Dunn has written an interesting account of the relations between the Roman and other bishops at the time of St Cyprian. He has clearly sought to enter the much-discussed territory of the precise power of a Roman pontiff at this time, and comes to the unremarkable conclusion that the bishop of Rome did not possess jurisdiction in the later sense: each bishop was inviolable in his diocese with his authority resting on his intercommunion with other bishops.

Clearly, in order to justify such a position, he has to argue against a variety of controversial views of Cyprian's true position. The first and central point is clearly the discussion of the two version of De unitate 4–5 and the manuscript variants of ch. 19 that are assigned to the ‘Primacy’ version (P) as opposed to the ‘received’ one (TR). Here the conclusion that he seeks is that the meaning of Matthew xvi.18–19 and John xx.21–3 for Cyprian is that every individual bishop fulfils the promises to Peter in his own Church.

In order to establish this point, Dunn now has to secure an argument, in which he follows Hall, to the effect that both P and TR relate essentially to the changed internal condition of the Church of Carthage between 251 and 252. Thus he believes that P was directed against Felicissimus and the laxist party in its insistence on unity with the bishop (Peter) in his diocese, whereas TR was directed against Novatian ‘when rival bishops were multiplying in numbers’ (p. 79).

Thus, by claiming that both recensions apply to internal divisions at Carthage, he can support his contention that every bishop, for Cyprian, is Peter in his particular Church. If, alternately, we accept Bévenot's interpretation that TR reflects the controversy with Stephen on rebaptism, of course the texts do represent some position on a specifically Roman primacy. I remain convinced that this is the case. John xx, Cant. 4 are used in the baptism controversy and not before, so that it does appear somewhat cavalier to conclude that ‘this does not mean that they were written at the same time’. There is no absolute necessity here but it is a strong indication that in fact they were.

Furthermore Novatian was, as Dunn is aware, a particularly Roman problem, which leads him to suggest that the De unitate was sent to Rome but only because its address to a particularly African problem could be modified so as to be of use to Cornelius as a help to him as bishop in his particular diocese. But the evidence of ep. lix.14.1 ought to be decisive against such a view since Cyprian describes to Cornelius how Fortunatus ‘the pseudo-bishop’ and Felicissimus the deacon had ‘dared to sail both to the chair of Peter and to the primordial (principalem) church from where the unity of the sacred bishops (unitas sacerdotalis) derived its origin (exorta)’. Here, clearly, Cornelius is not being described as an individual bishop in his own church and equal to all others.

It is here that I find Dunn at his most disingenuous when he described Cyprian's words as ‘more flattery than plain speaking’ (p. 95). He is right to deny any claim here to jurisdiction of one diocesan over another as would occur later. But that does not mean that Cyprian was not being moved by theological considerations, however embryonic and ill thought out. It is not simply the numerical size of the Roman Church when compared to the others that purely contingently gave it the weight that it had, as Dunn at one point hesitantly concedes (p. 55).

Dunn's approach seems to me in other words to ignore the logic of Cyprian's developing position. In De unitate 4.96 (TR version), Cyprian states that ‘the rest of the apostles … were endowed with equal honour and power, nevertheless (sed) the starting point (exordium) proceeds (proficiscitur) from an undivided whole (ab unitate), in order that the Church of Christ may be revealed to be one’. I have translated exordium here as ‘starting point’ even though I believe that there is a better translation. Another meaning of exordium is, according to the editors of the Oxford Latin dictionary, ‘the warp set up on a loom before the web is started’, and, by extension, the ‘weaving’ of fate. Cyprian's view of the Church is of a network of bishops bound together by the bonds of mutual recognition and inter-communion, as Dunn is aware, but the logic of such a network is that it has, like the weaving of a piece of cloth, a starting point in the process. This theological sense, derived from his theological model of the undivided ‘robe’ of Christ, is what for him necessitates Rome as the centre of the web as it were, holding its unity together.

Thus I believe that Cyprian, in his confrontation with Stephen, was wrestling with a problem of his own (and not, as Dunn believes, of Stephen's own) making. How is the episcopal college to react when the exordium behaves differently from how the model predicted that it would and ought to behave? I see no reason to see version TR as anything other than produced therefore in response to Stephen. Apart from the chronology of proof texts used by Cyprian, it is in the light of the re-baptism controversy that we should understand the ‘plea bargain in bad faith (perfida praevaricatio)’ of De unitate 4.126: the ‘deal’ for readmission of schismatics is exactly what this was: there was nothing comparable with either Novatian or the laxists.

Despite these reservations concerning the form in which he has chosen to express his main thesis, Dunn has made a useful contribution to many aspects of Cyprianic studies and is to be commended.