last month
next month

Centre for the Study of Manuscript Cultures (CSMC)

09/2015 manuscript  of the month

Was there a slave trade in southern Arabia in pre-Islamic times?

How two twin girls were sold according to a Sabaean document written on wood

Legal documents, trade agreements and other formal records documenting daily life in ancient southern Arabia have only been known to exist for a few decades. The inhabitants of this part of the Arabian Peninsula – Semitic-speaking tribes including the Sabaeans – did not write their correspondence in ink, but carved it on small pieces of wood, cigar-sized sections of either palm-leaf ribs or other kinds of wooden material. Virtually unlimited quantities of such matter were available as ‘waste’ from plantation work and were thus the simplest and cheapest materials on which to write that we know of, with the exception of potsherds. This unique manuscript culture existed in the southern part of the Arabian Peninsula from the early years of the first millennium BCE right up to the sixth century CE, directly prior to the onset of Islam. The document shown here on a piece of juniper not only demonstrates the nature of legally binding documents in southern Arabia, but is also extremely important in terms of the region’s social history. The text states that two girls – twins, in fact – were handed over into the ownership and control of members of their own family. But why their own family, of all people?

Let us take a look at the material used for the manuscript first of all. In ancient times, everyday correspondence in southern Arabia was carved on simple wooden sticks. This method of writing, which is unique to this part of the Middle East, is to some extent reminiscent of epigraphs, which is why these texts have also been classed as inscriptions in academic literature. However, they differ considerably from inscriptions chipped into stone and rock monuments in terms of the material on which they were written and the style of writing that was employed. Unlike the pointed geometrical letters found in inscriptions, the slanted and rounded shapes of the manuscripts in question made fluent writing easier, so it is quite appropriate to speak of a ‘cursive’ script in this case, even if the letters are not linked together in a regular way. This type of cursive script was no doubt notched into the surface of a soft, freshly cut wooden stick quite quickly using a sharp stylus.

Fig. 1: Mon.script.sab. 1, beginning of the document (lines 1–3) > Enlarge

The size of the writing area was defined by the diameter and length of the piece of wood used and varied considerably as a result. Despite its unusual thickness (almost 4 cm in this case), the item displayed here is certainly not one of the largest finds of its kind, the biggest of which are actually up to half a metre long and may contain much more writing. The writing area is often delimited by vertical lines. In the case of round pieces of wood like the one pictured here, a continuous horizontal line also marks the point where the text begins (see fig. 1). A symbol is often found in the right-hand margin next to the writing area, which possibly denotes the place where the text was originally written.

Fig. 2: Mon.script.sab. 1, continuation of the text (lines 4–7) > Enlarge

Fig. 3: Mon.script.sab. 1, continuation of the text (lines 7–10) > Enlarge

In summary, the content of the document (figs. 1–4) is as follows: a man called Saʿdṯawān from the clan of the Ḏū-Yaʿūd grants a member of another family, Wahbšabʿān from the clan of the Šahrum, possession of the two twin girls mentioned above, who are called ʾAmatʾabīhā (which translates as ‘Her father’s servant’) and Marsūʿat and are referred to as ‘daughters’ (bnt) of the same clan, the Šahrum. The next part of the text describes exactly what this transfer of ownership means: the new owners may ‘take possession of the girls and command them [to do things]’ (bʿl w-hwṣtn), ‘sell them’ (hšʾmn) or ‘barter them for something else’ (qyḍ). This legal status the girls now had, which incidentally also applied to any descendants they might have had one day, was due to their parentage: they were the children of a female ‘servant’ of the clan to which the donor and author of our document, Ḏū-Yaʿūd, belonged. In this case, however, the woman was not a ‘slave’ in the sense of personal, material ownership, but a member of a subordinate clan that was dependent to a certain extent on the author’s own family, which enjoyed more social prestige. Patron–client relationships of this kind – in which members of less influential clans joined higher-ranking clans as ‘servants’ and consequently came under their protection – were characteristic of ancient tribal society in southern Arabia and no doubt helped to create a degree of stability in its social and political fabric.

Fig. 4: Mon.script.sab. 1, final part of the text (lines 11–14) with the author’s signature at the end of the very last line > Enlarge

If members of two clans that were linked together this way happened to have sexual relations with each other, one question was bound to be asked: what was the social status of any children born as a result? In the patrilineal structure of ancient tribal society in southern Arabia, the descendants stemming from such a relationship – which was always between a man from a higher-ranking clan and a woman from a lower-ranking one as far as we can tell – were counted as belonging to the father’s family, which determined their destiny. Whenever the descendants were girls, who generally had a lower status in such a society anyway, the texts repeatedly speak of them being sold or given away by their father. As the document we see here also suggests, this could even be a way of repaying a debt to another person, for example. A Sabaean decree issued in the second century BCE to prevent practices of this kind from taking place actually goes as far as to forbid the killing of girls. The ‘sale’ of a daughter is therefore most likely to have been due to her family being in a difficult financial position at the time. This has nothing to do with trading them in the sense of selling slaves for a business, though, despite the commercial terminology used.

But let us return to our text now (fig. 5). The question that arose at the beginning still needs to be clarified, after all: why were the two girls handed over to members of their own family, the Šahrum clan, of all things? We will discover the answer to this if we take the social relations described above into account, which show that one family was dependent on another. As we have seen, the family of the two girls and the persons benefiting from the ‘sale’ were both ‘clients’ of the family to which the author and donor of the document belonged. This can only be the father of the two children, whom he would have had to have accepted into his family as that was the customary practice at the time. It seems the biological father only took a limited interest in the two illegitimate girls, which would explain why he passed their ‘ownership’ on to the mother’s own family – the Šahrum clan – in a documented legal act; they were the ones who had to look after the twins from then on.

Fig. 5: Mon.script.sab. 1, facsimile of
the document (with the written area
reproduced after creating an impression
of it) > Enlarge

The legal validity of the procedure is demonstrated by the usual formal aspects of the text: a phrase denoting the date is followed by the donor’s own signature in the very last line, for example. As the rather coarse style employed shows, not only the signature at the end of the line, but the name of the author preceding it (Saʿdṯawān) were written by the same hand. The document itself, however, was written meticulously by a professional scribe. Unlike numerous other documents from the same find (see fig. 6), the document was not annulled by scoring or notching at a later point in time, which means it would have continued to be valid for the twin girls’ descendants. If they had ever wanted to contest their social status, their father’s family would have referred them to the document, which was probably kept in the town’s archive.

There is no record to say whether an objection of this kind actually ever took place. The document itself survived for almost two thousand years in the dry desert climate of northern Yemen and was discovered by chance by late descendants of the Sabaeans in the twentieth century.

Fig. 6: Sabaean bond (Mon.script.sab. 252 = X.BSB 58), which was cancelled by notching the writing area once the legal matter described in it was dealt with. > Enlarge


ARBACH, Mounir (2005): ‘Le commerce d’esclaves en Arabie du Sud préislamique d’après une nouvelle inscription sabéenne du VIIe s. av. J.-C.’ In: A. V. Sedov / I. M. Smiljanskaja (eds.): Arabia Vitalis. Arabskij Vostok, islam, drevnjaja aravija. Sbornik statej posvjaščennyj 60-letiju V. V. Naumkina. Moskva: Rossijskaja Akademija Nauk / Moskovskij Gosudarstvennyj Universitet, 314–317.

MÜLLER, Walter W. (1983): ‘Ein Verbot, Töchter wegzugeben oder zu töten.’ In: Otto Kaiser et al. (eds.): Texte aus der Umwelt des Alten Testaments. Band 1 Lieferung 3: Dokumente zum Rechts- und Wirtschaftsleben. Gütersloh: Gütersloher Verlagshaus Gerd Mohn, 275–276.

MULTHOFF, Anne (2013): ‘“Und es sei aufgelöst und aufgehoben”. Zur Annullierung juristischer Urkunden im vorislamischen Südarabien.’ In: Françoise Briquel-Chatonnet / Catherine Fauveaud / Ivona Gajda (eds.), Entre Carthage et lʼArabie heureuse. Mélanges offerts à François Bron. Paris: De Boccard, 105–118.

STEIN, Peter (2010): Die altsüdarabischen Minuskelinschriften auf Holzstäbchen aus der Bayerischen Staatsbibliothek in München. Band 1: Die Inschriften der mittel- und spätsabäischen Periode. Tübingen / Berlin: Wasmuth.

Bavarian State Library, Munich, Germany
Shelf mark: Mon.script.sab. 1 = X.BSB 61
Material: juniper, inscribed all the way round; writing area at the top and at the sides delineated by carved lines.
Dimensions: length 12.8 cm; diameter 3.7 cm; Sabaean text written on 14 lines around the stick, which cover three quarters of the available writing area.
Provenance: central highlands of Yemen, c. mid-3rd cent. CE (dated using palaeographical criteria and an eponym mentioned in the text). The find was kept in the archive of the town of Naššān in ancient times (which is now known as as-Sawdāʾ) in the northern Wadi al-Ǧawf, where it was later excavated along with numerous other documents carved on wood.

Text by Peter Stein
© for fig. 1–4, 6: Bayerische Staatsbibliothek München;
fig. 5: Peter Stein
Reference note:
Peter Stein, “Was there a slave trade in southern Arabia in pre-Islamic times?”
In: Andreas Janke (ed.): Manuscript of the Month 2015.9, SFB 950: Hamburg,