Approaching Roman Religion 15The problem lies with determining details in the obvious absence of the books
themselves and the annoying presence of widely differing and wildly contradictory
testimonia (G. Rohde 1936). Specialist attempts to cut the theological Gordian knot
have produced solutions at once complementary and contradictory (Linderski 1985;
North 1998; Rüpke 2003). We can, however, assert certain things which bear directly
on Wissenschaftsgeschichte. First, the pontiffs were neither professional scholars nor
professional priests; thus whatever the books really contained, scholarly treatises they
did not. Second, factual information from the books did come into the hands of
scholars. Third, knowledge of the books’ contents was limited by considerations of
book production and literacy as well as the agreed-upon although not yet demon-
strated difficulties of moving the information via a pre-industrial transportation
system. At the one extreme, then, even if the books teemed with treatises and archival
religious information, the circulation of that information would be limited by phys-
ical factors. At the other extreme, there is no reason to think they teemed with such
information at all, in which case it made little difference whether or not they could
be circulated. In short, they resemble nothing more than a grandiose, possibly grand-
iloquent, version of the records which local temples kept throughout Italy and the
empire (MacMullen 1981: 11–12). This is not to minimize the importance of such
records; they could offer substantial libraries of religious information, a notable ex-
ample of which appears with the Serapeum of Alexandria and its religious education
activities in the late fourth centuryad (Watts 2006: 145 – 6, 189 – 90). But Alexandria
was of relatively easy access, while the libraries of small temples in lonely venues would
be far less accessible and, hence, their records’ influence would be rather less. Indeed,
the very existence of such libraries provides a further important reason for the Christians’
promiscuous destruction of temples in the later Roman empire. Thus on practical
grounds the pontifical books loom large, but on empirical grounds they loom larger
than they should, and this for scholarship of both ancient and modern times.
Further confounding factors begin with the third century ad. The “military
anarchy” between the death of Caracalla and the accession of Diocletian occasioned
an almost total dearth of contemporary scholarship apart from Dexippus, and a sharp
decline in epigraphic evidence. Cornelius Labeo, plausibly but not provably third-
century (HLL4.78), appears along with Censorinus (infra) as the only author with
provable interests in Roman religion. Labeo wrote on, among other topics, the Roman
calendar as well as the disciplina Etrusca, and possibly the Di Penates (Macr. Sat.
3.4.6), although this is disputed (Mastandrea 1979: 113 –19); his writings were notable
enough to earn the censure of Arnobius (2.15) and Augustine (Civ.2.11, 9.19) while
John Lydus, Servius, and Macrobius also utilized him; interestingly, Arnobius seems
to have known Varro only from handbooks while Augustine may well have had access,
at the least, to very substantial excerpts. Here we must not speculate freely. Varro’s
works remained very popular in late antiquity, Servius frequently cites them by title,
and thus Arnobius represents the exception while Augustine the rule. Contrariwise,
the works of Nigidius Figulus passed out of general interest after the second cen-
turyad, making it far more likely that later quotations attributed to him come from
handbooks. We can cautiously assert that there was an overall tendency toward that
which Isidore of Seville exemplifies, the promiscuous use of handbooks and lack of