by E R Dodds
In most of their descriptions of dreams, the Homeric poets treat what is seen as if it were "objective fact."8 The dream usually takes the form of a visit paid to a sleeping man or woman by a single dream-figure (the very word oneiros in Homer nearly always means dream-figure, not dream-experience).9 This dream-figure can be a god, or a ghost, or a preexisting dream-messenger, or an "image" (eidōlon) created specially for the occasion;10 but whichever it is, it exists objectively in space, and is independent of the dreamer. It effects an entry by the keyhole (Homeric bedrooms having neither window nor chimney); it plants itself at the head of the bed to deliver its message; and when that is done, it withdraws by the same route.11 The dreamer, meanwhile, is almost completely passive: he sees a figure, he hears a voice, and that is practically all. Sometimes, it is true, he will answer in his sleep; once he stretches out his arms to embrace the dream-figure.12 But these are objective physical acts, such as men are observed to perform in their sleep. The dreamer does not suppose himself to be anywhere else than in his bed, and in fact he knows himself to be asleep, since the dream-figure is at pains to point this out to him: "You are asleep, son of Atreus," says the wicked dream in Iliad 2; "You are asleep, Achilles," says the ghost of Patroclus; "You are asleep, Penelope," says the "shadowy image" in the Odyssey.13
All this bears little resemblance to our own dream-experience, and scholars have been inclined to dismiss it, like so much else in Homer, as "poetic convention" or "epic machinery."14 It is at any rate highly stylised, as the recurrent formulae show. I shall come back to this point presently. Meanwhile we may notice that the language used by Greeks at all periods in describing dreams of all sorts appears to be suggested by a type of dream in which the dreamer is the passive recipient of an objective vision. The Greeks never spoke as we do of having a dream, but always of seeing a dream— The phrase is appropriate only to dreams of the passive type, but we find it used even when the dreamer is himself the central figure in the dream action.15 Again, the dream is said not only to "visit" the dreamer etc.)16 but also to "stand over" him The latter usage is particularly common in Herodotus, where it has been taken for a reminiscence of Homer's "it stood at his head";17 but its occurrence in the Epidaurian and Lindian Temple Records, and in countless later authors from Isocrates to the Acts of the Apostles,18 can hardly be explained in this manner. It looks as if the objective, visionary dream had struck deep roots not only in literary tradition but in the popular imagination. And that conclusion is to some extent fortified by the occurrence in myth and pious legend of dreams which prove their objectivity by leaving a material token behind them, what our spiritualists like to call an "apport"; the best-known example is Bellerophon's incubation dream in Pindar, in which the apport is a golden bridle.19
But let us return to Homer. The stylised, objective dreams I have been describing are not the only dreams with which the epic poets are acquainted. That the common anxiety-dream was as familiar to the author of the Iliad as it is to us, we learn from a famous simile: "as in a dream one flees and another cannot pursue him—the one cannot stir to escape, nor the other to pursue him—so Achilles could not overtake Hector in running, nor Hector escape him."20 The poet does not ascribe such nightmares to his heroes, but he knows well what they are like, and makes brilliant use of the experience to express frustration. Again, in Penelope's dream of the eagle and the geese in Odyssey 19 we have a simple wish-fulfilment dream with symbolism and what Freud calls "condensation" and "displacement": Penelope is crying over the murder of her beautiful geese21 when the eagle suddenly speaks with a human voice and explains that he is Odysseus. This is the only dream in Homer which is interpreted symbolically. Should we say that we have here the work of a late poet who has taken an intellectual leap from the primitiveness of Rose's first stage to the sophistication of his third? I doubt it. On any reasonable theory of the composition of the Odyssey it is difficult to suppose that Book 19 is much later than Book 4, in which we meet a dream of the primitive "objective" type. Moreover, the practice of interpreting dreams symbolically was known to the author of Iliad 5, which is generally thought one of the oldest parts of the poem: we read there of an oneiropolos who failed to interpret his sons' dreams when they went to the Trojan War.22
I suggest that the true explanation does not lie in any juxtaposition of "early" and "late" attitudes to dream-experience as such, but rather in a distinction between different types of dream-experience. For the Greeks, as for other ancient peoples,23 the fundamental distinction was that between significant and nonsignificant dreams; this appears in Homer, in the passage about the gates of ivory and horn, and is maintained throughout antiquity.24 But within the class of significant dreams several distinct types were recognised. In a classification which is transmitted by Artemidorus, Macrobius, and other late writers, but whose origin may lie much further back, three such types are distinguished.25 One is the symbolic dream, which "dresses up in metaphors, like a sort of riddles, a meaning which cannot be understood without interpretation." A second is the horama or "vision," which is a straightforward preënactment of a future event, like those dreams described in the book of the ingenious J. W. Dunne. The third is called a chrematismos or "oracle," and is to be recognised "when in sleep the dreamer's parent, or some other respected or impressive personage, perhaps a priest or even a god, reveals without symbolism what will or will not happen, or should or should not be done."
This last type is not, I think, at all common in our own dream-experience. But there is considerable evidence that dreams of this sort were familiar in antiquity. They figure in other ancient classifications. Chalcidius, who follows a different scheme from the other systematisers,26 calls such a dream an "admonitio," "when we are directed and admonished by the counsels of angelic goodness," and quotes as examples Socrates' dreams in the Crito and the Phaedo.27 Again, the old medical writer Herophilus (early third century b.c.) probably had this type in mind when he distinguished "godsent" dreams from those which owe their origin either to the "natural" clairvoyance of the mind itself or to chance or to wish-fulfilment.28 Ancient literature is full of these "godsent" dreams in which a single dream-figure presents itself, as in Homer, to the sleeper and gives him prophecy, advice, or warning. Thus an oneiros "stood over" Croesus and warned him of coming disasters; Hipparchus saw "a tall and handsome man," who gave him a verse oracle, like the "fair and handsome woman" who revealed to Socrates the day of his death by quoting Homer; Alexander saw "a very grey man of reverend aspect" who likewise quoted Homer, and in Alexander's opinion was in fact Homer in person.29
But we are not dependent on this sort of literary evidence, whose striking uniformity may naturally be put down to the conservatism of Greek literary tradition. A common type of "godsent" dream, in Greece and elsewhere, is the dream which prescribes a dedication or other religious act;30 and this has left concrete evidence of its actual occurrence in the form of numerous inscriptions stating that their author makes a dedication "in accordance with a dream" or "having seen a dream."31 Details are rarely given; but we have one inscription where a priest is told in a dream by Sarapis to build him a house of his own, as the deity is tired of living in lodgings; and another giving detailed rules for the conduct of a house of prayer which are stated to have been received in sleep from Zeus.32 Nearly all the inscriptional evidence is of Hellenistic or Roman date; but this is probably fortuitous, for Plato speaks in the Laws of dedications which are made on the strength of dreams or waking visions, "especially by women of all types, and by men who are sick or in some danger or difficulty, or else have had a special stroke of luck," and we are told again in the Epinomis that "many cults of many gods have been founded, and will continue to be founded, because of dream-encounters with supernatural beings, omens, oracles, and deathbed visions."33 Plato's testimony to the frequency of such occurrences is all the more convincing since he himself has little faith in their supernatural character.
In the light of this evidence we must, I think, recognise that th
e stylisation of the "divine dream" or chrematismos is not purely literary; it is a "culture-pattern" dream in the sense I defined at the beginning of this chapter, and belongs to the religious experience of the people, though poets from Homer downwards have adapted it to their purposes by using it as a literary motif. Such dreams played an important part in the life of other ancient peoples, as they do in that of many races to-day. Most of the dreams recorded in Assyrian, Hittite, and ancient Egyptian literature are "divine dreams" in which a god appears and delivers a plain message to the sleeper, sometimes predicting the future, sometimes demanding cult.34 As we should expect in monarchical societies, the privileged dreamers are usually kings (an idea which appears also in the Iliad);35 commoners had to be content with the ordinary symbolic dream, which they interpreted with the help of dreambooks.36 A type corresponding to the Greek chrematismos also appears among the dreams of contemporary primitives, who usually attach special importance to it. Whether the dream figure is identified as a god or as an ancestor naturally depends on the local culture-pattern. Sometimes he is just a voice, like the Lord speaking to Samuel; sometimes he is an anonymous "tall man," such as we meet in Greek dreams.37 In some societies he is commonly recognised as the dreamer's dead father;38 and in other cases the psychologist may be disposed to see in him a father-substitute, discharging the parental functions of admonition and guidance.39 If that view is right, we may perhaps find a special significance in Macrobius' phrase, "a parent or some other respected or impressive personage." And we may further suppose that so long as the old solidarity of the family persisted, such maintenance of contact in dreams with the father-image would have a deeper emotional significance, and a more unquestioned authority, than it possesses in our more individualised society.
However, the "divine" character of a Greek dream seems not to depend entirely on the ostensible identity of the dream-figure. The directness (enargeia) of its message was also important. In several Homeric dreams the god or eidolon appears to the dreamer in the guise of a living friend,40 and it is possible that in real life dreams about acquaintances were often interpreted in this manner. When Aelius Aristides was seeking treatment in Asclepius' temple at Pergamum, his valet had a dream about another patient, the consul Salvius, who in the dream talked to the valet about his employer's literary works. This was good enough for Aristides; he is sure that the dream-figure was the god himself, "disguised as Salvius."41 It made, of course, some difference that this was a "sought" dream, even though the person to whom it came was not the seeker: any dream experienced in Asclepius' temple was presumed to come from the god.
Techniques for provoking the eagerly desired "divine" dream have been, and still are, employed in many societies. They include isolation, prayer, fasting, self-mutilation, sleeping on the skin of a sacrificed animal, or in contact with some other holy object, and finally incubation (i.e., sleeping in a holy place), or some combination of these. The ancient world relied mainly on incubation, as Greek peasants still do to-day; but traces of some of the other practices are not lacking. Thus fasting was required at certain dream-oracles, such as "Charon's cave" in Asia Minor and the hero-shrine of Amphiaraus in Oropus;42 at the latter one also slept on the skin of a sacrificed ram.43 Withdrawal to a sacred cave in quest of visionary wisdom figures in the legends of Epimenides and Pythagoras.44 Even the Red Indian practice of chopping off a finger joint to procure a dream has an odd partial parallel, which I will mention presently.45 There were also in later antiquity less painful ways of obtaining an oracle-dream: the dream books recommended sleeping with a branch of laurel under your pillow; the magical papyri are full of spells and private rituals for the purpose; and there were Jews at Rome who would sell you any dream you fancied for a few pence.46
None of these techniques is mentioned by Homer, nor is incubation itself.47 But as we have seen, arguments from silence are in his case peculiarly dangerous. Incubation had been practised in Egypt since the fifteenth century b.c. at least, and I doubt if the Minoans were ignorant of it.48 When we first meet it in Greece, it is usually associated with cults of Earth and of the dead which have all the air of being pre-Hellenic. Tradition said, probably with truth, that the original Earth oracle at Delphi had been a dream-oracle;49 in historical times, incubation was practised at the shrines of heroes—whether dead men or chthonic daemons—and at certain chasms reputed to be entrances to the world of the dead (necyomanteia). The Olympians did not patronise it (which may sufficiently explain Homer's silence); Athena in the Bellerophon story is an exception,50 but with her it may be a vestige of her pre-Olympian past.
Whether or not incubation had once been more widely practised in Greece, we find it used in historical times mainly for two specialised ends—either to obtain mantic dreams from the dead, or else for medical purposes. The best-known example of the former is Periander's consultation of his dead wife Melissa on a business matter at a necyomanteion, when an "image" of the dead woman appeared to Periander's agent, established her identity, prescribed cult, and insisted on satisfaction of this demand before she would answer his question.51 There is nothing really incredible in this story, and whether true or false, it seems in any case to reflect an old culture-pattern, out of which in some societies a kind of spiritualism has been developed. But in Greece the Homeric Hades-belief, as well as the scepticism of classical times, must have worked to prevent such a development; and in fact mantic dreams from the dead seem to have played only a very minor part in the Classical Age.52 They may have acquired more importance in some Hellenistic circles, after Pythagoreans and Stoics had brought the dead into more convenient proximity to the living, by transferring the site of Hades to the air. At any rate we read in Alexander Polyhistor that "the whole air is full of souls, who are worshipped as the daemons and heroes, and it is these who send mankind dreams and omens"; and we find a like theory ascribed to Posidonius.53 But those who held this view had no reason to seek dreams in special places, since the dead were everywhere; there was no future for necyomanteia in the ancient world.
Medical incubation, on the other hand, enjoyed a brilliant revival when at the end of the fifth century the cult of Asclepius suddenly rose to Panhellenic importance—a position which it retained down to the latest pagan times. About the wider implications of this I shall have something to say in a later chapter.54 For the moment we are concerned only with the dreams that the god sent to his patients. Ever since the publication in 1883 of the Epidaurian Temple Record,55 these have been much discussed; and the gradual change in our general attitude towards the nonrational factors in human experience has been reflected in the opinions of scholars. The earlier commentators were content to dismiss the Record as a deliberate priestly forgery, or else to suggest unconvincingly that the patients were drugged, or hypnotised, or somehow mistook waking for sleeping and a priest in fancy dress for the divine Healer.56 Few, perhaps, would now be satisfied with these crude explanations; and in the three major contributions to the debate which have been made in the present generation—those of Weinreich, Herzog, and Edelstein57—we can observe a growing emphasis on the genuinely religious character of the experience. This seems to me entirely justified. But there are still differences of opinion about the origin of the Record. Herzog thinks it is based in part on genuine votive tablets dedicated by individual patients—which might, however, be elaborated and expanded in the process of incorporation—but also in part on a temple tradition which had attracted to itself miracle stories from many sources. Edelstein, on the other hand, accepts the inscriptions as in some sense a faithful reproduction of the patient's experience.
Certainty in this matter is hardly attainable. But the concept of the culture-pattern dream or vision may perhaps bring us a little nearer to understanding the genesis of such documents as the Epidaurian Record. Experiences of this type reflect a pattern of belief which is accepted not only by the dreamer but usually by everyone in his environment; their form is determined by the belief, and in turn confirms it; hence they become
increasingly stylised. As Tylor pointed out long ago, "it is a vicious circle: what the dreamer believes he therefore sees, and what he sees he therefore believes."58 But what if he nevertheless fails to see? That must often have happened at Epidaurus: as Diogenes said of the votive tablets to another deity, "there would have been far more of them if those who were not rescued had made dedications."59 But the failures did not matter, save to the individual; for the will of a god is inscrutable—"therefore hath He mercy on whom He will have mercy." "I am determined to leave the temple forthwith," says the sick pimp in Plautus; "for I realise the decision of Asclepius—he does not care for me or want to save me."60 Many a sick man must have said that. But the true believer was no doubt infinitely patient: we know how patiently primitives wait for the significant vision,61 and how people return again and again to Lourdes. Often in practice the sufferer had to be content with a revelation that was, to say the least, indirect: we have seen how somebody else's dream about a consul could be made to serve at a pinch. But Aristides had also experienced, as he believed, the god's personal presence, and described it in terms that are worth quoting.62 "It was like seeming to touch him," he says, "a kind of awareness that he was there in person; one was between sleep and waking, one wanted to open one's eyes, and yet was anxious lest he should withdraw too soon; one listened and heard things, sometimes as in a dream, sometimes as in waking life; one's hair stood on end; one cried, and felt happy; one's heart swelled, but not with vainglory.63 What human being could put that experience into words? But anyone who has been through it will share my knowledge and recognise the state of mind." What is described here is a condition of self-induced trance, in which the patient has a strong inward sense of the divine presence, and eventually hears the divine voice, only half externalised. It is possible that many of the god's more detailed prescriptions were received by patients in a state of this kind, rather than in actual dreams.