Models of cumulative cultural evolution, by necessity, consider the information present in the
population before a learning event and often incorporate other factors, such as the origin , or
the fitness benefit of that information. These models often assume that certain information is
a prerequisite for other knowledge or that a skill may be sequentially refined through individual or
social learning , again usually at a constant rate. One interesting exception to this
rule is a model of cumulative cultural evolution described by Enquist et al. who allow a
nonconstant rate of invention for which they cite substantial empirical evidence. The analysis
in suggests that an increase in creativity with the amount of culture present led to the
exponential increases in human cultural output seen across domains; such increases could not
occur in their model with a constant rate of creativity. This model contrasts with others that
implicitly or explicitly assume that the amount of creativity might be inversely proportional to the
number of cultural traits already present. Although Enquist et al. regard creativity as
a function of the current number of cultural traits in the population , their model is consistent
with other descriptions of creativity as partly (but perhaps not primarily) directed towards
solving specific problems but mostly the product of copy error (for experimental investigations of
evolution driven by copy error, see ). In, novelty can be generated through a kind of
productive error in social transmission of a previously existing trait. The model captures some
important aspects of the relation between culture and creativity, in particular that the amount of
new culture created could depend in some way, either positively or negatively, on the amount of
culture already present (‘cognitive creativity’ in Dietrich's terminology).