Phoebus Apollo, the sun god, had a good eye for an inspiring location. According to the Homeric Hymn to him (not by Homer and probably dating from the sixth century BC) Apollo himself chose Delphi as the dwelling for his oracle. In order to establish it there, he first slew a resident serpent, Pytho, which had hitherto protected local deities. Apollo now replaced them. On account of his triumph, his oracle was known alternatively as the Pythian; the priestess through whom he spoke as the Pythia.

The site was not practical: it was not easily accessible to pilgrims who, over a millennium, came from the major cities of Greece and further afield to consult the oracle. It is remote, located on a mountainside below Mount Parnassus. It is prone to earthquakes and rockfalls. But it seems almost luminous in sunlight – appropriate for the temple to the sun god. For these very reasons it is awe inspiring.

The Homeric Hymn shows that the oracle’s cult was already well established, with a foundation myth, at an early date. As the site’s extent and grandeur reveals even in ruins, it became a major focus of devotion and remained so almost until it was suppressed by the Roman Church at the end of the fourth century AD. There are frequent references to it throughout ancient literature. The oracle might, for instance, be characterised as the principal dynamic of the Oedipus story. Sophocles’ Theban plays must often have been performed in the shrine’s amphitheatre, to audiences who had travelled there along the very road on which King Laius had been killed by the son whose fate the oracle foretold.

Because the oracle was consulted over such a long period about so many issues, public and private, we know a great deal about it. According to the historian and philosopher Plutarch, himself a priest of the shrine in the early second century AD, the Pythia was always uneducated and of local peasant stock. Her guttural utterances did not sound female. Perhaps because Apollo spoke through her, or perhaps because there were educated male priests such as Plutarch in attendance to interpret her responses, they were often recorded in hexameters – according to Pliny the Elder, the hexameter had been invented for this purpose. Hundreds of these oracular statements are preserved.

To what extent the Pythia was intoxicated by volcanic vapours issuing from a fissure beneath the temple and to what extent she could will herself into a trance without narcotic assistance remains a subject of controversy among scholars. The cult was of great interest to E.R. Dodds in his influential The Greeks and the Irrational (1950). He was strongly of the latter view as he sought, in an early exercise in interdisciplinarity, to apply lessons from modern anthropology and psychology to the phenomenon. But recent excavation has revealed that pockets of hallucinogenic gases are trapped in the rock beneath the temple – although the fissure recorded in antique accounts has disappeared, perhaps as a result of an earthquake. The ancient descriptions of the Pythia sitting on a tripod over the ‘chasm’ in the temple’s innermost sanctum, enveloped in rising fumes, are therefore rendered very plausible. (Dodds was sceptical; he discounted them because they were all late.)

The oracular pronouncements were not unsolicited prophecies, but responses to questions, often about prospective alternative courses of action, and many of them were not expressed in direct terms. They tended to be ‘Delphic’. Another name for Apollo was Loxias, ‘the ambiguous one’. The Pythia’s gnomic responses, in verse or prose, had to be interpreted by their recipients; hence one of the oracle’s mottos: ‘Know thyself’. They were usually compatible with various outcomes; their meaning would only be clarified when there had been an outcome. This flexible ambiguity
was not in itself a cause for cynical scepticism. But as early as Herodotus, there is a recorded instance of bribery to secure an advantageous pronouncement. The advantage was not of course that of the petitioner, but of one seeking to influence what petitioners were told. In this instance, the aim was to persuade Sparta to liberate Athens from Pisistratid tyranny, even though the Pisistratids were friends of Sparta. By subverting the process, Sparta was influenced to act contrary to its own interests.

The oracle became a major factor in relations between cities and in their internal political life. The reverence for the oracle was shared throughout Greece, by powers who were often at odds or even war with each other. Cities constructed treasuries on the site, and displayed outside them choice pieces of loot they had won in battle against each other or against barbarians. An intractable dispute between cities might be settled by turning to the oracle. In this regard, the oblique nature of the oracle’s pronouncements could be advantageous. Henry Kissinger did not invent constructive diplomatic ambiguity. Local representatives of each city not only dealt with pilgrims from that city but, being well informed about its affairs, could brief the priests who formulated the pronouncement. The difference between the oracle and modern fora for diplomatic settlement was that all parties were inclined to accept, even revere, the divine endorsement – though Herodotus reports that the Athenians were so alarmed by the Pythia’s first pronouncement before Salamis that they simply approached her a second time and secured a more favourable one.

With Roman annexation came plundering by Nero (who took exception to a disobliging oracle), bombastic restoration by Domitian and Hadrian, and the purloining by Constantine of a column of the monument to the victory over the Persians at Plataea in 479 BC. That he used it to grace the Hippodrome in Constantinople suggests that Delphi remained resonant for the first Christian emperor. The last extant oracular pronouncement, issued in response to an inquiry from Julian the Apostate, is poignant: ‘Tell the emperor that the Daedalic hall has fallen. Phoebus has his cell no longer.’

But even in ruins, Apollo’s earthly residence provokes wonder. And his oracle speaks to a range of current historical interests, some of which Dodds foresaw: religious belief, ritual, orality and gender. More than two and a half millennia on, we are unravelling its riddles still.

George Garnett is Professor of Medieval History at Oxford University, Fellow of St Hugh’s College and the author of The Norman Conquest in English History: Volume I: A Broken Chain? (Oxford University Press, 2021).