Ancient history can be disappointing. Everybody knows that an inscription is often broken at the place where it gives crucial information, and the evidence for oracular consultation at Dodona is a similar case. The following discussion is neither fully new nor fully original. Keeping at bay numerous modern hypotheses, which have too often been considered as the objective truth, I have rather tried to untangle the complex web of evidence from Dodona into a series of logical and simple questions that badly need answering before any further enquiry can be made. This is research in progress, in the context of the Dodona online project.Footnote ¹

Introduction: Puzzling Dodona

For the site and the oracle of Dodona, we have (1) substantial archaeological remains, (2) extensive excavations, (3) a large and ancient literary tradition and (4) myriad lead lamellae reporting the questions asked to the gods, dating from the late sixth century BCE to the Hellenistic (or even Roman) era. These are four pieces of a challenging puzzle, which, nevertheless, proves impossible to solve.

(1) The archaeological remains, including the ancient seat of the oracle, mainly date from the late fourth century and the Hellenistic period. We do not know whether the sack of the sanctuary by Aemilius Paulus in 168/7 BCE rang the death-knell of the oracle, but the sanctuary itself never really recovered.Footnote ² (2) Since the first campaigns of Carapanos between 1872 and 1877, the numerous excavations that took place mainly from the 1920s to the 1970s have been poorly published, to say the least.Footnote ³ More importantly for us, (3) the literary tradition is scattered over many centuries and has more to do with literary topoi than with direct testimony. As a result, for example, extant texts give us no clear information about the mode(s) of divination. Finally, a Gordian knot, (4): the lamellae barely give any hint about the way they were used. This will be the focus of this essay.

Epigraphic and Literary Traditions: The Flickering Image of the Oracle

The lamellae, containing the questions asked of the oracle, form an impressive bulk of material. Although they were known from the late nineteenth century, they only popped up in modern research a century later, with Éric Lhôte’s new publication of all the previously published texts, and with Esther Eidinow’s work about oracles, curses and risk in ancient Greece (which contained some new inscriptions, passed on to her by the late Professor Christidis). In 2013, came the long-awaited publication of a corpus, bringing scholars about 4000 new texts. Among those texts, which were often damaged or reduced to mere letters, about 1500 provide useful information about the oracular questions asked by the consultants – quite an extraordinary event in the fields of epigraphy and Greek religion.Footnote ⁴ Moreover, some 4000 other lamellae, collected from Sotiris Dakaris’ excavations and later campaigns, are still awaiting publication in the museum of Ioannina.Footnote ⁵ This is to say that there is no shortage of texts. Quite the contrary: if we consider the total of all epigraphically attested oracular consultations, the Dodonaean oracle represents almost 90 percent of the corpus. In other words, Dodona has delivered thirty times more usable inscriptions than Delphi,Footnote ⁶ but, as at Delphi, no inscription directly illuminates the mode of divination.

Unfortunately, the literary tradition of Dodona, although rich and ancient, going back to Homer and Hesiod, makes no mention at all of the use of lamellae.Footnote ⁷ This is not surprising. Except for the case of the oracle of Trophonios in Boiotia, described in an outstanding account by Pausanias, himself a consultant, no Greek oracle ever received a detailed literary report about its practical functioning. There are a few lines of Tacitus about the oracle at Claros, an obscure paragraph of Iamblichus for Didyma, many scattered and small details for Delphi, and so on: this literary scarcity forces us to rely on rare epigraphic data, such as a famous inscription detailing some aspects of the oracle of Apollo Koropaios in Thessaly.Footnote ⁸ In the absence of textual information, architectural plans and remains stay silent, and Didyma, one of the biggest and best preserved oracular sites, has given rise to many divergent interpretations,Footnote ⁹ all of them debatable. This is partly due to the nature of Greek literature, which so often spreads topoi rather than precise information. We are far better informed about the ideal type of a Greek oracle for a Greek mind than about factual details. Again, Delphi is among the best examples.Footnote ¹⁰ Its rich epigraphy does not provide a single mention of the prophet, who is attested by Herodotus and Plutarch, while the priests are not attested on stone before the third century BCE.

Thus, matching literature and epigraphy for oracles proves highly perilous: there is an abysmal gap between the questions and answers provided by Dodona according to the literary sources, and the questions asked on the lamellae.Footnote ¹¹ The former obviously belong to the topoi of oracles, related by authors often much later than the events they are supposed to report. The latter are inscriptions written specifically for the consultation, and belong to the highest level of ‘reality’. Finding a compromise in accordance with our own logic is a temptation, but probably not the best solution.

Actually, the Dodonaean oracle is less well known than the manteion of Delphi itself.Footnote ¹² Homer, Pindar, Sophocles and Euripides speak about male Selloi or (H)elloi, who would have been hypophetai, whatever that could mean;Footnote ¹³ Herodotus and Sophocles about dove-prophetesses;Footnote ¹⁴ and Strabo mentions symbola (gestures), such as those used at the oracle of Ammon in Libya.Footnote ¹⁵ Plato alone speaks about inspired prophetesses,Footnote ¹⁶ Ephorus about prophētis, a ‘female prophet’;Footnote ¹⁷ Callisthenes/Cicero about lots.Footnote ¹⁸ The tradition transmitted by Aeschylus tells about ‘an oak that speaks’ (the oak more than its leaves).Footnote ¹⁹ Ephorus tells us about sound oracles thanks to the bronze whip of a statue attached to a lebes, ‘a cauldron’;Footnote ²⁰ or through the never-ending tolling of bronze cauldrons set in a circle.Footnote ²¹ Others speak about doves,Footnote ²² or even dreams.Footnote ²³

All kinds of hypotheses have attempted to align this polymorphous information, but they resemble intellectual contortions rather than reality. Often a chronological perspective is adopted: for example, the Selloi/ (H)elloi of Homer and Pindar would have been replaced by the prophetesses at the time of Plato and so on (and Strabo even specifies that the charge of the Selloi was abolished); nevertheless, this looks like a desperate way to give a semblance of meaning to a hopeless oral and written tradition.Footnote ²⁴ To tell the truth, we simply do not know who was delivering the oracles to the consultants. We also do not know whether divination was ‘induced’, that is, obtained through the interpretation of signs (like the behavior of doves),Footnote ²⁵ or ‘inspired’ by the gods in the prophet’s mind in order to be pronounced as a text (as Plato).Footnote ²⁶

Back to the Lamellae: What Were They For? And the Question of ‘Archiving’

This long introduction was necessary to make it clear that the role of the lamellae during the consultation is anything but an easy problem. If we read a lamella (e.g. Lhôte Reference Lhôte2006: no 14): the Dodonaeans ask Zeus and Dione whether it is because of the impurity of some man that the god sends the storm’),Footnote ²⁷ it is too easy to declare that this text was the written enquiry and to simply move on. Some problems remain: when and for what reason was it written, and why and when was it left in place? Many solutions have been proposed, none is fully convincing.

Apparently, each sanctuary had its own way to record its activity, or not.Footnote ²⁸ I say ‘apparently’ because not all the documents produced in oracular sanctuaries have been preserved. If the oracles were written on media that have now disappeared, like linen, wood, papyrus, wax and so on, any compelling conclusion from the extant evidence may be seriously biased. At Lebadeia, for example, all the pilgrims had to record on a tablet (called a pinax) ‘what they heard and saw’ during their contact with the god (so, at least ‘Trophonios’ answer’, and maybe the question).Footnote ²⁹ None of those Lebadeian tablets has ever been unearthed, but Pausanias, who himself consulted the oracle, is a generally reliable source.Footnote ³⁰ At Dodona, the situation is quite different: only the questions were engraved, not on stone but on lead strips. If some answers too were written, they remain so rare that they obviously represent a kind of exception, to which no logical explanation has been found.Footnote ³¹

Some marks on the lamellae have long been considered as a proof of archiving in a special room of the sanctuary, in order to track back specific consultations. Such archiving would have required some specific markers, like a date (never attested in the lamellae), a name, a patronym, a geographic specification and/or the topic of the consultation, and so on. At Dodona, there are some names (‘Melas’, DVC 3820A), and, at best, we find the nature of the question (‘Mr/ Ms So-and-so about children’, e.g. DVC 3023B), but nothing substantial enough to constitute a real collection of archival information.Footnote ³²

This theory has limitations for other important reasons: (1) although the lamellae were found mainly around the oracular enclosure, many others were scattered all over the sanctuary.Footnote ³³ (2) The idea of an archiving process was attractive as long as scholars were convinced that the raison d’être of major Greek oracles was big politics, like international and civil wars, or foreign policy. Now, the lamellae of Dodona statistically confirm that less than two percent of the questions were political, most of them concerning domestic policy, especially about polis religion: only three are directly aimed at the conclusion of an inter-state alliance or internal appeasement (peace rather than war!).Footnote ³⁴ (3) It would have been logical to file the oracles according to the answers, to keep a record of what the god said, but those in Dodona are cruelly missing.

By contrast, the questions were mostly personal, and always concern, as it were, ‘minor’ problems (even if they were important for those who experienced them, e.g., about the opportunity to take part in military actions).Footnote ³⁵ What could be secret in such a question: ‘Do I have to marry [x]?’, with the very probable answer ‘yes’. One consultant out of five asks such a question. (4) Most of the lamellae were folded often more than once; some were also nailed, with the purpose of hiding the written face of the lead scroll, so that only the ones that bear a kind of small abstract on the verso could have been properly archived. (5) Most of the lamellae were palimpsests (from two to ten questions on the same lamella [e.g. M210]). They were then part of a recycling process, which contradicts the idea of filing, because no archivist would have been able to date the lamellae, for example, in order to release the oldest ones first for reuse.Footnote ³⁶ Usually, as far as a precise dating is possible, the inscriptions on the same lamella are close in time, but not always.Footnote ³⁷

The Uses of the Lamellae: A Tentative Approach

The way the lamellae were displayed in antiquity is unknown. If we can plausibly abandon the idea of any archiving, it seems that the inscriptions were mainly set around the oracle, where they were mostly found. Maybe some were left in the open wherever there was some free space, maybe on light structures that left no visible trace during the excavations, like wood.Footnote ³⁸

Two preliminary clarifications are necessary. (1) The fact that some questions were ‘hidden’ could fuel the old debate about the dishonesty of the oracular officials. From the seventeenth century, scholars considered oracles as forgeries by sanctuary priests, who needed to know about the question in order to provide an answer. Any attempt to conceal a question was thus interpreted as an attempt to avoid any corruption of process. Again, this was an acceptable idea for oracles that were thought to mainly deal with kings and States, but it is difficult to understand why there would be any potential rigging of answers about marriage, job, illness, small business matters and the like, from unknown people coming from far away. Anyway, a nailed lamella is easy to unfold and read. Prophets and prophetesses were definitely not appointed to cheat people; they were there to pass the divine answer to the consultant, because they trust their own oracular god, in this case the supreme god himself, Zeus. (2) A second bias has been to suppose a malevolent action by the consultant, as if Herodotus’ stories were simply descriptions of reality. In these stories, the treacherous consultants are systematically and severely punished by the god. For the same reasons, it would be inappropriate to consider writing the question to be a safeguard, for the sanctuary, to avoid nasty or unholy questions. Even the inquiries which asked: ‘Is it more advantageous to do what I have in mind?’ were not an indication of dubious practice. If ‘what somebody has in mind’ is not explicit, it is not because the pilgrim wanted to hide its question to the officials, nor because he took the god for a dummy. The divinity knows everything, especially Zeus, as is clear as early as Hesiod, then Theognis and Pindar.Footnote ³⁹ It is just because the space to write on a lamella was desperately small, for projects that could not be described in two or three lines,Footnote ⁴⁰ or for projects whose initial idea was clear but still needed some maturation. The consultant literally expected the omniscient god to ‘read his mind’.Footnote ⁴¹

Having said that, the main question is this: when and why were the lamellae used during the consultation? As nothing definite stands out from the literary tradition, the only certitude left is that the questions, written on the lamellae were asked of the gods. The lamellae were almost always written by the consultants themselves. There are so many different dialects, so many kinds of script and levels of language, that this can no longer be in dispute. The questions are formulaic, and the two main patterns are: ‘Is it better and more advantageous to do X?’ (or similar), and ‘To which gods do we have to offer prayers and sacrifices to get X?’ (or similar). These are shared throughout the Greek world from the earliest oracular texts.Footnote ⁴²

The question of literacy remains tricky.Footnote ⁴³ Especially for the late sixth century and the first half of the fifth century, we do not know how many Greeks could read and write. Nevertheless, the situation was probably not that problematic. The pilgrims going to Dodona, a sometimes long and quite expensive trip, were probably not the most illiterate of the Greeks. Less rich and less educated people could have delegated their question to another pilgrim going to Dodona for his own business.Footnote ⁴⁴ It was always possible for a pilgrim to find a fellow pilgrim to write the question. This could partly explain the oddness of many lamellae, in which, for example, some features of one dialect are mixed with another dialect; the writer would have tried to adapt himself to another dialect but made some mistakes. Sometimes the question on one side and the ‘abstract’ on the other are written by the same hand (as far as we can tell),Footnote ⁴⁵ sometimes not. This could be explained by the fact that a first person helped to write the question and, later, a second kind soul engraved the name and the abstract. It is also possible that a local official helped.Footnote ⁴⁶ As regards now the members of the sanctuary, were they able to read the questions of the consultants? Usually, priests, prophets and others officials, including women, came from elite classes. Their potential illiteracy would, in any case, not have been an issue, because the consultants, as we shall see later, probably had to pronounce their question aloud.

In conclusion, it is unlikely that illiteracy would have prevented pilgrims from fulfilling their oracular duties. Therefore, we could infer that (il)literacy was no big deal. It also means that any clues regarding the lamellae’s style of letters or dialect must be interpreted with some caution, since the writer would not necessarily have been the consultant. So any conclusions based on statistics concerned with what was written on the lamellae may be biased, and this more clearly demonstrates how these texts are fraught with uncertainty.

What Information Can We Draw from the Inscriptions Themselves? Some Basic Elements

The Dodonaean questions themselves, then, are our last, and also least, resort, since these are entirely open to interpretation and can spur the fiercest debates.

The god’s potentially orally-given answers to the questions on the lamellae were not nearly as important as what these questions themselves represent. Although the lamellae were left on site at the end of the process, we do not know if these lamellae were true dedications. They could play such a role, as a human gift at the end of the oracular procedure, an offering that honored the god, and would have reinforced sanctuary traditions.Footnote ⁴⁷ Moreover, as we will see later, certain lamellae left at the sanctuary were obviously not questions, but dedications, prayers, or calls to the gods to witness.

Answers could have been written, on lead or on other perishable material, which the consultants, when necessary, took with them when they left the place. The lamellae were so numerous, throughout time, that we would expect to have found at least some oracular lead lamellae elsewhere in the Greco-Roman world. Up to this day, however, no written answer or question from Dodona has been found outside the precinct of Dodona. In fact, writing the answer would have been pointless. Many of the private consultants came with a simple question, ‘Do I have to do this or not?’, and they did not need a reminder to remember the yes/no given by the god. Things were a little bit different when a consultant was someone else’s delegate, but it is not necessary to imagine longer or complicated responses. The questions, including the civic ones, were almost always answerable by yes or no.Footnote ⁴⁸ Written answers could be necessary if the oracle gave longer advice (e.g. for the ‘to which god’ questions).Footnote ⁴⁹

One question asked by Jessica Piccinini concerns potential traces of orality in the text of the lamellae. Are there ‘internal hints that the tablets were elaborated in writing only after the inquiry was expressed orally’; or ‘do the tablets contain extemporaneous expressions and, for example, words of mouth, which indicate an unprepared, unpolished and spontaneous words flowing as in any direct communication’.Footnote ⁵⁰ By her own admission, it is difficult to determine this with such very short texts, and she sends us back to considerations of literacy.

Most of the questions at Dodona are written in a definitely non-literary Greek, maybe closer to the spoken language of the consultants.Footnote ⁵¹ The usual Panhellenic formulas: ‘Is it better to do this (or no)?’, or ‘To which gods do we have to pray to get this?’, were usually followed by some words or a small sentence, which defined the question. Therefore, the typical question at Dodona is a mix of widespread formulaic habits and the presentation of the reason for the consultation. There do seem to be frequent verbal slips, but it is hard to tell how oral or literary they were. Some consultants seem to have been in a hurry; yet we do not know whether this haste is the result of external circumstances or the disorganization of some pilgrims – or even if that interpretation is correct.

The Role of the Lamellae during the Consultation

If we tentatively consider the lamellae as ‘offerings’ or something like them, did they play any role during the consultation itself? The functions are not mutually exclusive; indeed, they could be complementary. As regards the procedure, we can only speculate. The following proposals would work with inspired divination (divine inspiration in the prophet’s mind) as well as induced divination (divine clues through signs), whatever was the true Dodonean mode of revelation.

(1) Maybe the procedure was fully oral and direct, where the consultant went directly to the oracle, asked the question and, orally or sometimes by means of a technical process (e.g. lots, see below), got an answer. In other words, he/she asked the ‘prophet/ess’, who answered him/her. A lamella could be required from the beginning (e.g. because such is the habit of the place), or produced at the end in order to leave it as an ‘offering’ as well as a memory of the consultation.

(2) Or the procedure was fully written (as it seems to have been the case in Thessaly at Korope),Footnote ⁵² where the consultants were required to write their enquiry on a lamella, then gave it to an official (be it the ‘prophet/ess’ or any intermediary) who got the answer in the appropriate way. The intermediary would then deliver a written response to the consultant’s initial written question. A lamella was mandatory from the very beginning and was left at the end as a testimony of the consultation.

(3) Or else the procedure was a mix of (1) and (2), to varying degrees.

The uncertainty of these procedural proposals might be solved by returning to the lamellae themselves. Those lamellae that have a full question, either on its own or on the obverse with an abstract (e.g. ‘about a woman’ DVC 1010B, ‘about a child’ DVC 3666A) and/or a name (e.g. ‘Apellys’ DVC 85)Footnote ⁵³ do not help, since any of the proposed procedures could apply. More revealing are the many lamellae that only have a name or an abstract instead of a main question. If these represent the written question, then the prophet/ess would necessarily have had to have interpreted a great deal from the laconic lamella.

If, on the one hand, the written question has simply disappeared from the lamellae through damage or reuse, that could still fit our second proposal of a written exchange. A better understanding of whether that might have been the case would require a full autopsy of the evidence at the museum of Ioannina. Even so, since the majority of the lamellae are irremediably damaged, no further elucidation is reachable. If, on the other hand, the questions were never written on those lamellae, then it might mean that there was no need for them.

Some of these shorter inscriptions might imply that a complete question was not mandatory and the prophet/ess could infer a full question. For example, from ‘Agathon about children’, they might infer a question such as, ‘Will Agathon have children (with X)?’ or ‘Which god has Agathon to pray to in order to beget children?’ Therefore, such a lamella might satisfy both the first and second proposals, in terms of being oral or written.

When it comes to names only appearing on the lamellae, the first procedure becomes much more possible than the second. With the first procedure, the name is enough because the consultant will ask his question aloud, but the second, written procedure becomes impossible, because the prophet/ess would have had no verbal contact with the consultant and would never have been able to guess the question. So very short inscriptions, from names to short abstracts like ‘Is it?’, or ‘Isn’t it?’ (e.g. DVC 1040A), or ‘sacred’ (hiaros; e.g. DVC 3964B) make a fully written procedure impossible.

Our speculation has not been useless: we can propose a reasonable framework, although its details remain inaccessible:

1. (1) A question written on lead does not preclude the possibility that it was pronounced aloud; but a question reduced to one name or two words, or an abbreviation in writing, proves that some questions must have been spoken out loud. In the only Greek example where we hear about written questions sorted by lot from two sealed vases, the question to the Pythia was nevertheless asked aloud, as it is known from a well-preserved Athenian decree, corroborated by literary sources.Footnote ⁵⁴

2. (2) As orality is far more prevalent in Archaic and classical times, we could conclude that the majority of enquiries were made orally, and that at least part of the encounter with the sanctuary official(s) and the god was conducted orally.

3. (3) A lamella seems to have been part of the local practice, a bit like the pinakes in Trophonios’ oracle. At Lebadeia, the pinax came at the end of the process because its purpose was to illustrate the answer of the god (which possibly also required the writing of the question). At Dodona, only the question matters, and the time at which the lamellae was written during the consultation is simply unknown. Some pilgrims could be very diligent and write their question immediately, even if they were going to pronounce it aloud. Others will have been less careful, writing only the essentials, or nothing. Here, however, an important emotional factor might have come into play, given that the pilgrim was approaching the divine, and possibly expecting a moment of thambos (‘awe’).Footnote ⁵⁵ The written question could provide assurance that one would not forget the question once in front of the intimidating prophet/ess, during a high-intensity ritual. This moment does not appear tremendous for us, but for the Greeks, it definitely was. Zeus was the chief god, and the pilgrims nearly all made a long journey to ask their question. An oracular consultation is often close to an epiphany after a long pilgrimage.Footnote ⁵⁶ We cannot reject the idea that the lamella was only an ‘offering’ at the end of the visit, which was fully or mainly oral, but it would be surprising.

Some More Aspects: Isolated Letters and Double Display of a Quasi-Identical Question

Some lamellae display a letter.Footnote ⁵⁷ Since the archiving theory has been discarded, this letter is usually interpreted as the number of each consultant waiting in line for the oracle.Footnote ⁵⁸ If this is true, it provides another practical use for the lamellae, which would become indispensable at the very first stage of the consultation. For this supposition to be accurate, however, a number would have been written for each question, and many numbers would be recognizable on each palimpsest lamella. The sanctuary’s officials (e.g. the neokoros, the warden) never knew how many people would come during the day, except if the registration for the consultation was fixed at a given time, or by appointment. Anyway, all the lamellae without a number (the overwhelming majority, actually, 4153 of 4268) are then problematic.

A few lamellae display almost exactly the same question (e.g. DVC 2005A-2006A), and this could be considered as a case of more accurate rewriting. We can explain it according to our main two options. The consultant may have tried to be as precise as possible before asking the god, and thus made a second attempt at phrasing the question. This would be the proof that writing the question before the consultation was important. But the consultant, after asking the god, could also have written his/her question as precisely as possible, at the moment of the ‘offering’. There is no reason to reject the possibility that each consultant acquired a lamella at the beginning of the procedure, and wrote on it whenever he decided to do so. On that point, I note here that we have no proof that the questions were written verbatim, whether at Dodona or any other Greek oracle. This is obvious for the lamellae presenting only a name or an abstract: the name ‘Apellys’ is not a question you ask Zeus. For complete questions however, slight differences could surely be possible, as long as the message was respectful of the charis (mutual kindness/respect) owed to the gods. I will return to these almost identical questions later in this essay, because they could also be the sign of a lot-based procedure.

Abecedaries: Another Mantic Medium?

Seven inscriptions found among the lamellae were abecedaries, that is, the engraving of an alphabet. Some of them were probably complete (although today some letters are missing). Others appear to have been partial at the time they were engraved, and the question of whether they were originally complete or were left incomplete cannot be answered, because the lamellae are so badly preserved.

Complete abecedaries:

The engraved alphabet follows the rim of the lamella, with a bend at a right angle at the end of the text. The three first letters are missing on the left.Footnote ⁵⁹

An isolated one-legged Π appears just below Σ and Τ, and just beside it, if it is really a letter and not a scratch, a dotted koppa, Ϙ (2278B).Footnote ⁶⁰

(Maybe) incomplete abecedaries:

With an intentional (and extremely rare) ligature between Δ and Ε, that means that the two letters are joined (). On the same face, Διώνα[ν (Diona[n) could show another ligature between Δ and Ι, but this one could be incidental.Footnote ⁶¹

Both sequences of letters are on the same lamella, obverse and reverse; they could be from the same hand. As the lamella was turned over on its horizontal axis, we do not know how much of the object has been lost, because both faces begin with A on the left.Footnote ⁶²

As this sequence of letters is not attested in any Greek word, this is surely the fragment of an alphabet. The reading of this sequence depends on the Greek alphabet it is written in, but Lhôte in CIOD is in favor of the Thessalian alphabet.Footnote ⁶³

There are 3 cm left after the E, enough to write more letters, but nothing more appears on the lead.Footnote ⁶⁴ If the person who wrote Theodotos also wrote the alphabet, this means that he was able to write more than a few letters.

The meaning of abecedaries in such a context is not an easy issue. We can first think of the difficulty of writing on such lamellae. Lead is a ductile metal, but its surface is not like papyrus, stone or terracotta. I myself have tried to write on lead with a long sharp nail. To say the least, the results were not artistic, despite the fact that I am not totally clumsy. Literate people were not used to it, and thus consultants needed a bit of practice, the more so because many lamellae were very small, not exceeding a few centimeters across. Writing in very small characters renders the exercise more complicated, especially at a time when glasses did not exist. For some lamellae, could it have been that they were written before the lead was cut into a strip?Footnote ⁶⁵ The best way to learn the skills would have been to write some lettersFootnote ⁶⁶ or a name, or an alphabet, complete or not. The five letter As we find on DVC 910 A could be a trace of such a practice.Footnote ⁶⁷ One could engrave a full alphabet, or just the beginning of it, or some sequence of choice. The writing of some names (but a tiny minority) could also be indicative of such a fast training in engraving on lead.

Other scholars may say that mastering the alphabet was a skill worthy of being offered to the gods. At the sanctuary of Zeus on Mount Hymettos, for example, many abecedaries and inscriptions (of which the writers were proud) from the seventh century BCE have been brought to light.Footnote ⁶⁸ One can wonder, however, if this newly acquired competence in the high Archaic period kept its special attraction later in time, and it would be too easy to argue that Dodona was only a backward area.Footnote ⁶⁹ Anyway, the very idea of an offering of letters is no more than a guess. Still others may consider the abecedaries as a sign of functional illiteracy. The person who was not able to write nevertheless flaunted his ability to write the alphabet, or at least some letters.

So, offerings or not? Things are not simple, because Greek culture was not prone to clear-cut differentiation: just because Dodona is an oracle it does not follow that its activity must be only oracular. An oracular sanctuary is, first, a sanctuary, open to anybody eager to get closer to the gods, with no obligation to consult Zeus and Dione. Other ritual behaviors were welcome, like vows or prayers, which are similar to an oracular process. In Dodona, it means that nonoracular declarations and oracular questions could end up on the same medium, a lead lamella. Furthermore, the abecedaries would not be the only ‘offerings’ left in Dodona’s precinct. Some texts on the lamellae are obviously not questions asked to Zeus. For example, DVC/CIOD 2482BFootnote ⁷⁰ is a declaration about a loan; DVC/CIOD 2653 A about concerns regarding a house (?); DVC 34B and SGDI 1596–1597 (= Eidinow Reference Eidinow2007: 122, no 1–2) are clearly prayers.

In the context of oracles, an abecedary could act as a written question for poorly educated consultants, a bit like marking with a cross in order to sign a document. In this case, the god, who knows everything, would directly read in the alphabet the question orally asked to him by the consultant. We cannot discard this explanation, which could support our hypothesis of an oral consultation. The consultant would have enquired aloud while writing down just what he was able to write.

Abecedaries could also assume a more mantic meaning. What would be the point of presenting a fragmentary alphabet to Zeus, up to gamma or epsilon for example? I wonder if those abecedaries hid a particular type of question. Letters used as numbers would refer to potential mantic answers – not specified on the lamella for lack of space but asked orally, and written on a separate document by the consultant. For example, the five letters in DVC 1709B, ΑΒ̣ΓΔΕ, would give Zeus the opportunity to choose between five proposals. This kind of formulation would prove efficient for drawing lots between several pre-set answers, but an oral answer would be just as effective.

DVC’s edition foregrounds at least seven lamellae mentioning, beyond any doubt, a procedure of drawing lots.Footnote ⁷¹ The casting of lots could also better explain the meaning of other inscriptions,Footnote ⁷² thus at least partly acknowledging Cicero’s story about a monkey that turned over the vase and scattered the lots before a Spartan consultation.Footnote ⁷³ Not all the questions in Dodona were answered by lot,Footnote ⁷⁴ but some procedures involving drawing lots were obviously in operation.

What is true for partial abecedaries could also be true for the complete ones: each of the twenty-four lettersFootnote ⁷⁵ would correspond to twenty-four numbers on a (now lost) list, containing names, proposals/solutions, gods and heroes to pray to, and so on. This could explain the isolated numbers found on many lamellae, usually interpreted as the number of each consultant ‘in line for the oracle’. A set of twenty-four potential answers is huge, but under the Roman Empire, some dice oracles included verses extracted from Homeric poetry and its twenty-four ‘books’.Footnote ⁷⁶ It also brings to mind the famous Delphic consultation of Kleisthenes around 508 BCE. According to the Aristotelian Athenaion Politeia (21.6), the legislator asked the Pythia to choose a hero for each of his ten new Attic tribes, from a list of a hundred suggested names. The historicity of the story is not certain, but in any case, at the end of the fourth century, this kind of oracular process looked plausible to a peripatetic philosopher,Footnote ⁷⁷ and ultimately shows that our suggestion is not without sense.

But such a twenty-four-answer inquiry was surely not the rule. The usual scenarios might have involved two to five/six different suggested answers, and the most frequent case would have been the A or B, or the A and non-A alternatives. This would explain the huge number of As (thirty-one) and Bs (seventeen), far ahead of other letters,Footnote ⁷⁸ for inquiries up to five pre-set responses. This idea does not work, however, if the two- and three-letter groups are taken into consideration.Footnote ⁷⁹ At this point, no definitive explanation is possible, but it is nonetheless worth exploring potential solutions, in order to encourage further research.

Back to Lamellae in a Lot Oracle Context

The habit of writing, on two lamellae, the two alternatives of a question, A and non-A, is not unknown in the Greco-Roman world. In Egypt, this was the principle of the ‘ticket-oracle’:Footnote ⁸⁰ consultants would write the same question on two tickets, not necessarily word for word, the first one in the positive mode, the second in the negative, each one ending with the same formula, ‘give me this one’. Then, after the process of drawing lots, whatever it was, the consultant received a ticket back, requiring him to fulfill or to abandon his plan. In mid-fourth-century Athens, a decree follows the same process for an upcoming Delphic consultation. The Assembly write on two distinct tin lamellae the positive and negative versions of the same question, about the cultivation of an area of Eleusinian land sacred to Demeter and Kore. Each lamella is to be rolled up, wrapped in wool, then cast in a bronze vase that will be closed and shaken. Then one of the rolled lamellae will be taken out and put into a silver vase, the second one in a vase of gold, both of which will be sealed. This being done, the Assembly will send a delegation to Delphi, where the Pythia, when asked orally about cultivation, will give Apollo’s response by designating the silver vase or the gold one.Footnote ⁸¹

In Dodona, examples of an explicitly duplicated question are exceptional. Most of the enquiries, as we have said, imply a closed answer of yes or no: ‘Is it better and more advantageous to do X?’ This type of question, an alternative between A and not-A, is in its principle close to cleromantic divination, which works whatever was the operating mode of the sanctuary, be it by lots or through ‘inspired’ or ‘inductive’ answers.

Could it be that these duplicate lamellae are the sign of a general ‘ticket-oracle’ mantic mode at Dodona? The answer is no. This would mean that each lamella found around the seat of the oracle was the one that was left in the sanctuary, while the second would have been thrown away, or re-melted, or carried away by the consultants. In any case, there should be a quite even proportion of positive and negative questions, and this is not the case.Footnote ⁸² The hypothesis then is not valid and, consequently, the thousands of lamellae found in situ may not have been used for operational cleromantic purposes. In other words, the majority of Dodona’s lamellae were not used in any lot device.

This does not mean that the method of the ‘ticket oracle’ did not exist at Dodona. For orally asked questions, it is possible that an oracle official could have operated a cleromatic device, in a kind of black-or-white bean method, to generate an answer to the many enquiries that present simple alternatives. As there is no mention of any ‘black bean system’ in the lamellae, however, and given the fact that an oral answer was perfectly conceivable and in accordance to Greek mentalities, it remains safer to consider that answering by lot was probably not the usual method of divination in Dodona.Footnote ⁸³

Some exceptional lamellae, however, display a duplicated text that is strongly reminiscent of the ‘ticket-oracle’, as has been made clear by Robert Parker. The question is whether those lamellae themselves actually played a role in a cleromantic process, given that a lot device, of which we have no other trace besides these few lamellae, would have to return one of the two answers, yes or no. Speaking about DVC 2005 A and 2006 A (about the theft of a pig), the former positive, the latter negative, Parker specifies: ‘The editors suggest that the enquirer was dissatisfied with his initial formulation and tried again. But the switch from positive to negative suggests rather that both possible answers were written on the same lead strip, (…) and that in this case, for whatever reason, a separation leading to the return of one failed to occur.’Footnote ⁸⁴ I fully agree with the first argument. The slightly different formulation of the two texts refers to the same question, and in that respect, is close to the ticket-oracle system. But I remain skeptical about the second part of the argument, because the lamella has not been cut into two strips, and could never have been used in any draw procedure.

Three other lamellae provide duplicated texts written on the same strip of lead either head-to-tail,Footnote ⁸⁵ or following each other,Footnote ⁸⁶ and only once written on each side of a lamella.Footnote ⁸⁷ Their questions are close to the ticket-oracle type, but not in a way that could be used as such, because none of them was cut in half. We will never know how such types of question were asked nor how the oracle answered. It could have been orally, by stating A or non-A, for example; or by drawing lots, using a ‘black or white bean’ technique to reach the same result. Again, however, this small step in the understanding of the operation of the oracle does not help us pin down the exact moment when those lamellae, close to ticket-oracles, were written down. Consultants could have written them at the beginning, middle or end of their consultation, and not all consultants would have written them down at the same point.

This has been a long enquiry about a simple problem. I am afraid I have not delivered many strong conclusions, but knowing the limits of our knowledge is better than building brilliant ideas on a bed of sand. Any type of evidence is more complex than it looks at first sight, and more difficult to understand, but it is also richer in perspectives and full of hints leading to other questions.

It is probably pointless to try to establish the canonical way of consulting Zeus at Dodona – and perhaps there is no straightforward answer. Oral questions seem to have been the rule, like spoken answers; the existence of a lot procedure, however, is also attested, which does not preclude orality in the process. Indeed, many mantic procedures could have been in use at Dodona, simultaneously and/or successively, for its thousand years of existence.