3

The Apophthegmata Regum et Imperatorum and Plutarch’s Roman Lives

‘Plutarch seems to have kept some “commonplace book” in his philosophical studies (Mor. 464f, cf. 457d), but that tells us little of his methods in the Lives.’ That casual footnote was written over twenty years ago,¹ and reflected the little thought I had then given to the problem. Since then Luc van der Stockt and his Leuven colleagues, especially Birgit van Meirvenne, have greatly deepened our understanding of those ὑπομνήματα in the Moralia.² In this chapter I hope to be a little less casual.

In chapters 1 and 2 I was sceptical about the possibility that Plutarch took many notes as he read. I preferred to think that he had a single source open in front of him as he composed, and would supplement this, sometimes extensively, from his memory of earlier, wider reading. When we can compare his writing either directly with his source-material or (more often) with other writers who depend on those sources, we can often see that he owes most of his material and his narrative articulation to one source at a time. If he had taken full notes, he would more naturally and more frequently weave material together into an independent pastiche, owing no more to any one previous version than to any other.³

In the case of the late Republican Lives, there is often good reason to think that this one main source was the ‘Polliosource’ which he shares with Appian – either Pollio himself, or a source following Pollio closely.⁴ The overlap with Appian is often very close, so close that we have to assume that both authors are following that source’s articulation and language fairly faithfully, though of course both impose their own emphases and interpretative tweaks. Given that closeness, I was sceptical of the notion that Plutarch would have taken extensive notes on the Pollio-source itself.⁵ That possibility cannot be totally excluded, for it is a tricky business to specify what note-taking methods make sense and what do not. Many scholars have known students who solemnly regard ‘note-taking’ as a matter of copying out long chunks of Syme or Badian or Millar word for word. Many of us have done it ourselves – and, if we had research assistants as Plutarch may have done,⁶ we might do it, or get it done, a little more. But it is still hard to believe that someone as productive and experienced as Plutarch would find it a rational method. It would be so much easier to have his Pollio-source open before his eyes as he composed, and supplement it from memory.

Notes, I suggested in chapters 1 and 2 come in at a later stage. Some sort of ὑπóμνημα, what Lucian calls ‘a draft, a body of material which is still inartistic and uncoordinated’, seems to have been a regular stage of composition as a penultimate draft; then, as Lucian goes on to explain, in the final version the writer should ‘impose structure, give the work its beauty, use diction to add colour, and give form and rhythm’ (How to write History 47–8). It is this ὑπóμνημα which Plutarch would have been writing or (more likely) dictating as he held the Pollio-source open before him. At that stage he would already have given thought to the narrative strategy of the Life and the pair, and so the draft would presumably already reflect the interpretative emphases on which he had decided; but the last stage of stylistic beautification would still be in the future. Doubtless this model is too schematic. An author with Plutarch’s stylistic sense could hardly avoid some ‘beautification’ every time he put pen to paper, or dictated a word; and, given the speed with which he must have produced, it would be odd if some of his organizational or thematic ideas were not last-minute, involving him in larger-scale rethinking even in his final draft.⁷ But, as a simplified model, it is still one which I find attractive.

It is also close to the picture of a ὑπóμνημα which van der Stockt develops for the Moralia: ‘a more or less elaborate train of thought, involving material previously gathered and certainly written in full syntactical sentences... On the other hand, the ὑπóμνημα does not yet display literary finish (cf. Plutarch’s apology for the lack of καλλιγραφία in 464f–5a). In short, it probably took the form of a rough draft (cf. Pelling 1979, 94–5 (= ch.1 of this volume, pp. 23–4), on historical hypomnemata) not of an “Endfassung, die Reinschrift” (Dorandi 1997, 9).’⁸ Van der Stockt argued this on rather different grounds from the ones I used myself, and if either of us is right then it may give extra support to the other. However, van der Stockt’s arguments also confirm that I was indeed too casual in suggesting that those ‘commonplace books’ told us nothing about the production of the Lives. On the contrary, they may tell us a great deal.

Thanks to van der Stockt and van Meirvenne, we can now see more clearly what sort of ‘notes’ or ‘commonplace book’ this must have been: not merely compilations from his reading, as scholars used to believe, but trains of thought that occurred to him at any time, what he thought about as he walked or talked or bathed. We can see how the same reflections tend to be reused, with sentences or anecdotes or ideas slightly re-sorted, in recurrent ‘clusters’. We can see too the skill with which each cluster rearranges to produce an argumentative thrust which suits its context – and sometimes also, even more revealingly for his methods, some times when the jumble does not quite fit, when a train of thought which made excellent sense in one context, and presumably in the draft itself, produces inconsequentialities or loose ends when transposed into its new context. A thought which is marginal in train-of-thought A can become the whole point of the exercise in train-of-thought B; but when it does some of the other aspects of train-of-thought A may also find their way into B even though they are no longer very relevant. Van der Stockt 1999a has a splendid example from How to Tell a Flatterer from a Friend, when the topic of ‘silent flattery’ leads on, very strangely, to two anecdotes which do not illustrate this topic at all: one of Apelles the painter who rebuked the rich Megabyzus for talking about line-drawing and chiaroscuro, things of which he knew nothing; the second of Solon’s frank speaking in front of Croesus (58b–59a). What those stories really illustrate is bold frankness, παρρησία: that is only marginally relevant to the context in How to Tell a Flatterer, but the point is more apposite in the other context where Plutarch is deploying the same cluster of ideas, in Tranquillity of Mind (471d–2b). The Megabyzus anecdote there recurs in the course of an argument that we should all adjust our hopes to our capacities, and it could not be more pertinent. Van der Stockt reasonably concludes that both passages are based on a ὑπóμνημα where the argument was similar to that in Tranquillity of Mind, and in How to Tell a Flatterer Plutarch has not properly smoothed the cluster into a new argumentative context.⁹

Here again we may be able to find something similar in the Lives. In chapter 2 I tried to identify several items in Plutarch’s later accounts of the Catiline conspiracy, those in Crassus, Caesar, and Cato Minor, which Plutarch would have known when writing Cicero but omitted from the final version of that Life. One of these is the story of the young men who formed Cicero’s bodyguard and attacked Caesar as he left the senate (Caes. 8.3–4): that item is also mentioned by Suetonius (Div. Iul. 14.2) in a section which is otherwise very close to Plutarch’s Cicero and surely derives from the same source. Even at the time of Cicero, too, Plutarch knew a source which stressed Cato’s statesmanship in the aftermath of the conspiracy (Cic. 23.5–6). It would be strange if that source had omitted Cato’s corn-dole; but that is mentioned in Caesar (8.6–7) and in Cato Minor (26.1), not in Cicero itself. And it is likely that Plutarch also knew of the senate-meeting a few weeks after the execution, when Caesar delivered a speech in self-defence. This was not very relevant for Cicero and he omitted it, but he used it in the later Caesar (8.5).¹⁰

There are several ways in which we might explain Plutarch’s method here. One might be that he simply remembered these instances from his earlier reading when he came to compose the later Lives, and silently integrated the material with whatever source was then before his eyes.¹¹ But in chapter 2 I also wondered whether Plutarch included those items in the ὑπóμνημα he drew up for Cicero, and went back again to look at that ὑπóμνημα when he prepared the later Lives. There might be an indication of this, I suggested, in that story of Caesar being threatened by Cicero’s young bodyguards. The account in Caesar goes on:

They say that Curio threw his toga around him and bundled him away, and Cicero too shook his head when the young men looked towards him: perhaps he was afraid of the people, perhaps he simply thought that such a murder was wholly illegal and unjust. But, if this story is true, I cannot understand why Cicero did not mention it in his account of his consulship [i.e. the περί ὑπατείας]. He was certainly accused later of missing an ideal opportunity of foiling Caesar because of his terror of the ordinary people. (Caes. 8.3–4)

This train of thought is hard to follow. The omission of the incident from the περί ὑπατείας is not hard to explain at all. Indeed, we should expect Cicero to have omitted it, precisely for the reasons which Plutarch here goes on to give: Cicero, if he was criticized later for letting Caesar off the hook, would hardly wish to remind contemporaries of the incident.¹²

This looks very much like the sort of inconsequentiality which van der Stockt has identified in Moralia: Plutarch might have noted at a drafting stage that the incident was omitted from περί ὑπατείας, then incorporated that reflection into his final draft without thinking through the implications. If so, the drafting stage was surely that of the Cicero ὑπóμνημα, a time when there is good reason to think he had just read the περί ὑπατείας.¹³ Then, at the time of Caesar, his thoughts were more on Caesar as the great man of the people and on the opponents who lost their chance: that would be the time when he added the further thought that Cicero ‘was certainly accused later of missing an ideal opportunity of foiling Caesar because of his terror of the ordinary people’.

So, as in the van der Stockt case, we must accept the inconsequentiality and note that even the sage of Chaeronea can occasionally nod, can show some human frailty. We can understand the nod and the frailty a little better if we accept that he worked in the way that van der Stockt, van Meirvenne, and I all propose.

So my reconstruction coincides quite closely with that of van der Stockt and van Meirvenne. What it does not coincide with, however, is an alternative picture which Philip Stadter develops in a fascinating paper.¹⁴ Stadter goes back to the Apophthegmata, in particular the Apophthegmata Laconica, and argues that those collections allow us a glimpse of Plutarch in action. For him the Apophthegmata Laconica represent part of the note-taking for the Lives, a ὑπóμνημα or set of ὑπομνήματα initially used for the final writing-up of the Lives, then separately gathered into a self-standing collection. His picture for the Apophthegmata Regum et Imperatorum is a little different, for he there thinks that our version represents a selection from an earlier, more extensive anecdotal collection, one which would have been similar to the Laconica.¹⁵ Stadter suggests that such a collection may have followed the same lines as we can now trace for Philodemus at Herculaneum. First Plutarch would make a collection of Lesefrüchte, presumably garnered as he read and therefore ordered in the sequence of reading; then they could be snipped and re-ordered in the sequence of the Life itself, which would explain why many of the Apophthegmata orderings are the same as those of the corresponding Life.¹⁶ If so, then these snippets are unmistakably closer to ‘notes’ than anything of which we have yet found traces.

Stadter is absolutely right to point to the importance of the Apophthegmata, and, although I shall suggest a different picture of their relevance, there are two important points of common ground.

First, there must indeed be some relationship between the Apophthegmata and the Lives. There is the point of shared ordering which he stresses. Admittedly, this is more visible in the Apophthegmata Laconica than in the Apophthegmata Regum et Imperatorum, where there are many individual variations: the Alexander and Agesilaus collections in the Apophthegmata Regum et Imperatorum show a very different ordering from the Lives, while the Roman ones are much closer; but in any case there are still enough similarities to demonstrate contact. There is also the extreme closeness in wording between the Apophthegmata and the Lives, often in cases where Plutarch’s own characteristic vocabulary-choices or stylistic patterns are in point. We do need some explanation, and that explanation must be a systematic one: we cannot for instance assume that the closeness comes because the Lives and Apophthegmata happen in virtually every case to have the same sources (what a coincidence that would be!), and that the slight deviations are the few exceptions where the sources happen to be different. Even the deviations, as we shall see, come in cases where the other features of an anecdote show the same closeness between Life and Apophthegmata as we generally find. We must look for a general explanation which is powerful enough to explain both the usual closeness and the rare deviations.

Secondly, Stadter does not claim that such a collection would have been the only form of note-taking or preparation for the Life. Even in the cases of the stories themselves, the Lives frequently have more detail. In some cases we could doubtless explain this by Plutarch’s own engaging habit of ‘helping a story along a little’, introducing pieces of circumstantial detail which are owed to his own creative imagination rather than to his sources.¹⁷ There are still too many cases where the additional facts in the Life overlap with detail found in other, non-Plutarchan accounts of the same events. If Stadter’s overall picture is right, then perhaps Plutarch found himself remembering these extra details as he came to write up the Lives; perhaps he went back to the sources again, inspired by the aide-mémoire of these ὑπóμνημα-like notes; or perhaps he had a second, parallel, narrative ὑπóμνημα, one which might include more details of these episodes as well as a full-scale narrative skeleton for the Life. And there are many whole Lives that have no counterpart collections, including some figures who were not bad at the one-line jest – Cato the younger, Gaius Gracchus, Brutus.

So Stadter does not claim that this is more than a part of the Life’s preparation; still, if he is right it is an important part, and it would indeed make it more likely that notes of other sorts played a part in that preparation. For there are many tracts of the Lives which have no such apophthegma-rich episodes, and therefore have no counterparts in the Apophthegmata collection at all; yet, on a simple view, it would be surely be those non-apophthegmatic sequences which would require notes more, not less. The whole point about an apophthegm is that it is memorable, that one can remember it anyway. The intricacies, say, of Lucullus’ or Pompey’s eastern campaigns are cases where even the most retentive of readers might need something to jog the mind. If it was worth noting down apophthegmata to aider his later mémoire, then it was going to be more worthwhile still to jot down other things as well.

So if Stadter is right, it is most important. But is he right? This procedure has a counter-intuitive ring to it, especially if there are going to be other notes as well: would it really help composition to keep these two separate files? It is true that we should avoid being too aprioristic about this; what seems to me a ‘natural’ way of working may not seem natural to the scholar working in the next office, still less to anyone in a different technological and mental world. But it is worthwhile to play with an alternative reconstruction, which would still leave the Apophthegmata related to the Lives. I suggest that they are subsequent to the Lives, not part of their preparation: a collection based on Plutarch’s work for the Lives, but garnered from those Lives or the work for them, not for them. This is indeed the procedure which Plutarch himself hints at, if the dedicatory epistle to Trajan is genuine: there he says that ‘you’ already have the Lives, the ‘composition’ (σύνταγμα) ‘on the most distinguished leaders, law-givers, and rulers of the Greeks and Romans’ (172c). That epistle has often been taken as spurious, but Mark Beck has recently made a strong case for its authenticity.¹⁸ In any case, I shall argue that comparison of the works themselves suggests that the Apophthegmata are the later of the two.

This may seem to revert to a view which has long been discredited, for it was once thought that the Apophthegmata collection could have been made from the Lives, perhaps by Plutarch or more likely by someone after his death; and, if the second of those views was right, that this could help to explain what was seen as their low stylistic level. Volkmann seemed to have exploded that view:¹⁹ there are too many anecdotes in the Apophthegmata which do not figure in the Lives, and also vice versa; there is sometimes divergence of details between the two works; and yet some of those details in the Apophthegmata seem to be authentic, even though any collector could not have got them from the Lives.²⁰ There is also a difficulty in the perceived mediocrity of style: why should a collector so systematically have sacrificed the narrative richness of the Lives?²¹ But those difficulties can be met if we suppose that the collector used not, or not only, the finished Lives, but the ὑπομνήματα which Plutarch made for those Lives. We have already seen that such drafts may have included material which was not included in the final version of the Lives themselves.²²

Let us take a few test-cases. The first few examples will show some difficulties in the Stadter reconstruction – or at least show that, if the Apophthegmata did serve as ‘notes’ for a Life, the final version in that Life can be based on them only to a limited degree. They will also show how difficult it is to believe the converse thesis, that the Apophthegmata versions are simply ‘based on’ the Lives. That suggests that we should look for another explanation, presumably that both Apophthegmata and Lives are based on something else. We need a ‘systematic explanation’, as we have already seen, and that ‘something else’ cannot be defined as a series of shared sources which both versions happen to have in common; that would be too much of a coincidence, and anyway Plutarch would hardly follow anyone else’s words so closely and so regularly as we would have to assume. It is better to think of this ‘something else’ as some large-scale gathering of material by Plutarch himself, and that presumably points to some sort of preparation or note-taking. The next few examples will explore how detailed these ‘notes’ would be, and how close to the final versions which appear in the Lives. The answer will be ‘pretty close’: that fits the picture of a ὑπóμνημα, a penultimate draft, on the lines I sketched in chapter 1 and van der Stockt and van Meirvenne posit for the Moralia. The final cases will bring out some further features of the Apophthegmata collection itself, suggesting that the stories are more thoroughly recast and carefully ordered than is usually acknowledged.

Case 1: Lucullus apophthegm 1 = Lucullus 27.8–9. The Romans are about to fight Tigranes on a black-letter day, October 6th, the anniversary of the disaster at Arausio. Lucullus replies that ‘I will make this day too a lucky one for Rome’.

The apophthegm begins with some details of the forces on both sides, 150,000 men for Tigranes, 10,000 hoplites and 1,000 cavalry for Lucullus. Those figures have equivalents in the Life, but there they are more detailed, and not given in the immediate vicinity of the apophthegm itself. A chapter earlier Tigranes’ numbers were given as 20,000 archers and slingers, 55,000 horsemen including 17,000 cataphracts, 150,000 hoplites (so the same number as in the apophthegm), and 35,000 labourers (26.7–8). Lucullus then has ‘24 cohorts, comprising no more than 10,000 legionaries, and all the cavalry and about 1,000 slingers and archers’ (27.2).

This is not a case where Plutarch might be imaginatively expanding his notes when he comes to write up the Life. The numbers are too detailed for that, and too integrated into the surrounding narrative; and anyway it looks as if Plutarch draws these numbers from a source, as those given by App. Mith. 85.328 for Tigranes are not too different (250,000 infantry and 50,000 cavalry: Appian may well be rounding).²³ Nor does it seem likely that Plutarch added them from his memory, any more than he could have written up the rest of the battle from memory: once again, there is far too much detail. It is more likely either that he went back to the original source when composing Lucullus, or – if Stadter is right – that the numbers formed part of the second, parallel narrative ὑπóμνημα which he postulates.

In that case, however, (a) it seems hard to see why Plutarch should have added these numbers to the apophthegm-ὑπóμνημα as well, at least if he was still envisaging that ὑπóμνημα primarily as ‘preparatory notes’ for the Life. He would not even be using the numbers in the same context in the final version; he knew he would have to check back elsewhere for the intricate details of the engagement itself; and it would be more sensible to rely on finding those numbers again in his source, or in the narrative-ὑπóμνημα, rather than duplicating them here. And (b) it raises the question how important the apophthegm-ὑπóμνημα would be anyway as part of the preparatory process. When Plutarch came to write up the Life, all the details of the engagement would have to come from elsewhere; and the only element captured in that apophthegm-ὑπóμνημα would be the apophthegm itself, precisely the item which was most memorable and for which an aide-mémoire would be least necessary.

If we take the alternative view – that the Apophthegmata were based on a (single) draft for the Life – then there is no problem. The collector, Plutarch or another, could readily extract the important figures from the preceding story before going on to give the punch-line.

Case 2: Pompey apophthegm 3 = Pompey 10.11–13. Pompey was in a position to execute all the inhabitants of a Sicilian town. The ‘demagogue’ Sthenius told him that it was unjust to punish everyone when he, Sthenius, was the only one responsible. Pompey spared both town and demagogue.

In the Apophthegmata the story is told of the Mamertines, in the Life of Himera. ‘Himera’ seems right: in Cicero’s Verrines (2.2.110–13) the place is ‘Thermae’, i.e. Himeraean Thermae. That makes it difficult to think that the Life version is based on the Apophthegmata, as a reconstruction on Stadter’s lines implies. We should have to assume that, when he came to the Life, Plutarch realized that he had got it wrong in the Apophthegmata and silently corrected it. That is not impossible, for Plutarch may well have had a more accurate version before his eyes when composing the finished version of the Life (though if so this again raises the question of how crucial the apophthegm-ὑπóμνημα would be for the writing of that final version). But it again seems easier to assume that the Apophthegmata version is the later, and that the error crept in as the Apophthegmata were put together. The collector of the Apophthegmata, Plutarch or another, may have been misled by the prominence of the Mamertines elsewhere in that chapter of the Life, and may simply have misremembered.²⁴

Plutarch told the story again in Advice on Public Life (815e), and there again he erroneously has the Mamertines. Like many of the examples in Advice on Public Life, he may well be drawing this from his reading for the Lives, but doing so from memory. If he misremembered it then, it makes it more likely that it was he himself who was doing the collecting for the Apophthegmata, and misremembered it similarly again.

Case 3: a similar instance is Cicero apophthegm 7 = Cicero 26.11. Metellus Nepos had a reputation for bolting: once (so the Life explains) he had departed prematurely from his tribunate at Rome to join Pompey in the East, then left Pompey and returned just as abruptly. When Metellus buried his teacher of rhetoric, he put a stone figure of a crow on the grave. ‘How appropriate,’ said Cicero, ‘for he taught you to fly rather than to speak.’

In the Life the teacher’s name is given as Philagrus; in the Apophthegmata it is Diodotus. Presumably Philagrus is the original version, and ‘Diodotus’ has come in through some confusion with Cicero’s own teacher Diodotus (Cicero, Brutus 309).²⁵ Indeed, the phrasing of the Apophthegmata version does not make it quite clear if the man is Metellus’ teacher or Cicero’s or both.

Once again, it is possible but unlikely that the Life ‘corrected’ the Apophthegmata version, as Plutarch realized that he had made a mistake. It is more likely that the Life came first, and the Apophthegmata version introduced the mistake.

Case 4: Pompey apophthegm 8 = Pompey 33.8. Phraates of Parthia sent to Pompey and proposed the Euphrates as the boundary of the two realms. Pompey replied that the Romans would define the boundary as whatever was just.

That is the version in the apophthegm; the version in the Life is fuller, and the point is a double one. There Phraates sends both (a) to demand back his son-in-law Tigranes and (b) to suggest the Euphrates boundary; Pompey replies (a) that the boy belonged to his father rather than his father-in-law, and (b) that the boundary would be the just one.

The Tigranes point is evidently the more context-specific, and one needs to know the story-line in order to give it point; one can understand why Plutarch would use it in the Life, but the Apophthegmata would drop that element and concentrate on the simpler and more punchy point about the Euphrates. What, again, is more difficult is to assume that the Apophthegmata version came first, and that Plutarch elaborated the Life version on its basis. If Plutarch in note-taking had wanted to make certain that he would remember the Tigranes point as well as the Euphrates one, he would surely have included both in any ὑπóμνημα-draft.²⁶

Case 5: the next apophthegm is a similar instance, but the other way round: Pompey apophthegm 9 = Pompey 48.7. Lucullus had embarked on his luxurious retirement; Pompey gibed that luxurious living was more unseemly for a man of his age than playing politics. Lucullus replied that elderly good living was more fitting than to take offices which unsuited one’s age – an evident dig at Pompey’s own grabbing of offices against the usual age-restrictions.

That is the version of the Apophthegmata. The Life makes it simpler, with just the gibe of Pompey, ‘good living is more unseemly for an old man than politics’. One can see why. By this time of his career, Pompey is in his mid-forties, and his precocious office-holding is no longer particularly pertinent; and in the narrative structure of the Life the inappositeness would be clear.

Plutarch also uses the story in two other contexts, where we can see the same pattern. In the narrative context of Lucullus 38.5 he is content with the single gibe of Pompey at Lucullus’ luxury: there again we have seen so much of Pompey in the Life that one could no longer count his office-hunting untimely. In Should an Old Man Take Part in Politics? 785f–6a he keeps the double one, and Lucullus hits at Pompey’s ‘untimely love of office and honour’ (τῳ̑ δὲ Πομπηίῳ φιλαρχίαν ἐγκαλου̑ντα καὶ φιλοτιμίαν παρ’ ἡλικίαν) just as Pompey hits at Lucullus’ disgraceful old age.

If we take this case and the last one together, we see the difficulty in assuming either that the Life version is ‘based on’ the Apophthegmata or the other way round. In the one case the Life has the double-point and the apophthegm the single, in the other it is the reverse. If, however, we say that both are based on the same original notes, then both cases are equally explicable. In each case the notes will have had the double-point version, and the final drafts, whether of Life or of Apophthegmata, will have kept or abandoned the points which were appropriate.

That interpretation can account for any case where the one version is fuller than the other. Normally the Life is the fuller, and there are too many instances here to need illustration. More interesting are the cases where the Apophthegmata version has a crucial element missing from the Life. One of those is:

Case 6: Flamininus apophthegm nr. 4 = Flamininus 17.7–8. Antiochus was invading with a vast and varied force. Flamininus told of a dinner he had once eaten where there were an amazing number of different sorts of meat; all turned out to be pork, differently spiced. All these different sorts of Antiochus’ armoury – javelin-carriers, cataphracts, foot-companions, horse-cavalry, and so on – were just the same, all mere Syrians.

Where did this dinner take place? The Life does not say; the apophthegm does – in Chalcis. That is right: this was the version of Livy (35.49.6) and presumably of Polybius.²⁷ Presumably, once again, Plutarch had that detail in his notes, but streamlined it away when he came to produce the version in the Life.

That is also interesting stylistically, for in the Life the story is told as one of an apomnemoneumata-cluster – a collection of brief stories, out of time, illustrating the man’s wisdom or wit. Several Lives have such clusters, notably Themistocles, Pericles, Cicero, and Cato Maior. Stories which figure in those clusters are much less written up than those which are integrated into the narrative movement of Lives. Indeed, in the present case there is not much difference in stylistic level or elaboration between the Apophthegmata version and that in the Life. The Life suppresses some of the more technical vocabulary (καταφράκτους and ἀφιπποτοξὀτας); the Apophthegmata version spreads a little more on the intimidating approach of Antiochus, something which the Life has already made clear in the preceding narrative; the Life has a little more on the narrated dinner-party – Flamininus ‘taking his host to task for the variety of the meat-dishes and asking in amazement where he had got hold of such a varied diet’. But, taken as a whole, one would be hard-pressed to find any systematic difference in level.

That may be significant when we consider whether the Apophthegmata collection itself is really so mean and ill-finished a production as scholars have often thought. Its level is precisely what is appropriate to a genre like this, just as it is to those passages in the Lives which generically approximate to such a collection.²⁸ The watchwords are economy, directness, and simplicity, with everything subordinate to the forceful direct speech itself. And in that case the view of for instance Simon Swain, holding that the work is simply too mediocre to dedicate to Trajan, looks less plausible.²⁹

Let us return to the question we asked at the outset: how similar was Plutarch’s preparation for the Lives to the procedure which van der Stockt and van Meir-venne are uncovering in the Moralia? If Stadter is right about the Apophthegmata Laconica, then we seem to be dealing with ‘notes’ of a very different sort from the ὑπομνἠματα which van der Stockt and van Meirvenne are investigating. If I am right, however, the procedures may be much more similar. Certainly, the comparison of Apophthegmata versions and Life versions has much in common with the team’s comparison of Moralia clusters, and here too we may be detecting an underlying ὑπóμνημα which two passages share. But perhaps this is to go too fast; all we have so far shown is that there is ‘something else’, some sort of notes which underlie both versions. In the next few examples we shall try to detect how close those ‘notes’ are to any finished version, though it would not be surprising if there was no clear or single answer: many scholars are used to having penultimate drafts which are fairly finished in one paragraph and disgracefully scrappy in the next.

Relevant here is another feature which reminds us of the Moralia findings of van der Stockt and van Meirvenne: the odd loose ends of thought or argument. We have already seen one case of this in Caesar, and I suggested that this instance might go back to a draft in which the item figured in a different train of thought.³⁰ If so, that at least points to ‘notes’ where full trains of thought were written down: this is not just a scrappy collection of stray facts. We can also find cases in the Apophthegmata where the story would be difficult to follow if we just had the apophthegm alone, but we can make sense of it when we place it in the narrative context. Those again look like cases where the original shaping of the ‘notes’ is still directing the way the story comes out in this finished form. Consider for instance:

Case 7: Caesar apophthegm nr. 3 = Caesar 10. Caesar’s wife Pompeia has become involved in the Bona Dea scandal, and Caesar divorces her because his wife should be above suspicion.

The apophthegm introduces the story very allusively. Pompeia is ‘the subject of scandal because of Clodius’; Clodius is ‘on trial for this’ – for what? Hardly for sexual ‘scandal’! We need the Life to tell us: for ἀσεβεια (as Plutarch, not very precisely, renders the Latin incestus).³¹ Even Caesar’s divorce is introduced fairly obliquely. He is called as witness at the trial, and said nothing discreditable about his wife. ‘Why then did you divorce her?’ asks the prosecutor – yet this is the first we have heard of the divorce. That is not a natural way to tell the story, even after we have taken into account ancient narrative’s tendency to delay crucial details. It all looks like an over-casual abbreviation of a complex original, made more difficult because the apophthegm belongs right at the end of the sequence. The scandal, the nature of the charge, the divorce would all have figured earlier in the ὑπóμνημα version, and the collector here, picking up the story only in the final stages, has not worked hard enough to clarify what has already happened.³²

Can we detect anything more about the form such a ὑπóμνημα would take? Stadter’s argument from the ordering of the Spartan anecdotes³³ makes it likely that any such ὑπóμνημα would come after the main lines of a Life’s narrative strategy had been fixed: that argument holds good even though we have different ideas of how the ὑπóμνημα would look.

Stadter implies that, though the ordering reflects a reasonably late stage of the process, the actual shaping of the anecdote would not. For him the Apophthegmata version would still be in the form in which Plutarch had gathered it from his preliminary reading, snipped out from its original Lesefrüchte setting and repositioned in its right place for the Life. I agree with Stadter about the ordering, but will go further, and argue that we can sometimes see reflections of the Life’s narrative strategy in how the story is told, not merely where. If so, that suggests that they are more than snippets from the original reading.

Case 8: Caesar apophthegm nr. 1 = Caesar 2, Caesar and the pirates. We should also notice the additional detail of Crass. 7.5, surely acquired during Plutarch’s reading for Caesar but suppressed here as distracting in both Life and apophthegm: at one point Caesar exclaimed, ‘Crassus, how delighted you’ll be to hear of my capture!’ Plutarch may well have included it in any notes or draft that he took for Caesar, knowing that it might be useful for Crassus even if he could not use it in Caesar. But we cannot be sure of that: the item may simply have been remembered by Plutarch when he came to write up Crassus, whether or not he had ever noted it down.

If we return to the version in the Apophthegmata, we should first notice the dating: ‘when he was fleeing from Sulla when he was still a meirakion (a man in his late teens), ὅτε Σύλλαν ἒφευγεν ἒτι μειράκιον ὤν. That coincides with the date given in Caesar itself: in ch.1 Caesar has his brush with Sulla ‘when he was not quite yet a meirakion (οὔπω πάνυ μειράκιον ὤν, 1.3),³⁴ and has his encounter with the pirates after escaping. But this date is almost certainly wrong. There are good reasons for placing this capture in late 74 or early 73.³⁵ We can also see why Plutarch might have deliberately misdated it, bringing it before several episodes which in fact it followed. This allows him to gather together all Caesar’s early eastern experiences: these then culminate in his rhetorical training at Rhodes (ch.3); and that training duly sets him up to return to Rome and win glorious successes in the Antonius and Dolabella trials (Caes. 4.1–4) – trials which in fact took place in 77 and/or 76, some time before the pirate episode.³⁶ Suetonius has the sequence right (Div. Iul. 2–4), yet his presentation of the material is otherwise very close here to Plutarch’s own, and it must be highly probable that Plutarch himself has shifted the pirate episode away from its proper context. That makes it most interesting that the Apophthegmata version also has that misdating. This suggests that this ordering already figured in the notes or draft: this therefore already presumes an elaborate degree of narrative articulation.

We might also notice the detail in the two versions. There is a good deal of circumstantial material: the ‘one friend and two attendants’ who accompanied Caesar (2.2 ~ Suet. 4.1), and also the ‘thirty-eight days’ which he spent with the pirates (2.3 ~ prope quadraginta dies, Suet. 4.1). Both these items figure in the Life but not in the Apophthegmata:³⁷ presumably they were there in the draft, and were kept in the Life version but dropped in the Apophthegmata. In those two cases the items are shared with Suetonius, and hence we can presume that they go back to a source. Other points look much more like Plutarch’s own imagination – Caesar’s impromptu poetry recitals, for instance, and the abuse of the pirates for being such unsophisticated listeners; or his demands for silence whenever he went to bed. Neither features in any of our other versions of the story, and both will strike seasoned Plutarch-watchers as exactly the sort of minor detail which he happily adds.³⁸ Both these details occur in the Apophthegmata version as well as the Life. If they are indeed his own, it again suggests that the draft included a lot of the creative elaboration which would figure in the final version, even if it lacked its final stylistic polishing.

Case 9: Caesar apophthegm nr. 4 ~ Caesar 11. Caesar is reading of Alexander’s achievements, and bursts into tears. He tells his friends that ‘at this age Alexander had conquered the world, and I have so far achieved nothing’.

This is another case where Plutarch is heavily at work in the Life – not surprisingly, given this pairing of Alexander and Caesar. There is once more a question of dating, for again Plutarch has moved the story. Suetonius and Dio date it to Caesar’s Spanish quaestorship in 69–8,³⁹ and that must be the original version: then Caesar was genuinely ‘at the age when Alexander had conquered the world’. Plutarch moves it eight years later to Caesar’s Spanish proconsulship in 61–60. Again, one can see why. This is just before Caesar returns to Rome and launches on the political path which will take him to glory and power – and to be a match for Alexander. The Apophthegmata also misposition the story in the same way: it is put after, not before, the story of Clodius and Pompeia, dated to late 62. We must again agree with Stadter that the reordering of material was already in the ὑπóμνημα, and that the ordering of the final narrative has already been considered before this ὑπóμνημα takes shape.

The details are also significant. Plutarch’s version is irregular: Suetonius and Dio both have Caesar weeping, not because he is reading about Alexander, but because he has seen a statue of Alexander in the temple of Hercules at Cadiz Those extra details have point, for Cádiz was conventionally one end of the civilized world, and Hercules had traversed that world to west just as Alexander had to east – and just as Caesar had not. Plutarch had his reasons for that alteration, as we shall see in a later chapter.⁴⁰ Once more, that ‘reading’ version is shared by Life and by apophthegm, and we should infer that most of Plutarch’s tampering had already taken place by the time of the draft.

That point can be elaborated to a more general one if we consider the apophthegmata proper, the pieces of direct speech which constitute the episodes’ punch-lines. So many of these are extraordinarily similar, often verbatim, in Apophthegmata and in Life; yet here again Plutarch-watchers will know that he readily adapts direct speech from his source if it is less forceful than his own tastes would demand. Russell showed that in his analysis of Coriolanus, bringing out how Dionysius’ laboured rhetoric is regularly improved by Plutarch into a more effective phrase or sentence.⁴¹ Yet in the Apophthegmata it is Plutarch’s phrasing which we regularly find, not anything more cumbersome. If the Apophthegmata are keeping the language of the draft, this confirms that a lot of the improvement on his sources had already happened by the time that draft was composed.

All this suggests that these ‘notes’ for the Lives were indeed quite close to the sort of ὑπομνήματα which van der Stockt and van Meirvenne have posited for the Moralia – a full, not particularly scrappy, account, already showing a good deal of thoughtfulness and shape. There is however one big difference. In a Moralia cluster the original ὑπóμνημα could have taken shape at any time: perhaps when Plutarch was collecting his thoughts for the first essay where a cluster appears, perhaps simply in a ‘commonplace book’ at an earlier stage, perhaps even when he was writing a quite different essay which has not survived. In the Lives we can be more certain about the stage when those notes took shape: it was presumably during the preparatory work for a particular Life or group of Lives.

It is interesting here to look again at the ordering of the stories, the clue which triggered Stadter’s analysis. The collections which are most distant from the Lives’ ordering are those where Plutarch may have been able to rely most on his memory for a rich fund of witty remarks: Alexander, Agesilaus (in the Apophthegmata Regum et Imperatorum though not in the Laconica), and among the Romans Cato Maior and Cicero. The ones which are closest are the cases where most ‘research’ would be needed, directed and focused preparatory reading: not merely most of the Roman ones, but also Phocion. Those are precisely the cases where the ὑπóμνημα would be the obvious place to collect the results of that reading. The collector of the Apophthegmata – Plutarch or another – would accordingly know that this was the place to find an even richer collection of material than in the finished Life.

It would be fascinating to know whether each Life had its own ὑπóμνημα, or if Plutarch would sometimes combine his preparation for several Lives into a single draft: given the close similarities when the same events are described in several Lives, this second alternative must be a possibility. If he ever proceeded in that way, then he would surely do so when he was writing the great Lives of the end of the Republic. That is particularly so if I was right in chapter 1 to argue that six of those Lives – Caesar, Pompey, Cato Minor, Brutus, Antony, and Crassus – were prepared together as a single project. If he did draft in this way, then that multiple-Life ὑπóμνημα would be almost a draft Plutarchan history of the whole period, though one rich in biographical excursions – a fascinating prospect indeed.

In chapter 1 I left it open whether he proceeded like this.⁴² I will once again leave it open here, though I now incline to think that he did not. One test-case, though an indecisive one, is:

Case 10: Caesar apophthegm nr. 11 = Caesar 44.8 = Pompey 69.6–7. At Pharsalus Pompey ordered his troops to stand and receive the charge of the Caesarians. Caesar remarked that this was a bad error, with Pompey losing the tightening of the muscles and the zing and spirit that men generate during a charge (τὸν ἐξ ἐπιδρομη̑ς μετ’ ἐνθουσιασμου̑ τόνον καὶ ῥοί̑ζον).

That is the phrasing of the apophthegm. It is similar to that in the Pompey, which does not, sadly, have ‘zing’ (the nearest equivalent I can find for ῥοί̑ζον) but does have τòν ἐξ ἐπιδρομη̑ς τόνον (‘the tightening which comes from the charge’) and goes on to talk of τἡν ...ἐνθουσιασμου̑ καὶ φορα̑ς άντεξόρμησιν (‘the responding surge of spirit and momentum’); the language is much elaborated there. It is rather more distant from the wording in Caesar, which makes the same point but uses rather different images – the ‘clash together’ (σύρραξιν) which adds force to blows, the ‘joining to inflame’ (συνεκκαίει) the spirit which is ‘fanned up’ (ἀναρριπιζόμενον) by the encounter. It looks as if the language of apophthegm and Pompey is closer to what Plutarch found in his source, presumably the usual Pollio-source; in the closely parallel passage Appian also uses the τονος metaphor (‘tightening’, BC 2.79.330).

The natural implication is that this was also the language that stood in the ὑπóμνημα – but which ὑπóμνημα is that? If it were a simple ὑπóμνημα for Caesar, then reused for the Caesar apophthegmata, we would have the paradox of Pompey being closer to the Caesar ὑπóμνημα than Caesar is itself. That would be odd. Should we then prefer to think of a single shared ὑπóμνημα for both Lives, which for his own reasons he chose to follow closely in Pompey and more distantly in Caesar? That is tempting – but the argument is inconclusive. If there were different ὑπομνήματα for each Life, we could equally presume that the two drafts for Caesar and for Pompey might each include the story in similar form, and that he kept to that form more closely in the one case than in the other.

On the whole there is less evidence for a single, multi-Life ὑπóμνημα than we might have expected if Plutarch had worked in that way. This is a period where so many great men’s stories intersect, with none of them reluctant to launch a good line when one offered. If X says something witty about Y, it might be usable in either X’s or Y’s Life, but it can only appear in X’s collection of Apophthegmata. If Plutarch was working from a single multi-Life ὑπóμνημα, what we would expect would be a fair degree of cross-over: that is, cases when a man’s apophthegm cropped up in another Life rather than his own, but – as the collector worked through a single ὑπóμνημα sequentially – still figured in his own Apophthegmata. In fact we get a few cases, such as Crassus using the Caesar apophthegm in the pirate-episode,⁴³ but only a few, and those can be explained by Plutarch’s use of memory, or by his inclusion of the apophthegm in more than one ὑπóμνημα. It probably is more likely that each Life had a single ὑπóμνημα, and the Apophthegmata-collector worked through each of these drafts in turn as he gathered each man’s best remarks.

One further point about this case. This story eventually derives from Caesar’s Commentarii, where he makes the same point, talking of the natural excitement which is inflamed as one enters battle: quod nobis quidem nulla ratione factum a Pompeio uidetur, propterea quod est quaedam animi incitatio atque alacritas naturaliter innata omnibus quae studio pugnae incenditur (‘This seems to me a mistake by Pompey, for there is a certain stimulation of the spirit and an innate eagerness which is inflamed by the enthusiasm for the fight’), BC 3.92.⁴⁴ So Caesar made this criticism in writing, and in both Caesar and Pompey Plutarch makes this clear by using present tenses, ‘he says’ (φησιν) in Caesar and ‘he criticizes’ (αἰτια̑ται) in Pompey.⁴⁵ So indeed does Appian, ‘he blames’ (καταμἐμφεται).⁴⁶ But the Apophthegmata version gives the impression of an oral comment, perhaps on the battlefield itself, perhaps a repeated remark⁴⁷ – ‘he said’ or ‘would say’, ἔλεγε. We could put this down as simple confusion; but it is more likely that this version is deliberately recasting it as an apophthegm, making it an anecdote rather than a source-citation and thereby aligning it more closely with the rest of the collection. If that is so, the recasting into Apophthegmata-form begins to seem quite radical.⁴⁸

These last few cases have been drawn from the later group of six Lives, which I think were prepared together; even here, then, we should probably imagine one ὑπóμνημα for each Life rather than a multi-Life shared version. If this is right, it will make Plutarch’s procedure similar to that which we would anyway posit for the earlier Cicero, where (except for some minor overlap with Lucullus, prepared at around the same time) he was working only on a single Life from the period. In other ways, though, the preparation for Cicero looks rather different. The Life itself has several apomnemoneumata-clusters, especially in chs. 25–6 and 38; once again, as in the case of Flamininus (case 6 above), the stylistic level in these clusters is not very different from that in the Apophthegmata collections, and in some cases it is the Apophthegmata version which is the more elevated of the two.⁴⁹

The choice of story, however, varies quite a lot between Life and Apophthegmata. If we compare apophthegms 4–13 with the overlapping collection in Cicero 25–6, there are eight stories which are told in both, together with a further two about Verres which were used earlier in the Life; a further thirteen are used in the Life but not in the Apophthegmata. If we then compare apophthegms 15–19 with the similar cluster in Cicero 38, two are told in both, four just in the Life, and three just in the Apophthegmata. The obvious explanation is that the ὑπóμνημα gathered a large number of such stories, and that later two separate, independent selections were made, one by Plutarch for the various clusters in the Life and one for the Apophthegmata. (Nor is it difficult to guess where most of this material originally came from: Tiro collected three books of Cicero’s witticisms in his de iocis, and also wrote a biography which presumably included much of the same material.⁵⁰) That would suggest that in this case the ὑπóμνημα was rather less close to the final version in the Life than we found likely with the later Lives.

If we look at those stories which appear in the Apophthegmata but not in Cicero itself, one case is particularly interesting:

Case 11: Cicero apophthegm 15 = Pompey 63.2. Cicero blamed Pompey for leaving the city and ‘imitating Themistocles rather than Pericles’.

This does not appear in Cicero, but it does crop up in the later Pompey. The original source is Cicero himself, who uses the Themistocles comparison in two letters to Atticus (7.11(134).3 and 10.8(199).4). The Pompey version duly uses the present tense, Cicero ‘blames’ Pompey (αἰτια̑ται), again indicating that this is drawn from Cicero’s own writings; and again the Apophthegmata version has the past tense ‘blamed’, ἐμέμψατο, turning it into an anecdote and making it parallel to the rest of the Apophthegmata.

This use of Cicero’s own writings – primary sources – is striking. In chapter 1 I argued that Plutarch conducted plenty of this sort of research for Cicero itself, but nothing more when he came to write the later Lives; I suggested that he found his Pollio-source so full that he dispensed with any further reading of this type.⁵¹ I also suggested, though, that when writing the later Lives he could naturally look back at his earlier research, and exploit again some of that earlier reading in primary sources.⁵² We saw earlier a case where he seems to have done exactly that in Caesar, integrating with his later material something from his earlier reading in Cicero’s own work.⁵³ This now seems to be a second case, something gleaned from that primary research at the time of Cicero, included in the ὑπóμνημα but not in the final version of the Life, and then drawn again from that ὑπóμνημα for use both in Pompey and in the Apophthegmata collection for Cicero himself.

If this general picture is right, then several further questions can be asked – even if we can only attempt the sketchiest of answers here.

What, for instance, may have determined the omission of certain good stories from the Apophthegmata? ⁵⁴ Why should the most famous of all Caesar’s lines be omitted, his hoc uoluerunt (‘they would have it thus’) on the battlefield of Pharsalus? Plutarch makes a great deal of that in Caesar itself, even quoting Pollio on the issue whether Caesar delivered it in Latin or Greek (46.2); and so Pollio too included the dictum, and we can be sure that Plutarch read his Pollio-source before constructing the ὑπóμνημα for Caesar.⁵⁵ And a scan of the other Lives shows many other memorable remarks not used in the Apophthegmata. It is hard to think that all of them were absent from the ὑπóμνημα that, if I am right, they shared with their respective Lives.

Two patterns emerge that might give us a clue. The first is a tendency to drop stories which are morally discreditable. Take Pompey. In the Life we have the story of his arrogance to the Mamertines: ‘will you not stop reading out laws to us when we have swords at our side?’ (10.3). We have his disingenuous hypocrisy when voted a command for which he had intrigued, ‘O for all these pointless trials! It would have been better to be a nobody, if I am never to cease my soldiering...’ (30.7): Plutarch there goes on to make his disapproval clear. We have Pompey’s response in 59 when Caesar asked for his support: ‘Indeed, and I will meet those who threaten swords by bearing both sword and shield’ (47.7). Plutarch adds that this was the most disgraceful thing he had ever said. We have Pompey refusing to listen to Hypsaeus when he buttonholes him on his way to dinner: ‘you’re spoiling my meal, but achieving nothing else’ (55.10). And we have him so confident that ‘when I stamp on the soil of Italy, infantry and cavalry will spring up’ (57.9 = Caes. 33.5). None of these is at all improving; none should appear in a collection which a reader might find usable for his own purposes – or indeed a dedicatee, for one would hardly recommend to Trajan that he should treat a supplicant by telling him that he was spoiling the imperial dinner.

There are times, too, when one can see a story being bowdlerized in the Apophthegmata, even when it is included:

Case 12: Pompey apophthegm 12 = Pompey 51.7–8. Pompey and Crassus are asked after Luca if they will stand for the consulship. Pompey says that perhaps he will, perhaps he won’t; Crassus replies ‘more diplomatically’, saying that he will do whatever is in the public interest. Marcellinus attacks Pompey vehemently; Pompey calls him the most unjust man alive, as he owes it to Pompey that he has become eloquent rather than dumb, and a man who can vomit his food instead of a starveling.

That is how it is told in the Life. The Apophthegmata version drops the ‘perhaps he will, perhaps he won’t’ remark, which Plutarch clearly found offensive (in the Life Crassus is explicitly the ‘more diplomatic’). It also presents an extraordinarily simplified version of the politics. This is the point ‘when the rift with Caesar was becoming clear’: that is a travesty of the implications of Luca, when the whole point was that any rift seemed to have been healed.⁵⁶ Marcellinus himself is then ‘someone who seemed to have been brought on by Pompey, and had then gone over to Caesar’. Yet this is the consul of 56, Cn. Cornelius Lentulus Marcellinus, who ‘gave general support to the optimate position against Clodius and the so-called First Triumvirate’⁵⁷: he was not remotely a Caesarian. Even the apophthegmatic tradition did not usually make him into one: Valerius Maximus has him encouraging the Romans to ‘applaud, Quirites, while you can; soon you will not be able to do so unscathed’ (6.2.6), again the sort of thing an optimate would say, not a Caesarian. But here the Apophthegmata turn Marcellinus into a time-server faithlessly switching his allegiance when the ‘rift’ comes, someone whom Pompey could properly rebuke for ingratitude, rather than an uncompromising defender of the free state.

This taste for the morally improving is indeed a tendency, no more. There are certainly some apophthegms which are less creditable, especially with Cicero: his jest for instance, at Voconius and his ugly daughters, ‘against Phoebus’ will he once begat his brood’ (12 = Cic. 27.4). But even in that case there is a slight tweak. In the Life Cicero ‘spoke up’ with the remark (ἀνεφθέγξατο), in the Apophthegmata he ‘said it softly to his friends’. And in Cicero 25 Plutarch explicitly stigmatises several stories as sacrificing ‘propriety’ (τò πρέπον), stories like his riposte when Crassus had commented on how short-lived his family had always been, and added ‘now I wonder why I said that?’ ‘Because’, said Cicero, ‘you knew it would go down well with the audience, and were playing the demagogue.’ A whole group of these ‘indecorous’ stories, some half dozen of them (25.1–26.2), is then dropped from the Apophthegmata.⁵⁸

So moral improvement may be one reason for dropping a dictum from the Apophthegmata. There are also cases where it may be a question of generalizability: when an apophthegm is closely tied to a particular set of circumstances, it is more likely to be dropped. That may be for presentational purposes, for the more one has to put a story in context, the more likely one is to lose one’s audience. Or it may be a question of applicability: if the aim is really to provide apophthegmata which could be appropriated for a reader’s own purposes, then the more context-specific a story the less usable it is likely to be. That helps to explain the selectivity in the Phraates anecdote (Case 4, above): a reader – or a dedicatee, if that dedicatee is an emperor – might well find a time to say that ‘the Romans will use whatever boundary is just’, but there would not be many opportunities to say that a hostage belonged to his father, not his father-in-law. No wonder the Apophthegmata kept only the more generalizable as well as the more pungent part of the remark.

This too may be a broader point. Does it help to explain why the younger Cato and Brutus are given no Apophthegmata collections? They had their epigrammatic moments too, but the points they made would be less applicable, or at least less tactfully applicable, to the circumstances with which the users of the Apophthegmata would have to deal.

Both our tendencies – to keep the morally creditable and the generalizable – may help to explain the dropping of hoc uoluerunt. The idea that ‘it was their fault’ might seem a less than dignified comment on a great national tragedy; and a reader in the early second century would rarely find the right occasion to say anything like that – even, or perhaps especially, if that reader was Trajan.

There is a good deal more to ask about the Apophthegmata. We could explore their organization, basically but not entirely chronological as it is: there are times when one can see some artistry in the ordering within the Apophthegmata, just as we can with the ordering of those apomnemoneutic clusters in the Lives.⁵⁹ We could do more to analyse their style, especially in those cases where the Apophthegmata versions are just as finished as, perhaps even more finished than, their counterparts in the Lives.⁶⁰

Perhaps most rewarding of all, we could trace through the implications of this study for the literary criticism of the Lives. If the Apophthegmata versions allow us to reconstruct more of what stood in Plutarch’s penultimate draft for the Lives, that offers new, rich possibilities for exploring what he added in that ultimate draft. Once again, there is great potential here for further analysis of the type which van der Stockt’s Leuven group have made their own, and it will illuminate the peculiarities – and peculiar skill – of both Apophthegmata and Lives.⁶¹

Particularly rich will be those cases where we can also see a story being exploited in the Moralia. A prime instance will be Caesar’s bold but unsuccessful voyage in a small boat in the winter of 49–8 BC, eager as he was to get back to Italy and link with his troops (Caesar apophthegm nr. 9 = Caesar 38). That story is told in On the Fortune of the Romans as well (319b–d), and it looks as if all three versions are based on the same ὑπóμνημα. That case cries out for a cluster analysis.⁶² It would also, incidentally, carry important implications for the dating of On the Fortune of the Romans.⁶³

But it is time to finish by drawing out a few, brief further implications.

First, though I have been defining my suggestions in opposition to Stadter’s, it should be stressed that we are looking at different Apophthegmata. He concentrates on the Laconica, I on the Regum et Imperatorum, and there do seem clear differences between the collections: the differences in the ordering of the Spartan anecdotes within the Regum et Imperatorum are enough to show that. It is logically possible that both of us are right, for our respective collections⁶⁴ – or indeed that both of us are wrong. If either of us is right on the collection we have taken, it makes it more likely that the same reconstruction would work for the other collection as well, but that is not a logical necessity.

Secondly, I have so far avoided committing myself on whether the collection of Apophthegmata was made by Plutarch himself or by a later ‘editor’. But perhaps that is too cautious. If this reconstruction is right, then that collector was much more likely to be Plutarch: the person who knows his way best about his notes and drafts is always the author himself.

If so, then a third conclusion follows too: simply that Plutarch regarded such a collection as a sensible artistic thing to do. If it were a later collector, it might be merely an exercise of literary curiosity, born of the delights of access to a wonderful archive: but Plutarch himself was hardly likely to be so impressed by his own filing cabinets. That, in its turn, makes it more likely that this was indeed a collection meant for a purpose that the Lives itself could not serve, such as providing some leisure-reading for a busy man who would not have time to read through the longer work.

A man, indeed, like Trajan... And even if we do not take the dedication too simply, that dedication is still pointful. It may be less an indication of Trajan as ‘real reader’ or target audience, rather a signal to the real readers, whoever they may be, that this is the quintessential ideal reader, a man of affairs, a man with little time on his hands, but a man who still appreciates the wit of the ancients when it is presented to him in convenient form; a man, in short, whom they would feel both intrigued and flattered to have as their potential fellow-audience.⁶⁵

Notes

¹ In the original 1979 version of ch.1, p. 29 n. 3.

² See esp. van der Stockt 1999a, where the implications for Plutarch’s working methods are drawn out very clearly: see also van der Stockt 1999b and forthcoming, van Meirvenne 1999 and 2001, and van der Stockt and van Meirvenne, forthcoming.

³ So I argued at pp. 19–26.

⁴ Pollio-source: pp. 12–13.

⁵ Probably no extensive note-taking on the Pollio-source: pp. 22–3.

⁶ Research assistants: p. 24.

⁷ Thus Stadter, forthcoming, convincingly detects a case of last-minute reordering of anecdotes in Agesilaus, after the ὑπóμνημα stage. The Apophthegmata Laconica generally present Agesilaus anecdotes in the same order as in the Life, but three of the forty are out of order: two are at the beginning of the Life (ch.2), one at the end (ch. 36). The Apophthegmata probably retain the order of the ὑπóμνημα, and Plutarch decided in his final draft to move these three items to positions of special prominence. Stadter and I differ on the nature of that ὑπóμνημα, as will become clear, but this suggestion can hold on either model.

⁸ Van der Stockt 1999a, 595.

⁹ Van der Stockt 1999a, esp. 583–4. He finds other inconcinnities too in the How to Tell a Flatterer passage; his argument is subtle and convincing, but too complex to summarize adequately here. Notice also Ingenkamp, forthcoming, who finds a trace of a similar ὑπóμνημα cluster at Tranquillity of Mind 11, just before that passage; van der Stockt 1999b, arguing for another just after, involving Tranquillity of Mind 13, and van der Stockt and van Meirvenne, forthcoming for another in Tranquillity of Mind 16 and possibly 7. – Tasos Nikolaidis (to whom I am most grateful for a detailed critique of this chapter) puts to me that the How to Tell a Flatterer passage makes better sense than van der Stockt suggested. Nikolaidis puts weight on the transitional sentence at 58c–d, which talks of the excessive respect paid to wealth and outward appearance. That is relevant both to the Megabyzus story, with children initially impressed until the man opens his mouth, and to Solon and Croesus, where Solon’s frankness contrasts with the sycophancy which the king normally received. That may well be the underlying train of thought, the one which on this reconstruction figured in the ὑπóμνημα; I am not convinced that it reduces the inconcinnity in the text, for this is a very general point about flattery, one which does not link closely with what precedes or what follows, nor forms a very satisfactory transition between them.

¹⁰ On Caes. 8.3–4 see pp. 48 and 53.

¹¹ For this and other possibilities, including that of the ὑπóμνημα, pp. 52–3.

¹² p. 48 and n. 22, quoting Hardy 1924, 104 n. 1.

¹³ pp. 46–9.

¹⁴ Stadter, forthcoming. His paper and mine were delivered at the same conference at Leuven in June 2001, and will both appear in the proceedings of that conference (van der Stockt and Stadter (eds.) forthcoming).

¹⁵ On the face of it this brings his view of the Regum et Imperatorum closer to mine, as both of us think that Apophthegmata and Lives are both based on the same, earlier ὑπóμνημα. But our reconstructions of that ὑπóμνημα are very different, for in this case too his ὑπóμνημα is purely anecdotal, mine more complete and finished.

¹⁶ As in the case of Agesilaus and the Apophthegmata Laconica (but not the Agesilaus collection in Regum et Imperatorum), with the interesting exceptions noted above: see n. 7.

¹⁷ See pp. 94–5 and ch.5.

¹⁸ Beck (forthcoming); see also Fuhrmann 1988, 5–6. Volkmann 1869, 216–17 found the notion of the Lives as a single σύνταγμα ridiculous, and used this as an argument against authenticity. Ironically, recent scholarship has developed the idea that different Lives mutually cohere, especially in the pairings but also more widely: cf. pp. 44 n. 169 and 187–8; Mossman 1992, 103–4 on the likelihood that Pyrrhus and the other Hellenistic Lives presume a knowledge of Alexander; Harrison 1995 on the possibility that Agesilaus–Pompey and Demetrius–Antony are written against the background of Alexander–Caesar.

¹⁹ Volkmann 1869, 210–34, followed by Ziegler 1949, 226–8.

²⁰ For one such instance see case 6, pp. 74–5; more generally, Volkmann 227 on the extra details in the Apophthegmata, and 228–30 on divergences between Apophthegmata and Lives. If the argument of the present paper is accepted, those divergences admit of varying explanations. In some cases, they will be simple errors which crept into one work or the other after the ὑπóμνημα stage (e.g. cases 2, 3, with further instances in n. 26); in other cases, deliberate manipulation of detail in one of the finished versions (e.g. cases 10, 12, with further instances in nn. 58–9), something which today we will more readily acknowledge than in the days of Volkmann. However, this evidently needs more systematic analysis.

²¹ Ziegler 1949, 227 put this even more strongly than Volkmann: ‘Wie kann man glauben, da£ ein Mann, der die Dicta aus den Schriften P.s heraussuchte und zusammenstellte, den Text in so vielen Fallen willkürlich und meist im Sinne einer Verschlechterung geândert haben sollte?’ If my argument is correct, this underrates the artistic finish of the Apophthegmata, and fails to pay attention to the differing stylistic levels appropriate to such gnomologies: see p. 75.

²² See pp. 67–8. This reconstruction is not very different from the underlying view of Fuhrmann 1988, 7: ‘Sans doute a-t-il utilisé des notes qu’il avait prises au cours de ses travaux ou simplement de ses lectures; sa tâche aura constitué ici à compléter, à mettre en ordre...’ But I will try to give a fuller account of the nature of these ‘notes’. Ziegler 1949, 227 thinks that Plutarch constructed his Lives ‘aus einer Sammlung von Apophthegmen..., die mit der uns unter seinem Namen erhaltenen nahe verwandt, aber nicht identisch, sondern umfassender war als sie’. If I am right, the ὑπóμνημα for the Lives would indeed be much ‘fuller’, and would not be a ‘collection of Apophthegmata’ in any real sense: Ziegler’s view is in fact closer to Stadter’s, at least for the Reg. et Imp. I feel that the Apophthegmata are less mediocre a work than Ziegler thought (see last note). Stadter is more inclined to accept the mediocrity: cf. below, n. 29.

²³ They may well be too large: Memnon FGrH 434.57.4 gives a total of 80,000 men, Phlegon FGrH 257.12.10 40,000 infantry and 30,000 cavalry. But that does not affect the source-question here: indeed, the way that Appian and Plutarch inflate to more or less the same degree makes it more likely that the two notices have the same provenance.

²⁴ So e.g. Fuhrmann 1988, 304. Münzer, R-E iiia. 2335–6 thought that ‘Sthenius’ or ‘Stenius’ might have been a descendant of the Mamertines of Messene, and is followed by Heftner 1995, 108. That would explain the confusion, but seems to me over-generous to Plutarch.

²⁵ So Münzer, R-Exix. 2108 nr. 2.

²⁶ For some other cases where the Life version cannot realistically be based on the Apophthegmata, see case 8 (Caesar and the pirates, where the Life has some details shared with Suetonius but absent from the Ap. version, p. 77); case 10 (Caesar’s remark about Pharsalus, where the Ap. version would give the impression that it was something Caesar said at the time rather than included in his Commentarii, p. 80; Caesar apophthegm 8 = Caesar 35, where Metellus is ἔπαρχος του̑ ταμιείου in Ap. (‘in charge of the treasury’, presumably therefore quaestor), correctly tribune in Life; Aemilius apophthegm 9 = Aem. 35–6, where his second son dies five days after the triumph in Ap., three days after in the Life: three days will have been the version of Plutarch’s source (cf. Livy 45.40.7, Val. Max. 5.10.2) – though the explanation here may lie in textual corruption of numbers.

²⁷ So Briscoe 1981, 213 on the Livy passage. Some other cases where the Ap. version includes a detail lacking in the Life, or seems closer to Plutarch’s source-material: Lucullus apophthegm 2 = Lucullus 28, where the two versions are close, but the Ap. also has the cry that ‘it is more trouble to despoil them than to defeat them’; Caesar apophthegm 10 = Caesar 39.8 = Pomp. 65.8, where both Lives have Caesar’s comment as .. .αί τòν νικω̑ντ’ είχον; the Ap. has ...ἀλλὰ τòν εἰδότα νικα̑ν οὐκ ἔχοὐσιν, less pungent but apparently closer to the version Plutarch would have found in his sources (cf. App. BC 2.62.260, ...εἰ τòν νικα̑ν ἐπιστάμενον είχον, Suet. Div. Iul. 36, ...negauit eum uincere scire). See also Volkmann 1869, 227, with n. 20 above.

²⁸ On the gnomological genre cf. Engels 1993, Gemoll 1924.

²⁹ Swain 1990d, 247–8. Stadter, forthcoming, similarly assumes stylistic mediocrity, but prefers to use it as an indication of Plutarch’s confidence of his relationship with the emperor, someone to whom he did not ‘need to use his best rhetorical technique’. But there are genres when the ‘best rhetorical technique’ would itself commend an unpretentious, unelaborated style.

³⁰ Above, pp. 67–8.

³¹ The charge of incestus assimilated, so it seems, the religious outrage to the defilement of the Vestals’ chastity: so Moreau 1982, 83–9, followed by Tatum 1999, 74–5. Plutarch refers to ἀσέβεια at both Caes. 9.6 and Cic. 28.4. This was the normal Athenian charge for behaviour prejudicial to the sanctity of festivals: Harrison 1971, 62–3.

³² A few more awkwardnesses and loose ends in Apophthegmata which may be survivals from the fuller draft: Pompey nr. 1, where ‘he was hated as much as his father was loved’ is awkwardly abrupt: the equivalent in the opening chapter of the Life introduces his father more smoothly. Caesar apophthegm 8 = Caes. 35, which sets the story (Metellus and the treasury) in context by saying ‘when Pompey had fled to the sea’: that detail is accurate, for Pompey has not yet sailed at this stage (cf. Caes. 35.2), but it has real point only for those who know the context. We might have expected a vaguer ‘from Rome’.

³³ See p. 68 and also n. 7 above.

³⁴ The contradiction between  is more apparent than real. For Plutarch Octavian is a μειράκιον at 18 (Cic. 44.1, 45.2, 45.5) and a year later οὔπω πάνυ μειράκιον (Brut. 27.3). In Caes. 1 Caesar is about 18. The pirate episode is anyway a little later.

³⁵ See p. 93 and n. 5.

³⁶ See p. 93.

³⁷ Thus this is another case where the Life cannot simply be based on the Apophthegmata version: cf. above, n. 26.

³⁸ See pp. 94–5, 152–6.

³⁹ Suet. Div. Iul. 7, Dio 37.52.

⁴⁰ On Plutarch’s displacement and treatment of the Cadiz story see p. 257.

⁴¹ Russell 1963, esp. 22 and 26 = Scardigli 1995, 359–60, 368.

⁴² p. 23.

⁴³ p. 76. Another will be case 11, the use in Pompey of Cicero’s Themistoclean apophthegm.

⁴⁴ Despite that ‘inflamed’ (incenditur), Caesar’s language is not close enough to either Plutarch version to indicate what was the original phrasing in (presumably) the Polliosource.

⁴⁵ Cf. Frazier 1988b for the use of the present tense in citations.

⁴⁶ Appian oddly says that Caesar makes this criticism ‘in his letters’, ἐν ται̑ς ἐπιστολαι̑ς: there may be some confusion here arising from the phrasing, whatever it was, in the source. At Caes. 22.2 Plutarch quotes the Commentarii again, there BG 4.11–13, and once more Appian quotes the same passage (Celt. fr. 18): the pattern of shared quotation confirms that both are using their usual shared Pollio-source. In that passage Appian has Caesar writing , Plutarch ἐν ται̑ς ἐφημερίσι – a way of rendering Commentarii which became fairly regular, and is even found as the title of Caesar’s works in some of their manuscripts: Rüpke 1992, 202.

⁴⁷ The imperfect ἔλεγε might indicate a repeated remark, as e.g. with Pericles apophthegm nr. 1, Epaminondas nr. 18, Cato Maior nr. 23; but there are many cases when it clearly does not, e.g. Cyrus the younger nr. 1, Philip nr. 11, Alexander nr. 23, Antigonus nr. 16, Scipio Aemilianus nr. 20.

⁴⁸ Cf. Volkmann 1869, 213 and 219, commenting on the way in which Ap. frequently turns anecdotes – ‘strategemata et πολιτεύματα’ – into apophthegms. Volkmann was appalled; we might admire.

⁴⁹ e.g. Cicero apophthegm 5 has two instances of oratio recta where the equivalent in the Life only has one, but 6 is the other way round; 14 puts into oratio recta a remark which the Life left in indirect speech. There are other differences, for instances the tendency of Apophthegmata to miss out the names of incidental characters whereas the Life includes them (cf. e.g. nrs. 17–18); but that is generically natural (the Apophthegmata deal in what is generalizable rather than specific), and does not imply any difference in stylistic level.

⁵⁰ Cf. Quint. 6.3.5, where ... Tiro aut alius, quisquis fuit, qui tris hac de re libros edidit (‘Tiro, or whoever else it was who published three books on this matter’) need not be taken as indicating genuine doubt about authorship: Quintilian archly affects doubt that anyone so close to Cicero could have done him such a disservice. On Plutarch’s use of Tiro see p. 16; Moles 1988, 29 and 155 on Cic. 5.6.

⁵¹ pp. 16–17.

⁵² The prime case here is the use of the Second Philippic in Antony: pp. 17–18, 94–5,and Pelling 1988, esp. 26–7, 30, 137–40, 169–71.

⁵³ See pp. 67–8.

⁵⁴ Something which worried Volkmann (1869, 213, 219–21): cf. Fuhrmann 1988, 4–5.

⁵⁵ Though it is undeniably strange that Appian does not tell this story, e.g. at BC 2.82. (It is less strange that Plutarch omits it in Pompey 72–3: his narrative focus has by then moved to Pompey and to his camp.) A hypersceptic might wonder if Pollio mentioned this not in his History, but in, say, a separate letter; that suggestion won some approval in discussion at the Leuven conference (n. 14), and Tasos Nikolaidis suggests to me that it may go back to oral tradition. I think it more likely that Appian left the item out for his own reasons, not wishing to give space at this point to Caesar’s justifications. If so, that is most interesting for Appian’s technique and ‘ideology’: I intend to discuss this elsewhere.

⁵⁶ Cf. Fuhrmann 1988, 305.

⁵⁷ Broughton, MRR ii. 207.

⁵⁸ Some other morally discreditable stories which are dropped: several cases in Sulla when he is ungracious, whether to Athens (Sulla 6.5, 24.9) or to Mithridates (24.2) or to Halieis (26.7); several cases concerning Lucullus’ elderly self-indulgence, Lucullus 39.5, 39.6, 41.2, and especially 41.3 (‘Lucullus is dining with Lucullus’). Another case of mild bowdlerizing is Caesar apophthegm nr. 8 = Caesar 35, where some of the earlier parts of the story are cut away in the Apophthegmata, especially Caesar’s remark that there is a time for laws and a time for weapons. One case of alteration rather than omission, which may be explained either by moral considerations or by generalizability: Pompey apophthegm 7, Pompey burning Sertorius’ correspondence unread even though it implicated leading figures in Rome. In the Apophthegmata his motive is to give those figures ‘a chance to repent and become better men’, in the Life a more hard-nosed ‘fearing that he might stir up more violence than the wars that had ceased’.

⁵⁹ For instance Pompey apophthegm nr. 7 is out of order (it relates to Sertorius, but is put after the story of Pompey’s return from Spain in nr. 6), The Life’s order must be the chronological one, but the Apophthegmata order makes good artistic sense: nrs 4, 5, and 6 are all dealing with Pompey’s stupendously precocious rise. Caesar apophthegms 4–5, first the Alexander anecdote, then the ‘tiny city’ (πολίχνίον) in the Alps, are put in reverse order in Caesar 11: in the Life the Alexander anecdote leads well into his ‘thrusting on the outer sea’ in 12.1 but in Ap. the πολίχνιον leads just as well into his return to Rome. Then apophthegm nr. 7, the rather weak ‘risky and great enterprises must be done, not thought about’, is dropped from the Life, but in Ap. it gives a neat transition to nr. 8, the Rubicon. Aemilius apophthegms nrs. 6–8 correspond to Aem. 28.9, 34.3–4, 28.11–12, but the Ap. ordering again gives elegant transitions: in 6 Perseus is defeated, in 7 he is a prisoner, in 8 Aemilius is distributing the proceeds.

⁶⁰ To the cases in n. 49 could be added e.g. Pompey apophthegm nr. 2 = 10.14; Flamininus nr. 2 = 13, where the Life does write up but it is not clear that the Ap. version is lower-level.

⁶¹ To take some examples from Pompey: apophthegm nr. 5a keeps the narrative focus on Servilius throughout, whereas the Life switches the emphasis to begin with the soldiers, and a neat link from the preceding item of the procession. nr. 6 also has shift of emphasis: in Ap. the climax is, appropriately, the direct speech itself, in the Life the onlookers’ reaction to it. nr. 11 is similar. nr. 15 deals with his death: in Ap. it reads as if he is killed immediately after expressing his fears, whereas in the Life there is a whole page of extra action, including some further direct speech of Pompey himself (to Septimius: ‘have you not served with me?’). The extra pathos is right for the Life, but the speed of the Ap. version is also appropriate to give a resonant conclusion to the Pompey collection.

⁶² Van der Stockt (forthcoming) makes some initial steps to extending the Moralia approach to the Lives, agreeing that when similar material occurs in Lives and in Moralia the same ὑπομνήματα may be used, and may influence presentation. In that paper he raises the possibility that, as P. researched for the Lives, he ‘added concrete examples taken from his research for his Lives’: in that case, presumably added to his notes on moral topics. (How, incidentally, in technological terms? Might this be a case where a new paragraph might be cut and pasted into a pre-existing roll?) In any case, this is a slightly different picture from that of the Lives’ ὑπóμνημα which I am developing here, but I do not see why they should be incompatible.

⁶³ Namely, that On the Fortune of the Romans must have been written after the ὑπόμνημα for Caesar, and after Plutarch had first encountered the Pollio-source (who clearly included this story: cf. App. BC 2.57.235–8). This would put it late, well into the period when Plutarch was writing the Parallel Lives.

⁶⁴ Ziegler 1949, 228 interestingly found the Laconica even less impressive than the Regum et Imperatorum: ‘Der Charakter als bloEe, ungesichtete, nicht irgendwie durchgearbeitete Materialsammlung tritt in dieser Schrift [sc. Laconica ]noch starker hervor als in den Reg. et imp. ap. .’ If he is right, that could be used to support a suggestion that the apophthegms are less worked over in the Lac. than in Reg. et Imp., and may reflect a more primitive stage of production. Stadter too thinks that Reg. et Imp. is ‘modestly embellished’, more so than Lac.; in discussion at Leuven he put considerable weight on this point.

⁶⁵ Cf. p. 270 on the significance of Sosius Senecio as dedicatee of the Parallel Lives and of Polycrates as dedicatee of Aratus.