This is a potentially fascinating topic: the sensitivity to initial conditions often associated with chaos also applies to systems with multiple basins of attraction.

This view seems more plausible than asserting the gross events of history are directed by the flapping of every butterfly's wings.

At any given moment, there are several major themes that history may explore. Which of these is explored can be determined by the smallest events, but the small events do not shape the major themes.

In meteorological terms, the flapping of a butterfly's wings in Brazil does not cause a tornado in Texas. Rather, it may select the Texas tornado from among many possibilities.

Multiple attractors, with fractal basin boundaries: that would be an interesting study of history.

In addition, the hierarchical structure of many social systems suggests a possible fractality of political systems. Federal is to states, as state is to counties, as county is to cities, as city is to neighborhoods, as neighborhood is to households, as household is to families, as family is to individual. Are these just loose metaphors, or do they represent functional relationships across several scales?

Here are three examples.

John Gaddis discusses scale-invariant aspects of history.
Robert Adams suggests chaos as an explanation for the political turmoil of ancient Sumer.
Building on work of Richardson, D. Roberts and D. Turcotte study power-law relations in wars.