¹ This paper owes its origin to Professor Arnaldo Momigliano of University College, London. He provoked it by voicing a contrary opinion, and then generously encouraged the revisions which made my disagreement clear. I am also indebted to Professors G. W. Bowersock and W. V. Clausen of Harvard for having read and criticized an earlier draft.

² At the risk of inconveniencing the reader, I will acknowledge my sources for the following arguments at the outset rather than footnote each point separately. To Dessau belong points 6, 7, 9 and 10 (in so far as it treats of the formulas on food, drink and sex). He also pointed out the first three examples of stylization given in argument 11, and noted the Nebenvitae as a peculiarity held in common by the six Scriptores (the first part of my point 13). These contributions are to be found in the second part of his first essay, ‘Über Zeit und Persönlichkeit der S.H.A.’, Hermes 24, 1889, 378–392Google Scholar. cf. also his second paper, ‘Über die S.H.A.’, Hermes 27, 1892, 590–605Google Scholar.

Of the remaining points, several are commonplaces too widely recognized to be named any one man's property. Many or all of the passages on which they are based have been cited to serve various ends by a number of scholars, and especially by the Tendenz-seekers. This is true of points 1, 3, 8 and 5, although for calling attention to the last (on the nomen Antonini) one might perhaps give special credit to Tropaea, G., Rivista di Storia Antica 4, 1899, 233–245Google Scholar. A few of the passages gathered under point 2 have been noted by Manni, E., La Parola del Passato 8, 1953, 73–74Google Scholar. Points 4, 12 and the main part of 13 are, so far as I know, new and mine. But even the borrowed arguments have been enlarged and adjusted to suit my own purposes.

The texts which illustrate each point I have collected independently, including in some cases passages which the less partial reader may judge irrelevant.

³ So that each vita may be promptly connected with its Scriptor, I have initialed them according to a system whereby S = Aelius Spartianus, C = Julius Capitolinus, G = Vulcacius Gallicanus, L = Aelius Lampridius, P = Trebellius Pollio, and V = Flavius Vopiscus. Ascriptions are made as given in the titles and subscriptions of the manuscripts.

⁴ (Spartianus) Hadrian 4, 9; 6, 2; 7, 4; 7, 9; 8 passim; 18, 1; 22, 4; Julianus 4, 1; Severus 7,5; 9,3; (Capitolinus) Pius 6, 5; Marcus 10, 2–9; 11, 2; 25, 6; 26, 13; 29, 4; Pertinax 6, 2; 9, 9; 13, 2; Albinus 13, 5–10; Maximus and Balbinus 13; 17, 2; (Gallicanus) Cassius 8, 7; 12, 4; 12, 9; (Lampridius) Alexander 1, 5; 10, 7; 18, 2; 19. 1–4; 21, 5; 24, 1; 43, 1–2; 46, 5; 49, 2; 52, 2; (Pollio) Valeriani Duo 5, 3–4; Claudius 2, 8; (Vopiscus) Aurelian 13, 4; 21, 6; 26, 8; 40, 1; 49, 7; Tacitus 2, 2; 8, 4–6; 12; 15, 2; 18–19; Probus 6, 3; 11, 1–4; 13, 1; 15, 4; Carus 18, 4. The instances could be multiplied if one were to include corollary passages, where condemned rulers are described as abusing the Senate. I will admit that Dio and Herodian match many of the statements quoted here, but it would have been impossible to define where the HA speaks on its own authority.

⁵ (Spartianus) Hadrian 6, 5; 7, 6; 10, 1; 13, 4; Severus 8, 4, ‘accusatos a provincialibus iudices probatis rebus graviter punivit’; Niger3, 6; 7, 2; 10, 6; (Capitolinus) Pius 4, 10; 6, 1, ‘necumquam ullo laetatus est lucro, quo provincialis oppressus est’; 7, 1; 10, 7; Marcus 17, 1 and 4; 21, 9–10; (Gallicanus) Cassius 4, 2; 14, 8, ‘audisti praefectum praetorii…subito divitem factum. unde, quaeso, nisi de visceribus rei publicae provincialiumque fortunis ?’; (Lampridius) Alexander 15, 3, ‘dicens malum publicum esse imperatorem qui ex visceribus provincialium homines non necessarios nec rei publicae utiles pasceret’; 32, 4; 45, 7; 50, 2; (Pollio) 30 Tyrants 18, 6, ‘videsne ut ille provinciates non gravet ?’; 18, 8; 18, 10; Claudius 12, 4; (Vopiscus) Aurelian 7, 5, ‘de praeda hostis, non de lacrimis provincialium habeant’; 39, 5; Probus 23, 2.

⁶ (Spartianus) Hadrian 21, 2–3; (Capitolinus) Pius6, 4; 7, 7; 11, 1; Marcus 7, 1; Verus 8, 6; 9, 3; Pertinax 13,9; Gordiani Tres 23, 7; 24,1–25,7; 27, 7; 31, 1; (Gallicanus) Cassius 14, 2–8; (Lampridius) Heliogabalus 6, 1–4; 10, 4; 11, 1; 15, 1; Alexander 15; 19, 3; 23, 1–8; 29, 4; 34, 3; 35, 6; 41, 3; 45, 4–5; 65, 4–5; 66–67; (Pollio) 30 Tyrants 10, 10; 18, 11; (Vopiscus) Tacitus 6, 5–7; Aurelian 43; Carus 16, 3–8.

⁷ (Spartianus) Hadrian 21, 9; Severus 4, 1; 19, 10; Niger 3, 1; 12, 6; Caracalla 1, 3–5; Geta 6, 8; (Capitolinus) Marcus 8, 3; 18, 1; 28, 7; Pertinax 13,4; Albinus 3,5; 7, 2–4; 9, 6; 12, 1; 12, 9; 13, 3; Macrinus 5, 3; 10, 2–3; Maximini Duo 4, 4; 8, 8; 11, 1; Gordiani Tres 4. 5; 5. 5; 22, 5; 28, 4; 30, 8; 31, 4–6; Maximus and Balbinus 6, 5; 7, 2; 7, 6; 9, 4–5; 15, 2; (Gallicanus) Cassius 2, 3; 2, 8; 6, 5; 7, 7–8; 8, 7; (Lampridius) Diadumenus 3, 3; 7, 4; Heliogabalus 13, 3; 17, 7; Alexander 4, 5; 48, 5; 50. 3; 59. 1; (Pollio) Gallieni Duo 15, 3; 30 Tyrants 3, 6; 12, 16; 23, 2; Claudius 6, 5; 18, 4; (Vopiscus) Aurelian 21, 8; 50. 5; Probus 8, 1; Quadrigae 5, 3; 10, 3; Carus5, 4; 6, 2.

⁸ Severus 10, 4–6; 19, 3; 20, 3; 21, 11; Niger 12, 6; Caracalla 8, 10; 9, 2; Geta 1, 1–7; 2, 2–5; 3, 5–8; Macrinus 3; 6, 6–7; 7, 5–8; 14, 1–2; Diadumenus 1; 2, 3; 2, 10; 3, 1; 6–7; 9, 4; Heliogabalus 1, 5; 2, 4; 3,1; 9, 2; 18, 1; 33, 8; 34, 6; Alexander 1, 1; 2, 2; 7, 2–6; 8–10, 6; Gordiani Tres 3, 3; 4, 7–8.

⁹ Certain other tendencies seem to develop without close relation to the system of Scriptores. I have in mind the increase in references to dubious inscriptions (Niger [S] 12, 5; Gordiani Tres [C] 34, 2–6; Valeriani Duo [P] 8, 3; Gallieni Duo [P] 19, 4; 30 Tyrants [P] 7, 2; 11, 5; 22, 13; 33, 4–5; Probus [V] 21, 4; Carus [V] 5, 4; 11, 3) and to coinage of questionable authenticity (Diadumenus [L] 2, 6; Heliogabalus [L] 2, 4; Alexander [L] 8, 3; 25, 9; 39, 7; Gallieni Duo [P] 12, 1; 30 Tyrants [P] 26, 2; 31,3; Claudius [P] 14, 3; 17, 7; Aurelian [V] 9, 7; 12, 1; Probus [V] 4, 5; Quadrigae [V] 2, 1; 15, 8). Acclamations also seem to be introduced with increasing frequency after the Commodus.

¹⁰ For a list of the documents, which are too numerous to cite here, one may consult Ch. Lécrivain, Études sur l' Histoire Auguste, Paris 1904, 45–51. For discussion of them, see E. Diehl's RE article ‘Historia Augusta’, col. 2085; Homo, L., Revue Historique 151, 1926, 161–198Google Scholar; 152, 1926, 1–31; and pages 52 to 99 of Lécrivain's book.

¹¹ ‘Biberius Caldius Mero’, Tiberius 42, 1.

¹² (Spartianus) Severus 14, 13; Niger 12, 6; (Capitolinus) Pertinax 1, 1; Macrinus 14, 2; (Gallicanus) Cassius 1, 7; 9, 7; (Lampridius) Commodus 17, 11; Alexander11, 4; (Pollio) 30 Tyrants 10, 4–7; 33,2; Claudius 5,4; (Vopiscus) Tacitus 6, 4; 16, 6; 17, 1; Probus 4, 1; 4, 4; 10, 4; 21,4; Carus 8, 5. A fuller list of puns is given by Dessau, Hermes 1889, 384–5; this collection is supplemented by Hohl, E., Klio 11, 1911, 292Google Scholar, note 3.

¹³ Spartianus discusses the etymology of the name Caesar at Helius 2, 3–5 (a passage modelled on Suetonius, Galba 3, 1) and recites the long roll of distinguished men who left degenerate sons at Severus 21. At Caracalla 7, 3–5 he offers a frivolous account of the male deity Lunus. Capitolinus discourses on the nature of a tacitum consultum at Gordiani Tres 12, and on the origin of the hecatomb at Maximus and Balbinus 11, 4–7, in neither case reliably. Lampridius reviews the original names of all the Antonine emperors at Diadumenus 6, 5–10, and touches upon the history of the city Oresta at Heliogabalus 7, 6–10. Pollio calls attention to his close knowledge of a lexical distinction at Claudius 3, 3, while Vopiscus presents a superficial account of precedents for the interregnum of Tacitus 1.

¹⁴ (Spartianus) Severus 2, 7: origin of the custom by which legates ride in carriages; (Capitolinus) Marcus 12, 12: on the custom of spreading nets under acrobats; Maximini Duo 28, 9: origin of the saying ‘caliga Maximini’; Maximus and Balbinus 8, 5: why the emperors give gladiatorial shows as they set out for war; (Lampridius) Heliogabalus 19, 4: survivals of certain innovations by Heliogabalus; Alexander 26, 9: the ‘ad Mammam’ at Rome; 41, 4: origin of the custom whereby prefects borrow table-service; (Pollio) Gallieni Duo 20, 4: why soldiers wear sword-belts when they dine with the emperor.

¹⁵ (Spartianus) Hadrian 25, 10; (Capitolinus) Verus 2, 7; Albinus 11, 8; Macrinus 11, 3; 11, 7; 14, 3–4; Gordiani Tres 20, 6; (Lampridius) Diadumenus 7, 4; Alexander 3, 4; (Pollio) 30 Tyrants 11, 6; (Vopiscus) Aurelian 6, 6; Carus 11, 1.

¹⁶ Suetonius does have one instance in which a Greek verse is translated into Latin, but as a look at the text (Tiberius 53, 1) will show, it is not really parallel. And paraphrases of a Greek original, though not necessarily of Greek verse, are found at Julius 81, 1; Nero 20, 1; 20, 2; 33, 1; 45, 2. The translations are not in the form of Latin verse, as in the HA.

¹⁷ Niger [S] 8, 6; 12, 6; Macrinus [C] 11, 4; 14, 2.

¹⁸ Dessau had guessed that the Scriptores had done no less than devise these lines for their own criticism, and then presented them as Latin translations because they doubted their competence to make Greek verses (Hermes 1889, 383–4 and note 1). E. Klebs accepted his explanation, and found the procedure such a disturbing point in favour of unity that he felt obliged to deny the HA a patent for it: ‘ein findiger Kopf unter den Vorläufern unserer Biographen kam auf den Einfall seine eigenen Machwerke als Übersetzungen besserer griechischer Originale auszugeben, um auf diese Weise die Kritik von vornherein zu entwaffnen’, Rh. Mus. 47, 1892, 23. Not that there is any evidence to support his interpretation.

¹⁹ cf. Julius 53, 1, ‘parcissimum vini’; Augustus 76, 1, ‘cibi minimi erat atque vulgaris fere’; 77, 1, ‘vini quoque natura parcissimus’; Claudius 33, 1, ‘cibi vinique…appetentissimus’; 33,2, ‘libidinis in feminas profusissimae, marum omnino expers’; Galba 22, ‘cibi plurimi…libidinis in mares pronior’; Vitellius 13, 3, ‘homo non profundae modo sed intempestivae quoque ac sordidae gulae’; and Domitian 22, ‘libidinis nimiae’.

²⁰ (Spartianus) Niger 6, 6, ‘vini avidus, cibi parcus, rei veneriae nisi ad creandos liberos prorsusignarus’; Severus 19, 8; Caracalla 9, 3; Geta 4, 1; (Capitolinus) Maximus and Balbinus 6, 1, ‘fuit cibi avidus, vini parcissimus, ad rem veneriam rarus’; 7, 6; Albinus 11, 4; 13, 1; Macrinus 13, 4; Maximini Duo 28, 2; Gordiani Tres 6, 6; 19, 1–3; (Gallicanus) Cassius 3,4, ‘avidus vini item abstinens, cibi adpetens, et inediae patiens, Veneris cupidus et castitatis amator’; (Lampridius)Alexander 37, 11; 39, 2; (Pollio) Claudius 13, 5, ‘unica castimonia, vini parcus, ad cibum promptus’; (Vopiscus) Aurelian 6, 1, quoted above. These were among the many passages which Mommsen was prepared to surrender to the domain of his diaskeuasts (Hermes 25, 1890, 278 = Gesammelte Schriften VII, 348). But on this point at least we can insist that his position is untenable. Comment upon a man's appetites constituted a typical chapter in the Suetonian biography, on whose model our vitae were obviously conceived from the very first.

²¹ (Spartianus) Geta 1, 1; (Capitolinus) Albinus 1, 5; Gordiani Tres 3, 1; Maximus and Balbinus 4, 5; (Lampridius) Diadumenus 7, 2; Alexander 6, 1; 29, 1; 45, 1. F. Leo (Die griechisch-römische Biographie nach ihrer litterarischen Form, Leipzig 1901, 274) had noticed this borrowing.

²² cf. Peter, H., Die S.H.A.: Sechs litterargeschichtliche Untersuchungen, Leipzig 1892, 238Google Scholar.

²³ Cassius 6, 7 and 9, 5; Quadrigae 1,1. The last is the key text on the absence of Nebenvitae before the HA.

²⁴ for example, Helius [S] 1, 3; Cassius [G] 3, 1; Niger [S] 1, 1–2; 9, 1–4; Geta [S] 1,2; Macrinus [C] 1, 1; Quadrigae [V] 1, 1; 15, 9.

²⁵ The HA also casts out a few pieces of evidence implying that the Nebenvitae were not an innovation. At Cassius 5, 1 Gallicanus invokes the authority of Aemilius Parthenianus, ‘qui adfectatores tyrannidis iam inde a veteribus historiae tradidit’. But Parthenianus is otherwise unknown; the context—two chapters of anecdotes illustrating Cassius' severity—warns against trusting the HA; and, finally, if such a book really existed, it would have been cited for badly-needed information about other pretenders; in any case it would have been named at Quadrigae 1, 1. Let us ignore Parthenianus and consider Junius Cordus, who is said to have written lives ‘eorum imperatorum quos obscuriores videbat’ (Macrinus 1, 3). The possibility that Cordus existed has been regarded with distaste since Mommsen (Hermes 1890, 271–2 = Ges. Schr. VII, 342), but apart from that, since it is the life of Macrinus for which Cordus is being cited, and since Macrinus is a canonical emperor, we are not strictly obliged to believe that Cordus wrote lives of Caesares and tyranni. A third stumbling block awaits in the Helius, where the authority for a group of conventional anecdotes is foisted upon certain unnamed writers who have composed biographies of Aelius— ‘huius voluptates ab iis qui vitam eius scripserunt multae feruntur’ (Helius 5, 3). Here I can say only that I disbelieve Spartianus. A last objection might arise out of Capitolinus' preface to the Gordiani Tres: fuerat quidem consilium, venerabilis Auguste, ut singulos quosque imperatores exemplo multorum libris singulis ad tuam Clementiam destinarem. nam id multos fecisse vel ipse videram vel lectione conceperam.

But since Spartianus' Helius, Niger and Geta; Gallicanus' Cassius; Lampridius' Diadumenus; and Capitolinus' own Verus and Albinus precede this statement, it is quite possible that the remark aims at none other than the Scriptores Historiae Augustae.

Except for these four unreliable statements, all evidence about the treatment of Caesares and tyranni points clearly to an innovation on the part of the HA.

²⁶ For the sake of example, I give one such vita from each: the Helius by Spartianus, the Albinus by Capitolinus, the Cassius by Gallicanus, the Diadumenus by Lampridius, the 30 Tyrants by Pollio, and the Quadrigae by Vopiscus. Of course there are many others.

²⁷ That is to say, according to the manuscripts Capitolinus wrote the Verus; however, Spartianus implies at Helius 2, 9 that he was going to write it.

²⁸ For a contrary, but less comprehensive view, see S. Frankfurter, Eranos 1893, 228–232. According to Hyart, Ch. (Latomus 21, 1962, 675Google Scholar), H. Szeleste also decides that the various prefaces of the HA do not point to multiple authorship, as Frankfurter thought. In her discussion (Meander 16, 1961, 535–544Google Scholar) she expresses the view that ‘les introductions et les épilogues que l'on trouve dans l'Historia Augusta ne permettent pas de conclure à la pluralité des auteurs’. Her article is in Polish; I have not read it.

²⁹ This was the argument which Dessau proposed again and again to the traditionalists (Hermes 1889, 385–6; 1892, 591); I have never seen a fair reply to it. The next three arguments are, I think, mine, at least as they are presented here.

³⁰ Spartianus at Helius 1, 1, Gallicanus at Cassius3 3, Lampridius at Heliogabalus 35, 1–2 and Alexander 64, 2, and Capitolinus at Maximini Duo 1,3.

³¹ Hermes 1889, 392.

³² Hohl put forward this explanation for the employment of several names, Neue Jahrbücher für das klassische Altertum, Geschichte und deutsche Literatur 33, 1914, 703–5Google Scholar.

³³ ‘Die S.H.A.’, Rh. Mus. 47, 1892, 1–52Google Scholar; 5155–549. All page-numbers in my text refer to this article.

³⁴ E. von Wölfflin (Sitzungsberichte der königl. bayer. Akad. der Wissenschaften, philosophisch philologische Klasse, 1891, 465–538) is sometimes mentioned with Klebs for his contributions to the linguistic defence. But for three reasons he need not be considered here. In so far as he discussed distinctions between Scriptores, he leaned most heavily on their statements about themselves. But after Dessau had contested the honesty of these witnesses, it was not a rebuttal to stress their claim to be quite distinct individuals.

Secondly, Wölfflin limited himself for the most part to a study of the work done by Pollio and Vopiscus. He did not attempt to demarcate the work of Spartianus and Capitolinus, and had nothing to say about Lampridius and Gallicanus. Therefore he could hardly begin to establish a six-part system of peculiar traits.

Finally, he belongs with Mommsen's school rather than beside Klebs, in that he insisted on extensive contamination of many biographies by the editor of the collection. It was, in fact, only to serve this thesis that he availed himself of linguistic arguments at all.

³⁵ One might begin with the fourth chapter (‘Parallelstellen zu Angaben anderer Viten des Corpus’) of K. Hönn's study Quellenuntersuchungen zu den Viten des Heliogabalus und des Severus Alexander im Corpus der S.H.A. (Berlin and Leipzig, 1911, 107–169Google Scholar), which catalogues some of the parallels between Lampridius and other Scriptores.

³⁶ Hermes 1892, 601–603. It is gratifying to note that the conservative Peter reached somewhat the same conclusion about variations in approach by comparing the series Maximini Duo, Gordiani Tres, and Maximus and Balbinus with Herodian. On page 88, Sechs Untersuchungen, he writes: ‘es fällt eben leichter nach einer fremden Sprache die Hauptgedanken frei wiederzugeben als genau zu übersetzen, zumal wenn der Inhalt stark gekürzt werden soll. Gemeinsam ist daher dieser Behandlungeiner griechischen Vorlage und dem wörtlichen Abschreiben einer lateinischen nur die Bequemlichkeit’.

³⁷ For example, Mommsen (Hermes 1890, 251 = Ges. Schr. VII, 323), Baynes, N. (The Historia Augusta: its Date and Purpose, Oxford, 1926, 75Google Scholar) and, most recently, R. Syme (‘The Bogus Names in the Historia Augusta’, Bonner HA-Colloquium, 1964–65, Antiquitas, Reihe 4, 3, 270).

³⁸ Syme compares this ‘child of Salonae’ with the Saloninus whom Servius asserts to be the object of Vergil's Fourth Eclogue, and rejects both as inventions (‘Bogus Names’, Banner HA Colloquium 1964–65, 262).

³⁹ op. cit. 113, and note 1; 137–9; and especially 146, note 2. Compare also Leo, Gr.-röm. Biographie, 296 and 298.

⁴⁰ This is largely the fault of Mommsen, who introduced the theory as a logical stop-gap in his argument (Hermes 1890, 228–292 = Ges. Schr. VII, 302–362).

⁴¹ Rh. Mus. 1892, 518, note 2; 529.

⁴² Hermes 1890, 279–280 = Ges. Schr. VII, 350.

⁴³ Leo, op. cit. 283.

⁴⁴ Mommsen, Hermes 1890, 243 ff. = Ges. Schr. VII, 316 fF.; Wölfflin, Sitzungsber. d. königl. bayer. Akad. d. Wiss., philos.-philol. Kl.. 1891, 516–8; W. Soltau, Philologus, n.F. 28, 1917, 397–8. Wölfflin later decided that Capitolinus must have been the editor after all (Literarisches Centralblatt 1893, cols. 120–1).

⁴⁵ Peter, Sechs Untersuchungen 25, note 1; Mommsen, op. cit. 243 ff.; Wölfflin, op. cit. 516; Soltau, op. cit. 397–8.

⁴⁶ Homo, L., Revue Historique 132, 1919, 37–8Google Scholar (summary).

⁴⁷ op. cit. 496; 516.

⁴⁸ op. cit. 142; 146, note 4.

⁴⁹ H. Zernial's recent study of clausulae (Über den Satzschluss in der H.A., Deutsche Akademie der Wissenschaften zu Berlin: Schriften der Sektion für Altertumswissenschaft, 1956) should not be brought into a discussion of authorship. Zernial took uniform authorship for granted from the start, as he says himself, and interested himself instead in the preponderance of accentual over quantitative clausulae. Furthermore, since he did not propose to analyse all clausulae, but only samples from the more rhetorical passages, it might be felt that his results were biased toward the second part of the HA. See also his reviewer G. Bendz (Gnomon 1960, 247–251).