How do cumulative cultural evolution and individual learning differ? In an abstract computational sense, both are optimisation processes that search a space of possible explanations and previous work has identified deep parallels in the mathematical models used to describe them (Suchow, Bourgin, & Griffiths, 2017). However, there are obvious differences as well: for example, individual learning involves a single agent characterised by one set of prior beliefs, representational capabilities, and so forth, while cultural evolution involves multiple agents who may vary along these factors. We argue that this difference implies that the process of cumulative cultural evolution should involve searching a more restricted set of hypotheses and converge on simpler ones. In two iterated category learning experiments, we test this prediction and find that transmission chains composed of single individuals, who learn based on their previous performance, consider both a wider variety and more complex categorisation schemas than do chains involving multiple people.