The opening sentence of Thucydides’ work announces that he ξυνέγραψε the war between the Athenians and the Peloponnesians. His use of ξυγγράφειν, applied in the classical period to the composition of historical narratives, arguments, speeches, technical manuals, memoranda, draft proposals, architectural schedules, and the like, but not to drama, poetry, or fiction, reminds us that the study of Thucydides has one dimension which the study of (e.g.) Homer or Euripides lacks. In respect of any author we have to begin with the questions, ‘What did he say?’ and ‘What did he mean?’ The procedures designed to answer the first question are subsumed under ‘textual criticism’, those concerned with the second under ‘translation’ and ‘interpretation’. The division of labour is necessarily inexact, since difficulties in translation often make us ask, ‘Did he really write that?’, and, conversely, suspicion of the text or choice between variants can seldom claim to be rational unless the meaning is treated as the vital consideration. ‘Interpretation’, taking ‘What did he mean?’ beyond the point to which the translator has taken it, investigates the associations which words and ideas had for the writer and his audience, and it merges into the question, ‘Why did he write that, in that connection, at that time?’ In the case of a historian we can ask—indeed, we cannot help asking, unless we are sadly lacking in curiosity—the further, and separate question, ‘Is it true?’ If we ask, ‘Did Heracles really bring Alcestis back from the tomb?’ we are posing a question which does not extend beyond the intentions of Euripides; but as soon as our reading of Thuc. i. 139. I prompts us to ask, ‘Did the Spartans really say that if the Megarian Decree were annulled there would be no war?’, we realize the magnitude of the extra dimension which (whatever else it may lack) the study of a historian possesses.

Right at the start of Book II we are faced with the statement that the Thebans attacked Plataea ‘when the archonship of Pythodorus at Athens had two months still to run, in the sixth month after the battle at Potidea, and at the beginning of spring’. ‘Two months’ cannot be reconciled with ‘at the beginning of spring’, given the facts of the Athenian civil calendar, and ‘the sixth month’, though defensible, is not very easy to reconcile with the narrative of Book I. Many times in reading Thucydides, if we care whether what he says is true, or even coherent, we have to question, and sometimes reject, the text. Numerical words give the most trouble, for an obvious reason: ‘sixth’, ‘ninth’, ‘tenth’, etc. are all equally grammatical and may all make equally good sense until investigated more deeply than we can expect of a copyist; Polybius (xii. 4a) knew that. There are, however, difficulties of other kinds. Did Thucydides say of Gylippus in vi. 104. I ‘after negotiating with Thurii and renewing the citizenship given to his father’ or (a variant reading) ‘after negotiating with Thurii in accordance with the citizenship once given to his father’? It makes some difference to our notions about Greek grants of citizenship. And in vi. 88. 4 leaving of πολλοί or emending to ού πολλοί makes a difference to our picture of the balance of power between Syracuse and Athens.

Nearly four centuries after Thucydides wrote his famous characterization of stasis (iii. 82 f.), Dionysius of Halicarnassus (On Thucydides 29–33) subjected it to a severe stylistic criticism (words such as obscure’ and ‘contorted’ occur frequently, and some judgements on individual sentences express baffled exasperation). In particular, Dionysius confesses himself (32) unable to make any satisfactory sense of iii. 82. 7:

Dionysius does not command respect as a historian or as a critic of historiography; and we may even think him a captious critic of style when we have decided that the sentence quoted above means:

Thucydides refers in ii. 65. 12 and v. 26. 1, 3 f., to the total duration and ending of the Peloponnesian War (cf. i. 23. 1, its ‘great length’). His work, as we have it, breaks off abruptly while narrating the events of the autumn of 411, and it is clear that no continuation beyond that point attributable to Thucydides himself was known to the Greeks of the fourth century. It would therefore seem prima facie that Thucydides embarked on the writing of his work after the surrender of Athens in 404 but died before he could complete the task. He tells us in i. 1. 1 that he started to write a history of the war as soon as it broke out; but since the intention to write the history of a war which has only just begun is most naturally realized by making jottings as it proceeds, correcting and revising them as new information becomes available, and only when the war is over converting this material into a continuous readable narrative, i. 1. 1 is compatible with the hypothesis that the whole surviving work of Thucydides was written, in the order and in the form in which we have it, after 404. Apart from explicit references (mentioned above) to the end of the war, there are other passages which tend to support this hypothesis: in iv. 81. 2 a forward reference to the war after 413, in v. 1 a reference back to iii. 104. Then there are passages containing ‘only’ or a superlative: vii. 44. 1, ‘the only occasion during this war . . .’, vii. 87. 5, ‘the greatest achievement in this war’.

Utterances reported in direct speech abound in Herodotus: often conversations or isolated dicta, sometimes debates on major questions of policy or strategy, e.g. vii. 8-11. In Thucydides short utterances are very rare, but speeches play a very important role; the occasions range from ‘full-dress’ debates on momentous issues (e.g. i. 67–88) to the encouragement given by commanders to troops before battle (e.g. ii. 87–9).

Thucydides tells us (i. 22. 2 f.) his procedure in respect of what he calls . He claims not to have obtained information ‘from anyone who happened to be available’ but to have described events of which he himself was a witness and, for the rest, to have scrutinized his information with the utmost care. This was a laborious process, since accounts were often contradictory; memory has gaps, and a man’s picture of a past event is coloured by his loyalties (i. 22. 3). Documentation of the kind which a modern historian would expect to use was hardly available to Thucydides. Treaties were recorded in writing, and so were Athenian decrees and some balance-sheets; but it is questionable whether the letter written by Nicias in 414 (vii. 8, 10–15) was kept in an archive long enough for Thucydides to read it, and the figures given in iii. 87. 3 for cavalry and hoplite deaths in the plague were probably taken from a speech made by a general in the assembly. There is no evidence for the existence in the fifth century of written operation orders, situation reports, or war diaries, and it is to be presumed that what Thucydides has to tell us about operations is based on oral information from people who had taken part or had themselves received information from participants.

A historian who undertakes to tell his readers what happened is under no obligation to tell them also what could or should have happened, and they may prefer to make up their own minds about that. Thucydides is notably sparing of personal judgements on political and military choice, the significance of an item in a causal sequence, or moral implications. His fullest expression of opinion on a military decision is in vii. 42. 3, where he comments in parenthesis on the view taken by Demothenes of the situation which he found at Syracuse on arrival with the Athenian reinforcements. That we are concerned here with Thucydides’ own opinion, not simply with Demosthenes’, is clear from the finite tenses used throughout the parenthesis; compare viii. 96. 4, where Thucydides dwells on what would have happened if the Peloponnesian fleet in 411 had gone straight for the Piraeus, and contrast the parenthesis ού γἁρ ἃν τὁν ἒκπλουν έπιβουλεῦσαι (vii. 51. 1), in a statement of Syracusan plans. Given that when he wrote vii. 42. 3 Thucydides believed that Nicias and Lamachus ought to have embarked on the investment of Syracuse in the autumn of 415, it seems that he then thought their anxiety over lack of cavalry unjustified. Why did he not say so in vi. 70. 3–71. 2? Again, it is a fair inference from vii. 42. 3 that he then thought Lamachus’ proposal for immediate attack the best of the alternatives discussed by the Athenian generals at Rhegium (vi. 47–51). Perhaps—whatever the order in which he wrote the passages concerned—he changed his mind (cf. p. 15). The principle on which he selected points for comment on what might have been is not intelligible to us; it seems that having made his selection of data for presentation, as it were, in the indicative, he left it to us, for the most part, to construct our own unfulfilled conditionals. This is sometimes easy (e.g. vii. 50. 4).