The contingency symmetry bias (affirming the consequent fallacy) as a prerequisite for word learning: A comparative study of pre-linguistic human infants and chimpanzees

Abstract

Humans are known to possess an “affirming the consequent fallacy,” which assumes that a learned contingency holds true even when the order is reversed. In contrast, non-human animals do not fall for this fallacy, as they do not have the contingency symmetry bias. Importantly, language is founded on the symmetrical relationship between symbols and referents, and the contingency symmetry bias plays a key role in word learning. A critical problem for the ontogenesis of language is whether the contingency symmetry bias has been acquired through the experience of word learning or if it is present before infants begin word learning. Using a habituation switch paradigm, 8-month-old human infants and adult chimpanzees were familiarized with two object-then-movement sequences, whereby Object A (or B) was always paired with Movement A (or B). At test, the order of the contingency was reversed. The infants showed surprise when observing the violation of the object-movement pairings in the reversed sequence (Experiment 1). In contrast, despite the chimpanzees being able to detect the violation of the pairings in the original direction (Experiment 2a), they did not discriminate the learned and novel pairings when the order of the contingency was reversed (Experiment 2b). The results suggest that the contingency symmetry bias is a uniquely human cognitive bias, one which plays a critical role for language acquisition ontogenetically. This contingency symmetry bias likely gives humans a great advantage, by enabling them to rapidly expand their knowledge without direct training and making them strikingly different from other animal species. (250 words).