Answer by Aldo Corcella
To answer this question we have to delve into some technicalities of ancient book production and sift through patchy documentation that requires a certain amount of educated guesswork. First of all, we should keep in mind that the division into nine books can hardly be traced back to the author: when Herodotus refers to sections of his work, he uses the term logoi (‘tales, discourses’), intended as content-based units and briefer than the books of our tradition. Some have attempted to single out 28 original logoi within the Histories, but the results are not fully convincing. In any case, the general consensus is that the division into nine books was introduced at a later time in an authoritative edition of the work. The extension of the books corresponds to the standard length of bookrolls after the end of the 4th century BCE, probably based on editorial practices consolidated within the context of the Library of Alexandria.
The first explicit testimony of this division into nine books is found in Diodorus Siculus (1st century BCE), who says: ‘of the historians, Herodotus, beginning with the period prior to the Trojan War, has written in nine books a general history of practically all the events which occurred in the inhabited world’ (The Library of History 11.37.6, transl. Oldfather). Before Diodorus, however, some other sources apparently point to this same division. Especially important is a fragment from the commentary on Herodotus by the grammarian Aristarchus of Samothrace, director of the Alexandrian library in the middle of the 2nd century BCE, which reads: ‘Aristarchus’ Commentary to Herodotus’ Book 1’. The last words commented on are from chapter 215 of our editions, which suggests that Aristarchus worked on a Book 1 which had more or less the same length as ours.
We do not know whether the remaining division of books in the Alexandrine edition of Herodotus corresponded to that of the medieval manuscripts which later informed modern editions of the text. In the following centuries, however, passages from Herodotus’ work were usually cited with the indication of book numbers according to a division which corresponds exactly to the one we are familiar with. The oldest of such citations is in the so-called Lindos Chronicle (an inscription dated to 99 BCE), while others are in Strabo, Theo, Flavius Josephus (1st c. BCE – 1st c. CE) and the usage remains consistent up to Late Antiquity. A notable exception is Pausanias, who prefers to introduce passages from the Histories by reproducing Herodotus’ way of denoting his own text through such formulas as ‘in the logos referring to Croesus’.
Since the Hellenistic age, then, nine were Herodotus’ books, just as nine were the Muses. Yet the first testimonies that Herodotus’ books were named after the Muses date to the 2nd century CE. In his treatise How to write history (166 CE), Lucian of Samosata says that ‘Herodotus’ work was so highly admired that his books were called Muses’ (chapter 42). He repeats this idea in another work of uncertain date, Herodotus, or Aetion: ‘Herodotus enchanted the public to such an extent that his books were called Muses, also being nine’. A hypercritical scholar could observe that Lucian says the books were called Muses, but he does not properly attest that each book was called after a particular Muse – he could, in other words, just mean that the work as a whole was called Muses. The same could be said about the testimony of the Byzantine lexicon Suda (10th century), likely derived from the Biographical Dictionary of Learned Men of Hesychius of Miletus (6th century): ‘His books are entitled Muses’. Admittedly, single books are never named after the Muses when identifying a quote from the Histories in antiquity: to my knowledge, formulas like ‘Herodotus in Thalia’ are attested only in late Byzantine times and again in the Renaissance. However, it is very likely that both Lucian and Hesychius were acquainted with editions where each book was named after a Muse. This is all the more probable in the case of Photius: in the 9th century, he described the edition he read as ‘Herodotus’ nine books of Histories, according to the number and name of the nine Muses’ (The Library 60, 19b16-7).
In any case, the medieval manuscripts do bear the name of a Muse at the head of each book: for instance, in the ms. Laur. plut. 70.3 (early 10th century) we read ‘Clio 1’ at the beginning of book 1, and each of the following books is introduced by the name of a Muse following the order of Hesiod’s canonical list in Theogony 77-79: Clio, Euterpe, Thalia, Melpomene, Terpsichore, Erato, Polymnia, Urania, Calliope; and more or less the same happens in the ms. Vat. gr, 2369 (second half of the 10th century). Paratextual traces in the manuscripts just mentioned suggest that the names of the Muses were derived by medieval copists from bookrolls. This in turn implies that each book occupied a different roll and was named after the Muses already before the 4th century CE, when the normal form for books became the codex.
An epigram in the Palatine Anthology seems indeed to evoke a precise one-to-one correspondence between Herodotus’ books and the Muses: ‘Herodotus entertained the Muses, and each, in return for his hospitality, gave him a book’ (AP 9.160, transl. Paton). Unfortunately, the author of the epigram is unknown and its date uncertain. It was already known to Choricius of Gaza (6th cent. CE), however, who echoed it in his Apology for the Mimes (chapter 148), while in the same period the poet Christodorus imagined that Herodotus dedicated his work ‘to the nine Muses’ (AP 2.377-81). Perhaps Choricius read the epigram at the beginning of his edition of Herodotus. The distich, in other words, could be an example of the so-called ‘book epigrams’ (brief poems written in the liminal sections of manuscripts describing and praising the works and authors to follow) written either at the beginning of a roll with a copy of Book 1 or, if composed in Late Antiquity, on the opening page of a codex containing the whole Histories. A similar epigram (AP 7.17) describes the poetess Sappho taking a flower from each of the Muses ‘to lay beside my nine flowers of song’ – likely a reference to the Alexandrian edition of Sappho’s poems in nine books. In this case, we know that the epigram was written by Tullius Laurea, one of Cicero’s freedmen, in the 1st century BCE. Because of the similarities between the two epigrams, some scholars think that the epigram on Herodotus dates to the same period, but the hypothesis is not entirely convincing. In any case, the naming of a work consisting of nine books after the nine Muses in that period would be nothing unusual: a prose work entitled ‘9 books of Muses’ by a grammarian called Aurelius Opillus is to be dated between the 2nd and the 1st centuries BCE (Suetonius, On grammarians 4.7), while more uncertain is the dating of the ‘nine books named after the Muses’ by the rhetorician Bion (Diogenes Laertius 4.58) and of the naming of Aeschines’ letters as ‘Muses’ due to their number (Photius, The Library 61, 20a5-9); in Hadrian’s time a rhetorician called Cephalion composed an historical work divided into ‘nine books named after the nine Muses: Clio, Thalia, Polymnia, Melpomene, Terpsichore, Euterpe, Calliope, Erato, Urania’ (Photius, The Library 68, 34a4-34: the order of the names, different from that attested for Herodotus, could point to a variation of the model found in Herodotus’ editions).
To conclude, not only was Herodotus’ work called Mousai (‘the Muses’) in antiquity to indicate the work in its entirety; it would also seem that each of the nine books was already then named after a Muse. This practice might have already been introduced, outside of Alexandria, during the 3rd or in the first half of the 2nd century BCE, but hardly before that period (Barry Baldwin’s hypothesis that a contemporary of Herodotus, or Herodotus himself, might have ‘unofficially’ named the Histories after the Muses is unlikely as the division into nine books appears to have been an Hellenistic innovation). Given the current state of the evidence or until proven otherwise, the most sensible answer to the question is the one given by Wolfgang Aly in 1909: ‘The edition with the names of the Muses is probably to be dated between Aristarchus’ Commentary and Cephalion’ – that is, between the middle of the 2nd century BCE and the first decades of the 2nd century CE. What we know for certain is that by 166 CE the naming of Herodotus’ work after the nine Muses, which probably originated in a rhetorical milieu, was widespread enough that Lucian could mention it as a well-known fact without further comment.
Bibliographical references:
W. Aly, ‘Ein Beitrag zur Textgeschichte Herodots’, Rheinisches Museum 64 (1909): 591–600
B. Baldwin, ‘Herodotus and Tacitus: Two Notes on Ancient Book Titles’, Quaderni Urbinati di Cultura Classica N.S. 16 (1984), 31-34
A. Corcella, ‘L’opera storica di Teopompo e le realtà librarie del IV secolo a.C.’, Quaderni di storia 77 (2013): 69–118
A. Corcella, ‘Herodotus and the Textual Tradition’, Syllogos – Herodotus Journal 2 (2023), 86–106
J. Damaggio, ‘Aurelius Opillus’, in Grammatici disiecti, consulted online on 4 June 2024, https://doi.org/10.58079/p44s; first published online: 2016
K. Demoen, ‘Epigrams on Authors and Books as Text and Paratext’, in M. Kanellou, I. Petrovic, and C. Carey (eds.), Greek Epigram from the Hellenistic to the Early Byzantine Era (Oxford 2019), 66–82
C. Higbie, ‘Divide and Edit: A Brief History of Book Divisions’, Harvard Studies in Classical Philology 105 (2010): 1–31
F. Montana and E. Esposito, ‘Herodotus’, in Commentaria et Lexica Graeca in Papyris Reperta (CLGP) 1.2.6: Galenus-Hipponax (Berlin 2019), 17–89
G. Squillace, ‘Kephalion (93)’, in Brill’s New Jacoby; consulted online on 4 June 2024 http://dx.doi.org/10.1163/1873-5363_bnj_a93; first published online: 2016
D. Yatromanolakis, ‘Alexandrian Sappho Revisited’, Harvard Studies in Classical Philology 99 (1999): 179–195
Herodotus Helpline 