ago LAW AND FREEDOM IN HISTORY

the same way in the central and eastern provinces until after the death of
Justinian in A.D. 565. Similarly, a Han Empire which had met with its
second stroke in A.D. 184, and which broke up thereafter into 'the Three
Kingdoms', managed to reconstitute itself for a moment in the empire of
the United Tsin (imperabat A.D. 280-317) before going into its final dis-
solution in the fourth century of the Christian Era.

In cases in which the dissolution of a broken-down civilization has
been followed by the emergence of an affiliated civilization there has
usually been an interval of some three hundred years between the date of
the declining civilization's last, and fatal, stroke and the earliest date at
which the rising civilization becomes visible above the historical horizon.
Considering that this intervening chronological interregnum is apt to be
filled to some extent by a 'die-hard' epilogue to the history of the mori-
bund antecedent civilization's universal state,1 we have to reckon that
the process of disintegration may exceed its minimum Time-span of
some eight hundred years, running from the initial breakdown of the
society itself to the second breakdown of its universal state, by trespas-
sing on a subsequent interregnum which gives scope for an epilogue of
any length up to a limit of three hundred years. The maximum Time-
span of the disintegration process thus turns out to be of the order, not
of eight centuries, but of eleven.

Within the minirnuni Time-span of eight hundred years, the six inter-
vals between turning-points, into which the three complete cycles of
Rout-and-Rally can be analysed, fall into two groups, consisting of three
intervals each, which are of approximately equal aggregate length. The
first run of four centuries, constituting the Time of Troubles, is occu-
pied by a down-swing between the original rout and the first rally, an
up-swing between the first rally and the ensuing relapse, and a down-
swing between this relapse and the second rally. This second rally, which
occurs at the half-way point on the eight-hundred-years-long total course,
brings with it the establishment of the universal state, and the second
batch of four centuries, during which a pax oecumenica prevails, is
occupied by an up-swing between the second rally and a further relapse,
a down-swing between this further relapse and a final rally, and an up-
swing between this final rally and a final unretrieved relapse in which
the fabric of a long-since-disintegrating society now dissolves in a social
interregnum. While either group of three intervals between turning-
points is thus apt to occupy a span of four centuries in the aggregate,
there is no indication in the historical evidence of any corresponding
tendency, within each of these four-hundred-years-long spans, for the
three intervals occupying it to be uniform with one another in their
length. On the contrary, it looks as if the chronological articulation of
these intervals within any four-hundred-years-long span were highly
elastic; for in any one example of the series they are apt to differ from
one another in duration, and these differences in their duration which
thus present themselves within the history of each disintegrating civili-
zation also seem to differ, from case to case, in the ratio in which they
stand to one another. There does not seem to be any uniform ratio be-
1 See V. vi. 210-13.