The chronology of the final books of City of God: data and hypotheses

A case-study in Augustinian chronology, this article seeks to lay out, in full, the evidence for the dating of the last five books of De ciuitate dei. Modern chronologieshave proceeded backwards from the writing of Retractationes by 427 and forwards from the manifestly erroneous chronological data given in De ciuitate dei 18.54, which places it ‘about thirty years’ after March 399. While the resulting dates are broadly correct, this reasoning overlooks an important set of homiletic data. Following a now-neglected argument, the Maurists and Tillemont derived a date from De ciuitate dei 22.8 and a series of sermons (320–323) connected to the cult of St. Stephen. Supplemented with information provided in S. 355–356 and 319 and buttressed by new manuscript readings, that argument remains valid. Except perhaps for a portion of book 22, De ciuitate dei will have finished either by summer of 423 or by summer of 426, before Augustine’s semi-retirement that September. Per the later (and usual) dating, book 18 will belong in 423 or 424, which makes its vague chronology and omission of the (otherwise apropos) destruction of the temple of Caelestis in 421 especially problematic. The article therefore suggests four ways of explaining the oddities surrounding book 18: Augustine’s authorial choice, a shorter chronology (split in two due to uncertainty over Paschal dates), and a hiatus in composition ca. 421.