things that happened” (poeticis magis decora fabulis quam incorruptis rerum gestarum
monumentis, Pref.6). Here he is setting up a strategy of skirmishing with opposing
genres that will carry on strongly into the first book.^54
One of the factors that make this topic so difficult is that ancient writers intelli-
gently anticipate such moderns as Hayden White by systematically running to-
gether “content” and “form.”^55 Mu'qo"and fabula(“myth” and “fable”) are terms
that apply both to subject matter and to genre, so that any ancient discussion of
these topics keeps sliding — productively, but to the eyes of many moderns, con-
fusingly — from one category to another.^56 If you are distinguishing between “his-
tory” and mu'qo"or fabula,you are distinguishing not just between the historically
verifiable and the fabulous, or nonverifiable, but also between what belongs in his-
toriography and what does not: you are negotiating a generic as well as an episte-
mological boundary. Censorinus’s account of Varro’s divisions of time is once
again highly revealing. The second epoch is called “mythical” in terms of what we
might call “form,” “because in it many fabulous things are reported” (quia in eo
multa fabulosa referuntur); but he describes the historical period in terms of both
content and form, without tilting the balance definitively either way: this period is
“historical,” he says, “because the things that were donein that period are contained
in true histories” (quia res in eo gestae ueris historiis continentur, DN21.1).
Historians, then, fenced offmyth from their work in various ways, and one of
their reasons for doing this — or perhaps we should say one of their strategies for
doing this — was based on the idea that the times of myth were beyond the pale in
terms of chronology. For the historians there is no chronology of myth, no set of
interlocking synchronistic data that make a system; there is no “canon,” as they put
it.^57 “Mythical time had neither depth nor breadth,” says Veyne: “One might as
well ask whether the adventures of Tom Thumb took place before or after Cinder-
ella’s ball.”^58 This is — rather typically — overstated: Veyne himself immediately
concedes that the heroes had genealogies, which give both depth and breadth. At
least for the historical tradition, however, his large statement is broadly true.
Diodorus Siculus states the principle very explicitly, saying that he “cannot se-
curely divide up the times before the Trojan War because of the fact that no reli-
able chronological system has been transmitted” (tou;" me;n pro; tw'n Trwikw'n ouj
diorizovmeqa bebaivw" dia; to; mhde;n paravphgma pareilhfevnai peri; touvtwn
pisteuovmenon, 1.5.1).^59 The fundamental attitude is there from the start of histori-
ography. Finley is right to claim that Herodotus already had a chronological
scheme that he “refused to ruin... by incorporating the mythical events”; he
“made no effort to assign dates to the undatable myths.”^60 Sellar and Yeatman
Dividing Up the Past. 79