Preparedness in cultural learning

Abstract

It is clear throughout Cognitive Gadgets (2018) Heyes believes the development of cognitive capacities results from the interaction of genes and experience. However, she opposes cognitive instincts theorists to her own view that uniquely human capacities are cognitive gadgets. Instinct theorists believe that cognitive capacities are substantially produced by selection, with the environment playing a triggering role. Heyes's position is that humans have similar general learning capacities to those present across taxa, and that sophisticated human cognition is substantially created by our socioculturally transmitted environment. It is a core strategy of Heyes (Cognitive gadgets: the cultural evolution of thinking. Harvard University Press, Cambridge, 2018) to provide evidence of learning altering a cognitive capacity to conclude that a capacity is a cognitive gadget and not an instinct. We draw on recent work on the evolution of learning preparedness to examine the adequacy of this strategy. In particular, we analyse experimental evolution work showing how selection affects cognition within the laboratory. First, this work reveals that change due to learning can still be retained under genetic assimilation. This suggests that domain-specific adaptation can coexist with learning, moderate nativism, an option missed by the instinct versus gadget distinction. Second, we describe the conditions that select for increased preparedness in learning: certainty, reliability, and particular costs. We consider how these conditions can be used when conducting evolutionary reasoning about cognition, applying them to the important capacity for imitation. We find that the conditions lend theoretical support to moderate nativism about the capacity to imitate, which is supported by psychological evidence.

Sinking in: the peripheral Baldwinisation of human cognition

“…By contrast, there is empirical support for the idea that motivational [28] and attentional processes [25,27] have been Baldwinised in ways that could facilitate imitation. Humans have higher social motivation than chimpanzeesfor example, we are more likely to choose social over asocial activities [109] and, although many species are more attentive to biological than nonbiological motion [110], from infancy humans are especially attentive to faces [111][112][113] and hands [114], particularly hands in motion [115,116].…”

“…The Baldwin hypothesis has been invoked to explain a wide range of behavioural traits across the animal kingdom [12][13][14][15], including prey capture in archer fish [16] and the song of zebra finches [17,18]. In the case of humans, Baldwinisation is thought to have played a key role in the emergence of language [19][20][21][22][23][24] and imitation [25][26][27][28].…”

“…The associative sequence learning (ASL) model of imitation [100][101][102] has provided more fertile ground for Baldwin hypotheses [25][26][27][28]. The ASL model suggests that imitation depends on two sets of horizontal processesthat encode sequences of observed action components and mediate motor learning, respectivelyand on a large repertoire of matching vertical associations, bidirectional excitatory links, each connecting a visual representation of an action component with a motor representation of a similar component (e.g., the sight of winking with a motor programme for winking).…”

“…The ASL model suggests that imitation depends on two sets of horizontal processesthat encode sequences of observed action components and mediate motor learning, respectivelyand on a large repertoire of matching vertical associations, bidirectional excitatory links, each connecting a visual representation of an action component with a motor representation of a similar component (e.g., the sight of winking with a motor programme for winking). Two hypotheses suggest that Baldwinisation might have occurred for horizontal [27] or vertical [28] processes, and two more propose Baldwinisation of attention to body movements [25] or motivation to imitate [28].…”

“…The second Baldwin hypothesis is also incompatible with current data. In principle, Baldwinian evolution could have made it easier to learn matching than non-matching vertical associations [28], for example it may be easier for agents to associate the sight of a hand opening with performance of hand-opening than with performance of hand-closing. However, training experiments have shown that non-matching vertical associations can be learned with remarkable speed [106][107][108].…”