Throwing dice to generate oracles in Roman times

My last post here looked at some examples of Roman 20-sided dice with numerals on them, almost certainly used to create oracles, to discover the future.  There is some literary evidence of this sort of practice, and I want to review it here.

In Pausanias’ Description of Greece 7.25.10, written in the 2nd century AD, we have an account of how a temple would use a random number generator, plus a list of oracular replies, to allow an enquiry of the god.  Here is the passage, which I take from F. Graf, “Rolling the dice for an answer” in S. I. Johnston & P. T. Struck, Mantikê: Studies in Ancient Divination, Brill (2005), p.51-97; p.62.  Graf rightly observes that this approach cannot have been all that common, because Pausanias has to explain the procedure:

When one descends from Bura [in Achaea] towards the sea, there is the Buraikos river and a not large image of Herakles in a grotto; he too is called Buraikos, and he offers an oracle from a list (pinax) and from astragaloi. Whoever intends to consult the divinity, prays in front of the image, and after the prayer, he takes up four astragaloi (plenty of them are lying around Herakles) and rolls them on the table. For any combination of the astragaloi, the inscription in the list gives an easily accessible explanation of the combination.

astragali are literally knucklebones, perhaps with numbers inscribed on the sides, but no doubt dice could be and were used in the same way.  The astragalos was thrown, and a number from 1 to 6 produced.  Graf adds:

This description contains the two main elements that make this type of oracle function: astragaloi, and a list of answers.

Pausanias’ list is lost, but in the Anatolian inscriptions, we possess an entire set of them; we just have to add the several astragaloi that were thrown, the combination of which led to the answer.

The monuments referred to by Graf are the main subject of his excellent if rather dense paper.  They is a set of 17 “large and impressive” inscriptions on stone blocks, about six-foot tall and two-foot wide, in Lycia.  The ruins of Termessos contain a number of these texts.  The inscriptions consist of lists of answers, arranged in ascending order.  Graf gives one example.  The enquirer asks about a voyage that he is intending.  The priest throws the astragalos 5 times.  Here is one of the possible results:

Three “chians” and a six and the fifth a four: Sail wherever you wish; you will return full of joy, for you have found and accomplished everything that you are cherishing in your mind. Cypris likes you, the daughter of Zeus who likes to smile.

The “chios” is the technical term for throwing a “1” with an astragalos, so the text means that, if you throw three ones, a six, and a four, then this answer is the one.

Graf’s paper includes an appendix with a translation of the main text preserved on these inscriptions.  There are various resources around the web on astragalomancy.

So these two sources indicate the approach: throw the dice, or knucklebones, and look up the result in the tables of answers.  A table, perhaps of stone, with the knucklebones on it sat in front of the stones with the answers.

A further literary source for this is a scholion on Pindar, Pythian Odes, poem 4, line 337.[1]  Mopsus is “drawing lots” to find out the “will of heaven”.  Here is the text of the ode, for context:

But, when the flower of the seamen came down to the shore of Iolcus, Jason numbered them and praised them, every one; and, to aid him, Mopsus, after inquiring the will of heaven by noting the flight of birds and by drawing lots (κλάροισιν), right gladly gave the host the signal to set forth.

The scholia are at Perseus here.  As with Pausanias, this refers to a “holy table” in the temple, on which the astragali lay, ready to be used.  Graf again helpfully gives a translation of the key bits:

a. …καὶ ἐν τοῖς ἱεροῖς ἀστράγαλοι κεῖνται, οἷς διαμαντεύονται βάλλοντες αὐτούς.

there are astragaloi in the sanctuaries with which they take oracles by throwing them

b. κλάροισιν: ἰστέον ὅτι κλήροις τοπρὶν ἐμαντεύοντο, καὶ ἦσαν ἐπὶ τῶν ἱερῶν τραπεζῶν ἀστράγαλοι, οἷς ῥίπτοντες ἐμαντεύοντο.

… “there were astragaloi on the sacred tables with which they used to take oracles by throwing them”;

So this confirms the picture that we have already formed.

The inscriptions from Lycia make clear that several throws of the dice or knucklebones are combined in order to get the god’s answer.  The 20-sided Roman dice in the last post would look much more impressive than a few knucklebones, while providing the same function.  The large size and impressive appearance of these dice may be important.  Divination is a form of charlatanry, aimed at convincing the client of something that the diviner knows that he does not know.  So it is important for the diviner to appear impressive.  In the same way the use of 20-sided dice in modern Dungeons and Dragons is not just for convenience; the same end could be produced by several 6-sided ordinary dice. But using these unusual dice does give an air of something special and different.

This is not just my imagination.  The divination process could indeed be deliberately dressed up to be more complex than it needed to be, as is clear from the way that another ancient oracle handbook, the Sortes Astrampyschi is structured.  There the name of a deity is associated with the outcome, but by an unnecessarily complex series of dice throws.

Impressive-seeming objects from Egypt, used to communicate with the gods, immediately reminds us of the magical papyri.  It may be asked whether there is any connection.

A collection of papyrus books, some in demotic, some in Greek, containing magic spells and rituals was uncovered somehow at Luxor in the 1820s and passed into the hands of an Armenian adventurer calling himself Jean d’Anastasi.  No doubt the books came from the tomb of some Graeco-Egyptian priest.  Egypt was famous in antiquity for its magicians.  The existence of dice, also in both languages, combined with our literary testimony above, suggests that the 20-sided dice were not made for games of chance, but rather for use as tools in  ancient magical procedures, such as divination.

We do not possess any oracle book that expects the use of one or more 20-sided dice.  But we have seen at the Anatolian temple an example where the oracles needed to combine one or more dice to get a wider range of results.  Twenty-sided dice are another way to achieve the same end.

Other books of divination using lots or random numbers also exist.  In his 1913 book, Greek divination; a study of its methods and principles, (online here) W. Halliday discusses a great number of them, running into the middle ages.  I won’t go into any of these here.

It is interesting to reflect that these oracular books, and these 20-sided dice, may have been part of the professional equipment of a temple, or possibly the toolkit of an ancient magician.

Share
1. [1]In the Loeb this is p.218-9 – find it at Loebolus here.

6 thoughts on “Throwing dice to generate oracles in Roman times”

1. CATHY GRISHAM says:

Fascinating. Similar techniques are in still in practice today in many sacred spots throughout the Far East. I recall doing this in Japan and China both, but you use a number generated by answering some questions first.

2. This is most interesting! You don’t have any more details on this?

3. Ronan says:

Many temples in the Far East use a similar procedure in which one consults the oracles by means of shaking a bamboo cylinder filled with bamboo strips inscribed with a number until one falls out. One then consults the relevant strip of paper, usually stored in a cabinet to the side, to obtain a prophetic poem, plus some practical indications relating to one’s query. The full procedure is detailed here: https://en.wikipedia.org/wiki/Kau_chim

4. Thank you, yes. The first is the Minas Nerpal publication of the demotic dice. The second was new to me, a lost dice from Etruria. When I can I will look at these again.