¹ See, e.g., the surveys of Tod, M. N., OCD 394Google Scholar s.v. Epigraphy, Greek; Larfeld, W., Griech. Epigraphik ³ (Munich, 1914), 7Google Scholar; Guarducci, M., Epigrafia Greca (Roma, 1967), i. 27Google Scholar.

² We need not linger over the claims of Acusilaus of Argos to be regarded as an epigraphical pioneer: τ⋯ γ⋯ρ Ἀκουϲιλάου νοθεύεται (Acusil. FGrHist 2 T 7), and the bronze tablets of genealogical content dug up by his father (FGrHist 2 T 1) belong with the Trojan War diaries of Dictys and the golden tablets which provided Joseph Smith with scriptural foundation for the peculiar doctrines of the Mormon church; see further Speyer, W., Bücherfunde in der Glaubenswerbung der Antike (Göttingen, 1970), esp. 43ffCrossRefGoogle Scholar.

³ In particular, Hippias and Hellanicus.

⁴ Volkmann, H., ‘Die Inschriften im Geschichtswerk des Herodot’, Convivium (Festschr. K. Ziegler) (Stuttgart, 1954), 41–65Google Scholar.

⁵ Panofsky, H., Quaestionum de historiae Herodoteae fontibus pars prima (diss. Berlin, 1885)Google Scholar; Sayce, A. H., The Ancient Empires of the East, Herodotos i–iii (London, 1883)Google Scholar; ‘The season and extent of the travels of Herodotos in Egypt’, JPh 14 (1885), 257–86Google Scholar.

⁶ Similarly disquieting conclusions are suggested in a series of articles by O. K. Armayor, bearing mainly on the extent of Herodotus' travels: see in particular TAPA 108 (1978), 1ff.Google Scholar, HSCP 82 (1978), 45ff.Google Scholar, 84 (1980), 53ff.

⁷ Listed according to the order in which I shall discuss them.

⁸ I have omitted from this list the statue at Metapontum ⋯πωνυμίην ἔχων Ἀριϲτέω (4.15.2–4); it is not quite clear whether Herodotus means that there was an inscription on the statue, or simply that it was generally known as the statue of Aristeas, but the latter seems much more likely. Volkmann, who does not discuss the passage, evidently did not think it referred to an inscription.

⁹ An extraordinarily harsh anacolouthon: ɸάμενον…λέγον Madvig (adopted by Legrand in the Budé edition).

¹⁰ For a similar refusal to name names cf. 2.123.3; 4.43.7.

¹¹ For details see Meiggs-Lewis, pp. 57ff.

¹² It is interesting (though more significant in relation to Thucydides) that though Herodotus speaks of Pausanias' arrogance towards the allies (8.3) and of his ambition to rule all Greece (5.32), he says nothing about the couplet (subsequently erased by the Spartans), engraved on the monument by the regent's instructions, in which he was described as Ἑλλήνων ⋯ρχηγόϲ (Th. 1.132.2–3).

¹³ In addition to the well-known stelae from the Isthmus of Suez we may compare the inscriptions on the colossal monolithic statue of Darius discovered at Susa in 1972 (Journal Asiatique 260 [1972], 235 ff.Google Scholar, Cahiers DAFI 4 [1974], 73–160Google Scholar). The cuneiform inscriptions, on the right-hand side of the robe, run thus: ‘A great god is Ahuramazda who created this earth, who created this sky, who created man, who created happiness for man, who made Darius king. This is the stone statue which Darius the King ordered to be made in Egypt so that he who should see it in time to come should know that the Persian man has taken Egypt. I am Darius, the Great King, the King of kings, King of lands, King on this great earth, son of Hystaspes, an Achaemenid. Darius the King says: May Ahuramazda protect me and all my works’. On the left side of the robe a much longer hieroglyphic text describes Darius in purely Egyptian terms. A convenient brief account of the statue is given by Cook, J. M., The Persian Empire (London, 1983), 58, 238 n. 3, plate 15Google Scholar; for more detailed discussion see Root, M. C., The King and Kingship in Achaemenid Art (Acta Iranica 3e sér., Vol. ix, Leiden, 1979), 61ff., 68ffGoogle Scholar. It is worth noting that though Herodotus refers to Darius' concern for the Red Sea canal (2.158.1; 4.39.1) he does not mention the inscriptions erected there. (Fehling's objections (102 n. 13) to treating the Suez canal stelae as a parallel seem to me hypercritical.)

¹⁴ Thucydides (4.50.2) describes a Persian letter as written in Ἀϲϲύρια γράμματα; pace Gomme, the script used must have been Aramaic, though whether Thucydides realised this is another matter. For a discussion of the phrase's connotations see Nylander, C., ‘ΑССΥΡΙΑ ΓΡΑΜΜΑΤΑ: remarks on the 21st “Letter of Themistocles”’ OAth 8 (1968), 119ffGoogle Scholar.

¹⁵ So, e.g., Hignett, C., Xerxes' Invasion of Greece (Oxford, 1963), 351Google Scholar.

¹⁶ Page 193–4.

¹⁷ Page (231), discussing Herodotus' use of ⋯πιγράψαϲ in connection with the epitaph for Megistias (7.228), takes it for granted that the verb merely means ‘that the subject made arrangements for, and paid the cost of, the composition and the inscribing’. Still, in the case of Megistias, common sense suggests that Simonides, the subject of ⋯πιγράψαϲ, did in fact compose the epigram.

¹⁸ It has been suggested that Herodotus drew extensively on this picture for his description of Xerxes' host (7.61ff.); it would be pleasant to suppose that prolonged study of the painting impressed the accompanying epigram on his memory.

¹⁹ See further Mitchell, B. M., ‘Herodotus and Samos’, JHS 95 (1975), 75–91CrossRefGoogle Scholar, Tölle-Kastenbein, R., Herodot u. Samos (Bochum, 1976), 9ff., 104ffGoogle Scholar.

²⁰ Meiggs-Lewis no. 15, pp. 28f., Page, 191–3, Hansen, P. A., Carmina Epigraphica Graeca (Berlin, 1983), no. 179, pp. 99fGoogle Scholar. The first two lines (according to the later, Herodotean order) are cited in a papyrus commentary of the late first century a.d. (P. Oxy. 2535); no other published papyrus covers the passage.

²¹ Page well emphasises that it is certainly not to be defended (as Friedlaender attempted) by the alleged parallel from a now lost inscription of much later date, published by Meritt, , Hesperia 16 (1947), 289Google Scholar, from a copy made by Sir George Wheler during his visit to Athens in 1676, ⋯χνυθ⋯ν τόδε δ⋯ρον ὑπ⋯ρ τάɸον εἵϲατο μήτηρ, where Peek's ⋯χνυόεν (GVI No. 238, p. 62) is surely right.

²² Though this interpretation is virtually renounced in the Supplement, where ⋯χνυόεντι is recommended as probably the correct reading here.

²³ P. 182, 3. LSJ's ascription to Callimachus of the line quoted to illustrate the meaning reflects Schneider, not Pfeiffer; the reference should now read Suppl. Hell. 1031.

²⁴ It is of course conceivable (though it may not be thought very probable) that Herodotus knew the verses from an oral tradition which preserved the correct form of the epithet.

²⁵ For similar errors in fifth-century inscriptions see Hansen, CEG nos. 167 (Chios), 380 (Arcadia)Google Scholar, with notes ad loc.

²⁶ So Meiggs-Lewis. On the pre-Mnesiclean propylaea see Dinsmoor, W. B. Jr, Propylaea, i (Princeton, 1980)Google Scholar.

²⁷ ‘Herodotus in Athens?’ in Greece and the Eastern Mediterranean in Ancient History and Prehistory: Studies presented to Fritz Schachermeyr (Berlin, 1977), 246–65 (esp. 259f.)Google Scholar.

²⁸ I must admit to some scepticism regarding the greater part of Herodotus' alleged mileage, but the point is not material to the argument here.

²⁹ Still, we may smile at Jacoby's confident claim (233) that ‘Eine der sichersten Tatsachen in H's Leben ist seine intime Verbindung mit Athen and im besondern mit dem Kreise um Perikles’.

³⁰ Macan² on 8.55. The problems presented by Herodotus' description of the location of the Aglaurion have recently been discussed by Dontas, G. S., ‘The true Aglaurion’, Hesperia 52 (1983), 48ff.CrossRefGoogle Scholar, but his attempt to defend Herodotus is unconvincing: see further Lewis, D. M., Postscript in Burn, A. R., Persia and the Greeks ² (London, 1984), 607–8Google Scholar.

³¹ We may compare his mistake about the orientation of Thermopylae: see below, p. 289.

³² Macan¹ i.lxi.

³³ Cf. 1.66.1; 5.66.1; 7.156.1–2.

³⁴ Cf. Macan²ad loc.: ‘These inscriptions were read by the Ionians on the very next day…were they ever read by any other mortals thereafter? How many times the inscription was reproduced by Themistokles Herodotus does not specify, but he gives the ipsissima verba, which “cut the record” for argument and rhetoric in inscriptions, a veritable “sermon in stones”; had Herodotus copied the same? Did “Ionians” remember and report them at home? Is the anecdote an Athenian invention? The appeal reads in any case more like a letter or an oral address than like a hastily cut inscription, and that in duplicate’.

³⁵ We naturally wonder whether Herodotus had evidence that Themistocles specifically mentioned the Carians or whether Halicarnassian patriotism led him to invent a detail which implies that Carian commitment to the Persian cause was, or might reasonably have been thought to be, only half-hearted.

³⁶ The different versions of the ‘Oath of Plataea’ show in how offhand a manner Greeks could treat texts while professing to quote them (see further Burn, op. cit. 512ff.), but that kind of inaccuracy is to be explained by a greater reliance on memory.

³⁷ Herodotus' use of speeches has most recently been examined by Hohti, P., The Interrelation of Speech and Action in the Histories of Herodotus (Helsinki, 1976)Google Scholar, where references to earlier discussions may be found. We should note that this device occasionally enlivened the dry narrative of Hecataeus (FGrHist 1 F 30: Ceyx's agonised rejection of the Heracleidae).

³⁸ If our MSS are to be trusted, there are a few exceptions. τάδε is used to introduce a speech instead of τοιάδε at 3.29.2 and 5.112.1; both are short passages, and possibly Thucydides believed that he could virtually reproduce what was said. ὧδε is used to introduce the speech of the ephor Sthenelaidas (1.85.3), remarkable for its blunt laconism, and Archidamus' solemn appeal to gods and heroes before Plataea (2.74.2).

³⁹ The technique is of course open to objection where Herodotus might or might not have had reasonably reliable information as to who said what (as with the various deliberations on the Greek side in 480 and 479).

⁴⁰ Krueger suggested that this sentence is an interpolation; it is certainly clumsily expressed (see Macan ad loc.), but seems an unlikely addition; see Masaracchia ad loc.

⁴¹ On the Thermopylae epigrams see Page 195–6, 231–4; he well emphasises that Herodotus evidently did not believe either the first or the second poem to be the work of Simonides.

⁴² Stein's ⋯γεμόνα is very attractive.

⁴³ JHS 53 (1933), 72f.Google Scholar; this misconception has dictated Herodotus' figure for Greek fatalities at Thermopylae (8.25.2).

⁴⁴ The epic reminiscences of Simonides' epigram may put us in mind of the Iliadic seer Merops of Perkote ὃϲ περ⋯ πάντων | ᾔδεε μαντοϲύναϲ, οὐδ⋯ οὓϲ παῖδαϲ ἔαϲκε | ϲτείχειν ⋯ϲ πόλεμον ɸθιϲήνορα. τὼ δέ οἱ οὔ τι | πειθέϲθην. κ⋯ρεϲ γ⋯ρ ἄγον μέλανοϲ θανάτοιο (Il. 2.831ff., 11.329ff.). Megistias, who persuaded his son to leave but himself stayed to die a glorious death, may thus be judged happier than the Homeric prophet on two counts.

⁴⁵ In marked contrast to the apathetic resignation which Herodotus would have us believe typical of Mardonius' officers (9.16.3–5).

⁴⁶ Cf. Lewis, D. M., Sparta and Persia (Leiden, 1977), 148ffGoogle Scholar.

⁴⁷ We might wonder if the first audience of Sophocles' Ajax were struck by a certain parallelism between, on the one hand, the differing roles of Athens and Sparta in defeating the Persians and, on the other, the contrast, already present in the Iliad but accentuated by Sophocles, between the soldierly qualities of Odysseus and Ajax.

⁴⁸ ‘Hier ist kein Zweifel an eigener sorgfältiger Besichtigung des Schlachtfeldes. Er hat den steinernen Löwen des Leonidas in situ gesehen (7.225.2) and die Epigramme für die Gefallenen selbst von den Stelen abgeschrieben (7.228)’ (272).

⁴⁹ ‘It was not Herodotus' custom to read and copy inscriptions, and it is not known whether he ever saw the actual epigrams at Thermopylae. If he did see them, it appears improbable that he made copies of them for use in his History.’

⁵⁰ As Bury, argued, BSA 2 (1895–1896), 83Google Scholar.

⁵¹ 7.176; cf. 199, 200.1, 201. His orientation is also badly in error at 9.14: see Macan ad loc.

⁵² Cf. A. R. Burn, op. cit. 380–1: ‘Herodotos…had clearly been through the pass, making careful notes, but lacking either the time or the inclination to leave the road… Since he consistently refers to the coast-road as running north and south, with the sea “east” and the mountains “west” of it, it seems that he must have passed only during the middle hours of a day's journey, with the sun high enough overhead to give him no reminder of his bearings’ (cf. p. 414). The attempt to save Herodotus' reputation for autopsy by the hypothesis that he was pressed for time is familiar to students of his account of Upper Egypt.

⁵³ To support his thesis that Herodotus here relies on hearsay Page offers an ingenious (but to my mind unconvincing) argument based on his view of the relative merits of Herodotus' ῥήμαϲι πειθόμενοι and the reading given by later authors πειθόμενοι νομίμοιϲ (Lycurg. in Leocr. 109, D.S. 11.33.2; cf. Cic. Tusc. 1.101 ‘dum sanctis legibus obsequimur’); judging the latter intrinsically superior, he draws the necessary conclusion that it must therefore be supposed to have stood on the stone, while Herodotus' version represents a corruption. But ῥήμαϲι πειθόμενοι has a strong claim to primacy as the lectio difficilior; most critics have also thought it was in itself better. Page himself followed Herodotus, ' version when he published Epigrammata Graeca (Oxford, 1975), 18Google Scholar.

⁵⁴ My approach to these dedications owes much to the valuable discussions of Kleingünther, A., Πρ⋯τοϲ Εὑρετήϲ (Philologussuppl. xxvi, i. 1933), 60–5 and Fehling, 102–4Google Scholar. This area would be a quicksand for the non-epigraphist without Miss Jeffery's survey of the origins and diffusion of archaic Greek script; for developments since 1961 see Heubeck, A., Schrift (Archaeologia Homerica, iii (X), Göttingen, 1979), 73ff.Google Scholar, Jeffrey, , CAH iii ²1.819ff.Google Scholar, Johnston, A., ‘The extent and use of literacy: the archaeological evidence’, The Greek Renaissance of the Eighth Century B.C.: Tradition and Innovation ed. Hägg, R., Acta Instituti Atheniensis Regni Sueciae Ser. 4, xxx (Stockholm, 1983), 63–8Google Scholar.

⁵⁵ For some alleged religious διδαϲκάλια cf. 2.49.3.

⁵⁶ Herodotus does not ascribe the invention of a writing-system to Cadmus; we do not know what views he held about the origins of script in general or of the Phoenician system in particular.

⁵⁷ The antiquity of Cadmus' connection with Phoenicia is debatable, and some have supposed that it originated with the logographers: see Latte, , RE xx 1470Google Scholar s.v. Kadmos (4), Gomme, , JHS 33 (1913), 53ff., 223ff.CrossRefGoogle Scholar Herodotus is in fact our earliest explicit witness (but cf. Pherecydes, , FGrHist 3 F 21Google Scholar), though the manner in which he refers to Cadmus' Phoenician origins implies that the idea was familiar and, indeed, uncontroversial (cf. 2.49.3, 4.147.4). See further the judicious discussion by Edwards, R. B., Kadmos the Phoenician (Amsterdam, 1979), esp. 45ff.Google Scholar

⁵⁸ Ionic: ɸοινικήια SIG ³ 38 B 37–8 (Meiggs-Lewis no. 30, from Teos, 480–50); further fragments are published by Herrmann, P., Chiron 11 (1981), 1–30Google Scholar, in which the new compound ɸοινικογραɸέω (d 19–21) should be noted; Aeolic: ɸοινικόγραɸοϲ as the title of an official at Mytilene: IG xii. 2.96–7 (Hellenistic); Cretan: see Jeffery, L. H. and Morpurgo-Davies, A., ‘ΠΟΙΝΙΚСΤΑС and ΠΟΙΝΙΚΑΖΕΝ: BM 1969.4–2.1, a new archaic inscription from Crete’, Kadmos 9 (1970), 118–54CrossRefGoogle Scholar; their interpretation of ποινικάζεν and ποινικαϲτάϲ as verb and noun of agent from this root was called in question by Beattie, A. J., Kadmos, 14 (1975), 8–47CrossRefGoogle Scholar, who proposed an etymology from ποινή, but his arguments are well rebutted by P., G. and Edwards, R. B., Kadmos 16 (1977), 131–40Google Scholar. Xenophon's ɸοινικιϲτήϲ (Arab. 1.2.20) is best interpreted as ‘scribe, secretary’ (pace LSJ); see Lewis, , Sparta and Persia, 25 n. 143Google Scholar; cf. Hsch. s.v. ⋯κɸοινίξαι.

⁵⁹ It is not clear whether this designation is Herodotus' own invention, nor whether it should be understood as ‘Letters of Cadmus’ or as ‘Letters of the Cadmeans’.

⁶⁰ ⋯ών has often been suspected, but a convincing emendation has not been suggested.

⁶¹ 1.51.3–4: see above, p. 280.

⁶² Ap. 1.10–12.

⁶³ Wolf, F. A., Prolegomena ad Homerum (Halle, 1795), §55Google Scholar.

⁶⁴ Thus Biesantz, H., Minoica: Festschrift J. Sundwall (Berlin, 1958), 50–60Google Scholar.

⁶⁵ So, among others, Hammond, N. G., A History of Greece to 322 B.C. ² (Oxford, 1967), 654Google Scholar, Fritz, K. v., Die gr. Geschichtsschreibung (Berlin, 1967), Anm.-Bd. 212 n. 114Google Scholar. For an account of the discovery see Touloupa, E., Kadmos 3 (1964), 25ff.CrossRefGoogle Scholar; a detailed study has been published by Porada, E., ‘The cylinder seals found at Thebes in Boeotia’ (with contributions by J. A. Brinkman and H. G. Güterbock), AOF 28 (1981/1982), 1–78Google Scholar. One of the seals (conveniently illustrated in Guarducci, M., Epigrafia Greca, i. 45Google Scholar fig. 2) bears the name of Kidin-Marduk, an official of the Kassite king Burnaburiash II, who ruled c. 1359–33; the Mycenaean pottery found in the vicinity of the seals cannot be much later than 1230.

⁶⁶ Some Semitic epigraphists favour a twelfth-century date: see Naveh, J., AJA 77 (1973), 1–8CrossRefGoogle Scholar, Early History of the Alphabet (Jerusalem, 1982), 175ff.Google Scholar, Cross, F. M., BASOR 238 (1980), 1ff.Google Scholar, Puech, E., RBi 90 (1983), 365–95 (esp. 391ff.)Google Scholar; for a more moderate view see Isserlin, B. S. J., CAH iii ²1.816–18Google Scholar. In defence of a mid-eighth-century date see Jeffery, CAH iii ² 1.819ff.Google Scholar, Heubeck, op. cit. 73ff.

⁶⁷ This appears to have been Tod's view (loc. cit. n. 1).

⁶⁸ ‘Letters of Cadmus’, AJP 56 (1935), 5ff.Google Scholar; Mantiklos inscription: Jeffery 90–1, Hansen CEG no. 326 (where further bibliography may be found).

⁶⁹ He does not comment on a further difficulty raised by his reconstruction, which presents the inscription running retrograde; this is what we should expect (see Jeffery 43ff.), but Herodotus' catalogue of the differences between Greek and Egyptian custom (2.36.4) shows that he was unaware that Greek script could run right to left.

⁷⁰ See in particular the final comparative Table of letters.

⁷¹ What he claims to have learnt there does not inspire confidence; cf. Fehling 88.

⁷² It is not difficult to account for Herodotus' belief that flying snakes were to be found in Egypt (see Lloyd ad loc.), but it is impossible to suggest what he might have seen that he could reasonably have mistaken for heaps of snake skeletons (ϲωρο⋯…τ⋯ν ⋯κανθέων); snakes do not congregate to die; see further Fehling 20–3. Herodotus relates further odd encounters with bones on the battlefields of Pelusium and Papremis (3.12). The case against the 345 wooden statues is more complicated: see Fehling 59–65.

⁷³ On the Greeks' own views about the origins of writing see Jeffery, L. H., ‘Ἀρχαῖα γράμματα’, Europa, Festschr. Ernst Grumach (Berlin, 1967), 152ff.Google Scholar, Heubeck, op. cit. 105ff.

⁷⁴ FGrHist I F 20, 9 F 3 (Anaximander), 687 F 1 (Dionysius of Miletus); all from the sch. on Dion. Thrac. p. 183, 1 (Hilgard). Given the differences between Greek and Egyptian writing, this view seems at first sight far-fetched. But certainly some centuries later what is now generally supposed to have been a Linear B tablet was submitted to an Egyptian priest for decipherment (Plut. de gen. Soc. c. 5 (577e–f), and perhaps a similar find in the archaic age was taken as evidence that a script based on Egyptian hieroglyphic was formerly used in Boeotia. Prometheus' claim to have taught men writing (A. PV 460f.) is not relevant; though it implies the advent of literacy before Io (and therefore five generations before Danaus) the Titan need not be supposed to confine himself to the development of Greek civilisation.

⁷⁵ See Jacoby on Hecataeus FGrHist I F 20; but cf. Edwards, op. cit. 26ff., whose genealogical tables show that many thought them to be cousins.

⁷⁶ Stesich. fr. 213; Gorg. Pal. 30; E. fr. 578. It is not entirely easy to square Palamedes' claims to have invented writing with the mechanics of Odysseus' plot to destroy him (as related by Hygin. fab. 105; cf. [Apollod.] Vat. Epit. 3.8, probably following Euripides), involving as it did forged correspondence with Priam; see further RE xviii. 2501f. It would be interesting to know what those who held this view made of the ϲήματα λυγρά of Il 6.168.

⁷⁷ Apart from Phaedra's letter (856ff.) two references to literature on mythological and religious topics should be noted (451, 954). Euripides exploited Theseus' literacy in the famous fragment in which an illiterate peasant describes the name written on the hero's sails (fr. 382) — in the Ionic alphabet, which was not to be officially adopted at Athens until after Euripides' death; similar conceits in Agathon F 4 (Snell) and Theodectas F 6 (Snell).

⁷⁸ Tr. 1166ff.; cf. 47 (if this is genuine), 157–8; F 514. Many suppose that the Poimenes can be dated to the 460s on the strength of P. Oxy. 2256 fr. 3. 4; ‘sed res incertissima est’ (Radt).

⁷⁹ Grassl, H., ‘Herodot u. die griech. Schrift’, Hermes 100 (1972), 169–75Google Scholar, suggests that this digression had a further topical relevance, inasmuch as Herodotus by adducing evidence that the Ionians had been the first Greeks to take over the Phoenician script might be supposed to be offering his support to those who pressed for the adoption of the Ionic alphabet at Athens.

⁸⁰ This section finds an odd sequel in [Arist.] Mir. c. 133 (843b), where we are told of the unusual palaeographical expertise displayed by the guardians of the temple of Apollo Ismenius when they succeeded in deciphering as a dedication by Heracles an inscription from Hypata in Thessaly which had baffled local talent.

⁸¹ The account of the Lydian alphabet in RE (xiii. 2157) adequately demonstrates this; for further details see Gusmani, R., Lydisches Wörterbuch (Heidelberg, 1964), 28–9Google Scholar, Ergänzungsbd. (1980), i. 18Google Scholar. Most Lydian inscriptions belong to the fourth century; it would not be surprising if the proportion of unfamiliar letters was somewhat lower earlier.

⁸² Given the close ties which had existed between Greece and Lydia in the sixth century, many Greeks might well have been able to do rather better than that, even in Herodotus' day; but one should be cautious in such estimates, in view of the notorious Greek aversion to learning foreign languages.

⁸³ It is strange that Herodotus seems not to consider the possibility of any link between Croesus and the fabulously rich Pythius, son of Atys, to whom he has just introduced us (27–9): Atys, we may recall, was the name of Croesus' heir apparent, who had married shortly before he was killed (1.34.3), and it is tempting to see in Pythius a grandson of Croesus (as was first suggested by Urlichs, L., RhM N.F. 10 (1856), 26–7)Google Scholar, his name commemorating the king's dealings with Delphi. (The story is of course a doublet of the tale of Darius and Oeobazus (4.83–4) and ends with a rite (39) which Herodotus did not understand.)

⁸⁴ No trace of writing is reported on the broken remains of what appears to be one of these οὖροι: see Hanfmann, G., Sardis from Prehistoric to Roman Times (Princeton, 1983), 56Google Scholar.

⁸⁵ See above pp. 281f.

⁸⁶ Particularly if the Persian king invariably drank the water of the R. Choaspes (1.188.1).

⁸⁷ On the Suss statue see above, n. 13.

⁸⁸ See Unger, E., ‘Die Dariusstele am Tearos’, AA 1915, 3–17Google Scholar. General Jochmus was told at Pinarhisar in 1847 that a stone with letters ‘like nails’ had stood there until fairly recently (JRGS 24 [1854], 43–4Google Scholar).

⁸⁹ See further Friedrich, J., Die Welt als Geschichte 2 (1936), 108Google Scholar. The main interest of this report lies in the hope which it inspires that Persian inscriptions may be found in the Balkans, a hope encouraged by the discovery at Gherla in Rumania of a clay-tablet apparently preserving the draft of a building inscription of Darius I; see Harmatta, J., AAntHung 2 (1954), 1–16Google Scholar, Mayrhofer, M., Supplement zur Sammlung der altpersischen Inschriften (Vienna, 1978), 16Google Scholar; it is of course possible that the tablet originated a very long way from the Transylvanian vegetable-garden where it was dug up.

⁹⁰ Contrast Herodotus' summary of the Persian curriculum (1.136.2) ἱππεύειν κα⋯ τοξεύειν κα⋯ ⋯ληθίζεϲθαι.

⁹¹ So Spiegelberg, W., Die Glaubwürdigkeit von Herodots Bericht über Ägypten im Lichte der ägyptischen Denkrmäler (Heidelberg, 1926), 25–6Google Scholar (English translation by Blackman, A. W., The Credibility of Herodotus' Account of Egypt in the Light of the Egyptian Monuments [Oxford, 1927], 25Google Scholar). S. pointed out that Herodotus' account of Sethos (Sethon ?) (2.141), who was miraculously delivered from Sennacherib by small rodents with an appetite for leather, rests in part on a similar misunderstanding: the statue of a man holding a mouse (sacred to Horus of Letopolis and most likely marking the person depicted as that god's high priest) has given a strange turn to the tradition familiar to us from the Old Testament (ii Kings xix. 8ff.) of a divine visitation which confounded the apparently invincible Assyrian host. (Herodotus' version finds a curious parallel in Khotanese cult-legend: see SirStein, M. Aurel, Ancient Khotan [Oxford, 1907], i. 119–20Google Scholar.)

⁹² The horse and rider motif is well documented on seals, coins, and luxury objects: see Farkas, A., ‘The horse and rider in Achaemenid art’, Persica 4 (1969), 57–76Google Scholar, Root, M. C., The King and Kingship in Achaemenid Art, 129fGoogle Scholar. In an Aramaic letter, dated to the reign of Darius II, the satrap Aršāma orders a representation of a horse and its rider from a craftsman named Hinzani (see Grelot, P., Documents araméens d'Égypte [Paris, 1972], no. 70, pp. 318–19Google Scholar; Driver, G. R., Aramaic Documents of the Fifth Century B.C. [Oxford, 1957], Letter ix, pp. 32, 71–4Google Scholar), but, though it has often been assumed that Hinzani was a sculptor, there is no indication of the medium in which he worked, and he might have been a painter or seal-cutter: see Roaf, M., Iran 18 (1980), 72CrossRefGoogle Scholar.

⁹³ Herodotus is not very informative about Egyptian writing. He tells us that there are two scripts, known as sacred and common (ἱρά and δημοτικά, 2.36.4), and describes the inscription on the Karabel relief, which he wrongly takes to be Egyptian (see below pp. 300ff.), as written in γράμματα ἱρ⋯ Αἰγύπτια; but he does not elaborate on the nature of the two writing systems and the differences between them. (He should of course have distinguished three scripts, hieroglyphic, hieratic, and demotic: see further Lloyd ad loc.)

⁹⁴ My discussion of Herodotus' account of Sesostris is heavily indebted to Fehling (esp. 15–17, 98–101) and to the important article by Armayor, O. K., ‘Sesostris and Herodotus' autopsy of Thrace, Colchis, inland Asia Minor, and the Levant’, HSCP 84 (1980), 53–74Google Scholar.

⁹⁵ See further Lloyd, A. B., ‘Nationalist propaganda in Ptolemaic Egypt’, Historia 31 (1982), 33ff. (esp. 37–40)Google Scholar; however, Lloyd seems to me to assume too readily that Herodotus faithfully mirrors Egyptian tradition at all points.

⁹⁶ Herodotus takes pains to stress that these chapters form a continuous and unitary account: note the repeated ἔλεγον (both with and without οἱ ἱρέεϲ) at 2.99.2, 100.1, 101.1, 102.2, 107.1, 109.1, 111.1, 112.1, 113.1, 116.1, 118.1, 120.1, 121.1, 122.1, 124.1, 127.1, 129.1, 136.1, 139.1.

⁹⁷ The question is discussed by Lloyd, , Introduction 89–114Google Scholar, who does not regard the obvious historical ignorance of Herodotus' informants as grounds for denying them clerical status, or even for adopting the compromise favoured by, among others, Spiegelberg, that Herodotus tended to mistake the verger for the Dean. But a more serious objection lies in their religious ignorance, well exemplified in the moving story of Mycerinus, which is largely based on misunderstanding of rituals in honour of Osiris (129.3–130, 132–133), raising the further question of what sort of priests ever professed to serve gods who punish a just king for his just dealing. See further Fehling 54ff.

⁹⁸ Tacitus, in his account of the origins of Serapis (Hist. 4.83), notes that even in the Hellenistic age Egyptian priests knew little of foreign parts, with particular reference to their ignorance of the Black Sea area (‘Ponti et externorum parum gnaris’) – practically the only point in this Königsnovelle to inspire confidence.

⁹⁹ ‘Herodotus seems to be referring to actual monuments in Thrace, but unluckily he does not say what they were; the conclusion he bases on them is obviously false, but this does not prove that he had not seen them’ (How and Wells).

¹⁰⁰ Μέχρι μ⋯ν τούτου ⋯ψιϲ τε ⋯μ⋯ κα⋯ γνώμη κα⋯ ἱϲτορίη τα⋯τα λέγουϲά ⋯ϲτι, τ⋯ δ⋯ ⋯π⋯ το⋯δε Αἰγυπτίουϲ ἔρχομαι λόγουϲ ⋯ρέων κατ⋯ τ⋯ ἤκουον· προϲέϲται δέ τι αὐτοῖϲι κα⋯ τ⋯ϲ ⋯μ⋯ϲ ⋯ψιοϲ. It should be noted that this last clause upsets the balance of the sentence and badly obscures the distinction drawn in its earlier part; the same expression recurs at 2.147.1, where it is perfectly in order. I wonder if it was added here by a scribe who observed that Herodotus did not in fact confine himself to the traditions of the native priesthood.

¹⁰¹ Though a Greek traveller already convinced that Sesostris had passed that way might have persuaded himself that among the numerous megalithic monuments of south-eastern Bulgaria were weathered Egyptian stelae, I suspect the answer to this enigma lies not in archaeology but in lost literature. Since the origins of Colchis, which Herodotus regards as Sesostris' foundation (2.103.2–105), were relevant to the saga of the Argonauts, I wonder whether Herodotus here develops something drawn from Hecataeus, who dealt with the Argonauts in his Genealogies (FGrHist 1 F 17, 18) – perhaps combining Hecataean speculation with Egyptological information derived from elsewhere.

¹⁰² Did Herodotus imagine that Sesostris' next advance would have taken him into Greece, so that the lack of any evidence of his presence along the route leading via Tempe to Thermopylae could be deemed a strong argument from silence?

¹⁰³ Sesostris' campaign is seen as an isolated expedition; though he brings home some prisoners (2.107.1; 108.1), he makes no attempt to consolidate his conquests.

¹⁰⁴ On the Egyptian reliefs from Nahr-el-Kelb see Weissbach, F. H., Die Denkmäler u. Inschriften an der Mündung des Nahr-el-Kelb (Berlin u. Leipzig, 1922), 17–22Google Scholar. There are other possibilities, such as the commemorative stelae of Seti I and Rameses II from Beth-Shan and the fragment of a monumental stele from Megiddo bearing the name of Sheshonk I: see Pritchard, J. B., Ancient Near Eastern Texts relating to the Old Testaments ³ (Princeton, 1969), 253–5, 263–4Google Scholar.

¹⁰⁵ For detailed discussion of this monument see Cook, J. M., ‘The reliefs of “Sesostris” in Ionia’, Türk Arkeoloji Dergisi vi. 2 (1956), 59–65Google Scholar, Bittel, K., ‘Karabel’, MDOG 98 (1967), 5–23Google Scholar; there is a good photograph in Macqueen, J. G., The Hittites and their Contemporaries in Asia Minor (London, 1975), Plate 4Google Scholar.

¹⁰⁶ Above, pp. 283–5.

¹⁰⁷ ‘It is certain that neither the road from Smyrna to Sardis, nor that from Ephesus to Phocaea could have gone through this pass, which is very far from the proper track’ observed SirRamsay, William, JHS 2 (1882), 53Google Scholar; see further Armayor, op. cit. 69ff.

¹⁰⁸ On hieroglyphic Hittite see Gurney, O. R., The Hittites ² (Harmondsworth, 1980), 128ff.Google Scholar, Plate 26, Macqueen, op. cit. 23ff.

¹⁰⁹ 2.125.6; 136.4; 141.6.

¹¹⁰ Bergk introduced Sesostris by conjecture in Hipponax fr. 42: an ingenious guess, but far from certain.

¹¹¹ We shall not have to look far for a parallel: compare the blatantly circular chronological argument employed against the attribution of Mycerinus' pyramid to the courtesan Rhodopis (2.134.2–3).

¹¹² See further Lloyd, G. E. R., Magic, Reason and Experience (Cambridge, 1979), 15ff., 86ff.Google Scholar

¹¹³ Egypt is clearly treated as a special case, on account of the Egyptians' own care to preserve knowledge of the past: cf. 2.77.1: αὐτ⋯ν δ⋯ δ⋯ Αἰγυπτίων οἳ μ⋯ν περ⋯ τ⋯ν ϲπειρομένην Αἴγυπτον οἰκέουϲι, μνήμην ⋯νθρώπων πάντων ⋯παϲκέοντεϲ μάλιϲτα λογιώτατοί εἰϲι μακρῷ τ⋯ν ⋯γὼ ⋯ϲ διάπειραν ⋯πικόμην. See further von Leyden, W. M., ‘Spatium historicum’, DUJ n.s. 11 (1949–1950), 89–104Google Scholar (= Herodot: eine Auswahl aus der neueren Forschung ed. Marg, W. [Darmstadt, 1962], 169–81Google Scholar).

¹¹⁴ From this I believe it follows that we should be wary of arguments from Herodotus' silence where inscriptions are involved (as, for instance, from his evident ignorance of the archon-list, publicly displayed from c. 425).

¹¹⁵ For a rather different view of the relationship between Herodotus' account and the inscription recorded by Pausanias see Fehling 124.

¹¹⁶ The difficulty of properly delimiting the respective contributions of oral tradition and epigraphic evidence emerges clearly in the interesting article by Virgilio, B., ‘Atleti in Erodoto: tradizione orale e (possibile) tradizione epigrafica’, RIL 106 (1972), 451–68Google Scholar.

¹¹⁷ Discussion of the diffusion of literacy in Greece owes much to Havelock, E. A.: see, in particular, Preface to Plato (Oxford, 1963)Google Scholar, Prologue to Greek Literacy (Univ. of Cincinnati, 1971)Google Scholar, The Literate Revolution in Greece and its Cultural Consequences (Princeton, 1982)Google Scholar.

¹¹⁸ See, by way of introduction, Goody, J. (ed.), Literacy in Traditional Societies (Cambridge, 1968)Google Scholar, esp. J. Goody and I. Watt, ‘The consequences of literacy’, 27–68, Henige, D., The Chronology of Oral Tradition (Oxford, 1974)Google Scholar, Oral Historiography (London, New York and Lagos, 1982), esp. 1–22, 80–105Google Scholar, Ong, W. J., Orality and Literacy (London, 1982)CrossRefGoogle Scholar.

¹¹⁹ Goody well illustrates this last point with the Tiv genealogies from Nigeria (even more relevant to the study of Homer and Hesiod than to Herodotus). These stretch back some twelve generations to an eponymous ancestor, and serve as mnemonics for systems of social relations. At the turn of the century they were recorded by the British administration, but forty years later these written pedigrees occasioned considerable dissension: ‘the Tiv maintained that they were incorrect, while the officials regarded them as statements of fact, as records of what actually happened, and could not agree that the unlettered indigenes could be better informed about the past than their own literate predecessors’.

¹²⁰ Goody, op. p cit. 34.

¹²¹ However, even for those who hold a more conventional view of Herodotus' conception of his task consideration of the various hazards to be encountered by modern researchers in African history (on which see Henige, Oral Historiography passim) should demonstrate that the exploitation of oral tradition is not the straightforward matter that some Herodotean scholars appear to suppose.

¹²² Hydriotaphia ch. 5 (the five languages being Greek, Latin, Hebrew, Egyptian and Arabic); Simon. PMG 581 (quoted ad init.).

¹²³ Shakespeare, , Sonnet 81Google Scholar.

* A version of this article was read to the Oxford Philological Society in November 1983. It benefited considerably from the ensuing discussion and has since been much improved by the kindness of various friends, in particular Peter Hansen, Roger Moorey, John Penney, Susan Sherwin-White, my husband, Martin West, and, above all, of David Lewis, who drew my attention to much that I would otherwise have missed and patiently extricated me from numerous perplexities.