Perceptual learning is the ability to improve perception through practice. Perceptual learning is usually specific for the task and the stimulus features trained with. For example, training orientation discrimination with vertically oriented stimuli does not improve performance when the same stimuli are rotated by 90°. Often, perceptual learning is assumed to occur within the early sensory areas or when mapping sensory evidence onto decisions. Motor responses, involved in all perceptual learning experiments, are thought to play no role in the learning process. Here, we show that this is not true by providing evidence that perceptual learning is specific for the type of motor response. Interestingly, this specificity can be overcome by the use of a double training protocol. In a second project, we show that, unlike in the somatosensory and auditory system, there is surprisingly no common factor in vision. Finally, we challenge classic models of perceptual learning by showing that conditions under which perceptual learning should not occur as predicted can nevertheless occur. Taken together, our results indicate that the complexity of perceptual learning mechanisms goes much beyond classic models and current theories of perceptual learning.