In contrast to the viewpoint-dependence of within-modal recognition of previously unfamiliar objects, their cross-modal recognition is viewpoint-independent (Lacey et al., 2007, PLoS ONE 2:e890). Since within-modal viewpoint-independence characterizes recognition of highly familiar objects, we tested whether such viewpoint-independence can be acquired by learning, and if so, transferred cross-modally. Participants were exposed to objects in sets of four, after which recognition was assessed for the objects in the set, in unrotated and rotated orientations. Baseline visual and haptic within-modal trials were followed by a series of within-modal learning trials, either visual or haptic. To test cross-modal transfer, visual learning was followed by a final haptic trial and vice-versa. This design contrasts with studies that used a series of visual-haptic/haptic-visual trials (Norman et al., 2008 Perception 37:185-196) which may demonstrate learning to switch between modalities rather than the transfer of learning across modalities. Initial analyses suggest that viewpoint-independence is not easily learned and that within-modal learning does not transfer cross-modally. This implies that visual and haptic viewpoint-dependent representations are separate from each other and from cross-modal representations.