ABSTRACT
Children's major difficulty with shapes is in distinguishing them on the basis of touch. Experiments have confirmed that children find it no harder to compare shapes across vision and touch than to compare them tactually. As soon as children can recognize shapes by touch they are also able to translate the tactual information into its visual equivalent and vice versa. A cross-modal dictionary, whatever form it takes, must be able to connect particular entries from one modality with those from another, and whether the child uses absolute or relative codes is a factor which must affect its nature. The discovery that some animals have a cross-modal dictionary at their disposal makes it seem more likely that humans might also be able to recognize objects cross-modally before they can describe the objects in words. Some animals can recognize some objects across different modalities without the help of language. The extreme form of the language hypothesis is no longer tenable.
