Abstract
Thirty six Year 2 children were each presented with two out of a set of six realistic
models of various physical angle contexts and asked (1) to indicate which of a set
of 10 abstract angle models could also represent those contexts,' (2) to represent
the contexts in drawings; and (3) to indicate· whether they recognised any
similarities between the two contexts. Previous papers have reported results for the
turns, slopes, rebounds and corners contexts. The present paper reports on the
crossings and bends contexts and summarises the specific features of each of the 6
contexts which appear to hinder the recognition of similarities. The basic
hypothesis of abstraction theory - that an abstract angle can only be recognised in
a particular context when angle-related similarities between that context and other,
superficially different contexts are also recognised - is then examined.