Late Roman Census Inscriptions: Dating and Context

Kyle Harper (University of Oklahoma)

There is a remarkable dossier of Greek inscriptions from the later Roman empire which records in great detail the fiscal liabilities on a number of rural properties. The inscriptions, manifestly coeval, come from nine cities, all in the eastern Aegean or western Asia Minor. Their content is a rather forbidding jumble of abbreviations and fractions, and they have been little used outside the specialist literature on the algebra of late Roman taxation. No analysis has taken account of all the inscriptions, and not since an article by A. H. M. Jones in 1953 has anyone considered their value as evidence for social and economic history. Important new fragments, recently published, have not yet been incorporated into a broader understanding of this unique set of documents. The inscriptions are overdue for a reappraisal.

A new evaluation of the inscriptions should begin by reconsidering their date and context. The conventional wisdom is that the stones are Diocletianic, simply because his reign was the occasion for a dynamic reform of Roman fiscal assessment. But on close inspection, there are considerable reasons to believe they must be later. In this paper I will argue, using the evidence of scripts, onomastics, language, and prosopography, that the stones were engraved in the later fourth century. I propose that the reign of Valens marks the most likely context for the inscriptions. The consequences of this new dating are significant, not least because it allows us to reassess the purpose of the stones. I argue that the impetus for creating a monumental record of tax liability was not the census of Diocletian. Rather, the creation of large, easily-audited tax records should be related to specific political tensions during the reign of Valens, when the imperial center tried to wrest control over the process of taxation from the town councils.