ii The objectivity of large-scale historical inquiry 261
and appear willful and capricious. Nevertheless, the principles of objec-
tivity applicable to the details still apply to the syntheses, only on a more
subtle basis. In particular, any one elaboration can yield answers only to
those questions which it poses - not only to those general questions it
starts out with, but to those particular questions in terms of which it is
built up; that is, its range of relevance is limited by its particular struc-
ture.
Since any one inquiry must have a determinate structure, as we
shall see, any one inquiry is limited in the account it can give of the
general field it treats with. Accordingly, a large number of complemen-
tary studies is necessary to a balanced view of a given age or time, or
even a given personality; yet each of them can be criticized with regard
to its objectivity within the framework it has adopted - so that though
there may be an indefinite number of different but complementary objec-
tive studies in a particular field, there may be an infinitely larger number
of unsatisfactory studies in the same field that must be excluded.
Besides selection and ordering, categorizing also limits one's possi-
bilities: periodization, grouping of people, or lines of activity - the
choice of landmarks, the subordination of events, the coordination of
conditions for purposes of comparison - the very grouping of materials
into concepts, and the weighting of these concepts in ordering one's
interpretation - all can vary, yet remain valid for one or another pur-
pose.
A study which averages out the reactions of peasants and arti-
sans as "the lower classes" will read differently - answering different
questions - from one which makes the distinction at the expense of
clarity with regard to their common responses; a study arranging politi-
cal events by reigns will legitimately read differently from one ordering
them by cycles of prosperity.
It is thus unlikely that any one elaboration can even at best represent
all viewpoints, assuming that would be desirable. Accordingly, historical
writings will vary not only through the error and exploratory unsound-
ness at the frontiers of research, but in the most established areas, where
many works must complement each other. As the perfect roundness of a
circle can only be approximated to, in our measuring, by the infinite
sides of a polygon, so in history - and perhaps in other fields as well, as
they become more complex - a balanced view of a field will only be
multiply achieved. Bury's notes, superseding details of Gibbon's Rome,
corrects errors in view of better research, but it is otherwise with the
difference between a modern historian who sees in Byzantium the effort
to create a Medieval Christian society of a new sort, and on the other
hand Gibbon's triumph of barbarism and religion - an orientation which
continues to have its own validity.