The first-century-AD philosopher Onasander (Strategikos 1.13, 15) comments that a general must be a good and effective speaker in order to encourage his men before battle, and ancient historians on occasion have commanders address large audiences of soldiers in an age when there was no artificial amplification of the human voice. In 1993, and reiterated in 1998, Mogens Hansen argued that the extensive pre-battle ‘speech’ found in our sources was, if given at all, only ‘a few apophthegms that could be shouted by the general as he traversed the line, or a speech made to the officers only who passed it on to the soldiers’. He accepts that a general often addressed his troops in camp or in parade formation; his objection solely relates to those speeches ostensibly given immediately before battle to troops already arrayed for combat. For Hansen, ‘a brief exhortation of the troops, unit by unit, while the commander walked along the front is probably all the historical reality [one] may expect to find behind the fully-fledged speeches reported by the historians’, these being a ‘literary composition and not the historian's report of a speech which had actually been made’. In Hanson's view, unless the harangues were given to troops in camp, or to individual units in battle formation, or, perhaps, to small military forces, these speeches could not have been given as reported. With respect to the last possibility, Hansen only rather reluctantly acknowledges that Thrasybulus' speech delivered to approximately 1000 Athenian insurgents in 404 BC, and reported by Xenophon (Hell. 2.4.10–12), ‘is possibly right’.