Kaldellis and Kruse are meticulous scholars who have read the evidence regarding the commanders of late Roman field armies in the east closely and interpreted it aggressively, rejecting orthodoxy on the Notitia Dignitatum. Four short chapters covering 361–395, 395–450, 450–506 and 506–630 ce are supported by four long appendices, three on the prosopography of Roman generals (at Adrianople, under Theodosius I, from Arcadius to Heraclius) and one on the date of the eastern Notitia. The book presents two major arguments, that the system of regional and praesental generals described in the eastern Notitia was not put in place until the 440s and that the eastern chapters of the Notitia therefore date to the 440s.

The eastern Notitia recorded two commanders of imperial field armies (praesentales) and three regional commanders, in Thrace, Illyricum and Oriens. It is traditionally dated to c. 395, for example in A.H.M. Jones's Later Roman Empire (1964), PLRE and A. Demandt's 1970 piece (RE Suppl. 12, 553–790), with the military arrangements projected backwards to the mid-fourth century. K. and K. reject this orthodoxy. Chapter 1 starts in 361, so, unfortunately, omits discussion of arrangements under Constantius II, although they do see ‘a number of quasi-formalized regional commands, specifically for Gaul, Illyricum, and Oriens’ (p. 99, cf. p. 8). K. and K. argue that the two magistri who usually accompanied emperors had the titles magistri equitum, magistri peditum or magistri equitum et peditum until early in the reign of Theodosius I, but were then called magistri utriusque militiae or magistri militum. This is based on CT 12.1.113, which referred to magistri equitum et peditum in 386 and CT 1.21.1 of 393 directed to magistri utriusque militiae (p. 17). As they acknowledge, there are exceptions to this change in title: for example, Socrates, HE 6.6.2, describes Gainas as ‘stratēlatēs of both cavalry and infantry’, Basiliscus, in an inscription from Philippopolis (AE 1948.43 = LSA 367), was magister equitum peditumque, and on a consular diptych Justinian was magister equitum et peditum praesentalis. As they show clearly, Roman writers were often vague about military offices, with Ammianus Marcellinus calling Victor in 378 both magister equitum (31.12.6) and comes (31.13.9) and describing Equitius in 370 as per Illyricum eo tempore magistrum armorum (29.6.3). Inscriptions were also inconsistent, with the same Equitius being magister equitum peditumque in the mid-360s (CIL 3.10596) and magister utriusque militiae in 370 (CIL 3.5670a).

Based on the lack of titles matching the Notitia, K. and K. reject magistri with formal regional commands in the fourth century in Thrace and Illyricum, although they accept a Theodosian magister utriusque militiae per Orientem, Addaeus, based on CT 16.8.9 of 24 September 393. This regional specification falls between two other laws to Addaeus addressing him as magister utriusque militiae, CT 1.5.10 of 12 January 393 (O. Seeck, Regesten, 100, argued this should be June) and CT 6.24.6 of 3 October 395. Despite this they argue that the laws’ failure to specify regions in titles is significant (even though most laws in the CT, CJ or Novellae also omit regions for praetorian prefects). These are familiar difficulties. Orthodoxy until now has interpreted less precise titles using the framework of the Notitia; K. and K. prefer to argue that ‘the headings reflect a consistent reality’ (p. 25).

Chapter 2 (395–450 ce) shows a preference for generic unassigned magistri militum rather than poorly specified regional and praesental magistri. ‘Even though we have clear ad hoc precedents for both the MMO and the MMI in the late fourth century (and for the MMT in 412), it is not necessary to assume that these assignments anticipated or were directly formalized into the arrangements found in the Notitia’ (pp. 122–3). K. and K. argue that the structure seen in the Notitia was created late in the reign of Theodosius II in response to the Huns, and thus dismiss the 412 attestation of Constans as magister militum per Thracias in CT 7.17.1 (pp. 32–3) as exceptional. Chapter 3 (450–506 ce) shows how much better the literary evidence for generals’ titles becomes after 450 when the loose writing of many earlier authors was replaced with authors such as Malchus or Malalas, who were prepared to use technical language that matches the Notitia. K. and K. show that the first evidence for praesentalis in eastern titles appeared in the 440s although they accept that ‘the practice of appointing paired sets of magistri equitum and peditum (“masters of cavalry” and “of infantry”) was standard in the late fourth century and continued into the fifth’ (p. 131).

Following their argument that the evidence for both regional and praesental magistri militum does not match the Notitia until the 440s, K. and K. suggest that this was the date of the Notitia's eastern chapters and state that the onus probandi is now on scholars who follow the traditional dating (pp. 25, 154, cf. 176). Appendix 4 analyses the dating of the eastern chapters of the Notitia only, correctly noting that, ‘if we had the ability to verify its contents, we would not depend on it nearly so much’ (p. 157). No new evidence is presented, but as with regional magistri the authors draw different conclusions from many previous scholars. Perhaps the most significant contribution is the argument that in attempting to date the Notitia, we should be looking for a terminus ante quem, which for K. and K. is set by the regiment of the Leontoclibanarii, known from 487 but assumed to be founded or renamed by Leo I (457–474) and not recorded in the Notitia (p. 152).

Although discussing problems with dating the eastern chapters of the Notitia, K. and K. say little about the dating of the western chapters, MS transmission or its purpose as a separate or a combined or an illustrated document. The orthodox view of the Notitia is that it is a document of two parts, the western chapters frequently and inconsistently emended and mostly dating from the 420s, the eastern chapters a static document sent to the west c. 395/400. It is a very difficult text; most of the dating issues were outlined by Jones in 1964 in an appendix to Later Roman Empire; C. Zuckerman (Antiquité Tardive 6 [1998]) added some nuances; and more will be apparent when the 2019 Freiburg conference on the Notitia is published. K. and K. deal briskly with the western provenance of the earliest (ninth-century) manuscript evidence, wondering about an eastern origin like the Theodosian Code (p. 156). This creates the difficulty of explaining how the eastern Notitia representing the 440s became associated to a western Notitia conventionally dated to the early 420s (p. 178). Despite much discussion, it remains difficult to know what the Notitia was for, and there are good arguments that what we have is as much an ideological work as a practical one, with a western assemblage from early in the reign of Valentinian III, as argued by P. Brennan in a 1996 Entretiens Fondation Hardt article (pp. 153, 177).

In conclusion, the major problem tackled by K. and K. is whether scholarly orthodoxy has forced the highly inconsistent evidence for late Roman generals’ titles into an anachronistic framework for the late fourth and early fifth centuries. K. and K. are well aware of the flexibility and inconsistency of the late Roman state (pp. 34, 62, 79, 96–7, 158). They also have an excellent command of the evidence and its limits: ‘our ancient sources are simply too limited for historians to expect consistent attestations of all but the most basic features of the ancient world’ (p. 32). Nonetheless, they argue that, unless a general's post is described with a regional suffix matching the Notitia, then this is not proof of the existence of that position. This stimulating monograph identifies many problems regarding the organisation of late Roman field armies, but its arguments on regional commands and thus on dating the eastern Notitia are not compelling.