Skip to main content

Since the beginning of the western tradition of historiography, the idea of “eyewitness history” has held a central importance. Herodotus at various points emphasizes his autopsy – not of events, obviously, since those he narrates took place at an earlier time, but of places and objects (Darbo-Peschanski 1987). Thucydides is able to position himself closer to his subject, given the contemporary events he writes about, and he emphasizes this fact in his opening sentence (1.1). Despite this difference, both historians utilize a narrative style which elides any conceptual distance between the reader and the action: they write, “Such and such people did X,” without constantly reminding their audience of the source(s) of their knowledge for these events. For both authors, this represents a narrative fiction: even Thucydides makes only one explicit claim to witnessing an event himself (the loss of Amphipolis under his watch, 4.104-7; see Dewald 2006).

That fiction allows historiography to take the form it does, of a generally continuous story told by a generally omniscient narrator (de Jong 2013). But the historian can attempt to get the reader closer to the action, as it were, in various ways. One method of creating a more vivid sense of “being there” in ancient historiography was direct speech (a technique borrowed from the epic tradition). In this paper, I propose to examine a subspecies of this device for putting the reader on the scene: the brief dialogue, whether in direct or indirect speech (or a mixture).

Herodotus provides numerous potential examples; here I will focus on two of the more striking ones, which foretell the outcome of major battles and include named sources: the conversation of Demaratus and Dicaeus at Eleusis upon seeing a portent on the eve of Salamis (8.65); and the banquet at Thebes during which Thersander of Orchomenos hears from his Persian couch-companion that doom awaits them (9.16). In both cases, Herodotus reports conversations whose interlocutors were initially under instructions not to reveal their contents. While his named sources may in part reflect a concern for credibility, they also place him (and his reader/listener) at one or two removes from the action.

I will compare these with an essentially unique passage in Thucydides, the brief, rapid dialogue between an Ambraciot herald and the victorious Acarnanians after the disaster suffered by the Ambraciots in northwest Greece (3.113; see Stahl 2003, 129-58). Unlike the much longer and more famous Melian Dialogue in Book 5, the direct speech of this chapter is not introduced as such, but develops out of an initial exchange in indirect speech. The short scene is followed by the narrator’s claim that the Ambraciot losses were the worst to befall any single city within such a short time during the whole of the war. Although it is not a private conversation, Thucydides’ decision to “record” it helps mark the event as tragic in nature (Lapini 1991).

In what ways do these brief dialogues differ from basic historical narrative? What is the effect of the historian’s choice to present these scenes in dialogue/conversation form, rather than his authorial voice? What do these maneuvers reveal about the nature and expectations of ancient historical writing?