Yet from their agent-based model, Mesoudi and O'Brien argued that the same attribute patterns were consistent with other transmission models besides indirect bias, i.e., an equifinality thesis.
Returning to Bettinger and Eerkens (1999), it is good to know that guided variation characterized the western Great Basin mid-Holocene point sequence but indirect bias characterized the eastern sequence.
The complex inference was that people in the central region adopted bow-and-arrow technology as an integrated whole, a package deal (therefore, indirect bias), whereas people to the west assimilated the technology piecemeal (therefore, guided variation or trial-and-error adaptation).
The disadvantage of indirect bias over direct bias is that indirect bias allows suboptimal traits to accumulate when exhibited by the most-knowledgeable individual.
Indirect bias involves copying successful individuals, whereas direct bias involves copying effective traits.
In general, while unbiased transmission (i.e. random copying) might explain changes in the distribution of certain cultural traits [30], cumulative cultural evolution requires direct or indirect bias.
For indirect bias, new individuals copy all of the traits exhibited by the single individual in the previous generation who has the highest individual fitness, Zi.
On the other hand, the awarding of honours (e.g. Nobel prizes) to single individuals may encourage a prestige-based form of indirect bias.
With respect to real-life scientific and technological change, we might ask whether past or present educational and apprenticeship systems better resemble either direct or indirect bias.
It is also possible that the scientific method contains formal mechanisms (e.g. anonymous peer review) for avoiding the accumulation of suboptimal traits via indirect bias.
