Hostname: page-component-cd9895bd7-lnqnp Total loading time: 0 Render date: 2024-12-25T08:18:59.804Z Has data issue: false hasContentIssue false

Inconsistency and Lassitude: the Shield Emblems of the Notitia Dignitatum

Published online by Cambridge University Press:  24 September 2012

Robert Grigg
Affiliation:
University of California, Davis

Extract

It has usually been held that the shield emblems in the Notitia Dignitatum (Not. Dig.) were based upon an official pictorial register or pattern book, containing the unit emblems of the late Roman army. Thought to have been based upon an official source, as the text was, the shield emblems of the Not. Dig. are imagined to have been accurate in the original manuscript. It was only later, according to this view, that errors crept in during the transmission of the text and illustrations, so that the emblems now appear to be somewhat debased. For example, it is held that some of them no longer accompany the titles for which they were apparently intended.

Shifts in the relationship between the emblems and titles have long been noted. But there are other, more fundamental, inconsistencies that have escaped the attention of scholars. These previously led me to raise doubts about the truth of the conventional view above and to entertain the possibility ‘that the artist's sources were so impoverished that he was reduced to relying upon his own powers of invention’. I should now like to explain in greater detail my reasons for rejecting the conventional view and advancing the alternative explanation that the shield emblems of the Not. Dig. were largely ad hoc fabrications. The consequences for our understanding of the Not. Dig. and of the art of the later Roman Empire are obviously considerable.

Type
Research Article
Copyright
Copyright © Robert Grigg 1983. Exclusive Licence to Publish: The Society for the Promotion of Roman Studies

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Berger, P., The Insignia of the Notitia Dignitatum (1981), 4457Google Scholar.

2 With some authors this is merely implicit: e.g., Bury, J. B., JRS X (1920), 132Google Scholar, and Byvanck, A. W., Mnemosyne Ser. 3, VIII (1940), 195Google Scholar. It is an implicit premise in Seeck's discussion of the shield emblems: see Seeck, O., Hermes IX (1875), 232 f.Google Scholar, and id., Notitia Dignitatum (1876), XX f., as well as Hoffmann, D., Das spätrömische Bewegungsheer und die Notitia Dignitatum (1969) I, 7 f., 14, 163Google Scholar, who claims that the shield emblems have been debased, yet bases important conclusions upon them. Berger, , Insignia, 4457Google Scholar, has expressly affirmed this belief and set forth a series of arguments designed to confirm it. I shall examine those arguments at the appropriate places below. Berger, , Insignia, 161a, 168Google Scholar, also speculates that the register containing the model emblems may have been an ‘illustrated scroll’.

Authors researching other topics have used the shield emblems in ways that presuppose their accuracy: see nn. 14, 15, and 17 below, as well as T. Mommsen, CIL III, no. 6194, and Delbrueck, R., Probleme der Lipsanothek in Brescia (1952), 79 fGoogle Scholar.

3 Seeck, , Hermes IX (1875), 232 f.Google Scholar; id., Notitia, XX f.; Hoffmann, , Bewegungsheer I, 14, 163Google Scholar.

4 JRS LXIX (1979), 111.

5 Aside from the particular issues discussed here, there are two broad questions concerning later Roman art on which the results of my study bear. One is the kind and degree of evidential value that images in late Roman art characteristically possess. This question is often raised in connection with attempts to reconstruct the appearance of buildings or monuments on the basis of late Roman images. See, e.g., Duval, N., Cahiers Archéologiques XV (1965), 247–54Google Scholar. The other question is the extent to which late Roman painters made use of copy or pattern books. For the few fragmentary examples that survive, see Scheller, R., Survey of Medieval Model Books (1963), 5 f., 45 f.Google Scholar, and Kitzinger, E., The Place of Book Illumination in Byzantine Art (1975), 109 f., 119Google Scholar.

6 Though clearly a desideratum, no corpus of Roman shields or shield bosses exists. I have had to rely on scattered publications that are much too numerous to list exhaustively here. Some of the most important of these are Gansser-Burckhardt, A., Das Leder und seine Verarbeitung im römischen Legionslager Vindonissa (1942), 74Google Scholar; Toynbee, J. M. C., Art in Britain under the Romans (1964), 299Google Scholar; Hübner, E., Archaeologisch-epigraphische Mittheilungen aus Oesterreich-Ungarn II (1878), 105–19Google Scholar; and Klumbach, H., Bayerische Vorgeschichtsblätter XXV (1960), 125 ffGoogle Scholar.

7 Unfortunately there is no comprehensive publication for the shields discovered at Dura-Europos. See Cumont, F., Fouilles de Doura-Europos (1922–1923) (1926), 261–3, 323–37Google Scholar; id., Syria VI (1925), 1–15; The Excavations at Dura-Europos, Preliminary Reports (hereafter Reports), First Season, Spring 1928, ed. Baur, P. and Rostovtzeff, M. (1929), 1618Google Scholar; Reports, Second Season, 1928–1929, ed. Baur, P. and Rostovtzeff, M. (1931), 75, pl. xxviGoogle Scholar; Reports, Seventh and Eighth Seasons, 1933–1934, ed. Rostovtzeff, M., Brown, F, and Welles, C. (1939), 326–69Google Scholar.

8 Reports, Sixth Season, 1932–1933, ed. Rostovtzeff, M., Bellinger, A., Hopkins, C., and Welles, C. (1936), 456–66Google Scholar, pls. xxv, xxvi.

9 Thanks to a variety of sources, not least of which are the legionary coins minted under Gallienus, Victorinus and Carausius. For the legionary series on the coins of these emperors, see Oman, C., Numismatic Chronicle, 4th Series, XVIII (1918), 8096Google Scholar; id., Numismatic Chronicle, 5th Series, IV (1924), 53–68; and Webb, P., The Roman Imperial Coinage v, pt. 1, 34, 92–7Google Scholar; v, pt. 2, 384 f., 388 f, and 468–70. More generally, see von Domaszewski, A., Abhandlungen zur römischen Religion (1909), 3 ff.Google Scholar, Renel, C., Cultes militaires de Rome: Les ensignes (1903), 73 ff.Google Scholar, and E. Ritterling, RE XII, 1371–6.

10 A good bibliography for all aspects of the Column of Trajan is found in Rossi, L., Trajan's Column and the Dacian Wars, trans, rev. Toynbee, J. M. C. (1971), 231–3Google Scholar. One may now add to it Florescu, F., Die Trajanssäule, I: Grundfrage und Tafeln (1969)Google Scholar, and Gauer, W., Untersuchungen zur Trajanssäule, I: Darstellungsprogramm und künstlerischer Entwurf (Monumenta Artis Romanae, 13) (1977)Google Scholar. I base my statement about the absence of legionary badges from the legionary scuta both on Rossi's discussion (Column, 108 ff.) and my own examination of the plates in Hartleben, K. Lehmann, Die Trajanssäule (1926)CrossRefGoogle Scholar.

11 The titles and stations of some of the units in the Comitatus and Pseudocomitatus in the Not. Dig. have helped to link them to former legions whose badges are known. For this, see Hoffmann, , Bewegungsheer I, 188, 227 ffGoogle Scholar. Sources concerning these badges are given above in n. 9 and in Parker, H. M. D., The Roman Legions, 2nd ed. (1958), 106, 116Google Scholar, n. 2, 262 f., 269 f. Below I list the ten units in the Not. Dig. falling into this category, giving (a) the title in the Not. Dig., (b) legionary title, (c) emblem in the Not. Dig., and (d) legionary badge:

(1) (a) Primani, (b) Leg. I Italica, (c) knotted rope (?), Not. Or. vi, 5, (d) boar or hippocamp.

(2) (a) Quinta Macedonica, (b) Leg. V Macedonica, (c) radial thunderbolts (?), Not. Or. vii, 4, (d) eagle.

(3) (a) Decima Gemina, (b) Leg. X Gemina, (c) radial thunderbolts (?), Not. Or. vii, 7, (d) bull.

(4) (a) Tertiodecimani, (b) Leg. XIII Gemina, (c) rampant quadruped, Not. Or. viii, 6, (d) lion.

(5) (a) Quartodecimani, (b) Leg. XIV Gemina, (c) eagle, Not. Or. viii, 7, (d) Capricorn.

(6) (a) Secundani, (b) Leg. II Adiutrix, (c) essentially blank, Not. Or. ix, 13, (d) Pegasus.

(7) (a) Minervia, (b) Leg. I Minerva, (c) essentially blank, Not. Or. ix, 15, (d) Minerva or ram.

(8) (a) Octaviani, (b) Leg. VIII Augusta, (c) four radially arranged peltae, Not. Occ. v, 10, (d) bull.

(9) (a) Secundani Italiciani, (b) Leg. II Italica, (c) wheel cross, Not. Occ. v, 86, (d) she-wolf and twins.

(10) (a) Tertiani, (b) Leg. HI Italica, (c) blank, Not. Occ. v, 88, (d) stork.

Among these ten comparisons, there is only one instance of a possible correspondence: the shield emblem of the Tertiodecimani (no. 4) shows a rampant quadruped, which conceivably was derived from the lion of Leg. XIII Gemina.

Ignoring this evidence, Berger, , Insignia, 44 f.Google Scholar, assumes without argument that the shield emblems and legionary badges were identical. Thus she believes that one can confirm the accuracy of the shield emblems in the Not. Dig. by comparing them with the legionary badges used under the early and middle Empire. Although aware of the failure of one of the shield emblems to match the appropriate legionary badge (Insignia, 228, n. 54), Berger, (Insignia, 48)Google Scholar claims to have found a confirming instance in the tombstone of Valerius Thiumpus. Berger's argument is borrowed from Mommsen (see n. 2 above), but she has overlooked Hoffmann's critical assessment of it—Bewegungsheer II, 88, n. 189 (‘fraglich’).

12 Couissin, P., Les Armes romaines (1926), 393 f.Google Scholar, and the editors of Reports, Seventh and Eighth Seasons, 330, argue that the rectangular scutum was gradually abandoned in the course of the second century A.D.

13 For these two columns, see Kollwitz, J., Öströmische Plastik der theodosianischen Zeit (1941), 3 ff.Google Scholar; Giglioli, G., La colonna di Arcadio a Costantinopoli (1952)Google Scholar; Becatti, G., La colonna coclide istoriata (1960), 83 ffGoogle Scholar.

14 In the reliefs decorating the arch, which was erected c. 300, some of the shields carried by Roman soldiers bear images either of Hercules; a rampant lion, rendered in profile; or an eagle holding a thunderbolt in its talons. Kinch, K. F., L'arc de triomphe de Salonique (1890), 16 f., 19 f.Google Scholar, thought that the eagle and the lion referred to Legio V Macedonica and Legio XIII Gemina respectively; he also thought that some of the shields referred to the Iovii and Herculii, units probably created during the Tetrarchy; as he noted, in the Not. Or. v, 3, 4, and Not. Occ. v, 2, 3, their shields are decorated with eagles. Kinch evidently regarded the shield emblems as reliable evidence, notwithstanding the fact that one might have expected the emblem of the Herculii to be either Hercules or one of his well-known attributes. Alföldi, , Germania XIX (1935), 324Google Scholar, n. 6, modifies Kinch's suggestion by regarding the eagle as the emblem of the Iovii and the image of Hercules as the emblem of the Herculii. The lion he connects with a Germanic unit bearing the name Leones; this unit, formed under Caracalla, Alföldi supposes was not merely a contingent of Germanic bodyguards, but a regular military unit that likely survived into the fourth century (ibid., 324 f.).

15 Germania XIX (1935), 324 ff.Google Scholar, and Dumbarton Oaks Papers XIII (1959), 171 ffGoogle Scholar.

16 Hoffmann, , Bewegungsheer II, 63Google Scholar, n. 431, 82, n. 63, and Laubscher, H. P., Der Reliefschmuck des Galeriusbogens in Thessaloniki (1975), 16 f., 47 fGoogle Scholar.

17 See, e.g., L'Orange, H. P. and von Gerkan, A., Der spätantike Bildschmuck des Konstantinsbogens (1939) 63 ff.Google Scholar; E. Polaschek, RE XVIII, 1109; Ross, M., Dumbarton Oaks Papers XIII (1959), 181Google Scholar; MacMullen, R., Art Bulletin XLVI (1964), 442Google Scholar; Hoffmann, , Bewegungsheer I, 133 ff.Google Scholar; and Berger, , Insignia, 45 ffGoogle Scholar.

18 Alföldi, , Germania XIX (1935), 326Google Scholar. See Not. Or. vi, 9, and Not. Occ. v, 14, for the emblem of the Cornuti—and possibly Not. Occ. v, 25, if one accepts the view (see Seeck, , Notitia, XXIVGoogle Scholar) that this was the shield originally intended for the preceding titulus. Interestingly, the Cornuti Seniores, Cornuti Iuniores emblems in Not. Occ. vi, 6, 7, are completely different from these. And none of the Cornuti shields featuring the peltate form also features a Victory, as is seen on the shield on the pedestal of the Arch of Constantine. That any of these peltate forms were intended to have goats' heads as terminals is doubtful, although a fur pattern is evident in Not. Or. vi, 9. Not. Occ. v, 15, lacks zoomorphic terminals altogether.

Anxious to defend the proposition that the shield emblems are accurate, Berger, Insignia, 45 f., identifies the Victory accompanying Not. Occ. v, 24, as the correct emblem of the Cornuti also. Berger has evidently overlooked Seeck's suggestion (Notitia, XXIV) that the emblem originally intended for the Cornuti now accompanies Not. Occ. v, 25.

19 Grigg, , JRS LXIX (1979), III fGoogle Scholar. E.g., many of the titles, by virtue of their reference to things after which the units were named, suggest images that the artist would likely have conceived as appropriate. Those units with the titles Ioviani (or Iovii), Herculiani, Martii, Victores, Leones, Cornuti, Lanciarii, Balistarii, Minervii, Sagittarii, Armigeri, and Scutarii might all have been given plausible looking emblems solely on the basis of the underlying reference of the title (as distinct from the use of these terms to refer to military units). For this reason, it is a mistake to assume, as does Berger, , Insignia, 48 f.Google Scholar, that a plausible match between a unit's title and its shield emblem authenticates the accuracy of the emblem.

20 JRS LXIX (1979), 111.

21 The comparison is between the second page of Not. Or. v and the second page of Not. Occ. vi.

22 Not. Or. v, 17.

23 Not. Occ. vi, 21, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 38, 39. 40.

24 I list here the essentially blank shields by chapter, giving the ratio of their number to the total number of shields in the chapter: Not. Or. v, 17 (1/24); Not. Or. vi, 17 (1/22) (shields vi, 24, 25, cannot be classified; shield vi, 12, which is blank in Seeck's edition, is blank in neither the Oxford nor the Paris MS of the Not. Dig.); Not. Or. vii, 9, 10 (2/21); Not. Or. viii, 12, 20 (2/21); Not. Or. ix, 7, 12, 13, 15, 16 (5/15); Not. Occ. v, 4, 5, 9, 11, 12, 18, 23, 26, 44, 47, 48, 52, 54, 55, 56, 57, 62, 66, 67, 70, 71, 72, 73, 76, 77, 84, 85, 87, 88, 89, 99, 101, 104, 105, 106, 108, 111, 115, 119, 120, 123 (41/123); Not. Occ. vi, 2, 3, 5, 6, 7, 20, 21, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 37, 38, 39, 40 (21/39).

25 JRS LXIX (1979), 111, fig. 2.

26 If I have applied my standard uniformly, then the following distribution of awkward emblems is significant: Not. Or. v, 7, 12, 20, vi, 2, 11, 13, 20, 21, vii, II, 12, viii, 5, 11, 21, ix, 14; Not. Occ. v, 27, 28, 32, 34, 37, 41, 42, 43. 63, vi, 13, 15. The disparity in proportions that emerges when one takes the number of shields in each chapter into consideration is so great as to rule out coincidence as a cause at a very high level of confidence (employing the Chi-square distribution, it is over the 99 per cent level).

27 Not. Occ. v, 6, 7, 8, for the wheel cross; Not. Occ. v, 14, 15, 16, 17, for the shaft and crescent.

28 Not. Occ. v, 27, 28, 32, 34, 37, 41, 42, 43.

29 There is wide agreement that the two halves of the Not. Dig. were composed at different times. The estimated dates given by Jones, A. H. M., The Later Roman Empire 284–602 (1964) 11, 1417Google Scholar, are fairly standard: ‘the Eastern section was revised … at a date not long after 395 …’; on p. 1423 he writes, ‘the military lists [in the western section] have then been revised after a fashion down to 420, perhaps to the end of Honorius's reign. The date is unlikely to be much later, since the units named after Valentinian III are so few’. Hoffmann, , Bewegungsheer I, 52 f.Google Scholar, provides a terminus ante quem of May 394 for the eastern military lists, Not. Or. v, vi, vii, viii, and xi; Not. Or. ix, however, he dates from before 410. The western lists, according to Hoffmann, (Bewegungsheer I, 22Google Scholar) must stem from the reign of Honorius (395–423) and were composed directly after his death. Ward, J., Latomus XXX (1974), 434Google Scholar, gives the date of c. 394 for the eastern half and c. 430 for the western half.

30 Hoffmann, , Bewegungsheer I, 26Google Scholar (for a list of those units), 28, 526.

31 Not all of the units in Hoffmann's list are represented by an emblem.

32 According to Berger, , Insignia, 56 f.Google Scholar, ‘another argument for the authenticity of the shield devices likewise derives from cross-checking the internal evidence [the reader should understand that Berger is here appealing to the criterion of internal consistency]: various divisions of the same military unit display similar elements on their shields even if those devices are found on different pages or in different parts of the Notitia’. Berger cites only the correspondence between the Primi Theodosiani (Or. v, 23) and the Secundi Theodosiani (Or. vi, 23) to support her conclusion.

33 See Hoffmann, , Bewegungsheer I, 11 f.Google Scholar, who distinguishes between units that are tactically related and those related by means of similar titles that are combined with terms of distinction, such as seniores and iuniores. For units bearing these particular terms of distinction, see Tomlin, R., American Journal of Philology XLIII (1972), 253–78CrossRefGoogle Scholar. Tomlin, loc. cit., surveys the varied sources of regimental titles in the Roman army. Hoffmann and Tomlin both deal with the creation of new units by the process of division and detachment from older units. On this also see Jones, , Later Roman Empire I, 680 ffGoogle Scholar. General discussions of the naming of Roman legions are found in Parker, , Legions, 261Google Scholar, and Webster, G., The Roman Imperial Army of the First and Second Centuries A.D., 2nd ed. (1979), 109Google Scholar.

34 Hoffmann and Tomlin in n. 33 above explain the designation seniores-iuniores as evidence, in normal circumstances, of a genetic relationship between the units bearing these titles.

35 e.g., Prima Armeniaca-Secunda Armeniaca (Not. Or. vii, 13, 14) and Lanciari Seniores-Lanciari Juniores (Not. Or. v, 2, ix, 16).

36 Seeck, , Notitia, xxGoogle Scholar; Hoffmann, , Bewegungsheer I, 14, 163Google Scholar. Both Seeck and Hoffmann propose to use the present distribution of the shield emblems as evidence that in some places they have been shifted out of their proper sequences. They also use them to infer links between several units, for which see especially Hoffmann, , Bewegungsheer II, 1Google Scholar, n. 34, 2, n. 67.

37 I count a total of ten nominally related units in Not. Or. viii and ix, not one of which has its emblem co-ordinated with its nominally related sister unit: Not. Or. viii, 4, 5, 6, 7, 8, 9, 15; Not. Or. ix, 3, 4, 16.

38 Hoffmann, , Bewegungsheer I, 13 fGoogle Scholar.

39 ibid., 13.

40 Here and in the Appendix, I do not classify tituli 8 and 21, Balistarii Seniores and Balistarii Theodosiaci, as nominally related units. The unit properly related to the Balistarii Seniores is the Balistarii Iuniores, Not. Or. viii, 15.