6

Truth and fiction in Plutarch’s Lives

As for Solon’s meeting with Croesus, some scholars fancy that they have disproved this on chronological grounds. Yet, when a story is so famous and well-attested, when (more important) it fits Solon’s character so well, and is so worthy of his wisdom and largeness of spirit, I am not prepared to reject it because of the so-called rules of chronology. So many scholars are continually revising these rules, and still there is no agreement on how the inconsistencies are to be reconciled.

(Plutarch, Solon, 27.1)

That passage captures a lot about Plutarch’s approach to history, and to truth. Such an approach to chronology is not ridiculous. In this particular case the chronological experts doubtless had it right and Plutarch was wrong; but there was a wider sense in which he was right to be sceptical of the chronologists, even if this made him credulous of the story. The chronologies of early Greek, Persian, and Lydian history were hardly secure, and the different systems were not easy to harmonize. Plutarch knew the problems very well.¹ It would have been rash for him to build large consequences on such speculative grounds, particularly as the meeting of Solon and Croesus afforded one of the more firmly attested synchronizations for the experts to build their systems around; and, even on the experts’ own terms, the case against the story was scarcely clear-cut.² Plutarch was not indifferent to chronology,³ but he liked it to rest on firmer grounds.

Still, his cavalier attitude here certainly gives one pause. He makes no attempt to grapple with the chronological problem himself, he simply discards it; his placing of the story implies that he accepted the traditional dating shortly after Solon’s reforms, a context which made the chronological problem more acute;⁴ and, after he has made the reasonable point that the story is supported by a wealth of evidence, we are surprised that he goes on to put more weight on the way it fits Solon’s character – precisely the reason, of course, why it was doubtless made up. Momigliano has argued that ‘the borderline between fiction and reality was thinner in biography than in ordinary historiography’,⁵ and that ‘the biographers felt in principle much freer than the historians in their use of evidence’.⁶ This, perhaps, is one instance where Plutarch shows less concern to investigate historical truth than we should like.

Still, Momigliano’s thesis is a large one, and deserves more critical discussion than it has received; and the enquiry will now be more complex because of well-argued attempts to demonstrate that ‘ordinary historiography’ itself showed less regard for the truth than Momigliano implies, and perhaps assumed a different concept of ‘truth’ from our own.⁷ Yet the comparison of biography and historiography can still be a rewarding one, and Plutarch is the only Greek political biographer who allows serious analysis – really, indeed, the only substantial figure of whom we even know.⁸ He clearly thought hard not only about biography, but about how history should be written too: On Herodotus’ Malice is sufficient demonstration of that, and the Lamprias catalogue attests a work on How We Are to Judge True History (Lampr. cat. 124). Our enquiry should be profitable.

It will be in three parts, though they overlap. The first will discuss Plutarch’s historical criticism, the criteria he uses for accepting or rejecting material as plausible or implausible. The second will revert to a topic already treated in chapter 4, what he does with a story once he has decided to include it – how far he can add circumstantial detail, how far he can change small details of it, how far he can shift around its chronology, how far he can strain its interpretation to suit his thesis, and so on. The third will return to Momigliano’s formulations, and explore the relation between Plutarch and historiography.

I. Plutarch’s critical ability

Even by modern standards, many of Plutarch’s historical arguments are quite impressive. Take the discussion of Aristides’ wealth at Arist. 1.⁹ He discusses the various arguments used by Demetrius of Phalerum to suggest that Aristides was quite well off, his tenure of the eponymous archonship, his ostracism, and the suggestion that he set up a choregic monument. He counters the last point by suggesting that friends might have subsidized him, as they did with Plato and Epaminondas in similar cases; then he adds the epigraphic argument which he draws from Panaetius – that the letter-forms show that this inscription is of the wrong date, and must have been dedicated by a different Aristides. We would of course think the second point decisive, and wonder why he spent time on the first; still, both the points are reasonable, and he need not have been wholly confident that Panaetius had his letter-forms correctly dated. Against the other points, the archonship and the ostracism, he shows fairly easily that others were ostracized without being rich or noble (Damon), and cites Idomeneus to demonstrate that Aristides was elected archon, not drawn by lot from the pentakosiomedimnoi. Once again, he might have done a little better (e.g. by quoting Aristotle’s Constitution of Athens),¹⁰ but not much: he is using his wide reading and general knowledge very effectively. And, most important, he cared: this is worth two pages of fairly hard argument, and in the first chapter of a pair, where he usually prefers a more gripping and less learned opening – indeed, something more like the reflections on justice in ch. 6, which could easily have been made into a prologue.¹¹

Of course, there is a sense in which the scholarship suits Plutarch’s rhetorical concerns as well. Aristides’ poverty will be important to his literary presentation of the pair, both confirming the famous incorruptibility and making it more remarkable; the elder Cato’s justice, austerity, and domestic management will be rather more qualified and problematic.¹² The point will be important enough for Plutarch to return to it at the close of the Life, with another, balancing display of intelligent learning: Craterus cannot have good authority for his claim that the elderly Aristides was condemned for corruption, for, if he had, it would have been in his manner to quote it, and the item would anyway have featured in the canon of stories of the ungrateful demos (Arist. 26.2–5). Aristides’ poverty is also confirmed by the fate of his children and grandchildren after his death, and again an impressive parade of learning confirms the point (27). But, strictly in terms of forceful argumentation, the scholarly technique is not ideal: Craterus’ obloquy was obscure enough to be ignored rather than countered, and the initial points would have been more effective if they had been briefer.

More important, nothing suggests that the rhetorical point was here in conflict with what Plutarch believed to be true. We often, and fairly, concentrate on the ways in which rhetoric could distract writers from historical truth, but there were other ways in which it positively helped: in fostering a powerful memory, for instance,¹³ and encouraging wide and cultured reading; in providing the techniques to impose a clear, ordered structure on chaotic source-material; in affording a sensitivity to bias; in encouraging discrimination among more or less reliable witnesses; in providing arguments from eikos which could stimulate a sceptical approach to unlikely stories. The Aristides affords a good instance of this, with Plutarch’s rhetorical expertise helping his critical alertness to the truth. Some modern theorists have stressed the analogy between historical research and forensic technique, with historians asking questions of ‘witnesses’ and elaborating techniques to check their testimony:¹⁴ in this case, at least, one sees what they mean.

There are plenty of other examples of good argument: the remark at Crass. 13 that one cannot trust Cicero when he incriminates Caesar and Crassus in a work he published after both were dead;¹⁵ the discussion of Phocion’s social standing at Phoc. 4.1–2, including the sensible remark that, if Phocion had been of low birth, he would not have attended the Academy or indulged in other privileged pursuits in his youth; the use of more reliable chronology to demonstrate that Stesimbrotus had got his dates wrong in making Themistocles the pupil of Melissus and Anaxagoras (Them. 2.5). He is alert to bias as well, as befits one who is familiar with the rhetorical arts of misrepresentation: Oppius cannot be trusted when he writes about Caesar’s friends or enemies, so one cannot believe his story of an atrocity of Pompey during the Sullan period (Pomp. 10.9); Antiphon’s slanderous stories about Alcibiades should be discounted, given the conventions of invective (Alc. 3.2); Andocides’ anti-democratic bias and Phylarchus’ sensational style discredit their stories about Themistocles’ death (Them. 32.4); Duris of Samos can always be relied upon to magnify his country’s sufferings (Per. 28.3); a story of Phylarchus would not be worth accepting but for the fact that it is supported by Polybius, for Phylarchus would always bend the truth in Cleomenes’ favour (Arat. 38.12);¹⁶ and Theopompus is more reliable when praising people than when blaming them, so his version of Lysander’s financial integrity deserves some respect (Lys. 30.2).¹⁷ And he certainly knows when truth is particularly hard to attain, whether because of the remoteness of the period or because it was so heavily overlaid by propaganda (Per. 13.16, Thes. 1.2–5, Lyc. 1.1, Mor. 326a).

He knows the value of contemporary and eyewitness sources, too – that often emerges.¹⁸ Of course, some contemporary sources could be rather embarrassing to have:

It is time for me to appeal to the reader for indulgence, as I treat the events that Thucydides has already handled incomparably: in this part of his narrative he was indeed at his most emotional, vivid, and varied. But do not assume that I am as vain as Timaeus, who thought he would outdo Thucydides in brilliance and show Philistus to be totally vulgar and amateurish... Of course, it is not possible to omit the events treated by Thucydides and Philistus, for they include material which gives an especially clear notion of the man’s character and disposition, so often revealed by his many calamities. But I have summarized them briefly and kept to the essentials, just to avoid the charge of total negligence. I have tried instead to collect material which is not well known, but scattered among other authors, or found on ancient dedications and decrees. Nor is this an accumulation of useless erudition: I am conveying material which is helpful for grasping the man’s nature and character.

(Nic. 1.1, 5.)

It is Thucydides’ artistic qualities, we noticed in the last chapter (p. 117), that make him inimitable: he is so emotional, vivid, varied, brilliant. But, as we also noticed, Plutarch will compensate not by literary virtuosity and not just by altering to a biographical focus: he will find out new facts. All this study of ‘other authors’ and ‘ancient dedications and decrees’ sounds like serious historical enquiry committed to the truth; and indeed Nicias is quite full of such out-of-the-way evidence, including epigraphic material, and he prefers to pursue this approach rather than strain every last ounce from Thucydides.¹⁹ He quite often uses inscriptions elsewhere, too.²⁰ One need not overdo it: there are of course spectacular weaknesses as well, especially when he argues from silence;²¹ his attempts to reconstruct the political climate of a different age can be disquietingly simple;²² and he does not always seem to us to give weight to the right evidence or arguments – in different ways, Arist. 1 and Solon 27 were both examples of that. Even in Nicias Gomme is probably right, and we should not really think of ‘research’: the collection of material was doubtless too unsystematic.²³ But it is not unfair to speak of ‘scholarship’, the thoughtful application of extensive learning in the interest of getting facts right; and, without overdoing it, we can at least say that he is as critical, as intelligent, and as committed to the truth as most ancient writers about the past²⁴ – when he wants to be.

There is evidently a sting in that tail, but we have already seen enough of Plutarch’s commitment to the truth to justify two additional points. First, a good deal of this political biography is very committed to getting it right. Now that evidently contrasts with literary biography, which is so often ‘representative’ and fictional;²⁵ it also probably contrasts with any other models of Greek biography that Plutarch knew. As Momigliano brought out,²⁶ in formal terms the genera proxima to biography were, first, encomium, and secondly the biographical novel on the model of Xenophon’s Cyropaedia. Both, evidently, were quite popular genres in the Hellenistic period, and both certainly indulged in fictional elaboration, particularly of matters such as their subjects’ childhood.²⁷ Momigliano was quite right to observe that the fictional quality of such works could well topple over into biography itself. In the case of literary biography, that indeed happened; but Plutarch’s biography was generally different – even though it was usually public figures, not intellectuals, that those mendacious genera proxima treated, and we might consequently expect political biography to be the more susceptible variety. Perhaps Plutarch had Hellenistic forerunners who were equally historically committed, but I doubt it;²⁸ more likely, he was largely creating his own genre, and creating it in a mould which was unexpectedly truthful, a closer cousin to historiography than to the formally more similar encomium or biographical novel.²⁹

Consider, for instance, his use of Onesicritus’ How Alexander Was Brought Up, apparently a charming, romantic, and extravagant work written on the model of the Cyropaedia.³⁰ Plutarch seems to draw on this a little in the first few chapters of Alexander,³¹ but the remarkable thing is that he does not draw on it more. As Hamilton brings out clearly, he is much more restrained in the Life than in his essays (or speeches) ‘On Alexander’s Fortune’. In the Life, for instance, he does not develop the motif of the ‘philosopher in arms’ civilizing the barbarous world, though we know this featured in the work of Onesicritus (FGrH 134 fr. 17), and we might expect it to be to Plutarch’s taste: it was certainly fundamental to the essays, particularly the first.³² Yet he had sufficient knowledge of the historical tradition to be sceptical. He felt that this sort of material was inappropriate for the sort of biography he was writing, however suitable it might be for a rhetorical essay: and it is pretty remarkable that he did, and that he prevented his biography from drifting into the fictional conventions of its genera proxima. In this as in other ways, Plutarch’s closest precursor was perhaps Nepos, as Geiger has suggested:³³ for Nepos was not, it seems, mendacious, and certainly not sensational. Still, in that case our respect for Plutarch’s massive originality will be all the greater, for the gulf between the two is so immense; and it is still remarkable that Plutarch should have turned to that model, if model it was, rather than any other. For instance, he clearly knew hagiographic literature such as Thrasea Paetus’ Life of Cato.³⁴ but that was not to be his style. There were so many forces which were driving Plutarch, if he wrote biography, into writing in the manner of the Scriptores Historiae Augustae, fictional and extravagant. It is striking that he, like his contemporary Suetonius, chose to write something much more historically faithful.

The second point here is the range of information he adduces – inscriptions, speeches, letters, oral traditions, comic poets, as well as mainstream historians. We are so used to this that we almost take it for granted, and are indeed accustomed to complain about the way he takes the comic poets too seriously;³⁵ but more striking is the fact that he uses them at all. This sort of range does not regularly appear in historiography: some historians would use out-of-the-way material some of the time, of course, but this is still much more the sort of polymathy that we associate with antiquarianism. For instance, Millar commented on Cassius Dio’s strange use of Cicero’s speeches: ‘This illustrates a curious, but important, feature of ancient historiography – while it was possible to use Cicero’s speeches for putting together a speech “by Cicero”, it was not possible to use them to provide evidence for the main narrative; that was supplied by the narrative sources alone.’³⁶ That is surely right about Dio, and probably about Appian too. But Plutarch does use Cicero quite extensively as a historical source, especially the Second Philippic in Antony (we shall come back to that).³⁷ In this respect Plutarch is closer to our modern idea of respectable historical activity than most of the historians were, and there is indeed an essential historical seriousness about what he does. Once again, the parallel with his contemporary Suetonius is suggestive. For all the difference in literary form, there is still a similar use of antiquarian material and method, and a similar commitment to getting things right.³⁸

To go back to that earlier sting in the tail: Plutarch is quite good at historical argumentation when he wants to be. Sometimes – not necessarily very often or extensively – he does not. His biography is a most flexible genre. Sometimes he is very interested in history, in isolating the historical forces that carried Caesar to power, for instance, or Themistocles, or that made and then broke the Gracchi; sometimes his interest in history is much slighter.³⁹ This can be seen most easily in his reluctance to give historical background: there is far more on the historical setting, for instance, in the Caesar than in the Antony, to take two figures who dominated their respective decades. It can be seen in the seriousness of the political analyses he gives, with more trivial versions emerging in some Lives than in others, and it can also be seen in the depth or shallowness of his critical discussions. When his mind is not fundamentally concentrated on the history, his standards of criticism tend to relax. For instance, he knows perfectly well that invective cannot be trusted: as we have seen, he rejects Antiphon’s attack on Alcibiades on those grounds (Alc. 3; above, p. 145). At times, he is therefore critical of Cicero’s invective against Antony – but, if it suits him to stress Antony’s luxury, he will use Cicero.⁴⁰ If the propaganda which reviles Cleopatra fits his general picture of the queen, he will use it, then briefly confess it may be false (Ant. 58.4–59.1). When he comes to discuss less ‘historical’ areas or people – not necessarily those living in mythical times, but rather those who had become enveloped in the sort of tradition which made it difficult to write a more ‘historical’ biography – he can be much less critical of his material and perpetrate much more precarious arguments. Sertorius yearns for the Isles of the Blessed; Antony comes to hate Octavian more fiercely because he always loses at dice; the young Cassius bloodied the nose of the young Faustus Sulla, hence the usual view of Cassius’ motive for tyrannicide must be wrong. We should believe the miraculous stories of Rome’s foundation, despite the criticism of sceptics, for the Roman state would hardly have advanced to such might if it did not have a wondrous and divine origin.⁴¹ Indeed, that argument (or rather lack of it) about Solon and Croesus also belongs in this list. All this does strike a very different tone from the acute, sceptical argumentation discussed earlier; but, for one reason or another, he thought that the sort of material he was treating in these less historically committed passages did not lend itself to argumentation of such rigour. In Aristotelian terms, one should only seek the ἀκρίβεɩα appropriate to the ὕλη, the precision appropriate to the material the writer was describing.

Even here Plutarch is not so far removed from the historians. Livy, for example, knew that people added venerability to antique stories by including the supernatural, and proclaimed his indifference to such material; but, he goes on to say, the Romans went on to win such martial glory that they of all people might naturally be permitted to trace their descent from Mars (praef. 6–7). That is almost exactly the same point as Plutarch makes in Romulus, though Livy puts it a little more sardonically; and it is hardly the style of argument which recurs in his later books. And even Thucydides concludes his Archaeologia with the claim that he has relayed the facts with sufficient accuracy, given their antiquity (ὡς παλαɩὰ εἶναɩ, 1.21.1):⁴² it would be unreasonable, clearly, to expect the same rigour or precision as with more recent and verifiable events. Thus in the Archaeologia he has strained to extract conclusions from Homer, with a surprising lack of scepticism;⁴³ ‘a funny kind of history’,⁴⁴ indeed, and one which is not going to be prominent in the rest of his work. It is unthinkable, for instance, that he would use Euripides as evidence for Athenian war-weariness, or exploit Aristophanes on Cleon; with harder history, that was not his way.

There were various reasons why Plutarch might favour a different style of argument and degree of rigour. It might be the period, as with Romulus; it might be the type of figure – Sertorius really did belong in a different world from Sulla, Pompey, or Caesar, even if he lived in the same times; or it might be the style of Life – Antony or Brutus do have rather different, less ‘historical’, styles than many of the others. It is certainly hard to find arguments as unrigorous or material as implausible as this in Lives like Caesar, or Gracchi, or Themistocles. This also helps to explain a feature which disturbed Momigliano, Plutarch’s readiness to write confidently about figures like Romulus, instead of confessing his inability to judge deficient evidence.⁴⁵ That is true – though it is also true that Plutarch begins Thes.–Rom. by giving a general caveat and asking for indulgence;⁴⁶ but in any case that does not mean that he was as uncritical when he wrote of the later, more solid figures. His genre was flexible enough to accommodate both sorts of figure, but he knew he could not ask the same questions with the same rigour about them all. The use of evidence may ‘lack consistency’, as Momigliano protests;⁴⁷ but not irrationally so. And even when we do find those less rigorous elements, his approach is usually not wholly indifferent to the truth, not simply ahistorical. He may be accepting material too credulously, but in each case it is to support a view that he genuinely holds, and often for good reason: the rest of the Life usually makes that plain. He really believes that Antony and Cleopatra were infatuated and extravagant, that Sertorius was a dreamer, Solon a courageous sage, and Cassius a fierce hater of tyrants: he just allows himself to make the points a little easily. He is not presenting a false picture, just helping his truth along a little.

So far we have been applying our own standards of what a historian ought to be about. Plutarch’s canons were rather different, and he conveniently sets them out at the beginning of On Herodotus’ Malice.⁴⁸ For instance, the historian should not use severe words if milder ones should suffice; he should not drag in irrelevant material which is discreditable to a character, but should not suppress anything that puts him in a good light; when two versions of an incident are current, he should select the more creditable one; when the motivation of an action is disputed, he should give a character the benefit of the doubt; he should not include stray abuse and then admit it to be false, for such innuendo can only detract from a character’s moral stature; and so on. How far did Plutarch the biographer practise what he preached for historians?⁴⁹

The answer is at first sight surprising. The principles certainly strike us as ‘moralistic’: a character must be given the benefit of any ethical doubt, and his moral stature must not be gratuitously impugned. And yet it is (to put it crudely) the most ‘historical’ Lives which are most scrupulous in observing those canons: those Lives such as Caesar or Themistocles, where the moralism is typically subtle and muted, and where the interest in exploring historical background is unusually intense. The moralism of other Lives is more open, for instance in Cato Minor or Antony: yet it is those Lives which more often flout the principles.

In Caesar Plutarch thus makes little of discreditable facts, for instance Caesar’s sexual habits; indeed, he avoids malicious comment even when it would have been justified, for example in his treatment of Caesar’s vulgar demagogy, or his extravagance, or his debts.⁵⁰ Caesar’s ambition was for tyrannical power (Caes. 57.1, 69.1), and the theme is important to the Life:⁵¹ Plutarch of course strongly disapproved,⁵² but in Caesar he kept his disapproval to himself. If there is ever a word of criticism, it tends to be a mild one: for instance, the description of the Anticato is much more temperate in Caesar (54.3–6) than in Cato Minor (11.7–8, 36.5, 54.2). The principles of On Herodotus’ Malice can sometimes be seen working on a smaller scale. When Caesar refuses to give evidence against Clodius but insists on divorcing his wife, Plutarch’s explanation is very proper: some say that Caesar was sincere, others that he was courting the goodwill of the masses (9.10). His historical thesis would really welcome the second interpretation, for he is stressing Caesar’s scheme of rising to power as the people’s champion;⁵³ but he includes the more favourable explanation as well. Again, at Them. 31.5 he strains plausibility in his attempt to find a creditable motive for Themistocles’ suicide; Gracchi 22.6–7 takes the most generous view possible of Gaius’ reasons for entering popular politics.

So far the principles really do seem to obtain; but other Lives are different. Antony retails that long catalogue of abuse which Calvisius cast at Cleopatra – then adds that ‘Calvisius was doubtless lying’ (58.7–59.1). Precisely the sort of innuendo which On Herodotus’ Malice had outlawed. The principles can always be ignored when Plutarch is painting a villain, a Cleon or a Sulla or (in his decline) a Marius. Cleon opposed the peace-proposals in 425 because of his personal enmity with Nicias (Nic. 7.2): that is not what Thucydides had said (4.21), for all his stress on that enmity (4.27.5) and his distaste for Cleon. Sulla had claimed that he indulged the captured Athens because of the world’s debt to Athens’ past: Plutarch dismisses this, for his Sulla is simply ‘sated with vengeance’ (Sulla 14.9). Hardly ‘giving Sulla the benefit of the ethical doubt’. In other Lives Plutarch praises Sulla’s constitutional settlement,⁵⁴ but not in Sulla itself: scarcely ‘including everything to Sulla’s credit’. The second half of Marius is similar in texture: Plutarch regularly assumes the worst about Marius’ motives, often rejecting the man’s own explanations as disingenuous or silly;⁵⁵ and the close of this Life wallows in the slaughter, wilfully exonerating Cinna to emphasize Marius’ own role.⁵⁶

But it is not just ethical outcasts who are denied the benefit of any ethical doubt. When the moralism is crude, even the most favourable Lives can show little respect for the principles. Cato Minor is favourable enough, and Plutarch indignantly rejects the story that Cato was once drunk as praetor (44.2). Yet he still includes that story; it adds nothing, the innuendo can only detract from the subject’s character, and the precepts of On Herodotus’ Malice should have led him to suppress it completely: the historian, we recall, should not include abuse which he then admits to be false. Pericles similarly includes various slanders about the genesis of the war (esp. 31), just as earlier it has mentioned and dismissed Idomeneus’ allegation that Pericles murdered Ephialtes (10.7) and Duris’ charge of Athenian atrocities at Samos (28); and Plutarch has also assumed the worst motivation for appointing Lacedaemonius to the Corcyrean expedition (29.1–2).⁵⁷ At Cic. 29.1–2 he allows that Cicero’s allegations against Clodius were fair and true, but that is not (says Plutarch) why he made them: he was simply trying to put himself right with Terentia. Much of Pompey, too, is sympathetic; but Plutarch rejects out of hand the excuses made by his friends for his conduct in 59 (Pomp. 47.7–8). He includes the attacks made on Pompey for marrying Cornelia, and indeed affords them quite an expansive treatment (55.1–5): a more sympathetic approach would have been possible, and just – and indeed would in some ways have suited the Life, with its engaging stress on Pompey’s uxoriousness. Plutarch then concentrates on the bad aspects of the administration in 52, the favouritism towards Scipio and Plancus and the arrogant treatment of Hypsaeus: then he curtly adds that ‘in all other respects he restored order well’ (τὰ δ’ ἂλλα καλως ἂπαντα κατέστησεν εις τάξιν, 55.11). The balance of the whole treatment could have been more generous. None of this is untruthful: Plutarch really did believe that Cleon was petty, Sulla and Marius murderous, Cato drunken, Cicero henpecked, and Pompey impolitic. But there is surely a certain lack of charity.

And perhaps these variations among the Lives are not so surprising after all. The Lives are sometimes a little removed from historiography, sometimes closer; and On Herodotus’ Malice was giving precepts for historians. Plutarch is alert and sensitive to the possibilities of his genre, and when he was writing a ‘historical’ Life he himself became far more of a historian, keeping more closely to the norms he thought appropriate to that genre. Those norms included both truthfulness and ethical generosity: when he felt it appropriate, he would pursue the truth thoughtfully and rigorously, but also with a proper readiness to err on the side of charity. At other times, he could be more relaxed.

II. Plutarch’s manipulation of his narrative

Quite evidently, Plutarch does not take over his historical material blindly: he does interesting things with it. Sometimes he criticizes it explicitly, as we have seen, making it clear why he is favouring one version or rejecting another; more often, he simply tacitly rewrites it, elaborating, reordering, giving different emphases, often revising the detail. We examined some of these rewritings in chapters 4 and 5. Can we say more about what Plutarch thought he was doing when he recast material in this way?

If we could tackle Licinius Macer or Valerius Antias, I think they would each indignantly deny that their detailed and circumstantial narratives were all made up. I think they would argue that it must have been like that; their justification would be ‘it stands to reason’, with a priori probability – or what seemed to them to be probability – totally outweighing the lack of actual evidence.

Thus Peter Wiseman.⁵⁸ Whatever the case with Licinius or Valerius, what would Plutarch have said? When we can detect him adapting or improving his material, would he have given that explanation, ‘it must have been like that’? Or would it have been a more pungent and robust ‘don’t be a pedantic bore: I’m just making it into a better story’? Perhaps it is a bit of both – so much a bit of both that it is hard to think that Plutarch would have drawn a hard-and-fast line between cases where he was sacrificing the truth and cases where he was reconstructing it. In that case, Wiseman’s analysis may not be quite the right way to look at it, and we may instead have to develop a concept along the lines of ‘true enough’, more true than false. Still, his remark gives a helpfully sharp question, and we shall ask it.

First, though, we should notice what Plutarch does not do. His source-material sometimes left vast gaps, for instance the hole in Themistocles from 493 to 483. Some Lives indeed were veritable string vests, more hole than substance – Phocion, Aristides, Philopoemen, Poplicola, Artaxerxes, even Crassus. But Plutarch does not fill them with fabrication. Normally, we can see, he likes to pair two Lives of roughly similar length, but there are times when he cannot. He simply does not have enough material on Cimon, Phocion, Fabius, Marcellus, or Agesilaus to make them weighty enough matches for their pairs; he does not make it up. And this reluctance to fabricate is particularly plain in his treatment of boyhood and youth.⁵⁹ So many Lives have virtually nothing before adulthood: Antony, Nicias, Phocion, Camillus, Flamininus, Marcellus, Fabius, Timoleon, and more. We could all make up a few good stories about a schoolboy Antony or Nicias; Plutarch does not. Similarly with death: we would be glad to know more about how Camillus or Flamininus met their ends, but Plutarch lets the gaps stand. These are precisely the sort of holes which literary biography would strain to fill; nor would encomium or the biographical novel have been silent. Plutarch is different. When he adds circumstantial detail, it is in much more limited ways: the circumstances of a conversation, perhaps, with a cautious plotter, a boisterous lover, a playful maid, a sneering innkeeper, or an officious servant. He can add a man to hold out his toga for Antony’s vomit; a new river or hill to make sense of a campaign; new circumstances for Coriolanus’ friend, once rich, now in penury.⁶⁰ But the big invention is not in his style.

He does do something to fill childhood gaps, though: the most usual sort of item we find is what one could call ‘the routine generalization’, passages such as Timoleon 3.4–5.

Timoleon was patriotic and unusually gentle – except that he nourished a peculiarly intense hatred for tyranny and for evil people. In his military campaigns he showed such a finely balanced character that he displayed great understanding as a young man and great bravery in his old age.

(Tim. 3.4–5)

Again one notices the lack of the anecdotal register, the reluctance to make up stories.⁶¹ Sometimes the material is more circumstantial, but still not particularly anecdotal: in adolescence Agis abandoned the foppery which typified his degenerate Sparta, and turned to the old ways of austerity (Ag.–Cl. 4.1–2); Coriolanus’ youthful rivals ‘excused their inferiority by attributing it all to his physical strength’ (Cor. 2.2). We shall look at such passages more carefully in a later chapter (ch. 14, pp. 308–10), where we shall see that it is unlikely that much of this material stood in Plutarch’s sources: he is simply retrojecting important aspects of the men’s later careers. Indeed, Wiseman’s ‘must have been like that’ principle so far works perfectly well. Plutarch is not fabricating. He is simply inferring what sort of youth it must have been who grew up into the man he knew; just as in our earlier examples he was inferring how a slave or innkeeper must have behaved, how a battle must have been fought, and what must have happened to Antony’s vomit. This is not fiction or invention. It is creative reconstruction.

This type of inference can be more far-reaching. Consider the beginning of the Antony, where Plutarch does not know much about Antony’s youth. What he does know, he seems to draw from Cicero’s hostile portrait in the Second Philippic. Cicero there described Antony’s early involvement with the young nobleman Curio. Plutarch reshapes it.⁶² Cicero had stressed the erotic aspects of the relationship, likening Antony to a male prostitute; Plutarch omits that. Perhaps that is restraint, perhaps scepticism: as we have seen, he knew what invective could get up to. Cicero had given no hint that Curio was the leading partner, and had represented Antony as just as depraved as Curio. Plutarch prefers to make Curio a corrupting influence, subtly tempting Antony into submissiveness. Now, in these early chapters of Antony we should not be too ready to assume that Plutarch is simply giving Antony the benefit of the doubt: he is not usually so generous, at least in this part of the Life. But he is concerned to develop the notion of the man as brilliant but passive, the susceptible victim of others’ wiles: first his wife Fulvia will render him ‘submissive’ (χειροήθης, 10.6, the same word as here), then of course Cleopatra and her flatterers. Plutarch is preparing the way here. That is both a literary and a historical point. Told this way, the story better prepares for the later developments; but it is also more plausible that Antony should have been behaving like this. His whole later life shows his passivity and susceptibility, so even in his youth he was very likely similar: ‘it must have been true’.

Still, a few chapters later we find Plutarch suppressing any mention of Antony campaigning with Caesar in Gaul, something he must surely have known: that allows him to pretend that it was Curio, once again, who brought Antony over to Caesar in 50 (5.2). Later in the same chapter he gives Antony a speech which his source (it seems) ascribed to Caesar himself (5.10).⁶³ In these cases one’s nose twitches a little more. Did Plutarch really believe that ‘it must have been’ like that, so clearly ‘must have been’ that Antony could not have been at Caesar’s side in Gaul, whatever his sources said, or that Caesar could not have delivered his speech? Hardly; but it again suited the characterization to make Antony the passive one, and he was prepared to bend the truth in a small way. Similarly, Plutarch seems to have rearranged Caesar’s early adventures in order to place his period of study in Rhodes just before his first rhetorical successes in Rome (Caes. 3–4).⁶⁴ It is an elegant arrangement, allowing Plutarch to group together all the foreign adventures before his narrative focus returns decisively to Rome; it would have been far more disruptive to have Caesar’s first steps in Rome, then move off to Greece, then back to Rome again. It is pleasing logically as well. Caesar learns his rhetoric, then comes back and applies it. But did Plutarch really think ‘it must have been true’, just because it is all so neat? Surely not. Here he was adapting the truth for literary purposes, and he knew it.

That is particularly clear when he gives contradictory accounts of the same events in different Lives. The debates of December 50 and early January 49 are described inconsistently in Pompey, Caesar, and Antony ; the Lupercalia incident of 44 is interpreted differently in Antony and in Caesar ; Antony’s behaviour at Philippi comes out differently in his own Life and in Brutus. Crassus (7.7) presents a summary of Roman politics over thirty years which represents Pompey as the steady representative of the established order and Caesar as head of the populares; compared with Pompey and Caesar, a ludicrous oversimplification. Clodius is in one Life an independent figure, in another a subservient follower of the triumvirs’ will. Plutarch’s view of the origins of the Roman civil war is subtly different in different Lives.⁶⁵ Yet all these Lives seem to have been prepared at the same time, and with the same material: he simply cannot have thought them all true.

Coriolanus is interesting here. As we have already seen, there seems an unusually large degree of retrojection here – the envy of his youthful rivals, and the way they made excuses to themselves. There is also considerable tidying of detail: it looks as if Coriolanus’ first military service has been transferred to the completely wrong battle, Lake Regillus, simply because the Dioscuri appeared at the battle of Regillus – all an appropriately miraculous and charged setting.⁶⁶ One suspects that such tidying was deliberate, though one cannot be certain: Plutarch might even have claimed that ‘it must have been true’, though possibly he would have found that a little strained. But what is most interesting is the treatment of Coriolanus’ relationship with his mother.⁶⁷ Marcius set himself new targets in courage and achievement:

Others do this in search of fame and glory; Marcius aim’ was his mother’s approval. Nothing could make him feel more honoured or happier than for her to hear him being praised, or see him being crowned, or to embrace him in tears of delight... Marcius felt he owed his mother the joy and gratitude which would normally fall to a father as well, and could never be satisfied with giving Volumnia pleasure or paying her honour. He even chose his wife according to his mother’s wishes and request, and he continued to live in the same house with her even after his wife had borne their children.(Cor. 4.5–7)

We shall look more closely at that passage in chapter 14 (pp. 309–10), and see how extensively Plutarch is here elaborating his source Dionysius. Matters such as the psychological reconstruction, ‘it was his mother’s will and choice that dictated the marriage’, seem to be his own inference; and this is a very different Volumnia, much less resistible and less limp than the one Dionysius had portrayed. Plutarch seems to be reading back from the final scene, where Coriolanus collapses before her pressure: what sort of man must he have been to react like this, what sort of relationship must widowed mother and orphaned son have had? What sort of person must Volumnia have been? Plutarch comes up with this, a psychologically deep and not unconvincing portrait. And could he have answered, ‘it must have been like that’? I think he could: it was his way of getting at the truth, to apply later actions and events to reach a portrait which made psychological sense.⁶⁸

So what do we conclude about Plutarch’s attitude to the truth? He does not always behave as we would, certainly; he tidies and improves, and in some cases he must have known he was being historically inaccurate. But the process has limits, and the untruthful tidying and improving is never very extensive. The big changes, the substantial improvements tend to come where he could genuinely claim – ‘yes, it must have been like that’. Nor is this one of the features which vary significantly among the Lives: the examples both of reconstruction and of untruthful improvement can be drawn as readily from Caesar as from Antony, from Themistocles as from Theseus. Even when he was at his most historically interested and alert, there was nothing in such rewriting to make him ashamed.

Yet it is not wholly satisfactory simply to conclude that Plutarch is usually reconstructing the truth, but sometimes consciously sacrificing it. It is so hard to believe that he thought he was doing anything totally different in the two cases: the techniques he employs in making Curio lead Antony into licentiousness (reconstructing truth), and then over to Caesar (sacrificing it), are simply too similar, and Plutarch surely did not regard them as belonging in different categories. It is better to start from the basic point that the process has limits: he will not bend the truth too far, and the big changes are indeed classifiable as ‘creative reconstruction’. At some times – when, for instance, he simplifies his whole political interpretation in Crassus, or when he transfers a speech in Antony – he seems to us to be going a long way; but he is never there falsifying things that are central to the particular Life, even if they were central to others. He is perhaps giving his subject a bigger role, but a role which helps to support his characterization of the man as a whole, Crassus as an oscillator between political extremes, Antony as submissive and passive: once again, he is only helping the truth along a little, allowing himself some licence to support a picture generally true. It is not, I think, that the concept of truth was itself different;⁶⁹ if it had been, such disquisitions as that in Aristides 1 would be hard to fathom. It is simply that the boundary between truth and falsehood was less important than that between acceptable and unacceptable fabrication, between things which were ‘true enough’ and things which were not. Acceptable rewriting will not mislead the reader seriously, indeed readers will grasp more of the important reality if they accept what Plutarch writes than if they do not. Truth matters; but it can sometimes be bent a little.

III. Plutarch and historiography

When defended in such theoretical terms, Plutarch’s habits admittedly seem to invite Dover’s disdain for ‘a pretentious kind of falsehood’ which, he thinks, ‘theGreeks had more sense than to call...“ideal truth” ‘;⁷⁰ just as Plutarch’s variations in critical rigour seem to imply an approach which a modern biographer would find bizarre. But perhaps Plutarch’s mode of thought is less alien to us than we readily admit. On a cool estimate historians and biographers still regard different sorts of rigour as appropriate for different sorts of material. Take childhood anecdotes, for instance: can we really believe that the infant Hugh Gaitskell startled a strange lady by singing from his pram, ‘Soon shall you and I be lying / Each within our narrow tomb’? Yet that is what the standard biography claims, without reserve. Or that young Florence Nightingale took a morbid pleasure in sewing back together the dolls which her healthy elder sister had torn apart? A family friend was most indignant at that suggestion! Or that Franklin Roosevelt had such demonic authority and confidence that he knew a bird would wait on a tree while he walked several hundred yards for a gun?⁷¹ The three biographies in question have different styles; but in each case this is not the sort of material which would have crept into the adult chapters. ‘Helping the truth along’, indeed.

And a whole new aspect of this discourse has now crept in. Modern writers have big ideas, grand interpretative themes, which they wish to present: so of course did the ancients. But the ancients presented these through careful and supple narrative technique, while moderns prefer to set out their ideas and arguments more directly, in passages of analysis rather than narrative.⁷² In such passages the licensed overstatement is extremely familiar – the sort of remark which emerges in examination papers with ‘discuss’ after it. Consider, for instance, the following passage of A.J.P. Taylor:⁷³

In his lazy fashion, Baldwin truly represented the decade. The ordinary Englishman, never attending church or chapel, probably without a Bible in his house, expecting his children to swallow unquestioningly a Christian education, was an exact parallel to the statesman who made speeches supporting the League of Nations and never thought of asking the chiefs of staff how the League could be supported. Façade became reality for a generation trained in cinema palaces. Churchill really thought that there was a glorious Indian empire still to be lost; Baldwin really imagined that he was defending democratic virtues; Left-wing socialists really anticipated a Fascist dictatorship in England. So the watchers in the cinemas really felt that life was going on among the shadows of the screen. Of course no one supposed that the tinny words would take on substance or that even the most menacing figures among the shadows could reach out and hit the audience on the head. That is what happened before the decade ended. The pretence turned out to be no pretence. Or perhaps it merely eclipsed the real life underneath.

It is very familiar in style, especially in first and last paragraphs of chapters (and that is a last paragraph). Nor is it bad writing, even if we feel there are too many ‘trulies’, ‘exacts’, and ‘reallies’. But some of those sentences are not true, at least not literally, really, or exactly true. Never attending church or chapel? An exact parallel? Never thought? And so on. Some, of course, we would unambiguously class as metaphors, such as the cinematic analogy; but the whole passage in fact has a metaphorical quality.⁷⁴ It is a different sort of ‘licence’ from that which a Plutarch, a Livy, or a Tacitus would employ, but it could well be categorized in the same terms. The author helps the truth along, the sentences are true enough, it would be obtuse to draw a boundary between those sentences which are just about true (‘Churchill really...’) and those which are just about not (‘The ordinary Englishman...’); the reader will understand more by accepting them all than by disbelief. ‘A pretentious kind of falsehood’, perhaps, but one we instinctively understand and applaud. This is good, stimulating, provocative history.

We are already comparing Plutarch with modern historians; what, finally, about the relationship with ancient historiography? It has not been possible here to carry through anything but a fitful comparison, but only a little has been found to support Momigliano’s contention. True, in the first section we found that some Lives were closer to historiography than others, and in those others the truth is pursued in a less rigorous way: some slightly less censored material can slip in. But even there Plutarch’s commitment to the truth was seen to be generally insistent, intelligent, and impressive. And when we consider the rewriting of material, it is clear that the same approach can be extended to the historians. Not, of course, that every historian would exercise his licence in the same way and with the same limits: doubtless Timaeus assumed different limits from Polybius, Antias from Sisenna, perhaps even Herodotus from Thucydides. On the whole Plutarch seems to belong with the more scrupulous group; and we can certainly see him operating in a similar way to the great historians who survive. With Livy or Tacitus, for instance, one will again find many examples which fit Wiseman’s ‘must have been true’ rubric – but some, usually less important, cases which do not. Almost at random, let us take a few instances from the first pages of Livy.

It is not hard to detect that Livy has rewritten the vulgate version of the Aeneas legend, exaggerating the speed of Rome’s early growth and suppressing the story of bad feeling between Lavinia and Ascanius.⁷⁵ ‘Must have been true’? Perhaps: Livy might indeed have assumed that Rome’s origins must have been in keeping with her later glories – though the concern for a rapid, nervous, and appropriately elevating introduction will also have played a part. But then a few chapters later one can detect further rewriting of the detail of Romulus’ campaigns, carefully allocating one domestic event to follow each invasion: Caenina invades, and a temple is founded to Jupiter Feretrius; Antemna invades, and the offer of citizenship is made; Crustumerium invades, and colonies are sent (1.10.1–11.4). Livy’s source-material was much more messy, it seems: two invasions, then all the domestic decisions, then a last invasion.⁷⁶ Did Livy really think his smoother sequence ‘must have been true’, because events in life are always smooth, never messy? Surely not: this was sacrificing the truth,⁷⁷ because in this case sequence hardly mattered: it was a ‘relatively harmless literary device’,⁷⁸ no more. A little later he seems to have rewritten the first steps of Tullus Hostilius’ reign to make him more ferox than his sources claimed: it is he, not Cluilius, who now stirs up the war with Alba.⁷⁹ ‘Must have been true’? Yes, in this case it must: Livy had a clear idea of Tullus’ character from later events, and could naturally assume that he must have been ferox here too.⁸⁰ But would Livy really have thought his techniques in the second case were different in kind from those in the first and third? Again, surely not. In small ways, his licence ran to improving the truth; but in general, the same techniques produced reconstruction rather than sacrifice.

More substantial and contentious instances may be drawn from Tacitus. Let us assume (though it cannot be quite demonstrated) that, in recounting the death of Augustus and accession of Tiberius, Tacitus borrows some details from events forty years later, when Claudius died and Nero succeeded: in particular, the detail of the barricades in the streets (Ann. 1.5.4) may well be imported from those later events (cf. Ann. 12.68.3). Tacitus points the parallel between the two sequences by a series of verbal repetitions.⁸¹ Possibly he ‘may intend to suggest that Tiberius’ accession was as questionable, disreputable, and indeed criminal as Nero’s’,⁸² but that is unlikely to be the most helpful approach: this is not simply a blackening device. It is surely clear that Tacitus is interested in the character of the principate itself, not just the principes, and is concerned to analyse it, not simply denigrate. He points the patterns of behaviour which are imposed on every new princeps alike: that is one reason why so many of the specific traits of Tiberius’ principate are foreshadowed in the retrospect of Augustus – the ruthless elimination of rivals, the personal grudges, the bloodiness, the luxus, the choice of a successor who is even worse (1.9–10). And now each accession necessarily proceeds on the same pattern. That is a point worth making for its own sake, and as important to Nero’s principate as to Tiberius’. (It is sad that we lack the opening of Gaius’ reign, and of Claudius’: very likely some aspects of this pattern would have been even more insistent.) If the narrative here is given extra details to make the parallels closer, could Tacitus have felt ‘it must have been true’? Yes, he could: if he concluded that events genuinely repeated themselves in the same pattern, that in itself was grounds for inferring what must have happened, whether or not his source chanced to mention it. They must have wheeled out the barricades, because that is what they always do. He was probably right. It is not quite the modern historian’s way, at least not with factual detail;⁸³ but it is not an absurd way.

Yet, once again, we find the same paradox as with Plutarch: Tacitus’ techniques here are so similar to those he employs elsewhere, in cases where the ‘must have been’ analysis seems inadequate. Let us take another instance which Woodman has brilliantly illuminated,⁸⁴ the description of Germanicus’ visit to the remains of Varus’ camp (Ann. 1.61–2), immediately followed by the German attack on Caecina (1.63–7). Woodman has shown that the accounts are very close to two passages in the earlier Histories, 2.70 and 5.14–15: much too close for coincidence. The events can hardly have been so similar to each other in fact, nor in Tacitus’ sources for the two periods. It is then highly unlikely that much of the material in Ann. 1 is drawn from his sources for AD 15: whether or not Woodman’s term of ‘self-imitation’ is the best way to describe it,⁸⁵ there must certainly be a high degree of imaginative free composition in the Annals passage; and very likely in the Histories passages too.⁸⁶ In this case Tacitus can hardly be intending to emphasize the parallels between AD 15 and events sixty years later: Woodman is surely right, and this would be pointless.⁸⁷ It is rather that Tacitus finds all these sequences independently worth dramatic elaboration, and in each case sets about it in a similarly enthusiastic way.

We can also understand why Tacitus thinks this scene worth elaborating. It is not just a question of rhetorical effect and entertainment, though those interests doubtless played a part: this is a most critical moment, with Caecina himself in danger of re-enacting Varus’ disaster. Well indeed may Caecina dream of Varus’ ghost beckoning him to follow,⁸⁸ or Arminius cry out that this is a second Varus delivered into his hands; and the description of Varus’ camp even seems to be finessed to make the parallel closer.⁸⁹ Indeed, it is not extravagant to think of the sequence as recreating a picture of Varus’ disaster itself, much as the early books of the Iliad subtly re-enact the earlier events of the war.⁹⁰ The effect is worth gaining, for Varus’ defeat is so important in the background to Germanicus’ campaign: this is the very lowly base from which his glorious successes of the next book will rise; and, just as suggestively, it conveys the dangers of the old-fashioned type of war which is Germanicus’ speciality, and so alien to Tiberius.⁹¹ But ‘must’ all Tacitus’ detail have been true, in his eyes? Surely not: once we have to postulate a third category, not quite ‘true’, not ‘false’, but ‘true enough’, that surely is where this case belongs. If pressed, Tacitus would have had to admit that this was imagination, not fact – but he would have been surprised to be pressed. It was a reasonable exercise of his licence to convey points of serious historical interest and insight. And it is again hard to believe that he would have thought he was doing anything very different here and in the case of Tiberius’ accession. There too he is manipulating and creating detail, in that case to bring two events into clearer connection with one another, here simply to make the moment more arresting and thought-provoking. As it happens, we would say that one was reconstructing, one was sacrificing, truth: our boundary falls between the two. The ancients’ limit fell a little further on, and in some cases quite a long way further on; but it does not mean that it was ahistorical. Tacitus had a historical point to make, and employed his licence to make it more clearly.

Two last points, both suggesting ways in which this analysis of Plutarch suggests morals for historiography. First, we have seen that Plutarch’s critical alertness is a variable rather than a constant: in less historical and more fluid areas he could be distinctly less rigorous than when writing of a Themistocles or a Caesar. Yet all these Lives belong in the same series. Herodotus’ logoi and Livy’s pentads and decades are also fused into a single work; possibly with, in some senses at least, a greater unity than Plutarch thought appropriate for his series – but there too it is dangerous to assume that their principles or practice were always on the same lines. In describing Egypt, doubtless, Herodotus goes a long way to tidy his material. When he writes of Helen, he surely does claim that Egyptian sources have told him a story which in fact owes more to his own creative art (2.112–20):⁹² it is a Greek story, clearly owed to the Odyssey and to Stesichorus (and just possibly Hesiod, though fr. 358 MW is poor evidence). Certainly, Greek travellers or immigrants could have taken the story with them generations before, and Herodotus could have heard it out there⁹³ (though it is more difficult to believe that it was told him by his notorious priests⁹⁴) – assuming, of course, that he went there at all.⁹⁵ But he could not have heard it like this – it is too close to the Greek literary forebears, too clean, too ungarbled and uncontaminated by disparate elements. The only real Egyptian feature is the bias: the hostility to Paris and Menelaus, who both abuse the laws of hospitality, and the contrast with the upright Egyptian king Proteus. But whatever Herodotus has done here and in similar cases, we should not necessarily infer that he proceeded similarly in later books, when he was dealing with more solid material. The difference between Helen and Xerxes was as great for him as the difference between Theseus and Caesar for Plutarch, or between Romulus and Pompey for Livy. Nor should we assume that Tacitus would naturally place his inventive limits in the same place with a romantic figure like Germanicus, in a remote and eerie forest, and with the grinding detail of senatorial business in Rome. There is more than a little of Plutarch’s Sertorius about Tacitus’ Germanicus, living in the same period but a different world from his harder-headed peers, and one which demanded description in a different style.⁹⁶ Even with Helen, too, we should not assume that Herodotus’ procedure was simply fictional and ahistorical. If he did hear a garbled, contaminated Egyptian version, would he necessarily think it improper to compare it with his Greek literary authorities, and ‘correct’ the Egyptian version in their light? The aspects of the Egyptian version confirmed by Greek sources would seem more credible, and more worth retaining. If he sensed a garbling, it would be tactful to remove it: ‘this is what they must have meant.. ? Again, not a wholly absurd procedure,⁹⁷ but one which could easily leave a version more Greek than Egyptian – except, of course, for that irreducible bias.

Finally, a more general reflection. It is correct and important to observe the frequency in the ancient historians of what we might call free composition: Woodman is absolutely right to stress the central nature of inventio in theory and practice. But it need not follow that the procedure has no limits. Plutarch, as we have seen, would fabricate more than we would, but there remains a stern divide between his creations and those of literary biography; there are still so many things which he would not invent. Reality can be bent, but not too far. We naturally place our limits and distinctions in a slightly different place; we too have our licences to improve, but exercise them rather differently. What is important is to try to discriminate exactly where the ancient limits were drawn, where historians would fabricate and where they would stop. There is still a gulf between Thucydides and pseudo-Callisthenes, between history and fiction. And that distinction, it seems to me, may be drawn because history will only invent and improve (a) within certain limits, however startling we may sometimes find them; (b) sometimes to overrule and reject certain well-attested ‘hardcore’ facts,⁹⁸ but more usually to supplement them and give them added clarity and vitality; and (c) ultimately with a historical purpose: that is, historians would hope to delight and divert their audience, but to do so by deepening their insight, helping them to understand events as they really happened and people as they really behaved. In this, Plutarch and the historians are at one.

Notes

¹Cf. e.g. Numa 1–2, Cam. 22.2 and especially Them. 27.2, ‘Thucydides’ version [that Themistocles met Xerxes’ son rather than Xerxes himself] seems to me to correspond better with the chronological data, though even these are in some considerable confusion.’ In that passage I am inclined to read oὐδ’ αὐτοίς ἀτρέμα σύντέταραγμένοις with Cobet, Flacelière, and Ziegler³: Plutarch’s usage of άτρέμά does not allow us to extract the required sense from σύνταττομένοις or σύντέταγμένοις, pace LSJ⁹, Frost, and Ziegler⁴.

²Heracleides Ponticus allowed Solon to live on well into Peisistratus’ reign (Sol. 32.3): admittedly, Heracleides himself may have been influenced by the desire to accommodate the Croesus-story, but this will not have been clear to Plutarch. Even the more moderate estimate of Phanias of Ephesus placed Solon’s death in 560/59 (ibid.). Croesus succeeded to the Lydian throne fourteen years before the fall of Sardis, i.e. c. 561/60 (Weissbach, R-E Suppl. v (1931) 457), and Plutarch will at least have known the synchronization with Peisistratus (Hdt. 1.59.1). If he thought about it, he could reasonably wonder if the error might lie in associating the meeting with Solon’s ten-year travels after his lawgiving (Hdt. 1.29.2–30.1, etc.): why could Solon not have made the trip in his notoriously sprightly old age (Sol. 31.7), as Diog. Laert. 1.50 seems to have inferred? After all, on one view he was in Cyprus when he died (Diog. Laert. 1.62, Val. Max. 5.3. ext. 3, Suda s.v. Σόλων, cf. Sol. 32.4), and some talked of voluntary exile under Peisistratus (Diog. Laert. 1.51–4, Suda, (Dio Prus.) 37.4, Gell. 17.21, cf. POxy. 4.664.9–10): he was not immobile – or so Plutarch might infer. Duff 1999, 312 therefore overstates when he talks of Plutarch knowing ‘clear chronological evidence to the contrary’ of the version he gives: the chronological indications tended one way, the other considerations (the ‘witnesses’ as well as the appropriateness) tended the other, and Plutarch could reasonably regard the issue as unstraightforward. We of course approach such questions rather differently, and regard all such data as totally unreliable, ‘representative’ rather than authentic: cf. esp. Lefkowitz 1981, 45. Such an approach is as dismissive as Plutarch’s attitude to the chronologists, and both have much good sense to commend them.

³Them. 27.2 is itself enough to show that (n. 1): cf. e.g. Arist. 5.9–10, Them. 2.5, Per. 27.4, with Hamilton 1969, xlvi–xlvii and Gomme in HCT i. 58 and n. 3 (less misleading than Hamilton’s criticism suggests); Frazier 1988b and 1996, 36–7; Duff 1999, 312–4.

⁴Cf. n. 2.

⁵Momigliano 1993, 56.

⁶Momigliano 1985, 87.

⁷See particularly Woodman 1988, esp. the Epilogue (197–215), and Kraus and Woodman 1997. Woodman’s 1988 book begins with criticism of Momigliano’s assumptions about historiography (as expressed elsewhere, Momigliano 1978); cf. also his p. 213 n. 17. Of other recent literature, Fox 1993, Moles 1993b, Wiseman 1979, 1981, and 1993, and, very differently, Brunt 1980, are especially thought-provoking.

⁸On the unlikely possibility of extensive Hellenistic political biography see Geiger 1985 and n. 28 below.

⁹Well discussed by Hamilton 1969, xlviii.

¹⁰A point fairly made by Gomme HCT i. 76, and Hamilton.

¹¹I return to Aristides 1 at pp. 269 and 367–8, and discuss the implications for Plutarch’s proemial and terminal self-characterization.

¹²Cf. esp. the thoughtful reflections in the Synkrisis (Cato Mai. 30(3)–31(4) ), and below, pp. 200–1, 225, 235–6 n. 124, and 275. As often, the first Life of a pair is the more straightforward, while the second presents a morally interesting variation: cf. ch. 16, esp. pp. 356–9.

¹³Hamilton 1969, xxi–xxii, is again good on this.

¹⁴See esp. Fogel and Elton 1983, 13–15, 21–2, 49–50, 90–5. I discuss this analogy, in particular its implications for ‘evidence’, more fully in Pelling 1997c, 213–24. – There is of course a blunter sense in which all historical writing is necessarily rhetorical: a point especially associated with Hayden White, e.g. 1973 and 1978: cf. n. 74 below. White gives disquietingly little attention to this gathering and criticism of evidence, as Momigliano 1981 fairly observes; but when Momigliano retorts that ‘rhetoric has long been for the historian an effective (never essential) device to be used with caution’, that too is surely a misunderstanding of rhetoric. Cf. Woodman 1988, 88 and 108 n. 72, and for a thorough narratological discussion of the ‘essential separation’ between historiography and fiction, even historical fiction, see Cohn 1999, esp. 18–37, 109–31, and 150–62 (quotation from p. 157).

¹⁵On this ‘work published after both Caesar and Crassus were dead’ see pp. 44 n. 172,47–8, 50, 67–8.

¹⁶On a literal interpretation, this principle would seem to be equivalent to discarding Phylarchus completely and simply following Polybius. But by looking at Plutarch’s practice one can see what he meant. For instance, at Phil. 5 and Ag.–Cl. 44–6 he describes the fall of Megalopolis in 223, and clearly draws on Phylarchus: that seems established by comparison of the tenor of his versions with Polybius’ criticisms at 2.61–2. Thus Plutarch’s generous attitude to Cleomenes and his stress on Megalopolis’ wealth both seem distinctively Phylarchan. But Plutarch knows Polybius’ account too (cf. e.g. Ag.–Cl. 46.5), and carefully avoids committing himself to the details which Polybius criticized. Thus Polybius (2.62) trenchantly attacked Phylarchus’ figure of 6000 talents for Cleomenes’ booty; Plutarch simply has a cautious ‘he acquired considerable wealth’ (χρημάτων εὐπορήσαντι, Phil. 5.5: cf. Ag.–Cl. 46.1). He apparently feels he can retain other details which Polybius did not criticize (or criticized so obliquely that the point could easily be missed: contrast Ag.–Cl. 45 and Plb. 2.55.8 (with Walbank’s note), the embassy of Lysandridas and Thearidas). When Polybius is being so captious, he could reasonably be regarded as ‘confirming’ Phylarchus’ other details if he did not explicitly attack them. Cf.

also Pelling 1997a, 108, 116–17.

¹⁷On this passage see now Schepens 2001, esp. 542–4, though I am not wholly convinced by Schepens’ wider conclusions about Theopompus’ portrayal of Lysander.

¹⁸Cf. e.g. Pomp. 72.4, Cim. 4.5, Gracch. 4.6, Mar. 25.6, Ant. 77.3, Aem. 15.5 with 16.3; Cf. the extensive reading in contemporary, often non-chronological material before writing Cicero (ch. 1, pp. 16–18. Further instances in Hamilton 1969, xlvii, who here seems to me right against Gomme HCT i. 58–9.

¹⁹Cf. then 3.3, 3.7–8 (inscriptions); 4.5–8. 8.3–4, 11.7 (comic poets, not very shrewdly exploited); 4.2 (dialogue of Pasiphon); 10.1, 11.10 (Theophrastus); 17.4 (Euripides); 23. 8 (Philochorus); 28.5 (Timaeus). For other, unattributed non-Thucydidean material, cf. Gomme, HCTi. 71–2. Not straining every ounce from Thucydides: cf. esp. 6.4, where he might have said more about both Mende (for instance mentioning Nicias’ wound, Thuc. 4.129.4) and Cythera (where the fifth column might interestingly have presaged events at Syracuse, Thuc. 4.53–4); note also the omission of the Melos campaign of Thuc. 3. 51, even though he goes on to make something of it in the comparative epilogue (Crass. 36(3).5: below, p. 361). See also Frazier 1996, 32–4, with good remarks on the proem and its commitment to finding out interesting new detail.

²⁰e.g. Ages. 19.10, ‘I discovered in the Laconian archives’ (Λακωνικαὶ ἀναγραφαί...) with Shipley 1997, 244 ad loc.; Arist. 5.10, 10.9–10, 19.7, Cim. 13.5, Sol. 11.1–2. See

p. 40 n. 120.

²¹Cf. esp. Caes. 8.4 with pp. 47–8, 50, 67–8, Alc. 32.2, Alex. 46.3, Cic. 49.4.

²²On Greek history, cf. Gomme HCT i. 59–61, 73–4, with pp. 134 and 141 n. 62 above; on Roman, ch. 9 below.

²³Gomme, HCT i. 76. But elsewhere ‘research’ is not an unfair description: we saw in ch. 1 that he undertook a quite extensive and systematic course of reading when preparing some Roman Lives.

²⁴Cf. Wiseman’s generalizations about historians’ critical research, in his chapter titled ‘Unhistorical thinking’: ‘The historians of Greece and Rome [except for Thucydides] did not “put their authorities to the question”. They did not have the questions to put, because they were incapable of the “historical imagination” needed for the historian to relive for himself, as Collingwood puts it, the states of mind into which he inquires... “Evidence” as the main preoccupation of the historian is a modern concept... for people brought up on the techniques of rhetoric, the first plausible story was good enough... [Livy and Dionysius] could assess the accuracy of what their sources told them only by the rhetorician’s criterion of inherent probability’ (Wiseman 1979, 42, 47, 48, 50). All these are arguably overstatements (cf. e.g. Rawson 1985, 217 n. 16): but scarcely extreme ones. In this company Plutarch can hold his head high.

²⁵See above all Lefkowitz 1981; Fairweather 1984. This distinction between literary and political biography is a rough one, and is clearly unsatisfactory in the formal terms in which it was articulated by Leo 1901: for a succinct statement of the reasons, cf. Momigliano 1993, 87–8, and see also below, n. 28 and p. 330 n. 6. But, however rough, the distinction can remain useful in illuminating such central points of difference as length, style of presentation, and focus of interest, as well as this issue of truthfulness (so Geiger 1985, 18–29): we shall see that in ch. 14 as well as in this chapter.

²⁶Momigliano 1993, ch. 4.

²⁷Cf. ch. 14, pp. 303–4; more on childhood at pp. 153–6.

²⁸Geiger 1985, 30–65 discusses whether there was much Hellenistic biography that might be called ‘political’, and argues that there was not: I still sympathize with that sceptical approach despite the powerful points made by Moles 1989. I would however now put matters a little differently, in particular emphasizing the difficulty of drawing strict generic boundaries: cf. my sketch in OCD³ s.v. ‘biography, Greek’. (Cf. Gentili and Cerri 1988, 61–8 and 82–5; Burridge 1992, Frazier 1996, esp. 9, and Duff 1999, 17, take a similar view.) If that approach is right, it makes the present points more striking still: despite biography’s close relationship to encomium and the Cyropaedia-type novel, Plutarch does not take over the more casual approach of those genres to truthfulness.

²⁹Cf. Geiger 1985, 9–29, 114–15.

³⁰At least, Diog. Laert. 6.84 specifically links the two, and what we know of the content supports the connection. Cf. esp. Strasburger, R-E xviii (1939) 464–5; Brown 1949; Hamilton 1969, lvi–lvii; Momigliano 1993, 82–3. Geiger 1985, 48–9 n. 43 is too

cautious: cf. Moles 1989, 232.

³¹Hamilton 1969, liii, lvi–lvii.

³²Cf. Hamilton 1969, xxiii–xxxiii, lvi–lvii.

³³Geiger 1985, esp. 117–20.

³⁴Geiger 1979, and 1985, 60–1, 120; ch. 1, p. 13.

³⁵In itself of course, a very fair point: cf. Pelling 2000, 128, and especially Gomme, HCT i. 60, 69–70. The way Plutarch uses such material is not quite ours, and in particular he tends to use quotations in an out-of-time way, sometimes using them to illustrate events years before the remarks were actually made: see the insightful remarks of Frazier 1988b.

³⁶Millar 1964, 54–5. Tony Woodman points out to me that historians, notably Velleius, regularly echo the phraseology of Cicero’s speeches in their narratives, both when describing the same events as Cicero and elsewhere. That is rather different from using Cicero as a historical ‘source’ in Plutarch’s manner; but it is a valuable reminder of the problems of definition involved in categorizing what is and what is not a ‘source’.

³⁷See p. 154; cf. pp. 17–18, 94–5, and Pelling 1988, 26–7.

³⁸Cf. Frazier 1996, 39. Nepos too made some use of such material: Geiger 1985,109–10.

³⁹I elaborate this point elsewhere, especially in chs. 4 and 9.

⁴⁰The Second Philippic is explicitly criticized at Ant. 6.1, and similar material is rejected at Ant. 2.2–3; but other passages, especially in Ant. 9–13 and 21, are largely based on the Philippic : see pp. 94–5 and Pelling 1988, 26–7, and nn. ad locc.

⁴¹Sertorius and the Isles of the Blessed, Sert. 8.2–9.2; Antony, Octavian, and dice, Ant. 33; Cassius and Faustus Sulla, Brut. 9.1–4; gods and Rome’s foundation, Rom. 8.9, with pp. 185–6 below. The Romulus passage has something in common with Dionysius of Halicarnassus’ approach to early Roman history, which is informed by ‘a desire to give an account which made sense in terms of what happened later in Rome’s history’, and envisages that coherence in moral terms: ‘in his conception of Rome’s development he blurs the distinctions between what is true and what is good, between the morally praiseworthy and an unbiased reading of the evidence’ (Fox 1993, quotations from 34 and 47). Plutarch’s remark can also be related to an assumption that important themes, especially those concerned with Rome, are associated with divine interventions (Swain 1989c, esp. 286–7) – but it is still not the sort of argument or material which he would readily develop in treating later periods. See also the discussion of truthfulness in Theseus in the next chapter.

⁴²I am grateful to David Lewis for reminding me of this passage. I elaborate this point in the next chapter, pp. 174–5.

⁴³Cf. Gomme, HCT i. 109–10 on 1.9.4. It is piquant to find Thucydides hailed as the grandfather of cliometrics on the basis of 1.10 (Fogel, in Fogel and Elton 1983, 69): not in fact the style of evidence or argument where Thucydides is at his surest.

⁴⁴Finley 1985, 12, of attempts to extract history from unlikely poetic sources.

⁴⁵Momigliano 1985, 87.

⁴⁶I return to this in ch. 7.

⁴⁷Momigliano 1985, 88.

⁴⁸855a–6d: Cf. esp. Homeyer 1967; Wardman 1974, 189–96.

⁴⁹The question was usefully asked by Theander 1951, 32–7: he found that Plutarch did largely practise what he preached, at least in his own narrative of those controversial events discussed in the body of On Herodotus’ Malice. Many of those are of course treated in Themistocles, one of the more historically alert Lives (‘had we only the Themistokles.. .we should have a good opinion of Plutarch’s learning and not a bad one of his judgement’, Gomme, HCT i. 61): thus Theander’s conclusion cannot be extended to other Lives without qualification, as we shall see. Wardman 1974, 191–6, also overstates the absence of ‘malice’ (i.e. the кακοήθεια of On Herodotus’ Malice) in the Lives. – For an interesting application of a similar approach to Dionysius of Halicarnassus, cf. Fox 1993. He argues that in treating early Roman history Dionysius does pursue the historical principles he expresses in the theoretical works, however strangely they may seem to us to blur the distinction between morality and truthfulness (above, n. 41). In many ways those principles resemble Plutarch’s, demanding (at least provisionally) an assumption of true nobility in the characters and events he describes.

⁵⁰e.g. Caes. 4.4–9, 5.8–9, 11.1: for Plutarch’s disapproval elsewhere, cf. pp. 105 and 114 n. 62.

⁵¹See pp. 5–6, 104–5 and 207–8.

⁵²Cf. esp. Otho 17.11, Pyrrh. 14; Wardman 1974, 53–5, 109–10, 217.

⁵³4.8, 5.8–9, 6.9, etc.: p. 104.

⁵⁴Lucull. 5.5, Cic. 3.3, 10.2, Pomp. 5.4–5, 81(1).2.

⁵⁵Mar. 29.4, 30.5–6, 31.2–5, 32.2, 33.2, 34.7, 41.6.

⁵⁶Mar. 43.7, 44.9–10.

⁵⁷In Pelling 2000, 107, I argue that Plutarch’s reconstruction here may rest on no more than a critical and imaginative reading of Thucydides’ text. If that is so, his assumption – we might say ‘fabrication’ – of so uncharitable a motive is even more striking.

⁵⁸Wiseman 1981, 389, followed by Woodman 1988, 93.

⁵⁹I expand this point in ch. 14, esp. pp. 308–10.

⁶⁰Ant. 13, 10.8–9, Caes. 9.2–10.11. Mar. 44, Ant. 9.6, 48.6, 49.2, 76.1–3, Caes. 19.10, Cor. 10.5. On the Mar. case cf. Carney 1960, 28–9; on Cor., Russell 1963, 25 = Scardigli 1995, 365; on Ant. and Caes., p. 94 and Pelling 1988, 33–6, together with notes ad locc.

⁶¹More on this in ch. 14, esp. pp. 307–15 (where I return to these ‘routine generalizations’).

⁶²Ant. 2.4–8, reshaping Cic. Phil. 2.44–8.

⁶³Cf. Pelling 1988, 127, 130 on both passages. Plutarch probably also transfers certain actions from Curio to Antony (above, pp. 107–8), but that is too controversial to serve as evidence here. Even if he did, he could reasonably infer that Antony and Curio would have followed a similar line (as many modern scholars have been happy to accept): ‘it must have been true’.

⁶⁴See pp. 76–7, 93.

⁶⁵See pp. 96–102, 107–8, 207–10; and Pelling 1988, 126–30, 144–5, and 171–3. There is a similar difference of interpretation between Phil. 5.3–4 and Ag.–Cl. 45.4–8:Pelling 1997a, 117–9.

⁶⁶Russell 1963, 23–4 = Scardigli 1995, 362–3.

⁶⁷I return to this passage, and discuss it from different angles, in chs. 14 (pp. 309–10)and 18 (pp. 395–6).

⁶⁸Just as Erikson explicitly started from later events to reconstruct the childhood of Martin Luther, in his case stressing an abnormal relationship with his father. ‘A clinician’s training permits, and in fact forces, him to recognize major trends even when the facts are not all available; at any point in a treatment he can and must be able to make meaningful predictions as to what will prove to have happened; and he must be able to sift even questionable sources in such a way that a coherent predictive hypothesis emerges’; the reconstruction of Luther’s youth should proceed similarly. (Erikson 1958, 50, cf. esp. 37.)

⁶⁹This is one point on which I disagree with Woodman (1988), though in practice the difference is partly semantic. Woodman argues for a different concept of truth, one more closely connected with impartiality (esp. 73–4, 82–3) and plausibility (87, 92–3); I prefer to think of a similar concept of truth, but one which was pursued and presented with different narrative conventions and licences. Woodman’s position is neater, but I prefer mine for reasons analogous to those given above. Like Plutarch (above, pp. 143–6), historians discuss the difficulties of recovering the truth; bias certainly figures among these to a greater degree than we might expect (Woodman 1988, 73–4), but historians knew that there were other problems too, especially those concerned with the nature of documentary evidence. It is hard to see why, for instance, the loss of records in the Gallic sack (Livy 6.1, cf. On the Fortune of the Romans 326a) or the confusion of the early fasti (Livy 2.21.4) or the secrecy of imperial records (Dio 53.19) should be a hindrance to recovering truth, if truth be interpreted in terms of impartiality and plausibility. Such passages defy interpretation unless truth meant something close to what we mean in our less theoretically agonized moments today, events as they really happened, history wie es eigentlich gewesen ist. (For similar points cf. Moles 1993b, esp. 114–18.) It does not follow, of course, that truth was pursued in the same way then as now, or with the same rigour in detail: though we too have our unrigorous licences, as I shall argue below. For similar talk of ‘licence’ cf. Morgan 1993, 187, arguing that some of the novel’s fictional licences originate in historiography.

⁷⁰Dover, HCT v. 396 n. 2, interestingly criticized by Woodman 1988, 197–215. It will be clear that I am in sympathy with most but not all of Woodman’s points.

⁷¹For the first two instances (from Williams 1979 and Strachey 1918), cf. ch. 14,pp. 317–21; for the last, Davis 1973, 83–4.

⁷²True, narrative history is drifting back into fashion, or rather respectability: it can never be out of fashion as long as people like stories, and care whether they are true. Cf. esp. Stone 1985. But a feature of the new narrative historians is that ‘analysis remains as essential to their methodology as description, so that their books tend to switch, a little awkwardly, from one mode to the other’ (Stone 1985, 91, cf. 75). Witness Stone himself. His next chapter begins ‘one of the more striking features of Christianity has been its perennial tendency to fission... There are two ways of looking at this crisis of European civilization...’, and so on. That is very different from the ancient style. Or consider Schama’s acclaimed history of the French Revolution (1989). This makes great play with its return to a narrative strategy, justifying it (esp. xvi, 6) on the interesting grounds that the agents themselves narrativized their own actions and related them to preceding narratives (especially those, as it happens, of Plutarch’s Lives: Schama 1989, 32, 169, 171, cf. below, p. 238). Its blurb advertises it as ‘a return to the magnificent tradition of the epic narrative’ – and not unfairly, for Schama indeed narrates with brilliant panache, with a particular gift for the biographical vignette. But Schama too finds room for analysis (e.g. 62 on the ‘politicization of the money crisis’), and much of the narrative brilliance lies in his use of survey (notably 123–31 on aristocratic ballooning) to give an arrestingly unexpected insight. Hardly the ‘tradition of the epic narrative’, though none the worse for that.

⁷³Taylor 1965, 319–20.

⁷⁴One is again reminded of Hayden White’s emphasis on ‘metaphor’ as one of his four modes of discourse, the others being metonymy, synecdoche, and irony: cf. White 1973 and 1978, esp. 72–4, 252–3. But White, as usual with his borrowings of figures of speech, means something rather different, and indeed rather peculiar: for instance, he strangely associates ‘metaphor’ with the sixteenth and seventeenth centuries. Others evidently find this schematism more illuminating than I do.

⁷⁵Cf. esp. 1.3.4; the battle between Ascanius and Mezentius after Aeneas’ death was famous (Cato, Origines, frs. 9–11 P = 9–11 Ch.), and Livy’s claim that Mezentius and the Etruscans were acquiescent on Aeneas’ death is especially startling. 1.3.1 is similarly very bland, in view of Cato’s story of Lavinia’s flight from Ascanius to the woods (fr. 11 P, cf. Dion. Hal. AR 1.70.2–3).

⁷⁶The fuller version of Dion. Hal. AR 2.32–7 probably gives a fair idea of what Livy found in the source or sources they (surely) share. All such speculations are admittedly suspect; but at least one can here understand why Livy should have smoothed and abbreviated a version similar to that of Dionysius, but not why Dionysius should have confused and roughened a version with Livy’s articulation but more detail. Cf. Burck 1934, 143, and on the methodological principle 5–6.

⁷⁷Presuming, of course, that Livy thought his source’s version true. But Tony Woodman may well be right when he suggests to me that Livy accepted his source’s version on the same terms as he expected readers to accept his own: that is, he assumed that predecessors had smoothed events so thoroughly that their sequence or detail did not deserve any particular respect. That interpretation does not fundamentally affect the point I am making here. Livy would still be showing the same respect or disrespect for accurate sequence: on Woodman’s view, he would simply be taking for granted the same assumptions in others.

⁷⁸Cornell 1986, 73, who uses the phrase of small-scale fictitious inventions, acceptable enough (he thinks) provided they did not do violence to the traditional facts. In this case, they did not do much violence, and that was enough. Cf. Wiseman 1993, 134, 142.

⁷⁹Dion. Hal. AR 3.2–5 again seems to come from the same origin. Note especially the contrast between Dion. Hal. 3.5.1 and Livy 1.23.4 on Cluilius’ death: for Dionysius it is genuinely a divine visitation, ‘because he stirred up a war between mother-city and colony that was neither just nor necessary’; for Livy Tullus claims that this was divine intervention, but that is simply a piece of his fierce bluster. Livy’s portrayal fits his individual portrayal of Tullus so closely that we are again surely dealing with his own rewriting.

⁸⁰On Tullus’ ferocia, Ogilvie 1965, 105–6, though this seems less traditional a characteristic than Ogilvie implies. Indeed, it is hard to trace much colour at all in Tullus’ characterization before Livy: there is little in Dion. Hal., and Cic. De rep. 2.31–2 does not suggest ferocia; not much can be extracted from Piso frs. 10, 13 P. = 17, 20 F. (discussed by Forsythe 1994, 195–201, 216–20).

⁸¹On all this cf. Martin 1955; Goodyear 1972, 125–9 and bibliography there cited. Add now esp. Martin 1981, 162, and Woodman 1998, 23–39, esp. 35.

⁸²Goodyear 1972, 126; cf. e.g. Woodman 1979, 154.

⁸³With mental detail, our conventions are of course different: what ‘Livia’ (or Tiberius, or Caesar, or Pericles) ‘must have intended’ is often inferred on similarly slender grounds, and ones which postulate no better and no worse a model of continuity in human experience.

⁸⁴Woodman 1979, and 1988, 168–79. Much ofWoodman’s analysis is rightly followed by Goodyear 1981.

⁸⁵West and Woodman themselves suggest an alternative explanation in their Epilogue (1979, 195–6).

⁸⁶‘Much of what we are told at 64–5 may have happened not in AD 15, but in AD 70’ (Goodyear 1981, 108): an over-sanguine view. It is more likely that it did not happen at all: if Tacitus could ‘invent’ the one, he could invent the other as well.

⁸⁷Or so I wrote in 1990. I am now less sure: it may be that the parallel is suggestive for a future audience who might read Annals and Histories consecutively, and the Annals passages are crafted as prequels for the (already written) Histories sequence. The point might be that the dangers of barbarians capitalizing on Roman dissension are real in 15 AD,but for the moment averted (a point which could be related to other Germanicus-themes which I have explored in Pelling 1993 and 1997b); as dissension worsens, (a) the sites of past disaster are now from civil wars – Cremona, Hist. 2.70 – rather than foreign, and (b) the German-Roman encounters are even more perilous than in 15 AD.But even if some such analysis might hold, it would be hard for Tacitus to think that it ‘must have been true’, that events must have run so similar a course simply because the parallel might be thematically suggestive. The main run of the argument here would still hold.

⁸⁸For more about this dream, together with Germanicus’ further dream at 2.14, see Pelling 1997b, 206–9.

⁸⁹1.61.2 seems to suggest two camps for Varus (cf. Koestermann 1956, 443 n. 32 and Goodyear 1981, 95–6 ad loc.): not quite what we would have inferred from the description of Varus’ march at Cassius Dio 56.19–21, but suggestively close to Caecina’s efforts at 1.63.4–64.3 and 1.65.7. Hence read prima with MSS at 1.61.2, not Koehler’s primo (approved by Woodman 1979, 232 n. 2).

⁹⁰Virgilian influence is of course important too: ‘with this identification of Caecina with Varus, the present and past merge into one, as they do so often in the Aeneid...’,Woodman 1988, 174, cf. 169.

⁹¹More on this theme in Pelling 1993.

⁹²Fehling 1989, 59–65, to which this whole paragraph is heavily indebted, and which seems to me to survive the virulent critique of Pritchett 1993, 63–71; but I take a less sceptical view than Fehling of Herodotus’ commitment to the truth, even if his ‘licences’ sometimes startle us.

⁹³Cf. Henige 1982, 81 ff.: a particular danger for modern oral historians is that ‘a great deal of testimony obtained from informants is really feedback; that is, it originated as information that entered the society and was absorbed into its traditions because it proved useful or entertaining’. Armayor 1978, 65, here seems to me too sceptical. Lienhardt 1985, 147, reports a case in the 1870s of simplified Christian eschatological doctrine being fed back to a missionary-traveller, who was less surprised than he ought to have been: a suggestive parallel.

⁹⁴On these problematic priests cf. Fehling 1989, 71–7; Armayor 1978; West 1985, 298 and n. 97. On the other side, Lloyd 1975, 89–114; Pritchett 1993, 75–85.

⁹⁵Not, of course, that extensive use of literary sources need in itself suggest lack of autopsy: such arguments could easily be made to suggest that Caesar never went to Germany. Nor can I resist quoting Nigel Williams, who mentions ‘a Scottish writer who has recently completed a book about Peru. “My most important source”, he told me, “was not my experience of the place but The Ladybird Book of Peru. ” ’ (N. Williams 1993, 95).

⁹⁶On Sertorius, above, p. 149; on Germanicus’ world, Pelling 1993.

⁹⁷Nor very different, for instance, from the procedure followed by a biographer of a living person today, if he found after an interview that his subject’s own reminiscences were contradicted in unimportant detail by a documentary record. In such cases, a certain amount of discreet tidying and cleaning up is part of the job. Cf. Henige 1982, 66–73, on the dangers of imposing one’s own acculturated interpretations on material gathered in the field, and in particular the difficulties of integrating oral and written material. West 1985, 304–5, has some very valuable remarks on the relevance of such contemporary experience for our view of Herodotus. – This paragraph and the first part of this note, written in 1990, may owe more than I then realized to personal childhood experience. Perhaps I may fill that in here. My father was a sports journalist, who frequently interviewed not very articulate informants. He would regularly frame questions in the form ‘would you say, Billy, that the cross came over exactly where you wanted it, but it was still tricky to get the deflection...’: this then appeared in print as ‘Billy L. said, “The cross.” ‘, to everyone’s satisfaction. I am less clear than West 1991, 148 and n. 28, that such ‘leading questions’ make a caricature of the information-gathering procedure: this may be a case where the legal analogy is misleading (cf. above, p. 145 and n. 14). My father would also, I think, have felt free to correct misremembered details, if (say) full-back Dewi J. said that he had scored twenty-five points the previous season against Ebbw Vale when it was in fact against Cross Keys. Dewi ‘would have meant’ Cross Keys, and the informant would certainly not have resented the tacit correction. I add this titbit because journalistic parallels have been exploited so fruitfully by Wiseman (1993, 139–40) and especially Woodman (1998, 1–20 and elsewhere). Doubtless there is also a case for picturing ‘Herodotus the journalist’ to go along Redfield’s ‘Herodotus the tourist’ (Redfield 1985). But Woodman’s points have largely been concerned with sensationalism and fictionality: it is worth remembering those aspects of a journalist’s technique which do have a concern for truth, even if they would not pass muster in a court-room.

⁹⁸I draw the phrase ‘hard-core facts’ from Woodman 1988, esp. 88–94, with whom I here wholly agree. I am less certain that we are at one on the first criterion, where he would assume different and more generous limits (I cannot for instance follow him in his treatment of Thucydides’ plague, 32–40); or on the third, where his emphasis suggests a different balance between diversion and historical analysis. But his Preface (pp. xi–xii) does enter a caveat, and he readily accepts that diversion and ‘deepening insight’ often go hand in hand.