Rebecca Lawson, University of Liverpool

Two studies examined how training context influences the naming of plane-misoriented pictures of familiar objects. The disadvantage for naming misoriented views of an object can reduce if the object is seen repeatedly in training. In the first study, training reduced the misorientation disadvantage for pictures of objects only if training presented misoriented pictures of the objects and not if training presented misoriented words which named the objects or if training presented upright pictures or words. In the second study, if shape discrimination in training was easy, performance improved in training but not subsequently when the difficulty manipulation was removed. The misorientation disadvantage was thus sensitive to the nature of the training context (whether upright or misoriented pictures or words were presented) but not to the difficulty of that context (whether useful orientation-invariant information was easy or hard to extract). People only appear to extract orientation-invariant information under tightly restricted conditions. However once this strategy is triggered, it seems equally effective regardless of the difficulty of discrimination.

Visual Cognition, (2003), 10, 795-821