The most famous – and most discussed – ancient statement on speeches in historiography is probably Thucydides 1.22.1, but Polybius’ discussion of speeches in Timaeus in Book 12 of his Histories follows closely.Footnote ¹ Although Polybius’ criticism of Timaeus has been fruitfully studied from very different angles, the meaning and implications of many of his statements are still debated.Footnote ²

This paper is part of an ongoing project that systematically (re-)examines the role of speeches in both the methodological and the narrative passages of Polybius' Histories. This project was prompted by a chapter on the speeches in the third book of Polybius' work, published in a volume on speeches in ancient historiography in 2010.Footnote ³ Originally, the present paper was to become the second major step of this undertaking, a comprehensive discussion of Polybius’ criticism of the speeches in Timaeus in Book 12, which represents the core of his theoretical discussion of speeches in historical narrative. This large-scale investigation came to an early halt at the programmatic passage 12.25a.4–5, which opens Polybius’ criticism and is key to understanding his approach to speeches in historical narrative, but is also one of the most controversial statements of the entire Histories. Since none of the available interpretations of this passage seemed satisfactory (on the contrary, many, including Walbank's, turned out to be at odds with Greek grammar, as I will demonstrate below), it seemed necessary critically to re-examine the syntax, semantics and grammar of this crucial passage on its own and, if possible, to clarify its meaning, before any comprehensive review of Polybius’ discussion of speeches in Book 12 could be carried out. What follows is therefore a philological analysis with a deliberately narrow focus on 12.25a.4–5; its purpose is to provide more reliable foundations for a comprehensive discussion, which will be undertaken in a separate study.

12.25a.4–5 is part of Polybius' general criticism of Timaeus for deliberately misrepresenting the historical events, a flaw, he notes, strikingly at odds with Timaeus’ constant castigations of the alleged unreliability of other historical accounts. Polybius says he will now provide the ultimate evidence for Timaeus’ deliberate falsification of the past (ταῦτα παρ’ ἀλήθειαν ἐν τοῖς ὑπομνήμασι κατατέταχε Τίμαιος, καὶ τοῦτο πεποίηκε κατὰ πρόθεσιν, 12.25a.4) which will convince even the most stubborn admirers of Timaeus’ work.Footnote ⁴ This evidence, Polybius says, is the speeches in his work, and 12.25a.5 gives the reason (γάρ) for this statement (a translation of the passage is provided at the end of this article):

It is the beginning of the passage (οὐ γὰρ τὰ ῥηθέντα γέγραφεν, οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, ἀλλὰ προθέμενος ὡς δεῖ ῥηθῆναι) which has attracted most critical attention. In particular, scholars have found difficulty with the apparent tautology of οὐ … τὰ ῥηθέντα γέγραφεν and οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν and tried to eliminate this problem by arguing that Polybius is distinguishing three different methods of rendering speeches in historiography, namely a verbatim reproduction, an accurate summary, and free invention by the historian.

Reviewing the arguments on which this assumption of such a tripartite structure of Polybius’ thought is based, I will demonstrate that the interpretation of τὰ ῥηθέντα as meaning a reproduction of the actual speech, and of ὡς ἐρρήθη κατ’ ἀλήθειαν as referring to an ‘accurate résumé of the actual speech’ (Walbank) encounters serious difficulties. On the one hand, these difficulties arise from fundamental issues of Greek semantics, which will be discussed in the first part of my argument; on the other, discussions have often focussed too narrowly on the beginning of the passage without giving due attention to its relation to the points Polybius makes in the remaining part of the sentence. However, as I will argue, this is essential to an adequate understanding of Polybius’ criticism. Therefore, the findings of the first part of my argument will be complemented by the first in-depth analysis of the structure and semantics of Histories 12.25a.4–5 as a whole. Combined with a comprehensive analysis of Polybius’ use of τὰ ῥηθέντα throughout his work, these considerations will provide a strong argument in favour of abandoning the idea that Histories 12.25a.4–5 represents a general statement on historical method and on different ways of rendering speeches in historical works.Footnote ⁵ Rather, as I will argue, Polybius is concerned here exclusively with the direct speeches in Timaeus’ work and their lack of historical reliability.

The most influential interpretation of 12.25a.4–5 is, no doubt, the one proposed by Walbank in his Historical Commentary. Walbank translates the passage as ‘what was said, nor the real sense of what was said’, which he paraphrases as ‘i.e. neither a transcript nor an accurate résumé of the actual speech’. At the same time, he rejects the alternative reading ‘what was said, nor the form in which it was actually said’, which was proposed by Welzhofer in 1880.Footnote ⁶ Similarly, Pédech takes τὰ ῥηθέντα γράφειν to refer to a reproduction of the historical speeches in direct speech, while explaining (like Walbank) ὡς ἐρρήθη κατ’ ἀλήθειαν as referring to an accurate résumé in reported speech of what was said.Footnote ⁷

An alternative explanation of τὰ ῥηθέντα is suggested by Mohm, who paraphrases it as ‘a reproduction of the argument of the speech and its structure, the words being those of the historiographer’ (‘ein Nachzeichnen des Gedankengangs and des Argumentationsablaufs der Rede, wobei die sprachlichen Formulierungen auf den Geschichtsschreiber zurückgehen’).Footnote ⁸ Finally, Nicolai objects to Pédech's interpretation. Referring to 12.25a.5 (quoted above), 25b.1, 25b.4 and 36.1.7,Footnote ⁹ he claims that Polybius never distinguishes between a verbatim reproduction and a summary of a speech.Footnote ¹⁰ Instead, he suggests that ὡς ἐρρήθη κατ’ ἀλήθειαν is merely an explication of τὰ ῥηθέντα and that Polybius employed this slightly tautological expression to emphasise even more effectively that Timaeus did not even intend to provide verbatim reproductions of speeches.Footnote ¹¹

All interpreters of this passage focus on one problem, namely the perceived tautology of οὐ … τὰ ῥηθέντα … οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, and all interpreters, except for Mohm and, probably, Welzhofer, unanimously take τὰ ῥηθέντα to mean ‘the speeches as they were given’, that is, as an expression synonymous with οἱ ῥηθέντες λόγοι.Footnote ¹² If τὰ ῥηθέντα refers to the speeches as they were actually pronounced by the historical actors, then ὡς ἐρρήθη κατ’ ἀλήθειαν becomes a problem because it seems superfluous.

Only Nicolai, as pointed out above, accepts the idea that this tautology might be deliberate and that ὡς ἐρρήθη κατ’ ἀλήθειαν does not introduce a new point but is added to reinforce Polybius’ previous statement, οὐ … τὰ ῥηθέντα γέγραφεν. The other commentators cited above, by contrast, sought to solve the tautology by assigning a new meaning to ὡς ἐρρήθη κατ’ ἀλήθειαν. They regarded the phrase as the middle element of an enumeration of three different methods of representing speeches in historical narratives:Footnote ¹³ first, Timaeus could have given the verbatim reproduction of the speeches, τὰ ῥηθέντα, which he did not do; alternatively, he could have given an accurate résumé of what was said, ὡς ἐρρήθη κατ’ ἀλήθειαν – a reading which is obviously influenced (and, in this case, led astray) by Thucydides’ famous ἐχομένῳ ὅτι ἐγγύτατα τῆς ξυμπάσης γνώμης τῶν ἀληθῶς λεχθέντων (1.22.1);Footnote ¹⁴ finally, the third option – and the one which Timaeus chose – the historian could simply compose the speeches himself according to the rules of rhetorical compositions, προθέμενος ὡς δεῖ ῥηθῆναι.Footnote ¹⁵

As I will now demonstrate, the interpretation of ὡς ἐρρήθη κατ’ ἀλήθειαν as ‘a résumé of the actual speech’ (Walbank) is ruled out by Greek language use. The main problem is that ὡς ἐρρήθη cannot refer to the content of a speech to the exclusion of its form or style; but this is what a ‘résumé of a speech’ must mean, summarizing only the arguments (possibly in the same order as they were proffered by the speaker) but not the exact words of the speaker, as explicitly stated by Pédech (‘au style indirect’).Footnote ¹⁶ But for ὡς ἐρρήθη to refer exclusively to the content of a speech, one would have to substitute ὡς with οἷα οr ἅ.Footnote ¹⁷ Such a distinction is to be expected especially in a rhetorical context such as the present one: at least since Aristotle, authors of rhetorical handbooks clearly distinguished what was said from how it was said, the first referring to the content of the speech, the πράγματα, the second to its form, the λέξις. Aristotle makes a point of differentiating between ἃ δεῖ λέγειν and ὡς δεῖ εἰπεῖν (Rhetorica 3.1403b16) and, shortly afterwards, contrasts the ‘convincing’ (πιθανόν) that is based on the facts (πράγματα) with the ‘convincing’ that is based on the form or style of the speech (τῇ λέξει).Footnote ¹⁸

A close look at the structure of Polybius’ argument supports this interpretation: first, the poignant parallel of ὡς ἐρρήθη κατ’ ἀλήθειαν and the following ὡς δεῖ ῥηθῆναι renders it highly improbable that ὡς in ὡς ἐρρήθη κατ’ ἀλήθειαν should be taken as limited to either the style or the content of the speeches: since ὡς in ὡς δεῖ ῥηθῆναι clearly refers to both style and content of the speeches in Timaeus’ narrative indiscriminately, the same must be assumed for ὡς in ὡς ἐρρήθη κατ’ ἀλήθειαν, lest the parallel become meaningless.

Second, there is a clear parallelism between the beginning of 12.25a.4–5 and its end: the bipartite expression which concludes the passage, ἀπόδειξιν τῆς ἑαυτοῦ δυνάμεως ποιούμενος, ἀλλ’ οὐκ ἐξήγησιν τῶν κατ’ ἀλήθειαν εἰρημένων, not only summarizes Polybius’ main point, that the speeches in Timaeus are the historian's own composition and not an accurate representation of what was really said.Footnote ¹⁹ The two parts of this phrase also resume chiastically Polybius’ opening statement, with ἀπόδειξιν τῆς ἑαυτοῦ δυνάμεως ποιούμενος referring back to προθέμενος ὡς δεῖ ῥηθῆναι and οὐκ ἐξήγησιν τῶν κατ’ ἀλήθειαν εἰρημένων to οὐ … τὰ ῥηθέντα γέγραφεν, οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν. Since Polybius does not make any distinction between the form and the content of the speeches here, it seems safe to assume the same for τὰ ῥηθέντα and ὡς ἐρρήθη κατ’ ἀλήθειαν at the beginning of the passage.

These considerations lead us to the first conclusion of this argument, that ὡς ἐρρήθη κατ’ ἀλήθειαν cannot mean an ‘accurate résumé’ (in indirect speech) of a historical speech and, therefore, cannot refer to the content of a speech to the exclusion of its form: ὡς in ὡς ἐρρήθη and ὡς δεῖ ῥηθῆναι can refer either to the style (λέξις) of a speech only (which does not make much sense in the present context),Footnote ²⁰ or to both style and content of a speech together (the speech ‘as it was uttered’), but never to the content of a speech alone.

Τὰ ῥηθέντα, by contrast, can denote a verbatim reproduction of an original speech, but it can equally be used to refer to the opposite notion, a mere summary of the content of a speech, regardless of the speaker's actual words.Footnote ²¹ This is how Mohm understood the term,Footnote ²² and this also explains why Welzhofer took ὡς ἐρρήθη to mean ‘the form in which it was said’ without further discussion.Footnote ²³ Hence, τὰ ῥηθέντα can be used both in the sense of ‘the exact speech that was given in style and content’, and in the sense of ‘the content/ main points of what was said’.

Therefore, and this is the second conclusion, if we wanted to uphold the assumption that Polybius is, in fact, distinguishing three different ways of rendering speeches in historical works – accurate reproduction, summary of the content and the historiographer's own compositions – it is ὡς ἐρρήθη, and not τὰ ῥηθέντα, as Walbank and Pédech claimed, which would have to be taken to designate the first alternative; τὰ ῥηθέντα, by contrast, which Walbank and Pédech had taken to refer to a verbatim reproduction of a speech, would then have to be taken to denote an (accurate) summary or résumé of the content of a speech alone. On this basis, Polybius' sentence would translate thus:

At first sight this might seem a satisfactory explanation, but a closer look reveals some difficulties with this, too, which are worth exploring further. First, the use of οὐδέ is problematic. If the above interpretation were correct, Polybius’ arrangement of the three methods of rendering speeches in a historical narrative would be somewhat odd. We expect a climax, starting with the best possible and, from the point of view of historical accuracy, most desirable option, a verbatim reproduction of the speeches, then proceeding to the less desirable but still acceptable option of an accurate summary of the content, and culminating in the absolutely unacceptable method adopted by Timaeus, free composition of speeches by the historian. As it stands, however, the order of the alleged first and second elements of the list is reversed.

We would then have to assume that οὐδέ implies a sense of climax along the lines of ‘Timaeus does not provide a summary of the actual speeches, let alone a verbatim reproduction …’. The difficulty with this assumption is that οὐδέ does not have this meaning. To be sure, it can be used ‘with a sense of climax, “nor even”’,Footnote ²⁴ as in the example cited by Denniston, Plato, Respublica 347d1–2: οὐκ ἔχοντες ἑαυτῶν βελτίοσιν ἐπιτρέψαι οὐδὲ ὁμοίοις. The difference between this and our passage is that in Respublica 347d1–2 οὐδέ introduces an element of inferior degree to the element introduced by οὐ: ‘they cannot socialize with people who are better than themselves nor even with those who are their equals’. This makes sense because οὐδέ, connecting two negative terms with a climax, would naturally be expected to introduce a term that is even ‘more negative’ than, or inferior to, the first one, as in βελτίοσιν and ὁμοίοις: the climax is a negative one. Οὐδέ in Polybius 12.25.a.4, by contrast, requires a ‘positive climax’ because the verbatim reproduction of the original speech (ὡς ἐρρήθη κατ’ ἀλήθειαν) would be even more authentic, and hence desirable, than a (faithful) summary (τὰ ῥηθέντα). The sense required for οὐδέ is therefore ‘let alone’, rather than ‘nor even’,Footnote ²⁵ for which Denniston, not surprisingly, does not cite any passages, because in Greek such a ‘positive’ climax after a negative expression is regularly introduced with μὴ ὅτι.Footnote ²⁶

We are thus left with the alternative either to accept that Polybius might have phrased his statement differently than we would expect (which is certainly an option) or to look for a different explanation. If we opt for the second option (which I do), the above considerations make it seem doubtful that we are dealing with three different elements of an enumeration. This suggests that we should explore different possible meanings of the passage.

This interpretation receives further support from a close examination of the specific uses Polybius makes of τὰ ῥηθέντα when referring to speeches of historical actors; to my knowledge, such a systematic study has never been undertaken. A TLG search for τὰ ῥηθέντα in the Histories shows that Polybius applies the term to refer to (1) a summary mention of a speech without any further specifications regarding content and form, as in 4.26.8Footnote ²⁷ and 28.16.9;Footnote ²⁸ (2) speeches rendered in indirect speech, as in 3.29.1,Footnote ²⁹ 11.9.8,Footnote ³⁰ 15.24.3,Footnote ³¹ 18.8.8,Footnote ³² 23.18.1;Footnote ³³ and (3) speeches rendered in direct speech, as in 5.37.12.Footnote ³⁴ This indeterminacy of the term shows that Polybius uses τὰ ῥηθέντα to refer to speeches neutrally and in a general sense as a historical event (‘the things said’/ ‘the speech’) without any inherent specification as to whether a speech is actually represented in a historical work at all and, if it is, in what way the historiographer processed the event. Particularly significant here is 21.22.1, where Polybius uses τὰ ῥηθέντα to refer to a speech of Eumenes which he had reported previously partly in indirect (21.19.1–5) and partly in direct speech (21.19.6–21.11). Whether the historiographer rendered ‘the things said’ in direct or indirect speech, mentioned them only in passing as a statement of fact, namely that a historical actor gave a speech, or provided a mere summary of the contents, becomes clear only through the context in which the expression is used.

In other words, τὰ ῥηθέντα is not a (quasi-)technical term describing any specific way of representing speeches in historiography, and there is no reason to suppose that 12.25a.4 is the only passage in which Polybius expected his readers to understand it as such. If τὰ ῥηθέντα does not designate any specific way of rendering speeches but requires specification from the context, it cannot stand alone representing the first element of an enumeration which, in turn, leads to the conclusion that the idea of a tripartite structure of Polybius' argument should be abandoned.

This suggests that οὐ τὰ ῥηθέντα γέγραφεν means that Timaeus ‘has not written down what was said’ in as general and as neutral a sense as possible: Polybius begins his argument with a statement of the general problem, namely that there is a discrepancy between the historical event, in this case the documented speeches given by the historical actors, τὰ ῥηθέντα, and the way in which this event is represented in Timaeus’ account (γέγραφεν). The concrete form in which Timaeus (mis-)represents these facts is not yet an issue; it is in the following part of the sentence, starting with οὐδέ, that it becomes apparent through ὡς ἐρρήθη κατ’ ἀλήθειαν (an expression, as argued above, referring to both style and content), that Polybius is here referring to the direct speeches in Timaeus (and these only): all direct speeches in Timaeus are compositions of the historian rather than as-faithful-as-possible reproductions of the original speeches. Again, examination of the overall structure of the passage provides additional support for these conclusions: as pointed out above, there is a noticeable parallelism between ἀπόδειξιν τῆς ἑαυτοῦ δυνάμεως ποιούμενος and προθέμενος ὡς δεῖ ῥηθῆναι on the one hand, and οὐκ ἐξήγησιν τῶν κατ’ ἀλήθειαν εἰρημένων and οὐ … τὰ ῥηθέντα γέγραφεν, οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν on the other. Summing up his main points, Polybius’ statement at the end of his argument also distinguishes between two different ways of rendering direct speeches only: they are either an accurate reproduction of the original speeches, which is what serious historiographers should aspire to, or compositions of the historiographer, which is the method adopted by Timaeus and severely criticized by Polybius.

Hence Nicolai was right in surmising τὰ ῥηθέντα and ὡς ἐρρήθη κατ’ ἀλήθειαν referred to the same notion, and as such could be seen as tautological, but that there was some sort of difference between them. We are now able to explain that this difference consists in the fact that Polybius specifies the implication of the general expression οὐ τὰ ῥηθέντα γέγραφεν by adding the explicatory οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν.

Such a specific use of οὐδέ after a more general element introduced by οὐ, which is sometimes, as in the present case, followed by ἀλλά, and sometimes not, occurs also elsewhere throughout the Histories; it can be seen as yet another instance of Polybius’ tendency to combine expressions which seem to be almost synonymous in order to create as precise an expression as possible.Footnote ³⁵ The following examples demonstrate that, as at 12.25a, the terms thus linked are usually not strictly speaking synonymous and the resulting phrase tautological, but that the second term generally specifies or adds some additional nuance to the (often more general) first term; often, the second term also describes the result or consequence of the first. In all those cases in which this construction is followed by ἀλλά the stylistic effect is a particular emphasis on the element which ἀλλά introduces. As such, the expression is a climactic one, but the emphasis does not increase steadily from the first to the last element; rather, the explication of the first two elements prepares the introduction of and lays stress on the third one.

Examples of this usage include 1.65.3 (Καρχηδονίους δὲ [sc. ἐξεδέξατο] κατὰ τὸν αὐτὸν καιρὸν οὐ μικρὸς οὐδ’ εὐκαταφρόνητος ὁ πρὸς τοὺς ξένους καὶ τοὺς Νομάδας καὶ τοὺς ἅμα τούτοις ἀποστάντας Λίβυας [sc. πόλεμος]);Footnote ³⁶ 3.57.4–5 (ἡμεῖς … τοῦτο τὸ μέρος … παρελείπομεν … πρῶτον μὲν οὐ βουλόμενοι παρ’ ἕκαστα διασπᾶν τὴν διήγησιν οὐδ’ ἀποπλανᾶν ἀπὸ τῆς πραγματικῆς ὑποθέσεως τοὺς φιληκοοῦντας, δεύτερον δὲ κρίνοντες οὐ διερριμμένην οὐδ’ ἐν παρέργῳ ποιήσασθαι τὴν περὶ αὐτῶν μνήμην, ἀλλὰ κατ’ ἰδίαν καὶ τόπον καὶ καιρὸν ἀπονείμαντες τῷ μέρει τούτῳ …; the reader being led astray [ἀποπλανᾶν] would, in fact, be the effect of ‘breaking up’ [διασπᾶν] the narrative, just as a ‘scattered’ [διερριμμένην] mention of such things would result in them appearing ‘subordinate’ [ἐν παρέργῳ] in different narrative contexts);Footnote ³⁷ 10.7.2 (οὐ κατεπέπληκτο τοὺς Καρχηδονίους οὐδ’ ἥττητο τῇ ψυχῇ);Footnote ³⁸ 15.13.3–4 (τοῖς δὲ μισθοφόροις τῶν Καρχηδονίων οὐ συνεγγιζόντων οὐδὲ παραβοηθούντων, ἀλλ’ ἀποδειλιώντων ταῖς ψυχαῖς, πέρας ἐνέκλιναν οἱ βάρβαροι);Footnote ³⁹ 27.9.5 (ποτὲ δὲ καὶ χλευάζειν ἐγχειροῦσι τὸν ἕτερον, οὐ μισοῦντες οὐδὲ καταγινώσκοντες, ἀλλὰ παραδόξως τε συμπαθεῖς γινόμενοι …); 11.4.2 (οὐ γὰρ νῦν πρῶτον οὐδὲ δεύτερον ποιούμεθα πρὸς ὑμᾶς τοὺς ὑπὲρ τῆς εἰρήνης λόγους, ἀλλ’ ἐξ ὅτου τὸν πόλεμον ἐνεστήσασθε).

Comparison with these passages shows that the same syntactic structure and stylistic effect underlie Polybius' οὐ … τὰ ῥηθέντα γέγραφεν οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, ἀλλὰ προθέμενος ὡς δεῖ ῥηθῆναι …Footnote ⁴⁰ Accordingly, not too much emphasis should be placed on the explicatory οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν. As mentioned above, this element is more than just a tautological apposition because it specifies Polybius’ general criticism, that ‘Timaeus has not recorded the things said’, by explicating the consequences of this methodological failure: the style and content of the direct speeches in Timaeus’ work are inauthentic. But in order to appreciate the stylistic effect of the phrasing it is necessary to regard the first and the second elements as being closely interrelated: grammatically, both are objects dependent on γέγραφενFootnote ⁴¹ which are combined as a pair by οὐδέ.Footnote ⁴² This prepares the introduction of Polybius’ actual point: Timaeus’ historical work is condemnable not simply because the speeches are inauthentic. Timaeus deliberately falsified history by substituting for the true speeches his own rhetorical showpieces, designed solely to exhibit his rhetorical ability; in so doing, he made it impossible for the reader to achieve any clear and reliable understanding of what happened and why, which Polybius regards as the essence of any respectable historical writing.Footnote ⁴³

So far, this discussion has demonstrated the difficulties inherent in two generally accepted and interrelated assumptions about Polybius’ criticism of speeches in Timaeus, namely (1) that the passage constitutes a general statement on historical method regarding different ways of representing speeches in historical works, and (2) that Polybius distinguishes three different ways of rendering speeches in historiography, ordered as a climax descending from the most desirable to the least acceptable.

By contrast, the in-depth discussion of the semantics and syntactic structure of the passage has revealed two things. Firstly, unlike Thucydides 1.22.1, Histories 12.25a.4–5 is not a general theoretical statement on different methods of rendering speeches in historiography,Footnote ⁴⁴ the most reproachable of which is illustrated in the case of Timaeus, and we should not read Polybius’ argument in the light of the statement of his predecessor. On the contrary, 12.25a.4–5 is concerned only with the direct speeches in Timaeus’ work specifically. Polybius is not giving a lesson in historiographical method in general but is criticizing one particular aspect of Timaeus’ work, which in his opinion best illustrates the unreliability of the work as a whole. And the direct speeches are particularly suitable for that purpose because every reader can compare the original speech (as far as it was accessible, but that does not concern us here) with what Timaeus has written. And Timaeus’ procedure is so very problematic because, if accurately reproduced, direct speeches can offer the readers an immediate and authentic access to the past and the minds of the historical actors themselves. In stark contrast, the speeches in Timaeus’ work grant the readers access to Timaeus’ mind only and thus provide no valuable insight into the reasons and factors that determined the course of the past. This explains why Polybius thought that the (direct) speeches in Timaeus were particularly suitable to prove the general unreliability of his work as a whole (12.25a.1–4).

Secondly, instead of a tripartite structure we are, in fact, dealing with a bipartite structure which also occurs elsewhere in the Histories. This overall structure is designed to lay stress on the second element (introduced by ἀλλά) by creating a strong contrast with the first one. This contrast owes its effectiveness to an emphasis on the first element which is created by its subdivision into two closely interrelated, yet not simply tautological, components by means of οὐ … οὐδέ.

The following is a translation that makes no attempt to smooth over Polybius’ somewhat cumbersome diction, but instead adheres as faithfully as possible to the original grammar and syntax (the impersonal aorist passive verbs are particularly difficult to render): ‘Not the things that were spoken has he recorded and how it was really spoken, but presenting us with how it should have been spoken …’.

As mentioned above, scholarly interest in 12.25a.4–5 has focussed almost exclusively on the first part of the sentence, while the rest of Polybius’ statement has been neglected. This has obfuscated the fact that Polybius’ remarks about the (lack of) reliability of Timaeus’ speeches is, in fact, not an independent point made for its own sake. They are part of a more general criticism of characteristics of Timaeus’ narrative that are centred on, but not limited to, the way in which Timaeus handles direct speeches. In the remainder of this article I will demonstrate how Polybius’ criticism of the direct speeches in Timaeus relates to the remaining part of the sentence and Polybius’ statements made there.

Concentrating on defining how οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν relates to what precedes it, scholars have disregarded the fact that Polybius is, in fact, making two points about the way in which Timaeus handles speeches in his work, each point being represented by one of the two main verbs, namely γέγραφεν in οὐ … τὰ ῥηθέντα γέγραφεν and ἐξαριθμεῖται in πάντας ἐξαριθμεῖται τοὺς ῥηθέντας λόγους καὶ τὰ παρεπόμενα τοῖς πράγμασιν. These two main points are connected by οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, ἀλλὰ προθέμενος ὡς δεῖ ῥηθῆναι. In what follows, particular attention will be paid to the function of this juncture.

Before doing so, however, we need to consider one preliminary issue concerning the meaning of πάντας … τοὺς ῥηθέντας λόγους that was raised by Walbank. In his comment on the phrase, Walbank explains it as: ‘“all these speeches”, that is, all the speeches according to the historian's version, not the real speeches, which P. insists were not recorded … P. does not express the distinction very clearly’.Footnote ⁴⁵ I will argue now that this explanation is unconvincing for several reasons and that we should take ῥηθέντας to refer to the actual, historical speeches and not, as Walbank suggests, to the speeches as invented by Timaeus.

To begin with, the aorist participle ῥηθέντας makes sense only if Polybius is referring here to speeches that were actually uttered in the past, and it is in this sense that Polybius uses combinations of passive participles of verbs of speaking, with or without κατ’ ἀλήθειαν, ἀληθῶς, or similar adverbial expressions, both throughout his work and in our passage itself. Nicolai, in fact, quotes three such instances, erroneously referring to them as evidence that Polybius did not distinguish between knowledge of the original speeches and the way these are rendered in the historical account.Footnote ⁴⁶ Yet in 12.25b.1, τῆς ἱστορίας ἰδίωμα τοῦτ’ ἐστὶ τὸ πρῶτον μὲν αὐτοὺς τοὺς κατ’ ἀλήθειαν εἰρημένους, οἷοί ποτ’ ἂν ὦσι, γνῶναι λόγους, the added οἷοί ποτ’ ἂν ὦσι makes it unmistakeable that Polybius is here referring to the original speeches as given by the historical actors, and the contrast of τοὺς ῥηθέντας λόγους and ψευδῆ δ’ ἀντὶ τούτων ἐπιχειρήματα καὶ διεξοδικοὺς … λόγους in 12.25b.4, as well as that of <τὰ> κατ’ ἀλήθειαν ῥηθέντα <καθ’> ὅσον οἷόν τε πολυπραγμονήσαντας διασαφεῖν and οὔτε (sc. πρέπει) τοῖς ἱστοριογράφοις ἐμμελετᾶν τοῖς ἀκούουσιν οὐδ’ ἐναποδείκνυσθαι τὴν αὑτῶν δύναμιν in 36.1.7, have the same effect. To this list we can add the final part of our passage, which was discussed above, where again the contrast with ἀπόδειξιν τῆς ἑαυτοῦ δυνάμεως makes it clear that τῶν κατ’ ἀλήθειαν εἰρημένων refers to the original speeches as they were actually uttered by the historical actors, and not to the speeches as they were composed by the historiographer. For the representation of speeches in historical works, by contrast, Polybius uses γράφειν, as at the beginning of our passage, where both terms are combined in οὐ [...] τὰ ῥηθέντα γέγραφεν.

Secondly, the verb ἐξαριθμεῖται itself precludes that Polybius is referring with ῥηθέντας to the speeches composed by Timaeus. Ἐξαριθμέω means to ‘recount’, that is, give a full enumeration of a given set of objects without any exceptions, and in this sense it often (as in the present passage) connotes a tiresome, unnecessary process.Footnote ⁴⁷ By definition, the process of ‘re-counting’ requires a pre-existing number of items which someone chooses to enumerate completely instead of selecting only the most significant elements. These considerations lead us to the conclusion that πάντας ἐξαριθμεῖται τοὺς ῥηθέντας λόγους refers to the speeches that were actually uttered by the historical actors.Footnote ⁴⁸

In order to understand the function of this expression, it is important to note that we are already in the second part of Polybius’ argument in which he is concerned with yet another flaw of Timaeus’ historical work. In the first part of the sentence, οὐ γὰρ τὰ ῥηθέντα γέγραφεν, οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, ἀλλὰ προθέμενος ὡς δεῖ ῥηθῆναι, Polybius criticizes the general lack of authenticity of the direct speeches in Timaeus’ work. Now he moves on to his second point of criticism, namely the number of speeches included by Timaeus: Timaeus’ fault is not only to have composed the speeches in his work himself. He also did not distinguish between relevant and irrelevant speeches that were given in a certain historical situation but included each and every speech of any historical actor because this provided him with even more opportunities to substitute his own compositions for original speeches and thus to show off his rhetorical skills.Footnote ⁴⁹

This is in keeping with Polybius’ criticism of the negative influence of rhetorical education on the reliability of historiography elsewhere in the Histories: rhetorical declamation induces authors such as Timaeus to show off their rhetorical skills by composing declamatory show pieces – instead of including the original speeches, even though they might have been of mediocre quality – as well as showing off their historical knowledge by adding as many circumstantial details as possible, relevant or not. By contrast, Polybius repeatedly stresses that it is a core competence of the true historiographer to identify, and report, only those speeches that had a decisive impact on the course of events, for example, at 12.25i.5 (τὸ μὲν ματαίως καὶ ἀκαίρως [καὶ] πρὸς πάντα πάντας διεξιέναι τοὺς ἐνόντας λόγους contrasted with τοὺς ἁρμόζοντας καὶ καιρίους ἀεὶ λαμβάνειν), 36.1.7 (the historian's duty is <τὰ> κατ’ ἀλήθειαν ῥηθέντα <καθ’> ὅσον οἷόν τε πολυπραγμονήσαντας διασαφεῖν, καὶ τούτων τὰ καιριώτατα καὶ πραγματικώτατα) and 2.56.10.Footnote ⁵⁰

In the latter passage Polybius objects to historiographers who make use of these rhetorical practices (τοὺς ἐνδεχομένους λόγους ζητεῖν καὶ τὰ παρεπόμενα τοῖς ὑποκειμένοις ἐξαριθμεῖσθαι) in order to evoke pity and horror, effects normally associated with tragedy (καθάπερ οἱ τραγῳδιογράφοι), instead of doing what they are supposed to do, namely ‘to report what was actually done and said, even if it happens to be entirely mediocre’ (τῶν δὲ πραχθέντων καὶ ῥηθέντων κατ’ ἀλήθειαν αὐτῶν μνημονεύειν πάμπαν, <κ> ἂν πάνυ μέτρια τυγχάνωσιν ὄντα). Comparison of Polybius’ objections to the practice of these ‘tragic’ historiographers in this passage with his criticism of Timaeus at 12.25a.5 is helpful further to elucidate the latter passage. The methods of Timaeus and the ‘tragic’ historiographers have in common that both deflect from the actual events and apply rhetorical methods of amplification instead: they invent speeches and include as much circumstantial description as possible.

Yet two important differences need to be stressed: first, both Timaeus and the ‘tragic’ historiographers composed the speeches in their works themselves. But Timaeus included all speeches that were given in a specific situation indiscriminately, that is, all the speeches in Timaeus were actually given at some point, it is only their style and content that are inauthentic, and many of them are irrelevant. The ‘tragic’ historiographers, by contrast, also include speeches that were never actually given, but could have been given in a certain situation (τοὺς ἐνδεχομένους λόγους),Footnote ⁵¹ that is, not only content and style of their speeches are inauthentic but also the occurrence of the speeches itself. Second, the ‘tragic’ historiographers apply the rhetorical techniques for the specific aim of inducing a particular emotional response in their recipients, while Timaeus applied them simply to show off his rhetorical education (ὥσπερ ἀπόδειξιν τῆς ἑαυτοῦ δυνάμεως ποιούμενος, 12.25a.5).

Returning to 12.25a.5, we can now identify two interrelated elements that constitute Polybius's criticism of Timaeus. Firstly, the direct speeches in Timaeus are inauthentic: Timaeus wants to show off his rhetorical abilities and therefore composes the speeches himself. And secondly, owing to the same flaw, Timaeus replaces each and every historical speech indiscriminately with one of his own, which he supplements with an unnecessary wealth of details about their circumstances. But many of these speeches along with the circumstances in which they occurred are historically irrelevant and should therefore never have been included. In contrast to the ‘tragic historians’, however, Timaeus does not seem to have invented speeches which have never been given at all. Whether this is accurate or not, we have no means to determine; at any rate, Polybius does not criticize Timaeus for inventing speeches that were never given.Footnote ⁵²

In the two ὡς-sentences Polybius has created a sophisticated transition between these two points. From the point of view of grammar, each of these ὡς-sentences belongs to one of the main verbs that form the core of Polybius’ statement: as the above discussion has shown, (οὐδ’) ὡς ἐρρήθη κατ’ ἀλήθειαν and (οὐ) τὰ ῥηθέντα both represent direct objects dependent on γέγραφεν, while ὡς δεῖ ῥηθῆναι depends on προθέμενος which, in turn, is governed by ἐξαριθμεῖται.

By way of contents and phrasing, by contrast, this seemingly clear-cut attribution is deliberately undercut; several stylistic properties invite the reader to perceive them as forming a pair despite the fact that they depend on different finite verbs and, therefore, grammatically belong to different parts of the sentence: both are introduced with ὡς, contain a passive aorist form of λέγειν, and consist of the same number of words (if we include προθέμενος); moreover, the chiasmus of κατ’ ἀλήθειαν and δεῖ poignantly highlights the key notions of historical accuracy as opposed to composition by the rules of declamation. Finally, the adversative ἀλλά, which connects the two parts of the sentence, does not connect the two main verbs directly but sets up an opposition between the objects of the first main verb, οὐ τὰ ῥηθέντα and οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, and the object of the participle, the second indirect question, ὡς δεῖ ῥηθῆναι; all this contributes to the two indirect questions acquiring the weight of almost a third point besides the main verbs, the principal carriers of meaning of the two parts of the sentence.

The result is a third centre of meaning between the two verbs, which has more weight than a mere apposition but does not constitute a fully independent, additional, point of its own. Rather, it bridges the main points by simultaneously explicating the first one and facilitating the transition to the second. Apart from the aforementioned stylistic features, the participle plays a crucial role in this process: after the first part of the sentence, the reader might have expected the main point of the passage to be the opposition between recording speeches as they were actually given and composing speeches according to the rules of school rhetoric. Instead of ὡς δεῖ ῥηθῆναι depending on a participle, we would have expected the indirect question to depend on a finite verb in order to complete the opposition: ‘not the things that were spoken has Timaeus recorded and how it was really spoken, but he presents us with how it should have been spoken …’

Instead, by making ὡς δεῖ ῥηθῆναι depend on a participle, Polybius lays more emphasis on it than on the preceding οὐδ’ ὡς ἐρρήθη κατ’ ἀλήθειαν, which, being the explication of οὐ τὰ ῥηθέντα, has no verbal form of its own to support it; at the same time, however, ὡς δεῖ ῥηθῆναι is not on the same syntactic level as οὐ τὰ ῥηθέντα, which directly depends on the first finite verb (γέγραφεν) and together with it constitutes the first main and independent point. Rather, it is subordinate to, and so prepares the reader for, the introduction of Polybius’ second main point, which is represented by the second finite verb (ἐξαριθμεῖται). Consequently, there is a certain tension between grammar and style in Polybius’ sentence: the style mitigates the grammatical association of each of the two elements of the middle part of the sentence with Polybius’ first and second main point, respectively. Polybius thus prevents the two finite verbs from being the only carriers of meaning and the sentence from falling into two clearly distinguished parts. Instead, this construction allows him to tie the two main verbs closely together by way of an elaboration of his first point which, at the same time, introduces his second point by illustrating it (προθέμενος ὡς δεῖ ῥηθῆναι), thus forcefully expressing the close interrelation of what he regards as Timaeus’ key faults: the lack of authenticity of the direct speeches combined with his inability (refusal?) to limit the direct speeches in his work to the historically relevant ones only.

I bring this article to a conclusion by proposing a full translation of Histories 12.25a.4–5 which takes into account the results of the preceding discussion: