[There is] a case study by Oliver Sacks about a man who was born blind and recovered his sight in middle age. He was unable to recognize his dog and cat (except by touching them) since they looked like different creatures at different angles. He also had to mark out a path through his apartment to walk along; if he strayed off the path he wouldn't be able to tell where he was. I think that this case might bring up interesting questions about what our innate perceptual knowledge is; do we have to learn how to put together all our different views of an object, or did this man start out knowing, but his skill became atrophied?
These cases have formed the basis for much debate about the innate origins of intermodal perception. The issue has its origins in the empiricist philosophers, specifically, Bishop Berkeley (along with Locke and Hume), who argued that because the different modalities registered such different kinds of sensations, it would never be possible for the blind man to perceive the world as the sighted man does. The cases of restored vision (many of which are reported in the late 19th, early 20th century) were first thought to be excellent test cases of nativist vs. empiricist theories (the former arguing for equivalence over modalities, due to innate constraints; the latter arguing for complete non-equivalence, because the origins of knowledge are the sensations). Unfortunately, it was soon recognized that the cases in which vision was restored later in life were contaminated, because lack of visual input (light) over the years actually led to degeneration of the visual system at a relatively low level (i.e. the retina, ganglion cells). So: the fact that the person with restored vision doesn't "recognize" his cat visually does not necessarily mean that shape constancy requires experience. Rather, it could just mean that the visual system was damaged.
If you are interested in evidence for the innate bases of intermodal perception and depth perception (the two key issues for empiricist vs. nativist debates), there is now good evidence from studies in infancy that BOTH exist in newborns!!