Our thesis is that proprioception is not a sixth sense distinct from the sense of touch, but a part of that tactile (or haptic) sense. The tactile sense is defined as the sense whose direct intentional objects are macroscopic mechanical properties. We first argue (against D. Armstrong, 1962; B. O'Shaughnessy 1989, 1995, 1998 and M. Martin, 1992, 1993,1995) that the two following claims are incompatible : (i) proprioception is a sense distinct from touch; (ii) touch is a bipolar modality, that intrinsically has both a subjective-bodily pole and objective pole. We then argue that the bipolarity of touch should be preferred over the introduction of a sui generis sense of the body. We try to revive Aristotle suggestion according to which the body is the tactile medium (like the air for sight). Since this medium is constantly changing its shape, we need some specific channel of information about its state : proprioception, functionally defined, is that part of touch which informs us about the state of this changing tactile medium. Though muscular and articular receptors are usually dedicated to inform us about the mechanical properties of the tactile media, and the skin receptors about the mechanical properties of the tactile objects, this is not essentially so. In weighting or wielding experiments we access the weight of external objects even when skin sensitivity is absent; in prosthetic touch, the skin receptors play the role usually assigned to muscle and articular receptors, namely to inform us about the mechanical state of the tactile medium. So proprioception, anatomically defined, can play both the role of informing us about the tactile medium, or about the tactile objects. That other sensory modalities also rely on proprioceptive information should be understood in terms of cross-modal dependencies: of sight, hearing, smell, taste...on touch.