To the Editors:

Peter Green finds much that is valuable in The Delphic Oracle (NYR, April 5), but disagrees on central issues, primarily on what responses reported for the archaic period we can consider authentic. The matter boils down to this: can we accept the responses reported by Herodotos as genuine pronouncements of the Pythia? These responses are mostly very different from the Historical responses (those that have contemporary attestation). No verifiable response of any oracular establishment, past or present, is anything like those that Herodotos reports for Delphi; it is always more like those that I classify as Historical. Therefore Green and others who want to “save the appearances” must demonstrate that the oracles which Herodotos quotes are authentic, in substance if not in form. Inspirational talk about archaic Delphi won’t do; we must have evidence that is cogent, not circular argument. The case for trusting Herodotos’ reports cannot be made from those reports alone. That is what Green does. He asserts two crises in response to which the Delphic Oracle changed its modes. First it gave Croesus assurance of victory, and he was defeated. Thereafter the Delphians considered the Persians invincible and so discouraged opposing them when Xerxes invaded in 480. But the Greeks defeated the Persians at Salamis and Plataia. So the Delphians gave up prediction as a bad job and spoke only safe directions and statements. That is, Green deduces the crises of the Oracle from Herodotos’ accounts and then uses the crises to verify the accounts. Furthermore we should notice that ten years before Salamis the Athenians and Plataeans routed the Persians at Marathon. Why did the Delphians continue to think them invincible?

To speak of “three centuries of Delphic involvement in city-state affairs” (the archaic period) is simply to reassert the conventional view. How much “involvement” is demonstrable? Even in the Quasi-Historical responses Delphi seldom takes the initiative. In reality, men went to Delphi for sanctions and directions, believing in the Pythian Apollo’s special authority at his sacred shrine. Zeus’s Oracle at Dodona was as highly esteemed as the Delphic throughout Greek history, although consultants received only yes/no answers or the name of a god to sacrifice to, determined by lot or a similar device.

Professor Green says that in respect to the Pythia’s frenzy, vapors, and chasm I am flogging dead horses. He has misapplied the metaphor: a flogger of a dead horse is trying to make the horse get up and go, not punishing a horse that he thinks alive. Green asks, “who, today, would argue that [the Pythia] was [seized with a frenzy]?” Well, look at Robert Flacelière’s Greek Oracles (1965), p. 50: “All our sources confirm that when the Pythia was prophesying she was in a state of frenzy.” None of them does (unless one says that Lucan confirms this), but Flacelière has written on Delphic subjects for forty years, and he mentions “noxious fumes” on the same page. Not only that, he says in the same book that “Phoebus Apollo…was also the sun-god” (p. 35). Green says that I seem “anxious to convince [readers] that Apollo was not a sun god” (also a dead horse), although I mention this matter only incidentally in the preface as an example of outdated beliefs that persist even among classical scholars. And it does persist.

Contemporary scholars like Dodds and Parke attribute possession or trance, if not frenzy (and this is a product of possession), to the Pythia, although Plutarch, a Delphic priest, says plainly that Apollo did not speak with the Pythia’s vocal chords: she spoke in her own words, reflecting the truth that Apollo revealed to her. Green apparently accepts possession, as well as a physical pneuma and a chasm (although he asserts that “advocates of rocky chasms…have been few and far between”): on his inspirational first page (12) he mentions a chasm (“the omphalos was yin to the chasm’s yang,” whatever that means—it sounds fine), and the hole over which the Phythia sat (a figment of fantasy).

I was careless when I set Dodds beside Rohde for the idea that the Dionysiac cult at Delphi inspired the alleged mediumship of the Pythia. There my topic is her supposed possession, and I should not have run it together with Dionysiac influence. I quote Dodds: “…Apollo’s Delphic utterances are always couched in the first person, never in the third.” I then show that in many responses the Pythia refers to Apollo in the third person: Dodds made an incautious statement. This was my criticism of Dodds, not what Green says it was. Dodds does distinguish between mantic and telestic (Dionysiac) mania, but the possession that he posits for the Pythia is Dionysiac; it is not the mantic manikê of Plato, which is not so described and is largely a play on words.


In fact, Green often misrepresents what I have to say or credits me with something that I have not said. For example, the reader of his review will suppose that I have imputed fraud and cynicism to the Delphic priesthood. In fact, I make it clear from the outset that I do not; I never mention it and it should be plain that I consider fraud unnecessary for the sort of response actually spoken.

It seems that I am a rationalist, showing “a fine rationalist frenzy,” but after reading the review one may ask, “Who’s in a frenzy now?” Green’s is an obscurantist frenzy, an outburst of that Delphic piety of which I speak in my preface. In his remarks on my objective criteria, which he denies are objective, he confuses “historical,” as I define the term, with “authentic.” An Historical response, in my classification, is simply one spoken within the lifetime of the earliest authority who reports it.

Green ends up granting that Herodotos’ oracles may not be the Pythia’s ipsissima verba, though he starts by accepting them as quoted. When all is said, he does no more than reassert the conventional view of the Delphic Oracle’s utterances and operations. He does not come to grips with the evidence that I present for my conclusions; and I am afraid that a reader of the review will not learn much about the book’s content. I expect more of the same kind of thing. As I say in my preface, “I am aware that my argument, however well-founded, will not prevail against the will to believe.”

Joseph Fontenrose

Department of Classics

University of California, Berkeley

Peter Green replies:

Professor Fontenrose’s categories are a kind of card-forcing. His “Historical” category, however much he may hedge his bets over authenticity, carries a specious air of trustworthiness; his “Quasi-Historical” category, despite carefully noted exceptions, does not. His “Legendary” grab-bag is even vaguer, still more dependent on “common-sensical” assumptions. Plenty of scope for subjective judgments here. If he were sorting, say, the Alexander-historians, every single one of our surviving sources would get the “Quasi-Historical” label, since “Historical” for him simply means “attested by a contemporary witness.” Since few contemporary references to the Pythia’s operations survive from the period 750-450, whereas those centuries offer plentiful material that must, in Fontenrose’s terms, be described as “Quasi-Historical” or “Legendary,” it is not hard to see how, using this model, much historical (small h) evidence can be discredited simply by its formal resemblance to these two latter categories. In other words, we have a built-in presupposition of spuriousness for the period of Delphi’s greatest influence and prestige.

Fontenrose takes me to task for not using the word “historical” (non-capitalized) in his own outré sense. But why on earth should I? The only way that Fontenrose can assert that “the Historical are commonplace directions and statements” is by setting up a model that relegates anything else to the “Quasi-Historical” or “Legendary” bins: this is where circularity creeps in. For Fontenrose, to demonstrate that the Herodotean oracles are authentic means conforming to the guidelines he has set up. Of course Herodotus is our only direct source for the Persian War oracles; what we have to do is to estimate the likelihood of their genuineness against a broad background of historical probability, not argue their invalidity on the basis of categories we ourselves have created. (Incidentally, I defy Fontenrose to point out a passage in my review where I actually accept the Herodotean oracles as quoted: the gist, in prose or verse, is what matters, a distinction for which the Troezen Decree provides illuminating parallels.) Fontenrose wants to cut Delphi loose from socio-historical developments altogether because this is the only way his arguments will work. If he wants to know why the Delphians continued to think the Persians invincible after Marathon, it is, clearly, because they could tell the difference (as Fontenrose seemingly can’t) between a minor campaign and a major invasion.

What I or Fontenrose may believe about the validity (as opposed to the modus operandi) of Delphi is strictly irrelevant in historical terms: what does matter—as I emphasized in my review—is the attitude of the Greeks themselves at any given time. Here we have a distinction of prime importance, which Fontenrose constantly confuses. What bothers him, seemingly, is the supposed “will to believe” among modern scholars rather than the degree and nature of belief at various stages in antiquity; and behind that there lies the simpliste rational objection that because oracles are superstitious nonsense they couldn’t possibly have “worked.” Whether we allow that last inference or not (I’d be inclined to argue it), all this is strictly beside the point. For the record, I believe in mantic possession (as did everyone in the ancient world) but not in gibberish; in the pneuma, but not in physical vapors; in a hole, but not in a rocky chasm. The pneuma (see below) was a quasi-physical conceptualizing of divine possession, while Georges Roux (Delphes, pp. 110-117) has shown that some kind of symbolic chthonian hole may well have existed under the tripod in the adyton.


To what extent Robert Flacelière qualifies as a “serious scholar” in the field of Greek religion is a matter for debate; certainly he can’t be blamed for the laxness of his translator. What he actually wrote (p. 71, French ed.) was: “Le délire de la Pythie est attesté par toutes nos sources,” and goes on to make it quite clear (cf. p. 51) that he is referring to Platonic mania, the utterance of enthousiasmos, a common meaning for délire: no frenzy, no gibberish. “Frenzy” may indeed be a “product of possession,” but is not (as Fontenrose implies) its inevitable consequence.

What Plutarch (Moral.397C, 404B, 414E are presumably the passages Fontenrose has in mind) is at pains to emphasize, as H.W. Parke (Greek Oracles, p. 79) clearly saw, is that Apollo did not occupy the Pythia’s body in any vulgar physical sense: Parke comments, accurately (a point I also made in my review), that “the ancient world found it extremely difficult to imagine a spiritual force.” Fontenrose’s quibbling over Dodds’s definition of mantic possession speaks for itself; and the reader of my review must be left to decide whether it gives the impression that Fontenrose “imputed fraud and cynicism to the Delphic priesthood.” On the one occasion where I discuss Fontenrose’s arguments in this context (p. 14, col. 2) I emphasize his total rejection of such a theory.

Delphi was there for consultation: “involvement” does not of itself imply initiative. That is as true of the Persian War oracles as of any other: the Athenians, for instance, wanted Delphi’s opinion, and got it. Like Dodona, like Didyma, the Pythia doubtless did issue a majority of responses that were run-of-the-mill; but this is no reason for treating the minority of significant exceptions as spurious. Here Professor Badian’s line of approach strikes me as far more useful. As he says, people who want to believe will, under normal circumstances, believe regardless, and can cheerfully accommodate, or explain away, any number of false responses: but if they are confronted with blatant, and public, political manipulation, then at the least they will stop asking religiously inspired political questions. I would argue, myself, that by the time of Philip’s démarche with the Chalcidians, politicization was already well advanced, so that consultation had become conscious propaganda rather than in any sense an act of faith; the same could just be true of the Thebans before Leuctra in 371 (Pausanias 4.32.5), though they did have a long-standing record of reliance on the Pythia in military matters, which persisted even after their annihilation by Alexander.

Professor Badian’s terminus ad quem, then, may be updatable (I suspect the truth to lie somewhere between his date, after 356, and that advanced by Parke, c. 431), but methodologically his arguments are of the greatest value. However, we still need a historical analysis of Delphi’s changing role in the affairs of the Greek city-states between the eighth and fifth centuries BC. Because of its preconceptions, Professor Fontenrose’s monograph not only fails to provide this, but in some respects is liable to prove a stumbling-block for future research.

This Issue

September 27, 1979