SOURCE: "The Sources: The Evidence for Written Sources," in his Herodotus, Twayne Publishers, 1982, pp. 142-53.

[Evans is professor of classics at the University of British Columbia, Canada. In the following excerpt, he finds little clear evidence that Herodotus relied on written sources. He describes Herodotus instead as an original researcher and interviewer whose History synthesizes the claims of a variety of mostly oral informants, including the guardians of official oral traditions, keepers of family genealogy, and individual storytellers.]

Some four centuries after Herodotus, another historian from Halicarnassus, Dionysius, briefly described the beginnings of historical research: "Before the Peloponnesian War [431-404 B.C.] there were many early historians in many places. Among them were Eugeon of Samos, Deiochus of Proconnesus, Eudemos of Paros, Democles of Phygele, Hecataeus of Miletus, Acusilaus of Argos, Charon of Lampsacus, and Amelesagoras of Chalcedon. A second group was born a little before the Peloponnesian War and were Thucydides' early contemporaries; these were Hellanicus of Lesbos, Damastes of Sigeum, Xenomedes of Ceos, Xanthus of Lydia, and many others." All of these wrote histories of individual tribes or cities, using records from temples or secular archives, and telling myths and folktales which, remarked Dionysius, who lived in the society that produced the emperor Augustus, "seem silly to presentday men." Herodotus, however, enlarged the historian's scope. "He chose not to write down the history of a single city or nation, but to put together many, varied events of Europe and Asia in a single comprehensive work."

This is a well-worn passage. Taken at face value, it indicates that before Herodotus there was a clutch of shadowy writers who wrote local histories. Herodotus himself mentions only one of them, Hecataeus of Miletus, who wrote a work in two books on historical geography (Periegesis), accompanied by a map that showed the world as a disc edged by Ocean, and another work on genealogies. He lived through the Ionian Revolt, and twice gave the rebels advice that they rejected. Herodotus says nothing pejorative about him, though once he tells a tale that has been taken as ridicule. In Egypt, Hecataeus gave the priests of Amon his own pedigree, beginning with a god and descending through sixteen generations, whereupon the priests showed him the series of statues of their high-priests, three hundred and forty-five in all, each representing a generation of men. The story illustrated the antiquity of Egypt, and it is not clear that Herodotus meant it as a joke at Hecataeus's expense, as some commentators have thought. In any case, the tale probably derived from Hecataeus's own Periegesis, and the joke, if joke it is, was told by the victim himself.

Herodotus had absorbed what Hecataeus wrote about Egypt, but too little survives to be sure how dependent Herodotus was. The description of the phoenix, the hippopotamus, and the hunting of the crocodile had parallels in Hecataeus. Hecataeus called Egypt "the gift of the Nile," and Herodotus approves of the description without mentioning its source. At Buto, Herodotus viewed the floating island of Chemmis, reported that he saw no sign that it floated, and was skeptical about floating islands in any case. However, a fragment of Hecataeus states that Chemmis, which he spelled "Chembis," floated.

All this indicates familiarity with Hecataeus but not dependence. Outside Egypt, Herodotus quotes Hecataeus only once, for a variant version of how Athens expelled the aborigine Pelasgians from Attica. Perhaps he found in Hecataeus his list of Persian satrapies and the myth of Scythian descent from a union between Heracles and a woman who was half snake. His disapproval of the Ionian Revolt may have come in part from him, for Hecataeus thought it was foolish, and so did Herodotus. But Hecataeus was a friend of Artaphrenes, the satrap at Sardis during the revolt, and his attitude can have been no secret. For that matter Hecataeus and Herodotus belonged to the same social class, and may have had more in common than an interest in historical geography. But Herodotus had progressive ideas about cartography, and was contemptuous of maps that showed the earth circular, with a Europe and Asia of equal size and all surrounded by Ocean. Hecataeus's famous map shared this disdain but not alone, for he was part of an Ionian tradition of cartography, and the concept of the circumambient Ocean went back to Homer.

Hecataeus is the most solid of Herodotus's predecessors, although evidence is lacking to substantiate theories of widespread borrowing. There is less to say about the other writers mentioned by Dionysius (and some he did not mention), and they are harder to date. Charon of Lampsacus is credited with a number of works, including one on Persia, two on Lampsacus, and a chronicle of the kings and ephors of the Lacedaemonians, based on Spartan sources. We cannot show that Herodotus read him, but it is at least possible, for though Jacoby, a generation ago, dated him to the last decades of the fifth century, too late for Herodotus, more recent scholarship is inclined to make him earlier. His Persika [On Persia] took up only two books; other than that, we can say only that it included an account of Mardonius's expedition into northern Greece in 492B.C. Charon's scope was narrower than that of Herodotus.

Xanthus of Lydia wrote a work On Lydia which Herodotus may have used; in the next century, the historian Ephorus claimed that Herodotus got his "startingpoints" (aphormai) from him, but his meaning is obscure. The History does begin with Lydia, and in that sense Xanthus may have provided Herodotus with a starting point. But surviving fragments show no parallels with Herodotus. He borrowed nothing from Xanthus that we can identify.

Even more shadowy is Dionysius of Miletus, who wrote a work on Persia down to Darius's death, and then followed it with a sequel. The one morsel of information that we know about his works reveals that he described the revolt of the magi, and named Smerdis's brother Panxouthes rather than Patizeithes, as Herodotus does. Dionysius of Halicarnassus overlooks him, which is a tribute to his obscurity. However, the king lists of Lydia and of Media in Herodotus do not synchronize, which has given grounds for suspecting that he took his lists from two separate sources, and faute de mieux, Dionysius of Miletus is a candidate for one of them. As for Acusilaos of Argos, a fragment has turned up in an Oxyrhynchus papyrus; it gives a straightforward account of the myth of the Lapith king Caeneus who was changed from a man into a woman. It reads like an entry in a mythology handbook: the sort of material purveyed by the logioi, or prosewriters, whose imaginary debate on the causes of the Persian War begins the History. Herodotus's researches were a different sort of thing.

Hellanicus of Lesbos was a prolific writer who was still working in the last decade of the fifth century, but a Persika by him could have antedated the History, although we cannot show that Herodotus used it. However, he demonstrates the availability of chronological sources, for three of his works betray a fascination with dates. One was a local history of Athens dated by archon years. A second was on the victors at the Cernean Festival in Sparta, and the third was on the priestesses of Hera in Argos, who held office for life; Hellanicus reckoned their tenure by years, thus producing a chronological framework. Another writer whom Herodotus may have used was Pherekydes of Athens, who wrote a work containing much information about the Philaids, the family to which Miltiades belonged, but what Herodotus has to say about Miltiades' ancestors might as well have been obtained from any well-informed member of the clan. We are far from being able to demonstrate Herodotus's dependence. Moreover, his failure to mention any prose writer except Hecataeus is remarkable, for he frequently cites poets, not merely Homer and Hesiod, but even one as récherché as Aristeas of Proconnesus, the author of the Arimaspeia. In part, this is because familiarity with the poets was the mark of an educated Greek, but that cannot be the whole reason. Herodotus, it is clear, regarded himself as an independent researcher whose analysis of the Persian War was an original achievement.

What is evident is that Herodotus did not conduct his researches in a vacuum. The fifth century was a period when records and genealogies were being ferreted out, and what may have been left to memory in the past was put in writing. Herodotus maintained the persona of an oral historian working with oral sources. He put down the stories that people told him, even when he did not accept their truth, and he would not impose his own critical judgment so far as to suppress variant versions— as Hecataeus did, for in his proem he made a claim for accuracy. All versions were evidence for Herodotus, and he would not neglect evidence in his research.

The Sources Cited

Halfway through the Egyptian logos, Herodotus announces that what he has written so far is based on observation (opsis), judgment (gnome), and interrogation (historie), but from that point on he will relate stories of Egypt, some told by the Egyptians themselves, and some by other nations about Egypt. Elsewhere, he prefaces various reports with phrases such as "the Greeks say," "the Lacedaemonians say," "the people living around Thermopylae say," to give only a few examples. Not infrequently he contrasts the reli-ability of these sources: the priests of Ptah in Egypt told how Psammetichus discovered that the Phrygians were the world's most ancient people; the Greeks, he adds, related silly stories about his experiment. Or at times he is precise about his limitations: he describes the golden statue of Marduk-Ba'al at Babylon, but he had not seen it himself, for Xerxes had removed it; he could only tell what the Chaldaeans said. Four times only he mentions informants by name, but all of these gave him private information. The sources designated as "Greeks," "Carians," "Scythians," and the like seem to purvey information which, if not official, was shared widely among the peoples named. It was part of the body of tradition that they preserved.

Can we take this view seriously? When Herodotus says of the Nile Delta that "the Ionians say" it is Egypt, and that its coast stretches from the so-called watchtower of Perseus to the salt marshes of Pelusium, can we believe that this was an oral tradition widespread in Ionia? It may be argued that behind many of these references to oral traditions there lies a literary source: Hecataeus took Egypt as the Delta, and so we can read "the Ionians say" as "Hecataeus writes." But Hecataeus's geographical notions, like his map, belonged to a school of Ionian savants. Herodotus was strictly accurate when he wrote "the Ionians say."

Or was the persona of an oral historian purely conventional, in which case we must treat these interjections as literary fictions? This is not to accuse Herodotus of dishonesty, but simply to suggest that he was not merely the father of history, but also of the literary conventions that affect historiography. Herodotus was capable of literary convention. But before we weigh his claim to be an oral historian, we must look at what oral history consists of.

The nature of oral tradition

The boundary between a literate and a nonliterate culture is notoriously difficult to define, and if it would be incorrect to denominate the Greece of Herodotus as nonliterate, neither would it be right to think that the elements of oral culture were dead. Oral history is transmitted by what we may visualize as a long series of interlocking conversations continuing from generation to generation, and each generation makes its own adjustments to the tradition. What is of social relevance is remembered and subjected to interpretation; what ceases to be relevant is forgotten.

Let us look first at problems of chronology. In Africa south of the Sahara, where the elements of oral culture are still alive, the sources of history for the precolonial period are generally professional storytellers and official keepers of state traditions, and usually they have at best a hazy notion of absolute chronology. However, they can, by means of genealogy, establish an area of time within which an event took place. In nearly every state headed by kings or chiefs, the custodians of traditions possess king lists, in what is supposed to be chronological order, and when these states came into contact with writing, such lists were often the first morsels of history to be written down. In them, kings generally appear as consecutive rulers, even when their reigns in fact overlapped. The Behistun inscription provides an example of this, for there Darius names nine ancestors who were kings before him, without indicating that, in part, they ruled in parallel lines of the Achaemenid house. The same sort of distortion affects Herodotus's dating of the fall of Media and the rise of Cyrus, for by his reckoning Cyrus became king in 558 B.C., and the act by which he secured the throne was the overthrow of the king of Media, Astyages, whose fall, therefore, belonged to the same year. The two are made to rule consecutively. In fact, Cyrus first succeeded his father, Cambyses I, as vassal king of Anshan, and only later, in 550, did he unseat Astyages.

However, some regnal lists can assign a sum of years to a reign with great accuracy, for statistics of that sort could be preserved with great care, sometimes by means of mnemonic devices. It follows that when we find a reign measured precisely in oral tradition, we should pay attention, for the statistic may depend on a reliable method of time reckoning. Herodotus provides an example: he assigns thirty-six years to the Pisistratid tyranny before its expulsion from Athens. If Herodotus knew the date of the expulsion—Thucydides gives 510 B.C.—then he knew that Pisistratus routed his opponents at Pallene and became tyrant in 546: a troublesome date, for Herodotus puts Cyrus's first conquest, his overthrow of Croesus, just after Pallene. Yet we must not discard the thirty-six years, for oral tradition tends to be exact about such things.

At the same time, genealogies are subject to adjustment, sometimes due to social considerations, sometimes from structural amnesia, a process that streamlines tradition by forgetting irrelevant details. Generations tend to lengthen with time, and genealogies preserved in oral traditions now, may list no more names than they did a century ago, though there are a hundred more years to account for. Herodotus's rule of thumb for a generation was three to a century, but when he listed the pedigrees of the two Spartan royal houses, he traced both back to Heracles, whom he placed nine hundred years before his time, and at three generations to a century the pedigrees cannot stretch so far. Consequently these generations must be calculated at forty years. Structural amnesia, combined with a determination to make Heracles the progenitor of the Spartan kings, has resulted in a generation too long to be probable.

It must ultimately have been the Spartans themselves— specifically the Spartan kings—who stretched these generations, for the royal houses derived prestige from descent from Heracles. Herodotus probably got these pedigrees in Sparta, for, as Plato noted, the Spartans had an extraordinary appetite for genealogies. There is no compelling reason to think that he got them through an intermediary, such as Hecataeus. But there is no such thing as a standard generation in Herodotus. His three generations to a century is a rule of thumb, and usually he follows the practice of the oral historian who deals with areas of time rather than precise years.

The methods of transmission

Oral traditions found in African states generally fall into two categories: official and private. The first category represents the "truth" about the past as the state recognizes it, and it is common to find professionals charged with its preservation. The Yoruba town of Ketu in Nigeria, for instance, had an hereditary official known as the baba elegum who knew the town history by heart. Rwanda, a kingdom until 1961, had an assortment of officials: genealogists who remembered pedigrees, memorialists who knew the important events of reigns, rhapsodists who preserved panegyrics on the kings, and the abiiru who kept the secrets of the dynasty. There are exceptions: Burundi, which ceased to be a monarchy in 1966, had no official traditions as such, but history was transmitted by songs, tales, and proverbs. But, in general, organized states had both official traditions and specialists who were charged with preserving them.

Private traditions are those transmitted by individual groups, such as families and clans, which may have a official status within the clan, but as far as the outside world is concerned, they are private, and are handed down with less care, though at the same time there is less motive to distort. However, oral traditions are never transmitted purely for the love of objective knowledge. Private traditions may be put to political and social uses less blatantly than official ones, but there is still the wish to put the clan's ancestors in a good light.

When Herodotus cites ethnic groups as sources, such as "the Carians" or "the Spartans," inter alia, the implication is that he is giving their official traditions. Thus the Cretans could say that the Carians once inhabited islands under Minoan suzerainty, but Carian tradition differed, and Sparta, Thera, and Cyrene all had traditions about the founding of Thera and Cyrene. More than once Herodotus notes the Spartan version of a morsel of Spartan history differed from that held by the rest of Greece. He saw himself as a reporter of traditions such as these, but he declined to vouch for their accuracy. Yet, the question arises: if Herodotus could draw on oral traditions, were there specialists in archaic Greece who preserved them, as there have been, and are, in other parts of the world?

The evidence is meager, but tantalizing. Aristotle names among the officials necessary for a city various registrars called "temple-remembrancers, archive keepers, and remembrances." The "remembrancer" (mnemon) was a registrar of property, but the name seems to indicate an official who relied simply on his memory at one time, before literacy became common. From Crete we have an inscription that sets forth the rights of one Spensithios, who was to be poinikastes (specialist in Phoenician letters) and mnemon of a Cretan city, and his descendants after him. Slight as it is, the evidence suggests that Herodotus found keepers of tradition in various cities and temples, and that these specialists could expound history. At Delphi, a hieromnemon could have told him of Croesus's dedications and other memorials—the erga ("works") of famous men—and would give official answers to queries about oracles that Delphi had given, although this need not imply that the oracles were kept on file. Temple memorialists could elucidate difficult inscriptions or describe important mementoes; one in the Theban temple of Ismenian Apollo may have given Herodotus a hand with "Cadmean letters" he found on some tripods dedicated there, and another at the Samian Heraeum may have expounded upon the picture of the bridge built by the engineer Mandrocles over the Bosphorus, and dedicated to Hera by Mandrocles himself.

In one instance we find Herodotus interviewing an official such as this. At Sais, in Egypt, he approached the temple of Neith to inquire about the source of the Nile. His interlocutor was the temple scribe, and if we may infer from evidence for Egyptian temples in the Hellenistic period, he was the keeper of traditions. The scribe of Neith purveyed myth, but he was performing his proper function of expounding priestly wisdom.

He was, however, a scribe rather than a "sacred remembrancer" (hieromemnon), for Egypt had long used writing, and literacy was a qualification for the priesthood. In contemporary Greece, too, many memorialists must have used writing, but they still acted as spokesmen for an undifferentiated group, such as the priests of Dodona, whose spokesman told Herodotus that the Pelasgians had no names for their gods until they got them from Egypt. Only occasionally did he single out an individual as a source, because his information was unique, or because—as in the case of the scribe of Neith—Herodotus did not believe him.

There were also private family traditions, usually favorable to the family in the sense that forgetfulness obscured what cast aspersions on its past, but less prone to official bias. The family status was what concerned such traditions. Thus, when Herodotus indicates that both the great political families of Athens of the fifth century, the Philaids and the Alkmaeonids, had been hostile to the Pisistratid tyrants, he is reflecting the traditions of the two families, which forgot that they had ever cooperated with the tyrants. Yet, family traditions could not be distorted outrageously without provoking disbelief; they had to compete with other family traditions, and with official tradition as well, which it was difficult to impose upon. Alkmaeonid and Philaid tradition notwithstanding, the archon list that was inscribed and set up in the marketplace of Athens about 425B.C. bore evidence that both families had cooperated with the tyrants. A better case is the competition between the two versions of how Athens was freed from her tyranny. The story with a degree of official sanction gave credit to the tyrannicides, Harmodius and Aristogeiton: a statuegroup of the pair was erected in the agora, and their descendants received special honors. The competing tradition, for which both Herodotus and Thucydides vouched, claimed that the tyrannicides were motivated by a private quarrel, that they killed the wrong man, and that Athens was delivered from the tyrant four years later through the efforts of the Alkmaeonids. We may safely assign this tradition to the Alkmaeonid family. Yet, in spite of this family's prominence, and the authority of Herodotus and Thucydides, it did not displace the official version, for the descendants of the tyrannicides continued to receive honors into the fourth century.

There are a number of instances in the History where Herodotus has clearly drawn upon family traditions. He knew what the Gephyraean clan, to which the tyrannicides belonged, said about its antecedents, and his defense of the Alkmaeonids against the charge of medism at the battle of Marathon belonged to Alkmaeonid tradition. His slighting references to king Cleomenes of Sparta must derive from the traditions of the house to which he belonged, for Cleomenes' heirs belonged to another branch that had no reason to cherish his memory. He may have dealt kindly with Artabazus, who fled ingloriously from Plataea, because one of his sources was Artabazus's family. The descendants of Demaratus, who governed three towns in the Caicus valley as Persian vassals, were probably the source for the tale of Demaratus's friend, Dicaeus, who had seen a foreboding vision prior to Salamis. The grandson of Zopyrus who won Babylon for Darius came to Athens as an exile and may have had tales to tell there. Family traditions were kept with less care than their official counterparts and faded badly after three generations, but they probably had much color, some of it political, and they were strong on genealogy.

Finally there were the legends of storytellers. Who knows what Herodotus learned at Halicarnassus, where the Oriental world marched with the Greek? As we have seen, he may have heard there about the turncoat Phanes, and the story of the Magian Revolt could have been based on one of the versions of the Behistun inscription that reached Ionia; fragments of an Aramaic version have turned up in Egypt. Herodotus made a point of disdaining idle tales, but a discriminating listener must have found a good many that were not idle.

Archives

In 403 B.C., Athens organized a central archive in the Metroon, the temple of the Mother of the Gods. Before that time, our information about her archives is scanty, though a late source states that they date back to the sixth century B.C. Outside Athens, our evidence is even scantier, but no doubt temples kept archives: inventories, for the most part, including lists of priests, but perhaps secular records too. A fair amount of archival material was preserved.

How accessible these records would be to the curious researcher, however, is another question. Unless a document was published, Herodotus would have been at the mercy of mnemones and hieromnemones, and as we suspect, many of these were still working within an oral tradition. Publication meant inscribing a document and setting it up in a public place. Herodotus cites twenty-four inscriptions, half of them Greek, and half non-Greek. Some he copied, for he gives the texts, but one at least, an inscription on the Great Pyramid at Gizeh, he paraphrased from memory. On the whole, he does not seem to have valued documentary evidence highly, though the reason may have been, not that it was unavailable, but that it was inaccessible.

Conclusion

We must conjecture, but we can be brief. Herodotus was gathering his source material at a time when a minor explosion of historical research was taking place, but he was in the vanguard. It is likely that he was widely read; he was not isolated from the intellectual milieu of his time, but we can prove that he read only one predecessor, Hecataeus, and we cannot demonstrate real dependence even on him. It is futile to conjure up shadowy predecessors whose works he may have copied.

Yet, he lived at a time when oral traditions were still preserved with care, and he probably gleaned much of his information from oral sources: from mnemones, hieromnemones, family traditions, and individuals with tales to tell. Even without the structures intended to preserve oral traditions, memories can remain fairly green for three generations. The Persian Wars were still within the three-generation span when Herodotus did his research, and indeed, there were still men alive who had witnessed the great invasion firsthand. The persona of an oral historian that Herodotus assumed must be taken seriously.