hymn
of Qāniya, a poem of twenty-seven lines rhyming on -ḥk, which is
addressed to the goddess Shams and dates from the end of the first
century ce.42 This text is also carved on a rock surface. It shows that
the Ancient South Arabian script was capable of rendering not only
prose but also rhymed poetry. The existence of a literary tradition,
thus, is not so much a question of script but rather a question of
writing materials: all of the extensive compositions just dealt with
have been recorded on stone blocks or rock surfaces, which provide
almost unlimited writing space, while the only format suitable for the
transport and storage of literature were wooden sticks with a mark-
edly limited surface. In principle it would of course have been possible
to record an extensive literary composition on a series of sticks tied
together, yet so far no evidence for this has been discovered.
There are no indicators of the existence of a literature of an epic,
mythological, or historical kind in pre-Islamic Arabia. Of course it
cannot be conclusively ruled out that these kinds of texts might have
been recorded on media not yet discovered, such as leather or palm
leaves. However, the evidence presented above suggests that if such
a literature existed, one might expect it to have been recorded at least
occasionally on rock and stone surfaces, too, or to have been cited
or alluded to in other inscriptions. The fact that this is not the case
warrants the assumption that literature of this kind did not exist in
written form, be it in South Arabia or elsewhere on the peninsula. In
fact, this fits quite well with what the Islamic sources have to say
about early Islamic teaching practices, which—just like the transmis-

42
  Cf. Robin, “Monuments,” 122–125; Beeston, “Antecedents,” 236ff.; Robin,
“Inscriptions,” 516–521.—In the meantime, two other poems in Sabaic language and
script have been identified: the rock inscription VL 24 = J 2353 from Wadi Shirjān,
which is called “chant” (s3mdt, cf. Robin, “Inscriptions,” 523, with n. 47) and is com-
posed of verses rhyming on –r (a fact that has recently been discovered by Anne
Multhoff), and a poem addressed to the God Almaqah, forming part of the dedicatory
inscription ZI 11.
 literacy in pre-islamic arabia 269

sion of poetry—appear to have been based mainly on oral

tradition.43

Who Did the Writing? The Question of Literacy in Pre-Islamic

Arabia

After having surveyed the material evidence remaining of pre-Islamic

Arabian writing, I will now turn to the question of who actually wrote
these documents. For even though the material presented above sug-
gests a fairly widespread use of writing at least in some parts of the
peninsula, this does not imply that a large number of people were
able to read and write. Statistical evidence on the rate of literacy in
pre-Islamic Arabia is of course unavailable, yet there are a number
of hints which might help us to approach the issue.
Of course, reliable information on the spread of reading and writ-
ing among the inhabitants of pre-Islamic Arabia is to be expected
above all from authentic firsthand sources. At first glance, the Arabian
Peninsula virtually appears like a stronghold of literacy: Thousands44
of rock graffiti, scattered over large areas of the peninsula and written
in numerous languages and dialects, testify to a centuries-old use of
writing even outside the elite of rulers and priests.45 However, in a
just published study entitled “Literacy in an oral environment,”
Michael C. A. Macdonald convincingly demonstrates that the wide-
spread use of a certain script by a group of people as a pastime, i.e.,
without any economic or communicative purpose (in contrast with
letters or legal documents), does not necessarily presuppose their
ability, or even interest in drawing up complex economic documents

43
  Cf. Schoeler, Charakter, 53–58, and Schoeler, “Schreiben,” 7–16. According to
Robin, “Inscriptions,” 559–560, the Bible—although it was the most important docu-
ment of the Christians and Jews who penetrated southern Arabia in the mid-first
millennium ce—was probably never translated into Sabaic. In contrast to this, the
victory inscription of the Abessinian king Kāleb Ellā Aṣbeḥā from Mārib, RIE 195
(Bernand et al., Recueil des inscriptions, 284ff.), which is as early as ca. 525 ce (for an
interpretation of this text, see Müller, “Bruchstücke”), contains numerous Ethiopic
quotations from the Bible.
44
  The Safaitic inscriptions alone amount to more than 20,000, while the Thamudic
ones number more than 11,000 (according to Macdonald, “Reflections,” 44–45).
45
  These short texts, which contain hardly more than personal names and the
respective genealogies, repeatedly hint at the nomadic life of their authors. Neverthe-
less, they do not contain a “message” proper, except for “I was here” (like the Arabic
tribal mark, or wasm).
270 peter stein

or letters.46 On the contrary, a nomad who has left, say, a Safaitic

inscription along his way through the Syrian desert in order to pass
the time may not have been able to use this script to any practical
purposes, since in the neighboring settlements, where such abilities
might have been necessary, very different systems of script and lan-
guage were used, namely Aramaic and Greek. Except for the markings
discussed above, there was thus no reason for these nomads to write
down their mother tongue, since they obviously did not engage in
any advanced mercantile activities, at least judging by their inscrip-
tions. Consequently, the authors of the Safaitic inscriptions are, so to
speak, literate persons in the midst of a non-literate society (thus
Macdonald, “Literacy”). The North Arabian epigraphic evidence
therefore does not warrant the assumption of literacy in the sense
that writing is used on a regular basis in order to perform a range of
communicative tasks within commercial and social life.
A similar picture can be sketched regarding South Arabia, which
is all the more remarkable since we actually do possess authentic
everyday documents in the form of letters and business documents
inscribed on wooden sticks. Of course, in contrast with the nomadic
north of the peninsula, we do have to assume here a basically literate
society in which the use of writing was part of ordinary business
activities, and even of private life. This does not, however, imply that
the ability to read and write was common knowledge among the
inhabitants of ancient South Arabia. Even though the numerous rock
graffiti of this region, which mostly contain personal names, do seem
to indicate a high degree of literacy, these graffiti display more or less
accurate imitations of the monumental script, normally carved by
professional stone masons, up to the latest period.47 Documents rel-
evant for everyday life, by contrast, were written in the minuscule

46
  I am indebted to the author for his kind permission to read and cite his ideas
while the article was still in press. In what follows, his elaborate argument—which is
based, among other things, on examples from both Europe and North Africa—can be
rendered only in a very condensed manner.
47
  Whether the authors of those graffiti were actually able to read the ubiquitous
monumental inscriptions remains a separate question. Basically, the representative
character of a monumental inscription is not diminished by the fact that most passers-
by are not able to read the text. There are numerous examples of monumental inscrip-
tions inside and outside South Arabia that are placed, for instance, in rather elevated
positions that make it impossible to read them from the ground. Hence, the purpose
of such inscriptions does not consist in being legible to everybody but rather in the
representative and public nature of its commanding position.
 literacy in pre-islamic arabia 271

script, which during the course of centuries had evolved into a com-
pletely cursive script of its own whose characters hardly show any
similarities to the monumental script of the same time (cf. fig. 1).48
Assuming widespread literacy in ancient South Arabia would thus
presuppose that its inhabitants had mastered two separate writing
systems, and consistently kept them apart. Apart from its lack of
practical plausibility, this assumption is disproven by some of the
details of content and structure exhibited by the minuscule
inscriptions.49
A centralized manufacture of the latter is suggested, first of all, by
some of their external characteristics,50 namely, the standardized shap-
ing of the writing surface with vertical borderlines and a symbol,51
(cf. fig. 2a and b), their generally precise style based on standardized
shapes,52 and, last but not least, the probable provenance of large
amounts of these sticks from one and the same deposit.53 Judging by
the content of these inscriptions, it also seems likely that they were
not written by private individuals but rather by professional scribes
specially trained for this purpose. Letters in particular generally refer
to their sender in the third rather than the first person and thus appear
to have been dictated to scribes, and also read by them after having
been delivered—a common practice among neighboring cultures.54

48
 See also Ryckmans, “Pétioles,” 32. All of the very few examples of minuscule
characters not inscribed on wooden sticks (cf. the references given in n. 28) do not
occur on rock!
49
  For the following, cf. in more detail Stein, Die altsüdarabischen Minuskelin-
schriften, 32-35.
50
 See already Ryckmans, “Pétioles,” 32–33.
51
  The background of these symbols still remains unclear; the obviously limited
repertoire of forms, however, and the fact that they occur both on letters (fig. 2a), and
on legal documents (fig. 2b), suggests that they did not originate in a private con-
text.
52
  Among the inscriptions known to date, there are only comparatively few cases
of an extraordinarily crude ductus.
53
  The bulk (if not all) of the hitherto known minuscule inscriptions on wooden
sticks from ancient South Arabia probably originates from one single spot near a
sanctuary in the ancient city of Nashshān (nowadays as-Sawdāʾ) in the Yemenite Jawf
(cf. J.-F. Breton in Ryckmans et al., Textes du Yémen, 3–4). Only a very small number
of Hadramitic inscriptions, mostly not yet deciphered, came to light by regular exca-
vations in the site of Raybūn in Hadramaut (cf. Frantsouzoff, “Hadramitic docu-
ments”).
54
  Although the Akkadian letters from Mesopotamia speak of the sender normally
in first person, the so-called Messenger Formula—ana X qibi-ma umma Y-ma, “To X
(= addressee) speak! Thus (says) Y (= sender): ...”—strongly hints at a comparable
practice. Similar formulas also occur in the Syrian cultural region (cf., for example,
272 peter stein

The comparatively rare instances where the sender of a letter identi-

fies himself as its writer either through his signature (four letters of
fifty that have so far been examined)55 or through his use of the first
person (three letters of fifty56) accord well with this hypothesis, for
professional scribes, too, might sometimes write letters of their own.
In this case, the identity of sender and scribe calls for the use of the
first person, in contrast with the large majority of letters written by
the scribe in the name of other, perhaps illiterate, persons. The
hypothesis that these scribes learned their skill within an organized
training system is supported by a number of exercise texts, which
reflect various stages of such a training curriculum.57
In view of the evidence presented above, there is no question of
widespread literacy among the inhabitants of pre-Islamic Arabia.
Basically, two levels of literacy must be distinguished. The first one,
which indeed extends throughout the entire peninsula and even
existed among nomads, consists in the ability to leave behind spon-
taneous and brief rock graffiti, which serve the sole purpose of passing
the time and fulfill no communicative function. These graffiti are
found even in the remotest area and use a writing material that is
abundantly at hand, namely, rock surfaces as they occur in their natu-
ral environment. In contrast with this, the second level of literacy
consists in a well-developed system of social and economic commu-
nication that was limited to a number of urban centers, located above
all in South Arabia. This latter form of literacy typically employs two
different types of script: A monumental one used for representative
purposes, which was written mainly on rock and stone, and an almost
completely different minuscule script that was used for matters of
everyday communication and written exclusively on handy wooden
sticks. A society that employs this latter type of literacy for purposes
of economic and social communication can indeed be called “literate.”
However, “literature” in the strict sense of the term was apparently
written in neither of these scripts but rather transmitted orally.

Schwiderski, “Präsenz des Boten,” and Tropper, “Briefformular,” 64–65). Further-

more, there are indications that the letter correspondence in ancient South Arabia
was organized between public “offices,” which means that the addressee was not given
his letter at home, but rather had it read to him by some official.
55
  For example, Mon. script. sab. 2: ʿbr yʿmr bn lbʾm 2 ʿm-n s2rḥm ḏ-qdḥn … … w-6
zbr s2rḥm, “To Yaʿmar, son of Labuʾum, from S2arḥum of (the family) Qadḥān: [...]
And S2arḥum has written (this)” (cf. fig. 3).
56
  For the evidence, see Stein, Die altsüdarabischen Minuskelschriften.
57
  Cf. for details Stein, Die altsüdarabischen Minuskelschriften, 38-39.
 literacy in pre-islamic arabia 273

To some extent, these conclusions may be extended to the urban

centers of the neighboring Ḥijāz. This would correspond to the Arabic
sources on the pre-Islamic and early Islamic periods.58 As Nöldeke
has demonstrated, what later Islamic authors have to say about the
Prophet in this regard is too tendentious and inconsistent to allow
an unequivocal judgement to be made.59 Nevertheless, Nöldeke con-
cedes that the Prophet (and several persons in his surroundings) may
have possessed the rudimentary kind of literacy necessary to conduct
commercial activities, while mastery of the more advanced skills nec-
essary to read literary works, for example, can be ruled out. From the
evidence presented above it is clear, however, that even if writing was
used for commercial purposes, this does not necessarily mean that a
large section of the population was able to read and write. Ancient
South Arabian documents suggest rather otherwise, since even per-
sons involved in business activities did not write by themselves but
employed professional scribes. Against this background, the existence
in Mecca and Medina of a number of people able to read and write,
as is alleged by the Islamic tradition, seems quite plausible. These
individuals may have been trained by members of the Jewish and
Christian communities, or may even have been part of them. Yet a
level of literacy beyond this can in all likelihood be ruled out.

58
  For the following, cf. Schoeler, “Schreiben,” 2f. with n. 8, and Robin, “Inscrip-
tions,” 557–561 (includes further references). Robin adduces an additional argument
for the largely oral character of the Prophet’s environment: the complete lack of con-
temporary written documents from the region— whether in the form of original
inscriptions or in the form of secondary hints at a literary tradition, such as transla-
tions of the Bible into the local languages.
59
  GdQ, vol. 1, 12–17.
 274
peter stein

Fig. 1. The development of the Ancient South Arabian monumental (musnad, left) and minuscule (zabūr, right) scripts from the Early
Sabaic to the Late Sabaic periods, i. e., from the eighth century bce to the sixth century ce.
 literacy in pre-islamic arabia

Fig. 2a. Example of a Sabaic letter (Mon. script. sab. 38).

275
 276
peter stein

Fig. 2b. Example of a Sabaic legal document (Mon. script. sab. 55).
 literacy in pre-islamic arabia

Fig. 3. Sabaic letter with signature of the sender (Mon. script. sab. 2)
277
278 peter stein

Bibliography

Sigla of inscriptions follow the list in Stein, Untersuchungen, 274–290. The publica-
tion of the inscriptions bearing the siglum “Mon. script. sab.” from the Bavarian
State Library in Munich is currently in progress (Stein, Die altsüdarabischen
Minuskelinschriften).
al-Ansary, A. R. Qaryat al-Fau: A Portrait of Pre-Islamic Civilization in Saudi Arabia.
Riyadh, 1982.
Beeston, A. F. L. “Musnad.” In EI2, vol. 7, 704–707.
Beeston, A. F. L. “Antecedents of Classical Arabic Verse?” In Festschrift Ewald Wagner
zum 65. Geburtstag, vol. 1, Semitische Studien unter besonderer Berücksichtigung
der Südsemitistik, edited by W. Heinrichs and G. Schoeler, 234–243. Beirut, 1994.
Beeston, A. F. L., M. A. Ghul, W. W. Müller, and J. Ryckmans. Sabaic Dictionary
(English-French-Arabic). Louvain-la-Neuve and Beirut, 1982.
Bernand, E., A. J. Drewes, R. Schneider, and F. Anfray, eds. Recueil des inscriptions
de l‘Éthiopie des périodes pré-axoumite et axoumite. Paris, 1991.
Blanck, H. Das Buch in der Antike. Munich, 1992.
Burton, J. “Muṣḥaf.” In EI2, vol. 7, 668–669.
EI2: The Encyclopaedia of Islam. New edition. Leiden, 1960–2002.
Frantsouzoff, S. A. “Hadramitic documents written on palm-leaf stalks.” Proceedings
of the Seminar for Arabian Studies 29 (1999): 55–65.
GdQ: Theodor Nöldeke. Geschichte des Qorāns. 2nd ed. 3 vols. (Vols. 1 and 2 revised
by Friedrich Schwally, vol. 3 revised by Gotthelf Bergsträsser and Otto Pretzl.)
Leipzig, 1909–1938.
Gerlach, I. “Der Friedhof des Awām-Tempels in Marib: Bericht der Ausgrabungen
von 1997 bis 2000.” Archäologische Berichte aus dem Yemen 9 (2002): 41–91.
Ghédira, A. “Ṣaḥīfa.” In EI2, vol. 8, 834–835.
Grohmann, A. Arabische Paläographie. Vol. 1. Vienna, 1967.
Harris, W. V. Ancient Literacy. Cambridge, 1989.
Jamme, A. Miscellanées d’ancient arabe III. Washington, 1972.
Leslau, W. Comparative Dictionary of Geʿez (Classical Ethiopic). Wiesbaden, 1991.
Macdonald, M. C. A. “Reflections on the linguistic map of pre-Islamic Arabia.”
Arabian Archaeology and Epigraphy 11 (2000): 28–79.
Macdonald, M. C. A. “Literacy in an Oral Environment.” In Writing and Ancient
Near Eastern Society: Papers in Honour of Alan R. Millard, edited by P. Bienkowski,
Ch. Mee, and E. Slater, 49–118. New York and London, 2005.
Macdonald, M. C. A. “Old Arabic (epigraphic).” In Encyclopedia of Arabic Language
and Linguistics, edited by K. Versteegh et al., vol. 3. Leiden 2008, 464–477.
Maraqten, M. “Writing materials in pre-islamic Arabia.” Journal of Semitic Studies
43 (1998): 287–310.
Müller, W. W. “Zwei weitere Bruchstücke der äthiopischen Inschrift aus Mārib.”
Neue Ephemeris für semitische Epigraphik 1 (1972): 59–74.
Müller, W. W. “L‘écriture zabūr du Yémen pré-islamique dans la tradition arabe.”
In Ryckmans et al., Textes du Yémen antique, 35–39.
Nebes, N. “Ein beschrifteter Goldanhänger aus dem Friedhof von Šaʿūb.” Archäolo­
gische Berichte aus dem Yemen 9 (2002): 227–231.
Pirenne, J. Les témoins écrits de la région de Shabwa et l’histoire (Fouilles de Shabwa I).
Paris, 1990.
Ricks, S. D. Lexicon of Inscriptional Qatabanian. Rome, 1989.
Robin, Ch. “Les plus anciens monuments de la langue arabe.” In L’Arabie antique
de Karib˒īl à Mahomet: Nouvelles données sur l’histoire des Arabes grâce aux
inscriptions, edited by Ch. Robin, 113–125. Aix-en-Provence, 1991.
 literacy in pre-islamic arabia 279

Robin, Ch. “Documents de l’Arabie antique III.” Raydān 6 (1994): 69–90.

Robin, Ch. “Les inscriptions de l’Arabie antique et les études arabes.” Arabica 48
(2001): 509–577.
Ryckmans, J. “Une écriture minuscule sud-arabe antique récemment découverte.”
In Scripta Signa Vocis: Studies about Scripts, Scriptures, Scribes and Languages in
the Near East, presented to J. H. Hospers by his pupils, colleagues and friends,
edited by H. L. J. Vanstiphout et al., 185–199. Groningen, 1986.
Ryckmans, J. “Pétioles de palmes et bâtonnets inscrits: un nouveau type de docu-
ments du Yémen antique.” In Ryckmans et al., Textes du Yémen antique, 25–34.
Ryckmans, J. “Pétioles de palmes et bâtonnets sud-arabes inscrits: notes de paléog-
raphie.” In Arabia Felix: Beiträge zur Sprache und Kultur des vorislamischen
Arabien. Festschrift Walter W. Müller zum 60. Geburtstag, edited by N. Nebes,
250–259. Wiesbaden, 1994.
Ryckmans, J. “Origin and evolution of South Arabian minuscule writing on wood
(1).” Arabian Archaeology and Epigraphy 12 (2001): 223–235.
Ryckmans, J., W. W. Müller, and Y. M. Abdallah. Textes du Yémen antique inscrits
sur bois (with an English Summary). Louvain-la-Neuve, 1994.
Schoeler, G. “Schreiben und Veröffentlichen: Zu Verwendung und Funktion der
Schrift in den ersten islamischen Jahrhunderten.” Der Islam 69 (1992): 1–43.
Schoeler, G. Charakter und Authentie der muslimischen Überlieferung über das Leben
Mohammeds. Berlin, 1996.
Schwiderski, D. “Die Präsenz des Boten in kanaanäischen und aramäischen Brief­
formularen des 1. Jahrtausends.” In Wagner (ed.), Bote und Brief, 103–112.
Schwiderski, D., ed. Die alt- und reichsaramäischen Inschriften. Vol 2. Texte und
Bibliographie. Berlin, 2004.
Sedov, A.W. “Die archäologischen Denkmäler von Raybūn im unteren Wādī Dauʿan
(Ḥaḍramaut).” Mare Erythræum 1 (1997): 31–106.
al-Selwi, I. Jemenitische Wörter in den Werken von al-Hamdānī und Našwān und
ihre Parallelen in den semitischen Sprachen. Berlin, 1987.
al-Shaḥri, ˓A. A. M. “Recent epigraphic discoveries in Dhofar.” Proceedings of the
Seminar for Arabian Studies 21 (1991): 173–191.
Sima, A. Die lihyanischen Inschriften von al-ʿUḏayb (Saudi-Arabien). Rahden, 1999.
Sima, A. “Kleinasiatische Parallelen zu den altsüdarabischen Buß- und
Sühneinschriften.” Altorientalische Forschungen 26 (1999): 140–153.
Sima, A. “Die hasaitischen Inschriften. ” In Neue Beiträge zur Semitistik: Erstes
Arbeitstreffen der Arbeitsgemeinschaft Semitistik in der Deutschen Morgenländischen
Gesellschaft vom 11. bis 13. September 2000 an der Friedrich-Schiller-Universität
Jena, edited by N. Nebes, 167–200. Wiesbaden, 2002.
Stein, P. “Schreibfehler im Sabäischen am Beispiel der mittelsabäischen
Widmungsinschriften.” Le Muséon 115 (2002): 423–467.
Stein, P. “The inscribed wooden sticks of the Bayerische Staatsbibliothek in Munich.”
Proceedings of the Seminar for Arabian Studies 33 (2003): 267–274.
Stein, P. Untersuchungen zur Phonologie und Morphologie des Sabäischen. Rahden,
2003.
Stein, P. “The Ancient South Arabian Minuscule Inscriptions on Wood. A New
Genre of Pre-Islamic Epigraphy.” Jaarbericht EX ORIENTE LUX 39 (2005):
181–199.
Stein, P. “Stein vs. Holz, musnad vs. zabūr – Schrift und Schriftlichkeit im vorisla-
mischen Arabien.” Die Welt des Orients 35 (2005): 118–157.
Stein, P. Die altsüdarabischen Minuskelinschriften auf Holzstäbchen aus der
Bayerischen Staatsbibliothek in München. Vol. 1. Die Inschriften der mittel- und
spätsabäischen Periode. Tübingen. Forthcoming.
280 peter stein

Tropper, J. “Das ugaritische Briefformular.” In Wagner (ed.), Bote und Brief,

63–74.
Wagner, A., ed. Bote und Brief: Sprachliche Systeme der Informationsübermittlung
im Spannungsfeld von Mündlichkeit und Schriftlichkeit. Frankfurt, 2003.
Wissmann, H. v. Die Geschichte von Sabaʾ II. Das Großreich der Sabäer bis zu seinem
Ende im frühen 4. Jh. v. Chr, edited by W. W. Müller. Vienna, 1982.