p. 492) where it is first defined as “a whole, comprehending in its compass a number of
parts,” from the Latin “complectere,” meaning “to encompass, embrace, comprehend,
comprise.” Among its partial synonyms is “complicated,” although, as Israel (2005)
points out, this comes from a different Latin root, “complicare,” meaning “to fold
together” or “interwoven” (OED, 1971, p. 493). Israel takes the strong position that this
latter is a merely epistemological concept while the former is fundamentally ontological,
complaining that such figures as von Neumann (1966) mistook them as identical,
although this is arguably an overly strong position.
Finally, a virtue of this general definition is that it also encompasses one of the
current cutting edge areas of economics, the behavioral and experimental approaches.
Many who follow these approaches do not consider the complexity view to be all that
relevant to what they do, Kenneth Binmore and Matthew Rabin for example. However, at
the foundation of behavioral economics is the concept of bounded rationality, introduced
originally by Herbert Simon. It is not just Simon, but many since who have seen
complexity as implying that rationality must be bounded (Sargent, 1993; Arthur, Durlauf,
and Lane, 1997), and thus is lying at the foundation of behavioral economics.6
The second definition is of dynamic complexity, which is arguably the most
widely used form in economics, even though this is not per se on Lloyd’s list of 45
definitions. Rosser (1999) has codified this as “broad tent complexity” and draws on Day
(1994) for its definition. “A dynamical system is complex if it endogenously does not
tend to asymptotically to a fixed point, a limit cycle, or an explosion” (Rosser, 1999, p.
170). Rosser then draws on the criticisms of “chaoplexity” by Horgan (1997) to argue
6 For a discussion of the relation of Sargent to Simon, see Esther-Mirjam Sent, 1997.