Causal Knowledge in Corvids, Primates, and Children: More Than Meets the Eye?

Amanda Seed*, Daniel Hanus, Josep Call

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapter

7 Citations (Scopus)


Previous experimental work has led several authors to conclude that only humans reinterpret 'first-order perceptual relations in terms of higher order role-governed relational structures,' meaning that among other things, non-human animals are not capable of physical reasoning based on abstract, unobservable object properties. Instead they must rely on first-order perceptual information to solve problems. Such an account could approximate behaviour driven by physical knowledge very well if natural selection has pre-prepared the learning animal to attend to perceptual features of the stimulus that are most likely to correlate with its functional properties in the natural environment. This chapter aims to challenge the 'relational reinterpretation hypothesis' put forward by Penn et al. (2008) by describing data from several other paradigms aimed at distinguishing between explanations based on surface-level perceptual characteristics and those in which object properties (such as solidity, continuity, weight, and rigidity) are represented at a deeper level of abstraction, where 'abstract' means that the information is not equivalent or reducible to concrete, analogue sensory input, but rather has undergone further processing in which meaning is extracted.

Original languageEnglish
Title of host publicationTool Use and Causal Cognition
PublisherOxford University Press
ISBN (Electronic)9780191731259
ISBN (Print)9780199571154
Publication statusPublished - 19 Jan 2012


  • Abstract
  • First-order perceptual relations
  • Object properties
  • Perceptual information
  • Physical reasoning
  • Relational reinterpretation hypothesis


Dive into the research topics of 'Causal Knowledge in Corvids, Primates, and Children: More Than Meets the Eye?'. Together they form a unique fingerprint.

Cite this