Abstract
If visual information for perception and action is processed by distinct pathways, how is it subsequently combined? In disorientation studies, Hermer & Spelke (Nature 1994) found an intriguing failure by disoriented children to search combining wall colours with cues about room layout. In the present studies we hypothesised that this result is not specific to disorientation, but represents a more general difficulty combining “ventral” and “dorsal” visual information in early childhood. 18–24 month olds seated at a table saw toys hidden in boxes distinguished by different colours (e.g. red, blue) and retrieval actions (e.g. push, pull). After a short delay, children had to discriminate between boxes to find the toy, based either on (i) action alone, (ii) colour alone, or on (iii) the conjunction of the two. Colour was poorly remembered on its own, but moreover when colour and action had to be combined, action information was retained but colour information was lost. At 30–36 months, although colours and actions were equally well remembered on their own, colour but not action information was lost in conjunction judgements. Studies in progress indicate that monochromatic textures are likewise lost in conjunction judgements with actions at 30–36 months, but faces are not. Adults remembering conjunctions of subtly different actions and colours do not seem to show the selective disregard of colour, either when tested like children with a free response time, or when pushed to make speeded responses. These results show that there are systematic biases in how “dorsal” and “ventral” visual cues held in children's working memory are matched to visual cues in the test phase; however this pattern differs for different “ventral” cues. One hypothesis is that which information children use or disregard is determined by uneven speeds of processing for different visual cues.
Supported by Economic and Social Research Council grant RES-000-23-1474