How specific is human cultural learning?

How specific is human cultural learning? To answer this question we need to know if nonhuman animals learn culturally and if so, how. This requires us looking over the fence. Since I consider cultural learning being a subcategory of social learning, I will review several studies on social learning in non-human animals. Interestingly, these studies share an irritating feature, they don’t fit (easily) into the current (dominant) theories of social learning research. Red-footed tortoises show clear evidence of learning a difficult spatial problem (detour task) by observing a conspecific model despite the fact that they are solitary. Archer fish, lizards and common marmosets exhibit high fidelity copying of movement patterns of a conspecific model, thereby solving the correspondence problem of imitation (transforming visual information into matching motor acts). Dogs seem able to perform — at least in ostensive-communicative contexts — selective imitation; they copy a conspecific’ actions
more often when those actions are the efficient solution to a problem than when they are not. However, the ability to interpret the ‘rational’ dimension of others actions has been considered a hallmark of human cultural learning. Together these data challenge some theories of social learning by suggesting that several core components of human cultural learning, such as high copying fidelity, intentional inhibition and selectivity, are shared by many species. What remains human-specific in social learning needs to be specified.