Anomalies paléographiques des chiffres sur les inscriptions africaines de l’antiquité tardive

The author presents the results of a survey for which he has accumulated the necessary data a number of years. Its starting point is a phenomenon which appears on African inscriptions from the 3rd century. In indicating hundreds, the ordinator [the man who planned the layout of the inscription], followed by the stonecutter, kept the minuscule form C from the draft text, probably to distinguish those letter Cs which represented figures from those which functioned as letters, and he did the same for the letter L, which could also be a figure, 50, or a letter. The beginning of this practice coincides with the change from the old form of minuscule (Old Roman Cursive) to the new form of minuscule (New Roman Cursive), a period during which the old form of the minuscule C continued to be used both in manuscripts and on inscriptions. The author demonstrates from a catalogue of these inscriptions (table 1 + 2, figs 3-10), the prototype being the Severan irrigation regulations from Lamasba, fig. 2, and from a catalogue of scripts in manuscripts drawn up from papyri, from ostraca, from wax, wood or bronze tablets, and from dipinti on amphorae (table 3, figs 11-12 and 14) the hypothesis that late antique script of manuscripts uses the same letter C for figures and the same archaic script. Having listed and classified the types of ductus (with 4, 3, 2 strokes, and finally a single one: see fig. 13) in the manuscripts (table 3), the author compares them with the forms encountered on African inscriptions (table 4). It appears that form III (a ligature with 2 strokes), attested as early as the Lamasba text and present everywhere during the 4th century, co-existed for a while with form II. The author is therefore able to suggest a chronological and geographical classification of three zones of workshops (table 5), which he then illustrates in statistical form (table 6). These observations enable us to qualify the hypothesis of the supposed ‘3rd-century change in scripts’. It must surely have been a process rather than an event, unequal in its effects and spread over a period of time. What is needed now is a research in other provinces. [J.-M. C., transl. by M. J. Jones]