Delphi: What the Pythia Actually Did

Before anything could happen, somebody had to check whether the god was in.

The test was a goat. Cold water was poured over it, and the animal was watched. If it shuddered, and not a twitch but a full-body tremble from the hooves up, Apollo was present and the day's business could proceed. If it stood there, everyone went home. Plutarch, who was a priest at Delphi and had seen this done, describes an occasion when the staff pushed ahead despite a bad omen and the Pythia became distressed and later died.

That is the ancient oracle of Delphi at ground level: a procedure, with a pass/fail gate at the start, staffed by people who knew what a bad day looked like.

The picture in your head is a painting

Most of what we all half-know about the Pythia comes from nineteenth-century art and the writers who fed it. A young woman half-clothed and contorted on a tripod, hair loose, breathing smoke from a crack in the earth, gasping out nonsense while robed men lean in to convert it into hexameters.

Nearly every element of that is either late, hostile, or invented.

The ancient evidence is thinner and stranger. Our fullest description comes from Plutarch, writing around 100 CE, which sounds excellent until you notice he was describing an institution roughly eight hundred years old and visibly in decline. He wrote an entire essay wondering why the oracles had gone quiet. He's a good witness to late Delphi. For classical Delphi, at its height, he's a man reporting on his own workplace's better days.

Nine days a year

The first thing that reorganises the picture is the calendar.

Delphi was not open. In the earliest arrangement the oracle could be consulted on one day a year: the seventh of the month Bysios, held to be Apollo's birthday. Demand later forced it up to the seventh of each month, but only for the nine months Apollo was considered resident. In winter he was away and Dionysus held the sanctuary, and there were no consultations.

Nine days a year. For the most authoritative religious institution in the Greek world, serving city-states across the Mediterranean.

Which produces the second thing: a queue, and a market in it. The right to consult first, promanteia, was a formal privilege, granted by Delphi to cities and individuals as an honour and recorded on stone. If you had done the sanctuary a service, you skipped the line.

There was also a fee. Consultants brought a pelanos, a sacrificial cake, whose price was fixed and published, with different rates for a city and a private person. Inscriptions preserve the tariffs. On top of that came an animal for sacrifice, which had to meet standards of its own.

None of this is a scandal. It's how a sanctuary with nine operating days and international demand rationed access, and it tells you the oracle was, among other things, a functioning organisation with revenue.

What she actually said

The most useful thing anyone has done with Delphi is a catalogue.

Joseph Fontenrose collected every oracular response attributed to Delphi in ancient literature and inscriptions, and sorted them by how well attested they are: whether a response is supported by contemporary evidence, or turns up only in a story told long afterwards, or belongs to legend outright.

The pattern that falls out is striking. The best-attested responses are overwhelmingly plain.

They approve or refuse a proposal. They pick between named options. They answer a question of the form "to which god should we sacrifice in order to get this result?", which was, apparently, one of the commonest things anyone asked. They sanction a colony, a cult, a change in practice.

They are not, on the whole, riddles.

The famous ambiguous oracles cluster in the other category. Croesus being told that if he attacked Persia he would destroy a great empire, and destroying his own: that reaches us through Herodotus, writing generations later, in a narrative built to show what happens to a man who mistakes his own certainty for a god's. The "wooden walls" told to Athens before Salamis is a better-supported case, but it comes with the same problem: we get it from a historian assembling a story about a war whose outcome he already knows.

A perfectly ambiguous prophecy is a wonderful thing to have in a narrative and a useless thing to receive as a client. The oracles that survive because they made a good story are not a representative sample of the oracles that were given.

A clearing house that everybody trusted

So what was it for, if it mostly answered yes?

Mainly, sanction. A Greek city considering something irreversible (sending several hundred citizens to found a settlement on a distant coast, adopting a foreign cult, altering an ancestral rite) needed the decision to be more than a vote that a faction could reopen next year. Delphi supplied that. The god had approved it.

Underneath that, there was something more practical going on. Delphi received questions from every corner of the Greek world, year after year, from people planning colonies. The sanctuary had, by accumulation, better information than anyone about which coastlines were settled, which were hostile, and which had failed. When it told a city where to plant a colony, the answer had a real evidence base behind it, even if nobody described it that way.

And the answers were remembered because they were recorded, in the way Rome kept its own oracular collection under lock and key. That's the ordinary mechanism behind an institution's authority: not that it was always right, but that it was consulted formally, answered formally, and kept.

The woman on the tripod

The Pythia herself is the part we know least about, which hasn't stopped anybody.

She was a local woman from Delphi. Not from a priestly dynasty as far as we can tell, not selected for education, not trained in verse composition. Diodorus reports that the role was eventually restricted to women over fifty, who dressed as girls, after a consultant abducted a young Pythia, and he's the only source for both the rule and the reason, so the story deserves a raised eyebrow.

What she plainly did was sit on a tripod in the inner chamber and answer questions from people who had queued, paid, sacrificed and been vetted. Whether she was in an altered state, and if so what caused it, is a genuinely open question, and one a geologist and an archaeologist tried to settle with chemistry, with mixed results.

The rest of the apparatus we can describe: the days, the fee, the priority list, the goat. An oracle that produced binding decisions for the Greek world ran on nine mornings a year, and started each of them by watching an animal to see whether it shivered.