[In the following essay, Plant addresses the influence of Antiphon and Gorgias on the rhetorical techniques used by Thucydides.]

In recent years, there has been considerable debate about the reliability of Herodotus: the attack on his honesty led by Fehling, the defence by Pritchett.¹ The debate, it seems, may have begun at least as far back as Thucydides,² but now Thucydides himself may have joined the school of liars. Badian has produced a new reading of Thucydides' description of the outbreak of the Peloponnesian War, arguing that Thucydides deliberately set out to mislead the reader, misrepresenting the Spartans as the instigators of the War and carefully masking the Athenians' own responsibility.³

But why is Badian's interpretation rejected as implausible?⁴ Is it because Thucydides is still seen as pioneer in the search for historical truth? While no longer viewed as the sort of ‘scientific historian’ championed by Cochrane, Thucydides is still often portrayed as somewhat unique in the ancient world for the care he took to discover what actually happened.⁵ In fact we have little contemporary evidence with which to mount a serious challenge to Thucydides' standard of accuracy, save for a few points of detail. Even there, I would venture, it has been impossible to prove him wholly accurate or inaccurate, and it is not my intention to try to do so. My subject is Thucydides' claim to accuracy in his description of his methods. For it is Thucydides' claim to truthful reportage, as much as anything, which makes Badian's thesis seem implausible. The importance that Hornblower ascribes to Thucydides' description of method serves as an example of the high esteem in which it has been held by Thucydides' readers. On the first chapter of Book One, he writes, ‘Thucydides' self-conscious discussion of the methods of his intellectual undertaking … is not the least of his intellectual achievements.’ He describes chapter 22 thus: ‘A very important chapter indeed. There is nothing like it in Herodotus. More remarkably, it is hard to parallel in any writer later than Thucydides …’⁶ But is Thucydides unparalleled? What precedents or antecedents are there for what he says about his methods in his preface (chapters 1-23), especially 1.22?⁷

Studies of Thucydides have already taught us not to see his work in isolation. For example, Cornford, Finley, and Macleod sought literary parallels for Thucydides' approach to history. More recently Woodman considered how literary and rhetorical concerns influenced Thucydides' composition.⁸ Hornblower, as we have seen, compares Thucydides with Herodotus. It is, however, to the sophists and the rhetoric of the fifth century, as exemplified by Gorgias and Antiphon, that I would like to turn, to put Thucydides' discussion of the methods of his history into some perspective. My thesis is that contemporary forensic oratory provides parallels for both Thucydides' avowed method and his avowal of method. I will argue that Thucydides took elements from the persuasive and self-justificatory oratory used in forensic rhetoric and used them to good effect in his work. These forensic techniques can perhaps be traced to the teaching of sophists like Gorgias.

It has long been recognized that Thucydides was well versed in contemporary developments in the art of rhetoric.⁹ Gorgias in particular has been identified as an influence. His contribution to Thucydides' style was noted in antiquity.¹⁰ The influence of Gorgias has been found not just in elements of style, but also in the intellectual underpinnings of Thucydides' work. Hunter has demonstrated that in his basic conception of the nature of logos and the power of fear on the mind, Thucydides displays close affinities with the philosophic position of Gorgias.¹¹ The sophistic influence on Thucydides is clearly strong. Indeed, it has been said that ‘Thucydides was … so thoroughly influenced by the sophists that it would only be a slight exaggeration to call him a sophist himself.’¹²

The way in which Gorgias and other sophists taught is significant. Along with the ideas expressed, the teaching method itself has been identified as contributing to aspects of Thucydides' work, particularly the speeches.¹³ For one method practised by the sophists was to teach by example, through model speeches or technai. These model speeches had both an epideictic and pedagogic function, designed as they were both to show off the master's rhetorical skill and to illustrate his ideas. We have evidence that Gorgias taught in this way from Aristotle¹⁴ and from two examples of such works that remain, the Helen and the Palamedes. These technai were written down and given to students to learn. Such speeches were typically forensic in character. The rhetorical teaching of the sophists like Gorgias had, of course, a ready purpose and market in forensic oratory—indeed their developments in argument were designed for just such a public forum (Isocrates 13.19-20). Antiphon is credited with being the first to write down forensic speeches for others to use.¹⁵ The extant corpus of his works includes twelve model speeches, grouped into three Tetralogies, each Tetralogy including speeches for the prosecution and for the defence.¹⁶ Cole argues that the complexity, compression, and frankness of the arguments in the speeches in Thucydides mean that they cannot have been made in the form he gives us on the occasions when he claims they were made.¹⁷ These elements, along with the neglect of any characterization of the speakers, the concentration on general rather than specific issues, and the pairing of opposite arguments (which are also typical of Thucydides' speeches) are characteristic of technai, model speeches. This leads him to the conclusion that Thucydides meant his own speeches to have just such a paradigmatic function: that they were meant as technai, models for statesmen to use in the appropriate circumstances.¹⁸

This is going too far, for Cole does not take enough account of the way in which the speeches are firmly attached to a narrative, and his outline sketch of the contents of various speeches is oversimplified to force them to fit his paradigmatic model. However, this is not to say that technai were not influential in determining the composition of the speeches by Thucydides, and where the speeches do reproduce something of a genuine speech, the original composer too may have been influenced by technai. That is after all what the technai were for!

Whether or not we accept Cole's thesis, or my modification of it, we must still accept the strong influence of contemporary rhetoric on Thucydides. This is apparent not just in the speeches, but also in the way in which Thucydides persuades us of the validity of his reconstruction of the past in his preface to Book One, chapters 1.3, 1.21, and 1.22. These chapters of Thucydides are regularly compared with Herodotus, but while this does identify differences, it does not help us to understand Thucydides' rhetoric.¹⁹ We need instead to consider an example of contemporary rhetoric, such as a fairly typical part of a defence speech by Antiphon (Choreutes 31):

εγo τοίνυν τούς τε λόγους ὑμι̑ν εἰκότας ἀποφαίνω, καὶ τοι̑ς λογοι̑ς τοὺς μάρτυρας ὁμολογου̑ντας καὶ τοι̑ς μάρτυsι τὰ εργα, καὶ τεκμήρια εξ αὐτω̑ν τω̑ν εργων, καὶ ετι πρὸς τούτοις δύο τo μεγίsτω καὶ ἰsχυροτάτω …

Moreover, I am presenting to you arguments which are probable, and evidence which is consistent with the arguments and facts which are consistent with the evidence, and proof from the facts themselves, and in addition to these the two greatest and strongest arguments …

The speaker is in self-conscious mode. He tells the jury what arguments he has used, specifying two of the most important ones, and points out how he has supported those arguments with evidence. The speaker thus supports his arguments by his self-conscious reference to them. In a sophistic model speech by Gorgias, we also find this sort of explicit reference to method in support of the argument (Helen 2):²⁰

του̑ δ' αὐτου̑ ἀνδρὸς λέξαι τε τὸ δέον ὀρθω̑ς καὶ ελέγξαι τοὺς μεμφομένους ‘Ελένην, … εγo δὲ βούλομαι λογιsμόν τινα τἳ̑ λόγῳ δοὺς τὴν μων κακω̑ς ἀκούουsαν παυ̑sαι τη̑ς αἰτίας, τοὺς δὲ μεμφομένους ψευδομένους επιδείξας καὶ δείξας τἀληθες [e] παυ̑sαι τη̑ς ἀμαθίας.

The same man should say what ought to be said and refute those who blame Helen … I want to support my argument with some reasoning and put a stop to the criticism of the woman who is being slandered, and by both demonstrating that those who criticize her are lying and revealing the truth to put a stop to ignorance.

Here along with his aim, Gorgias explains concisely the means by which he will achieve it, λογιsμόν τινα τἳ̑ λόγῳ δούς. Thus he reveals an explicit awareness of his own method. Notice that he has also included a general statement of methodological principle, του̑ δ' αὐτου̑ ἀνδρὸς λέξαι τε τὸ δέον ὀρθω̑ς καὶ ελέγξαι τοὺς μεμφομένους ‘Ελένην. This rhetoric strengthens the thesis about to be presented about Helen. For of course Gorgias says what ought to be said: he refutes those who blame Helen. The statement of methodological principle validates the argument to follow. The function of such consideration is persuasion. According to Plato, Gorgias said that the entire business of rhetoric was directed towards persuasion.²¹ In this case, the persuasion takes the form of the validation of an argument by reference to the means by which it was formed, along with the reverse, that is the refutation of a rival argument. It has a similar function in Thucydides.

One of the key elements which Thucydides stresses in his preface is the desire to discover what really was said or happened. This concern adds a degree of authority to his work. He emphasizes, in particular, the difficulty of this task. He identifies the difficulty of trusting the evidence (χαλεπὰ ὄντα παντὶ εξη̑ς τεκμηρίῳ πιsτευ̑sαι, 1.20.1), the difficulty of obtaining evidence about the distant past (τὰ γὰρ πρὸ αὐτω̑ν καὶ τὰ ετι παλαίτερα sαφω̑ς μεν εὑρει̑ν διὰ χρόνου πλη̑θος ἀδύνατα ην, 1.1.3; ηὑρη̑sθαι δε ἡγηsάμενος εκ τω̑ν επιφανεsτάτων sημείων ὡς παλαιὰ εεναι ἀποχρώντως, 1.21.1; ὅsον δυνατὸν ἀκριβείᾳ περὶ εκάsτου η]πεξελθών, 1.22.2), and the difficulty of remembering accurately what has been said or done in the more recent past too (χαλεπὸν τὴν ἀκρίβειαν αὐτὴν τω̑ν λεχθέντων διαμνημονεἅsαι εν έμοί τε εν αὐτὸς ἤκουsα καὶ τοι̑ς ἄλλοθέν ποθεν εμοὶ ἀπαγγέλλουsιν, 1.22.1). He also tells us what a difficult task it was (έπιπόνως δε ηὑρίsκετο, 1.22.3) to assess the evidence.²²

Connor reminds us to be suspicious of rhetorical embellishment in Thucydides' preface, and he is right. At 1.1.3 we are told that it was impossible to find out clearly (ἀδύνατα ην). Later the impossible becomes only difficult (χαλεπὰ ὄντα, 1.20.1); later still, however, Thucydides speaks more positively about his outline of the past: τοιαυ̑τα ἄν τις νομίzων μάλιsτα a διη̑λθον οὐχ ἁμαρτάνοι … ηὑρη̑sθαι δε ἡγηsάμενος έκ τω̑ν επιφανεsτάτων sημείων ὡς παλαιὰ εiναι ἀποχρώντως (1.21.1). Finally, τω̑ν τε γενομένων τὸ sαφες is considered to have been found with sufficient clarity that, it is implied, it can be a help in considering the future and satisfy the author's desire to create something useful (1.22.4).²³ By reminding us of the labour he put into finding out what really happened, Thucydides substantiates the results of that labour.

Of course, the assertion that the truth, true knowledge about the facts, should be searched for and presented to the jury is common in forensic rhetoric. In Choreutes 26 the speaker attempts to validate his case by a claim to have challenged the opponents to use the methods necessary for establishing the truth (εξ ων γε χρὴ ἄνθρωπον ὄντα τἀληθη̑ καὶ τὰ δίκαια πυνθάνεsθαι). The speaker impresses upon his listeners the veracity of his own case and discredits his opponents through reference to the methods that should have been used, and to which he, but not his opponent, paid attention.²⁴ The difficulty of obtaining the truth about the past is also a well-attested concept in Antiphon's rhetoric. In the Second Tetralogy 2.1 he begins the defence speech by pointing out the difficulty of knowing exact details about the affair: πραγμάτων … ων εγo χαλεπω̑ς μεν τὴν ἀκρίβειαν εγνων.²⁵ This phrase is reminiscent of Thucydides' own χαλεπὸν τὴν ἀκρίβειαν αὐτὴν … διαμνημονευ̑sαι (1.22.1).

We can also parallel in forensic rhetoric Thucydides' observations about the poor methods of others as a further element of rhetorical persuasion. In 1.20-1 Thucydides formulates rules for the general methodological approach of other people and he generalizes to support his argument. People accept hearsay (οἱ γὰρ ἄνθρωποι τὰς ἀκοὰς … ὁμοίως ἀβαsανίsτως παρ' ἀλλήλων δέχονται, 20.1). People (τω̑ν ἀνθρώπων) judge present war greatest, then revert to amazement; they are credulous not critical (21.2). Their methods produce mistaken opinions (οὐκ ὀρθω̑ς οἴονται, 1.20.3) and serve as a balance to emphasis the success of Thucydides' own methods at producing the right opinion (… οὐχ ἁμαρτάνοι, 1.21.1). It has been suggested that he has Herodotus in mind with this criticism, for he quotes two ‘errors’ found in Herodotus.²⁶ However, he does not mention him by name and Herodotus does not make the other ‘mistake’ Thucydides identifies, concerning Hipparchus and Hippias.²⁷ The criticism is deliberately general. Thucydides could have named a specific target for criticism here, as he singles out Hellanicus in 1.97.2, but instead he decided to make a general point about lack of critical ability, to compare his method with all others, rather than with one specific rival. The point made strengthens Thucydides' own claims to accuracy; his use of specific examples in 1.20.2-3 to make his general point is particularly effective.

In Antiphon, the speaker's care in finding the truth is often compared with the lack of effort made by the opponent to find out what happened (Prosecution for Poisoning 13):²⁸

Περὶ μὲν οην τούτων οὐκ ἄδηλον ὅτι αὐτοὶ εφευγον τω̑ν πραχθέντων τὴν sαφήνειαν πυθέsθαι …, ὥsτε sιωπώμενον καὶ ἀβαsάνιsτον αὐτὸ εα̑sαι εβουλήθηsαν … περὶ δὲ τω̑ν γενομένων πειράsομαι ὑμι̑ν διηγήsαsθαι τὴν ἀλήθειαν.

So concerning [the slaves] it is obvious that they shrank from finding out the clear facts about what had happened; … so they wanted to leave it unsaid and uninvestigated … But I will try to tell you the truth about what happened.

Antiphon also liked to make his speakers refer to universal truths in respect to method. An example comes in Choreutes 29-30:

πόθεν χρή, ωἄνδρες, e τἀληθη̑ πιsτὰ e τὰ μὴ ἀληθη̑ ἄπιsτα ποιει̑ν ἄλλοθεν e εκ τω̑ν τοιούτων; ὅπου μὲν γὰρ λόγῳ τις διδάsκοι περὶ τω̑ν πραχθέντων, μάρτυρας δὲ μη̑ παρέχοιτο, μαρτύρων ἄν τις τοὺς λόγους τούτους iνδεει̑ς φαίη εεναι· ὅπου δὲ μάρτυρας μεν παρέχοιτο, τεκμήρια δi αη τοι̑ς μαρτυρου̑sιν ὅμοια μὴ ἀποφαίνοι, ταὐτὰ ἄν τις εχοι εἰπει̑ν, εἰ βούλοιτο.

Gentlemen, how can the truth be confirmed and what is untrue refuted other than by such means as these? For if someone were to tell you about what happened through an argument and not provide witnesses, someone could say that these arguments were lacking witnesses, but if he were to provide witnesses, but not show any proofs to support the witnesses, someone could point this out [as a criticism] if he wanted to.

Here Antiphon's speaker anticipates the points he is about to make about his own method and thereby strengthens his case. For he establishes what the appropriate method is, then goes on to detail how he has followed that very method in the next sentence. The combination is a powerful rhetorical ploy, the ‘rule’ postulated legitimising the practice specified.²⁹

The most important principle which a speaker needed to assert was that he was telling the truth and that his opponent was not (e.g. Choreutes 9). Gorgias demonstrates this rhetorical topos in a model forensic speech (Palamedes 5):

‘′Οτι μὲν οην οὐ sαφω̑ς <εἰδoς> ὁ κατήγορος κατηγορει̑ μου, sαφω̑ς οιδα· sύνοιδα γὰρ εμαυτἳ̑ sαφω̑ς οὐδὲν τοιου̑τον πεποιηκώς· οὐδε οεδ’ ὅπως ἂν εἰδείη τις oν τὸ μὴ γενόμενον. εἰ δε οἰόμενος οὕτω ταυ̑τα εχειν εποιει̑το τη̑ν κατηγορίαν, οὐκ ἀληθη̑ λέγειν διὰ διssω̑ν ὑμι̑ν επιδείξω τρόπων.

So I know clearly that the prosecutor accuses me but does not know the matter clearly. For I know in myself clearly that I have not done anything like this; nor do I know how anyone could know that what has not happened has happened. But if he made the prosecution believing that these things were so, I will show you that he is not telling the truth in two ways.

Palamedes plays upon the ideas of knowing clearly and of telling the truth, structuring his defence speech around these two important principles and telling us so. So we find that the desire to find what really happened appears not just in Thucydides, but is (not surprisingly) a commonplace in the rhetoric of forensic oratory of the fifth century; so too is a statement of the method needed to discover it. We find this in both Gorgias and Antiphon. It is used as an element of persuasion, and opposes the counter-claim that the speaker is not telling the truth. This does not of course mean that the speakers always did tell the truth! The difficulty of finding out what happened is stressed; this emphasizes the care the speaker has taken to find the truth, and is compared with the opponent's lack of interest in what really happened, or his deliberate falsification. Thucydides, we find, is influenced by these forensic topoi in using them to justify his reconstruction both of the distant and more recent past and support its validity.

Thucydides makes use of another forensic technique too. For his methodological concerns in 1.1.3, and 1.20-2 focus on the use of evidence. Thucydides emphasizes that he will not simply repeat evidence: he must examine it carefully to see if he can trust it. He tells us that this was not a simple task as not all evidence could be trusted: χαλεπὰ ὄντα παντὶ εξη̑ς τεκμηρίῳ πιsτευ̑sαι (1.20.1). Thucydides provides examples of mistaken beliefs to illustrate his point. Others do not believe correctly (οἱ ἄλλοι ‘′Ελληνες οὐκ ὀρθω̑ς οἴονται, 1.20.3). Their search for the truth is careless, due in large part to the uncritical way in which they find and receive evidence (οὕτως ἀταλαίπωρος τοι̑ς πολλοι̑ς ἡ zήτηsις τη̑ς ἀληθείας, καὶ επὶ τὰ ετοι̑μα μα̑λλον τρέπονται, 1.20.3; cf. οἱ γὰρ ἄνθρωποι τὰς ἀκοὰς … ἀβαsανίsτως παρ’ ἀλλήλων δέχονται, 1.20.1). Thucydides criticizes logographoi and poets by observing that their accounts of the past cannot be proved wrong (ἀνεξέλεγκτα, 1.21.1). He will put no trust in stories which cannot have been tested and dismisses exaggerated stories written to entertain (1.21). This method has been seen as Thucydides' great legacy to historical writing.³⁰

Thucydides' methods again have forensic underpinnings: the concern with testing to find the truth is found in Antiphon (First Tetralogy 1.10):³¹

οἵ τε γὰρ επιβουλεύοντες ἀνεξέλεγκτοι ἂν εiεν, εἰ μήθ' ὑπὸ τω̑ν παραγενομένων μήθ' ὑπὸ τω̑ν εἰκότων εξελέγχονται.

For those who plan crimes would not be convicted, if they were not convicted by eye-witnesses and by what is likely.

Thucydides makes the point that such a method cannot be applied to the works of the logographers—his very language (ἀνεξέλεγκτα, 1.21.1) suggesting forensic inquiry. Indeed, in the context of his attempt to find out about the past (εὑρει̑ν, 1.1.2, 1.20.1, 1.21.1, 1.22.3) we find Thucydides uses the language of judicial inquiry. His vocabulary includes τεκμήριον, sημει̑ον, and μαρτύριον (and their cognates) ten times in his preface alone.³² From the outset Thucydides looks to support his opinion (νομίzω) with evidence (εκ τεκμηρίων, 1.1.3).³³ The language of inquiry, coming as it does from a forensic context, is a strong indication of the influence of forensic oratory on Thucydides' concern with and treatment of evidence.³⁴

Chapter 1.22 needs careful consideration. It is largely concerned with Thucydides' method of dealing with testimony, evidence as to what was said or done just before or during the Peloponnesian war. We should notice three important points in the context of this discussion:

(i) Thucydides implies that the speeches would have been based directly upon the testimony of witnesses, but that his and his witnesses' inability to remember precluded that sort of direct repetition of speeches (1.22.1). In this concern we find that the implicit ideal is an accurate account by an eyewitness with an accurate memory.

(ii) In describing his actual practice, Thucydides picks up this implicit ideal in the participial qualifying phrase εχομένῳ ὅτι εγγύτατα τη̑ς ξυμπάsης γνώμης τω̑ν ἀληθω̑ς λεχθέντων. Here τω̑ν ἀληθω̑ς λεχθέντων looks back to τὴν ἀκρίβειαν αὐτὴν τω̑ν λεχθέντων—τη̑ς ξυμπάsης γνώμης is presumably what is left of the speech when you have forgotten τὴν ἀκρίβειαν αὐτήν. It is difficult to be precise about the meaning of τη̑ς ξυμπάsης γνώμης itself. The phrase is vague, perhaps deliberately so. The stress on accuracy in the use of the superlative (ὅτι εγγύτατα) and the suggestion of completeness (ξυμπάsης) and truth (ἀληθω̑ς) disguise the necessarily subjective interpretation of τη̑ς γνώμης, the general sense, main thesis, or entire intention of the speaker or speech.³⁵

(iii) And further, despite his expression of the ideal, Thucydides admits to writing as he thought each speaker would have said τὰ δέοντα. His own contribution is admitted. The point just made about the limitation of memory cleverly justifies and indeed necessitates such a contribution to his work by the writer. The vital and difficult words in this clause are τὰ δέοντα: ‘what was most appropriate’ or ‘necessary’ or ‘what the situation demanded’, which scholars have tried to use to define the exact quantity and quality of Thucydides' contribution to the speeches.³⁶ τὰ δέοντα here is an admission by Thucydides of the importance of external factors in his composition of the speeches, factors to be taken into account beyond reporting what was known to have actually been said. Thucydides justifies this departure from the ideal of recording what was actually said, by pointing to the impossibility of that task, and at the same time he makes his recording of no more than the ξύμπαsα γνώμη of the speaker sound a good deal closer to the ideal of accuracy than it is, perhaps alluding with τὰ δέοντα to the sophistic practice, particularly favoured by Gorgias, of creating appropriate arguments for every and any occasion.³⁷

τὰ δέοντα seems to have been a term favoured by Gorgias to describe the composition of a speech appropriate to the occasion in both form and content, and it is sophistic in character. Gorgias uses the singular form in the Epitaphios: του̑τον νομίzοντες θειότατον καὶ κοινότατον νόμον, τὸ δέον εν τἳ̑ δέοντι καὶ λέγειν καὶ sιγα̑ν καὶ ποιει̑ν <καὶ εα̑ν> (fr. 82.B6 Diels-Kranz). Macleod drew attention to the parallel between τὰ δέοντα at 1.22.1 and what Gorgias says about his purpose in the Helen, where he speaks of his task as λέξαι τε τὸ δέον ὀρθω̑ς (Helen 2). Macleod also uses Isocrates 13.7-8 and Plato, Phaedrus 234e6 to demonstrate that τὰ δέοντα or τὸ δέον may not only refer to the content of a speech, but also to its form, the rhetorical elements of the speech.³⁸ Thucydides, on this reading, announces that he has composed what was rhetorically appropriate to make each speech as apt and effective as possible in the context in which is presented—in both form and content. Thucydides may have written τὰ δέοντα μάλιsτ' εἰπει̑ν at 1.22.1 both to signal this intention and his knowledge of the sophists' art of speech-writing.

Of course what he claims to do and what his practice actually was (or practices were, as he may have changed or evolved his methods as he worked) need not be the same. I want to focus on his claims and their effect: that Thucydides impresses the reader with the claim of accuracy, even at the point when he is describing the subjective input to his work. This is not an attempt to mislead the reader. Rather, Thucydides is intent upon convincing the reader of the truth of his work and he does so by clinging to the ideal found in his rhetorical models, even when describing his own contribution.

Thucydides is more straightforward in his account of his method of dealing with information about what was done during the war than what was said (1.22.2). He deliberately excludes the sort of subjective input he admits for the speeches, and so contrasts the two methods (the contrast highlighted by the μέν and δέ antithesis). He emphasizes the rigorous checking of information (ὅsον δυνατὸν ἀκριβείᾳ περὶ εκάsτου επεξελθών) and notes the sorts of difficulties he had to face (1.22.3). Here he identifies two reasons for his difficulties: the bias and poor memories of his informants. This is a similar concern with the quality of the evidence which we found in his attempt to find out about the distant past too.

Herodotus is often cited in comparison with Thucydides' expressed method at 1.22.2. His disdain of a ‘chance informant’ (οὐκ εκ του̑ παρατυχόντος) is compared with Herodotus' habit of repeating the first report to hand.³⁹ Hornblower thinks Thucydides has a broader target here and he is probably right.⁴⁰ Thucydides does not identify a particular rival, but condemns the method Herodotus (and no doubt others) used. Forensic inquiry, however, can provide us with examples of the principles of inquiry to which Thucydides alludes here. For Thucydides is commenting upon the way evidence is dealt with. He condemns people who do not test-by-torture the evidence they receive, using a forensic word ἀβαsανίsτως (1.20). The prosecutor in Antiphon, Prosecution for Poisoning 13 makes a similar charge of the defence (cited above). A forensic speaker needed to assess whether a witness was trustworthy or not, and commonly questioned the reliability of the evidence provided by witnesses, as Thucydides said he did.⁴¹ The word Thucydides uses for his investigation, επεξέρχομαι (1.22.2), is a word used to describe the process of investigating the testimony of witnesses by Antiphon too (Prosecution for Poisoning 6):

καίτοι αὐτὸ του̑το εχρη̑ν ὃ καὶ εγo προὐκαλούμην, προθυμηθη̑ναι, ὅπως τὸ πραχθέν ηἀληθω̑ς επεξελθει̑ν.

Although he should have welcomed what I challenged him to do, so that what was done might be investigated honestly.

Antiphon, in this prosecution speech, is criticizing the defendant for not allowing the investigation of the testimony of the slaves by torture to get to the bottom of the matter (sαφω̑ς εἰδέναι, Prosecution for Poisoning 6). Thucydides uses επεξέρχομαι again with a forensic connotation in Book 3, in the prosecution speech by the Corinthians against the Plataeans (3.67).⁴² We should also notice that Antiphon will often criticize his opponent for not torturing, which is Thucydides' point: others do not test their sources. In the Prosecution for Poisoning 5-13, for example, the prosecutor criticizes his half-brother, who is conducting the defence, for not using torture to settle the matter. The difficulties posed by poor memory or possible bias on the part of witnesses, points made by Thucydides in 1.22.3, were explicitly recognized in forensic rhetoric too. In his Art of Public Speaking (of which we only have a fragment preserved in Longinus), Antiphon points out that it is not natural to remember accurately (Antiphon fr. C3.1 [Blass: Maidment]).

Witnesses in court were notoriously unreliable. In Herodes 18-19 the speaker complains about the disadvantages he is suffering in presenting his case in court because the prosecution refused to let him out of prison on bail. He claims that this will make even his friends willing to tell lies for the prosecution. The possibility of false evidence from witnesses was itself a cliché in rhetoric: the idea is found, for example, in Andocides, Mysteries 7; Lysias 19.4; Demosthenes, Against Neaera 6; Against Callippus 1, 16; Against Aphobus III 5, 22-8; Gorgias, Palamedes 23, as well as Antiphon. A speaker will claim that he is presenting the facts (Choreutes 8) and presenting them correctly, whereas his opponent has created a tissue of lies (ibid. 9). In On the Choreutes 18, Antiphon argues that what happened is a matter for the truth and justice, not merely (false implied) words. So, speakers concluded, witnesses had to be checked carefully. Antiphon includes something of the process expected to find reliable testimony in, for example, Choreutes 23:

καὶ ἰέναι εκέλευον λαβόντα μάρτυρας ὁπόsους βούλοιτο επὶ τοὺς παραγενομένους, λέγων αὐτἳ̑ ὀνόματι εκαsτον, τούτους ερωτα̑ν καὶ ελέγχειν.

And I kept telling him to take as many witnesses as he wanted and go to those who had been there, name each of them, and question and examine them.

So we find that both Thucydides' concern with checking the testimony of witnesses and the expression of that concern are forensic in character, arising from elements of persuasion necessary to that genre. For witnesses were the primary means of establishing what happened in forensic reconstructions of the past. Antiphon makes the point about the primacy of evidence from witnesses well in Choreutes 30.⁴³ In concentrating his attention on his witnesses, Thucydides neglects to mention other forms of information which he (or perhaps his informants) used, such as inscriptions (6.54, 59.3, 55.1, 1.134.4), treaties (4.16, 5.18, 47, 77, 79, 8.18, 37, 58, cf. 1.44), and material evidence (1.8, 10, 2.15, 1.93.2). Nor does he discuss other principles of method which he used, such as argument based on τὸ εἰκός or ἡ φύsις (a principle he perhaps alludes to in 1.22.4).

τὸ εἰκός is used explicitly throughout the Archaeology (1.9.5, 1.10.2-4, 1.11) particularly in assessing written testimony about the past, such as Homer. It appears regularly in the speeches Thucydides records. Interestingly it also appears explicitly in the construction of the narrative at 8.46.5, 8.87.3, and 6.60.2. The use of the argument based on τὸ εἰκός was itself sophistic in character, and was a common element in forensic and other contemporary rhetoric.⁴⁴ Thucydides' use of an argument based upon ἡ φύsις, evidence from character, is thoroughly forensic in character too. This is found regularly in Thucydides' speeches, but not just in the speeches.⁴⁵ Thucydides refers to human nature as a constant in the narrative at 2.50, 3.82.2, 3.84.2, 4.108.4, and 5.68. The principle was no doubt of importance to Thucydides in re-creating the thoughts, motives, and intentions of characters in his history (e.g. his knowledge of Cleon's thoughts in 4.27-8).⁴⁶ So Thucydides does not discuss all his methods nor display a full awareness of them. Rather he concentrates on justifying his version of the events during the war from the careful use he has made of eyewitnesses. Both the justification itself and the focus on witnesses are forensic in character.

At 1.22.4 Thucydides includes an apology for the unromantic character of his work:

καὶ ες μὲν ἀκρόαsιν ἴsως τὸ μὴ μυθω̑δες αὐτω̑ν ἀτερπέsτερον φανει̑ται· ὅsοι δὲ βουλὴsονται τω̑ν τε γενομένων τὸ sαφὲς sκοπει̑ν …

And perhaps the unlegendary nature of my work will offer very little pleasure when it is heard; but all those who want to consider the clear facts of what happened …

This reiterates Thucydides' earlier contrast between his work, with its focus on the truth, and the compositions of the poets and logographers who were concerned more with entertaining than telling the truth, and had won their way into τὸ μυθω̑δες (1.21.1).⁴⁷ The apology for the lack of τὸ μυθω̑δες thus serves a rhetorical purpose in reinforcing the reader's (or listener's) view of the correctness of Thucydides' methods in achieving τὸ sαφές. There may, however, be further inspiration for this rhetoric from another forensic topos. That topos is the prejudice, common in Greek rhetoric, against a clever speaker. This prejudice is identified by Gorgias in Helen 13 in his discussion of how persuasion in a speech manipulates the mind of the listener:

… δεύτερον δὲ τοὺς ἀναγκαίους διὰ λόγων ἀγω̑νας, εν οiς εiς λόγος πολὺν ὄχλον ετερψε καὶ επειsε τέχνη γραφείς, οὐκ ἀληθείᾳ λεχθείς.

… and secondly forced trials by means of speeches in which one speech delighted and persuaded a large crowd because it was written with skill, not because it was spoken with truth.

A clever speaker, we are told, persuades through his skill at writing the argument, not by telling the truth (τέχνη γραφείς, οὐκ ἀληθείᾳ λεχθείς). This prejudice is commonly alluded to in Antiphon's forensic oratory. His speakers would try to convince the jury that they would be persuaded in the speaker's favour by the truth, not by mere rhetorical skills (e.g. Herodes 1-7). A clever speaker, in fact, needed to apologize for his rhetorical expertise, and argue that despite his skill he was telling the truth (Second Tetralogy 2.2). Notice in Helen 13 that pleasing the crowd (πολὺν ὄχλον ετερψε) is identified as an element of the artfully but not truthfully composed speech. That Thucydides was well aware of this prejudice is clear from a speech he attributes to Cleon at 3.38. There Cleon rebukes the assembly for treating serious political debate as if it were a public display by a sophist, and for enjoying the art of debate on display (38.7). He points out that τὸ εὐπρεπὲς του̑ λόγου will attempt to mislead (38.2). This concern is ironic, as Thucydides earlier characterized Pericles' successors, such as Cleon, as resorting to this rhetorical technique, in contrast to Pericles' own methods (2.65.8, cf. 10). The idea is found elsewhere (e.g. in debates in Herodotus), and is essentially rhetorical in character.⁴⁸ Thucydides, we find, makes a virtue out of the ἀτερπέsτερον nature of his narrative. By pointing out that his work may not be enjoyable (1.22.4) within the context of an assertion of the truth which the history contains, Thucydides has the antithesis between the true and the entertaining underlying his thought. The common rhetorical assertion that the entertaining is not true serves to reinforce the impression of the veracity of Thucydides' work, unentertaining as he claims it to be. Again a rhetorical topos, common to forensic oratory, is useful in understanding Thucydides' rhetoric.

So the works of Gorgias and Antiphon help to put Thucydides' methodological statements in some perspective, providing parallels to the sorts of things Thucydides says about his methods. Gorgias begins his speech about Helen by asserting that what is becoming to speech is truth (κόsμος … λόγῳ δε ἀλήθεια, Helen 1) and that he wanted to demonstrate what was true and bring an end to ignorance (εγo δὲ βούλομαι … δει̑ξαί τε τἀληθες καὶ παυ̑sαι τη̑ς ἀμαθίας, Helen 2). Yet he concludes this techne by admitting that the work is not serious—a mere amusement in fact (παίγνιον, Helen 21).⁴⁹ His claims to the truth are part of this rhetorical exercise, not to be taken seriously—except as a model for forensic rhetoric.

Thucydides' main purpose in the first twenty-three chapters of his history is not to give a summary of early Greek history, but to highlight the greatness of his war by contrasting it with the previous actions of the Greeks, as he tells us in Book One, Chapters 1.2 and 21, and rounds off in Chapter 23, where he belittles the greatest achievements and the disasters of former times.⁵⁰ His methodological statements, including Chapter 22, are made in support of this thesis, hence they are directed to a purpose other than exposition of his methods themselves. What we find when we compare elements that make up his methodological statements with the works of Antiphon and Gorgias, who were developing techniques for persuasive argumentation, is that they reveal common intellectual and rhetorical underpinnings.

This influence should not surprise us. For the influence of the sophists on Thucydides' work, Gorgias in particular, has long been recognized, and the forensic model was a form favoured by Gorgias for instruction.⁵¹ The model format provided not just the basis for the reconstruction of speeches, but the topoi to substantiate a reconstruction of past events. The substantiation of that reconstruction included such rhetorical topoi as the search for the truth, the difficulty of finding the truth, the writer's efforts to acquire and check evidence, the use of eyewitnesses as primary substantive evidence, and the failure of rivals to use appropriate methods.

The techniques of forensic inquiry, especially the use of witnesses, gave Thucydides a sound means of establishing what had happened. His application of such techniques to the writing of contemporary history made a significant contribution to the historian's art. His discussion of these methods is persuasive. Thucydides is concerned to convince us of the truth of his argument and manages to do this in a sophisticated way. This in itself does not mean that his reconstruction of the past is true, any more than the arguments in Antiphon and Gorgias are true, but it also does not mean that it is deliberately false! Thucydides is a clever exponent of the rhetorician's art, and as Gorgias warns us, we should always be careful of logoi written with skill.

Notes

1. D. Fehling, Herodotus and His Sources, trans. J. G. Howie (Leeds, 1989); W. K. Pritchett, The Liar School of Herodotos (Amsterdam, 1993).

2. See 1.20.3 with S. Hornblower, A Commentary on Thucydides vol. 1 (Oxford, 1991) ad loc.

3. E. Badian, From Plataea to Potidaea (Baltimore, 1993). See also J. L. Moles, ‘Truth and untruth in Herodotus and Thucydides’, in C. Gill and T. P. Wiseman (edd.), Lies and Fiction in the Ancient World (Austin, 1993), pp. 88-121.

4. J. L. Moles, JACT Review 2:14 (1993), 14-18, p. 16.

5. C. N. Cochrane, Thucydides and the Science of History (Oxford, 1929); for a more recent opinion, see T. P. Wiseman, Clio's Cosmetics: Three Studies in Greco-Roman Literature (Leicester, 1979), p. 41; for discussion of this point, see A. J. Woodman, Rhetoric in Classical Historiography (London, 1988), p. 5.

6. S. Hornblower, Commentary, pp. 6-7, 59.

7. For 1.1-23 as prooimion, see Dion. Hal. De Thuc. 20.

8. F. M. Cornford, Thucydides Mythistoricus (London, 1907); J. H. Finley, Thucydides (Cambridge, MA, 1942); C. Macleod, Collected Essays (Oxford, 1983), A. J. Woodman (n. 5).

9. Cf. Dion. Hal. De Comp. Verb. 22. For a full discussion, see T. Cole, The Origins of Rhetoric in Ancient Greece (Baltimore, 1991), pp. 73ff. For pathos and ethos in Thucydides, see C. Carey, ‘Rhetorical means of persuasion’, in I. Worthington (ed.), Persuasion: Greek Rhetoric in Action (London, 1994), pp. 26-45, at pp. 27, 33-5.

10. Marcellinus, Life of Thucydides 36, 51; Gorgias fr. 82.A35 Diels-Kranz; Dion. Hal. De Thuc. 19, 24; see A. W. Gomme, Essays on Greek History and Literature (Oxford, 1937), pp. 116-24.

11. V. J. Hunter, ‘Thucydides, Gorgias and mass psychology’, Hermes 114 (1986), 414-29.

12. M. Gagarin, ‘Probability and persuasion: Plato and early Greek rhetoric’, in Persuasion: Greek Rhetoric in Action (n. 9), pp. 46-68, at p. 65; see also A. W. Gomme, A Historical Commentary on Thucydides vol. 2 (Oxford, 1956), p. 131; but cf. K. J. Dover, Thucydides Book VI (Oxford, 1965), p. xvii.

13. Cole (n. 9), pp. 104-11.

14. Aristotle, On Sophistical Refutations 33. 183b36 = Gorgias fr. 82.B14 Diels-Kranz; see also Isocrates 13.12-13.

15. (Plut.) Vit. X Or. 4; Diod. ap. Clem. Alex. Strom. 1.365.

16. For examples of loci communes, probably derived from models: Choreutes 3-6, cf. Herodes 87-9; Herodes 14, cf. Choreutes 2; Prosecution for Poisoning 12f., cf. Herodes 38ff., Choreutes 27; see K. J. Maidment, Minor Attic Orators vol. I (Cambridge, MA, 1941), p. 338 for loci communes in Andocides, On the Mysteries.

17. Cole (n. 9), p. 104.

18. Cole (n. 9), pp. 104-11.

19. See, for example, Hornblower, Commentary, pp. 58-61. For a recent analysis of the problems encountered by scholars in reading 1.22, see C. Orwin, The Humanity of Thucydides (Princeton, 1994), pp. 207-12.

20. Text from D. M. MacDowell, Gorgias. Encomium of Helen (Bristol, 1982); Helen is Gorgias fr. 82.B11 Diels-Kranz.

21. 82.A 28 Diels-Kranz = Plato, Gorgias 453A.

22. In 1.22.1 the idea of the difficulty of obtaining a truthful report is maintained. Thucydides tells us that he has recorded the speeches for us, keeping as close as possible (ὅτι εγγύτατα) to the general intention (/purpose/gist) of what was actually said (τω̑ν ἀληθω̑ς λεχθέντων); the superlative conveys well Thucydides' commitment to accuracy while reiterating the impossibility of verbatim reproduction.

23. This point is made by W. R. Connor, Thucydides (Princeton, 1984), p. 27.

24. See also, for example, First Tet. 4.1, Herodes 86.

25. See also Herodes 86.

26. That is, that the Spartan kings had two votes (Hdt. 6.57) and Pitanate was a deme name applied to a division of the army (Hdt. 9.53, cf. 3.55); see also Thuc. 1.126.7, correcting Hdt 5.71 on Cylon's conspiracy; see A. W. Gomme, A Historical Commentary on Thucydides vol. 1 (Oxford, 1945), pp. 137-8.

27. See Hdt. 5.55.

28. See also Choreutes 47.

29. See also Choreutes 18-19.

30. ‘He set standards of research and accuracy for all time’, S. Hornblower, Thucydides (Baltimore, 1987), p. 30.

31. See also Choreutes 7, First Tet. 1.9.

32. Examples: 1.1, 1.2, 3.3, 6.2, 8, 9.4, 10, 20, 21, ibid. For discussion of the use of these terms in Thucydides, see Hornblower (n. 30), pp. 100-7.

33. The evidence he uses in his preface is wide-ranging: it may be literary, mythological, anthropological, archaeological, geographical, or political: Homer in particular (1.3, 1.9), traditional stories (1.9.2, 1.4, 1.13, 13.5, probably 1.2 as well), archaeological (1.8, 10), geographical (1.7), past and present social customs (1.5.2, 6), or resources (1.1): see Hornblower (n. 30), pp. 73-96.

34. Hornblower finds this vocabulary of evidence unsystematic and non-technical. His comparison between τεκμήριον, sημει̑ον, and μαρτύριον (e.g. at 1.6.2 and 2.15.3 and 1.8.1 respectively) reveals that the three terms are used to mean much the same thing by Thucydides; but the vocabulary of evidence was not strictly technical in the works of such orators of the fifth century as Antiphon either. Such technicality in the use of forensic language was a later development of the fourth century: see Hornblower (n. 30), pp. 100-9.

35. The meaning of τη̑ς ξυμπάsης γνώμης is unclear. Most editors accept ‘general sense’; G. E. M. de Ste Croix suggests ‘main thesis’ (The Origins of the Peloponnesian War [London, 1972]. pp. 7-10); E. Badian has recently reopened the debate, arguing for ‘the entire intention’ (‘Thucydides on rendering speeches’, Athenaeum 70 [1992], 187-90).

36. Thucydides uses τὰ δέοντα four times elsewhere in the context of knowing or advising or doing what political or military actions were demanded by the situation (1.60.5, 70.8, 138.3; 2.43.1). See also τὸ δέον 4.17.2, 5.66.3. Attempts by D. Rokeah to pin down one specific connotation for this phrase in Thucydides are inconclusive: see ‘τὰ δέοντα περὶ τω̑ν αἰεὶ παρόντων: speeches in Thucydides: factual reporting or creative writing?’ Athenaeum 60 (1982), 386-401 and ‘A note on Thuc. 1.22.1’, Eranos 60 (1962), 104-7.

37. For the connection between τὰ δέοντα and Gorgias' work on καιρός, see W. Vollgraff, L' Oraison funèbre de Gorgias (Leiden, 1952), pp. 20-8; see also Gorgias fr. 82.Ala Diels-Kranz (= Philostratus, Lives of the Sophists 11), and A20 (= Plato, Gorgias 447c).

38. ‘Form and meaning in the Melian Dialogue’, Historia 23 (1974), 385-400; G. Kennedy, The Art of Persuasion in Greece (Princeton, 1963), p. 48 also argues that τὰ δέοντα has a rhetorical sense in Thucydides.

39. Hdt. 2.123.1, 7.152.3; see A. W. Gomme (n. 26), ad loc.

40. Hornblower, Commentary ad loc.

41. E.g. Choreutes 28-30

42. Επεξέρχομαι has many senses, including that of prosecuting a legal case (Antiphon, Prosecution for Poisoning 1) and a military sense of ‘to make a sally’ (Thuc. 3.26, etc.); cf. Connor (n. 23), pp. 27-8.

43. Quoted above, pp. 66-7.

44. On τὸ εἰκός in Thucydides, see H. D. Westlake, Essays on the Greek Historians and Greek History (Manchester, 1969), pp. 55ff; cf. Hornblower (n. 30), pp. 106-7. See also J. H. Finley, ‘Euripides and Thucydides’, in Three Essays on Thucydides (Cambridge, MA, 1967), p. 9. On τὸ εἰκός in Greek rhetoric, see Gagarin (n. 12).

45. There are many examples in Thucydides' speeches: 1.76, 86.1, 3.67.2, 1.76.3, 5.105.1, 2, 3.39.5; 45.7; 4.19; 61.5; see de Ste Croix (n. 35), p. 29.

46. See Connor (n. 23), pp. 116-18.

47. For τὸ μυθω̑δες, see S. Flory, ‘The meaning of τὸ μὴ μυθω̑δες (1.22.4) and the usefulness of Thucydides' History', CJ [Classical Journal] 85 (1990), 193-208.

48. Herodotus (Demaratus to Xerxes): βαsιλευ̑, κότερα ἀληθείη χρήsωμαι πρὸς sὲ e ἡδονη (7.101.3). For the idea in Thucydides, see also 7.8.2, 14.4, 3.38.7, 40.3, 1.84.2, 2.65.8, 10, 6.83.2-3. For discussion of the ‘aesthetic of exhibition’ as an element of Greek rhetoric, see J. Poulakos, Sophistical Rhetoric in Classical Greece (Columbia, 1995), pp. 39-46.

49. For comic elements in Gorgias, see P. Harding, ‘Comedy and rhetoric’, in Persuasion: Greek Rhetoric in Action (n. 9), pp. 196-221, at p. 202.

50. See Woodman (n. 5), pp. 5-7.

51. In the absence of the analytical metalanguage of the fourth-century: Cole (n. 9), p. 93.