SOURCE: "Narrative Discourse in Thucydides," in The Greek Historians: Literature and History—Papers Presented to A. E. Raubitschek, Anma Libri, 1985, pp. 1-17.

[In the following excerpt, Connor argues that the predominant critical examination of Thucydides as a political scientist and a historical scientist neglects the strength of his narrative technique—and consequently misses "the pleasure of reading" his History.]

There are today many signs of a sea change in our understanding of the relationship between literature and history and hence in our understanding of the historians of the past and of historical writing in the present. Lawrence Stone drew attention to some of these signs a few years ago in an essay entitled "The Revival of Narrative" [Past and Present 85 (Nov. 1979)]. Stone argued that there was a "noticeable shift of content, method and style among a very tiny, but disproportionately prominent, section of the historical profession." The change was from what he called "structural" history to "narrative" history, that is to historical writing that is descriptive rather than analytical and whose central focus is on man and not on circumstances. He was not referring, of course, to the writing of antiquarians, or annalists, but to the shift from quantitative or "scientific" history toward another set of questions, especially those about the role of power and of the individual in history and also to the effort "to discover what was going on inside people's heads in the past, and what it was like to live in the past, questions which inevitably lead back to the use of narrative." Eric Hobsbawn in reply challenged many of Stone's conclusions but conceded "there is evidence that the old historical avant-garde no longer rejects, despises and combats the old fashioned 'history of events' or even biographical history, as some of it used to" [Past and Present 86 (1980)].

I suspect there is more of a change than Hobsbawn, and perhaps even than Stone, admitted. The signs multiply that a major change is under way, one with important implications for all who are concerned with history. We are witnessing, I believe, not just a resurgent academic appreciation of some traditional techniques of historical scholarship, nor a recognition that narrative theory affects historical writing as much as it does the novel, but a rethinking of some of the fundamental modes whereby our culture relates to the past. The issue, if I am correct, is not just the revival of narrative but a new and more experiential mode of historical understanding. This large claim is not to be argued in short compass. My aim here is more modest, to look at one author from antiquity, writing in what I believe was a period similar in one respect to our own— its rethinking of its relationship to the past and of the problem of writing about the past. Studying narrative discourse in Thucydides will not by itself clarify what is happening in our own culture, but it may contain a few hidden analogies to some of the changes going on right now.

We have now almost stopped talking about Thucydides as a "scientific historian." That analogy, borrowed from the enthusiasms of an earlier generation, had a long life in Thucydidean studies and caused much belief. It encouraged the notion that Thucydides was not so much a writer as a proto-political scientist and sent readers scurrying about to find in his work "laws" comparable to those found by natural scientists. Much attention was thus paid to passages that generalized about human nature or that expounded the so-called Law of the Stronger. Little attention was paid to the fact that these passages are almost always found in the speeches of the work, and that the structure of the debates and their setting within the narrative often subvert or modify the generalizations advanced by individual speakers. The search for the laws of a political science in Thucydides made him into a hard line Cold Warrior, teaching the lesson of the tough-minded pursuit of self interest and national interest. The attempt to make Thucydides into a "scientific historian," in other words, narrowed and distorted our understanding of the literary richness of the work. Still, the analogy did help us become aware of certain important features of the text, even if it did not go far toward explaining them. It drew attention to the restraint and austerity of Thucydides, the comparative infrequency of authorial interventions, and the avoidance of explicit judgments and evaluations. To be sure, it also tempted us to mistake these features for an attempt to write a purely "objective" or "value-free" history and to neglect the frequent and powerful indications of implicit value judgments throughout the Histories. The analogy to scientific history, in other words, did what analogies usually do—it opened our eyes to some features of the text and obscured some other features, equally important for a full and balanced appreciation.

Now that we have swung away from the view that Thucydides was a cold and detached observer and have begun to emphasize the elements of feeling, involvement, judgment, and pathos in his work, it is easy to be scornful of the old belief in a "scientific" Thucydides. But we learned a lot in that school, including the great debt Thucydides owed to the intellectual revolution of the mid fifth century B.C., especially to Hippocratic medicine and the early Sophists. Thucydides' work, we agree, was profoundly influenced—not molded or determined—but influenced by the thinking about myth, persuasion, and psychology that was going on during his childhood and youth. Out of that revolution Thucydides drew some of the elements that were to prove most important for his work. He combined a realistic, tough-minded psychology, the Hippoeratics' insistence on careful testing of observations and reported facts, and the argumentative techniques of the Sophists into a powerful machine for historical analysis.

We can best see this engine at work in the opening chapters of the Histories, the so-called "Archaeology," where it is applied to the legends of early Greece. If we look closely, we note a surprising contrast between Thucydides and the supposedly more credulous Herodotus. Thucydides turns out to be willing to accept a considerable amount of this legendary material, but only after it has come through his analytical engine. Along the way the variants in the stories are studied, the alleged motives of the actors are tested against his "modern" psychology and an interpretation is presented that is grounded in analogies from primitive cultures and arguments from probability.

Consider one example. Herodotus begins his history of the Persian wars by telling some legends about early hostilities between Greeks and barbarians. He includes two versions of the story of Paris' abduction of Helen, but then dismisses both: "Which of these two accounts is true I shall not trouble to decide. I shall proceed at once to point out the person who first within my own knowledge commenced aggressions on the Greeks, after which I shall go forward with my history …" (Herodotus 1.5, trans. G. Rawlinson). Soon we are studying the expansions of the Persian empire in the sixth century B.C.

Thucydides, by contrast, refines and then accepts legends about early Greece. The opening of his history, the "Archaeology," accepts the reality of the Greek expedition against Troy. But he drastically reinterprets traditional legends: "Agamemnon," he says, "seems to me to have assembled his expedition not so much because of the oaths which Tyndareus imposed upon the suitors of Helen [that is, that they should assist the successful suitor if anyone ever abducted Helen] but because he was the most powerful man of his day" (1.9.1). He leaves no room here for story-telling about the power of oaths or the chivalric loyalty of unsuccessful suitors. Power counts and Agamemnon had it; naturally then, others followed when he gave the order. If we look through the Histories we find that Thucydides accepts a surprising amount of legendary material but accepts it only after his new historical method has separated plausible versions from myth, sentimentality and downright falsehood.

This method, Thucydides' new historical engine, is one of the boldest and most powerful inventions of the intellectual revolution of the fifth century. We understand it and appreciate it thanks in large part to the phase of our own past that emphasized the "scientific" nature of Thucydides' work. But that emphasis did little to help us understand how this historical method functions within the text. For that we must turn to the aspect of Thucydides that has attracted so much attention in recent years—the nature of narrative discourse. Much interesting work has been or is being done in this rich field, but I shall concentrate on a very specific question, and a very difficult one: Why do we believe Thucydides' account? What makes him seem so persuasive and compelling? To phrase the question in this way is not to imply that all historians believe Thucydides all the time—far from it. But those critics who have challenged Thucydides most sharply will be the first to point out the extraordinary hold he has upon our thinking about the Peloponnesian War. Even when his account has received repeated and serious criticism, historians and laymen alike are reluctant to repudiate it. To be sure we try to utilize all our sources about antiquity, especially those by contemporary writers, but Thucydides enjoys, rightly or wrongly, an esteem not accorded to Ctesias, Xenophon; Appian, Suetonius, or even Herodotus, Polybius, Livy and Tacitus. Why is this? The reason is not that Thucydides' account has been tested against a large number of independently verifiable facts and found consistently reliable. Only rarely can his work be compared to contemporary documents, and when it is compared, as when we have an inscription, there is almost always a problem. The problems do not refute Thucydides; we simply lack solid, independent verification. It is then something else that causes the intensity of belief engendered by Thucydides.

What is this something? Surely it is in part the recognition that Thucydides, whatever his biases and faults, is a highly intelligent observer. But how do we know that? And how can we test that impression? There is no sufficient outside authority to which we can appeal. We have only the words of the text to rely upon. In other words, the narrative discourse of Thucydides itself establishes the authority of the writer and persuades us to listen with respect, if not total assent.

It achieves this hold, moreover, without using many of the conventions of scholarly history. Obviously no one would expect to find in his work the apparatus of modern historical research, but the contrast between Thucydides and Herodotus indicates how rarely Thucydides uses the devices by which Herodotus presented to his readers the problems of finding out about the past. Herodotus will commonly identify the places where he finds a serious difficulty; he will report alternative versions or views; he will cite the consideration that leads him to prefer one version to another and he will state his conclusion in language that expresses the degree of confidence he feels. He may make mistakes of fact or logic; he may even be quite silly, but the problematic of history is always before our eyes. As [R. W.] Macan said in his appreciation of Herodotus in the Cambridge Ancient History, "Where there is a variant, he will not suppress alternatives, or impose his own judgment upon posterity. Even when his own mind is made up, he will allow his informants, and his public, the benefit of the doubt." The historian and his reader are colleagues, sharing the problems and engaged in dialogue about their solution.

Thucydides' practice is quite the opposite. Through " most of his work, he avoids discussing the problems of history and presents a finished product. As Macan says, "The results of his method, which is to extract for his readers, to all generations, a clear and chronologized narrative, the precise sources of which are seldom even indicated, must be taken or left on his authority, and on his authority alone." Reader and author stand in a different relationship. They are not colleagues, but performer and audience, the writer who knows how to produce a polished work and the audience who appreciates its craftsmanship and reacts to its quality.

To Thucydides' detractors this is sufficient to condemn him for "brain-washing" or manipulation. We expect to be colleagues, especially if we are professional historians (not that Thucydides ever was), and feel cheated if we are not allowed to look over his shoulder at the reports and documents he is using. Thucydides' defenders, on the other hand, wax eloquent. Gomme, following Gilbert Murray, for example, wrote that Thucydides was "determined to do all the work himself and to present only the finished product to the public, as the artist does. Wren showed St. Paul's Cathedral to the world, not his plans for it; so does the painter his picture; so did Pheidias his sculpture" [The Greek Attitude to Poetry and History, 1954].

If we step aside for a moment from the speeches for the prosecution and the defense, we notice something that seems to me more important than praise or blame. Thucydides' avoidance or rejection of the conventions of historical argument make it all the more difficult to give a satisfactory answer to our original question: Why do we believe Thucydides? We do not believe him because he has identified and clarified the problems, cited his sources, gathered the evidence and established his conclusions with such plausibility that we are forced to assent. Perhaps we believe him for precisely the opposite reason—because he writes not as the scribes and Pharisees do, but with authority.

That authority derives, I believe, from three sources. To one I have already alluded: it is the demonstration of historical method in the "Archaeology." The opening twenty-three chapters are a short, highly selective inquiry into some aspects of the past and constitute an epideixis, a demonstration piece, showing what Thucydides' method can do. They constitute an implicit a fortiori argument. If Thucydides' powerful engine can extract such a compelling interpretation of the remote past, a fortiori it should be able to attain important results in interpreting and analyzing the recent past. That, I believe, is what Thucydides implies when, after the investigation of early Greece, he points out the difficulties of finding out about the remote past but goes on to affirm that anyone who accepts the approximations he has derived from the indicators (tekma ria) he has mentioned will not go astray (1.21.1). He then turns to the problems of reconstructing the events of the Peloponnesian War. The famous "programmatic" or "methodological chapter" (1.22) is not a comprehensive statement of his historical principles but an affirmation of difficulties overcome and hence of the enduring utility of his work.

The first source of Thucydides' authority then is the demonstration of his historical method. But once the engine has been displayed, it is locked up again. We may hear it rumbling away in the background somewhere; we are reminded of its existence from time to time. But we do not regularly see it collecting, analyzing, testing and selecting reports and data about events and turning them into finished historical narrative.

Thus for much of the work the historical method of Thucydides is out of sight, if not entirely out of mind. In these portions its effects are reinforced by two further sources of authority. One is Thucydidean "style," that formidable, overwhelming complexity that can shatter all the neat antilogies and balances of Greek and strain the language to its limit.

Once again it is important to ask the simple but fundamental question. Why is Thucydides' style so difficult? Is it, as Collingwood thought [in his The Idea of History, 1956], the result of Thucydides' bad conscience, his uneasiness at pretending to write history when he was really writing political science or theory? Or is the more conventional answer correct—that it is the result of the originality and subtlety of his ideas. Is there some gnostic message concealed in the complexity of his expression? If we look closely, we find, I believe, that neither of these answers is correct. The difficulties derive not from the author's psychic disquiet nor from hidden subtleties, but from a desire to affirm his respect for the complexity of historical events and human motives. In Thucydides we discover not an arcane philosophy but a style that replicates the intractability of historical experience. It assures the reader that the author will not oversimplify or reduce events to cliché, antithesis, or dogma.

This assurance is conveyed by, and much of the difficulty arises from, Thucydides' use of multiple viewpoints in narrating events. He will begin from one point of view, and switch, usually without warning or marker, to quite a different perspective. Often we end by viewing a single event from two or three different viewpoints. In the account of the third year of the war, for example, Thucydides tells of the consternation that swept through Athens when a Peloponnesian fleet appeared in the Saronic Gulf. The Peloponnesians had almost defeated the Athenian ships in the Corinthian Gulf. Then, at the end of the campaigning season, the Peloponnesian commanders decide to undertake one more operation. They will march their sailors overland, each carrying his oarlock and seat cushion, to Megara where forty ships are drawn up in their ally's dockyards. With these ships, they plan to make a surprise attack on the Piraeus.

Up to this point the narrative is straightforward, perfectly clear, even relatively easy Greek. Then follows a sentence of such contorted phraseology—not to mention its nine negatives in 35 words—that the critics have tried to emend it or delete it. Crawley's translation smooths out some of the difficulties but catches the main idea:

There was no fleet on the look out in the harbor [of the Piraeus] and no one had the least idea of the enemy attempting a surprise: while an open attack would, it was thought, never be deliberately ventured on, or if in contemplation would be speedily known at Athens. (2.93.3)

Crawley has added a crucial phrase to the Greek: "while an open attack would, it was thought, never be deliberately ventured upon …". Why did Crawley add this phrase? He recognized and marked for his reader what is implicit and hence a source of obscurity in the Greek: that Thucydides has shifted from reporting the attitude of the Peloponnesians to conveying the psychology of the Athenians. In the next sentence, when Thucydides shifts back to the plans of the Peloponnesians, the reader understands, thanks to this contorted sentence, both the Peloponnesians' feeling that their original plan was terribly risky, and the ironic fact that precisely because of that risk it might well have worked:

… arriving at night and launching their vessels from Nisaea, they sailed, not to Piraeus as they originally intended, being afraid of the risk, besides which there was some talk of a wind having stopped them, but to the point of Salamis …

The rest of the account of the operation continues this alternation between Peloponnesian and Athenian viewpoints. The effect is consistently ironic: by the time we hear of the Peloponnesian decision to abandon their original plan we know that from the Athenian point of view it might have worked; by the time we hear of the panic in Athens, we know that from the Peloponnesian point of view the plan was too risky to carry through. The irony is characteristic of Thucydides and so are the rapid changes of viewpoint, a major component of his style and an important contributor to this second source of his authority.

The richness of Thucydides' account comes sharply into focus if we compare this passage to the smooth and nicely balanced version of the episode supplied by Diodorus of Sicily. Diodorus reduces all to a single viewpoint, that of one Peloponnesian commander; he omits the vivid detail of the march overland with each rower carrying his oarlock and seat cushion, and by failing to mention the growing fears of the Peloponnesian commanders leaves his reader without explanation for the outcome of the operation, an attack on Salamis not Piraeus:

In this year, Cnemus, the Lacedaemonian admiral, who was inactive in Corinth, decided to seize the Piraeus. He had received information that no ships in the harbour had been put into the water for duty and no soldiers had been detailed to guard the port; for the Athenians, as he had learned, had become negligent about guarding it because they by no means expected any enemy would have the audacity to seize the place. Consequently Cnemus, launching forty triremes which had been hauled up on the beach at Megara, sailed by night to Salamis, and falling unexpectedly on the fortrees on Salamis called Boudorium, he towed away three ships and overran the entire island. (Diodorus 12.49.2-3, trans. C.H. Oldfather)

No one needs corroboration from contemporary documents or even a knowledge of the chronological and historiographical relationship between Thucydides and Diodorus to know which of these two accounts to prefer.

A third source of authority, however, may prove even more important. To call this the "experiential" or "participatory" aspect of Thucydides' work would be cumbersome, but the terms for all their awkwardness call attention to a feature often neglected in the work. We do not usually think of Thucydides as a writer who keeps drawing his readers into the narrative of events until they feel they are themselves present, actually experiencing them. But Thucydides achieves this implication of the reader to an extraordinary degree. We do not often let ourselves be caught up in the vicarious experience of the actions he describes, but we should. For every minute we spend searching for laws or theory or gnostic insights, we might well allot equal time and attention to Thucydides' ability to recreate events and moods.

To achieve this end Thucydides has many techniques, chief among them the dramatic interplay between abstraction and sudden flashes of vividness. His style aspires to a level of generality that brings out the similarity of one episode to another. It often verges on the formulaic. But darting through it are words, phrases, sometimes whole episodes whose extraordinary vividness creates the illusion that we are ourselves present, witnessing events. In the passage we have just examined, for example, Thucydides notes that each sailor carried with him on the march his oarlock and cushion. In recounting the initial attack on Plataea, he focusses on the spear point jammed into the lock of the gate to prevent escape (2.4.3). At the siege of Plataea we observe the careful planning of the escape, the counting of the bricks (3.20.3), the removal of the sandal from the right foot (3.22.2), the armament and name of the leader of those who first climbed the ladders  (3.22.3). In the account of the battle near Naupactus we watch an Athenian ship wheel rapidly about an anchored merchantman and ram its Leucadian pursuer amidship and sink it (2.91.3); on Sphacteria we experience the dust, the headgear pierced by arrows, the broken spears (4.34.3). Every attentive reader of Thucydides could expand the list and note the close bond between visual detail and the mood of the scene and the feelings of the participants. Vision in Thucydides is the privileged sense, most commonly invoked and most directly linked to the emotions.

Yet modern critics have said far less about the vivid side of Thucydides' style than about the complexities of his moral and political beliefs. In this respect, for all their obvious faults, the ancient critics are closer to the mark. Plutarch, for example, stressed Thucydides' use of enargeia, vividness, in his comments on Thucydides.

The most effective historian is the one who makes his narrative like a painting by giving a visual quality to the sufferings and characters. Thucydides certainly always strives after this vividness in his writing, eagerly trying to transform his reader into a spectator and to let the sufferings that were so dazzling and upsetting to those who beheld them have a similar effect on those who read about them. (Plutarch On the Fame of the Athenians, ch. 3 [Moralia 347A])

Hobbes [in his Preface to a translation of Thucydides, 1629] made a similar observation, but set it in a more provocative context. He praised Thucydides as "the most politic historiographer that ever writ. The reason wherefore I take to be this … he maketh his auditor a spectator." And how did Thucydides attain this most politic result? "The narrative," Hobbes says, "doth secretly instruct the reader and more effectively than can be done by precept." The political lessons and the utility of the Histories, in other words, derive not from Thucydides' explicit comments or implicit theorizing, but from the reader's own involvement in the work.

Once we recognize this, many features of the work become far more intelligible. We can worry less about the author's hidden theories and more about our own reactions to the text. And as we become more active participants in the events described, we can dispense with the narrator's guidance and explicit comments. The narrator can become self-effacing, speak in the third person, intervene only rarely with his own judgments and evaluations and let the reader do, in Henry James' phrase, "quite half the labor." He may pretend that the war itself is the narrator which reveals its own greatness to those who scrutinize events closely enough (1.21.2).

We recognize, of course, that this is pretense, or a game in which reader and author engage. History is not chronicle, and Thucydides was certainly not a modest registering machine duly recording each event in colorless exactitude. He is present, selecting, shaping, coloring, at every episode, every phrase, every subscript. We ancient historians and philologists have reached the age of discretion at which we can reasonably be expected to recognize that the shaping of the text by the author, even as the author pretends to remove himself from his story, is not a bad habit, unprofessional conduct on the part of the writer. It is one way a writer can accomplish an essential part of his purpose—the involvement of the reader in the events and the activation of the reader's own evaluative capacities.

The illusion created by Thucydides, then, is one of immediate presence, of our own participation in the events described. That illusion, in Thucydides' view at least, excludes another one, and a very near and dear one to our historical hearts. Thucydides avoids letting his reader think that he is in the archive selecting the documents, or in the author's study participating in the choice of one version over another. We are not his colleague. Instead, the documents have been gathered, the informants interrogated, the selection of alternative versions has been completed. This much of the work is done and the reader is presented with the final product and asked to respond to it.

In this respect Thucydides' practice contrasts very sharply with that of Herodotus. To many modern readers Herodotus is much more congenial. As we read his work we are constantly reminded of the difficulties he encountered in assembling and shaping his material— the legends and biased accounts presented to him as fact, the leg-pulling, the gaps and the polemics he encountered. We are at his side, sharing the decisions with him, and enjoying the process. By contrast Thucydides may seem to us, as he does to Truesdale Brown [in Historia 31 (1982)], to underestimate our intelligence:

… while breaking new ground in his scrupulous use of sources, he underestimated the intelligence of his readers. Having arrived at his own conclusions by a critical examination of the evidence he does not share the materials he rejects with the reader. Herodotus was less critical … he could not bear to omit anything which he felt might appeal to his readers.

Brown's comments call attention to a major contrast between the two writers. But his explanation leaves out of sight the possibility that there are different principles at work. Thucydides imposes a different division of labor between author and reader. The historian's job is to investigate, compile, select, edit and present. The reader's half, the greater half, is to react, to assess, and thereby to learn.

This feature of Thucydides' technique has several important consequences for our understanding of the Histories, two of which call for special comment. The first concerns the ease with which description of mood is confused with statement of fact. The second returns to the major topic of this paper, the establishment of Thucydides' authority.

First, mood. One of Thucydides' goals, we have suggested, was to create in his reader the illusion that he is himself present at events. Sometimes this goal leads to descriptions not of the event itself but of the reactions and feelings of those who were present at the event. The most famous of these passages is the description of the great naval battle in the harbor at Syracuse. The main portion of the description of the battle contains passages such as this:

… while the naval battle was hanging in the balance, the land army of each side experienced a great conflict and convergence of reactions. The group from the immediate area were eager to win, with a view to even greater glory; those who had invaded feared lest they should experience even worse than their present state. Since for the Athenians everything depended on the ships, their fear of the future was beyond any comparison and thanks to the uncertainty of the naval battle, uncertain too had to be the vision of it from the land. Since one could see only a small portion of the action, nor did all look at the same spot, those who saw in one engagement their own forces succeeding were encouraged and would turn to invocations of the gods not to deprive them of their safety. But those who looked at a defeat raised a ritual lament even while they were shouting and from the sight of what had happened lost their spirit even more than those who were in the engagement.… (7.71.1-3)

The passage is an excellent example of several Thucydidean techniques—shifting viewpoints, the emphasis on vision, the creation of mood. But William Scott Ferguson deplored this approach:

… Thucydides fails even to suggest the factors that determined the outcome. Instead, he dwells on certain typical incidents in the confused fighting that followed, and then turns our attention to the spectators on the shore, and leaves us to infer the manifold vicissitudes of the protracted struggle from the agony of fear, joy, anxiety.…

True enough. Thucydides' concern, however, was not to recover the tactics, such as they were, of this confused battle, but to record the changes in the morale of the Athenians, the crucial factor in the next stage of the operations. We have in this passage another kind of enargeia, a vividness not of precise details but of mood.

For the reader passages such as this pose special problems. One can, for example, mistake mood for fact. Again, Thucydides does not always stop to distinguish and to mark important differences. Sometimes it is not entirely clear whether he is telling us what the situation was or how it seemed to contemporary observers. At the beginning of book eight, for example, Thucydides discusses the situation in Athens when news came of the loss of the expedition in Sicily:

All things on all sides grieved them and there surrounded them in this situation fear and dismay of the very greatest sort. For since they had lost, both individually and as a city, many heavy-armed soldiers and cavalry and crack troops of a quality the match of which they did not see to be available, they were depressed. Likewise since they did not see sufficient ships in their dockyards nor money in their treasury, nor crews for their ships, they despaired of any salvation in the present situation. They believed that the enemies from Sicily would immediately sail with their fleets against the Piraeus, especially since they had conquered so decisively, and that their own more immediate enemies at that very time had made double efforts in full force and would bear down upon them from land and sea and that their own allies would revolt and join them. (8.1.2-3)

The passage tells us how the Athenians looked upon their situation in the bleak moments when the news about Sicily arrived. It is full of descriptions of feelings—how they saw things, what they believed would happen. But many excellent commentators take the passage as Thucydides' assertion of facts, and then point out, using evidence from other passages in Thucydides, that the situation was by no means as hopeless as this passage suggests. Meiggs, for example [in The Athenian Empire (1972)], writes "the empire, meanwhile, according to Thucydides, threatened collapse as the allies competed fiercely to be the first to revolt, now that Athens' power was broken. His detailed narrative does not fully bear out this gloomy analysis." And Andrewes in the new Oxford commentary on book eight (p.6), noting that Athenian despair about their navy, concludes, "clearly the decree of 431 (ii.24.2) had not been maintained to keep a reserve of a hundred triremes, the best of each year, in readiness with their trierarchs." Perhaps not, but as Andrewes points out, their despair about their finances makes no mention of the reserve fund of one thousand talents also established in 431 and still available for use in the post-Sicilian emergency (8.15.1). Thucydides says that the Athenians did not see the resources to deal with the present situation; he does not say there were no resources to see.

As we read on in book eight the facts gradually become clear and the mood of the Athenians gradually changes: there is a reserve fund; not all allies revolt; those that do revolt often act prematurely; Sparta is not effective in exploiting the situation and the Syracusans are not swift or decisive in their intervention. At Cynossema Athens wins a major victory. In the eighth book we trace an irregular movement from despair to growing confidence, from apparent defeat to the renewed efforts of the Athenians in a final, and even greater, struggle. If we understand Thucydides' emphasis on changing moods, this book becomes more intelligible and we are far less likely, here or elsewhere, to mistake description of moods for statements about facts.

In a second way too we can now better understand Thucydides' technique and why it has produced such intense conviction. At certain points the narrative creates in the reader the feeling of being directly present at an episode in the war. We are as far from the historians' study as we can possibly be; we are in the war itself. We see; we hear; we even know the plans and thoughts of the participants. The crucial elements are before us, not in pictorial fullness, as one might find in a Hellenistic historian, but through highly selective detail. As Lawrence Stone has said of Peter Brown, "The deliberate vagueness, the pictorial approach,… the concern for what was going on inside people's heads, are all characteristic of a fresh way of writing history." And like a pointilliste painting it draws us in, involves our minds in the process of creation, and wins our assent. Seeing is, after all, believing.

Those who have learned what history is from professional scholars and who know a good footnote when they see one, may find this a paradoxical conclusion. How can one believe a writer who, after a few opening chapters, simply bypasses the whole problematic of history and writes as if he knew precisely what went on in the war, and even in the participants' heads? If we had been weaned on Macaulay and Carlyle, things might look different. Those writers remind us that in History's house there are many rooms and many passageways. We lose something if we block off too many parts of the mansion or condemn too much of it too soon.

And what is it, precisely, we risk if we close the chambers Thucydides occupied? The loss of vicarious experience and of the sense of participating in a reality far different from our daily life, the very thing that makes the study of history so important for the growth of the mind and imagination. "But surely," one might object, "this vicarious experience can be obtained without sacrificing the constant gestures of respect for the problematic of history which are the marks of modern scholarly history." Perhaps, but one should not underestimate the difficulty. For many readers any pause over the problems of historical evidence and reconstruction, any worry about conflicting sources or assessments, any entry into the historian's study, shatters the illusion of participating in the past. The problematic of history impedes its experiential power. To talk about the problems of historical analysis imposes a chasm between reader and past event. Even when it produces conviction, a residue of doubt remains. For historical analysis is always based on the calculation of probabilities. We read the arguments and assent, but our language reminds us of the uncertainty. We say we are "almost one hundred percent sure" or that we are "halfway convinced" or that we have "found the preponderance of evidence" on one side of the matter.

The division of labor we have noted in Thucydides bypasses this problem. Thucydides may have his doubts and unresolved problems. But he keeps them to himself and lets the reader transcend them. The conviction which attends the reading of Thucydides' work is thus not related to the calculation of probabilities or the careful assessment of plausible solutions. We feel we have been there. The world Thucydides has described, the patterns of power and human conduct, are so consistent, so real, that we have no choice but to assent. We are moved by the greater logic that derives its power not from accumulated evidence or carefully constructed syllogisms but from the evocation of a coherent world. At length we feel, not that we have deduced the nature of that world, but that we have temporarily become part of it. Of course we then believe—not despite, but because of the fact that Thucydides does not write as the scribes and Pharisees do, but as one with authority.

What shall we then conclude about Thucydides' authority? Why do we believe him? We have seen three sources of it in the work, the first the powerful engine of historical analysis whose workings are best to be seen in the "Archaeology." The second is a style that affirms the author's respect for the complexity of historical events and that views the past from multiple perspectives. The third source, perhaps the most important, is the reader's feeling of experiencing the events described.

The sources do not all cohabit in blissful harmony. Indeed between the second and the third there is an inevitable tension—the rapid shifting of viewpoints risks a shattering of the experiential quality of the work. But Thucydides' style not only contributes to his authority but sustains the tension and transforms it into the uniquely powerful result we have all experienced in reading Thucydides. It accounts for something scholars of Thucydides experience but often fail to mention—Thucydides' appeal as a writer and the pleasure of reading him. If we concentrate too much on the scientist or philosopher, we lose sight of the vividness of his writing and the rich, demanding but rewarding experience of reading a great writer. Expecting profundity we can miss the color, the swift-paced action, the detail, the opportunity to see, to experience, to understand. To say this is not to make Thucydides into a simple writer, not to minimize the efforts he expects from his readers. But it is a reminder of the ability of this work to avoid the eventual emptiness of an exclusively analytical method and to resist with equal determination the tendency so evident in Hellenistic historiography to report anything that is sufficiently lurid and sensational. What makes Thucydides' work what it is—one of the unsurpassed and enduring achievements of prose narrative—is precisely this tension and interaction among different modes of narrative discourse.