Using Homer for Divination: Homeromanteia in Context

By Dr. Raquel Martín-Hernández / 03.28.2014
Postdoctoral Researcher
Universidad Complutense de Madrid


The verses of Homer were often interpreted in Antiquity as the voice of an oracle. But we have no material evidence about the use of Homeric verses as oracles before the third and fourth centuries CE. This paper is focused on the use of Homeric works in divination in Roman times and consists in a comparative study of the three Homeromanteia that have been preserved. A comparison of the content and the layout of these three texts leads us to a series of questions focused on how this kind of lot-divination was transmitted in Antiquity, and on the idea of “sacred” books as direct links to oracular answers. The fact that one of the Homeromanteia appears in a Greek magical handbook allows us to stress the anomaly of this type of  divination, bibliomancy, when compared to all other divinatory recipes preserved in the Greco-Egyptian magical handbooks. In my presentation I will pay special attention to the layout of the Homeromanteion in the magical handbook, in order to provide new answers regarding the uses and the users of this kind of divination. This paper is part of a larger project on the transmission of Greek magical knowledge in books.


§1  Much has been said about the uses and abuses of the Homeric texts; about whom their users were, how they were used, why, when, how, and where they were performed, and who the audience of the Iliad and the Odyssey was in Classical and Roman times, among other questions. From the use of Homer’s works as school-texts for writing and learning to the use of his poems by scholars who tried to figure out their enigmatic and hidden meaning, a wide range of research possibilities appear before our eyes.

§2  Focused in the chronological period between the 2nd and the 5th century CE, I will offer in this article an overview of how the Homeric verses were used for divination purposes through the study of the so-called Homeromanteion, a lot-divination text made up by a total of 216 Homeric verses. I will attend both the particularities of the text and the layout of the books where the three Homeromanteia are preserved. A close look to the information given by a papyrological approach will provide new evidence to add to the current studies of the text in order to get a more complete panorama of this particular literary piece.

§3  Special attention will be paid to the Homeromanteion in P.Lond. 121, better known as PGM VII[1], due to the fact that my research project centers on the study of this particular bookroll. The text will be analysed in light of the other Homeromanteia and the Greek and Demotic divinatory prescriptions preserved in the corpus of magical papyri from Egypt. Through this analysis I will point out that the Homeromanteion in PGM VII is a kind of ‘alien’ among the divination recipes of magical handbooks from Egypt, but it will be explained that its inclusion in this particular book is very well motivated[2].

The Use of Authorities in Divination

Bust of Aristophanes

§4  Before going into detail, let us start by seeing how the verses of respected voices were used for divination. The use of “sacred” books and authoritative texts for providing oracular answers to people when needed has been widely studied[3]. According to S. Augustine and S. Jerome[4], the Bible was used in their time for divination purposes in the same way books by Virgil and Homer were[5]. However, the use of “sacred books” for divination is not only a matter of Late Antiquity or the Middle Ages. The belief that Homeric verses provided oracular responses is already apparent in the fifth century BCE. Aristophanes uses the authority of Homer in such a manner in his comedy Peace[6]:

{ΙΕ.} Ποῖον γὰρ κατὰ χρησμὸν ἐκαύσατε μῆρα θεοῖσιν;
{ΤΡ.} Ὅνπερ κάλλιστον δήπου πεπόηκεν Ὅμηρος·
«Ὣς οἱ μὲν νέφος ἐχθρὸν ἀπωσάμενοι πολέμοιο
Εἰρήνην εἵλοντο καὶ ἱδρύσανθ’ ἱερείῳ.
Αὐτὰρ ἐπεὶ κατὰ μῆρ’ ἐκάη καὶ σπλάγχν’ ἐπάσαντο,
ἔσπενδον δεπάεσσιν, ἐγὼ δ’ ὁδὸν ἡγεμόνευον·»
χρησμολόγῳ δ’ οὐδεὶς ἐδίδου κώθωνα φαεινόν.
{ΙΕ.} Οὐ μετέχω τούτων· οὐ γὰρ ταῦτ’ εἶπε Σίβυλλα.
{ΤΡ.} Ἀλλ’ ὁ σοφός τοι νὴ Δί’ Ὅμηρος δεξιὸν εἶπεν·
«Ἀφρήτωρ, ἀθέμιστος, ἀνέστιός ἐστιν ἐκεῖνος,
ὃς πολέμου ἔραται ἐπιδημίου ὀκρυόεντος.»
{ΙΕ.} Φράζεο δή, μή πώς σε δόλῳ φρένας ἐξαπατήσας
ἰκτῖνος μάρψῃ

Hierocles: Say, what oracle authorized you to burn thighs for the gods?
Trygeas: The very fine one that Homer composed, of course.
“Thus casting away the detestable vapour of warfare,
they opted for Peace and with a victim established her.
And when the thighs were burnt and the innards devoured,
they poured libation from cups, and I led the way.”
but to the oracle monger on one passed a gleaming goblet!
Hierocles: That’s nothing to me; Sibyl did not say it!
Trygeas: But here’s something the sage Homer said that, by god, is well put:
“Clanless, lawless, heartless is that man
who lusts for the horror of warfare among his own people.
Hierocles: Take heed, lest a kite somehow beguile your wits by deception and snatch up[7].

§5  Even if the lines quoted by Aristophanes are not really Homeric verses and they were invented ad hoc for comic purposes[8], we have to assume that the audience could not understand the joke if they were not already familiar with such a practice.

§6  This is not the only case in which Homeric verses are involved in divinatory matters in Classical times. In Plato’s dialogue Crito, Socrates tells Crito a dream. A woman dressed in white addressed him with a Homeric verse[9]:

Ὦ Σώκρατες, ἤματί κεν τριτάτῳ Φθίην ἐρίβωλον ἵκοιο

“Oh Socrates, on the third day you would come to the fertile Phthia.”

§7  Both Crito and Socrates interpret without any shadow of a doubt the clear prophecy offered by the verse: the death of the philosopher is already close.

§8  The use of verses by famous poets as oracles is not only limited to Homer in Classical times[10]. We may recall the Herodotean passage in which the historian speaks about the exile of Onomacritus due to an accusation of inserting verses of his own into the oracles written by Musaeus, literature linked to Orphic poetry[11].

§9  Aristophanes, once again, offers other examples of the use of hexametrical oracles pretending to be the word of an authoritative voice, this time the voice of Bakis. Oracles attributed to Bakis are used for comic purposes in PaxBirds, and Knights[12]. These situations provide us with a wonderful example of how popular the consultation of books with hexametrical oracles was in Classical times and also about the poor reputation of people who used them badly.

§10  In Roman times, the use of Homer for divination is even clearer[13]. Dio Casio, in his work on the history of Rome, refers to an oracle of Zeus that used to provide verses by Homer and other known poets as oracles[14].

§11  Finally, the Pseudo-Plutarchean work de Vita Homeri[15] ends with a sentence that clearly speaks about the use of the Homeric works as oracles in contrast to other uses of his poetry.

πῶς δὲ οὐκ ἂν πᾶσαν ἀρετὴν ἀναθείημεν Ὁμήρῳ, ὅπου καὶ ὅσα αὐτὸς μὴ ἐπετήδευσε, ταῦτα οἱ ἐπιγενόμενοι ἐν τοῖς ποιήμασιν αὐτοῦ κατενόησαν; καὶ χρῶνται μέν τινες πρὸς μαντείαν τοῖς ἔπεσιν αὐτοῦ, καθάπερ τοῖς χρησμοῖς τοῦ θεοῦ, ἄλλοι δὲ ἑτέρας ὑποθέσεις προθέμενοι ἁρμόζουσιν ἐπ’ αὐτὰς τὰ ἔπη μετατιθέντες καὶ συνείροντες.

How then could we possibly not attribute every virtue to Homer, when those who have come after him have even found in his poetry all the things he did not himself think to include? Some use his poetry for divination, just like the oracles of a god, while others put forth entirely different subjects and ideas and fit the verses to them, transposing them and stringing them together in new ways[16].

§12  The passages quoted above illustrates perfectly the capacity of hexametrical poetry, attributed to well-known poets and especially to Homer, to be re-contextualized for divination in private and public spheres. Thanks to them, different ways of mediation of Homeric poetry in divination can be noticed: the use of a book, the meaning of a dream, and the performing of an oracle, and different levels of interpretation: by a professional, by the reader/listener himself, by another person who may or not be a scholar.  The adaptability of Homeric poetry for divination is, at this point, more than clear.

The Homeromanteia

Frammento di codice papiraceo con Homeromanteion e Catabasi orfica, II o II- III sec. d.C. (P.Bon. 3+4, MALTOMINI, ZPE 85 (1991), 239-43, taf. V) / Biblioteche nell’Antichità

§13  The Homeromanteion is a certain divinatory text that uses a selection of Homeric verses for offering automatic oracular answers by lot. The text was preserved in three different manuscripts: P.Lond. 121 (i. e. PGM VII), P.Oxy. 56.3831, and P.Bon. 3, listed here by the quantity of text preserved[17]. The name Homeromanteion is already attested as the title for the text in the copy of PGM VII, but according to the Oxyrhynchus papyrus, it could be also called ‘the Scimitar’[18].

§14  Let us describe briefly the different books where the Homeromanteia are preserved. P.Bon. 3 (=Suppl.Mag. II.77) is part of a quite damaged papyrus codex of miscellaneous content dated palaeographically to the 2nd-3rd century CE[19]. The Homeromanteion was preceded in the book by a hexametrical katabasis[20] related to Orphic literature, but it was written by a different hand[21].

§15  The Homeromanteion from Oxyrhynchus (P.Oxy. 3831), however, consists of a single papyrus leaf from a pocket-size codex that can be palaeographically dated to the 3rd-4th century CE[22]. The handwriting is very fast, well trained, and difficult to read because of the use of multiple abbreviations. Only one page is preserved, and it is not possible to guess the content of the complete book (if it was once a book and not an isolated sheet[23]). The text is actually incomplete. The scribe gave up writing the text after verse number αγς (136)[24] and it is not possible to provide an explanation for such a problem.

§16  The Homeromanteion preserved in PGM VII is the only example written on a scroll, and the only one that is related without any doubt to the world of divination in magic. Measuring 2.33 meters long, the PGM VII is an opisthograph bookroll, containing a handbook of magic. The Homeromanteion is located at the beginning of the roll and it is almost complete. It was written in a cursive script, quite accurate, that can be palaeographically dated to the end of the fourth century CE[25].

§17  These three Homeromanteia belong more or less to the same period, around 3rd-4th century CE. The Homeromanteion in PGM VII was edited firstly in 1893 and re-edited after 1931[26]. The discovery of P.Bon. 3 and P.Oxy. 3831 contributed significantly to a better understanding of this text, because they preserve the first part of the text, precisely the most damaged part of PGM VII. The verses of P.Oxy. 3831 were confirmed by P.Bon. 3, a fact that reveals that the three copies of the text were exact, having the same verses in the same order, a particularity that leads us to hypothesize about the existence of only one tradition of the text. Thanks to the Oxyrhynchus text, on the one hand, we know how the Homeromanteion was used, because it preserves the instructions for use that were not preserved (or they were never written) in the other two copies. Besides, the text from Oxyrhynchus makes clear that the Homeromanteion only could be performed in specific days and hours of these days, information that  decisively links the Homeromanteion of PGM VII with the following spell on the same column entitled “ἡμερομαντίαι καὶ ὧραι”[27].

§18  According to the instructions for use, the Homeromanteion was performed in the following way: firstly, the practitioner has to consult the table of days and hours in which the oracle can be performed. When the day and hour are favourable for divination, the consultant has to utter a prayer addressed to Apollo (preserved in P.Oxy. 3831 and fragmentarily in PGM VII) composed indeed by five Homeric verses (the last one modified[28]). While reciting the prayer, the consultant is supposed to think about the question he/she wants answered. Then a dice has to be thrown three times[29]. This process gets as a result a number of three digits that must be located throughout the ordered series of numbers followed by a Homeric verse. The numbers are arranged in a series of six numbers of three digits each, separated by lectional signs[30]. The arrangement of the numbers is clearly established, starting with ααα (111) and ending (only preserved in PGM VII) with ςςς (666), in the following order: ααα, ααβ, ααγ… αβα, αββ, αβγ… βαα, βαβ… ςςς, a system that covers all the possibilities of throwing a dice three times. The number obtained by the consultant leads him/her to the Homeric verse that would be the oracular answer.

§19  As in other different lot divination texts, like the well known Sortes Astrampsychi[31], and the Christian lot-divination books like the Sortes SanctorumSortes SangallensesSortes Apostollorum, etc., the mechanism of the oracle is based on chance by a mediating element that led the consultant to the oracular answer[32]. The relation of the Homeromanteion with the sortes books has been stressed by all the scholars who have studied the text, but the differences should also be pointed out.

§20  The Homeromanteion is based in the power of Homeric verses to be interpreted as an oracle, a belief deeply rooted in Classical times. The authority of the voice of Homer as an inspired poet would guarantee the effectiveness of the oracle book[33]. Homer was not a simple poet; indeed he was seen as a god in Late Antiquity. A sentence on an ostracon containing a school exercise from Karanis, a city of the Arsinoite nome in Egypt, dated to the 3rd century CE, makes clear how strong  feelings about Homer in this time could be: θεός, οὐκ ἄνθρωπος ὁ Ὅμηρος: “Homer is a god, not a human being”[34], a sentence that must be linked to the idea held by some Neoplatonists of Homer as a divinity[35]. The Homeromanteion, as a divination text already fixed, is dated to a chronological frame where, in the words of M. Beard[36], “books with pedigrees could be plausibly viewed as direct links to the supernatural worlds.” As the Bible and the Torah, the Homeric work is considered a sacred book, a text of extraordinary authority that can be re-contextualized and can be used to communicate with the gods. Therefore, the Homeromanteion must be studied stressing its status as a “sacred” book, or at least, as a text related to a sacred book.

§21  Regarding the performance of the oracle, its simplicity must be emphasized. To obtain an oracular answer for the Homeromanteion is very easy and, as W. E. Klingshirn says in relation to the Sortes Sangallenses, quoting D. Potter’s work on the Sortes Astrampsychi, “the actual oracle-monger had only a very small role: he acted as the god’s agent interpreting the wisdom of the sage Astrampsychus”[37]. In the case of the Homeromanteion the role of the “oracle-monger” is even smaller. Only an interpreter is needed to elucidate the meaning of the Homeric verse as an answer for the question, and for these matters, in the 4th century CE, not only oracle-mongers, but also learned men could do it.

§22  In contrast to the preserved books of sortes, the Homeromanteion does not present a number of fixed questions[38]. This particularity is shared with the engraved minor-Asiatic sortes[39] dated to the 2nd century CE. Yet, in contrast to these inscriptions, the oracular answer offered by the Homeromanteion is an isolated sentence, a Homeric verse, and not a small text in hexameters ruled over by a tutelary divinity[40]. In the case of the Homeromanteion, only one main divinity and, probably, a secondary one, controls the oracle: Apollo, the god to whom a prayer must be uttered at the beginning of the oracular procedure, and Athena, whose name appears when reading the first letters of the five first verses as an acrostic[41].

Homeromanteia and the use of Homeric verses in the 2nd-4th century CE

PGM VII, Segno e Testo 13, 2015, pp. 147-164 | Raquel Martín-Hernández

§23  In her article on the Homeromanteion, A. Karanika asks: “why someone, presumably the ‘author’ of the document, would produce this document, as opposed to another that uses a different selection of lines?”[42]. This question is indeed interesting, and I will try to offer some thoughts on it, although it must be noted that the three preserved Homeromanteia contain the very same version of the text. So, although we may hypothesize about the existence of different versions of the text, we do not have the evidence yet.

§24  For the moment, it is not possible to find a pattern that explains why the verses of the Homeromanteion were arranged in the order they appear, but it is a fact that they were not chosen randomly[43].

§25  Regarding their content, the verses that compose the Homeromanteion belong to most of the books of the Iliad and the Odyssey[44]. An important number of verses belong to the Doloneia and have to do with the hero Diomede[45]. The relation of this hero with Apollo and Athena, the two divinities that seem to rule over the oracle, is also an important connecting factor[46].

§26  Besides, a significant number of verses appear in other literary compositions dated to the Roman Period. Some of them were already proverbs and they were used by other writers in their own works[47]. The relatively large number of verses of the Homeromanteion that are used by Eudocia Augusta in her Homerocentos (39 verses) has been stressed by Karanika, who speculates about the existence of textbooks or summaries of Homeric works[48]. In order to support this assumption, I will suggest a new literary work that can be added for comparison in this sense. I have noticed that the Homeromanteion shares also a relatively important number of verses with the Life of Homer by Pseudo-Plutarch[49], a text that, as we have already pointed out, ends with a reflection on the use of Homeric poetry both for divination and for centos. This coincidence opens new ways for a better understanding of the reception of Homer in Roman Period, and the use of his poems by the scholars. This fact provides new proof to reinforce the idea that the “summaries” of Homer were used for erudite compositions and also, if the dating of the Pseudo-Plutarchean work to the late second and 3rd century CE is correct, calls for a reconsideration of the dating of this phenomenon.

§27  A. Karanika proposes a way to understand the Homeromanteion focused on the act of writing and reading aloud[50]. On the other hand, A. Zografou has stressed that the verses of the Homeromanteion are, mostly, “ceux qui ont une valeur prophétique déjà dans l’épopée” and “ceux qui ont un caractère gnomique ou proverbial propre à un oracle”[51], emphasizing that some of the verses belong to Homeric λόγοι, a fact that she connects with the performative aspect of the mantic poetry[52]. I would add to this observation that an interesting number of verses are not only part of λόγοι, but also contain the very first verse of a hero’s speech, a feature that could be linked to the act of memorizing texts.

§28  A close look at the layout of the Homeromanteion in PGM VII (and also of the rest of the papyri) could be helpful to reinforce this approach. The Homeromanteion preserved at the PGM VII was written carefully by the scribe who also used apostrophes to notice the elision of vowels. From my point of view, this care in the act of writing is an important feature that speaks to the literacy of the scribe, his knowledge of scribal techniques, and his concern for the correct performance of the text in the practice of the divination ritual[53].

§29  In addition, the Homeromanetion in PGM VII is followed in column 5 by the Demokritos’ “table gimmicks” (Δημοκρίτου παίγνια). The layout of these texts in the papyrus reveals that the texts are connected; maybe they were copied from the same book. The paragraphos[54] is used to separate the different constituent parts of the whole recipe (series of verses in Homeromanteion, small recipes in Demokritos’ table gimmicks), and the same decorative border preceded by a coronis is drawn at the end of both texts[55]. In this sense, the two spells are individualized in the papyrus by the scribe, while being connected with the rest of the brief magical recipes in the first part of the papyrus according to my division of the handbook’s content[56].

§30  The Demokritos’ table gimmicks and related literature have been studied as tricks intended for performance at the symposium, with a view to how the magicians operated in their social context[57]. On the assumption that these small recipes are related somehow to the Homeromanteion, perhaps, as I said above, by thinking that they were written in the same miscellaneous book from which the scribe of PGM VII is copying, it is not very daring to think that the symposium could be the right environment for performing such a learned divinatory procedure, or at least, to think that these texts were intended to be useful to the same audience/readers[58].

§31  In sum, the Homeric verses of the Homeromanteion are mostly well known, sometimes even proverbs, which were used by other authors in their own literary works, a coincidence that could be related to the existence of Homeric “summaries” for the use of scholars. In this sense, the Homeromanteion has been defined as a “centolike incantation”[59], but the whole text is not intended to be read line by line. Each verse must be understood as a complete text with its own meaning, a kind of micro-text that must be understood as the oracle answer for a formulated question[60]. The way in which the verses were written in PGM VII and the way the text is linked to other literature, in addition to the evidence that one of the Homeromanteia is preserved in a “pocket-size codex”, designed to be portable literature, pushes us to assume a connection with the ambiance of the symposium and the performance of magic in this milieu. It is easy to assume that the Homeromanteion was specially designed to be attractive for a select audience or consumer, a consumer familiar with reflecting on the meaning of Homeric verses, living in the Roman period, a reflection that leads us to question:  what was the intention of the scribe of PGM VII for including this text in his magical handbook.

The Homeromanteion of PGM VII in relation to the divination recipes on the PGM, an alien?

§32  The corpus of Greek and Greek-Demotic magical papyri preserves an incredible large number of recipes to obtain personalized answers from the gods. The different recipes are based on direct communication. This communication is obtained by a direct vision of the god (σύστασις and αὐτοψία), by natural elements like fire and water (pyromancy, lychnomancy, hydromancy), by dreams, by a medium, and by a dead (necromancy)[61]. The practice is normally a self-experience or it is limited to the practitioner and the consultant (and, in some cases, to the medium). D. Frankfurter asserts that “in the context of the priestly culture itself[62], the promulgation of these divination spells in ritual corpora typical of priests suggests that they represent a distinctive development within the priesthood: an interest (or a diversification of much older priestly interests) in personally ‘beholding’ deities that had traditionally served as ritual analogies for healing and cult performance”[63]. He also understands the divination recipes in the corpus of PGM as “a ‘freelance’ version of the temple oracles.” The professional could perform a private complex ritual that could be seen as a “miniaturization” of the divinatory process in a temple[64]. But can this explanation be applicable to the Homeromanteion?

§33  J. Dieleman[65] defines the content of the magical books from the Theban library as “the result of a desire to collect and combine ritual texts of different origins. The spells in their present state testify clearly several phases of editing, thereby demonstrating that the ancient redactors were highly skilful philologists and proficient in both Egyptian and Greek.” These large handbooks of magical knowledge are grimoires, companions in which the scribes, the magicians, try to accumulate all sort of magical recipes from different sources[66] that sometimes are even discussed and completed with other sources. The integration of the Homeromanteion in the PGM VII must be studied in this light. In the fourth century CE, the Homeromanteion seems to be a consolidated divination text, transmitted at least since the end of the second century CE. As it has been stressed above, the text has a kind of autonomy that allows it to form parts of different types of books.

§34  Let us see how the arrangement of the text in PGM VII provides us with more information about it. The Homeromanteion occupies the beginning of the recto side of the papyrus (the preserved part), i.e. a prominent place[67]. The end of the text is marked out not only with the customary paragraphos at the end of the last number of the last series, but also with a decorated border occupying the width of the column. The border is preceded by a damaged coronis that corresponds to the one at the end of the border. This kind of border is not without parallels[68]. Below the frieze, located on the middle of the column with larger letters, a title was written down by the scribe: τέλος ἔχει τῶν ἐπῶν Ὁμηρομαντείου ἐπ΄ ἀγαθῷ. The title is located in the middle of the column and is adorned with some dashes placed above and below the words, as it is the common way of writing out titles in Homeric books, especially in bookrolls[69]. The expression τέλος ἔχει + genitive is actually the complete version of writing end-titles in Homeric books according to a study by Schironi[70]. On the other hand, the expression ἐπ’ ἀγαθῷ and the ornaments is reminiscent of the decorated title preserved in the miscellaneous codex from Montserrat (forth century CE) at the end of Hadrianus’ tale[71]. The final title, clearly related to Homeric books, is the only final title preserved in the papyrus, and it clearly distinguishes the complete text from the rest of spells on the papyrus. This feature may indicate the importance given to Homer even in magic[72] and, besides, it gives us a clue about the transmission of the recipes. In this case, the title and the expression ἐπ’ ἀγαθῷ connect the divinatory text with the other Homeromanteia, especially the one preserved in Bologna, written in a book of miscellaneous content. The text was arranged in the papyrus in a “literary way”, stressing its relation to the traditional Homeric literature.

§35  Finally, it should be considered why this particular text was written in this particular magical handbook. In order to give a tentative answer to this question we should turn to D. Frankfurter’s concept of “stereotype appropriation”[73], that is, how the spells are inserted in the Greek Magical Papyri by force of the audience. According to this approach, the Egyptian priests had to find new ways to reach clients from the Hellenized population, and appealing to a Greek authority (or pseudo-authority) of the spells was only one of the ways to achieve it. The Homeromanteion is clearly a “Hellenistic” product. It was gathered in this particular book, which has a special “Hellenistic” flavour in comparison with the rest of manuals. In this book, recipes that claim to be magical texts by Pythagoras, Democritus, and even Orpheus were compiled.

§36  Magical texts in handbooks are always linked to the problem of tradition and innovation in a time of great change. In this context, the Homeromanteion is a wonderful example of how the Greek tradition is still alive, and even reinforced, in order to have its own space in the multicultural Egyptian magical wisdom.


Suppl.Mag. = Daniel, R., and Maltomini, M., eds. 1990-1991. Supplementum Magicum, II vols. Papyrologica Coloniensia vol. XVI.1 and 2. Opladen.

OF = Bernabé, A. 2004-2007. Orphicorum et Orphicis similium testimonia et fragmenta. Poetae Epici Graeci. 3 vols., München-Leipzig.

PGM = Preisendanz, K., and Henrichs, A. eds. 1974. Papyri Graecae Magicae. Die Griechischen Zauberpapyri, II vols. Stuttgart.

The papyri are quoted following the Checklist:


Alvar Nuño, A. 2006. “ΑΝΕΡΡΙΦΘΩ ΚΥΒΟΣ. Áyax y Aquiles tiran los dados.” MNHH 6:15-32.

Bakker, E. J. 2005. Pointing at the Past: From Formula to Performance in Homeric Poetics, Cambridge.

Bain, D. 1998. “Salpe’s ΠΑΙΓΝΙΑ: Athenaeus 332A and Plin. N. H. 28.38.” CQ 48 1:262-268.

Beard, M. 1991. “Writing and Religion: Ancient Literacy and the Function of the Written World in Roman Religion”, Literacy in the Roman World (ed. J. H. Humphrey) 33-58. JRA Suppl. 3. Ann Arbor.

Betz, J. D. ed. 1986. The Greek Magical Papyri in Translation including the Demotic Spells, Chicago.

Browne, G. M. 1970. “The Composition of the Sortes Astrampsychi.” BICS 17:95-100.

Collins, D. 2008a. “The Magic of Homeric Verses.” CPh 103, 3:211-236.

Collins, D. 2008b. Magic in the Ancient Greek World. Malden (MA).

Costanza, S. 2009. Corpus palmomanticum graecum. Papyrologica Florentina 39. Firenze.

Dieleman, J. 2005. Priest, Tongues and Rites. The London-Leiden Magica Manuscripts and Translation in Egyptian Ritual (100-300 CE). Leiden.

Faraone, Ch. A. 2011, “Hexametrical Incantations as Oral and Written Phenomena.” Sacred Words: Orality, Literacy and Religion, Orality and Literacy in the Ancient World 8 (eds. A. P. M. H. Lardinois, J. H. Blok and M. G. M. van der Poel) 191–204. Leiden.

Frankfurter, D. 1998. Religion in Roman Egypt. Assimilation and Resistance. Princeton.

Frankfurter, D. 2005. “Voices, Books, and Dreams: The Diversification of Divination Media in Late Antique Egypt.” Mantikê. Studies in Ancient Divination(eds. S. I. Johnston and P. T. Struck) 233-254. Boston – Leiden.

Gallo, I. 1986. Greek and Latin Papyrology. London.

Gil, J. and Torallas Tovar, S. 2010. Hadrianvs. P.Monts. Roca III. Barcelona.

Graf, F. 2005. “Rolling the Dice for an Answer.” Mantikê. Studies in Ancient Divination (eds. S. I. Johnston and P. T. Struck) 51-97. Boston – Leiden.

Heinevetter, F. 1912. Würfel- und Buchstabenorakel in Griechenland und Kleinasien. Breslau.

Johnson, W. A. 2004. Bookrolls and scribes in Oxyrynchus. Toronto.

Johnston, S. I. 2008. Ancient Greek Divination, Malden (MA).

Karanika, A. 2011. “Homer the Prophet: Homeric Verses and Divination in the Homeromanteion.” Sacred Words: Orality, Literacy and Religion. Orality and Literacy in Ancient World vol. 8, Mnemosyne Supplements (eds. A. P. M Lardinois, J. H. Block and M. G. M. van der Poel) 255-277. Leiden.

Keney, J. J. and Lamberton, R. 1996. Plutarch. Essay on the Life and Poetry of Homer. Atlanta.

Kenyon, F. 1893. Greek Papyri in the British Museum. Catalogue with Texts, I. London. 83-115.

Klingshirn, W. E. 2002. “Defining the Sortes Sanctorum: Gibbon, Du Cange, and Early Christian Lot Divination.” JECS 10:77-130.

Klingshirn, W. E. 2005. “Christian divination in Late Roman Gaul: The Sortes Sangallenses.” Mantikê. Studies in Ancient Divination (eds. S. I. Johnston and P. T. Struck) 99-128. Boston – Leiden.

Kotansky, R. “Commentary on PGM VII 167-186.” The Greek Magical Papyri in Translation including the Demotic Spells (ed. H. D. Betz) 119-120. Chicago – London.

Kraus, T. J. 2010. “Die Welt der Miniaturbücher in der Antike und Spätantike.” Studien zum Neuen Testament und seiner Umwelt 35:79-110.

Lamberton, R. 1986. Homer the Theologian. Neoplatonist Allegorical Reading and the Growth of the Epic Tradition. Berkley – London.

Lamberton, R. 1992. “The Neoplatonists and the Spiritualization of Homer.” Homer’s Ancient Readers. The Hermeneutics of Greek Epic’s Earliest Exegetes. (eds. R. Lamberton and J. J. Keaney) 115-133. Princeton.

Maltomini, F. 1991. “P.Bon. 3+4: una nota codicologica.” ZPE 85:239-243.

Maltomini F. 1995. “P. Lond. 121 (= PGM VII), 1–221: HOMEROMANTEION.” ZPE 106:107–122.

Martín Hernández, R. forthcoming a. “Two More Verses for the Homeromanteion (PGM VII).” ZPE.

Martín Hernández, R. forthcoming b. “The Use of Lectional Signs in PGM VII.”

Martínez, J. 2011. “Onomacritus the Forger, Hipparchus’ Scapegoat?.” Fakes and Forgers of Classical Literature, (ed. J. Martínez) 217-226. Madrid.

Mastrocinque, A. 2002. “The Divinatory Kit from Pergamon and Greek Magic in Late Antiquitiy.” JRA 15:173-187.

Mastrocinque, A. 2003. “Dinners with the Magus.” MHNH 3:75-94.

McNamee, K. 2007. Annotations in Greek and Latin text from Egypt. American Studies in Papyrology vol. 45. New Haven.

Montevecchi, O. and Pighi, G. B. 1947. “Prima ricognizione dei papiri dell’Università di Bologna.” Aegyptus 27: 80-83.

Nagy, G. 1996. Poetry as Performance. Homer and Beyond. Cambridge – New York.

Nagy, G. 2002. “The language of Heroes as Mantic Poetry: Hypokrisis in Homer.” EPEA PTEROENTA. Beiträge zur Homerforschung. Festschrift für Wolfgang Kullmannn zum 75. Geburtstag.” (eds. M. Reichel and A. Rengakons) 141-150. Stuttgart.

Nollé, J. 2007. Kleinasiatische Losorakel. Astragal- und Alphabetchresmologien der hochkaiserzeitlichen Orakelrenaissance. München.

Olson, S. D. 1998. Aristophanes. Peace. Oxford.

Parsons, P. 1989. “3831. Homer Oracle.” The Oxyrhynchus Papyri 56: 44-48. Oxford.

Parsons, P. 2012. “Homer: papyri and performance.” I papiri omerici. Atti del convegno internazionale di studi, Firenze, 1-10 Giugno 2011. (eds. G. Bastianini and A. Casanova)  17-27. Firenze.

Potter, D. 1994. Prophets and Emperors: Human and Divine Authority from Augustus to Theodosius. Cambridge (MA).

Schironi, F. 2010. To Mega Biblion: Book-Ends, End-Titles and Coronides in Papyri with Hexametric Poetry. American Studies in Papyrology vol. 48. Durham (NC).

Schwendner, G. 2002. “Under Homer’s Spell.” Magic and Divination in the Ancient World (eds. L. Ciraolo and J. Seide) 107-118. Leiden.

Stewart, R. 1995. “The Textual Transmission of the Sortes Astrampsychi.” ICS 20:135-147.

Stewart, R. 2001. Sortes Astrampsychi vol. 2. Munich-Leipzig.

Suárez de la Torre, E. 2011. “Versos homéricos en los papiros mágicos griegos.” Ἀντίδωρον. Homenaje a Juan José Moralejo (eds. M. J. García Blanco et alii) 527-543. Santiago de Compostela.

Suárez de la Torre, E. 2013. “Divination et magie. Remarques sur les papyrus grecs de l’Égypte gréco-romaine.” Kernos 26:152-172.

Turner, E. G. 1977. The Typology of the Early Codex. Pennsylvania.

van der Horst, P. W. 1998. “Sortes: Sacred Books as Instant Oracles in Late Antiquity.” The Use of Sacred Books in Ancient World (eds. L. V. Rutgers et alii) 143-174. Leuven.

Wessely, C. 1893. Neue griechische Zauberpapyri, Denkschriften der kaiserlichen Akademie der Wissenschaften in Wien, philos.-hist. classe XLII. 2. Wien.

Zografou, A. 2013. “Un oracle homérique de l’Antiquité tardive : un livre-miniature à usage oraculaire.” Kernos 26:173-190.


[1] The text of the papyrus has been studied with the help of high quality digital images. As part of my complete project, I have carried out a new edition of the text that will be available on-line through i-Mouseion project of the Center for Hellenic Studies.

[2]I will keep aside in this study the discussion of the use of Homeric verses for different magical purposes, especially for healing. Verses of Homer appear in two other magical handbooks, but they are not related to divination; PGM IV 468-474 and 820-833, and PGM XXIIa. For the use of Homer in magic see Collins 2008a and 2008b, Suárez de la Torre 2011.

[3] See e. g. Beard 1991 and van der Horst 1998.

[4] Augustine, Epist. 55.20.37. Jerome, Epistula ad Paulinum Nolanum 53, 7 (CSEL 54, 453). See Klingshirn 2002: 82-84.

[5] The use of the Bible for divination was not only conducted by secular people, but also by members of the clergy on the light of Canon 16 of the council of Vannes, dated to the 462 and 468 CE: aliquanti clerici student auguriis et sub nomine confictae religionis quas sactorum sortes uocant…hoc quicumque clericus detectus fuerit uel consulere uel docere ab ecclesia habeatur extraneus. “Some clergy are devoted to the interpretation of the sign, and under the label of what pretends to be religion, what they call Saints’ Lots…any cleric found either to have consulted or expounded this should be considered estranged from the church” (Concilia Galliae, A. 314-A. 506 [CCSL 148:156]). Text provided by Klingshirn 2005: 100. For the use of the Bible for divination see Klingshirn 2005.

[6] Aristophanes Peace 1089-1094.

[7] Translation by A. H. Sommerstein 1998.

[8] The verses are a mixture of Homeric verse-fragments and free composition in Homeric dialect. See the commentary on the verse by Olson 1998.

[9] Plato Crito 44b. The Homeric verse is Iliad IX 363.

[10] The different attestations show the relation between oracles and hexametrical poetry, a relation widely studied. In PGM I, 328-333, the ἐποποιία (divination through hexameters) is listed among other types of divination. For the use of hexameters in different magical practices see Faraone 2011.

[11] See Herodotus VII 6 2 (OF 1109) and Pausanias I 22 7. On Onomacritus’ forgeries see Martínez 2011 with bibliography. On the other hand, other commentaries related to Orphic literature could be interpreted in the same light; as verses used for divination. Scholium to Euripides’ Alcestis 968 (II 239, 3 Schwartz), and Scholium to Euripides’ Hecuba 1267 (I 89, 12 Schwartz = OF 813 II), speak about verses by Orpheus that serve as oracles in a Dionysian temple. The word describing the media in which the Orphic verses were written is πίναξ, the same word used by Pausanias in his description of the oracular text at the temple of Heracles, a lot oracle played with astragaloi (Pausanias VII 25 10).

[12] Aristophanes Peace 1070-1072, Birds 970-975, Knights 124-149. Herodotus (VIII 20 77), Pausanias (IV 27 4) and Plutarch (Mul. Virt. 243B and De Pyth. Or. 399A) refer to Bakis as a compiler of oracles. Sometimes in the sources his name appears in plural as a synonym for ‘soothsayer’ (like Sibyl/Sibyls). His name uses to be linked with Delphi. On Bakis see e.g. Fonterose 1978: 145-165. I would like to thank Prof. Richard Martin for pointing this information out to me.

[13] See van der Horst 1998: 163 and Zografou 2013: 177 in relation to the following passages quoted.

[14] In Dio Cassius 79.8.5-6 the oracle of Zeus Belus gives response to Emperor Severus in a first moment with 2 verses of the Iliad (II 478 and 479) and, in a second visit to the oracle, with one verse of Euripides Phoenissae (20). In Dio Cassius LXXIX 40 3, the author refers again to the oracle of Zeus Belus and this time, again, two Homeric hexameters configure the oracular answer (Iliad VII 102-103).

[15] Pseudo-Plutarch Life of Homer  II 218 4. On the Life of Homer see Keney and Lamberton 1996.

[16] Translation by Keney and Lamberton 1996 cit.

[17] In chronological order they are P.Bon. 3, P.Oxy. 3831, and PGM VII.

[18] The alternative name is strange. It could be related to the magical recipe named “The sword of Dardanos” in PGM IV 1716-1870. See Schwendner 2002: 108-109 and commentary of the magical recipe in Betz 1986 ad loc.

[19] See Montevecchi and Pighi 1947, and Maltomini 1991. P.Bon. 3 = Suppl.Mag. II.77.

[20] OF 717 (pages 272-287).

[21] For details see Maltomini 1991.

[22] Parsons 1994. See Turner 1977 typology 22 and 29, and Kraus 2010. It is interesting to note that other divinatory texts are preserved in pocket-side books. A handbook of Palmomancy (P.Ryl. 1.28), attributed to Pseudo-Melampous, is a pocket-codex dated to the fourth century CE. See Costanza 2009. Another miniature book, in which a divinatory text of sortes is preserved, is about to be published by Anne Marie Luijendijk. This publication will be very useful to our work on the material transmission of magical knowledge and the use of these texts by monger-diviners.

[23] It is easier to assume that the page belonged to a codex, because the text stars in verso side and follows in recto side, something normal in codices depending on the arrangement of the different sheets to make the codex, but strange in a document written in one sheet.

[24] The numbers are written until the end of the column (last number is αεα), but the corresponding verses are not written. It could be hypothesized that there was not a necessity for the scribe (the owner and/or the user) of the book to write the complete oracle, on the idea that he knew the complete oracle by heart, but the hypothesis cannot be prove by material evidences.

[25] The date of the papyrus is discussed and it fluctuates from the end of the 3rd century to the 5th. The different editors of the text propose: Wessely 1893 3rd-4th CE, Kenyon 1893 and Preisendanz 1931 3rd CE, and Maltomini 1995 4th-5th CE.

[26] See footnote 24 for bibliography.

[27] See PGM VII lines 155-167. It is important to point out that the two calendars are not matching in the selection of favourable days and hours.

[28] The first 3 verses are consecutive, Iliad XVI 514-516. Forth and the fifth verses are not; Odyssey i 174 and Odyssey xviii 113. The person of the verb in the last verse was modified in order to fit in the prayer, a practice that reminds the use of Homeric verses for the composition of centos.

[29] This oracle could only be performed with dices because the numbers 2 and 5 are listed in the series of numbers, so it was not possible to be performed with knucklebones (ἀστράγαλοι), an irregular piece used in other lot-divination systems that only have 4 positions with the values 1, 3, 4 and 6.

[30] In PGM VII a single paragraphos at the end of each series of number and, in colums 1, 2 and 3 two simple paragraphoi, one beneath the number and one beneath the first word of the Homeric verse. In P.Oxy. 3831 a single paragraphos, and P.Bon. 3 a forked paragraphos. On the lectional signs of PGM VII see Martín-Hernández forthcoming b.

[31] On the Sortes Astrampsychi see Browne 1970, Stewart 1995 and 2001.

[32] On the different systems of randomization in lots oracles see Graf (2005) p. 61.

[33] On the hidden meaning of Homeric verses and the relation between the performative aspect of the poetry and divination see Nagy 2002. On Homer and performance see e.g. Nagy 1996 and Bakker 2005.

[34] O.Mich. 3.1100, 10-11 = TM 64126. Edition by H. C. Youtie and J. G. Winter, Michigan Papyri, VIII: Papyri and Ostraca from Karanis, second series, Ann Arbor, 1951, p. 206. Image:

[35] See Lamberton 1986: 1-43, Lamberton 1992, and van der Horst 1998: 160.

[36] Beard 1991.

[37] See W. E. Klingshirn 2005: 102 quoting D. Potter 1994: 25.

[38] It can be argued that the catalogue of question is not preserved in any of the manuscripts, but according to the mechanics of the oracle a catalogue of question is not necessary. Besides, P.Oxy. 3831 makes clear that the preserved text is the beginning of the book because the leaf was numbered with an α, and the instructions makes clear that the question must be thought by the consulter: εὔξασθαι ἐν σεα̣υτῶι̣ πρ.[ / ἃ βούλει, “pray inwardly for what you want.” See Parsons 1989.

[39] See Heinevetter 1912, Graf 2005, and Nollé 2007.

[40] In this case the Minor-asiatic sortes inscriptions are closer to the Sortes Astrampsychi and related sortes books in which each decad or dodecad is ruled over a divinity. The Christian variations address the groups of answers to God, Angels, the Virgin Mary, Ancient Testament wise men and women, and other powerful beings of the Christian pantheon.

[41] See Zografou 2013: 183. Both gods are very important from a Greek oracular tradition point of view. Apollo is the god of divination par excellence, and Athena is related to lot-divination according to Zenobius 5.75.5-8. For the relation of Athena with sortes in the context of the Iliad and vase paintings see Alvar Nuño 2006. It must be pointed out another possibility of reading the acrostic. The first 8 verses can be read as Ἀθήναειος, the well known name Ἀθήναιος with the very common typo ει > ι. This reading, while interesting because it could be interpreted as the name of the composer of the text, is, however, less suggestive. The relation of the selected Homeric verses in the text with the goddess Athena and the role of the goddess as patron of the sortes divination make the first interpretation preferable.

[42] Karanika 2011: 267

[43] Only the first 5 verses seem to have been ordered with an intention, because we can read the name Athena in the acrostic. An important study to analyze the selection of verses has been made by Zographou 2012.

[44] Only the books Iliad XIII and XXIII, and Odyssey xv are not represented in the Homeromanteion. But we cannot assert that these books were not used in the Homeromanteion, because ten verses are still unidentified and the verse corresponding to number ββς could be Iliad XXIII 771.

[45] It must be noticed that the Homeric verses appearing in PGM IV 468-474 and 820-833 are also related to Diomedes and the Doloneia.

[46] The connection is pointed out by Zographou 2013.

[47] Athenaeus, Dio Cassius, Porphyry, Plutarch, etc. use verses that appear in the Homeromanteion, and some of them are quoted in the gnomological compendia.

[48] See Karanika 2011: 269 n. 46 for a list of the coincident verses in the Homeromanteion and the in works by Eudocia Augusta. The coincidence of the verses between the two texts has been related to the date of the composition of both texts (Karanika 2011: 269), but it is important to remind that the Homeromanteion in P.Bon. 3 is dated to the 2nd-3rd century CE.

[49] As far as I know, no scholar has pointed out this interesting coincidence. The verses of the Homeromanteion that appear in the Life of Homer are 19 (or 20). The list of coincidental numbers are: ααβ, αγς?, αδγ, αδς, αεβ, αςδ, αςςβαα, βαε (according to the new identification of the verse in my edition), βςδ (twice in Life of Homer), γδα, γες, γςα, δγδ, δςγ, δςεδςςεαδ, ½ εαε. I have highlighted some numbers in bold to stress that they are consecutive numbers in the Homeromanteion.

[50] Karanika 2011: 268.

[51] Zografou 2013: 180.

[52] See Nagy 2002.

[53] On the relation of punctuation and performance in Homeric papyri see Parsons 2012.

[54] The paragraphos is a lectional sign that consists in a horizontal line inserted in the left margin between two lines of the text to separate sections. See Gallo 1986: 90.

[55] The Homeromanteion and the Paignia were marked out with a special decoration that has parallels in literary books, especially in anthologies and miscellaneous books, a fact that relates the text in PGM VII with the other Homeromanteia, one written in a miscellaneous book, the other one in a pocket-size codex.

[56] All of the texts belonging to this first part (from the beginning of the roll to column 6) are brief recipes, not very sophisticated to be performed, and belonging to the kind of recipes related to the “paignia literature.” See Bain 1998, and Martín-Hernández forthcoming b.

[57] See e.g. Bain 1998 and Mastrocinque 2003. For the literary examples of these tricks see Kontansky 1986.

[58] For divination in learned environment see e.g. Ammiamus Marcellinus XXIV 1 29-32. On the Trial of Hilarius and Patricius and the sources transmitting the episode see Mastrocinque 2002: 178, n. 14.

[59] Karanika 2011: 266, n. 40.

[60] Some of the verses of the Homeromanteion are modified in order to be understood without the context.

[61] For a classification see e.g. Frankfurter 1998: 179-184, and 2005, Johnston 2008: 153-175, and Suárez de la Torre 2013: 166-167.

[62] In the idea that the PGM are the product of Egyptian priests.

[63] D. Frankfurter 1998: 181.

[64] See Johnston 2008: 153-179.

[65] Dieleman 2005: 11

[66] The above discussion about the relation among recipes in the first part of PGM VII supports this approach.

[67] It is highly plausible that the preserved papyrus corresponds mostly to the complete book by looking at the disposition of the text in verso side and the addition of the spells by the second scribe. The text in P.Oxy. 3831 preserves the number α written on the top of the sheet, probably an indication of the first page of the codex, a feature that suggests the same idea of a prominent position of the text in the different books.

[68] The border remembers the ornamented title of a Demosthenes’ book, P.Flor. 4, and the one drawn at the end of Iliad 14 in the Morgan Homer (P.Amh. inv. G 202). Borders like this one appear also for separating different poems, see e.g. P.Oxy. 13.1614. The border marks the end of the first Olympian.

[69] See Schironi 2010: 23. This type of ornamentation is well documented in scrolls and also in codices, for example P.Oxy. 10.1231 fr. 56; P.Bodmer14-15 ( (first dated to the second century, second to the 2nd-3rd century CE), P. Lond. Lit. 132 (Papyrus Arden). See plate 16 in Johnson 2004. See P. Lond. Lit. 30 (Odyssey) plate IV – V in McNamee 2007.

[70] See Schironi 2010: 21-24

[71] Gil and Torallas Tovar 2010: 30. The expression has been connected with the Latin feliciter.

[72] In this sense we have to remind that the Homeric verses in PGM IV were written by the scribe of the magical codex with larger letters.

[73] Frankfurter 1998: 224–237.

Martín-Hernández, Raquel. “Using Homer for Divination: Homeromanteia in Context.” CHS Research Bulletin 2, no. 1 (2013).