Old Arabian script was used from perhaps as early as the eighth century BCE to as late as the fourth century CE, undergoing alterations as it spread across the vast Arabian Peninsula. Of the regional varieties, it is those conventionally termed Thamudic and Safaitic that appear in North Arabia. Occasionally, finds have been made of Old Arabian texts in the Levant of the Hijaz variety called Dedanite and Lihyanite, but the vast majority are concentrated at al-῾Ula (Dedan) in the Hijaz. The terms Thamudic and Safaitic are commonly used as conventional labels for these North Arabian inscriptions. They are in a South Semitic script and appear mainly on the desert fringe of the Levant. It is commonly assumed that the script evolved from South Arabian through the influence of Dedanite and Lihyanite, with the various script types of Thamudic developing from those earlier writing systems and ending with the Safaitic texts. The language of the Thamudic and Safaitic texts represents the linguistic family of what has been called Early North Arabian. The most distinguishing linguistic feature of these dialects is the use of the definite article h-.

Thamudic texts are found throughout the Arabian Peninsula and represent the oldest of the North Arabian texts. The concentration of such texts in the northern Hijaz led initially to their ascription to the tribe of Thamud, known from the Assyrian annals and the Qur'an. Attempts have also been made by some scholars to associate the users of the script with the descendents of the Midianites. Neither theory has gained acceptance. The earliest forms of the script seemed to have emanated from the North Arabian oases of Tayma' and Dedan in the Hijaz, from which they spread elsewhere. [See Tayma'; Dedan.] In 1937, Fred V. Winnett proposed classifying the texts into five types labeled A–E, suggesting a sequential chronological and geographic development for the script. The problem with this classification is that the categories are not ironclad: some texts exhibit the characteristics of more than one group. Primitive forms also appear with more evolved forms of the script, defying any simple chronological arrangement of the texts. In 1970, Winnett revised the categories into geographic designations, with Najdi (B), Hijazi (C, D), and Tabuki (E) replacing the older categories. Both systems are still utilized for convenience, in full recognition that further modification is needed. [See the biography of Winnett.] The now more than fifteen thousand recorded Thamudic texts are mainly graffiti, containing the name and patronymic of an individual, sometimes with a petition to a deity or an individual and accompanying phrase. Their laconic nature makes dating precarious.

Safaitic inscriptions primarily appear in southern Syria, northern Jordan, and northern Saudi Arabia, but a few isolated finds have been made in western Iraq, on Lebanon's coasts and Biqa῾ Valley, and as far afield as Pompeii, Italy. The name Safaitic is derived from the ṣafa, the basalt desert southeast of Damascus, Syria; the term is, however, a misnomer resulting from the initial discovery of the texts in the region In 1857. Most of the fourteen thousand published texts are concentrated in the basalt desert, or ḥarra, of northeast Jordan and neighboring southern Syria. Literally thousands more exist in this region and need to be recorded. The main contributors to the decipherment of the texts were Joseph Halévy In 1877, F. Praetorius In 1883, and E. Littmann In 1901. The latter was the first to recognize that the alphabet consisted of twenty-eight letters comparable to the Arabic alphabet. Some of the Thamudic E, or Tabuki, texts have a more marked similarity with the orthography of Safaitic than the common varieties of Thamudic and have been designated South Safaitic by some. The texts frequently contain extensive geneaologies that trace the lineage back to an eponymous ancestor; they sometimes include an individual's social group or tribe ('l), along with various formulae and religious petitions, with occasional allusions to some historical event. The date is normally of an unidentifiable or ambiguous nature.

Few of the Safaitic texts can be dated with any precision, but they seem generally to begin as early as the first century BCE and cease probably by the fourth century CE because they reflect no evidence of any Christian or Islamic influence. The overwhelming impression is that they mainly stem from the first century CE, based on references to the Herodian dynasts, Nabatean kings, and Roman authorities, as well as the “Jews” (yhd), Itureans (yẓr), and Persians (mḏẖy). Because most of the Thamudic and Safaitic inscriptions are found in the remote desert regions of North Arabia, some distance from settled regions and villages, they have traditionally been perceived as the product of a nomadic population. Many of the texts speak of seasonal migrations with flocks and herds and many are accompanied by rock-art drawings of camels and horses, suggesting that pastoralists are the authors. Others, however, speak of attachments to villages and express intimate knowledge of political developments and imperial authorities, implying a population that was not just on the fringe of the sedentary population. The deities mentioned in the texts also include the Syrian Ba῾al Shamin and Nabatean Dushara—cults associated with the shrines and temples of the settled population. Additional texts are being found scattered along the villages and settlements of the Transjordanian plateau and southern Syria.


  • Graf, David F. “Rome and the Saracens: Reassessing the Nomadic Menace.” In L'Arabie préislamique et son environnement historique et culturel: Actes du Colloque de Strasbourg, edited by Toufic Fahd, pp. 341–400. Leiden, 1989. Adventuresome reappraisal of the traditional “nomadic” interpretation that needs to be refined in light of newly published texts.
  • Harding, G. Lankester. An Index and Concordance of Pre-Islamic Arabian Names and Inscriptions. Toronto, 1971. Complete listing of the publications and onomasticon of the texts.
  • Macdonald, M. C. A. “Nomads and the Ḥawrān in the Late Hellenistic and Roman Periods: A Reassessment of the Epigraphic Evidence.” Syria 70 (1993): 303–413. Defense of the traditional “nomadic” view (formulated by the pioneer interpreters of the texts) that ignores the implications of many of the texts.
  • Winnett, Fred V., and G. Lankester Harding. Inscriptions from Fifty Safaitic Cairns. Toronto, 1978. Valuable collection and discussion of many of the texts.

David F. Graf