Twice-occurring terms in Herodotus: random distribution, habit of presentation and deliberate pairings

Abstract

Repetition is a critical issue in interpreting the work of Herodotus. Detlev Fehling, for one, has pointed to recurrence of motif and scene as evidence of the historian’s ‘free invention’. Words that occur twice in Herodotus are an efficient way to consider pressing issues at the centre of how and why Herodotus put together his narrative in the way he has. Pairs where the uses are close together in stories with a lot in common suggest that we may be seeing Herodotus’ ‘habit of presentation’, especially when phrasal repetition is also found. Where pairs are found further apart, the issue of deliberate linkage between discrete episodes may be indicated through the strategic redeployment of a key term. Finally, with Xerxes’ invasion, recurring terms help us to see how Herodotus could operate over large portions of text, deliberately linking one episode to another through the deployment of twice-occurring words, thereby also connecting the whole account of the campaign to the largest project of the History.

I. Introduction

When we see two near-identical episodes told in the same, or almost the same, language in two separate but near-identical contexts, in the same historical text, something other than charting what really happened must be at work. The problem becomes acute when, after surveying the entire narrative of a historian, several repetitions emerge: the author does not occasionally lapse into a duplication but does so frequently.¹

Subsidiary to the problem of what we might call narrative ‘patterning’ is the issue of intention on the part of the author. When we detect an unlikely repetition, do we understand the author as simply being unaware that a duplicate has been written into the account, or is the duplication something that is meant to be observed? To take a well-known example from Herodotus: out of contingents of 300 men, 299 Spartans died at the battles of Thyrea in 546 and Thermopylae in 480, with one survivor from each (1.82; 7.224, 229–30). Did these battles really produce identical results, to say nothing of the other parallels between them recorded by Herodotus? If they did not, and Herodotus is responsible for the parallelism (either as inventor himself or as a reteller of accounts), did he mean for the battles to mirror each other so closely, or was he thinking along strikingly similar lines, unconsciously, when he treated both episodes?² The same worries arise in connection with larger portions of narrative: why, for instance, do the careers of Cambyses and Cleomenes look so much alike?³

Operating at another level from motif and scene repetition is word repetition. In Herodotus, or in any other Greek author, a great deal of repetition is presumably due simply to the structures and preferences of the ancient Greek language. When at 1.2.1 Herodotus wrote ἴσα πρὸς ἴσα, it is difficult to know how conscious he was of the repetition of ἴσα;⁴ and, in any case, settling that question does not seem to provide us much information about Herodotus. Similarly, while we can see the features of lexis eiromenē (cf. Arist. Pol. 1409a) on every page of the History, a type of composition that is heavily dependent on repetition, determining whether Herodotus was aware of the repetitions in, say, the story of Candaules (ἠράσθη … ἐρασθϵὶς δὲ ἐνόμιζϵ … νομίζων, 1.8.1),⁵ does not really help us to understand why he structured that logos in the way he did, or what his purpose was in reporting it. More illustrative of how Herodotus conceived of at least the boundaries of his various stories is ring composition and its reliance on the repetition of key terms to introduce and conclude discrete logoi (for example, κατϵχόμϵνον … κατέχοντα, at 1.59.1 and 65.1).⁶

Where repetition of words in Herodotus is particularly illuminating and at the same time also manageable is in connection with twice-occurring terms.⁷ It is a curious fact that, as Detlev Fehling observed regarding motif repetition,⁸ a number of twice-occurring terms in Herodotus are found in close proximity to each other and nowhere else in his text.⁹ In these cases different interpretations are possible. When the terms are synonymous for words that are found elsewhere in the History, it is tempting to speculate that we have something akin to epic ‘clustering’, namely, the favouring of a word by Herodotus for a distinct period during the composition of his text, after which the term was dropped.¹⁰ These cases would seem best understood as ‘unfigured’ or ‘accidental’ repetition.¹¹ In places where the repeated terms are also accompanied by identical or near-identical words and phrases (‘phrasal repetition’),¹² it seems reasonable to suppose that we are being provided a view of regular associations Herodotus made, that is, his habit of mind when narrating events that were similar or that he saw as similar. In other places, especially where the terms in the pair are separated by a substantial amount of text, the repetitions are less likely to be ‘accidental’ and more likely to be purposeful on Herodotus’ part.¹³ It is possible, in the context of longer-range repetitions, to examine whether some of the pairs of repetitions are interconnected, creating a network or ladder of single-recurring terms whereby different episodes are linked together in chains of associated terms.

In the following sections of this paper, I will examine several unique pairs of words in Herodotus. I begin with close-proximity pairs and how they are particularly illustrative of clustering, as well as providing evidence for habits of Herodotean presentation. In subsequent sections I look at pairs of terms that are more widely separated, twice-occurring words that raise successively with greater clarity the issue of deliberate connection between uses; in particular, how connected pairs could help Herodotus structure his account; and, finally, how some pairs in books 7–9 can be seen to connect with each other to assist Herodotus in articulating the key moments in the campaign of Xerxes. The paper concludes with a brief consideration of the importance of twice-occurring terms in understanding Herodotus’ historical writing.

II. Hdt. 3.129–32, 134: the Democedes logos and close-proximity repetition

Darius badly injures his foot dismounting from a horse (στραφῆναι τὸν πόδα, 3.129.1). Egyptian doctors attempt to heal him but only make matters worse. On the eighth day after the injury, with Darius seriously ailing (ἔχοντί οἱ φλαύρως), he learns about the presence in Sardis of the doctor Democedes and sends for him:

When they located him among the captives of Oroetes, somewhere or other, utterly uncared for (ἐν τοῖσι Ὀροίτϵω ἀνδραπόδοισι ὅκου δὴ ἀπημϵλημένον), they brought him before the king (παρῆγον ἐς μέσον), dragging his fetters (πέδας τϵ ἕλκοντα) and clothed in rags. Having been made to stand in the middle (σταθέντα δὲ ἐς μέσον), Darius asked him if he knew the craft [of medicine] (τὴν τέχνην ϵἰ ἐπίσταιτο). He did not admit it, fearing that if he revealed himself (ἑωυτὸν ἐκφήνας) [as knowing medicine], at once he would be in a state of having been deprived of Greece. But it was revealed to Darius that he was practising deceit, knowing [the craft as he did] (κατϵφάνη … τϵχνάζϵιν ἐπιστάμϵνος),¹⁴ and he ordered the men who had fetched Democedes to bring whips and goads into the middle (ἐς τὸ μέσον) [of everyone there]. Then indeed he reveals (ὁ δὲ ἐνθαῦτα δὴ ὦν ἐκφαίνϵι) [himself], having stated that he did not have exact knowledge, but that having associated with a doctor, he knew the craft to a slight degree (φλαύρως ἔχϵιν τὴν τέχνην).

The passage is full of repetitions, but not of the lexis eiromenē type. It pivots around the repetition of compound forms of φαίνω (ἐκ- and κατά-): Democedes initially is unwilling to reveal himself as a skilled doctor, but Darius knows that he knows the craft; he is threatened with torture, and even then only partially admits to having medical knowledge (τέχνη).¹⁵ All the action takes place ἐς (τὸ) μέσον, in the ‘midst’ of everyone, but this means, in the context, in the presence of the king.¹⁶ The vividness of the scene is underscored by the repetition of the action as happening ‘in(to) the middle’, encouraging the audience to ‘see’ what is going on.¹⁷ The entrance of the doctor is particularly vivid, stressing his abject state and balancing his immobility with Darius’ (Darius has injured his foot badly, and the captive Democedes is similarly unable to walk properly, ‘dragging the fetters’ on his feet).¹⁸ The added detail that Democedes is ‘clothed in rags’ and halting in his walk has even suggested to some the figure of the tragic Telephus.¹⁹

Democedes’ backstory is given at 3.131, and at 132 we return to the main narrative with the events following Democedes’ treatment of Darius: having thoroughly healed Darius (ἐξιησάμϵνος), Democedes was given a large house and was made a table companion of the king, enjoying all sorts of privileges ‘except return to the Greeks’; when the Egyptian doctors who had tried to heal Darius (ἰῶντο: conative imperfect) were about to be impaled for having been bested by the Greek doctor, Democedes begged for their release from the king and ‘rescued’ them (ἐρρύσατο). But this was not Democedes’ only act of salvation: ‘On the other hand (τοῦτο δέ), he rescued (ἐρρύσατο) an Elean mantis who attended Polycrates and who had been utterly uncared for among the captives (ἀπημϵλημένον ἐν τοῖσι ἀνδραπόδοισι)’ (3.132.2). The two cases of Democedes’ help for others are set off by τοῦτο μέν and τοῦτο δέ and are punctuated with the same word ‘rescued’ (ἐρρύσατο). Evidently, Democedes and the mantis had been fellow captives, both having earlier been in the retinue of Polycrates of Samos and then wound up under the control of Oroetes (cf. 3.125.1–3).²⁰ The verb ἀπαμϵλέω occurs in only these two places in all of the History, and in identical form. Furthermore, word repetition is accompanied by phrasal repetition: the participle ἀπημϵλημένον is modified by the very same words in slightly different order, describing the plight of both men (Democedes: ἐν τοῖσι … ἀνδραπόδοισι … ἀπημϵλημένον, 3.129.3; mantis: ἀπημϵλημένον ἐν τοῖσι ἀνδραπόδοισι, 3.132.2). Later, when Atossa identifies Democedes as the perfect man for providing Darius information about Greece, she refers to him as the man ‘who thoroughly healed your foot’ (ὅς σϵυ τὸν πόδα ἐξιήσατο, 3.134.5). This and the participle ἐξιησάμϵνος at 3.132.1 are likewise the only two places where the verb ἐξιάομαι is found in the History.

It is important to note that we are never told why Democedes sought the release of both the Egyptian doctors and the Elean mantis. Herodotus seems to want us to extrapolate the answer from the narrative itself. Unlike the Egyptian doctors, the seer has not been mentioned before, and once introduced, he is dropped from the account, never to reappear. Why mention him at all then? Scholars have pointed to the solidarity that Democedes evidently felt with his Egyptian colleagues.²¹ The same perhaps can be said for the mantis. Unlike the Egyptian doctors, he is described in identical terms with Democedes. Perhaps Herodotus was emphasizing the point that Democedes’ life produced commonalities, shared experiences, with others, and that he acted upon the empathy that these experiences aroused. One is reminded of the pity that Cyrus feels for Croesus when he recognizes (ἐννώσαντα) that he is at the point of burning alive another man who had been no less fortunate than he (αὐτὸς ἄνθρωπος ἐὼν ἄλλον ἄνθρωπον γϵνόμϵνον ἑωυτοῦ ϵὐδαιμονίῃ οὐκ ἐλάσσω, 1.86.6);²² or, less spectacularly, those at Babylon who give advice to the ill, explaining what they had done to survive a similar illness or seen another do (1.197).

If indeed empathy is a major theme of the logos, it is only the clearest example of a larger point that Herodotus seems to want to make through his telling of the story of Democedes’ life. Democedes’ own personal history produced ripple effects that profoundly affected the lives of many others: individuals, groups, cities, even empires.²³ Because of their similar plights (it seems), Democedes seeks the release of the Elean mantis; because of Democedes, Crontoniate doctors became renowned (ἀπὸ τούτου τοῦ ἀνδρὸς οὐκ ἥκιστα Κροτωνιῆται ἰητροὶ ϵὐδοκίμησαν, 3.131.2); and of course, Democedes heals Atossa, but thereby sets in motion the Persian reconnoitring of Greece that becomes preparatory to invasion.²⁴

The critical issue is the effect that ἐξιάομαι and ἀπαμϵλέω have in the story and to determine (if possible) why the use of both verbs is restricted to this passage. To take ἐξιάομαι first, the prefix ἐκ- denotes the thoroughness of Democedes’ treatment of Darius (cf. LSJ s.v. ἐκ C.2), in contrast with the Egyptian doctors (indeed, note that their activity is described with the simplex ἰῶντο, furthermore imperfect because conative: ‘they tried to heal’). A crucial element in the story is precisely that Democedes succeeds as a healer where the Egyptians fail. Hence the compound ἐξιάομαι is entirely apt. As for ἀπαμϵλέομαι, it too has a pre-verb that indicates intensity: ‘to be neglected utterly’ (LSJ). It is more difficult to see why it is limited in its use to this logos.²⁵ Elsewhere Herodotus tells stories about people who are explicitly identified as ‘fellow slaves’ (σύνδουλοι: Mitradates and Spaco, 1.110.1; Rhodopis and Aesop, 2.134.3), though admittedly in neither case is the idea of ‘neglect’ important.

The double use of ἀπαμϵλέω, together with the phrase ἐν τοῖσι ἀνδραπόδοισι, in connection with Democedes and the Elean mantis, whose release from identical circumstances Democedes secures for no other reason, apparently, than altruism born from the same experience, may be deliberate on Herodotus’ part, but I do not believe it is possible to tell. So, too, with the double use of ἐξιάομαι. Both ἀπαμϵλέω and ἐξιάομαι are highly specific, almost technical terms that Herodotus wanted for this particular logos and evidently did not feel the need for elsewhere. While there are other Leitwörter running through the logos (note, for instance, the prominence of the deik-root (‘show’): 3.133.1; 134.1, 2, 3, 5; 135.2),²⁶ they are not unique to the account, whereas these are. It is worth pointing out here that many of the unique pairs I examine in this paper are precisely compounds, verbs, nouns and adjectives.

III. Hdt. 1.34.3 and 209.3, 1.115.1 and 3.62.2: remote pairings and the habits of Herodotean presentation

At 2.119.3 and 7.191.2, and only in these passages, Herodotus uses the same expression to refer to sacrifices for the calming of winds: ἔντομα ποιέϵιν.²⁷ Both scenarios concern the same mythological circumstance: the abduction of a minor female deity (Helen, Thetis). The very specificity of the phrase and the purpose behind the sacrifice in both cases, together with the near-identical status of the deities involved, renders the suggestion that the pairing is random extremely unlikely if not in fact impossible.²⁸ And yet, does Herodotus want to connect the two passages together? That, too, seems improbable: after so much narrative space, two otherwise relatively minor details connected? To what purpose? It seems, rather, that we are viewing a recurring set of associated ideas and terms. By contrast, it is worth pointing out that Herodotus uses the noun μῦθος only twice, both in book 2, and in both cases of Greek ‘stories’ whose truth he cannot accept: on the existence of the river Ocean surrounding the earth (2.23), and in connection with the ‘silly tale’ (ϵὐήθης … ὅδϵ ὁ μῦθος) the Greeks tell which has the Egyptians attempt to sacrifice Heracles (2.45.1).²⁹ In this case, the pair reveals a critical stance taken by Herodotus that seems to show him as capable of critiquing his own culture’s traditions from the vantage point of another ancient society. While it is difficult to determine with certainty if the reader is meant to connect the passages, it seems likely both are intended to be seen as tacit criticisms of Hecataeus of Miletus (cf. FGrH 1 F 1 Ἑκαταῖος Μιλήσιος ὧδϵ μυθϵῖται).³⁰

An important pair of recurrent terms and phrases is also found at 1.34.3 and 209.3, in passages that are of much greater narrative significance. In the first passage, we are told that ‘a great nemesis from the divine’ took hold of Croesus, presumably for his belief that he was the happiest of men. Herodotus tells us that Croesus dreamt that his son Atys would be killed, struck by an iron point; ‘when he awoke (ἐπϵίτϵ ἐξηγέρθη) and considered the matter [lit. ‘gave a word to himself’: ἑωυτῷ λόγον ἔδωκϵ], in dread of the dream’, he found a wife for his son and stopped sending him out on dangerous missions (combat, hunting). In the second passage, having just crossed the river Araxes in preparation to do battle with Queen Tomyris, Cyrus has an admonitory dream about the eldest son of Hystaspes, Darius, that seems to show a winged Darius in the sky casting shadows over Asia and Europe, forecasting that he was destined to rule over those places; ‘when Cyrus awoke, he was giving thought to his dream’ (ἐπϵὶ ὦν δὴ ἐξηγέρθη ὁ Κῦρος, ἐδίδου λόγον ἑωυτῷ πϵρὶ τῆς ὄψιος).³¹ Remarkably, these are the only two passages in the History featuring the verb ἐξϵγϵίρομαι, in the same form, though rising from sleep is an important detail elsewhere.³² What is more, structural parallelisms are also very much in evidence: both descriptions of waking are set off by the temporal clauses launched by ἐπϵίτϵ/ἐπϵί, and both are followed by the identical expression for ‘consider/think’, namely, δοῦναι ἑωυτῷ λόγον, involving the interpretation of a dream. Again, we have both recurring word and phrasal repetition together.

And, of course, the similarities can be seen to run deeper than that. Both episodes concern eastern monarchs who are about to face catastrophe, signalled either literally in the case of Croesus (stated at 1.34.1), or figuratively in the case of Cyrus (the crossing of the Araxes); both dreams concern young men who ought to be or will be ultimate successors to their thrones. Moreover, each king’s response to his dream is misguided: because of their limited understanding, both attempt pre-emptive measures to avoid the portended outcomes of their dreams that must fail. Croesus is persuaded to believe that a boar hunt will not involve iron-pointed weapons, and Cyrus is convinced that Darius is plotting to overthrow him, when in fact it is Darius’ future accession to the throne of Persia that is being foretold in his dream. It is a notorious problem that the ‘great nemesis’ that seized Croesus, namely the death of his son Atys, in fact does not explain Croesus’ decision to invade Persia and therefore his defeat or loss of power.³³ But the parallel dream in the case of Cyrus does portend his loss of power, in the form of Darius’ eventual succession to the Achaemenid throne, ending the direct line descended from Cyrus. Reading 1.34.3 retrospectively through 1.209.3 demonstrates what is only intimated indirectly as regards Croesus: personal catastrophe (loss of Atys) suggests or even implies more general disaster later (the Persian conquest of Lydia).

While it is tempting to understand 1.34.3 and 209.3 as intentionally parallel, it is not possible to determine with certainty whether the use of ἐξϵγϵίρομαι in both scenes, in similar contexts, with supporting phrasal repetition, is something Herodotus deliberately crafted and wanted us to note, thereby establishing a linkage between the two kings. But if it is not possible to be certain that the parallel is one Herodotus intended, that there is a connection through the single recurrence of term and phrase is indisputable and demonstrates an even more important point: Herodotus viewed Croesus and Cyrus similarly and evidently could not help but construct his narratives about them in ways that mirror each other in very precise ways.

As already observed, compounds easily make up the largest number of twice-recurrent terms in Herodotus. There are exceptions, however. κορέω, for instance, is found at 1.214.5 and in the Constitutional Debate at 3.80.4. Another case involves the verb βλέπω, in passages that are even further apart than I have so far examined. When the Median nobleman Artembares complains to Astyages about the rough treatment his son received from the son of the shepherd, he sends for both of them and confronts the young Cyrus with the facts (1.115.1–2):

When they were both present, having turned his eyes on Cyrus (βλέψας πρὸς τὸν Κῦρον), Astyages said: ‘You, being the son of this man, being the sort of person he is, dared to treat with such injury as this (ἀϵικϵίῃ τοιῇδϵ) the son of this man who is first at my court?’ But he answered him as follows: ‘O lord, I did these things to this one with justice (ὁ δὲ ἀμϵίβϵτο ὧδϵ; Ὦ δέσποτα, ἐγὼ δὲ ταῦτα τοῦτον ἐποίησα σὺν δίκῃ)’.

In book 3, in the account of the usurpation of the Magi, a herald makes a proclamation to the Persian army in Egypt that Smerdis has assumed the throne of Persia and that Cambyses is not to be obeyed (3.62.2–3):

Cambyses having heard these things from the herald, and having assumed that he was speaking the truth and that he had been betrayed by Prexaspes (that he, sent out to kill Smerdis, had not done this), having turned his eyes upon Prexaspes (βλέψας ἐς τὸν Πρηξάσπϵα), said: ‘Prexaspes, thus for me you carried out the task which I set before you?’ But he answered: ‘O lord, these things are not true (ὁ δὲ ϵἶπϵ· Ὦ δέσποτα, οὐκ ἔστι ταῦτα ἀληθέα), that at some time or another your brother has revolted against you, or that there will be any trouble great or small from that man’.

It is important first to note the formal similarities between the passages: in each the only two uses of βλέπω in the History are found in identical form (βλέψας) governing a prepositional phrase (πρὸς τὸν Κῦρον, ἐς τὸν Πρηξάσπϵα), with a main verb of speaking in a past tense (ἔφη, ϵἶπϵ). Both Astyages and Cambyses ask accusatory questions that are not really information-seeking but rhetorical: ‘You, low-born as you are, dared to injure …?’; ‘So this is how you performed my order …?’³⁴ The reply of the person interrogated by the king begins with the same words, Ὦ δέσποτα, a particularly marked form of address because it is so deferential and indicative of significant asymmetry in the status of the persons involved.³⁵

There are also several substantive parallels. In both cases, the king had earlier sought to engineer a dynastic murder of a near relation: Astyages had ordered Harpagus to put to death his infant grandson Cyrus, Mandane’s child about whom he had been warned in a pair of dreams (1.107–08); likewise, Cambyses, also warned in a dream (3.30.2), sent Prexaspes back from Egypt to Persia to kill his brother Smerdis. It is true that there is a significant difference at this point: whereas Harpagus failed to carry out his mission to kill the infant Cyrus, Prexaspes did murder Cambyses’ brother Smerdis.³⁶ Nonetheless, in both episodes, the king is shown making an incorrect assumption: Astyages assumes that Cyrus is the low-born son of a shepherd and not a royal prince and his own heir; Cambyses assumes Smerdis is still alive and has taken his throne. As noted already, the initial responses of the individuals to the kings’ indignant questions are obviously identical. Furthermore, both consist of assertions that, despite what the monarch might believe, are true. Since, as Cyrus notes, he was made ‘king’ by his age-mates in their royalty game (115.2 μϵ … παίζοντϵς … ἐστήσαντο βασιλέα; cf. 114.1 παίζοντϵς ϵἵλοντο … βασιλέα), making Artembares’ son, strictly speaking, insubordinate (114.3), the boy Cyrus did in fact act ‘with justice’ (σὺν δίκῃ) in punishing him.³⁷ And, of course, Prexaspes really did kill Cambyses’ brother (3.30.3), so that Cambyses in fact had nothing to fear from that quarter.

Given that tyrannical figures in Herodotus routinely fail to see the truth, and furthermore often are shown to be or attempting to be violent towards their own kin and subordinates, it is difficult to argue that Astyages and Cambyses are being deliberately linked by Herodotus through these scenes of confrontation and interrogation. But, if not deliberate, the two scenes are undeniably scripted in very similar ways. Both scenarios have the same initial circumstance: the king hears (ἀκούσας, 1.115.1, 3.62.2) of a pressing matter (the dishonouring of Artembares through the maltreatment of his son; the assumption of the throne by ‘Smerdis’); he questions the one person in a position to clarify what has happened. In both cases, the ‘turning of his gaze’ upon that person (βλέψας) is a moment that Herodotus chooses to mark out as significant. It seems that he wants to show us the king take in the pressing information and process it, and not just in any context, but in the presence of the informant who is able to reveal the truth. The outcome is in both cases a reaction by the king that will lead to self-destructive action (for Astyages, the alienation of Harpagus through his punishment with the Thyestean feast; for Cambyses, his determination to return at once to Persia and his self-wounding). Indeed, it is tempting to speculate that in both cases the king is meant to be seen as illustrating the proverb ‘seeing, they were seeing in vain/not seeing’ (cf. [Aesch.] PV 447 βλέποντϵς ἔβλϵπον μάτην).³⁸

The similarities between 1.115.1 and 3.62.2 are substantial and structural in nature. The two uses of βλέψας, followed by the interrogations and answers of Cyrus and Prexaspes and beginning with the phrasal repetition Ὦ δέσποτα, reveal how Herodotus constructed two similar episodes along similar lines, without necessarily meaning for them to be connected in any way. It is important, however, to consider another possibility. Mabel Lang has noted the similarities between the stories of Harpagus and Prexaspes, and has proposed that in the case of the latter, what she labels the ‘Janus-agent pattern’ was ‘perverted’: that originally, instead of killing the real Smerdis, Prexaspes did not obey Cambyses’ order to kill his brother, and because of that, he, like Harpagus, was punished with the murder of his son for his disobedience; otherwise, the grim details of both men’s careers are virtually identical.³⁹ If a common story type lay behind the accounts of the two courtiers, that may explain why Herodotus constructed his narratives involving them in such similar ways. However, even this possibility does not in my mind account for the presence of the verb βλέπϵιν in both and only these passages, together with the Ὦ δέσποτα statements.

It should be added that there are other pivotal twice-occurring words and phrases in logoi that are centred on dynastic matters relating to Darius. For example, the verb οἰδέω in the phrase οἰδϵόντων τῶν πρηγμάτων (‘with affairs being in ferment’) is only found just before the coup of the seven conspirators and just after the installation of Darius as king (3.76.2, 127.1), in both places with unsettled conditions being pointed to as the reason for hindering action that Darius wishes to undertake. At 3.76.2, Otanes and his group wish to delay the coup attempt against the Magi, whereas Darius and his supporters want to push forward as planned; at 3.127.1, Darius cannot move openly with an armed force against Oroetes, so he instead sends out an assassination team.⁴⁰ Moreover, in Darius’ appeal for a volunteer to do away with Oroetes, he points to Oroetes’ murder, literally his ‘making invisible’, of Mitrobates and his son (3.127.3 ἠίστωσϵ); the same rare verb is used earlier in Phaedime’s expression of certainty that, should she be caught feeling for the ears of the pseudo-Smerdis, the magus would ‘disappear her’ (3.69.4 ἀιστώσϵι). These are the only two occurrences of ἀιστόω in Herodotus.

IV. Hdt. 1.108.2 and 111.1, and 1.187.5 and 212.2–3: two stronger cases for deliberate echoing

I would like now to look at two cases of close-proximity recurrence of unique pairs, but ones that seem to provide more reason to believe that they are meant by Herodotus to be connected.

In the story of Cyrus’ origins, the adjective ἐπίτϵξ + ἐοῦσα (‘about to give birth’) is used of both his biological mother Mandane and his surrogate mother Spaco/Cyno and in only these two places in the whole of the History (1.108.2, 111.1). Three possibilities present themselves. (i) The two uses of ἐπίτϵξ are random. This is highly unlikely, given that the passages are so close together and concern two women who are so similar, namely, both maternal figures for the same individual, Cyrus. Furthermore, other women’s pregnancies and deliveries of offspring are prominently featured elsewhere in the History and do not have the term.⁴¹ (ii) The doublet is evidence of clustering as well as habitual thinking on Herodotus’ part, but not necessarily a deliberate pairing that he wanted the reader to see. This interpretation cannot be ruled out but seems less likely when we consider other features of the story. (iii) The pairing is deliberate because the logos concerning Cyrus’ origins, birth and near-destruction is precisely constructed around a series of parallels presented almost as diptychs: the two dreams of Astyages; the two missions to carry out the murder of the infant Cyrus, the one by Harpagus and the other by Mitradates; the exact synchronization of Mitradates’ concern for his wife Spaco/Cyno with hers for him (ἤσαν δὲ ἐν φροντίδι ἀμφότϵροι ἀλλήλων πέρι, 1.111.1); the sole occurrence in the account of the description of Cyrus’ birth and his miraculous survival of the phrase ‘the true/actual logos’ (ὁ ἐὼν λόγος, 1.95.1, 116.5).⁴² At the centre of the story, and an element that Eduard Fraenkel saw as crucial to the logic of its presentation, is the suppression until narratively important of the detail that Spaco/Cyno gave birth to a stillborn child while Mitradates was away.⁴³ It is difficult to see in this context where parallel presentation seems so much in the forefront of Herodotus’ narrative how the two uses of ἐπίτϵξ could not be deliberate, indeed carefully deployed by Herodotus.

The second set of examples has also to do with two female characters with a great deal in common. Recounting the accomplishments of the Babylonian queen Nitocris in book 1, Herodotus concludes his treatment of her reign by telling the story of her deception of Darius many years after her death and burial. Herodotus tells us that Nitocris created for herself a tomb above, or rather in, one of Babylon’s gates, and had carved into the outside of the crypt the following words (1.187.2):

If one of the kings of Babylon who comes after me is in need of money (ἢν σπανίσῃ χρημάτων), let him open my tomb and take however much money he needs; however, not being in need (μὴ σπανίσας) may he otherwise not open it, for it will not be well for him (οὐ γὰρ ἄμϵινον).

Herodotus makes a point of telling us that Nitocris’ tomb remained undisturbed until the reign of Darius. That king considered it terrible (δϵινόν) that he should not use the gate in which Nitocris was entombed (because a corpse was above a person’s head while passing underneath through the gate, 187.4), and that although money had been deposited there, the inscription forbade anyone taking it (μὴ οὐ λαβϵῖν).⁴⁴ Darius opened the tomb, but found no money, only another text with the following message: ‘were you not insatiate for money and sordidly avaricious, you would not be opening up tombs of the dead’ (ϵἰ μὴ ἄπληστός τϵ ἔας χρημάτων καὶ αἰσχροκϵρδής, οὐκ ἂν νϵκρῶν θήκας ἀνέῳγϵς, 187.5).

At the end of the same book, recounting Cyrus’ campaign against the Massagetae, Herodotus reports that Queen Tomyris’ son was captured by Cyrus after falling into a trap that involved getting him drunk on wine. When Tomyris learned what had happened, she sent a message to Cyrus: ‘O insatiate-for-blood Cyrus, don’t be encouraged by this matter that has happened’ (Ἄπληστϵ αἵματος Κῦρϵ, μηδὲν ἐπαρθῇς τῷ γϵγονότι τῷδϵ πρήγματι, 1.212.2), namely the wine-assisted capture of her son. She demands that Cyrus return her son, and concludes her message with a threat: ‘if you will not do these things, I swear by the sun, the ruler of the Massagetae, truly I will glut you with blood, even though you are insatiate’ (ἦ μέν σϵ ἐγὼ καὶ ἄπληστον ἐόντα αἵματος κορέσω, 212.3). In the ensuing battle the Persians are defeated and Cyrus killed. Tomyris orders a search for Cyrus’ body among the dead, and having found it, puts his severed head in a wineskin full of blood. Herodotus continues (214.4–5):

Defiling [him] (λυμαινομένη), she was saying over his corpse: ‘You destroyed me (σὺ μὲν ἐμέ) though alive and victorious in battle over you when you took/killed my child with deceit (παῖδα τὸν ἐμὸν ἑλὼν δόλῳ); you I, precisely as I warned, will glut with blood (σὲ δ’ἐγώ, κατά πϵρ ἠπϵίλησα, αἵματος κορέσω)’.

Although Herodotus knows many accounts of how Cyrus met his end, this is the one that is in his view most reliable, a point he also made in connection with his report of Cyrus’ birth (πιθανώτατος, 214.5; cf. 1.95.1).⁴⁵ Reciprocity is the key, emphasized by the pairing of pronouns close together at the start of succeeding sentences linked by men/de as subject and object, and then with their cases reversed (σὺ μὲν ἐμέ … σὲ δ’ἐγώ), expressing the requital Tomyris has exacted from her adversary.⁴⁶

The twice-occurring repetition to be examined is ἄπληστος. Although technically found three times in the History, two uses are close together and refer to the same person within the same context. Whereas some thematic continuity can be assumed for a unique pair within a single logos such as we have in ἀπαμϵλέω or ἐξιάομαι in the Democedes logos, the repetition of ἄπληστος raises the possibility of the same patterning or messaging by Herodotus in different sections of his History.

The parallel circumstances associated with the use of ἄπληστος are arresting. Nitocris and Tomyris are both eastern queens who communicate with a Persian king characterized as avaricious.⁴⁷ The communications themselves are in fact in each scenario a two-part message.⁴⁸ The first communication consists of a condition as well as a warning; the second confirms what the narrative makes clear, namely, that the Persian king is a wrongdoer in precisely the manner warned against by the queen. ἄπληστος is more obviously relevant in the story of Tomyris, for its original meaning (‘unfillable’) can be more easily connected to the liquid elements which are important in that logos, both wine and blood.⁴⁹ ἄπληστϵ at 1.212.2 is worth noting in particular. Vocative forms of adjectives are strongly marked,⁵⁰ and when they occur in Herodotus are often negative adjectives found in oracles or spoken by deities themselves, both addressing parties that have acted or tried to act against the divine will.⁵¹ This is a significant detail, not only because it lends divine weight to Tomyris’ voice, making her a quasi-oracular authority, but also because it connects her first message to Cyrus back to Nitocris’ first message to ‘a future king of Babylon’, namely, Darius: if that future king is not in need of money but opens up her tomb anyway, ‘it will not be well for him’ (οὐ γὰρ ἄμϵινον), a phrase that has also been interpreted as oracular in nature.⁵² It is worth noting here that the only other occurrences of this phrase in Herodotus’ History both come from the mouth of Darius later in book 3, are close together in the text and are both connected to his attempt to secure the throne of Persia (3.71.2, 82.5).⁵³ Similar is the phrase οὐ θέμιτον, found only at 3.37.3 and 5.72.3, the only two uses of the adjective and in both instances negated, with both referring to the unlawful entry into temple space by an impious king (Cambyses, Cleomenes).⁵⁴

It is difficult to escape the conclusion that the repeated use of ἄπληστος was deliberate on Herodotus’ part: as with the story of Cyrus’ two mothers, the term fits within a larger set of parallel details. What is more, while not proof in itself, it is easy to see how useful the repetition is to Herodotus’ narrative: the story of Nitocris’ posthumous rebuke of Darius, though later in time chronologically, prepares us textually for Tomyris’ vaunting over Cyrus.

V. Xerxes’ invasion of Greece: the interconnection of paired terms

In this section I will take up twice-occurring terms that are found in connection with Herodotus’ treatment of Xerxes. The reason for this focus on Xerxes and his activities is that, in Herodotus’ narrative of the monarch, sets of twice-occurring terms are particularly visible and can be seen to intersect, with one set becoming implicated in another, thus generating a larger network or ‘ladder’ of doublets. Additionally, Xerxes’ destruction of temples and divine statues became a topos in antiquity and seems to have generated something of a recognized and conventional vocabulary, helping to make clear cases where there were divergences from normally favoured terms. Such a situation leads inevitably to the issue of intention on Herodotus’ part in his use of these pairs of terms.

i. Hdt. 7.49.4 and 8.54

At 7.49.4 Artabanus attempts to make clear to Xerxes how the logistics of invading Greece with such a large force will themselves prove to be a major obstacle to his success. Land and sea will be Xerxes’ greatest enemies:⁵⁵

Land becomes hostile to you in the following way: if nothing wishes to be opposed to you (ϵἰ θέλϵι τοι μηδὲν ἀντίξοον καταστῆναι), to such a degree does the land become more hostile to you, namely, to the degree that you advance further, always led (lit. ‘stolen’) on to what lies next (τὸ πρόσω αἰϵὶ κλϵπτόμϵνος).⁵⁶ For humanity there is no satisfaction of success (ϵὐπρηξίης δὲ οὐκ ἔστι ἀνθρώποισι οὐδϵμία πληθώρη).

The gnomic sentiment with which this passage concludes focusses on the concept of ‘success’ (ϵὐπρηξίη); Artabanus even notes a little later that he is employing a piece of proverbial wisdom (τὸ παλαιὸν ἔπος, 7.51.3).⁵⁷ His words can be connected to the widely expressed sentiment that success in the form of ϵὐπραξία/ϵὐπραγία leads to hubris and massive change in fortune, from good to bad.⁵⁸ The only other place where ϵὐπρηξίη is found in the History is in connection with a communication between the same two men.⁵⁹ After Xerxes’ successful capture of Athens, marked by his burning of the Acropolis (ἐνέπρησαν πᾶσαν τὴν ἀκρόπολιν. σχὼν δὲ παντϵλέως τὰς Ἀθήνας, 8.53.2–54), Xerxes sent a messenger back to Artabanus in Susa ‘in order to announce his present success’ (ἀγγϵλέοντα τὴν παρϵοῦσάν σφι ϵὐπρηξίην, 8.54). It seems that Xerxes is deliberately responding to the caution of his uncle expressed at 7.49.4. He appropriates the key term of his uncle’s advice and throws it back at him: ‘here’s the good fortune you warned me about, but now it is in my hand’. But, of course, Artabanus will be shown to be correct in the end. The qualification of ϵὐπρηξίην, easy to read over, is particularly important in this regard. ‘Present success’ (παρϵοῦσα ϵὐπρηξίη) limits the concept; it makes ϵὐπρηξίη contingent, dependent on the moment.⁶⁰ And, of course, Xerxes’ success will indeed be fleeting, ruined by his defeat at Salamis which in essence forces his return to Persia, to say nothing of the Battle of Plataea later still. That there is a connection between uses of ϵὐπρηξίη seems inescapable, given the circumstances and that the same people are involved. I should add that while ϵὐπρηξίη is limited to these passages, allied concepts such as ϵὐδαιμονίη/ϵὐδαίμων are much more common and widely distributed in Herodotus.⁶¹

ii. Hdt. 8.102.3 and 7.8.β.2

I do not want to lose sight of the burning of the Acropolis as a significant moment for Xerxes, one that marks the high point of his invasion of Greece.⁶² Another single-recurrent pair of terms is to be connected precisely to this detail from Herodotus’ narrative of Xerxes’ expedition. In book 8, when Xerxes is contemplating abandoning his campaign for Persia after his defeat at Salamis, Queen Artemisia endorses the plan that he remove himself from the theatre of action, leaving Mardonius to prosecute the land war against the Greeks (8.102.3):

No account is taken of Mardonius, if he should suffer some loss (Μαρδονίου δέ, ἤν τι πάθῃ, λόγος οὐδϵὶς γίνϵται); for not even if they are victorious in some action do the Greeks win, having destroyed your slave. But you, having burnt Athens, for which reason you made the expedition, will march away (τῶν ϵἵνϵκα τὸν στόλον ἐποιήσαο, πυρώσας τὰς Ἀθήνας ἀπϵλᾷς).

Artemisia provides a compelling pretext for Xerxes to abandon his army in Greece by pointing out that he has fulfilled his stated goal for the expedition: the burning of Athens (πυρώσας τὰς Ἀθήνας). It goes without saying that Artemisia’s comments are also a tacit admission of defeat for Xerxes, inasmuch as they endorse his personal retreat to Persia. It is, therefore, deeply significant that the only other use in the entire History of the key verb πυρόω is at the start of book 7, when Xerxes announces to his council of war his intention to invade Greece, and specifically to burn down Athens in retaliation for the firing of Sardis (7.8.β.2–3):

I, on behalf of that man [Darius] and the rest of the Persians, will not cease until I capture and burn Athens (ἕλω τϵ καὶ πυρώσω τὰς Ἀθήνας), the very people who began doing unjust things towards me and my father (οἵ γϵ ἐμὲ καὶ πατέρα τὸν ἐμὸν ὑπῆρξαν ἄδικα ποιϵῦντϵς). First (πρῶτα) having come to Sardis … they burnt (ἐνέπρησαν) its groves and temples.

There can be little doubt, I think, that this first instance of πυρόω is meant to be linked to its second and only other use later:⁶³ Xerxes views the destruction of Athens by fire as the chief and crowning moment of his planned invasion, payback for the attack on Sardis and the burning of its ‘groves and temples’, and Artemisia reminds him of this fact when she provides him cover for his ignominious retreat to Persia. In essence she is saying: ‘You have achieved your stated goal; now you can leave’. What is more, in Xerxes’ accounting of events at 7.8.β.2, in speaking of the Athenians as initiators of the wrongs against Darius and himself, he uses language that connects his planned action against the Athenians to the largest and most comprehensive narrative arc of the History, namely, the conflict between Greeks and barbarians as defined by Herodotus at the beginning of book 1: ‘the man who I myself know first began unjust deeds against the Greeks (πρῶτον ὑπάρξαντα ἀδίκων ἔργων ἐς τοὺς Ἕλληνας), identifying this man, I will proceed with the rest of my account’ (1.5.3; cf. 1.130.3). The voice of Xerxes and that of the narrator align.⁶⁴

iii. Hdt. 8.53.2 and 9.76.1

In the narrative of the Persian capture of Athens, just before his notice that it was after that episode that Xerxes sent a messenger back to Artabanus reporting his ϵὐπρηξίη, Herodotus describes the final, desperate scene on the Acropolis (8.53.2):

When the Athenians saw that they [the Persians] had got to the top, some were throwing themselves down from the wall and perishing, while others were fleeing into the temple hall. Those of the Persians who had made the ascent first turned against the gates and, having opened them, were slaying the suppliants. When all had been overwhelmed by them (ἐπϵὶ δϵ σφι πάντϵς κατέστρωντο), they plundered the temple and burnt the entire Acropolis (τὸ ἱρὸν συλήσαντϵς ἐνέπρησαν πᾶσαν τὴν ἀκρόπολιν).

The pairing of συλήσαντϵς with ἐνέπρησαν is particularly worth noting, for Herodotus deploys it elsewhere (6.19.3, 101.3), even using exactly the same wording as 8.53.2 in one other passage (8.33 τὸ ἱρὸν συλήσαντϵς ἐνέπρησαν). Aeschylus, too, at Pers. 809–10, can write οἳ γῆν μολόντϵς Ἑλλάδ’ οὐ θϵῶν βρέτη | ᾐδοῦντο συλᾶν οὐδὲ πιμπράναι νϵώς.⁶⁵ Herodotus concludes his account of the Battle of Plataea with a similar capping scene of narrative description, creating a bridge to the episode involving the woman of Cos and Pausanias: ‘when the barbarians had been overwhelmed by the Greeks (ὡς δὲ τοῖσι Ἕλλησι ἐν Πλαταιῇσι κατέστρωντο οἱ βάρβαροι), then, of her own accord a woman approached them’ (9.76.1).⁶⁶

The repetition that deserves our attention here is κατέστρωντο.⁶⁷ It is important first to note that the verb καταστορέννυμι is extremely rare. Although as old as Homer, it is found only three times in the two epics, and then very infrequently in subsequent literature.⁶⁸ Secondly, it seems it was first used in a battle description by Herodotus.⁶⁹ And as it turns out, the particular deployment of κατέστρωντο at 8.53.2 was memorable enough to draw the attention later of Pollux, who grouped it together with other figurative synonyms for ‘killed’.⁷⁰ It continued to be extremely rare in this sense after Herodotus (cf. Paus. 7.15.9). The verb καταστορέννυμι occurs in fact three times in the History, the third case also being connected to Plataea, though in the active voice: Theban cavalry caught sight of the Megarian and Phliasian contingents pursuing the enemy in disorder and ‘overwhelmed’ (κατϵστόρϵσαν) 600 of them,⁷¹ leading Herodotus to comment, ‘these men perished doing nothing worth mentioning’ (οὗτοι μὲν δὴ ἐν οὐδϵνὶ λόγῳ ἀπώλοντο, 9.70.1). While similar in effect (the verb caps the scene and permits the narrator to take a bigger view of the significance of the moment), the Theban destruction of the Megarian and Phliasian troops is but a single episode in a much larger narrative panel, whereas the two uses of the passive κατέστρωντο are found at the conclusion of the accounts of both the capture of the Acropolis and the Battle of Plataea.

Indeed, it is important to observe the rhetorical effect of κατέστρωντο at 8.53.2 and 9.76.1. Both occurrences are found at the end of substantial narratives about key events from the story of Xerxes’ invasion (capture of Athens, the Battle of Plataea) and serve to bring each episode to a close: they ‘cap’ them.⁷² This capping is brought about largely through focussing the reader’s attention on the finality of the moment in question, the awful and irreversible fate that meets each collection of people. Herodotus frequently brings stories about ill-fated communities and individuals to a conclusion in such a fashion, that is, with a significant repetition: for cities/groups, compare 1.84.1 (the capture of Sardis), 4.11.4 (the self-destruction and burial of the royal tribe of the Cimmerians), 6.18 (the fall of Miletus);⁷³ for individuals, 1.45.3, 1.82.8 and 7.107.2 (the suicides of Adrastus, Othryadas and Boges), and 3.125 and 128.5 (the linked deaths of Polycrates and Oroetes). All of these passages are clearly closural in function and achieve their effect through the repetition of significant words (‘Adrastus, son of Gordias, son of Midas’ at 1.35.3 and 45.3;⁷⁴ Sardis ‘captured’ at 1.84.1 and 5), or with a strong deictic term signalling the end (for example, οὕτω ‘thus’ at 3.128.5, 4.11.4 and 7.107.2), and sometimes both (1.84.5 οὕτω δὴ Σάρδιές τϵ ἡλώκϵσαν).⁷⁵

It is important to note that the terms used of the two combatant groups at Plataea reported in 9.76.1 are Hellēnes and barbaroi. While it is true that Herodotus can use the words barbaros and Persēs/Mēdos interchangeably, very clearly he can also deploy barbaros in a more meaningful way, in particular when it is contrasted with Hellēn, as we see most notably in the Proem (ἔργα … τὰ μὲν Ἕλλησι τὰ δὲ βαρβάροισι ἀποδϵχθέντα). In the narrative of Marathon, for instance, barbaros is used with distinct frequency, and is explicitly contrasted with Hellēn at 6.112.3, in an observation made by the narrator that is clearly meant to highlight a moment of great significance.⁷⁶ The scenes featuring Pausanias that immediately follow the Battle of Plataea in book 9 stress the moral superiority of the Spartan commander and the Greeks as a whole over the Persians: Pausanias sees to the rescue and restoration to her home of the woman of Cos, and he refuses to allow the body of Mardonius to be mutilated (as the body of Leonidas had been), even commenting that such actions ‘are more fitting for barbarians to do than verily for Greeks’ (τὰ πρέπϵι μᾶλλον βαρβάροισι ποιέϵιν ἤ πϵρ Ἕλλησι, 9.79.1).⁷⁷ Given that the barbaros/Hellēn distinction seems most definitely to be in play at 9.76.1, it can be seen to connect with the largest and most important narrative arc of the History: the conflict of Greeks and barbarians announced at the start of the work. The seizure of the Acropolis and the Battle of Plataea are joined in Herodotus’ mind; the capture of the Acropolis is also explicitly linked to the Ionian Revolt; and so on. A ladder or chain of events becomes discernible. Herodotus’ narrative is in fact constructed precisely out of such ‘chains’ of interlocking events, often battles, that can be traced back to the very beginning of the History.⁷⁸ Remember that Xerxes’ intention to invade Greece and punish the Athenians for ‘having begun doing unjust deeds’ (ὑπῆρξαν ἄδικα ποιϵῦντϵς, 7.8.β.2) is phrased in language that can be connected to Croesus and to Herodotus’ declaration of the main topic of the History at 1.5.3.

These recurring pairs of words concern not a subsidiary storyline but the main narrative of the second half of the History. They are deployed against a backdrop that more generally depicts the Persians and Greeks as understanding the conflict between them as chiefly about the burning of temples and sacred images (5.102–05). Retribution is therefore due in precisely these terms in Darius’ and Xerxes’ understanding (note especially 6.101.3 τὰ ἱρὰ συλήσαντϵς ἐνέπρησαν, ἀποτινύμϵνοι τῶν ἐν Σάρδισι κατακαυθέντων ἱρῶν), both presented as obsessed by the need for revenge against the Athenians for burning temple space (5.105.2, 6.94.1, 7.4, 7.8.β.3).⁷⁹ Changing sides in the conflict is an impossibility for the Athenians ‘in the first place and most importantly’ (πρῶτα … καὶ μέγιστα) because of the enemy destruction by fire and demolition of ‘the statues and habitations of the gods’ (8.144.2 τῶν θϵῶν τὰ ἀγάλματα καὶ τὰ οἰκήματα ἐμπϵπρησμένα τϵ καὶ συγκϵχωσμένα cf. 143.2 ἐνέπρησϵ), crimes of Xerxes that Themistocles had earlier highlighted in a speech to the Athenians (8.109.3 ἐμπιπράς τϵ καὶ καταβάλλων τῶν θϵῶν τὰ ἀγάλματα)⁸⁰ and that Xerxes also referred to in a message conveyed by Mardonius to Alexander which was then reported to the Athenians (8.140.α.2 ἱρά … ὅσα ἐγὼ ἐνέπρησα).⁸¹ It is true that Herodotus can attribute to Darius and Xerxes a range of motives for their aggression against the Greeks of Asia and Europe and even refers to the burning of the temple of Cybebe (Cybele) at Sardis during the Ionian Revolt as a pretext ‘alleged’ by the Persians for the later ‘counter-burnings’ of Greek temples (τὸ σκηπτόμϵνοι οἱ Πέρσαι ὕστϵρον ἀντϵνϵμπίμπρασαν τὰ ἐν Ἕλλησι ἱρά, 5.102.1).⁸² Some detect in such passages Herodotus signalling a difference between a ‘pretext’ on the one hand and ‘real causes’ on the other, revealing that the Persians in particular used the burning of the temple at Sardis as justification for large-scale aggressive action against the Greeks, that is, that revenge was a cloak for imperial invasion.⁸³ Granting that Herodotus viewed some causes as more decisive than others, it is important to note, as Simon Hornblower does, that the use of terms such as σκήπτομαι and πρόφασις ‘does not automatically imply the falseness of the excuse or proffered reason’.⁸⁴

In this context it is good to remember that the Persian destruction by fire of Greek sanctuaries and statues, at Athens in particular, became a hallowed and living memory, a topos with a quasi-standard vocabulary featuring the verbs συλάω and especially ἐμπίπρημαι.⁸⁵ Hence, a deviation from those words such as πυρόω would presumably have been all the more noticeable.⁸⁶ Moreover, the burning of Sardis and the counter-burning of Athens were viewed as epochal moments by later ages: Philip II pointed to Persian wrongdoing as a reason for war against Persia (Polyb. 3.6.13 παρανομίαν) and Alexander the Great maintained that the burning of Persepolis was punishment for the sacking of Athens and the burning of the temples (Arr. Anab. 3.18.12 τάς τϵ Ἀθήνας κατέσκαψαν καὶ τὰ ἱϵρὰ ἐνέπρησαν; cf. Strabo 15.3.6; Diod. Sic. 17.72.2; Curt. 5.7.4; Plut. Alex. 38.3–4).⁸⁷ Given the centrality of the destruction of temples by fire in the narrative arc of Herodotus’ History, that is, as a triggering event of hostilities between Greeks and Persians from the Ionian Revolt and as the most important strategic and symbolic moment of Xerxes’ invasion in particular, it is important to observe the role that single-recurring terms play in this articulation of the conflict, especially in the second, Xerxes phase.

Indeed, taking stock of the recurrent pairs associated with Xerxes’ invasion of Greece, we can perhaps make out a larger point. While repeated, twice-occurring terms are to be found throughout the History, several seem to mark out the major plot developments in the story of Xerxes’ invasion and help to connect it to the largest narrative of the work: the conflict of the Greeks and barbarians. What is more, these unique pairs seem to be of the type where deliberate choice on Herodotus’ part must clearly be understood: they do not occur randomly throughout the History, and the words themselves represent common concepts that are found elsewhere but are expressed with different terms. If this is an accurate assessment, it is important to note further that these deliberate repetitions of significant terms create connections over very large portions of text. As such, they could also be seen as examples of Herodotus’ use of analepsis and prolepsis, in Irene de Jong’s analysis, ‘carefully insert[ed] … at places where they are most effective’.⁸⁸

VI. Conclusions: unique pairs and the challenge of Fehling

There will always be disagreement about whether one set of twice-occurring terms represents a deliberate doublet and another simply a function of Herodotus’ habit of presentation. It is likely that I have mischaracterized some of the unique pairs of terms examined in this paper. But it seems to me incontrovertible that there can only be three possibilities for these pairs: random occurrence, unconscious habit of presentation and deliberate pairing. Paying attention to the distribution of the terms in the text, whether the once-iterated words are accompanied by other parallels, and viewing each passage as a whole ought to make one interpretation more likely than another. Generally speaking, close-proximity pairs suggest clustering on Herodotus’ part and are often evidence of non-deliberate repetition. So, too, when phrasal repetition accompanies single-word reiteration. Both seem to be strong indicators that Herodotus had certain habits of presentation that he followed but of which he may not have been entirely or at all aware. On the other hand, when the twice-occurring terms, especially from different logoi, can be set beside other details that are parallel, then I think the balance must tilt in favour of deliberate repetition.

However these questions are decided, though, I believe it has been shown that in most cases the claim that the two uses are random and therefore meaningless is not sustainable. Unique pairs have a distinct heuristic value in studying Herodotus’ text. With great clarity and economy they bring to the fore central questions relating to how Herodotus constructed his narrative, whether deliberately or as a result of patterns of presentation to which he adhered, consciously or not. Fehling mounted an attack on the credibility of Herodotus that, for all its shortcomings, must be answered, or else his text’s many repetitions and regularities cannot but provoke our suspicion.⁸⁹ One of the central elements of Fehling’s critique is that Herodotus invented his material and as such produced numerous duplications throughout his account, as ‘liars’ tend to do. Twice-occurring terms help us to bring nuance to the study of repetition in Herodotus and thereby to salvage his reputation as an historian. Some of the cases examined in this paper show that his historical imagination did settle into patterns of language and so indirectly support Fehling’s position. But this predictability in choice of expression is so widespread and various that it could just as easily be explained as Herodotean habit of presentation, regardless of whether he was making up his account or recounting material he had obtained from informants, documents and autopsy. Furthermore, and perhaps more importantly, Fehling’s views do not allow for Herodotus deliberately to connect events that he saw as linked or parallel through the strategic deployment of rare terms, as he seems to have done in many cases.