In today's world, large-scale systems are frequently involved in level
s of complexity that have a serious and global impact on productivity.
A context model is set forth that facilitates a new definition of com
plexity, providing a background against which a science of complexity
can be developed, and describing an empirical process from which the c
omplexity of any particular situation can be quantified through a situ
ation complexity index. In moving toward a science of complexity, the
desirability of incorporating semiotics as a component of the science
is indicated, because of the contributions semiotics makes to understa
nding the foundations and ubiquity of modelling. In the development of
models, as semiotics indicates, the connection between what is being
modelled and the language of description in the model is critical. Str
uctural analysis clarifies the inappropriateness of models comprised o
nly of prose to convey a description of a complex situation. Illustrat
ive examples from the practice of interactive management (a system of
management that supports the development and interpretation of structu
ral models of complex situations, and design of improved systems) show
the significance of structural thinking as the primary intellectual m
ode required to manage or cope with complexity. Group activity that wo
uld otherwise be invalidated by spreadthink is converted into a powerf
ul approach to in-depth learning about a complex situation, which then
provides a well-supported foundation for an organized attack to bring
a complex situation under control.